基于3S技术的天山历史云杉林空间分布的提取

3S-Based Extraction of Spatial Distribution of Picea schrenkiana var.tianschanica in History in the Tianshan Mountains

运用遥感手段结合历史时期遥感影像数据,以天山云杉(Picea schrenkiana var. tianschanica)林生境特征为固定因子,结合植被指数分析、地形因子分析、主成分分析及面向对象的决策树分类的方法提取历史时期天山云杉林的空间分布信息,从而为历史资料缺失情境下的天然林资源保护工程实施效益评价提供支持。研究表明:①将天山云杉林的年龄特征作为固定因子,以现状年高空间分辨率遥感影像及森林资源二类调查数据作为本底资料,在面向对象分类方法支持下可以很好的从历史时期的遥感影像中提取出天山云杉林的历史空间分布信息,提取精度可达93. 3%;②在植被指数因子中,NDVI对天山云杉林指示性最好,并确定用于天山云杉林提取的最佳NDVI值域为[0. 35,0. 8];③地形因子及主成分分析方法可以大大压缩影像的冗余信息,在提升云杉林信息提取的精度的同时提高运行速度。从整体来看,利用历史时期遥感影像并结合天山云杉林的生境特征,可以很好的提取出历史时期云杉林空间分布信息,从而为资料缺失情境下的森林资源管理措施制定及应对气候变化提供数据支持。

The spatial distribution information of Picea schrenkiana var. tianschanica in the Tianshan Mountains in historical period was extracted based on the vegetation index,topographic factor,principal component analysis,the decision tree classification method and the habitat characteristics of P. schrenkiana in the study area using the remote sensing methods combined with the historical remote sensing image data. So as to provide support for the benefit evaluation of the natural forest resources protection project under the situation of missing historical data. Results showed that: ① the historical spatial distribution information of P. schrenkiana in the Tianshan Mountains could be extracted from the remote sensing images,the forest stand age of P. schrenkiana was set as a fixed factor,and the present remote sensing images with high spatial resolution and forest management investigation data were used as the background information. The accuracy of information extraction of P. schrenkiana in the study area could be as high as 93. 3%,and the remote sensing images can be used to extract the spatial distribution information of P. schrenkiana in the Tianshan Mountains;② In the vegetation index factors,the response of P. schrenkiana in the Tianshan Mountains to NDVI was the most sensitive,and the best NDVI range for extracting the information of P. schrenkiana in the Tianshan Mountains was [0. 35,0. 8];③ Topographic factor and principal component analysis method could greatly compress the redundant information of image,which improved the accuracy of information extraction of P. schrenkiana forest and improved the running speed. On the whole,the spatial distribution information of P. schrenkiana forest in the historical period can be well extracted by using the historical remote sensing images and combining with the habitat characteristics of P. schrenkiana forest in Tianshan Mountains,so as to provide data support for the formulation of forest resource management measures and the response to climate change in the context of data missing.