干旱区生态环境知识资源中心

The value of ecosystem services in the Sanjiang plain is affected by both land cover change within the plain and habitat loss and fragmentation in the plain. Remotely images provide a synoptic view of land cover patters, and landscape statistic and pattern metrics quantize the land cover change. This study attempts to develop a methodology to quantify spatial distribution and patterns of land cover changes using landcover metrics. We derived multi-temporal land cover maps from multi-temporal remotely sensed imagery, and analyzed the quantitative characteristics and land cover change trajectories using the landcover transition metric, and then described the spatial pattern of land cover change using the landscape metrics. A case study in the Sanjiang from 1999 to 2007 shows that the number of patches (NP) and the patches density (PD) disclose Cropland is more crumbling. All have lower area-weighted mean fractal dimension index (FRAC_AM), higher area-weighted mean contiguity index (CONTIG_AM), lower interspersion and juxtaposition index (IJI) and normalized landscape shape index (NLSI), indicating greater aggregation, less association with others and simpler and larger patches in shape, respectively, in a regional view. Specifically, during the study period, the human activities are dominant in environmental change, the Wetland and Rice show a vary difference percentage of landscape (PLAND) in 1999 to 2007, but the former is decreasing and the latter is increasing. The transitions from other land cover (especially the Wetland and the Forest) to Cropland (both arid and wet) constructed the primary land cover change trajectories, and the new Cropland are more fragmentation, less association with others and simpler patches in shape. The results suggest the spatial pattern metrics of land cover and the land cover transition metric can provide an appropriate quantitative measurement for better understanding of the distribution and changes of land cover and the dominant force that cause the change with time.