Applying the Rothamsted Carbon Model for Long-Term Experiments on Japanese Paddy Soils and Modifying It by Simple Tuning of the Decomposition Rate

We applied the Rothamsted Carbon Model (RothC), which was developed for simulating the soil organic carbon (SOC) turnover in non-waterlogged soils, for long-term experiments (16–22 years) on Japanese paddy soils and modified it to accurately simulate the changes in the content of SOC with time in paddy soils. The RothC underestimated the SOC content in all the nine plots at five sites: Gley Soils in Akita and Shimane, Gray Lowland Soils in Toyama and Mie, and Yellow Soil in Oita prefecture. This may be mainly due to the slow decomposition rate of organic matter during the rice-growing season, when submerged soils are waterlogged and subjected to anaerobic conditions. On the other hand, the decomposition of organic matter might be inhibited, not only during the submergence period but also throughout the year in paddy soils because of the difference in the composition of microorganisms between upland and paddy soils. Taking these possibilities of differences in the decomposition rate between upland soils and paddy soils into account, we changed the decomposition rates of the RothC during the submergence period (summer) and the period without submergence (winter), separately. We ran the model many times by changing the decomposition rates for summer and winter separately and tried to identify the optimum combinations of the values of the factors required to change the default decomposition rate, so that the modeled SOC content would be consistent with the values observed in nine plots at five experimental sites. To determine the optimum combinations of the values of the factors, we used two statistical indices, the root mean square error (RMSE), which represents the degree of coincidence, and the mean difference (M), which is a measure of model bias. We found that the optimum combination of the values of the factors required to change the decomposition rate was 0.2 in summer and 0.6 in winter. The modified RothC for paddy soils by simple tuning of the decomposition rate using the values of the factors (0.2 and 0.6) resulted in a much better performance than that of the original RothC for simulating the changes in the SOC content with time in Japanese paddy soils under various climatic conditions, types of soil texture, and management systems. This modified model can be used for the estimation of carbon loss from soils as well as for the planning of suitable organic matter management, at least in Japanese paddy soils.