Split-Spectrum InSAR for Reliable Ionospheric Correction: A Critical Review of Uncertainty, Validation, and Reproducibility

MOHIT SHEODE 

Split-spectrum interferometry provides a physically grounded means of separating dispersive ionospheric phase from non-dispersive deformation-related phase in Interferometric Synthetic Aperture Radar observations. Its practical reliability, however, depends on processing choices that are often reported incompletely and validated inconsistently. This critical narrative review examines the physical basis of range split-spectrum correction and synthesizes representative literature on sub-band design, sensor and acquisition-mode dependence, coherence loss, filtering, phase unwrapping, parameter errors, and time-series application. A structured reporting audit of 25 scholarly sources shows that sensor context is usually documented, whereas spectral design, reproducibility, and uncertainty are frequently only partially reported. The review identifies a central trade-off between increased sensitivity to frequency-dependent phase and reduced stability as usable sub-band bandwidth narrows. To make reliability quantitatively operational, it states the standard phase-separation coefficients, provides covariance-aware uncertainty propagation, and translates the method into a reproducible processing sequence. It proposes that an operational correction should be accompanied by three linked products: the estimated ionospheric phase, the corrected interferometric phase, and a spatially explicit uncertainty or reliability layer. A minimum reporting framework is developed for sensor context, spectral design, interferometric processing, ionospheric estimation, validation, reproducibility, and uncertainty. Cross-sensor benchmark datasets representing different radar bands, environments, coherence regimes, and acquisition modes are recommended. The resulting framework shifts evaluation from visual phase improvement toward transparent, quantitative, and reproducible evidence that the corrected deformation signal can be trusted.