灵空山自然保护区油松—辽东栎林多度-面积关系分析

通过4 hm2样地调查的数据资料,采用随机分布多度模型和聚集分布多度模型,在对山西灵空山海拔1 500～1 800 m的油松—辽东栎林物种多度及其水平空间分布分析的基础上,运用估计优度评价了2个模型预测分布多度的适宜性。结果表明:①在不同像元的30种木本植物中有20种的多度依次增加且所占的水平空间也依次扩展,有10种不表现为上述关系。②随着像元面积的扩大,遇到多度序列中面积小于上一个种时,多度—面积曲线呈现较大波动;剔除波动节点的物种时,多度—面积曲线的波动趋于平缓。对于同一个物种来说,像元面积越大,其物种所占面积也越大。③估计优度评价结果显示聚集分布多度模型用于预测多度—面积关系优于随机分布多度模型。④无论是随机分布多度模型还是聚集分布多度模型均依赖于m的取值,即物种在固定像元下所占像元数。对于分散程度较高的物种,采用2种模型进行预测时结果较精确,反之预测结果误差越大;在样地总面积一定时,像元面积越小,预测结果越精确。

Relationship between the Abundance and Area of Pinus tabuliformis and Quercus liaotungensis in Lingkong Mountain Nature Reserve

According to the investigation data of a 4 hm2 sample plot,the spatial distribution and species abundance of Pinus tabuliformis and Quercus liaotungensis at the altitude of 1 500-1 800 m in Lingkong Mountain were analyzed with the random distribution abundance model and aggregated distribution abundance model,the suitability of these two models to forecast the distribution abundance was evaluated by estimate goodness.The results show that,of 30 kinds of woody plants with different pixel values,the abundance of 20 kinds increases and their horizontal space expanses,other 10 kinds do not show the above-mentioned changes.With the expansion of pixel area,the abundance-area curve experiences great volatility when the area of abundance sequence is less than the last species;With the species on the fluctuations node excluded,the fluctuation in the curve tends to flatten.For the same species,the greater the pixel area,the larger the species occupied area.The evaluation results of estimate goodness show that the aggregated distribution abundance model is better than the random distribution abundance model to predict the relationship between abundance and area.Both the random distribution abudance model and the aggregated distribution abundance model are dependent on the values of "m",that is,the occupied pixel number of species under the fixed pixel.For the species with a higher dispersion,the forecast results will be more accurate by using both two types of models,otherwise,the forest errors will be greater;When the total plot area is certain,the smaller the pixel area,the more accurate the forecast results.