Tropospheric phase delays are a major source of error in interferometric synthetic aperture radar (InSAR) measurements, particularly in coastal areas with complex atmospheric dynamics. Traditional correction methods often assume isotropy in atmospheric turbulence, overlooking potential directional dependencies.
This study investigates the effectiveness of incorporating anisotropic models into tropospheric delay correction, focusing on geodynamically active coastal regions in southern Iran including the Makran coast, southern Zagros, and central Alborz as well as inland non-coastal areas such as South Khorasan.
An empirical variogram analysis of over 1,200 interferograms revealed significant directional dependence in tropospheric turbulence, most evident in coastal zones where moisture distribution and topography are highly heterogeneous. Among the tested variogram models exponential, Gaussian, and spherical, the exponential model consistently yielded the best fit, reducing the root mean square error (RMSE) by up to 56.8%.
Further evaluation using Least Squares Collocation (LSC) across four sampling scenarios demonstrated that anisotropic modeling significantly improves delay correction under sparse or irregular sampling conditions typical of coastal environments, achieving up to 15 cm improvement (50% RMSE reduction). In contrast, for densely and uniformly sampled regions, the gain was marginal (below 0.5 cm).
These findings underscore the importance of accounting for anisotropic atmospheric behavior in InSAR applications. The proposed anisotropic correction framework offers practical value for improving deformation monitoring in regions with challenging atmospheric conditions.