Integration of remote sensing and GIS for monitoring and evaluating agricultural land loss - A case study of Dongguan

This paper focuses on the monitoring and evaluation of agricultural land loss with the integration of remote sensing and GIS techniques. Firstly, the research provides a method for land use change detection to avoid the over-estimation of land use change in the classification of multi-date remote sensing images. Over-estimation of land use change frequently happens in the classification of remote sensing images because of the inadequate creation of classification signatures. The method can clarify some untrue land use changes obtained from conventional detection methods. This presents a quantitative method to analyze land loss based on more an 'objective' approach. The dynamic of landscape can be vividly captured and stored using the detection technique. The results of the detection reveal a fast urban growth pattern of Dongguan since the economic reform. The evaluation of the impacts of agricultural land loss is necessary for urban planning and policy making to resolve the intense land use conflicts in recent China. An evaluation system is provided with the integration of remote sensing and GIS in assessing agricultural land loss. The evaluation is mainly based on the concept of suitability loss, which is developed for the measurement of the impacts of land loss. The evaluation system consists of land development monitoring, land evaluation and GIS overlay analysis. The approach demonstrates that the indicator AREA (the total amount of land loss) is inadequate to identify the complete influences of agricultural land loss. It is illustrated that the indicator I_loss which integrates the spatial integrity of land development and land quality, is more adequate than the indicator AREA to address the impacts of land loss. If used, the indicator AREA will lead to the false perception of agricultural land loss because of the negligence of land quality. This research shows that the proposed indicator I_loss can be easily applied in the GIS environment.