Aboveground biomass determines canopy rainfall interception loss in Semiarid Grassland Communities

Canopy rainfall interception is one key hydrological process, affecting rainwater redistribution and effectiveness in semiarid regions. Canopy rainfall interception loss is jointly influenced by meteorology, vegetation and topography. The canopy water storage capacity (S), rainfall interception depth (I_m) and ratio (I_%) and vegetation characteristics, together with topographic factors of three grassland communities (dominated by Bothriochloa ischaemum, Lespedeza davurica and Artemisia gmelinii, respectively) were investigated on the Loess Plateau of China during the main growing season (June to September). Results showed that I_m ranged from 0.55 to 0.89 mm and I_% ranged from 6.14% to 12.1%, with the maximum values occurring in August for three communities, and A. gmelinii community had the largest I_m (0.89 mm) and I_% (12.1%). The I_m and I_% were positively correlated with aboveground biomass (AGB), coverage (Cov), leaf area index (LAI), community-weighted mean height (CWMH) and altitude (Alt), but negatively correlated with slope degree and rainfall intensity (RI). Hierarchical partitioning analysis (HPA) showed that AGB had the highest contribution for I_m (20.3%), while Alt had the highest contribution for I_% (18.2%). The regression models based on forward selection could effectively predict the values of I_m (R² = 0.802, RMSE = 0.049) and I_% (R² = 0.546, RMSE = 1.434). Topographic factors (altitude, slope degree and aspect) indirectly influenced both I_m and I_% by modulating vegetation characteristics (AGB, Cov, etc.). All these indicated that aboveground biomass mainly determines grassland community rainfall interception loss in the semiarid Loess Plateau.