I developed a weeding-duration model for Sakhalin fir （Abies sachalinensis （Fr. Schmidt） Masters） plantations that employs a generalized linear model. The number of years following planting that weeding is necessa...I developed a weeding-duration model for Sakhalin fir （Abies sachalinensis （Fr. Schmidt） Masters） plantations that employs a generalized linear model. The number of years following planting that weeding is necessary is the response variable, and elevation, slope steepness, maximum snow depth, annual precipitation, geology, soil, site index, slope aspect, and vegetation type are explanatory variables. Among the explanatory variables, geology, soil, slope aspect, and vegetation type are categorical data. A Poisson distribution is assumed for the response variable, with a log-link function. Elevation, slope steepness, maximum snow depth, annual precipitation, site index, and vegetation type had a significant effect on weeding duration. Among the eight models with the smallest Akaike information criterion （AIC）, I chose the model with no multicollinearity among the explanatory variables. The weeding-duration model includes site index, maximum snow depth, slope steepness （angle） and vegetation type as explanatory variables; elevation and annual precipitation were not included in the selected model because of multicollinearity with maximum snow depth. This model is useful for cost-benefit analyses of afforestation or reforestation with Abies sachalinensis.展开更多
文摘I developed a weeding-duration model for Sakhalin fir （Abies sachalinensis （Fr. Schmidt） Masters） plantations that employs a generalized linear model. The number of years following planting that weeding is necessary is the response variable, and elevation, slope steepness, maximum snow depth, annual precipitation, geology, soil, site index, slope aspect, and vegetation type are explanatory variables. Among the explanatory variables, geology, soil, slope aspect, and vegetation type are categorical data. A Poisson distribution is assumed for the response variable, with a log-link function. Elevation, slope steepness, maximum snow depth, annual precipitation, site index, and vegetation type had a significant effect on weeding duration. Among the eight models with the smallest Akaike information criterion （AIC）, I chose the model with no multicollinearity among the explanatory variables. The weeding-duration model includes site index, maximum snow depth, slope steepness （angle） and vegetation type as explanatory variables; elevation and annual precipitation were not included in the selected model because of multicollinearity with maximum snow depth. This model is useful for cost-benefit analyses of afforestation or reforestation with Abies sachalinensis.
