Model selection to assess the effects of age, mate replacements and diet on the productivity of breeding successful pairs.

Model definition enumerates the fixed effects considered in the GLMMs. All models included territory and year nested by population as random effects.*Best model AICc = 174.899.Note: OC = rabbit consumption; AR = partridge consumption; H′ = diet diversity; PSi = prey consumption specificity. Parameters’ interactions are denoted by (*), while (2) indicates a quadratic effect. ΔAICc refers to the difference in the corrected Akaike Information Criteria (AICc) between model i and the model with the lowest (AICc) (i.e. the best model). Models with ΔAICc cw) explains the probability that a given candidate model is the best of the proposed set, so the sum of all the models is 1.0. R2GLMM(m) estimates model fit using fixed effects only, while R2GLMM(c) estimates model fit including both fixed and random effects.本模型定义列出了广义线性混合模型（Generalized Linear Mixed Models，GLMMs）中所考量的固定效应。所有模型均将以种群嵌套的领地和年份作为随机效应。*最优模型的校正赤池信息准则（corrected Akaike Information Criterion，AICc）值为174.899。注：OC代表兔类取食量；AR代表鹧鸪取食量；H′代表饮食多样性；PSi代表猎物取食专一性。参数间的交互作用以(*)标注，而(2)表示二次效应。ΔAICc指的是第i个模型与AICc值最低的模型（即最优模型）之间的校正赤池信息准则差值。赤池权重（Akaike weights，cw）可用于衡量某一候选模型为所提议候选集合中最优模型的概率，因此所有候选模型的权重之和为1.0。R2GLMM(m)仅通过固定效应评估模型拟合度，而R2GLMM(c)则同时纳入固定效应与随机效应以评估模型拟合度。