该文根据“空间代替时间”的原则和对马尾松(Pinus massoniana)种群的重要值-平均胸径的变动趋势的分析，确定5个森林群落为广东省黑石顶自然保护区森林群落演替的一个时间序列；并在 此基础上，运用分形理论的计盒维数和信息维数，对演替过程中马尾松种群空间格局的动态进行了分形分析。结果表明，马尾松种群的计盒维数和信息维数2个参数值均呈递减趋势，与其个体数、重要值变动趋势一致。在群落的演替过程中，马尾松种群的空间占据能力不断下降，种群呈衰退趋势。随其在群落中的优势地位逐渐被其它种群取代，群落将由以马尾松占绝对优势的单优群落演替为多优常绿阔叶林群落。分析结果同时表明，分形分析是群落演替过程中种群空间格局动态研究的有效方法，分形维数则能反映种群格局的尺度变化规律。
Background and Aims The dynamics and fractal characteristics of population patterns during vegetation succession are rarely explored. Scale, pattern and process of ecological succession are three intertwined concepts in modern ecology. Succession research will inevitably involve scale and pattern analyses. Fractal theory can be employed as an effective tool for synergic analysis of scale, pattern and successional processes. There are few reports on pattern dynamics of Pinus massoniana, a species that declines during succession in southern China. Five forest communities were selected in Heishiding Nature Reserve, Guangdong Province for a case study. Our objectives were to determine 1) whether results obtained by employing fractal analysis are consistent with those from traditional research where space is substituted for time and 2) the pattern dynamics of P. massonianaand the ecological meaning of the change of fractal characteristics in succession.
Methods For the traditional space-for-time succession approach, we analyzed importance value (IV) and mean diameter at breast height (DBH) ofP. massoniana, and the selected forest communities were treated as a time series of five successional stages. The box-counting dimension (Dbox) and information dimension (Dinfo) from fractal theory were employed to analyze the pattern dynamics of P. massoniana.
Key Results Dbox and Dinfo ofP. massonianadecreased during succession, in accordance with number of individuals and IV, and the traditional method of substituting space for time was applicable in this study. This demonstrated deteriorating regeneration conditions and decreasing ability of spatial occupation of P. massoniana with succession—that P. massonianahad a declining population. With its loss of dominance, the community would turn from P. massoniana dominance into a multi-population dominated evergreen broadleaved forest community. The results of fractal analysis were not only consistent with those of the trend analysis of IV and mean DBH, but also yielded more ecological information a bout scale-related spatial pattern of the population, its degree of spatial occupation and its role in community and regeneration conditions.
Conclusions This study indicated that fractal analysis is an effective approach to study dynamic pattern analysis during forest succession, and the results were more reliable than those obtained by employing only IV and meanDBH in a traditional space-for-time approach.