Multi-Species Correlations
Observed correlations between atmospheric concentrations of two or more species represent additional information for improving surface flux estimates through coupled inverse analyses. These correlations can arise from having similar spatial and temporal flux distributions, from chemical mass balance, or from atmospheric transport processes. We have recently demonstrated this methodology using aircraft observations of CO₂ and CO over the western Pacific. We developed methodologies to quantify CO₂-CO correlations and subsequently used them in a coupled inverse analysis [1] and showed they improve a posteriori flux estimates of CO₂. We are extending this analysis to other datasets and other carbon compounds.

Model Parameter Estimation
Conventionally, chemical inverse modelling has focused on estimating surface fluxes of gases and particles [2,3,4]. At least for biospheric fluxes such estimates provides little additional fundamental understanding. Using biospheric models that include a small set of "tunable" physical parameters we are developing statistical tools to identify which physical parameters can be estimated reliably from a particular dataset.

Use of Satellite Data
Satellite observations of the tropospheric composition are beginning to revolutionise the way we think about Earth's climate [5,6,7]. Satellite instruments literally produce Gigabytes of data/day. One approach to synthesising this large amount of data is to statistically combine them with a global 3-D chemistry transport model using a Kalman filter. Together with Manuel Gloor (Dept of Geography, University of Leeds) we are developing numerically efficient tools with which to analyse large satellite datasets.