Models to predict changes in annual runoff with thinning and clearcutting of Japanese cedar and cypress plantations in Japan

Abstract

Forest management (thinning and clearcutting) can reduce evapotranspiration and increase catchment runoff. By summarizing data on the increase in annual runoff (ΔQ) due to forest management at various sites and analysing data using linear regression, traditional studies have reported large unexplained variability among data for different sites. To improve the predictability of ΔQ, it might be useful to model ΔQ for specific species and regions while considering underlying processes. This study performed such modelling for Japanese cedar and cypress plantations in Japan. Model 1 predicts ΔQ assuming that ΔQ equals the decrease in canopy interception loss (ΔEi), which was further modelled by stem density using 46 data for interception loss. Model 2 predicts the potential maximum of ΔQ (ΔQmax) assuming that ΔQmax equals the sum of ΔEi and the decrease in canopy transpiration (ΔEt). Here, ΔEt was calculated using a model developed in our previous study. ΔQ predicted using Model 1 approximated ΔQ observed for seven catchments, and the errors in prediction were less than those derived from traditional linear-regression analysis. ΔQmax predicted using Model 2 was greater than the observed ΔQ for all catchments. Thus, Models 1 and 2 would be respectively useful in assessing the effectiveness and limitations of managing Japanese cedar and cypress plantations to secure water resources, which have been controversial in Japan. Furthermore, the concept of the models gives implications for studies on other species and regions, because the models have demonstrated how to improve predictability of ΔQ considering underlying processes with the input of commonly available data.