林地是重要的林业生产资料,科学合理地划分林地等级,能够真实、全面地反映林地的质量差异,为政府宏观调控林地市场和加强林地管理提供决策依据。以标准林分蓄积量为森林潜在生长力评价要素,划分林地质量等级,构建与林地质量分等指标有关的森林潜在生产力评价模型,采用LM—UGO法估算参数,选用相关系数、均方差、残差平方和、卡方检验评价模型拟合优度,并以武夷山市和武夷山国家公园森林资源为案例开展林地质量分等应用。结果表明:1)建立的森林潜在生产力评价模型,相关系数0.966、均方差1.726、残差平方和286.095、卡方检验33.898,确定的年均气温、年均降水量、腐殖质层厚度、土层厚度、坡度、坡位等6个林地分等指标因子的权重分别为0.08,0.22,0.27,0.26,0.10,0.07。2)依据连年生长量和平均生长量,确定林地质量分等的基准年龄为20 a。3)武夷山市和武夷山国家公园林地质量等级以3,4等级占比较大,1等级最少,林地潜在生长水平处于中等水平;各等级面积占比与海拔(地貌)关系显著,呈幂函数或多项式关系。构建与林地质量分等指标有关的森林潜在生产力评价模型,可以避免层次分析法、灰色关联分析法等经典决策分析方法确定指标因子权重的主观性问题。

Forestland is an essential resource for forestry production.Scientific and reasonable classification of forestland grades can accurately and comprehensively reflect differences in forestland quality,providing a decision-making basis for government macro-control of the forestland market and strengthening forestland management.Taking standard stand volume as an evaluation element of forest potential growth power,we divided forestland quality grades and developed a forest potential productivity evaluation model related to forestland quality grading indicators.We estimated parameters using the LM-UGO method,evaluated model fitting goodness with fitting correlation coefficient,mean square error,sum of squares of residuals,and Chi-square test,and applied forestland quality grading to the forest resources in Wuyishan City and Wuyishan National Park as case studies.The results showed:1) The established forest potential productivity evaluation model had a correlation coefficient of 0.966,a mean square error of 1.726,a sum of squares of residuals of 286.095,and a Chi-square test value of 33.898.The weights of six forestland grading index factors including average annual temperature,average annual rainfall,thickness of humus layer,thickness of soil layer,slope,and slope position were 0.08,0.22,0.27,0.26,0.10,and 0.07 respectively.2) Based on the annual growth amount and average growth rate,the benchmark age for forestland quality grading was determined to be 20 years.3) In Wuyishan City and Wuyishan National Park,the proportions of grades 3 and 4 in forestland quality were relatively large,the proportion of grade 5 was 9.43%,grade 1 was the smallest,and the potential growth level of forestland was at a medium level.The area proportion of each grade was significantly related to altitude(topography),exhibiting a power function or polynomial relationship.Developing a forest potential productivity evaluation model related to forestland quality grading indicators can avoid the subjectivity problem of determining index factor weights by classical decision analysis methods such as the analytic hierarchy process and grey relational analysis method.