Geoderma (Nov 2024)
Using combustion analysis to simultaneously measure soil organic and inorganic carbon
Abstract
Soil organic carbon (SOC) and soil inorganic carbon (SIC) are of longstanding interest due to their relationship with other key soil properties and indications for soil health and carbon storage. At the USDA-NRCS Kellogg Soil Survey Laboratory (KSSL), total carbon (SOC + SIC) is determined via dry combustion analysis, while calcium carbonate (CaCO3) equivalent is determined via manocalcimetry. For calcareous (carbonate bearing) samples, SIC is estimated as 12 % of CaCO3 equivalent, while SOC is estimated as the difference between measured total carbon and estimated SIC. An alternative dry combustion method for the measurement of SOC and SIC pools was evaluated with the goal of directly measuring – not estimating – inorganic and organic carbon on calcareous samples. The alternative temperature ramp dry combustion (TRDC) method comprises two variants that differ in ramp cycle and carrier gases used. One variant operates under continuous oxygen and has temperature ramp plateaus of 400, 600 and 900 °C; thus, it is referred to as the non-gas switching variant or TRDCNGS. The other variant operates under oxygen until 400 °C, then switches to nitrogen gas for a ramp to 900 °C, then reintroduces oxygen at 900 °C; thus, it is referred to as the gas switching variant or TRDCGS. Both variants were applied in duplicate to 110 diverse samples, including 32 calcareous samples, from across the USA that had been previously characterized by the KSSL. Samples were selected to capture wide variability in carbon contents. Comparing carbon data outcomes with data from the legacy KSSL methods revealed the TRDCGS variant as best for calcareous samples, whereas the TRDCNGS variant was preferred for non-calcareous samples. A combination of the two method variants offers an accurate and direct measurement of SOC and SIC. For calcareous samples, mid-infrared (MIR) spectral analysis demonstrated TRDC method as slightly more accurate than legacy KSSL methods for estimating SOC and SIC.
