Objective  This study aims to investigate the spatial variation pattern and influencing mechanism of soil organic carbon (SOC) under complex surface conditions in the Qilu Lake watershed, and to provide a scientific basis for sustainable use and fine management of land resources in the local irrigation area.

  Method  Two hundred and sixteen soil samples were collected in the irrigation area of the Qilu Lake watershed to determine the SOC content, followed by the spatial autocorrelation analysis of its spatial pattern. Combined with a total of 15 potential natural and anthropogenic influencing factors, multiscale geographically weighted regression (MGWR) model was employed to explore the influencing factors and mechanisms of SOC spatial variation in the irrigation area of the Qilu Lake watershed.

  Result  The global and local Moran index results showed that SOC had a significant spatial autocorrelation and revealed a clustered distribution pattern. The samples with higher SOC content were concentrated in the lakeside farming area in the southwestern area of Qilu Lake, while those with lower SOC content were concentrated in the foothill area in the southwestern area of Qilu Lake. Strong spatial heterogeneity in SOC content was discovered in the farmland of southern Qilu Lake. The results of the multiscale geographically weighted regression model showed that soil moisture content, total soil porosity, vegetation coverage and spread index had significant positive effects on SOC (P < 0.05) at the local level. The distance from road, the proportion of maximum patch area and patch abundance density had significant negative effects on SOC (P < 0.05).

  Conclusion  The spatial differentiation of SOC in the irrigated area of the Qilu Lake watershed was influenced by a combination of soil environment, natural factors and human activities, and this influence had spatial heterogeneity.