Estimating Trace Element Emissions Using USGS Coal Data: Fid-Bau Portal

Estimating Trace Element Emissions Using USGS Coal Data

Tumati, Prasad R. / Bilonick, Richard A.

The COALQUAL data base1 was assembled by the United States Geological Survey (USGS) from publicly available coal deposit samples. The USGS collected the data without the benefit of a carefully considered sampling plan aimed at specific coal quality estimates. Therefore, the analysis of the collected data is problematical. This does not necessarily mean that the data set cannot provide useful estimates, only that the statistical techniques used would ordinarily be more complex.

Spatially weighted average trace element concentrations estimated for the entire coal field within western Pennsylvania differed substantially from simple equally weighted average concentrations. The percentage differences ranged from -36% (where the spatially weighted average exceeds the equally weighted average) to +42%. Regression models were constructed to predict the effects of coal washing (i.e., the removal of ash and sulfur) on trace element concentrations. The percent of trace element variance accounted for by ash and sulfur ranged as high as 89%. The predictions based on these models for the only two CONSOL mines in the region tended to be significantly closer to the actual values than those produced by the U.S. Environmental Protection Agency (EPA)2 using equally weighted averages.

Finally, the paper discusses the advantages of providing accurate estimates of trace element concentrations for coal burned at power plants which are part of a relatively small sample set. An accurate and reliable estimate of total national emissions from all power plants can be achieved by using stratified random sampling, where power plant boiler size is used to define the boundaries for each of a small number of strata.