Risk Probability Forecast Technology of the Gust in Beijing Winter Olympic Games Based on the Theory of Digital Image Processing

Abstract

AbstractGust, which usually means a sudden, brief increase in the speed of the wind, has a great impact on some weather-sensitive activities, like Beijing Winter Olympic Games. Gust forecast has been a big challenge because of its uncertainty in time and spatial scale. The risk probability forecast has been often used to describe this uncertainty. Traditionally, the risk probability forecast of gust was obtained by neighborhood probability method. This method cannot accurately describe the uncertainty of time and spatial scale of gust. To solve this problem, based on the mean filtering algorithm in the digital image processing and the neighborhood probability method in meteorology, this study developed a new risk probability forecast technology of the gust, which converts the binary probability of the neighborhood probability method into the grayscale probability. And this new technology is used to post-process the two-week forecast results from the Global/Regional Assimilation and Prediction System (GRAPES) model of the China Meteorological Administration. The risk probability forecast obtained by the new technology, the observation, and the original forecast result were compared. The results show that the new risk probability products can better describe the characteristics of the gust. And compared with the gust peak speed from the original deterministic forecast results, the new risk probability products can better show some high-risk probabilities of the gust peak speed, which means it can provide forecasters with more useful information. It is concluded that the new technology is effective and significantly improves the skill of gust forecasting. This technology has good prospects for various applications and developments in numerical weather prediction.KeywordsDigital image processingRisk probability forecast of gustNumerical weather prediction