Movie: Comparison between the Averaged Hour UAM-V and Observed Ozone Concentration Fields for Each OTAG Episode

Description:

Six panel movies, one for each OTAG episode, comparing the UAM-V basecase D2 average hour ozone to the average hour observed ozone.

Top Left Panel: Observed ozone monitoring sites from the AIRS, CASTnet, AND EMEFS networks.
Top Middle Panel: The raw AIRS, CASTnet, AND EMEFS averaged hour ozone spatially interpolated between monitoring sites (contoured).
Top Right Panel: The UAM-V basecase D2 average hour ozone concentrations at the first level of the coarse model grid
Lower Left Panel: The difference between the average hour of the UAM-V and observed ozone concentrations
Lower Middle Panel: The ratio of the average hour of the UAM-V and observed ozone concentrations
Lower Right Panel: The ratio of the average hour of the UAM-V and observed ozone concentrations at a different scale from the lower middle panel.

Purpose:

These movies are to aid in the identification of any large scale biases in the diurnal cycle and spatial pattern of the UAM-V modeled ozone concentration fields for each OTAG episode.

Method:

The UAM-V and observed averaged hour ozone values were calculated by averaging together all ozone values that occurred at 1 am, then all values that occurred at 2 am, etc. over the length of time for each episode. Therefore, there are 24 averaged hour values for each episode.

Contoured Observed Ozone: The observed ozone values from the AIRS, CASTnet, AND EMEFS networks were spatially interpolated between the monitoring sites (contoured) to the same grid used in the coarse grid version of the UAM-V model.
UAM-V Ozone concentration fields: The UAM-V ozone values are the result of the basecase D2 model runs. The ozone values are at the lowest layer of the coarse grid which is an average value between the surface and 100 m above the ground.

Interpretation:

NOTE: The interpretation below are those of Bret Schichtel. Anyone with alternative interpretations are strongly encouraged to submit their comments.

Examination of the movies shows that the UAM-V modeled ozone generally captures the large scale positioning and features of the ozone "blobs" for each episode. However, there are two apparent bias in the model data. First, there is a bias in the diurnal cycle with the model overestimating the night time and early morning observed ozone. In many areas, the model overestimates the observed data by more than a factor of 4 in the South during the morning hours when the observed ozone is the lowest. Second, during the 1988 and 1991 episodes, the model data has a spatial bias during all hours of the day. In both the difference and ratio plots a spatial gradient is evident with the model overestimating the observed values in the south and underestimating them in the north. This spatial bias is not apparent in the 1993 and 1995 episodes, but the observed ozone values in the north are lower than the 1988 and 1991 episodes.