Contrasting Responses of Winter Wheat to Waterlogged and Non-Waterlogged Coal Mining Subsidence Revealed by Integrated DInSAR and UAV Red-Edge Remote Sensing

Abstract

This study aimed to clarify the spatiotemporal differences in the growth responses of winter wheat (Triticum aestivum L.) to mining subsidence across different growth stages and hydrological conditions in high-groundwater-level Coal–Grain Composite Areas (CGCAs). A typical high-groundwater-level mining area in Jiaozuo City, Henan Province, China, was selected as the study area. Multi-temporal Differential Interferometric Synthetic Aperture Radar (DInSAR) and UAV red-edge multispectral imagery (UAV reMSI) were integrated. Sentinel-1A data were used to extract surface deformation and delineate subsidence areas. Four vegetation indices, NDVIre, SR, OSAVI, and NDSI, were selected from 13 candidate indices, and multidimensional spectral information was integrated using principal component analysis (PCA) to construct the Coal–Grain Composite Area Winter Wheat Remote Sensing Ecological Index (CGW-RSEI) for characterizing spatiotemporal differences in winter wheat growth conditions under different subsidence conditions. The results showed that: (1) Winter wheat exhibited a clear growth stage-dependent response to mining subsidence. Compared with Check Area 2 (CK2), winter wheat in the subsidence areas (SAs) exhibited an overall growth advantage during Length of Growing Season 1 (LOS1), became suppressed during Length of Growing Season 2 (LOS2), and showed distinct responses during the End of Growing Season (EOS). (2) Evaluation using fractional vegetation cover (FVC) showed that CGW-RSEI provided a relatively stable integrated characterization across different study areas and growth stages, with a mean R² of 0.832 across the four study areas, which was higher than those of the individual vegetation indices. (3) The spatial responses of winter wheat were associated with the hydrological conditions of the SAs. At the maturity stage, the non-waterlogged areas (SA1–SA3) exhibited better growth conditions near the subsidence centers than in the peripheral areas, with CGW-RSEI values approximately 0.05–0.14 higher, whereas the waterlogged area (SA4) showed the opposite gradient, with CGW-RSEI near the subsidence center approximately 0.05 lower than that in the peripheral area. (4) Clear differences were also observed in the spatial heterogeneity of winter wheat stands. At the heading stage, the coefficients of variation (CoVs) of the CGW-RSEI in SA1–SA3 reached their minimum levels, with a mean of 1.45%, whereas SA4 reached its maximum level across the growth stages (7.00%). Field-measured plant height CoV during the same period showed a similar spatial pattern among the study areas. These findings demonstrate that winter wheat responses to mining subsidence exhibit clear growth stage dependence and spatial differentiation, and that the integration of multidimensional remote sensing information can provide a relatively stable characterization of growth differences under contrasting subsidence-related hydrological conditions. This study provides an integrated remote sensing approach for monitoring winter wheat growth conditions and identifying mining subsidence impacts in high-groundwater-level CGCAs.