Mesoscale Spatial Variability and COAMPS Evaluation Using Measurements from a Tower Network

Many previous researches have shown values of the high-resolution models where improvements are seen in better defined and more realistic structures, which are evaluated subjectively. At the same time, high-resolution models also present a greater challenge for model evaluation when objective verification skills are used. The higher resolution models may score poorly due to timing and position errors even though the higher resolution runs clearly generated the ‘feature' that is closest to the observations. The objective of this research is to understand the spatial and temporal variability as seen from a dense mesoscale measurement network and to understand high-resolution mesoscale model performance using these measurements. The center of discussion of this research is mesoscale variability and the representation of such variability from a measurement network and from a high-resolution mesoscale model. The measurements we use are the five-minute averaged peak wind, temperature, and humidity from a network of 37 wind towers in a 30 X 40 km area at Kennedy Space Center/Cape Canaveral Air Force Station (KSC/CCAFS) from 2008, 17 of which have at least two levels of wind and temperature measurements which can be used for surface flux and surface roughness calculations. Our data analyses clearly shows the variation of surface roughness length with wind direction which corresponds well with the upwind surface characteristics. Between station correlations also revealed quantitatively the decreasing correlation with between-station distance for perturbations of different time scales and for different variables, which provided a guidance of the density of measurement network needed for adequately representing the spatial/temporal variability. Finally, fine resolution simulations in the KSC/CCAFS region during a tropical storm were made using Coupled Oceanic and Atmospheric Prediction System (COAMPS). Comparison with the tower observations suggested significant model uncertainty along the coastline, which may be related to the substantial gradient in wind and thermodynamic quantities at the land/sea boundaries.