Evaluating change in agricultural landscape pattern between 1980 and 2000 in the Loess hilly region of Ansai County, China

Sustainability has become an important discussion topic across the world since the late 1980s. However, there is still lack of knowledge about the landscape sustainability and its quantitative evaluation, especially at local and regional scale levels. The sustainable development of the Chinese Loess Plateau has always been a focus in China mainly because of the occurring severe soil erosion. This study investigates the changes of the landscape pattern and the changes of the ecological sustainability of the agricultural landscapes in Ansai County between 1980 and 2000, which is located in the northern part of the Chinese Loess Plateau. Using a landscape typology as a spatial reference framework, the concept of hemeroby was used, in a multidisciplinary approach, for the assessment of the ecological aspect of agricultural landscape sustainability. We combined expert judgement with a regression model and a Geographic Information System. Fourteen variables describing the landscape structure were chosen as predictors for hemeroby. The research showed that the major changes in the agricultural landscape pattern were caused not only by socio-economic factors but also by the government policies over the past two decades, such as the large-scale eco-environment conservation programmes starting in the late of the 20th century. One of these programmes is known as the Grain-for-Green Programme (initiated in 1999). Between 1980 and 2000, the proportion of cropland-dominated landscape sharply decreased while the percentage of the mixed farmland–grassland landscape and the mixed farmland–woodland–grassland landscape increased. Meanwhile, the ecological sustainability of the agricultural landscape as a whole tended to become more sustainable, except a few landscape types. Among variables describing landscape pattern, patch size standard deviation (PSSD), total edge (TE), mean shape index (MSI), landscape shape index (LSI) and Shannon's evenness index (SEI) were shown to be the significant predicators for hemeroby. Although some data limitation, landscape structure turned out to be a good predicator for land use intensity and ecological sustainability of agricultural landscape estimated by hemeroby. As a highly integrative indicator, the concept of hemeroby provides a framework for the comparison of different land use pattern.