We evaluated the performance of the WRF-ARW model (version 3.2.1) to predict the winds at the Alcantara Launch Center applying it operationally on rocket launching occasions. Seasonal analysis of the model behavior were performed against two radiosonde data sets, which represent the dry (strong winds) and rainy (weak winds) seasons. 72 hours forecasts were made from initial conditions provided by the Global Forecasting System (GFS/NCEP). Simulations configured for three nested grid domains, maximum 1 km horizontal resolution and 42 vertical levels were compared with observations at each 6 hours through the Willmott index of agreement (d). Initial tests of sensitivity were made for comparison and adjustment of different physical and dynamic parameters, like size of domains, number of vertical levels, spin-up time and Planetary Boundary Layer parameterizations. The model was able to represent the vertical wind profile, showing no difference in performance between the dry or rainy seasons and reaching maximum values of d in the order of 0.90. Overall, the model overestimated the layer's average wind components (U and V) up to 3.0 m/s with respect to the observations.
Wind Engineering; radiosonde; numerical simulation; rocket launching
