Landscape Pattern And Its Dynamic Change In Shaanxi From 2000 To 2010

In a geographic space, social-ecological systems (SESs) can be viewed as social-ecological landscapes (SELs) which are both characterized by hierarchical structure. Elements or patterns at different scales exhibit different characteristics and change dynamically at different speed. Typically, dynamic change of landscape pattern at different scales reflects interaction and its dynamics between anthropogenic activities and surrounding environment at different levels. The proportion and configuration of different land cover types are two fundamental components of landscape pattern. Shaanxi province has witnessed a rapid economic development and significant increase in urbanization level in recent years. Accordingly, the demand for resources and pressure on ecosystems has increased sharply. Meanwhile, to better the condition of eco-vulnerable areas and mitigate eco-environmental change, a series of eco-recovery projects came into effect such as the Grain to Green Project. Thus, taking Shaanxi as study area, combing theories and methods of landscape ecology together with GIS and spatial analysis methods, this research aims to study the dynamics of landscape pattern and forest landscape pattern in Shaanxi separately.Based on the land cover data of Shaanxi in 2000 and 2010, reclassifying land cover into three categories, i.e., urban land use, forest and grass and agricultural land use, utilizing landscape mosaic as an landscape composition indicator, this paper will analyze the dynamic change of landscape pattern in Shaanxi and its characteristics. Reclassifying forest land cover into three categories, i.e., patch, dominant and interior based on forest density (proportion of forest land cover) at different scales, this research will analyze the scale-dependent behavior and its dynamics of all forest pixels, forest gain pixels and forest lost pixels in Shaanxi and their spatial variation. The main conclusions are as following:(1) According to the results of landscape mosaic analysis, the proportion of mosaic type FF which is wholly occupied by forest and grass in surrounding neighborhood was the highest and increased significantly. Mosaic types which are occupied by forest and grass with a percentage larger than 60% in surrounding neighborhood tended to change to type FF, suggesting a reasonable vegetation recovery at this scale. Mosaic type AA, which is occupied by agricultural land 100%, and types occupied by agricultural land larger than 60% in neighborhood, exhibiting a more dynamic character compared to other types, showed a tendency to change to type FF or types occupied by forest and grass with a percentage larger than 60%, indicating a change from agricultural land to forest and grass. As the only one type which did not transit to any other type, UU which is occupied by urban land use 100% in neighborhood showed a large increase in its percentage in spite of a relatively low proportion. Types which occupied by urban land use larger than 10% in neighborhood tended to change to type UU or types occupied by urban land use larger than 60%, suggesting an urban sprawl and influences of human activities on surrounding land use.(2) Patch, dominant and interior forest pixels in Shaanxi, study area 1 which is mainly for forest in Loess Plateau and study area 2 which is mainly for forest in Qinling Mountain area showed a similar scale-dependent behavior even with a different proportion of each type. Generally, the proportion of interior forest was high at small scales and decreased as the scale increases. The percentage of patch and dominant forest was low at small scales and increased with the increase in scale. At large scales, most of forest land cover in Shaanxi and study area 2 were interior and dominant, whereas the proportion of patch forest was much higher in study area 1, implying a relatively high fragmentation.(3) At small scales, both forest gain pixels and forest lost pixels were mainly interior, in contrast, at large scales, patch and dominant forest accounted for large proportion of forest gain pixels and patch for forest lost pixels, indicating that forest lost tended to occur in area with a local high forest density and relatively high fragmentation at large scales. By contrast, constrained by its initial spatial pattern, forest gain pixels were largely patch in study areal and dominant in study area 2 at large scales. Besides, forest lost pixels tended to occur in dominant at large scales.Through a spatial-temporal analysis of landscape pattern and its dynamics, these results explored the interaction among different land use types and dynamic change of spatial pattern in Shaanxi during this 10 years from a spatial perspective. Multi-scale analysis of forest landscape showed the multiple spatial patterns of forest landscape and its dynamic characteristics. These results could help to provide information for research on dynamic change of systems, landscape planning and management, as well as policy making and evaluation of eco-recovery project, which are all support landscape sustainable development.