Evaluating Uncertainties in Terrestrial CO2 Flux Estimations: An Integrated Mesoscale Atmospheric and Biogeochemical Model Approach

Anthropogenic CO2 emissions have been partly counterbalanced by natural land and ocean biogeochemical processes. Attempts to pinpoint the terrestrial sink's causes and locations have been made using atmospheric inversions and terrestrial biogeochemical models (TBMs), but both approaches grapple with significant uncertainties. Atmospheric inversions are hindered by transport model uncertainties and sparse observations in critical areas globally, casting large uncertainties over terrestrial sink assessments. High-resolution inversions promise a nuanced understanding of the land sink's magnitude and trends in well-monitored regions, contingent on a meticulous evaluation of various regional uncertainties. TBMs face their own challenges with uncertainties arising from model parameters, structures, and initial conditions. Disparities observed in net ecosystem exchange estimates from multi-TBM intercomparison highlight our incomplete grasp of the mechanisms at play. In this study, we develop a comprehensive ensemble mesoscale atmospheric framework to scrutinize model uncertainties stemming from atmospheric transport, biospheric fluxes, fossil fuel emissions, and boundary inflow, examining their spatiotemporal patterns. Our findings underscore fossil fuel emissions as a predominant source of uncertainty in biospheric flux estimates. We observe that atmospheric transport uncertainties lessen over time, while uncertainties in large-scale CO2 boundary inflow significantly hinder precise regional flux quantification. Integrating process-based errors from the TBM CASA into our mesoscale atmospheric framework, we sought to refine and potentially optimize key parameters influencing ecosystem respiration and photosynthesis by leveraging atmospheric concentration and flux data. This combined analysis uncovers CASA's underestimation of net growing-season uptake, attributable to overlooked sink processes leading to an overestimated respiration rate. Our investigation underscores the imperative to assess inversion system components, including biospheric models, through the lens of multiple data streams, highlighting the complexity and interdependence of factors influencing CO2 flux estimates.