COS 49-7
Investigating stream ecosystem response to landscape change at environmentally meaningful time-scales: the Coweeta-LTER Hazard Site project

Tuesday, August 6, 2013: 3:40 PM
M101A, Minneapolis Convention Center
Jeremy C. Sullivan, Ecology, Odum School of Ecology, University of Georgia, Athens, GA
Edward Gardiner, Capitalize Climate Program Office
Paul Bolstad, Department of Forest Resources, University of Minnesota, St. Paul, MN
David Leigh, Geography, University of Georgia
Mark Scott, S.C. Dept. of Natural Resources
E. Fred Benfield, Department of Biological Sciences, Virginia Tech, Blacksburg, VA
Becky Bixby, Department of Biology, University of New Mexico, Albuquerque, NM
Ted Gragson, Anthropology, University of Georgia, Athens, GA
Catherine M. Pringle, Odum School of Ecology, University of Georgia, Athens, GA
Background/Question/Methods

The southern Appalachian region is an area of high biodiversity and endemism that is undergoing dramatic changes in land cover due to increasing human population and associated exurbanization. The goal of the Coweeta LTER Hazard SiteProject (CHSP) is to examine how two landscape trajectories, predicted by a hedonic land use model (forests converting to low density residential/exurban lands, and agricultural lands converting to urban lands), influence the ecological characteristics of streams. This research is based on a 30-year (2000-2030) proactive sampling strategy, whereby eight headwater streams and their associated watersheds (N=2 forested reference sites, N=3 urbanizing sites, and N=3 exurbanizing sites) are sampled every five years for physical (land cover and geomorphology), chemical (nutrients/ions), and biological (algae, macro-invertebrate, and fish) attributes.  We predict that sites undergoing more rapid land cover change will exhibit greater dissimilarity in these attributes through time, relative to forested streams.  Our approach aims to determine the appropriate sampling frequency and timescales for detecting ecological thresholds in urbanizing landscapes by comparing data produced by the 5-year sampling frequency of the CHSP to data produced by an annual (2 of 8 hazard sites are being sampled annually for a subset of CHSP variables from 2009-present) sampling schedule.    

 Results/Conclusions

Here we present evidence that streams in rapidly developing watersheds are exhibiting substantial shifts in biotic and abiotic characteristics relative to streams in forested watersheds. GIS analysis of land cover data support predicted  landscape trajectories, with urbanizing sites showing the greatest increases in watershed building density (44% ±9.07 from 2000-2010), followed by exurbanizing sites (increase of 12%±14.96 from 2000-2010).  Predictions of increasing stream community dissimilarity in more rapidly developing watersheds are supported by fish assemblage data (urbanizing sites, D2000, 2010=0.47±0.09; exurbanizing sites, D2000, 2010=0.36±0.05; forested sites, D2000, 2010=0.32±0.08).   Streams in more rapidly developing watersheds exhibited higher concentrations of NO3-, Ca2+, K+, and Na+,  with dramatic increases in NO3- concentrations between sampling dates in 2000 and 2010 (83% increase in urbanized sites vs. 31% increase in forested sites). Pebble counts show streams in urbanizing watersheds have smaller bed particles (D50=24.83mm± 6.69) than forested (D50=90.60mm± 12.16) and exurbanizing (D50=68.37mm± 13.87) sites.  Results suggest that streams in urbanizing watersheds exhibit relatively rapid ecological change.  Continued long term sampling (2015-2030) of the Coweeta LTER Hazard Sites will provide us with a more predictive understanding of the complex interrelationship between land cover change and stream ecology in this rapidly urbanizing region of the southern Appalachians.