Title: Evaluations of numerical weather prediction (NWP) models from the point of view of inputs required by atmospheric dispersion models

Authors: Steven R. Hanna, Ruixin Yang, Xiangtao Yin

Addresses: Institute for Computational Sciences and Informatics, MS 5C3, 103, Science and Technology I, George Mason University, Fairfax, VA 22030-4444, USA. Institute for Computational Sciences and Informatics, MS 5C3, 103, Science and Technology I, George Mason University, Fairfax, VA 22030-4444, USA. Institute for Computational Sciences and Informatics, MS 5C3, 103, Science and Technology I, George Mason University, Fairfax, VA 22030-4444, USA

Abstract: Numerical weather prediction (NWP) models are being used to provide inputs of wind fields, vertical temperature and stability structure, surface heat and momentum fluxes, mixing depths, and other parameters to atmospheric dispersion models. However, previous evaluations of NWP models tended to focus more on predictions of weather related parameters, such as rainfall and 500 mb height. An evaluation methodology that is focused more on the needs of dispersion models is proposed and tested on two NWP models (MM5 and RAMS) as applied to a light wind period in the United States. during the summer of 1995. It is seen that the root mean square error in hourly averaged wind-speed is about 2 m/s and in wind direction is about 50 degrees. It is also seen that there is about two to four times more kinetic energy in the spatial fields of observed wind-speeds as compared to predicted wind-speeds. This difference in kinetic energy may be partly attributed to sub-grid scale turbulence, since the model predictions represent averages over the 12 km grid square, while the observations were made at points.

Keywords: model evaluation; NWP models; uncertainty in models; wind velocity errors.

DOI: 10.1504/IJEP.2000.000530

International Journal of Environment and Pollution, 2000 Vol.14 No.1/2/3/4/5/6, pp.98-105

Published online: 18 Jul 2003 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article