| 基于卷积门控循环单元神经网络的临近预报方法研究 |
| |
| 引用本文： | 陈训来,刘军,郑群峰,李旭涛,刘佳,姬喜洋,陈元昭,叶允明.基于卷积门控循环单元神经网络的临近预报方法研究[J].高原气象,2021,40(2):411-423. |
| |
| 作者姓名： | 陈训来 刘军 郑群峰 李旭涛 刘佳 姬喜洋 陈元昭 叶允明 |
| |
| 作者单位： | 深圳市气象局,广东深圳518040;深圳南方强天气研究重点实验室,广东深圳518040;深圳南方强天气研究重点实验室,广东深圳518040;哈尔滨工业大学(深圳),广东深圳518055 |
| |
| 基金项目： | 广东省科技厅科技项目(2019B020208016,2014A020218014,2016A020223016);国家自然科学基金项目(41975124);广东省气象局重点研究项目(GRMC2018Z06);中国气象局预报员专项(CMAYBY2019-081)。 |
| |
| 摘 要： | 基于天气雷达资料的外推预报是灾害天气0~2 h临近预报基础,本文以业务应用为目标,应用广东省2015-2018年11部新一代多普勒雷达反射率拼图资料,研究了基于卷积门控循环单元神经网络ConvGRU的临近预报方法,采用多损失函数加权与分级加权的策略,基于ConvGRU框架建立三层自编码模型(Encoder-Decoder)的雷达回波临近预测模型,进行未来2 h逐6 min、连续20帧雷达回波图的预测,并与业务上已经应用的交叉相关法、光流法和粒子滤波法的临近预报结果对比,进行典型个例分析和长时间检验。结果表明,基于ConvGRU方法对强对流天气具有较好的预报效果,对雷达回波位置、强度和形状与实况更接近,表明深度学习方法通过对时间序列数据的学习,能较好地把握强回波区域的特征,在一定程度上能够相对比较准确地预报较强回波范围,但该方法预报雷达回波图像存在损失空间细节信息的局限,且对层状云降水的预报效果较差;ConvGRU方法的临界成功指数(CSI)和命中率(POD)评分高于传统的交叉相关法、光流法和粒子滤波法,且虚警率(FAR)评分为最小,在业务中具有广泛的应用前景。
|
| 关 键 词： | 临近预报 ConvGRU 雷达回波 |
A Study on Radar Echo Nowcasting Based on Convolutional Gated Recurrent Unit Neural Network |
| |
| CHEN Xunlai,LIU Jun,ZHEN Qunfeng,LI Xutao,LIU Jia,JI Xiyang,CHEN Yuanzhao,YE Yunming.A Study on Radar Echo Nowcasting Based on Convolutional Gated Recurrent Unit Neural Network[J].Plateau Meteorology,2021,40(2):411-423. |
| |
| Authors: | CHEN Xunlai LIU Jun ZHEN Qunfeng LI Xutao LIU Jia JI Xiyang CHEN Yuanzhao YE Yunming |
| |
| Institution: | (Shenzhen Meteorological Bureau,Shenzhen 518040,Guangdong,China;Shenzhen Key laboratory of severe weather in south China,Shenzhen 518040,Guangdong,China;Harbin Institute of Technology(Shenzhen),Shenzhen 518055,Guangdong,China) |
| |
| Abstract: | At present,the extrapolation forecast based on radar echoes is the mainstay of disaster weather 0~2 hours nowcasting. This paper proposes a convolutional gated recurrent unit neural network(ConvGRU)by using radar mosaics at 6 min intervals obtained from the radar images provided by 11 doppler radars in Guangdong Province from 2015 to 2018. Through the automatic learning of massive data,the inherent characteristics of the data and the contained physical laws can be discovered using the proposed network. A multi-loss function weighting and hierarchical weighting strategy are proposed. Based on the ConvGRU framework,a three-layer self-encoding model(Encoder-Decoder)is built for training to establish a radar echo prediction model which predicts radar echoes for 20 consecutive frames in the next 2 hours by 6 minutes. The results are compared with the operationally applied methods including tracking radar echoes by correlation(TREC),optical flow,and particle filter using typical case analysis and long-term verification. All the subjective and objective evaluation results indicate that the proposed ConvGRU method shows better forecasting performance in severe convective weather systems in predicting radar echo position,intensity and shape than other methods. These results indicate that the deep learning method can better grasp the characteristics of the strong echo area,and predict the strong echo accurately to a certain extent by automatic learning of time-series radar echo data. For the long-term evaluation results,the ConvGRU method has higher critical success index(CSI)and probability of detection(POD)scores than those of the traditional TREC,optical flow and particle filtering methods,and has the lowest false alarm rate(FAR)scores among all methods,suggesting it could be widely used in operational applications. However,the deep learning-based method has the limitation of losing spatial detail information in radar echoes due to the upsampling and down-sampling operators,and the prediction performance of stratiform cloud precipitation is relatively poor. |
| |
| Keywords: | Nowcasting ConvGRU radar echo |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
