吉林蛟河阔叶红松林样地种-面积关系

Species-area relationships within sample plot in a broad-leaved Korean pine forest at Jiaohe,Jilin Province

种-面积关系是群落生态学的核心问题之一,是生物多样性尺度转换的重要依据。利用吉林蛟河阔叶红松林30 hm^2的样地数据,采用随机取样与巢式取样方法,分别在10、20、30 hm^2尺度上建立对数模型(Logarithmic function)、幂函数模型(Power function)和逻辑斯蒂模型(Logistic function)拟合局域种-面积关系,并利用赤池信息准则(AIC)进行拟合结果优度检验。结果表明,取样方法对种-面积关系的构建有显著影响,随机取样优于巢式取样。种-面积关系的构建与尺度(取样上限)密切相关:在小尺度上(10 hm^2),对数模型与逻辑斯蒂模型拟合效果优于幂函数模型;在中尺度和大尺度上(20、30 hm^2),相对于对数模型和幂函数模型,逻辑斯蒂模型能更好地拟合阔叶红松林的种-面积关系。据AIC值可知,随机取样下的逻辑斯蒂模型拟合效果最好,是拟合30 hm^2阔叶红松林样地种-面积关系的最适模型。因此研究时需要根据区域森林群落的实际情况选择种-面积模型。

The species-area relationship （SAR） is a core component of community ecology, and is an important basis for biological diversity scaling. The SAR is used to describe community types and can solve many ecological problems, such as the determination of minimum sampling areas in a community. Therefore, it is of great importance to diversity conservation. Recently, a number of studies have demonstrated substantial uncertainties in selecting the best SAR model for a data set. In the present study, a 30-hm2 permanent forest plot was established in a broad-leaved Korean pine forest in Jiaohe, Jilin Province, China. All trees with diameters at breast height （DBH） i〉 1 cm were tagged and the height, DBH, and crown diameter of these trees were measured and recorded. We established a logarithmic model, a power function model, and a logistic model using the 30-hm2 sample plot to simulate the SAR of a broad-leaved Korean pine forest. We examined how SARs simulated by logarithmic, power function, and logistic models differed after random sampling or nested sampling methods had been used to collect data, and how this difference was affected by sampling scales （broad, moderate, and fine scales）. The Akaike Information Criterion （AIC） value was used to compare the goodness-of-fit for each SAR model. The results showed that the sampling method had a significant influence on the SAR, and that the goodness-of-fit for random sampling was better than that for nested sampling at all sampling scales. The establishment of a species-area relationship was closely related to the sampling scales, and the logarithm and logistic models were superior to the power function model at the fine scale （ 10 hm2 ）. At the moderate and large scales （20 hm2 and 30 hm2, respectively）, the logistic model better fitted the species-area relationship for broad-leaved Korean pine forest than did the logarithm and the power function models..A comparison of the different models showed that the logistic model with random sampling produced an optimal fit for the species-area relationship within the 30 hm2 broad-leaved Korean pine sample area （AIC = 76.91 ）, and that the appropriate minimum sampling area was 10 hm2. We concluded that both sampling scale and sampling method had significant influences on the SAR. The scale effect on the SAR is closely related to the community species distribution pattern, and the impacts may result from habitat heterogeneity and successional stage. Habitat heterogeneity and community succession stage might have influenced the number of regional species and species composition, and these different species distribution patterns were reflected in the different SAR curves. Therefore, in practical applications, the variation in the actual community structure and environments within the sampling area should be fully considered. Further work needs to consider the actual situation of the local forest community to simulate the species-area relationship models