泾河沿岸景观格局梯度变化及驱动力分析

Gradient variations in landscape pattern along the Jinghe River and their driving forces

以MODIS影像数据为基础,利用Fragstats为工具,对泾河主干沿岸做10 km缓冲区(Buffer) ,从景观水平和类型水平上,分析河岸带景观格局的梯度变化,同时对产生这种梯度变化的驱动力进行分析。结果表明泾河自上游而下,景观水平上的景观指数有3种不同的变化类型:(1)上升型(2 )下降型(3)无明显趋势,表现出斑块数量、丰富度、连接度的递增,形状指数、分维数和斑块大小的递减;类型水平上以农田、草地和农草交错类为主,这3种类型的景观指数表现突出,具有明显的峰值。景观格局的梯度变化受到温度、降水、土壤等影响,同时,人为干扰也是不可忽视的重要因素,经济发展,人口增加使景观类型及其分布格局发生巨大的变化。景观格局的数量化对生态系统评估、监测起重要的作用。

Jinghe river is originated at the Liupan Mountains of Shannxi province in Northwest China, and is one of the secondary tributaries of the Yellow River. The Jinghe River basin is located on the Loess Plateau which is notorious for its widespread problem of soil degradation and wind erosion. The native ecosystems of the Jinghe River basin exhibit gradient changes from the upper ranges to the lower ranges, and all are rather vulnerable to human disturbances because of their fragile environment. The goals of this study are two-fold: (1) to quantify the gradient variations in land use and land cover pattern along the main channel of the Jinghe River using landscape metrics, and (2) to explore the driving forces that may have caused the gradient changes in this riverine landscape. Based on MODIS remote sensing data for the year of 2002, we identified nine land use and land cover types using unsupervised classification techniques, and georeferenced them with global position system (GPS) points. The most dominant land use and land cover types are farmlands and grasslands. Forests, although rare, are also found in this region. To analyze the riverine landscape pattern, we created a buffer zone along the main channel of the Jinghe River with a buffer width of 10 km. The land use map and the river buffer map were overlaid, and the analysis was conducted with a moving window (10×15 km) starting from the upper edge of the river buffer area downward. A selected group of landscape metrics was used to quantify the changing pattern of land use and land cover in the study region. All of the analyses were performed using a geographic information system (GIS) and the landscape pattern analysis software, FRAGSTATS. Our results showed that the landscape metrics exhibited three different kinds of variations at the landscape scale: (1) increasing, (2) decreasing, and (3) no consistent trend. In particular, patch number, patch richness, and patch connectivity all increased from the upstream to the downstream of the Jinghe River. The area-weighted shape index, the area-weighted fractal dimension, and patch size decreased. Grasslands were the most dominant land use type in the upper ranges, characterized by convoluted, ambiguous boundaries. In contrast, farmlands dominated the downstream region and, like most other human-created patches, had regular shapes and clear-cut boundaries. At the patch type level, the most dominant classes were farmland, grassland and the ecotone between the two classes. Farmland was the only dominant class in the Jinghe River buffer area, a land use type that has less capability for conserving water and soil than forest, shrubland, and grassland. In general, the values of landscape indices for grassland, farmland, and the ecotone between the two types showed distinctive peaks along the landscape gradient. The spatial variations of the riparian landscape along the river were affected by a suite of natural and anthropogenic factors, such as temperature, precipitation, soil, and various human activities. Human population density was one of the most important factors. Rapid population growth and economic development may have significantly influenced the spatial pattern of the riverine landscape. Quantifying the landscape pattern and its change should provide useful information for effectively monitoring, assessing, and managing these ecosystems.