Tuesday, June 5, 2007
468

Use of Geospatial Data to Predict Downstream Impacts of Coal Mining in an Appalachian Watershed

Brent R. Johnson, Adam Haas, and Ken M. Fritz. U.S. Environmental Protection Agency, Office of Research and Development, National Exposure Research Laboratory, Cincinnati, OH 45268

Mountaintop removal and valley filling is a method of mining coal that results in burial of Appalachian headwater streams.  Leaching of fill material often results in elevated ion concentrations below fills.  A primary objective of this study was to quantify downstream extent of mining disturbance using geostatistical tools.  We measured water chemistry parameters (D.O., pH, temperature, conductivity) in situ throughout the partially mined Buckhorn Creek watershed, eastern KY, in summer 2005 (n = 239 sites) and spring 2006 (n = 498 sites).  Sites were located 100-500 m apart and covered a maximum flow-connected distance of 25 km.  Spatial dependency was modeled with empirical semivariograms utilizing asymmetric stream distance.  Temperature semivariograms indicated spatial correlations of ca. 5 km.  Mined tributaries produced strong confluence effects on pH and conductivity that violated assumptions of geostatistical models.  A deterministic model was therefore developed to predict downstream values using measurements from tributary confluences weighted by drainage area.  The model showed strong predictive ability and was robust across a large range of drainage areas and values of conductivity and pH.  Such models may prove valuable to regulatory agencies in evaluating permit applications under Sections 401 and 404 of the Clean Water Act.