Subsetting Reduces the Error of MIR Spectroscopy Models for Soil Organic Carbon Prediction in the U.S. Great Plains

The high demand for soil organic carbon data to support soil health and climate change mitigation efforts must be met with rapid, accurate, and inexpensive measurement methods. Mid-infrared spectroscopy is a promising complement to conventional soil carbon analysis; however, its practicality depends on the construction and efficient use of a soil spectral library. Subsetting is a calibration optimization technique that can reduce the model prediction error. Nevertheless, the effectiveness of different subsetting criteria has yet to be well explored. The objective of this study was to assess whether several subsetting criteria would result in calibration models with reduced error in the prediction of soil organic carbon content, compared to calibration models constructed from a full spectral dataset. A mid-infrared spectral library composed of soil samples from Nebraska and Kansas was subset by (i) soilscapes, (ii) presence/absence of carbonates, (iii) a combination of soilscape and presence/absence of carbonates, and (iv) wetlands. Partial least squares regression was used to construct calibration models for each subset and the full set. Predictive performance of the subset models was compared to that of their corresponding full set model using several statistical metrics. In addition, several thresholds to rate model performance were used to assess the desirability and reliability of the subset models. Subsetting by soilscape reduced model error by 13 to 55% compared to their full set model counterpart. Subsetting by the presence/absence of carbonates reduced model error by 21 and 46%. Five of the eight models for the combination subsets reduced the model error by 14 to 51%. Subsetting by wetlands reduced model error by 22 and 56%. These results suggest that the tested subsetting criteria were generally effective in improving model performance for soil organic carbon prediction through mid-infrared spectroscopy.