---

作 者 ：李晓兵, 陈云浩, 余弘婧

期 刊 ：植物生态学报 2005年 29卷 2期 页码：185-196

关键词：NECT；土地覆盖；数据集；

Keywords:NECT, Land cover, Multi-temporal NOAA/AVHRR NDVI images,

摘 要 ：着重探讨了建立中国东北样带（Northeast China transect，NECT）土地覆盖特征数据集的原理、方法及其在全球变化研究方面的重要应用。NECT土地覆盖特征数据集是以多时相的1 km分辨率的NOAA/AVHRR归一化植被指数NDVI（Normalized difference vegetation index） 数字影像为基础，同时采用高程、气候、土壤、植被、土地利用、土地资源、生态区域、行政边界、经济、社会等多源数据作为数据源，并经过标准化处理（如数字化、空间插值、几何配准、投影转换）集成而成。在土地覆盖特征数据集的主要应用方面，如：1）利用多时相、1 km分辨率的NOAA/AVHRR影像完成了中国东北样带土地覆盖分类图。一级分类系统包括森林、草原、荒漠和沙地、灌丛、农田、混合覆盖类型、城镇和水体等8类，二级分类体系包括12类。经过地面采样进行精度检验，分类精度达到81.61％。2）对主要植被类型的植物生长季变化进行的研究。利用多时相的遥感影像构造了能够反映植被年际、季节生长变化的遥感植被指数NDVImax、NDVI变幅xam以及NDVI的标准偏差x′s等，分析这3个参数1983～1999年的17年中的变化情况。该数据集的建立是研究该样带土地覆盖特征及其变化规律的基础，对基于样带的全球变化研究有重要的意义。

Abstract:In this paper, the principles and methodologies for developing data set of land cover characteristics for the Northeast China transect (NECT) are discussed. Data set was developed using multi-temporal NOAA/AVHRR NDVI images with 1-km spatial resolution. Elevation, climate, soil, vegetation, land use, land resource, ecoregions, political boundaries, economic data, and social data were included as data layers, and all data layers were standardized and then integrated by digitization, spatial interpolation, geometrical registration, and projection transformation. Using this data set, several land use characteristics were mapped and analyzed. 1) Land cover mapping: multi-temporal NOAA/AVHRR NDVI images with 1-km spatial resolution were adopted to classify the land cover map of NECT. The first classification included 8 land cover types, forest, grassland, desert, shrub, cropland, mixed type, building area and water bodies. There were 12 land cover types in the second classification. Classification accuracy was 81.61% determined by ground truthing. The land cover patterns reflected the integrated physical geographical characteristics of NECT. 2) Land cover parameters were calculated using multi-temporal remotely sensed reflectance data that included annual NDVI maximum (NDVImax), seasonal NDVI amplitude (xam) and annual NDVI standard deviation (x′s). Strong inter-annual and seasonal changes in vegetation growth for different land cover types were apparent. NDVImax, xam and x′s showed an increasing trend with an increase in annual average temperatures for meadow steppe, typical steppe and desert steppe. 3) Two methods were used to evaluate seasonal changes in the length of the growing seasons over time. The length of the growing seasons of the meadow steppe and typical steppe lengthened by (9±2) days and (11±3) days, respectively, from 1983 to 1999. Similar change was not detected for the desert steppe. These results are in accordance with field records. This data set is a foundation for future research on land cover characteristics and their changes along the NECT. It also is an important contribution to international research on large-scale gradients as well as to global change research based upon the NECT.

---