Modeling the thermic soil temperature regime boundary of the eastern United States

Within U.S. soil taxonomy,soil temperature regimes (STRs) are recognized and incorporated in the classification and mapping of soils. Although knowledge of STR can inform crop selection, yield estimates, and nutrient management decisions, detailed and current maps of STRs derived from measured data are not available for most of the United States. Soil temperature measurements at the prescribed depth of 50 cm were summarized for 69 Soil Climate Analysis Network (SCAN) stations in the eastern United States. Mean annual soil temperature (MAST) was calculated for the 2005-2015 period from monthly averages calculated from daily soil temperature observations. A prediction equation for MAST was developed using mean annual air temperature (MAAT) from the PRISM (Parameter-elevation Regressions on Independent Slopes Model) group as a predictor: MAST = 2.49 + (0.91 x MAAT). The adjusted r(2) was .96 with a RMSE of 0.74 degrees C and mean absolute error of 0.59 degrees C based on 10-fold cross-validation. Orographic influence is incorporated in PRISM MAAT, yielding a simple model. The MAST surfaces were calculated by applying the prediction equation to MAAT grids for the 1895-2015 period. The yearly MAST surfaces allow estimation of STR over time and for various summary periods including the 1895-2015 record, 30-yr normal, and probabilities based on individual years for the 1895-2015 period. Based on this investigation, the mesic-thermic STR boundary occurs north of the presently recognized location, is dynamic, and is supported by existing native plant communities associated with thermic STR.