Speaker
Description
Long-wavelength synthetic aperture radar (SAR) missions provide a unique opportunity to investigate changes within penetrable natural media. This study introduces a closure-phase-based framework for evaluating scene stability and scattering complexity using multi-temporal SAR observations.
For long-wavelength SAR systems, measurements over natural media such as forests, snow, and ice can contain contributions from multiple scattering components distributed within the penetrable medium. Under such conditions, nonzero interferometric closure phase is not necessarily indicative of temporal variation in SAR signal but can also arise from the vertical distribution and superposition of multiple scattering contributions. This study introduces a model-driven closure-phase stability index to distinguish expected scene-dependent closure behavior from inconsistencies across multi-baseline SAR observations.
The proposed formulation is derived through a Taylor expansion of closure phase with respect to vertical wavenumber, allowing known acquisition-geometry terms to be separated from coefficients characterizing the scattering structure of the resolution cell. The model is jointly fitted to the observed closure phases from all available acquisition triplets. The mismatch between the observed and model-consistent closure phases is then quantified through a circular residual and formulated as a pixel-wise scene stability index. Low values indicate closure behavior consistent with a common scattering structure across the acquisition stack, whereas larger values indicate increasing departure from this consistency. The proposed methodology is particularly relevant to long-wavelength missions such as BIOMASS P-band and NISAR L-band and is applicable across different multi-baseline configurations, from the relatively large perpendicular baselines of BIOMASS tomography to the smaller baseline separations typically encountered in NISAR observations. This flexibility allows closure-phase consistency to be investigated across different vertical-wavenumber sampling regimes, while the penetration capability of long-wavelength SAR provides sensitivity to changes in the internal scattering characteristics of vegetation, snow, ice, and other natural media.
Experimental analyses are conducted using BIOMASS observations over forest environments and NISAR observations over a glacier region in Nepal. Of particular interest, the availability of NISAR observations before and across the 26 August 2026 rock–glacier collapse provides an opportunity to investigate whether changes in interferometric and closure-phase consistency preceded the catastrophic failure. The NISAR-derived closure-phase stability index reveals localized changes in the scattering consistency of the glacier region prior to the subsequent failure, which may not readily apparent from SAR intensity observations alone. These results demonstrate the potential of closure-phase model consistency as a complementary diagnostic for detecting changes in the scattering state of penetrable natural media.