Spatial and temporal modeling of hourly ozone levels for assessment of GChM modeling

In order to understand the variation of observed ozone levels over a region and a period of time, chemical transport models such as the Global Chemistry Model (GChM) are important. Currently, atmospheric chemists compare their GChM computer simulated ozone levels with the observation closest to the model node. Because of spatial, temporal, and measurement variations in the observations, this simple method of comparison is limited for assessing the performance of GChM modeling. We propose in this research two statistical methods to interpolate/extrapolate ozone levels at the model nodes from observed hourly ozone levels over August 30 and August 31, 1995 in eastern North American with longitude -71 to -75 and latitude 40 to 43. One method is the kriging, another is a two-step procedure: kriging followed by temporal analysis to consider spatial as well as temporal variability. All methods of assessment will be compared and contrasted.