关键词: 中性理论模型, 生态位模型, 取样尺度, 统计模型, 温带森林

Abstract:

Aims Our objective was to explain processes that dominate species-abundance distribution pattern and mechanism of community assembly in temperate forests.
Methods We used three 5.2-hm ² permanent plots established in secondary Populus davidiana-Betula platyphylla forest, secondary conifer and broad-leaved mixed forest and Tilia amurensis-Pinus koraiensis mixed forest in Changbai Mountains. Within each plot, we randomly selected 500 subplots within 260 m × 200 m at the scales of 10 m × 10 m, 30 m × 30 m, 60 m × 60 m and 90 m × 90 m. We calculated the mean value of species-abundance distributions taken from the 500 subplots as the observed species-abundance distribution. We estimated the fitted species-abundance distributions by neutral, log-normal, Zipf, broken stick and niche preemption models at different scales. Simulation effects were tested by Chi-square test, Kolmogorov-Smirnov (K-S) test and Akaike Information Criterion (AIC). For the neutral model, we first estimated two parameters θ and m and then simulated 600 species-abundance distributions. The average of these 600 species-abundance distributions was the best-fit result of the neutral model. We employed the 95% confidence envelopes that were approximated by the 2.5 and 97.5 percentiles of the abundances of species of rank i = 1 to S over the 600 simulations to test goodness-of-fit for the neutral model. All of the computations were conducted in R 2.14.1 with UNTB and VAGEN packages.
Important findings The neutral model fit species-abundance distribution at different successional stages of conifer and broad-leaved mixed temperate forest communities. All five models fit the observed value at the 10 m × 10 m sampling scale, and the goodness of fit of the log-normal, Zipf, broken stick and niche preemption models were better than that of the neutral model. That means at small sampling scale the species-abundance distribution is dominated by neutral process and niche process; however, the niche process is much more important. At other sampling scales (30 m × 30 m, 60 m × 60 m, 90 m × 90 m), the neutral model was the best-fit model. As the sampling scales increased, other models gradually dropped out. At the sampling scale of 90 m × 90 m, none of the models fit well, except for the neutral model. This suggests for north temperate forests in the Changbai Mountains that the random process represented by neutral model is the main ecological process that determines the species-abundance distribution pattern at middle and large sampling scales and that the species-abundance distributions at different sampling scales are likely dominated by different ecological processes.

Key words: neutral theory model, niche model, sample scale, statistical model, temperate forest