Integrating closure phase residuals and environmental information in SBAS-InSAR: a case study of the eastern Valley of Mexico

Non-zero closure phase (NCP) residuals can bias small-baseline subset interferometric synthetic aperture radar (SBAS-InSAR) time-series deformation estimates, particularly where scattering conditions vary in space and time. This study develops an environment-informed SBAS-InSAR framework for the eastern Valley of Mexico. A CatBoost model and explainable-response analysis are used to quantify the nonlinear associations between absolute NCP residuals and geometric, land-cover, and environmental variables. These results are fused with interferometric coherence to construct a Comprehensive Environmental Evaluation Index (CEEI), which guides interferometric-network optimization and, together with land-cover information, defines spatially adaptive weights for time-series inversion. The CEEI-guided network removes 68 high-risk interferometric edges while retaining a connected network of 522 pairs and reduces the root-mean-square error of the NCP residuals by 11.88%. Independent GNSS validation shows that the joint CEEI–land-cover weighting reduces the deformation RMSE from 15.45 to 12.06 mm, corresponding to a 21.9% improvement, and decreases the deformation-velocity bias from 10.02 to 6.46 mm/year. These results demonstrate that integrating closure-phase diagnostics with environmental information improves the robustness of SBAS-InSAR deformation estimates in heterogeneous environments.