You are here:
Spatial prediction models for the probable biological condition of streams and rivers in the USA
Citation:
Fox, EricW, R. Hill, S. Leibowitz, Tony Olsen, AND M. Weber. Spatial prediction models for the probable biological condition of streams and rivers in the USA. International Statistical Ecology Meeting, Seattle, WA, June 28 - July 01, 2016.
Impact/Purpose:
This abstract concerns the development of spatial statistical models for the probable biological condition of rivers and streams within the USA. The application of these models is the creation of maps displaying the predicted probability that streams and rivers within the USA are in good condition. The maps may have very useful applications towards watershed restoration and conservation efforts.
Description:
The National Rivers and Streams Assessment (NRSA) is a probability-based survey conducted by the US Environmental Protection Agency and its state and tribal partners. It provides information on the ecological condition of the rivers and streams in the conterminous USA, and the extent to which they support healthy biological condition. An important problem is the prediction of stream integrity at new, unsampled locations. Using random forests (Brieman, 2001) we develop a model to predict the probability that a stream is in good (or conversely poor) biological condition. The model is fit to categorical response data consisting of 1365 NRSA survey sites and their designation as being in good or poor condition according to the macroinvertebrate Multimetric Index (MMI). The predictor data consist of over 200 GIS-level catchment and watershed variables from the EPA’s Stream-Catchment Dataset (Hill et al., 2015). The out-of-bag performance of the random forest classifier is evaluated with classification rates, the area under the curve, and other graphical summaries. We find that the random forest model performs remarkably well according to these diagnostics, and is well-suited for modeling stream condition classes with a large predictor set. We also address issues with variable selection and model stability, and compare other statistical modeling approaches for modeling MMI with random forests. The application of building the random forest model is the creation of maps displaying the predicted probability of good biological condition for all catchments within the NRSA sampling frame. The maps may have very useful applications towards watershed restoration and conservation efforts.