Preprints
https://doi.org/10.5194/hess-2023-282
https://doi.org/10.5194/hess-2023-282
18 Dec 2023
 | 18 Dec 2023
Status: this preprint is currently under review for the journal HESS.

Closing the data gap: runoff prediction in fully ungauged settings using LSTM

Reyhaneh Hashemi, Pierre Javelle, Olivier Delestre, and Saman Razavi

Abstract. Prediction in ungauged basins (PUB), where flow measurements are unavailable, is a critical need in hydrology and has been a focal point of extensive research efforts in this field over the past two decades. From the perspective of deep learning, PUB can be viewed as a scenario where the generalization capability of a pretrained neural network is employed to make predictions on samples that were not included in its training data set. This paper adopts this view and conducts genuine PUB using long short-term memory (LSTM) networks. Unlike PUB approaches based on k-fold training-test technique, where an arbitrary catchment B is treated as gauged in k−1 rounds and as ungauged in one round, our approach ensures that the sample for which the PUB is conducted (the UNGAUGED sample) is completely independent from the sample used to previously train the LSTMs (the GAUGED sample). The UNGAUGED sample includes 379 catchments from five hydrological regimes: Uniform, Mediterranean, Oceanic, Nivo-Pluvial, and Nival. PUB predictions are conducted using LSTMs trained both at the regime level (using only gauged catchments within a specific regime) and at the national level (using all gauged catchments). For benchmarking the performance of LSTM in PUB, four regionalized variants of the GR4J conceptual model are considered: spatial proximity, multi-attribute proximity, regime proximity, and IQ-IP-Tmin proximity, where IQ, IP, and Tmin are the indices defining the five hydrological regimes. To align with the study's fully ungauged context, the IQ index, which is also an input feature for the LSTMs, and the regime classification, crucial for the REGIME LSTMs, are reproduced under ungauged conditions using a regime-informed neural network and an XGBoost multi-class classifier respectively. The results demonstrate the overall superior performance of NATIONAL LSTMs compared to REGIME LSTMs. Among the four regionalization approaches tested for GR4J, the IQ-IP-Tmin proximity approach proves to be the most effective when analyzed on a regime-wise basis. When comparing the best-performing LSTM with the best-performing GR4J model within each regime, LSTMs show superior performance in both the Nival and Mediterranean regimes.

Reyhaneh Hashemi, Pierre Javelle, Olivier Delestre, and Saman Razavi

Status: open (until 13 Mar 2024)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CC1: 'A curious interpretation', Daniel Klotz, 26 Jan 2024 reply
    • AC1: 'Reply on CC1', Reyhaneh Hashemi, 14 Feb 2024 reply
      • CC2: 'Reply on AC1', Daniel Klotz, 15 Feb 2024 reply
        • AC2: 'Reply on CC2', Reyhaneh Hashemi, 23 Feb 2024 reply
Reyhaneh Hashemi, Pierre Javelle, Olivier Delestre, and Saman Razavi
Reyhaneh Hashemi, Pierre Javelle, Olivier Delestre, and Saman Razavi

Viewed

Total article views: 458 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
351 92 15 458 6 7
  • HTML: 351
  • PDF: 92
  • XML: 15
  • Total: 458
  • BibTeX: 6
  • EndNote: 7
Views and downloads (calculated since 18 Dec 2023)
Cumulative views and downloads (calculated since 18 Dec 2023)

Viewed (geographical distribution)

Total article views: 442 (including HTML, PDF, and XML) Thereof 442 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 25 Feb 2024
Download
Short summary
Here, we have tackled the challenge of estimating water flow in areas without direct measurements, a crucial task in hydrology. We have applied deep learning techniques to a large sample of French catchments with various hydrological regimes. We have also compared our approach with traditional methods. We found that incorporating more data improves the accuracy of our deep learning predictions. Notably, our method outperforms traditional approaches in certain regimes, though not universally.