A Deep Learning Method for Bias Correction of ECMWF 24–240 h Forecasts

Abstract

Correcting the forecast bias of numerical weather prediction models is important for severe weather warnings. The refined grid forecast requires direct correction on gridded forecast products, as opposed to correcting forecast data only at individual weather stations. In this study, a deep learning method called CU-net is proposed to correct the gridded forecasts of four weather variables from the European Centre for Medium-Range Weather Forecast Integrated Forecasting System global model (ECMWF-IFS): 2-m temperature, 2-m relative humidity, 10-m wind speed, and 10-m wind direction, with a forecast lead time of 24 h to 240 h in North China. First, the forecast correction problem is transformed into an image-to-image translation problem in deep learning under the CU-net architecture, which is based on convolutional neural networks. Second, the ECMWF-IFS forecasts and ECMWF reanalysis data (ERA5) from 2005 to 2018 are used as training, validation, and testing datasets. The predictors and labels (ground truth) of the model are created using the ECMWF-IFS and ERA5, respectively. Finally, the correction performance of CU-net is compared with a conventional method, anomaly numerical correction with observations (ANO). Results show that forecasts from CU-net have lower root mean square error, bias, mean absolute error, and higher correlation coefficient than those from ANO for all forecast lead times from 24 h to 240 h. CU-net improves upon the ECMWF-IFS forecast for all four weather variables in terms of the above evaluation metrics, whereas ANO improves upon ECMWF-IFS performance only for 2-m temperature and relative humidity. For the correction of the 10-m wind direction forecast, which is often difficult to achieve, CU-net also improves the correction performance.

Forecast calibrations of surface air temperature over Xinjiang based on U-net neural network

“…It retains the convolution and pooling layers from CNN to extract the main features of the raw data, and adds skip connections, which can identify and retain features at different spatial scales. So far, U-net has been widely applied to the convection prediction, statistical downscaling and the model forecast postprocessing (Sha et al, 2020a(Sha et al, , 2020bDupuy et al, 2021;Han et al, 2021;Lagerquist et al, 2021).…”

Deep-learning-based post-processing for probabilistic precipitation forecasting

“…Inspired by this, an increasing number of studies are being performed that apply advanced DL models in the contexts of weather forecasting (McGovern et al, 2017), climate projection (Reichstein et al, 2019), and Earth system science (Schultz et al, 2021). Specifically, as discussed by Düben et al (2021), there are many potential applications of DL in each component of the workflow for NWP, such as data assimilation (e.g., Hatfield et al, 2021), physical parameterization (e.g., Han et al, 2020), statistical downscaling (e.g., Sha et al, 2020), and post-processing (e.g., Han et al, 2021).…”

“…Cho et al (2020) assessed various machine learning (ML) models for the bias correction of extreme air temperatures and found that ML-based models have greatly improved R 2 values and reduced bias. Han et al (2021) further applied a U-shaped NN (U-Net) with encode and decode layers into postprocessing for the 2-m temperature, 2-m relative humidity, 10-m wind speed, and 10-m wind direction and obtained remarkable improvements.…”

Precipitation Retrieval From Fengyun-3D MWHTS and MWRI Data Using Deep Learning

“…The above studies showed that emerging machine learning techniques had been successfully and extensively applied in precipitation rate retrieval, increasingly becoming an important direction of precipitation estimation. Nevertheless, few studies used modern Deep Learning (DL) techniques [28], where multiple processing layers representing multiple levels of abstraction exist between the input and the output of a DL neural network (NN) [29][29] [30]. In addition, there is a key problem in the current deep learning algorithm for precipitation retrieval: Although the trained model may have high accuracy and beat previous benchmarks, they are only applied to some specific data sets.…”