A Research On Prediction Of Typhoon Extreme Precipitation Based On A Dynamic Similitude Method

Precipitation and high winds brought by typhoons can cause serious damage to society.Extreme precipitation events occur frequently in our country,and extreme precipitation disasters are more severe.Therefore,effective forecasting and early warning of typhoon extreme precipitation have a Great significance on China's anti-disaster reduction.This paper uses a combination of dynamics and statistics to study the probabilistic forecasting methods of extreme precipitation events caused by typhoons,and calculates and evaluates 25 typhoons that landed in China between 2012 and 2015,and do the sensitivity test of optimal similar number of members and effective probability threshold.Sensitivity tests were performed on the number and effective probability thresholds.The experimental results show that the method of forecasting the probability of extreme precipitation in typhoon has certain forecasting skills and can accurately forecast most stations that produce extreme precipitation,but the forecast range is slightly larger and the empty reporting rate is higher.The sensitivity test of similar members shows that 100% of the similar member ratio has a better forecast effect;on this basis,the best effective probability threshold of 24 h forecast is between 0% and 1.5%,and the best effective probability threshold of 48 h forecast is between 0% and 1%.Since different typhoon has the difference between the landing time and the starting time,the forecasting time will be 24 hours ahead of schedule in order to explore the effect of the reporting time on the forecasting method.The forecast effect is similar to that of the original forecast,but there are more empty reports and a slightly higher hit rate.In business precipitation forecasting,six precipitation levels are used universally.Therefore,the thresholds of 50 mm,100mm,and 250 mm are used to predict the probability in this paper,and compare the results with the original threshold.The analysis shows that the forecasting effect with 50 mm as the threshold is the best,followed by 100 mm and 250 mm as the threshold,and the 100 mm forecasting effect is close to the original forecast,indicating that the average of 95% extreme precipitation threshold at all sites is close to 100 mm.Finally,this method is used to conduct prediction tests on “Sudirro”,“Meddem” and “Fidte”,and compare forecasting effect by changing different regions and datas.The results show that this method has better results for the three typhoon forecasts and a higher hit rate;after reducing the forecast area,BS increases,the airtime rate decreases slightly,the hit rate remains unchanged,so the overall forecast effect is not much difference with the original forecast.In order to improve the forecasting method,Sudiro was selected as an example.After similar typhoons were selected for the first time,the secondary distance difference is added,and eleven similar typhoons are obtained.At this time,the hit rate is the same as the original forecast,but the empty report is significantly reduced.The forecast effect has obviously improved.