Prior Information
The core of the G3W is model ingestion of observational data to estimate and reduce the uncertainty in the GHG fluxes. This analysis depends critically on the quality of ancillary data, prior flux information, and estimates of their uncertainty.
This prior information is critical, as the prior uncertainty estimates determine the range in which the optimized fluxes are allowed to deviate from the prior fluxes. A too-wide uncertainty might lead to too many degrees of freedom and, for example, unrealistic spatial or temporal distribution patterns and temporal variability. Multiple sources of prior data (ensemble of convenience) or upscaled flux measurements (validation against ground truth) may be used to provide information on the spatial and temporal error covariance structures of the prior flux uncertainty. For terrestrial CO2, a priori biogenic fluxes must include both a realistic representation of the diurnal cycles of photosynthesis and respiration, and explicit representation of heterotrophic respiration, as it is explicitly used by some inverse systems.
Some prior flux estimates used by inverse modelling are now available in just a few weeks and open the possibility of corresponding low-latency inversion products, but this should be extended to all required prior data.
The global modelling/ production centres that will participate in G3W should as much as possible follow independent methods from each other to generate and use the prior information for their data assimilation systems to ensure independence of the products.
Action P1: Identify data-stream needs on prior emission and absorption of CO2.
The activities include creating an inventory of necessary variables for atmospheric inversion and 3D state assimilation, defining temporal and spatial resolution and latency requirements, developing a data convention, coordinating an evaluation of priors and data requirements, and establishing provider commitments. Success will be measured by the availability of a list of requirements and comprehensive documentation of variables to the operational centers.
Action P2: Identify data-stream needs on prior emission and absorption of CH4.
The activities include creating an inventory of required variables for atmospheric inversion and 3D state assimilation, defining temporal and spatial resolution and latency, developing a data convention, coordinating an evaluation of priors and data requirements, and establishing provider commitments. Success will be measured by the availability of a list of requirements and comprehensive documentation of variables for the operational centers.
Action P3: Identify data-stream needs on prior emission and absorption of N2O.
The activities involve creating an inventory of variables required for atmospheric inversion and 3D state assimilation, defining their temporal and spatial resolution and latency, developing a data convention, coordinating an evaluation of priors and data requirements, and establishing provider commitments. Success will be measured by the availability of a list of requirements and comprehensive documentation of variables for the operational centers.
Action P4: Characterize the various fluxes hosted by the repository across timescales.
The activities include gathering existing information on uncertainty variances and spatiotemporal correlation structures of hosted datasets, identifying knowledge gaps, issuing recommendations to fill these gaps, and making this information clearly available to users. Success will be measured by the public availability of a synthesis of known error variances and knowledge gaps.
The description of this block is based on the G3W implementation plan. This document, approved at the 78th WMO Executive Council, prevails.