Suitability evaluation of the rural settlements in a farming-pastoral ecotone area based on machine learning maximum entropy
Haitao Zhou,Xiaodong Na,Lin Li,Xiaoli Ning,Yanru Bai,Xiaodong Wu,Shuying Zang
DOI: https://doi.org/10.1016/j.ecolind.2023.110794
IF: 6.9
2023-08-14
Ecological Indicators
Abstract:The suitability evaluation of rural settlements is the core foundation of planning and layout optimization. Settlements in a farming-pastoral ecotone are migrative, dynamic, and diverse, and thus their suitability changes constantly. However, our limited understanding of factors that drive this dynamic process and affect the suitability of humane settlements in the farming-pastoral ecotone hindered the high-quality development of human settlements in such areas. Here we selected the ethnic minority border area of Dalham Maomingan United Banner (DMUB) in the farming-pastoral ecotone of Northern China to evaluate the suitability of its rural settlements. A data-driven machine learning maximum entropy (Maxent) method was applied to the rural settlement datasets of DMUB in years 1996, 2010, and 2020, as well as to 13 influencing factors derived from optical images and topographical ancillary data, demonstrating that the Maxent model can quantitatively measure the contribution and importance of each factor and its variation over time. Furthermore, the results showed that factors such as distance to cultivated land, population density, and distance to road had a great influence on the early-stage distribution of settlements. However, the importance of cultivated land gradually decreased with the significantly increased effect of grassland in the later period. The influence of road factors fluctuated first increasing and then decreasing. The Maxent model was also used to automatically determine suitable range for each factor according to the response curve: elevation falling between 1450 and 1650 m approximately, slope being <7°, the aspect range about 75°-225°, the optimal distance to town and hospital being within 3000 m, and vegetation cover about 0.60–0.75. Such multi-period suitability evaluation indicated that the suitable area gradually decreased, and the settlement fragmentation was serious. The settlement suitability has been dynamically transformed, but mostly toward the unsuitable development. This study provides a decision-making basis for the site selection and planning layout of rural settlements and for the livability assessment of villages in the farming-pastoral ecotone.
environmental sciences