快速城市化地区土地覆盖景观特征的粒度效应

Effects of spatial grain size on landscape pattern of land-cover types in the rapidly urbanized region

土地覆盖及其变化研究一直是全球变化研究的焦点,而土地覆盖作为生态系统的重要组成成分,具有明显的空间尺度(包括空间粒度和幅度)属性,不仅如此,尺度问题还成为景观空间研究的关键环节之一。因此,研究了解土地覆盖的尺度效应具有重要的理论意义与实际意义。基于高分辨率SPOT遥感数据,对广州市土地覆盖的景观特征进行类型及景观水平的粒度效应研究。结果表明,粒度增加对优势的景观类型有加强作用;粒度效应在类型和景观水平上都存在“临界阈”现象,粒度10、32和128是“临界阈”粒度。不同土地覆盖类型、不同特征指数的粒度响应存在差异:绿地和水域是对粒度响应最敏感的两种土地覆盖类型,而园地的斑块密度和面积百分比对粒度变化不敏感。景观水平上,斑块密度和平均形状指数的粒度响应最强烈,多样性指数则最弱,反映不同粒度形成的景观的异质性不同。

The concept of scale is of great importance and also a scientific frontier in ecology, landscape ecology, and geography. The study of scale changing involves how theories and methods under a different scale. Scale changing, therefore, affects many fields and remains to be thoroughly studied. Being basic spatial unit, different grain sizes affect the results of spatial analysis. Scale response of landscape characteristics of spatial features becomes one of the most important issues in geographical and ecological researches. The SPOT image of remote sensing of central Guangzhou on November 7, 2002 was translated as the basic vectorgraph of land cover types. The basic cell size （the original pixel） of the translated vectorgraph was 5m. A series of vectorpraphs with different grid cells were constructed by assembling the basic cells using the software Grid module of Arc/info. In these vectorgraphs, different grid cells consisted of different basic cells, which the scale was represented by the pixels in one side of the square grid cell, including 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 16, 24, 32, 64, 128, 256 and 512 pixels/basic cells. The scale effect of spatial grain sizes was analyzed under two levels of land-cover types and landscape in the study area. Results showed that dominant types of landscape were strengthened with increasing grain size. Areas of greenland and urban land, the highest two in eight types of land covers, increased with the increasing grain sizes. The increasing area ratio of urban land was largest, from 32.88% to 37.32% with the grain sizes changing from 4 × 4 to 16 × 16 pixels. Area ratios of farmland, woodland, scrubland and road ＆ square decreased with the increasing grain sizes. Among four types, the area ratio of farmland decreased most rapidly from 11.75 % to 7.54 % with grain sizes changing from 4 × 4 to 16 × 16 pixels. The area ratios of orchard and water did not change significantly with different grain sizes. There were obvious scale effects to respond to different grain sizes in land cover types and landscape. The responding extents were varied with different land cover types and landscape. In the same type, the responding extents within different grain sizes were different. Among landscape indices, the patch density （ PD）, MESH, and DIVISION of land cover were most sensitive in responding to different grain sizes. Some thresholds to respond grain sizes of some land cover types were revealed, such as grain sizes 10, 32 and 128 in this study. The results showed that landscape heterogeneities were related to grain sizes. Indices of landscape were variable under different grain sizes, such as indices to describe composing, ratio, shape, and congregating. These indices were divided into three types： stable, decreasing, and fluctuating （increasing then decreasing） with the grain sizes. The indices describing dimensional fractions of landscaoe （FRAC-AM and PAFRAC） varied.