Objectives This study aimed to investigate calibration methods for improving the accuracy of the RothC model in paddy fields of Northeast China, and to evaluate the spatiotemporal dynamics of soil organic carbon (SOC) under different straw return scenarios in typical rice-planting areas of Northeast China, thereby providing a basis for enhancing the soil fertility improvement effect of straw return.

Methods The study area was the 850 Farm located in Qiqihar City, Heilongjiang Province. Soil organic matter data for this farm were obtained from HWSD2, land use data from the 2020 CLUD datase, and soil type information from the 1:4000000 Soil Map of China compiled by the Institute of Soil Science, Chinese Academy of Sciences. The RothC model was used to predict the dynamic changes of the soil organic carbon pool in paddy fields at both the experimental site and regional scales. The site-scale model was calibrated and validated using measured data collected from the field and long-term positioning observation data. Finally, by integrating rice planting area, soil properties, and meteorological data of Qiqihar City, the spatiotemporal changes of SOC under different straw return scenarios at the regional scale were simulated.

Results The tested models included the RothC model and its two modified versions, RothC_p and RothC_0.6. Each model was calibrated using three methods (M1, M2, M3), resulting in nine simulation models. All models and calibration methods overestimated the observed trends to varying degrees. The RothC_0.6 model, which represents the anaerobic environment of paddy fields by fixing the moisture factor, aligned more closely with the actual decomposition processes in cold-region paddy fields. Consequently, it performed best in simulating local paddy soil organic carbon dynamics, especially under the treatment with only inorganic fertilizer and no straw return (NPK), where simulation accuracy was extremely high (normalized root mean square error, nRMSE < 1%). For the treatment with inorganic fertilizer plus straw return (NPKS), the initial simulations overestimated SOC (nRMSE: 11.21%−12.12%). However, after semi-quantitatively adjusting the carbon input based on actual field management practices, model performance was substantially improved (nRMSE decreased to 1.42%−2.34%). Therefore, the RothC_0.6 model calibrated by the transfer function method was selected for subsequent regional-scale simulations. Twenty-year regional scenario simulations in Qiqihar showed that Remaining root stubble (CK), with limited carbon input, could not compensate for mineralization losses, dramatically decreasing SOC stock by 10.49 Tg relative to the baseline. Remaining root stubbles and shallow rotary tillage in Spring (CQX) sustained SOC stock with only a 0.7 Tg reduction, as carbon input roughly balanced decomposition. In contrast, full straw return significantly increased SOC stock. Among the straw return practices, All stubbles were cut and buried into soil and watering in Autumn (QSQM), with its semi-quantified carbon input closer to the actual amount and a more favorable decomposition environment during the fallow period, demonstrated significantly better carbon sequestration than All the straw and root stubbles remained in field and deep ploughing in Autumn (QSF), increasing SOC stock by 3.93 Tg (QSF) and 13.83 Tg (QSQM) relative to the baseline. The spatial distribution of SOC density in Qiqihar paddy fields was generally lower in the southwest and higher in the northeast, and straw return further enhanced the spatial heterogeneity of SOC density.

Conclusions The RothC_0.6 model calibrated by the transfer function method (M3) achieved a mean deviation (MD) of −1.24, a normalized root mean square error (nRMSE) of 2.34%, and a consistency index (d) of 0.45. This method is computationally simple, efficient, and suitable for simulating SOC dynamics in cold-region single-cropping paddy fields with no straw return. Regional simulations based on the calibrated model revealed that Remaining root stubble (CK) and Remaining root stubbles and shallow rotary tillage in Spring (CQX) are insufficient to maintain the existing SOC stock, while full straw return significantly increases SOC stock by enhancing the absolute content of the humus carbon pool. The spatial pattern of SOC density in Qiqihar paddy fields is characterized by lower values in the southwest and higher values in the northeast. The northwestern region of Qiqihar has higher carbon sequestration potential than other areas, and straw return further increases the spatial heterogeneity of SOC density. Therefore, priority should be given to promoting full straw return in Northeast China paddy fields, and differentiated soil fertility management strategies are recommended for the high-potential northwestern region. Among all practices, All stubbles were cut and buried into soil and watering in Autumn (QSQM) achieved the best carbon sequestration effect. This study provides a scientific basis and decision support for straw resource management and black soil carbon sequestration optimization in cold-region paddy fields of Northeast China.