基于生态过程模型和森林清查数据的森林生长量估算对比研究

利用遥感驱动的生态过程模型-Boreal Ecosystem Productivity Simulator （BEPS）、2001-2006年国家森林资源连续清查数据（一类清查-样地尺度）和2003-2009年森林资源规划设计调查数据（二类调查-区域尺度），分别计算江西省吉安市的森林生态系统生长量，从不同空间尺度和森林类型对3种数据源估算的森林生长量进行了分析。结果表明，样点尺度上，BEPS模型模拟的森林生长量（4.18 Mg·hm^-2·a^-1）低于群落生长量（5.86 Mg·hm^-2·a^-1），与乔木层生长量（4.29 Mg·hm^-2·a^-1）基本一致，模型模拟结果与两者的拟合R2分别为0.48和0.43。区域尺度上，BEPS模型模拟、二类调查数据计算的群落及乔木层生长量分别为4.65、4.36和3.34 Mg·hm^-2·a^-1，BEPS模型估算的吉安市各县森林总生长量与二类调查数据计算的群落、乔木层生长总量拟合R2分别达0.84和0.83。一类清查数据计算结果高于二类清查数据计算结果，BEPS模型模拟森林生长量分别与基于一类清查数据计算的乔木层生长量及二类调查数据群落生长量较为一致。从研究区两种主要森林类型来看，常绿阔叶林年平均生长量高于常绿针叶林，常绿针叶林与模型估算结果差异小于常绿阔叶林。最后利用模型估算了研究区2001-2010年平均生长量，为认识研究区的森林生长空间分布差异及更新森林生物量提供支持。

Comparison of estimated forest biomass increment rate based on a process-based ecological model and forest inventory data

Based on a remote sensing driven, processed-based ecological model （Boreal Ecosystem Productivity Simulator, BEPS）, National Forest Continues Inventory （NFCI） data （surveyed in 2001 and 2006） and Forest Management Inventory （FMI） data （surveyed in 2003 and 2009）, the forest annual biomass increment rate in Ji’an city was calculated, respectively. The differences among the forest biomass increment rates estimated using these three different methods were analyzed for various spatial scales and forest types. Our results showed that at the plot scale the forest biomass increment rate （4.18 Mg·hm^-2·a^-1） simulated by the BEPS model was close to the overstory biomass increment rate （OBIR, 4.29 Mg·hm^-2·a^-1）, and lower than the biome biomass increment rate （BBIR, 5.86 Mg·hm^-2·a^-1） estimated using the NFCI data. The R2 of OBIR and BBIR simulated by the BEPS model against estimates with the NFCI data were 0.43 and 0.48, respectively. The regional mean forest biomass increment rate simulated by the BEPS model was 4.65 Mg·hm^-2·a^-1 while the OBIR and BBIR estimated using the FMI data were 4.36 and 3.34 Mg·hm^-2·a^-1, respectively. The R2 values of county-level forest biomass increment rate simulated by the BEPS model against OBIR and BBIR were 0.84 （p〈0.01） and 0.83 （p〈0.01）, respectively. The OBIR and BBIR estimated using the NFCI data were higher than those estimated using the FMI data. The forest biomass increment rate simulated by BEPS modeled was close to the OBIR value based on NFCI and the CBIR value based on FMI. As to two dominant forest types, evergreen broadleaf forests grow faster than evergreen coniferous forests. The agreement between simulated and estimated forest biomass increment rate is better for evergreen coniferous forests than for evergreen broadleaf forests. This study confirms that the output from BEPS can be used for updating forest biomass.