Stemflow estimation models for Japanese cedar and cypress plantations using common forest inventory data

Although stemflow (SF) had been regarded as a small portion of the gross rainfall (GR), recent studies have revealed that, depending on the forest stand structure, the SF/GR ratio should not be neglected. This study derived SF/GR estimation models using common forest inventory data. A set of SF/GR ratio and forest inventory data (stand density (SD), total basal area (BA), mean diameter at breast height ( D B H ¯ ), mean tree height ( H ¯ ), canopy cover (CC), and leaf area index (LAI)) was collected from previous studies of Japanese cedar and cypress plantations. To further investigate the relation between SF/GR ratio and forest stand structures, we examined additional stand-structure variables (mean basal area ( B A ¯ ), mean stem surface area ( S A ¯ ), and total stem surface area (SA)) derived from the inventory data, and the stand-scale funneling ratio (FRstand) evaluating the efficiency of funneling rainwater. Among all the stand-structure variables, SD exclusively determined the SF/GR ratio, providing the best-fitting positive single linear regression equation as a density-based SF/GR model with a root mean square error (RMSE) of 2.4%. Although this model is useful for practical forest water management because it requires only SD which is the most basic forest inventory data, it has a weak point in meticulous forest water management because it cannot reflect the effect of tree growth on SF/GR ratio. Thus, we developed a size-based SF/GR model (RMSE = 2.0%) based on the strong relationship between the FRstand and D B H ¯ . This model is applicable to meticulous forest water management because it reflects the effects of not only SD but also tree growth by DBH on SF/GR ratio. These models derived from the common forest inventory data are potentially applicable to the evaluation and control of SF in forest water management.