Emily Watson May 5,2011 Downstream Strategies & FHPs Create - - PowerPoint PPT Presentation

Tennessee/Cumberland Aquatic Habitat Assessment

Emily Watson May 5,2011

Downstream Strategies & FHPs

 Create spatially-explicit habitat assessment models for

each of the Midwestern FHPs…and now SARP

 Create an integrated GIS decision support tool  Create a regional representation of habitat condition

Modeling Approach

Landscape Predictor Data

• Natural and Anthropogenic
• Local vs. US Network vs. DS

Network vs. Regional Stream or Lake Response Data

• Environmental Data
• Fish Data
• Assemblage
• Abundance
• Presence-Absence
• Index of Biotic Integrity
• Community Metrics

Model Results

• Response variable predictions @

1:100K SLW scale

• Predictor variable importance

weightings

• Stressor-Response functions
• Estimates of model uncertainty

Post-Modeling Results

• Cumulative Natural Habitat

Quality Index (CHQI)

• Cumulative Anthropogenic

Stressor Index (CASI)

• CHQI and CASI accumulated

from 1:100K SLW up to HUC12.

INPUT OUTPUT

BOOSTED REGRESSION TREES

Local vs. Network vs. Regional

Network area draining to focal SLW (network variable) Local area draining to focal SLW (local variable) Dam downstream of SLW (regional variable) Ecoregion (regional variable) Local area draining to focal SLW (local variable)

Example: Great Lakes Basin BKT

 Brook Trout Habitat Condition

 FHPs determined HUC-8 watersheds where brook trout are

likely to be found

 FHP provides a list of top stressors or influences for the

probability of brook trout (predictor variables)

 Collect regional fish sampling data (response variables),

provided by FHP

 Run and apply model  Map and interpret results

Response Variable: BKT

Variable Variable Description Relative Influence Cumulative Percent

POPDENS Local population density 9.044 100 MINELEVSMO Minimum stream elevation 7.732 91 LU_WETPC Network wetland land cover (percent) 6.838 83 SOIL_C_PC Network soil class C land cover (percent) 6.757 76 SLOPE Watershed slope 6.277 70 SOIL_A_PC Network soil class A land cover (percent) 6.090 63 LU_FORPC Network forest land cover (percent) 4.570 57 IMPERVSC Network impervious surface cover (percent) 4.097 53 ROADCRC Network density of road crossings 3.601 49 LU_FORP Local forest land cover (percent) 3.192 45 AREA_SQKMC Upstream drainage area 3.187 42 CATTLE Local density of cattle 3.054 39 LU_DEVPC Network developed land cover (percent) 2.926 36 GAP_TEMP Predicted thermal regime (cold, cool, warm) 2.321 33 WATER_QW Local groundwater withdrawal amount 2.222 30 SOIL_B_PC Network soil class B cover (percent) 2.093 28 WATER_SWC Network surface water withdrawal amount 1.967 26 LU_AGPC Network agriculture land cover (percent) 1.941 24

Analogous to a Standardized Regression Coefficient Variables with top 75% of Influence; Red = Anthropogenic

Predictor Variable Weights

Probability of Brook Trout

Natural Habitat Quality Index

Predictor-Response Functions

• 1.5
• 1.0
• 0.5

• 4
• 3
• 2
• 1

(1.00) (0.87) (0.81) CHQI is calculated for each 1:100K Segment Level Watershed Independent functional relationship between the fish response variable and natural landscape attributes.

• Min. Elevation

Soil C % Land Cover Average Slope

• Marg. Effect on Logit

Anthropogenic Stress Index

Predictor-Response Functions

CASI is calculated for each 1:100K Segment Level Watershed

• 1.0
• 0.5

• 1

(1.00) (0.76) (0.5)

• Pop. Density (#/km2)

Wetland Cover (%) Forest Land Cover (%)

• Marg. Effect on Logit

Independent functional relationship between the fish response variable and anthropogenic stressors.

NHQI vs. CASI

Restoration Priorities Protection Priorities

Restoration/Protection Priorities

Decision Support Tool

 Integrated within ArcMap 9.3  Index calculator  Comprise programming model for weighting decisions based on current

conditions

 Downstream future conditions based on user input and model output.  Reporting all the data with one click, producing customized report of

catchments and watershed conditions and contributing factors.

• 1

• 1.0
• 0.5

 Great Lakes Basin FHP

 Brook trout, walleye, lake sturgeon, darters (guild)

 Driftless Area Restoration Effort

 Brook/brown trout, sculpin, Am. Brook lamprey, smallmouth bass

 Fishes and Farmers of the Upper Mississippi River

 Fish/mussels species richness, % native species

 Great Plains FHP

 Topeka shiner, others to be determined

 Midwest Lakes

 Sportfish indices, keystone species, water quality

Potential FHPs Response Variables

Moving Forward: Ohio River Basin - Predictor Variables

Predictor Variable Natural/Anthropogenic Surface & Ground Water Use Estimates A Land use/Land cover (2001) A National Inventory of Dams A TIGER Roads US Census/ESRI (2000) A Agriculture Census (2000): Cattle A Mines & Mineral Plants A Superfund Sites A Toxic Release Inventory A NPDES A Climate Data N Elevation N Soil Data N National Wetland Inventory N National geologic data N

• 1. Large & Great Rivers Signature Fish (Rivers Index)
• 2. Medium/Small Rivers, Streams, & Headwaters

Signature Fish (Streams Index)

• 3. Presence/Absence of Fluvial Mussels
• 4. Macroinvertebrate Index/Metric
• 5. Modified Index of Centers of Diversity for Fish

(MICD)

• 6. Fish IBI scores

Moving Forward: Ohio River Basin – Response Variables

 1st three Response Variables submitted to DS on April 1st

 Signature Fish Abundance Metrics

 Rivers index (Paddlefish, Blue Suckers, Sauger, Sturgeon, etc)  Streams Index (Darters, SM Bass, SP Bass, etc)

 MICD for Fish

 Still in the works

 Presence/Absence of Fluvial Mussels (by May?)

 Macroinvertebrate Index/Metric Scores (data in, not processed)

 Fish IBI Scores (raw data available)

 May initially be limited to 5 Response Variables

Moving Forward: Ohio River Basin – Response Variables

Response Variables: Development

 Requested fish data from throughout the entire basin

 State WQ agencies targeted

 21,008 events entered into Access DB

 ORBFHP area only

 11,712 “Community” samples

 Excludes targeted or non-standardized samples  Excludes events without abundance data

 9,368 events matched to GIS stream reach COMID  5,566 events marked as most recent event for a COMID

Data Sets Compiled

 IN DEM – 1996-2007  IL EPA – 1990–2004  KY DOW – 1999-2010  MD DNR – 1995-2010  NC DWQ – 1998-2009  NY DEC – 1990-2009  OH EPA – 2000-2007  PADEP – 2000-2010  VA DEQ – 1993-2010  WV DNR – 1990-2006  NKY SD1 – 2001-2010  ORSANCO – 2000-2010  USGS – 1993-2009  U OF S. MISS – 1990-1995

Data Sets In Progress

 IN DEM – 2008-2010  TN Wildlife Res Agcy – 1994-2008  Tulane – 1998-2000  NC State – 2003  U of Florida – 2005-2010  Yale – 2006-2008  PA Fish & Boat Co.  PADEP EMAP-GRE co-op – ’07-’10  Eastern Ky U  TN Valley Authority  USEPA NRSA

Data Sets Requested

 Pennsylvania  Kentucky large rivers  Green River basin

Gaps

Modified Index of Centers of Diversity

 Scores sites based on highest abundances of the rarest

species relative to all sites in the basin

density of a species ata site * 10,000 sum of densities of the same species at all sites Relative Density total number of species fromall sites sum of the relative densities for all species at a site MICD

MICD

Signature Fish Species Index

Large & Great River spp (i.e., Rivers Index)

 Classified all spp by stream size based on entire dataset

using NHD+ stream orders

 Determined which species nearly exclusively occur at 6th

• rder or greater sites

 Eliminated more tolerant species

 Classified 38 spp as Signature Spp for Large & Great Rivers

 Need to be more restrictive?  Can distribute list of species to anyone interested

Signature Fish Species Index

Rivers Index



Used individual box plots of abundances for each species to create scores of 1-5 for each species’ abundance at a given site



Scores for each spp were then added together to give

• verall score for each site

• 5

5 10 15 20 25 30 35 SumOfCount

upper non-outlier range 75% 25% median

Signature Fish Species Index

Rivers Index



Modifications



Extremely rare spp score 5 if >=1 individual



American eel, lake sturgeon, paddlefish, slender madtom, slender chub, alligator gar,



Rare spp score 3 if 1 ind and 5 if >1 ind



Shoal chub, shovelnose sturgeon, silver lamprey, spotted gar, blacknose shiner, silverband shiner, pugnose minnow,

Rivers Index

Signature Fish Species Index

Medium, Small, & Headwater Streams (i.e., Streams Index)



Species chosen based primarily on strategic plan:



Smallmouth bass, spotted bass, Etheostoma spp, and Percina spp



Headwater species added



Scored same as Rivers Index



Scored using all Etheostoma spp as one taxa and all Percina spp as another



Also true for Phoxinus spp, Cottus spp, & Clinostomus spp

Streams Index

Best Sites – All 3 Response Variables

 Review draft models – WILL NEED FAST TURN-AROUND

 Determine effectiveness of response variables  Discuss need for Blocking Variables

 Upper vs lower basin

 Add new Predictor Variables if available

 Active River Area – available for upper basin

 Revise existing response variables

 Incorporate new data – try to eliminate <2000 (759 of 5566)  Modify processes where needed

 Submit new response variables

 Mussels  IBIs? Macroinvertebrates?

Next Steps