Comparative Assessment of Various Machine Learning-Based Bias Correction Methods for Numerical Weather Prediction Model Forecasts of Extreme Air Temperatures in Urban Areas
Comparative Assessment of Various Machine Learning-Based Bias Correction Methods for Numerical Weather Prediction Model Forecasts of Extreme Air Temperatures in Urban Areas
复制标题
DOI：
10.1029/2019ea000740
复制
发表时间：
2020-04-01
影响因子：
3.1
通讯作者：
Cha, Dong-Hyun
中科院分区：
文献类型：
--
作者：
Cho, Dongjin;Yoo, Cheolhee;Cha, Dong-Hyun
预报最高和最低气温对于减轻热浪和热带夜晚等极端天气事件的损害至关重要。数值天气预报(NWP)模式已被广泛应用于气温预报，但由于其网格分辨率较低和缺乏参数化，普遍存在系统偏差。利用随机森林(RF)、支持向量回归(SVR)、人工神经网络(ANN)和多模式集成(MME)对韩国首尔地区次日最高气温和最低气温(TMaxt+1和Tmint+1)的本地资料同化和预报系统(LDAPS；a Local NWP Model Over Korea)模式输出进行了修正。用14个LDAPS模式预报数据、现场观测的日最高、最低气温和5个辅助数据作为输入变量。结果表明，LDAPS模型对TMAXT+1的预测的R-2为0.69，偏差为-0.85摄氏度，RMSE为2.08摄氏度，而通过后向预测验证，所提出的模型的R-2从0.75提高到0.78，偏差从-0.16提高到-0.07摄氏度，RMSE从1.55提高到1.66摄氏度。对于Tmint+1的预测，LDAPS模型的R-2为0.77，偏差为0.51℃，RMSE为1.43℃，而偏差修正模型的R-2值为0.86-0.87，偏差为-0.03-0.03℃，RMSE为0.98-1.02℃。通过后向预测和留一站交叉验证，MME模型具有更好的泛化性能。
Forecasts of maximum and minimum air temperatures are essential to mitigate the damage of extreme weather events such as heat waves and tropical nights. The Numerical Weather Prediction (NWP) model has been widely used for forecasting air temperature, but generally it has a systematic bias due to its coarse grid resolution and lack of parametrizations. This study used random forest (RF), support vector regression (SVR), artificial neural network (ANN) and a multi-model ensemble (MME) to correct the Local Data Assimilation and Prediction System (LDAPS; a local NWP model over Korea) model outputs of next-day maximum and minimum air temperatures ( Tmaxt+1 and Tmint+1) in Seoul, South Korea. A total of 14 LDAPS model forecast data, the daily maximum and minimum air temperatures of in-situ observations, and five auxiliary data were used as input variables. The results showed that the LDAPS model had an R-2 of 0.69, a bias of -0.85 degrees C and an RMSE of 2.08 degrees C for Tmaxt+1 forecast, whereas the proposed models resulted in the improvement with R-2 from 0.75 to 0.78, bias from -0.16 to -0.07 degrees C and RMSE from 1.55 to 1.66 degrees C by hindcast validation. For forecasting Tmint+1, the LDAPS model had an R-2 of 0.77, a bias of 0.51 degrees C and an RMSE of 1.43 degrees C by hindcast, while the bias correction models showed R-2 values ranging from 0.86 to 0.87, biases from -0.03 to 0.03 degrees C, and RMSEs from 0.98 to 1.02 degrees C. The MME model had better generalization performance than the three single machine learning models by hindcast validation and leave-one-station-out cross-validation.