Skip to main content
SpringerLink
Log in

Improved forecasting via physics-guided machine learning as exemplified using “21·7” extreme rainfall event in Henan

  • Article
  • Published:
Science China Earth Sciences Aims and scope Submit manuscript

Abstract

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.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  • “58·7” Research Group. 1987. Meteorological analysis of the “58·7” heavy rain (in Chinese). Atmos Sci, 11: 100–107

    Google Scholar 

  • “75·8” Research Group. 1977a. Analysis of the causes of the extraordinary heavy rain on “75·8” (Part II) (in Chinese). Meteor, 3: 6–8

    Google Scholar 

  • “75·8” Research Group on Rainstorm. 1977b. Analysis of the causes of the extraordinary heavy rain on “75·8” (Part I) (in Chinese). Meteor, 3: 3–5

    Google Scholar 

  • “75·8” Research Group. 1979. Dynamic analysis of the “75·8” rainstorm in Henan (in Chinese). Acta Meteor Sin, 1: 45–55

    Google Scholar 

  • Agel L, Barlow M, Feldstein S B, GutowskiJr. W J. 2018. Identification of large-scale meteorological patterns associated with extreme precipitation in the US northeast. Clim Dyn, 50: 1819–1839

    Article  Google Scholar 

  • Buhe C L, Zhuge A R, Xie Z W, Gao Z T, Lin D W. 2022. Water vapor transportation features and key synoptic-scale systems of the “7·20” rainstorm in Henan province in 2021 (in Chinese). Chin J Atmos Sci, 46: 725–744

    Google Scholar 

  • Chattopadhyay A, Nabizadeh E, Hassanzadeh P. 2020. Analog forecasting of extreme-causing weather patterns using deep learning. J Adv Model Earth Syst, 12: 2019MS001958

    Article  Google Scholar 

  • Chen Y, Huang G, Wang Y, Tao W, Tian Q, Yang K, Zheng J, He H. 2023. Improving the heavy rainfall forecasting using a weighted deep learning model. Front Environ Sci, 11: 1116672

    Article  Google Scholar 

  • Collow A M G B, Bosilovich M G, Koster R D. 2016. Large-Scale influences on summertime extreme precipitation in the Northeastern United States. J Hydrometeorol, 17: 3045–3061

    Article  Google Scholar 

  • Cui X, Yang Y. 2022. Tracing and quantitative contribution analysis of the water vapor source for the “21·7” heavy rain in Henan (in Chinese). Chin J Atmos Sci, 46: 1543–1556

    Google Scholar 

  • Dai K, Zhu Y J, Bi G G. 2018. A review of statistical post-processing techniques for quantitative precipitation forecasting with ensemble models (in Chinese). Acta Meteor Sin, 76: 493–510

    Google Scholar 

  • Ding Y H. 2015. A study on the “75·8” heavy rainstorm in Henan: A review and comment (in Chinese). Acta Meteor Sin, 73: 411–424

    Google Scholar 

  • Ding Y H. 2014. The contribution of Mr. Tao Shiyan to the study of the conditions and mechanisms of rainstorms in China (in Chinese). Chin J Atmos Sci, 38: 616–

    Google Scholar 

  • Ding Y H. 2019. The development process and important progress of rainstorm theory in China (in Chinese). Heavy Rain Disast, 38: 395–406

    Google Scholar 

  • Dowdy A J, Catto J L. 2017. Extreme weather caused by concurrent cyclone, front and thunderstorm occurrences. Sci Rep, 7: 40359

    Article  CAS  Google Scholar 

  • Ebert E E. 2001. Ability of a poor man’s ensemble to predict the probability and distribution of precipitation. Mon Wea Rev, 129: 2461–2480

    Article  Google Scholar 

  • He L F, Chen S, Guo Y Q. 2020. Obtervation characterittict and synoptic mechanitmt of typhoon Lekima extreme rainfall in 2019 (in Chinese). J Appl Meteor Sci, 31: 513–526

    Google Scholar 

  • Hess P, Boers N. 2022. Deep learning for improving numerical weather prediction of heavy rainfall. J Adv Model Earth Syst, 14: e2021MS002765

    Article  Google Scholar 

  • Karpatne A, Watkins W, Read J, Kumar V. 2017. Physics-guided Neural Networks (PGNN): An application in lake temperature modeling. arXiv preprint, arXiv:1710.11431.

  • Kim T, Ho N, Yun S.2022. Benchmark dataset for precipitation forecasting by post-processing the numerical weather prediction. arXiv preprint, arXiv: 2206.15241.

  • Klocek S, Dong H, Dixon M, Kanengoni P, Kazmi N, Luferenko P, Sharma Z, Weyn J, Xiang S. 2021. MS-nowcasting: Operational precipitation nowcasting with convolutional LSTMs at microsoft weather. arXiv preprint, arXiv: 211109954.

  • Leinonen J, Hamann U, Germann U, Mecikalski J R. 2021. Nowcasting thunderstorm hazards using machine learning: The impact of data sources on performance. Natural Hazard Earth System Sci Discuss, 1–28

  • Lei Y S. 1981. Synthetic analysis of meridional persistent heavy rainstorm (in Chinese). Acta Meteor Sin, 39: 168–181

    Google Scholar 

  • Li Z C, Chen Y, Zhang F H, Sun J, Wang Y D, Fu J L, Yu C. 2015. Consideration by “75·8” extreme heavy rainfall event in Henan (in Chinese). Meteor Environm Sci, 38: 1–12

    Google Scholar 

  • Li H, Wang X M, Zhang X, Lu L Y, Xu W M. 2018. Analysis on extremity and characteristics of the 19 July 2016 severe torrential rain in the north of Henan province (in Chinese). Meteorol Bull, 44: 1136–1147

    Google Scholar 

  • Liang X D, Xia R, Bao X, Zhang X, Wang X M, Su A F, Fu J L, Li H R, W C, Y M, Hu H, Yang J, Liu L, Xu H X, Zhu C W, Liu B Q, Hu N, Wang X F, Chen G M, Chen L J. 2022. Overview and preliminary exploration of multiscale characteristics of the extreme heavy rain process in Henan in July 2021 (in Chinese). Chin Sci Bull, 67: 997–1011

    Google Scholar 

  • Liu D X, Qiao S J, Zhang Y Q, H N, Wei J L, Zhang R K, Huang P. 2019. Survey of data sampling methods for imbalance classification (in Chinese). J Chongqing Univ Technol (Nat Sci), 33: 102–112

    Google Scholar 

  • Luo Y L, Sun J S, Li Y, Xia R D, Du Y, Yang S, Zhang Y C, Chen J, Dai K, Shen X S. 2020. Science and forecasting of heavy rainfall in China: Research results of 40 years of reform and opening-up (in Chinese). Acta Meteor Sin, 78: 419–450

    Google Scholar 

  • Markus R, Gustau C, Bjorn S, Jung M, Denzler J, Carvalhais N, Prabhat N. 2019. Deep learning and process understanding for data-driven Earth system science. Nature, 566: 195–204

    Article  Google Scholar 

  • Meng Z, Zhang F, Luo D, Tan Z, Fang J, Sun J, Shen X, Zhang Y, Wang S, Han W, Zhao K, Zhu L, Hu Y, Xue H, Ma Y, Zhang L, Nie J, Zhou R, Li S, Liu H, Zhu Y. 2019. Review of Chinese atmospheric science research over the past 70 years: Synoptic meteorology. Sci China Earth Sci, 62: 1946–1991

    Article  Google Scholar 

  • Ng C P, Zhang Q, Li W, Zhou Z. 2022. Contribution of thunderstorms to changes in hourly extreme precipitation over China from 1980 to 2011. J Clim, 35: 4485–4498

    Article  Google Scholar 

  • Nie Y, Sun J. 2022. Moisture sources and transport for extreme precipitation over Henan in July 2021. Geophys Res Lett, 49: e2021GL097446

    Article  Google Scholar 

  • Pan H, Chen G. 2019. Diurnal variations of precipitation over North China regulated by the mountain-plains solenoid and boundary-layer inertial oscillation. Adv Atmos Sci, 36: 863–884

    Article  Google Scholar 

  • Pan Y, Gu J X, Yu J J, Shen Y, Shi C X, Zhou Z J. 2018. Experimental fusion method of high-resolution multi-source precipitation observation products in China (in Chinese). Acta Meteor Sin, 76: 755–766

    Google Scholar 

  • Pathak J, Subramanian S, Harrington P, Raja S, Chattopadhyay A, Mardani M. 2022. FourCastNet: A global data-driven high-resolution weather model using adaptive fourier neural operators. arXiv preprint, arXiv: 2202.11214v1.

  • Qi D R N, He L F, Wang X M, Chen S. 2022. “7·20” Fine observation of extreme rainstorm in Henan province and its thermodynamic causes (in Chinese). J Appl Meteorol Sci, 33: 1–15

    Google Scholar 

  • Ran L Q, Li S Y, Zhou Y S, Yang S, Ma S P. 2021. Observation and analysis of extreme rainstorm dynamic, thermal and water vapor characteristics of “7·20” extreme rainstorm in Henan province in 2021 (in Chinese). Chin J Atmos Sci, 45: 1366–1383

    Google Scholar 

  • Rao C H, Bi X X, Chen G H, Yu Z F. 2022. Offshore Typhoon had a Significant Impact on “21·7” Simulation of the effects of water vapor flux and dynamic and thermal conditions during extreme rainstorm in Henan province (in Chinese). Chin J Atmos Sci, 46: 1577–1594

    Google Scholar 

  • Shen X S, Wang J J, Li Z C, Chen D H, Gong J D. 2020. Independent innovation and development of Numerical Weather Prediction in China (in Chinese). Acta Meteor Sin, 78: 451–476

    Google Scholar 

  • Shi C, Zhou L H, Fan L, Zhang W X, Cao Y C, Wang C, Xiao F, Lv G L, Liang H. 2022. Analysis of observational data using Beidou/GNSS “21·7” extreme rainstorm process in Henan (in Chinese). Chin J Geophys, 65: 186–196

    Google Scholar 

  • Su A F, Lv X N, Cui L M, Li Z, Xi L, Li H. 2021. The Basic observational analysis of “7·20” extreme rainstorm in Zhengzhou (in Chinese). Heavy Rainstorm Hazards, 40: 445–454

    Google Scholar 

  • Sun J S. 2017. Differences and relationships between short-term heavy precipitation and heavy rainfall (in Chinese). Heavy Rainstorm Hazards, 36: 498–506

    Google Scholar 

  • Sun J S, Lei L, Yu B, Ding Q L. 2015. The fundamental features of the extreme severe rain events in the recent 10 years (in Chinese). Acta Meteor Sin, 73: 609–623

    Google Scholar 

  • Sun J, Zhang F Q. 2017. Daily extreme precipitation and trends over China. Sci China Earth Sci, 60: 2190–2203

    Article  Google Scholar 

  • Tao S Y. 1980. Heavy rain in China (in Chinese). Beijing: Science Press. 225

    Google Scholar 

  • Tewari A, Zollhöfer M, Garrido P, Bernard F, Kim H, Pérez P, Theobalt C, Kurth T, Hall D, Li Z, Azizzadenesheli K. 2018. Self-supervised multilevel face model learning for monocular reconstruction at over 250hz. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2549–255

  • Wang X K, Cui C G, Wang J Y, Yang H, Zhou W. 2022. “21·7” Diagnostic analysis of water vapor and jet stream characteristics of heavy rainstorm in Henan (in Chinese). Chin J Atmos Sci, 48: 533–544

    Google Scholar 

  • Wu J S, Jin L, Wang L Z. 2006. Research on genetic algorithm evolutionary design of BP neural network weather forecasting modeling (in Chinese). J Trop Meteorol, 22: 411–416

    Google Scholar 

  • Xie Y, Franz E, Chu M, Thuerey N. 2018. TempoGAN: A temporally coherent, volumetric GAN for super-resolution fuid fow. ACM Trans Graph, 37: 1–15

    Article  Google Scholar 

  • Xu J, Li R, Zhang Q, Chen Y, Liang X, Gu X. 2022. Extreme large-scale atmospheric circulation associated with the “21·7” Henan flood. Sci China Earth Sci, 65: 1847–1860

    Article  CAS  Google Scholar 

  • Yang Z P. 2015. Research on imbalanced data classification method based on neural networks (in Chinese). Doctoral dissertation. Shanghai: East China University of Science and Technology

    Google Scholar 

  • Yang S N, Duan Y H. 2020. Extremity analysis on the precipitation and environmental field of typhoon Rumbia (1818) (in Chinese). J Appl Meteorol Sci, 31: 290–302

    Google Scholar 

  • Yao X P, Li R Y. 2023. Progress in research of the July 2021 extreme precipitation event in Henan province China (in Chinese). Acta Meteor Sin, 81: 853–865

    Google Scholar 

  • Yin J, Gu H, Liang X, Yu M, Sun J, Xie Y, Li F, Wu C. 2022. A possible dynamic mechanism for rapid production of the extreme hourly rainfall in Zhengzhou City on 20 July 2021. J Meteorol Res, 36: 6–25

    Article  Google Scholar 

  • Yin S, Li W, Chen D, Jeong J H, Guo W. 2011. Diurnal variations of summer precipitation in the Beijing area and the possible effect of topography and urbanization. Adv Atmos Sci, 28: 725–734

    Article  Google Scholar 

  • Zhang X, Wang X M, Li H, Lv L Y. 2020. Construction of extreme rainstorm index based on environmental parameters and its application (in Chinese). Meteorol Bull, 46: 898–912

    Google Scholar 

  • Zhang X, Yang H, Wang X M, Shen L, Wang D, Li H. 2021. “21·7” Characteristics of extreme heavy precipitation and analysis of circulation anomalies in Henan (in Chinese). Chin J Atmos Sci, 44: 672–687

    Google Scholar 

  • Zhang X U, Yang Y, Chen B, Huang W. 2021. Operational precipitation forecast over China using the weather research and forecasting (WRF) model at a gray-zone resolution: Impact of convection parameterization. Weather Forecast, 36: 915–928

    Google Scholar 

  • Zhao P J, Shao Y X, Zhang X. 2019. Comparative analysis of the environmental field after landfall of similar path typhoons “Capricorn” and “Umbia” (in Chinese). Meteorol Environ Sci, 42: 17–28

    Google Scholar 

  • Zhao S X, Sun J H. 2019. Research progress on rainstorm mechanism and forecasting in China and reflections on related issues (in Chinese). Heavy Rainstorm Hazards, 38: 422–430

    Google Scholar 

  • Zhong Q, Sun Z, Chen H, Li J, Shen L. 2022. Multi model forecast biases of the diurnal variations of intense rainfall in the Beijing-Tianjin-Hebei region. Sci China Earth Sci, 65: 1490–1509

    Article  Google Scholar 

  • Zhong S, Zhuang Y, Hu S, Chen Z, Ding W, Feng Y, Deng T, Liu X, Zhang Y, Xu D. 2021. Verification and assessment of real-time forecasts of two extreme heavy rain events in Zhengzhou by operational NWP models. J Tropical Meteor, 27: 406–417

    Article  Google Scholar 

  • Zhou K H, Sun J S, Zheng Y G, Zhang Y T. 2022. Quantitative precipitation forecast experiment based on basic NWP variables using deep learning. Adv Atmos Sci, 39: 1472–1486

    Article  Google Scholar 

  • Zhou Z H. 2016. Machine Learning (in Chinese). Beijing: Tsinghua University Press

    Google Scholar 

  • Zhu K, Zhang C, Xue M, Yang N. 2022. Predictability and skill of convection-permitting ensemble forecast systems in predicting the record-breaking “21·7” extreme rainfall event in Henan Province, China. Sci China Earth Sci, 65: 1879–1902

    Article  Google Scholar 

  • Zhu Y, Luo Y. 2015. Precipitation calibration based on the frequency-matching method. Weather Forecast, 30: 1109–1124

    Article  Google Scholar 

Download references

Acknowledgements

This research was supported by the National Key R&D Project (Grant No. 2021YFC3000903), the National Natural Science Foundation of China (Grant Nos. 42275013, 42030611, 42075002), 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).

Author information

Authors and Affiliations

  1. China Meteorological Administration Training Center, Beijing, 100081, China

    Qi Zhong & Xiuping Yao

  2. State Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing, 100081, China

    Qi Zhong

  3. Zhejiang Meteorological Observatory, Hangzhou, 310017, China

    Zhicha Zhang

  4. Hebei Provincial Weather Modification Center, Shijiazhuang, 050020, China

    Shaoyu Hou

  5. Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, 100029, China

    Shenming Fu

  6. National Meteorological Center, Beijing, 100081, China

    Yong Cao

  7. School of Computer Science, Chengdu University of Information Technology, Chengdu, 610225, China

    Linguo Jing

Authors
  1. Qi Zhong
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Zhicha Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Xiuping Yao
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Shaoyu Hou
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Shenming Fu
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Yong Cao
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Linguo Jing
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Qi Zhong or Xiuping Yao.

Ethics declarations

Conflict of interest The authors declare that there are no conflicts of interest.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhong, Q., Zhang, Z., Yao, X. et al. Improved forecasting via physics-guided machine learning as exemplified using “21·7” extreme rainfall event in Henan. Sci. China Earth Sci. 67, 1652–1674 (2024). https://doi.org/10.1007/s11430-022-1302-1

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11430-022-1302-1

Keywords

Search

Navigation