Qimin Deng , Peirong Lu , Shuyun Zhao , Naiming Yuan
{"title":"U-Net: A deep-learning method for improving summer precipitation forecasts in China","authors":"Qimin Deng , Peirong Lu , Shuyun Zhao , Naiming Yuan","doi":"10.1016/j.aosl.2022.100322","DOIUrl":"10.1016/j.aosl.2022.100322","url":null,"abstract":"<div><p>A deep-learning method named U-Net was applied to improve the skill in forecasting summer (June–August) precipitation for at a one-month lead during the period 1981–2020 in China. The variables of geopotential height, soil moisture, sea level pressure, sea surface temperature, ocean salinity, and snow were considered as the model input to revise the seasonal prediction of the Climate Forecast System, version 2 (CFSv2). Results showed that on average U-Net reduced the root-mean-square error of the original CFSv2 prediction by 49.7% and 42.7% for the validation and testing set, respectively. The most improved areas were Northwest, Southwest, and Southeast China. The anomaly same sign percentages and temporal and spatial correlation coefficients did not present significant improvement but maintained the comparable performances of CFSv2. Sensitivity experiments showed that soil moisture is the most crucial factor in predicting summer rainfall in China, followed by geopotential height. Due to its advantages in handling small training dataset sizes, U-Net is a promising deep-learning method for seasonal rainfall prediction.</p><p>摘要</p><p>本研究应用了名为U-Net的深度学习方法来提高中国夏季 (6–8月) 降水的预报技能, 预报时段为1981–2020年, 预报提前期为一个月. 将位势高度场, 土壤湿度, 海平面气压, 海表面温度, 海洋盐度和青藏高原积雪等变量作为模型输入, 本文对美国NCAR气候预报系统第2版 (CFSv2) 的季节性预报结果进行了修正. 结果显示, 在验证集和测试集上, U-Net平均将原CFSv2预测的均方根误差分别减少了49.7%和42.7%. 预报结果改善最大的地区是中国的西北,西南和东南地区. 然而, 同号率和时空相关系数没有得到明显改善, 但仍与CFSv2的预测技巧持平. 敏感性实验表明, 土壤湿度是预测中国夏季降雨的最关键因素, 其次是位势高度场. 本研究显示了U-Net模型在训练小样本数据集方面的优势, 为我国汛期季节性降雨预测提供了一种有效的深度学习方法.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 4","pages":"Article 100322"},"PeriodicalIF":2.3,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49004690","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A spatiotemporal 3D convolutional neural network model for ENSO predictions: A test case for the 2020/21 La Niña conditions","authors":"Lu Zhou , Chuan Gao , Rong-Hua Zhang","doi":"10.1016/j.aosl.2023.100330","DOIUrl":"10.1016/j.aosl.2023.100330","url":null,"abstract":"<div><p>Many coupled models are unable to accurately depict the multi-year La Niña conditions in the tropical Pacific during 2020–22, which poses a new challenge for real-time El Niño–Southern Oscillation (ENSO) predictions. Yet, the corresponding processes responsible for the multi-year coolings are still not understood well. In this paper, reanalysis products are analyzed to examine the ocean–atmosphere interactions in the tropical Pacific that have led to the evolution of sea surface temperature (SST) in the central-eastern equatorial Pacific, including the strong anomalous southeasterly winds over the southeastern tropical Pacific and the related subsurface thermal anomalies. Meanwhile, a divided temporal and spatial (TS) 3D convolution neural network (CNN) model, named TS-3DCNN, was developed to make predictions of the 2020/21 La Niña conditions; results from this novel data-driven model are compared with those from a physics-based intermediate coupled model (ICM). The prediction results made using the TS-3DCNN model for the 2020–22 La Niña indicate that this deep learning–based model can capture the two-year La Niña event to some extent, and is comparable to the IOCAS ICM; the latter dynamical model yields a successful real-time prediction of the Niño3.4 SST anomaly in late 2021 when it is initiated from early 2021. For physical interpretability, sensitivity experiments were designed and carried out to confirm the dominant roles played by the anomalous southeasterly wind and subsurface temperature fields in sustaining the second-year cooling in late 2021. As a potential approach to improving predictions for diversities of ENSO events, additional studies on effectively combining neural networks with dynamical processes and mechanisms are expected to significantly enhance the ENSO prediction capability.</p><p>摘要</p><p>2020–22年间热带太平洋经历了持续性多年的拉尼娜事件, 多数耦合模式都难以准确预测其演变过程, 这为厄尔尼诺-南方涛动(ENSO)的实时预测带来了很大的挑战. 同时, 目前学术界对此次持续性双拉尼娜事件的发展仍缺乏合理的物理解释, 其所涉及的物理过程和机制有待于进一步分析. 本研究利用再分析数据产品分析了热带东南太平洋东南风异常及其引起的次表层海温异常在此次热带太平洋海表温度(SST)异常演变中的作用, 并构建了一个时空分离(Time-Space)的三维(3D)卷积神经网络模型(TS-3DCNN)对此次双拉尼娜事件进行实时预测和过程分析. 通过将TS-3DCNN与中国科学院海洋研究所(IOCAS)中等复杂程度海气耦合模式(IOCAS ICM)的预测结果对比, 表明TS-3DCNN模型对2020–22年双重拉尼娜现象的预测能力与IOCAS ICM相当, 二者均能够从2021年初的初始场开始较好地预测2021年末 El Niño3.4区SST的演变. 此外, 基于TS-3DCNN和IOCAS ICM的敏感性试验也验证了赤道外风场异常和次表层海温异常在2021年末赤道中东太平洋海表二次变冷过程中的关键作用. 未来将神经网络与动力 模式模式间的有效结合, 进一步发展神经网络与物理过程相结合的混合建模是进一步提高ENSO事件预测能力的有效途径.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 4","pages":"Article 100330"},"PeriodicalIF":2.3,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48586843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Lin Ouyang , Fenghua Ling , Yue Li , Lei Bai , Jing-Jia Luo
{"title":"Wave forecast in the Atlantic Ocean using a double-stage ConvLSTM network","authors":"Lin Ouyang , Fenghua Ling , Yue Li , Lei Bai , Jing-Jia Luo","doi":"10.1016/j.aosl.2023.100347","DOIUrl":"10.1016/j.aosl.2023.100347","url":null,"abstract":"<div><p>Accurate forecasting of ocean waves is of great importance to the safety of marine transportation. Despite wave forecasts having been improved, the current level of prediction skill is still far from satisfactory. Here, the authors propose a new physically informed deep learning model, named Double-stage ConvLSTM (D-ConvLSTM), to improve wave forecasts in the Atlantic Ocean. The waves in the next three consecutive days are predicted by feeding the deep learning model with the observed wave conditions in the preceding two days and the simultaneous ECMWF Reanalysis v5 (ERA5) wind forcing during the forecast period. The prediction skill of the <span>d</span>-ConvLSTM model was compared with that of two other forecasting methods—namely, the wave persistence forecast and the original ConvLSTM model. The results showed an increasing prediction error with the forecast lead time when the forecasts were evaluated using ERA5 reanalysis data. The <span>d</span>-ConvLSTM model outperformed the other two models in terms of wave prediction accuracy, with a root-mean-square error of lower than 0.4 m and an anomaly correlation coefficient skill of ∼0.80 at lead times of up to three days. In addition, a similar prediction was generated when the wind forcing was replaced by the IFS forecasted wind, suggesting that the <span>d</span>-ConvLSTM model is comparable to the Wave Model of European Centre for Medium-Range Weather Forecasts (ECMWF-WAM), but more economical and time-saving.</p><p>摘要</p><p>海浪预报对海上运输安全至关重要. 本研究提出了一种涵盖物理信息的深度学习模型Double-stage ConvLSTM (D-ConvLSTM) 以改进大西洋的海浪预报. 将D-ConvLSTM模型与海浪持续性预测和原始ConvLSTM模型的预测技巧进行对比. 结果表明, 预测误差随着预测时长的增加而增加. D-ConvLSTM模型在预测准确度方面优于前二者, 且第三天预测的均方根误差低于0.4 m, 距平相关系数约在0.8. 此外, 当使用IFS预测风替代再分析风时, 能够产生相似的预测效果. 这表明D-ConvLSTM模型的预测能力能够与ECMWF-WAM模式相当, 且更节省计算资源和时间.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 4","pages":"Article 100347"},"PeriodicalIF":2.3,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44406695","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"3D DBSCAN detection and parameter sensitivity of the 2022 Yangtze river summertime heatwave and drought","authors":"Zhenchen Liu , Wen Zhou , Yuan Yuan","doi":"10.1016/j.aosl.2022.100324","DOIUrl":"10.1016/j.aosl.2022.100324","url":null,"abstract":"<div><p>Spatially and temporally accurate event detection is a precondition for exploring the mechanisms of climate extremes. To achieve this, a classical unsupervised machine learning method, the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm, was employed in the present study. Furthermore, the authors developed a 3D (longitude–latitude–time) DBSCAN-based workflow for event detection of targeted climate extremes and associated analysis of parameter sensitivity. The authors applied this 3D DBSCAN-based workflow in the detection of the 2022 summertime Yangtze extreme heatwave and drought based on the ERA5 reanalysis dataset. The heatwave and drought were found to have different development and migration patterns. Synoptic-scale heatwave extremes appeared over the northern Pacific Ocean at the end of June, extended southwestwards, and covered almost the entire Yangtze River Basin in mid-August. By contrast, a seasonal-scale drought occurred in mid-July over the continental area adjacent to the Bay of Bengal, moved northeastwards, and occupied the entire Yangtze River Basin in mid-September. Event detection can provide new insight into climate mechanisms while considering patterns of occurrence, development, and migration. In addition, the authors also performed a detailed parameter sensitivity analysis for better understanding of the algorithm application and result uncertainties.</p><p>摘要</p><p>极端气候事件的精准识别是机理分析的重要前提. 本研究借助无监督机器学习中经典的DBSCAN密度聚类算法, 发展了在三维 (经度-纬度-时间) 空间内进行目标事件识别和参数敏感性分析的研究方案. 在2022年长江全域高温伏秋旱事件识别中的应用表明, 本次天气尺度极端热浪和季节尺度重旱事件的产生发展, 空间传播模式不同. 天气尺度热浪信号自6月底从北太平洋向西南方向延伸, 直至8月中旬覆盖长江全域; 季节重旱信号于7月中旬从孟加拉湾陆面区域向东北向延伸, 直至9月中旬覆盖长江全域. 同时, 本研究中亦进行了相关参数敏感性的详细分析, 对算法应用, 结果理解亦有帮助.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 4","pages":"Article 100324"},"PeriodicalIF":2.3,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45194770","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Nan Zhao , Baoxiang Huang , Xinmin Zhang , Linyao Ge , Ge Chen
{"title":"Intelligent identification of oceanic eddies in remote sensing data via Dual-Pyramid UNet","authors":"Nan Zhao , Baoxiang Huang , Xinmin Zhang , Linyao Ge , Ge Chen","doi":"10.1016/j.aosl.2023.100335","DOIUrl":"10.1016/j.aosl.2023.100335","url":null,"abstract":"<div><p>Oceanic eddies are an omnipresent phenomenon of seawater flow and critical in transporting oceanic energy and material. Consequently, mastering and comprehending the characteristics of ocean eddies through detecting and recognizing eddies contributes to the understanding of oceanography. In traditional oceanography, a series of methods to identify eddies with physical or geometric characteristics have been developed. Deep learning frameworks have recently been applied in the eddy detection field. In this paper, a Dual-Pyramid UNet architecture that combines a pyramid split attention (PSA) module and atrous spatial pyramid pooling (ASPP) is proposed to identify oceanic eddies from remote sensing data. The encoder and decoder parts can effectively integrate low-level and high-level features, thus ensuring that feature information is not lost in large quantities after the nonlinear connection mode. In addition, the PSA and ASPP modules are introduced into the encoding, decoding, and skip connections to enhance feature extraction. Experiments were implemented in two typical study areas—the North Atlantic and South Atlantic. The recognition results demonstrate that Dual-Pyramid UNet can outperform four other competitive AI-based methods, especially for eddy edges and small-scale eddies.</p><p>摘要</p><p>海洋涡旋是大洋中重要的组成部分, 对海洋能量和物质的输送至关重要. 海洋涡旋的检测和表征无论是对于海洋气象学, 海洋声学还是海洋生物学等领域都具有重要的研究价值. 本文基于UNet架构, 并结合金字塔分割注意力(PSA)模块和空洞空间卷积池化金字塔(ASPP)构造了Dual-Pyramid UNet模型, 以平面异常和海表面温度数据中进行海洋涡旋的识别. 实验在北大西洋和南大西洋两个涡旋活跃区域进行并选用多个评价指标对识别结果进行评价以证明模型的优异性能.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 4","pages":"Article 100335"},"PeriodicalIF":2.3,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44472848","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Deep learning in extracting tropical cyclone intensity and wind radius information from satellite infrared images—A review","authors":"Chong Wang , Xiaofeng Li","doi":"10.1016/j.aosl.2023.100373","DOIUrl":"10.1016/j.aosl.2023.100373","url":null,"abstract":"<div><p>Tropical cyclones (TCs) seriously endanger human life and the safety of property. Real-time monitoring of TCs has been one of the focal points in meteorological studies. With the development of space technology and sensor technology, satellite remote sensing has become the main means of monitoring TCs. Furthermore, with its superior data mining capability, deep learning has shown advantages over traditional physical or statistical-based algorithms in the geosciences. As a result, more deep-learning algorithms are being developed and applied to extract TC information. This paper systematically reviews the deep-learning frameworks used for TC information extraction and then gives two typical applications of deep-learning models for TC intensity and wind radius estimation. In addition, the authors present an outlook on the future perspectives of deep learning in TC information extraction.</p><p>摘要</p><p>热带气旋 (TC) 严重危害人类生命和财产安全, TC的实时监测一直是研究热点, 随着空间和传感器技术的发展, 卫星遥感已成为监测TC的主要手段. 此外, 深度学习具有卓越的数据挖掘能力, 在地球科学中的表现优于基于物理或统计的算法, 越来越多的深度学习算法被开发和应用于TC信息的提取, 本文系统地回顾了深度学习在TC信息提取中的应用, 并给出了深度学习模型在TC强度和风圈半径提取中的应用. 此外, 本文还展望了深度学习在TC信息提取中的应用前景.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 4","pages":"Article 100373"},"PeriodicalIF":2.3,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42205029","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A deep learning–based U-Net model for ENSO-related precipitation responses to sea surface temperature anomalies over the tropical Pacific","authors":"Yuchao Zhu , Rong-Hua Zhang","doi":"10.1016/j.aosl.2023.100351","DOIUrl":"10.1016/j.aosl.2023.100351","url":null,"abstract":"<div><p>SST–precipitation feedback plays an important role in ENSO evolution over the tropical Pacific and thus it is critically important to realistically represent precipitation-induced feedback for accurate simulations and predictions of ENSO. Typically, in hybrid coupled modeling for ENSO predictions, statistical atmospheric models are adopted to determine linear precipitation responses to interannual SST anomalies. However, in current coupled climate models, the observed precipitation–SST relationship is not well represented. In this study, a data-driven deep learning-based U-Net model was used to construct a nonlinear response model of interannual precipitation variability to SST anomalies. It was found that the U-Net model outperformed the traditional EOF-based method in calculating the precipitation variability. Particularly over the western-central tropical Pacific, the mean-square error (MSE) of the precipitation estimates in the U-Net model was smaller than that in the EOF model. The performance of the U-Net model was further improved when additional tendency information on SST and precipitation variability was also introduced as input variables, leading to a pronounced MSE reduction over the ITCZ.</p><p>摘要</p><p>SST–降水反馈过程在热带太平洋ENSO演变过程中起着重要作用, 能否真实地在数值模式中表征SST–降水年际异常之间的关系及相关反馈过程, 对于准确模拟和预测ENSO至关重要. 例如, 在一些模拟ENSO的混合型耦合模式中, 通常采用大气统计模型 (如经验正交函数; EOF) 来表征降水 (海气界面淡水通量的一个重要分量) 对SST年际异常的线性响应. 然而在当前的耦合模式中, 真实观测到的降水–SST统计关系还不能被很好地再现出来, 从而引起 ENSO模拟误差和不确定性. 在本研究中, 使用基于深度学习的U-Net模型来构建热带太平洋降水异常场对SST年际异常的非线性响应模型. 研究发现: U-Net模型的性能优于传统的基于EOF方法的模型. 特别是在热带西太平洋海区, U-Net模型估算的降水误差远小于EOF模型的模拟. 此外, 当SST和降水异常的趋势信息作为输入变量也被同时引入以进一步约束模式训练时, U-Net模型的性能可以进一步提高, 如能使热带辐合带区域的误差显著降低.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 4","pages":"Article 100351"},"PeriodicalIF":2.3,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45065441","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yue Chen , Xiaomeng Huang , Jing-Jia Luo , Yanluan Lin , Jonathon S. Wright , Youyu Lu , Xingrong Chen , Hua Jiang , Pengfei Lin
{"title":"Prediction of ENSO using multivariable deep learning","authors":"Yue Chen , Xiaomeng Huang , Jing-Jia Luo , Yanluan Lin , Jonathon S. Wright , Youyu Lu , Xingrong Chen , Hua Jiang , Pengfei Lin","doi":"10.1016/j.aosl.2023.100350","DOIUrl":"10.1016/j.aosl.2023.100350","url":null,"abstract":"<div><p>A novel multivariable prediction system based on a deep learning (DL) algorithm, i.e., the residual neural network and pure observations, was developed to improve the prediction of the El Niño–Southern Oscillation (ENSO). Optimal predictors are automatically determined using the maximal information for spatial filtering and the Taylor diagram criteria, enabling the best prediction skills at lead times of eight months compared with most operational prediction models. The hindcast skill for the most challenging decade (2011–18) outperforms the multi-model ensemble operational forecasts. At the six-month lead, the correlation (COEF) skill of the DL model reaches 0.82 with a normalized root-mean-square error (RMSE) of 0.58 °C, which is significantly better than the average multi-model performance (COEF = 0.70 and RMSE = 0.73°C). DL prediction can effectively alleviate the long-standing spring predictability barrier problem. The automatically selected optimal precursors can explain well the typical ENSO evolution driven by both tropical dynamics and extratropical impacts.</p><p>摘要</p><p>本文基于残差神经网络和观测数据构建了一套深度学习多因子预报测模型, 以改进厄尔尼诺-南方涛动(ENSO)的预报. 该模型基于最大信息系数进行因子时空特征提取, 并根据泰勒图的评估标准可自动确定关键预报因子进行预报. 该模型在超前8个月以内的预报性能要优于当前传统的业务预报模式. 2011–2018年间, 该模型的预报性能优于多模式集成预报的结果. 在超前6个月预报时效上, 模型预报相关性可达0.82, 标准化后的均方根误差仅为0.58°C, 多模式集成预报的相关性和标准化后的均方根误差分别为0.70和0.73°C. 该模型春季预报障碍问题有所缓解, 并且自动选取的关键预报因子可用于解释热带和副热带热动力过程对于ENSO变化的影响.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 4","pages":"Article 100350"},"PeriodicalIF":2.3,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"47797791","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yi Fang , Yunfei Wu , Fengmin Wu , Yan Yan , Qi Liu , Nian Liu , Jiangjiang Xia
{"title":"Short-term wind speed forecasting bias correction in the Hangzhou area of China based on a machine learning model","authors":"Yi Fang , Yunfei Wu , Fengmin Wu , Yan Yan , Qi Liu , Nian Liu , Jiangjiang Xia","doi":"10.1016/j.aosl.2023.100339","DOIUrl":"10.1016/j.aosl.2023.100339","url":null,"abstract":"<div><p>Accurate wind speed forecasting is of great societal importance. In this study, the short-term wind speed forecasting bias at automatic meteorological stations in Hangzhou, Zhejiang Province, China, was corrected using an XGBoost machine learning model called WSFBC-XGB. The products of the local NWP (numerical weather prediction) system were used as the inputs of WSFBC-XGB. The WSFBC-XGB-corrected results were compared with those corrected using the traditional MOS (model output statistics) method. Results showed that WSFBC-XGB performed better than MOS, with the root-mean-square errors (RMSEs)/accuracy rates of the wind speed forecasting (ACCs) of WSFBC-XGB being reduced/ promoted by 26.1% and 7.64%/35.6% and 7.02% relative to NWP and MOS, respectively. The RMSEs/ACCs of WSFBC-XGB were smaller/higher than those of MOS at 90% stations. In addition, the mean decrease in impurity method was used to analyze the interpretability of WSFBC-XGB to help users gain trust in the model. Results showed that the four most important features were the wind speed at 10 m (47.35%), meridional component of wind at 10 m (12.73%), diurnal cycle (9.97%), and meridional component of wind at 1000 hPa (7.45%). The WSFBC-XGB model will help improve the accuracy of short-term wind speed forecasting and provide support for large-scale outdoor activities.</p><p>摘要</p><p>准确的风速预报具有重要的社会意义. 在本研究中, 使用名为WSFBC-XGB的XGBoost机器学习模型对中国浙江省杭州市自动气象站的短期风速预报误差进行校正. WSFBC-XGB使用本地数值天气预报系统的产品作为输入. 将WSFBC-XGB校正的结果与传统MOS(模型输出统计)方法校正的结果进行了比较. 结果表明: WSFBC-XGB预报风速的均方根误差(RMSE)/准确率(ACC)分别比NWP和MOS降低/提高了26.1%和7.64%/35.6%和7.02%; 对于90%的站点WSFBC-XGB的RMSE/ACC均小于/高于MOS. 此外, 采用平均杂质减少法对WSFBC-XGB的可解释性进行分析, 以帮助用户增加对模型的信任. 结果表明: 10米风速(47.35%), 10米风的经向分量(12.73%), 日循环(9.97%)和1000百帕风的经向分量(7.45%)是前4个最重要的特征. WSFBC-XGB模型将有助于提高短期风速预报的准确性, 为大型户外活动提供支持.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 4","pages":"Article 100339"},"PeriodicalIF":2.3,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42468550","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Projected relationship between ENSO and following-summer rainfall over the middle reaches of the Yangtze River valley based on CMIP6 simulations","authors":"Yue Sui , Guoping Xie","doi":"10.1016/j.aosl.2023.100374","DOIUrl":"10.1016/j.aosl.2023.100374","url":null,"abstract":"<div><p>El Niño–Southern Oscillation (ENSO) events have a strong impact on the middle reaches of Yangtze River summer rainfall (YRSR). The authors project how this impact might vary in the future. As in the historical period of 1979–2014, the middle reaches of the Yangtze River are projected to experience positive (negative) precipitation anomalies in post-El Niño (post-La Niña) summers in 2015–2100 under SSP5-8.5, and three related physical processes—namely, ENSO–tropical Indian Ocean (TIO) sea surface temperature (SST), TIO SST–Philippine Sea convection (PSC), and PSC–YRSR relationships—continue to have an impact. First, because the above three processes are projected to be significant, the ENSO–YRSR relationship is significant at the 90% confidence level over most periods of 2015–2100 under SSP5-8.5, according to the median of five reasonable CMIP6 models and the median of 20 reasonable EC-Earth3 runs. Second, due to the aforementioned three stronger relationships, the ENSO–YRSR relationship is projected to be somewhat stronger in 2015–2100 than in 1979–2014, based on both the median of all 30 CMIP6 models and the median of all 56 EC-Earth3 runs. Third, the ENSO–YRSR relationship projected by the median of five reasonable models remains stronger than that projected by the median of all CMIP6 models, owing to the former's stronger TIO SST–PSC and PSC–YRSR relationships under SSP5-8.5. Additionally, future assessments of uncertainty of projections in ENSO–YRSR relationship are still necessary.</p><p>摘要</p><p>厄尔尼诺—南方涛动 (ENSO) 对长江中游夏季降水 (YRSR) 的年际变化影响较大. 基于CMIP6模式数据, 本文预估了未来ENSO与长江中游夏季降水关系的变化. 与1979∼2014年相似, SSP5-8.5高排放情景下ENSO−YRSR的关系仍表现为前冬发生厄尔尼诺 (拉尼娜) 后, 长江中游夏季降水为正异常 (负异常). 同时, 仍然受三个物理过程影响:前冬ENSO影响次年夏季印度洋海温 (ENSO−TIO SST), 印度洋海温异常进而影响菲律宾对流 (TIO SST−PSC) , 菲律宾对流对长江中游夏季降水产生影响 (PSC−YRSR). 例如, (1) 5个CMIP6好模式的中位数和20个EC-Earth3好子集的中位数均预估ENSO−YRSR在2015∼2100年大部分时段保持显著正相关关系, 因为上述三个物理过程的相关关系在未来也显著. (2) 30个CMIP6模式的中位数和56个EC-Earth3子集的中位数预估ENSO−YRSR关系略有增强; 主要是因为上述三个物理过程在未来变强. (3) 5个CMIP6好模式的中位数预估ENSO−YRSR关系仍强于30个CMIP6模式的中位数结果, 主要是因为前者预估的TIO SST−PSC和PSC−YRSR关系更强. 未来将关注ENSO−YRSR预估的不确定性来源.</p></div>","PeriodicalId":47210,"journal":{"name":"Atmospheric and Oceanic Science Letters","volume":"16 6","pages":"Article 100374"},"PeriodicalIF":2.3,"publicationDate":"2023-05-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46308738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}