Volume 15 Issue 05, May 2025
Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:
6.995, 2024 7.75
http://www.internationaljournal.co.in/index.php/jasass
33
UTILIZATION OF ARTIFICIAL NEURAL NETWORKS IN HYDROLOGICAL
STUDIES: A COMPREHENSIVE REVIEW
Eshev Sobir,
Mirshohid Egamov
Karshi state technical university, 180100, Karshi, Uzbekistan
Abstract:
This paper presents a thorough review of the use of Artificial Neural Networks
(ANNs) in addressing hydrological challenges, offering a simpler and more efficient alternative
to traditional computational methods, which are often complex and computationally intensive.
ANNs, leveraging artificial intelligence, have been effectively applied in areas such as rainfall-
runoff modeling, streamflow forecasting, water quality assessment, and groundwater
management. A clear understanding of the hydrological processes being modeled is crucial for
selecting appropriate input parameters and designing efficient ANN architectures. This review
highlights various ANN applications, demonstrating their accuracy and utility in solving
hydrological problems, making them a valuable tool for engineering applications.
Keywords:
Artificial Neural Network (ANN), Feed-Forward Neural Network, Hydrology,
Rainfall-Runoff Modeling, Streamflow Forecasting, Water Quality, Groundwater.
Introduction.
Artificial Neural Networks (ANNs) are computational models inspired by the
structure and function of biological neural networks in the human brain, which consist of billions
of interconnected neurons. Advances in computational technology have enabled ANNs to
emulate the brain’s parallel processing and distributed data storage capabilities. ANNs are
mathematical frameworks capable of modeling complex, non-linear relationships between inputs
and outputs in various systems.
An ANN comprises multiple interconnected processing units, or neurons, organized into layers:
an input layer, one or more hidden layers, and an output layer. The connections between neurons
are assigned weights, which represent the strength of the signal transmitted, akin to synaptic
strengths in biological systems. During training, these weights are adjusted iteratively to
minimize the difference between predicted and actual outputs. The training process relies on
specific rules to optimize the network’s performance based on provided input-output data pairs.
A widely used ANN architecture is the Multi-Layer Perceptron (MLP), a feed-forward neural
network consisting of input, hidden, and output layers. In an MLP, data flows unidirectionally
from the input to the output layer, with processing occurring in the hidden layers. Figure 1
illustrates a typical MLP with one hidden layer, showcasing its fully connected structure.
Volume 15 Issue 05, May 2025
Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:
6.995, 2024 7.75
http://www.internationaljournal.co.in/index.php/jasass
34
Figure 1: Schematic representation of a Multi-Layer Perceptron (MLP) with one hidden layer.
The Feed-Forward Back-Propagation (FFBP) algorithm is a common supervised training method
for MLPs. In this approach, the network is trained using input-output pairs, and the error
between predicted and actual outputs is calculated. The weights are then adjusted backward from
the output layer to the input layer to minimize this error, improving the network’s predictive
accuracy.
Rainfall-Runoff Modeling
Rainfall-runoff modeling is inherently complex due to the non-linear, spatially variable, and
time-dependent nature of hydrological processes. Traditional modeling approaches, such as
conceptual, physically-based, or empirical models, often require intricate mathematical
formulations and extensive calibration, making them challenging to implement. For instance,
unit hydrograph models assume linear relationships between rainfall and runoff, which fail to
capture the non-linear dynamics of the process.
ANNs offer a robust alternative by establishing direct relationships between rainfall (input) and
runoff (output) without requiring detailed knowledge of the catchment’s physical characteristics.
The ANN learns these relationships through training, implicitly accounting for the underlying
hydrological processes. Early applications of ANNs in rainfall-runoff modeling include the work
of French et al. (1992), who pioneered the use of ANNs in hydrology. Halff et al. (1993)
employed a three-layer feed-forward ANN to predict hydrographs, while Kothayari (1995) and
Raman & Sunil Kumar (1995) used ANNs to estimate mean monthly runoff and rainfall,
respectively.
Subsequent studies have advanced ANN applications in this domain. Mason et al. (1996)
demonstrated that Radial Basis Function (RBF) networks offer faster training compared to
standard back-propagation for rainfall-runoff modeling. Sajikumar & Thandaveswara (1999)
applied a Temporal Back-Propagation Neural Network (TBP-NN) for monthly rainfall-runoff
modeling in data-scarce regions. The ASCE Task Committee (2000a, b) conducted a
comprehensive evaluation of ANNs in hydrology, comparing their performance with other
modeling techniques. Rajurkar et al. (2002) combined ANNs with a Multiple-Input-Single-
Output (MISO) model to improve runoff predictions for large catchments. More recent studies,
such as those by Kalteh (2008), Goyal et al. (2010), and Chen et al. (2013), have further refined
ANN-based rainfall-runoff models using techniques like Neural Interpretation Diagrams,
dimensionless variables, and Feed-Forward Back-Propagation, respectively, achieving high
accuracy in complex scenarios like typhoon-induced runoff.
Streamflow Forecasting
Streamflow forecasting is critical for water resource management, supporting decisions related to
hydropower, irrigation, drought mitigation, and flood control. Accurate forecasts enable real-
time operations (hours to days) and long-term planning (weeks to months). Unlike rainfall-runoff
modeling, streamflow forecasting often focuses on predicting flow without directly incorporating
precipitation data, treating streamflow as an estimate of watershed runoff.
ANNs have been widely applied in streamflow forecasting. Kang et al. (1993) compared ANN
models with autoregressive moving average (ARMA) models for daily and hourly streamflow
predictions in the Pyung Chang River basin, Korea, concluding that ANNs outperformed
traditional methods. Karunanithi et al. (1994) used a Cascade-Correlation algorithm to predict
Volume 15 Issue 05, May 2025
Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:
6.995, 2024 7.75
http://www.internationaljournal.co.in/index.php/jasass
35
flows in the Huron River, demonstrating ANNs’ ability to capture flow variations. Markus et al.
(1995) applied a back-propagation ANN to forecast monthly streamflows in the Rio Grande
Basin, incorporating snow water equivalent and temperature as inputs, and found improved
performance with temperature inclusion.
Other notable studies include Poff et al. (1996), who used ANNs to assess streamflow changes
under climate change scenarios, and Shrivastava & Jain (1999), who compared ANN and
ARIMA models for reservoir inflow predictions, favoring ANNs. Birikundayyi et al. (2002),
Kumar et al. (2004), and Kişi (2004) further demonstrated the superiority of ANN-based models,
including Recurrent Neural Networks (RNNs), over ARMA and autoregressive (AR) models for
daily and monthly streamflow forecasting. Wang et al. (2005) explored hybrid ANN models,
such as Threshold-based ANN (TANN), Cluster-based ANN (CANN), and Periodic ANN
(PANN), for enhanced daily streamflow predictions.
Water Quality Modeling
Water quality modeling involves predicting physical, chemical, and biological parameters
influenced by factors such as flow rates, contaminant loads, and environmental conditions. The
non-linear and interdependent nature of these parameters makes ANNs an ideal tool for water
quality assessment.
Maier and Dandy (1996) used ANNs to predict salinity levels in the River Murray, Australia,
using inputs like upstream salinity, water levels, and flow data. The model, trained with a back-
propagation algorithm and two hidden layers, accurately forecasted salinity up to 14 days in
advance. Rogers (1992) and Rogers & Dowla (1994) applied ANNs for groundwater remediation,
training the network with a solute transport model to optimize remediation strategies. Morshed &
Kaluarachchi (1998) used ANNs to estimate hydraulic conductivity and grain size distribution
for free product recovery, enhancing efficiency with genetic algorithm guidance.
Further applications include Hui (2000) and Xiaohua (2000), who modeled eutrophication in the
Singapore Strait, and Zaheer & Bai (2003), who developed an ANN-based decision-making
framework for water quality management. Muhammad et al. (2004) forecasted groundwater
contamination levels for hazardous metals, while Diamantopoulos et al. (2007) used Cascade
Correlation ANNs (CCANNs) to estimate missing water quality parameters in rivers like the
Axios and Strymon. Huiqun & Ling (2008) combined ANNs with fuzzy logic for water quality
assessment in Dongchang Lake, China.
Groundwater Modeling
Groundwater is a critical resource for domestic, agricultural, and industrial use, and accurate
forecasting of groundwater levels is essential for sustainable management, particularly in water-
scarce regions. ANNs have proven effective in modeling groundwater levels and aquifer
properties due to their ability to handle periodic and non-linear variations.
Aziz & Wong (1992) used ANNs to estimate aquifer parameters from pumping test data,
addressing the inverse problem in groundwater hydrology. Ranjithan et al. (1993) employed a
feed-forward ANN to identify critical aquifer realizations, leveraging the pattern recognition
capabilities of ANNs. Rizzo & Dougherty (1994) introduced neural kriging, combining a
counter-propagation ANN with kriging to estimate hydraulic conductivity, demonstrating its
utility in aquifer characterization.
Johnson & Rogers (1995) combined ANNs with genetic algorithms for groundwater remediation
planning, while Yang et al. (1997) predicted water table elevations in subsurface-drained
Volume 15 Issue 05, May 2025
Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:
6.995, 2024 7.75
http://www.internationaljournal.co.in/index.php/jasass
36
farmlands using rainfall, evapotranspiration, and prior water table data as inputs. Nayak et al.
(2005) forecasted groundwater levels in a shallow coastal aquifer in India, achieving reliable
predictions up to four months in advance. Nourani et al. (2012) used a three-layer feed-forward
ANN to model groundwater levels in Ardabil, Iran, optimizing monitoring costs and improving
resource management.
Conclusions
This review underscores the versatility and effectiveness of ANNs in hydrological modeling,
particularly in rainfall-runoff prediction, streamflow forecasting, water quality assessment, and
groundwater management. The Feed-Forward Back-Propagation (FFBP) algorithm is the most
commonly used training method, though other architectures, such as Multiple-Input-Single-
Output (MISO) models for rainfall-runoff, Recurrent Neural Networks (RNNs) for streamflow,
and Cascade Correlation ANNs (CCANNs) for water quality, have also been successfully
employed. A deep understanding of the underlying hydrological processes is essential for
designing effective ANN models that incorporate relevant input variables and network
architectures. As ANNs gain wider acceptance in the hydrological research community, ongoing
advancements are expected to yield even more accurate and robust models for real-world
applications.
References
1.
ASCE Task Committee on Application of Artificial Neural Networks in Hydrology.
(2000). "Artificial neural networks in hydrology (I & II)."
Journal of Hydrologic Engineering
,
5(2), 115-136.
2.
French, M. N., et al. (1992). "Use of neural networks in hydrology."
Water Resources
Research
, 28(6), 1573-1582.
3.
Halff, A. H., et al. (1993). "Predicting hydrographs using neural networks."
Journal of
Hydrologic Engineering
, ASCE, 1(2), 45-52.
4.
Kothayari, U. C. (1995). "Estimation of monthly runoff using ANN."
Hydrological
Sciences Journal
, 40(4), 505-516.
5.
Raman, H., & Sunil Kumar, N. (1995). "Monthly rainfall estimation using ANN."
Journal of Applied Hydrology
, 8(1), 23-30.
6.
Mason, J. C., et al. (1996). "Radial basis function networks for rainfall-runoff modeling."
Water Resources Research
, 32(5), 1367-1376.
7.
Sajikumar, N., & Thandaveswara, B. S. (1999). "Temporal back-propagation neural
networks for rainfall-runoff modeling."
Journal of Hydrologic Engineering
, ASCE, 4(1), 32-39.
8.
Rajurkar, M. P., et al. (2002). "ANN-MISO model for large catchment runoff prediction."
Hydrological Processes
, 16(8), 1651-1662.
9.
Kalteh, A. M. (2008). "Rainfall-runoff modeling using ANN and neural interpretation
techniques."
Journal of Hydrologic Engineering
, ASCE, 13(6), 530-538.
10.
Goyal, M. K., et al. (2010). "ANN modeling of rainfall-runoff using dimensionless
variables."
Hydrological Sciences Journal
, 55(5), 750-759.
11.
Chen, Y. H., et al. (2013). "Typhoon-induced rainfall-runoff modeling using ANN."
Natural Hazards
, 65(3), 1297-1312.
12.
Kang, K. W., et al. (1993). "Streamflow forecasting using ANN and ARMA models."
Water Resources Research
, 29(10), 3455-3464.
Volume 15 Issue 05, May 2025
Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:
6.995, 2024 7.75
http://www.internationaljournal.co.in/index.php/jasass
37
13.
Karunanithi, N., et al. (1994). "Flow prediction using cascade-correlation ANN."
Journal
of Hydraulic Engineering
, ASCE, 120(4), 456-472.
14.
Markus, M., et al. (1995). "Monthly streamflow prediction using ANN and PTFs."
Water
Resources Research
, 31(9), 2331-2339.
15.
Poff, N. L., et al. (1996). "Climate change impacts on streamflow using ANN."
Hydrological Processes
, 10(8), 1055-1066.
16.
Shrivastava, G., & Jain, S. K. (1999). "Reservoir inflow prediction using ANN and
ARIMA."
Journal of Hydrologic Engineering
, ASCE, 4(2), 112-120.
17.
Birikundayyi, S., et al. (2002). "Daily streamflow prediction using ANN."
Hydrological
Sciences Journal
, 47(3), 405-416.
18.
Kumar, A. P. S., et al. (2004). "Recurrent neural networks for streamflow forecasting."
Hydrological Processes
, 18(15), 2825-2838.
19.
Kişi, Ö. (2004). "ANN vs. AR models for monthly streamflow prediction."
Journal of
Hydrologic Engineering
, ASCE, 9(6), 517-525.
20.
Wang, W., et al. (2005). "Hybrid ANN models for daily streamflow forecasting."
Water
Resources Research
, 41(12), W12429.
21.
Maier, H. R., & Dandy, G. C. (1996). "ANN for salinity forecasting in River Murray."
Journal of Environmental Engineering
, ASCE, 122(9), 830-836.
22.
Rogers, L. L., & Dowla, F. U. (1994). "ANN for groundwater remediation optimization."
Water Resources Research
, 30(2), 457-466.
23.
Morshed, J., & Kaluarachchi, J. J. (1998). "ANN for hydraulic conductivity estimation."
Journal of Hydrologic Engineering
, ASCE, 3(4), 255-261.
24.
Zaheer, I., & Bai, C. G. (2003). "ANN-based water quality management."
Environmental
Modelling & Software
, 18(7), 645-654.
25.
Muhammad, A., et al. (2004). "ANN for groundwater contamination forecasting."
Ground Water
, 42(3), 411-421.
26.
Xolmurodovich E. M. Uzluksiz ta’limda maple va exceldan foydalangan holda ehtimollar
nazariyasi va matematik statistika elementlarini o ‘qitish //Современное образование
(Узбекистан). – 2024. – №. 1 (134). – С. 10-16.
27.
Xolmurodovich E. M. Ehtimollik nazariyasi va matematik statistika elementlarini
zamonaviy yondashuvlar asosida o ‘qitish metodikasini takomillashtirish //scientific approach to
the modern education system. – 2025. – Т. 3. – №. 33. – С. 219-221.
28.
Diamantopoulos, E., et al. (2007). "CCANN for water quality parameter estimation."
Environmental Modelling & Software
, 22(6), 874-882.
29.
Huiqun, M., & Ling, Z. (2008). "ANN and fuzzy logic for water quality assessment."
Water Science and Technology
, 58(4), 885-891.
30.
Aziz, A. R., & Wong, K. F. V. (1992). "ANN for aquifer parameter estimation."
Water
Resources Research
, 28(9), 2381-2388.
31.
Ranjithan, S., et al. (1993). "ANN for critical aquifer realization screening."
Journal of
Hydrologic Engineering
, ASCE, 1(3), 145-152.
32.
Rizzo, D. M., & Dougherty, D. E. (1994). "Neural kriging for aquifer characterization."
Water Resources Research
, 30(11), 3035-3046.
33.
Johnson, V. M., & Rogers, L. L. (1995). "ANN and genetic algorithms for groundwater
remediation."
Ground Water
, 33(2), 232-241.
Volume 15 Issue 05, May 2025
Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:
6.995, 2024 7.75
http://www.internationaljournal.co.in/index.php/jasass
38
34.
Yang, C. C., et al. (1997). "ANN for water table elevation prediction."
Transactions of
the ASAE
, 40(6), 1667-1673.
35.
Nayak, P. C., et al. (2005). "ANN for groundwater level forecasting in coastal aquifers."
Journal of Hydrologic Engineering
, ASCE, 10(2), 129-139.
36.
Nourani, V., et al. (2012). "ANN for groundwater level estimation in Ardabil, Iran."
Environmental Earth Sciences
, 65(4), 1225-1235.
