Abstract

Extensive sampling networks have been established to determine the space–time distributions of greenhouse gases so that these data can be used to provide information about the sources and sinks. However, the problem of deducing sources and sinks from concentration data is an ill-conditioned (and often underdetermined) problem and as such is subject to large amplification of errors in observations or models. Various techniques that have been introduced to address this problem are reviewed. Particular attention is given to techniques of Bayesian synthesis inversion that can provide estimates of the uncertainty for the sources that are deduced.

The Context

The atmospheric budgets of radiatively active gases such as carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) remain subject to very considerable uncertainty. This uncertainty exists despite a wide range of observational and theoretical approaches that have been used in attempts to resolve the ambiguities. One particularly important approach has been to estimate the strengths of the various source and sink processes by using trace gas concentrations from global sampling networks. The principle is that the spatial distributions of concentrations constrain the spatial distributions of sources and sinks. These constraints in turn imply constraints on the possible source and sink processes.

In principle, determining sources and sinks from the spatial distributions of trace gas concentrations (technically called inversion) provides a “snapshot” of the distribution of sources, regardless of the processes that produce the sources.