CONSISTENT SAMPLING METHODS FOR COMPARING MODELS TO CO2 FLASK DATA

We address the issue of how to compare atmospheric carbon dioxide (CO2 ) observations from the National Oceanic and Atmospheric Administratio n/Climate Monitoring and Diagnostics Laboratory (NOAA/CMDL) air sampli ng network to model output in a consistent method for model evaluation and inverse problems. Two steps are needed to ''sample'' the model us ing the same methodology that is applied to the flask samples: (1) ens ure that the wind direction is one that is deemed acceptable in the fi eld and (2) reject outliers. Since a ''clean air sector'' has only bee n determined for four sites, we compute a statistical analysis of the wind direction when samples are collected at 38 NOAA/CMDL CO2 collecti on sites operational from 1993-1995. We compare this to the wind direc tions from the European Centre for Medium-Range Weather Forecasts (ECM WF) assimilated wind fields for the same 3 year time period. From this analysis, we deduce which of the sites are selecting for wind directi on. We then analyze model output to compare the effects on monthly mea n concentration from this two-step methodology. We find that at the ma jority of sites the removal of outliers is particularly important. Fin ally, we explore the impact of sampling frequency in the model. At mos t sites, the sampling frequency does impact the model results. We conc lude that it is best to use high-frequency model output rather than sa mpling with the same low-frequency as the flask network. The bias resu lting from low-frequency sampling is potentially large, and if done in both the model and the observations, the comparison error is exacerba ted. We estimate this bias error for use in error estimation and inver se problems.