Using improved Gray Clustering Method to evaluate the degree of damage to arable lands in mining areas

China has limited cultivated land and huge population, making China's per-capita cultivated land far smaller than most other countries. Although it brings enormous economic wealth to the country, mineral resources mining has also destroyed a large fraction of cultivated lands. Hence for the efficient protection of cultivated lands, it is critical to evaluate (in terms of quantity and quality) cultivated land damage via mining operations in China. Grey Clustering Method can be used in such evaluations in terms of fuzzy mathematical theory. Although widely used in other fields, its application in evaluation the degree of damage caused by mining operations to cultivated lands has been highly limited. The task of this paper was to analyze the degree to which arable lands had been damaged by mining operations. The study aimed to more objectively and accurately assess the services of arable lands and the measures for reclamation and protection. The classic Gray Clustering Method was used to determine the weights of the indicators for damaged arable lands, with appropriate adjustments to relate the threshold of the indicators to cluster objects. In this paper, effective soil thickness was set to dominant restrictive indicator. Then the relative restrictive indicators were integrated into the Gray Clustering Method to construct a GCM_DR (Grey Clustering Method Dominant Restrictive Indicator and Relative Restrictive Index) which was then used to evaluate the extent of damage of arable lands in mining areas. The model was used to evaluate some damaged arable lands in the high-dive mining area and the results compared with other evaluation methods. The results showed that the degree of damage to a total of eight evaluated units of land in two collapsed basins were respectively II, III, IV, V, I, II, Ⅲ and Ⅲ. This was approximately consistent with the results of other methods, suggesting a high reliability of the proposed method. The set of dominant restrictive indicators that accounted for the combined factors and dominant factors of the model somewhat reduced assessment process and improved evaluation efficiency. Compared with the classical Gray Clustery Model, the improved model emphasized relative restrictive index which was more sensitive. The model allocated higher weights to indexes with greater degree of damage so that the clustering coefficient closely reflected the direction of severe damage. In terms of weight ratio, the improved model emphasized the relative degree of damage expressed by the clustering coefficient. Thus to some extent, it reduced the irrationality of weight distribution caused by different hierarchical spans. This made GCM_DR model more applicable in evaluating the degree of damage to cultivated lands in mining areas. The model was suitable for determining the state of arable lands, establishing land reclamation measures and consolidating land protection programs.