R-language scripts for RIVPACS-type predictive modeling

             A RIVPACS-type predictive model predicts the taxonomic assemblage of macroinvertebrates, fish, or periphyton that one would expect to find in an aquatic ecosystem, if that ecosystem were in a minimally-disturbed “reference” condition. The expected assemblage is then compared with the assemblage that is observed by sampling the ecosystem. Discrepancies between the two assemblages indicate the degree of ecosystem stress or impairment.

A full discussion of predictive models is provided by the Western Center for Monitoring and Assessment of Freshwater Ecosystems

This page describes scripts for building and applying predictive models, written in the R computing language.

The scripts were written by John Van Sickle, under the support of the US Environmental Protection Agency, National Health and Environmental Effects Research Laboratory, Western Ecology Division. The scripts (Version 4.2.1) were finalized Jan. 1, 2011. They are designed for experienced R programmers. Users will need to modify some scripts to suit their particular data sets.

To obtain the scripts free, contact vansicklej@peak.org

Script features include:

1. Creation and manipulation of site-by-taxa data matrices, including random subsampling to a fixed count.
2. Choice of different dissimilarity measures and clustering algorithms, including flexible-beta clustering and options for dendrogram pruning..
3. Choice of all-subsets or stepwise model selection methods for discriminant function models.
4. Can build and use a Random Forest model, rather than discriminant functions, to predict site group membership.
5. Predictions for new sites, including assessment of site outlier status.
6. Can save final model as an R object and export to users, with a stand-alone script for assessing new sites. Can also export your model for submission to the Western Center for Monitoring’s web-accessible modeling system.
7. Calibration and predictions for null models.
8. O/E and BC indices.
9. Detailed comparisons of expected and observed taxa at user-selected sites.

For additional information see the following articles:

Van Sickle, J. 2008. An index of compositional dissimilarity between observed and expected assemblages. Journal of the North American Benthological Society 27, 227-235.

Van Sickle, J., D.P. Larsen and C.P. Hawkins. 2007. Exclusion of rare taxa affects performance of the O/E index in bioassessments. Journal of the North American Benthological Society 26, 319-331.

Van Sickle, J., David D. Huff, and C.P. Hawkins . 2006. Selecting discriminant function models for predicting the expected richness of aquatic macroinvertebrates. Freshwater Biology 51, 359-372.

Van Sickle, J., C.P. Hawkins, D.P. Larsen and A.T. Herlihy. (2005). A null model for the expected macroinvertebrate assemblage in streams. Journal of the North American Benthological Society 24, 178-191.