Skip to main content
SpringerLink
Log in

High temporal resolution rainfall–runoff modeling using long-short-term-memory (LSTM) networks

  • Original Article
  • Published:
Neural Computing and Applications Aims and scope Submit manuscript

Abstract

Accurate and efficient models for rainfall–runoff (RR) simulations are crucial for flood risk management. Most rainfall models in use today are process-driven; i.e., they solve either simplified empirical formulas or some variation of the St. Venant (shallow water) equations. With the development of machine-learning techniques, we may now be able to emulate rainfall models using, for example, neural networks. In this study, a data-driven RR model using a sequence-to-sequence long-short-term-memory (LSTM) network was constructed. The model was tested for a watershed in Houston, TX, known for severe flood events. The LSTM network’s capability in learning long-term dependencies between the input and output of the network allowed modeling RR with high resolution in time (15 min). Using 10-year precipitation from 153 rainfall gages and river channel discharge data (more than 5.3 million data points), and by designing several numerical tests, the developed model performance in predicting river discharge was tested. The model results were also compared with the output of a process-driven model gridded surface subsurface hydrologic analysis (GSSHA). Moreover, physical consistency of the LSTM model was explored. The model results showed that the LSTM model was able to efficiently predict discharge and achieve good model performance. When compared to GSSHA, the data-driven model was more efficient and robust in terms of prediction and calibration. Interestingly, the performance of the LSTM model improved (test Nash–Sutcliffe model efficiency from 0.666 to 0.942) when a selected subset of rainfall gages based on the model performance, were used as input instead of all rainfall gages.

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

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7

Similar content being viewed by others

References

  1. Anctil F, Lauzon N, Andréassian V, Oudin L, Perrin C (2006) Improvement of rainfall–runoff forecasts through mean areal rainfall optimization. J Hydrol 328(3–4):717–725

    Article  Google Scholar 

  2. Bedient PB, Hoblit BC, Gladwell DC, Vieux BE (2000) Nexrad radar for flood prediction in houston. J Hydrol Eng 5(3):269–277

    Article  Google Scholar 

  3. Bengio Y, Simard P, Frasconi P et al (1994) Learning long-term dependencies with gradient descent is difficult. IEEE Trans Neural Netw 5(2):157–166

    Article  Google Scholar 

  4. Borah DK, Ahmadisharaf E, Padmanabhan G, Imen S, Mohamoud YM (2018) Watershed models for development and implementation of total maximum daily loads. PhD thesis, American Society of Civil Engineers

  5. Brays Bayou (2019). https://www.hcfcd.org/projects-studies/brays-bayou/. Accessed 7 Mar 2019

  6. Breiman L (2001) Random forests. Mach Learn 45(1):5–32. https://doi.org/10.1023/A:1010933404324

    Article  MATH  Google Scholar 

  7. Brunner GW (1997) HEC-RAS (river analysis system). In: North American water and environment congress and destructive water. ASCE, New York, pp 3782–3787

  8. Daniel EB, Camp JV, LeBoeuf EJ, Penrod JR, Abkowitz MD, Dobbins JP (2011) Watershed modeling using GIS technology: a critical review. J Spat Hydrol 10(2):13–28

    Google Scholar 

  9. Dawson CW, Wilby R (1998) An artificial neural network approach to rainfall–runoff modelling. Hydrol Sci J 43(1):47–66. https://doi.org/10.1080/02626669809492102

    Article  Google Scholar 

  10. Doshi-Velez F, Kim B (2017) Towards a rigorous science of interpretable machine learning. Preprint arXiv:1702.08608

  11. Downer CW, Ogden FL (2004) GSSHA: model to simulate diverse stream flow producing processes. J Hydrol Eng 9(3):161–174. https://doi.org/10.1061/(ASCE)1084-0699(2004)9:3(161)

    Article  Google Scholar 

  12. Downer CW, Ogden FL (2006) Gridded surface subsurface hydrologic analysis (GSSHA) user’s manual; version 1.43 for watershed modeling system 6.1. Technical report, Engineer Research and Development Center, Vicksburg, MS, Coastal and Hydraulics Laboratory

  13. Duan Q, Sorooshian S, Gupta V (1992) Effective and efficient global optimization for conceptual rainfall–runoff models. Water Resour Res 28(4):1015–1031

    Article  Google Scholar 

  14. Farabet C, Couprie C, Najman L, LeCun Y (2013) Learning hierarchical features for scene labeling. IEEE Trans Pattern Anal Mach Intell 35(8):1915–1929

    Article  Google Scholar 

  15. Faur C, Cougnaud A, Dreyfus G, Le Cloirec P (2008) Modelling the breakthrough of activated carbon filters by pesticides in surface waters with static and recurrent neural networks. Chem Eng J 145(1):7–15

    Article  Google Scholar 

  16. Fry T, Maxwell R (2018) Using a distributed hydrologic model to improve the green infrastructure parameterization used in a lumped model. Water 10(12):1756

    Article  Google Scholar 

  17. Furl C, Ghebreyesus D, Sharif H (2018) Assessment of the performance of satellite-based precipitation products for flood events across diverse spatial scales using gssha modeling system. Geosciences 8(6):191

    Article  Google Scholar 

  18. Gers FA, Schmidhuber J, Cummins F (1999) Learning to forget: continual prediction with LSTM. In: 9th international conference on artificial neural networks, pp 850–855

  19. Harris County Flood Warning System (2019). https://www.harriscountyfws.org/. Accessed 7 Mar 2019

  20. Hettiarachchi P, Hall M, Minns A (2005) The extrapolation of artificial neural networks for the modelling of rainfall–runoff relationships. J Hydroinform 7(4):291–296

    Article  Google Scholar 

  21. Hochreiter S (1998) The vanishing gradient problem during learning recurrent neural nets and problem solutions. Int J Uncert Fuzziness Knowl Based Syst 6(02):107–116

    Article  Google Scholar 

  22. Hsu KI, Gupta HV, Sorooshian S (1995) Artificial neural network modeling of the rainfall–runoff process. Water Resour Res 31(10):2517–2530

    Article  Google Scholar 

  23. Inc ST (2002) Channel interconnected and model pond routing version 3.10 (2002). http://www.streamnologies.com/products/icpr/icpr.htm. Accessed 7 Mar 2019

  24. Johannet A, Borrell Estupina V, Pistre S et al (2015) Neural networks for Karst groundwater management: case of the Lez spring (Southern France). Environ Earth Sci 74(12):7617–7632

    Article  Google Scholar 

  25. Kan G, Yao C, Li Q, Li Z, Yu Z, Liu Z, Ding L, He X, Liang K (2015) Improving event-based rainfall–runoff simulation using an ensemble artificial neural network based hybrid data-driven model. Stoch Env Res Risk Assess 29(5):1345–1370. https://doi.org/10.1007/s00477-015-1040-6

    Article  Google Scholar 

  26. Kao IF, Zhou Y, Chang LC, Chang FJ (2020) Exploring a long short-term memory based encoder–decoder framework for multi-step-ahead flood forecasting. J Hydrol 2020:124631

    Article  Google Scholar 

  27. Kingma DP, Ba J (2014) Adam: a method for stochastic optimization. Preprint arXiv:1412.6980

  28. Kratzert F, Klotz D, Brenner C, Schulz K, Herrnegger M (2018) Rainfall–runoff modelling using long-short-term-memory (LSTM) networks. Hydrol Earth Syst Sci Discuss. https://doi.org/10.17605/OSF.IO/QV5JZ

    Article  Google Scholar 

  29. Le XH, Ho HV, Lee G, Jung S (2019) Application of long short-term memory (LSTM) neural network for flood forecasting. Water 11(7):1387

    Article  Google Scholar 

  30. LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521(7553):436

    Article  Google Scholar 

  31. Lindström G, Johansson B, Persson M, Gardelin M, Bergström S (1997) Development and test of the distributed HBV-96 hydrological model. J Hydrol 201(1–4):272–288

    Article  Google Scholar 

  32. Lipton ZC, Berkowitz J, Elkan C (2015) A critical review of recurrent neural networks for sequence learning. Preprint arXiv:1506.00019

  33. Lü H, Hou T, Horton R, Zhu Y, Chen X, Jia Y, Wang W, Fu X (2013) The streamflow estimation using the Xinanjiang rainfall–runoff model and dual state-parameter estimation method. J Hydrol 480:102–114. https://doi.org/10.1016/j.jhydrol.2012.12.011

    Article  Google Scholar 

  34. Mhammedi Z, Hellicar A, Rahman A, Kasfi K, Smethurst P (2016) Recurrent neural networks for one day ahead prediction of stream flow. In: Proceedings of the workshop on time series analytics and applications—TSAA’16, pp. 25–31. https://doi.org/10.1145/3014340.3014345. http://dl.acm.org/citation.cfm?doid=3014340.3014345

  35. Nayak P, Sudheer K, Ramasastri K (2005) Fuzzy computing based rainfall–runoff model for real time flood forecasting. Hydrol Process Int J 19(4):955–968

    Article  Google Scholar 

  36. Nerrand O, Roussel-Ragot P, Personnaz L, Dreyfus G, Marcos S (1993) Neural networks and nonlinear adaptive filtering: unifying concepts and new algorithms. Neural Comput 5(2):165–199

    Article  Google Scholar 

  37. NOAA National Centers for Environmental Information (NCEI) (2018) US billion-dollar weather and climate disasters. https://www.ncdc.noaa.gov/billions/. Accessed 23 May 2018

  38. Pappenberger F, Beven KJ, Hunter N, Bates P, Gouweleeuw B, Thielen J, De Roo A (2005) Cascading model uncertainty from medium range weather forecasts (10 days) through a rainfall–runoff model to flood inundation predictions within the European flood forecasting system (EFFS). Hydrol Earth Syst Sci Dis 9(4):381–393

    Article  Google Scholar 

  39. Paszke A, Chanan G, Lin Z, Gross S, Yang E, Antiga L, Devito Z (2017) Automatic differentiation in PyTorch. In: 31st conference on neural information processing systems (Nips), pp 1–4. https://doi.org/10.1017/CBO9781107707221.009

  40. Silva-Araya W, Santiago-Collazo F, Gonzalez-Lopez J, Maldonado-Maldonado J (2018) Dynamic modeling of surface runoff and storm surge during hurricane and tropical storm events. Hydrology 5(1):13

    Article  Google Scholar 

  41. Somu N, MR GR, Ramamritham K (2020) A hybrid model for building energy consumption forecasting using long short term memory networks. Appl Energy 261:114131

    Article  Google Scholar 

  42. Srinivasulu S, Jain A (2006) A comparative analysis of training methods for artificial neural network rainfall–runoff models. Appl Soft Comput 6(3):295–306

    Article  Google Scholar 

  43. Sudheer KP, Gosain AK, Ramasastri KS (2002) A data-driven algorithm for constructing artificial neural network rainfall–runoff models. Hydrol Process 16(6):1325–1330. https://doi.org/10.1002/hyp.554

    Article  Google Scholar 

  44. Sutskever I, Vinyals O, Le Q V (2014) Sequence to Sequence Learning with Neural Networks. In: Ghahramani Z, Welling M, Cortes C, et al (eds) Advances in Neural Information Processing Systems 27. Curran Associates, Inc., pp 3104–3112

  45. Talchabhadel R, Shakya NM, Dahal V, Eslamian S (2015) Rainfall–runoff modelling for flood forecasting (a case study on west rapti watershed). J Flood Eng 6(1):53–61

    Google Scholar 

  46. Taver V, Johannet A, Borrell-Estupina V, Pistre S (2015) Feed-forward vs. recurrent neural network models for non-stationarity modelling using data assimilation and adaptivity. Hydrol Sci J 60(7–8):1242–1265

    Article  Google Scholar 

  47. Torres JM, Bass B, Irza N, Fang Z, Proft J, Dawson C, Kiani M, Bedient P (2015) Characterizing the hydraulic interactions of hurricane storm surge and rainfall–runoff for the Houston–Galveston region. Coast Eng 106:7–19. https://doi.org/10.1016/j.coastaleng.2015.09.004

    Article  Google Scholar 

  48. van Gerven M, Bohte S (2018) Artificial neural networks as models of neural information processing. Frontiers Media, New York

    Book  Google Scholar 

  49. Wu Y, Liu Z, Xu W, Feng J, Palaiahnakote S, Lu T (2018) Context-aware attention LSTM network for flood prediction. In: 2018 24th international conference on pattern recognition (ICPR), pp 1301–1306

  50. Young CC, Liu WC (2015) Prediction and modelling of rainfall–runoff during typhoon events using a physically-based and artificial neural network hybrid model. Hydrol Sci J 60(12):2102–2116

    Article  Google Scholar 

  51. Yuan X, Chen C, Lei X, Yuan Y, Adnan RM (2018) Monthly runoff forecasting based on LSTM–ALO model. Stoch Environ Res Risk Assess 32(8):2199–2212

    Article  Google Scholar 

  52. Zarzar CM, Hosseiny H, Siddique R, Gomez M, Smith V, Mejia A, Dyer J (2018) A hydraulic multimodel ensemble framework for visualizing flood inundation uncertainty. JAWRA J Am Water Resour Assoc 54(4):807–819

    Article  Google Scholar 

  53. Zhijia L, Lili W, Hongjun B, Yu S, Zhongbo Y (2008) Rainfall–runoff simulation and flood forecasting for Huaihe Basin. Water Sci Eng 1(3):24–35

    Google Scholar 

Download references

Acknowledgements

This research was funded by the Severe Storm Prediction, Education and Evacuation from Disasters Center (Grant No. R09252) and the National Oceanic and Atmospheric Administration (Grant No. NA18NOS0120158). Their support is gratefully acknowledged.

Author information

Authors and Affiliations

  1. Oden Institute for Computational Engineering and Sciences, University of Texas at Austin, Austin, 78712, TX, USA

    Wei Li, Amin Kiaghadi & Clint Dawson

Authors
  1. Wei Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Amin Kiaghadi
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Clint Dawson
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Amin Kiaghadi.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

See Figs. 8, 9 and Table 4.

Fig. 8
figure 8

Land use of study area

Full size image
Table 4 Mannings n values for various land covers to use for GSSHA simulation
Full size table
Fig. 9
figure 9figure 9

Comparison of observation, 10-gages LSTM, and 153-gages LSTM. A close look at the lower discharge region would reveal the unphyscial oscillation of 153-gages prediction, indicating overfitting

Full size image

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, W., Kiaghadi, A. & Dawson, C. High temporal resolution rainfall–runoff modeling using long-short-term-memory (LSTM) networks. Neural Comput & Applic 33, 1261–1278 (2021). https://doi.org/10.1007/s00521-020-05010-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00521-020-05010-6

Keywords

Search

Navigation