Comparative modeling of the effect of thinning on canopy interception loss in a semiarid black locust (Robinia pseudoacacia) plantation in Northwest China

• Published in: Journal of hydrology (Amsterdam) Vol. 590; p. 125234
• Main Authors: Ma, Changkun, Luo, Yi, Shao, Mingan
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.11.2020
• Subjects: Loess plateau; Rainfall interception; Revised Gash Analytical Model; Thinning; WiMo model; WiMo model; Revised Gash Analytical Model; Thinning; Loess plateau; Rainfall interception
• ISSN: 0022-1694; 1879-2707
• DOI: 10.1016/j.jhydrol.2020.125234

[Display omitted] •Thinning reduced interception loss >6% of gross rainfall annually.•The RGAM and WiMo models accurately predicate seasonal interception loss.•The RGAM model was sensitive to in both leafed and leafless seasons.•The optimized RGAM model performed better than the WiMo model. Canopy interception loss is a key component of forest hydrological cycle that determine the amount of net rainfall reaching forest floor together with drought/climate stressors affecting dryland forest plantations. A good understanding of the relationship between canopy interception loss and forest management such as thinning is important for improved watershed management and ecological services. In this study, we measured event-based rainfall partitioning for a thinned (TH, with 38% basal area removed) and a control (CT) Robinia pseudoacacia forest plantation during leafed and leafless seasons in 2015 in the semiarid Loess Plateau region in China. Interception loss from both forest plots were simulated using the Revised Gash Analytical Model (RGAM) and the WiMo model. The results showed that observed annual throughfall, stemflow and interception loss were respectively 80.8%, 1.7% and 17.5% of the gross rainfall under CT. The corresponding values under TH were 87.9%, 0.8% and 11.3%, respectively. The RGAM and WiMo models were well calibrated and validated using filed data collected in leafed and leafless seasons. The analyses suggested that models accurately predicated interception loss under both CT and TH conditions and captured seasonal variations in canopy and meteorological parameters. The RGAM model was most sensitive to the ratio of mean evaporation rate to mean rainfall intensity, and canopy storage capacity in leafed season, and to the ratio of mean evaporation rate to mean rainfall intensity in leafless season. Moreover, 37.2% and 42.1% of the interception in leafed season evaporated from the canopy respectively during rainfall event and after rainfall. The corresponding values for leafless season were respectively 49.3% and 22.4%. Overall, the performance of the optimized RGAM and WiMo models were satisfactory with respect to modeling error (−6.9 − −2.9%) and Nash-Sutcliffe model efficiency (0.80–0.94), although that of the optimized RGAM model was slightly superior to WiMo model. The models would facilitate the implementation of water-oriented management in semiarid forest plantations through a more accurate simulation of the impact of thinning on interception loss.