As a natural disaster,extreme precipitation is among the most destructive and influential,but predicting its occurrence and evolution accurately is very challenging because of its rarity and uniqueness.Taking the exam...As a natural disaster,extreme precipitation is among the most destructive and influential,but predicting its occurrence and evolution accurately is very challenging because of its rarity and uniqueness.Taking the example of the“21·7”extreme precipitation event(17–21 July 2021)in Henan Province,this study explores the potential of using physics-guided machine learning to improve the accuracy of forecasting the intensity and location of extreme precipitation.Three physics-guided ways of embedding physical features,fusing physical model forecasts and revised loss function are used,i.e.,(1)analyzing the anomalous circulation and thermodynamical factors,(2)analyzing the multi-model forecast bias and the associated underlying reasons for it,and(3)using professional forecasting knowledge to design the loss function,and the corresponding results are used as input for machine learning to improve the forecasting accuracy.The results indicate that by learning the relationship between anomalous physical features and heavy precipitation,the forecasting of precipitation intensity is improved significantly,but the location is rarely adjusted and more false alarms appear.Possible reasons for this are as follows.The anomalous features used here mainly contain information about large-scale systems and factors which are consistent with the model precipitation deviation;moreover,the samples of extreme precipitation are sparse and so the algorithm used here is simple.However,by combining“good and different”multi models with machine learning,the advantages of each model are extracted and then the location of the precipitation center in the forecast is improved significantly.Therefore,by combining the appropriate anomalous features with multi-model fusion,an integrated improvement of the forecast of the rainfall intensity and location is achieved.Overall,this study is a novel exploration to improve the refined forecasting of heavy precipitation with extreme intensity and high variability,and provides a reference for the deep fusion of physics and artificial intelligence methods to improve intense rain forecast.展开更多
The characteristics and dynamics associated with the distribution, intensity, and triggering factors of local severe precipitation in Zhejiang Province induced by Super Typhoon Soudelor(2015) were investigated using m...The characteristics and dynamics associated with the distribution, intensity, and triggering factors of local severe precipitation in Zhejiang Province induced by Super Typhoon Soudelor(2015) were investigated using mesoscale surface observations, radar reflectivity, satellite nephograms, and the final(FNL) analyses of the Global Forecasting System(GFS) of the National Center for Environmental Prediction(NCEP). The rainfall processes during Soudelor's landfall and translation over East China could be separated into four stages based on rainfall characteristics such as distribution, intensity, and corresponding dynamics. The relatively less precipitation in the first stage resulted from interaction between the easterly wind to the north flank of this tropical cyclone(TC) and the coastal topography along the southeast of Zhejiang Province, China. With landfall of the TC in East China during the second stage, precipitation maxima occurred because of interaction between the TC's principal rainbands and the local topography from northeastern Fujian Province to southwestern Zhejiang Province. The distribution of precipitation presented significant asymmetric features in the third stage with maximal rainfall bands in the northeast quadrant of the TC when Soudelor's track turned from westward to northward as the TC decayed rapidly. Finally, during the northward to northeastward translation of the TC in the fourth stage, the interaction between a mid-latitude weather system and the northern part of the TC resulted in transfer of the maximum rainfall from the north of Zhejiang Province to the north of Jiangsu Province,which represented the end of rainfall in Zhejiang Province. Further quantitative calculations of the rainfall rate induced by the interaction between local topography and TC circulation(defined as "orographic effects") in the context of a one-dimensional simplified model showed that orographic effects were the primary factor determining the intensity of precipitation in this case,and accounted for over 50% of the total precipitation. The asymmetric distribution of the TC's rainbands was closely related to the asymmetric distribution of moisture resulted from changes of the TC's structure, and led to asymmetric distribution of local intense precipitation induced by Soudelor. Based on analysis of this TC, it could be concluded that local severe rainfall in the coastal regions of East China is closely related to changes of TC structure and intensity, as well as the outer rainbands. In addition, precipitation intensity and duration will increase correspondingly because of the complex interactions between the TC and local topography, and the particular TC track along large-scale steering flow. The results of this study may be useful for the understanding, prediction, and warning of disasters induced by local extreme rainfall caused by TCs, especially for facilitating forecasting and warning of flooding and mudslides associated with torrential rain caused by interactions between landfalling TCs and coastal topography.展开更多
基金"The pre-warning and prediction system for unexpected geological calamities in Zhejiangprovince and demonstration of its application - A "provincial key project from the science and technologybureau of Zhejianga key project "the study on forecasting system for heavy rains in Zhejiang province"
基金funded by the National Key Research and Devel-opment Program of China[grant number 2017YFC1501405]the National Natural Science Foundation of China[grant numbers 41975180,41705119,and 41575131]the National Center of Meteorology,Abu Dhabi,AE(UAE Research Program for Rain Enhancement Science)。
基金supported by the National Key R&D Project(Grant No.2021YFC3000903)the National Natural Science Foundation of China(Grant Nos.42275013,42030611,42075002)+2 种基金the CMA Innovation Foundation(Grant No.CXFZ2023J001)the Open Grants of the State Key Laboratory of Severe Weather(Grant No.2023LASW-B05)the Key Foundation of Zhejiang Provincial Department of Science and Technology(Grant No.2022C03150)。
文摘As a natural disaster,extreme precipitation is among the most destructive and influential,but predicting its occurrence and evolution accurately is very challenging because of its rarity and uniqueness.Taking the example of the“21·7”extreme precipitation event(17–21 July 2021)in Henan Province,this study explores the potential of using physics-guided machine learning to improve the accuracy of forecasting the intensity and location of extreme precipitation.Three physics-guided ways of embedding physical features,fusing physical model forecasts and revised loss function are used,i.e.,(1)analyzing the anomalous circulation and thermodynamical factors,(2)analyzing the multi-model forecast bias and the associated underlying reasons for it,and(3)using professional forecasting knowledge to design the loss function,and the corresponding results are used as input for machine learning to improve the forecasting accuracy.The results indicate that by learning the relationship between anomalous physical features and heavy precipitation,the forecasting of precipitation intensity is improved significantly,but the location is rarely adjusted and more false alarms appear.Possible reasons for this are as follows.The anomalous features used here mainly contain information about large-scale systems and factors which are consistent with the model precipitation deviation;moreover,the samples of extreme precipitation are sparse and so the algorithm used here is simple.However,by combining“good and different”multi models with machine learning,the advantages of each model are extracted and then the location of the precipitation center in the forecast is improved significantly.Therefore,by combining the appropriate anomalous features with multi-model fusion,an integrated improvement of the forecast of the rainfall intensity and location is achieved.Overall,this study is a novel exploration to improve the refined forecasting of heavy precipitation with extreme intensity and high variability,and provides a reference for the deep fusion of physics and artificial intelligence methods to improve intense rain forecast.
基金supported by the Huadong Regional Meteorological Science and Technology Innovation Fund Collaborative Project (Grant No. QYHZ201404)the Development of Social Welfare Project of Zhejiang Province (Grant No. 2013C33037)+2 种基金the Science Foundation of Zhejiang Province (Grant No. LY18D050001)United States Office of Naval Research Project (Grant No. N000140910526)the Development of Social Welfare Key Project of Zhejiang Province (Grant No. 2017C03035)
文摘The characteristics and dynamics associated with the distribution, intensity, and triggering factors of local severe precipitation in Zhejiang Province induced by Super Typhoon Soudelor(2015) were investigated using mesoscale surface observations, radar reflectivity, satellite nephograms, and the final(FNL) analyses of the Global Forecasting System(GFS) of the National Center for Environmental Prediction(NCEP). The rainfall processes during Soudelor's landfall and translation over East China could be separated into four stages based on rainfall characteristics such as distribution, intensity, and corresponding dynamics. The relatively less precipitation in the first stage resulted from interaction between the easterly wind to the north flank of this tropical cyclone(TC) and the coastal topography along the southeast of Zhejiang Province, China. With landfall of the TC in East China during the second stage, precipitation maxima occurred because of interaction between the TC's principal rainbands and the local topography from northeastern Fujian Province to southwestern Zhejiang Province. The distribution of precipitation presented significant asymmetric features in the third stage with maximal rainfall bands in the northeast quadrant of the TC when Soudelor's track turned from westward to northward as the TC decayed rapidly. Finally, during the northward to northeastward translation of the TC in the fourth stage, the interaction between a mid-latitude weather system and the northern part of the TC resulted in transfer of the maximum rainfall from the north of Zhejiang Province to the north of Jiangsu Province,which represented the end of rainfall in Zhejiang Province. Further quantitative calculations of the rainfall rate induced by the interaction between local topography and TC circulation(defined as "orographic effects") in the context of a one-dimensional simplified model showed that orographic effects were the primary factor determining the intensity of precipitation in this case,and accounted for over 50% of the total precipitation. The asymmetric distribution of the TC's rainbands was closely related to the asymmetric distribution of moisture resulted from changes of the TC's structure, and led to asymmetric distribution of local intense precipitation induced by Soudelor. Based on analysis of this TC, it could be concluded that local severe rainfall in the coastal regions of East China is closely related to changes of TC structure and intensity, as well as the outer rainbands. In addition, precipitation intensity and duration will increase correspondingly because of the complex interactions between the TC and local topography, and the particular TC track along large-scale steering flow. The results of this study may be useful for the understanding, prediction, and warning of disasters induced by local extreme rainfall caused by TCs, especially for facilitating forecasting and warning of flooding and mudslides associated with torrential rain caused by interactions between landfalling TCs and coastal topography.
