Great River Ecosystems Reference Condition Workshop, January 10-11, 2006, Cincinnati, OH ~ Abstracts

A reference condition approach for the Great Rivers of the Central Basin: The
Ohio, Missouri and Upper Mississippi Rivers

Ted Angradi, US EPA, Office of Research and Development, Duluth, MN
David Bolgrien, Terri Jicha, Brian Hill, Mark Pearson, and Debra Taylor, US EPA,
Office of Research and Development, Duluth, MN

Empirical bioassessment of rivers is based on the comparison of conditions of sampled
sites to conditions at sites in the same or comparable resource considered to be in
"reference" condition. In the Environmental Monitoring and Assessment Program for
Great River Ecosystems (EMAP-GRE) we use "least disturbed condition" (LDC) as our
definition of reference. Our underlying (optimistic) assumption is that although none of
the Great Rivers of the Central Basin is in pristine condition, and there are no other
closely comparable large and pristine Central Basin rivers, there is variation in condition
(or degree of impairment) along each river that provides the scope for bioassessment.
We use a 3-phased approach to obtaining our set of internal least disturbed sites on each
river. We use a GIS model to find locations on each river with the highest probability of
being in LDC condition. We use a natural gradient approach to filter out the best (and
worst) sites from all the sampled sites. Finally, we will verify our reference site selection
using metrics based on biotic assemblages (e.g., fish, macroinvertebrates). Our GIS
models allows us to score every potential sample location on each river based on the
proximity to upriver and local human disturbances including tributaries, dams, NPDES
permits, urban areas, river crossings, and floodplain land use. Tributary influence on the
mainstem is weighted by tributary watershed land use. Model outputs are used to define
LDC candidate reaches that are then randomly sampled using a probability design.
Model-suggested, and all other sites that are actually sampled are filtered by scoring each
site based on multiple (>12) abiotic metrics relative to a natural gradient for each metric
(river mile as a surrogate for watershed area). Abiotic filtering metrics include water
chemistry (e.g., nutrients, chloride), habitat (woody debris, sediment toxicity, riparian
vegetation) and landscape metrics (scores from the GIS proximity model). Comparing
biotic metrics between filtered LDC sites and the entire population of probability sites
will provide a test of the efficiency of our approach. This abstract does not necessarily
reflect EPA Policy.

Dr. Angradi is research biologist with the EPA Mid-Continent Ecology Division
Laboratory in Duluth, MN. He has been involved in large river research for about 18
years, and has worked on the Snake River in Idaho, the Colorado River in Arizona, and
the Missouri River in North Dakota.

-------