With the Southern Loess Plateau as the object of study,we select the nonbiological factors( physical factors),biological factors and human factors that affect the landscape of arable land to build indicator system. Using GIS,we perform the visualization expression and hierarchical storage of influencing factors to build 1 km × 1 km integrated vector and raster database of arable land landscape pattern and its influencing factors. Using spatial regression analysis,we determine the quantitative relationship between arable land landscape and its influencing factors. The results show that the arable land in the Southern Loess Plateau is mainly distributed in the regions with high temperature,great average annual precipitation,high altitude,high soil N content,small slope,GDP per unit area of land,low ≥10℃ accumulated temperature,and short distance away from the rivers and roads. The study provides a scientific basis for clarifying the relationship between arable land landscape and its influencing factors.
To study the dynamic changes of land use and predict the future land use scenarios based on the current land use,this paper uses Cellular Automata- Markov( CA- Markov) model to simulate the landscape pattern in 2030. The results show that in the study area during the period 1980- 2005,grassland and construction land increased,and woodland increased slightly; waters and unused land decreased,and arable land underwent dramatic changes. The simulation precision of CA- Markov model is 87. 28%,indicating that the use of it for simulation is reliable. The land use of the study area will be changed greatly in the future. This method provides a reference for the regions to carry out land prediction,and the research results can provide a basis for the study of optimization of land.
