Determining Uncertainties Associated With Retrieved AOD Based on the Dark Object Method: A Case Study With GF-1/WFV Satellite Data

The retrieval of aerosol optical depth (AOD) from remote sensing data using the dark object (DO) method, which utilizes the measured radiance in areas with dark underlying scenes, such as dark vegetation or shadowed areas, is widely applied in atmospheric science. While many studies have aimed to improve its accuracy and validate its effectiveness, comprehensive uncertainty analysis associated with AOD retrieved based on the DO method remains limited. Therefore, this study focuses on addressing the uncertainty issues in AOD retrieval, using shadowed regions as a case study based on the DO method. We developed a method suitable for AOD retrieval from China's GF-1/WFV satellite data, which lacks shortwave infrared band information, and analyzed the retrieval process to identify the sources and propagation of uncertainty in the AOD retrieval process. Following metrological principles, we performed thorough calculations and analysis of the uncertainties associated with key influencing factors throughout the retrieval process, including with the retrieval algorithm, input data and related aspects. Furthermore, by considering the interrelationships among influencing factors and synthesizing uncertainties, we traced the propagation of uncertainty from GF-1/WFV Level 1 products to the final AOD products. The AOD retrieval results indicate strong consistency with AERONET ground observations. We characterized the uncertainties associated with AOD retrieval using an uncertainty tree diagram, quantifying specific uncertainty results. This provides insights into satellite-based AOD retrieval methods and investigates the traceability of China's land satellite AOD products.