Spatiotemporal monitoring of grasshopper habitats using multi-source data: Combined with landscape and spatial heterogeneity

Grasshoppers, as pernicious pests, pose a formidable threat to the advancement of agriculture and animal husbandry. Their presence can elicit a cascade of environmental challenges, underscoring the pressing need for effective control measures. However, grasshopper development is an intricate process influenced by diverse environmental factors with varying weights across regions, making it difficult to prevent and control. Therefore, this study focused on the prevalent infestation region, Xilingol, and selected Dasyhippus barbipes as the research subject because of its highest density and largest damaged area. Initially, according to the development mechanisms of D. barbipes, 31 habitat factors from five categories (meteorology, vegetation, soil, topography, and ecology) were selected; then, difference tests, correlation analysis, importance tests, and principal component analysis were applied to construct representative indicators for monitoring the habitat of D. barbipes (HDB). Subsequently, employing the occurrence data of D. barbipes from 2018 to 2023, a spatial pattern analysis was conducted to explore the hotspot aggregation area (HAA) and spatiotemporal characteristics of D. barbipes. Finally, considering landscape and spatial heterogeneity, the Landscape-based Geographically Weighted Logistic Regression (L-GWLR) model for HDB was constructed to achieve adaptive changes in factor weights across regions. The indicators included minimum temperature during the egg stage, precipitation, and soil temperature during the spawning stage, slope, fractional vegetation coverage in the nymph stage, soil moisture in the 1st to 3rd nymph instar, patch area, and gyration radius. The spatial pattern analysis revealed a significant spatial autocorrelation in the distribution of D. barbipes at a 90 % confidence interval (z > 1.65 and p < 0.1), and HAAs were concentrated in West Ujimqin, XilinHot, and ZhengLan. The habitat monitoring results demonstrated the superior performance of the L-GWLR model over models neglecting landscape or spatial heterogeneity. These findings provide essential support for the environmentally friendly scientific control of grasshoppers, contributing significantly to the sustainable development of agriculture and animal husbandry.