Relative importance of climate vs local factors in shaping the regional patterns of forest plant richness across northeast China

Northeast (NE) China covers three climatic zones and contains all the major forest types of NE Asia. We sampled 108 forest plots in six nature reserves across NE China to examine the influence of climate and local factors (canopy seasonality, successional stage, topography and forest structure) on geographic patterns of plant richness. We analyzed the relative effects of different factors at two spatial scales: the regional scale (across both latitude and altitude) and the local scale (along the altitudinal gradient within site). Our results showed that the relative importance of climate vs local factors differed remarkably depending on scale and functional group. While total and tree species richness were mainly limited by climate, herb and shrub richness was more related to local factors (especially at the local scale). In the climatic factors, heat sum was the major correlate of tree, shrub and total species richness, while herb richness was more associated with winter coldness. Precipitation was not a limiting factor for forest plant richness in NE China. Climate accounted for 34! 76% of variation in richness at the regional scale, but explained only 0!44% at the local scale. Among the local factors, shrub species richness was sensitive to seasonal canopy openness, with higher richness in deciduous forests than in the evergreen needle-leaf forest. On the other hand, herb richness was sensitive to forest successional stage, with higher richness in middlesuccessional forests than in the early and late-sucessional forests. Local topography (aspect and position on slope) and forest structure (tree density) also showed remarkable influence on species richness. Our results suggest the importance of including local factors when examining large scale diversity gradient (especially for understory species), and the necessity of comparing diversity patterns among functional groups at different spatial scales.