Quantifying Change in Wetland Extent & Wetland Quality Effects on a Western and Clark s Grebe Breeding Population Eagle Lake, CA: 1998-2010 Renée E. Robison 1, Daniel W. Anderson 2,3, and Kristofer M. Robison 1 1 2 3
Outline Background Focal Species Study Site Hypothesis Methods Results Discussion 2
Focal Species Clark s Grebe (Aechmophorus clarkii) Western Grebe (Aechmophorus occidentalis) 3
Western and Clark s Grebe Ecology Colonial nesting species Inland lakes and reservoirs Highly wetland dependent Flooded wetland vegetation Nests placed in water >25cm deep Floating nest structures jeffrichphoto.com 4
Study Site EAGLE LAKE, CALIFORNIA Lassen County 27km (17mi) N of Susanville Closed-drainage system Lake level highly variable Run-off Precipitation Evaporation Ground-flow 5
Study Site EAGLE LAKE, CALIFORNIA Intermountain west population Grebe nesting locations Spaulding Stone s Bay Grebe monitoring 1998-2010 6
Hypothesis Habitat Availability Lake Level Habitat Quality Populationlevel breeding success 7
Outline Background Methods Data Collection Remote Sensing Statistical Analysis Results Discussion 8
Methods - Data Collection FIELD SURVEY Pelagic-strip transects Out to 200m LAKE LEVEL DATA Provided courtesy of Lassen County Public Works 9
Methods - Remote Sensing IMAGERY ACQUISITION LandSat Thematic Mapper (TM) 4 5 USGS Earth Explorer website Annual peak nesting dates + 2 weeks Dependent on: Availability Cloud cover 10
Methods - Remote Sensing PRE-PROCESSING Generate composite rasters (TM bands 1 7) Display using natural color band combination (R: 5, G: 4, B:3) Digitize lake surface area polygons Clip rasters to surface area polygons using Raster Processing Clip tool 11
Methods - Remote Sensing IMAGE ANALYSIS Classification Using ArcMap s Image Analysis tool Normalized Difference Vegetation Index (NDVI) NDVI estimates: Relative biomass Density Intensity 12
Methods - Remote Sensing POST-PROCESSING Using ArcMap s Spatial Analyst Reclassify tool Manual classification with 2 classes Non-wetland Wetland Raster Conversion Conversion Tool s Raster to Polygon tool Wetland polygons 13
Methods - Remote Sensing HABITAT INDICES Habitat Availability Index % Wetland Entire lake surface Clip NDVI raster to wetland polygons Add Field in clipped NDVI attribute table Habitat Quality Index Mean of NDVI pixel values Entire wetland extent surface Field Calculator (Value*Count) Column statistics 14
Methods - Statistical Analysis INFORMATION-THEORETIC APPROACH N = 11 11 a-priori candidate models Ranked via Akaike Information Criterion (AIC c ) Lowest AIC c Highest Akaike weight (w i ) Lowest Evidence Ratio VARIABLES Response Productivity (YY Ad -1 ) Predictors Lake Level (m AMSL) Habitat Availability Habitat Quality Year 15
Outline Background Methods Results Inter-annual changes Predictor variables Response variable Model Selection Discussion 16
Lake Level (m AMSL) Results - Lake Level 1557.5 1557 1556.5 1556 1555.5 1555 1554.5 1554 1553.5 (No Data) 1553 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Year Data courtesy of Lassen County Public Works 17
Results Lake Level 18
Habitat Availability Index Results - Habitat Availability 0.045 0.04 0.035 0.03 0.025 0.02 0.015 0.01 0.005 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Year 19
Results - Habitat Availability 20
Habitat Availability Animation 21
Habitat Quality Index Results Habitat Quality 120 100 80 60 40 20 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Year 22
Productivity (YY Ad -1 ) Results - Productivity 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 (No Data) 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Year 23
Results Model Selection Candidate Models AIC c (1) Habitat Availability -4.67 (2) Habitat Quality -3.80 (3) Year, Habitat Availability, Interaction 1.11 (4) Lake Level, Habitat Availability, Interaction 1.31 (5) Habitat Quality, Habitat Availability, Interaction 2.76 (6) Year 2.93 (7) Lake Level 3.72 (8) Year, Habitat Quality, Interaction 4.31 (9) Lake Level, Habitat Quality, Interaction 5.50 (10) Year, Lake Level, Interaction 9.57 (11) Year, Lake Level, Habitat Availability, Habitat Quality 17.25 24
Results Model Selection < 95% CREDIBILITY SET OF MODELS PREDICTING PRODUCTIVITY (YY AD -1 ) Model AIC c AIC c w i w i Ratio Evidence (1) Habitat Availability -4.67 0.55 (2) Habitat Quality -3.80 0.87 0.35 0.90 1.54 (3) Year, Habitat Availability, Interaction 1.11 5.78 0.03 0.93 18.01 25
Results - Habitat Availability Productivity = 0.59+-0.75*Exp(-97.33*[% Wetland]) Reproductive rates increase exponentially with increasing wetland habitat availability. 26
Results Habitat Quality Productivity = 0.14 + 0.0054*Habitat Quality Reproductive rates are positively correlated with more robust wetland vegetation. 27
Results HABITAT AVAILABILITY VS. PRODUCTIVITY LAKE LEVEL VS. HABITAT AVAILABILITY 28
Results HABITAT QUALITY VS. PRODUCTIVITY LAKE LEVEL VS. HABITAT QUALITY 29
Outline Background Methods Results Discussion Conclusions 2012-2015 Drought Management Recommendations 30
Discussion Our study Habitat availability and quality Lake level Drought continues No nesting at Eagle Lake since 2011 No habitat available Consider over-water nesting species Allow optimal habitat availability Not strict stabilization 31
Acknowledgements Jeff Davis, Frank Gress, Khem So, John & Tracy Crowe, & John Eadie American Trader Trustee Council Kure/Stuyvesant Trustee Council Henry A. Jastro & Peter J. Shields Research Award 32
Questions? Renée E. Robison rrobison@colibri-ecology.com 33