aus dem fachbereich geowissenschaften der universität...

aus dem Fachbereich Geowissenschaftender Universität Bremen

Nr.85

Ratmeyer, Volker

UNTERSUCHUNGEN ZUM EINTRAG UND TRANSPORT

LITHOGENER UND ORGANISCHER

PARTIKULÄRER SUBSTANZ

IM ÖSTLICHEN SUBTROPISCHEN NORDATLANTIK

Berichte, Fachbereich Geowissenschaften, Universität Bremen, Nr. 85,161 S., Bremen 1996

ISSN 0931-0800

Die "Berichte aus dem Fachbereich Geowissenschaften" werden in unregelmäßigen Abständen vom

J'achbereich 5, Universität Bremen, herausgegeben.

Sie dienen der Veröffentlichung von Forschungsarbeiten, Doktorarbeiten und wissenschaftlichen Beiträgen. die

im Fachbereich angefertigt \Vurden.

Die Berichte können bei:

Frau Giscla Boelen

Sonderforschungsbereich 261

Universität Bremen

Postfach :130 440

D 28.334 BREMEN

Tdelon: (49) 421 218-4124

Fax: (49) 421 218-3116

angefordert werden

Zitat:

Ratmeyer, V

Untersuchungen zum Eintrag und Transport lithogener und organischer partikulärer Substanz im östlichen

subtropischen Norclatlantik.

Berichte, Fachbereich Geowissenschaften, Universität Bremen, NI'. 85, 161 Seiten, Bremen, 1996.

ISSN 0931-0800

Untersuchungen zUlnEintrag und Transportlithogener und organischer partikulärer Substanz

im östlichen subtropischen Nordatlantik

Dissertationzur Erlangung des Doktorgrades

am Fachbereich Geowissenschaftender Universität Bremen

vorgelegt vonVolker Ratmeyer

Bremen 1996

Danksagung

Mein besonderer Dank gilt Herrn Prof. Dr. Gerold Wefer für die Vergabe und Betreuung der

vorliegenden Dissertation. Sein Interesse am Fortgang dieser Arbeit und seine vielfaltigen

Anregungen waren eine wesentliche Unterstützung.

Herzlich danken möchte ich Gerhard Fischer für zahlreiche tiefgreifende und fruchtbare

Diskussionen, die wesentlich zur Freude an dieser Arbeit und deren Gelingen beigetragen haben. Für

ständige Hilfsbereitschaft und tatkräftige Unterstützung vor allem während der Entwicklung des

Kamerasystems bedanke ich mich bei Götz Ruhland.

Prof. Dr. Wolfgang Balzer danke ich für kritische wissenschaftliche Diskussionen sowIe für die

Übernahme des Co-Referates für diese Dissetiation.

Für das außergewöhnlich angenehme und unkomplizierte Arbeitsklima möchte ich allen Kollegen und

Mitarbeitern der Arbeitsgruppe Meeresgeologie am Fachbereich Geowissenschaften der Universität

Bremen danken. Matiin Cepek, Nils Andersen und Gerrit Meinecke danke ich für zahlreiche

Inspirationen durch wissenschaftliche wie auch nichtfachliche Gespräche und Diskussionen.

Weitreichende Unterstützung erhielt ich von Gerhard Kulm und Bernhard Diekmann durch

Einweisung am Laser-Granulometer und zahlreiche Diskussionen über die Korngrößenanalyse. Ulrike

Wyputta danke ich herzlich für die Berechnung der atmosphärischen Trajektorien, sowie Susanne

Neuer fLir wissenschaftliche Diskussionen und die konstruktive Arbeit am Manuskript. Für wetivolle

Hilfe auf See danke ich besonders Uwe Rosiak, Ronald Heygen, Helmuth Meyer und Peter Gerchow.

Ein herzliches Dankeschön gilt meinen Eltern für ihr Verständnis und ständiges Interesse während

einer langen Zeit wissenschaftlicher Ausbildung.

Mein abschließender Dank gebührt meiner Frau Charlatte, die mich während der Promotion vor allem

durch ihre liebevolle Aufmunterung immer wesentlich unterstützte.

Inhaltsverzeichnis

Seite

1. Einleitung 1

1.1. Einführung 1

1.2. Zielsetzung der Arbeit 6

1.3 Durchgeftihlie Arbeiten 8

1.4 Arbeitsgebiet. 10

1.4.1 Ozeanzirkulation 10

1.4.2 Atmosphärischen Zirkulation und Staubtransport... ll

l.5 Methodik 12

1.5.1 PaliikelkaJllera 13

1.5.2 Bearbeitung der Sinkstoffproben 21

1.5.3 Bearbeitung der atmosphärischen Daten 23

2. Veröffentlichungen 25

2.1 Manuskript 1: 25

A high resolution camera system (paJ'Ca) for imaging partic1es in the ocean: System design

and results from profiles and a three-month deployment


Lithogenic particle fluxes and grain size distributions in the deep ocean offNW Africa:

implications for seasonal changes of aeolian dust transport and mass fluxes


Impact of mineral dust on deep-ocean particle flux in the eastern subtropical

Atlantic Ocean


Deep water particle flux in the Canary Island region: Seasonal variation

in relation to long-term satellite derived pigment data and lateral sources

3. Ergebnisse der einzelnen Arbeiten 123

4. Zusan1111enfassung 128

5. Schlußfolgerungen und Ausblick 132

6. Literatur 135

7. Anhang 141

Einleitung

1. Einleitung

1.1 Einfühnmg

Der Stoffllaushalt der Ozeane wird reguliert durch eine Vielzahl unterschiedlicher Austausch- und

Transportvorgänge, von denen einige als maßgeblich f1ir das globale Klimageschehen angesehen

werden. Dem Kohlenstoffkreislauf wird dabei eine besondere Bedeutung zugeschrieben, obwoh I

bislang keine Einigkeit besteht über die Art der Wechselwirkungen zwischen Kohlendioxidgehalt der

Atmosphäre und dem Klima. Messungen an Eiskernen zeigen maßgebliche Schwankungen des

atmosphärischen C02-Gehaltes während der letzten 160.000 .Jahre (Barnola et al. 1987), die

wesentlich auf den Gasaustausch zwischen Atmosphäre und Ozean zurlickgef1ihrt 'werden. Vor dem

Hintergrund des Kohlenstoffkreislaufs als möglicher Klimaregulator nimmt das Inventar an

partikulärem Material im Ozean eine zentrale Stellung ein. Partikulür vorliegendes Material entsteht

im Meer primär durch Phytoplanktonwachstum und durch den Eintrag aus den angrenzenden

Kompartimenten der Biosphäre, sekundär als Produkt der marinen Nahrungskette und durch Erosion

bereits sedimentierter Partikel. Neben den im Meerwasser gelösten Nährstoffen spielen Partikel eine

wichtige Rolle als Träger verschiedener chemischer Elemente, die in dieser Form eine hohe zeitliche

und räumliche Variabilität aufweisen. Weder die Bildung noch der Eintrag von partikulärem Material

verlaufen kontinuierlich. Aufgrund jahreszeitlicher klimatischer Schwankungen kommt es zu

episodischen Ereignissen in verschiedenen Regionen des Ozeans, wie z.B. Planktonblüten 111

Hochproduktionsgebieten oder die saisonale Bildung von Meereis in den Polargebieten. Die dabei

entstehenden Wechselwirkungen zwischen Ozean und Atmosphäre wirken langfristig wiederum

regulierend auf das globale Klimageschehen. Unter diesen Wechselwirkungen gehört der Austausch

von C02 zwischen Atmosphäre und Ozean zu den Wichtigsten. Die biologische Bildung partikulären

Materials in der Deckschicht des Ozeans bestimmt - neben physikalischen Parametern wie

Wassertemperatur und Zirkulation - maßgeblich den Kohlendioxidaustausch zwischen Ozean und

Atmosphäre. Durch Photosynthese und Entstehung organischer Substanz in Form mariner

Organismen wird dem Oberflächenwasser C02 entzogen. Da Atmosphäre und Oberflächenwasser

jedoch bestrebt sind, bezüglich ihres Gasgehaltes im Gleichgewicht zu stehen, wirkt die Senkung des

C02-Partialdruckes (PC02) im Oberflächenwasser ebenfalls verringernd auf den C02-Gehalt der

Atmosphäre. Der Einbau von Kohlenstoff in Kalkschalen mariner Organismen hingegen erhöht den

C02-Gehalt im Oberflächenwasser und damit auch den der Atmosphäre. Da beide Prozesse

gegenläufig sind, ist nicht das Maß der biologischen Produktivität als PC02-Regulator

ausschlaggebend, sondern vor allem die Menge und Zusammensetzung des organischen Materials, das

dem Oberflächenwasser durch Absinken entzogen wird. Von besonderem Interesse ist dabei das

2

Verhältnis von organischem zu karbonatischem Kohlenstoff (rain ratio) in dem sinkenden lind

schließlich sedimentierenden Material (Zusammenfassung in Berger und Keil', 1984).

Bisher gibt es erst wenige Abschätzungen darüber, welcher Anteil der durch diese "biologische

Pum pe" (Broecker 1982) ursprüngl ich exportierten organ ischen Substanz tatsäch Iich die tie fcrcn

Wasserschichten und den Meeresboden erreicht. Nur etwa ein zehntel der primäl- 1111

Oberflächenwasser produzierten Substanz wird aus der photischen Zone in tiefere Wasserschichtcn

cxportiert. Weiterer Zerfall, Abbau und Remineralisation während des Absinkens bis ZlIlll

Meeresboden reduzieli den Gehalt an partikulärem organischen Kohlenstoff um weitere 90 (%. In

Küstenozeanen akkumulieren schließlich nur etwa I%, im offenen Ozean hingegen nur ca. 0.03 (:';) der

ursprünglichen primär produzierten Biomasse im Sediment (Berger et al. 1989). Für die Interpretation

von den im Sediment überlieferten Signalen zur Rekonstruktion paläoklimatischer Veränderungen,

aber auch für das tiefergehende Verständnis des heutigen Klimageschehens, ist deshalb die Kenntnis

von Effizienz und Mechanismen der "biologischen Pumpe" sehr wichtig.

Partikelbildung und Partikelfluß

Zwei grundlegende Quellen tragen zur Bildung des paliikulären Kohlenstoffinventars im Ozean bei:

Dies ist sowohl das Wachstum mariner photosynthetisierender Organismen, zum anderen der Eintrag

organischen und anorganischen Materials aus kontinentalen Herkunftsgebieten. Der externe Eintrag

wird bestimmt durch Flußeintrag, Küstenerosion, Eintrag von eistransportiertem Material und

atmosphärischem Transpoli. Hinzu kommt bereits sedimentielie Substanz, die durch Erosion und

Resuspension wieder in den TranspOlikreislauf zurückgefühli wird. Der Einfluß beider Primärquellen

ist regional sehr unterschiedlich, global überwiegt die marin produzierte Biomasse den Eintrag vom

Land um ein Vielfaches. Es wird jedoch diskutieli, daß die Primärproduktion in oligotrophen

Meeresgebieten in besonderem Maße von der Zufuhr limitiereneder Mikronährstoffe wie z.B. Eisen

abhängt (Martin und Fitzwater 1988), die durch windtransportierten Staub kontinentaler Herkunft

angeliefert werden können.

Der größte Teil des vertikalen Transportes biogener und anorganischer Partikel durch die Wassersäule

verläuft über biologische Prozesse (Fowler und Knauer 1986; Wefer 1993). Diese Annahme basiert

weitgehend auf der Tatsache, daß die meisten biogenen wie terrigenen Partikel zu klein sind, um ohne

weitere Vergrößerung oder Beschwerung in größere Ozeantiefen und schließlich zum Meeresboden zu

gelangen. Einen Überblick über die Sinkgeschwindigkeiten einzelner mariner Paliikel in Abhängigkeit

ihrer Größe geben Degens und Ittekott (1984). Einzelne Partikel mit einem Durchmesser von unter 30

~Lm haben nach dem Stoke'schen Gesetz Fallzeiten von weniger als 10m Tag- 1, hierunter fcillt ein

Großteil mariner Organismen wie Coccolithophoriden, Diatomeen, Silikoflagellaten sowie nahezu der

gesamte, über den Wind eingetragene Staub (bei Transpoliweiten über 500 km; Schütz 1980,

Einleitung 3

Westphal et al. 1987). Die kurze zeitliche Verzögerung zwischen saisonalen Partikelsignalen, die 111 il

Sinkstoff-Fallen 111 verschiedenen Wassertiefen gemessen wurden, deuten jedoch au f

Sinkgeschwindigkeiten von 100-150 md-I (Lalllpitt 1985; Asper et al. 1992), 280 m d- I (Fischer et

a1. 1996a), bis hin zu 200-2000 m d- I (Wcfcr 1993). Grund für diese hohen Werte sind Prozesse, die

zur Koagulation von kleinen Partikeln zu größeren Aggregaten ("marine snow", definiert als amorphe

Aggregate mit einem Equivalentdurchmesser > 0.5 mm; Fowler und Knauer 1986) und damit auch zu

höherem spezifischen Gewicht führen. Eine Zusammenfassung über die unterschiedlichen

Enstehungsprozesse geben Bruland (1989), Wefer (1993) sowie Alldredge und Jackson (1995). Neue

experimentelle Studien unter kontrollierten Bedingungen zeigen zudem die Bedeutung von

polysacehariden Diatomeenausscheidungen (RiebeseIl 1991) für die Aggregation von Phytoplankton.

Als eigenständige transparente exopolymere Partikel (TEP) sorgen sie gleichzeitig für die

Anreicherung und Koagulation suspendierter Partikel außerhalb der Diatomeenflocken (Passow und

Alldredge 1995). Neben der Verklumpung durch Diatomeen (Alldredge und Gotschalk 1989) und

Coccolithophoriden (Honjo 1982) wird das aktive "packaging" durch Abweidung des Phytoplanktons

und Ausscheidung in Form von Kotpillen durch das Zooplankton als hauptsächlicher biologischer

Aggregationsprozeß angesehen (Gersonde und Wefer 1987 Wefer 1993). In größeren Wassertiefen

vor allem in höheren Breiten kommt es durch tägliche Wanderung des Zooplanktons auch zu

mehrfacher Aufbereitung und erneuter Ausscheidung von Faeces (v. Bodungen et al. 1987). Dies kann

zur Vermischung und simultanen Sedimentation von suspendiertem, vergleichsweise alten Material

mit frischer organischer Substanz aus der photischen Zone führen. Inwieweit mikrobiologische

Aktivität durch Bakterienbesiedlung auf Aggregaten eine weitere Koagulation begünstigt oder

verhindert, ist bisher nicht gesichert. Neuere experimentelle Untersuchungen zeigen jedoch deutlich

die Signifikanz von Bakterienbesiedlung für den Kohlenstoffumsatz in Aggregaten. Smith et al.

(1995) fanden 40-60% des Kohlenstoffs fixiert durch Bakterien während der Dauer ell1er

experimentellen Diatomeenblüte.

Neben den biologischen Aggregationsvorgängen spielen physikalische Koagulationsprozesse elI1e

wichtige Rolle (Alldregde und McGillivary 1991). Obwohl Koagulation z.B. von Tonpartikeln « 2

11m) schon durch van der Waals-Kräfte (Russell 1980) oder Ladungsungleichgewichte der

Tonminerale ( McCave 1985) ermöglicht wird, ist die Ausfällung kleinster Partikel durch den Einbau

in biogene Aggregate der entscheidende Prozess (Stolzenbach 1993). Wesentlich ist demnach die

Interaktion von biogenen mit terrigenen und suspendierten Partikeln. MacIntyre et al. (1995) fanden

z.B. Ansammlung von stark porösen Aggregaten an Dichtegradienten in den oberen 100 m der

Wassersäule, von denen ein hoher Prozentsatz durch Turbulenz vor Ort entstand, ein weiterer Anteil

lateral eingeschichtet wurde. Asper et al. (1992) zeigen einen maßgeblichen Unterschied der

Massenbilanzen zwischen veliikalen Verteilungsprofilen mariner Aggregate und Patiikelfluß

Messungen mit Sinkstoff-Fallen in Wasseliiefen bis 3500 m im Panama-Becken. Lateral

4

eingeschichtete und suspendierte partikuläre Substanz prägt demnach zwar entscheidend die

Partikelverteilung in der Wassersäule, für den vertikalen Stofftransport ist jedoch das pri miire,

episodische Signal schnell sinkender, und daher seltener anzutreffender Aggregate aus dem

Oberflächenwasser verantwortlich. Letzteres führt neben einem effektiven Ausfiltern suspencl icrter

Terrigenpartikel zusätzlich zu einer Besehwerung der biogenen Aggregate, und damit zu erhöhter

Sinkgeschwindigkeit ("particle loading", Ramaswamy et al. 1991), verringerter Verweilclauer lind

dam it geringerer Rem ineral istion in der Wassersäu le, letztl ich also zu einer höheren Effiziell z der

"biologischen Pumpe" (Haake und Ittekott 1990). Dieses "Nebeneinander" verschiedener PanikeI

Populationen (Asper et al. 1992) und deren Wechselwirkung hat auch Konsequenzen für die vertikale

Verteilung gelöster Stoffe in der Wassersäule.

Obwohl 1I1 den letzten Jahren zahlreiche neue Erkenntnisse zur Bildung und zlIm

Koagulationsverhalten mariner Aggregate gewonnen wurden (Zusammenfassung in Alldredge und

Jackson 1995), sind wesentliche Fragen zum Transport und zur Quantifizierung des Massenilusses

partikulärer Substanz durch die Wassersäule noch nicht beantwortet. Langzeitmessungen des

Partikelflusses liegen u.a. aus der Sargasso-See (z.B. Deuser 1986; Jickels et al. 1987), dem Nord

(z.B. Honjo und Manganini 1993) und Südatlantik (z.B. Wefer und Fischer 1993; Fischer et al. 1996),

Pazifik (z.B. Asper et al. 1992), den Polargebieten (z.B. Fischer et al. 1988; Berger und Wefer 1990)

sowie aus dem Indischen Ozean (z.B. Ittekott et al. 1991) und der Arabischen See (z.B. Ramaswamy

et al. 1991) vor. Diese Messungen basieren fast ausschließlich auf der Probenahme mit Sinkstoff

Fallen, die über mehrere Jahre in verschiedenen Wassertiefen verankert wurden. Nahezu alle Studien

kommen zu dem Ergebnis, daß die zeitliche Variabilität des Partikelflusses bis in die Tiefsee eng an

die saisonalen Muster der biologischen Aktivität im Oberflächenwasser gekoppelt ist. Das Maß der

Saisonalität im Vergleich zwischen Primärproduktion und bodennahem Partikelfluß bestimmt dabei

nach Berger und Wefer (1990) wesentlich die Höhe der Exportproduktion und ist demnach ein

wichtiger Parameter bei der Rekonstruktion der Effizienz der "biologischen Pumpe".

In dieser Dissertation wird der Vergleich von Primärproduktion und Paliikelflüssen in der Tiefsee um

die Kanarischen Inseln diskutiert. Obwohl die Amplitude der gemessenen Flüsse entsprechend der

stark unterschiedlichen Saisonalität in verschiedenen Meeresregionen stark schwankt, besteht eine

enge Kopplung der Flüsse mit der Primärproduktion auch in sehr niedrigproduktiven Gebieten (z.B. in

der Sargasso-See; Deuser et al. 1981). Neue Untersuchungen mit alternativen Methoden werfen

jedoch die Frage auf, ob die mit Fallen gemessenen Werte tatsächlich eine gute Abschätzung der

Massenbilanz - und damit der Effizienz der "biologischen Pumpe" - ermöglichen. Zwar ist man sich

weitgehend einig, daß der mit Sinkstoff-Fallen gemessene Patiikelfluß zum größten Teil über den

Transpari durch große Aggregate und Kotpillen geschieht. Fischer et al. (1996) fanden z.B. extrem

kurzzeitige Absinkereignisse mit einer Dauer von wenigen Stunden bis zu 6 Tagen mit jeweils hohen

5Eil/leitung-----------------------------------,~----

Anteilen an Kotpillen verschiedener Größe im subtropischen Nordatlantik ancl der Station CI

(Kanarische Inseln). Die direkte Messung von Aggregatverteilungen und Sinkgeschwindigkeiten in

sill! wurde bisher jedoch nur in vergleichsweise wenigen Studien durchgeflihrt (z.B. Kajihara 197 I;

Shanks und Trent 1980; Silver und Alldredge 1988; Honjo et aI. 1984; Maclntyre et aI. 1995), Die

Schwierigkeit solcher Beobachtungen ist vor allem inder äußerst fragi len Natur mariner Aggregate

begründet, die - außer den stabilen und kompakten Kotpillen - in der Regel in den Sinkstoff-Fallen

nicht erhalten bleiben. Über ihre tatsächliche Verteilung in der Wassersäule, ihre Größen lind

Sinkgeschwindigkeiten ist daher nur sehr wenig bekannt (z.B. Alldredge und Gotschalk 1988; Gentien

et aI. 1995; Asper et aI. 1992; MacIntyre et aI. 1995; Syvitski et aI. 1995). Um zuverlässige Daten vor

allem über die in-silll Größe und Sinkgeschwindigkeit zu erhalten, ist in Ergänzung zur Beprobung

mit Sinkstoff-Fallen eine nicht-destruktive Beobachtungsmethode erforderlich. Hierzu eignen sich vor

allem optische und fotografische Methoden (Honjo et aI. 1984; Asper 1987; Lampitt et aI. 1993;

Costello et aI.I 989; Gardner und Walsh 1990). Erst eine Kombination verschiedener Methoden kann

zu einem weitergehenden Verständis des Vertikaltransports führel1.

Das Nährstoffangebot in verschiedenen Ozeantiefen hängt ebenfalls wesentlich von der

Sinkgeschwindigkeit und dem Zerfall fallender und suspendierter Aggregate ab, wobei dem Eintrag

frischen Materials aus dem Oberflächenwasser eine höhere Bedeutung zukommt als der Zufuhr durch

resuspendierte, lateral advektierte Substanz. Neben der hohen Remineralistionsrate in der photischen

Zone erfolgt der Nährstoffeintrag im Oberflächenwasser hauptsächlich über fluviatilen, glazialen oder

atmosphärischen Transport vom Kontinent. Mit jährlich etwa 20 x 109 t (Milliman und Syvitski 1992)

wird durch die Flüsse bei weitem die größte Menge anorganischer, also lithogener Substanz in die

Ozeane geschüttet. Der globale atmosphärische Eintrag hingegen wird auf etwa 1.4 bis 2.0 x 109 t

geschätzt (Duce et aI. 199 I; Schütz und Sebert 1987). Obwohl er nur etwa ein zehntel der durch

Flüsse transportielien Menge beträgt, spielt der atmosphärische Staubtransport durch wesentlich

größere Transpoliweiten von mehreren tausend Kilometern eine bedeutendere Rolle für den

Stoffhaushalt der Ozeane (Sarnthein et aI. 198 I; Duce et aI. 199 I; Kremling und Streu 1993). Den

Extremfall bildet der Austrag von Aerosolen aus der Takla-Makan/Gobi-Wüste über den Pazifik

hinweg bis zu 12000 km weit in arktische Breiten (Prospero 1990). Etwa die Hälfte bis zu zwei

Dritteln des globalen Staubtransportes gehen von den ariden Gebieten Nordwestafrikas aus über den

subtropischen und tropischen Atlantik bis in die Karibik (Prospero 1990), sowie in Richtung Europa

über das Mittelmeer (Buat-Menard et a1. 1989). Die Bedeutung dieses mineralischen Aerosols für den

globalen Stoffkreislauf zeigen auch neuere Arbeiten über den Eintrag in entfernte kontinentale

Ökosysteme wie das Amazonasbecken (Swap et a1. 1992). Aktuelle Messungen bestätigen den

globalen Einfluß des Staubtransportes und zeigen eine deutliche, bisher stark unterschätzte Kopplung

zwischen mineralischen Hintergrund-Aerosol und dem Strahlungsverhalten, der Albedo und dem

Wärmebudged der Atmosphäre zum Teil mit wesentlichem anthropogenem Einfluß aus extensiver

6

Landwirtschaft in ariden Gebieten (Li et al. 1996; Tegen et al. 1996). Nach Staubmessungen und

Modellberechnungen (Schütz 1980, Westphal et al. 1987) fallen über dem Atlantik bis zu 80 % der

jährlich bis zu 260 x 106 t aus den Liefergebieten Sahara und Sahel exportierten Staubes innerhalb der

ersten 1000 km aus. Der signifikante Einfluß dieser Staubmengen und die enge Kopplung VOll

atmosphärischen Zirkulationsmustern und Staubmessungen mit dem Partikelfluß in der Tiefsee vor

Nordwestafrika werden in dieser Arbeit gezeigt.

1.2 Zielsetzung der Arbeit

Die im Rahmen der vorliegenden Dissertation angefertigten Arbeiten sollen Kernaspekte zum

Partikeltransport und Patiikelfluß anhand von Meßergebnissen detailliert beschreiben und einen

Beitrag zum besseren Verständnis der zugrundeliegenden Mechanismen leisten. Die vorliegende

Arbeit basiert daher auffolgenden methodischen und wissenschaftlichen Ansätzen:

1) Um die räumliche und zeitliche Verteilung, das Größenspektrum und die Konzentration mariner

Aggregate in-silu zu erfassen, wurde ein fotografisches Kamerasystem (ParCa) in Anlehnung an

das fotografische Prinzip von Honjo et al. (1984) entwickelt und gebaut. Technisches Ziel war die

Entwicklung des in-silu arbeitenden Systems höchstmöglicher optischer Auflösung bei hohem

Probenvolumen, zum profilierenden und verankerten Einsatz in der Tiefsee. Zur Messung von

veliikalen Aggregatvelieilungen mußte das System vom Schiff aus nach vorgegebenen

Tiefenabständen ausgelöst werden. Hierzu war die Entwicklung einer computergestützten I-Iard

und Software-Schnittstelle notwendig, die an Deck die erforderliche Steuerung des Kamerasystems

über den schiffseigenen Einleiterdraht übernimmt. Für den Einsatz in einer Langzeitverankerung

mußte ein druckdichter autarker Steuercomputer entwickelt werden, der einen stromsparenden

Betrieb nach veränderbaren Programmabläufen gewährleistet. Hierfur wurde die Elektronik von

Kamera und Blitzen geändert und angepaßt. Wissenschaftliches Ziel war die Erfassung von

Aggregatvelieilungsmustern auf Veliikalprofilen im äquatorialen Atlantik sowie im verankerten

Einsatz in etwa 1000 m Wasseliiefe bei den Kanarischen Inseln. Parallel zu Sinkstoff-Fallen

wurden Kamerasystem und ein Influx-Strömungsmesser eingesetzt, der neben Strömung,

Temperatur und Druck über optische Sensoren Chlorophyllgehalt und Mie-Rückstreuung erfaßte.

Anhand des Vergleichs der Datensätze war zu prüfen, ob sich zeitliche Muster im Patiikelsignal

auch bei der Verwendung optischer Methoden nachweisen lassen, und welche Aussagen sich aus

eventuellen Unterschieden in den Messungen fur den Patiikeltransport in der Wassersäule ergeben.

2) Die mit Sinkstoff-Fallen gemessenen Paliikelflüsse an drei verschiedenen Stationen vor NW

Afrika wurden auf saisonale Muster und interanuelle Variationen untersucht. Hierbei sollte die

Signifikanz der Kopplung an Parameter im Oberflächenwasser und in der Atmosphäre erarbeitet

werden. Schwerpunkte lagen dabei auf dem Vergleich mit Variationen im Staubeintrag,

Einleitung 7

atmosphärischen Zirkulationsmustern aus zwei isobaren Schichten der unteren Atmosphäre,

Oberflächenwassertemperaturen und Pigmentdaten aus Satell itenbeobachtungel1. Die Stationen in

unterschiedlichen Bereichen des Auftriebsgebietes vor Nordwestafrika sollten hinsichtlich der

Partikelflußmuster verglichen werden, um an hand von Daten aus jeweils zwei Fallentiefen

Aussagen über den Vertikaltransport zu treffen und den Einfluß von Lateraladvektion und

Resuspension einzuschätzen. Die lithogene Fraktion wurde dabei besonders auf ihre

Korngrößenzusammensetzung untersucht. Der Vergleich von Stationen und Wassertiefen sollte

schließlich in ein schematisches Transportmodell eingehen, um den Einfluß des über die

Atmosphäre eingetragenen Staubes zu bilanzieren und mit Literaturdaten von Akkumulationsraten

im Sediment zu vergleichen. Als methodischer Ansatz sollte dabei die Eignung der

Korngrößenvelieilung in den Sinkstoffproben fUr die Charakterisierung des Vertikaltranspolies

und des Staubeintrages geprüft werden.

Die wissenschaftliche Fragestellung der vorliegenden Arbeit läßt sich in den folgenden 5 Einzelfragen

zusammenfassen:

• Welche Muster im Partikelfluß lassen sich an den Verankerungsstationen differenzieren, und

in welchem Zusammenhang stehen sie zur Saisonalität von Produktion und externem Eintrag

im Oberflächenwasser?

• Wie hoch ist der Anteil der lithogenen Fraktion in den Sinkstoffen, und welche Bedeutung

hat er für den Gesamt-Partikelfluß und die Sedimentbildung an den unterschiedlichen

Stationen?

• Stehen lithogene PaIiikelflüsse in der Tiefsee mit der atmosphärischen Zirkulation in

Zusammenhang und gibt es eine signifikante Kopplung zwischen Partikelflüssen und

atmosphärischen Staubeinträgen?

• Lassen sich über die lithogene Fraktion in den Sinkstoffen und deren Korngrößenverteilung

Aussagen über Transportmechanismen in den Kompartimenten Ozean und Atmosphäre

treffen, wenn ja welche?

• Welche Partikel zeichnen sich für den Veliikaltransport verantwortlich, und wie läßt sich der

Anteil suspendierter und lateral advektierter Partikel in der Wassersäule abschätzen?

1.3. Durchgeführte Arbeiten

Die im Rahmen der Dissertation angefertigten Arbeiten wurden in Form von 4 Manuskripten zur

Veröffentlichung bei internationalen Fachzeitschriften eingereicht. Manuskript 1 ist bereits

8---------------------- ---,-~-,----------------

erschienen, lind Manuskripte 2 und 4 wurden zur Vcröffentlichung akzeptiert. Die einzelnen Arbeitcll

behandeln die unter 1,2 beschriebene Zielsetzung nach folgendcn Aspekten:

Manuskript I:

Ratmeyer V. and G. Wefer (1996) A high resolution camera system (ParCa) 1'01' imaging p:lrticlcs in

the ocean: System design and reslilts from profiles amI a three-month dcployment..JolIl'llo! (!l MUl'Ille

Research, 54. 589-603.

Das Manuskript beinhaltet eme detaillierte technische Beschreibung des methodischen Ans:lt/cs

fotografischer in-si/li Partikelmessung sowie die Diskussion wührend verschiedener Einsiitze

gewonnen Ergebnisse. Nach einem kurzen Überblick über die heute angewendeten Methoden zur

zerstörungsfreien Messung von Partikeln und Aggregaten wird das methodische Konzept des Systcms

vorgestellt. Anschließend erfolgt eine Beschreibung und Diskussion der gemessenen Aggregatgrößen

Verteilungen auf zwei Vertikalprofilen im westlichen und östlichen tropischen Atlantik. sO\vie in der

Sinkstoffallen-Verankerung Cl4 während der Monate Juni bis September 1994. Die Ergebnisse

werden im Vergleich zu parallel gemessenen Daten von M ie- Rückstreuung und Ch lorophyllgeha It

d iskutieli. Die während der Verankerung gewonnenen Ergebnisse werden zudem direkt mit den

Partikelflüssen aus Sinkstoffproben verglichen. Abschließend wird die Vergleichbarkeit der

unterschiedlichen Methoden und ihrer Aussagekraft für die Verteilung und das Verhaltcn

unterschied Iicher Partikelpopulationen d iskutieli.

Manuskript 2:

Ratmeyer V., G. Fischer and G. Wefer. Lithogenic particle fluxes and grain size distributions offNW

Africa: lmplications 1'01' seasonal changes 01' aeolian dust input and downward transport.

Im Druek bei Deep-Sea Research.

Neben den in Manuskript I dargestellten Partikelflüssen aus der Verankerung CI4 werden Amplitude

und saisonale Variation der Iithogenen Partikelfl üssen an drei versch iedenen Verankerungsstationen

(Cl, CH, CV) beschrieben. Zur Messung der Korngrößenverteilung an Sinkstoffen wird eme

Fehlerabschätzung mittels Vergleich verschiedener Methoden und Proben durchgeführt. Die

Korngrößenverteilung der karbonatfreien Fraktion in den Sinkstoffen wird beschrieben und im

Vergleich zu den Flußmustern aus verschiedenen Wassetiiefen diskutieti. Anhand des Vergleiches der

versch iedenen Stationen werden aus den Ergebnissen Rückschlüsse auf untersch ied Iiche

Transpotimechanismen in der Wassersäule gezogen. Anschließend erfolgt der Vergleich der

Sinkstoffdaten mit atmosphärischen Zirkulationsmustern im Hinblick auf den Eintrag terrigener

Partikel aus windtransportieliem Staub. Verschiedene Datensätze zur Beschreibung der

atmosphärischen Situation (Modelltrajektorien, Windmessungen) an den einzelnen Stationen werden

9

beschrieben und auf ihre Korrelation mit Fluß- und Korngrößcnparametern der Sinkstoffe geprüft.

Aus dem Vergleich der Sinkstoffdaten, atmosphärischcr Daten und Litcraturwcrte zu Staubtransport

und Scdimentakkumulationsraten wird abschlicßcnd cin schematisches Modell entworfen, das

mögliche Transportwege und Stoffflüsse zwischen den Kompartimenten Atmosphäre, Wassersäule

und Sediment im Arbeitsgebiet beschreibt und bilanzicrt.

Manuskript 3:

Ratmeyer V., W. Balzer, G. Bergametti, I. Chiapello, G. Fischer and U. Wyputta. Impact of mineral

dust on deep-ocean particle flux in the eastern equatorial Atlantic Ocean.

Eingereicht zur Veröffentlichung bei Geophysical Research Letters.

Basierend auf den Ergebnissen der Arbeiten zu Manuskript 2 wird eme enge Kopplung

atmosphärischer Zirkulationsmuster mit den Lithogen- und AI-Flüssen von Sinkstoffen sowie deren

Korngrößenverteilungen aus 1000 m Wassertiefe diskutiert. Hierzu werden an der Station CV

gemessene Sinkstoffdaten mit atmosphärischen Trajektorien aus zwei repräsentativen Druckniveaus

verglichen und simultan erfaßten atmosphärischen AI-Konzentrationen von der Kapverdischen Insel

Sal gegenübergestellt. Als Schlußfolgerungen werden Aussagen über die Bedeutung des

atmosphärischen Staubtransportes für den Paliikeleintrag in den subtropischen Nordatlantik und die

Sedimentbildung in diesem Gebiet getroffen.

Manuskript 4:

Neuer S., V. Ratmeyer, R. Davenpoli, G. Fischer and G. Wefer (1997) Deep-water particle flux in the

Canary Island region: Seasonal variation in relation to long-term satellite derived pigment data and

lateral sources.

Im Druck bei Deep-Sea Research.

Der Beitrag zu diesem Manuskript bestand 111 der Messung der Lithogenflüsse und

Korngrößenvelieilungen an der Verankerungsstation CI, sowie in der Diskussion von Parametern der

Korngrößenverteilung im Hinblick auf ungestörten versus lateral beeinflußten Veliikaltransport an

dieser Station. Im Vergleich mit den übrigen Partialflüssen der Komponenten Corg, Ckarb und Opal

konnten basierend auf dem in Mauskript 2 durchgefliluien Stationsvergleich Aussagen über die

Korrelation von Stofflüssen und Transportmechanismen zwischen 1000 und 3000 m Wasseliiefe

getroffen werden. Die Ergebnisse werden historischen CZCS-Daten zur Pigmentverteilung aus den

angrenzenden Küstenauftriebsgebieten gegenübergestellt und im Hinblick auf laterale Advektion aus

westwälis driftenden Auftriebsfilamenten diskutiert.

]0

1.4 Arbeitsgebiet

Das folgende Kapitel soll in Ergänzung zu den Manuskripten einen kurze Einführung in die

ozeanographischen und atmosphärischen Charakteristika des Arbeitsgebietes vor Nordwestafl'ika

geben. Die Verankerungsstationen liegen zwischen 10° und 300 N Im Grenzbereich des

nordwestafrikanischen Küstenauftriebs zum subtropischen Nordatlantik. Dieses Gebiet zeichnet sich

durch ozeanische und atmosphärische Besonderheiten aus, die an dieser Stelle im LJberblick

dargestellt werden sollen. Eine detaillierte Beschreibung findet sich im Manuskript 2 (Seite 43).

1.4.1 Ozeanzirkulation

Vier dom inierende Komponenten bestimmen das Muster der Oberllächenzirkulation im östl ichen

subtropischen Nordatlantik. Dazu gehören der Kanaren-Strom (CC), der Nord-Äquatoriale

Gegenstrom (NECC), der Nord-Äquatorial-Strom (NEC) sowie randlich die nördlichsten Ausläufer

des Benguela-Stroms bzw. Süd-Äquatorial-Stroms (SEC) (Abb. 1a) (siehe auch Stramma und Siedler

1988; Klein und Siedler] 989). Alle Verankerungsstationen im Arbeitsgebiet liegen im Bereich dcs

Kanaren-Stroms bzw. dessen Ausläufern zum Nord-Äquatorial-Strom südlich etwa 15°N. Die enge

Kopplung der ozeanischen Zirkulation mit den stetigen meridionalen Passatwinden (s.u.) zeigt sich in

dem nahezu ganzjährigen, etwa 50 bis 70 km breiten Küstenauftrieb mit maximaler Intensität bei etwa

200 N (Cape Blanc) (Mittelstaedt 1991). Stark abgeschwächt und mit saisonalem Muster setzt sich

diese Auftriebszone nach Süden bis etwa 10-12°N, 20 0 W (südlich Cape Verde) fort, wo nur noch

zwischen November und Februar Tiefenwasser aus dem Bereich des Süd-Atlantischen Zentral

Wassers (SACW) bis in das Obernächenwasser aufsteigt (Hughes und Balion ]984). Entsprechend

dieser Situation besteht ein starker Gradient von Biomasse und Partikelfracht zwischen Auftriebszone

und offenem Ozean nahezu entlang der gesamten nordwestafrikanischen Küste bis etwa 300 N.

Filamente aus den Auftriebsgebieten vor Cape Ghir, Cape Bojador und Cape Blanc driften jedoch aus

dem Schelfbereich bis zu mehrere hundert Kilometer westwärts und bilden so einen wichtigen

überregionalen TranspOlimechanismus von Biomasse und Nährstoffen in Richtung des offenen

Ozeans (Nykjaer 1988; Nykjaer und Van Camp 1994).

Einleitung 1]

I/1

'11I

)

15°20°

/

25°W

b

30°

35or;:===;==~~==~"=;=====;====~

N

20°

25°

15°

10°

35° I(li) \C'c,_'_,'" --.--,

N f)' ai

30° r

11

\Ji1l. § )1\

10t <1-----<>., i':;J~ '~\'"1, ....-- ..,i :.I! i, ~c200011 "r"("'~NEC:C' ~,I"kL.~2 '/ 'I lu/Li i ~,~._,-~~

25°W 20° 15° 10°

Abb. 1: (a) Oberflächenströmungen und Bathymetrie vor der nordwestafrikanischen Küste,Eingezeichnet sind der Kanarenstrom (CC), der Nord-Äquatorial-Strom (NEC) und der NordÄquatoriale Gegenstrom (NECC), (b) Schematische Karte der Hauptwindrichtungen imPassatniveau (bis etwa 2 km Höhe, Druckniveau 850 hPa) und im Saharan Air Layer (SAL, 3-6km Höhe, Druckniveau 500 hPa). Die gestrichelte Linie markiert die nördlichste Ausdehnungder Intertropischen Konvergenz-Zone im Juli!August (ITCZ).

1.4.2 Atmosphärische Zirkulation und Staubtransport

Neben der nahezu ganzjährigen Nährstoffzufuhr aus den Auftriebsgebieten besteht im Arbeitsgebiet

ein wichtiger Nährstoff- und Partikeleintrag aus der Atmosphäre. Durch extensive Erosion in

Küstengebieten Marokkos und Mauretaniens sowie aus der Sahel-Region und der Sahara wird jährlich

etwa ein Drittel der globalen atmosphärischen Staubfracht über den östlichen subtropischen Atlantik

transportiert. Der Staub wird entweder über den Nord-Ost-Passat oder in den 3 - 6 km hohen

Luftmassen des Saharan Air Layer (SAL) aus den Liefergebieten im Kontinentinneren transportiert.

Dort entstehen heiße, hochenergetische Winde ("Harmattan") auf Oberflächen-Niveau mit extremen

Staubkonzentrationen, die über thermische Konvektion große Mengen Staub in den sogenannten

"African Dust Plume" zwischen 1 und 6 km Höhe injezieren (Kalu 1979; Tetzlaff und Wolter 1980).

Maximale Staubexporte werden im Sommer (Juni bis August) erreicht, wenn sich westwärts gerichtete

12

Winde in dieser Höhe mit erhöhter Geschwindigkeit im sogenannten "African Easterly Jet" zwischen

17° und 21 oN konzentrieren (Abb. Ib). Auf dem Weg über den subtropischen bis tropischen Atl<lI1tik

crreichen zwischen 5 und 20 (Yo der ursprünglichen Staubfracht nach eimvöchigem Transport

schließlich elie Karibik und elcn slidamerikanischen Kontincnt. Nach Modcllberechnungen und

optischen Messungen (Carlson and Prospero 1972; Schütz et al. 198 I; Westphal ct al. 1987) ist dei' im

SAL transpOitierte Stau b gegenü bcr dcm Passat-transportiertcn dom in icrend.

Aufnieelrigen Druckniveaus in Höhcn bis zu 2 km bildet dcr Passat das wichtigste TransportmediullJ.

Gesteuert durch den saisonalen Wcchsel elcr Intertropischen Konvergcnz-Zonc (ITCZ), erreicht der

Nord-Ost-Passat seine maximalc Ausdehnung und Staubfracht in elen Wintermonaten zwischcn

November und März. Als Übcrgangszone zwischen Passat- unel Monsunwindcn crreicht elie 1TCZ ihre

nödlichste Position bei etwa 12°N (im IGistenbereich) im Juli/August. Nördlich der direkt VOll der

ITCZ beeinflußten Zone besteht ein nahezu kontinuierlicher, ganzjährigcr StaubtranspOlt durch den

Passat zwischen etwa 15° und 25°N, dcr vor allem im Sommer einen Großteil seiner Staubfracht aus

dem elarüberliegenelen SAL bezieht.

Stärkste saisonale Schwankungen der Luftbewegungen beider Druckn iveaus entstehen Im

Arbeitsgebiet in der südlichen Region zwischen 10° und etwa 15°N, also im nördlichsten direkten

Einflußbereich der ITCZ. Radiosonden-Messungen (Chiappello et al. 1995) unel Trajektorien

Berechnungen (Manuskripte 2 und 3 dieser Arbeit) zeIgen hier gegensätzliche winterliche und

sommerliche Muster in Passatniveau und SAL.

1.5 Methodik

Entsprechend der unterschiedlichen Arbeitsansätze zur Erfassung und Quantifizierung des

Partikeltranspoltes mit Kamerasystem und Sinkstoff-Fallen kamen grundlegend verschiedene

Methoden zur Anwendung. Der Einsatz der Partikelkamera erfolgte profilierend auf den Reisen M22

1 und M23-3 mit FS Meteor bis zu einer Tiefe von 1000 m im äquatorialen Atlantik, sowie in den

Monaten Jun ibis Septem bel' 1994 bei den Kanarischen Inseln in der Sinkstoffallen-Verankerung CI4

in 995 m Wasscrtiefe. Gleichzeitig wurden Sinkstoffproben aus verankerten Sinkstoff-Fallen

zwischen 900 und 4000 m Wassertiefe an 3 Stationen vor NW-Afrika untersucht (Abb. 2). Zum

Vergleich der Sinkstoffdaten mit Oberflächensedimenten wurden schließlich Proben des obersten

Zentimeters aus Multicorer- Kernen eier verschiedenen Verankerungsstationen bearbeitet.

Einleitung-----='----------------_.--~.~-~._~~~-~_._---~.~~.._..._.,----- 13

40oN

Azoren

30 CI Q"y/Kanarisd)() fnseln

Cape GlJir

Cape Bojador

20

10

CS T

p .. :

KapverdIsehe Inseln • b

CV T

10 20 oE

0 - - - - - - -0-0

GeoB 2217

GeoB 2212

1050 0 W 40 30 20 10 0

Abb. 2: Karte mit den Verankerungsstationen CI, CB und CV sowie Partikelkamera-Einsätzenan den Stationen GeoB 2212 und GeoB 2217. Verankerungspositionen sind durch ein Dreieck,Kameraeinsätze durch ein Karo gekennzeichnet.

1.5.1 Partikelkamera

Sowohl die technische Entwicklung wie auch der Einsatz der Partikelkamera zur Messung der in-silu

Partikelvelteilung waren ein Ziel dieser Arbeit. Aus diesem Grunde wurde - neben den

Meßergebnissen - eine detaillierte Systembeschreibung des Kamerasystems und seiner technischen

Komponenten veröffentlicht (Seite 25). Bezüglich der Arbeitsweise des Systems und eigentlichen

Methodik der fotografischen in-situ-Pmtikelmessung wird auf diese Arbeit verwiesen. Im Rahmen

dieses Kapitels wird in Ergänzung dazu ein Überblick über die zum Bau und Einsatz notwendigen

Entwicklungsschritte, die fotochemische Entwicklung und die digitale AusweItung des Bildmaterials

gegeben.

Systementwicklung

Entsprechend der Forderung nach höchstmäglicher optischer Auflösung zur Messung auch kleinster

Partikel < 100 11m wurde ein Mittelformat-Fotosystem statt einer Videokamera gewählt. Da die

Wahrscheinlichkeit, große sinkende Aggregate in der Wassersäule anzutreffen entsprechend ihrer

hohen Sinkgeschwindigkeit gering ist (Asper et al. 1992), mußte ein optimales Volumen

Auflösungsverhältnis gefunden werden. Da das maximale Probenvolumen durch die Blitzstärke und

14

eIie Brennweite eIer Kollimatorlinsen eIer Beleuchtungseinheit vorgegeben war, wurde eIas Problc~nl

über eine Belichtung mit Mehrfachaufnahmen gelöst. In eier verankerten Version erlauben

Mehrfachbelichtungen die Fotografie eines Volumens eIer n-fachen Größe eIes Originalvolul1lens mit

Intervallzeiten größer als eIie kleinste angenommene Partikel-Sinkgesehwindigkeit, wobei !l die

Anzahl der Bel ichtungen pro Bi leI ist. ZusätzIich wurde versucht, l1l it Mehrfach bel ichtu ngen

Transportwege einzelner Partikel festzuhalten, was jedoch aufgruneI zu langer BlitzlaeIezeiten nicht

erfolgreich war. Im profilierenden Einsatz wureIe eIie Mehrfachbelichtungsoption nicht genutzt. Zur

Steuerung eIes Systems wureIen zwei Microcol1lputer (basiereneI auf PIC-Prozessoren) eingebau t, die

während des profi lierenden Einsatzes die Koml1lun ikation mit dem Decksrechner au I' dem Seil iCf

übernehmen sowie Auslösesignal uneI Bildnummer an Kamera uneI Blitze weiterleiten. Zur Anpassung

der käuflichen Komponenten Kamera und Blitze an die Steuercomputer wie auch zur Modifikation auf

Mehrfachbelichtung waren umfangreiche Änderungen an der Elektronik bzw. Neuentwicklungen von

mehreren elektronischen Schaltkreisen notwendig. Zur Triggerung des Systems im profilierenden

Einsatz wurde eine Software in der Programmiersprache Turbopascal (mit wesentlichem Beitrag von

G. Ruhland, Bremen) und später unter Visual Basic geschrieben, die das System über einen pe an

Deck des Forschungsschiffes automatisch nach Zeit- oder Tiefenintervallen steuert sowie Bild- und

Schiffsdaten protokollieJi. Während des verankerten Einsatzes übernahmen die im System

eingebauten Computer die gesamte Steuerung einschließlich der Stromversorgung nach frei

programmierbaren Steuerdaten. Der Einsatz während der Meteor-Reise M22- I im September/Oktober

1992 diente weitgehend der technischen Entwicklung und dem Test des Systems unter

Hochseebedingungen. Während der Reise M23-3 im März 1993 konnten auf mehreren Profilen

erfolgreich Bildserien aufgenommen werden. Der verankerte Einsatz lieferte ebenfalls das gewünschte

Bildmaterial. Das System ist in eIer jetzt vorliegenden Form für den mehnnonatigen Einsatz in einer

Wassertiefe bis zu 6000 mausgelegt.

Bildbearbeitung

Die Entwicklung des Film-Materials erfolgte in den meisten Fällen direkt an Bord. Filmlängen von

maximal 10m konnten am Stück entwickelt werden, größere Längen (30 m) wurden entsprechend

gekürzt. Die Entwicklung erfolgte mit Kodak HC-IIO Entwickler in 5 I Tanks, beim Kodak

Technical-Pan Film zu niedrigem Kontrastindex bei 400 ISO, beim Kodak Tri-X-Pan Film zu

mittlerem Kontrast bei 800 bis 1600 ISO. Abzüge zur Dokumentation wurden in verschiedenen

Formaten auf kunststoffbeschichtetem Positiv- und Negativpapier hergestellt. Digitalisierung,

Aufbereitung und Analyse der Bilder erfolgte anschließend direkt vom Negativ am PC unter

Verwendung des Bildanalysesystems Optimas 5.2. Abbildung 3 zeigt den hierzu installieJien

Systemaufbau mit PC, modifiziertem Vergrößerungsgerät und CCD-Videokamera.

Einleitung

Bildanalysesystembestehend aus:

1 PC, hochauflösender Monitor 17"2 Fotografisches Vergrößerungsgerät3 Einstellbarer verlängerter Tubus4 CCD-Kamera 768 x 512 Pixel

1

(3

4

15

Software:

Optimas 5.2 BildanalysesystemOptimas All C-CompilerMicrosoft Excel Tabellenkalkulation

Auschnitt:A Höchste Auflösung

Kleinste Partikel

Gesamtbild:B Geringste Auflösung

Makro-Partikel

Abb. 3: Schematischer Aufbau des Bildanalysesystems. Das Negativ wird direkt auf denBildwandler der Videokamera (Sony 1/2" CCD) projeziert. Der verlängerte Tubus zwischenNegativ und Kamera ermöglicht Abbildungsmaßstäbe von 8: 1 bis zu 2: I mit einem 80 mmProjektionsobjektiv (Apo Rodagon-N). Die Digitalisierung erfolgt im PC (Pentium P60, 16 MBRAM) über eine Framegrabberkarte mit einer Auflösung von 768 x 512 Bildelementen (Pixeln)und einer Farbtiefe von 24 Bit (Image VGAplus 24).

Das nacheinander über mehrere Vergrößerungsstufen eingelesene Bild wurde unter Optimas

digitalisiert und zur weiteren Bearbeitung auf Festplatte gespeichert. Gegenüber alternativen

Erfassungsmethoden wie Scanner oder Mikroskop mit Videoaufsatz konnte mit diesem Aufbau das

gesamte Größenspektrum der Partikel von 80 bis über 50000 IAm in repräsentativen Bildausschnitten

erfaßt werden, indem zunächst das Gesamtbild, dann mehrere Ausschnitte auf den Bildwandler der

Kamera projeziert wurden.

Zur Flächen- und Größenanalyse der digitalisierten Bilder können verschiedene Verfahren

angewendet werden, die über eine Reihe von Routinen digitaler Bildaufbereitung schließlich ein

binäres und damit zählbares und statistisch auswertbares Bild liefern. Zur Binarisierung muß ein

Schwellenwert ("Threshold") auf der Grauwert- oder Farbskala des Bildes als Grenze zwischen

Objekten und Bildhintergrund definiert werden. Diese Schwellwert-Festlegung stellt das eigentliche

methodische Problem der automatischen Partikelanalyse dar. Die Binarisierung wird um so

16

schwieriger, je höher die Grauwertvarianz des Bildhintergrundes im Vergleich zu den zu messendcn

Vordergrundobjekten (Partikel) ist, da sich bei sehr ungleichem Hintergrund kein eindimensionaler

Schwellwert festlegen läßt. Verschiedene Methoden werden zur Lösung dieses Problems bei

ähnlichen Aufgaben der Patiikelfotografie in der Mikrobiologie eingesetzt. Dort sind Verfahren zur

Grenzfindung durch mathematische Ableitungen von Grauwertprofilen (Sieracki et al. 1989) sowie

die Hintergrundabschwächung und Objektseparation durch Differenzoperatoren (z.B. Schröder ct al.

1990) standarisiert. Innerhalb eines bestimmten Größenspektrums erlaubt vor allem die Filterung mit

Operatoren wie der "Mexican Hat"-Matrix eine effektive Hintergrundabschwächung bei gleichzeitiger

Anhebung des Vordergrundes trotz ursprünglich flacher Kantengradienten (Abb. 4).

Gauss Operator (G)

(

" 6 4 ')4162416 <IG= 62436246

4162416 <I

1<16" 1

Originalbild (optische I Gauss'sche Unschärfe)

,"

"Mexican Hat" - Operator

(

.'.2.3.2.' )·2·24.2.2

D = .34244·3·2·24·2·2_\.2·3·2_1

Differenziertes Bild mit "Mexican Hat"-Operator

Abb. 4: Beispiele fl.ir Differenzoperatoren zur Bildaufbereitung. Die Filtermatrix besteht auseinem Kernwert und der Restmatrix. Auf eine gleichgroße Bildmatrix angewendet, verändert sieden Kernweli der Bildmatrix nach den Eigenschaften (Grauwerten) seiner Nachbarwerte. Beieinfachen Differenzoperatoren wie dem Gauss-Filter oder dem "Mexican Hat" wird derKernwert der Bildmatrix durch Multiplikation und Aufsummierung seiner Nachbarwerte mitden äquivalenten Werten der Filtermatrix neuberechnet. Beim Gauss Operator führt dies zueiner Verunschärfung des Bildes vergleichbar mit optischer Unschärfe (a), beim "Mexican Hat"zu dem gegenteiligen Effekt mit starker Versteilung der Vordergrundgradienten bis hin zu einerringförmigen Depression (b). Im gezeigten Beispiel (Foraminiferen mit etwa 200 11mDurchmesser, aufgenommen in einem beleuchteten Volumen von etwa 38 I) wird dadurch dieHelligkeit bzw. der Grauwert des Objektes ohne Vergrößerung seines Durchmessers sehr starkvom Hintergrund abgehoben und somit meßbar.

Einleitung 17

Je breiter das Größenspektrum der Partikel jedoch ist, desto fehlerhafter wird diese Methode

(Schröder et a!. 1990; Psenner 1991) aufgrund zwangsläufig variierender Größenverhältnisse

zwischen Operatormatrix und den zu messenden Objekten. Für die Messung von marinen Partikeln

eines vergleichsweise breiten Größenspektrums - zum Tei I nahe an der Auflösunggrenze des

Erfassungssystems - erwiesen sich beide Methoden als unzureichend. Eine weitere Schwierigkeit war

die Eliminierung großer Partikel, die außerhalb des Tiefenschärfebereichs der Partikelkamera randlich

mitbeleuchtet wurden. Sobald diese Objekte hinter dem Probenvolumen lagen, bildeten sie einen stark

vom Hintergrundspektrum des Restbildes abweichenden Grauwert. Die Separation kleinerer, korrekt

beleuchteter Objekte innerhalb des Probenvolumens vor einem solchen Objekt war mit

Differenzoperatoren wie dem "Mexican Hat" nicht durchführbar. Eine vergleichsweise einfache, aber

sehr effektive Methode zur Bildaufbereitung ohne Differenzfilterung ist die Hintergrundseparation

durch Dilatation und anschließende Subtraktion (Abb 5a-e). Ein ähnliches Verfahren wendete Psenner

(1991) zur Messung von Bakterien auf Aggregaten an, jedoch mit optischer Unschärfe durch

Defokussierung des Mikroskopes. Diese Methode ist weitgehend größenunabhängig, erfordert

allerdings mehrere Arbeitsschritte bis zum binären Bild. Die Dilatationsmatrix (5x5) erwies sich als

ausreichend für alle Aggregate bis zu einer maximalen Achsenlänge von 25 mm. Größere Aggregate

wurden von Hand selektiert, Bilder mit erkennbarem Zooplankton wurden aussortiert. Das so

aufbereitete Bild wurde vom Original subtrahiert, das Restbild mit den zu messenden Partikeln

binarisieli und ausgewertet. Das Ergebnis ist bezüglich der Vordergrundoptimierung vergleichbar zur

Filterung mit dem "Mexican Hat" bei optimalen Größenverhältnissen (Abb. 6).

Aus den identifizierten Partikeln wurden Achsenlängen, Equivalentdurchmesser und Flächenwerte

extrahieli und automatisch in eine Excel-Tabelle exportiert. Aufgrund der großen Zahl zu messender

Bilder wurde im Rahmen der unter OPTIMAS compilierbaren Programmiersprache (ALI-C) ein

Programm zur automatischen Aufbereitung und Analyse nach dem oben beschriebenen Prinzip

erstellt. Nach mehreren Testserien erwies sich die Automation als sehr zuverlässig, da sich die

Lichtverhältnisse auf den Bildern während eines Einsatzes nicht änderten. Es wurden insgesamt etwa

1500 Einzelbilder digitalisiert und bearbeitet.

18

Digitalisierter Ausschnitt (2)

Digitalisierung vom Negativmit SIW GGD-Kamera

60

unebenes Hintergrundprofil

Digitalisierter Ausschnitt (3)768 x 512 Pixel

max. Auflösung: 80 ~m



Negativ 60 x 60 mm



20 40Profil-Länge (X-Intervall)

255Grauwert

~ ~

1~~ ~t.

•

[>

Ausschnittsvergrößerung (3)

Bildhintergrund nach Dilatation durch Maximum Operator

Maximum Operator [F)

(

00000)o 0 0 00F =Max 00000

o 0 0 00o 0 0 0 0

Grauwertverteilungdes Bildhintergrundes

Abb. 5a-c: Im Rahmen dieser Arbeit angewendetes Verfahren zur Hintergrundsubtraktion.Dargestellt ist die Digitalisierung verschiedener Ausschnitte des Negatives (a) mit einem großenSpektrum an PaIiikelgrößen und -formen. Durch die Streuwirkung kleinster PaIiikel, die aufdem Negativ nicht mehr aufgelöst werden können, und durch das Wasser selbst besteht einschlechtes Kontrastverhältnis mit sichtbaren Hintergrundschwankungen, die die Messungverfälschen können (b). Durch optische oder digitale Verunschärfung bzw. durch Dilatation miteinem Maximum-Operator wird der Hintergrund mit seinen Unebenheiten separiert. (c). ImGegensatz zur Verunschärfung wird bei dieser Dilatation nicht nur das Kernelement derBildmatrix geändert, sondern der Maximalwert innerhalb der Bildmatrix gesucht und - nachAddition mit den Filterwerten ("Offset") - auf die gesamte Matrix ausgeweitet. Dies fuhrt durchVergrößerung der hellen Bereiche zur Überlagerung und damit Eliminierung allerVordergrundpixel und eignet sich so zur Separation einer Hintergrund-Grauverteilung.

Einleitung 19

Ergebnis der Hinlergrundsublraktlon

\ ...

•

~B

o Grauwert

20 40Profi~LAnge (X-Intervall)

60

B

Blnarisienmg nadl Schwarz-lWeißpunkt (Threshold)Flächenldenliflkalion

60

ESO - Verteilung gezahlter FlAcI1en

15 30 45gezAMe flAche ft.

Extraktion von FlächeMaximale und minimaie AchseEquivalenldurchmesser (ESO)Formparameter (Streckung)

Cl 1.5 r----~---~------,~

i~ 1.0

H8-0.5j! 0.0

0

Blnarisierte Bnmap zur Flächenauszählung

Abb.5d-e: (d) Nach Subtraktion des Hintergrundes und Grauwertspreizung bleiben fastausschließlich Vordergundpixel übrig, die nun einen wesentlich höheren Abstand zumHintergrund haben. (e) Nach Definition des Schwellwertes (Threshold) erfolgt dieBinarisierung, d.h. die Trennung des Bildes nach Vorder- und Hintergrund mit zweiGrauwerten. Die anschließende Analyse der Vordergrundpixel markiert Einzelflächen undliefert Größen- und Formparameter der Partikel.

20

1 2

A Originalbild

flacher Gradient

steiler Gradient

0

40

80 A~ 120;::J

l" 1600

200 8240

0 20 40 60 80 100 120 140

Pixel-Inervall

J1l 0~ 2 I

Abb. 6: Anwendung der Dilatation und Hintergrundsubtraktion auf zwei extrem vergrößerteBildausschnitte. (A) Im Ausschnitt 1 ist ein relativ gut kontrastiertes Einzel-Objekt erkennbar,Auschnitt 2 zeigt drei einzelne Objekte vor einem dunklen, stark unscharfen und sehr schwachkontrastierten Hintergrund. (B) Die Subtraktion hebt in beiden Bildern den Vordergrund auf einmeßbares Niveau nehezu ohne räumliche Verzerrung. Auch in Ausschnitt 2 können die Partikelnun deutlich vom unscharfen Bereich getrennt werden. Die Effizienz der Hintergrundabschwächung und Gradient-Versteilung ist vergleichbar mit der Differenzfilterung durch einenoptimal dimensionierten "Mexican Hat" Operator.

Einleitung

Vergleich der Bilddaten mit SinkstojjjJroben

21

wobei

Da mit der Paltikelkamera nur Aggreagtkonzentrationen und Größenverteilungen gemessen werden

konnten, war ein direkter Vergleich mit den Partikelflüssen aus den Sinkstoffproben nicht möglich.

Um dennoch eine Abschätzung der Vergleichbarkeit der mit beiden Methoden gemessenen Daten zu

erhalten, wurde der Gesamtfluß Ftotal der Sinkstoffe unter Annahme einer geschätzten

Sinkgeschwindigkeit Vs in eine Konzentration Ctotal bzw. CAgg umgerechnet:

Ctotal = Ftotal * Vs-I (mg m-3)

Vs = 100 md-I .

Die Trockengewichte MAgg der ausgezählten Aggregate wurden nach Alldregde und Gotschalk

(1988) berechnet:

MAgg C!-tg) = 0.0018 * D1.25

mit D = Durchmesser eines einzelnen Aggregates (mm).

Die Massenkonzentration ergab sich somit aus der Anzahl n der Aggregate pro Liter Probenvolumen

multipliziert mit ihrem Gewicht im Volumen von einem Kubikmeter:

CAgg = 1000 * n * M (mg m-3)

Die Anwendung dieser Berechnung diente zur Abschätzung der Vergleichbarkeit beider Methoden

und dem Vergleich der Ergebnisse mit Literaturwerten.

1.5.2 Bearbeitung der Sinkstoffproben

Zur Messung der Korngrößen und Iithogenen PaJiikelflüsse an 3 Tiefseestationen (CIl-4, CBl-4,

CVl-2) vor NW-Afrika wurden 349 Sinkstoffproben aus insgesamt 10 Verankerungen untersucht. Ein

Schema zur Probenaufbereitung und Bearbeitung der Sinkstoffe zur Korngrößenanalyse ist in Abb. 7

dargestellt. An der CB-Station gemessene Partikelflüsse wurden bereits von Fischer et al. (1996a)

veröffentlicht. Außer der Aufbereitung für die Korngrößenanalyse ist eine detaillierte Beschreibung

der Probenaufbereitung bei Fischer und Wefer (1991) nachzulesen. Zur Charakterisierung der

Korngrößenverteilung wurden die aufbereiteten Proben mit einem CIS-GALAI Laser Granulometer

gemessen. Eine Fehlerabschätzung erfolgte zusammen mit der Messung von Oberflächensedimenten

mit den Geräten MICROMERITICS Sedigraph und FRITSCH Laser Diffraktometer. Die Aufbereitung,

Fehlerbetrachtung und eine abschließende Bewertung der Anwendbarkeit der verschiedenen Systeme

für die Analyse Sinkstoffproben ist detailliert in Manuskript 2 (Seite 43) beschrieben. Die Messungen

mit Laser GramJlometer und Sedigraph erfolgten am Alfred-Wegener-Institut für Polar- und

Meeresforschung in Bremerhaven.

22

Da der Laser Granulometel' als e1l1z1ges Instrument die Korngrößen-Messung in den gering

konzentrielien Partikelproben ermöglichte, soll an dieser Stelle kurz auf die Funktionsweise dcs

Gerätes eingegangen werden. Eine detaillierte Beschreibung des Meßprinzips findet sich in

Aharonson et al. (1986). Das Meßprinzip des Laser Granulometers basiert auf der Detektion dcr

Abschattungszeiten und -intensitäten eines schnell rotierenden HelNe Lasers durch Partike I mit

mehreren Photozellen. Da die 111 der Meßzelle suspendierten Partikel praktisch keine

Eigengeschwindigkeit gegenüber dem Laser haben, ist die Überstreichungs- und dalll it

Abschattungszeit auf der Photozelle direkt proportional zur Größe des Partikels. Für die genaue

Messung muß sich das Paliikel im Fokus des Laserstrahis befinden, da sonst zu große Schatten eine

falsche Größe ergeben. Da jedoch die Schattendichten (Randschärfen) von verschieden fokussierten

Partikeln detektieli werden, können nicht fokussierte Messungen erkannt und eliminiert werden. tim

reproduzierbare Meßergebnisse zu erhalten, ist die Probe zur Vermeidung zu vieler Überlagerungen

im Fokusbereich des Lasers sehr gering konzentriert zu halten. Die Detektionsgrenze liegt daher weit

unter der eines Sedigraphen oder Laser Diffraktometers. Zusätzlich zum Meßsystem ist das Gerüt mit

einer Video-gestützten Bildanalyse ausgestattet, die jedoch im Rahmen der durchgeführten Arbeiten

ausschließlich zur Überprüfung des Dispergierungsgrades der Proben genutzt wurde. Für die Messung

der Sinkstoff-Proben wurde eine geschlossene 2.0 ml Küvette als Meßzelle verwendet, da die Proben

für eine DurchflußzeIle zu gering konzentrieli waren. Aus der homogenisielien Gesamtprobe mußten

daher mit der Pipette Teilproben zur Messung entnommen werden. Durch die geringe

Probenkonzentration war der mögliche Fehler der Messung gegenüber der Gesamtvelieilung in der

Ausgangsprobe vergleichsweise groß. Die Messung von bis zu 5 Teilproben mit jeweils 10

Meßzyklen pro Sinkstoffprobe erwies sich jedoch als repräsentativ. Die vergleichende Messung von

Oberflächensedimenten mit Laser-Granulometer, Laser-Diffraktometer und Sedigraph zeigte eine

signifikante Unterschätzung des Feinkornanteils bei Proben mit hohem Ton- und Feinsiltgehalt durch

beide optischen Verfahren. Die Höhe und Lage des Modalwertes der Korngrößenverteilung in Proben

mit hohem Feinkornanteil verschob sich jedoch deutlich in den Feinsiltbereich um 10 flm. Für die

Interpretation der Messungen wurde daher das Verhältnis der Fraktionen 6-11 flm und 20-63 flm

sowie der Mittlere Durchmesser (Mean) als Kriterium bewertet, im Gegensatz zu den sonst üblichen

Parametern Modalweli, SOliierung und Schiefe. Für die Betrachtung der Flüsse wurde die Fraktion 6

63 flm auf 100% normiert. Eine detaillielte Beschreibung der zur Fehlerabschätzung durchgefülllien

Messungen findet sich in Manuskript 2 (Seite 43).

Einleitung

Bearbeitung der Sinkstoffprobenzur Korngrößenanalyse

1/16 bis 1/5 Split pro Probe

23

Dispergierungirn UltraschIlbad rnin. 2 Min

2 rnl Teilprobenentnahrne

Karbonatlösung,Aufoxidierung des Cocg

Neutralisation:

Waschen:

Lagerung:

vor der Messung:

Messung:

DZugabe von

20% iger Essigsäure7% igern HP2

DDispergierung rnit Ultraschall,

rnikroskopische Kontrolle

DZugabe von

2% iger NH3- Lösung

DAbsieben

der Fraktion> 63 ~rn

(keine der Probenenthielt rneßbare Fraktion

> 63 ~rn)

Ddurch Zentrifugieren

bei 4000 rprnin deinonisiertern Wasser

Dbei 4°C

D

DLaser-Granulometerunter Videokontrolle

des Dispergierungsgrades

D

48 bis 60 hbei 20°C

2 bis 3 x

2x15Min

10 Meßzyklen,bis zu 5 Teilproben

Analyse: Mittelwert der Häufigkeitsverteilungenaus allen Teilproben pro Probe;Umrechnung in Volumenverteilungen;Berechnung der Mean-, Median und Modalwerte;Berechnung der normierten Größenklassen6-11 ~m, 11-20~m, 20- 63~mbezogen auf die Fraktion 6-63~m.

Abb. 7: Schema zur Aufbereitung der Sinkstoffproben zur Korngrößenanalyse. Oberflächensedimentproben wurden in gleicher Weise behandelt, die Messung erfolgte jedoch zusätzlich anSedigraph und Laser-Diffraktometer.

1.5.3 Bearbeitung der atmosphärischen Daten

Isobare 4-tägige Zieltrajektorien aus den Drucklevels 850 hPa und 500 hPa wurden von U. Wyputta,

(AG Paläozeanographische ModelIierung, Bremen), nach der Methode von Petterssen (1956)

berechnet und freundlicherweise zur Auswertung zur Verfügung gestellt. Die Auswertung erfolgte

anhand manueller Auszählung und Klassifizierung der Herkunft und des Pfades isobarer Windfelder

24

nach ozeanischem und kontinentalem Ursprung. Die Häufigkeit von Trajektorien kontinentaler

Herkunft pro Zeitintervall wurde als Maß für mögliche Staubfracht angenommen. Zwei Datensätze

standen zur Verfiigung: Ein Modelldatensatz mit einer Laufzeit von 21 Jahren mit tägl iehen

Trajektorien aus beiden Drucklevels wurde zum Vergleich der monatlichen Mittelwerte des

PaJiikelflusses und der atmosphärischen Zirkulationsmuster verwendet. Im Manuskript 2 (Seite 43)

liegt dieser Datensatz den Berechnungen zu Saisonalitätsindizes an den drei Verankerungsstationen

sowie dem TranspOlimodell zugrunde. Eine zweite Berechnung von täglichen Trajektorien auf der

Grundlage von gemessenen meridionalen und zonalen Windkomponenten (Daten vom European

Centre for Medium Range Weatherforecasts, ECMWF) zwischen 1992 und 1994 an der Station CV

wurde rur den direkten Vergleich von Korngrößen, Partikelflüssen und atmosphärischem Staubeintrag

im Manuskript 3 (Seite 91) verwendet.

Zur Beschreibung der oberflächennahen Winde wurden darüber hinaus Meßdaten des Deutschen

Wetterdienstes, Hamburg, von den Stationen Las Palmas, Safi, Nouadhibou, Sal und Dakar

verwendet.

Veröffentlichungen 25

2. Veröffentlichungen

Nachfolgend sind die im Rahmen dieser Dissertation veröffentlichten bzw. zur Veröffentlichung

eingereichten Manuskripte aufgeführt.

2.1 Manuskript 1:

A high resolution camera system (ParCa) for imaging particles in the occan: System design and

results from profiles and a threc-month deploymcnt

by V. Ratmeyer and G. Wefer

(veröffentlicht bei Journal 0/Marine Research, 54, 589-603, 1996)

Inhalt Seite

Abstract 27

1. Introduction 27

2. Methods 29

3. ReslI1ts and Discussion 33

a. Particle abundance profiles 33

b. PaJiicIe abundances and concentrations

during long-term deployment 35

References 40

V Ratmeyer and G. Wefer: A high resolution camera ,\ystem...

loumal 0/ Marine Research, 54, 589-603, 1996

A high resolution camera system (ParCa) for imagingparticles in the ocean: System design and results

from profiles and a three-month deployment

by Volker Ratmeyer1 and Gerold Wefer1

ABSTRACTFor direct optical measurement of abundance, concentration and size distribution of marine

particIes, a high-resolution camera system (ParCa) was designed to improve on similar systemsused by Honjo ef al. (1984), Asper (1987) and others. Imaging a probe volume of up to 37 1,smallest particles with diameters of 50 !-Lm can be counted. The images provide information onparticle size, shape and abundance either during profiling through the water column or whilemoored in a certain depth over time. Depth profiles were acquired between fall 1992 and latespring 1993 on R. V. Meteor cruises M22-1 and M23-3 at 6 stations in the equatorial AtlanticOcean and off the west African shelf. The images show variable particle and aggregateconcentrations through 550 m of the water column, with highest concentrations in the upper80 m. A distinctive change in the depth of the upper chlorophyll maximum from about 75 m inthe Brazil Basin to about 50 m in the Guinea Basin was measured with the attached INFLUXcurrent meter (Krause and Ohm. 1996) and is as weIl represented in the particle abundancesof two selected profiles. In contrast, both profiles show a second particle abundance maximumbetween 100 and 250 m, which is not visible in the chlorophyll-a and backscatter signal of theINFLUX sensors. Total particle abundance maxima raise from 677 counts per liter in thecentral Brazil Basin to 991 counts in the Guinea Basin, corresponding to marine snowabundances of 57 and 127 counts per liter, respectively.

In order to compare high-resolution data on particle concentration and f1ux through time,ParCa was also deployed on a sediment-trap mooring at 995 m depth in the Canary Basinbetween June and September 1994. First results show similar trends in sediment-trap derivedf1uxes of particulate matter from 2.8 to 67.2 mg m- 2 d-I and equivalent spherical volumes ofpartieIes with diameters > 0.5 mm from 0.98 to 4.13 mm3 1- I.

1. Introduction

Most marine particles and aggregates are produced in the upper few hundredmeters of the ocean, and they are known to be the most important carbon andelement carriers from the ocean's surface to the deep sea. Recent investigations ofparticle ftuxes throughout the world's oceans (Honjo et al., 1982; Deuser, 1986;Wefer, 1989; Fischer et al., 1996) show the necessity to obtain detailed knowledgeabout the various origins, transportation, alteration processes and pathways of

1. Fachbereich Geowissenschaften. Universität Bremen, Klagenfurter Str., 28359 Bremen, Germany.

589

27

590 Journal 01 Marine Research [54,3

particles through the water column. Most of the common hydrocasting and f1uxrecording methods allow the record of composition, seasonality patterns and f1uxthrough time, but significantly less data are available concerning in-silu size, settling

behavior and high-frequency changes in the abundance of particles and aggregatesthrough the water column (Lampitt et al., 1993; Syvitski et al., 1995; MacIntire et al.,

1995).Although it has been proposed that particle f1ux and sedimentation in the deep

ocean is mainly driven by large, rapidly-sinking aggregates in the size class of marinesnow (defined as aggregated particles of a diameter greater than 0.5 mm) (McCave,1975), evidence in the form of in-situ observation of abundance and sinking speed ofparticles in all size cIasses is based only on a small number of observations. This ismainly due to the extremely delicate nature of marine aggregates whieh are generallynot preserved in trapped material. Except for the fast sinking and stable fecal pellets,most forms of marine aggregates are very fragile and thus require the use of

nondestruetive methods for observation and very advanced sampling teehniques.Successful approaches incIude the use of submersibles (Kajihara, 1971; Silver andAlldredge, 1981), SCUBA techniques (Shanks and Trent, 1980), various photographie systems and video observation teehniques (Honjo et al., 1984; Lampitt, 1985;

Asper, 1987; Hecker, 1990; Heffler et al., 1991; Asper et al., 1992; Gorsky et al., 1992;Lampitt et al., 1993; MacIntire et al., 1995; Syvitski et al., 1995), holographie imagingtechniques (Costello et al., 1989) and the use of transmissometers and ftuorometers(Gardner and Riehardson, 1992; Krause and Ohm, 1995), or eombinations of those.

Although each technique has its clear advantages, only a few ean be used to observe abroad spectrum of particle sizes, their sinking speed and eomposition in situ

espeeially over long time periods and at great water depths.

The particle camera system (ParCa) presented in this paper improves on thephotographic method developed by Honjo et al. (1984) and provides in-silu information on the abundanee, eoneentration and size speetra of marine particles andaggregates while using the highest resolving optics and film format available forunderwater observations. ParCa was used between fall 1992 and late spring 1993 atseveral stations in the equatorial Atlantie Oeean and off the west Afriean shelf forparticle size and abundance measurements through the upper 550 m of the watercolumn. During these eruises the system was used to a water depth of 600 m.Between June and September 1994, ParCa was deployed in a sediment-trap mooringat 995 m depth in the Canary Basin.

In this paper, we describe the technieal setup of the ParCa system and show thefirst results on particle size, abundance and estimated spherical volumes obtainedduring the first deployments in different areas of the AtJantic Ocean. A detaileddiscussion of these data in comparison to partiele sinking speeds and f1uxes measuredwith sediment traps will be presented and discussed eJsewhere.

V Ratmeyer and G. Wefer: A high resolution camera syste/~/l~.,,~~._~~ 29_

1996J Ratmeyer & Wefer: ParCa system design & results 591

2. Methods

Improving on similar photographic devices uscd by Honjo et al. (1984), Lampitt(1985) and Asper (1987) the system presented here combines the advantage of highresolution photographic observation techniqucs with the necds of illuminating largeprobe volumes to image a broad spectrum of particle sizes, ranging from 0.05 mm to0.4 m. It provides a sampling capacity of up to 650 images, which allows a highsampling frequency during moored deployments and gives a high spatial resolutionof the particle distribution when lowered through the water column. To augment theuse of sediment traps, particle concentration can be measured even in areas ordepths with high lateral transport.

To obtain oceanographic data and additional optical measurements during profiling, an INFLUX current meter with CTD-sensors, fluorometer and backscatteringsensor (Krause and Ohm, 1996) was mounted to the system. Both optical sensors hadthe same source of excitation (flashlight, broad band up to 520 nm), the detectionrange was 680 ± 10 nm far fluorescence and broad band for backscattering.Sampling volume was 1 ml for the fluorescence and 20 I far the backscattering sensor.Additional instrument specifications are given in Krause and Ohm (1996).

ParCa can be operated remote-controlled when lowered through the watercolumn or autonomous when attached to a mooring during long-time deployment.Maximum operating depth is 6000 m.

a. System setup. To get the highest possible optical resolution in the images, ParCaconsists of a submersible photographie camera system (Fig. 1a) instead of a videosetup. Both camera (PHOTOSEA 70) and strobe lights (PHOTOSEA 1500SD) arecommercially available, though camera and strobe electronics had to be modified forlong-time deployment. The main advantage of the camera is the 6 x 6 cm negativeformat which provides about four times higher resolution than similar systems basedon 35 mm film or highest-resolution video. The camera can hold up to 650 framesusing thin-based film.

The best optical setup of camera and light source for in-silu imaging of highlyhydrated partieles and aggregates was described by Honjo et al. (1984). To caIculateparticle abundance and concentrations, it is necessary to know the volume of waterrepresented in the images. Thus, the light source at the focal distance of 60 cm isfixed orthogonal to the optical axis of the camera (Fig. 1b). It consists of two sets ofhighly refractive acrylic fresnel lenses-the collimator-in front of the two strobelights. This setup produces a collimated light beam covering the area of 60 x F cm,where F is the thickness of the light beam, given by the collimator's aperture and thedepth of field of the viewing optics. Typical aperture setting was f-stop 11, resulting inF = 12 cm depth offield and a 12 cm wide light slab, which was verified using a metriccalibration setup and a test tank. The viewed probe volume due to the lens viewingangle of 50° was then 44 x 44 x 12 cm, or 24 1. Maximum depth of field was F =

lounIa! 01 Marine Research

30--------------------

592 [54,3

A

8

System setup (moored version):

1: Camera PHOTOSEA 702,3: Strobes PS 1500S04: Collimator5: Nylon 0.1 mm6: Controller & battery7: Buoancy package

f

view trom top

200cm7

view trom side

Figure 1. Schematic view (A) and optical setup (B) of the ParCa system used in this study.

19 cm using f-stop 16, providing 37 I maximum probe volume. Maximum in-süuresolution is 0.05 mm, using KODAK Technical Pan film, and around 0.08 mm usingKODAK Tri X Pan Film. The Tri X Pan allows forced film processing at 800 to 1600ASA and can therefore be used with higher f-stop settings due to greater depth offield and larger probe volumes.

V. Ratmeyer and G. Wefer: A high resolution camera system... 31

1996] Ratmeyer & Wefer: ParCa system design & results 593

The power-supply, camera and f1ashlights are controlled by a PIC-processor basedmicrocomputer which is installed in aseparate pressure housing. Both software and

the electronic interface between controller, camera and external PC were developed

at the Marine Geology Group at the University of Bremen. During vertical profiling,

exposure is triggered either by the controller-unit or via cable by a PC on deck of theresearch vesseI. For long-time and moored deployment, a set of contral options can

be programmed: start-time of exposure program, time interval between single

frames, time interval between multi-exposures on a single frame, multi-exposure/

single exposure rate, total number of frames to be exposed and number of test runs

when power is switched on. Because a major limitation of the system is given by thecamera's film-capacity, any combination of single and multiple exposures may be set.

Using the multi-exposure option with interval times greater than the assumed

slowest sinking time of particles through the camera's field of view, the final image

gives the particle distribution inside a water volume as large as n-times the illumi

nated volume, where n is the number of exposures in one single frame. This approachis only used during moored deployment.

Frame number and number of exposures per frame are recorded on each image by

a second PIC-processor and LED's inside the camera. Power is supplied by a

pressure resistant, non liquid, rechargeable 24V/38Ah lead acid battery (manufac

tured by DeepSea Power & Light). AII components of the ParCa system are instaIIedinside a hot-galvanized steel frame, whose dimensions are 2 m x 1.2 m x 0.8 m

(Fig. 1a). Total weight in air is 310 kg, weight in water is approximately 250 kg. While

the system is attached to a mooring, two buoyancy packages each holding 4 glass

spheres are mounted to the frame.

b. Image digitizing and analysis. Quantitative analysis of particle abundance and sizeis performed by digital image analysis. The image analysis setup consists of a

modified photographic enlarger, a Sony single-chip CCD color camera, a Sony

Trinitron monitor for image control and an IBM compatible Pentium personalcomputer with 17" color display. Images are captured using an 8 bit grayscale video

framegrabber (MuTech Image VGA plus) inside the computer, aIIowing aresolutionof 768 x 512 pixels. A menu-driven software package (BioScan Optimas) is used far

capturing, processing and analysis of the images. Its powerful C-library of imageanalysis related algorithms and functions was used to create complex macros forimage enhancement and process automatization. A ftowchart showing the procedureof data acquisition, image-analysis and data processing is given in F..-igure 2. Imagesshowing zooplankton were excluded from measurements.

c. Image captwing. A variable tube extends the optical pathlength between negativeand lens of the enlarger and reduces negative size to the Y2 inch diameter of thccamera's CCD chip. This allows the entire image to be digitized but provides only

594 Journal ofMarine Research [54,3

moored deployment

exposure program:altemating single- and

n multi-exposuresper time interval

profiling deploymentexposure program:single exposuresper depth interval

. . " ., ;

:.'.

". ... . \ "

'..,.'" '\.; ,~.

, ,

. ....... .'------.-"-----'·01.-------

measurement of:

partieIe abundance,part/eie size and shapein n light slab volumes

computation of:partieIe volumes,partieIe concentration,partieIe flux through time

comparison with:

INFLUX data ofchl-a,optical backscatter,water temperature,salinity,

sediment trap data ofpartieIe ftux,partieIe composition,

image digitizingentire image:

partieIes > 1.0 mm3 subsampies of 1.5 I:

part/eies 0.05 - 1.0 mmstorage on harddisk

image enhancementglobal background correction,background subtraction,gray-slope filtering,removing single foreground pixels,thresholding

image analysis

further dataprocessing:

compar/son ofboth approaches,

seasonal and highfrequency variationof dITferent partieIepopulations andsize classes,

lateral and verticaldistribution pattems

measurement of:

partieIe abundance,partieIe size,aggregate shape

computation of:depth profiles ofpartieIe volumes,partieIe concentration,

comparison with:

INFLUX data ofchl-a,optical backscatter,water temperature,salinity

Figure 2. Flowchart showing the procedure of image acqulSltlOn, image-analysis and dataprocessing used at the Marine Geology Group at the University of Bremen.

V Ratll7eyer and G. Wefer: A high resolution camera .system... 33

1996] Ratmeyer & Wefer: ParCa system design & reslllts 595

low optieal resolution. Because the optical resolution of the digitizing system islimited by hardware to 768 x 512 pixels, different magnifieations have to be set formeasuring different size cIasses of partieles in the original image. A transparentoverlay is used to define subsampies in the photographie frame. This method allowsthe deteetion of rare partieles in the size cIass of marine snow as weIl as the use of thehigh resolution to record smallest particIes with diameters of 0.5 mm.

d. Image processing. Even though illumination in the ParCa system is collimated,slight gray-shading of the background and sometimes illumination of partieles out offoeus is not avoidable. Thus, each image is enhanced digitally to provide the bestaecuraey far measuring particle size and concentration. This is done by automatedbackground correction, background subtraction, removing single foreground pixelsand thresholding fareground pixels. Particles out of focus can cIearly be recognizedand are removed using a gray-slope filter, which identifies unsharp particIes from thegradient of background to thresholded foreground gray values. For all imagesexposed on the same profile, edge-values for background correction and thresholding are not changed due to equal illumination during deployment time. Every step ofimage enhancement can be controlled by overlaying the thresholded fareground, theenhanced image and the originally grabbed frame.

e. Image analysis. ParticIes are identified automatically after image enhancement isperfarmed. Data of particIe area, major and minor axis length are directly exportedinto a Microsoft Excel worksheet. All particIes are assumed to lie along themid-plane of the illuminated probe volume. Equivalent spherical diameter andvolume are caIculated for each individual particle, assuming spherical shape. Sevensize classes « 0.1 mm (A), 0.1-0.25 mm (B), 0.25-0.5 mm (C), 0.5-0.75 mm (D),0.75-1.0 mm (E), 1.0-2.0 mm (F) and > 2.0 mm (G)) are defined to characterize thesize distribution. Particles > 1.0 mm diameter are recorded using the entire photographie frame, whereas the smaller size classes are measured using subsampies of theimage representing an in-süu volume of 1.5 liters.

f Sediment-trap sampies. During moored deployment, sampies were collected withcone-shaped sediment traps with 0.5 m2 openings (Aquatec, Kiel, FRG) at 923 and3070 m water depth. SampIe cups were poisoned with mercuric chloride prior to andafter recovery. The collected material was treated as described in Fischer and Wefer(1991). Only material from the upper trap was used far comparison with data fromthe ParCa system at 995 m and the INFLUX system at 1015 m water depth.

3. ResuJts and discussion

a. Particle ablllzdance profiles. Six vertical profiles through the upper 550 m of thewater column were completed during R.V. Meteor M22-1 and M23-3 in fall 1992 and

34

• "marine snow" ( :> 0.5 mm )o partieles :> 0.05 mm

I- 3-point runnlng average

backscalter (%)

0.090 0.100

back.scaller (OA,)

particle abundance ( n j1 ) chJ..a ( mg m-3) water tempo (0C)o 300 600 0 20 40 so 0.0 1.2

partide abundance (n i1) chl-a (m;;J m-3) water tempo ce)300 600 900 0 50 100 150 0.0 1.2 20

Figure 3. VerticaI particle abundance profiles (A, B) and INFLUX derived data of opticaIbackscatter, chlorophylI-a and water temperature (a, b) observed at two stations (site GeoB2212: 04°00.4/5; 25°37.1'W; site GeoB 2217: QOoOO.5'N, 100 49.8'W) in the equatorialAtlantic Ocean during R. V. Meteor cruise M 23-3 in March 1993.

250

bBaA

spring 1993 on oceanographic sampling sites and mooring stations in the equatorialAtiantic Ocean and off the West African shelf. Here we describe two of thesepartide abundance profiles obtained with the ParCa system on cruise M23-3 atsampling sites GeoB 2212 (04°00.4'S; 2S037.1/W) and GeoB 2217 (00000.5'N;10049.8'W) aboard R.V. Meteor in the equatorial Atlantic Ocean, both completed inMarch 1993 (Fig. 3). Site GeoB 2212 was located in the central Brazil Basin, whereassite GeoB 2217 was situated within the equatorial Guinea Basin. The profiles showvariable distributions of particles and aggregates between the ocean's surface and550 m water depth, with major concentrations occurring in the upper 80 meters.Though a significant part of the photographed particles belongs to the size dass ofmarine snow, typical flocs and fractal shaped aggregates reported elsewhere (Syvitskyet al., 1995; Lampitt, 1985; Logan and Wilkinson, 1990) were not found during eitherprofile. Both images and INFLUX-derived data of fluorescence (chlorophylI-a) andoptical baekscatter show a shift of the chlorophylI-a maximum eorresponding to theupward shift of the thermodine from a water depth of 75 m in the west (site GeoB2212, Fig. 3a) to 50 m in the east (site GeoB 2217, Fig. 3b). Total number ofparticlesreached from 125 (512 m) to 677 (10 m) counts per liter in the oligotrophie BrazilBasin (Fig. 3A), and from 281 (461 m) to 991 (36 m) in the equatorial upweIIing areaof the Guinea Basin (Fig. 3B). Particles and aggregates in the size class of marinesnow (> 0.5 mm) show abundances of 8 to 57 and 34 to 127 counts per liter, withdistinetive peaks observed at the thermocline-base, corresponding to the ehloro-

V Ratll7eyer and G. Wefer: A high resolution camera system .. 35

1996] Ratmeyer & Weler: ParCa system design & reslilts 597

phyll-a maximum observed with the INFLUX-current meter. In contrast, bothprofiles show distinctive particle abundance maxima below 100 m depth, which arenot represented in the INFLUX measurements of backscatter, chlorophyll-a andwater temperature. While two peaks in the central Brazil Basin at 220 and 305 mseem to consist of the entire measured particle size spectrum, the broad maximumaround 150 m depth in the Guinea Basin is not reflected by the marine snowabundance. These peaks are not seen in backscattering at either site.

b. Partiele abundances and concentrations dllring long-tenn deployment. Für instrument testing and data acquisition, ParCa was deployed at 995 m water depth inSummer 1994 on the sediment-trap mooring CI4 located in the Canary Basin(29°09.1 'N; 15°26.5'W). After three month of deployment between June and September, the system was recovered. During this time, at alternating 7.5 and 15 hourintervals one double-exposed frame was exposed.

Figure 4 shows the results of the ParCa deployment (Fig. 4A, B) in comparison tothe optical INFLUX data (Fig. 4C) from 1015 m depth. Figure 5 shows the total fluxrecord of the simultaneously deployed upper sediment trap (Fig. 5A) at 923 m, andover sampling-time of 4.3 days integrated particle volumes of the ParCa-derived datafrom 995 m water depth (Fig. SB, C). Current speeds at 1015 m depth in the areabelow camera and trap show both a high (diel) and low frequent variation but neverexceeded 10 cm S-1 (Fig. 5D). Thus, we assurne that hydrodynamic biases at thetrapping site were minimal (Baker et al., 1988; Gust et al., 1992).

Two major events occur in the total flux record derived from the sediment trap at925 m depth (Fig. 5A). The first broad maximum in the second half of June containsthe highest flux of particulate matter with 67.2 mg m- 2 d- 1• A second, much smallerevent was recorded with a total flux of 26.9 mg m- 2 d- 1 between July 17th and 21st insampie 10. SampIe 15 contains fish-parts and swimmers and was therefore excludedfrom further caIculations. The lowest total particle flux during deployment time wasrecorded in sampIe 19 with 2.8 mg m- 2 d- 1• In comparison, the ParCa-deriveddistribution of partieles > 0.05 mm shows a more differentiated abundance patternwith high temporal variations of 47 to 164 particles per liter du ring the first 50 days ofdeployment. This is mainly driven by partieles < 0.5 mm with abundances between38 and 151 counts 1-1 (Fig. 4A). In contrast, particles> 0.5 mm account for most ofthe caIculated volume distribution, as shown in (Fig. 4B). Highest particle volumes ofup to 6.48 mm3 1-1 oeeur within the first flux event during the seeond half of June. Inmid-July, the seeond flux maximum is represented by values of up to 5.6 mm3 I-Ibetween the 13th and 17th of July. Figure 5B and SC show volume data of marinesnow sized particIes, integrated over 4.3 days, while Figure SC gives the results of thedata eorreeted bya three-day offset in the start time of the eamera. Due to a readoutfailure it is likely that a eertain start time offset took plaee, but unfortunately there isno way to quantify this preeisely. A delay greater than three days is not possible

36

598 Journal 01Manne Research fS4.3

Canary BasinMooring CI4

Particle Camera (ParCa), 995 m

o 0

o 0

...... ,.' ... '.~~'-!.• .:..'~.-.!. ..!. ••• ~ •••••••••• s'.

85756555

• "marine snrm' (> 0.5 rrrn )

o pa1ides 0.05 - 0.5 rrrn

453515

B .~::=. ,', ,', •...........•... ',' ..' ." ... ,,' -·lI'· ", ',: I •

0' . 0 .,' ) ~

oday

INFLUX - Currentmeter, 1015 m

uppar: optical backscallarlowar: chlorophyll - a 0.300,03 C

:;;-rg~~.,i>l~g-~~

.0

0.02

day 5 1509,06.94

25 35 45 55 65 75

0.20

85

02,09,94

Figure 4. Results of the ParCa system (A, B) and the optical sensors of the simultaneouslydeployed INFLUX currentmeter (C) at mooring-station CI 4 (29°09.1'N; 15°C26.5'W) inthe Canary Basin, between lune and September 1994. Gray areas refer to sediment trapderived peaks in total particle flux as shown in Figure 5.

because the program of 420 exposures was correctIy finished when the system wasrecovered on 4 September, 1994. Both figures show the sirnilar trends compared tothe flux record of the sediment trap; but, in detail, the data corrected by thethree-day offset provide the better fit (Fig 6B). This result supports previous findingsthat relatively rare aggregates and large particles contribute to most of the particleflux recorded with sediment traps (e.g. Deuser er al., 1981; Asper, 1987; Gardner andWalsh, 1990; Walsh and Gardner, 1992).


1996] Ratmeyer & Wefer: ParCa system design & results 599

Canary BasinMooring CI4

-- --~--f ---r---:I I I I, ,

I Sll/lllIing-lnterval: 4.3 days I

I integratioo-interval: 4.3 days I

Upper Sediment - Trap, 923 m

Partide Carrera (ParCa), 995 m

150 A

~ 100)( '7..,::J

""~~ ~- E 50

10 0.., ~1'7!ll

5- CI)

~ 5::J<J

0day 5 15 25 35 45 55 65 75

6 B

-8 ~ '7_3'E .g "'E

[> E

0

6

-8 ~ '7_3'E.a "'E

[g E

0day

Figure 5. Sediment trap derived total particle fluxes (A), integrated particle volumes calculated from ParCa images (B, C) and currentspeeds recorded with the INFLUX currentmeter (D) at mooring-station CI 4 (29°09.l'N; 15°26.5'W) in the Canary Basin, betweenJune and September 1994.

38

600 faumal 0/Manne Research [54,3

5.00 -"',.-'

A4.00

f/)

~::J

3.00g '7-

~ ",-.. .. ..

17j ~ ..2.00 ....

Q)C .. Y:: 0.028 x+ 1.423'C

E .. ..'" ..

1.00 ..r 2:: 0.327n:: 18

0.00

0.00 20.00 40.00 60.00 80.00

5.00

~ B0 4.00f/) ..>.ro

"0 ..C0+ ---. 3.0000" 'r"l- ..~ "'E0 E 2.00 .. y:: 0.046 x + 1.045>~

...... ..Cf/) .. ..<D 1.00

..C .. r 2:: 0.731'e ..E n:: 18

0.00

0.00 20.00 40.00 60.00 80.00

total flux in sediment trap

(mg m-2 cr1 )

Figure 6. Correlations of ParCa-derived marine snow volume concentration and total particlef1ux recorded with the upper sediment trap at station CI 4 in the Canary Basin. (A) shows"raw" particle volumes as plotted in Fig. SB versus total particle f1ux, (B) gives the result ofadding a 3-day offset to particle volume data (see text) as plotted in Figure Sc.

In contrast to the two depth profiles described before, optical backscatter andchlorophylI-a concentration do not follow any of the trends of the total particleabundance (Fig. 4C), and both show significantly lower values. Very low backscattervalues around 0.05% have previously been reported at this site between July andSeptember in almost the same water depth (900 m) by Fischer et al. (1996) betweensedimentation events. This may be partly due to the much lower abundance ofparticles during thedeployment at 1000 m depth compared to the other dataobtained in the upper 600 m of the water column. Furthermore, the deployment time

V. Rat1l1eyer and G. Wc;fi:r: A high resolution call7era ,\ysfell7... 39

1996J Ratmeyer & Wefer: ParCa system design & results 601

is known to be a low production period. During spring, much higher total ftuxes of upto 128 mg 01- 2 d- 1 are reported from previous studies at the same site, whereshort-term sedimentation pulses could be recorded with sediment traps as weIl aswith the chlorophyll-a sensor of the INFLUX currentmeter (Fischer et al., 1996).

However, during the ParCa deployment single peaks in the optical backscatter occuronly during or in between the two described particle flux events, whereas chlorophylldoes not show any significant variance.

In addition to previous work with sediment trap moorings the study described inthis paper offers for the first time a direct comparison of sediment trap derived ftuxmeasurements with optical data of chlorophyll-a, Mie-backscatter and in-situ accessed images of marine particles. This approach provides mare information onin-situ particle dynamics and composition of sinking material and may lead to a moredetaiIed understanding of mass ftuxes in open ocean conditions.

The resuIts described here indicate that none of the observation methods usedduring the profiles or moored deployment shows the entire particle community:while the camera system does not resolve the distribution of smallest particlescontributing to the backscatter signal of the currentmeter, the INFLUX sensors danot reftect changes in the distribution of partieles greater than 0.05 mm. On the otherhand, the simultaneous use of both systems in addition to sediment traps providesinformation on the different behavior of different size cIasses of marine particles,indicating that the smaller size classes of marine particles show less variance throughdepth and time than large aggregates and marine snow (Gardner and Walsh, 1990;Lampitt et al., 1993; Syvitski et al., 1995). This is of particular interest far abundanceand mass esimates of slowly versus rapidly sinking particulate matter (Alldredge andSilver, 1988; Gardner and Walsh, 1990; Syvitski er al., 1995). The comparable trendsin particle abundance, marine snow volume concentrations and total particle fluxesreported here allow a use of this dataset for further computation of in süu measuredparticle distributions versus particle fluxes and sinking speeds, which will be presented in a following paper.

Acknowledgments. We thank the masters and the crews of the research vesseIs FS Polarsternand FS Meteor for the their professional help during deployment and recovery of the ParCasystem and the mooring array. Special thanks are due to G. Fischer far many hours of creativediscussion, carefuI reading of the manuscript and making helpful suggestions, and to G. Ruhland for innumerable comments and help during development of the ParCa system. We arealso grateful to G. Krause far providing the INFLUX systems used during profiles anddeployment, and to P. Gerchow, R. Heygen, U. Rosiak, R. Plugge and H. Pabst for technicalassistance at sea. Laboratory work performed by V. Diekamp and C. Slickers is gratefullyacknowledged. We would also like to thank the two anonymous referees for their commentswhich improved the paper. This research was funded by the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 261 at Bremen University, Contribution 113).

40

602 louma! 01 Marine Research [54,3

REFERENCESAlldredge, A. L. and M. W. Silver. 1988. Charaeteristies, dynamies and signifieanee of marine

snow. Prag. OeeanogI. 20, 41-82.Asper, V. L. 1987. Measuring the flux and sinking speed of marine snow aggregates. Deep-Sea

Res., 34, 1-17.Asper, V. L., S. Honjo and T. H. Orsi. 1992. Distribution and transport of marine snow

aggregates in the Panama Basin. Deep-Sea Res., 39, 939-952.Baker, E. T., H. B. Milburn and D. A. Tennant. 1988. Field assumption of sediment trap

efficiency under varying flow eonditions. J. MaI. Res., 46, 573-592.Costello, D. K., K. L. Carder, P. R. Betzer and R. W. Young. 1989. In situ holographie imaging

of settling particles: Applieations for individual particle dynamics and oeeanic flux measurements. Deep-Sea Res., 36, 1595-1605.

Deuser, W. G. 1986. Seasonal and interannual variations in the deep-water particle fluxes inthe Sargasso Sea and their relation to surface water hydrography. Deep-Sea Res., 33,225-246.

Deuser, W. G., E. H. Ross and R. F. Anderson. 1981. Seasonality in the supply of sediment tothe deep Sargosso Sea and implications for the rapid transport of matter to the deep ocean.Deep-Sea Res., 28,495-505.

Fischer, G., G. Krause, S. Neuer and G. Wefer. 1996. Short-term sedimentation pulsesrecorded with a chlorophyll sensor and sediment traps in 900 m water depth in the CanaryBasin. Limnol. Oceanogr., (submitted).

Fischer, G. and G. Wefer. 1991. Sampling, preparation and analysis of marine particulatematter, in The Analysis and Characterization of Marine Particles, D. C. Hurd and W. Spencereds., AGU, Washington, 391-397.

Gardner, W. D. and M. J. Richardson. 1992. Particle export and resuspension fluxes in thewestern North Atlantic, in Deep-Sea Food Chains and the Global Carbon Cycle, G. T.Rowe and V. Pariente, ed., Kluwer Academic Publishers, 339-364

Gardner, W. D. and 1. D. Walsh. 1990. Distribution of macroaggregates and fine-grainedpartieles across a continental margin and their potential role in fluxes. Deep-Sea Res., 37,401-411.

Gorsky, G., C. Aldorf, M. Kage, M. Picheral, Y. Garcia and J. Favole. 1992. Verticaldistribution of suspended aggregates determined by a new underwater video profiler. Ann.Inst. Oceanogr., 68, 275-280.

Gust, G. R, R H. Byrne, R. E. Bernstein, P. R Betzer and W. Bowles. 1992. Particle fluxesand moving fluids: experienee from synchronous trap collections in the Sargasso Sea.Deep-Sea Res., 39, 1071-1083.

Hecker, B. 1990. Photographie evidence for the rapid flux of partieles to the sea floor and theirtransport down the continental slope. Deep-Sea Res., 37, 1773-1782.

Heffler, D .E., J. P. M. Syvitski and K. W. A<;prey. 1991. The floc camera, in Principles,Methods and Application of Particle Size Analysis, J. P. M. Syvitski, ed., CambridgeUniversity Press, New York, 209-221.

Honjo, S., K. W. Doherty, Y. C. Agrawal and V. L. Asper. 1984. Direct optical assessment ofJarge amorphous aggregates (marine snow) in the deep ocean. Deep-Sea Res., 31, 67-76.

Honjo, S., S. Manganini and J. J. Cole. 1982. Sedimentation of biogenie matter in the deepocean. Deep-Sea Res., 29, 609-625.

Kajihara M. 1971. Settling velocity and porosity of 1arge suspended particles. J. Oceanogr. Soc.Japan, 27, 158-162.

Krause, G. and K. Ohm. 1996. The INFLUX current meter and its liSC togcther with asediment trap. Deep-Sea Res., (submitted).


1996) Ratmeyer & Wefer: ParCa system design & results 603

Lampitt, R. S. 1985. Evidence for the seasaoni deposition of detritus to the deep-sea floor andits subsequent resuspension. Deep-Sea Res., 32, 885-897.

Lampitt, R. S., W. R. Hillier and P. G. Challenor. 1993. SeasonaI and dieI variation in the openocean concentration of marine snow aggregates. Nature, 362] 737-739.

Logan, B. E. and D. B. Wilkinson. 1990. Fractal geometry of marine snow and other biologiealaggregates. Limnol. Oceanogr., 35] 130-136.

MacIntire, S., A. L. Alldredge and C. C. Gotschalk. 1995. Accumulation of marine snow atdensity discontinuities in the water column. Limnol. Oceanogr., 40] 449-468.

McCave, 1. N. 1975. Vertical flux of partieies in the ocean. Deep-Sea Res., 22,491-502.Shanks, A. L. and J. D. Trent. 1980. Marine now: Sinking rates and potential role in vertical

flux. Deep-Sea Res., 27A, 137-143.Silver, M. W. and A. L. Alldredge. 1981. Bathypelagic marine snow: deep-sea aIgaI and

detrital community. J. Mar. Res., 39] 501-530.Syvitski, J. P., K. W. Asprey and K. W. G. Leblanc. 1995. In-situ characteristics of particies

settling within a deep-water estuary. Deep-Sea Res. II, 42] 223-256.Walsh, 1. D. and W. D. Gardner. 1992. A comparison of aggregate profiles with sediment trap

fluxes. Deep-Sea Res., 39] 1817-1834.Wefer, G. 1989. Particie flux in the ocean: Effects of episodic production, in Productivity of the

Oceans: Present and Past. W. H. Berger, V. S. Smetacek and G. Wefer, eds., Wiley andSons, 139-153

Receivcd : 15 August, 1995; revised; 14 November, 1995.

Ratlllcycr ct al.: Lithogenic particle fluxes and grain sizc distributions...

2.2 Manuskript 2:

Lithogenic particle fluxes and grain size distributions in the deep ocean offNW Africa:

implications for seasonal changes of aeolian dust transport and mass fluxes

byV. Ratmeyer, G. Fischer and G. Wefer

43

Abstract 45

Introduction 46

Regional Settings 47

1. AtInospheric regime and dust transpoli 47

2. Ocean Circulation 49

Methods 50

Results 56

Discussion 68

1. Currents and trapping efficiency 68

2. TranspOli processes and pathways 68

3. Comparison of sinking pmiiculate matter

and surface sediments 74

4. AtInospheric input oflithogenic particles 75

Conc1usions 82

SUlnlnary 83

References 85

(im Druck bei Deep-Sea Research)

Inhalt Seite

Ratllleyer ct a!.: Lithogenic particle fluxes and grain size distributioJls...

Lithogenic particle fluxes and grain size distributions in the deep ocean off NW Africa:

Implications 1'01' seasonal changes 01' aeolian dust input and downward transport

Volker Ratmeyer, Gerhard Fischer and Gerold Wefer

ABSTRACT

45

Between 1988 and 1994, twenty time-series sediment traps were deployed at different water depths in

the Canary Island region, off Cape Blanc (Mauretania), and off Cape Verde (Senegal). Lithogenic

paIticle fluxes and grain size distributions ofthe carbonate-free fraction ofthe trapped material show a

high impact of dust transported either in the NE trade winds 01' the Saharan Air Layer (SAL). Highest

annual mean lithogenic fluxes (31.2 - 56.1 mg m-2 d-I) were observed at the Cape Blanc site, and

1argest annual mean diameters (> 6 11m) were found off Cape Verde (14.5 - 16.9 11m) and off Cape

B1anc (15.2 - 16.7 11m). Lowest annual lithogenic fluxes (11.4 - 21.2 mg m-z d-I) as weH as smaIIest

mean diameters (13.5 - 13.7 11m) occurred in the Canary Island region. A significant correlation of

organic carbon and lithogenic fluxes was observed for all sites. Off Cape B1anc, fluxes and mean

diameters correlated welI between upper (around 1000 m depth) and lower traps (around 3500 m

depth) were found only, indicating a fast and mostly undisturbed downward transport of particulate

matter. In contrast, major correlation of fluxes without corresponding mean diameters occurred in the

Canary Island region, which translates into a fast vertical transpolt and scavenging of lateraHy

advected material with depth at this site. The seasonality of lithogenic fluxes was highest in the

Canary Island region and off Cape Verde, reflecting strang seasonal patterns of atmospheric

circulation with highest occurrence of continental winds in the trade wind layer during winter. In

addition, grain size statistics reflect a dominant change of dust transport in the trade winds during

winter/spring and transport in the SAL during summer 1993 at the Cape Verde site. Highest lithogenic

fluxes during winter were correlated to mean diameters around 10 - 13 11m, whereas lower fluxes

during summer consisted of coarse grains around 20 11m. Annual mean dust input was caIculated from

lithogenic fluxes in the range of 0.7 to 1.4 x 106 y-1, roughly confirming both sediment accumulation

rates and atmospheric model caIculations repOlied previously from these sites.

46

INTRODUCTION

Litl10genic particles In many regions 01' the world's ocean are one 01' the major components 01'

particle flux to the deep-sea. Lithogenic partieles enter the marine environment either through aquatic

01' aeolian transport, where the aquatic processes contribute by rar the lat'gest amounts 01' mineral

detritus to the oceans: the annual global river discharge 01' sediments has been calculated as 20 billion

tons (bt) per year (Milliman and Syvitski, 1992), followed by the flux 01' ice-rafted debris into the

polar seas (1.3 bt i l; Honjo, 1993). Global coastal erosion by ,vave and current action contributes

about 0.5 bt i l to the ocean's mineral particle interior. Estimations 01' the global aeolian dust input to

the oceans range from 2.0 bt i 1 (Schütz ancl Sebert, 1987) to 1.4 bt i l (Duee et al., 1991). Two major

global sources eontribute approximately 90 % 01' the global atmospherie dust budged: Saharan dust

outbreaks over the Atlantie Oeean (0.6 to 1.0 bt i l; Schütz and Sebert, 1987) and the Mediterranean

Sea, and Asian dust from northern China towards the North Paeifie Oeean (0.2 - 0.35 bt i l; Prospero,

]990).

Although the totalmass 01' wind transported lithogenie partieles towards the global oeean is about

one order less than the material transported by hydralIlie processes, its eontribution to the sediment

flux in the deep-sea is signifieantly higher sinee river transported material is effieiently trapped on the

continental shelves and slopes (Sarnthein et al., ]982). Reeent studies have shown that the

atmospherie lang range transport 01' mineral dust has a signifieant influenee on the produetivity 01'

both continental and marine ecosystems (e.g. Duce, ]986; Martin and Fitzwater, ]988). For instanee,

the Amazon rainforest at least partly depends on minerals and trace elements transported from NW

African deserts over the Atlantic during massive aeolian dust outbreaks (Swap et al., 1992).

Lithogenic components obtained from deep-sea sediment cores may be used as important tracers for

elimatic and paleoclimatic variations on time sca]es from 103 to 106 years (e. g. Koopmann, 1981;

Sarnthein et al., 1982; Gingeie, 1991; Duce, 1996) and help to reconstruct glacial I interglacial

climatic changes.

However, the atmospheric input of mineral dust to the marine environment, the lateral and veliical

transportation of lithogenic paJiicles in the ocean and their alteration by marine organisms are still a

subject of discussion, especially on time scales from days to decades. On shorter time scales,

environmental effects - such as dust storms, extensive rainfall, seasonal variations of atmospheric and

oceanic circulation - have a major influence on the abundance and flux of marine paJiicles and

lithogenic components with associated chemical elements (Kremling and Streu, 1993; Quetel et a1.,

1993). To quantify these fluxes, the pathways of lithogenic paJiicle transport and deposition have to be

bettel' understood. This includes the evaluation of processes like physical aggregation and selective

fallout of distinctive size fractions during atlnospheric transportation, particle loading through

zooplankton grazing and biological aggregation in the ocean's surface layer (Ramaswamy et al., 1991;

Stolzenbach, 1993), episodic versus continuous removal of particles from the upper water column into

Ratmeyer et a1.: Lithogenic particle fluxes emd grain size distributions... 47

the deep-sea (Wefer, 1989; Fischer et al., 1996b), and resuspension 01' lateral advection of slow

sinking partieies in the deep-sea.

Direct measurements of short-term variations 111 composition and abundance of mineral matter

during these different stages of transport and deposition, however, are limited. Successful approaches

include the determination of aeolian element fluxes (e.g. Buat-Menard 1989; Kremling and Streu,

1993), marine element and particle fluxes (e.g. Duce et al., 1991; Deuser, 1986; Wefer, 1989),

aggregate abundances (e.g. Honjo et aI., 1984; Asper et al., 1992; MacIntyre et al. 1995; Ratll1eyer

and Wefer, 1996), as weIl as the use of different theoretical models regarding aerosol (e.g. Schütz et

aI., 1981; Westphal et aI., 1987) and oceanic aggregate transportation (e.g. Jackson, 1990; RiebeseIl

and Wolf-Gladow, 1992; Risovic, 1993). But, due to the complexity of interactions all10ng thc

cOll1partments, further work and sufficiently more data are required to quantify both the transportation

processes and the mass flux of partieies in the ocean.

In this paper we describe lithogenic particle flux and gram Slze distributions obtained from

sediment trap deployments, in order to determine shOlt-term variations in flux and transportation of

lithogenic matter off NW Africa, one of the major global sources of mineral aerosoI. The main aims

are (1) to report and discuss the lithogenic fluxes and grain size distributions at different sites and

depth levels as parameter for particle alteration through vertical transportation, (2) to compare thc

different seasonal lithogenic flux patterns to the input of mineral components through atmospheric

processes, and (3) to evaluate the connection between input, sedimentation and deposition of

lithogenic particles for the sedimentary record offNW Africa.

REGIONAL SETTINGS

1. Atmospheric regime and dust transport

Aeolian dust transport over the NW African continent and the eastern tropical Atlantic occurs

either above the trade wind inversion within the Harmattan wind system, 01' by strong trade winds

blowing towards south-easterly directions along the coast (Fig. 1a). The mid-tropospheric Harmattan

winds originate over the central Sahara, where intensive insolation creates strong winds and heavy

convection in the dry surface air layer. Dust partieIes are lifted to altitudes higher than 6 km - known

as the "African Dust Plume" (Kalu, 1979), and move westward towards the tropical North Atlantic

above a strong trade wind inversion at about 1.5 km above sea level. Blowing at altitudes between 1

and 5 km, the westward moving Harmattan or "Saharan Air Layer" (SAL) reaches its maximum

strength and highest dust load during summer, then focusing into the "African Easterly Jet" between

1r N and 21 0 N. The almost continuous westward flow of these air and dust masses finally reaches

the Caribbean Sea after one week at altitudes between 1.5 and 3 km.

A second, north-turning part of the SAL carries dust from the Sahara and Sahel zone in an

antieyclonic flow along the African coast towards the Canary Islands, preferably forming a trajectory

48

at about 3 km above sea level in summer (Bergametti et al., 1989; Tetzlaff and Wolter, 1980). A major

climatic boundary, the Inteliropical Convergence Zone (ITCZ), separates the northeast trade winds

from the southeast monsoon. Due to the seasonal movements of the adjaeent high pressure cells, the

ITCZ moves between 2° N and 12° N above the eastern tropieal Atlantic, reaehing its northernmost

position during July/August. The periodical evolution of atmospherie easterly waves throughout the

lower and mid-troposphere during summer enables the dust transport aeross the ITCZ from souree

regions in the Sahel area towards the Canary Islands (Teztlaff and Wolter, 1980) and off the westcoast

of Africa (Karyampudi and Carlson, 1988).

101520

Madeira Island! '>VR~

4000-~ i

cv•

Canary Islands

25 W

Cape Verde Islands

-~I

30

25

15

20

10

1015

~\SAL

2025 W

15

10

20

25

351~'" , 11

N11fb\ I "i \:Y ce/ "ii

l30 Ii",ll

l..

c' "/11

"11 ~ ~/"II20i~ / ce 11~ \ ~

Jl j ~ \,l 1,1

1

1

:,11

1

' NEe / t, 11

'" I!:::- ~ ~II,1'..---7 '11:: ..---7 NEe~"i.11!L...- I .l=~~'===="J.

25 W 20 15 10

Fig. 1: (a) Attllospheric circulation and dust transpOli over the eastern subtropical Atlanticduring summer. During winter, trade winds turn to more easterly directions on the entire coast,and dust transport in the SAL becomes less intense. (b) Oceanic surface circulation off NWAfrica. Abbreviations: CC - Canary Current; NEC - North Equatorial Current; NECC - NorthEquatorial Counter Current. (c) Bathymetly and sediment trap sampling sites.

Ratllleyer et al.: Lithogenic particle fluxes and grain size distributions... 49

Although the high-level dust transport in the SAL was long proposed to be the main process of dust

export towards the eastern tropical Atlantic, recent studics (Bergametti et a!., 1989; Kremling and

Strcu, 1993; ChiapeIIo et a!., 1995) show the presencc of significant amounts of mineral aerosol in thc

10w level trade winds. In contrast to the high level winds, only sholi-range transport occurs in thc

trade wind layer above the north-eastern tropical Atlantic, which has strong implications for thc

interpretation of deep-sea sediment records off NW Africa (Sarnthein et a!., 1982). Mineral aerosol

around the Canary Islands can reach concentrations between 10 to 1000 Ilg m-3 at surface level

(Jickels et a!., 1987). Other allthors repOli highest monthly mean dust concentrations ranging from 90

to 120 ~tg m-3 at the Cape Verde Islands (ChiapeIIo et a!., 1995) to 22 Ilg m-3 in the low level winds at

Barbados in the western tropical Atlantic (Prospero and CarJson, 1972). Dust observed over thc

equatorial Atlantic Ocean was identified by the use of trajectory-computations (e.g. Coude-Gaussen ct

a!., 1987; Bergametti et a!., 1989; Swap et a!., 1992), elemental composition andmineralogy (Coudc

Gaussen et a!., 1987; Grousset et a!., 1989) and pollen identification in marine sediments

(Hooghiemstra and Agwu, 1986; Dupont and Agwu, 1991).

Source regions were coastal areas in southern Morocco and the southern Sahara and Sahel. As a

function of dust concentration in the different air layers over the continent and the ocean, turbidity and

solar radiation measurements give implications on the climatic effects of atmospheric dust loads at

recent times. Over the tropica1 and subtropical North Atlantic Ocean, mineral dust was found to be the

dominant light scattering aerosol, suggesting to be a highly effective climate forcing agent in areas of

high atmospheric dust load (Li et a1., 1996). In addition, increasing atmospheric heating rates fairly

correspond to maximum atmospheric dust load at latitudes between 10° and 30° N (Tegen et a!.,

1996).

2. Ocean Circulation

The three sampling sites referred to in this study are located within the Canary Current, which

flows southward along the NW-African coast as far as about 15°N and turns westward towards the

open tropical Atlantic forming the North Equatorial Current (NEC) (Fig. 1b) (Stramma and Siedler,

1988; Klein and Siedler, 1989). A strong coupling exists between the surface wind layer and the ocean

eirculation system, including a tight relation of the seasonal ITCZ movements and the latitudinal

extension ofthe equatorial eurrents as weIl as the year-round occurrenee in a 50 to 70 km wide coastal

band ofupweiling offthe coast (Mittelstaedt, 1991; Van Camp et a!., 1991).

Most intense upweIling oeeurs throughout the year off Cape Blane araund 20° N, where the

southward flowing North Atlantie Central Water (NACW) meets the less saline and nutrient-rich

north-flowing South Atlantie Central Water (SACW) between 150 and 600 m water depth, providing

part of the source water for upweIling (Hughes and Bmion, 1974). Further south off Cape Verde,

around 12° N and 20° W, the North Equatorial Under Current (NEUC) arrd the North Equatorial

50

Counter Current (NECC) build another upwelling cell which injects colder water into the westward

f10wing NEC (I-lagen and Schemainda, 1984). In this area, upwell ing is mainly present during

November through February (Mittelstaedt, 1991).

Phytoplankton biomass and the associated abundance of particlcs are high on the shelf along thc

entire coast duc to the influence of upwelling, ancl decrcase signincantly towards the open oeean.

However, off Cape Ghir as weIl as off Cape Blanc large filaments of chlorophyll-rich water are

advected several hundred kilometers offshore beyoncl the shelf edge, transpoliing colder waters with

enhanced pigment and particle concentrations towards the open ocean (Gabric et a!. 1993; Davenport

et a!., subm.). In general, primary production along the northwest African coast is mainly linked to

regions of intense upwelling, while the abundance and concentration of pigments and biogenie

particles off NW Africa is as well determined by the lateral and vertical transportation through

offshore advection of coastal waters.

METHODS

Lithogenic particle fluxes and grain size distributions of particulate matter were analyzed in 349

sampIes from 20 sediment traps deployed at 3 different sites off NW Africa (Fig. 1c.). Cone-shaped

multisampIe sediment traps with 0.5 and 1.17 m2

(only CB2) openings and 13, 20 01' 22 sampie cups

were used. Positions, depths, and deployment data of the traps are shown in Tab. I. In addition,

AANDERAA current meters of the type RCM 7 (CB 1,2) and RCM 8 (CB3,4, CV 1,2), and INFLUX

current meters (CIl-4) were installed 20 m underneath the upper traps during all deployments.

Prior to and after deployment, sampIe cups were poisoned with mercuric chloride, and density of

the sampIe water was increased by adding NaCI suprapur. After trap recovery, sampIes were wet

sieved (1 mm), splitted, treated and analyzed as described in Fischer and Wefer (1991). The lithogenic

fraction was calculated as (after Wefer and Fischer, 1993):

Lith = Total flux - (opal flux + carbonate flux + 2 Corg flux)

For grain size measurements on the lithogenic part of the sampIes, 1/16 to 1/5 split of the sampIe

was treated for 48 hours with 20% acetic acid and 7% H20 2 at 20°C to remove carbonate and organic

matter. Ultrasonic dispergation was used to destroy aggregates and fecal pellets prior, during and after

chemical treatment and prior to analysis.

Due to the relatively low opal content of the total trapped matter, no opal leaching was done in

order to prevent clay mineral alteration. The remaining carbonate-free fraction was neutralized with

2% NH3 and wet sieved. None of the sampIes contained a measurable amount of lithogenic grains>

63 Jlm. The fraction < 63 ~Lm was washed and centrifuged at 4000 rounds per minute with deionized

water and stored at 4°C.

Ratmeyer et al.: Lithogenic particlefluxes aml grain size distributions... 51

a) sediment trap mooringsTrap name Trap type Position Depth Trap depth Sampling Sampling intervalSite / No. (opening) Lat Lon (m) (m) period (sampies x days)

CI1 SMT 29°06.8'N 15°26.8W 3605 1000/3078 25.11.91 - 20 x 15(0.5 m2) 25.09.92

CI2 SMT 29°07.0'N 15°24.2'W 3606 1036/3067 01.10.92 - 20 x 10(0.5 m2) 09.04.93

CI3 SMT 29°09.1'N 15°26.5'W 3606 1026/3086 12.04.93 - 20 x 21(0.5 m2) 07.06.94

CI4 SMT 29°09.6'N 15°26.7'W 3625 923/3070 09.06.94 - 20 x 4(0.5 m2) 02.09.94

CS 1 Mark VI 20 0 55.3'N 19°44.5'W 3646 2195 22.03.88 - 13 x 27(0.5 m2) 08.03.89

CS 2 Mark V 21°08.7'N 20 0 41.2'W 4092 3502 15.03.89 - 22 x 17(1.17m2) 24.03.90

CS 3 SMT 21°08.3'N 20 0 40.3'W 4094 730/3557 08.04.90 - 19 x 22(0.5 m2) 12.06.91

CS 4 SMT 21°09.0'N 20 0 41.0'W 4108 733/3562 03.05.91 - 20 x 10(0.5 m2) 19.11.91

CV 1 SMT 11°29.0'N 21°01.0'W 4968 1003/4523 05.10.92 - 2x5+(0.5 m2) 04.04.93 18 x 10

CV2 SMT 11°29.0'N 21°03.5'W 4973 975/4435 07.04.93 - 20 x 25(0.5 m2) 25.08.94

b) surfaee sedimentsCore Mooring site Position Depth Sampie depth

Lat Lon (m) (em)

MUC 2913-4 CI 29°06.6'N 15°26.7'W 3621 0-1

MUC 2912-2 CS 21°16.8'N 20 0 42.5'W 4136 0-1

MUC 2911-2 CV 11°30.2'N 21°02.4'W 5020 0-1

Tab. 1: Positions, specifications and sampling intervals of the sediment trap sites.

Grain size analysis was performed on a CIS-GALAI laser partic1e analyzer. The advantage of this

instrument is the very low optical detection limit not depending on optical properties of the pmiicles.

The principle of operation of the CIS-GALAI is based on the relationship between the measurable

transition times of a constantly rotating ReINe laser beam and the beam intensity during detection of a

particle. A pmiicle in focus of the fast rotating beam produces a detection pulse, basically its shadow,

on a photosensitive sensor array installed across the sampIe cel\. Since the particle's movement is

practically zero compared to the fast rotating laser beam, duration of the detection pulse is uniquely

related to the pmiic1e's size (summary in Aharonson et a\., 1986). The system provides microscopic

video capabilities in combination with an image analysis system to measure partic1e shape and detect

aggregation during measurement. Rowever, in this study microscopic capabilities were only used to

control dispergation ofthe sampIe.

Due to the low partic1e concentration in most samples a 2.0 ml sampIe cell with magnetic stirrer

had to be used instead of a flow-through cel\. Prior to measurements, the whole sampIe was

homogenized with a magnetic stirrer and by ultrasonic dispergation. A 2 ml subsampIe was then

52

transferred with a pipette into the measurement cell. For each samplc, at Icast two measurcmcnts WCIT

performed using up to 5 different subsampies. Each measurel1lcnt consisted of a set of I() stacked

internal detection cycles based on a minimum number of counlcd particlcs. Analysis resolution W;IS

0.5 ~lm, with a minimum detectablc grain size 01'2.0 ~lm. Although grain si7c data from sediment trap

sam pies are available for the size spectrum 2.0 - 63.0 ~1l11, volllllle distribution spectra and grain si/.c

statistics were computed only for thc fraction 6.0 - 63 pm. McCavc et al. (1986) and McCave cl al.

(1995) found a significant underestimation ofthe clay-fraction « 2 pm) concurrent with a COl1stant

"ghosel-peak in the fraction 2 - 10 pm as a general problem of the laser difTraction method (see also

Bayvel and Jones, 1981: Dodge, 1984). Although the instruments based on laser diffraction are not

directly comparable to the CiS-GALAI particle analyzer used in this study, \ve decicled to excludc thc

smallest silt-sizecl fraction (2-6 ~1l1l) of particles from the clataset. In order 10 compare sediment trap

derived data with previously recorded grain size distributions from surface sediments offNW Africa

by Koopmann (1981) and Sarnthein et al. (1982), the fraction 01' 6-63 ~lm - instead 01' 10-63 ~lm lIscd

by McCave et al. (1995) - was computed for grain size distribution analysis ancl interpretation 01'

vertical transport patterns. Duc to the very low absolute particic conccntration in most sampies, the

clay fraction (grains< 2 ~lm) could not be quantitatively separated by decantation 01' filtration. Thus, a

direct measurement ofthe total size fraction was not possible.

a) surface sediment / artificial sampies

instrument MUC 2911-2 MUC 2912-2 MUC 2913-4 Glas beads(Site CV) (Site CB) (Site CI) (artificial)

CIS-GALAI mode (IJm) 8 23 19 26/53% <61Jm 16 7 7

FRITSCH mode (IJm) 3 19 18 24/52% < 2IJm 23 17 18% <61Jm 53 41 43

Sedigraph % < 2IJm 65 51 49% <61Jm 79 68 68

b) sediment trap sampies

instrument CB1 #5 CV10 #17 CV1 u #3

CIS-GALAI mode (IJm) 21 10 21% <61Jm 19 31 19

FRITSCH mode (IJm) 20 4 19% <21Jm 15 21 15% <61Jm 35 51 38

Tab. 2: Grain size data from surface sediments, sediment trap sampies and an artificial glasssampie measured with different instruments.

In order to estimate the average fine silt and clay content in the deep water column at the different

sites, surface sediment sam pIes from all mooring positions were analyzed. Sampies were taken from

the uppermost centimeter of multi-corer derived sediment cores. They were treated in the same way as

the sediment trap sam pies. Analysis was performed using a MICROMERITICS Sedigraph 5100, a

Ratmeyer et al.: Lithogenic particle fluxes and grain size distributions... 53

FRITSCH laser diffraction particle sizer (size range 0.5 - I 00 ~lm) and tbe CIS-GALAI particle analyzer

(size range 2 - 63 ~m). However, a direct comparison of the Sedigraph and tbe laser sizers is not

possible due to different analysis techniques and detectable size ranges (see also Singer et al., 1988;

McCave et al., 1995). AdditionaIly, only 3 sediment trap sampies contained enough lithogenic

particles to fit the detection sensitivity of the FRITSCH instrument. Tbus, data obtained from the

Sedigraph were solely used to estimate tbe clay and fine silt content in comparison to theFRITSCH

laser sizer.

100.010.01.0

/~ ", ', ,001.02.0304.05.060

asl'd<JnIl'<",

---///-------'

.-'

0.1

MUC 2912-2 (Site CS)"t2.._··········~1 d >6 ~m$0 ./ n = 23j" .:. /<. r2 = 0.627

5.0 Li" 30 '.)<~- 20 ~•• '/

4.0 ':: /' '. ~J

6.0

2.0

1.0

3.0

0.0

b

100.0

a MUC 2911-2 (Site CV)6.0

Ta d>6lJrn

~ 4.0

n = 20

5.0~[30 , ••. , " ••"c:.

r2 =0.594

~4.0

i 2.0 /~c 10 -,'Q):::J 00

g~ 00 10 20 3.0 4.0 5.0 60

3.0 asYdurl!~~ <", ,Q)

E:::J 2.0Ö>

1.0

-',.

0.0

0.1 1.0 10.0

C MUC 2913-4 (Site CI)

100.010.0grain size

(j.Jm)

1.0

1.0

2.0

0.0

d glass spheres

3.0

100.01.0 10.0grain size

(j.Jm)

"lLJ d>6~m5.0 n =23j" .... .' r2 = 0.708

§g "t' .;' :., .9····i 20 •

1.0 •

00 ,.'

00 1.0 2.0 3.0 ~_O 5.0 e.oos"*",,

''''

6.0

5.0

>-u4.0c

Q):::J

g'O'3.0~C

Q)

E:::J 2.0Ö>

1.0

0.0

0.1

Fig. 2: Comparison of grain size distributions ofthree surface sediments (a - c) and an artificialglass sphere sampie (d) analyzed with a CIS-GALAI particle analyzer (solid lines) and aFRITSCHlaser diffraction particle sizer (dotted lines).Gray shaded areas give the percentage of silt sizegrains< 6 ~m measured with the FRITSCH instrument (except glass spheres sampie). Small plotsshow the correlation of percentages obtained with both methods (solid dots for grains of 6 - 63~m, open dots for grains of2 - 6 ~m)

54

a 6.0 CV1u Nr. 3

5.0'~\\,

~,, ,

4.0, ,, ,

~, ,, ,

5f~,,

ol=~ 3.0 \Ql ,E

2.0\

:J ,~ \,1.0

"

0.0

0.1 1.0 10.0 100.0

b 6.0 CV10 Nr. 17

5.0

~ ,,... ' .. -....."

4.0 ,~

,,!l;;?

,3.0

,,ol=~ ,Ql ,,§ , .'

2.0 .'

~51 %

1.0

0.0

0.1 1.0 10.0 100.0

C 6.0 CB1 Nr. 5

5.0

r 4.0

l~ 3.0GI

§ 2.0~

1.0

0.0

0.1 1.0grain size

(jJm)

10.0 100.0

Fig. 3: Comparison of grain size distributions of three sediment trap sampies analyzed with aCIS-GALA! particle analyzer (solid lines) and a FRITSCH laser diffraction particle sizer{dottedlines). Gray shaded areas give the percentage of silt size grains< 6 IJm measured with theFRITSCH laser sizer.

Results ofthe comparison are plotted in Figure 2 (a-c: surface sediment sampies, d: artificial glass

bead sampie; M. Breitzke, Bremen, unpublished data) and Figure 3 (sediment trap sampies). Table 2

shows the modal sizes and grain size statistics derived from the' different instruments. The results

clearly show a significant underestimation of the fine-grained fraction measured with the laser

diffraction and the laser transition techniques. The FRITSCH laser "sees" between 20 and 30 % less

particles < 6 IJm compared to the Sedigraph, and only about 1/3 of the clay fraction. The CIS-GALAI

instrument, however, detects even less particles < 6 IJm compared to both alternative instruments,

RatlJleyer et al.: Lithogenic particle fluxes and grain size distributions ... 55

whi le für grains> I0 ~tm a good correlation was obscrved for sam pies with relatively low contcnts of

grains< 6 ~m (Iess than 50 %).

A good resolution of modal size and abundance was found for relatively coarse sampies (mode >.

10 ~lm; sampies GeoB 2912-2, GeoB 2913-4, CBl #5, CVlu #3, CBl #5 and glass beads; Fig. 2b, c,

d; Fig. 3a, c), whereas finer sampies show both an underestimation of the coarse fraction as weIl as a

shift of the main modal grain size towards smalleI' diameters on theCIS-GALAI (sampie GeoB 29 I 1-2

and CVl 0 #17; Fig. 2a, Fig. 3b).

In contrast, the FRITSCH laser resolves modal grains in the silt fraction less precise than the Crs

GALAf (see glass beads sampie, Fig.3d), which confirms previous findings by Singer et a!. (J 988) anel

McCave et al. (1995). For the discussion of grain sizes in sediment trap sampies, the modal diameter

measured with the CIS-GALAI must be carefully interpreted. Modal sizes around I0 ~lm probably

reflect high contents of sm aller grains rather than a real mode at 10 ~lIn, whereas larger modes arouncl

20 ~lm reflect real values resulting from highest abundances of grains with diameters aroulld 20 ~lll1.

Due to the extremely low concentration of paliicles in most sam pies, however, theCIS-GALAI particle

analyzer was the ollly usable instrument. As discussed above, data of the coarse fraction > 10 ~m are

usable for the interpretation of downward transport, which as weil fits to the grain size range of a

significant amount of atmospheric dust off NW Africa alld thus provides information of dust input to

the ocean.

Wind data were obtained by the Global-Telecommunications-System of the World Meteorological

Center, and sea surface temperatures were derived from the IGOSS-NMC data set (weekly mean SST

data blended from ship, buoy, and bias-corrected satellite data on a one degree grid; see Reynolds anel

Smith, 1994). 1'0 evaluate seasonal changing provenance of the low level winds, daily trajectories at

the 850 hPa level were computed for alI sites over the last year of a 21-year model run (U. Wyputta,

Bremen, unpub!. data). Computation of 4-day backward trajectories was performed after Petterssen

(1956), using a two-dimensional isobaric model based on a model developed at the Norwegian

Meteorological Institute (NM!) (Eliassen, 1976). The wind fields used for trajectory calculation were

generated with the ECHAM 3 atmosphere circulation model (Lorenz et a!. 1996).

56

RESULTS

Continuous lang-term records are available for the c1eep sediment traps in the Canary Islancl regicJIl

(Cl, 1991 - 1994), off Cape Blanc (CB, 1988 - 1991) ancl off Cape Verde (CV, 1992 - 1994). 'rimc

series of data from traps at shallow depths are partly interrllpted or incomplete due to malfunctions.

Total and pariial particle flllxes measured off Cape Blanc were pllblished by Wefer ancl Fischer (1993)

and Fischer et al. (1996a). Patiicle fluxes at the CI site north of the Canary Islands are presented by

Neuer et al. (in press). Grain size statistics and relative abundances were computed and normalized for

the volume distribution of silt-grains of 6 - 63 ~lm. Lithogenic particles> 63 ~lm were not found in any

sampIe. In addition to the analysis of sinking particlilate matter, a grain size analysis of multicorer

derived sediment sampIes was undeliaken to determine the entire grain size spectrum of the surfacc

sediments at the mooring sites CI, CB and CV (see Methods).

1. Sire CI (Canary lS'land.s)

Paliicle flux

At about 1000 m water depth, lithogenic fluxes ranged from 0.7 to 43.6 mg m-2

d-I• Highest

lithogenic fluxes occurred in spring 1993 (Fig. 4d), strongly correlated to the highly seasonal total

particle flux (Neuer et al., in press). Though amplitudes differ considerably between years, spring anel

early summer maxima indicate a seasonal variation both in total and lithogenic fluxes during 1992 anel

1993. In 1992, lithogenic particles contributed the largest part to the total pmiiculate matter during

sampling time, with maxima of 67 % in May and 72 % in June. From 1992 to 1993, the relative

contribution of lithogenic to total particulate matter decreased, which is also reflected by low

Lith,/Corg ratios during this period (Fig. 6).

Lithogenic paliicle fluxes obtained with the deep traps at3000 m varied from 1.2 to 61.8 mg 111-2

d-I,

which was significantly higher than the fluxes measured at 1000 m. Throughout the entire sal11pling

time, lithogenic fluxes showed astrang correlation to the total pmiicle fluxes, both in time and

amplitude (Neuer et al., in press). As observed in the upper traps, the seasonal flux pattern is clearly

visible (Fig. 5d), with almost steady recurring maxima in spring through a11 years, and additional

peaks in late summer and fall.

UnIike the data from 1000111, maxima ofthe Lith./ Cora. ratio correspond to lithogenic particle flux<0

peaks in MaylJune 1992, November 1992, May 1993 and JulylAugust 1994, whereas the main flux

peaks in spring 1992, 1993 and 1994 show a stronger correspondence to organic carbon fluxes (Fig.

6). This indicates a higher contribution of lithogenic matter at greater water depth during the summer

seasons. Where data are available for both elepth levels between 1991 and 1993, temporal variations of

lithogenic patiicle fluxes appeal' to be weil corre1ated between 1000 and 3000 m water depth

(Fig.13a).


Canary Islands (CI1-4, upper traps)

30 J25

20

15

:: j10

a 1992 1993 1994

60b

5.5

50 4.5

.~'E40

.5 :1 %I!-'" 30

20

mean ciameterOJm}

10

02.09.1994

1994

no sampet Ivltilable

lithogenic tlux 0

organk: Carbon flux •

c

d60

Fig. 4: (a) Sea surfaee temperatures, surface wind velocities and direetions at site CI north ofthe Canary Islands. Gray shaded are strongly seasonal periods of highest wind velocitieseorresponding to strictly NNE directions. (b) Grain size spectra, equivalent spherical meandiameters, (e) fluxes of grains in size class 6 - 11 Jlm and 20 - 63 Jlm, (d) lithogenic fluxes andorganic carbon fluxes (Corg) obtained from sediment traps CIl - CI4 (1000 m water depth).

58

Canary Islands (CI1-4, lower traps)

Fig. 5: (a) Grain size spectra and equivalent spherical mean diameters, (b) fluxes of grains insize classes 6 - 11 J!m and 20 - 63 J!m, (c) lithogenic fluxes and organic carbon fluxes (Corg)obtained at site CIl - CI4 (3000 m water depth).

60a

50

40

~--ä §:

'" 30

20 ]mean dameter

(~m)

10

30bJh

~M _

20'e:Z ."ß..: "eHg',!!~ 10~. ~

go

60 C~'"-8:-- 40

i'"'e.ll g'

20,~

11.12.1991

1993

5.5

4.5

%

Grain size

The size spectrum of lithogenic particles is generally characterized by a high content of small

grains< 11 J!m (Fig. 4b). Mean grain diameters at 1000 m varied between 10.7 J!m (February 1992)

and 15.5 J!m (May 1992) (Fig. 4b). The largest measured grains had diameters of 51.5 J!m (December

1991). Highest abundances over the entire deployment time were observed for the size class 6 - 11 J!m

during January/February 1992 and 1993 and in the end of June 1994, respectively. Two distinctive

peaks are visible in the time series of the ratio of grains< 11 J!m to grains 20 - 63 J!m (Fig. 6), both

occurring in the winter/spring season in 1992 and 1993 when fluxes were highest, too. The fraction of

large grains with diameters between 20 and 63 J!m showed the opposite trend (Fig. 4c).

The size fraction > 6 J!m in the lower sediment traps at 3000 m is characterized by mean values

between 11.0 J!m (May 1994) and 17.0 J!m (May 1992) (Fig. 5a), which occurs to be higher than

observed around 1000 m water depth. Annual mean grain diameters varied between 14.9 J!m (1992),

13.5 J!m (1993) and 12.8 J!m (1994, Jan. - Sept.). Largest grains had diameters up to 55.0 J!m. As

Ratmcycr el Cl!.: Lithogcnic particlejluxcs (md grain size distributions... 59

described for the upper traps, grains of the size class 6 - ll pm were most abundant ranging between

38.3 % in June 1992 and 69.8 % in May 1994.

Unlike in the upper traps, a general trend towards highcr abundance of thc fine-grained fraction

occurrcd between April 1992 anel June 1994. In contrast, the frilCtion of grains with diameters between

20 anel 63 11m showeel a elecreasing trend over this perioel. In both size classes, short-term abundance

variations up to 20 % occurreel during winter/spring seasons in 1992 anel 1993, anel towarels the enel of

sampling time between July anel September 1994.

02.091994

1994

upper !raps:--- I'

lower traps: ------.'---------_.----'

Site CI20.0

~ E~ "-i'lcoullvaNN 10.0~-'0 EQ.?-~:h

00

20.0

2'0

~g~ ~ 10.0C xv ~

8'~

~

00

Fig. 6: Ratios of fluxes of grains in the size class 6 - 11 11m versus 20 - 63 ~lm and Lith.lCorgratios for both depth levels at site CI.

A comparison of mean diameters between both trap levels reveals no significant correlation (r2 =

0.089, n = 45; Fig.13b), though the trend from December 1991 through June 1992 appears to be

comparable, including the spring maximum in the < 11 11m fraction. From spring 1993 through fall

1994, a negative correlation of lithogenic fluxes versus the ratio of size classes 6-11 11m anel 20-63 11m

is visible (Fig. 6). This indicates a higher contribution of coarser material at greater water depths

during periods of high fluxes than observeel for the previous year.

2. Site CB (Cape Blanc)

Data are available from April 1990 through November 1991 for the shallow sediment traps around

700 m water depth (Fig. 7d), and for the entire deployment time from April 1988 through November

1991 for the deep traps deployed between 2195 (CB 1, 1988) and 3562 m (Fig. 8).

60

Cape Blanc (CB3-4. upper traps)

1991199019891988a

:: j20

15

:]10

IIIIIIII

III

, I I I

W IS" 14.;} nIf~If"" ':' a .,~ ....~~xt'f: uJi:~~~II~fJl.""f*'! I'~.W: ~ ~ :s ------,---~_:.--{_.~f; ,----;'-;-.-;! -~'-- ---.-!- .:'::- -'-:. --'; -1--~ -~- ~T --', ---._!~ -.~. ---"'-'- -,t -( _C - - L"'_'__'_:_ -- -~---;.-- - _. -, - - --,

E -L;'~~······:···:!!I"':·w·'.~--·-.1••'-""'.'.:'!-'-')~""·· ~:·~W·~""'··"'lfi···':·"',;'~-··'··fl>:~'.'_.·:."'.'&::.lo!oljl:.._:~:.•~-.: ~·ij!!I~~···:·,...·i+l:~!.,l······:·:··:·.ii.-~'.~••:....·····IL:···········i~.-~i~I~L$.iI~:·~,~·:··:.<:····L·······..:,,:'$:..:..l,.:' ••..._.. - .,-,!.c..::;n~-T-1

b 60 5.5

no ....... _

60 Ch.U, 40

l!.ll 'El!lr 20

)'ll -0

160 }d 80

i ,t, 120 '. 6O~

:j .il 0'E 60 40 3 ~J! .~ ~ ~

40 20-"

no oampllo __

0 0

1988 1989 1990 1991

Fig. 7: (a) Sea surfaee temperatures, surfaee wind veloeities and direetions at site CB off CapeBlane. Gray shaded are seasonal periods of highest wind velocities. (b) Grain size spectra andequivalent spherical mean diameters, (c) fluxes of grains in size class 6 - 11 J-tm and 20 - 63 J-tm,(d) lithogenic fluxes and organic carbon fluxes (Corg) obtained from sediment traps CB3 - CB4(700 m water depth).

Ratmeyer et al.: Lithogenic particle jluxes and grain size distributions... 61

Grain size

Mean diameters of grains sampled with the upper traps around 700 m ranged from 11.5 to 18.5

)lID, showing an overall declining trend from April 1990 through August 1991 (Fig. 7b). Maximum

diameters of 53.3 )lm were observed in August/September 1990 and July/August 1991. From spring

1990 through February 1991, grains were characterized by mean diameters> 15 ~lm. Largest meall

diameters around 18 )lm were observed during May and August through October 1990. The

corresponding grain size spectrum shows a broad distribution of mediull1 abundances of grain sizes

between 10 and 26 )l1l1. During these months, grains ofthe size class 20 - 63 )l1l1 contributed over 50

% to the lithogenic partic1e flux.

In contrast, weil sorted, high abundances of grains< 11 )lm occurred during the late spring and

summer 1991 (Fig. 9). Grains< 11 ~lm contributed 50 to 65 % to the Iithogenic fluxes in summer

1991, with two small peaks ofmean diameters (14.7 and 14.4 )lm) corresponding to lithogenic particle

flux maxima observed at the same time.

Cape Blanc (CB1-4, lower traps)

60a

5.5

50 4.5

40

30

20 Jmeandiameter(11m)

10

60

3.5

%

2.5

1.5

0.5

30

20 '3 ()'" 03 <l

I, E'10 5- )(

19.11.1991

1991

lithOgeni<::f1U-;~)

~aniC Carbon flux .! I

199019891988

c

18.04.1988

80

40

120

160

Fig. 8: (a) Grain size spectra and equivalent spherical mean diameters, (b) fluxes of grains insize classes 6 - 11 )l111 and 20 - 63 )l111, (c) lithogenic fluxes and organic carbon fluxes (Corg)obtained at site CBl -CB4 (2195 m - 3562111 water depth).

62

Mean diameters obtained with the lower traps ranged from 12.1 to 19.5 pm, where highest values

were found in June and October 1988 and from August through October 1990 (Fig. 8a). Smallcsl

l1lean diameters were observed in the interl1lediate years during April 1989 ancl June 1991. Smallcst

l1lean diameters occurred during 1991, inclicating a d istinct year-to-year variance as observed in Iota I

and partial particle fluxes (Fischer et al., 1996a). On a lang-term scale, distinctive variations in grain

size are visible in 2-year intervals, whereas a clear seasonal pattern cannot be resolved from bolh

l1lean diameters and size spectra. In camparisan to the upper sediment traps, grain size distributions

appeal' to be less sOl"ted. Peaked size abundances occurred only during fall 1988 and June 1991 for

grains around 20 pm and grains< 11 pm, respectively.

Site es10.0

0.0

"//_J\ _\

20.0

10.0

0.0

18.04.1988

1988

Fig. 9: Ratios of fluxes of grains in the size class 6 - 11 pm versus 20 - 63 ~lm and Lith.lCorgratios for both depth levels at site CB.

Where sampies from the upper traps (1990-91, Fig. 7a) are available, both trends and amplitudes of

mean diameters and frequency distributions are weil correlated to the lower traps (Fig.8a, 13 d). In

contrast, lithogenic particle fluxes show less significant correlations (Fig. 13c).

3. Sire CV (Cape Verde)

The particle flux pattern recorded with both upper (at"ound 1000 m) and lower (around 4500 111)

sediment traps at CV is dominated by a distinctive decrease of lithogenic fluxes from the period

before October 1993 to the time from November 1993 through August 1994 (Table 3;Fig. 11 c). At

1000 m depth this change was even stronger than observed in the deep traps, resulting in extremely

low concentrations of particulate matter in the sampie cups. Because quantitative sampie splitting and

analysis was impossible due to extremely low pmiicle concentrations, partial fluxes and ratios could

not be analyzed from this depth for the period after October 1993.


Cape Verde (CV1-2, upper traps)

..

1994

°.°.

1993a

I

1

III1

..j.-. I. -.

I1 Ij , , .. .I ,.. •. " ., .. " .•..

W ' . • ,. . "I ~I' . I

SE---~b·F~~~;;~~;;;:·~~·;·;~·~~~~~·

,,~ 30 1cE~8~ I G' 25

8.8'-!~ 20.: ..(Q! 15

:: j. ~ .'.:,t. .'.

10

o

~ 30

20 ~ 0

3 JlN ~

10 ~ ><

1994

Ilthogenic t1ux 0

organlc Carbon ftU)( •

1993

,,0."

..;,'" \\'b- ~ ~~ ,P

b- ~ --

.-6o

60

90

11.10.1992

120 ; d

.,

::~J.,

60b 5.5

50 4.5

1140 3.5

'iij 'E.5 :L %E-'" 30 2.5

20 ] 1.5mean diameter

(~ml

10

Fig. 10: (a) Sea surface temperatures, surface wind velocities and directions at site CV, offCape Verde. Gray shaded are seasonal periods of highest wind velocities and correspondingNNE directions. (b) Grain size spectra and equivalent spherical mean diameters, (c) fluxes ofgrains in size dass 6 - 11 I!m and 20 - 63 I!m, (d) lithogenic fluxes and organic carbon fluxes(Corg) obtained from sediment traps CVl - CV2 (l000 m water depth).

64

Particle flux

Total fluxes at 1000 m depth ranged from 0.6 to 304.8 mg m-2 d-I. Lithogenie fluxes ranged from

11.3 to 128.7 mg m-2 d-I for the sampling period from Oetober 1992 through Oetober 1993. Both

lithogenie and organie earbon fluxes show strongly eorresponding seasonal trends to the total partiele

flux pattern during the first year of sampling (Fig. 1Oe). Peaks occurred during winter ancl spring,

whereas lower fluxes with less strong amplitude variations are observed for summer and early fall

1993. Highest lithogenic fluxes of 128.7 mg m-2 d-I were recorcled in Mareh 1993.

In contrast to the declining lithogenic fluxes from winter 1992 through summer 1993, the average

contribution of lithogenic particles to the total partiele flux increasecl from the winter season (Oetober

1992 through Mareh 1993) through summer (April through September 1993). This trend is as weIl

reflectecl in the Lith.l Corg ratio (Fig. 12). From November through August 1994 total particle fluxes

showecl a similar seasonal pattern, but with significantly lower amplitucles.

Cape Verde (CV1-2, lower traps)

60

a50

40

30

20 Jmean diameter(11 m)

10 -

60

J

b~ Ec"-

~~ ~ 40

~~ ':'E

~~ E 20§'~ ..-go I

.. -. \ i

"0\''/ c' v

.- - ...

%

2.5

1.5

0.5

120

90

60 ~

30 i

c

•.• --+.. .. ....

--j

lithogenic flux

organic Carbon flux • ~30

20'3 0

~ ~;, E

d;.:~l10 ~ x

11-10.1992

1993 199425.08-1994

Fig. 11: (a) Grain size spectra ancl equivalent spherical mean diameters, (b) fluxes of grains insize classes 6 - 11 11m and 20 - 63 11m, (c) lithogenie fluxes and organic carbon fluxes (Corg)obtained at site CV 1- CV2 (4000 m water clepth).

Ratmeyer er (11.: Lirhogenic p({rticlefluxes ({nd grain size distributions... 65

'fotal particle fluxes obtained from the deep traps at 4500 m ranged from 0.7 to 401 mg m-2

d-I.

Lithogenie fluxes varied between 0.5 and 104.9 mg m- 2 d-I. As observed at 1000 m depth, a

signifieant deeline offluxes oeeurred from 1992 to 1994 (Fig. I Je). Highest lithogenie fluxes oecurred

during winter 1992 and spring through May 1993 without clear defined peaks. The contribution of

1ithogenie particles to the total flux at this depth was higher than at 1000 m (40.6 to 70.9 %). Highesl

pereentages oeeurred during summer 1993 through Oetober 1994. The Lith.l Corg ratios did not shO\v

seasonal trends as observed for the first year of sampling at 1000 m, though values were eonstantly

higher (Fig. 12). Comparing upper and lower sediment traps, lithogenie particIe flllxes do not show

any signifieant eorrelation (Fig.13e), though general trends in annual mean values exist.

Grain size

Grain size analysis of sampies from 1000 m depth was possible only for the first year of sampling.

Mean equiva1ent spherieal diameters of the size elass 6 - 63 ~lm ranged from 11.8 to 18.7 ~lm. Like

observed for the lithogenie particle flux, a seasona1 pattern is visible in the time series of the grain size

speetrul11 and the eomputed arithmetie mean diameters, reflecting a winter/spring and a summer

season (Fig. lOb). In eontrast to related lithogenie flllxes, a distinetive inerease of mean grain sizes

oeeurred between March and April 1993, and from August through Oetober 1993. In general, mean

diameters during winter and spring 1992/1993 refleeted high abundanees of grains< 11 Fm,

espeeially in March 1993, A sharp inerease ofmean diameters from 12.5 ~m (March 1993) to 18.7

~m (lune 1993) was eaused by a diserete shift of the size-speetrum from highly abundant paliieles <

11 ~m to high abundanees of grains> 20 ~m. Also, the upper limit of the size speetrum inereased

sharply from 35 to 49.5 ~m. During summer, the mean diameter deereased slowly from July through

Getober 1993, reaehing a minimum of 14.7 ~m in September.

Site CV

25.08.1994

upper traps:-tower traps: ------

19941993

1] •.•••••jn/~./ ••..••••.s==un .....nunnnnnu .....u.un

I I

11.10.1992

10.0

~ E

~f.§ 5.0q.Q ..-

e;i;0.0

20.0

rg~ ~ 10.0

l~0.0

Fig. 12: Ratios of fluxes of grains in the size cIass 6 - 11 ~m versus 20 - 63 ~lm and Lith.lCorgratios for both depth levels at site CV.

66

Note that the corresponding size spectrum again behaved discrctc: dccrcasing mcan diameters are

caused by a step-like downward shift 01' abundances 01' grains around 20 ~lm vcrsus grains< 1 1 pm,

rather than a continuous shift 01' grain diameters through intermediate size classes (Pig. lOb). 'fhe

seasonal change between winter/spring and summer situation is as weil represented in the ratio 01' the

size classes 6-11 versus 20-63 ~lm (Fig. 12). During the entire sampling time, the grain size üequency

distribution shows a weil sorted spectrum with high abundances in narrow size classes.

Mean grain diameters obtained at 4500 m depth were generally higher than those observed at 1000

m. Values ranged from 14.5 to 20.7 ~lm, caused by an overall high abundance 01' grains with Slzes

from 10 to 20 ~lIn in almost all sampies. A similaI' trend 01' increasing mean diameters as observed in

the upper traps is visible, where smallest sizes around 15 ~lm are observecl in spring 1993 and largest

diameters arouncl 19 ~lm occurrecl cluring late summer and early fall (Fig. 11 a). A final clecrease

towarcls mean diameters 01' 15 ~lm oecurred in May 1994. This trencl goes along with changes in the

sorting 01' the size spectrum. During winter and spring 1993, a broacl spread abundances 01' grains

between 6 ancl 28 ~lm occurred, while cluring summer 1993 through spring 1994 grains between 20

and 28 ~lm clominatecl the size spectrum. Lithogenic particle f1uxes anclmean diameters dicl not show

any correlation between the two clepth levels (Fig. 13e). Insteacl, a weak correlation 01' mean cliameters

appears to be present, though it cloes not reach a significant conficlence (Fig. 13f).

4. Comparison bet,veen the different sites

Significant clifferences in the lithogenic f1uxes ancl grain size clistributions occurrecl between the

three sites. Annual ancl overall mean lithogenic f1uxes ancl equivalent spherical mean diameters are

given in Table 3. Highest annual lithogenic f1uxes (31.2 - 56.1 mg m-2 cl-I) were observecl at the Cape

Blanc site (Fischer et al., 1996a). Largest annual mean diameters were recorcled at site CV (14.5 - 16.9

pm) and site CB (15.2 - 16.7 pm). Lowest annual1ithogenic f1uxes (11.4 - 21.2 mg m-2

cl-I) as weil as

smallest observeclmean diameters (13.5 - 13.7 pm) occurrecl in the Canary Island region.

Significant correlations between lithogenic ancl organic carbon f1uxes were observecl at all sites.

Lith./Corg ratios were generally lower at shallower compared to deeper trap levels (Fig. 14a, c, e).

Greatest clepth-re1ated clifferences in Lith./Corg ratios occurrecl at site CV. At site CB, the gradient 01'

Lith,/Corg ratios computecl from 1988 to 1990 differ consiclerably from those found cluring 1991 at the

3500 m level (Fig. 14c). The other sites show more consistent graclients.

To determine which grain sizes contributed most to the lithogenic material, f1uxes were plotted

against mode cliameters ofthe fraction > 6 ~lln 1'01' each site (Fig. 14b, d, f). Differences between sites

occurred both in mocle diameter and relation 01' upper versus lower traps. Mocle diameters were

generally larger in the lower traps at either site. The laI'gest difference in modal diameter between

depth levels occurrecl at site CV, the lowest was observed at site CB. In the Canary Island region,

grains with diameters 01' 6-11 pm generally contributecl to almost the total amount 01' lithogenic f1uxes

Ratmeyer ef aI.: Lithogenic particlefluxes alld grain size distributions... 67

at both depth levels (Fig. 14b). In contrast, sites CB and CV show larger mode diameters with maxima

of24.3 /-lm (CB) and 26.1 ~lm (CV) (Fig. 14d,f). According to weil correlated grain size distributions

and mean diameters observcd off Cape Blanc (Fig. 13d), thc range of Lith. flllx/mocle diam. ratios is

eomparable for both trap levels (Fig. 14c1).

Site CI Site CS SiteCV

120,0

e

80.00.0

0.0

n:::: 29

,'=0.047

40.0

"

80.0

1200

1200

c

, ", .40.012D.O80.0

---.-

40.0

n = 44

~ = 0.367

0.0

. "00 ~._,.LC.__'----'-''--_~ ,

80.0

120.0 .------..- •••.•.•••..•.......•..

Iowertrapslilhogenic particle flux

(m;j rn'cf')

lower1rapslilhogenic partide flux

(m;j rn' d')

lowertrapslithogenic partic!e ftux

(m;j rn' cf')

20.0 ~------..- •...... - ..•.-.-..--.----. 20.0 .' .... - •••..-.-•• _ ••.•..--.--.-•.-~-...•••- .•.•••• 20.0 , --..- ---._-•..... - .•• - •• --- __ co

~ = 0.665

;,E1lc

" .g-E~iP 15.0

n = 45

~ = 0.089

, I "t. ,

-- .~.- ~..-:.~ ~-~ .. '0.' ." .

b n = 36

. ,,,,/'

/',,/<

/,/,:;/," ,

/', '

d n = 29

~ = 0.279

15.0

10.0 ,__. ~__ ..__..•c • ..... l

15.0

lower trapsequivalent spherical mean diameter

(~m)

lowertrapsequivalent spherical rrean diameter

(~m)

10.0 L ~ ---'-._._.•..•.•.•._ •.~._ .••..•..•.......•. ) 10,0 •....._._. • __. .L - ••.

20.0 10.0 15.0 20.0 10.015.0

lov~er trapsequivalent spherical rr.ean diameter

(~m)

Fig. 13: Correlation of lithogenic f111xes and equivalent spherical mean diameters at sites CI (a,b) CB (c, d) and CV (e, 1) obtained from upper versus lower traps. Solid regression linesindicate a significant correlation between depth levels. Significant correlations at the 95%confidence level are found at site CI (a, f1uxes) and at site CB (d, mean diameters).

Low and intermediate lithogenic f111xes are mostly represented by grains with mode diameters< 11

flm, which is comparable to the f1ux amplitude and modal grain sizes observed at site CI. In the lower

traps, large mode diameters corresponded to significantly higher f1uxes, whereas samples from upper

traps did not show this trend. In contrast to sites CI and CB, sampies from site CV showed clearly

different patterns for both trap levels. At 1000 m, 11ighest lithogenic particle f1uxes during 1993

represented smallest grains with mode diameters < 11 /-lm. These are the highest lithogenic f1uxes

recorded at the upper trap level of either site. The lower traps showed no relationship between f1uxes

and mode diameters (Fig. l4e).

30020,0100

./

'00

Site CV

",,'. '1:ßOO :~d.

1;>00

1600

c

400300

Site CS

20.050040.030.020.0

68

Site CI

160,0 160.0

a

120.0 1200

J~

ti ",. 800 ! ~ H00

r40.0 40.0

Corg flux

(rrgm'd')b. upper traps

• () lower traps

Corg flux

(rrg m1 6')

Carg fllfx

{~m'd-'}

1600

d

120,0

1,6

-300

, ,

"0.0

20015.010.0

00

5025.020.015.0

1200

160,0

b

15.0 20.010.0

400

1200

equivalent spherical mode diameter(~m)

equivalent spherical mcx:le diameter(~m)

equivalent spherical mode diarneler(~m)

Fig. 14: Lithogenic fluxes versus organic carbon flux ratios at the different sites CI (a), CB (c)and CV (e), and fluxes versus equivalent spherical mode diameters (b, cl, f) from upper (dashed)and lower (solicl) depth levels. Open dots show lower trap sampies which have no equivalentsampIe from the upper trap. Organic carbon values were not corrected for dissolution losses.

DISCUSSION

1. Currents and trapping efficiency

Regarding the efficiency of trapping sinking particulate matter, horizontal current velocities have

to be evaluated in order to determine a possible hydroclynamic bias caused by speeds> 12 em s·1

(Baker et al. 1988). No significant influence of currents on total and patiial patiicle fluxes at the sites

CI and CB could be interpreted from current records due to low current velocities mostly below 10 CIl1

s·l, and due to the lack of correlation to any of the described parameters including grain size

distributions obtained from the trap sampies (Neuer et al. in press, Fischer et al. 1996a). A similar

situation appears für the site off Cape Verde, where currents between 1.5 to 10 cm s·l and 1.5 to 15

Cl11 s·l were recorded during winter 1992 and summer 1993, respectively.

2. Transport processes and pathways

Both lithogenic flux data and grain size frequency distributions show different veriical and lateral

trends and correlations at the sites studied (Fig. 13,15). Differences occur between depth levels at a

given site as well as between sites at a ceriain water depth. A parameter for these variations is given

by the correlation of either lithogenic fluxes or grain sizes between depth levels (Fig. 13). The

Rallllcyer 1'1 al.: Lilhogenic parliclc.flllxes al/(/ grain size dislrihlll ions... 69

sign iIIcance of correlation may provide in format ions on both transportat ion ,1I1d alteration 01' the

material through changes of its composition. Weil correiatedflllxes indicate high particle sinking

speeds anel large aggregate sizes, resulting in a fast removal of particles frolll the upper ocean, anel

thus short reelunelance times available for disaggregation anel partial decomposition.

In contrast, well-correlateel grain size spectra anel Illean diwne/ers may provide a measure for thc

ielentiflcation 01' a given particlc population at different depth levels, indicating that no signilicanl

alteration of the initial composition took place eluring transportation. Ilowever, the effeet of

hyclrodynamic processes such as lateral currents have to be carefully evaluated duc to the scnsitivity

01' silt-sized grains in the range of 10 - 63 ~lm ("sortable silt" after McCave et al., 19(5) to

fractionating and sorting during lateral transportation (McCave et al., 1995, Kranek and Milligan,

1991). Furthermore, the signi fieance 01' coagulation and aggregation must be estilllated from eitherill

si/li measurements or from parameters describing the probability of biogenic transportation, such as

CIN-ratios or lithogenie versus organic carbon fluxes.

ßecause a signitlcant correlation of lithogenic ancl organic carbon fluxes was found for all studied

sites (Fig. 14 a, c, e) (see also Wefer and Fischer, 1993; Fischer et al. 1996a: Neuer et al. in press), wc

assume that the vertical partiele transport above 1000 m water depth was Illainly driven through

biogenie aggregation. The lithogenic particlc input from dust fallouts may thereby have important

eff'ects on the sedimentation processes. High amounts of relatively heavy quartz grains may inerease

sinking speeds 01' large aggregates (marine snow) as weil as feeal pellets through partiele loading

(Ramaswamy et al., 1991), ancl eonsequently reduee organie matter degradation (Haake anel Ittekott,

1(90).

North 01' the Canary Islanels anel off Cape Blanc, ei/her lithogenic fluxes or equivalent spherieal

me,1I1 diameters show a sign ificant eorrelation between depths (Fig. 13 a,b,c,eI). In contrast, off Cape

Verde no correlation was found tor both parameters (Fig. 13 e,f). These findings reflect elifferent

vertical particlc transportation anel alteration processes in the deep water column. In detail, we

assumed the following situations for the three studieel sites:

a) Site CB (Cape Blanc)

A sign itlcant correlation of the gram size distri bution between upper (1000 m) and lower traps

(3500 m) was faune! only at this site. Mean eliameters of the fraction > 6 ~lm correlateel weil when')

rcgressed linearily (r~ = 0.665, n = 36, sec Fig. 13d). Although the size spectra show less-sorteel size

distributions for the lower traps (Fig. 8a), even eletailed patterns in the spectra are recognizable when

comparing the two depth levels, especially during summer 1990 anel summer 1991 (Fig. 7b, 8a).

COlllparablc patterns in sam pies obtained in almost 2.5 km vertical distance frOl11 each other point

to no alteration during sinking. Furtherlllore, the time lag between dcpth levels occurreel to bc in the

range 01' 10- 20 days, translating into relatively high particle sinking speeds bet,veen 150 anel almost

70

280 m d- I, which confirms prevlous cstimates reported by Fischer et al. (1996a). However, il

comparably good correlation of lithogenic flllxes is missing, althollgh the seasonal pattern in botl1

depth levels is visible between 1990 and 1991. In addition, during 1990 ancl 1991 astrang decreilse 01'

thc organic carbon-flllx by abollt 40 % bctween trap depths was reported by r"ischer et al. (1996a).

This is as weil represented in the Lith./Corn-flllX ratios (Fig. 9, Fig. 14e) ancl indicates ei ther ab

degradation of organic carbon or increased scavenging of lithogenic particles with depth. 'l'hese

eontradictionary findings - comparable grain size spectra indicating no major particle alteration on the

one hand, and strongly decreasing Corg val lies with depth on the other hand - afford a furtllel'

disetlssion ofthe mechanisms ofparticle transportation thrallgh the water colllmn.

Paliicle loading of biogenic aggregates by lithogenic components increases particle sinking rates

(Ramaswamy et al., 1991) and may transfer a given grain size distribution from the "catchment area"

tovvards greater depths, if aggregates remain undestroyed. Large, fast sinking aggregates (marine

snow) are discllssed to scavenge sllspended, laterally advected particles (Asper et al., 1992:

Stolzenbach, 1993) during their transport through the water column, thereby increasing the total

particle flux with depth. In case of an enhanced content of lithogenic grains in the advected material,

scavenging would certainly change the grain size distribution of lithogenic components in the sinking

matter found at greater water depths.

Such a change is not reflected in the data recorded at the CB site. However, lithogenic fluxes were

lower by a factor of two between summer 1990 and summer 1991 in the upper traps, correspond ing to

a low size class ratio of the 6-11 flm versus 20-63 flm size class (Fig. 9). During the following

summer 1991, both flux and grain size changed significantly, leading now to corresponding

amplitudes of fluxes and low Lith./ Corg ratios at both depth levels. These changes implie a more

efficient biogenic removal of the lithogenic fraction from the surface during summer 1991 than

before. From summer 1990 through spring 1991, higher Lith./ Corg ratios and coarser grain sizes were

found at both depth levels (Fig. 9), together with higher lithogenic fluxes at the lower depths (Fig. 8c).

Here we suggest a less efficient removal from the surface level and possible scavenging of relatively

coarse, single lithogenic grains from the same source during vertical transpoli.

This explanation was also suggested by Kremling and Streu (1993), who found AI as tracer for

aeolian input not correlated to the partietI1ate P-flux at a sediment trap site 500 km west of site CB.

Because size spectra show similar patterns between depths during this period as weIl, increasing

fluxes with depth may be caused by the higher catchment volume of the lower traps (Siegel et al.,

1990) rather than the additional lateral input of advected or resuspended material with a different

composition. Finally, adegradation of organic carbon within the sinking aggregates through

dissolution and microbiological activity would certainly lead to pmiial disintegration of sinking

aggregates and arelease of particle components into the water column (Alldredge et al. 1990).

However, a significant removal of partial lithogenic size fractions caused by different settling rates of

Ratmeyer et al.: Lithogenic particle jluxes and grain size distributions... 71

sl11all versus large grains is not visible in the grain sizc spcctra. Although certainly a major process

during aggregate sinking, the effect of carbon dcgradation with dcpth on thc grain size distribution

seems to be less strang than through additional input of ncw material by scavcnging of advccted or

resuspcnded matter, which is discussed below for the CI sitc.

Duc to the sensitivity of the grain size distribution for alteration, the strang correlation of grain

sizcs observed between depths leads to thc conclusion that at site CB no major alteration of thc

sinking material by means of an additional input through resuspended or laterally advected particles

took place. Thus, we suggest that site CB may be characterized as a reference site for almost nOI1

disturbed, rapid veliical transport of paliicles through the water column.

b) Site CI (Canary Islands):

The significant correlation of jluxes between upper and 100ver traps in the Canary Islancl region (1'2

= 0.37, n = 44, see Fig. 13a) indicates rapid sinking ofthe bulk particulate matter. This observation is

also documented by the correlation of lithogenic, organic carbon and other paliial fluxes (Fig. 6, Fig.

14a; Neuer et al., in press). However, this result gives no idea on the abundance and behavior of

different paliicle populations, e.g. sinking versus suspended matter. In contrast to paliicle fluxes, thc

missing correlation of grain size mean diameters of the silt fraction between 1000 and 3000 111

indicates a different material composition at a given time at both depth levels (Fig. 13b). This result is

opposite to Cape Blanc. Additionally, Neuer et a!. (in press) reported constantly higher fluxes for thc

lower traps, with lithogenic material following this trend.

From other open ocean sites, several authors (e.g. Walsh and Gm'dner, 1992; Honjo et al., 1992;

Ittekott, 1991) reported increasing particle and element fluxes with depth in the Sargasso Sea and thc

Gulf of Mexico, respectively, and assumed that lateral advection of paliicles from distant sources may

have caused this trend, Due to the regional settings close to the NW-African continent this assumption

may be valid for the CI site as weIl. Neuer et a!. (in press) discuss the advective transpOli of

phytoplankton biomass towards the CI-site through large filaments, advected from the NW African

continent. During the winter flux peaks at both depth levels, the C/N-ratios in the trapped material

were typical for fresh phytoplankton detritus. In contrast, higher differences between depth levels

were found during summer, translating into langer particle residence times and slower sinking

velocities.

According to Ramaswamy et al. (1991) and Asper et al. (1992), a seasonal differing interaction of

sinking aggregates with lithogenic components and laterally advected organic flocs would explain the

increasing pmiicle fluxes with depth and seasonal changing C/N-ratios. This assumption is suppolied

by the grain size variations. Grain sizes differ significantly between 1992 and spring 1993 with depths.

The far less sorted grain size spectrum in the 10wer traps does not show any peaks of fine-grained

material recorded during 1992 with the upper traps (Fig. 4b, 5a). The coarsening of grain sizes on the

72

one hand, and the missing correlation of amplitudes on the other hand point to some kind of alteration

of the lithogenic fraction during sinking.

The lower traps show a different relationship of lithogenic flux versus mean diameter comparcel to

the upper traps (Fig. ISa), which is mainly related to flux peaks recoreled during early spring 1992 anel

summer/fall 1992. Due to the close coupling of organic carbon, carbonate ancl lithogenic fluxes. it

seems unlikely that the lithogenic fraction solely receives these major changes in composition. Wc

assume an alteration of the sinking matter through scavenging of material derived from laterall.y

advected filaments.

According to Siegel et aI. (1990), small aggregates with sinking speeds of 1-35 m er lwould havc a

lateral distribution range of up to 500 km, leading to an enhanced record of particles from distanl

sources depending on the sampling depth. In addition, the coarsening of grain sizes could be causccl

by the input of larger grains scavenged from atmospheric fallout regions closer to the NW-African

continent during the phytoplankton bloom. If a lateral transportation of these filaments occurs, larger

grains compared to the atmospheric fallout at the distal CI-site should be involved into the westward

aclvection of aggregates and finally reach greater clepths and offshore areas. One etfect would be the

change ofthe grain size distribution ofthe entire particle population at the trapping site.

c) Site CV (Cape Verde):

In contrast to the situation observed in the Canary Island region, data from the CV site showed

neither a correlation of lithogenic fluxes nor a significant correlation of grain size means between the

two trap depths (Fig. 13e,f). However, the weak correspondence of mean grain diameters may be

caused by lower settling velocities towards greater depths. The overall trend visible in the upper traps

occurred also at the deep level, as shown in the plots of grain size spectra (Figs. lOb, 11 a).

Comparing fluxes and grain sizes, a correlation occured between the mean diameter and the Lith./

Corg ratio at the upper depth level: during winter when high fluxes were corresponding to highly

abundant grains< 11 /1m, a weIl developed coupling of lithogenic and organic carbon fluxes occurreel,

together with relatively low Lith./ Corg ratios (arounel 5; Fig.12). In contrast, during spring ancl

summer 1993 the contribution of large lithogenic grains increased due to a major increase of the Lith./

Corg ratio by almost a factor of two. The coarsening of litbogenic grains between spring anel summer,

however, corresponels as weil to a major seasonal change in atlnospheric transportation of mineral

elust from low level to high level winds at these latitueles, as discussed below. Altbough a correlation

of grain size anel Lith./ Corg ratio is not as clearly visible at the other sites, a closer inspection of tbe

grain size spectrograms shows a similar trend at CI (Iower traps, Fig. 5a) eluring May 1992 anel

OctoberlNovember 1992 as weIl as general trenel at CB (upper traps, Fig. 7b) from 1990 to 1991. In

contrast to the situation at site CI, elecreasing litbogenic fluxes were recordeel for the lower traps.

Ratmeyer et al.: Lilhogenic particle jluxes and grain size distributions... 73

These results may reflect a different paliicle transport mechanism through the water column than

described for the nOlihern sites, indicating that some kind of alteration occurred during transport.

Sc,lVenging of laterally advected, differently composed material at greater depths seems to be unlikely

duc to the missing increase 01' fluxes with depth compared to site CI. Since the lithogenic fluxes die!

not correlate between the different depth levels, we aSSlllne that the deeper traps received material

from a larger 01' a different source area compared to the upper traps.

Further interpretation 01' transpoli mechanisms appears to be not reasonable due to missing data of

deep current velocities as weil as data from the upper traps. FOIiunately, data on atl110spheric dust

concentration are available for the deployment period 01' the CV sediment traps (G. Bergametti, pers.

comm.; Chiapello et al., 1995), and allow a comparison 01' lithogenic fluxes and grain sizes from 1000

m water depth and wind-transported dust measured at Sal, Cape Verde Islands, which will be

presented elsewhere (Ratmeyer et al., subm.).

Site CI Site CS

20.0

b

Ci ",,10;

~'. ~

.... " ..". 1>." ~l:>

15.0

equivalent spherical rrean clarreter(~m)

0.0 , ..,"_"...~_.~ ..__.L...__._. __ .... _._... .. __'

20.0 10.0

160.0

a120.0

80.0

,f

0'40.0

15.0

equivalent spherical mean diarreter(~m)

160.0

120.0

~ ~-u ".~ ~E 80.0

'",s E::J -

40.0

,0.0

("

10.0

Site CV

160.0 ;.-.-~~~~-- ... _.--~~~~___,

c120.0

"0" • t>,;

80.0 ., •116 • /

j' ••

40,0 ";

20.015.0

equivalent spherical rrean diarreter(~)

0"

[lc> ;0.0 ,---.---~~_.~~~-.,.."..'-~-~

10.0

Fig. 15: Provinces 01' lithogenic flux versus mean diameter 1'01' the three sites obtained fromboth depth levels. In plots 1'01' sites CI (a) and CV (c), dotted lines indicate high abundances ofcoarser grains in the lower traps. At site CB (b), values cover a similar range of ratios wheredata are available 1'01' both depth levels. Open dots show lower trap sampIes which have noequivalent sampIe from the upper trap.

74

3. CompariscJ/1 ofsinking particulate matter and sm/ace sediments

A comparison of sediment trap derived grain sizc statistics (fraction 6-63 ~lm) with those obtained

from surface sediments (uppermost centimetel' of multicorer-derived sedimentcores) recovered at the

mooring sites and data repOlied by Koopmann (1981) and Sarnthein et al. (1982) shows Table 4.

Modal diameters appeal' to fit weIl to published data, with highest values at site CB off Cape Blane.

SmaIlest modes were found for the CV site, however, with a significant difference of absolute müdes

detected by the CJS-GALAI and the FRITSCH instrument due to the high contents of clay (65 %, see

Table 2). The percentages ofthe fraction 6-63 ~m of the totallithogenic fraction also confirm findings

by Koopmann (1981) for surface sediments, again with highest abundances off Cape Blanc (Tab. 4).

A different situation, however, is reflected in the sediment trap data. Here, Im'gest average modal

grain sizes are found off Cape Verde, amI smallest values occurred at the CI site. Although our data

represent average particle fluxes integrated over 2 to 4 years sampling time, the effects of high

seasonal and interannual variations may have a major effect on grain size distributions. Thereforc,

they cannot be directly compared to the upper surface sediments, which comprise arecord of several

1000 years. The range of modal diameters at sites CI and CV (16 to 20 ~lIn), however, corresponds to

the range observed in surface sediments. Instead, lithogenic fluxes recorded in the lower traps at cither

site appear to reflect the sedimentation rates observed in the sediments off NW Africa by Koopmann

(1981) for the last 4000 years. Again, highest values are observed for site CB with an annual mean

flux of 56 mg m-2 d-I (Tab. 3), corresponding to accumulation rates of 28 - 56 mg m-2 d- I,

interpolated for this site by Koopmann (1981). Sites CI and CV recorded lower mean fluxes between

21 and 28 mg m-2 d- I, respectively, which is confirmed by sediment accumulation rates between 28

and 14 mg m-2 d- I (Koopmann, 1981).

a) sediment trap sampies (Iower traps)

annual mean size > 6 IJm CI (Canary Islands)

mode (IJm) 8mean (IJm) 14

b) surface sediments (this study)

silt sized fraction> 6 IJm MUC 2913-4

mode (IJm) 19 (18)*mean (IJm) 20 (18)*% of silt « 63 IJm) 32**

c) surface sediments after Koopmann (1981)

silt sized fraction > 6 IJm around CI

mode (IJm) 16-20% of silt « 63 IJm) 30-35

CS (Cape Slanc)

1617

MUC 2912-2

23 (19)*12 (17)*

32**

around CS

20-2535-40

CV (Cape Verde)

1917

MUC 2911-2

8 (17)*18(17)*

21**

around CV

16-2020-30

* data from FRITSCH laser diffraction particle sizer** data from Sedigraph analysis

Table 3: Comparison of annual mean and modal grain sizes and abundances ofthe lithogenicfraction> 6 ~m in lower sediment trap sampies and surface sediments.

Ralllleycr ct al.: Lithogcnic particlefluxes and grain size distributions... 75

While sediments reflect accumulation rates witll inverse correlation to distance from the coast

rather than latitudinal variations (Koopmann, 1981), present lithogenic fluxes through the water

colullIn are sensitive to spatial differences along the latitudinal axis by I11cans of scasonality and mass

input, and thus show a higher variability between sites than surfüce sediments. In addition, higher

accumulation rates measured with sediment traps at greater depths llIay at least partly result from thc

trapping 01' resuspended material, which then is recorded as part of the total vertical flux towards the

deep-sea (see also Gardner ,md Richardson, 1992).

site depth (m) 1988 1989 1990 1991 1992 1993 1994 total

a) annual mean lithogenic particle flux (mg m-2 d-1)

CI 1000 8.33 14.40 11.363000 20.32 24.90 18.47 21.23

CS 700 37.83 24.55 31.193500 79.98 55.63 58.17 30.50 56.07

CV 1000 61.92 n.d. (61.92)*4500 49.44 7.34 28.39

b) annual mean equivalent spherical mean diameter (Jlm)

CI 1000 13.60 13.49 13.543000 14.91 13.48 12.78 13.72

CS 700 16.89 13.55 15.223500 18.19 15.57 17.34 13.99 16.72

CV 1000 14.50 n.d. (14.50)*4500 16.71 17.14 16.92

* lithogenic fluxes at CV (upper trap) were under detection limit during 1994

Table 4: Annualmean fluxes and grain diameters at the different sampling sites.

4. AtlJ10sphcric input 01 lithogcnic particlcs

The trade winds off NW Africa are the main driving force 01' coastal upwelling and provide a

major part 01' mineral dust transported towards the open ocean. ThllS, they have a great impact on

flllxes 01' patiiclliate matter in the subtropical and tropical north Atlantic. Figures 4a, 7a and 10a show

trade wind velocities and directions measured at Las Palmas (Canary Islands), NOlladhiboll (Cape

Blanc, Mauretania) and Sal (Cape Verde Islands) as well as sea surface temperatures during sampie

collection periods at the sediment trap sites CI, CB and CV, respectively.

Strong seasonal patterns are visible in strength and direction 01' the trade winds during the entire

sampling period for all sites. Significant differences between sites can be recognized following a

general trend from northern towards sOllthern positions: While temperatures generally increased, mean

wind speed decreased both in amplitude and variability. Highest scasonal changes in wind speed

occurred at the Canary Islands, whereas highest intcrannual changes in both wind speed and sea

surface temperature were measured at Cape Blanc. At the same site strongest SST-gradients occurred

76

between seasons, probably reflecting intense upwelling 01' coleler waters dming spring anel fa 11 (sec

also Fischer et al., 1996a).

Average wind speeds al1Cl sea smface temperature oeeurreel to be in phase at the Canary Islands

anel ehanged towards opposite phases at the Cape Yerde Islands. Wind elireetions turned inereasingly

towards easterly and north-easterly d ireetions arOlll1d Sal Is land, wh ile off Cape Blane and arolillei Ihe

Canary Islands dominant westerly and northwesterly components are visible clLlring intennediilte

seasons. Although all sites belong to the same atmospherie and oeeanic subsystem off NW Africa,

these data show that signif~eantdifferences exist in the atmospherie anel ocean surl:lee environ!1lents at

the sampling sites as \vell as the transport 01' dust throughout the year.

a) Site CI (Canary Islands)

At site CI, where highest changes In temperature and surface winds oecur throughout the year,

strongest seasonality eould be observeel in the total and partial partiele fluxes, with highest maxima

occurring in the deep traps 10 days to one month after the phytoplankton bloom took place in Sllrfaee

waters during late winter from 1991 through 1993 (Neuer et al., in press). A similar seasonal pattern

ean be found in the surfaee wind speeds and direetions, showing both lowest wind speeds and highest

oeeurrenee 01' omni-directional winds during winter and late spring (FigAa). During the remaining

months 01' the year, strong and stable north-easterly trade winds with a small westerly component

prevail in March/April and from July through September throughout the entire sampling period (gray

shaded areas in Fig. 4a). From air-impaetor measurements and trajeetory-analysis at the pressure

levels between 850 al1CI 925 hPa, several authors report dust transport towards the Canary Islands

during either spring or summer situations from different sources in Maroeco and the Sahel area

(Coude-Gaussen et al., 1987; ßergametti et al. 1989).

To eompare marine lithogenic fluxes with the atmospherie eireulation patterns 111 the trade wind

layer, daily 850 hPa trajectories were examined for eontinental versus oeeanic provenanee in order to

get a monthly based index 01' possible dust transport (Fig. 16a-e, upper). In addition, lithogenie lluxes

were eompllted as monthly mean values 01' upper verSllS lower traps (Fig. 16a-e, lower). Our results,

however, indieate the highest eontribution 01' dust at thc CI site during winter and spring, whcn the

provenanee 01' low level winds derived from trajeetory analysis oeeurs to be mainly eontinental (Fig.

16a, upper). The oeCUITenee 01' eontinental winds eorresponds to maxima in the monthly mean

lithogenic llux reeord during February, Mareh al1Cl Oetober for both trap levels (Fig. 16a, lower). A

maximum in the 100ver traps in July/August, however, is neither retleeted in the trajeetories nor in the

upper traps, whieh eonfirms the assumption 01' an additional, non-atmospherie input 01' mineral matter

through lateral adveetion in the deeper water eolumn, as diseussecl above.

Thc supply 01' additional material during summer/fall as visible j~'om monthly mean values in the

lower traps fits to observations reported by Davenport et al. (unpubl. data), who found extended

Rallllcycr cl a!.: Lilhogcnic particlc fluxcs and grain size dislribul ions... 77

filament activity outside the primary upwelling zone ofT Cape Ghir and inereased chlorophyll bioll!a~.;s

to levels of winter bloom values at the CI site (see Neuer et al., in press) during summcl'. J-lowevcl',

although fluxes observed at the CI site appeal' to correspond to the highly seasonal traclc winds, grain

size cl istri butions from both clepth levels do not provide a corrclation with atmospherie tem pora I

patterns (Fig. 16a).

Site CI

Fig. 16: Comparison ofmonthly mean lithogenic fluxcsand occurrence 01' eontincntalwinds at 850 hPa per month atsite CI. Wind valucs areeomputed from dailytrajeetories over onc ycar (U.Wyputta, unpubl. data) andrefleet the timc, whcre lowaltitude winds had a eontinentalrather than ocean ie provcnance,transporting dust towards thesampling site.

\

\

\\

\[;j

RI

II

I

I

'0

hl\

\

\\

\\ /

U

mI

I

II

I

L1

4.0

2.0

5.0

3.0

1.0

00

0.0

40.0

20.0

30.0

10.0

upper traps: -s-tower traps: --0 ."

b) Site CB (Cape Blane)

Off Cape Blanc, the seasonality of fluxes is also eorrelated to upwelling and high biomass

production when strong trade winds from NNW to NNE prevail (Fig. 7a, see also Fischer et al.,

I996a) rather than to the enhanced dust input (Kremling and Streu, 1989). Although aeolian dust input

appears to be present throughout the year, biomass production and vertical patiicle export to the deep

ocean by biogenie transportation obviously overcomes initial surface aerosol input patterns. This fact

is supported by previous findings of Sarnthein et al. (1982) who prapose the distinction of the

different dust transport layers from the sedimentary record at these latitudes only by compositional

features instead 01' grain sizes. Westphal et al. (1987) repoli a significant occurrence of Itcoastal

sources lt of dust around Cape Blanc, which translates into an additional factor controlling grain size

distributions.

Frall! trajectory analysis, a frequent occurence of continental winds at the 850 hPa level was found

throughout the year, rcaehing abundances between 3 and 4 weeks per month during spring ancl fall

(Fig. 17, upper). A correlation of continental winds and monthly mean fluxes is not visible throughout

the ycar (Fig. 17 lower). This observation may partly be caused by the continuous high level dust

trallsrorl in thc SAL at pressure levels around 50ü hPa, with maxima during summer (Semmel hack,

1934~ Samthcill cl al., 1982). We conclude that the trigger 1'01' sedimentation and removal of lithogenic

78

particles from the upper ocean at the CI and CB site is coupled with biologieal proeesses rathel' than

lithogenie particle input. This eonelusion is eonfirmcd by a study on short-tcrnl scdimentation (Fischer

ct aL, 199Gb), where pu Ises with durations in the range of hours to days were observed at thc CI si tc, II'

episodie lithogenie inputs would eause aggregation and rapid particle sinking, a bctter corrclation 01'

lithogenie fluxcs and atmospherie input patterns should be obtained.

Thc mean grain size observed in the lower traps at CB was the highest reeorded at either sitc.

confirming findings by Lepplc (1975) and Westphal et al. (1981) who observed lat'gest modal

diameters by mass (25 - 30 ~lm) of eoastal dust sourees for Cape Blane on the T\1auritanian coasL

Temporal differences in atmospheric dust eoneentrations on the eontinent along the eoast are as weil

reported by d'Almeida (1987) for two stations in Mauritania and Senegal, with highest vertieal optieal

turbidity during April/May in Boutlimit, Mauretania, and a different pattern with maxima betwecn

.Tuly and September observed in Dakar, Senegal. Although these sites are not direetly comparable to

the oeeanic reeord of both aerosol and marine partiele fluxes, they indieate a eertain variability of dust

expOli from eontinental sourees betwecn latitudes on a short time seale.

Site CB

5.0

4.0

3.0

2.0

1.0

0,0

100.0

80.0

60.0

40.0

20.0

0,0

//, /

El

upper traps: _tower traps: --8--·

,, ,b- - -0- - -E( /

Fig.17: Comparison of monthly mean lithogenic fluxes anel occurrence of continental winels at850 hPa per month at site CB.

c) Site CV (Cape Verele)

In contrast to the situation at site CB, grain sizes recorded at the CV site refleet changes in the

meteorological pattern elue to more contrary seasonal situations in the traele winels anel the SAL.

However, the significant decrease of fluxes after October 1993 was exclueleel from the eliscussion.

Mean anel modal grain sizes of the lithogenic fraction at 1000 m water elepth increase from values

arounel 12 ~lm eluring November 1992 through March 1993 to 18 J.lm in spring anel early summer

Ratll1eyer et a1.: Lithogellic particle fluxes alld grain size distributions... 79

1993. These values confirm findings of Jaenicke and Schütz (1978) for mineral aerosolmodal grain

sizes around 10 to 30 ~lm observed between 500 and 1000 km off the coast as weIl as resu lts from

model calculations by Westphal et al. (1987). From beginning of March through September 1993,

grains with diameters of 20 ~m and more dominate the grain size spectrum recorded in the lower traps

and contribute most to the total lithogenic fraction. In fact, this happens at the time when surface

winds measured at Sal, Cape Verde Islands, turn from mean ENE to NNE directions (Fig. 10a).

Results from trajectory analysis for site CV show highest abundance of low altitude winds as weil,

with continental provenance during winter and spring (Fig.18, upper), and a significant decrease

during late spring and summer. During this time, high-altitude winds in this area are known to turn

strictly from westerly to easterly directions and develop the "African Easterly Jet" above 2 km altitude

during summer (Carlson and Prospero, 1972; Chiapello et al. 1995). Due to this seasonal change in the

atmospheric vertical structure, the transportation of continental dust changes as weIl. Low-Iayer

easterly trade winds are known to transpoli high amounts of dust during winter and early spring (90

~g m-3: Chiapello et al., 1995; 350 ~g m-3: Westphal et al., 1987). Instead, dust transpOli from late

spring through summer is dominated by the easterly h igh-altitude winds. This change of the transport

from low layer to high altitude winds shoLJld infiuence the grain size distribution. Previous studies

showed the possible presence of grains up to 90 ~lm (Glaccum and Prospero, 1980) and 200 flm

(Jaenicke and Schütz, 1967) in the atmospheric surface level at distances > 1000 km off shores.

Assuming a comparable initial grain size distribution with a certain component of coarse grains, low

layer transport as reported during winter provides a larger fraction of grains for sedimentation

compared to high level transport.

During summer, when highest dust transpoli occurs at altitudes up to 6 km, only a small fraction

consisting of pmiicIes large enough to fall rapidly tb rough both wind layers may reacb the surface at a

given distance from land (Bergametti, pers. comm.). Smaller grains are transported further westward

and finally reach the western tropical Atlantic and the Caribbean Sea (Prospero et al., 1981). In

addition, those small grains falling through the low layer winds receive a higher horizontal stress

compared to coarse grains, and could therefore be diluted in the sinking fraction reaching surface level

(Jaenicke and Schütz, 1978). Apparently, this process results in higher fiuxes and smaller grains

entering the oceans surface during winter, corresponding to the trend of monthly mean atInospheric

dust concentrations at surface level on Sal Island from 120 to 70 flg m-3 during winter (Chiapello et

a1., 1995). Monthly mean lithogenic fiuxes computed for both trap levels at CV confirm this trend,

although only the upper trap record shows a time lag of fiuxes compared to trajectory provenance

(Fig. 18).

80

Site CV

5.0(l)0 .-.c c

~ cu 4.0~ c

Q)U) >_

3.0TI 0 U) -c~x

.~~~-J'l~ 2.0oc~

Q)(l)

~~ 1.0

80.0

100.0X:J

C ~ 80.0

~.g ~E 1ij" 60.0.?- Q. EE.S:! 0} 40.0§ ~ EE E~ 20.0

0.0

G - -ß- _-Cl

J

8/ \

I

I

I

I

[1/

/

/

'u

upper traps: -tIlower traps: --0--·

Fig. 18: COl11parison of 1110nthly l11ean lithogenic fillxes and occurrence of continental winds at. 850 hPa per month at site CV.

The direct cOllpling of Coru and lithogenic particle fillxes indicates a rapid sinking of paliicles anclb

efficient removal of airborne lithogenic material from sllrface waters as discllssed above. This resllit

affords the presence of a mainly llnaltered grain size distribution at the upper trap level. We conclllde

that the coarsening of grain sizes and the decreasing trend of lithogenic fiuxes at 1000 m wate I' depth

from March to April 1993 refiects the change of atmospheric dust transpoli from low layer trade

winds to easterly jet winds at high altitudes. Although standing in contrast to repolis of Sarnthein et al.

(1982) for sediment measurements at higher latitudes nOlih of Cape Verde, this assumption enables a

direct comparison of aerosol concentration measurements, air mass trajectory analysis and sediment

trap derived lithogenic fiuxes and grain sizes for the years 1992 and 1993.

d) Seasonal ity

In order to provide a parameter for the differences in seasonality of dust carrying winds and

lithogenic fluxes, monthly mean lithogenic fluxes of the lower traps were computed as clll11ulative

percentages over one year after the method of Berger and Wefer (1990). To obtain an index for

seasonality of organic fluxes, these authors define the length ofthe time it takes to generate one half of

the annual production as "production half-time", using data sorted from highest to lowest. We think

that the latter method is as weil valuable to obtain a measure of seasonality of lithogenic particle

fiuxes. To compare seasonal indices, continental trajectory abundances were as weil computed in the

same way (Fig. 19).

Ratmcycr ct al.: Lithogcnic partic!c f7uxcs alld grain sizc distributions... 81

10 12

Sile CLSite CS:Site CV: --+-

~lil(2)

6543210

60

40

20

SI

20

month

80

100

a

80

~ 60 -~

~c _

f~40 -

c8

100 -,--------------="""b Lithogenic Fluxes

(Iower traps)

month10 12

SI 6 5 4 3 2 1 0

Fig. 19: "Half-time" values and seasonal index (SI) of lithogenic fluxes (after Berger andWefer, 1990; see text) recorded at the lower depth levels at the different sites, in comparison tocontinental trajectory abundances.

The results provide a similar shape ofthe cumulative flux plots (Fig. 19b), but closer examination

shows distinct differences between study sites: highest seasonal indices (SI) of 2.2 and 1.9 are found

for sites CI and CV. Site CB has a significantly 10wer seasonal index of 1.0. This result confirms the

idea of highest variations in flux and seasonality of lithogenic matter at sites CI and CV with 50% of

the annual flux recorded during 3.8 to 4.1 montlls (Fig. 19b), corresponding to a comparable

seasonality in abundance of continental winds between 3.4 and 4.1 montlls (Fig. 19a), respective1y.

Off Cape Blanc, however, both fluxes and winds show 1ess seasona1ity, reflecting a higher year-round

continuity of mineral dust provision towards the ocean within the trade wind layer. Using the

classification of Berger and Wefer (1990) for expoli flux types, the temporal situation regarding the

82

trade wind-related mineral dust transport and lithogenic particle flux in the ocean would be describcd

as "medium seasonal" for sites CI and CV, and as "weakly seasonal" to almost "constant" for site eH.

CONCLUSIONS

Our results enabled us to estimate manne lithogenic particle fluxes received from different

atmospheric levels, in order to determ ine the impact of trade wind-driven (short-range) versus high

altitude related (Iong range) transport of dust and input to the ocean. Because only three datasets of the

marine lithogenic particle flux were available, the following simplifications were made: The arca 01'

mass flux estimation was set from 300 N to IooN, covering a region between 300 and 400 km

(equivalent to site CI), and500 to 600 km (corresponding to sites CB and CV) offshore. Three equal

area subregimes corresponding to the three trap sampling sites were assumed, each covering abotlt

800 x 100 km2.

A schematic model of the mass fluxes and transpol1 pathways off NW Africa is presented in Fig.

18. Numbers and pathways are deduced from sediment trap fluxes, sediment accumulation rates

(Koopmann, 1981; Sarnthein et al., 1982), 850 hPa trajectories (U. Wyputta, Bremen) ancl

atmospheric dust data (Prospero and Carlson, 1972; Lepple, 1975; Schütz et al., 1981). For

calculations of yearly fluxes in the three sub-areas, annual mean sediment trap derived lithogenic

fluxes were used (see Tab. 3). The range of initial atmospheric dust-exp0l1 from the continent towarels

the ocean was assumed to be 120 - 520 x 106 t y-l (after Schütz et al., 1990), with a contribution of 24

- 104 x 106 t y-l in the nOl1hern area. This assumption translates into a total yearly dust input of 1.0

4.3 x 106 t y-l (Canary Islands) and 2.0 - 8.7 x 106 t y-l (Cape Blanc, Cape Verde).

The seasonal variance of SAL versus trade wind-transported dust could not be resolved from the

Schütz model, but seasonal changes derived from trajectory computations (see Fig. 16a-c) anel grain

size analysis are qualitatively shown by arrow-broadness in Fig. 18. However, these values are

roughly confirmed by sediment trap-derived fluxes of 0.7 x 106 t y-l (3000 m water depth, Canary

Island area) and 0.9 x 106 t y-l (1000 m water depth; Cape Blanc area, Cape Verde area). A fUl1her

quantification of dust transpol1 in the different air layers may in future be resolved from Aerosol

Optical Depth (AOD) measured with the satellite-based AVHRR sensor (Advanced Very High

Resolution Radar), and from clay mineralogy ofthe trap sampies.

At site CI, the mass flux at 3000 m water depth was chosen due to the additional input of laterally

advectedmaterial at this depth (0.3 x 106 t y-l) probably originating from SUl-face waters closer to the

coast (as discussed above; Neuer et al., in press). Water column-fluxes are generally lower by a factor

of 2, assllll1ing minimum dust input at surface level. This is confirmed by corresponding surface

sediment accumulation rates of 0.4 - 0.8 x 106 t y-l (site CI and CV) and 0.8 - 1.6 x 106 t y-l (site

CB) calculated by Koopmann (1981). He suggested an error of 25 to 50 % con1pared to sUl'face dust

input, depending partlyon the opal content, and partlyon the missing carbonate fraction in the

Ratmeyer ef al.: Lithogenic particle fluxes and grain size distributions... 83

sediment data (0 - 25 %). Due to a similar preparation of sampIes during this study, this may be true

for sediment trap data as weIl.

Skm

SAL

JVL?~4/

Trade winds

OceanSurface

1000m

Sediment

Canary Islands Cape Blanc

800 km

Cape Verde

1,.-------------,

800 km

I} Dala !rom Schülz el al. (1981), Prospero & Carlson (1972), Lepple (1975)2) Half cf dust input was assumed for the Cl site after Schütz et al. (1981)3} Dala from Koopman (1981), Samlhein el al. (1982)

Numbers are Mio t I year

Fig. 20: Sehematie model for the transport of lithogenie matter off NW Afriea, between 10° Nand 30° N. Numbers are 106 t y-l, and are ealculated for three equal-area subregimes: 300 400 km off eoast (Canary Islands), 500 - 600 km off eoast (Cape Blane, Cape Verde), eaehcovering 800 x 100 km2. Atmospherie patterns are presented using arrows for the SAL and thetrade wind layer. At sites CI and CV, trade winds eontribute most to the dust input at surfaeelevel during intermediate seasons and during winter. See text for further diseussion.

SUMMARY

Lithogenie particle fluxes and grain size distributions at three deep-sea sediment trap mooring sites

off NW Afriea were examined and eompared to the sedimentary reeord and to seasonal atmospherie

variations in the trade wind layer. Results show that clearly different situations regarding both the

atmospherie input of mineral dust as weIl as the downward transport through the water eolumn were

found for the different sites, although a11 sites belong to the same atmospherie and oeeanie eireulation

system. Lithogenie fluxes and grain size distributions were found to be not neeessarily eorrelated

between depth levels of several thousand meters distanee.

84

Two contrary situations were observed: comparable temporal flux patterns observed at site C:I

indicate an efficient and fast downward transport 01' particulate matter, while the lack of correlation in

grain sizes in combination with the significant increase 01' fluxes with depth suggests an add itional

partieJe input from distant sources through lateral adveetion. In eontrast, the situation found at site Cl3

indicates an almost non-disturbed veriical particle transport, which is reflected in similar grain size

patterns between depths. These results show that grain size distributions obtained from sediment trap

sam pIes can be used to estimate the significance of alteration during vertical transpali, and enable the

identification of a given particle population at different depth levels.

The comparison oftrap sampIes with surface sediments reflects, of course, the higher temporal anel

spatial variability of tluxes and grain sizes sinking through the water column, but still shmvs rather

similar accumulation rates. Lateral variations between sites were also found for the variabi Iity 01'

tluxes and grain sizes in comparison to atmospheric circulation patterns, with strongest coupl ing 01'

dust transpoli and marine particle flux observed at site CV during the first year of sampling. Here,

grain sizes and tluxes show a eontrary trend towards coarse grains and lower fluxes between winter

and summer season, retleeting the change of atmospheric dust transport from low altitude trade winds

to the high-level African Easterly Jet.

In terms of seasonality, the sites cr and CV show comparable seasonal trends of lithogenic particle

tlux and provenance of continental trade winds, while site CB obviously receives both higher ancl

more continuous tluxes throughout the year from the trade wind and SAL transported dust. Thus, the

significance of dust transport in the different air layers during seasons is highest for the CV site and

lowest for the CB site. Although our resltlts confinn many studies during the last 20 years regarding

the tlux of dust into the NE subtropical Atlantic Ocean, they also show that the transpoli of dust within

the different air layers has a significant impact on the distribution of different size fractions of

lithogenic particles in the water column.

Ratmeyer et al.: Lithogenic particlefluxes amI

ACKNOWLEDGEMENTS

size distributions... 85

We thank the masters anel the erews of the research vessels FS Polarstern anel FS Meteor for their

professional help eluring eleployment ancl recovery of the 11l00ring arrays. Special thanks are elue to G.

Kulll1, Alfreel Wegener Institute for Polar anel Marine Research, Bremerhaven, for introeluction on the

CIS-GALAI paJiicle analyzer, helpful eliscussions eluring analysis anel carefully reaeling the manuscript.

B. Diekmann gave many suggestions which greatly improveel the manuscript. We thank M. Breitzke

for lIsing the FRITSCI-I laser sizer anel provieling elata for the instrument comparison, as weIl as S.

Neuer and R. Davenport for helpful discussions. Special thanks are elue to U. Wyputta for trajectory

computing, and to G. Bergametti (Paris) for helpful comments and discussion. We fUJiher thank the

two anonymous reviewers, and V. Diekamp, R. Fröhlking and C. Riauelel for laboratory work. This

research was fundeel by the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 261 at

Bremen University, Contribution 125).

REFERENCES

Aharonson N., N. Karasikov, M. Roitberg and 1. Shamir (1986) GALAI-CIS- J - A novel approach toaerosol particle size analysis. Journal 01' Aerosol Science, 17, 530-536.

Allelredge A. L., T. C. Granata, C. C. Gotschalk anel T. D. Dickey (1990) The physical strength ofmarine snow and its implications for particle disaggregation in the ocean. Limnology andOceanography, 35, 1415-1428.

Alldredge A. L., U. Passow anel B. E. Logan (1993) The abundance and significance 01' a class oflarge, transparent organic pmiicles in the ocean. Deep-Sea Research, 40, 1131-1140.

Alldredge A. L. and M. W. Silver (1988) Characteristics, dynamics and significance 01' marine snow.Progress in Oceanography, 20, 41-82.

d'AJmeiela G. A. (1987) On the variability 01' desert aerosol radiative characteristics. Journal 0]Geophysical Research, 92D, 3017-1026.

d'Almeida G. A. and R. Jaenicke (1984) Sahara desert dust transpOJi to the south. Proc. 11th Int. Conf.on Atm. Aerosols, Condension and !ce Nuclei, 3-8 Sep., Budapest, Preprint Vol. T, pp. 185-190.

Asper V. L., S. Honjo and T. H. Orsi (1992) Distribution and transpOJi ofmarine snow aggregates inthe Panama Basin. Deep-Sea Research, 39, 939-952.

Baker E. T., H. B. Milburn and D. A. Tennant (1988) Field assessment 01' sediment trap efficiencyunder varying flow conditions. Journal 0]Marine Research, 46, 573-592.

Bayvel L. P. and A. R. Jones (1981) Electromagnetic scattering and its applications. Applied Science,London, 289 pp.

Bergametti G., L. Gomes, G. Coude-Gaussen, P. Rognon and M. N. Le Coustumer (1989) Africandust observed over Canary Tslands: Source-regions identification and transport pattern 1'01' somesummer situations. Journal 0] Geophysical Research, 94,14855-14864.

Berger W. H. alld G. Wefer (1990) Export production: seasonality and intermittancy, anelpaleoceanographic implications. Palaeogeography, Palaeoclimatology, Palaeoecology, 89, 245254.

86

Buat-Menarel P., J. Davies, E. Remouelaki, J. C. Miquel, G. Bergametti, C. E. Lambert, U. Etzat, C.Quetel, J. La Rosa anel S. W. Fovvler (1989) Non-steady-state biological removal of atmosphericparticles from Mediterranean surface waters. Nature, 340, 131-134.

Carlson T. N. and J. M. Prospero (1972) The large-scale movement ofSaharan air outbreaks over thcnOlihern equatorial Atlantic. Journal o[Applied Meteorology, 11,283-297.

Chiapello 1., G. Bergametti, L. Gomes, B. Chatenet, F. Dulac, 1. Pimenta ancl E. Santos Suares (1995)An additionallow layer transport of Sahelian and Saharan dust transport over the North-EasternTropical Atlantic. Geophysical Research Letters, 22, 3191-3194.

Coude-Gaussen G., P. Rognon, G. Bergametti, L. Gomes, B. Strauss,.I. M. Gross anel M. N. LeCoustumer (1987) Saharan dust on Fuerteventura Island (Canaries): Chem ical and m ineralogi ca Icharacteristics, air mass trajectories, and probable sources. Journal o[ Geophysical Research, 92,9753-9771.

Darwin C. (1846) An account ofthe fine dust which often falls on vessels in the Atlantic Ocean.Quarterly Journal ojthe Geological Society ojLondon, 2, 26-30.

Davenport R., S. Neuer, A. Hernandez-Guerra, M.-.I. Rueda, O.L1inas, G. Fischer anel G. Wefer(subm.) Seasonal anel interannual pigment concentrations in the Canary Islands region from CZCSdata and comparison with observations from ESTOC time-series station. submitted to the Journaljor Remote Sensing.

Deuser W. G. (1986) Seasonal and interannual variations in the deep-water particle fiuxes in theSargasso Sea and their relation to surface hydrography. Deep-Sea Research, 33, 225-246.

Dodge L. G. (1984) Calibration ofthe Malvern particle sizer. Applied Optics, 23, 2415-2419.

Dupont L., and C. O. C. Agwu (1991) Environmental control ofpollen grain distribution patterns inthe Gulf of Guinea anel offshore NW-Africa. Geologische Rundschau, 80, 567-589.

Duce R. A (1986) The impact of atmospheric nitrogen, phosphorous anel iron species on marinebiological productivity. In: The role ojair-sea exchange in geochel71ical cycling, P. Buat-Menard,editor, NATO ASI Series C, Vol. 185, Dordrecht, Reidel, pp. 497-529.

Duce R. A, P. S. Liss, 1. T. Merill, E. L. Atlas, P. Buat-Menard, B. B. Hicks, 1. M. Miller, J. M.Prospero, R. Arimoto, T. M. Church, W. Ellis, J. N. Galloway, L. Hansen, T. D. Jickels, A H.Knap, K. H. Reinhardt, B. Schneider, A Soudine, .I. J. Tokos, S. Tsunogai, R. Wollast and M.Zhou (199 I) The atmospheric input oftrace species to the world ocean. Global Biogeochel71icalCycles, 5, 193-259.

Eliassen A (1976) The trajectory model: a technica1 description. Norwegian Institute of Air Research,Kjeller, Norway.

Ehrenberg C. G. (1862) Erläuterungen eines neuen wirklichen Passatstaubes aus dem atlantischenDunkelmeere vom 29. Oci. 1861. Monatsberichte der Königlich Preussischen Akademie derWissenschaften, Berlin, 202-224,.

Fischer G., B. Donner, V. Ratmeyer, R. Davenport and G. Wefer (I 996a) Distinct year-to-yearparticle fiux variations off Cape Blanc during 1988- I99 I: Relation to d 180-deduced sea-surfacetemperatures and trade winds. Journal ojMarine Research, 54, 73-98.

Fischer G., G. Krause, S. Neuer and G. Wefer (l996b) ShOli-term sedimentation pulses recorded witha fiuorescence sensor and sediment traps in 900 m water depth in the Canary Basin. in press byLimnology and Oceanography.

Fischer G. and G. Wefer (1991) Sampling, preparation and analysis of marine particulate matter. In:The analysis and characterization ojmarine particles. D. C. Hurd, W. Spencer, editors, AGU,Washington. pp. 391-397.

Ratllleyer et al.: Lithogenic particle fluxes ami grain size distributions... 87

Gardllcr W. D. and M. J. Richardson (1992) Particle export and rcsuspension fluxes in thc westernNorth Atlantic. In: Deep-Seafood chains (md the global carbon cycle. G. T. Rowc and V. Pariente,cditors, Kluwer, Washington. pp. 339-364.

Gabric A. J., L. Garcia, L. Van Camp, L. Nykjaer, W. Eifler and W. Schrimpf (1993) Offshore cxportor shelf production in the Cape Blanc (Maurctania) filament as described from coastal zone colorscanner imagery. Journal ofGeophysical Research, 98, 4697-4712.

Gingeie F. (1991) Zur klimaabhängigen Bildung biogener und terrigener Sedimente und ihrerVeränderung durch die Frühdiagenese im zcntralen und östlichen Südatlantik. Berichte,Fachbereich Geowissenschaften, Univ. Bremen, Germany, 26, 202 pp.

Glaccum R. A. and J. M. Prospero (1980) Sahal'an aerosols over the tropical North Atlantic.Mineralogy. Marine Geology, 37, 295-321.

Grousset F., P. Buat-Menard, D. Boust, R.-C. Tian, S. Baudei, C. Pujohl, C. Vergnaud-Grazzini(1989) Temporal changes of aeolian Saharan input in the Cape Verde abyssal plain since the lastglacial period. Oceanologica Acta, 12, 177-185.

Haake B. and V. Ittekott (1990) Die wind getriebene "biologische Pumpe" und Kohlenstoffentzug imOzean. Naturwissenschaften, 77, 75-79.

Hagen E. and R. Schemainda (1984) Der Guineadom im ostatlantischen Stromsystem. Beiträge zurMeereskunde, 51, 5-27.

Honjo S., K. W. Doherty, Y. C. Agrawal and V. L. Asper (1984) Direct optical assessment of largeamorphous aggregates (marine snow) in the deep ocean. Deep-Sea Research, 31,67-76.

Honjo S. and K. W. Doheliy (1988) Large scale apeliure in time-series sediment traps: Design,objectives, construction and app1ication. Deep-Sea Research, 35, 135-149.

Honjo, S., D. W. Spencer and W. D. Gardner (1992) A sediment trap intercomparison study in thePanama Basin. Deep-Sea Research, 39,333-358.

Honjo S. (1993) From the Gobi to the bottom ofthe North Pacific. Oceanus, 35, 45-53.

Hooghiemstra H. and C. O. C. Agwu (1986) Distribution of palynomorphs in marine sediments: Arecord for seasonal wind patterns over NW Africa and adjacent Atlantic. Geologische Rundschau,75,81-95.

Hughes P. and E. D. Barton (1974) Stratification and water mass structure in the upwelling area offl10lihwest Africa in AprillMay 1969. Deep-Sea Research, 21, 611-628.

Ittekott V. (1991) Particle flux studies in the Indian Ocean. EOS, 72, 527-53 O.

Jackson G. A. (1990) A model for the formation of marine algal flocs by physical coagulationprocesses. Deep-Sea Research, 37, 1197-1211.

Jaenicke R. and L. Schütz (1967) Studien zur oberen Grenzgröße des natürlichen Aerosoles. Contrib.Atmos. Phys., 40, 129-143.

Jaenicke R. and L. Schütz (1978) Comprehensive study ofphysical and chemical properties ofthesurface aerosols in the Cape Verde Island region. Journal of Geophysical Research, 83, 35853599.

Jickels T. D., T. M. Church and W. G. Deuser (1987) A comparison ofatmospheric inputs and deepocean particle fluxes for the Sargasso Sea. Global Biogeochemical Cycles, 1,117-130.

Kalu A. E. (1979) The African dust plume: Its characteristics and propagation across West Africa inwinter. In : Sahal'an dust: mobilization, transport, deposition. C. Morales, editor, SCOPE 14,Wiley, New York, pp. 95-118.

Klein B. and G. Siedler (1989) On the origin ofthe Azores Current. Journal ofGeophysical Research,94,6159-6168.

88

Koopmann B. (1981) Die Sedimentation von Saharastaub im subtropischen Nordatlantik wiihrend derletzten 25.000 Jahre. Meteor Forschungs-Ergebnisse C, 35, 23-59.

Kremling K. and P. Streu (1993) Saharan dust influenced trace element fluxes in deep North Atlanticsubtropical waters. Deep-Sea Research, 40, 1155-1168.

Kranck K. and T. G. Mil1igan (1991) Grain size in oceanography. In: Principles, methods, emdapplication ofparticle size analysis. 1. P. M. Syvitski, editor, Cambridge university Press,Cambridge, pp. 332-345.

Lepp1e F. K. (1975) Eolian dust over the North Atlantic Ocean. (Ph.D. thesis), 270 pp., University ofDelaware, Newark.

Li X., H. Maring, D. Savoie, K. Voss and J. M. Prospero (1996) Dominance ofmineral dust in acrosollight-scattering in the NOlih Atlantic trade winds. Nature, 380, 416-419.

Lorenz, S., B. Grieger, P. Helbig and K. Helierich (1996) Investigating the sensitivity oftheAtmospheric General Circulation Model ECHAM 3 to paleoclimate boundary conditions.Geologische Rundschau, in press.

MacIntire S., A. L. Alldredge and C. C. Gotschalk (1995) Accumulation of marine snow at densitydiscontinuities in the water column. Limnology emd Oceanography, 40, 449-468.

Martin J. H. and S. E. Fitzwater (1988) Iron deficiency limits phytoplankton growth in the north-eastPacific subantarctic. Nature, 331, 341-343.

McCave 1. N., R. 1. Bryant, H. F. Cook and C. A. Coughanowr (1986) Evaluation of a laserdiffraction-size analyzer for use with natural sediments. Journal ofSedimentmy Petrology, 56,561-564

McCave 1. N., B. Manighetti and S. G. Robinson (1995) Sortable silt and fine sedimentsize/composition slicing: Parameters for palaeocurrent speed and paleoceanography.Palaeoceanography, 10, 593-610.

Milliman J. D. and J. P. M. Syvitski (1992) Geomorphic/Tectonic contraI of sediment discharge to theacean: The importance of small mountainous rivers. Journal ofGeology, 100, 520-540.

Mittelstaedt E. (1972) Der hydrographische Aufbau und die zeitliche Variabiltät der Schichtung undStrömung im nordwestafrikanischen Auftriebsgebiet im Frühjahr 1968. MeteorForschungsergebnisse, A, 11,1-57.

Mittelstaedt E. (1991) The ocean boundary along the nOlihwest African coast. Prog. Oceanography,26,307-355.

Neuer S., V. Ratmeyer, R. Davenpoli, G. Fischer and G. Wefer (1997) Deep water particle flux in theCanary Island region: Seasonal variation in relation to lang-term satellite derived pigment data andlateral sources. in press by Deep-Sea Research.

Petterssen S. (1956) Weather analysis and forecasting. McGraw-Hill, NewYork.

Prospero J. M. and T. N. Carlson (1972) Vertical and areal distribution of Saharan dust over thewestern equatorial Atlantic acean. Journal of Geophysical Research, 77, 5255-5265.

Prospero 1. M., R. A. G laccum and R. T. Nees (1981) Atmospheric transport of soil dust from Africato South America. Nature, 289, 570-572.

Prospero 1. M. (1990) Mineral-aerosol transport to the North Atlantic and North Pacific: the impact ofAfrican and Asian sources. In: The long range atmospheric transport ofnatural and contaminantsubstances. A. H. Knap, editor, Kluwer Acad.Publ., The Netherlands. pp. 59-86.

Quetel C. R., E. Remoudaki, J. E. Davies, J. Miquel, S. W. Fowler, C. E. Lambert, G. Bergametti andP. Buat-Menard (1993) Impact of atmospheric deposition on particulate iron flux and distributionin northwestern Mediterranean waters. Deep-Sea Research, 40, 989- 1002.

89Rotll1eyer et al.: Lithogenic particlefluxes (md grain size distrihutions...---------

Ramaswamy V., R. R. Nair, S. Manganini, B. Haake anel V. {ttckott (1991) Lithogenic fluxes to thec1cep Arabian Sea measurecl by sediment traps. Deep-Sea Research, 38, 169-184.

Ratmeyer V. and G. Wefer (1996) A high resolution camera systcm (Pm'Ca) for imaging particles inthe ocean: System design and results from profiles anel a three-month c1eployment. Journal ofMarine Research, 54,589-603.

RiebeseIl U. and D. A. Wolf-Gladow (1992) The relationship bctween physical aggregation ofphytoplankton anel particle flux: a numericalmodel. Deep-Sea Research, 39, 1085-1102.

Risovic D. (1993) Two-component model of sea particle size distribution. Deep-Sea Research, 40,1459-1473.

Russel W. B. (1980) Review ofthe role of colloidal for'ces in the rheology of suspensions. Journal ofRheology, 24, 287-317.

Sarnthein M., 1. Thiede, U. Pflaumann, H. Erlenkeuser, D. Fütterer, B. Koopmann, H. Lange anel E.Seibold (1982) Atmospheric and oceanic circulation patterns off Northwest Africa during the past25 million years. In: Geology ofthe Northwest African continentalmargin. U. Rad, K. Hinz, M.Sarnthein, E. Seibold, eelitors, Springer, Berlin, pp.545-604.

Schütz L., R. Jaenicke anel H. Pietrek (1981) Saharan elust transport over the North Atlantic ocean. In:Desert dust: Origin, characteristics and effects on man. T. L. Pewe, editor, Geol. Soc. of Amer.,spec. paper, 186, 87-100.

Schütz L. anel M. Sebeli (1987) Mineral aerosol anel source ielentification. Journal ofAerosol Science,18, 1-10.

Schütz, L., J. M. Prospero, P. Buat-Menard, R. Harris, R. A. C. Carvalho, N. Z. Heidam, A. Cruzadoanel R. Jaenicke (1990) The long range transport of mineral aerosols: Group Repoli. In: The longrange atmospheric transport ofnatural and contaminant substances. A. H. Knap, editor, KluwerAcad. Pub!., The Netherlands. pp. 197-229.

Semmelhack W. (1934) Die Staubfalle im nordwest-afrikanischen Gebiet des Atlantischen Ozeans.Annalen der Hydrographie, 62, 273-277.

Siegel D. A., T. C. Granata, A. F. Michels and T. D. Dickey (1990) Mesoscale eddy diffusion, particlesinking, and the interpretation of sediment trap data. Journal ofGeophysical Research, 95, 53055311.

Singer, J., J. B. Anderson, M. T. Leelbetter, K. P. N. Jones, I. N. McCave and R. Wright (1988) Theassessment of analytical techniques for the analysis of fine-grained sediments. Journal ofSedimentary Fetrology, 60, 658-679.

Stolzenbach K. (1993) Scavenging of small partieles by fast-sinking porous aggregates. Deep-SeaResearch, 40, 359-369.

Stramma P. and G. Siedler (1988) Seasonal changes in the North Atlantic SubtropicaI gyre. Journal ofGeophysical Research, 93, 8111-8118.

Swap R., M. Garstang, S. Greco, R. Talbot and P. Kallberg (1992) Saharan dust in the Amazon Basin.Tellus,44B,133-149.

Tegen I., A. A. Lewis and I. Fung (1996) The influence on climate forcing of mineral aerosols fromdisturbed soils. Nature, 380, 419-422.

Tetzlaff G. and K. Wolter (1980) Meteorological patterns and the transport of mineral elust from theNorih African continent. Faleoecology ofAfrica, 12,31-42.

Van Camp L., L. Nykjaer, E. Mittelstaeelt and P. Schlittenharelt (1991) Upwelling and bounelarycirculation off nOlihwest Africa as depicted by infrared and visible satell ite observations. Frag.Oceanogr., 26, 357-402.

90

Walsh 1. D. and W. D. Gardner (1992) A comparison of aggregate profiles with sediment trap fluxcs.Deep-Sea Research, 39, 1817-1834.

Walther J. (1924) Das Gesetz der Wüstenbildung in Gegenwart und Vorzeit. 4. Edition, Quelle undMeyer, Leipzig, 421 pp.

Wefer G. (1989) Particle flux in the ocean: Effects of episodic production. In: Productivity in theocean: Present andpast. W. H. Berger, V. S. Smetacek ancl G. Wefer, editors, Wiley, New York.pp. 139-153.

Wefer G. and G. Fischer (1993) Seasonal patterns ofvertical particle flux in equatorial and coastalupwell ing areas of the eastern Atlantic. Deep-Sea Research, 40, 1613-1645.

Westphal D. L., O. B. Toon and T. N. Carlson (1987) A two-dimensional nUillerical investigation 01'the dynam ics and m icrophysics of Saharan dust storms. Journal ofGeophysical Research, 92D.3027-3049.

Wooster W. S., A. Bakun and D. R. McLain (1976) The seasonalupwelling cycle along the easternboundary of the North Atlantic. Journal 01Marine Research, 34, 131-141.

RatlJlcycr et al.: Impact 01mineral dust on deep-ocean particle jlux...

2.3 Manuskript 3:

Impact of mineral dust on deep-ocean particle flux in thc castern subtropical Atlantic Ocean

by V. Ratmeyer, W. Balzer, G. Bergametti, 1. Chiapello, G. Fischer and U. Wyputta

(zur Veröffentlichung eingereicht bei Geophysical Research Letters)

91

93Ratmeyer ct al.: Impact ofmincral dust on dccp-ocean partic1eflux...--------------

Impact of mineral dust on deep-ocean particle flux in the castern subtropical

Atlantic Occan

V. Ratmeyer, W. Balzer, G. Bergametti, I. Chiapello, G. Fischer and U. Wyputta

Atmospheric transport of mineral aerosol from continental sources has an essential impact on

both the global atmospheric heat budget l and the input of particles and trace elements into the

world ocean2-4

. In addition to the aeolian supply of nutrients and their effect on marine

biomass production5,6, scavenging of heavy dust grains by organic particles sufficiently

increases aggregate settling speed? and thus enhances the efficiency of the biological pumps.

While the physical link between the lower atmospheric and the upper ocean circulation is

widely understood9,1O, it is not known to what extent vertical fluxes of pariiculate matter in the

deep ocean reflect seasonal or short-term changes of the atmospheric dust input. Here we

report seasonallithogenic particle and Al fluxes from a deep-ocean sediment trap deployment

during 1992 and 1993 off NW Africa, concurrent with atmospheric Al concentrations and

backward trajectories from two barometric levels in the lower and mid troposphere, We find

that Al fluxes, lithogenic particle fluxes and grain size distributions in the eastern subtropical

Atlantic Ocean are linked to airmass pathways and surface mineral aerosol concentrations,

yielding close temporal coupling between atmospheric dust transpOli and deep-ocean pariicle

fluxes.

A strong correspondence of particle fluxes in the deep-sea with seasonal changes of surface

winds and biomass production in the upper ocean was found in many regions of the world

oceanl1,12. The downward transport of particulate organic carbon, namely the efficiency of the

biological pump, is highly variable and depends essentially on the settling velocities of

sinking paliiculate matter. They are to a large extent controlled by biologically triggered

aggregation processes like coagulation of phytoplankton cells l3, by zooplankton grazing, and

by scavenging of lithogenic grains by sinking organic particles. This has strong implications

for the atmospheric CO2 level14, for the formation of deep-sea sediments l5 and for the

reconstruction of paleoclimate l6. To understand the particle fluxes in the water column, and to

estimate the efficiency of the biological pump, a more detailed knowledge of transport

processes and controlling factors is essentials. Areas where high productivity coincide with

heavy atmospheric dust input, such as the eastern tropical Atlantic, are prominent sites for

94

studying how atmospherically controlled dust deposition affects the biological pump. Along

the northwest African coast lG (1 J0_300N) nutrient upwelling leacls to high biologieal

productivity, ancl the extensive year-rouncl atmospheric dust transport from the dry areas of

the coastal Morocco, Sahel and Sahara regions towarcls the subtropical anci tropical Atlantie li

18 contributes one thircl to half of the estimatecl 2 billion tons of global aeolian mineral dust

input to the worlcl ocean l9. The stucly area is situated within the first 1500 km off the Afriean

coast where most of the dust leaving the Saharan sources is assumecl to settle out.

Atmospheric Al-concentrations of the mineral aerosol (Fig.1 a) were measured between 1992

and 1994 in the surface layer of Sal Island, Cape Verde Archipelago and were found to exhibit

a strong seasonal pattern. Monthly averaged concentrations were highest from December

(4600 ng m,3) through January (15300 ng m,3), and decreased significantly during the summer

months (minimum in June: 290 ng m'3). The same seasonal change was recorded in sediment

trap sampIes at the CV site off Cape Verde between Gctober 1992 and Gctober 1993. Here,

changes of the particulate Al-flux at 1000 m water depth are similar to the pattern of

atmospheric Al-concentrations (Fig. 1b). High mean fluxes of up to 8.1 mg m,2 ci' I Al

(February 1993) prevailecl during winter and spring, while low mean fluxes of Al down to 0.8

mg m,2 d'] (August 1993) occured during summer. Concurrently, the Al fluxes show a

significant conelation (r2=0.79, n=58) with the lithogenic partic1e fluxes during the entire

sampling period (Fig. 1e). Fe/Al-ratios in trapped material varied from 0.43 (August 1993) to

0.58 (Getober 1992), similar to the range of atmospheric ratios between 0.41 - 0.63 from

January and September 1993.

In contrast to the Al fluxes, the flux of coarse lithogenic patiic1es in the size c1ass 20-63 Ilm

shows a reverse pattern (Fig. 1d). Coarse grains contributed generaIly less than 12 mg m,2 cl,l

(20-30 % ofthe measured fraction 6-63 Ilm) to the lithogenic flux during winter, but increased

significantly from April 1993 through July 1993 to fluxes between 15 and 32 mg m,2 d,l (30

55 %). Possible changes in the vertical transport from the ocean's surface to the trap depth,

however, do not provide a satisfying explanation. Grganic carbon and lithogenic partic1e

fluxes correlated weIl during both seasons (r2=0.65, n=28), indicating a rapid and efficient

biogenic partic1e transport during the entire deployment time. Instead, we suggest that the

seasonal pattern of coarse grain fluxes is mainly caused by the input of mineral dust from

Ratmcycr ct al.: Impact ofmincral dust on dccp-occan particleflux... 95

Season

15000

JAN FES MAR APR MAY JUN JULocr NOV DECo -l---}==~--,.---.----r---,--\==:r==4==~==='---J.--

5000

10000

a

0.0

120x::>-=(l)~ 90

Ut '0ro

ECl. 60u

OJ'e(l) EOJ - 300

oS::::;

0

40

12.0 1992 1993

"- - 8.0~ U

x "::> E

b-= OJ

::;: E 4.0

c

d

E::1.(') 30<!J0 ~

N Cuc

E0 20U 0>

~ E'0 - 10x::>Li:

2827o-t-'--------+------------------------{

Sampie 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1920 21 22 23 24 25 26

ocr

Fig. 1: (a) AI concentrations measured at surface level on Sal Island, Cape Verde Archipelago(l6°45'N; 22°57'W): monthly mean values are calculated from discontinuous daily dataobtained during 1992 and 1993. Dust sampies at Sal Island were obtained by bulk filtration onNuclepore@ 0.4 /1m pore size using a flow rate of about 1 m3 h- 1 (24h/sample). Aluminumparticulate concentrations in atmospheric dust were determined directly on the filter by X-Rayspectrometry fluorescence (Siemens SRS 303)20,25.(b) Aluminum fluxes measured in sedimenttraps at 1000 m water depth near the Cap Verde (CV: 11°29.0'N; 21°01.0'W). AI wasdetermined by AAS after subjecting the trapped particles to a microwave oven pressuredigestion using HF/HN03. (c) Lithogenic patiicle fluxes recorded in the sediment trap sampies.Prior to and after deployment, sampie cups were poisoned with mercuric chloride. After traprecovery, sampies were wet-sieved (l mm), splitted, treated and analyzed as describedelsewhere26. The lithogenic fraction was calculated as: Lith = Total flux - (opal flux +carbonate flux + 2 Corg flux). (d) Lithogenic particle fluxes of size class 20-63 /1m. For grainsize measurements on the Iithogenic fraction, 1/16 to 1/5 split of the sampie was treated withacetic acid and H20 2 to remove carbonate and organic matter. The remaining fraction wasneutralized with NHJ and wet sieved. None of the sampies contained a measurable amount oflithogenic grains> 63 /1m. Grain size analysis was performed on a CIS-GALAI laser particleanalyzer27. Percentages were normalized to the fraction 6-63 ~tm.

96

different atmospheric pathways, which reflect a seasonal change of atmospheric circul ation

patterns in this region. Atmospheric dust transport in the area is known to occur either in thc

trade winds below 2 km altitude (pressure levels araund 850 hPa)20 or in the Saharan Air

Layer (SAL)21 above the trade wind inversion (around 500 hPa). Among the different regions

all along the coast, the study area exhibits the strangest seasonal patterns for both atmospheric

transport layers (U. Wyputta, Bremen, unpubl. data). Strangly differing situations were found

during January and July 1993, when airmasses in both atmospheric layers followed opposite

directions towards the CV site at both times (Fig. 2).

Fig. 2: Baekward trajeetoriesto the CV site of ainnasspathways in the trade windlayer (850 hPa, solid line) andthe Sahm'an Air Layer (SAL)(500 hPa, dashed line) InJanuary (a) and July (b),1993. Transpoti of dust fromeontinental sourees duringwinter oeeurs primarily at lowaltitudes (a), while duringsummer dust is earried byhigh altitude winds within theSAL (b). Daily atmospheriebaekward trajeetories wereeomputed for site CV fromOetober 1992 throughOetober 1993 at 850 and 500hPa barometrie levels with aglobal 2-dimensional isobaricmodel. The preeursor of thismodel was developed at theNorwegian MeteorologicalInstitute (NMI)28. The modeleomputes 4-day isobariebaekward trajeetories fromzonal and meridional windeomponents29. Wind datawere obtained from theEuropean Centre for MediumRange Weather Foreeasts(ECMWF).

20' E0' 10'

- 850 hPa--- 500 hPa

10'20'30'

July 19931

Ia January 19931

50' W 40'

40'N

- 850 hPa--- 500 hPa

30'

20'

10'S '---__L-...L-_'-----__L- L--_---'__---'------'_----'

40' .---,---------:-,------,------r-r---.,---,..,..------,-,,-~,_.__,

N

30'

10'S

50' W 40' 30' 20' 10' 0' 10' 20' E

Rat1l1eyer et al.: Impact ofmineral dust on deep-ocean particleflux... 97

During intermediate times, a dominant increase of continental airmass trajectories at the 500

hPa level occured between winter/spring 1993 and summer 1993 (Fig. 3a). Since air masses

with continental provenance suggest possible dust transpOli, highest amounts of dust within

the SAL should be exported from March through October 1993 (Fig. 3a), coinciding with

relatively low dust transport in the trade winds (Fig. 3b). Frequent trade winds of continental

origins indicate that dust transport during winter is mainly routed through the lower

atmosphere (Fig. 3b).

a

b

Season winter / spring

1992 19931.0

0.5

0.0 -+-'---,----,c-'--------'-t---'--.------f--.J,.L---,----r--,------,------,--,----'

1.0

0.5

0.0 -t--L--,---,---,------,---L...i--,-----..,----,---,------,-----,--,.--J

c

20.0

10.0 +--------+-----------------------,

OCT

Fig. 3: Comparison of atmospherie trajeetories (a, b) and sediment trap (e) data from site CVbetween Getober 1992 and 1993. Frequeney index of eontinental backward trajeetories (a) atthe 500 hPa and (b) at the 850 hPa pressure level: 1.0 stands for eontinental provenanee andpossible dust transport during all days of a sampling period. Numbers were eounted fromdireetional plots, classifying sources as eontinental versus oeeanie. (c) Equivalent spheriealmean diameters of the fraetion 6-63 11m of lithogenic partieles trapped at 1000 m water depth ..

These major changes in atmospheric circulation have strong implications for the manne

lithogenic particle inventory. We assume that the intital grain size distribution of the dust

(dominated22 by particles between 1 and 10 /lm, but with significant admixture of grains< 40

/lm) is the same for both atmospheric transport layers. Consequently, low altitude atmospheric

98

transport leads to a significant fallout of coarsc grains during wintcr bcforc rcaching thc C:V

site. High altitude transport during summer causcs a di1Tcrcnt situation: at the samc dislancc

from land, only grains large and heavy enough to havc sufficient deposition vclocitics rcach

the ocean's surface. The resulting mean grain sizc of thc particle fraction entering thc occan i:,

shifted towards thc coarsc end, and fluxcs are lowcr duc to thc fact that only a small part 01'

thc mineral aerosol reach the ocean at thc CV sitc. A further rcduction of finc grains may

occur whcn low layer winds blowing from thc oppositc directions winnow out thc fine

material from the üllling fraction. Concurrent with the increase of continental winds at high

altitudes (around 500 hPa), mean grain sizcs in the trappcd lithogenic particles incrcase

bctween spring and summer 1993 (i'om 11 to 19 ~lm (Fig. 3c). Thc link betwccn atmospheric

transport and grain sizc is corroboratcd by the good corrclation of the percentage of Cl sizc

fraction in thc trapped lithogenic particles versus the frequency of continental winds at both

pressure levels: highly abundant fine grains (size fraction 6-11 ~lm) correspond to üequent

continental trade winds (Fig. 4a), while a high concentration of coarse grains (size fr8ction 20

63 ~tm) correlates with frequent continental winds in the SAL (Fig. 4b).

n -= 12

r 7;;:; 0 62

4.0 503.02.0

200 '--__L......__L......__L __--'-__

0.0 10

bE:l 60.0

70.0

J,;;;a.~ ~ 50.0oe'" '"M.~aa;C <Il 40.0 -~.~o:§Ci 30.0;F

n::; 12

r7

~ 0.59 I

4.0 5.03.02.0

0.0 ,~~---,

a500 -----

0.0 1.0

E::t 40.0

'"<D

~-;g-1II b 30.0"Öc~ E'Vif)

~ ~ 20.0a; .S8'~15 10.0;F

Occurence of almospheric trajectoriesWiUl contlnenlal provenance at 500 hPa

(weeks f mon1h)

Occurence of atmospheric trajectorieswilh continental provenance at 850 hPa

(weeks I month)

Fig. 4: (a) Linear regression between the abundance 01' lithogenic grains in the size class 20-63fll11 1'1'0111 1000 111 wate I' depth (plotted as % of the 1'raction 6-63 fll11) and the Illonth Iy occurence01' continental winds within the 500 hPa pressure level (plotted as weeks per Illonth in 1993) (b)Linear regression between the abundance 01' lithogenic grains in the size class 6-1 1 fll11 1'rolll1000 III water depth (plotted as % 01' the 1'raction 6-63 fll11) and the Illonthly occurence 01'continental winds within the 850 hPa pressure level (plotted as weeks per Illonth in 1993).

We conclude that two general atmospheric transport situations are reflected in the lithogcnic

particle signal recorded in the deep ocean at the CV site:

Ratllleyer ef a1.: Impact ofmincra1 dust 011 deep-ocean particlef1ux... 99

a) During winter and spring, dust is mainly transported in continental trade winds and causes

high atmospheric Al and dust concentrations at the surface level. Dust fallout leads to a high

input of lithogenic particles to the upper ocean and high Al and lithogenic particle fluxes at

1000 m water depth. The flux of coarse grains and the resulting mean diameters of the

lithogenic particle fraction are low at this time of the year.

b) During summer, continental winds at high altitudes in the SAL are the main agent of dust

transport. Only a small fraction of dust particles settles out from the transport layer bringing

about low Al and dust concentrations at surface level. Consequently, lower Al and lithogenic

particle fluxes are observed in the trap concurrent with higher fluxes of coarse grains and

larger mean diameters as compared to the winter situation.

Our results imply that different fractions of dust enter the pelagic ocean depending on the

altitude level in which atmospheric transport occurs. From continental sources, proximal and

distant ocean regions within the transport path receive different fractions of dust, not only

depending on the distance from the source but also on the seasonal change of the atmospheric

circulation. This may have implications for the interpretation of deep-sea particle fluxes and

the sedimentary record l5. Because Saharan dust has been traced over the entire tropieal

Atlantic Ocean finally reaching the Caribbean18 and the Amazon Basin23, it is likely that open

ocean pmiicle fluxes mayas weIl react to shOli-term dust fallout from atmospheric long range

transport. Regarding mass fluxes and sediment accumulation rates, proximal deep ocean areas

receive a lügher impact of low altitude, short range-transpOlied dust, probably concurrent with

a higher temporal variability and shorter response time compared to the impact of the fraction

transported at high altitude. If an atmospheric signal like the lithogenic grain size spectrum is

transported to greater water depths, changes of grain size and element distributions should

allow a more accurate reconstruction of atmospheric transport patterns and climatic changes

during the last glacial and interglacial periods.

Given the strong coupling of atmospheric circulation, dust transport and particle fluxes in the

high productivity areas of the eastern subtropical Atlantic Ocean, our results indicate that a

significant pOliion of lithogenic particles is rapidly exported from the ocean's surface through

biogenic downward transport7• Although we could show the seasonal effect of different

atmospheric transport pathways on the deep-ocean lithogenic particle fluxes, more data are

needed to estimate the influence of different size classes of dust on the export of organic

carbon from the surface ocean. In addition to the recently proposed impact of dust on radiative

100

foreing and atmospherie light seattering14, the dust may have a signifieant effeet on global

climate by indireetly enhaneing the effieieney of the oeeanie biologieal pump. It may be 01'

even more signifieanee for today' s climate beeause a substantial part of the reccnt

atll10spherie dust load is suggested to be of anthropogenie origin from disturbed soils by

extensive land use in semi-arid regions l.

RatJlleyer et al.: Impact ofmineral dust on deep-ocean particleflux...

REFERENCES

101

1. Tegen 1., A. A. Lewis and 1. Fung (1996) The influence on climate forcing of mineral aerosolsfrom disturbed soils. Nature, 380, 419-422.

2. Duce R. A., P. S. Liss, J. T Merill, E. L. Atlas, P. Buat-Menard, B. B. Hicks, J. M. Miller, J. M.Prospero, R. Arimoto, T. M. Church, W. Ellis, J. N. Galloway, L. Hansen, T D. Jickels, A. H.Knap, K. H. Reinhardt, B. Schneider, A. Soudine, J. J. Tokos, S. Tsunogai, R. Wollast and M.Zhou (1991) The atmospheric input of trace species to the world ocean. Global BiogeochemicalCycles, 5, 193-259.

3. Bergametti G., L. Gomes, E. Remoudaki, M. Desbois, D. Martin and P. Buat-Menard (J989)Present transpoli and deposition patterns of African dust to the Northwestern Mediterraneall. In:Paleaoclimatology and paleometereology; modern and past patterns ofglobal atmospherictransports, M. Leinen and M. Sarnthein, eds., Kluwer, Dordrecht, pp. 227-252.

4. Buat-Menard P., J. Davies, E. Remoudaki, J. C. Miquel, G. Bergametti, C. E. Lambert, U. Etzat,C. Quetel, J. La Rosa and S. W. Fowler (1989) Non-steady-state biological removal ofatmospheric particles from Mediterranean surface waters. Nature, 340, 131-134.

5. Maliin J. H. and S. E. Fitzwater (1988) Iron deficiency limits phytoplankton growth in thenorth-east Pacific subantarctic. Nature, 331, 341-343.

6. Donaghay, P. L., P. S. Liss, R. A. Duce, D. R. Kester, A. K. Hanson, T Villareal, N. W.Tindale and D. J. Gifford (1991) The role of episodic atmospheric nutrient inputs in thechemical and biological dynal11ics of oceanic ecosystems. Oceanography, 4, 62-70.

7. Ramaswal11Y V., R. R. Nair, S. Manganini, B. Haake and V. Ittekott (1991) Lithogenic fluxes tothe deep Arabian Sea measured by sediment traps. Deep-Sea Research, 38, 169-184.

8. Fowler S. W. and G. A. Knauer (1986) Role of large particles in the transport of elements andorganic compounds through the oceanic water column. Prog. Oceanography, 16, 147-194.

9. Philander S. G. H. and R. C. Pacanowski(1980) The generation of equatorial currents. JournalofGeophysical Research, 85,1123-1136.

10. Mitte1staedt E. (1991) The ocean boundary along the northwest African coast. Prog.Oceanography, 26, 307-355.

11. Wefer G. and G. Fischer (1993) Seasonal patterns of vertical particle flux in equatorial andcoastal upwelling areas ofthe eastern Atlantic. Deep-Sea Research, 40, 1613-1645.

12. Deuser W. G. (1986) Seasonal and interannual variations in the deep-water particle fluxes in theSargasso Sea and their relation to surface hydrography. Deep-Sea Research, 33, 225-246.

13. Alldredge A. L. and G. A. Jackson (1995) Aggregation in marine systems. Deep-Sea Research11,42, 1-7.

14. Berger W. H. and R. S. Keil' (1984) Glacial-Holocene changes in atl110spheric C02 and thedeep-sea record. In: Climate Processes and climate sensitivity, Hansen, J. E and Takahashi, T,eds., AGU Geophysical Monograph 29,5,337-351.

15. Sarnthein M., J. Thiede, U. Pflaumann, H. Erlenkeuser, D. Fütterer, B. Koopmann, H. Langeand E. Seibold (1982) Attnospheric and oceanic circulation patterns off Northwest Africaduring the past 25 million years. In: Geology ofthe Northwest African continental margin. U.Rad, K. Hinz, M. Sarnthein, E. Seibold, editors, Springer, Berlin, pp.545-604.

16. Berger W. H., V. S. Sl11etacek and G. Wefer (1989) Ocean productivity and paleoproductivityan overview. In: Productivity ofthe Ocean: Present and Past, Berger, W. H., V. S. Smetacekand G. Wefer, eds., Wiley, pp. 1-34.

102

17. Schütz, L., J. M. Prospero, P. Buat-Menard, R. HaITis, R. A. C. Carvalho, N. Z. Heidal11, A.Cruzado and R. Jaenicke (1990) The long range transport of mineral aerosols: Grollp Reporl.In: The long range atmospheric transport ofnatural (md contaminant substances. A. 1-1. Knap.editor, Kluwer Acad. Pub!., The Netherlands. pp. 197-229.

18. Prospero J. M. (1990) Mineral-aerosol transport to the North Atlantic and North Pacific: thcimpact of African and Asian sources. In: The long range atmospheric transport ofnatural ({/l(lcontaminant substances. A. H. Knap, editor, Kluwer Acad.Pub!., The Netherlands. pp. 59-86.

19. d'Almeida G. A. and R. Jaenicke (1984) Sahara desert dust transport to the south. Proe. J 1th In1.Conf. on Atm. Aerosols, Condension and lee Nuelei, 3-8 Sep., Budapest, Preprint Vol. I, pp.185- J90.

20. Chiapello 1., G. Bergametti, L. Gomes, B. Chatenet, F. Dulac, J. Pimenta and E. Santos Suares(1995) An additional low layer transport of Sahelian and Saharan dust transport over the N orthEastern Tropical Atlantic. Geophysical Research Letter,I', 22, 3191-3194.

21. Semmelhack W. (1934) Die Staubfälle im nordwest-afrikanischen Gebiet des AtlantischenOzeans. Annalen der Hydrographie, 62, 273-277.

22. Schütz L. and M. Sebert (1987) Mineral aerosol and source identification. Journal ofAerosolScience, 18,1-10.

23. Swap R., M. Garstang, S. Greco, R. Talbot and P. Kallberg (1992) Sahat'an dust in the AmazonBasin. Tellus, 44B, 133-149.

24. Li X., H. Maring, D. Savoie, K. Voss and J. M. Prospero (1996) Dominance ofmineral dust inaerosollight-scattering in the NOlth Atlantic trade winds. Nature, 380, 416-419.

25. Losno et a!. (1987) Determination of optima conditions for atmospheric aerosol analyses by XRay flllorescence, Environ. Techno!. Lett., 8, 77-87.

26. Fischer G. and G. Wefer (1991) Sampling, preparation and analysis of marine particulatematter. In: The analysis and characterization ofmarine particles. D. C. Hurd, W. Spencer,editors, AGU, Washington. pp. 391-397.

27. Aharonson N., N. Karasikov, M. Roitberg and J. Shamir (1986) GALAI-CIS-l - A novelapproach to aerosol palticle size analysis. Journal of Aerosol Science, 17, 530-536.

28. Eliassen A. (1976) The trajectory model: a technical description. Norwegian Institute of AirResearch, Kjeller, Norway.

29. Petterssen S. (1956) Weather analysis and forecasting. McGraw-Hill, NewYork.

Neuer et a1.: Deep-tvater particle flux in the CanclIY [sland region...

2.4 Manuskript 4:

103

Decp-water particle flux in the Canary Island region: seasonal trends in relation to long-tenn

satcllite derived pigment data and lateral sources

by S. Neuer, V. Ratmeyer, R. Davenpoli, G. Fischer and G. Wefer

(im Druck bei Deep-Sea Research)

Inhalt Seite

Abstract 105

Introduction 105

Oceanographic setting 106

Methods , 108

Results 109

Currents 109

Seasonality 01 sw1ace pigment data 109

Seasonality 01deep-sea particle flux 11 0

Discussion 114

Conclusions 118

References 119

Neuer ef a1.: Deep-water particlejlux in fhe CanalY Jsland region... 105

Deep-water particle flux in the Canary Island region: seasonal trends in relation to long-term

satellite derived pigment data aIHllatcral sources

SUSa11l1e Neuer, Volker Ratmeyer, Robert Davenport, Gerhard Fischer and Gerold Wefer

ABSTRACT

We present a three-year record of deep-water particle flux at the recently initiated ESTOC time-series

station (European Station for Time-series in the Ocean, Canary lslands) locatecl in the eastern

subtropical NOlih Atlantic gyre. Particle flux was highly seasonal with flux maxima occurring in late

winter and early spring. A comparison with historic CZCS (Coastal Zone Color Scanner) data show

that these flux maxima occurred about one month after maximum chlorophyll was observed in surface

waters in a presumed primary source region located 100 km2 northeast of the trap location. The main

components of the paliicles collected with the traps were mineral particles and carbonate, both

correlating tightly with organic matter sedimentation. Mineral paliicles in the sinking matter are

indicative ofthe high aeolian input from the African desert regions.Comparing particle fluxes in I km

and 3 km depth, we find that paliicle sedimentation increased substantially with depth. Yearly organic

carbon sedimentation was 0.6 gm-2 at I km depth compared to 0.8 gnr2 at 3 km depth. We

hypothesize that higher phytoplankton biomass observed fUliher nOlih could be a source of laterally

advecting particles that interact with fast sinking particles originating from the primal'y source region.

This hypothesis is also suppolieel by the different size distribution of lithogenic matter found in both

trap depths.

INTRODUCTION

Deep water paliicle flux is closely linked to upper water biogeochemistry. It does, however, not only

reflect the balance between production and loss rates in the euphotic zone (e.g. Michaels and Silver,

1988; Peineli et al. 1989, Newton et al., 1994) but is also influenced by midwater processes such as

remineralisation, (dis-) aggregation and lateral advection (e.g. Banse, 1990, Asper et al., 1992b;

Wefer, 1993). '1'0 understand seasonal and interannual variability of particle flux, and to distinguish

autochtonous from lateral sources, particle trap studies need to be combined with investigations of the

large scale anel temporal variability of relevant upper water column processes, for example the

seasonal development of phytoplankton biomass. Sateliite observations give a synoptic view of

processes in the upper tens of meters of the ocean at length scales commensurate with the source

regions of paliicle traps moored at depth (several tens to hundreds of square kilometers, Siegel et al.

1990). These observations combined with particle flux studies can be used to elaborate on the

106

temporal and spatial eoupling ofthe phytoplankton development with particle flux at c!epth (Deuser ct

al. 1990).

Particle stuc!ies at the ESTOC @uropean ~tation far Iime-series in the Ocean, Canary Islancls)-stat ion

locatecl 100 km north of Gran Canaria startec! in late 199 I, anc! the month ly sam pling sched ule 01'

hydrographic and biogeochemical parameters was initiated at the station in 1994 (L1inas ct al., ] 9()/j).

The purpose of this European time-series station is it to understand the seasonality of long-term

changes in biogeochemical parameters, particle fluxes and current regimes on the castern side of the

North Atlantie subtropieal gyre.

In this study we present a three year reeord of deep water partiele flux investigated at the ESTOC

station. We compare the sesonality of particle flux to the seasonality of long-term pigment data

derived from historie CZCS observations in the region. Due to the particular setting of ESTOC in a

region of high ae01ian dust input and in proximity to an upwelling mal'gin, special attention is given to

the role of lithogenie material eollected by the partiele traps. We find high seasonality in particlc

fluxes at this station, closely linked to the end of the prominent winter bloom signal observed in

historie CZCS images. The dominant flux eomponents were carbonate and mineral particJes whieh

were closely related to the sedimentation of organic matter. A notable inerease of particie

sedimentation with depth indicates the importance of laterally advected paliiculate matter at this site.

OCEANOGRAPHIC SETTING

The Canary Islands are situated 100 to 600 km west ofthe NW African coast in the eastern extensions

of the subtropieal North Atlantic gyre at a latitude of 27-28° N. The Canary Islands region is part of

the recirculation regime linking the Gulf stream with the nOlih equatorial current via the Azores and

Canary currents. One branch ofthe Azares current flows to the west of the Canary Islands and another

branch turns southward east of Madeira, crosses the Canary arehipelago and joins the Canary Current

flowing southward along the NW African coast. (Stramma and Siedler, 1988; Klein and Siedler,

1989). Along the coast, along-shore upwelling and related currents interact with the Canary Current.

Owing to the seasonal variability of the trade winds, upwelling along the NW African coast north of

25° N is strongest during the summer months (van Camp et aI., 1991). While the upwelling zone itself

is restricted to a coastal band of about 50-70 km (Mittelstaedt, 1986), filaments of upwelled water

have been observed to adveet several 100 km westward (Nykjaer, 1988). These filaments are created

in the current field of the Canary Current at specific sites along the coast and can be observed in

AVHRR (Advanced Very High Resolution Radiometer) satellite images because of their colder sea

surfaee temperature and in CZCS images because oftheir 11igher pigment concentrations compared to

ambient levels. For the Canary archipelago, Cape Ghir to the north and Cape Yubi and Cape Bojador

located east and south of the islands are especially important sites where upwelled water can be

Neuer et al.: Deep-water particle flux in the CanmJ! Island region... 107

advected seaward as filaments. Off Cap Ghir, prominent filaments of several 100 km length can be

observed mainly during fall after a time lag of about three month after the onset of strongest upwelling

(van Camp et al, 1991; Nykjaer and van Camp, 1994). The filaments off Cape Yubi and Cape Bojador

influence mainly the south of the islands (Hernanelez-Guerra et al., 1993). So far no evidencc is

provided that any of the filaments impact directly on the region north of the islands where ESTOC is

located (Oavenport et al., subm.).

The Canary islands are located in the path of long range atmospheric dust transport from the NW

African continent over the tropical Atlantic ocean towards the Caribbean Sea. Estimations of total

atmospheric dust export at the NW African coast between 100 anel 300N towarels the northern

equatorial Atlantic range from 480 x 106 tons y-I (Lepple 1975) to about 170 x 106 tons y-I (Schütz

et al. 1990). More than two thirds of the aerosol falls out within the first 1000 km, resulting in

sediment deposition rates of up to 20 cm per 1000 years within the first 250 km off the coast

(Mieldleton et al.; 1986). Throughout the year, aeolian dust transport at these latidudes over the

tropical Atlantic occurs either above the trade wind inversion within the Hannattan wind system, or by

strong traele winds blowing towards south-westerly directions along the coast. Ouring summer, patt of

the "Saharan Air Layer" carries dust northward from the Sahara and Sahel in an anticyclonic flow

along the African coast towards the Canary Islanels, preferably forming a trajectory at about 3 km

above sea level (Kalu, 1977; Tetzlaff and Wolter, 1980). Ouring winter and spring, strong NNE trade

winds are most responsable for dust transport from coastal source regions in Morocco and the eastern

Sahara (Chiapello et al., 1995). Mineral aerosol at surface level around the Canary Islands can reach

concentrations of 10 - 1000 /lg m-3 (Jickels et al., 1987), with highest peaks from spring to fall when

heavy dust outbreaks from the continent are most common (Bergametti et al., 1989).

Phytoplankton biomass in the Canary Island region is low during most of the year and exhibits a

strong seasonal cycle in both production and biomass (Braun et al., 1985; Braun and Real, 1986; Oe

Le6n and Braun, 1973). The formation of a seasonal thermocline between 60 and 90 m, mostly

associated with a deep chlorophyll maximum, prevents the passage of nutrients into the euphotic zone

during most of the year. Ouring late fall and winter, mixing of the water column down to about 200 m

results in an increase of new nutrients in the mixed layer. Fernandez de Puelles and Garcfa-Braun

(1989) found that the homogenization of the upper 200 m coincided with the onset of the

phytoplankton production cycle. At this time during winter, highest rates of primaJ'y production were

observed, followed by an increase of microzooplankton biomass. This increase suggests that

phytoplankton biomass is grazed at a substantial rate. The annual rate of primary production was

determined to amount to 120 gm-2yr- 1 (Braun et al., 1982; Fernandez de Puelles and Garcia-Braun,

1989), which is higher than a large scale estimate of 60-90 gm-2yr- 1 (Berger et al., 1989). An increase

108

111 primary productivity can at times also be observed in fall due to vertical advection following

surface water cooling (off Tenerife; Fernill1dez de Puelles and Garcfa-Braun, 1989).

METHODS

Particle flux Particle flux was determined at the ESTOC Station (water depth 3600 m, Fig. 1) using

20 cup particle traps of the Kiel type (AQUATEC, opening 0.5 m2) in four mooring deployments with

varying length and sampling intervals (CI 1-4, Table 1).

/

/; ,~~f···-

, ESTOC~ .. i

(,;./i~':6P!~, 1."Canary Islands

Cape Yubi

Fig. 1: Mooring location at the ESTOC station and the presumed source region for partielescollected by the particle traps.

The sampling cups of the particle traps were poisoned prior and after deployment with mercury

cloride in a 40 psu density gradient (Suprapur NaCI). The > 1mm fraction was not further analysed

and mainly contained pteropod shells, insignificant amounts of amorphous aggregates and

occasionally fish rests. Sampies were omitted if during sampie preparation fish remains were detected.

The < 1mm size-fraction was analysed according to Fischer and Wefer (1991). Briefly, the freeze

dried sampies were weighted to determine total flux; carbonate and organic carbon were determined

from non-acidified and acidified sampies analysed in a CHN-analyser (Hereaus) and biogenic opal

was determined by the automated wet-Ieaching method (Müller and Schneider, 1993). The lithogenic

fraction was calculated as:

Lith = Total flux - (opal flux + carbonate flux + 2 Corg flux).

Neuer el al.: Deep-waler parlicleflux in Ihe Cal1wy Islamf region... 109

SampIes were not corrected for dissolution loss, which for biogenic opal was estimatecl previously to

amount to 1-7% on an annual basis (Fischer et al. 1996). However, as dissolution processes occur

rapidly within the cups they will affect all sampies equally (Fischer ct al. 1996).

Mooring Trap depth (m) Total collection period Deployment Sam pieIkm 3km Start End (Days) interval (Days)

CI 1 1006 3084 25.11.1991 25.09.1992 305 15CI2 1036 3067 01.10.1992 09.04.1993 190 10CI3 1026 3086 12.04.1993 07.06.1994 430 2\CI4 923 3070 09.06.1994 02.09.1994 86 8

Tab. 1: Trap depths, mooring intervals and collection periods of the particle trap deploymentsCI 1-4 at the ESTOC station. No data are available from the 1 km trap of Cl 3 because of lossof the flotation buoy.

CZCS-derived pigment data: Monthly composite images for the period 1979-1985 were analysed from

the OCEAN Level 3 Coastal Zone Color Scanner database for northwest Africa supplied by the Joint

Research Centre, Ispra, Italy (Davenpoli et al., subm.). Pixel resolution was equivalent to 1 km x 1

km. The most likely source region of material arriving in the 1 km traps was identified to be a 100

km2 area northeast ofthe ESTOC Station (Fig. 1). This source region was estimated assuming a mean

particle sinking rate of 100 m d- 1 and an average of the mesoscale current speed for the southward

flowing Canary current of 5-1 0 cm s-l (Molina and Laatzen 1986, Müller and Siedler 1992).

RESULTS

Currents: Current meter measurements (RCM 8 AANDERAA and INFLUX current meters, Krause

and Ohm, unpub1. manuscript) from depths 20 m below the sediment traps were available for CI 1-4

(upper traps) as weU as for CI3 and CI4 (Iower traps). At depths around 1000 m, velocities ranged

from 1 to 11 cm s-l, with directions varying between 10° and 170°. Slightly higher velocities were

recorded for the lower depths, with maxima of 15 cm s-l in November 1993 and oscillating directions

from NNE to SSW. Regarding the paliicle flux patterns and trap efficiencies we assume that

hydrodynamic biases were negligible due to low current velocities mostly below 10 cm s-l (Baker et

a1. 1988, Gust et a1. 1992) as weU as the lack ofcorrelation to any ofthe described flux parameters.

Seasonality of surface pigment data: Fig. 2 shows the phytoplankton pigment concentration in the

source box for the period 1979-1985. High pigment levels indicative of bloom situations were

observed each winter with maximum values in January with large interannual variability ranging from

0.05 mg m-3 in January 1982 to 0.19 mg m-3 in January 1980. The 10west anl1ual pigment

concentrations of about 0.03 mg m-3 were observed each year during the summer montlls. Only in

August/September 1980, unusual peak concentrations were observed reaching 0.06 mg m-3.

110

J M M J S N J M M J S N'J M M J S N' J M M J S N J M M J S N J M M J S N J M M J S N

Fig. 2: Monthly averaged chlorophyll values from 1979 until 1985 derived from CZCS data fürthe 100 km2 source region depicted in Fig. 1.

1985198419831982198119801979

(1)0.15

(J

.g::J<Il

roc0.~ 0

E~ 0> 0.10(1) E(Jc0(J

c(J)

E0>a::

0.05

0.00

0.20

Seasonality 01deep sea particleflux: As a complete record of particle fiux during the mooring period

is available only from the traps located at 3 km depth,we will first present them; lateron we will

compare these fiux data with those obtained in 1 km depth. Patiicle fiux at 3 km depth is highly

seasonal with the largest amount of particles collected at the end of February and beginnning of March

(Fig. 3a). Total fiux observed in late winter ranged from 118 mg m-2d- 1to a maximum of 188 mg m

2d- 1 (Fig. 3 and Table 2). During winter 1993, three distinct fiux maxima were measured, one al ready

in January . During the remaining months, particle flux was highly variable with occasional peaks

during late spring, summer and fall. The fiuxes remained, however, below the high values recorded

during late winter, and ranged between 49 mg m-2d- 1 and a maximum of 88 mg mg m-2d- 1 (August

1994 , Fig. 3, Table 2). The sedimentation of organic carbon, carbonate, biogenic opal and lithügenic

material followed the total patiiculate fiux pattern (Fig. 3 b, c, d, e). These biogenic and Iithogenic

fiux components peaked during the late winter as weil as during summer-fall sedimentation maxima.

Comparison 01particle flux in different depths Peak fluxes of total particulate matter determined at

1000m depth (Fig. 4) were for the most pati coincident with those found at 3 km depth (FIG. 3), albeit

lower in magnitude. While the late winter sedimentation event in 1993 was about 70 % in magnitude

compared to the reslIlts determined with the deep trap, fiuxes in spring of 1992 were almost one forth

Neucr cl a!.: Deep-waler parliele jlux in Ihe Canary Jsland regiol1. ..

A) Late winter/ spring1000 mYear 1992 1993 1994Peak a a b cSampIe interval Mar. II -26 Jan. 23- Feb.2 Feb.2 1-Mar.2 Mar. 2-12

Flux (mg m-2d- 1) 49 J28 J I 1 I J8C/N 8.0 8.4 8.5 7.6Opal/total 0.04 0.07 0.15 0.05Carb/total 0.51 0.55 0.49 0.45Lith/total 0.31 0.26 0.23 0.37

]000 mPeak a' a' b' c' a'SampIe interval Mar. J J-26 Feb.2-11 Feb.21-Mar.2 Mar. 2-12 Feb.ll.-Mar.4

Flux (mg m-2d- 1) J88 147 165 118 122CIN 9.0 8.9 8.6 9.J 10.2Opal/total 0.04 0.03 0.07 0.05 0.02Carb/total 0.52 0.51 0.52 0.46 0.60Lith/total 0.33 0.39 0.31 0.40 0.27

B) Summer/Fall1000 mYear 1992 1993 1994Peak b c d bSampIe interval May 26-Jun. J J Oct 20-30 June 4-25 Jun. 26-July 4

Flux (mg m-2d- 1) 25 34 31 49CIN 6.7 7.4 7.2 11.1Opal/total 0.03 0.03 0.01 0.04Carb/total 0.19 0.50 0.21 0.21Lith/total 0.65 0.35 0.65 0.54

3000 lJ1

Peak b' c' d' b'SampIe interval July 11-26 Oct.20-30 June 4-25 July 30-Aug. 7

Flux (mg m-2d- 1) 88 59 49 81CIN 12.2 10.3 9.4 10.5Opal/total 0.04 0.03 0.02 0.02Carb/total 0.33 0.47 0.56 0.49Lith/total 0.48 0.44 0.36 0.40

111

Tab. 2: Total fluxes and ratios ofthe flux maxima determined with the upper (1 km) and deeper

(3 km) paliicle traps in late winter spring and in summer/fall. See Figs. 3 and 4 for the

designation ofthe flux peaks.

of the fluxes recorded with the deep trap. Most sampling intervals during which peak fluxes were

recorded coincided in time. A time lag between upper and lower traps could, however, be observed for

the following peaks:( 1) in summer of 1992 (b and b' in Figs 3 a and 4, time-lag about one month), (2)

112

111 spring of 1993 (a and a', time-lag about 10 days) and (3) In summer of 1994 (b anel b', time-lag

about olle month).

A

3000 m

200

1992

a'b'

1993 1994

C 120

x 80" -'" .v u~ ':E0

E~Ü 40

B

-x •~ <-:'0

e> Eo '"ü E

o

~ ~;;; Ea. '"o E

12

12

b'

s

E80

~u: ;,0

.~ 'E 40

'"E5 -

1992 1993 1994

Fig. 3: Particle flux at the ESTOC station determined with the 3000 m particle traps. a. totalparticulates, b. organic carbon, c. carbonate, d. opal and e. lithogenic matter.

Neuer el Cl!.: Deep-wCller par{icleflux in {he Can(l}J! Jsland region. .. 113

The fraction of opal of total particle flux reached highest values during the late winter sedimentation

events (up to 0.15, Table 2) in both traps. Values were mllch lower (0.1 to 004) during summer.

Carbonate amounted to about half of the total flux, with some exceptions c1uring summer in the I km

trap when it was as low as 20% (Table 2). Lithogenic matter contribllted 35% to 65% of total flux

during the summer/fall peaks, compared to 23% to 40% in late winter/spring in both traps.In general,

lithogenic matter appears to contribllte more to total flux in the upper traps in late winter/spring, but

more in the lower traps in summer. These observations point to a highcr input of lithogcnic material to

the sinking flux in summer, combined with more scavenging in the deeper water layers possibly

caused by a 100ver sinking speed. CIN ratios ofthe organic matter in the upper traps were similar in all

seasons (7-8, with the exception of the summer peak in 1994) and reflect the ratios typically found for

phytoplankton detritus. The ratios of carbon to nitrogen were higher in the lower traps compared to the

1 km traps, but the difference was consistently greater in summer (Table 2).

Fig. 4: Total paliiculate flux at the ESTOC station determined with the 1000 m particle traps.

Table 3 shows correlation coefficients of the linear regressions of carbonate, biogenic opal and

lithogenic material with organic carbon. All components in both trap depths were tightly correlated

with organic carbon flux. The lowest correlation coefficient was found for opal, probably due to its

varying flux representation in the different seasons (Table 2). Flux of organic carbon explained more

than 70 % of the variability of the sedimentation of carbonate and lithogenic material (Table 3). The

close correlation of carbonate and lithogenic matter with organic carbon is also shown when

regressing carbonate with lithogenic material (Fig. 5a). While the correlation coefficient is higher for

the paliiculate matter collected in 3 km depth compared to 1 km depth, the slopes of both linear

relationships are almost equal. The linear regression of lithogenic matter versus opal, on the other

114

hand, has smaller correlation coefficients, and the slopes c1iffer by almost Cl f~lctor 01' two (Fig. Sb).

Again, the correlation is tighter in 3 km depth than in 1 km deptl1.

Depth lkm 3 km

CC<lrb 0.77 O.7SOpal 0.59 0.69Lith O.S3 0.74

Tab. 3: Correlation coefftcients 01' linear regression 01' carbonate (C carb), biogenic opal ancl

lithogenic material (Lith) vs flllX 01' organic carbon 1'01' particles collected at the two mooringdepths. n equals 47 1'01' 1 km (except opal where n = 43) ancl 77 1'01' 3 km regressions.

80.040.0

Lithogenic particle flux

(mg m"d')

00

20.0 ,-~.---_ _---.. --_ _]

~ 100Jm B I

15.0 I • =m Ir i

I I10.0 , I

r ,I5.0 l~

; .., I

I",,;-,;···i;.:.:·'. . J0.0 ~ ~, __ I~~__~ _

-~ u"" "ro E0. '"o E

80.040.0

Ulhogenic particle flux

(mg m'd')

100Jm

• 300Jm

40.0

120.0 ,

000 ~f

Fig. 5: (a) Linear regressions 01' carbonate versus lithogenic flux for particles collected with the1 km (open circles) and 3 km (solid circles) traps. Regression equations: 1 km: y = 1.27x ··1.39,1'2=0.57, n=47,' 3 km: y = 1.19x + 1.61,1'2=0.73, n=77.(b) Same 1'01' opal flux regressed onlithogenic flux. Regression equations: 1 km: y = 0.19x - 0.72,1'2=0.34,11=43,'3 km: y = 0.1lx - 0.55,1'2=0.64,11=77

DISCUSSION

Particle flux at the ESTOC station is highly seasonal, reflecting the seasonality 01' the phytoplankton

development as depicted in the long-term record 01' surface pigment data. A close correlation between

the maximum pigment values of the winter bloom in the Canary Islancls region and the particle flux

maxima at depth can be inferred from the seasonal patterns in both data sets in our study. The historie

CZCS images from 1979 until 1985 presented above show maximal pigment values in surface waters

in December and January which decline at the end of January and the beginning of February (Fig. 2).

ThllS, the end of the late winter spring bloom in surface waters occurs about one month before the

highest flux peaks are observed in the traps.

A coupling of surtace processes and sinking particle fluxes has been observed previously in a large

scale synoptic study comparing CZCS satellite and particle flllX c1ata in the Sargasso Sea (Deuser et

Neller et a!.: Deep-water particleflux in the Canary Jsland region. .. 115

al., 1990). There, a 1.5 1110nth lag betwecn enhanced pigment values in the surfacc in the source region

(a 200 km2 square northeast of the trap position) and arrival of particles in 3200 m depth was

observed. A close coupling of surface water processes and particle Jlux in the Sargasso Sea was also

observed by Asper et al. (1992 a) who found that primary productivity signals in the euphotic zone

were transferred rapidly to the deep sea and resulted in corresponding deep ocean /lux maxima. At thc

NABE (North Atlantic Bloom Experiment, 47° N, 20° W) site, maximum chlorophyll and primary

production preceded the par1icle Jlux peak determined in 3100 m depth by 25 to 35 days (Newton ct

al., 1994). Thus, the close coupling of seasonal phytoplankton maxima and particle Jlux peaks founcl

at the ESTOC site is characteristic of many areas in the Atlantic and is mediated by rapid transfer of

surface water production to the deep ocean.

In addition to the strong seasonality of particle /lux observed at the ESTOC site, observations with a

continuously recording Jluorometer (attached to an INFLUX current meter, Krause et al., unpubL

manuscript) have shown that pat1icle /lux during the winter sedimentation maximum at the ESTOC

station can be episodical, occurring at times in the time-range of days (Fischer et al., in press).

Furthermore, when studying the nature of the particulate matter in the traps, these authors found

coocolithophorids (coccoliths and coccospheres) to constitute a dominant pali ofthe pal1iculate matter

either contained in fecal pellets or in aggregates. The close correlation of organic carbon and

carbonate sedimentation in both trap depths confirms the important role of coccolithophorids as

primary producers and for the expOl1 /lux in the Canary lslands region during all seasons.ln contrast,

diatoms (which constituted the majority of the biogenic opal measured) were much more important

export producers during the bloom events in winter than during the rest of the year.

A comparison with a pal1icle /lux study conducted by Honjo and Manganini (1993) further north

along the 20° W meridian as pat1 ofNABE reveals striking differences in POC /lux and importance of

different flux constituents (Table 4) compared to the ESTOC site. First of all, yearly POC /lux

determined in 1 km depth successively increases towards northern latitudes, from twice the ESTOC

integrated yearly /lux ofO.5 gm-2 at the Azores front (at 34°N, 1 gm-2) to thrice the ESTOC value at

48° N (1.5 gm-2). Deeper, the difference of the POC flux becomes sma11er, also because of the

increased particle /lux measured in 3 km depth at the ESTOC station (Table 4). Carbonate

sedimentation is also higher nOl1h of the ESTOC site but contributes a high proportion to total

paJ1iculate Jlux at a11 three sites, especia11y at 34° N. The importance of biogenic opal as constitutent

of the sinking particulate /lux also increases with increasing latitude. The most striking difference,

however, is the importance of lithogenic particle /lux. While in lügher latitudes the proportion of

lithogenic matter to the sinking /lux averaged over one year is between 12 and 22 %, it is between 40

to 46 % at the ESTOC station.

116

Depth (m)upper

lower

ESTOC 34°N 48°NPOC 0.50 1.00 1.48Carb 2.42 (0.37) 13.1 (0.68) 11.0 (55.3)Opal 0.19 (0.03) 1.64 (0.08) 3.5(0.18)Lith 3.04 (0.46) 2.7(0.14) 3.2 (0.16)

POC 0.84 0.86 1.00Carb 9.39 (0.51) 12.9 (0.61) 15.4 (0.59)Opal 0.59 (0.04) 2.0 (0.09) 5.6(0.21)Lith 7.25 (0.40) 4.6 (0.22) 3.2 (0.12)

Tab. 4: Comparison of yearly integrated fluxes (mg m-2 y-l) and fractions oftotal particle fl ux(in brackets) at the ESTOC station (29°07' N, 15°27' W, this study, from 1992 for 1000 111

"upper" - and mean of 1992 and 1993 for "Iower") with those determined at 34°N and 48°N(both 21°W, from 1989 and 1990, Honjo and Manganini 1993) in 1 km depth (upper) and in600 (ESTOC) 01' 700 m (34°N and 48°N) above bottom (Iower).

Partiele flux at the ESTOC-site is likely influenced by its particular setting elose to the NW African

upwelling margin and underneath the atmospheric dust plume originating from the desert regions of

NW Africa. The contribution of lithogenic matter to the pmiicle flux is high, probably influcncing

physical aggregation, pmiicle size and sinking speed of the particles (e.g. Buat-Menard 1989). The

settling of organic carbon in the water column was found to be elosely linked to lithogenic partiele

flux (Deuser et a\., 1983; Ramamswany et a\., 1991; Wefer and Fischer, 1993; Ratmeyer et al. subm.)

and scavenging by fast sinking pmiicles is probably the main vertical transpoli mechanism of the

smaller mineral size fraction (e.g. Kremling and Steu 1993).

The elose correlation to organic matter flux at the ESTOC station suggests that lithogenic particles are

either scavenged by aggregates 01' are incorporated into fecal pellets by feeding zooplankton, either in

the sm"face layer 01' during descend through the water column. These aggregates and fecal pellets

themselves probably contain coccolithophorid remains 01' even originate from coccolithophorid

aggregation in the surface layer (Fischer et a\., in press) causing the tight correlation of biogenic

carbonate with lithogenic matter. The fact that peak fluxes of all components show similar 01' equal

timing in 1 km and 3 km depth indicates that partieles sank rapidly through the water column, in the

order of 100 md-I. In contrast to the elose correlation of the flux data, comparing the grain sizes of the

silt fraction shows no correlation in the mean diameters between 1 km and 3 km (Ratmeyer et al.

subm) at the ESTOC station. Although the range of grain sizes is comparable in both depths (annual

mean grain diameters between 13.5 and 14.9 Ilm), the seasonal variation of grain sizes differs and the

grain size spectrum of the lower traps appears to be less sorted. Additionally, high lithogenic fluxes

recorded in the lower traps in spring 1993, spring 1994 and fall 1994 are linked to a light coarsening

of grain sizes. Ratmeyer et a\. discuss the significance of correlation as a measure fo the alteration of

the sinking partieles. According to these authors, weil correlated grain size spectra indicate little

alteration of the sinking material by disaggregation 01' lateral advection. This case was observed off

Neuer et a1.: Deep-lvater particleflux in the Canary lsland region... 117

Cape Blane. The pOOl' correlation of grain sizes obscrved at the ESTOC-site, however, points to an

input of different material in the two trap depths 01' an alteration by dilution 01' addition of certain size

fractions during sinking. Since particle fluxes were consistently higher in the 3 km traps, it appears

1ikely that partiClllate matter was advected laterally into the deeper trap from regions of higher export

procluction.

According to Siegel et al (1990), the source area of a trap at 3 km depth should be restrictecl to about

100 km 2 for paliicles with a sinking speed of about 100 md-I. Particles with a slower sinking rate, i.e.

slowly sinking aggregates, however, would have length sc<des of several hundreds of kilometers.

Higher phytoplankton biomass can be found further north, upstream of the ESTOC station along the

NW African coast (Davenport et al. subm.). A possible source region of enhanced phytoplankton

biomass is located off Cap Ghir where prominent filaments have been observed to advect several 100

km seaward. These fi laments do not reach into the 100 km2 source region depieted in Fig. I but come

as close as 400 - 500 km to the ESTOC-station (Davenport et al. subm.). Outside of the primaI'y

upwelling zone of Cap Ghir, CZCS pigment data show that the phytoplankton winter bloom is by far

the largest signal during the year with pigment values about twice those observed during the ESTOC

winter bloom (Davenpoli et al. subm.). During summer/fall, filament activity can increase ambient

chlorophyll to levels that about equal those ofthe ESTOC winter bloom values. We do not believe that

material from the primal'y upwelling zone could influence the ESTOC trap as the contribution of

biogenic opal to the sinking flux in the deep sea is not enlarged compared to the 1 km trap. However,

it is known that as upwelled water advects away from the coast, a succession from diatoms to

flagellates and cooclithophorids occurs (for the NW African upwelling zone, Richert 1975). Thus, the

particulate matter reaching the source region of the deep trap might originate from plankton

populations that are not very different from open ocean populations encountered in the ESTOC region.

This might explain the similar composition of the particulate material in both trap depths. Thus, in

addition to the autochtonous production leaving the primary source region of the CI trap as rapidly

sinking particles (with a low ClN-ratio), slower sinking aggregates (with a higher CIN-ratio) could be

advecting laterally from areas closer to the NW African coast, explaining both the higher amplitude of

the winter bloom sedimentation peaks and the f1ux peaks in summer/fall in the deep trap. This laterally

advected material could be interacting with the faster sinking paliicles and explain the coincidence of

particle flux peaks in both traps.

Increased particle flux with depth is also observed at other sites. An additional lateral component was

discussed to cause l1igher particle fluxes at depths in the Sargasso Sea (Asper et al. 1992a). In the

North Atlantic, Honjo and Manganini (1993) hypothesize that particles in shallower depth had not yet

aggregated sufficiently to have "gained enough vertical settling speed to enable them to be eaught by a

sediment trap". This view is also supported by Asper et al. (1992b). From observations in the Panama

118

basin the latter atlthors conclude that a distinction is nccessary between suspended and setlling

material in order to aCCOllI1t for "both rapid, seasonal sinking ane! also for lateral transport". Th us, in

addition to the time-varying export flux from the surface layer, aggregation with suspended maller

originating from closer to the NW African coast may explain increases of flllxes with depth at the

ESTOC site.

CONCLUSIONS

Results from particle flux stlle!ies at he ESTOC time-series station ine!icate that the seasonal signal of

upper water column processes is readily transferred to the deep ocean as seen in all biogenie anel

lithogenic flux components. The notable increase of pmticle flux with depth is likely caused by thc

interaction of fast sinking partieies originating from a primary source region close to the ESTOC

station with those advected laterally from closer to the NW African upwelling margin. This has

important implications for the fooe! supply to the benthos. In a distance of 400 km from the coast, the

deep-sea community receives more organic material than if it was e!ependent only on the export

production leaving the overlaying oligotrophie water column. Furthermore, the unusually high

contribution of lithogenic matter at this site likely influences particle distribution and sinking rates and

makes this an important site for the study of the coupling of atmospheric dust input with particle flux

processes. This research was funded by the Deutsche Forschungsgemeinschaft

(Sonderforschungsbereich 261 at Bremen University, Contribution 153).

ACKNOWLEGDEMENTS

We thank the captains and the crew of RV Meteor and Polarstern for their professional help in

deploying and recovering the mooring arrays as weil as G. Ruhland, K. Dehning and U. Rosiak for

technical assistance. We are indepted to V. Diekamp and K. Slickers for the laboratory analyses. G.

Krause provided the INFLUX current meter data. The comments of F. Garcia-Pichel improved on the

manuscript. We thank the two anonymous referees for improving comments.

Neuer et al.: Deep-water particle jlux in the Canary lI'land region...

REFERENCES

119

Asper, V. L., W. G. Deuser, G. A. Knauer and S. E. Lohrenz (1992 a) Rapid coupling of sinkingparticle fluxes between surface and deep ocean waters. Nature, 357,670-672,

Asper, V. L., S. Honjo and T. H. Orsi (1992) Distribution and transport of marine snow aggregates inthe Panama basin. Deep-Sea Research, 39, 939-952.

Baker, E. T., H. B. Milburn and D. A. Tennant (1988) Field assessment of sediment trap efficiencyunder varying flow conditions. Journal ofMarine Research, 46, 573-592.

Banse, K. (1990) New views on the degradation and disposition of organic particles as collected bysediment traps in the open sea. Deep-Sea Research, 37, 1177-1195.

Bergametti, G., L. Gomes, G. Coude-Gaussen, P. Rognon and M. N. L. Coustumer (1989) Africandust observed over Canary Islands: source-regions identification and transport pattern for somesummer situations. Journal ofGeophysical Research, 94, 14855-14864.

Buat-Menard, P., J. Davies, E. Remoudaki, J. C. Miquel, G. Bergametti, C. E. Lambert, U. Etzat, C.Quetel, J. L. Rosa and S. W. Fowler (1989) Non-steady-state biological removal of atmosphericparticles from Mediterranean surface waters. Nature, 340, 131-134.

Chiapello, 1., G. Bergametti, L. Gomes and B. Chatenet (1995) An additionallow layer transport ofSahelian and Saharan dust transport over the North-Eastern Tropical Atlantic. Geophys.Geophysical Research Letters, 22, 3191-3194.

Davenpoli, R., S. Neuer, A. Hernandez-Guerra, M.-J. Rueda, O. Llinas, G. Fischer and G. Wefer(subm.) Seasonal and interannual pigment concentrations in the Canary Islands region from CZCSdata and comparison with observations from the ESTOC time-series station. Journalfor RemoteSensing,

Deuser, W. G., F. E. Muller-Karger, R. H. Evans, O. B. Brown, W. E. Esaias and G. C. Feldman(1990) Surface-ocean color and deep-ocean carbon flux: how c10se a connection? Deep-SeaResearch, 37,1331-1343.

Deuser, W. G., E. H. Ross and R. F. Anderson (1981) Seasonality in the supply of sediment to thedeep Sargasso Sea and implications for the rapid transfer of matter to the dep ocean. Deep-SeaResearch, 28,495-505.

Fischer, G., B. Donner, V. Ratmeyer, R. Davenport and G. Wefer (1996) Distinct year-to-year partic1e

flux variations off Cap Blanc during 1988-1991: Relation to ö180-deduced sea-surfacetemperatures and trade winds. Journal ofMarine Research, 54, 73-98.

Fischer, G., G. Krause, S. Neuer and G. Wefer (in the press) Short-term sedimentation pulses recordedwith a fluorescence sensor and sediment traps in 900 m water depth in the Canary basin. Limnologyand Oceanography

Fischer, G. and G. Wefer (1991) Sampling, preparation and analysis ofmarine particulate matter. In:The Analysis and Characterisation ofMarine Particles . Hurd, D. c., Spencer, D. W., Editors,Geophysical Monographs, pp. 391-397.

Gust, G., R. H. Byrne, R. E. Bernstein, P. R. Betzer and W. Bowles (1992) Partic1e fluxes and movingfluids: experience from synchronous trap collections in the Sargasso Sea. Deep-Sea Research, 39,1071-1083.

Hern<lndez-Guerra, A., J. Arfstegui and L. Nykjaer (1993) Phytoplankton pigment patterns in theCanary Islands area as determined using Coastal Zone Colour Scanner Data. InternationalJournalfor Remote Sensing, 14, 1431-1437.

Honjo, S. and S. J. Manganini (1993) Annual biogenic partic1e fluxes to the interior ofthe NorthAtlantic Ocean; studied at 34°N 21 °W and 48°N 21 °W. Deep-Sea Research, 40,587-607.

120

.liekeis, T. D., T. M. Church and W. G. Deuser (1987) A comparison of atmosphcric inputs and clccp"ocean particle fluxes for the Sargasso Sea. Global Biogeochelll. Cycles, 1, I 17-130.

Kalu, A. E. (1979). The African dust plume: its characteristics and propagation across West Africa inwinter. In: Saharan dust: Illobilization, transport, deposition. C. Morales, Editor, .lohn Wiley, NewYork, pp.95-118.

Klein, B. and G. Siedler (1989) On the origin ofthe Azores current. Journal ofGeophysical Research.94, 6159-6168.

Kremling, K. and P. Streu (1993) SahaI'an dust influenced trace element fluxes in dcep North Atl anticsubtropical waters. Deep-Sea Research, 40, 1155-1168.

Lcpple, F. K. (1975) Eolian dust over thc North Atlantic Ocean. PhD Thesis, University ofDelaware,Newark, 270 pp.

L1inas, 0., A. R. d. Leon, G. Siedler and G. Wefer (1994) The ESTOC time series started operation.Intern.ationalWOCE Newsletter, 17,20.

Michaels, A. F. and M. W. Silver (1988) Primary production, sinking fluxes and thc microbial foodweb. Deep-Sea Research, 35,473-490.

Middleton, J. T., A. S. Goudie and G. L. Wells, Ed. (1986) The frequency and source areas of duststorms. In: Aeolian GeolllOlphology, W. G. Nickling, Editor, Allen and Urwin, New York, pp. 237259.

Mittelstead, E.(1986) Upwelling regions. In: Oceanography, Landholt-Börnstein New Series, ].Sündermann, Editor, Springer Verlag, New York, pp. 135-163.

Molina, R. and F. L. Laatzen (1986) Hidrologia en la region comprendida entre las Islas Canariasorientales, Marruecos y las Islas Madeira. Campana 'Norcanarisa 1'. Bol. Inst. Ocean. Esp., 3, 1-16.

Müller, P. and R. Schneider (1993) An automateel leaching method for the determination of opal insediments and particulate matter. Deep-Sea Research, 40, 425-444.

Müller, T. J. and G. Siedler (1992) Multi-year current time-series in the eastern NOith Atlantic Ocean.Journal ofMarine Research, 50,63-98.

Newton, P. P., R. S. Lampit, T. D. Jickells, P. King and C. Boutle (1994) Temporal and spatialvariability ofbiogenic particle fluxes during the JGOFS northeast Atlantic process studies at 47 N20 W. Deep-Sea Research, 41, 1617-1642.

Nykjaer, L. (1988) Remote sensing applied to the NOIihwest African upwelling area. Ph.D. Thesis,University ofCopenhagen, pp. 144.

Nykjaer, L. anel L. V. Camp (1994) Seasonal and interannual variability of coastal upwelling alongnorthwest Africa and Portugal from 1981 to 1991. Journal ofGeophysical Research, 99, 1419714207.

Peinert, R., B. v. Bodungen and V. S. Smetacek (1989) Food web structure and loss rate. In:Productivity ofthe ocean: present andpast . Berger, W. H., V.S. Smetacek and G. Wefer, Editors,Wiley, Chichester, pp. 35-48.

Ramaswamy, V., R. R. Nair, S. Manganini, B. Haake and V. Ittekot (1991) Lithogenic fluxes to thedeep Arabian sea measured by sediment traps. Deep-Sea. Research, 32, 363-368.

Ratmeyer, V., G. Fischer and G. Wefer (subm.) Lithogenic particle fluxes and grain size distributionsin the deep ocean offNW Africa: implications for seasonal changes of aeolian dust transport andmass fluxes.

Richert, P. (1975) Die räumliche Verteilung und zeitliche Entwicklung des Phytoplanktons mitbesonderer Beruucksichtigung der Diatomeen, im N.W. afrikanischen Auftriebsgebiet. PhDThesis, Universität Kiel, 141 pp.

Neuer el al.: Deep-waler parlicle flux in the CanCllY IYland region... 121

Schütz, L., R. Jaenicke and H. Pietrek (1981) Sahar'an dust transport over the North Atlantic acean.Geol. SOG. ojAmer., 186, 87-100.

Siegel, D. A., T. C. Granata, A. F. Michaels and 1'. D. Dickey (1990) Mesoscale eddy diffusion,particle sinking, and the interpretation of sediment trap data. Journal ojGeophysical Research. 95,5305-5311.

Stramma, L. and G. Siedler (1988) SeasonaI changes in the North Atlantic Subtropical gyre. JournalofGeophysical Research, 93,8111-8118.

Tetzlaff, G. and K. Wolter (1980) Metereological patterns and the transpOli of mineral dust from theNorth African continent. Paleoecology ojAjrica. 12, 31-42.

Van Camp, L., L. Nykjaer, E. Mittelsteadt and P. Schlittenhard (1991) UpweIling and boundarycirculation off Northwest Africa as depicted by infrared and visible satellite observation. Progressin Oceanography, 26,357-402.

Wefer, G. (1993) Formation and composition ofmarine particulates. In: The Global Carbon Cycle.Heimann, M., Editor, Springer Verlag, Berlin Heidelberg, pp. 505-530.

Wefer, G. and G. Fischer (1993) SeasonaI patterns ofverticaJ particle flux in equatoriaI and coastaIupwelling areas ofthe eastern Atlantic. Deep-Sea Research, 40,1613-1645.

Ergell/lisse

3. El'gebnisse der einzelnen Al'beiten

Gegl icdert nach den vorgestclltcn Manuskriptcn wcrden nachfolgcnd dic Ergcbnissc dcr einzelncn

Arbeiten vorgestellt.

A high resolution call1era .Iystelll {ParCa}for illlagillg partie/es ill the ocean: 5)istelll design emd

resultsji-olll profiles and a three-lIIonth deployment

123

Währcnd des Einsatzes dcs Kamerasystcms auf zwei ausgesuchten Profilcn an dcn Stationen GeoB

2212 (04°00.4'S; 25°37'W) und Geol3 2217 (00000.5'N; 10049.8'W) im tropischen Ost- und

Weslatlantik bis 600 m Wasserticfe wurden ill-situ Aggregatverteilungen und -größen gemessen.

Höchstc Aggregatkonzentrationcn in beiden untcrsuchten Größenklassen (Partikel < 0.5 111m, und

"marine snow" > 0.5 mm) wurden dabei in den obersten 80 m gefundcn, jcdoch zeigten sich

signifikante Maxima auch in größeren Tiefen. Aus den zeitgleich mit dcm Intlux-Ströl11ungsmessers

gemesscnen Temperaturen und Chlorophyll-a-Gehalten geht eine Anhebung der Thermokline von

etwa 75 m auf 50 m zwischen westlicher und östlicher Station hervor, die sich ebenfalls in der

Konzentration vor allem größerer Aggregate widerspiegelt. Absolute Konzentrationen zwischen

beiden Stationen differieren etwa um den Faktor 2, mit höchsten Werten an der östl ichen Station im

Guinea Becken um 991 I-I für die gesamte Fraktion. Maxima der Aggregate in der Größenklasse

"marine snow" variieren zwischen 57 und J27 I-I zwischen beiden Stationen. Im Vergleich zu den

optischen Sensoren des Intlux-Strömungsmessers zeigen sich - bis auf die Übereinstimmung im

Bereich dcr Thermokline - keine korrelierbaren Muster. Obwohl große Schwankungen der

Aggrcgatkonzcntration über die gesamte Profiltiefe gemessen wurden, blieben die Signale für

Chlorophyll und Mie-Rückstreuung unterhalb der Thermokline nahezu gleich. Diese Abweichung der

Ergcbn issc zeigt die stark begrenztc Vergleichbarkeit beider Methoden im Hinbl ick au f die

Interprctation der Messungen für dic gesamte Partikel fracht im Ozean. Offensichtlich werden Partikel,

die innerhalb des Auflösungsbereiches der Kamera liegen, von den optischen Sensoren nicht erfaßt.

Diese jedoch lasscn Aussagen über die Trübe, also den Gehalt an feinsten, suspendierten Partikeln zu,

die ihrerseits von der Kamera nicht aufgenommen werden können, Ein wesentlicher Grund für diese

Unterschiede ist - neben der optischen Auflösung dcr Kamera - das vergleichsweise sehr kleine

Meßvolumen der optischen Sensoren. Da sich im offenen Ozean die Aufenthaltswahrscheinlichkeit

der Partikel in einem gegebenen Volumen umgekehrt zu deren Größe verh~i1t (Asper et al. 1992),

benötigt man ein bcstimmtes Mindestvolumen zur Erf~lssung größerer Aggrcgate. Da der größte Teil

des Vertikaltransportcs jedoch über solche Aggregate geschieht, wurdc das beleuchtete Volumen des

Kamerasystems zu Lasten der optischen Auflösung cntsprechend groß dimensionicrt. Es war daher

möglich, im parallelcn Einsatz zu Sinkstoff-Fallen gleichzeitig den Partikel fluß und dieill-sitl! Größen

124

der sinkenden und suspendierten Aggregate zu messen. Als wesentliches Ergebnis dieses verankerten

Einsatzes (Station CI4; 29°09.1'N; 15°26.5'W; 995 m WassCliiefe) zeigt sich eine deutl ich bessere

Vergleichbarkeit dieser Methoden gegenüber den optischen Sensoren. Zeitliche Variabiltiüt im

Partikelfluß und Änderung der Konzentration von Aggregaten> 0.5 mm korrelieren signifikant über

einen Zeitraum von etwa 3 Monaten. Chlorophyll und Rückstreuung zeigen dagegen auch hier keine

vergleichbaren Trends. Ein wichtiges Ergebnis dieses methodischen Vergleiches ist daher die

Bestätigung der Annahme, daß eine enge Beziehung zwischen Vorkommen größerer Partikel (> 0.5

111m) in der Wassersäule und dem Vertikaltransport besteht.

Manuskripte 2 und 3

Lithogenic particle fluxes emd grain size distributions offNW Africa: Implicationsfor seasonal

changes ofaeolian dust input emd down'Yvard transport

und

Impact ofmineral dust on deep-ocean particle flux in the eastern equatorial Atlantic Ocean

Sinkstoffproben von 3 Stationen (CI: 29°09'N; 15°26'W; CB: 21 0 09'N; 20 0 41'W; CV: 11°29'N;

21°04'W) vor NW Afrika wurden auf lithogene Partikelflüsse und Korngrößenverteilungen hin

untersucht und mit Sedimentakkul11ulationsraten aus der Literatur sowie atmosphärischen

Zirkulationsmustern und Staubmessungen verglichen. Ein signifikanter Einfluß atmosphärischen

Staubeintrages zeichnete sich in allen untersuchten Proben ab. Höchste Lithogenflüsse wurden im

Jahresmittel mit 31.2 - 56.1 mg m-2 d- l an der Station CB vor Cape Blanc gemessen. Geringste

Flüsse (Jahresmittel 11.4 - 21.2 mg m-2 d- l ) sowie kleinste mittlere Mean-Durchmesser (Jahresmittel

13.5 - 13.7 mg m-2 d- 1) wurden nördlich der Kanarischen Inseln an der Station CI gefunden. Höchste

Anteile an groben Körnern in der Fraktion 6-63 Jlm sowie größte mittlere Mean-Durchmesser

(Jahresmittel 14.5 - 16.9 flm) konnten an der Station CV vor Cape Verde nachgewiesen werden.

Bezüglich des Vertikaltransportes durch die Wassersäule wurden zwei grundsätzlich verschiedene

Situationen gefunden: Die deutliche Korrelation der Flußratenvariationen aus flachen (etwa 1000 m)

und tiefen (etwa 3600 m) Sinkstoff-Fallen and der Station CI zeugen von einem schnellen und

effektiven Vertikaltransport durch die gesamte Wassersäule. Nicht korrelierbaren Korngrößenspektren

zwischen beiden Tiefen deuten in Kombination mit einem signifikanten Anstieg der Flüsse mit der

Tiefe jedoch auf einen zusätzlichen lateralen Eintrag von Material mit höherem Anteil an gröberer

Fraktion, vermutlich durch lateral verdriftende und absinkende Filamente mit hoher Biomasse aus

kontinent-nahen Gebieten (siehe auch Manuskript 4). Im Gegensatz dazu wurden an der Station CB

gut korrelierende Flußraten in Kombination mit signifikant korrelierenden Korngrößen zwischen

beiden Fallentiefen (etwa 700 und 3500 m) gefunden, was auf einen weitgehend ungestörten

TranspOli ohne lateral advektiertes Material schließen läßt. Allerdings ist auch hier ein Anstieg der

Ergebnisse 125

Flüsse mit der Tiefe zu erkennen, was jedoch nicht im Widerspruch zu ungestörter Sedimentation

steht: Da die tieferen Fallen einen wesentlich größeren Einzugsbereich und damit eine höhere

Fängigkeit besitzen als die flachen Fallen (siehe auch Siegel et al. 1990), sollte gerade bei ungestötte,"

Sedimentation der Fluß mit der Tiefe zunehmen. Da die Korngrößenspektren zudem vergleichbar sind

und somit auf weitgehend gleiche Zusammensetzung des Materials in beiden Tiefen hinweisen, ist der

Eintrag von Material aus lateraler Advektion sehr unwahrscheinlich. Dieser Vergleich zeigt, daß in

Gebieten mit hohem Staubeintrag grundsätzliche Aussagen über den Vertikaltransport im Wasser

gemacht werden können, solange Korngrößen des eingetragenen Staubes signifikante Anteile an

Grobfraktion (> 10 ~lm) haben. Für distale Regionen ist die Verwendbarkeit von Korngrößen als

Transportindikator fraglich, da ab einer Entfernung von etwa 1500 km von der Küste deutl iche

Unterschiede in der Zusammensetzung des nun sehr feinkörnigen Staubes auch bei großen

Transportweiten nicht mehr zu erwarten sind (Westphal et al. 1987). Die deutlichste Anbindung der

Lithogenfraktion an den atmosphärischen Eintrag besteht an der Station CV. Im Gegensatz zur Station

CB zeigt sich hier zwar kein ungestörter Vertikaltransport innerhalb der Wassersäule, der extreme

saisonale Wechsel im atmosphärischen Eintrag prägt jedoch deutlich die lithogenen Flußraten wie

auch das Korngrößenmuster in etwa 1000 m Wassertiefe zwischen Oktober 1992 und Oktober 1993.

Mit Hilfe täglich berechneter atmosphärischer Trajektorien aus den Druckniveaus 850 und 500 hPa

konnten die Windmuster mit kontinentaler Herkunft ausgezählt und als Maß für möglichen

Staubtranspali mit den Sinkstoffen verglichen werden. Es zeigt sich eine signifikante Korrelation

monatlicher Mittelwerte zwischen dem Anteil and Feinkornfraktion (6-1 0 ~m) und 850 hPa

Trajektorien, sowie zwischen Grobfraktion (20-63 ~m) und 500 hPa Trajektorien. Da die beiden

Druckniveaus als repräsentatives Niveau für das Passatstockwerk (850 hPa) und den Saharan Air

Layer (SAL, 500 hPa) gewählt wurden, ist ein Vergleich mit den saisonalen Mustern dieser für den

Staubtransport verantworlichen Luftmassen möglich. In den Sinkstoffen zeichnet sich der winterliche

Staubeintrag über den Passat mit hohen Flüssen an Feinkornfraktion ab, während geringere Flüsse und

grobe Korngrößen im Sommer über den SAL eingetragen werden. Dieses Muster findet seine

Bestätigung in Messungen auf der Kaperdischen Insel Sa1, wo höchste Staubkonzentrationen mit den

assoziielien Elementen Al und Fe in den Monaten Dezember bis Februar dokumentieli sind (siehe

auch Chiapello et al. 1995). Die bezüglich Staubeintrag und Vertikaltransport untersuchten Muster

wurden mit Literaturdaten zu Sedimentakkumulationsraten (Koopmann 1981; Sarnthein et al. 1982)

von Oberflächensedimenten sowie zu atmosphärischer Staubfracht (Schütz et al. 1981; Carlson und

Prospero 1972; Lepple 1975) im Arbeitsgebiet vor NW Afrika verglichen. Die Berechnung des

jährlichen Staubeintrages aus den mit den Sedimentfallen gemessenen Flüssen (in Teilgebieten von

80.000 km2 im Abstand von 300 bis 500 km vor der Küste) ergaben Werte von 0.7 bis 1.4 x 106 t y-1

und bestätigen sowohl die Dimension der im Modell berechneten Staubfracht wie auch die

Sedimentakkumulationsraten (Abb. 8).

126

Kanarische InselnCI

Cape BlancCB

Cape VerdeCB

800 km

11

17' N800 km

1.4

0.8 - 1.6"

Resuspension

11

24' N

[~::o.])::=======~ko~ntinu'~iegrligch~======~

800 km

5 km

1000 m

3000 m -i'---:t:===~

Sediment I 004 - 0.8" I

OzeanDeckschicht

SAL (500 hPa)

Passat (850 i)Pai

"Daten aus Schütz et al. (1981), Prospero & Carlson (1972), Lepple (1975)~l Die Hälfte des Eintrages wurde für Cl angenommen nach Schütz et al. (1981)~'Daten aus Koopman (1981), Sarnthein et al. (1982)

Zahlen in Mio t IJahr

Abb. 8: Schematisches Bilanzmodell zu Staubeintrag und Veliikaltransport vor NW Afrika (ausManuskript 2: Ratmeyer et al., subm.). Die Bilanzierung beruht auf gemessenenLithogenflüssen im Vergleich zu Literaturdaten aus Schütz et al. (1981)1,2, Prospero undCarlson (1972)1, Lepple (1975)1, Koopmann (1981)3 und Sarnthein et al. (1982)3. Die alsEintragsgebiete angenommenen Regionen haben eine Ausdehnung von jeweils 80.000 km2. DieZah1enwelie sind Transport- und Flußraten in Millionen Tonnen pro Jahr für das jeweiligeTei 1gebiet.

Manuskript 4

Deep-yvater particleflux in the Canary Island region: Seasonal variation in relation to long-term

satellite derived pigment data and lateral sources

Sinkstoffproben der Station CI nördlich der Kanarischen Inseln zeigen ell1e hohe saisonale

Variabiltität des GesamtpaIiikelflusses in beiden Fallentiefen (etwa 1000 und 3000 m), mit

Hauptmaxima (bis zu 188 mg m2 d- 1 in 3000 m Tiefe) im Spätwinter und Frühjahr sowie geringeren

Maxima (zwischen 49 und 88 mg m2 d- 1) im Sommer und Herbst. Neben den hohen Lithogenflüssen

bildet Karbonat den Hauptteil des Gesamtflusses, wobei der Gruppe der Coccolithophoriden die

wichtigste Rolle als Karbonatträger zukommt. Eine enge Kopplung aller Partialflüsse zwischen beiden

Wassertiefen spricht für biogenen Transpoli und ein schnelles Absinken des partikulären Materials

während der gesamten Probenahme zwischen 1991 und 1994, mit einem zeitlichen Verzug zwischen

etwa 10 Tagen und einem Monat. Die Gesamt- und Corg- Flüsse zeigen allerdings eine deutliche

Ergebnisse 127

historischen CZCS-Daten (Coastal Zone Color Scanner, 1979-85) zeigt ll1 den Sinkstoffen ell1e

Korrelation mit zeitlichen Verzug von etwa einem Monat zur maximalen PlanktonblUte im

Oberflächenwasser eines nordöstlich der CI Station liegenden Produktionsgebietes. Dieses Gebiet

wird daher als Hauptlieferant des gemessenen Corg- und Karbonatflusses angesehen. Höhere

Planktonbiomassen können aus Satellitendaten weiter nördlich, entlang der Afrikanischen Küste

ersehen werden. Die Beobachtung von Filamenten aus Hochproduktionsgebieten wie z.B. vor Cape

Ghir zeigte eine mehrere hundert Kilometer reichende westwärts-Drift bis zu einer Entfernung von

etwa 500 km an die Station CI. Besonders im Sommer und Herbst werden in diesen Filamenten

Produktionsmaxima erreicht, die an Amplitude denen des angenommenen Liefergebietes der CI

Station gleichkommen. Die Zunahme der FlUsse mit der Tiefe an der CI Station, die Unterschiede in

den Korngrößen des lithogenen Materials und die erhöhten ClN-Raten in den tiefen Fallen besonders

im Sommer sprechen fur einen zusätzlichen, lateral advektierten Eintrag aus entfernteren Gebieten. Da

nicht davon auszugehen ist, daß die Quelle dieses Materials unmittelbar in den küstennahen Regionen

zu suchen ist, bieten die erwähnten Filamente eine Erklärungsmöglichkeit für erhöhte Flüsse und

erhöhte C/N-Werte. Denkbar ist eine laterale Advektion nach Absinken der Filamente von der

Oberfläche, gekoppelt mit effektivem Vertikaltransport an der CI Station und Ausfallung des

suspendierten Materials durch schnell sinkende Aggregate zwischen den beiden Fallentiefen. Ein

solcher Prozeß kann die Unterschiede, aber auch das nahezu zeitgleiche Flußmuster in beiden

Fallentiefen erklären.

128____________________________~__o •

4. Zusammenfassung

Untersuchungen zum partikuliircn Stofftransport im tropischcn und subtropischcn Atlantik wurden all

Sinkstoffproben und mit in-si/li Bcobachtungcn durchgcflihrt. Der wesentliche Teil dieser Arbeiten

bczog sich dabei auf dcn östlichcn subtropischcn Nordatlantik zwischen 10° und 3()o N vor deI

nordwestafrikan ischcn Küstc. Durch dic Nähe zum Klistenau rtricb und durch intcns iven aeo li scl1l'n

Staubtransport in dieser Region konnten für dcn partikulären Stofftransport im Ozean rcpräscntativc

Prozcsse nachgewiesen wcrdcn. Durch die Nutzung von atmosphiirischcn Modell rechnungen

(Z ieltrajektorien aus versch iedenen Druckn ivcaus) und Satell itendaten (P igm entvcrtei Iungcn) konnl CIl

wichtige Initialparameter des partikuliiren Stoffeintrags und -Iransports cinbezogcn und nil die

Interpretation der Ergcbnissc vcrwcndct werdcn. Daraus crgab sich ein zwischen den Kompartimenten

Atmosphäre, Ozcan und Sediment übergreifendes Bild, das sowohl eine Abschätzung der

Massenbilanzen terrigenen Materials ermöglicht als auch wichtige Prozesse des 'Transports ill den

Ozean und durch die Wassersäule aufzeigt und zu erklärcn vcrsucht. Dic nachfolgcnd

zusammengefaßten Ergebnissc sind an hand eines schematischen Modells mit den für die bearbeiteten

Stationen charakteristischen Transportmechanismen in Abbildung 9 dargestellt. Als Grundlage für

diese Darstellung dient das im Rahmen des Manuskriptes 2 erarbeitete Transportmodell aus dcm

Vergleich der lithogenen Partikelllußmessl1ngen mit Literaturdaten zu Staubeintrag und

Sedimentakkumulationsraten (Abb. 8).

Entsprechend der in Kapitel 1.2 formulierten Fragestellung lassen sich die Ergebnisse dieser Arbeit

wie folgt zusammenfassen:

• Die an den untersuchten Stationen CI, CS und CV gefundcnen saisonalen A1us/er im Partikelfluß

unterscheiden sich sowohl in ihrer jahreszeitlichen Zuordnung wie auch in der Abhängigkeit zu

Produktion im Oberflächenwasser und atmosphärischem Eintrag. Stärkste saisonale Muster im

Gesamtpartikeltluß wie auch in allen Partialtlüssen wurden an der Station Cl gefunden, mit

Hauptmaxima in Spätwinter und Frühjahr und geringeren Maxima vvährend Sommer und Herbst.

An dieser Station zeigte sich auch die engste Kopplung der FlUsse in beiden Ilntersuchten

Wassertieren (etwa 1000 und 3000 m). Der Verglcich mit der Saisonalität von historischen CZCS

Pigmentverteilungen ergab die Lage des wahrscheinlichen Liefergebietes für das saisonale Muster

verantwortl ichcn organ ischen Materials in einem etwa 100 km2 großen Areal nordöstl ich der

Station. Im Gegensatz zur Station CI ist bei den anderen untersuchtcn Stationen keine derart klarc

Korrclation dcr Flüsse in untcrschiedlichen Tiefen möglich. Geringste saisonale Variationen zeigen

im Vergleich die Flüsse dcr Station CS. Obwohl dic Amplitude der Flüsse in den oberen Fallen

(etwa 730 m) stark schwankt, ist eine striktjahreszeitliehe Zuordnung dcr Maxima wie bei CI nicht

mögl ich. Die bekanntcn und beobachteten Randbed ingungen bestätigen diese ß iId: der

Klistenuuftrieb erreicht sowohl seine höchste Amplitude als auch Kontinuität bei etw 20 0 N, und

Zusammenfassung 129

auch der Staubeintrag unterliegt den geringsten saisonalen Wechseln auf dieser Breite. Deutlich ist

jedoch eine interannuelle Variation, die sich neben den Variationen in den organischen

Partialflüssen auch Im Lithogenfluß zeigt und mit der Abkühlung der

Oberflächenwassertemperaturen korrelieli (Fischer et al. 1996). Die an der Station CV gemessenen

Flüsse zeigen eine recht hohe Saisonalität mit deutlichen Maxima des Lithogenflusses im Winter

und Frühjahr. Allerdings wurde hier eine extreme Abnahme des Gesamtflusses zwischen den

Jahren 1993 und 1994 in beiden Fallentiefen beobachtet, deren Ursache nicht geklärt ist. Eine enge

Anbindung ergibt sich jedoch an die Saisonalität der atmosphärischen Zirkulation und den damit

verbunden Staubeintrag (s.u.).

• Der Anteil an lithogener Fraktion in den Sinkstoffen ist an allen untersuchten Stationen signifikant

lind beträgt zwischen 23 und 78 % am Gesamtfluß. Höchste Lithogenflüsse wurden im Jahresmittel

mit 31.2 - 56.1 mg m-2 d-l an der Station CB vor Cape Blanc gemessen. Geringste Lithogenflüsse

(Jahresmittel 11.4 - 21.2 mg m-2 d- l ) wurden nördlich der Kanarischen Inseln an der Station CI

gefunden. Die Bedeutung dieses Anteils fur die Sedimentbildung ist als wesentlich einzuschätzen.

Die aus den Sinkstoffproben errechneten Werte für den jährlichen Staubeintrag bestätigen

überraschend gut die in der Literatur veröffentlichten Akkumulationsraten.

• Eine enge Anbindung der Lithogenflüsse an den atmosphärischen Staubtransport konnte vor allem

an der Station CV nachgewiesen werden. Hier zeigt sich der saisonale Wechsel des

StaubtranspOlies in Passat und "Saharan Air Layer" (SAL) sowohl in den Korngrößen wie auch in

den Lithogenflüssen in etwa 1000 m Wassertiefe. Hohe Flüsse korrelieren im Winter mit hohen

Anteilen an Feinfraktion (6-11 ~Lm) und dem Eintrag über den Passat (0-2 km Höhe), geringe

Flüsse spiegeln im Sommer den Eintrag einer groben Fraktion (20-63 ~m) aus dem SAL (3-6 km

Höhe) wieder. Aus monatlichen Mittelwerten kontinentaler atmosphärischer Trajektorien im

Passat-Niveau (um 850 hPa) konnten ebenfalls Vergleiche zwischen Lithogenfluß und

Staubeintrag an den beiden nördlichen Stationen durchgeführt werden. Es zeigt sich e1l1e

korrelierende Saisonalität beider Parameter fur beide Stationen, jedoch mit höherem Index für die

Station CI. Ähnlich der Situation bei CV spricht dies ebenfalls für einen bevorzugten Eintrag

lithogenen Materials aus dem Passat. An der Station CB zeigt sich die geringste Saisonalität

sowohl der Flüsse wie auch der Passatwinde kontinentaler Herkunft. Auch ist hier bekannt, daß der

Staubtransport über den SAL nahezu ganzjährlich verläuft. Beides spricht letztlich auch fur eine

Kopplung von Staubtranspoli und Partikelflüssen, auch wenn dies nicht an saisonal ausgeprägten

Mustern festzumachen ist.

• Die geringe Saisonalität und damit auch geringe Variabilität der Bildung und des Eintrages

paJiikulären Materials begünstigen die Station CB fur die Betrachtung des Partikeltransportes

innerhalb der Wassersäule. Obwohl keine sehr gute Kopplung der Flüsse aus beiden bearbeiteten

130

Wassertiefen vorliegt, zeigt sich ellle sehr gute Übereinstimmung der Korngrößen der

karbonatfreien Fraktion. Da der Parameter "Korngrößen" eine Verteilung innerhalb dcr

Lithogenfraktion darstellt, spricht eine gute Übereinsti ITImung für das Vorhandensein des gle ichcn

Materials in beiden Wasseliiefen, also auch für einen ungestölien Transport ohne laterale Zufuhr

advektielien Materials oder partielle Auflösung und selektive Freisetzung bestimmter Fraktioncn.

Im Gegensatz dazu spiegelt die Übereinstimmung der zeitlichen Flußmuster ohne Korrelation der

Korngrößen zwischen beiden Fallentiefen an der Station CI eine andere Situation wieder. Hier legt

die Zunahme der Flüsse mit der Tiefe und das Fehlen der Korngrößenkorrelation den Eintrag von

zusätzlichem, lateral advektiertem oder suspendieliem Material nahe. Durch das Ausfiltern lateral

eingeschichteten Materials - z.B. aus absinkenden ehemaligen Auftriebsfilamenten mit hoher

organischer Fracht - durch schnell sinkende Aggregate aus der Deckschicht ändert sich sowohl dic

Zusammensetzung des sinkenden Materials als auch sein Fluß.

• Für das Vorhandensein hoher Anteile suspendierten Materials an der Station CI sprechen auch die

Ergebnisse des methodischen Vergleichs zwischen Partikelkamera und Sinkstoff-Fallen. Die

zeitliche Korrelation der Sinkstoff-Flüsse mit den Konzentrationen an Aggregaten der

Größenklasse "marine snow" (> 0.5 mm) ist signifikant. Da mit der Falle nur die Flüsse, mit der

Kamera nur die Konzentrationen an Aggregaten gemessen wurden, ist der Vergleich nicht direkt

durchführbar. Unter Verwendung der Formeln von Alldredge und Gotschalk (1988) zur

Umrechnung von Aggregatgröße in Dichte und Gewicht, sowie unter Annahme einer

hypothetischen Sinkgeschwindigkeit von 100 m d- I , die etwa aus dem zeitlichen Versatz zwischen

Partikelsignalen zwischen oberer und unterer Falle geschätzt werden kann, ergeben sich

Unterschiede um den Faktor 50 bis 100 im errechneten Fluß zwischen beiden Geräten. Die

geringeren Flüsse wurden dabei mit der Falle gemessen. Obwohl die angenommene

Sinkgeschwindigkeit sicher keinen genauen Wert liefern kann, sind die Unterschiede sehr hoch.

Dieses Ergebnis bestätigt die in anderen Studien ermittelten Ergebnisse mit vergleichbaren

Diskrepanzen (z.B. Asper et al. 1992) und wirft die Frage auf, welche Bedeutung das nicht in den

Fallen gemessene Material für den Stoff11aushalt in der Wassersäule wie auch den Veliikaltransport

hat. Bezeichnend ist, daß an dieser Station in größeren Tiefen ein hoher Anteil suspendielien

Materials zu erhöhten Flüssen führt, die dann mit den Fallen gemessen werden können. Ungeklärt

bleibt jedoch, in welchen Tiefen dieses Material eingeschichtet wird, und wie lange es suspendieli

bleibt. Das Ergebnis zeigt, daß zum detaillierten Verständnis der TranspOlivorgänge in der

Wassersäule vor allem die Differenzierung verschiedener Aggregatpopulationen mit ihren

unterschiedlichen Charakteristika und Sinkverhalten in der Wassersäule notwendig ist. Dazu

erscheint die Anwendung und der Ausbau von zerstörungsfreien in-silu Methoden wie

Kamerasystemen und optischen Sensoren als ein möglicher Weg, vor allem im Vergleich mit

Verfahren wie der Probenahme mit Sinkstoff-Fallen und in-silu Pumpen.

Zusammenfassung

500 hPa

850hPa ~

1000m

3000m

CI .{

······<·sö·~mer

CV

Produktion A

131

Abb. 9: Schematisches Szenario der aus den Ergebnissen dieser Arbeit resultierendenTransportmechanismen des Staubeintrages und des Vertikaltransportes an den unterschiedlichenStationen CI, CB und CV.

CI: Ausgehend von einem Gebiet erhöhter Primarproduktion (Produktionsgebiet A) werden diePartikelflüsse in den oberen 1000 m von dem Export (A') aus dem Produktionsgebiet A genährt.Durch absinkende Auftriebsfilamente (Produktionsgebiet B) mit hohem Biomassegehalt wirdpartikuläre Substanz in der Tiefe lateral verdriftet (B'). Der bodennahe Partikelfluß enthältKomponenten aus beiden Quellen mit ihrer jeweiligen Staubfracht, jedoch durch sein höheresvertikales Einzugsgebiet (Siegel et aI. 1990) einen n-fachen Anteil von A' im Vergleich zurTiefe bei 1000 m (in der Summe nA' + B'). Der Staubeintrag ist geprägt von Passattransportiertem, feinkörnigen Material.

CB: Anders als bei CI gibt es keine Zufuhr von zusätzlichem Material aus anderen Quellen alsProduktionsgebiet A. Dementsprechend enthalten Flüsse durch die gesamte Wassersäule nur dieKomponente A'. Der Staubtransport verläuft ohne starke saisonale Wechsel in beidenStockwerken über das Jahr kontinuierlich.

CV: Die Situation bei CV ist stark vom saisonalen Wechsel des Staubeintrages aus beidenDruckniveaus geprägt. Größte Staubmengen werden im Winter durch den Passat eingetragenund prägen die Flüsse und Korngrößenverteilungen in den oberen 1000 m der Wassersäule. Mitgrößerer Tiefe verringert sich der Partikelfluß durch laterale Abfuhr oder partielleRemineralisation (A ''), sodaß in bodennahen Tiefen ein geringerer Fluß herrscht (in der SummeA'-A'').

132

5. Schlußfolgerungen und Ausblick

Die Analyse von Sinkstoffen im Hinblick auf Flußraten und Korngrößen der Lithogenfraktion konnte

zusammen mit in-sitll Beobachtungen zur Aggregatverteilung für die Interpretation von

Partikeitransportprozessen herangezogen werden. Dieser Ansatz erwies sich trotz ll1ethod ischer

Schwierigkeiten als erfolgreich. Er zeigte jedoch auch, daß zum detaillierten Verständins dieser

Prozesse vor allem die in-SÜll Messung bisher nur selten erfaßter Parameter notwendig ist, besonders

hinsichtlich des Vergleiches mit verschiedenen Methoden gewonnener Daten. So ist z.B. die

Sinkgeschwindigkeit einzelner Aggregate ein Schlüsselparameter, der das Verhalten unterschied Iicher

Aggregatpopulationen in der Wassersäule wesentlich beeinflußt. Die zur in-SÜll Erfassung von

Sinkgeschwindigkeiten notwendigen technischen Entwicklungen stellen sicherlich eine Schwiel"igkeit

dar, ihre Bedeutung für ein weitergehendes Verständnis des Stofftransportes im Ozean steht jedoch

außer Zweifel.

Die Bedeutung des windtransportierten Staubes für die Stofff1üsse im subtropisch-tropischen Atlantik

kann ebenfalls als wesentlich betrachtet werden. Obwohl der direkte Einfluß der Staubfracht auf das

Absinken organischer Partikel nicht beobachtet werden konnte, legen die gefundenen Kopplungen der

Partikelflüsse in der Tiefsee mit den atmosphärischen Transportmustcrn eine Beschleunigung des

Absinkens organischer Aggregate durch die Beschwerung mit Staub nahe. Wichtig zum weiteren

Verständnis dieser Kopplung wäre eine zeitlich wesentlich höher aufgelöste Messung bei

gleichzeitiger Erfassung von Pigmentverteilung in der Deckschicht und der Staubfracht. Die

bearbeiteten Stationen, vor allem vor Cape Verde, eignen sich für einen solchen Ansatz. Die

notwendige Technologie zur zeitgleichen Erfassung von Pigmentvelieilungen mit dem SeaWifs

Satelliten ist ebenfalls kurz vor dem Einsatz und verspricht die baldige Möglichkeit vergleichender

Messungen. Auch für die paläoklimtische Interpretation von Sedimentkernen könnten sich so neue

Erkenntnisse vor allem aus den saisonalen Wechseln versch iedener atmosphärischer TranspOliniveaus

ergeben, da sich hier sehr kurzfristige Variationen zunächst unabhängig von der Küstenentfernung im

Korngrößenspektrum niederschlagen.

Neueste Messungen in Verbindung mit atmosphärischen ModelJberechnungen zeigen die Bedeutung

des mineralischen Aerosols für das Strahlungsverhalten und Wärmebudged der Atmosphäre, was im

Gegensatz zu den Effekten von C02 und S042- lange unterschätzt wurde. Hinzu kommt die Tatsache,

daß ein wesentlicher Teil des weit transportierten Mineralstaubes heute aus der Erosion

landwiIischaftlich überbelasteter Böden in ar iden Gebiet stammt (Li et al. 1996; Andrae 1996). Von

Messungen in der Sahelzone ausgehende Schätzungen deuten fast die Hälfte des globalen

Mineralstaubes als anthropogen. Ausgehend von dieser Überlegung ergibt sich auch für die Kopplung

der Staubfracht an den KohlenstofftranspOli im Ozean die Frage, inwieweit anthropogen verursachter

Staub zur Effizienz der biologischen Pumpe beiträgt. Nicht zuletzt daher erscheint es sinnvoll, einer

Ausblick 133

weiteren Abschätzung der Kopplung zwischen Staubeintrag und Partikelfluß auch In distalen,

quellfernen Gebieten wie dem Westatlantik oder der Karibik nachzugehen.

Literatur

6. Litcl'atur

135

Aharonson N., N. Karasikov, M. Roitberg and 1. Shamir (1986) GALAI-CIS-l - A novel approach toaerosol particle size analysis. Journal of Aerosol Science, 17, 530-536.

Alldreelge A. L. and G. A. Jackson (1995) Aggregation in marine systems. Deep-Sea Research Part IJ,42, 1-7.

Alldredge A. L. and C. Gotschalk (1989) Direct observations of the mass flocculation of diatomblooms. Deep-Sea Research, 36, 159-171.

Alldregde A. L. und P. McGillivary (1991) The attachment probabilities of marine snow and theirimplications for particle coagulation in the oceans. Deep-Sea Research, 38, 431-443.

Andrae M. 0. (1996) Raising dust in the greenhouse. Nature, 380,389-390.

Asper V. L. (1987) Measuring the flux and sinking speed ofmarine snow aggregates. Deep-SeaResearch, 34,1-17.

Asper V. L., S. Honjo and T. H. Orsi (1992) Distribution and transport ofmarine snow aggregates inthe Panama Basin. Deep-Sea Research, 39, 939-952.

Barnola J. M., D. Raynaud, Y. S. Korotkevich and C. Lorius (1987) Vostok ice core provides160.000-year record of atmospheric C02. Nature, 329, 408-414.

Berger W. H. and R. S. Keir (1984) Glacial-Holocene changes in atmospheric C02 and the deep-searecord. In: Climate processes and climate sensitivity. AGU Geophysical Monographs, 29, 337-351.

Berger W. H. anel G. Wefer (1990) Export production: seasonality and intermittancy, andpaleoceanographic implications. Palaeogeography, Palaeoclimatology, Palaeoecology, 89, 245254.

Berger W. H., Smetacek V. and G. Wefer (1989) Ocean productivity and paleoproductivity - Anoverview. In: Productivity in the ocean: Present andpast. W. H. Berger, V. S. Smetacek and G.Wefer, editors, Wiley, New York, pp. 139-153.

Bodungen v. B., G. Fischer, E. M. Nöthig and G. Wefer (1987) Sedimentation ofkrill faeces duringspring development of phytoplankton in Bransfield Strait, Antarctica. In: Particle flux in the ocean,Degens E. T., E. T. Izdar and S. Honjo, eds., Mitt. Geol. Paläont. Inst., Univ. Hamburg, SCOPEUNEP Sonderband, 62, 243-257.

Broecker W. S. (1982) Ocean chemistry during glacial time. Geochimica et Cosmochimica Acta, 46,1689-1705.

Bruland K. W., P. K. Bienfang, J. K. B. Bishop, G. Eglington, V. Ittekott, R. Lampitt, M. Sarnthein, 1.Thiede,1. Walsh, G. Wefer (1989) Flux to the seaflor. In: Productivity in the ocean: Present andpast. W. H. Berger, V. S. Smetacek and G. Wefer, editors, Wiley, New York, pp.193-215.

Buat-Menard P., J. Davies, E. Remoudaki, J. C. Miquel, G. Bergametti, C. E. Lambert, U. Etzat, C.Quetel, J. La Rosa and S. W. Fowler (1989) Non-steady-state biological removal of atmosphericparticles from Mediterranean surface waters. Nature, 340, 131-134.

Carlson T. N. and 1. M. Prospero (1972) The large-scale movement of Sahm'an air outbreaks over thenorthern equatorial Atlantic. Journal 01Applied Meteorology, 11,283-297.

Chiapello 1., G. Bergametti, L. Gomes, B. Chatenet, F. Dulac, J. Pimenta and E. Santos Suares (1995)An additionallow layer transport of Sahelian and Saharan dust transport over the North-EasternTropical Atlantic. Geophysical Research Letters, 22, 3191-3194.

Costello D. K., K. L. Carder, P. R. Betzer and R. W. Young (1989) In situ holographic imaging ofsettling pmiicles: applications for individual particle dynamics and oceanic flux measurements.Deep-Sea Research, 36,1595-1605.

136

Degens E. T. and V. fttekott (1984) A new look at clay-organic interactions. Mitt. Geol. Paläont. Inst.,Univ. Hamburg, 56, 229-248.

Deuser W. G. (1986) Seasonal and interannual variations in the deep-water particle fluxes in theSargasso Sea and their relation to surface hydrography. Deep-Sea Research, 33, 225-246.

Duce R. A., P. S. Liss, 1. T. MeriII, E. L. Atlas, P. Buat-Menard, B. B. Hicks, J. M. Miller, J. M.Prospero, R. Arimoto, T. M. Church, W. Ellis, J. N. Galloway, L. Hansen, T. D. Jickels, A. H.Knap, K. H. Reinhardt, B. Schneider, A. Soudine, J. J. Tokos, S. Tsunogai, R. Wollast ami M.Zhou (1991) The atmospheric input of trace species to the world ocean. Global BiogeochemicolCycles, 5,193-259.

Fischer G. and G. Wefer (1991) Sampling, preparation and analysis of marine particulate matter. In:The analysis and characterization 01marine partie/es. D. C. Hurd, W. Spencer, editors, AGU,Washington. pp. 391-397.

Fischer G., D. Fütterer, R. Gersonde, S. Honjo, D. Ostermann and G. Wefer (1988) seasonalvariability of particle flux in the WedeII Sea and its relation to ice cover. Nature, 335, 426-428.

Fischer G., B. Donner, V. Ratmeyer, R. Davenport and G. Wefer (1996a) Distinct year-to-yearpatiicle flux variations off Cape Blanc during 1988-1991: Relation to d 180-deduced sea-surfacetemperatures and trade winds. Journal 01Marine Research, 54, 73-98.

Fischer G., G. Krause, S. Neuer and G. Wefer (1996b) Short-term sedimentation pulses recorded witha fluorescence sensor and sediment traps in 900 m wafer depth in the Canary Basin. in press byLimnology and Oceanography.

FO\ovler S. W. and G. A. Knauer (1986) Role of large particles in the transpoli of elements and organiccompounds through the oceanic wafer column. Prog. Oceanography, 16,147-194.

Gardner W. D. and 1. D. Walsh (1990) Distribution of macroaggregates and fine-grained particlesacross a continenta1 mm'gin and their potential role in fluxes. Deep-Sea Research, 37, 401-411 .

Gentien P., M. Lunven, M. Lehaitre and J. L. Duvent (1995) ln-situ depth profiling of pmticle sizes.Deep-sea Research, 42, 1297-1312.

Gersonde R. and G. Wefer (1987) Sedimentation ofbiogenic siliceous particles in Antarctic waters(Atlantic sector). Marine Micropaleontology, 11, 311-332.

Haake B. and V. Ittekott (1990) Die windgetriebene "biologische Pumpe" und Kohlenstoffentzug imOzean. Naturwissenschaften, 77, 75-79.

Honjo S. (1982) Seasonality and interaction of biogenie and lithogenic particulate flux at the PanamaBasin. Science, 218, 883-884.

Honjo S., K. W. Doheliy, Y. C. Agrawal and V. L. Asper (1984) Direct optical assesment of largeamorphous aggregates (marine snow) in the deep ocean. Deep-Sea Research, 31, 67-76.

Honjo, S. and S. J. Manganini (1993) Annual biogenie particle fluxes to the interior ofthe NorthAtlantic Ocean; studied at 34°N 21 °W and 48°N 21 °W. Deep-Sea Research, 40,587-607.

Hughes P. and E. D. Balion (1974) Stratification and wafer mass structure in the upweIIing area offnOlihwest Africa in ApriIlMay 1969. Deep-Sea Research, 21, 611-628.

Ittekott V., R. R. Nair, S. Honjo, V. Ramaswamy, M. Bartsch, S. Manganini and B. N. Desai (1991)Enhanced particle fluxes in the Bay ofBengal induced by injection offreshwater. Nature, 351,385-387.

Jickels T. D., T. M. Church and W. G. Deuser (1987) A comparison of atmospheric inputs and deepocean particle fluxes for the Sargasso Sea. Global Biogeochemical Cycles, 1,117-130.

Kaj ihara M. (1971) Settl ing velocity and porosity of large suspended particles. Journal 01 theOceanographic Society 01Japan, 27, 158-162.

Literatur 137

Kalu A. E. (\979) The African dust plume: Its characteristics and propagation across West Africa inwinter. In : Sahm-an dust: mobilization, transport, deposition. C. Moralcs, editor, SCOPE 14,Wiley, New York, pp. 95-118.

Klein B. and G. Siedler (\ 989) On the origin of the Azores Current. Journal ofGeophysical Research,94, 6159-6168.

Kremling K. and P. Streu (1993) Sahat'an dust influenced trace element fluxes in deep North Atlanticsubtropical waters. Deep-Sea Research, 40, 1155-1168.

Koopmann B. (1981) Die Sedimentation von Saharastaub im subtropischen Nordatlantik während derletzten 25.000 Jahre. Meteor Forschungs-Ergebnisse C, 35, 23-59.

Lampitt R. S. (1985) Evidence for the seasonal deposition of dtritus to the deep-sea floor and istsubsequent resuspension. Deep-Sea Research, 32,885-897.

Lampitt R. S., W. R. Hillier and P. G. Challenor (1993) Seasonal and diel variation in the open oceanconcentration of marine snow aggregates. Nature, 362, 737-739.

Li X., H. Maring, D. Savoie, K. Voss and 1. M. Prospero (1996) Dominance ofmineral dust in aerosoilight-scattering in the North Atlantic trade winds. Nature, 380, 416-419.

McCave I. N., P. F. Lonsdale, C. D. Holister and W. D. Gardner (1985) Sediment transport over theHatton and Gardar contourite drifts. Journal of Sedimentology and Petrology, 50, 1049-1062.

MacIntyre S., A. L. Alldredge and C. C. Gotschalk (1995) Accumulation ofmarine snow at densitydiscontinuities in the water column. Limnology and Oceanography, 40, 449-468.

Mmiin J. H. and S. E. Fitzwater (1988) Iron deficiency limits phytoplankton growth in the nOlih-eastPacific subantarctic. Nature, 331,341-343.

Milliman J. D. and 1. P. M. Syvitski (1992) Geomorphic/Tectonic control of sediment discharge to theocean: The importance of small mountainous rivers. Journal ofGeology, 100, 520-540.

Mittelstaedt E. (1991) The ocean boundary along the northwest African coast. Prog. Oceanography,26,307-355.

Nykjaer, L. (1988) Remote sensing applied to the Northwest African upwelling area. Ph.D. Thesis,University of Copenhagen, pp. 144

Nykjaer, L. and L. V. Camp (1994) Seasonal and interannual variability of coastal upwelling alongnOlihwest Africa and POliugal from 1981 to 1991. Journal ofGeophysical Research, 99, 1419714207.

Passow U. and A. L. Alldregde (1995) Aggregation of a diatom bloom in a mesocosm: The role oftransparent exopolymer pmiicles (TEP). Deep-Sea Research Part 11, 42, 99-109.

Petterssen S. (1956) Weather analysis and forecasting. McGraw-Hill, NewYork.

Prospero J. M. (1990) Mineral-aerosol transport to the North Atlantic and North Pacific: the impact ofAfrican and Asian sources. In: The long range atmospheric transport ofnatural and contaminantsubstances. A. H. Knap, editor, Kluwer Acad.Publ., The Netherlands. pp. 59-86.

Psenner R. (1991) Detection and sizing of aquatic bacteria by means of epifluorescence microscopyand image analysis. Image enhancement and analysis, Oct. 1991, 13-15.

Ramaswamy V., R. R. Nair, S. Manganini, B. Haake and V. Ittekott (1991) Litll0genic fluxes to thedeep Arabian Sea measured by sediment traps. Deep-Sea Research, 38, 169-184.

Ratmeyer V. and G. Wefer (1996) A high resolution camera system (ParCa) for imaging particles inthe ocean: System design and results from profiles and a three-month deployment. Journal ofMarine Research, 54, 589-603.

RiebeseIl U. (1991) Particle aggregation during a diatom bloom. I. Physical aspects. Marine EcologyProgress Series, 69, 281-291.

138

Russel W. B. (1980) Review ofthe role of colloidal forces in the rheology 01' suspensions. Journal 0/Rheology, 24, 287-317.

Sarnthein M., J. Thiede, U. Pflaumann, H. Erlenkeuser, D. Fütterer, B. Koopmann, H. Lange and [~.

Seibold (1982) Atmospheric and oceanic circulation patterns ofT Northwest Africa during the past25 million years. In: Geology ofthe NortfHjlest Aji'iean eontinental margill. U. Rad, K. Hinz, M.Sarnthein, E. Seibold, editors, Springer, Berlin, pp.545-604.

Schröder 0., C. Krambeck and H. .I. Krambeck (1990) Einflüsse digitalerBildsegmentierungstechniken auf die quantitative Analyse fluoreszierenden mikrobiellenPlanktons. In: Pillmann W. ancl A. .laeschke, eds., Informatikfiir elen UIIHjleltsehutz, Springer,Berlin, pp. 827-836.

Schütz L. (1980) Long range transport of desert dust with special emphasis on the Sahara. In:Aerosols: Anthropogenie anel natural, sources anel tramjJort, T. J. Kneip and P. J. Lioy, ec!s., Ann.of the New York Acadademy of Science, 338, pp. 515-532.

Schütz L. and M. Sebert (1987) Mineral aerosol and source identification. Journal ofAerosol Science,18,1-10.

Schütz L., R. Jaenicke and H. Pietrek (1981) Sahat'an dust transport over the North Atlantic ocean. In:Desert dust: Origin, characteristics anel effects on mall. T. L. Pewe, editor, Geo!. Soc. of Amer.,spec. paper, 186, 87-100.

Shanks A. L. and J. D. Trent (1980) Marine now: Sinking rates and potential role in vertical flux.Deep-Sea Research, 27A, 137-143.

Sieracki M. E., E. Reichenbach and K. L. Webb (1989) Evaluation of automated threshold selectionmethods for accurately sizing microscopic fluorescence cells by image analysis. AppliedEnvironmental Microbiology, 49, 2762-2772.

Silver M. W. and A. L. Alldredge (1981) Bathypelagic marine snow: deep-sea algal and detritalcommunity. Journal ofMarine Research, 39,501-530.

Smith D. C., G. F. Steward, R. A. Long and F. Azam (1995) Bacterialmediation of carbon fluxesduring a diatom bloom in a mesocosm. Deep-Sea Research Part 11,42, 75-97.

Stolzenbach K. (1993) Scavenging of small partieies by fast-sinking porous aggregates. Deep-SeaResearch, 40, 359-369.

Stramma P. and G. Siedler (1988) Seasonal changes in the North Atlantic Subtropical gyre. Journal ofGeophysical Research, 93, 8111-81 18.

Swap R., M. Garstang, S. Greco, R. Talbot and P. Kallberg (1992) Saharan dust in the Amazon Basin.Tellus , 44B, 133-149.

Syvitski J. P., K. W. Asprey and K. W. G. Leblanc (1995) In-situ characteristics ofparticles settlingwithin a deep-water estuary. Deep-Sea Research 11,42,223-256.

Tegen 1., A. A. Lewis and 1. Fung (1996) The influence on climate forcing of mineral aerosols fromdisturbed soils. Nature, 380, 419-422.

TetzlaffG. and K. Wolter (1980) Meteorological patterns and the transport ofmineral dust from theNorth African continent. Paleoecology ofAfrica, 12, 3 1-42.

Wefer G. (1993) Formation and composition of marine particulates. In: The global carbon cycle, M.Heimann, ed., NATO ASI Series, Vo!. 115.

Wefer G. and G. Fischer (1993) Seasonal patterns of vertical particle flux in equatorial and coastalupwelling areas ofthe eastern Atlantic. Deep-Sea Research, 40, 1613-1645.

Literatur 139

Westphal D. L., O. B. Toon and T. N. CarIson (1987) A two-dimensional numerical investigation ofthe dynamics and microphysics of Saharan dust storms. Journal 01 Geophysical Research, 92D,3027-3049.

Anhang

7. Anhang

Nachfolgend sind die im Rahmen dieser Dissertation verwendeten Daten aufgelistet.

141

142

Tab. 1: Aggregatverteilungen Profil GeoB 2212: Häufigkeiten und Größenklassen

Bild Nr Tiefe Ausschnitte Anzahl A B C D E F G(m) pro Liter: <0.1 mm < 0.25 mm < 0.5 mm < 0.75 mm < 1.0 mm <= 2.0 > 2.0 mm

5 0.00 5 238.64 56.24 84.36 71.44 22.04 3.8 0.76 06 2.50 3 374.68 101.08 125.40 107.92 25.08 10.64 3.80 0.767 5.00 3 380.76 110.20 118.56 114.00 27.36 6.84 3.80 0.009 10.00 3 677.16 223.44 248.52 150.48 40.28 12.92 1.52 0.00

10 12.50 3 677.16 223.44 248.52 150.48 40.28 12.92 1.52 00012 1750 3 483.36 146.68 164.92 114.76 41.04 12.92 3.04 00014 22.50 3 466.64 133.76 173.28 109.44 38.76 6.08 5.32 0.0016 27.50 3 411.16 111.72 148.96 10716 33.44 7.60 2.28 0.0022 51.00 3 279.68 73.72 101.84 78.28 21.28 2.28 2.28 0.0024 57.00 3 309.32 96.52 102.60 86.64 19.00 3.80 0.76 0.0026 63.00 5 299.44 7676 125.40 68.40 2356 456 0.76 00028 69.00 5 348.84 90.44 117.04 96.52 34.20 9.88 0.76 0.0030 75.00 5 362.52 92.72 122.36 108.68 28.88 6.08 3.80 00032 81.00 5 338.96 66.88 115.52 111.72 2736 10.64 6.84 0.0034 87.00 5 306.28 72.20 95.76 90.44 29.64 12.92 5.32 0.0036 93.00 5 247.76 44.84 103.36 78.28 15.96 4.56 0.76 0.0037 96.00 5 288.04 5852 114.00 87.40 22.04 5.32 0.76 0.0041 101.12 5 283.48 63.84 85.88 95 31.92 6.84 0 047 122.24 5 465.12 141.36 165.68 126.16 24.32 3.80 3.04 0.7652 139.84 5 520.60 178.60 177.08 128.44 28.88 4.56 3.04 0.0057 157.44 5 492.48 155.80 195.32 115.52 21.28 4.56 0.00 0.0062 175.04 5 492.48 155.80 195.32 115.52 21.28 4.56 0.00 0.0067 192.64 5 414.20 127.68 148.20 117.04 19.00 1.52 0.76 00070 203.20 5 537.32 172.52 200.64 135.28 24.32 4.56 0.00 0.0073 213.76 5 357.96 83.60 129.96 11552 24.32 3.80 0.76 0.0076 224.32 5 639.92 197.60 240.16 162.64 34.20 4.56 0.76 0.0079 234.88 5 393.68 121.60 141.36 119.32 11.40 0.00 0.00 0.0082 245.44 5 351.12 91.96 126.92 108.68 19.00 3.04 1.52 0.0085 256.00 5 312.36 69.16 114.76 110.96 12.92 4.56 0.00 0.0088 266.56 5 304.76 75.24 123.12 91.20 11.40 3.80 0.00 0.0093 284.16 5 272.84 68.40 101.84 84.36 18.24 0.00 0.00 0.0096 294.72 5 251.56 75.24 94.24 71.44 10.64 0.00 0.00 0.0099 305.28 5 280.44 51.68 103.36 99.56 22.04 3.04 0.00 0.76

102 315.84 5 167.20 43.32 57.76 57.76 6.84 1.52 0.00 0.00105 326.40 5 176.32 39.52 68.40 55.48 12.16 0.76 0.00 0.00108 336.96 3 186.96 44.84 68.40 62.32 11.40 0.00 0.00 0.00111 347.52 3 239.40 60.80 95.00 72.96 9.12 1.52 0.00 0.00114 35808 3 415.72 115.52 171.00 10108 25.08 2.28 0.76 0.00117 368.64 3 385.32 102.6 133 123.12 21.28 3.8 1.52 0120 379.20 3 291.84 59.28 119.32 89.68 19.00 3.04 1.52 0.00123 389.76 5 378.48 126.16 13376 88.16 28.88 1.52 0.00 0.00126 400.32 5 265.24 53.20 99.56 89.68 17.48 3.80 1.52 0.00129 410.88 5 265.24 53.20 99.56 89.68 17.48 3.80 1.52 0.00132 421.44 5 246.24 49.40 91.20 83.60 20.52 1.52 0.00 0.00135 432.00 5 229.52 36.48 95.00 84.36 10.64 2.28 0.76 0.00138 442.56 5 248.52 50.16 95.00 85.12 15.96 1.52 0.76 0.00141 453.12 5 206.72 37.24 71.44 73.72 22.04 1.52 0.00 0.76144 463.68 5 205.20 47.88 77.52 65.36 11.40 0.00 3.04 0.00147 474.24 5 199.88 38.76 70.68 69.16 19.76 1.52 0.00 0.00150 484.80 5 159.60 45.60 58.52 38.00 14.44 3.04 0.00 0.00152 491.84 5 154.28 28.88 56.24 51.68 17.48 0.00 0.00 0.00155 502.40 5 142.88 30.40 56.24 44.84 9.12 2.28 0.00 0.00158 512.96 5 107.92 14.44 39.52 35.72 15.96 2.28 0.00 0.00160 520.00 5 144.40 2584 57.00 47.88 12.16 152 0.00 0.00

Anhang

Tab. 2: Aggregatverteilungen Profil GeoB 2212: marine snow und 1-pixel-Filter

Bild Nr Tiefe Anzahl Anzahl Anzahl +1pix Anzahl +1pix(m) <0,5 mm > 0,5 mm > 0,5 mm gesamt

5 0.00 212.04 26.6 3344 306.666 2.50 3344 40.28 3914 377.727 5.00 342.76 38 46.36 528.969 10.00 62244 54.72 54.72 677.16

10 12.50 62244 54.72 55.86 5802612 17.50 426.36 57 5358 47514 22.50 41648 50.16 46.74 438.916 27.50 367.84 43.32 34.58 3454222 51.00 253.84 25.84 24.7 294.524 57.00 285.76 23.56 26.22 304.3826 63.00 270.56 28.88 36.86 324.1428 69.00 304 44.84 41.8 3556830 75.00 323.76 38.76 41.8 350.7432 81.00 294.12 44.84 46.36 322.6234 87.00 2584 47.88 3458 277.0236 93.00 22648 21.28 24.7 267.937 96.00 259.92 28.12 3344 285.7641 101.12 244.72 38.76 35.34 374.347 122.24 433.2 31.92 34.2 492.8652 139.84 484.12 3648 31.16 506.5457 15744 466.64 25.84 25.84 4924862 175.04 466.64 25.84 23.56 453.3467 192.64 392.92 21.28 25.08 475.7670 203.20 50844 28.88 28.88 447.6473 213.76 329.08 28.88 34.2 4989476 224.32 6004 39.52 2546 516879 234.88 382.28 114 1748 372482 24544 327.56 23.56 20.52 331.7485 256.00 294.88 1748 16.34 308.5688 266.56 289.56 15.2 16.72 288.893 284.16 2546 18.24 1444 262.296 294.72 240.92 10.64 18.24 26699 305.28 254.6 25.84 17.1 223.82

102 315.84 158.84 8.36 10.64 171.76105 326.40 1634 12.92 12.16 181.64108 336.96 175.56 114 11.02 213.18111 347.52 228.76 10.64 19.38 327.56114 358.08 387.6 28.12 27.36 400.52117 368.64 358.72 26.6 25.08 338.58120 379.20 268.28 23.56 26.98 335.16123 389.76 348.08 304 26.6 321.86126 400.32 24244 22.8 22.8 265.24129 410.88 24244 22.8 2242 255.74132 42144 224.2 22.04 17.86 237.88135 432.00 215.84 13.68 15.96 239.02138 442.56 230.28 18.24 21.28 227.62141 453.12 1824 24.32 19.38 205.96144 463.68 190.76 1444 17.86 202.54147 474.24 178.6 21.28 19.38 179.74150 484.80 142.12 1748 1748 156.94152 491.84 136.8 1748 1444 148.58155 50240 13148 114 14.82 1254158 512.96 89.68 18.24 15.96 126.16160 520.00 130.72 13.68 13.68 1444

143

144

Tab. 3: Aggregatverteiiungen Profil GeoB 2217: Größenklassen

Tiefe Ausschnitte Anzahl A B C 0 E F G(m) pro Liter: < 0.1 mm < 0.25 mm < 0.5 mm < 0.75 mm < 1.0mm <= 2.0 mm > 2.0 mrn

0 5 724 406 85 120 70 21 10 124 5 674 442 74 87 35 21 2 13

12 5 463 242 64 99 33 17 5 316 5 450 221 68 101 39 9 7 520 5 603 348 54 102 65 21 6 728 5 493 246 56 106 58 17 6 436 5 991 603 89 174 81 25 13 644 5 875 511 89 160 78 21 9 752 5 998 588 108 162 83 30 17 1060 5 360 188 53 81 30 2 5 168 5 465 281 46 83 39 13 2 176 5 490 302 62 88 26 3 8 184 5 477 297 48 73 45 11 3 092 5 296 132 41 66 33 19 2 3

100 5 427 235 57 87 39 3 4 2120 5 461 276 52 89 35 7 1 1

134.65 5 721.24 228.76 259.16 174.8 48.64 7.6 1.52 0.76150.97 5 687.04 193.80 254.60 192.28 38.76 4.56 3.04 0.00171.37 5 783.56 209.00 299.44 227.24 34.20 11.40 2.28 0.00195.85 5 590.52 134.52 239.40 178.60 32.68 3.80 1.52 0.00216.25 5 590.52 134.52 239.40 178.60 32.68 3.80 1.52 0.00228.49 5 408.12 82.84 145.16 133.00 44.08 1.52 1.52 0.00240.73 5 421.8 92.72 140.60 141.36 42.56 3.80 0.76 0.00248.89 5 342 85.88 98.80 12236 30.40 4.56 0.00 0.00265.21 5 421.8 90.44 155.04 126.16 44.08 6.08 0.00 000273.37 5 442.32 11020 162.64 120.08 41.80 7.60 0.00 0.00285.61 5 428.64 84.36 161.12 137.56 38.00 7.60 0.00 0.00301.93 5 424.84 8664 158.08 129.20 39.52 9.12 2.28 0.00314.17 5 490.96 11020 166.44 159.60 44.08 9.12 1.52 0.00326.41 5 383.04 86.64 144.40 117.80 28.12 6.08 0.00 0.00338.65 5 380.76 66.12 142.88 126.92 35.72 6.84 2.28 0.00350.89 5 366.32 72.20 126.16 118.56 37.24 8.36 3.80 0.00363.13 5 377.72 80.56 148.96 106.40 33.44 7.60 0.76 0.00371.29 5 314.64 79.80 101.08 87.40 36.48 6.08 3.80 0.00387.61 5 348.84 77.52 133.00 92.72 38.00 6.08 1.52 0.00395.77 5 288.8 79.80 85.88 80.56 32.68 8.36 0.76 0.76412.09 5 319.96 72.96 113.24 91.96 31.16 6.84 3.80 0.00424.33 5 288.04 60.8 98.04 98.04 21.28 9.12 0.76 0436.57 5 294.88 59.28 98.80 101.08 26.60 7.60 0.76 0.76452.89 5 288.04 59.28 94.24 87.40 32.68 11.40 3.04 0.00461.05 5 281.2 75.24 89.68 82.08 25.84 6.08 2.28 0.00473.29 5 281.2 75.24 89.68 82.08 25.84 6.08 2.28 0.00485.53 5 322.24 62.32 118.56 95.76 34.96 7.60 3.04 0.00497.77 5 359.48 8436 123.88 109.44 31.92 9.88 0.00 000510.01 5 449.16 100.32 150.48 136.04 50.16 6.84 5.32 0.00522.25 5 433.96 88.16 149.72 143.64 37.24 13.68 1.52 0.00534.49 5 376.2 85.88 114.00 120.08 38.76 12.92 4.56 000

Anhang

Tab. 4: Aggregatverteilung Profil GeoB 2217: marine snow und 1-pixel-Filter

Tiefe Anzahl Anzahl Anzahl +1pix Anzahl +1pix(m) <0,5 mm > 0,5 mm > 0.5 mm gesamt

0 611 113 92 6994 603 71 64.5 568.5

12 405 58 59 456.516 390 60 79.5 526.520 504 99 92 54828 408 85 105 74236 866 125 120 93344 760 115 127.5 936.552 858 140 89 67960 322 38 46.5 412.568 410 55 46.5 477.576 452 38 48.5 483.584 418 59 58 386.592 239 57 52.5 361.5

100 379 48 46 444120 417 44 51.26 591.12

134.65 662.72 58.52 52.44 704.14150.97 640.68 46.36 47.12 735.3171.37 735.68 47.88 42.94 687.04195.85 552.52 38 38 590.52216.25 55252 38 42.56 499.32228.49 361 47.12 47.12 414.96240.73 374.68 47.12 41.04 381.9248.89 307.04 34.96 42.56 381.9265.21 371.64 50.16 49.78 432.06273.37 392.92 49.4 47.5 435.48285.61 383.04 45.6 48.26 426.74301.93 373.92 50.92 52.82 457.9314.17 436.24 54.72 44.46 437326.41 348.84 34.2 39.52 381.9338.65 335.92 44.84 47.12 373.54350.89 316.92 49.4 45.6 372.02363.13 335.92 41.8 44.08 346.18371.29 268.28 46.36 45.98 331.74387.61 303.24 45.6 44.08 318.82395.77 246.24 42.56 42.18 304.38412.09 278.16 41.8 36.48 304424.33 256.88 31.16 33.44 291.46436.57 259.16 35.72 41.42 291.46452.89 240.92 47.12 40.66 284.62461.05 247 34.2 34.2 281.2473.29 247 34.2 39.9 301.72485.53 276.64 45.6 43.7 340.8649777 317.68 41.8 52.06 404.32510.01 386.84 62.32 57.38 441.56522.25 381.52 52.44 54.34 405.08534.49 319.96 56.24 5624 376.2

145

146

Tab. 5: Aggregatverteilung während der Verankerung C14: Häufigkeiten und Größenklassen

Bild Nr Tag Ausschnitte Anzahl A B C D E F Gpro Liter: <0.1 mm < 0.25 mm < 0.5 mm < 0.75 mm < 1.0mm <= 2.0 mm > 2.0 mm

6 1.88 5 50.54 11.78 14.82 17.48 4.56 1.9 0 010 3.13 5 59.66 14.82 18.62 18.24 5.32 2.28 0.38 0.0013 4.06 5 62.70 11.78 22.42 23.18 4.18 1.14 0.00 0.0019 5.94 5 87.78 22.04 28.50 29.64 6.46 1.14 0.00 0.0022 6.88 5 87.78 22.04 28.50 29.64 6.46 1.14 0.00 0.0028 8.75 5 136.80 58.90 33.82 31.92 9.50 2.28 0.38 0.0033 10.31 5 147.44 57.00 36.10 41.80 9.88 2.28 0.38 0.0038 11.88 5 130.34 50.54 37.24 29.64 11.02 1.52 0.38 0.0043 13.44 5 104.88 48.26 22.80 22.04 7.22 3.42 1.14 00045 14.06 5 118.56 55.10 27.36 22.42 10.64 2.28 0.76 0.0050 15.63 5 85.5 23.94 26.98 23.56 10.26 0 0.38 0.3854 16.88 5 114.76 36.48 33.44 33.06 950 1.14 1.14 0.0058 18.13 5 112.48 45.98 28.12 26.98 9.12 2.28 0.00 0.0060 18.75 5 126.16 60.80 28.50 25.08 9.12 2.28 0.38 0.0065 20.31 5 140.60 53.20 39.90 34.96 9.88 2.66 0.00 0.0068 21.25 5 106.02 41.42 25.46 19.00 13.68 5.32 1.14 0.0073 22.81 5 98.80 33.06 26.60 28.88 8.74 1.14 0.38 0.0075 23.44 5 83.60 22.42 23.94 21.28 11.40 3.42 1.14 0.0078 24.38 5 77.14 17.48 22.80 26.22 9.50 1.14 0,00 0.0084 26.25 5 158.46 76.00 3686 30.02 10.26 3.80 1.14 0.3888 27.50 5 101.08 25.46 32.68 30.40 11.40 1.14 0.00 0.0093 29.06 5 71.06 11.78 20.14 28.88 7.98 1.90 0.38 0.0095 29.69 5 63.84 12.16 21.28 21.66 6.84 1.52 0.38 0.0098 30.63 5 72.20 11.78 27.74 23.18 8.36 1.14 0.00 0.00

105 32.81 5 64.98 12.54 23.94 20.52 7.60 0.38 0.00 0.00110 34.38 5 61.56 10.26 21.09 2622 3.42 0.57 0.00 000112 35 5 59.85 11.4 17.67 21.66 8.55 0.57 0 0115 35.94 5 118.56 55.10 27.36 22.42 10.64 2.28 0.76 0.00120 37.50 5 81.51 15.39 31.35 21.66 11.97 1.14 0.00 0.00125 39.06 5 64.6 11.4 20.9 19.76 8.74 1.52 2.28 0130 40.63 5 4674 10.26 17.48 10.64 6.46 0.76 1.14 000135 42.19 5 72.58 1368 26.98 25.08 5.32 1.52 0.00 0.00140 43.75 5 119.32 43.70 34.20 26.98 12.54 1.90 0.00 0.00145 45.31 5 119.32 43.70 34.20 26.98 12,54 1.90 0.00 0.00150 46.88 5 163.78 66.88 44.46 38.00 11.78 2.28 0.38 0.00155 48.44 5 163,02 82.08 40.66 27.74 9,12 3.42 0.00 0.00160 50.00 5 163,02 72.58 40.66 37.24 9.50 2.28 0.76 0.00165 51.56 5 68.40 12.92 19.38 25,08 9.50 1.52 0,00 0.00170 53.13 5 60.80 12.92 17.48 22.42 6.08 1.90 0.00 0.00175 54.69 5 63.46 15.58 26.6 14.44 5.7 1.14 0 0181 56.56 5 39.90 7.22 1330 13.30 4.94 1.14 0.00 0.00185 57.81 5 47.12 10.64 1330 13.30 7.60 1.90 0.38 0.00190 59.38 5 45.98 10.64 11.02 16.34 5.70 2.28 0.00 0.00195 60.94 5 45.98 10,64 11.02 16,34 5.70 2.28 0.00 0.00200 62.50 5 47.88 9.88 19.00 13.68 4.94 0,38 0.00 0.00205 64.06 5 55.86 10.26 19.38 18.62 6.08 1.52 0.00 0.00210 65.63 5 52.44 10.64 15.96 15.20 7.60 1.90 1.14 0.00215 67.19 5 46.36 9.50 14.06 15.58 6.84 0.38 0.00 0.00220 68.75 5 51.68 11.78 15.96 14.44 6.84 2.66 0.00 0.00225 70.31 5 4560 7.60 10.26 16.34 8.74 2.28 0.38 0.00230 71.88 5 62.70 12.54 21.28 19.38 7.22 1.90 0.38 0.00235 73.44 5 52.82 10.26 19.38 14.06 7.60 1.52 0.00 0,00240 75.00 5 53.96 11.40 16.34 19.38 6.08 0.38 0.38 0.00245 76.56 5 61.56 14.44 14.82 24.70 5.32 2.28 0.00 000250 78.13 5 61.56 11.78 17.48 21.28 9.88 1.14 0.00 0.00255 79.69 5 63.84 14.06 19.38 19.00 988 1.14 0.38 0.00260 81.25 5 55.86 10.26 19.00 19.00 7.22 0.38 0.00 0.00264 82.50 5 53.20 10.26 14.06 17.10 988 1.14 0.76 0.00270 84.38 5 53.96 15.20 16.34 16.72 494 0.76 0.00 0.00

Anhang

Tab. 6: Aggregatverteilung während der Verankerung C14: marine snow und 1-pixel-Fliter

Bild Nr Tag Anzahl Anzahl Anzahl +1 pix Anzahl +1pix<0,5 mm > 0,5 mm < 0,5 mm > 0.5

6 1.88 44.08 6.46 47.88 7.2210 3.13 51.68 7.98 54.53 6.6513 4.06 57.38 5.32 68.78 6.4619 5.94 80.18 7.6 80.18 7.622 6.88 80.18 7.6 102.41 9.8828 8.75 124.64 12.16 129.77 12.3533 10.31 134.9 12.54 126.16 12.7338 11.88 117.42 12.92 105.26 12.3543 13.44 93.1 11.78 98.99 12.7345 14.06 104.88 13.68 89.68 12.3550 15.63 74.48 11.02 88.73 11.454 16.88 102.98 1178 102.03 11.5958 18.13 101.08 11.4 107.73 11.5960 18.75 114.38 11.78 121.22 12.1665 20.31 128.06 12.54 106.97 16.3468 21.25 85.88 20.14 87.21 15.273 22.81 88.54 10.26 78.09 13.1175 23.44 67.64 15.96 67.07 13.378 24.38 66.5 10.64 104.69 13.1184 26.25 142.88 15.58 115.71 14.0688 27.50 88.54 12.54 74.67 11.493 29.06 608 10.26 57.95 9.595 29.69 55.1 8.74 58.9 9.1298 30.63 62.7 9.5 59.85 8.74

105 32.81 57 7.98 57.285 5.985110 34.38 57.57 3.99 54.15 6.555112 35 50.73 912 77.805 11.4115 35.94 104.88 13.68 86.64 13.395120 37.50 68.4 13.11 60.23 12.825125 39.06 52.06 12.54 45.22 10.45130 40.63 38.38 8.36 52.06 7.6135 42.19 65.74 6.84 85.31 10.64140 43.75 104.88 14.44 104.88 14.44145 45.31 104.88 14.44 127.11 14.44150 46.88 149.34 14.44 149.91 13.49155 48.44 150.48 12.54 150.48 12.54160 50.00 150.48 12.54 103.93 11.78165 51.56 57.38 11.02 55.1 9.5170 53.13 52.82 7.98 54.72 7.41175 54.69 56.62 6.84 45.22 6.46181 56.56 33.82 6.08 35.53 7.98185 57.81 37.24 9.88 37.62 8.93190 59.38 38 7.98 38 7.98195 60.94 38 7.98 40.28 6.65200 62.50 42.56 5.32 45.41 6.46205 64.06 48.26 7.6 45.03 9.12210 65.63 41.8 10.64 40.47 8.93

215 67.19 39.14 7.22 40.66 8.36220 68.75 42.18 9.5 38.19 10.45225 70.31 34.2 11.4 43.7 10.45230 71.88 53.2 9.5 48.45 9.31235 73.44 43.7 9.12 45.41 7.98240 75.00 47.12 6.84 50.54 7.22245 76.56 53.96 7.6 52.25 9.31250 7813 50.54 11.02 51.49 11.21255 79.69 52.44 11.4 50.35 9.5260 81.25 48.26 7.6 44.84 9.69264 82.50 41.42 11.78 44.84 8.74270 84.38 48.26 5.7 24.13 2.85

147

J48

Tab. 7: Aggregatverteilung während der Verankerung C14: Volumina und Größenklassen

Bild Nr Tag Ausschnitte Volumen A B C 0 E F Gmm3/1 <0.1 mm < 0.25 mm < 0.5 mm < 0.75 mm < 1.0mm <= 2.0 mm > 2.0 mm

6 1.88 5 1.67 0.00 0.05 0.48 0.57 0.57 0.00 0.0010 3.13 5 2.44 0.00 0.07 0.51 0.63 0.84 0.39 0.0013 4.06 5 1.47 0.00 0.09 0.68 0.36 0.34 0.00 0.0019 5.94 5 2.07 0.00 0.09 0.89 0.74 0.35 0.00 0.0022 6.88 5 2.95 0.01 0.10 0.65 1.20 0.56 0.43 00028 8.75 5 2.99 0.01 0.10 0.80 1.19 0.68 0.22 0.0033 10.31 5 3.38 0.01 0.11 1.15 1.16 0.60 0.35 0.0038 11.88 5 3.17 0.01 0.10 084 1.36 0.52 0.33 0.0043 13.44 5 3.37 0.01 0.06 064 0.78 1.22 0.67 0.0045 1406 5 7.29 0.00 0.09 0.71 0.66 0.55 0.24 5.0350 15.63 5 5.39 0.00 0.08 072 1.22 0.00 0.20 31654 16.88 5 3.80 0.01 0.13 0.88 1.18 0.47 1.14 0.0058 18.13 5 2.68 0.01 0.09 0.72 1.04 0.82 0.00 00060 18.75 5 2.91 0.01 0.09 0.69 1.20 0.66 0.26 0.0065 20.31 5 2.85 0.01 0.14 0.94 1.11 0.66 0.00 0.0068 21.25 5 5.03 0.01 0.07 0.48 1.72 1.67 1.08 0.0073 22.81 5 2.54 0.01 0.08 0.85 1.07 0.32 0.20 0.0075 23.44 5 4.16 0.00 0.08 0.68 1.48 1.20 0.71 0.0078 24.38 5 2.63 0.00 0.07 0.81 1.29 0.45 000 0.0084 26.25 5 6.54 0.01 0.12 0.75 1.21 1.18 099 2.2888 27.50 5 2.65 0.00 0.11 0.82 1.39 0.33 0.00 0.0093 29.06 5 2.71 0.00 0.06 0.84 0.95 063 0.22 0.0095 29.69 5 2.10 0.00 0.08 0.67 073 0.38 0.24 0.0098 30.63 5 2.31 0.00 0.09 0.63 1.13 0.46 0.00 0.00

105 32.81 5 1.52 0.00 0.09 0.55 0.76 0.12 0.00 0.00110 34.38 5 1.22 0.00 0.07 0.65 0.37 0.13 0.00 0.00112 35.00 5 1.85 0.00 0.05 0.62 1.01 0.16 000 0.00115 35.94 5 3.98 0.01 0.09 0.56 1.40 0.80 1.14 0.00120 37.50 5 2.50 0.00 0.10 0.57 1.46 0.36 0.00 0.00125 39.06 5 5.45 0.00 0.07 0.54 1.06 0.54 3.24 0.00130 40.63 5 2.48 0.00 0.06 0.27 0.77 0.29 1.08 0.00135 42.19 5 1.67 0.00 0.09 0.57 0.54 0.47 0.00 0.00140 43.75 5 2.81 0.01 0.12 0.53 1.58 0.56 000 0.00145 45.31 5 2.14 0.01 0.09 0.72 0.92 0.40 0.00 0.00150 46.88 5 3.11 0.01 0.13 1.00 1.14 0.62 0.20 0.00155 48.44 5 2.98 0.01 0.13 0.73 1.03 1.08 0.00 0.00160 50.00 5 3.50 0.01 0.13 0.93 1.19 0.70 0.54 0.00165 51.56 5 2.34 0.00 0.06 0.68 1.11 0.49 0.00 0.00170 53.13 5 1.86 0.00 0.06 0.56 0.70 0.53 0.00 0.00175 54.69 5 0.94 0.00 0.07 0.21 0.45 0.19 0.00 000181 56.56 5 1.35 0.00 0.05 0.33 0.58 0.38 0.00 0.00185 57.81 5 2.16 0.00 0.04 0.36 0.87 0.60 0.29 000190 59.38 5 1.81 0.00 0.04 0.46 0.61 0.70 0.00 0.00195 60.94 5 1.60 0.00 0.06 0.40 0.85 0.29 0.00 0.00200 62.50 5 1.29 0.00 0.06 0.48 0.60 0.15 0.00 0.00205 64.06 5 1.73 0.00 0.07 0.44 0.74 0.47 0.00 0.00210 65.63 5 2.74 0.00 0.06 0.34 0.83 0.47 1.04 0.00215 67.19 5 1.21 0.00 0.05 0.44 0.63 0.09 0.00 0.00220 68.75 5 2.00 0.00 0.05 0.41 0.82 0.72 0.00 0.00225 70.31 5 2.57 0.00 0.03 0.41 1.06 0.80 0.26 0.00230 71.88 5 2.21 0.00 0.08 0.54 0.79 0.56 0.23 0.00235 73.44 5 1.71 0.00 0.06 0.38 0.91 0.36 0.00 0.00240 75.00 5 1.56 0.00 0.06 0.47 0.70 0.11 0.23 0.00245 76.56 5 2.08 0.00 0.05 0.68 0.61 0.73 0.00 0.00250 78.13 5 2.15 0.00 0.06 0.60 1.20 0.29 0.00 0.00255 79.69 5 2.49 0.00 0.06 0.46 1.24 0.36 0.37 0.00260 81.25 5 1.56 0.00 0.06 0.56 0.83 0.10 0.00 000264 82.50 5 2.74 0.00 0.05 0.46 1.20 0.41 0.61 000270 84.38 5 1.35 0.00 0.06 0.46 0.58 0.25 0.00 0.00

Anhang

Tab. 8: Aggregatverteilung während der Verankerung C14: marine snow und 1-pixel-Fliter (Volumina)

Bild Nr Tag Anzahl Anzahl Anzahl +1pix Anzahl +1 pix<0,5 mm > 0,5 mm < 0,5 mm > 0.5

6 1.88 0.53 1.14 0.56 1.5010 3.13 0.58 1.85 0.68 1.2713 4.06 0.77 0.69 0.88 0.8919 5.94 0.99 1.08 0.87 1.6422 6.88 0.76 2.19 0.83 2.1428 8.75 0.91 2.09 1.09 2.1033 10.31 1.27 2.11 1.11 2.1738 11.88 0.95 2.22 0.83 2.4443 13.44 0.71 2.66 0.76 4.5745 14.06 0.81 6.48 0.81 5.5350 15.63 0.81 4.58 0.91 3.6854 16.88 1.01 2.78 0.91 2.3358 18.13 0.82 1.87 0.81 1.9960 18.75 0.80 2.11 0.94 1.9465 20.31 1.08 177 0.82 3.1268 21.25 0.56 4.47 0.75 3.0373 22.81 0.94 1.60 0.85 2.5075 23.44 0.76 3.39 0.83 2.5778 24.38 0.89 1.74 0.89 3.7084 26.25 0.89 5.65 0.91 3.6988 27.50 0.93 1.72 0.92 1.7693 29.06 0.91 1.80 0.83 1.5795 29.69 0.75 1.35 0.74 1.4798 30.63 0.73 1.58 0.68 1.23

105 32.81 0.64 0.88 0.68 0.69110 34.38 0.72 0.50 0.70 0.83112 35.00 0.68 117 0.66 2.25115 35.94 0.65 3.33 0.66 2.58120 37.50 067 1.82 0.64 3.33125 39.06 0.61 4.84 0.47 3.49130 40.63 034 2.14 0.50 1.57135 42.19 0.66 1.01 0.66 1.57140 43.75 067 2.14 0.74 1.73145 45.31 0.82 1.32 0.98 1.64150 46.88 1.14 1.97 1.00 2.04155 48.44 0.87 2.12 0.97 2.27160 50.00 1.08 2.42 0.91 2.01165 51.56 0.74 1.60 0.69 1.41170 53.13 0.63 1.23 0.46 0.94175 54.69 0.29 0.65 0.34 0.81181 56.56 0.39 0.96 0.39 1.36185 57.81 0.40 1.76 0.45 1.54190 59.38 0.50 1.31 0.48 1.23195 60.94 0.46 1.14 0.50 0.94200 62.50 0.54 0.75 0.53 0.98205 64.06 0.51 1.22 0.46 1.78210 65.63 0.40 234 0.45 1.53215 67.19 0.49 0.72 0.47 1.13220 68.75 0.45 1.54 0.45 1.83225 70.31 0.45 2.12 0.53 1.85230 71.88 0.62 1.58 0.53 1.43235 73.44 0.44 1.27 0.48 1.15240 75.00 0.53 1.03 0.63 1.19245 76.56 0.74 1.34 0.70 1.41250 78.13 0.66 1.49 0.59 1.73255 79.69 0.52 1.97 0.57 1.45260 81.25 0.63 0.94 0.57 1.58264 82.50 051 2.23 0.51 1.53270 84.38 0.52 0.83 0.26 0.41

149

150

Tab. 9: Volumenverteilung von marine snow während Fallenproben-Inlervall

FallenprobeCI4

1234567891011121314151617181920

Volumenmm3/1 >0.5

1.2301.6382.1394.1282.5542.2462.6300.9882.109

2.6631281.810

2.2674531.1579291.3628651.0662621.4266661.7492721.2148061.4655211.528639

+ 3 TageOffseI

234567891011121314151617181920

Volumenmm3/1 >0.5

1.322.144.582.133152.450.992.112.661.812.040.941.290.981.531.751.211.470.76

Anhang 151

Tab. 10: Flußraten (mg / m' d) von Gesamt, Lithogen und organischem Kohlenstoff in den oberen Fallen der Station CI

Falle Probe Zeit intervall Einwaage Total Flux % Lithogen Lith Flux Flux 6·11 Flux 11·20 Flux 20·63 Corg Flux Ratio Ratio(mg) 6·11/20·63 Lilh/Corg

CI10 1 11.12.91 18.55 9.73 66.30 6.45 3.59 1.85 1.01 0.50 3.56 12.852 26.12.91 16.50 8.66 49.04 4.25 1.90 1.45 0.90 0.43 2.12 9.813 10.01.92 21.10 11.07 67.10 7.43 4.40 2.00 1.03 0.45 4.26 16.624 25.01.92 29.75 15.61 61.70 9.63 4.89 3.32 1.42 1.76 3.46 5485 10.02.92 29.60 15.53 44.52 691 5.00 1.60 0.31 2.14 16.01 3.226 2502.92 37.35 19.59 59.33 1162 6.94 3.42 1.26 0.88 5.53 13.217 11.03.92 60.60 31.79 36.75 1168 7.00 3.86 0.82 2.25 8.53 5198 26.03.92 92.60 48.58 30.99 1505 8.32 4.83 1.90 3.51 4.37 4.299 11.04.92 25.40 13.32 53.09 707 4.34 211 0.62 0.94 7.04 7.4910 26.04.92 36.55 19.17 54.08 1037 5.84 3.25 1.28 1.30 4.55 7.9711 11.05.92 52.95 27.78 67.70 1880 1.80 10.4712 26.0592 34.00 17.84 5766 10.28 4.39 3.47 2.42 1.71 1.81 6.0013 11.06.92 46.90 24.60 64.82 15.95 7.46 5.55 2.94 1.61 2.54 9.9114 26.06.92 33.70 17.68 71.91 1271 6.63 3.98 2.10 1.74 3.16 7.3215 110792 11.75 6.16 67.24 4.14 1.88 1.33 0.93 0.46 2.01 9.0416 26.07.92 16.10 8.45 73.07 6.17 3.70 1.79 0.68 0.46 5.41 13.3317 110892 6.30 3.30 48.52 160 0.78 0.57 0.26 025 2.99 6.4618 26.08.92 16.90 8.87 40.02 355 1.69 1.17 0.69 117 2.45 3.0319 10.09.92 18.90 9.91 71.66 710 4.14 2.07 0.90 0.50 4.61 14.0820 25.09.92 930 4.88 35.72 174 0.86 0.57 0.32 0.50 2.66 3.49

CI20 1 1110.92 26.15 22.02 3184 7.01 4.08 2.14 0.79 2.24 5.14 3.122 20.10.92 29.40 24.76 38.98 9.65 5.32 3.05 1.28 1.29 4.17 7.493 30.1092 40.85 34.40 34.72 1194 5.89 3.85 2.20 2.28 2.69 5.244 08.11.92 40.05 33.73 11.75 3.96 1.96 1.08 092 2.23 2.14 1.785 18.11.92 32.55 27.41 38.14 10.45 4.53 3.48 2.44 1.76 1.85 5.936 2711.92 15.00 12.63 5.84 0.74 0.34 0.23 0.17 0.87 2.00 0.857 0712.92 22.85 1924 3492 672 3.47 212 1.13 1.23 3.07 5.458 1612.92 1550 13.05 32.62 4.26 1.89 1.38 0.99 0.93 1.91 4.599 26.12.92 10.95 9.22

10 040193 16.10 13.56 22.25 302 1.72 0.85 0.44 1.38 3.93 2.1911 14.01.93 17.35 14.61 17.98 263 1.27 0.87 0.50 1.11 255 2.3612 23.01.93 19.55 16.46 29.03 478 2.17 1.65 0.96 1.28 2.26 3.7313 02.02.93 15235 128.29 25.81 3311 16.31 11.64 5.16 7.90 3.16 4.1914 110293 82.55 6952 28.88 20.08 12.92 5.51 164 4.47 787 4.4915 21.02.93 48.65 40.97 24.84 1017 5.96 2.89 1.33 2.78 4.49 3.6616 02.03.93 131.45 110.69 23.06 2552 13.95 7.49 4.09 7.13 3.41 3.5817 12.0393 140.35 118.19 36.90 4361 24.65 13.52 5.44 7.94 4.53 5.4918 21.03.93 22.10 18.61 21.89 407 1.96 1.13 0.98 1.60 2.00 2.5419 3103.93 116.05 97.73 34.15 33.37 17.38 10.51 5.48 7.48 3.17 4.4620 09.04.93 3165 26.65 32.10 8.55 4.06 2.69 1.80 1.73 2.25 4.96

CI3m 1 23.04.93 53.80 25.62 38.89 9.96 4.37 3.52 2.07 1.48 2.11 6.732 14.05.93 63.00 30.00 34.17 10.25 5.74 3.28 1.23 3.17 4.68 3.233 04.06.93 53.40 25.43 43.51 11.06 6.19 3.63 1.25 2.10 4.94 5.274 25.06.93 28.30 13.48 33.42 4.50 2.67 118 0.65 1.66 4.10 2.725 16.07.93 22.40 10.676 06.08.93 25.05 11.93 47.73 5.69 3.15 1.84 0.70 2.37 4.48 2.40

2708.93CI40 1 13.06.94

2 17.06.94 6.81 7.923 21.06944 25.06.94 16.15 18.785 3006.946 04.07.94 42.15 49.017 08.07.948 13.07.94 21.10 24.539 170794

10 21.07.94 15.80 18.3711 26.07.9412 30.07.94 13.20 15.3513 03.08.9414 07.0894 11.35 13.2015 12.08.9416 16.08.94 57.75 67.1517 2008.9418 25.0894 4.45 5.1719 2908.9420 02.0994 8.60 10.00

CI30 1 23.04.932 1405.933 04.06.934 25.06.935 16.07.936 060893

152

Tab. 11: Korngrößendaten der karbonat- und Corg-freien Fraktion 6-63 [Jm in den oberen Fallen der Station

Falle Probe Max (~m) Md (50%) 01 (25%) 03 (75%) Mean >6 Mode >6 6-11 ~m 11-20 ~m 20-63 ~m

>6 (~m) >6 (~m) >6 (~m) (~m) (~m) (%) (%) (%)

CI10 1 31 10.61 15.64 11.21 13.16 6.26 55.68 28.68 15.64

2 51.5 12.32 17.57 12.60 15.42 8.33 44.82 34.09 21.093 37.5 10.12 14.15 10.30 12.88 6.26 59.20 26.90 13.904 35.5 11.09 16.49 10.87 13.38 8.95 50.83 34.47 14.705 22.5 8.78 14.64 8.14 10.65 6.26 72.36 23.12 4.526 28.5 9.92 14.57 10.01 12.22 6.26 59.74 29.45 10.817 25.5 9.89 15.93 10.06 11.92 6.26 59.95 33.02 7.038 38 10.47 16.37 10.35 12.82 7.76 55.27 3208 12659 27.5 9.74 15.75 9.68 11.90 6.26 61.41 2987 872

10 31.5 10.31 1620 10.45 12.82 6.26 56.32 31.30 12.391112 51 12.85 16.53 13.17 15.53 8.33 42.70 33.77 23.5313 45.5 12.14 16.99 12.19 14.48 6.73 46.76 34.82 18.4314 40 11.07 14.98 11.52 13.87 6.73 52.19 31.31 16.5015 43.5 12.37 15.77 1305 15.34 8.33 45.31 3217 22.5116 31 10.01 14.52 10.93 12.64 6.26 59.94 28.98 110817 43 11.48 16.35 11.55 13.90 8.33 48.46 35.34 162018 38.5 11.68 16.87 12.39 14.59 8.33 47.63 32.95 19.4119 33.5 10.04 16.98 10.32 12.62 8.33 58.21 29.17 12.6220 47 11.34 15.65 11.80 14.21 6.73 49.12 32.44 18.45

CI20 1 32.5 10.22 16.00 9.85 12.46 6.73 5816 30.52 11322 36.5 10.40 16.26 10.42 12.84 7.76 55.13 3164 13.233 35.5 11.38 15.50 1180 13.99 6.73 49.35 32.28 18.384 41.5 11.54 15.87 13.28 14.83 6.73 49.51 27.37 23.125 39.5 12.80 16.47 12.99 15.32 8.33 43.35 33.27 23.386 48 11.95 17.60 13.07 15.14 8.95 46.16 30.71 23.127 37.5 11.16 16.97 11.46 13.71 7.22 51.70 31.48 16.828 45.5 12.56 17.89 12.95 15.42 7.76 44.37 32.39 23.249 51 11.09 15.87 11.54 14.19 6.73 50.18 32.13 17.69

10 36 10.24 15.95 10.61 13.14 6.73 57.12 2833 14.5511 40 11.44 16.61 12.04 14.59 8.33 48.16 32.94 18.8812 41.5 12.13 18.05 12.36 14.64 8.95 45.31 34.60 20.0913 35 11.13 16.59 11.22 13.63 8.33 49.24 35.17 15.5914 33.5 9.44 15.31 9.08 11.62 6.73 64.37 27.45 81815 41.5 9.86 15.12 10.49 12.85 8.33 58.55 28.41 13.0416 42.5 10.63 16.47 1111 13.56 6.26 54.64 29.35 160117 35 10.54 16.39 10.20 12.80 8.33 56.53 30.99 12.4818 40 11.78 16.68 13.23 15.10 8.95 48.20 2776 24.0419 48 11.05 15.08 11.28 1387 7.76 52.08 31.50 16.4220 47 11.81 16.59 12.58 15.09 6.73 47.41 31.50 2109

CI3m 1 38.5 12.03 16.55 12.47 14.73 12.79 43.90 35.33 20.772 35.5 10.21 16.17 10.34 12.61 7.76 56.03 31.99 11.973 37.5 10.38 16.14 10.41 12.70 7.76 55.91 32.77 11.324 49 9.92 15.42 10.40 13.23 7.76 59.33 26.21 14.465 41 9.82 14.42 9.79 12.28 7.76 59.81 2983 10356 39 10.31 17.19 1013 12.84 7.76 55.32 32.33 12.34

CI40 1 37 11.03 15.19 11.39 13.44 6.73 51.79 33.39 14.822 40.5 11.19 16.18 11.43 13.93 8.95 50.13 34.48 15.403 35.5 10.23 1701 10.45 12.98 8.95 55.42 31.35 13.234 29 972 15.27 9.40 11.93 8.33 60.54 3023 9.225 30 9.65 14.68 9.34 11.64 8.33 63.40 29.04 7.566 29.5 9.30 15.82 8.98 11.36 7.76 64.60 28.26 7.147 40.5 8.98 14.75 862 11.34 8.33 66.71 25.99 7.308 42 11.81 17.32 12.75 15.07 8.33 47.36 31.56 21.089 47 1189 1794 12.66 15.17 9.61 45.82 32.68 21.50

10 33 10.39 16.93 10.28 1282 8.33 55.33 31.33 13.3411 32 10.41 16.90 10.26 12.64 8.33 55.89 32.88 11.2312 38.5 11.46 15.94 12.03 14.33 8.33 49.98 31.23 187913 43 10.73 16.96 1130 13.74 6.73 53.36 29.49 171614 41 10.37 16.26 11.16 13.49 8.33 55.58 28.00 16.4315 415 12.02 16.77 11.67 14.20 8.33 45.62 37.11 172716 44.5 11.75 16.52 12.61 15.06 8.33 48.05 30.45 215017 36.5 10.24 16.54 10.22 12.76 8.33 57.81 30.15 12.0418 42 10.37 16.34 10.68 12.99 6.26 55.97 31.21 12.8119 40.5 10.33 16.78 10.10 12.83 8.33 57.47 2986 12.6720 36 9.75 14.84 9.59 1236 8.33 61.47 26.33 1220

CI30 1 32.5 10.27 16.19 9.98 12.26 7.76 57.22 33.32 9.452 37.5 12.00 17.94 12.15 14.56 8.33 45.37 35.78 18853 30 10.51 15.67 10.65 12.85 6.26 55.32 29.94 14.744 23 8.83 14.04 8.28 10.57 6.26 69.55 26.36 4.095 28.5 8.99 14.53 8.57 11.18 7.76 65.90 27.24 6.876 53.5 15.90 19.46 15.66 18.90 21.07 32.24 31.92 35.84

Anhang 153

Tab. 12: Flußraten (mg I m' d) von Gesamt. Lithogen und organischem Kohlenstoff in den unteren Fallen der Station CI

Falle Probe Zeitintervall Einwaage Total Flux % Lithogen Lith Flux Flux6·11 Flux 11·20 Flux 20·63 Corg Flux Ratio 6- Ratio(mg) 11120-63 LithlCorg

CI1u 1 11.12.91 44.65 23,42 37.29 8.73 4.23 2.72 1.79 0.93 2.36 9,40

2 2612.91 6.25 3.28 36,46 1.20 0.71 0.31 0.17 0.20 423 5.85

3 10.01.92 13.05 6.85 31,46 2.15 1.03 0.65 0,47 0.37 2.22 5.87

4 25.01.92 56.10 29,43 35.79 10.53 4.86 3.57 2.11 1.34 2.30 7.875 10.02.92 42.70 22,40 38.19 8.55 4.05 3.11 1,40 0.91 2.90 9.40

6 25.02.92 126.50 66.36 44.90 29.79 14.21 10.37 5.21 2.54 273 11.72

7 11.03.92 90.80 4763 33.65 16.03 8.68 5.23 2.11 2.50 4.11 6,42

8 2603.92 358.20 187.91 32.89 61.80 31.74 21.87 8.19 9.36 388 6.61

9 11.04.92 76.35 40.05 39.31 15.74 7,48 5.32 2.94 1.52 2.54 103810 26.04.92 82.60 43.33 40.11 17.38 8.29 5.78 3.31 1.57 2.50 11.0711 11.05.92 133.05 69.80 47.04 32.83 13.06 10.80 8.98 1.99 1,45 16.5112 2605.92 112.95 59.25 4764 28.23 1131 8.95 7.97 2.00 1,42 14.1513 11.06.92 69.15 36.28 47.81 17.34 7.58 5.98 3.79 1.20 2.00 14,42

14 26.06.92 78.05 40.94 44.23 18.11 6.94 5.75 5.42 185 1.28 9.7815 11.07.92 137,40 7208 45.66 32.91 14.26 11.22 7,43 6,41 1.92 5.1316 26.0792 166.85 87.53 4790 41.92 1781 13.20 10.92 6.69 1.63 6.2717 11.08.92 69.55 36,49 42.81 15.62 6.78 5.16 3.68 1.90 1.84 8.2418 26.08.92 37.35 19.59 41.55 8.14 365 2.70 1.79 0.99 2.04 8.2519 1009.92 49.90 26.18 44.37 1161 5.75 3.91 1.96 1,47 2.94 7.9220 2509.92 38.15 20.01 39.26 7.86 3,46 2.59 1.81 0.98 1.91 7.98

CI2u 1 11.10.92 61.35 51.66 46.50 24.02 11.27 7.82 4.93 2.07 2.29 11592 20.10.92 71.35 60.08 47.02 28.25 1400 9.36 4.90 1.97 2.86 14333 30.10.92 69.70 58.69 44.29 26.00 11.71 8,45 5.84 1.88 2.01 13.814 08.11.92 54.10 45.56 41.58 18.94 9.24 5.77 3.93 1,45 2.35 13.075 1811.92 57.95 48.80 42.77 20.87 9.35 7.32 4.20 1,49 2.22 14.026 27.1192 36.15 30,44 44.74 1362 6.62 4.24 2.76 103 2,40 13.267 07.12.92 43.50 36.63 43.55 15.95 8.64 4.58 2.73 1,40 3.16 11,408 16.1292 3685 3103 4202 13.04 5.89 3.88 3.27 1.05 1.80 12.369 2612.92 33.65 28.34 40.16 11.38 5.66 3.59 2.13 0.91 2.65 12.5210 04.01.93 31,40 26,44 36,41 9.63 5.15 3.15 132 1.08 3.89 89211 140193 52.55 4425 35.36 15.65 7.32 5.13 3.19 2.00 2.29 7.8412 23.01.93 97.10 81.77 41.08 33.59 15,41 11.16 7.02 3.63 2.20 92613 02.0293 156.95 132.17 38.78 51.26 21.17 18.21 11.88 5.23 1.78 9.8014 1102.93 174.55 146.99 38.67 56.84 26.18 1910 11.57 5.73 2.26 9.9215 21.02.93 115.05 96.88 40.84 39.57 18.07 13.10 8,40 4,44 2.15 89216 02.03.93 195.70 164.80 3127 51.54 24.92 16,40 10.21 8.36 2,44 6.1617 12.03.93 139.50 117,47 40.26 4730 23.83 15.77 770 5.09 309 9.2918 21.03.93 122.00 102.74 42.31 43,47 21.52 13.34 8.61 4.35 2.50 10.0019 31.03.93 128.55 108.25 32.26 34.92 18.62 11,44 4.86 5.50 3.83 6.3520 09.04.93 78.35 65.98 37.82 24.95 12.57 8.21 4.17 2.74 3.01 9.11

CI3u 1 2304.93 46.55 22.17 29,40 6.52 3.93 1.92 0.67 0.96 5.91 6.762 14.05.93 112.95 53.79 39.85 21,43 11.96 6.57 2.90 2.04 4.13 10.503 0406.93 103.75 49,40 42.01 20.76 10,41 6.66 3.68 1.74 2.82 11.924 25.06.93 103.40 49.24 35.87 17.66 9.83 4,48 3.35 1.63 2.93 10845 16.0793 97.05 46.21 34.22 15.81 8.80 4.83 219 1.52 4.02 10,406 06.08.93 82.35 39.21 37.28 14.62 8,44 4.78 1,40 1,40 6.01 10,487 27.08.93 64.95 30.93 32,43 10.03 5.50 3.23 1.30 1.23 4.22 8.198 17.0993 51.20 24.38 28.18 6.87 3.83 1.98 1.06 0.87 3.61 7.889 08.10.93 49.80 2371 31.87 7.56 4.76 1.85 0.94 0.90 5.ü7 8.3910 29.1093 53.35 25,40 2956 7.51 4.54 219 0.79 1.30 5.78 5.7711 19.11.93 2230 10.6212 10.12.93 25.80 122913 3112.93 93.75 44.64 2311 10.32 6.54 2.73 1.05 3.30 6.22 3.1314 21.01.94 75,45 35.93 23.86 8.57 5.28 2.39 0.90 1.77 5.84 4.8515 1102.94 153.80 73.24 21.27 15.58 9.58 3.77 2.22 3.38 4.31 4.6116 04.03.94 255.50 121.67 26,48 32.22 17.68 9.54 4.99 6.83 3.55 4.7217 25.03.94 108.15 51.50 18,47 9.51 5.89 2.97 0.65 2,40 9.11 3.9718 15.0494 102.20 48.67 23.69 11.53 7.20 3.13 1.20 3.01 5.99 38319 06.05.94 177.55 34.14 30.73 10,49 7.33 2,40 0.77 2.16 9.54 4.85

CI4u 1 13.06.942 17.06.94 17.95 20.87 44.66 9.32 5.78 2,42 1.13 1.73 512 5,403 21.06.944 25.06.94 25.90 3012 4006 1206 7.58 3.22 1.27 2.08 5.95 5.795 30.06.946 04.07.94 25.50 29.65 50.70 15.03 9.80 3.91 1.32 2.32 7,45 6,487 08.07948 13.07.94 51.30 59.65 43.87 26.17 16.12 7.02 3.03 3.38 5.33 7.739 17.07.94

10 21.07.94 52.35 60.87 42.95 26.15 1638 7.37 2.39 3.33 6.84 7.8511 26.079412 3007.94 58,40 67.91 46.44 31.53 17.97 8.94 4.63 3.69 3.88 8.5513 03.08.9414 07.08.94 69.85 8122 40.38 32.80 17,46 8.82 6.51 3.55 2.68 9.2415 12.08.9416 16.08.94 47.60 55.35 39.16 21.68 11.59 6.99 3.10 2.16 3.73 10.0117 20.08.9418 25.08.94 44.15 51.34 39.76 20,41 10.26 6.09 406 234 2.53 8.7219 29.08.9420 02.09.94 29.60 34,42 36,43 12.54 6.31 3.95 2.27 1,47 2.78 8.55

154

Tab. 13: Korngrößendaten der karbonat- und Corg-freien Fraktion 6-63 fJm in den unteren Fallen der Station CI

Falle Probe Max (~m) Md (50%) 01 (25%) 03 (75%) Mean >6 Mode> 6 6-11 ~m 11-20 ~m 20-63 ~rn

>6 (~m) >6 (~rn) >6 (~m) (~rn) (~m) (%) (%) (%)

CI1u 1 47.00 11.68 16.41 12.47 15.02 6.26 48.39 3109 2052

2 40.00 9.75 14.64 10.42 12.99 6.73 59.62 2628 14.10

3 46.00 11.83 16.27 12.69 15.18 6.73 47.99 30.35 21.66

4 38.50 1187 17.72 12.26 14.70 8.95 46.10 33.85 2005

5 42.50 11.76 16.72 1155 14.41 8.33 4729 36.39 16.32

6 36.50 11.73 17.23 11.81 14.25 8.33 4770 34.81 17.49

7 44.50 10.57 16.01 10.52 13.12 8.33 54.15 32.66 13.19

8 37.50 11.23 15.36 10.70 13.30 8.33 51.37 35.38 13.25

9 37.00 1175 17.59 11.91 14.26 833 4752 33.78 18.70

10 43.50 11.72 17.12 12.09 14.39 6.26 47.69 33.27 1904

11 50.00 13.67 18.24 13.82 16.43 8.95 3977 32.89 2735

12 55.00 13.68 17.72 14.29 17.00 8.33 40.06 31.71 2823

13 49.00 12.53 16.46 12.83 15.79 8.33 43.71 34.45 21.84

14 44.50 13.83 17.96 14.54 16.95 22.63 38.33 31.74 29.93

15 46.00 12.46 16.34 13.03 15.53 8.95 43.33 34.09 22.57

16 43.00 12.79 16.35 13.63 15.68 8.33 42.48 31.48 2604

17 42.50 1279 16.75 13.10 15.43 21.07 43.40 33.06 23.53

18 54.00 12.48 1783 12.94 15.81 776 44.85 33.16 2199

19 38.50 10.98 16.37 11.56 13.66 1109 49.51 33.63 16.85

20 40.50 12.37 16.58 12.88 15.24 10.32 44.05 32.90 23.05

CI2u 1 43.00 12.08 16.75 12.51 14.77 6.73 4693 32.55 20.53

2 38.00 11.43 16.24 11.60 13.89 6.73 4954 33.12 17.35

3 46.00 12.23 15.84 12.94 15.07 8.33 4503 32.52 22.45

4 4400 11.55 15.97 12.44 14.76 8.95 48.79 30.45 2076

5 39.50 1226 15.98 12.38 14.72 8.33 4478 35.09 20.13

6 39.00 11.44 1759 1217 14.67 8.95 4861 31.15 2024

7 3750 10.68 1615 11.34 13.44 6.73 54.16 28.71 1713

8 44.00 12.54 17.17 1355 15.66 6.73 4518 29.74 2509

9 38.00 1157 15.79 12.01 14.37 7.76 49.72 31.55 18.73

10 36.00 1073 1618 10.84 13.19 8.33 5351 32.75 1374

11 40.50 11.97 17.42 12.49 14.86 8.95 46.77 32.82 20.42

12 42.00 11.91 17.75 12.45 14.70 8.33 45.87 33.23 20.90

13 41.50 12.77 1736 1313 15.36 1701 4130 35.52 23.18

14 4000 11.91 17.25 12.47 14.75 8.95 4606 33.59 20.35

15 39.00 12.09 16.73 12.61 15.05 8.33 4566 33.10 21.24

16 39.00 11.64 1589 12.34 14.49 7.76 4835 31.83 19.82

17 42.50 11.23 16.32 11.22 14.13 8.33 5037 33.35 16.28

18 42.00 1135 16.08 12.13 14.64 8.33 49.51 30.68 19.81

19 37.50 10.68 16.17 10.83 13.14 8.95 5332 32.77 1391

20 40.00 1102 15.76 11.53 13.85 9.61 5039 32.89 1672

CI3u 1 37.50 9.85 14.85 9.71 12.18 7.76 60.34 29.46 10.20

2 38.00 10.43 16.45 10.72 13.06 8.33 55.81 30.66 13.53

3 36.00 11.29 15.35 11.62 13.95 6.73 50.14 32.11 1775

4 39.00 10.36 15.66 11.96 13.70 6.73 55.64 25.39 1897

5 37.00 10.31 16.00 10.70 13.10 6.73 55.64 30.53 1383

6 33.50 1023 15.57 9.99 12.33 6.26 57.73 32.66 9.60

7 3550 10.41 16.53 10.72 13.01 6.73 54.81 32.19 13.00

8 32.00 10.31 15.59 1082 13.03 6.73 55.78 2879 15.43

9 35.00 9.79 15.66 9.79 12.44 6.26 63.05 24.52 12.44

10 33.00 9.86 14.96 10.15 12.38 6.26 60.42 2911 10.46

11 37.50 9.68 15.62 9.68 12.14 8.33 62.20 27.19 10.62

12 33.00 9.54 16.05 9.74 12.20 7.76 60.89 27.91 11.21

13 29.00 9.42 14.86 9.79 11.91 6.26 63.40 26.41 10.19

14 36.50 9.63 15.76 9.80 12.35 6.73 61.55 2791 10.54

15 41.00 9.54 15.56 10.30 12.80 6.73 61.52 24.21 14.27

16 41.50 10.61 1663 10.97 13.60 6.26 54.89 29.63 15.48

17 25.50 9.61 14.09 1000 11.85 6.26 61.93 31.27 6.80

18 35.50 9.77 14.91 9.47 12.12 6.26 62.45 2714 10.42

19 30.50 9.03 14.61 8.37 11.04 6.26 69.84 2283 7.32

CI4u 1 37.00 9.54 15.86 11.04 12.98 6.73 61.42 22.48 16.11

2 34.00 9.64 15.48 9.96 12.32 7.22 61.97 25.94 1209

3 32.50 9.79 14.34 9.75 12.33 6.73 62.67 24.71 12.62

4 34.00 9.71 1539 9.65 12.00 6.26 62.80 26.66 10.55

5 32.50 9.53 15.06 9.35 11.99 6.73 63.31 25.99 10.70

6 36.50 9.24 14.67 9.33 11.79 7.22 65.22 2602 8.75

7 31.00 9.19 16.59 9.05 11.59 6.26 6476 26.23 9.01

8 35.00 9.68 14.44 9.70 12.19 7.76 6162 26.81 1156

9 4·150 9.37 15.02 9.58 12.18 7.22 6364 25.40 10.96

10 39.00 9.61 15.41 9.42 11.94 6.26 6265 28.19 9.16

11 36.50 10.09 15.85 10.78 1291 8.33 5783 28.23 13.94

12 35.50 10.49 16.17 10.78 13.07 6.26 56.98 28.34 14.68

13 36.50 10.19 14.79 11.02 13.19 6.26 58.87 25.30 15.83

14 42.50 11.02 15.56 12.06 14.28 8.33 53.24 2691 1985

15 40.00 10.38 15.60 11.10 1353 6.73 5712 27.20 15.69

16 35.00 10.74 1710 10.78 13.29 7.76 53.45 32.23 14.32

17 38.50 11.00 15.19 1202 14.13 8.33 51.39 2923 19.38

18 41.00 11.32 15.25 12.13 14.71 8.33 50.27 29.85 19.88

19 3850 10.90 17.15 11.29 13.72 7.76 53.07 30.04 16.89

20 4100 11.23 15.42 11.67 14.36 6.73 50.36 31.53 1811

Anhang

Tab. 14: Flußraten (mg I m2 d) von Gesamt, Lithogen und organischem Kohlenstoff in den oberen Fallen der Station eB

Falle Probe Zeitintervall Einwaage Total Flux % Lithogen Lith Flux Flux 6-11 Flux 11-20 Flux 20-63 Corg Flux Ratio Ratio(mg) 6-11/20-63 Lith/Corg

C830 1 29.04.90 723.55 269.23 39.40 106.07 31.24 38.92 35.91 14.45 0.87 7.342 20.05.90 81365 302.75 2033 6155 17.16 22.21 22.18 12.91 077 4773 11.06.90 295.10 109.80 3873 42.52 14.41 15.41 12.70 6.41 1.14 6.634 0207.90 232.35 86.46 33.29 28.78 1306 10.21 551 4.31 2.37 6.675 24.07.90 404.55 150.53 1238 18.63 9.03 6.15 3.45 7.42 2.61 2.516 14.08.90 187.50 69.77 22.54 15.73 6.66 5.49 3.58 3.57 186 4417 05.09.90 169.95 63.24 31.44 19.88 6.70 6.82 6.36 3.43 1.05 5.798 26.09.90 289.75 107.81 3861 41.63 12.71 14.99 13.93 6.19 0.91 6.739 18.10.90 194.70 72.45 3723 26.97 8.09 9.95 8.93 4.37 0.91 6.1810 08.11.90 199.85 74.36 2765 20.56 7.38 7.57 5.61 5.67 1.32 3.6211 30.11.90 223.30 83.09 42.73 35.51 12.75 12.96 9.80 6.22 1.30 5.7012 21.12.90 292.55 108.86 3313 36.07 14.81 13.35 791 8.06 1.87 4.4713 12.01.91 163.20 60.73 2288 13.90 5.56 4.60 374 4.71 1.49 2.9514 0202.91 213.50 79.44 26.20 20.82 8.63 7.35 4.84 9.43 1.79 2.2115 24.02.91 304.50 113.30 1758 19.92 8.55 7.67 3.71 14.79 2.30 1.3516 17.03.91 114.95 42.77 30.11 12.88 6.15 4.18 255 7.51 2.41 1.7';17 08.04.91 108.65 40.43 33.36 13.49 6.47 4.30 2.72 7.71 2.38 17518 29.04.91 54.20 20.17 2462 4.97 2.78 1.54 0.65 4.31 4.28 1.1519

C840 1 13.05.91 29.40 2352 22.00 5.17 311 1.38 0.68 3.62 4.56 1.432 23.05.91 71.75 57.40 2421 13.90 7.56 4.49 184 7.61 4.10 1.833 02.06.91 116.50 93.20 1063 9.91 4.40 3.57 1.94 1158 2.27 0.864 12.06.91 154.55 123.64 25.70 31.77 16.17 9.47 6.13 14.59 2.64 2185 22.06.91 179.90 143.92 3272 47.09 27.84 1320 6.05 15.77 4.60 2.996 02.0791 203.95 163.16 1958 31.94 20.90 8.16 2.88 13.44 725 2.387 12.07.91 103.25 82.60 32.26 26.65 17.29 6.76 2.60 8.86 6.65 3.018 22.07.91 78.80 6304 3798 2394 13.84 7.59 2.52 6.93 5.49 3.469 0108.91 245.90 196.72 2968 58.38 37.11 16.65 4.62 23.44 804 2.4910 11.08.91 200.15 160.12 30.93 49.53 30.83 13.77 4.92 18.65 6.27 2.6611 21.08.91 402.35 321.88 3390 109.12 56.84 32.58 1970 48.41 2.89 2.2512 31.08.91 241.40 193.12 31.80 61.42 38.32 18.43 4.67 35.69 8.21 1.7213 10.09.91 101.85 81.48 3511 28.61 15.97 9.54 3.10 10.23 5.15 2.8014 20.09.91 91.20 72.96 2186 15.95 8.67 5.15 2.13 7.78 4.07 2.0515 30.09.91 99.00 79.20 17.40 13.78 7.77 4.14 1.88 6.55 4.14 2.1016 10.10.91 93.90 75.12 8.44 634 3.59 2.05 0.70 4.18 512 1.5217 20.10.91 35.30 28.24 3253 9.19 5.24 2.98 0.97 4.50 5.42 2.0418 30.10.91 14.15 11.32 17.85 2.02 1.07 0.69 0.26 1.20 4.06 1.6819 09.11.91 13.75 11.00 2809 3.09 1.60 0.99 051 1.28 3.14 2.4120 19.11.91 24.25 19.40 23.79 4.61 2.45 152 0.65 1.64 3.78 281

155

156

Tab. 15: Korngrößendaten der karbonat- und Corg-freien Fraktion 6-63 IJm in den oberen Fallen der Statio

Falle Probe Max (~m) Md (50%) 01 (25%) 03 (75%) Mean >6 Mode >6 6-11 ~m 11-20 ~m 20-63 ~m

>6 (~m) >6 (~m) >6 (~m) (~m) (~m) (%) (%) (%)

CS 30 1 4301 16.63 20.31 15.31 18.37 19.62 29.45 36.69 33852 46.19 16.85 21.45 15.46 18.48 21.07 27.89 36.08 36033 37.29 14.94 19.28 14.53 16.96 15.84 33.90 36.24 29.864 32.33 12.14 16.16 12.28 14.49 8.33 45.39 35.47 19.135 4301 11.70 16.51 12.11 14.39 7.76 48.45 33.02 18546 53.27 12.75 17.78 13.08 16.17 8.33 42.34 34.90 22.767 49.61 15.32 18.94 14.87 18.04 17.01 3372 34.31 31978 53.27 15.92 20.16 1499 18.11 21.07 30.53 36.00 33.479 4301 16.00 19.98 15.32 18.28 18.27 29.98 36.90 33.12

10 46.19 14.15 18.98 14.05 16.88 13.73 35.91 36.81 27.2811 46.19 14.39 19.06 14.23 17.01 19.62 35.90 36.49 276112 40.05 12.85 18.08 12.82 15.53 1475 41.07 37.01 21.9313 43.01 13.77 17.88 13.73 16.43 21.07 40.01 33.11 26.8814 49.61 12.93 17.96 13.27 15.88 12.79 41.48 35.29 23.2315 40.05 12.72 17.69 1230 15.08 1962 42.89 38.50 18.6116 37.29 11.80 16.52 12.43 14.40 8.33 47.76 32.43 19.8217 43.01 11.78 16.31 12.07 14.97 8.33 4798 31.87 20.1518 37.29 10.30 15.69 10.40 12.98 8.33 55.95 30.96 130919 49.61 13.27 17.69 13.00 1583 21.07 40.50 36.80 22.70

CS 40 1 40.05 9.94 15.24 1016 12.64 6.73 60.11 26.72 13.172 32.33 10.66 15.03 10.52 1316 10.32 54.39 32.33 13.273 46.19 11.99 16.55 12.40 14.78 1191 44.36 36.07 19564 43.01 1126 15.16 11.96 14.41 8.33 50.89 29.81 19.305 30.10 9.86 14.33 10.12 12.38 673 59.12 2804 12846 40.05 9.51 15.07 8.79 11.76 6.73 65.44 25.54 9.037 5327 9.11 14.58 9.15 1204 6.73 64.89 25.36 9.758 34.72 10.35 16.75 9.98 12.57 9.61 57.79 31.69 10.529 32.33 9.78 14.51 9.18 11.78 8.33 63.57 28.51 7.91

10 5327 9.25 15.52 9.29 12.62 7.76 62.26 2781 9.9311 40.05 11.19 16.37 1162 14.42 8.33 52.09 29.85 18.0612 28.03 9.80 15.43 9.41 11.87 7.76 62.39 30.01 7.6013 32.33 10.40 16.74 10.46 12.76 8.33 55.83 33.34 108314 37.29 10.66 15.86 10.70 13.28 8.33 54.36 32.30 133415 3729 10.29 15.72 1072 13.37 7.76 56.36 30.02 13.6216 37.29 10.28 16.19 10.57 13.00 6.26 56.56 32.39 11.0517 43.01 10.36 17.07 10.01 12.79 8.33 57.05 32.43 10.5218 34.72 10.98 15.00 10.80 1325 7.76 52.70 34.30 12.9919 43.01 11.07 15.23 11.50 13.94 8.95 51.65 31.90 16.4420 43.01 10.88 16.13 10.83 13.53 7.76 53.07 32.88 1405

Anhang 157

Tab. 16: Flußraten (mg I m' d) von Gesamt, Lithogen und organischem Kohlenstoff in den unteren Fallen der Station es

Falle Probe Zeitintervall Einwaage Total Flux % Lithogen Lilh Flux Flux6-l1 Flux 11-20 Flux 20·63 Corg Flux Ratio 6- Ratio(mg) 11/20·63 Lith/Corg

CSl 1 18.04.88 1012.00 299.85 52.34 15693 46.87 57.98 52.08 11.89 0.90 13.202 15.05.88 471.00 139.56 49.99 69.76 2199 25.23 22.54 5.13 0.98 13.593 11.06.88 472.50 140.00 40.74 5704 15.18 2043 2144 481 071 11.874 08.07.88 487.50 14444 4752 68.64 22.16 24.93 21.56 5.35 1.03 12.835 04.08.88 1219.50 361.33 34.60 12502 43.13 43.08 38.81 23.75 1.11 5.266 31.08.88 823.00 243.85 32.24 78.62 25.54 29.00 24.08 7.65 1.06 10.287 27.09.88 346.50 102.67 40.33 4140 12.97 14.97 1346 2.96 0.96 13.998 24.10.88 609.00 18044 3705 66.85 19.81 23.66 23.38 4.73 0.85 14.129 2011.88 789.00 23378 38.21 89.34 2442 31.85 33.07 10.03 0.74 8.9010 17.12.88 367.50 108.89 4245 4622 14.57 16.09 15.55 4.08 0.94 11.3411 13.01.89 365.00 108.15 41.73 45.13 1589 16.31 12.93 4.03 1.23 11.2012 09.02.89 598.50 177.33 40.33 71.52 2638 2741 17.73 7.31 149 9.7813 08.03.89 474.50 140.59 38.97 5479 2249 2117 11.12 5.37 2.02 10.20

CS2u 1 01.04.89 63640 127.98 32.06 41.03 19.50 15.32 620 3.10 3.14 13.222 18.04.89 705.25 141.83 3637 5158 23.92 19.77 7.89 449 3.03 11.483 05.05.89 790.10 158.89 34.12 54.21 30.55 16.62 7.04 545 4.34 9.954 22005.89 685.50 137.86 3783 52.15 22.06 2043 9.66 4.47 2.28 11.665 08.06.89 512.10 102.99 39.39 40.56 18.01 1548 7.07 3.19 2.55 12.716 25.06.89 447.00 89.89 39.65 35.64 1543 13.20 702 2.82 2.20 12.657 12.07.89 449.05 90.31 36.23 3272 14.68 12.37 5.67 2.58 2.59 12.688 29.07.89 528.85 106.35 37.18 39.54 18.62 1444 649 3.03 287 13.079 15.08.89 658.50 13243 31.38 41.55 19.83 14.58 7.15 3.66 2.77 11.3610 0109.89 1223.20 245.99 38.88 9565 36.68 3340 25.57 7.65 143 12.5011 18.09.89 1125.15 22627 34.12 77.21 2745 26.39 23.37 5.84 1.17 13.2312 05.10.89 94040 189.12 32.70 61.83 23.21 21.03 17.60 5.75 1.32 10.7513 22.10.89 87525 176.02 31.88 56.11 22.77 18.75 14.59 4.68 1.56 11.9914 08.11.89 786.75 158.22 41.19 65.18 24.90 2121 19.07 4.84 1.31 134615 25.11.89 82000 164.91 44.32 7309 2970 24.38 19.00 5.00 1.56 14.6116 12.12.89 775.50 155.96 46.12 71.93 2743 26.10 1840 490 149 14.6917 29.12.89 567.95 114.22 44.81 51.19 18.72 18.98 1349 3.15 1.39 16.2718 15.01.90 50015 100.58 43.93 44.19 15.73 14.66 13.80 3.11 1.14 14.2019 0102.90 990.15 199.13 3929 78.24 2745 2644 24.36 7.03 113 11.1320 18.02.90 1083.90 217.98 42.90 93.52 32.10 31.90 29.52 7.76 1.09 12.0621 07.03.90 910.15 18304 35.17 64.37 25.50 1942 1945 5.47 1.31 11.7722 24.03.90 62740 126.17 37.74 47.62 20.71 16.32 10.59 3.83 1.96 12.42

CS 3u 2 2005.90 77045 286.68 2886 82.75 29.63 30.17 22.96 801 1.29 10.333 10.06.90 440.05 163.74 43.46 71.16 24.23 23.86 23.07 5.58 1.05 12.754 02.07.90 374.70 139.42 4426 61.71 19.82 22.31 19.59 4.36 1.01 14.165 23.0790 475.50 176.93 30.36 53.72 20.87 1828 14.57 4.59 143 11.706 14.08.90 45855 170.62 29.94 51.08 17.67 17.91 15.50 4.81 1.14 10.627 04.09.90 331.90 123.50 42.52 52.51 17.15 17.97 17.38 3.97 099 13.228 26.09.90 384.15 142.94 41.79 59.73 19.51 19.58 20.64 4.89 0.95 12.209 17.1090 350.15 130.29 43.89 57.19 20.26 18.81 18.12 5.04 1.12 11.35

10 08.11.90 299.90 111.59 46.69 52.11 18.58 17.55 15.97 4.90 1.16 10.6411 29.11.90 24340 90.57 49.16 4453 1748 15.96 11.09 4.16 1.58 10.7212 21.12.90 32620 121.38 44.00 53.40 21.16 17.68 14.56 6.43 1.45 8.3113 11.01.91 351.60 130.83 43.30 56.64 23.57 20.54 12.54 5.67 1.88 9.9914 02.02.91 393.75 146.51 38.63 56.60 2460 20.38 11.62 10.52 212 5.3815 23.02.91 371.95 138.40 35.30 48.86 20.21 17.02 11.62 11.99 1.74 4.0716 17.03.91 244.35 90.92 4385 39.87 18.31 1332 8.25 5.93 2.22 6.7317 07.04.91 189.80 7062 38.39 27.11 1168 9.35 6.08 4.72 192 5.74

CS 4u 1 13.05.91 10.20 8.16 28.50 233 1.30 0.74 0.29 0.67 4.45 3462 23.05.91 59.55 47.64 40.44 19.27 8.11 6.33 4.82 3.52 1.68 5473 02.06.91 17125 137.00 32.37 44.35 23.14 15.18 6.03 11.75 3.84 3784 12.06.91 110.35 88.28 37.25 3289 17.67 10.18 504 7.35 3.51 4475 22.06.91 66.80 5344 37.52 20.05 1153 6.09 242 4.25 4.76 4.726 02.07.91 130.05 104.04 36.39 37.86 23.06 10.66 4.14 7.81 5.57 4.857 1207.91 9745 77.96 36.19 28.21 16.29 8.51 341 4.86 4.78 5.818 22.07.91 49.15 39.32 39.39 1549 728 5.00 321 3.03 227 5.129 0108.91 147.35 117.88 35.91 4233 23.50 13.17 5.66 7.76 4.15 546

10 11.08.91 17870 142.96 33.14 47.37 25.31 14.56 7.50 11.57 337 4.1011 2108.91 127.15 101.72 38.78 3945 19.77 12.62 7.05 9.48 2.80 4.1612 31.0891 209.60 16768 38.52 64.59 32.85 20.63 11.11 22.75 2.96 2.8413 1009.91 85.90 68.72 34.78 23.90 13.82 7.38 271 8.54 5.10 2.8014 20.09.91 106.80 8544 30.30 2589 13.55 7.96 4.38 7.57 3.09 3.4215 30.09.91 88.15 7052 3472 24.49 12.24 8.28 3.96 5.02 3.09 4.8816 10.10.91 97.25 77.80 30.17 2347 1196 7.69 3.82 4.54 3.13 5.1717 20.10.91 41.50 33.20 34.21 11.36 5.53 3.79 2.03 197 2.72 5.7618 30.10.91 48.05 3844 33.95 13.05 6.21 4.65 2.19 2.16 2.83 6.0519 0911.91 3385 27.08 34.62 9.38 4.67 3.09 162 1.59 2.89 5.8820 19.11.91 25.25 20.20 3828 773 3.93 2.65 1.15 1.34 3.42 5.77

158

Tab. 17: Korngrößendaten der karbonat- und Corg-freien Fraktion 6-63 IJm in den unteren Fallen der Station CB

Falle Probe Mex (~m) Md (50%) 01 (25%) 03 (75%) Mean >6 Mode> 6 6-11 ~m 11-20 ~m 20-63 ~m

>6 (~m) >6 (~m) >6 (~m) (~m) (~m) (%) (%) (%)

CS 1 1 49.61 15.92 20.44 15.14 1803 1962 2986 36.95 33.19

2 49.61 1579 20.06 1474 17.85 21.07 31.53 36.17 32.31

3 5721 17.17 21.57 15.97 19.52 22.63 26.60 35.81 3759

4 43.01 15.57 19.57 14.92 1769 24.30 32.28 36.31 31.41

5 43.01 15.18 19.08 14.59 17.46 19.62 34.50 34.46 31.05

6 49.61 15.31 19.24 14.52 17.42 21.07 32.49 36.89 30.62

7 40.05 15.91 20.32 14.92 17.59 2263 3133 3616 32.51

8 57.21 16.59 2012 15.71 18.77 19.62 29.64 35AO 34.97

9 46.19 1709 2121 15.88 1918 22.63 2733 35.65 37.01

10 53.27 15.91 20.60 15.29 1839 24.30 3153 3482 33.65

11 49.61 14.70 1918 14.30 1712 2107 35.20 36.14 28.66

12 43.01 13.82 18.48 13.43 15.95 13.73 36.88 38.32 2480

13 43.01 1278 17.19 12.41 15.33 11.09 4105 38.65 20.30

CS 2u 1 43.01 11.78 18.08 11.46 14.12 9.61 4753 37.35 15.12

2 40.05 1184 16.92 11.54 1414 8.95 4638 38.33 15.29

3 49.61 10.31 16.67 10.39 13.22 8.33 56.35 30.66 12.99

4 34.72 1268 18.19 12.24 14.63 18.27 42.31 39.18 18.52

5 37.29 12.08 17.02 11.77 14.45 8.33 44.40 38.16 17.44

6 40.05 1272 17.41 12.55 14.93 19.62 43.28 37.02 19.70

7 37.29 1212 16.82 12.06 14.41 11.09 44.86 37.80 17.34

8 34.72 1184 16.18 11.73 14.11 9.61 4708 36.51 16.41

9 34.72 1185 17.26 11.83 14.18 8.33 47.71 35.09 1720

10 46.19 13.66 1769 13.91 16.44 13.73 38.35 34.92 26.73

11 53.27 1479 1883 14.57 17.40 22.63 35.55 34.18 30.26

12 57.21 14.10 17.66 14.38 17.15 19.62 37.54 34.00 28.46

13 49.61 13.27 17.58 13.87 16.60 11.09 40.57 33.42 2601

14 5721 1390 17.37 14.58 1717 22.63 38.20 32.54 29.26

15 46.19 13.43 17.84 13.72 16.29 21.07 40.64 33.36 26.00

16 46.19 1371 1806 13.89 16.74 18.27 38.14 36.28 25.58

17 49.61 14.07 18.37 13.79 16.94 19.62 36.57 37.08 26.35

18 53.27 14.74 18.87 15.02 17.94 2107 35.59 33.18 31.23

19 53.27 15.14 18.79 14.92 17.77 19.62 35.08 33.79 31.13

20 49.61 14.97 18.72 15.09 17.75 18.27 34.33 34.11 3157

21 46.19 13.86 17.76 15.00 17.20 8.33 39.61 30.17 30.21

22 46.19 1274 16.33 12.85 15.80 8.33 43.49 34.27 22.24

CS 3u 2 46.19 14.41 18.87 14.05 16.63 2107 35.80 36.46 27.74

3 46.19 15.29 18.54 15.19 17.84 22.63 34.06 33.52 32.42

4 46.19 15.39 19.63 15.25 17.71 19.62 3212 36.14 3174

5 46.19 13.45 1736 13.91 16.58 2263 38.84 34.04 27.12

6 46.19 14.81 18.71 14.59 17.52 21.07 34.59 35.07 30.34

7 61.45 15.34 19.23 15.30 18.24 22.63 32.67 34.23 33.10

8 49.61 15.75 19.36 15.75 18.57 22.63 32.67 32.77 34.56

9 49.61 14.84 19.03 14.89 17.50 22.63 35.42 32.89 31.68

10 40.05 14.81 19.20 14.60 17.23 22.63 35.66 33.69 30.65

11 4301 13.54 1751 13.49 16.17 15.84 39.25 35.85 24.90

12 4619 13.86 1784 14.16 16.72 8.33 39.63 3311 27.26

13 37.29 13.10 17.56 12.83 15.40 18.27 41.61 36.25 22.13

14 37.29 12.61 16.82 12.56 14.91 11.09 43.46 36.01 20.53

15 43.01 13.08 16.80 13.25 15.32 21.07 41.37 34.84 23.78

16 40.05 11.93 16.37 12.14 14.89 10.32 45.92 33.40 20.68

17 37.29 12.51 16.96 12.77 15.01 12.79 43.08 34.49 22.43

CS 4u 1 37.29 10.39 16.95 10.76 13.05 6.73 55.69 31.80 12.51

2 61.45 13.23 17.13 13.56 16.49 8.33 42.10 3287 25.03

3 32.33 1124 15.49 11.39 13.45 8.33 52.17 34.23 13.60

4 34.72 10.79 15.68 11.16 13.27 6.73 53.72 30.96 15.31

5 34.72 10.39 15.87 10.26 12.66 6.73 57.52 30.39 1208

6 24.30 9.80 14.40 10.01 12.06 6.73 60.91 28.16 10.93

7 34.72 10.29 16.26 10.01 12.65 8.33 57.74 30.18 12.08

8 49.61 1201 17.51 12.45 14.92 15.84 47.03 32.27 20.71

9 37.29 10.39 15.59 10.72 13.01 8.33 55.51 31.10 13.38

10 34.72 10.90 16.06 11.24 13.58 833 53.42 30.74 15.84

11 43.01 11.33 15.36 11.93 14.22 6.73 50.13 31.99 17.88

12 40.05 11.20 15.26 11.83 14.09 8.33 50.86 31.95 1719

13 32.33 10.55 16.40 10.12 12.53 7.76 57.80 30.88 1132

14 40.05 11.22 16.13 11.71 13.95 8.33 52.33 3076 16.91

15 43.01 11.33 16.08 11.53 14.12 8.33 49.99 33.82 16.18

16 34.72 11.05 16.54 11.79 13.95 8.95 50.97 32.77 16.26

17 37.29 11.62 16.29 11.65 14.07 7.76 48.71 33.40 17.89

18 4005 11.86 17.22 11.51 14.27 8.33 47.60 35.60 16.80

19 37.29 11.43 15.69 11.85 14.15 7.76 49.76 33.00 17.2420 37.29 11.07 16.89 11.24 13.74 10.32 50.83 34.31 14.86

Anhang 159

Tab. 18: Flußraten (mg I m2 d) von Gesamt, Lithogen und organischem Kohlenstoff in den oberen Fallen der Station CV

Falle Probe Zeitintervall Einwaage Total Flux % Lithogen Lith Flux Flux6-11 Flux 11-20 Flux Corg Flux Ratio Ratio(mg) 20-63 6-11120-63 Lilh/Corg

CV10 1 11.10.92 33.55 48.80 38.75 18.91 10.70 5.48 2.73 6.96 3.92 2.722 20.10.92 75.60 63.66 35.73 2275 13.37 6.88 2.50 5.77 5.34 3.943 30.10.92 159.90 134.65 4296 57.85 23.77 19.87 14.21 12.94 1.67 4.474 08.1192 208.30 175.41 51.09 8962 37.49 2892 23.22 14.22 1.61 6.305 1811.92 152.70 128.59 47.48 61.06 3235 19.00 9.71 10.31 3.33 5.926 27.11.92 219.55 184.88 50.24 9288 4817 33.24 11.47 14.47 4.20 6.427 07.12.92 162.40 136.76 47.57 65.06 38.67 1917 7.21 11.52 5.36 5.658 1612.92 202.65 170.65 4074 6952 3702 23.77 8.73 13.43 4.24 5.189 26.12.92 123.45 103.96 31.17 32.40 17.51 11.02 3.88 7.06 4.52 4.5910 04.01.93 157.85 132.93 38.13 50.69 27.17 1685 6.67 9.92 4.07 5.1111 14.01.93 209.80 176.67 4107 72.55 39.95 24.90 7.70 14.15 5.19 5.1312 23.01.93 228.85 192.72 48.43 93.34 5488 30.81 7.64 13.57 7.18 6.8813 02.02.93 262.10 220.72 4497 99.26 52.82 35.74 10.69 1664 4.94 5.9714 11.02.93 266.10 224.08 3697 8285 48.00 28.47 6.38 16.80 7.53 49315 21.02.93 244.85 206.19 33.77 69.63 3819 24.03 7.41 12.99 5.15 5.3616 02.03.93 176.35 148.51 4189 6221 3800 19.44 4.76 10.56 7.98 5.8917 12.03.93 361.85 304.72 42.25 12875 68.56 45.44 14.74 20.25 4.65 6.3618 21.03.93 177. 75 149.68 3420 51.20 28.66 16.91 5.63 11.54 5.09 4.4419 31.03.93 210.90 177.60 21.88 3885 15.82 1452 8.52 14.83 1.86 2.6220 04.04.93 91.35 162.40 38.97 6328 30.84 22.44 1000 16.25 3.08 3.89

CV20 1 20.04.93 312.15 135.72 58.25 79.06 26.85 25.98 26.23 6.84 1.02 11.552 1505.93 188.60 75.44 51.85 39.12 11.91 11.82 15.39 4.81 0.77 8.143 0906.93 322.10 128.84 58.48 75.34 23.22 21.94 30.18 7.74 0.77 9.734 04.07.93 184.40 73.76 5301 3910 1369 1229 13.13 4.32 1.04 9.065 29.0793 228.90 91.56 5129 46.96 17.36 15.14 14.47 5.63 1.20 8.346 23.08.93 5135 20.54 5487 1127 4.25 3.88 3.14 1.47 1.35 7.687 17.09.93 119.50 47.80 3091 14.78 6.90 4.69 3.18 2.71 2.17 5.458 1210.93 309.35 123.74 47.07 5825 26.40 19.77 12.08 6.58 2.18 8.859 06.11.93 1105 4.42

10 01.12.93 5.40 2.1611 2612.93 4.95 1.9812 20.01.94 5.70 2.2813 14.02.94 2.20 0.8814 11.0394 6.60 2.6415 05.04.94 420 1.6816 30.04.94 5.40 2.1617 25.05.94 3.05 1.2218 19.06.94 1.40 0.5619 14.07.94 2.00 0.8020 25.08.94 1.65 0.66

Tab. 19: Korngrößendaten der karbonat- und Corg-freien Fraktion 6-63 ",m in den oberen Fallen der Station CV

Falle Probe Max (~m) Md (50%) 01 (25%) 03 (75%) >6 Mean >6 Mode >6 6-11 ~m 11-20 ~m 20-63 ~m

>6 (~m) >6 (~m) (~m) (~m) (~m) (%) (%) (%)CV10 1 30.10 10.44 15.82 10.81 13.04 673 56.59 28.99 14.42

2 28.03 10.06 15.45 980 12.17 6.73 58.77 30.23 11.003 37.29 13.27 17.31 13.41 1554 1701 41.09 34.35 24.564 43.01 13.36 1636 1362 15.95 833 41.83 32.27 25.915 34.72 10.98 16.72 11.61 13.51 6.73 52.98 31.11 15.916 34.72 11.09 1518 10.85 1305 8.95 51.86 35.79 12.357 32.33 982 14.45 9.83 12.27 6.73 59.45 29.46 11.098 28.03 10.75 16.88 10.58 12.78 6.73 53.25 34.20 12.559 32.33 10.65 17.04 10.78 12.96 6.73 54.04 3400 11.9610 32.33 10.67 17.21 10.90 13.01 8.33 53.60 33.25 13.1511 34.72 10.29 15.69 10.19 12.64 8.95 55.07 34.32 10.6112 3010 9.95 15.98 9.30 11.90 895 58.80 33.01 8.1913 37.29 10.64 1511 10.24 1269 895 53.22 36.01 10.7714 30.10 10.15 16.39 9.18 11.84 895 5794 34.37 7.7015 28.03 10.32 15.75 1021 12.57 6.73 54.84 34.51 10.6416 28.03 9.80 15.43 923 11.77 961 6109 31.26 7.6517 40.05 10.62 15.52 10.14 13.00 8.33 5325 3530 11.4518 28.03 10.29 16.52 10.08 12.51 8.33 55.98 33.03 10.9919 4005 12.77 18.04 12.69 1500 2430 40.71 37.37 21.9320 34.72 11.41 16.55 11.45 13.77 7.76 48.74 35.46 15.80

CV20 1 43.01 15.75 19.16 14.68 17.22 22.63 33.96 32.86 33.182 4961 1751 20.20 15.67 18.74 2430 30.45 30.21 39.343 43.01 1727 19.67 1589 1859 22.63 3082 29.12 40.064 43.01 15.47 18.38 1495 17.42 21.07 35.01 31.42 33.575 46.19 14.53 18.04 1455 17.11 19.62 3696 32.24 30.806 37.29 14.18 17.99 14.06 16.19 2430 37.70 34.47 27.837 3472 12.30 16.25 12.71 14.71 7.76 46.70 3177 21.538 37.29 12.17 17.61 12.60 14.92 7.76 4532 3394 20.74

160

Tab. 20: Flußraten (mg / m2 d) von Gesamt, Lithogen und organischem Kohlenstoff in den unteren Fallen der Station CV

Falle Probe Zeitintervall Einwaage Total Flux % Lithogen Lilh Flux Flux 6-11 Flux 11-20 Flux 20-63 Corg Flux Ratio Ratio

(mg) 6-11/20-63 Lith/Corg

CV1u 1 11.10.92 124A3 180.99 61.88 111.99 42.78 40.33 28.87 9.52 1A8 11.7G

2 20.1092 179.00 150.74 57.27 86.33 33.18 32.14 21.01 6.08 1.58 1419

3 30.10.92 245.85 207.03 56A8 116.93 43.29 43A1 30.23 10.06 1A3 11.6~'

4 08.1192 157.18 13236 56.95 75.37 22.35 29.32 23.70 5.92 0.94 12.74

5 18.11.92 23215 195A9 53.69 104.97 35.68 37.03 32.26 7.55 1.11 13.90

6 27.11.92 173.52 14612 57.11 83A5 26A7 29.97 27.01 5A6 0.98 15.27

7 07.12.92 146.52 123.39 57.89 71.42 30.16 25.01 16.25 4.87 1.86 14.67

8 1612.92 216.19 182.05 54.71 9960 32.70 37A1 29A9 577 1.11 17.269 26.12.92 172.00 14484 4979 72.12 26.08 24.91 21.13 4.81 1.23 14.9B

10 04.01.93 127.91 107.71 5138 55.35 21AO 19.85 14.10 3.76 1.52 14.7·1

11 14.0193 219.66 184.97 43.79 80.99 31A5 30.71 18.83 4.85 1.67 16.6912 23.01.93 145.72 122.71 44.37 54.44 19.83 20.38 14.23 3.82 1.39 14.24

13 02.02.93 177.90 149.81 55.87 83.70 36.19 29.91 17.60 5.38 2.06 15.5714 1102.93 200.07 168A8 54.92 92.52 38.50 3234 21.69 6.64 178 13.9~1

15 21.02.93 201.07 169.32 46.80 79.25 30A2 31.81 17.01 6.03 1.79 13.1416 02.03.93 174.99 14736 45.21 66.62 28.59 25.12 12.91 4.25 2.22 -15.6717 12.0393 223.66 188.35 47.95 9031 37.30 34.62 18.38 7.06 2.03 12.8018 2103.93 142.21 11976 40.59 48.61 19.57 18.84 10.20 4.81 1.92 10.'11

19 31.03.93 207.00 174.31 45.15 78.70 3117 2977 17.76 7.29 1.75 108020 04.04.93 106AO 18916 4768 90.20 34.52 32.98 22.69 8.37 1.52 10.78

CV2u 1 20.04.93 179AO 7800 60.62 4728 17.28 16.99 13.01 3.02 1.33 15.65

2 1505.93 401.05 160A2 62.17 99.73 31.14 33.57 35.02 6.11 0.89 16.313 09.06.93 91A5 36.58 61.28 22.42 8.83 8.33 5.26 150 1.68 14.924 040793 8175 32.70 66.26 21.67 6.62 6.92 8.12 1A5 0.82 14.905 29.07.93 19.95 7.98 70.04 5.59 1.68 1.69 2.22 OAO 0.75 14086 2308.93 36.30 14.52 68.31 9.92 2.90 313 3.88 0.67 0.75 14.787 17.09.93 14935 59.74 66.37 3965 9.11 12.01 18.52 245 0.49 16.198 12.10.93 2700 10.80 66.70 720 2.31 2.34 2.55 0.50 0.91 14.329 06.11.93 1685 6.74 60.91 4.11 135 1.44 1.33 0.30 1.02 136410 011293 2625 10.50 63.52 6.67 1.93 222 2.52 OA7 077 14.1811 2612.93 9.55 3.82 70.97 271 0.88 0.91 0.92 017 0.96 15.5512 2001.94 360 1A413 1402.94 2475 9.90 63.24 626 2.05 1.99 222 0.36 0.92 17.3714 1103.94 13.20 5.28 6076 3.21 101 106 1.14 0.21 0.88 14.9615 05.04.94 4350 17.40 63.62 1107 2.88 3.81 4.38 0.73 066 15.2416 3004.94 2285 9.14 63.11 577 1.56 1.98 2.23 0.37 0.70 155817 2505.94 210 0.8418 19.06.94 180 07219 14.0794 3895 15.58 6414 9.99 2.93 3.48 359 0.67 0.82 148220 25.08.94 2955 11.82 65.67 7.76 2.25 2A5 3.05 OA9 0.74 15.89

Anhang 161

Tab. 21: Korngrößendaten der karbonat- und Corg-freien Fraktion 6-63 ~n1 in den unteren Fallen der Station CV

Falle Probe Max (~m) Md (50%) 01 (25%) 03 (75%) Mean >6 Mode> 6 6-11 ~m 11-20 ~m 20-63 ~m (%)>6 (~m) >6 (~m) >6 (~m) (~m) (~m) (%) (%)

CV1u 1 43.01 13.80 17.27 13.56 16.06 1701 38.20 36.02 25.78

2 37.29 13.53 17.23 13.39 15.73 15.84 3843 37.23 24.33

3 40.05 13.94 17.36 13.41 15.90 21.07 3702 37.13 25.854 40.05 15.49 20.11 14.29 1721 22.63 29.65 38.90 31.445 49.61 14.88 18.71 14.59 17.12 24.30 33.99 35.27 30.73

6 46.19 15.59 19.50 14.84 17.57 22.63 31.72 35.92 32377 49.61 12.64 17.64 12.81 15.22 2107 42.23 35.01 22.76

8 43.01 14.76 18.76 14.18 16.86 22.63 32.83 37.56 29.619 43.01 14.38 18.37 14.36 16.75 26.10 3616 34.54 29.30

10 46.19 13.58 18.17 13.48 16.12 1109 38.66 35.86 25.4811 34.72 13.43 17.92 12.96 15.29 21.07 38.83 37.92 23.2412 40.05 13.81 18.14 13.44 15.96 2107 36.42 37.44 26.1413 46.19 12.25 16.68 12.65 15.09 19.62 43.24 35.73 210314 34.72 12.93 17.34 12.90 15.17 8.33 41.61 34.96 23.4415 34.72 13.25 17.44 12.61 15.29 1475 38.38 40.15 21.4716 34.72 12.43 16.88 12.03 14.51 17.01 42.92 37.71 19.3717 40.05 12.60 17.39 12.39 14.90 1191 41.31 38.34 20.36

18 43.01 12.86 18.47 12.46 15.03 11.91 40.26 38.76 20.9819 37.29 13.00 1815 1280 15.22 17.01 39.60 37.83 22.5720 37.29 13.22 17.04 13.42 15.58 24.30 38.27 36.57 2516

CV2u 1 40.05 14.51 17.78 14.03 16.76 1827 36.54 35.93 2753

2 53.27 16.06 19.68 15.14 17.88 22.63 31.23 33.66 35.12

3 40.05 13.35 18.57 12.97 15.35 22.63 39.38 37.18 23444 46.19 16.69 19.75 15.92 18.81 22.63 30.5B 31.96 37.475 53.27 17.10 20.32 1636 19.03 24.30 30.00 30.24 39.76

6 46.19 17.03 20.60 16.05 18.72 22.63 29.27 31.59 39.147 53.27 1907 23.41 17.26 2067 24.30 22.98 30.30 46.728 49.61 15.80 18.92 15.44 18.14 21.07 32.11 32.45 35.439 43.01 15.43 19.34 14.97 17.45 22.63 32.76 34.96 32.2810 49.61 16.91 20.72 15.84 18.83 24.30 28.93 33.32 377511 43.01 15.62 18.83 15.22 17.87 18.27 32.58 33.50 33.9212 53.27 14.46 18.40 14.88 1732 18.27 36.26 32.87 30.8713 57.21 15.61 18.74 15.36 17.89 24.30 32.70 31.81 35.4814 46.19 15.87 1965 15.37 18.02 24.30 3142 3307 35.5115 53.27 16.98 21.73 16.06 18.99 21.07 25.98 3445 39.5616 46.19 17.01 21.60 15.79 18.85 21.07 27.12 34.28 38.6017 40.05 12.97 17.50 13.47 15.80 8.33 4090 33.82 25.281819 3729 12.61 16.82 12.56 14.91 1109 2931 34.80 359020 43.01 13.08 16.80 13.25 15.32 21.07 29.03 31.62 39.35

Tab. 22: Anzahl kontinentaler 850 hPa-und 500 hPaTrajektorien pro Monat bzw. Probenintervall

Monat CI # cont. Trajekt. CS # cont. Trajekt CV # cont. Trajekt. Datum Tage CV 850 hPa CV 500 hPa CV 850 hPa CV500#Traject. #Trajecl. Index hPa Index

Jan 12 21 30 11. Okt 92 6 6 5 1.00 0.83Feb 29 26 30 20. Okt 92 9.5 5 7 0.53 0.74Mar 23 16 27 30. Okt 92 9.5 9 7 0.95 0.74

Apr 13 12 22 08. Nov 92 9.5 8 5 0.84 0.53

May 6 22 17 18. Nov 92 9.5 9 1 0.95 0.11Jun 0 11 18 27. Nov 92 9.5 9 2 0.95 0.21Jul 4 24 5 07. Dez 92 9.5 9 2 0.95 0.21Aug 5 17 7 16. Dez 92 9.5 9 0 0.95 0.00Sep 19 24 11 26. Dez 92 9.5 8 0 084 0.00

Oct 27 30 19 04. Jan 93 9.5 6 1 0.63 0.11Nov 19 21 30 14. Jan 93 9.5 9 2 0.95 0.21Dec 21 18 30 23. Jan 93 9.5 7 0 0.74 0.00

02. Feb 93 9.5 7 0 0.74 0.00

11. Feb93 9.5 5 0 053 0.0021. Feb 93 9.5 9 0 0.95 0.0002. Mrz 93 9.5 0 0 0.00 0.0012. Mrz 93 9.5 4 4 0.42 04221. Mrz 93 9.5 6 4 063 0.4231. Mrz 93 9.5 9 6 0.95 0.6304. Apr 93 4.5 1 0 022 0.00

20. Apr 93 15.5 10 1 065 0.06

15. Mai 93 25 15 14 0.60 0.56

09. Jun 93 25 21 22 0.84 0.8804. Jul93 25 14 25 0.56 1.0029. Jul 93 25 4 24 0.16 0.96

23. Aug 93 25 8 23 0.32 0.9217. Sep 93 25 14 19 0.56 0.7612. Okt 93 25 23 23 0.92 0.92

ErkHinlllg

Hiermit erkläre ich, daß ich die vorgelegte Dissertation ohne unerlaubte fremde Hilfe verfaßt und

keine anderen als die angegebenen Quellen und Hilfsmittel benutzt habe.

Alle WÖlilich oder inhaltlich den angegebenen Quellen entnommenen Stellen sind als solche kenntlich

gemacht.

Bremen, im November 1996

v --:~~

(Volker Ratmeyer)

In dieser Reihe bereits erschienen:

NI'. 1

NI'. 2

NI'. 3

NI'. 4

NI'. 5

NI'. 6

NI'. 7

NI'. 8

NI'. 9

NI'. 10

NI'. 11

NI'. 12

NI'. 13

NI'. 14

NI'. 15

NI'. 16

NI'. 17

Nr.18

Nr.19

Wefer, G., E. Suess und FahrtteilnehmerBericht über die POLARSTERN-Fahrt ANT IV/2, Rio de Janeiro - Punta Arenas, 6.11. - 1,12,1985.60 Seiten, Bremen, 1986,Hoffmann, G.Holozänstratigraphie und Küstenlinienverlagerung an der andalusischen Mittelmeerküste,173 Seiten, Bremen, 1988. (vergriffen)Wefer, G. und FahrtteilnehmerBericht über die METEOR-Fahrt M 6/6, Libreville - Las Palmas, 18.2. - 23,3.1988,97 Seiten, Bremen, 1988.Wefer, G., G.F. Lutze, T.J. Müller, O. Pfannkuche, W. Schenke, G. Siedler, W. ZenkKurzbericht über die METEOR-Expedition Nr. 6, Hamburg - Hamburg, 28.10.1987 - 19.5.1988.29 Seiten, Bremen, 1988. (vergriffen)Fischer, G.Stabile Kohlenstoff-Isotope in partikulärer organischer Substanz aus dem Südpolarmeer(Atlantischer Sektor). 161 Seiten, Bremen, 1989,Berger, W.H. und G. WeferPartikelfluß und Kohlenstoffkreislauf im Ozean.Bericht und Kurzfassungen über den Workshop vom 3,-4. Juli 1989 in Bremen.57 Seiten, Bremen, 1989.Wefer, G. und FahrtteilnehmerBericht über die METEOR - Fahrt M 9/4, Dakar - Santa Cruz, 19.2. - 16.3.1989.103 Seiten, Bremen, 1989.Kölling, M.ModelIierung geochemischer Prozesse im Sickerwasser und Grundwasser.135 Seiten, Bremen, 1990.Heinze, P.-M.Das Auftriebsgeschehen vor Peru im Spätquartär. 204 Seiten, Bremen, 1990. (vergriffen)Willems, H., G. Wefer, M. Rinski, B. Donner, H.-J. Bellmann, L. Eißmann, A. Müller,B.W. Flemming, H.-C. Höfle, J. Merkt, H. Streif, G. Hel1weck, H. Kuntze, J. Schwaar,W. Schäfer, M.-G. Schulz, F. Grube, B. MenkeBeiträge zur Geologie und Paläontologie NorddeutscWands: Exkursionsführer.202 Seiten, Bremen, 1990.Wefer, G. und FahrtteilnehmerBericht über die METEOR-Fahrt M 12/1, Kapstadt - Funchal, 13.3.1990 - 14.4.1990.66 Seiten, Bremen, 1990.Dahmke, A., H.D. Schulz, A. Kölling, F. Kracht, A. LückeSchwermetallspuren und geochemische Gleichgewichte zwischen Porenlösung und Sedimentim Wesermündungsgebiet. BMFT-Projekt MFU 0562, Abschlußbericht. 121 Seiten, Bremen, 1991.Rostek, F.Physikalische Strukturen von Tiefseesedimenten des Südatlantiks und ihre Erfassung inEcholotregistrierungen. 209 Seiten, Bremen, 1991.Baumann,M.Die Ablagerung von Tschernobyl-Radiocäsium in der Norwegischen See und in der Nordsee.133 Seiten, Bremen, 1991. (vergriffen)Kölling, A.Frühdiagenetische Prozesse und Stoff-Flüsse in marinen und ästuarinen Sedimenten,140 Seiten, Bremen, 1991.SFB 261 (Hrsg.)1. Kolloquium des Sonderforschungsbereichs 261 der Universität Bremen (14.Juni 1991):Der Südatlantik im Spätquartär: Rekonstruktion von Stoffhaushalt und Stromsystemen.Kurzfassungen der Vorträge und Poster. 66 Seiten, Bremen, 1991.Pätzold, J., T. Bickert, L. Brück, C. Gaedicke, K Heidland, G. Meinecke, S. MulitzaBericht und erste Ergebnisse über die METEOR-Fahrt M 15/2, Rio de Janeiro - Vitoria,18.1. - 7.2.1991. 46 Seiten, Bremen, 1993.Wefer, G. und FahrtteilnehmerBericht und erste Ergebnisse über die METEOR-Fahrt M 16/1, Pointe Noire - Recife,27.3. - 25.4.1991. 120 Seiten, Bremen, 1991.Schulz, H.D. und FahrtteilnehmerBericht und erste Ergebnisse über die METEOR-Fahrt M 16/2, Recife - Belern, 28.4. - 20.5.1991.149 Seiten, Bremen, 1991.

Ne 20

NI'. 21

Nr.22

Nr,23

NI'. 24

NI'. 25

NI". 26

Nr.27

Nr.28

Nr,30

Nr.31

Nr.32

NI'. 33

NI'. 34

Nr. 35

Nr.36

Nr.37

Nr,38

Nr,39

Nr.40

Bcrncl', 11.Mechanismen der Sedimcntbildung in der fram-Strafk, im Arktischen Ozean und in derNorwegischen Sec. 167 Seiten, Bremen, 1991.Schncidcr, R.Spiilquarliire Produktivitätsänderungcn im ösllichen Angola-Becken: Reaktion auf Variationenim Passat-Monsun-Windsystem und in der Advektion des Bengucla-Küstenstroms.198 Seiten, Bremen, 1991. (vergrifTen)Hehheln, D.Spätqw1I1äre Stratigraphie und Paliiozeanographie in der Fram-Straße. 174 Seiten, Bremcn. I<)<) I.Lückc, A.Umsetzungsprozesse organischer Substanz während der Frühdiagenese in ästuarinen Sedimenten137 Seiten, Bremen, 1991.Wefcl', G. und f'ahrtteilnchmcrBericht und erste Ergebnisse der METEOR-Fahrt M 20/1, Bremen - Abidjan. 18.11.- 22.12 I<)') I.74 Seiten, Bremeu, 1992.Schulz, B.D. und Falll·ttcilnchmcl'Bericht und erste Ergebnisse der METEOR-Fahrt M 2012, Abidjan - Dakar, 27.12.1991 - 3.2. 1')92.173 Seiten, Bremen, 1992.Gingeie, Je.Zur klimaabhängigen Bildung biogener und terrigener Sedimente und ihrer Veränderung durch dieFrühdiagenese im zentralen und östlichen Südatlantik. 202 Seiten, Bremen, 1992.Bickcrt, T.Rekonstruktion der spätquartären Bodenwasserzirkulation im östlichen Südatlantik über stabileIsotope benthischer Foraminiferen. 205 Seilen, Bremen, 1992. (vergriffen)Schmidt, H.Der Bengucla-Strom im Bereich des Walfisch-Rückens im Spätquartär. 172 Seiten, Bremen. 1992.Mcinecke, G.SpätquaI-tiire OberDächenwassertemperaturen im östlichen äquatorialen Atlantik.181 Seiten, Bremen, 1992.Bathmann, u., U. Bleil, A. Dahmke, P. Miillcr, A. Nelll'korn, E.-M. N<ithig, M. Olcsch,J. Pätzold, H.D. Schulz, V. Smetacek, V. Spieß, G. Wefer, U. WillemsBericht des Graduierten Kollegs. Stoff-Flüsse in marinen Geosystemen.Berichtszeitraum Oktober 1990 - Dezember 1992. 396 Seiten, Bremen, 1992.Damm, E.Fruhdiagenetische Verteilung von Schwermetallen in Schlicksedimenten der westlichen Ostsee.115 Seiten, Bremen, 1992.Alltia, E.E.Sedimentology, MorphodYllamics anel Facies Association of a mesotielal Barrier IslandShoreface (Spiekeroog, Southern North Sea). 370 Seiten, Bremen, 1993.Duillkcr, J. und G. Wefer (Ursg.)Bericht über den I. JGOFS-Workshop. 1./2. Dezember 1992 in Bremen. 83 Seiten, Bremen, 1993.Kasten, S.Die VeI1eilung von Schwermetallen in den Sedimenten eines stadtbremischen Hafenbeckens.103 Seiten, Bremen, 1993.Spieß, V.Digitale Sedimentographie. Neue Wege zu einer hochauflösenden Akustostratigraphie.199 Seiten, Bremen, 1993.Schillzel, U.Laborversuche zu fIühdiagenetischen Reaktionen von Eisen (III) - Oxidhyelraten inmarinen Sedimenten. 189 Seiten, Bremen, 1993.Sieger, RCoTAM - ein Modell zur ModelIierung des Schwermetalltransports in Gn1I1dwasserieitern.56 Seiten, Bremen, 1993. (vergriffen)Willems, B. (Ed.)Geoscientific Investigations in the Tethyan Himalayas. 183 Seiten, Bremen, 1993.Bamel', K.Entwicklung von Laborversuchen als Grundlage für die ModelIierung des Transporlyerhallensvon Arsenat, Blei, Cadmium und Kupfer in wassergesättigten Säulen. 147 Sei tell, Bremen, 199.3.Sieger, R.Modellienmg des Stofftransports in porösen Medien unter Ankopplung kinetisch gcsteuertcrSorptions- und Redoxprozesse sowie thermischer Gleichgewichte. 158 Seitell, Brcl1lcn, I')'),

Nr.41

Nr. 42

Nr. 43

Nr.44

Nr.45

Nr.46

Nr.47

Nr.48

Nr.49

Nr.50

Nr.51

Nr.52

Nr.53

Nr.54

Nr.55

Nr.56

Nr.57

Nr.58

Nr.59

Nr.60

Nr.61

Thießen, W.Magnetische Eigenschaften von Sedimenten des östlichen SüdatIantiks und ihrepaläozeanographische Relevanz. 170 Seiten, Bremen, 1993.Spieß, V. und FahrtteilnehmerReport and preliminary results ofMETEOR-Cruise M 23/1, Kapstadt - Rio de Janeiro, 4.-25.2.1993.139 Seiten, Bremen, 1994.BleiI, U. und FahrtteilnehmerReport and preliminary results ofMETEOR-Cruise M 23/2, Rio de Janeiro - Recife, 27.2.-19.3.1993133 Seiten, Bremen, 1994.Wcfer, G. und FahrtteilnehmerReport and preliminary results ofMETEOR-Cruise M 23/3, Recife - Las Pahnas, 21.3. - 12.4.199371 Seiten, Bremen, 1994.Giesc, M. und G. Wefer (Hrsg.)Bericht über den 2. JGOFS-Workshop. 18../19. November 1993 in Bremen.93 Seiten, Bremen, 1994.Balzer, W. und FahrtteilnehmerReport and preliminary results ofMETEOR-Cruise M 22/1, Hamburg - Recife, 22.9. - 21.10.1992.24 Seiten, Bremen, 1994.Stax,RZyklische Sedimentation von organischem KoWenstoff in der Japan See: Anzeiger fürÄnderungen von Paläoozeanographie und Paläoklima im Spätkänozoikum.150 Seiten, Bremen, 1994.Skowronek, F.FfÜhdiagenetische Stoff-Flüsse gelöster Schwermetalle an der Oberfläche von Sedimentendes Weser Ästuares. 107 Seiten, Bremen, 1994.Dersch-Hansmann, M.Zur Klimaentwicklung in Ostasien während der letzten 5 Millionen Jahre:Terrigener Sedimenteintrag in die Japan See (ODP Ausfahrt 128). 149 Seiten, Bremen, 1994.Zabel, M.FfÜhdiagenetische Stoff-Flüsse in Oberflächen-Sedimenten des äquatorialen undöstlichen Südatlantik. 129 Seiten, Bremen, 1994.BleiI, U. und FahrtteiInehmerReport and preliminary results of SONNE-Cruise SO 86, Buenos Aires - Capetown, 22.4. - 31.5.93116 Seiten, Bremen, 1994.Symposium: The South AtIantic: Present and Past Circulation.Bremen, Gerrnany, 15 - 19 August 1994. Abstracts. 167 Seiten, Bremen, 1994.Kretzmann, U.B.57Fe-Mössbauer-Spektroskopie an Sedimenten - Möglichkeiten und Grenzen.183 Seiten, Bremen, 1994.Bachmann, M.Die Karbonatrampe von Organya im oberen Oberapt und unteren Unteralb (NE-Spanien,Provo Lerida): Fazies, Zyklo- und Sequenzstratigraphie. 147 Seiten, Bremen, 1994. (vergriffen)Kemle-von Mücke, S.Oberflächenwasserstruktur und -zirkulation des Südostatlantiks im Spätquartär.151 Seiten, Bremen, 1994.Petermann, H.Magnetotaktische Bakterien und ihre Magnetosome in Oberflächensedimenten des Südatlantiks.134 Seiten, Bremen, 1994.Mulitza, S.Spätquartäre Variationen der oberflächennahen Hydrographie im westlichen äquatorialen Atlantik.97 Seiten, Bremen, 1994.Segl, M. und FahrteilnehmerReport and preliminary results ofMETEOR-Cruise M 29/1, Buenos-Aires - Montevideo,17.6. - 13.7.1994.94 Seiten, Bremen, 1994.Bleil, U. und FahrtteilnehmerReport and preliminary results ofMETEOR-Cruise M 29/2, Montevideo - Rio de Janiero15.7. - 8.8.1994.153 Seiten, Bremen, 1994.Henrich, Rund FahrtteilnehmerReport and preliminary results ofMETEOR-Cruise M 29/3, Rio de Janeiro - Las Palmas11.8. - 5.9.1994. Bremen, 1994. (vergriffen)Sagemann, J.Saisonale Variationen von Porenwasserprofilen, Nährstoff-Flüssen und Reaktionenin intertidalen Sedimenten des Weser-Ästuars. 110 Seiten, Bremen, 1994. (vergriffen)

Nr.62

Nr.63

Nr.64

Nr.65

Nr.66

Nr.67

Nr.68

Nr.69

Nr.70

Nr.71

Nr.72

Nr.73

Nr.74

Nr.75

Nr.76

NI'. 77

Nr.78

Nr.79

Nr.80

Giese, M. und G. WeferBericht über den 3. JGOFS-Workshop. 5.16. Dezember 1994 in Bremen. 84 Seiten, Bremen, 1995.Mann, U.Genese kretazischer Schwarzschiefer in Kolumbien: Globale vs. regionale/lokale Prozesse.153 Seiten, Bremen, 1995. (vergriffen)Willems, H., Wan X., Yin J., Dongdui L., Liu G., S. Dürr, K-U. GräfeThe Mesozoie development ofthe N-Indian passive margin and ofthe Xigaze Foreare Basin insouthern Tibet, China. - Excursion Guide to IGCP 362 Working-Group Meeting"Integrated Stratigraphy". 113 Seiten, Bremen, 1995. (vergriffen)Hiinken, U.Liefergebiets - Charakterisierung proterozoischer Goldseifen in Ghana anhand vonFluideinschluß - Untersuchungen. 270 Seiten, Bremen, 1995.Nyandwi, N.The Nature of the Sediment Distribution Patterns in ther Spiekeroog Backbarrier Area,the East Frisian Islands. 162 Seiten, Bremen, 1995.Isenbeck-Schröter, M.Transportverhalten von Schwermetallkationen und Oxoanionen in wassergesättigten Sanden.- Laborversuche in Säulen und ihre ModelIierung -182 Seiten, Bremen, 1995.Hebbeln, D. und FahrtteilnehmerReport and preliminary results of SONNE-Cruise SO 102, Valparaiso - Valparaiso, 95134 Seiten, Bremen, 1995.Willems, H. (Sprecher), U. Bathmann, U. Blei!, T. v. Dobeneck, K Herterich, B.B. Jorgensen,E.-M. Nöthig, M. Olesch, J. Pätzold, RD. Schulz, V. Smetacek, V. Spieß. G. WeferBericht des Graduierten-Kollegs Stoff-Flüsse in marine Geosystemen.Berichtszeitraum Januar 1993 - Dezember 1995. 45 & 468 Seiten, Bremen, 1995.Giese, M. und G. WeferBericht über den 4. JGOFS-Workshop. 20.12]. November 1995 in Bremen.60 Seiten, Bremen, 1996. (vergriffen)Meggers, H.Pliozän-quartäre Karbonatsedimentation und Paläozeanographie des Nordatlantiks unddes Europäischen Nordmeeres - Hinweise aus planktischen ForaminiferengemeinschafteIl.143 Seiten, Bremen, 1996. (vergriffen)Teske, A.Phylogenetische und ökologische Untersuchungen an Bakterien des oxidativen und reduktivenmarinen Schwefelkreislaufs mittels ribosomaler RNA.. 220 Seiten, Bremen, 1996. (vergriffen)Andersen, N.Biogeochemische Charakterisierung von Sinkstoffen und Sedimenten aus ostatlantischenProduktions-Systemen mit Hilfe von Biomarkern. 215 Seiten, Bremen, 1996.Treppke, U.Saisonalität im Diatomeen- und Silikoflagellatenfluß im östlichen tropischen und subtropischenAtlantik. 200 Seiten, Bremen, 1996.Schiiring, J.Die Verwendung von Steinkohlebergematerialien im Deponiebau im Hinblick auf diePyritvelwitterung und die Eignung als geochemische Barriere. 110 Seiten, Bremen, 1996.Pätzold, J. und FahrtteilnehmerReport and preliminary results of VICTOR HENSEN cruise JOPS II, Leg 6,Fortaleza - Recife, 10.3. - 26.3. 1995 and Leg 8, Vitoria - Vitoria, 10.4. - 23.4.1995.87 Seiten, Bremen, ]996.Blei!, U. und FahrtteilnehmerReport and preliminary results ofMETEOR-Cruise M 34/1, Cape Town - Walvis Bay, 3.-26.1.1996.129 Seiten, Bremen, 1996.Schulz? H.D. und FahrtieilnehmerReport and preliminary results ofMETEOR-Cruise M 34/2, Walvis Bay - Walvis Bay, 29.1.-18.2.96133 Seiten, Bremen, 1996.Wefer, G. und FahrtteilnehmerReport and preliminary results ofMETEOR-Cruise M 34/3, Walvis Bay - Recife, 21.2.-17.3.1996.168 Seiten, Bremen, 1996.Fischer, G. und FahliteilnehmerReport and preliminary results ofMETEOR-Cruise M 34/4, Recife - Bridgetown, ]9.3.-15.4.1996.105 Seiten, Bremen, 1996.

NI'. 81

NI'. 82

NI'. 83

NI'. 84

Kulbrok,F.Biostratigraphie, Fazies und Sequenzstratigraphie einer Karbonatrampe in den Schichten derOberkreide und des Alttertiärs Nordost-Ägyptens (Eastern Desert, N'Golfvon Suez, Sinai).153 Seiten, Bremen, 1996.Kasten, S.Early Diagenetic Metal Enrichments in Marine Sediments as Documents ofNonsteady-StateDepositional Conditions. Bremen, 1996.Holmes, M.E.Reconstruction of Surface Ocean Nitrate Utilization in the Southeast Atlantic Ocean Basedon Stable Nitrogen Isotopes. 113 Seiten, Bremen, 1996.Rühlemann, C.Akkumulation von Carbonat und organischem Kohlenstoff im tropischen Atlantik:Spätquartäre Produktivitäts-Variationen und ihre Steuerungsmechanismen.139 Seiten, Bremen, 1996.