Seasonal Analysis of Polarimetric Responses Utilizing Ku-Band GB-SAR Time Series Data

Abstract

This study evaluated the seasonal dynamics of $H/\bar {\alpha }$ target decomposition parameters derived from the spatially and temporally averaged coherence matrix of fully polarimetric Ku-band GB-SAR time series data over a one-year analysis period across three distinct land cover types in a residential area in Hokkaido, Japan. Overall, mean values of $H_{spatial}/\bar {\alpha }_{spatial}$ and $H_{temp}/\bar {\alpha }_{temp} $ parameters were identified across different $H/\bar {\alpha }~2$ D space zones for each land cover type, indicating that they exhibited a specific characteristic associated with their scattering mechanism. Subsequently, seasonal changes were observed to shift their seasonal mean values, with the transitions occurring within the same $H/\bar {\alpha }~2$ D space zone for the built-up land cover type. As for tree land cover type, zonal alterations were more prevalent in $H_{spatial}/\bar {\alpha }_{spatial}$ than in $H_{temp}/\bar {\alpha }_{temp}$ for each seasonal transition. However, only the $H_{spatial}/\bar {\alpha }_{spatial}$ seasonal mean values for the shrub land cover type demonstrated zone changes. Afterward, correlations between precipitation and $H_{spatial}/\bar {\alpha }_{spatial}$ parameters were more pronounced in vegetated-related land cover types than in man-made ones. Meanwhile, snowfall had stronger relationships with built-up and tree areas, whereas shrub areas showed weaker correlations. Concurrently, statistically significant positive correlations were identified between $H_{temp}$ and precipitation across all land cover types. Nevertheless, the $\bar {\alpha }_{temp}$ yielded mixed results. For snowfall, built-up areas demonstrated more pronounced responses than trees and shrubs. Finally, outputs of this work could enhance knowledge of the underexplored yet compelling implementation of Ku-band GB-SAR time series data for continuous land cover monitoring activities.