Correction of Non‐Geophysical Errors in SAR Doppler Shift for Ocean Surface Current Retrieval

Abstract Synthetic aperture radar (SAR) data contains Doppler information on ocean surface movement, and the Doppler centroid anomaly (DCA) can be used to quantitatively retrieve ocean surface currents in the line‐of‐sight of radar. However, non‐geophysical errors (NGE) in Doppler information significantly affect retrieval accuracy. Generally, existing methods rely on large‐area land data that are spatially and temporally close to the imaging region for NGE correction. However, these methods inherently limit the applicability and accuracy of ocean surface current retrieval. To address this issue, this study systematically analyzes the variation patterns of NGE in Gaofen‐3 (GF‐3) satellite data over the Amazon rainforest from 2019 to 2024 and proposes a novel NGE fitting method based on Doppler shift variation characteristics. The proposed method formulates the NGE fitting equation by estimating the range bias in advance and then calculating the constant offset based on nearby land regions. This study validates the proposed method using GF‐3 data over the Agulhas region and compares the results with HYCOM and SWOT data. The results indicate that the proposed method effectively corrects the NGE and significantly reduces the retrieval error of ocean surface currents. In addition, the proposed method allows flexible selection of small‐area land data in temporal and spatial proximity for NGE correction, thereby reducing the land region requirements for NGE fitting and further expanding the applicability of SAR‐based ocean surface current retrieval.