TexAQS 2006 Twin Otter TOPAZ Comparisons
Results from 8 AQ forecast models compared with TOPAZ ozone lidar cross-sections
Details of the TOPAZ lidar instrument. The Twin Otter typically flew 3100 to 3500 meters ASL with the nadir viewing lidar collecting profiles every 10 seconds (~600 meter horizontal resolution) and vertical resolution of 90 meters. PBL heights from the Twin Otter are determined form an automated wavelet analysis of the aerosol backscatter returns. For the model comparisons the aircraft is flown through each model domain, and the O3 profiles are averaged over model grids in the horizontal and vertical. There is often missing values in the lidar returns, due to clouds, aircraft turns, or other causes. Observed lidar data are averaged into model grids only if more than 50% of the possible lidar data is available for that grid cell. A lidar value is considered only if its vertical overlap with a model grid is more than 50% (for either the model or the lidar grid). No interpolation is performed in either the vertical or horizontal, only nearest-neighbor determination go into the averaging. Once the lidar O3 is averaged over a model grid structure, the plots for both model and observations are made through bilinear interpolation on an identical 200 by 200 mesh.
A description of the forecast models, along with pertinent surface and airborne evaluation statistics, can be found in: McKeen et al, JGR-Atmospheres, vol. 114, doi 10.1029/2008JD011697. Hourly model output is used in all cases. Model results show the O3 cross-sections, PBL heights, and cloud edges. For cloud edges the 1.0×10-5g/kg contour is plotted for those models that supplied condensed water information (AURAMS, BAMS, CMAQ-5x and WRF/Chem models. For the BAMS and STEM models equivalent potential temperature from supplied T, p, and Qv fields are used to determine PBL heights. For the CHRONOS model no PBL or met fields were supplied, only relative humidity. For this model the level of maximum RH gradient is shown. This corresponds closely to O3 gradients on some days (e.g. 8/14/06) but not on all days, and thus is not suitable for direct comparisons.
O3 averages for models and observations, averaged from 0-1500 meters AGL are also shown in the plots. Model averages are determined only for grid cells that contains O3 lidar entries. This the lowest model levels are usually missing in the averages.