Summary of model fit measures.

Top row shows the full model with seven predictors, four climatic factors (JJA, DJF, CWD, PPT) and three fixed factors that did not vary under future climate scenarios (Soil depth, Wind speed, March radiation), with 36 parameters (linear, quadratic, 2-way interaction) for each of 21 (n-1) vegetation types. BIC value for the full model is 577096. Next seven rows show models with one factor at a time removed, sorted by decreasing ∆BIC. Last two rows show models with only the climate factors, or only the fixed factors. Average max probability is the mean value of the highest probability vegetation type in each pixel. Proportion correct is the proportion of pixels in which the maximum probability type was equal to the observed type.Summary of model fit measures.

首行展示包含7个预测变量的全模型，该模型涵盖4个气候因子（JJA、DJF、CWD、PPT）与3个在未来气候情景下保持恒定的固定因子（土壤深度、风速、三月辐射量）；针对21种（n-1）植被类型，每种类型均设置36个参数，包括线性项、二次项与双向交互项。全模型的贝叶斯信息准则（Bayesian Information Criterion，BIC）值为577096。后续7行依次展示每次移除单个因子后的模型，按ΔBIC降序排列。最后两行分别展示仅包含气候因子的模型，以及仅包含固定因子的模型。平均最大概率指各像元中概率最高的植被类型的均值。分类正确率指最大概率植被类型与观测类型一致的像元占比。模型拟合度量指标汇总。