Litter carbon inputs to the mineral soil of Japanese Brown forest soils: comparing estimates from the RothC model with estimates from MODIS

摘要

The level of organic carbon found in soil is the result of the balance between litter input to the soil and decomposition. Litter input to the soil is closely related to net primary production (NPP); at equilibrium, the NPP is equal to the litter input to soil. Plant litter input to a depth of 30 cm in the mineral soil was estimated for Japanese forest using the Rothamsted Carbon model (RothC) and an average value of soil organic carbon (SOC) content, and was compared with estimated litter inputs from the NPP dataset from Moderate Resolution Imaging Spectroradiometer (MODIS). A Monte Carlo uncertainty analysis of the input SOC was also conducted in order to reveal the sensitivity and uncertainty of the model to input SOC. The litter carbon input calculated using RothC and that derived from MODIS NPP were positively correlated, but the mean estimated litter input from RothC was 17.2% smaller than that estimated from MODIS. Mapping the normalized difference revealed spatial biases in the difference. The discrepancy was probably because of the different temperature controls for the MODIS algorithm and the RothC model, and also our simple assumption in the RothC calculation. This comparison reveals a close link between litter inputs estimated from SOC data and litter inputs estimated from satellite-based NPP data. The discrepancies between the estimates merit further study.