关键词: 粮食生产, 时空立方体, 时空演变, 时空地理加权回归, GIS

Abstract:

[Objective] Sichuan province, recognized as a strategic core region for China's food security, exhibits spatiotemporal dynamics in grain production that have significant implications for regional resource allocation and national food strategies. Previous studies have primarily focused on spatial dimension of grain production, treating temporal and spatial characteristics separately. This approach neglected the non-stationary effects of time and failed to integrate the evolution patterns of time and space in a coherent manner. Consequently, the interrelationship between the temporal evolution and spatial distribution of grain production was not fully elucidated. A spatiotemporal integrated analysis framework along with a method for extracting spatiotemporal features was proposed in this study. The objective is to elucidate the spatial pattern temporal variations in Sichuan province, thereby providing a scientific basis for regional food security management. [Methods] The study was based on county-level panel data of Sichuan province spanning the years 2000 to 2019. Multiple spatiotemporal analysis techniques were employed to comprehensively examine the evolution of grain production and to identify its driving mechanisms. Initially, standard deviation ellipse analysis and the centroid migration trajectory model were applied to assess the spatial distribution of major grain-producing areas and their temporal migration trends. This analysis enabled the identification of spatial agglomeration patterns and the direction of change in grain production. Subsequently, a three-dimensional spatiotemporal framework was constructed based on the space-time cube model. This framework integrated both temporal and spatial information. Hotspot analysis and the local Moran's I statistic were then utilized to systematically identify the distribution of cold and hot spots as well as spatial clustering patterns in county-level grain output. This approach revealed the spatiotemporal hotspots, clustering characteristics, and the evolving trends of grain production over time. Finally, a spatiotemporal geographically weighted regression model was employed to quantitatively assess the influence of various factors on grain production. These factors included natural elements (such as topography, climate, and soil properties), agricultural factors (such as the total sown area, mechanization level, and irrigation conditions), economic factors (such as per capita gross domestic product and rural per capita disposable income), and human factors (such as rural population and nighttime light intensity). The analysis elucidated the spatial heterogeneity and evolution of the principal driving forces affecting grain production in the province. [Results and Discussions] A high-yield core area was established on the eastern Sichuan plain, with its spatial distribution exhibiting a pronounced northeast-southwest orientation. The production centroid consistently remained near Lezhi County, although it experienced significant shifts during the periods 2000－2001 and 2009－2010. In contrast, the grain production levels in the western Sichuan plateau and the central hilly regions were relatively low. Over the past two decades, the province demonstrated seven distinct patterns in the distribution of cold and hot spots and three clustering patterns in grain production. Specifically, grain output on the Chengdu Plain continuously increased, the decline in production on the western plateau decelerated, and production in the central region consistently decreased. Approximately 64.77% of the province exhibited potential for increased production, particularly in the western region, where improvements in natural conditions and the gradual enhancement of agricultural infrastructure contributed to significant yield growth potential. Conversely, roughly 16.93% of the areas, characterized by complex topography and limited resources, faced potential yield reductions due to resource scarcity and restrictive cultivation conditions. The analysis further revealed that agricultural factors served as the dominant determinants influencing the spatiotemporal characteristics of grain production. In this regard, the total sown area and the area of cultivated land acted as positive contributors. Natural factors, including slope, soil pH, and annual sunshine duration, exerted negative effects. Although human and economic factors had relatively minor influences, indicators such as population density and nighttime light intensity also played a moderating role in regional grain production. The maintenance of agricultural land area proved crucial in safeguarding and enhancing grain yields, while improvements in natural resource conditions further bolstered production capacity. These findings underscored the inherent spatiotemporal disparities in grain production within Sichuan province and revealed the impact of agricultural resource allocation, environmental conditions, and policy support on the heterogeneity of spatial production changes. [Conclusions] The proposed spatiotemporal integrated analysis framework provided a novel perspective for elucidating the dynamic evolution and driving mechanisms of grain production in Sichuan province. The findings demonstrated that the grain production pattern exhibited complex characteristics, including regional concentration, dynamic spatiotemporal evolution, and the interplay of multiple factors. Based on these results, future policies should emphasize the construction of high-standard farmland, the promotion of precision agriculture technologies, and the rational adjustment of agricultural resource allocation. Such measures are intended to enhance agricultural production efficiency and to improve the regional eco-agricultural system. Ultimately, these recommendations aim to furnish both theoretical support and practical guidance for the establishment of a stable and efficient grain production system and for advancing the development of Sichuan province as a key granary.

Key words: grain production, spatio-temporal cube, spatial-temporal evolution, GTWR, GIS