Evaluation (CarbonTracker-CH4)

Diagnostics are available that are useful for evaluating the quality of CarbonTracker-CH4 emission estimates and distributions. Links and brief descriptions are provided below.

Comparison of Posterior and Observed CH4 at Observation Sites Used in the Assimilation

The posterior mole fractions are expected to be in better agreement with the observations than those calculated using prior flux estimates. Also, the differences between posterior and observed mole fractions should largely be within the limits of the assumed model-data mismatch error, indicating consistency of the statistical assumptions used in the assimilation.

Comparison of Posterior and Observed CH4 for Independent Aircraft Profiles (Not Used in the Assimilation)

The posterior CH4 profiles are expected to be in better agreement with the observations than those calculated using prior flux estimates. Agreement of posterior CH4 profiles with observations is a also a good test of model transport since emissions at the surface must be transported to measured altitudes. Biases may indicate problems with surface emission estimates or transport processes that mix trace species out of the atmospheric boundary layer.

Large-Scale Averages and Growth Rates of CH4

The CarbonTracker-CH4 should be able to correctly capture large-scale gradients and average abundances of CH4. It should also be able to track changes and trends in growth rates of CH4, especially since growth rates are directly related to source and sink processes.