Choosing a numerical weather prediction model to forecast zonda wind

SANTOS R.; HOFER A.; NORTE F.; ARANEO D.; SIMONELLI S.

Montevideo

Conferencia; WCRP Conference for Latin America and the Caribbean: Developing, linking and applying climate knowledge; 2014

WCRP

Zonda wind is very often a severe weather event that affects Mendoza province. This windstorm produces a lot of damages not only in metropolitan areas but also in the Andes foothills where fires sometimes are reported. The current strategies for Zonda objective forecasting are two: probabilistic and forecast of wind intensities based on numerical weather prediction models (NWP). The objective of this work is to explore the skills of two NWP models to forecast Zonda events over Mendoza province. Two NWP models are used namely the Global Environmental Model (GEM) and Weather Research and Forecasting model (WRF). Both models are initialized with NCAR Gobal Tropospheric Analysis (fnl). GEM is a global model with a horizontal variable grid, which is capable of forecasting windstorms, in many cases up to three days in advance; nevertheless GEM presents a bias in the forecasted maximum temperature. WRF is a current state of science atmospheric model, which presents an acceptable skill in forecasting the timing of the onset of Zonda events with a proper representation of the temperature field. Nevertheless, it is worth noting that GEM has a faster numerics than WRF, producing a six-day forecast in four-hour integration time, making GEM model as a candidate to be used in operational forecast.