A Review of AI Methods for the Prediction of High-Flow Extremal Hydrology

Hamitouche, Mohamed; Molina, José-Luis

doi:10.1007/s11269-022-03240-y

Cited by 12 publications

(5 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are many review studies on the different hydrological sub-areas in the literature, such as in Ecohydrology [9][10][11][12][13][14], Hydropower [15], Hydroinformatics [16][17][18][19][20][21][22][23][24][25], Gral Hydrology [26][27][28][29][30], Hydrology and Climate Change [31][32][33][34][35], Stochastic Hydrology [36][37][38][39], Forecasting and Uncertainty [40][41][42][43][44][45][46][47][48][49], Nival Hydrology [50][51][52][53], GIS Hydrology and Remote Sensing [54][55][56][57]…”

Section: Introductionmentioning

confidence: 99%

“…In contrast to these articles, "A review of AI methods for the prediction of highflow extreme hydrology" [19] presents a review with a qualitative approach to artificial intelligence methods for the prediction of extreme hydrological events, which is quite exhaustive in terms of the description of the different techniques. However, similar to previous studies, it provides a critical shortcoming in the methodology used for identifying and evaluating reference studies as it does not present a systematic protocol for the literature review.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Toward Systematic Literature Reviews in Hydrological Sciences

De León Pérez,

Acosta Vega,

Salazar Galán

et al. 2024

Water

View full text Add to dashboard Cite

Systematic literature reviews can provide an objective global overview of background research on hydrological questions. This study presents a methodology to ensure rigor, traceability, and replicability in evaluating state-of-the-art hydrological topics. The proposed methodology involved systematic, objective, and explicit steps. Strategies with defined selection criteria were used to search the relevant literature comprehensively and accurately on hydrological uncertainty as a case study. The most pertinent documents were filtered to build a critical state-of-the-art synthesis for evaluating their quality and relevance. This methodology allows systematic literature analysis to provide an objective summary of the evidence, with a structured procedure for state-of-the-art reviews, which promotes transparency in the search, unlike typical review papers that lack reproducible methodologies that may lead to obtaining a potentially subjective reference selection. Consequently, the proposed methodology improves reference reliability and study reproducibility while generalizing the methodology. This procedure has proven effective and practical for building state-of-the-art research in the study case (State-of-the-art Uncertainty in Hydrological forecasting from 2017 to 2023) and can be considered a relevant tool in hydrology and other STEM disciplines, providing a rigorous and transparent approach that enhances evidence quality and reliability.

show abstract

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

Toward Systematic Literature Reviews in Hydrological Sciences

De León Pérez,

Acosta Vega,

Salazar Galán

et al. 2024

Water

View full text Add to dashboard Cite

show abstract

“…ML algorithms can be classified into several categories, including decision trees and ensemble methods, artificial neural networks, support vectors machines, etc. [38]. The long short-term memory (LSTM) model has shown satisfactory forecasting performances in drought forecasting owing to its the long-term dependency problem [39,40].…”

Section: Introductionmentioning

confidence: 99%

Forecasting the Propagation from Meteorological to Hydrological and Agricultural Drought in the Huaihe River Basin with Machine Learning Methods

Hao,

Yan,

Chiang

2023

Remote Sensing

View full text Add to dashboard Cite

Revealing the mechanism of hydrological and agricultural drought has been challenging and vital in the environment under extreme weather and water resource shortages. To explore the evolution process from meteorological to hydrological and agricultural drought further, multi-source remote sensing data, including the Gravity Recovery and Climate Experiment (GRACE) product, were collected in the Huaihe River basin of China during 2002–2020. Three machine learning methods, including long short-term memory neural network (LSTM), convolutional neural network (CNN), and categorical boosting (CatBoost), were constructed for hydrological and agricultural drought forecasting. The propagation time from meteorological drought to surface water storage and terrestrial water storage drought, evaluated by the standardized precipitation evapotranspiration index, was 8 and 11 months with Pearson correlation coefficients (R) of 0.68 and 0.48, respectively. Groundwater storage drought was correlated with evapotranspiration and vegetation growth with a 12-month lag time, respectively. In addition, vegetation growth was affected by the drought of soil moisture at depths ranging from 100 to 200 cm with an 8-month lag time with an R of −0.39. Although the forecasting performances of terrestrial water storage drought were better than those of groundwater storage drought and agricultural drought, CNN always performed better than LSTM and CatBoost models, with Nash–Sutclife efficiency values during testing ranging from 0.28 to 0.70, 0.26 to 0.33, and −0.10 to −0.40 for terrestrial water storage drought, groundwater storage drought, and agricultural drought at lead times of 0–3 months, respectively. Furthermore, splitting training and testing data at random significantly improved the performances of CNN and CatBoost methods for drought forecasting rather than in chronological order splitting for non-stationary data.

show abstract

“…Over the past two decades, data-driven approaches based on artificial intelligence (AI), have gained drastically increasing interest from hydrologists [3], due to their significant contribution to improving the accuracy, versus the failure stories, of the classical and conventional methods in terms of spatial scale, time scale, the amount of data needed, facilities, the inability to handle nonlinear and nonstationary hydrological processes, and even in terms of accuracy, viewing the complexity of the equations governing the hydrological cycle's mechanisms which often require simplifications and theoretical assumptions leading to considerable errors and uncertainties [4].…”

Section: Introductionmentioning

confidence: 99%

“…However, despite the application of several ML techniques available in the literature, the gradient boosting (GB) approach has not been widely applied to predict daily flows [7]. Also, for hydrological extremes, GB and random forest (RF) are more often explored for qualitative predictions rather than quantitative predictions [4].…”

Section: Introductionmentioning

confidence: 99%