Abstract
Reliable meteorological forecasts of temperature and relative humidity are critically important to take necessary measures to avoid potential damage and losses. An operational meteorological forecast model based on the Weather Research and Forecast (WRF) model has been built in Xinjiang. Numerical forecasts usually have significant uncertainties and errors due to imperfections in models themselves. In this study, a straightforward automated machine learning (AutoML) approach has been developed to post-process the raw forecasts of the WRF model. The method was implemented and evaluated to post-process forecasts from 13 stations in northern Xinjiang. The post-processed temperature forecasts were significantly improved from the raw forecasts, with average RMSE values in the 13 stations decreasing from 3.24 °C to 2.34 °C by a large margin of 28%. As for relative humidity, the mean RMSE at 13 stations decreased from 19.54% to 11.54%, or it showed a percentage decrease of 41%. Meanwhile, biases were also significantly decreased, with average ME values being reduced from around 2 °C to ~0.33 °C for temperature and improved from −15.6% to ~0% for relative humidity. Moreover, forecast performance values after post-correction became much closer to each other than raw forecast performance values, improving forecast applicability at regional scales.
Highlights
With continuous developments in numerical weather forecasting technologies, numerical weather models have become a major tool for operational weather forecasting [1]
The post-correction improvement due to automated machine learning (AutoML) compared to SLR was −0.17 °C and −1.54% for T and RH root mean square error (RMSE) values, respectively; this was more significant than the improvements of of 0.04 °C and −0.56% caused by MLR
Post-correction performance was significantly improved from the original forecasts in almost all forecast processes, as shown in Figure 14 for station 51359
Summary
With continuous developments in numerical weather forecasting technologies, numerical weather models have become a major tool for operational weather forecasting [1]. This usually requires many trial-and-tests and various forms of expertise to tune up hyperparameters in order to determine how to construct an ensemble model from multiple candidate methods To solve this tuning problem, in recent years, automatic an machine learning (AutoML) framework has been developing rapidly in the discipline of data science. Even though this approach is popular in model optimization [18], computer visions [19], and other areas, its application in MOS to improve model forecasting performance has been scarce. This study used an AutoML-based ensemble model framework to post-process WRF temperature and relative humidity forecasts in the Urumqi-Changji-Shihezhi (UCS) region, the core area of the oasis economic zone, in 2019.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
