A Novel Approach to Map Soil Organic Carbon Content Using Spectroscopic and Environmental Data

In this study we present a methodology to estimate and map the content of Soil Organic Carbon (SOC) in topsoils from spectroscopic measurements (FTIR-ATR). We used the Walkley-Black method to determine the SOC content in 221 topsoil samples in Galicia. FTIR-ATR spectroscopic data was measured upon the same set of samples. A Random Forest (RF) model, linking SOC and FTIR-ATR data, revealed that the spectral band placed at 1697 cm-1, explains most of the variability of SOC. We crossed the spectroscopic data for this wavenumber with a number of raster environmental data (climate, land use and geology) using Partial Least Squares (PLS) to create maps depicting the spatial distribution of such band. A linear regression model (MLR), relating SOC as the dependent variable and the selected FTIR-ATR bands as independent proxy, shows a good predictive performance (r-squared =0.88; RSME = 2.14; ME = 0.05). We used this MLR, upon the PLS model, to generalize the distribution of SOC in our study area. This approach shows that FTIR-ATR data can be used to directly map SOC while minimizing analytical costs and time.