A model-based evaluation of inversions of atmospheric transport, using annual mean mixing ratios, as a tool to monitor fluxes of nonreactive trace substances like CO2 on a continental scale

The inversion of atmospheric transport of CO2 may potentially be a means for monitoring compliance with emission treaties in the future. There are two types of errors, though, which may cause errors in inversions: (1) amplification of high-frequency data variability given the information loss in the atmosphere by mixing and (2) systematic errors in the CO2 flux estimates caused by various approximations used to formulate the inversions. In this study we use simulations with atmospheric transport models and a time independent inverse scheme to estimate these errors as a function of network size and the number of flux regions solved for. Our main results are as follows. (1) When solving for 10-20 source regions, the average uncertainty of flux estimates caused by amplification of high-frequency data variability alone decreases strongly with increasing number of stations for up to similar to 150 randomly positioned stations and then levels off (for 150 stations of the order of +/-0.2 Pg C yr(-1)). As a rule of thumb, about 10 observing stations are needed per region to be estimated. (2) Of all the sources of systematic errors, modeling error is the largest. Our estimates of SF6 emissions from five continental regions simulated with 12 different AGCMs differ by up to a factor of 2. The number of observations needed to overcome the information loss due to atmospheric mixing is hence small enough to permit monitoring of fluxes with inversions on a continental scale in principle. Nevertheless errors in transport modeling are still too large for inversions to be a quantitatively reliable option for flux monitoring.