Estimating rainfall interception of xerophytic deciduous shrubs by static-and variable-parameter Gash models with stem- and leaf-dominated canopy water storage

The precise modeling of rainfall interception is important to understand the water balance of forest ecosystems, particularly in water-stressed restoration regions. However, xerophytic deciduous shrubs are currently underrepresented in rainfall interception models due to the significant variations in their canopy structures. Therefore, we estimated the rainfall interceptions of four representative deciduous shrubs (Spiraea pubescens, Potentilla fruticose, Hippophae rhamnoides, and Caragana korshinskii) in the semiarid region of northern China using the revised Gash model (RG model) with static parameters and a modified Gash model (MG model) with variable parameters. The results indicate that the MG model outperformed the RG model in studies of S. pubescens and P. fruticose, with regard to the modeling error (6.6%-7.9%), Nash-Sutcliffe model efficiency (0.68-0.86), and the root-mean-square error (0.17-0.71). However, the RG model was better than the MG model for H. rhamnoides and C. korshinskii, with regard to the modeling error (3.6%-6.3%), Nash-Sutcliffe model efficiency (0.82-0.94), and the root-mean-square error (0.17-0.66). The MG model was highly sensitive to changes in the leaf area index, canopy cover, mean evaporation rate, canopy storage capacity, and leaf storage capacity for S. pubescens and P. fruticose. Nevertheless, the RG model demonstrated high sensitivity to changes in the canopy storage capacity and stem storage capacity for H. rhamnoides and C. korshinskii. These indicated that canopy storage capacity has a significant effect on the rainfall interception capabilities of different shrub species, as indicated by both the RG and MG models. This critical indicator was determined by the water storage of stems for H. rhamnoides and C. korshinskii (ranged contribution of 59.5%-81.6%), whereas the water storage of leaves for S. pubescens and P. fruticose (with contributions of 51.3%-80.7%). Therefore, the optimized RG and MG models are recommended for estimating the rainfall interception of xerophytic shrubs with stem-dominated water storage and leaf-dominated water storage, respectively. This study categorizes the applicability of rainfall interception models in terms of important plant traits to precisely assess the water balance of xerophytic deciduous shrub ecosystems. In future studies, leaf configuration and hydrophobicity are highly recommended to be quantified for further improving rainfall interception models.