基于遥感的植被覆盖变化景观分析——以北京海淀区为例

Landscape Spatial-Temporal Pattern Analysis on Change in the Fraction of Green Vegetation Based on Remotely Sensed Data:A Case Study in Haidian District, Beijing

提出了基于土地利用分类的植被覆盖度计算的亚象元模型 ,针对不同的植被覆盖类型 ,综合利用“等密度模型”和“变密度模型”计算植被覆盖度 ,使其能有效地从遥感数据中提取植被覆盖信息。在 GIS的支持下 ,应用该模型对北京海淀区 1 975年、1 991年和 1 999年 5月份植被覆盖度进行计算 ,并着重对其景观变化过程进行了分析。研究结果表明 :1 975至 1 999年来 ,海淀区植被覆盖整体表现为增加趋势 ,但空间格局分布不均衡 ,全植被覆盖区和高植被覆盖区过分集中在西北部狭长地带 ,东南部则多为低植被覆盖区 。

Modern land surface parameterizations(LSP)in numerical weather prediction and general circulation models require specification of two major vegetation characteristics——vegetation type and vegetation amount. Vegetation amount is parameterized through the fraction of green vegetation and the leaf area index, the number of leaf layer of the vegetated part. The fraction of green vegetation, the radio of vegetation occupying a unit area, is a very important parameter in development of climatic and ecological models. However, on-ground fieldwork surveys of the fraction of green vegetation are time consuming and expensive and produce low-precision results. Estimation of the fraction of green vegetation using remotely sensed data may be a more efficient approach. The main goal of the current work was to explore the potential of deriving the fraction of green vegetation from normalized difference vegetation index(NDVI)using remotely sensed data, such as TM and MSS data. As each pixel of a satellite image represents a mosaic of structures on the ground surface, the sub-pixel models,including a “dense vegetation model” and a “nondense vegetation model”,for the fraction of green vegetation estimation have been developed. The choice between the “dense vegetation model” and the “nondense vegetation model” would depend upon the structure of vegetation, which is geographic region and vegetation type-specific. In this paper, a simplified approach, depending on the value of leaf area index, was used concerning the spatial resolution of satellite sensor data. The utility of different sub-pixel models for the fraction of green vegetation estimation based on land cover classification was also described. ;The accuracy of the model was checked, and the regression analysis was carried out to find the linear trend and bias of the curve from the 1:1 line. As a result, estimated fraction of green vegetation values are highly correlated with those observed at the ground (r 2=0.79, p<0.005, n=31). Comparison of the estimated fraction of green vegetation values and observed data demonstrated that this approach may be an important tool for solving monitoring of the regional fraction of green vegetation over large areas. The fraction of green vegetation of Haidian district in May, 1975, 1991 and 1999 were calculated using this method in Beijing. The landscape spatial-temporal patterns of change in the fraction of green vegetation and its driving processes during this period were also discussed, based on the change matrix in this paper. In the present study, this model seems to be a viable method for the fraction of green vegetation estimation over a regional area. The results of this study are valuable for conservation land planning and prediction of change in the fraction of green vegetation and can be reference for estimating the fraction of green vegetation in other areas.