Comparative modeling of the effects of intensive thinning on canopy interception loss in a Japanese cedar (Cryptomeria japonica D. Don) forest of western Japan

In the last several years, thinning has been conducted in coniferous plantations throughout Japan, mainly to secure water resources. Although canopy interception loss (I_c) is a major component of evapotranspiration, this is among the first studies that have examined changes in I_c while considering the effects of thinning in Japanese cedar (Cryptomeria japonica) forests, the most common type of plantation in Japan. We compared and contrasted three canopy interception models (Mulder, revised Gash, WiMo) before and after intensive thinning operations. The experimental Japanese cedar forest was intensively thinned with removal of 54% of all stems, reducing the basal area by 50% from 99.7 to 49.6m²ha^-1. Prior to thinning, all three canopy interception models performed well with root mean square errors for the Mulder, revised Gash, and WiMo models of 2.96, 3.59, and 3.81mm, respectively; while those after thinning were 1.32, 2.37, 2.38mm, respectively. In addition, all three canopy interception models performed better than the mean before thinning as evaluated by the Nash-Sutcliffe model efficiency with values 0.51 (Mulder), 0.28 (revised Gash), and 0.19 (WiMo) but did not perform as well after heavy thinning. These results suggest that canopy interception models might not be applicable after intensive thinning of Japanese cedar forests.