Quantifying production-living-ecology functions with spatial detail using big data fusion and mining approaches: A case study of a typical karst region in Southwest China

Abstract

Demands for increasing production and living functions often threaten ecological functions. Reliable quantification of land use multifunctionality (LUMF) contributes to identify and coordinate these conflicts. However, the comprehensive LUMF indicator system and fine spatial resolutions restricted each other. Previous assessments were limited to coarse grids or administrative units, thus failing to provide detailed spatial information and causing scale biases, particularly in landscapes with strong heterogeneity. Combining big data mining and fusion techniques, this study built a comprehensive and quantitative evaluation scheme of production-living-ecology (PLE) land functions with spatial detail. The scheme was applied in the karst region of southwestern China as a case study. Data on socioeconomic status, remote sensing, and location points were fused using a variable-boundary spatialized approach to quantify human activity extent and degree of convenience. Specific environmental characteristics, such as topography and karst rocky desertification, were integrated into the spatialized PLE functions. The results showed that the global distribution of the production and living functions presented some similarity, yet the high production and living functions were not completely aligned. Also, the spatial patterns of the living and production functions usually showed opposite formations to the ecology function. Unlike the global relationship, the enlarged PLE relationships in the study regions validated the importance and necessity of fine spatial information. Local relationships between the production-living and production-ecology functions were significantly different in areas with various economic development levels, particularly in built-up areas. Regions with strong synergistic relationships of PLE functions were rare and were usually found in sightseeing areas of ecology and agriculture. These results demonstrated the effectiveness of our proposed method of spatializing PLE functions with spatial detail. The findings provide detailed spatial information on local anthropogenic and physical differences and support for balancing PLE functions and policy-making about land use optimization in practice.