应用R-分布模拟杉木人工林林分断面积累积分布的研究

Application of R-distribution to model cumulative distribution of basal area of stands in Chinese fir plantations.

以杉木人工林未间伐林分断面积分布为研究对象，在分析3参数Weibull方程模拟分布数据的优缺点的同时，提出了一种由Richards方程所导出的新的分布函数——R-分布，并应用R-分布对不同年龄、不同密度及不同立地指数的杉木人工林林分断面积分布展开模拟．结果表明：①R-分布对林分断面积分布能实现高精度的模拟，且其参数具有深刻的理论涵义；参数P与林分年龄、平方平均直径呈紧密的正相关性，为分布的相对尺度参数；参数q与林分平方平均直径紧密相关，为分布的位置参数，指代林分分布数据的平均大小；参数r决定分布曲线的拐点，为分布的形状参数．②林分断面积累积分布曲线的拐点纵坐标主要存在于0．4～0．6之间，且多数在0．5以上；拐点横坐标及纵坐标受林分年龄、平方平均直径、林分密度及立地指数等林分因子的影响，但不同林分因子的影响程度不一样．③兼用PPM与PRM可能对R-分布的3个参数做出良好的预估．

Stand basal area distributions were obtained in order to investigate the R-distribution, a new function derived from the Richards function. After analyzing advantages and shortcomings of 3-parameter Weibull function, we applied the R-distribution to model basal area distributions in un-thinned stands of Chinese fir plantations at different ages, densities and site indices. We found that the R-distribution had high precision in modelling basal area distributions and its parameters appeared theoretically meaningful. The relative scale parameter p was highly correlated with stand age and quadratic mean diameter at breast height （DBH）; the location parameter q was highly correlated with quadratic mean DBH and the shape parameter r indicated the inflection point of the distribution. Ordinates of the inflection points of the cumulative distribution curve of stand basal areas mainly ranged between 0.4-0.6; most were found to be above 0.5. Both abscissa and ordinate of the inflection points were differently affected by age, quadratic mean DBH, density and site index. Based on parameter prediction method and parameter recovery method, the three parameters of the R distribution could be predicted well.