Resumen
Food security has been one of the greatest global concerns facing the current complicated situation. Among these, the impact of climate change on agricultural production is dynamic over time and space, making it a major challenge to food security. Taking the U.S. Corn Belt as an example, we introduce a geographically and temporally weighted regression (GTWR) model that can handle both temporal and spatial non-stationarity in the relationship between corn yield and meteorological variables. With a high fitting performance (adjusted R2 at 0.79), the GTWR model generates spatiotemporally varying coefficients to effectively capture the spatiotemporal heterogeneity without requiring completion of the unbalanced data. This model makes it possible to retain original data to the maximum possible extent and to estimate the results more reliably and realistically. Our regression results showed that climate change had a positive effect on corn yield over the past 40 years, from 1981 to 2020, with temperature having a stronger effect than precipitation. Furthermore, a fuzzy c-means algorithm was used to cluster regions based on spatiotemporally changing trends. We found that the production potential of regions at high latitudes was higher than that of regions at low latitudes, suggesting that the center of productive regions may migrate northward in the future.