Numerous shelter forests have been established to combat desertification in the Mu Us Sandy Land, China. Shelter forests modify the characteristics of the underlying surface and affect the regional water cycle by altering rainfall partitioning. Understanding the rainfall partitioning process and its controlling factors for indigenous and exotic species is crucial for vegetation restoration and sustainable soil water management. This study developed an event-based rainfall partitioning process for three typical shelter forests. Indigenous vegetation, Amygdalus pedunculata Pall. (A. pedunculata), and two exotic species, Amorpha fruticosa L. (A. fruticose) and Pinus sylvestris var. mongholica Litv. (P. sylvestris), were observed during the rainy seasons (July and August) of 2021 and 2022. The results showed that throughfall, stemflow, and interception loss constituted 71.01 %, 8.23 %, and 20.76 % of rainfall, respectively, for A. pedunculata. The corresponding values were 74.65 %, 8.47 %, and 16.88 % for A. fruticose and 73.27 %, 1.44 %, and 25.29 % for P. sylvestris. Compared with the introduced P. sylvestris, the shrub canopy showed a greater funneling ratio and was conducive to recharging soil water by precipitation. The amount and intensity of rainfall were significantly correlated with the rainfall partitioning characteristics, whereas the correlation between rainfall duration and partitioning was insignificant. Based on the results of the revised Gash model, the stemflow was primarily influenced by the percentage of rainfall diverted to the stemflow. The interception loss for P. sylvestris was primarily influenced by the canopy storage capacity. However, the canopy storage capacity and the ratio of mean evaporation rate to mean rainfall intensity had significant effects on the interception loss in A. pedunculata and A. fruticose. It is necessary to comprehensively consider the vegetation type (tree/shrub and indigenous/exotic species) and the corresponding rainfall partitioning characteristics of shelter forests for the scientific construction and management of shelter forests in the Mu Us Sandy Land.

Rainfall partitioning characteristics and simulation of typical shelter forest in Chinese Mu Us Sandy Land