temporal response patterns in fish and benthic ... · null hypothesis testing ho1: aquatic...
TRANSCRIPT
Temporal response patterns in fish Temporal response patterns in fish
and benthic invertebrate and benthic invertebrate
communities in streams of the communities in streams of the
North Central and Northeastern North Central and Northeastern
U.S.U.S.� Jason May-California
� Amanda Bell-Wisconsin
� Jonathan Kennen- New Jersey
� Dan Sullivan-Wisconsin
� Karen Beaulieu-Connecticut
US Geological Survey-WRD
National Water-Quality Assessment Program Surface-Water Regions (Major River Basins)
1
2
3
4
5
6
7
8
Null Hypothesis Testing
� Ho1: Aquatic assemblages will not exhibit significant changes in composition over time.
� Ho2: Fish and invertebrate assemblages will not differ in their response to environment degradation.
� Ho3: Aquatic assemblages response will not be related to temporal changes in environmental processes.
� Ho4 Natural variability does not influence temporal changes in assemblage composition.
� HA1: Sites which have been significantly altered over time periods prior to or exceeding the NAWQA sampling effort will show little or no trends in biotic assemblages and/or abiotic parameters.
Types of potential trends response
patterns/trajectories for specific
sites over time
� No response
� Upward/downward trends
� Recover trends
� Caveats
� Pre-disturbed prior our sampling
� Human influenced
� Antecedent Influenced conditions
Identify changes in Biota over time
(Temporal Separation)
Effects of Natural
Variability
2: BiologicalCondition
1:TaxonomicComposition
3: SpeciesTraits
Land UseHydrologyQW Habitat
Assemblage Trajectory over time – Examplehdsn12cs, 2/7/2009 8:41:34 PM
LISHA KILL NORTHWEST OF NISKAYUNA NYTransform: Square root
Resemblance: S17 Bray Curtis similarity
1993
1994
1995
2002
2003
2004
2005
2006
2D Stress: 0.01
Departure Distance (=temporal separation)
hdsn12cs, LISHA KILL NORTHWEST OF NISKAYUNA NYGroup average
20
06
20
05
20
03
20
04
19
94
19
93
19
95
20
02
Samples
100
80
60
40
Sim
ilari
ty
Transform: Square root
Resemblance: S17 Bray Curtis s imilarity
�Null hypothesis: “No tendency to show temporal separation”.�In general: assemblages with adjacent samples are closest in species composition, those further apart differ the most.
Same site
Inverts-significant
change over time
&
fish-N.S.
19 9 3
1 99 4
19 9 5
20 0 2
2 00 3
2 00 4
20 0 5
20 0 6
(√)-S tress : 0 .01LISHA-INVT
1 99 3
19 9 4
1 99 5
20 0 2
20 0 3
2 00 4
20 0 5
2 00 8
(√√)-Stress: 0.05LISHA-FISH
Data ElementsData Elements� 32 sites within NC-NE US (MRB 1 & 3), 5 larger river sites dropped after
some initial analyses
� Sites were required to have at least 6 years of sequential invertebrate and fish data to be included in these analyses- mean of 8 years of data max 11yrs
� A single year ‘best’ representative sample was selected for each site
� Biotic Data: � Inverts & fish assemblage , metrics, and species traits data
� Abiotic Data� Habitat information
� Chemistry� Field parameters, Major Ions, Nutrients, DOC and Pesticides
� Single sample prior ecology sample� Mean monthly average across the year of the ecology sample� Pesticide toxicity index (PTI) information
� Stream flow statistics that encompass timing, magnitude, duration, etc.
� GIS: (GIRAS, NLCDe 1992/1995, NLCDe 2001 )� PRISM-Climatic data precipitation and air temperature
Spatial Distribution of NC-NE Trend Sites
MRB1&3 Focus Sites For PresentationMRB1&3 Focus Sites For Presentation
Clinton River
Little Buck Cr.
Green River
Analysis road mapAnalysis road mapPreparatory phase:
-Data aggregation and QA
Exploratory phase:
-Data and range screening and outlier evaluation
-Data reduction, univariate, and multivariate analyses
-Initial model development
-Identifying subset of ancillary variables
Finalizing analyses phase:
-Final model development
-Evaluating potential effects of natural env. variability
Overview of presentation:Overview of presentation:� Highlight significant temporal changes in assemblage
composition
� Highlight the level of consistency in patterns among assemblage and environmental indicators
� Show example relations for inverts/fish with habitat, QW, and hydrology
� Show an example of response patterns within site type and address the question whether we can group responses by site type
Summary of Significant Assemblage Summary of Significant Assemblage
Trajectories (TRENDS) by Site TypeTrajectories (TRENDS) by Site Type
InvertebratesInvertebrates
URB AG REF INT Total
8/11 5/9 2/7 4/5 19/3219/32
FishFish
URB AG REF INT Total
6/11 5/9 4/7 0/5 15/3215/32
BothBoth
URB AG REF INT Total
5/11 3/9 2/7 0/5 10/3210/32
00.10.20.30.40.50.60.70.80.9
** **
****
****
**
****
** **
*
INVT
FISH**- p <0.05* - p<0.10
Urban Sites: Urban Sites: Strength of Strength of Assemblage Trajectories over timeAssemblage Trajectories over time
-0.2-0.1
00.10.20.30.40.50.60.70.8
**
**
** ****INVT
FISH**- p <0.05* - p<0.10
Reference Sites: Reference Sites: Strength of Strength of Assemblage Trajectories over timeAssemblage Trajectories over time
Urban Land Use Change at NCUrban Land Use Change at NC--NE SitesNE Sites
Urban site in suburban Indianapolis
� Changing from AG to URB LU- (1992-2001 NLCDe)
� Little Buck Creek (13%)
Urban Example: Little Buck Creek -45 km2
-headwater stream occasionally goes dry. -GW public supply well located near this site
1992
2000
20092005
Little Buck recent time series
LITTLE BUCK CREEK NEAR INDIANAPOLIS, INTransform: Square root
Resemblance: S17 Bray Curtis similarity
1993
1994
1995
1997
1998
2002
2003
2004
2D Stress: 0.03LITTLE BUCK CREEK NEAR INDIANAPOLIS, INGroup average
199
3
199
7
199
8
200
4
200
2
200
3
199
4
199
5
Samples
100
80
60
40
20
Sim
ilarity
Transform: Square root
Resemblance: S17 Bray Curtis similarity
Ancillary Variables Indicative of change over time:StreamFlow Variables:-Freq. low pulse spells(#events/yr)Habitat:-Wetted X AreaQW:pH
Little Buck Creek continued.
Little Buck examples of trends of select fish, invert, habitat, Little Buck examples of trends of select fish, invert, habitat,
hydrology parametershydrology parameters
GREEN RIVER AT STEWARTVILLE, MATransform: Square root
Resemblance: S17 Bray Curtis similarity
1993
1994
19952002 2003
20042005
2006
2007
2D Stress: 0.09
conn11cs, GREEN RIVER AT STEWARTVILLE, MAGroup average
2007
1993
1994
1995
2005
2006
2004
2002
2003
Samples
100
80
60
40
20
Sim
ilarity
Transform: Square root
Resemblance: S17 Bray Curtis similarity
Ancillary Variables Indicative of change over time:StreamFlow Variables:-Mean Annual Max daily flows (cfs)-Mean Min May flows over POR-Rise Rate (CFS/day)Habitat:Bank Vegetative Cover, Wetted shape, Froude
-107 km2
-Cold water-Gravel/Cobble-Some logging influences
Reference site example: Green River
Green River examples of trends of select fish, invert, habitat, Green River examples of trends of select fish, invert, habitat,
QW parametersQW parameters
CLINTON RIVER AT STERLING HEIGHTS, MITransform: Square root
Resemblance: S17 Bray Curtis similarity
1996
1997
1998
2002
2003
2004
20052006
2007
2D Stress: 0.03
Urban Example: Clinton RiverUrban Example: Clinton River
Ancillary Variables Indicative of change over time:StreamFlow Variables:-Mean Min May flows over PORHabitat:%Riffles, CV of Open Canopy, %Embedd, FroudeQW: Ave P, Ave_pH
Urban site in suburban Detroit
CLINTON RIVER AT STERLING HEIGHTS, MIGroup average
1996
1997
1998
2005
2006
2004
2002
2003
2007
Samples
100
80
60
40
20
Sim
ilarity
Transform: Square root
Resemblance: S17 Bray Curtis similarity
-803 km2
~80% flow from GW-Sewage treatment plant ~1km US-Highly permeable soils
Clinton River Invert. Clinton River Invert.
Richness BasedRichness Based--TraitsTraits
Number Generations per year
Quick Adult lifespan Collector-Filters
Dispersal Ability
Clinton River examples of trends of select fish, invert, Clinton River examples of trends of select fish, invert,
habitat, QW parametershabitat, QW parameters
Habitat
Fish
*Single QW
Inverts
Can we generalized site type response Can we generalized site type response
patterns?patterns?
� We can see some similarity in responses to ancillary variables across sites within a site type (e.g., URB or AG)
� But size, climate, and biogeographical differences represent a significant influence
� Our real power for understanding trends is best evaluated on a individual site basis
Urban InvertebratesUrban Invertebrates
AG InvertebratesAG InvertebratesHYDRO HABITAT QW
Important Variables Important Variables Important Variables
DH1, DH21, DH5, MA20, ML13 Riff%, BKVegCov, Frd, BO%, OTH% P, pH
DH21, FL3, ML5 HydRad, Depth SC, pH
FH9, FL3 W_cv, BFD, BKER, Emb SC
FL3, MA20, ML15 BKVegCov, Emb, Wetshap, Hydrad,SA% NH4
ML13, MH7 Riff%,RUN%, BFW, BKER, SA% NO2NO3
DH21, FL3, ML5 Pool%, D, BKER, SA%, OTH% NO2NO3
FH9, MA15, ML5 BFD, BKvegCov, HydRad, Si% pH
DH21, DL5, FH9, MA15, MA20 W_cv, BKvegCov, BKER,OC, Si% P, pH
ML5 W_cv, OC, WetArea, SI%, OTH% NH4
HYDRO HABITAT QW
Important Variables Important Variables Important Variables
DH1, DH21, DL5, TA2 BKER, Frd P, pH
MA20 Depth_ave NH4, P
RA1, RA3, TA2 Riff%, BFD, WetShape NH4, SC
MA15, MA20, ML5, MH7Riff%, Embed%, OC_Cv, WetSh, Frd
P, pH
DL5, DH21 Pool/riff, D, BKER, WetShape, BO% P, pH
DH1, RA1W/D, BKvegCov, BKER, CO%, OTH% NO2NO3, SC, pH
FL3, MA15, ML13, RA3 D_cv, WetArea, pH
DH21, MA20, ML15 W-cv, BFD, OC, CO_PCT NO2NO3, pH
DL, FH9, FL3, MA20 D,W/D, BKER, WP, BO_p NO2NO3, pH
DL5, MH6, ML13, ML3 Depth_cv NH4, pH
DH1, DH21, DL5, MA15, MA20Pool/riff, Si%, SA%, OTH% P
Summary of findingsSummary of findings--
NC/NE trend sitesNC/NE trend sites
� Established significant patterns of change in fish and invertebrate assemblages at 15 and 19 sites, respectively
� Ten of which are sites significant for both
� We delineated some of the potential mechanisms influencing these assemblages (i.e. habitat, QW, hydrology parameters accounting change)
� Noted some consistency of indicators over time
� Although we can see some similarities for sites within particular site type (AG or URB), the real strength of our analyses is to look at individual site patterns over time
Take home points Take home points
from these trends studies so farfrom these trends studies so far� Change in assemblage composition was evident, but response
differed relative to assemblage type.
� Hydrologic variability appears to be a major factor most of our analyses
� Difficulties with attributing the change in community patterns to ancillary variables.
� Accounting for climatic variability is very important
� Sample size is important-the more data more likely ability to detect change over time.