Visualizing the Instability Degree of Regional Cultivated Lands Using an Imagery-Based Model

Visualizing the regional cultivated land instability degree (CLID) can be used to determine data for local government´s effort on managing agricultural productions. An imagery-based model incorporating macroscaled contribution factors was performed in Zhangjiagang Region to calculate the local CLID in 2004. Using NDVI layer and recursive threshold procedure to extract the initial state, the model was operated by the weight of summing the neighborhood effects, 3 geographical factors and 2 social and economic factors of each cultivated land pixel in GIS. The impacts of social and economic factors on CLID were assessed using a potential model and illustrated in a GIS layer for the summing. Results indicated that the model could be used to rapidly and accurately calculate regional CLID through validation by a Reference Fuzzy Kappa value of 0.405 and a Model Fuzzy Kappa value of 0.424. This imagery-based model integrated both natural-geographical factors and social-economic factors to quantify the regional CLID value and the parameters used in the model were precise and easy to acquire. Understanding the regional CLID value may help the government to drive farmers to adopt various farming strategies or to effectively guide the farming behaviors in order to compensate food production pressures and thus lead sustainable and resource-saving development.