作　　者： ; ; ; ;

机构地区： 中山大学环境科学与工程学院

出　　处： 《环境科学学报》 2008年第4期778-785,共8页

摘　　要： 利用1990年、1997年和2004年3个时相的Landsat TM数据,基于CART分析算法,提取广州森林分类信息.利用样方调查与同期TM数据的光谱响应数据,分别建立了阔叶林、针叶林、园地的逐步回归模型,计算了广州森林碳储量,绘制了碳密度分布图.结果表明,广州森林覆盖率1997年比1990年显著增加,而2004年比1997年略有减少.3个时期广州森林碳储量、碳密度分别为5.93×106t、7.07×106t、7.64×106t与22.1t·hm^-2、22.7t·hm^-2、24.7t·hm^-2,均呈上升趋势.3个时期均有40%以上的森林碳密度低于25t·hm^-2,80%以上森林碳密度低于50t·hm^-2；但总体趋势是低碳密度森林比重减少,高碳密度森林比重增加,森林碳密度结构朝良性发展.在广州碳储量估算的基础上,计算MoranI系数与GearyC系数,表明研究区森林碳密度3个时期都存在正的自相关性,且分布格局由随机离散分布趋向于聚集分布.进一步利用半变异函数,对碳密度分布格局异质性进行分析,表明3个时期森林碳密度的空间分布格局异质性差异不大,1997年略强于其它时期；空间结构上均属于高度相关；空间自相关因素是森林碳密度空间异质性的主要因素；从1990-2004年,相关尺度增加,引起空间异质性的随机因素所占比重减少,自相关因素所占比重增加. Guangzhou is a rapidly developing area in China. Classification information of local forests was obtained from TM image data from 1990, 1997 and 2004 and analyzed based on Classification and Regression Tree （CART） analysis. Using field measurements and same term remote sensing information,along with stepwise regression analysis, the forest carbon models of coniferous forest,broadleaf forest and orchard were designed, and the forest carbon storage in those years was estimated. The results showed that the forest coverage ratio in Guangzhou increased from 1990 to 1997, but decreased from 1997 to 2004. The carbon density was 22. 1 t·hm^-2 in 1990, 22.7 t·hm^-2 in 1997 and 24.7 t·hm^-2 in 2004, and carbon storage was 5.93 ×10^6t in 1990, 7.70×10^6t in 1997 and 7.64×10^6t in 2004. The carhon storage and carbon density of forest increased over that time. The carbon density of Guangzhou forest was still low, but it is a huge potential carbon sink for the future. The area of the forest whose carbon density was less than 25 t·hm^-2 was more than 40% in the study period. Forests whose carbon density was less than 50 t·hm^-2 were distributed in the main area of Guangzhou. Spatial patterns of the carbon density were obtained by geostatistical analysis. The results show that the heterogeneity in spatial distribution of the Guangzhou forest carbon density did not change significantly, but the distance （range） of spatial dependence varied significantly. The distance （range） of spatial dependence increased, increasing from 807 m in 1990 to 1983 m in 2004. Analysis of spatial variation revealed that the spatial heterogeneity was largely caused by spatial autocorrelation （75.7% -82.0% ）. In conclusion, the results demonstrated that spatial autocorrelation is the primary factor which resulted in the spatial heterogeneity of the forest biomass pattern in Guangzhou.

关 键 词： 森林 碳储量 碳密度 遥感 演变 异质性 广州

领　　域： [环境科学与工程]