喀斯特山区耕地时空演变与影响因子分析

喀斯特山区因其耕地资源特殊性使得社会经济发展限制因素尤为突出，厘清其演变规律对区域社会-经济-生态的可持续发展具有重要影响。该研究以贵州省耕地为研究对象，结合GIS空间分析功能，采用核密度、Lorenz曲线和地理探测器等方法对2000、2005、2010、2015、2018年共5期耕地数据的空间分布和时空演变进行深度探究。结果表明：贵州省耕地"西北多、东南少"，空间非均衡分布，省会周围、生态脆弱敏感区及喀斯特地貌发育区域耕地总量较少；耕地分布呈现空间正相关性，高-高聚类区域自然条件比较相似，低-低聚类区域大多为典型的喀斯特地貌或其他生态敏感脆弱区；时间尺度上耕地总量和密度均在2005年出现递减拐点，2000-2018年以地类转出为主，主要流向为林地、草地和建设用地，动态变化程度随时间推移愈加强烈，到末期稍减；空间尺度上各期耕地增减都较为均衡，面积的减少量稳定，增加量则由高值骤降至低值，并在低值处趋于稳定；地理探测显示影响贵州省耕地空间分布的主要因子为人口、GDP，次要因子为路网密度、平均坡度，其中社会经济因素影响力大于自然因素。

Spatio-temporal evolution of cultivated land and analysis of influence factors in karst mountainous areas

Abstract: Cultivated land is a valuable ecological resource to ensure food security and high-quality development of the national economy. Therefore, the temporal and spatial evolution of a cultivated land and its influencing factors have become a critical way to improve the quality of agricultural development, and facilitate the sustainable use of a cultivated land, finally to promote the revitalization of rural industries in karst mountainous areas of southwest China. However, less previous studies focus on the coupling effects of cultivated land and socio-economic development, as well as the spatial heterogeneity of influencing factors. There is also a lack of long-term evolution analysis and detection of influencing factors for a cultivated land in western karst mountain areas at the spatial scale. Taking the cultivated land in Guizhou Province as the research object, an in-depth exploration of spatial distribution and spatiotemporal evolution were performed on the cultivated land data for the five periods in 2000, 2005, 2010, 2015, and 2018. The spatial analysis function of GIS was combined with the kernel density, Lorenz curve, and Geo Detector methods. The area was divided into the grids of 10 km×10 km, to explore the temporal and spatial evolution characteristics and influencing factors of a cultivated land in the karst mountainous areas. The kernel density was used to identify the change in the distribution density of a cultivated land in the mountainous areas. The Lorenz curve was applied to analyze the balance on the spatial distribution of a cultivated land. Moran index was used to clarify the spatial agglomeration characteristics of cultivated land area. The dynamics degree of a cultivated land was selected to express the rate of change in the amount and distribution of cultivated land resources. The transfer matrix of land use was applied to represent the transformation methods of a cultivated land and other land types. The Geo Detector was used to quantitatively evaluate the factors and influence intensity that affecting the distribution of cultivated land and the temporal and spatial evolution in the study area. The results indicated that: The cultivated land resources were found "more in the northwest and less in the southeast" in Guizhou, showing the unbalanced spatial distribution. The area density was in clusters, with the less total cultivated land around the provincial capital, ecological vulnerable zones, and karst area. The distribution of cultivated land within the province was positively correlated to space, where the high-high clusters was similar in natural conditions, while the most low-low clusters was the typical karst areas or other ecologically fragile areas. At the time scale, the total cultivated land and density showed a turning point in 2005, and decreased gradually thereafter. From 2000 to 2018, the land use transfer was mainly outwards to the woodland, grassland and land for construction, with the dynamic changes increasingly intense over time (slightly calmed at the end). At the spatial scale, the increase and decrease of cultivated land were rather balanced across the stages, indicating that the decrease in the area was stable, while, the increase reduced from the peak to the bottom, finally stabilized at a lower position. According to the Geo Detector, the primary influencing factors for the spatial distribution of a cultivated land in Guizhou were the population and GDP, whereas, the secondary factors density of road networks and average gradient. The socio-economic factors showed a greater influence than natural factors. A clear understanding can be achieved on the distribution pattern of cultivated land in the karst mountainous area, while, the temporal and spatial evolution with the driving factors can be served as a sound reference for the efficient utilization of a cultivated land and sustainable development of the region. The finding can provide an important guidance for the sustainable development and food security in karst mountainous areas.