Temperature effects on cropland soil particulate and mineral-associated organic carbon are governed by agricultural land-use types.

[Display omitted] • POC accounts for about a third of the total cropland SOC. • Cropland type is the dominant role of soil POC and MAOC. • Temperature effects on SOC fractions are cropland type-dependent. • Temperature effects decline after removing correlations with net primary productivity. • Main controls of soil POC and MAOC vary with cropland types. Agricultural land use and temperature change can profoundly affect soil organic carbon (SOC), but their combined effects on regulating cropland SOC fractions over broad geographic scales remain poorly understood. Based on 871 topsoil (0–20 cm) samples collected from the Sichuan Basin of China, we used random forest model and partial correlations analysis to explore how the functionally distinct SOC fractions, including particulate and mineral-associated organic C (POC and MAOC), are affected by both agricultural land use and temperature over a large scale. The results show that POC accounts for 32.54 % of the total cropland SOC and soils with high SOC contents have larger proportions of POC. Cropland type is the dominant control of the two SOC fractions and regulates the impacts of other factors. The contents of soil POC and MAOC in dryland are much lower than that in paddy field and dryland-paddy rotation land. Significant correlations between mean annual temperature and the two SOC fractions are only found in dryland and dryland-paddy rotation land, while mean annual temperature has no impact on both POC and MAOC in paddy field. Meanwhile, the correlation coefficients between mean annual temperature and POC and MAOC for dryland decreased by 61.90 % and 69.23 %, respectively, and for dryland-paddy rotation land decreased by 28.00 % and 14.29 %, respectively, after removing correlations with net primary productivity. Additionally, correlation between mean annual temperature and POC is slightly higher than that between mean annual temperature and MAOC in dryland, while the reverse in dryland-paddy rotation land. These results suggest that the temperature effects on POC and MAOC are cropland type-dependent. To accurately predicting future cropland SOC change and reliably evaluate the potential contribution of cropland soil C sequestration, fraction-specific control mechanisms related to cropland type should be sufficiently considered in carbon models. [ABSTRACT FROM AUTHOR]