/nt,^-a544 1, 6 4 - 7 V, lQgLfC€T-A G- - e_ t o,_ ,a&. t,

( ()1,\()51.s q:) 69 'l lee{l)( [:TA (i,,h^r lrrlr

69

GI.OBAL WARMING: THE PROCESS AND ITS ANTICIPATEDPHYTOCLIMATIC EFFECTS IN TEMPERATE AND

COLD ZONE'

L. Orl6ci

Department of Plant Scienccs. The University of Western Ontano, London. Canada N6A 587

Keywords: Alpine. Arctic. Borcal. Climate. Communities, Deciduous, Manabc, Migration, Models, Milankovitch. Plants.Tropical

Abstracl Postglacial vegetation dynamics reflects Slobal warming and cooling as thesc occur in normal progression throughprecessioncycle. ltalsoreflectsthcrmal influencesofaberrationssuchasvariationsinthcSun'sradiationandintheEarth'satmosphenc COt concentration. Although the present precession cycle is in the fina.l cooling phase as it approaches a newagc of glaciation, global warming caused by man-made increases in atmospheric COz and other greenhouse B,rses may tem-porarily, albcit substantially subvert the cooling clrmatic tend. The author uses the Manabe ocean-atmosphere model'sprediction of a l.5o global temgieralure nse in less than 7-decades to quantify the level of warming, and on this basis. he

undertakes an exanrination of the potential local effects over latitudes in eastern Nonh America aM elevations in Hawaii.His results indicate local thermal flux ranging from over l20 (Arctic Tundra ar 58o N) to under 40 (Warm TemperateEvergreenForesratl20N). Acomparativeevaluationoftheseleatjtheaurhortohypothesizrthatthepresenrvegerarionofthe cool temperate and cold region will disintegrate. When this happens, it will beget ttre rise of new plant communities thatwill be little more than chance assemblages of population remnants. species poor and ill-defined. But given enough time,migration, competition and environmental sorting will enhance community level differentiation and sharpen a new, climate-determined regional pattern. Neither the communities nor the zonal pattern will be comparable to those of today. TheTundra will lose much of its niche and the Boreal Forest will shrink to guantitatively insignificant isolated stands on azonalnorthern sites. Elements of the Cool Temperate Deciduous Foresl and to some extent thc Mixed Conifer-Northem HardwoodForesl will appear in pans of Labrador and adjacent maritime regions to the south. The Oak-Hickory Forest and Savanna,

now wedged between grassland and eastern deciduous forests from Minnesota to Oklahcma. and the Warm TcmpcrateEvergreenForcstinthesouthernCulfandAtlanticstateswill bemajorgainers. SuitableclemensoftheOat-HickoryForest*'rll disperse in north-east direction to form a broad nonh-south belt extending from subarctic westem Quebcc to th€ centralCreat Lakes *here thermal. soil moisture. and precipitation regimes will approximate the present conditions in the eastemplane states. Stgnrficantly, the Warm Tempcrate Erergreen Forest will expand north and north-east in step with the expand-tnginfluenceofthe Ilaritime-Tropical Airmass. OnthemountainsofHawaii,thcvegetationzoneswillshiftupwards,rcsult-in-e in the disappearance of the Tropical Alpine Tundra on ltlauna L,oa and Mauna Kea. and possibly the Tropical SubalpincScrub vegetation on the lesser high mountains. Regronal economies will disintegrate ard major dislocations of populationswill fan se!ere reqional conflicts.

Purpose and sites

The following sections address global warming. localthermal flux. and anlicipated collaleral effects in the wake

of 2.5o global temperaturel rise in less than 7-decades. Theauthor's intentions are lo examine isolated pieces of evi-dence and to piece together a geheral theory of vegetation

shi fts.

The localities under scrutiny include five sites from the

southern limits of the major vegetation zones in eastern

Norrh America. These are at Port Harrison (Quebec, ArcticTundra), Timmins (Ontario, Boreal Forest), Stratford (On-

I All tempcratures in Celsius dcgrces.

tario, M i xed Coni fer-Northem Hardwood Forest), Nashvi lle

(Tennessee. Cool Temperate Deciduous Forest), and Mobil(Alabama. Warm Temperate Evergre€n Forest). Two addi-

tional sites are in Tropical Alpine and Subalpine vegetation

on Mauna Kea in Hawaii.

The extended Port Harrison-Nashville line intercepts

four broad bioclimatic belts. In terms of averages, the Arctic

Airmass dominates nonh of the 48th parallel, and the

Maritime Tropical Airmass south of the 34th parallel. Be-

tween these extre mes. the Pacific Airmass shares dominance

with the Arctic Airmass above the 44th parallel and with the

/n

Teblc l. Elevation relatedrevisions.

tropical bioclimatic gradient on l\fauna Kea. teewerd slope. Data from Krajina 11$fl) *ith

Vegetation Elevation Annual precipitationm mean t. mm

oc

Alpine Tundra and BarrenSubalpine ScrublandSubhumid SavannaSubhumid Open ForestSemiarid Thorn ScrubSemiarid Savanna

3300 -2000 -r200-s00 -,r_

33002000r 200s00300

04.4

r0.024.023.124.6

- 5t0510 - 1020

1270 - 15201020 - t270640 - ta20

- 640

Maritime Tropical.Airmass below the 44th parallel. Thesemistationary tropical Pacific High located northeast ofHawaii, the southward shift of the northern storm tracks inthe winter, relatively low seasonal variation of the energyreceived from the sun, ocean and orographic effects - such as

the lag of seasonal higMow temperatures, average adiabaticlaps rate of 0.5o per 100 m. steep increase up to and rhen steepdecline of precipitation above the temp€rature inversion line(at about 1000 m elevation windward side). and leewardaridity - characterizes the island environment.

The precession cycle

The physical bases of global warming and cooling arefound, in part. in variation of the Earrh's orbital parametersthat Milankovitch ( I 94 I ) described and pielou ( l99l ) inrer-preted to an ecological audience. They are also found in var-iation of the sun's radiation and in changes of atmosphericcomposition, particularly in COz concentration (Mason1990, Mungal and Mclaren 1990. Manabe er al. 1990).

The Earth's orbital geometry undergoes long-term andshon-term cyclic variation with predictable consequencesfor the volume of ice on the earth's surface. A short cycle.called precession of the equinoxes, began t 8,000-years agoand will end in about 3,0O0 years. In this cycle, the timepoint of minimum disrance between eanh and sun isprecessed through the seasons, causing changes in the appor-tionment of incoming solar radiation between winter andsummer. When the earth is far from the sun in summer, thethermal contrast of the northern summer and winter isreduced (mild summer, mild winter). snow and ice accumu-lates, and the glacial age is tumed on. When the earrh is closeto the sun in summer, the seasonal thermal contrast sharpens(warm summer, cold winter), the winter.snow and ice mellsand an intergtacial period is turned on, provided that thestates of other parameters under the Milankovitch cycle per-mit this to happen.

At the beginning of the precession cycle, the global an-

nual average temperature was I lo, compared to the present

l50, and the Laurentide ice sheet exrended below rhe 40thparallel (Delcourr and Delcourr I987. Fig. l-4).: Thesouthem limit of the narrow Arctic Tundra belt lay jusr be lowthe 39th parallel. the Boreal Foresr had its sourhern limir jusrabove the 33rd parallel. and rhe Mixed Conifer-NonhemHardwood Forest occuned in a narrow belt between theboreal forest and the Warm Temperate Southem EvergreenForest. The Cool Temperate Deciduous Foresr did not yet

appear as a distinct vegetation formation. The precession

cycle has reached its thermal maximum (Hypsithermal. ex-pected highest contrast between summer and winter in ener-gy received from the sun) 10.000 years ago. then the slon,descent of the orbital parameters back to their Ice Age states

began. Interestingly, the global temperature peaked muchlater about 5.000 years ago. By that time. the sourhern limirof the Boreal Forest has been pushed north above the 49thparallel and the Cool Temperate Dec.iduous Foresr attainedits present northern limir near rhe 43th parallel. The presenr

southem limit of the Boreal forest is just above the .18rh

parallel. where the Arctic Tundra had its southem limir some9.500 years ago. The sourhern limit of rhe Arcric Tun&a isput at just above the 55rh parallel.

Precession dynamics was closely followed by substantial

altitudinal vegetalion shifrs on rhe high mounrains of Hawaii,

Traces of this are most clearly seen on Mauna Kea ( I9o 49'N). the highest volcano in Hawaii which attained its impres-

sive height {4000+ m) well before the present precession

cycle started. It had an ice cap about 60 m deep 20,000 years

ago (Cmikshank 1986) that extended down to the 3200 m

level. Coinciding with this, the Tropical Alpirr Tundncovered a broad belt from ice edge to the 2500 m level.

Today, the Alpine Tundra occupies elevations above 3300

m. By around 10,000 before present, the globd climate

warmed to I 2.5o and Mauna Kea's ice cap melted. The eleva-

tions that once supported the Alpine Tundra now suppon e

Tropical Subalpine Scrirbland vegetation. Table I describes

these formations.

2 All latitudinal limis arc givcn on the Port Harrison (Qucbcc) and Nashvillc (Tenncssee) linc

a.#;c ..'' d,.

t,,l tl

Table 2. Description of conditions et the lorler letitudey'eleuetion limits of vegelation zones. The soulhern limil of lhcWarm Temperatc Evergreen Forest is not climatic. \'dues underlined re extrapolations based on TFR = -

2.5753E+0.I2749X in which X represents the decimal equivalent of the locality's northern latitude. The correspondingvalue of the coeflicient of determination is 0.97. Regression does not include the l\tauna Kea sites. Abbreriatiors: TFR .thermal llux rale; LT. - local temperature rise; VP - expected annuel mean lemperalure. Data in rows 3, 4 are from \l'alteret al. (1975), except lasl column nhich are from Krajina (l%3).

ArcticTundra

BorealForest

MixedConifer -

NorthernHardwoodForest

CoolTemperateDeciduousForest

Warm

TempcrateEvergreenForest

Tropical TroprcalAlpine SubalpineTundra Scrubland(3.3@m) (1.000mt

Mauna Kea.Hawaii190 49'

5r0 1020

0 .1.4

t.8 r.8

4.5 4.5

4.5 8.9

l. Climatic station

2. Northern lat.3. Precip. mm

4. Mean ann. t. oC

5. TFR value

6. LT. value oC

7. VP value oC

Port Harrison.Quebec580 26',

372-7.5

!"9t2.34.8

Timmins.Ontario4go 3t'7ttr.33.7

9.2lo.5

Stratford,Ontano430 22',

773

8.32.8

7.015.3

Nashville.Tennessee

360 t0'I,t4415.62.1

5.320.9

Mobil.Alabama

3to 42'

r.439

r 9.8

1l3.E

236

The effect of greenhouse gases

Postglacial vegetation dynamics reflects not onlyprogression tfuough the Milankovitch precession cycle. butimportantly. noncyclic greenhouse warming linked toperiodic increases in atmospheric CO2 concentration. Clear-ly. CO2 concentration had a long-period of increase in 17.000to 10.000 years ago (Neftel et al. 1982) and other times. butthe rate of CO: increase has taken an alarming turn upwardsonly in recent decades (Bacastow and Keeling 198 I, Bolin1989. Mason 1990, McKay and Hengeveld 1990. Fig. 7).This happened because of massive man-made emissions.Along with CO2, other greenhouse gases (CFI1, N:O. CFCs)also increased.

Monographs on global warming point out that warcr vapor. carbondioxrde (CO:), nrcthanc (CHr). nitrous oxide (N:o). ozone (O-r). andchlorofluorocerbons (CFG) collectively kno*'n as ihc -grcenhousc

gascs- rcducc thc earth's hcat los to outer space and keep the surfaceof the carth narm by absorbing infra-rcd terrestrial mdiarion. Thcsccollectivcly pcvent about 9l9L of rnfra-rcd radiarion from escapingfrom tlrc canh's surfacr inlo outer spacc. Without grecnhousc gascs,

rhc global avcragc tcmperaturc would bc -t9", compared to thc prescnt+150. Waler vapor alrd CO: arc by fu thc rnost impodant infra.rcd ab-

sorbing agents, crcafl al wavelengths from 8 microns to 12 micmnswhere thc tfirc 8ae6 arc cfficicnt. On a molecular basis, CH.r is 2ltirncs more cffectivc than CO:, and thc CFCs ar€ ovcr 70O trnrs morceffcctivc than NHr (Mason 1990). In discussions of the Iopic. it rs

usually pointcd out that owing to alrcady clevued lcvcls. thc amos-pheric conctntralion of grccnhous€ gases shoutd havc causcd a l"temperature risc since thc turn ofthis Ccntury. but it did no(. bccaus€,

as ofrcn pointcd out llE ltErmal inertia of dccp occans prevented it.Mason (1990) givcs rcasons for this and exptains that a nct 0.5o

tcmpcnlurc risc in this Ccntury, calculatcd by Henderson-Sellcrs( I 990), is not a greeohousc signa.l.

According to b€st estimates, terrestrial temperatures willsoon begin a greenhouse rise, 0.5o by the year 20 t 0 and 20 -

3o by 2060 (l{anabe et al. 1990), assuming that greenhouse

gases continue to increase in a manner equivalenl to CO:concentration rising at a 17. compounded annual rate and

doubling by 2060. To put the magnitude of this temperaturerise into proper perspective. it will be sufficient to recall three

well known facts about the global average lemperature:

-- it is only a mere 4o higher today than it used to be at

the height of the last glaciation:

-- it probably never erceeded the l50 mark by more than

2.5o during recent inlergla,.'ial periods:

-- its oscillations remained within a 3.5o range over the

past 3.000.000 years (lv{cKay and Hengeveld 1990).

These show clearly Ihat a 20 - 3o temperature rise is rn-

deed extreme on the scale of historic events.

Local thermal flux

The general circulation models (Mason 1990) predict itin general terms and it can be seen from the contents c,lTable

2 in particular cases thal the rate at which a given global

temp€rature rise is linked to local temperature increases

depends on lalitude. The Arctic Tundra is an example. Atits southem limit near Port Harrison, the presenl annual mean

temperature is -7.5o. Some 9,50O years ago, its southemedge lay l0 degrees farther south in the Timmins area where

the present annual mean temperature is 1.3o. The thermal

distance of the historical and present southern limits is E.Eo.

Considering that the global average temperature was 12.6o

(this number by interpotation) 9,500 years ago and it is 150

at present, the thermal flux rate at Timmins is 8.8/2.4 = 3.7.

In other words the anticipated local temperature rise at Tim-mins in the wake of a lo rise in the global average tempera-ture is 3.7o. The anticipated total temperature rise under the

btt

12

Manahc modcl sccnario is 3.7x2.5o or 9.3n. and the an-ticipared annual mcan lcmpcrature is 10.6n. Similar catcula-tions yicld thc values in rows 5. 6. 7 of Table 2.

Anticipated vegetation

Thc organ and ccll-lcvcl nrr.chanisnrs of plant responseto cooling and warming are wc.ll undcrslood (see refercncesin Mucller-Dombois 1992). Also. vegetation models can belound based on anticipatcd tempcralures (Table 2y, precipita-tion levcls (Mason 1990. Figs. 6.7) and soil moisture condi-tions (Stcwart and Tiesst'n 1990. Fig. 58a) that may seft.e assources for migrant populations lo compose new plant com-munilics in the wake of greenhouse warming. The followingsynopsis describcs potc-ntial states on w,hich warming-in-duccd local vegetrtron dcvelopment should converge in thelong run:

Port Harrison (58o N). Daily I mm risc in precipirationis expccted in the immediate vicinily over Hudson Bay inwinter. and to the cast and north in summer. On this basis.anticipation of 500-600 mm annual precipirarion appearsrealistic. Soil moisture will decline by 201c to 3070 onupland sites. and the local temperarure will rise by more thanI 20 to a new annual average of over 40. These parameterspoint to the semiarid Oak-Hickory Forest model of westemMinnesota. Elements of this model. phoroperiod and soilquality permitting, will find suirable niche on upland sites.Tundra and boreal species may continue ro exist in cold. shel-tered sitcs on peat.

Timmins (48o N) Precipitarion remains unchanged(71 I mm1. although soil moisture losses will be substanrial(40% to 507c). The annual mean remp€rature will risc justover l0o. These values poinr to the Oak-Hickory Forest andSavanna model of eastern central Nebraska as possiblesource of migrants lor upland sites. Boreal and Mixed Con-ifcr-Northern Hard*ood Forest species will continue to existin cool sheltered siles on peat and clay.

Stratford (43o N) Some reducrion is expected in sum-mcr precipitation and a 307c to 40% reducrion in soil mois-ture. An anlicipated 700 mm annual precipiration and overl5o annual mean temperature identify the Oak HickorySavanna and Crassland models of cenrral Oklahoma as pos-sible species sources.

Nashrille (360 N). The average precipitarion will bemaintained. but the winters will be drier and the summerswetter under the increased influence of the Maritime Tropi-cal Airmass. These and the high lover 20o) annual meantemperature point to elements of the Warm TemperateEvergreen Forest model as possible componenfs of futurevegetation. Remnants of the present Temperate DeciduousForest will sunive in the high €levarion belt of the Ap-palachian Mountains.

Itobil (3 lo N). The high precipiration level and its man-ner of distribulion between seasons remain the same. but theannual mean tcmperarure will rise above 23o. These willdraw some new specres into the region without essentiallychanging the characler of the present vegetation.

Ilauna Kea (l90 N). The local remperature will iise to9o in the Subalpine zone. and if the same thermal nux rate( 1.8) holds. to about 40 in the Alpine Tundra. These will rrig-ger zonal migrations. The Subhumid Savanna invades the

present sites of the Subalpine vegetation which in turn. landmass and aridity permitting, invades the presenr altitudinalbelt of the Alpine Tundra.

Approach and accuracy

The choice of localities is incidental. beyond the inten-tion to span many degrees of northem latitudes and to havereliable information about them. The question of accuracy is

legitimate in connection rr'ith data sources. Admittedly, theyare not as optimal as one would prefer, it cannot be said thatthe baseline data ue critically faulty or even inadequate.The derivation of some historic temperature values by linearinterpolation matters. but it was performed within a rathertightly linear segment of the time/temperature graphpresen(ed in McKay and Hengeveld (1990, Fig. 3). Onemajor methodological deviation from previous practice, thatmay also bear on accuracy, is the use of historic increases in

the global annual mean temperaturc to determiDe local ther-mal flux rates. Th€ values in the last 3 columns of Table 2

are based on l5o as the global annual mean temperarure(Mason 1990), therefore, they are somewhat higher than

would be if the calcu lations were based on the l60 figure that

McKay and Hengeveld ( 1990) usc.

Overview and conjecfu res

The 2.5o global warming within seven decades wouldprobably have a less seuere effect on the vegetation. if it im-plied a uniform temp€rature rise over the earth's surface.

The calculations show a practically functional latitude-dependent trend of the expected local temperature rise. It isclear on historical grounds tha( even at the lesser thermal fluxrates, the vegetation response expected to be substantial. Allindications point in fact to a severe response in the high fluxrate areas, probably complete disintegration of the vegetation

in the cool temperate and cold region, including substantial

species losses. Disintegration will beget chance associ-

ationss of populations that can withstand the thermal stress

or have propagules equipped lo disp€rse quickly over great

distances, develop under soil conditions that are initially out

of phase with the changing climate, and flourish under the

given photoperiod. The initial population configurationswill converge, probably in the tinre span of centuries, on a

Thc low altilude vegetarion on thc Hrwaiian Ishnds is an example wherc haphazard assemblagcs of spcies comPosc corrununru6.

Significanlly, thcsc crmc into existencc by dcliberate or accidcntal transplrntation of spccies. Just onc nronomanir, H. Lyon. introduccd sonrc

l0.60cxaictreesandorhcr planrsroHawaiiinrhcspanof 27years. Thcscsurclyinchdedhundrcdsofspccicswhichmysourcc(Welcorncto Fostcr Carden. Self-Cuided Tour) docs not spccify.

\

mosaic of distinct plant communities and a climate-deter-

mined u.nique zonal pattern.

Two major directions of migration will result in the new

zonal pattern. One from the south-west will propel suitableelements of the Oak-Hickory Forest and Savanna. that today

stretches in a well-defined belt from Minnesotr to Ok-lahoma. between the Tall-Grass Prairie and the easlern

deciduous forests. north-east into a similar wide belt. extend-ing from eastern subarctic Quebec to the central Great Lakesand possibly some distance beyond to the south. The exactsouthern limit of this belt will depend on how far the

Maritime-Tropical Airmass of the Culf of Mexico will ex-tend its climatic influence nonh and nonheast. Port Har-

rison. Timmins. Stratford will be in different floristicprovinces of this belt. Another major direction of migrationpoints from the south-east Culf region to the north and nor-theast. This will draw the Warm Temperate Southeastern

Evergreen Forest farther inside the continent and up alongthe Atlantic Coast in step with the expanded dominince ofthe Maritime-Tropical Airmass. Warming will dislo,:ge the

Cool Temperate Deciduous Forest and the lvlixed Conifer-Norrhern Hardwood Forest from their present sites. Ele-ments of the Cool Temperate Deciduous Forest. and to a

lesser extent. elements of the Mixed Conifer-NorthemHardwood Forest. will find refugia in Labrador and adjacent

maritinre regions to the south and on high elevations in the

Appalachian Mountains. The Boreal Forest will end up

having no significant niches left in the region, except locallyon northern azonal sites. There are no altitudes high enough

to preserve the Boreal bioclimatic environment anywhere ineastern North-America, and low precipitation will prevent its

extension into the high Arctic.

The vegetation of Hawaii will begin a zonal march up

through elevation belts. The Tropical Alpine Tundra willdisappear. The lesser high mountarns will loose the TropicalSubalpine Scmbland.

Closing remarks

The 20th Century is ending with the global climatewarmer than at its beginning. A new greenhouse *'armingmay be just around the corner in the I lst Century. The majorcirculation models are rather clear on this when they forecast

a sharp temp€rature rise beginning in less than 2-decades.

Ironically, society is defenseless, since no ameliorative

measures can stop greenhouse warming immediately. They

can only reduce the rate of atmospheric deterioration that

could lead to more greenhouse warming.

Bioclimatic projections usually assume gradual change

at a slow rate on the time scale of organism longevity. The

post-glacial bioclimatic process is on average this kind. This

is quite apparent from the orderly march ofvegetation zones

through time and latitudes in the reflections of fossil pollen

and microtissue spectra. Significantly, this regularity is the

product of a convolution of natural processes. The an-

ticipated greenhouse warming is very different. It issues

from man-made emissions of carbon dioxide and other

greenhouse gases into the atmosphere and it can change the

7l

global temperalure very quickly in a matter of decades. Mat-ters are made even worse by ill-conceived forcstry practices.

All know the dangcrs in these. yet the deleterious praclices

not only continue but horribile dictu may even accelerate

through basically neSa(ive planning. The mosl typical case

of this is the narro*ing of the rotation cycle (Botkin et al.

1992) seen as an answer lo minimizing future business losses

of thc' forest industry in the wake of grecnhouse *'arming.

Will a 2.50 greenhouse * arming really happen * ithin the

sullc'sted time frame'? Its consistency with basic physical

principles and causal factors on trend with processes that did

lead to greenhouse warming at other times in the precession

cycle. are good reasons for society to starl planning. The

plans should consider as potential oulcomes. the disintegra-

tion of regional economies. major population shifts. and

inter- and intranational conflicts.

References

Bacastow. R. and C D Keeling. 1981. Atmospknc crrbon dtorrde

concentrltion and the obsen'ed ur-bom fractton ln: B Bohn

ted.). Carbon Cyclc Mulellrng. pp. 103-l 12. Wrley. Chrchester

Bolin. B. 1989 Changing climate. ln: L. Friday and R. Laskey tcds.r.

Thc Fragile Environrnent. pp. 127-141. Cambndge UntrerstllPress. licw York.

Bo&rn, D.B. rnd R.A Nishbet. l99l Projccted hanest of ttmtrrdunng global *arming. ln: A. Qureshi (ed.). Forests in a Chang-

ing Climcte. pp 26-5-19-1. Climate Institutc. J3l Fourth Sueet.

NE, Washington DC. 20001-5t2 1

Cruikshank. D.P. 1986. Mauna Kea A Guidc to ttr Upper Slopcs and

Obsen'ctones. Universrtv of Ha*ui. lnstitute of Astronomy.

Honolulu. Hawair

Detcoun. P A and H.R Dclcoun. 1987 Long-term Forcst Dynamics

of thc Temperate bnc Ecologrcrl Studies 63. Spnnger-Vcrlag.

Ne*'York.

Henderson-Scllers. A. 1990 l{odelling and monitonng "greenhouse'

warming Trcnds Ecol. and Evol. 5.27U775.

Kra.pna. V J 1963 Brogeocltmatic zoncs on thc Hauuran lslands

Neu'slcner of thc Ha*uian Botanical Socrcty 7:91-98.

l"lanabe. S.. K. Br.van and M.J. Spelman 1990. Transrent rcsponse of a

global ocean-atmospherc modcl to a doubhng of atmosphcnc car-

bon dioxrde. J Phys Oceanogr.20..'122'i19.

Mason. J. 1990. Th€ greenhouse cffect and global warrrung lnforma-

tion Officc. Bntrsh Coal. C.R.E. Stokc Orchard. Cheltenham.

Glouccstershrre. U.K.. CL51 1RZ.

McKay, C.A. and H Hengeveld. t990. Thc changing lmosphere ln:

C. Mungrlt and D.J. Mcl.aren (eds t. Planct Under Stress. pp' 46-

79. Oxford Univenity Prcss. Toronto.

Milankovitch, M. M. 1941. Canon of isolation ard thc lce Agc

problcm. Royal Serb Acad. Spec. Publ. 133

Mueller-Dombois. D. 1992. Potentiat cffects of the incrcase in carbon

dioxide and climate change on t}r dynamics of vegelation Watcr'

Air and Soil Pollurion 64:61-79.

Mungat, C. and D.J. Mclaren (cds.) 1990. Planct Under Stress. Oxford

University Press. London.

Neftel, A.. H. Ocschger. ct al. 19t2. lce corc sarnplc rEasurem€nts

givc atmosphcric COt content during rlr pas 40,00 years Na-

ture 295:22G223.

Pielou. E.C. 1991. After t}t lcc Agc. Univ. Chicago Prcss, Chicago'

Stesafl. J rnd H Tresscn 1990 Crxslands tnto gall and DJ Mcl_arcn. pp 1tt.206. Orforddc sU.nnisv cIrns: rCty pMrcusns_. London

Walter. H., E Harnrckell urd D. Mucllcr.Dombois. 1975. Climarcdiagram

Maps Springcr _Vcrlag, Ncw yort. Manuscript received; March 5, lW4