ISSN 1000-0526
CN 11-2282/P
CN 11-2282/P
Archive > Volume 45 Issue 10 > 2019,45(10):1352-1362. DOI:10.7519/j.issn.1000-0526.2019.10.002 Prev Next
Comparative Analysis of Aircraft Icing Forecasting Algorithms and Research on Ensemble Prediction Model
- Article
- Figures
- Metrics
- Preview PDF
- Reference
- Related
- Cited by
- Materials
Abstract:
Based on the aircraft icing data which were obtained from the artificial rainfall enhancement, this paper uses the Weather Research and Forecasting Model to simulate 51 aircraft icing processes, contrasts and analyses the prediction results of icing potential area and intensity forecasted by seven kinds of commonly-used icing forecasting algorithms, then utilizes the score weight integration method to establish the ensemble forecasting model of aircraft icing intensity, and tests its forecasting effect. The results show that (1) in forecasting an icing case that occurred in 4 April 2002, the forecasting effect of the false frost point temperature empirical method is consistent with the actual condition but there are great differences between the effect forecasted by the other icing algorithms and the observed condition. (2) After the statistical test for the 51 aircraft icing forecast effects, the prediction effect of the false frost point temperature empirical method is the best, whose accurate rate of icing intensity forecast is up to 72.55%, followed by the RAOB method, IC index method and I icing index method, but that of improved IC index method is the poorest, only 19.61%. (3) By comparing the forecasting effects of the ensemble forecasting models established by different icing algorithms, we find that when using IC index method, the false frost point temperature empirical method, and RAOB method to forecast, the forecast accuracy rate is the highest, which is 8% higher than the best forecast accuracy by a single forecasting algorithm and the false negative rate, weak rate and strong rate can all be controlled within 10%, and the false negative rate is reduced by 4%, the strong rate is reduced by 8%.
Keywords:
Project Supported:
Clc Number:
Mobile website
