Predicting soil carbon by efficiently using variation in a mid-IR soil spectral library

Abstract. Traditional laboratory methods of acquiring soil information remain important for assessing key soil properties, soil functions and ecosystem services over space and time. Infrared spectroscopic modelling can link and massively scale up these methods for many soil characteristics in a cost-effective and timely manner. In Switzerland, only 10 % to 15 % of agricultural soils have been mapped sufficiently to serve spatial decision support systems, presenting an urgent need for rapid quantitative soil characterization. The current Swiss soil spectral library (SSL; n = 4374) in the mid-infrared range includes soil samples from the Biodiversity Monitoring Program (BDM), arranged in a regularly spaced grid across Switzerland, and temporally-resolved data from the Swiss Soil Monitoring Network (NABO). Given the relatively low representation of organic soils and their organo-mineral diversity in the SSL, we aimed to develop both an efficient calibration sampling scheme and accurate modelling strategy to estimate soil carbon (SC) contents of heterogeneous samples between 0 m to 2 m depth from 26 locations within two drained peatland regions (HAFL dataset; n = 116). The focus was on minimizing the need for new reference analyses by efficiently mining the spectral information of SSL instances and their target-feature representations. We used partial least square regressions (PLSR) together with a 5 times repeated, grouped by location, 10-fold cross validation (CV) to predict SC ranging from 1 % to 52 % in the local HAFL dataset. We compared the validation performance of different calibration schemes involving local models (1), models using the entire SSL spiked with local samples (2) and 15 subsets of local and SSL samples using the RS-LOCAL algorithm (3). Using local and RS-LOCAL calibrations with at least 5 local samples, we achieved similar validation results for predictions of SC up to 52 % (R2 = 0.94–0.96, bias = −0.6–1.5, RMSE = 2.6 % to 3.5 % total carbon). However, calibrations of representative SSL and local samples using RS-LOCAL only required 5 local samples for very accurate models (RMSE = 2.9 % total carbon), while local calibrations required 50 samples for similarly accurate results (RMSE