Comparing Field Sampling and Soil Survey Database for Spatial Heterogeneity in Surface Soil Granulometry: Implications for Ecosystem Services Assessment

Lithospheric-derived resources such as soil texture and coarse fragments are key soil physical properties that contribute to ecosystem services (ES), which can be valued based on “soil” or “mineral” stocks. Soil survey data provides an inexpensive alternative to detailed field measurements which are often labor-intensive, time-consuming, and costly to obtain. However, both field and soil survey data contain heterogeneous information with a certain level of variability and uncertainty in data. This study compares the potential of using field measurements and information from the Soil Survey Geographic database (SSURGO) for coarse fragments (CF), sand (S), silt (Si), clay (C) and texture class (TC) in the surface soil (Ap horizon) for the 147-hectare Cornell University Willsboro Research Farm, NY. Maps were created based on following methods: a) utilizing data from the SSURGO database for individual soil map unit (SMU) at the field site and using representative or reported values across individual SMU; b) averaging the field data within a specific SMU boundary and using the averaged value across the SMU; and c) interpolating field data within the farm boundaries based on the individual soil cores. This study demonstrates the important distinction between mapping using the “crisp” boundary soil SSURGO databases compared to the actual spatial heterogeneity of field interpolated data. Maps of CF, S, Si, C, and TC values derived from interpolated field core samples were dissimilar to maps derived by using averaged core results or SSURGO values over the SMUs. Dissimilarities in the maps of CF, S, Si, C, and TC can be attributed to several factors (e.g., official soil series data being collected from “type locations” outside of the study areas). Correlation plot of clay estimates for each SMU showed statistically significant correlations between SSURGO and field-averaged (r = 0.823, p = 0.003) and field-interpolated clay (r = 0.584, p = 0.028) estimates, but no correlation was found for CF, S, and Si.