Application Of Ensemble Forecasts About The Surface Temperature And Precipitation Based On TIGGE Data

This paper based on the European Medium Range Weather Forecasts (ECMWF), the United Kingdom Meteorological Office (UKMO) and JMA of TIGGE data, length24h-168h, uses above the2-m ground temperatures and Precipitation as integrated applications and improved forecasting method research. First, I test and evaluate the European Medium Range Weather Forecasts (ECMWF),which predecessors have verified to be the best performance. Then, I use the dynamic performance-ranking method to select the optimized members from part of TIGGE data in East Asia made by the ECMWF and the UK Met Office, then, I compare the dynamic performance-ranking method with lagged average forecast method to see which can better show the advantages of ensemble forecasting.Finally,I use the ranking method to do some works for the Precipitation.The results show that the three-month daily temperature data can be considered as a normal distribution. The result that the effect of the control forecast is worse than ensemble one further makes it clear that ensemble forecasting has some advantages over the traditional deterministic one. With the extension of the forecast period, the forecast effect is decreased. The geographical distribution of the square root error shows that the mainly affected areas is the Qinghai-Tibet Plateau has the largest RMSE. Talagrand distribution shows there are some problems in current ensemble forecasting, that is, the divergence is not enough.In order to make better use of the ensemble forecasting, I select the optimized members by discarding the relatively poor members and selecting the better members from the ensemble forecasting system. This method is particularly applicable to the extreme weather forecast. This method is better than the simple ensemble one in terms of extreme weather in some regions with an accuracy of5℃. Comparing the results of the center of Europe and the United Kingdom Met office we can tell that the center of Europe has an advantage over the United Kingdom Meteorological Office.Lagged average forecast method (LAF) is another way to improve ensemble forecasting accuracy, it is to make an average record of forecast from different times at the same moment. Previous studies have shown that the earlier moments of the forecast reference also has value to the research. I use the data of European Centre for the LAF method study and compare it with the dynamic performance-ranking method only to find that two days LAF method is better than the3-5day LAF method for it makes up for the short comings of the ensemble means in certain areas where the record is too smooth. It has better effect in the geographical distribution of the square root error. But from an overall average of region and time, the effect is not as efficient as the simple ensemble one.Precipitation is one of the important factors affecting the weather, there are many factors affect the precipitation, including not only the large-scale atmospheric circulation, and also the MCC. Frist, talagrand distribution is used to evaluat the JMA, and then use the the ranking method to pick out the predicted better members,results show that this method have good performance results for rainfall intensity forecast, not the precipitation area forecast.