The Influence Of Scale Difference On Grading Cultivated Land Quality

With the rapid development of economy and the progress of the society, land resource management has transformed from quantity management to quantity-quality management, in which the accuracy of the cultivated land quality grade plays a vital role. Currently, the grading data of cultivated land quality at different scales are crucial for multiscale land resource management such as macro and fine administration. However, those grading data at different scales may differ from each other in terms of accuracy, quantity and spatial distribution because of the scale effect. Therefore, investigating the scale effect on cultivated land quality grade results, and further, studying its main influencing factors as well as proposing a pragmatic method to integrate those grading data at different scales will undoubtedly contribute substantially to the better management of regional land resources.This paper conducted a case study of the Donghai County in Jiangsu Province. Firstly, the scale effect on regional cultivated land quality grade results at three different scales (1:5000, large scale; 1:100,000, medium scale; 1:500,000, small scale) were investigated. The studied scale effect included the differences in terms of the landscape characteristic, grades structure and spatial distribution. Various methods including econometric models, spatial autocorrelation model and gradient analysis were adopted. Secondly, the influencing natural and socio-economic factors of such scale effect were studied by employing the Binary Logistic Model. Finally, by the share-scale idea, a comprehensive zoning based on the different influencing factors in different sub-regions was adopted to better facilitate the interconversion and aggregation of the grading data at different scales. The main results were as follows:(1) For the studied three different scales, the corresponding grading results, though in line with each other in general, differed in patch number, area as well as the spatial distribution of cultivated land in each quality grade. To be specific, the grading result at medium scale was more consistent with that at the refered large scale than that at small scale did. While in terms of each town’s regional average cultivated land quality grade, the grading results at small scale was more consistent with the refered large scale result than that at medium scale. In terms of spatial distribution characteristics such as spatial clustering and variation, our spatial autocorrelation and gradient analysis showed that the consistency between small scale and large scale was relatively better than the consistency between medium scale and large scale.(2) The influencing factors of cultivated land grading differentiation varied from scale to scale. The Binary Logistic Model analysis implied that the grading differentiation between large scale and medium scale was mainly affected by elevation, vegetation cover, patch cohesion, the proportion of cultivated land area and the ratio of the cultivated land units between two scales. While the increasing of elevation and vegetation cover played a significant role in promoting the difference, the increasing of patch cohesion, the proportion of cultivated land area and the ratio of cultivated land unit number between medium scale and large scale significantly refrained the difference. And of all the factors, patch cohesion was the most influential factor. For the cultivated land grading differentiation between large scale and small scale, it was mainly influenced by slope, road density and patch aggregation. While the increasing of slope and road density played a significant role in promoting the difference, the increasing of patch aggregation significantly refrained the difference.(3) Based on the influencing factors of the scale effect in different sub-regions, Donghai County was comprehensively zoned to, together with the share-scale idea, better facilitate the interconversion and aggeration of the grading results at different scales. Results demonstrated that our proposed method could effectively reduce the redanduncy due to the non-critical attributes and achieve the exact consistency in terms of the area of physical quality grade, utilization quality grade, economic quality composite grade between each scale. Moreover, it’s easier to understand and utilize our method, thus better serving the management of cultivated land quality.