A climate-sensitive mixed-effects tree recruitment model for oaks (Quercus spp.) in Hunan Province, south-central China

Abstract

Tree recruitment is an essential component and plays a fundamental role in forest management decision-making. However, the existing tree recruitment models for oak (Quercus spp.) species have been developed using traditional modeling functions, and have neglected the effects of climate, which do not apply to long-term projections of future forest composition under a climatic change scenario. In this study, we developed tree recruitment models based on the data of 548 permanent sample plots distributed in Hunan Province, south-central China. Five commonly used recruitment approaches (negative binomial model, zero-inflated Poisson, zero-inflated negative binomial model, Hurdle-Poisson model, and Hurdle-negative binomial model) were employed to analyze the data. Then the random effects of the regions were included to develop the nonlinear mixed-effects models and account for spatial and temporal correlations. Among the various candidate predictors tested including stand characteristics, site conditions, and climate, stand basal area (BA), quadratic mean diameter (QMD), basal area of the species (Bai), cosine of the slope combined with the natural logarithm of elevation (CIE), and mean annual temperature (MAT) had significant effects on oak recruitment. The zero-inflated negative binomial model was selected as the optimal model compared to the other four basic models. The performance of the models improved significantly when the random effects were added. These results and knowledge of adaptive management will be useful for developing practical forest management plans.