AWRA-L model for simulating interception of rainfall loss in large scale

Abstract: The application of the Gash model to simulate canopy interception is mostly on the study plot or a specified forest, and the application in a landscape scale is only reported in the Australian Water Resource Assessment system (AWRA-L). This study introduced the AWRA-L model, and aimed to simulate the interception loss of the natural secondary forest of birch (Betula platyphylla Suk.) during 2013 in China. The GLASS LAI data and submerging method were used to get the canopy capacity for different periods. The leaf area index (LAI) value obtained before germination was assumed to be the area of branches and trunks. While the LAI value obtained after germination was the area of the leaves, branches and trunks. The differences between the LAI values after and before germination were the area of leaves when the growing of branches and trunks was ignored. The submerging method showed that the water capacity of leaves and stems per unit area were 0.17 and 0.33 mm, respectively. The capacity of branches and trunks was 0.23 mm. The capacity of leaves was 0.04-0.51 mm. The mean canopy capacity was 0.60±0.14 mm. The intercept method showed the canopy capacity and the trunk capacity were 0.62 and 0.04 mm, respectively. The whole capacity of the birch forest in its over ground parts was 0.66 mm. The canopy capacity by the submerging method and the intercept method was not significantly different (P=0.23). Considering of the high canopy cover fraction of the birch forest, both the simulating results of AWRA-L model based on Van model and the AWRA-L(1979) model based on Gash (1979) model were discussed in the paper. For the 25 rain events measured during the experiment, the simulating errors of the cumulative interception loss obtained from the AWRA-L model and the AWRA-L(1979) model were -9.2 and -1.7 mm, respectively. The relative errors were -14.8% and -2.8%, respectively. Both the AWRA-L model and the AWRA-L(1979) model underestimated the interception loss. When the rain event on June 8 with the biggest simulating error was eliminated, the cumulative simulating errors of the AWRA-L model and the AWRA-L(1979) model were -4.4 and 1.5 mm, respectively. The relative errors were -8.7% and 3.0%, respectively. The mean errors of the simulating interception loss for single rain event by these two models were -0.15 and -0.03 mm, respectively. The simulating value and measured was not significantly different. The results illustrated that the canopy capacity obtained by the data of GLASS LAI and submerging method could be used to determine the canopy capacity and the results could be used for simulating of interception loss by the AWRA-L model. The AWRA-L model is reliable to simulate both the cumulative interception loss for a long period and the interception loss for a single rain event in a landscape scale.