Reconstructing two-decadal global daily high-resolution XCO₂ records using a hybrid Transformer–BiLSTM model

Abstract. Accurate and temporally continuous global observations of atmospheric carbon dioxide (XCO₂) are essential for climate monitoring and emission assessment. However, satellite-based XCO₂ observations are often spatially incomplete and temporally discontinuous, while existing products typically suffer from coarse spatial resolutions, hindering the detection of fine-scale emission changes. Here, we developed a novel spatiotemporal Transformer-BiLSTM deep-learning network, which integrates the local temporal feature extraction capability of bidirectional long short-term memory with the global spatial dependency modeling strength of Transformer via self-attention mechanisms. The network assimilates multisource data, from satellite observations, meteorological reanalysis, and precursor gases, to reconstruct global, daily, and seamless XCO₂ at 0.1° resolution from 2003 to 2022. Validation against independent Total Carbon Column Observing Network (TCCON) measurements shows excellent agreement, with a correlation coefficient (R²) value of 0.99, a root mean square error (RMSE) of 1.10 ppm, and a mean bias of 0.01 ppm. A subsequent bias correction scheme further improves cross-satellite consistency, achieving a cross-validation coefficient of determination (CV-R²) of 0.99 and an RMSE of 0.97 ppm. Our dataset enables accurate characterization of daily XCO₂ concentrations over global land surfaces, facilitating the detection of spatial heterogeneity associated with emission hotspots and point-source activities. The record reveals a persistent global increase in atmospheric XCO₂ over the past two decades, with a mean growth rate of 2.24 ppm/yr (p < 0.001). It reliably resolves global XCO₂ variability across a wide range of temporal scales, from day-to-day fluctuations to long-term trends. It consistently captures large-scale climate-driven signals, such as ENSO-related interannual variability, and short-lived XCO₂ enhancements associated with major wildfire events, demonstrating its capability to represent both persistent and episodic emission signals. This high-resolution, daily global XCO₂ product (GlobalHighXCO₂) provides a valuable benchmark for carbon cycle studies, atmospheric model evaluation, and emission monitoring, and is publicly available at https://doi.org/10.5281/zenodo.18220961 (Qu and Wei, 2026).