A global harmonized daily 0.1° XCH4 from GOSAT, GOSAT-2, and TROPOMI using machine learning-based bias correction and data fusion

This dataset provides the first global fusion of GOSAT, GOSAT-2, and TROPOMI satellite retrievals, generating a globally consistent daily 0.1° land-only XCH4 product for 2020–2023. Values are generated through a three-step machine learning approach: (1) sensor-specific bias correction using TCCON ground-based observations, (2) cross-sensor harmonization to GOSAT-2 as the reference sensor with the strongest post-correction TCCON agreement, and (3) priority-based fusion based on agreement with TCCON (GOSAT-2 → TROPOMI → GOSAT). Independent validation at withheld TCCON stations demonstrates robust generalization of the fused product (R² = 0.81, RMSE = 10.78 ppb), outperforming standard and operational bias-corrected satellite products and previously reported ML-based approaches. Regional assessments confirm 2.5–12% annual coverage improvements in data-sparse regions and 33–94% bias reductions compared to TROPOMI operational products across South Asia, the Amazon Basin, and Eastern Siberia. Associated preprint: Keya, J. N., Kim, Y., Choi, H., and Im, J.: Enhanced methane monitoring: A globally harmonized daily 0.1° XCH4 through machine learning-based fusion of GOSAT, GOSAT-2, and TROPOMI, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2026-1034, 2026. Each monthly file fused_xch4_YYYYMM.h5 contains all daily data for that month. Variable Dimensions Units Description fused_xch4 (1400, 3600) ppb Daily or monthly mean fused XCH4 gosat2_ml_bc_xch4 (1400, 3600) ppb ML bias-corrected GOSAT-2 XCH4 gosat_harmonized_xch4 (1400, 3600) ppb GOSAT XCH4 harmonized to ML BC GOSAT-2 scale tropomi_harmonized_xch4 (1400, 3600) ppb TROPOMI XCH4 harmonized to ML BC GOSAT-2 scale selected_sensor_id (1400, 3600) — ensor ID per pixel (1=GOSAT, 2=GOSAT-2, 3=TROPOMI) lat (1400,) ° 1D latitude array (60°S to 80°N) lon (3600,) ° 1D longitude array (180°W to 180°E) date scalar — Date string YYYYMMDD ----- TECHNICAL DETAILS ----- Spatial resolution : 0.1° x 0.1° Spatial coverage : 60°S–80°N,global land Temporal coverage : 2020–2023 (daily) Units : ppb (parts per billion) Satellites : GOSAT, GOSAT-2, TROPOMI Fusion priority : GOSAT-2 → TROPOMI → GOSAT Bias correction : ML-based (XGBoost/Random Forest), calibrated against TCCON. Validation : R² = 0.81, RMSE = 10.78 ppb (independent TCCON stations) File format : HDF5 (.h5) with gzip compression Grid reference : WGS84