A study of regional-scale variability of in situ and model-generated tropospheric trace gases: Insights into observational requirements for a satellite in geostationary orbit

We examine the results from a regional-scale chemical-transport model with 4-km resolution to determine the spatial variability of trace gases on this scale. Model-derived variability statistics are generated using 1st-order structure functions and then compared with in situ trace gas measurements from a series of aircraft campaigns. The variability of the observations and the model-derived concentrations are found to be in reasonable agreement for O3 and CO, but the model underestimates the observed variability of NO2. Variability statistics are then calculated for model-derived tropospheric column integrals. These integrals are calculated for 0–10 km (representative of the entire tropospheric column), 0–2 km (representative of the planetary boundary layer, PBL) and 2–10 km (representative of the free troposphere, FT). For each of the species examined, the variability of the tropospheric column is generally controlled by the variability in the lowest 2 km. The degree of control for each trace gas, however, is different. Whereas NO2 is completely dominated by PBL processes, CO variability in the FT contributes appreciably to the variability of the entire tropospheric column, suggesting that two independent pieces of information for CO would be most helpful for describing the variability of the entire tropospheric column. Likewise, the variability of an independent free tropospheric measurement of O3 would provide additional insight into the observed variability of the entire column, but the amount of additional information provided by a separate FT measurement is not as beneficial to what was found for CO. We provide additional analyses to quantify relationships that can be used to better understand the model-derived structure functions and their dependence on grid size and time of day. Lastly we present a practical example of how this information may be used for guidance in the development of science requirements for future satellite instruments since measurements from these instruments must be able to resolve smaller scale gradients to be used successfully for air quality applications.