Enhancing GOSAT methane observations over the high latitudes using a multi-objective genetic algorithm optimisation approach

Abstract. The Greenhouse Observing Satellite (GOSAT) is the world's first satellite mission dedicated for greenhouse gas monitoring and the University of Leicester has generated a well-validated, global data set of methane column dry-air mole fraction (XCH₄) which has been extensively used for regional and global methane emission attribution and trend analyses. However, satellite remote sensing for greenhouse gases is inherently challenging over high latitudes, due to severe cloud cover, high solar zenith angles and unfavourable surface conditions resulting in a major deficit of high-latitude winter data. A significant portion of otherwise successful retrievals are lost during the post-retrieval quality filtering process because the standard quality filtering, optimised for global data throughput, disproportionately affects, is overly restrictive for the high latitudes. Relaxing quality filters naturally leads to degradation in data quality due to increased contamination from clouds, aerosols and dark surfaces, and optimising quality filters is essential to obtain the best balance between data quality and data quantity. This study successfully improves the high-latitude throughput of the University of Leicester GOSAT Proxy XCH₄ dataset using a multi-objective genetic algorithm (GA) approach by optimising the post-retrieval filtering process, with the least impact on the data quality in comparison with the ground-based observations from Total Carbon Column Observing Network (TCCON) stations. We have found that the GA-optimised quality filtering can significantly increase the number of valid GOSAT methane observations over high latitudes by up to 20 % with a compromise of less than 1 ppb in single measurement precision. The optimisation enhances data throughput across the high-latitudes and preserves the statistical distribution and climatology of the original dataset. The optimisation process more than doubled the data in December, significantly contributing to mitigating the winter data-deficit in the high latitudes. This genetic algorithm optimisation approach holds potential for wider applicability including optimising the observation throughput for future satellite missions like CO2M, MicroCarb and GOSAT-GW.