PS 81-182 - PiSCES: Pi(scine) stream community estimation software: A tool for nationwide fish assemblage predictions

Friday, August 11, 2017
Exhibit Hall, Oregon Convention Center
Mike Cyterski1, John M. Johnston1, Mike O. Galvin1, Craig Barber1, Rajbir Parmar1, Amber R. Ignatius2 and Kurt L. Wolfe1, (1)Office of Research and Development, U.S. Environmental Protection Agency, Athens, GA, (2)Oak Ridge Institute for Science and Education, U.S. Environmental Protection Agency, Athens, GA

What species of fish might someone find in a local stream? How might that community change as a result of changes to characteristics of the stream and its watershed? PiSCES is a browser-based toolkit developed to predict a fish community for any NHD-Plus stream reach in the conterminous United States. The tool uses HUC-based distributional information for over 1,000 native and non-native fish species obtained from NatureServe, the USGS, and the Peterson Field Guide to Freshwater Fishes of North America. Occurrence envelopes, which define a range of conditions over which a species is most likely found, were developed based on statistical analysis of geographically diverse fish sampling information, and used to add refinement to community predictions beyond distributional polygons.


In addition to using current geographic distributions, the PiSCES tool predicts communities that reflect information on species rarity, stream size preferences, and occurrence envelopes for water quality metrics (pH, conductivity, turbidity, dissolved oxygen) and stream characteristics (drainage area, mean width, mean depth and slope). It can also create an abundance distribution for a predicted fish community based on maximum body sizes of community members. PiSCES is one component in an array of web-based services and science modules housed within the EPA QED (quantitative environmental domain), which is currently under development. PiSCES has been used to produce initial fish communities as input for bioaccumulation modeling within an integrated environmental modeling system.