Assessing the impact of the expanding continuous measurement network in North America on carbon budgeting with an atmospheric inversion

The lack of atmospheric CO2 measurements is considered one of the most important limiting factors for carbon budgeting at regional scales using inversion methods. Fortunately, from 2004 to 2008 the North American (NA) measurement network expanded from 10 to 35 continuous CO2 measurement locations or towers. It is assumed that these additional measurements along with improvements in modeling of atmospheric transport will help improve the ability to locate and quantify terrestrial CO2 sources and sinks. This work examines the extent to which the expanded network improves inversion results at a variety of spatial and temporal scales. The study specifically addresses (1) the impact of the increased spatial and temporal measurement coverage on the ability to infer sources and sinks for NA and (2) the relative influence of inversion setup choices, such as estimation scale, on inferred fluxes when compared to the impact of additional measurements. Geostatistical inverse modeling is used to inform the analysis because it relies more heavily on the information in the atmospheric observations relative to other inversion methods. Results show that the additional measurements have a larger impact on flux estimates at regional rather than continental scales. Increasing the spatial measurement coverage helps improve flux estimates when the underlying flux field is relatively uniform, whereas including more observations outside of well-mixed afternoon conditions are also necessary to improve flux estimates in highly variable flux areas in addition to increased spatial coverage. Results also show that, the scale at which fluxes are estimated and the way in which information on the spatial and temporal covariance of fluxes is incorporated into the inversion, have an impact that is comparable to that of increasing the number of observations used in the inversion. Overall, results provide guidance for increasing the NA measurement coverage in time and space, as well as for determining how to best use existing continuous measurements in atmospheric inversions to budget CO2 fluxes at different spatiotemporal scales.