Particulate Organic Matter at the Field Scale: Rapid Acquisition Using Mid‐Infrared Spectroscopy

Abstract

Modeling global C cycles requires in‐depth knowledge about small‐scale C stocks and turnover processes, yet different soil organic C (SOC) pools reveal considerable spatiotemporal heterogeneity at the field scale, which is scarcely known due to the considerable workload associated with traditional fractionation procedures. We investigated the potential of mid‐infrared spectroscopy combined with partial least squares regression (MIRS‐PLSR) for rapid assessment of different particulate organic matter (POM) pools and their spatial heterogeneity at the field scale. Locally calibrated prediction models estimated the contents of SOC, POM of three size classes (POM1: 2000–250 μm; POM2: 250–53 μm; and POM3: 53–20 μm), and lignin contents for 129 locations in a 1.3‐ha test field. Relations between the parameters were described using correlation analysis and fuzzy‐ κ statistics. All parameters were predicted successfully by applying local calibrations for MIRS‐PLSR ( R 2 = 0.77–0.96). The prediction model for POM1 chiefly relied on specific signals of lignin and cellulose; contents of POM2 were estimated by spectral bands assigned to degradation products as aliphatic C–H groups and aromatic moieties; carboxylic groups essentially contributed to the prediction of POM3. There was a close spatial relation between the coarse POM1 and POM2 fractions and lignin (κ = 0.77), which largely also explained variations in bulk SOC. In contrast, POM3 exhibited a less deterministic pattern in the field, thus contributing little to the spatial variation in SOC content.