Studi Literatur ML Untuk Prediksi Banjir
Studi Literatur ML Untuk Prediksi Banjir
Studi Literatur ML Untuk Prediksi Banjir
*Correspondence: [email protected]
Abstract: Floods are among the most destructive natural disasters, which are highly
contributed to risk reduction, policy suggestion, minimization of the loss of human life,
and reduction the property damage associated with floods. To mimic the complex
mathematical expressions of physical processes of floods, during the past two decades,
systems providing better performance and cost-effective solutions. Due to the vast
benefits and potential of ML, its popularity dramatically increased among hydrologists.
Researchers through introducing novel ML methods and hybridizing of the existing ones
aim at discovering more accurate and efficient prediction models. The main contribution
of this paper is to demonstrate the state of the art of ML models in flood prediction and
to give insight into the most suitable models. In this paper, the literature where ML
introduces the most promising prediction methods for both long-term and short-term
floods. Furthermore, the major trends in improving the quality of the flood prediction
well as climate scientists in choosing the proper ML method according to the prediction
task.
Keywords: flood prediction; flood forecasting; flash-flood model, big flood management;
hydrologic model; rainfall–runoff, hybrid & ensemble machine learning; artificial neural
networks (ANNs); support vector machines (SVM); natural hazards & disasters; adaptive
neuro-fuzzy inference system (ANFIS); decision trees (DT); internet of things (IoT);
random forest (RF); survey; classification and regression trees (CART), data science; deep
learning; big data; bagging, boosting, artificial intelligence (AI); soft computing; extreme
event management; time series prediction; multilayer perceptron (MLP); simulated
annealing (SA); multivariate adaptive regression splines (MARS), supervised learning
1. Introduction
Among the natural disasters, floods are the most destructive, causing massive
Governments, therefore, are under pressure to develop reliable and accurate maps of flood
risk areas and further plan for sustainable flood risk management focusing on prevention,
protection, and preparedness [1]. Flood prediction models are of significant importance
for hazard assessment and extreme event management. Robust and accurate prediction
analysis, and further evacuation modeling [2]. Thus, the importance of advanced systems
for short-term and long-term prediction for flood and other hydrological events is strongly
emphasized to alleviate damage [3]. However, the prediction of flood lead time and
condition. Therefore, today’s major flood prediction models are mainly data-specific and
involve various simplified assumptions [4]. Thus, to mimic the complex mathematical
expressions of physical processes and basin behavior, such models benefit from specific
techniques e.g., event-driven, empirical black box, lumped and distributed, stochastic,
Physically based models [6] were long used to predict hydrological events, such as
storm [7,8], rainfall/runoff [9,10], shallow water condition [11], hydraulic models of flow
[12,13], and further global circulation phenomena [14], including the coupled effects of
atmosphere, ocean, and floods [15]. Although physical models showed great capabilities
for predicting a diverse range of flooding scenarios, they often require various types of
development of physically based models often requires in-depth knowledge and expertise
numerous studies suggest that there is a gap in short-term prediction capability of physical
models (Costabile and Macchione [15]). For instance, on many occasions, such models
failed to predict properly [18]. Van den Honert and McAneney [18] documented the failure
and were not reliable due to systematic errors [20]. Nevertheless, major improvements in
physically based models of flood were recently reported through the hybridization of
In addition to numerical and physical models, data-driven models also have a long
tradition in flood modeling, which recently gained more popularity. Data-driven methods
moving average (ARMA) [24], multiple linear regression (MLR) [25], and autoregressive
integrated moving average (ARIMA) [26] are the most common flood frequency analysis
(FFA) methods for modeling flood prediction. FFA was among the early statistical
methods for predicting floods [27]. Regional flood frequency analyses (RFFA) [28], more
advanced versions, were reported to be more efficient when compared to physical models
data [29]. For instance, the climatology average method (CLIM) [28], empirical orthogonal
function (EOF) [30], multiple linear regressions (MLR), quantile regression techniques
(QRT) [31], and Bayesian forecasting models [32] are widely used for predicting major
floods. However, they were reported to be unsuitable for short-term prediction, and, in
this context, they need major improvement due to the lack of accuracy, complexity of the
usage, computation cost, and robustness of the method. Furthermore, for reliable long-
term prediction, at least, a decade of data from measurement gauges should be analyzed
for a meaningful forecast [32]. In the absence of such a dataset, however, FFA can be done
using hydrologic models of RFFA, e.g., MISBA [33] and Sacramento [34], as reliable
unavailable. In this context, distributed numerical models are used as an attractive solution
[35]. Nonetheless, they do not provide quantitative flood predictions, and their forecast
The drawbacks of the physically based and statistical models mentioned above
encourage the usage of advanced data-driven models, e.g., machine learning (ML). A
further reason for the popularity of such models is that they can numerically formulate the
flood nonlinearity, solely based on historical data without requiring knowledge about the
tools as they are quicker to develop with minimal inputs. ML is a field of artificial
intelligence (AI) used to induce regularities and patterns, providing easier implementation
with low computation cost, as well as fast training, validation, testing, and evaluation, with
high performance compared to physical models, and relatively less complexity [37]. The
continuous advancement of ML methods over the last two decades demonstrated their
for prediction with greater accuracy [42]. Ortiz-García et al. [43] described how ML
techniques could efficiently model complex hydrological systems such as floods. Many ML
algorithms, e.g., artificial neural networks (ANNs) [44], neuro-fuzzy [45,46], support
vector machine (SVM) [47], and support vector regression (SVR) [48,49], were reported as
effective for both short-term and long-term flood forecast. In addition, it was shown that
applications provided more robust and efficient models that can effectively learn complex
conclusion reported with regards to which models function better in certain applications.
In fact, the literature includes only a limited number of surveys on specific ML methods in
taken into consideration. The first is that they are as good as their training, whereby the
system learns the target task based on past data. If the data is scarce or does not cover
varieties of the task, their learning falls short, and hence, they cannot perform well when
they are put into work. Therefore, using robust data enrichment is essential through, e.g.,
retain the group characteristics [57], or recovering the missing variables using causally
The second aspect is the capability of each ML algorithm, which may vary across
different types of tasks. This can also be called a “generalization problem”, which indicates
how well the trained system can predict cases it was not trained for, i.e., whether it can
predict beyond the range of the training dataset. For example, some algorithms may
perform well for short-term predictions, but not for long-term predictions. These
characteristics of the algorithms need to be clarified with respect to the type and amount
of available training data, and the type of prediction task, e.g., water level and streamflow.
In this review, we look into examples of the use of various ML algorithms for various types
of tasks. At the abstract level, we decided to divide the target tasks into short-term and
structured ML methods as single methods and hybrid methods. Hybrid methods are those
Here, we should note that this paper surveys ML models used for predictions of
floods on sites where rain gauges or intelligent sensing systems used. Our goal was to
survey prediction models with various lead times to floods at a particular site. From this
perspective, spatial flood prediction was not involved in this study, as we did not study
This survey identifies the state of the art of ML methods for flood prediction where
peer-reviewed articles in top-level subject fields are reviewed. Among the articles
identified, through search queries using the search strategy, those including the
included in the review to identify the ML methods that perform better in particular
applications. Furthermore, to choose an article, four types of quality measure for each
article were considered, i.e., source normalized impact per paper (SNIP), CiteScore,
SCImago journal rank (SJR), and h-index. The papers were reviewed in terms of flood
variables, i.e., water level, river flood, soil moisture, rainfall–discharge, precipitation, river
inflow, peak flow, river flow, rainfall–runoff, flash flood, rainfall, streamflow, seasonal
stream flow, flood peak discharge, urban flood, plain flood, groundwater level, rainfall
stage, flood frequency analysis, flood quantiles, surge level, extreme flow, storm surge,
typhoon rainfall, and daily flows [59]. Among these key influencing flood resource
variables, rainfall and the spatial examination of the hydrologic cycle had the most
remarkable role in runoff and flood modeling [60]. This is the reason why quantitative
rainfall prediction, including avalanches, slush flow, and melting snow, is traditionally
used for flood prediction, especially in the prediction of flash floods or short-term flood
prediction [61]. However, rainfall prediction was shown to be inadequate for accurate
flood prediction. For instance, the prediction of streamflow in a long-term flood prediction
variables were considered in the [63]. Thus, the methodology of this literature review aims
to include the most effective flood resource variables in the search queries.
implement the complete list of search queries. Note that the ML methods for flood
prediction may vary significantly according to the application, dataset, and prediction
type. For instance, ML methods used for short-term water level prediction are significantly
different from those used for long-term streamflow prediction. Figure 1 represents the
organization of the search queries and further describes the survey search methodology.
The search query included three main search terms. The flood resource variables were
keywords for search queries mentioned above. Term 2 of search (<ML method1-m>)
a complete list of ML methods, from which the 25 most popular algorithms in engineering
applications were used as the keywords of this search. Term 3 included the four search
terms most often used in describing flood prediction, i.e., “prediction”, “estimation”,
“forecast”, or “analysis”. The total search resulted in 6596 articles. Among them, 180
original research papers were refined through our quality measure included in the survey.
T1 T2 T3
<V1> OR <Vn> <ML1> OR <MLn> <C1> OR <Cn>
AND AND
Q1-n
Section 3 presents the state of the art of ML in flood prediction. A technical description
on the ML method and a brief background in flood applications are provided. Section 4
presents the survey of ML methods used for short-term flood prediction. Section 5 presents
the survey of ML methods used for long-term flood prediction. Section 6 presents the
conclusions.
For creating the ML prediction model, the historical records of flood events, in
addition to real-time cumulative data of a number of rain gauges or other sensing devices
for various return periods, are often used. The sources of the dataset are traditionally
rainfall and water level, measured either by ground rain gauges, or relatively new remote-
real time. In addition, the high resolution of weather radar observations often provides a
more reliable dataset compared to rain gauges [63]. Thus, building a prediction model
based on a radar rainfall dataset was reported to provide higher accuracy in general [64].
Whether using a radar-based dataset or ground gauges to create a prediction model, the
historical dataset of hourly, daily, and/or monthly values is divided into individual sets to
construct and evaluate the learning models. To do so, the individual sets of data undergo
training, validation, verification, and testing. The principle behind the ML modeling
workflow and the strategy for flood modeling are described in detail in the literature
[48,65]. Figure 2 represents the basic flow for building an ML model. The major ML
algorithms applied to flood prediction include ANNs [66], neuro-fuzzy [67], adaptive
neuro-fuzzy inference systems (ANFIS) [68], support vector machines (SVM) [69], wavelet
neural networks (WNN) [70], and multilayer perceptron (MLP) [71]. In the following
presented.
Figure 2. Basic flow for building the machine learning (ML) model.
ANNs are efficient mathematical modeling systems with efficient parallel processing,
enabling them to mimic the biological neural network using inter-connected neuron units.
Among all ML methods, ANNs are the most popular learning algorithms, known to be
versatile and efficient in modeling complex flood processes with a high fault tolerance and
popular for modeling flood prediction since their first usage in the 1990s [73]. Instead of a
catchment’s physical characteristics, ANNs derive meaning from historical data. Thus,
ANNs are considered as reliable data-driven tools for constructing black-box models of
complex and nonlinear relationships of rainfall and flood [74], as well as river flow and
discharge forecasting [75]. Furthermore, a number of surveys (e.g., Reference [76]) suggest
ANN as one of the most suitable modeling techniques which provide an acceptable
[77,78] provided reviews on ANN applications in flood. ANNs were already successfully
used for numerous flood prediction applications, e.g., streamflow forecasting [79], river
flow [80,81], rainfall–runoff [82], precipitation–runoff modeling [83], water quality [55],
evaporation [56], river stage prediction [84], low-flow estimation [85], river flows [86], and
river time series [57]. Despite the advantages of ANNs, there are a number drawbacks
associated with using ANNs in flood modeling, e.g., network architecture, data handling,
and physical interpretation of the modeled system. A major drawback when using ANNs
is the relatively low accuracy, the urge to iterate parameter tuning, and the slow response
The feed-forward neural network (FFNN) [25] is a class of ANN, whereby the
network’s connections are not in cyclical form. FFNNs are the simplest type of ANN,
whereby information moves in a forward direction from input nodes to the hidden layer
and later to output nodes. On the other hand, a recurrent neural network (RNN) [91] is a
class of ANN, whereby the network’s connections form a time sequence for dynamic
temporal behavior. Furthermore, RNNs benefit from extra memory to analyze input
calculated using the propagation of the backward error gradient. In BP, there are more
phases in the learning cycle, using a function for activation to send signals to the other
nodes. Among various ANNs, the backpropagation ANN (BPNN) was identified as the
most powerful prediction tool suitable for flood time-series prediction [26]. Extreme
learning machine (ELM) [92] is an easy-to-use form of FFNN, with a single hidden layer.
Here, ELM was studied under the scope of ANN methods. ELM for flood prediction
recently became of interest for hydrologists and was used to model short-term streamflow
The vast majority of ANN models for flood prediction are often trained with a BPNN
[95]. While BPNNs are today widely used in this realm, the MLP—an advanced
representation of ANNs— recently gained popularity [96]. The MLP [97] is a class of FFNN
which utilizes the supervised learning of BP for training the network of interconnected
nodes of multiple layers. Simplicity, nonlinear activation, and a high number of layers are
characteristics of the MLP. Due to these characteristics, the model was widely used in flood
classes used in flood modeling, MLP models were reported to be more efficient with better
optimize [99]. Back-percolation learning algorithms are used to individually calculate the
propagation error in hidden network nodes for a more advanced modeling approach.
Here, it is worth mentioning that the MLP, more than any other variation of ANNs
(e.g., FFNN, BPNN, and FNN), gained popularity among hydrologists. Furthermore, due
to the vast number of case studies using the standard form of MLP, it diverged from
regular ANNs. In addition, the authors of articles in the realm of flood prediction using
the MLP refer to their models as MLP models. From this perspective, we decided to devote
The fuzzy logic of Zadeh [100] is a qualitative modeling scheme with a soft computing
technique using natural language. Fuzzy logic is a simplified mathematical model, which
works on incorporating expert knowledge into a fuzzy inference system (FIS). An FIS
further mimics human learning through an approximation function with less complexity,
which provides great potential for nonlinear modeling of extreme hydrological events
[101,102], particularly floods [103]. For instance, Reference [104] studied river level
forecasting using an FIS, as did Lohani et al. (2011) [4] for rainfall–runoff modeling for
hybrid of the BPNN and the widely used least-square error method [46]. The Takagi–
Sugeno (T–S) fuzzy modeling technique [4], which is created using neuro-fuzzy clustering,
based on the T–S FIS, first coined [67,77]. Today, ANFIS is known to be one of the most
reliable estimators for complex systems. ANFIS technology, through combining ANN and
fuzzy logic, provides higher capability for learning [101]. This hybrid ML method
corresponds to a set of advanced fuzzy rules suitable for modeling flood nonlinear
functions. An ANFIS works by applying neural learning rules for identifying and tuning
the parameters and structure of an FIS. Through ANN training, the ANFIS aims at catching
the missing fuzzy rules using the dataset [67]. Due to fast and easy implementation,
accurate learning, and strong generalization abilities, ANFIS became very popular in flood
modeling. The study of Lafdani et al. [60] further described its capability in modeling
short-term rainfall forecasts with high accuracy, using various types of streamflow,
rainfall, and precipitation data. Furthermore, the results of Shu and [67] showed easier
clustering algorithm, which led several rounds of random selection being avoided.
Wavelet transform (WT) [46] is a mathematical tool which can be used to extract
information from various data sources by analyzing local variations in time series [50]. In
transforms supports the reliable decomposition of an original time series to improve data
decomposes the original data into bands, leading to an improvement of flood prediction
lead times [106]. DWT decomposes the initial data set into individual resolution levels for
extracting better-quality data for model building. DWTs, due to their beneficial
characteristics, are widely used in flood time-series prediction. In flood modeling, DWTs
were widely applied in, e.g., rainfall–runoff [51[, daily streamflow [106], and reservoir
inflow [107]. Furthermore, hybrid models of DWTs, e.g., wavelet-based neural networks
(WNNs) [108], which combine WT and FFNNs, and wavelet-based regression models
[109], which integrate WT and multiple linear regression (MLR), were used in time-series
predictions of floods [110]. The application of WNN for flood prediction was reviewed in
Reference [70], where it was concluded that WNNs can highly enhance model accuracy. In
fact, most recently, WNNs, due to their potential in enhancing time-series data, gained
popularity in flood modeling [50], for applications such as daily flow [111], rainfall–runoff
Hearst et al. [115] proposed and classified the support vector (SV) as a nonlinear
search algorithm using statistical learning theory. Later, the SVM [116] was introduced as
a class of SV, used to minimize over-fitting and reduce the expected error of learning
which works based on the statistical learning theory and the structural risk minimization
rule. The training algorithm of SVM builds models that assign new non-probabilistic
binary linear classifiers, which minimize the empirical classification error and maximize
the geometric margin via inverse problem solving. SVM is used to predict a quantity
forward in time based on training from past data. Over the past two decades, the SVM was
also extended as a regression tool, known as support vector regression (SVR) [117].
SVMs are today know as robust and efficient ML algorithms for flood prediction [118].
SVM and SVR emerged as alternative ML methods to ANNs, with high popularity among
hydrologists for flood prediction. They use the statistical learning theory of structural risk
generalization and superior efficiency. Most importantly, SVMs are both suitable for linear
and nonlinear classification, and the efficient mapping of inputs into feature spaces [119].
Thus, they were applied in numerous flood prediction cases with promising results,
excellent generalization ability, and better performance, compared to ANNs and MLRs,
e.g., extreme rainfall [120], precipitation [43], rainfall–runoff [121], reservoir inflow [122],
streamflow [123], flood quantiles [48], flood time series [124], and soil moisture [125].
Unlike ANNs, SVMs are more suitable for nonlinear regression problems, to identify the
global optimal solution in flood models [126]. Although the high computation cost of using
SVMs and their unrealistic outputs might be demanding, due to their heuristic and semi-
black-box nature, the least-square support vector machine (LS-SVM) highly improved
LS-SVM involves solving a set of linear tasks instead of complex quadratic problems [128].
Nevertheless, there are still a number of drawbacks that exist, especially in the application
application in flood simulation. DT uses a tree of decisions from branches to the target
values of leaves. In classification trees (CT), the final variables in a DT contain a discrete
set of values where leaves represent class labels and branches represent conjunctions of
features labels. When the target variable in a DT has continuous values and an ensemble
of trees is involved, it is called a regression tree (RT) [130]. Regression and classification
trees share some similarities and differences. As DTs are classified as fast algorithms, they
became very popular in ensemble forms to model and predict floods [131]. The
classification and regression tree (CART) [132,133], which is a popular type of DT used in
ML, was successfully applied to flood modeling; however, its applicability to flood
prediction is yet to be fully investigated [134]. The random forests (RF) method [69,135] is
another popular DT method for flood prediction [136]. RF includes a number of tree
predictors. Each individual tree creates a set of response predictor values associated with
a set of independent values. Furthermore, an ensemble of these trees selects the best choice
of classes [69]. Reference [137] introduced RF as an effective alternative to SVM, which
often delivers higher performance in flood prediction modeling. Later, Bui et al. [138]
algorithm [139]. M5 constructs a DT by splitting the decision space and single attributes,
thereby decreasing the variance of the final variable. Further DT algorithms popular in
flood prediction include reduced-error pruning trees (REPTs), Naïve Bayes trees (NBTs),
strong background [140]. Thus, there is an emerging strategy to shift from a single model
of prediction to an ensemble of models suitable for a specific application, cost, and dataset.
ML ensembles consist of a finite set of alternative models, which typically allow more
flexibility than the alternatives. Ensemble ML methods have a long tradition in flood
prediction. In recent years, ensemble prediction systems (EPSs) [141] were proposed as
and weighting system [140]. Such a weighting procedure is carried out to accelerate the
performance evaluation process. The advantage of EPS is the timely and automated
performance of EPS, for flood modeling in particular, can be improved. EPSs may use
DTs, rotation forest (RF) bootstrap, and boosting, allowing higher accuracy and
robustness. The subsequent ensemble prediction systems can be used to quantify the
probability of floods, based on the prediction rate used in the event [142,143,144].
the applications of ensemble ML methods used for floods. EPSs were demonstrated to have
To improve the accuracy of import data and to achieve better dataset management,
the ensemble mean was proposed as a powerful approach coupled with ML methods
[140,141]. Empirical mode decomposition (EMD) [142], and ensemble EMD (EEMD) [143]
are widely used for flood prediction [144]. Nevertheless, EMD-based forecast models are
also subject to a number of drawbacks [145]. The literature includes numerous studies on
The most popular ML modeling methods for flood prediction were identified in the
previous section, including ANFIS, MLP, WNN, EPS, DT, RF, CART, and ANN. Figure 3
presents the major ML methods used for flood prediction, and the number of
corresponding articles in the literature over the last decade. This figure was designed to
160 ANFIS
140
120 MLP
100 WNNs
80
EPS
60
40 DT & RF & CART
20 ANNs
0
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 SVM & SVR
Year
Figure 3. Major ML methods used for flood prediction in the literature. Reference year:
SVMs, MLPs, DTs, ANFIS, WNNs, and EPSs are the most popular. These ML methods can
methods, i.e., ANFIS, WNNs, and basic EPSs, several different research strategies for
obtaining better prediction evolved [137]. The strategies involved developing hybrid ML
models using soft computing techniques, statistical methods, and physical models rather
than individual ML approaches, whereby the extra components complement each other
with respect to their drawbacks and shortcomings. The success of such hybrid approaches
motivated the research community to explore more advanced hybrid models. Figure 4
presents the progress of single vs. hybrid ML methods for flood prediction in the literature
over the past decade. The figure shows an apparent continuous increase and notable
progress in using novel hybrid methods. Through Figure 4, the taxonomy of our research
600
Number of Articles
500
400
300 Hybrids
Singles
200
100
0
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Year
Figure 4. The progress of single vs. hybrid ML methods for flood prediction in the
Furthermore, the types of prediction are often studied with different lead-time
predictions due to the flood. Real-time, hourly, daily, weekly, monthly, seasonal, annual,
short-term, and long-term are the terms most often used in the literature. Real-time
prediction is concerned with anywhere between few minutes and an hour preceding the
flood. Hourly predictions can be 1–3 h ahead of the flood forecasting lead time or, in some
cases, 18 h or 24 h. Daily predictions can be 1–6 days ahead of the forecast. Monthly
forecasts can be, for instance, up to three months. In hydrology, the definitions of short-
term and long-term in studying the different phenomena vary. Short-term predictions for
floods often refer to hourly, daily, and weekly predictions, and they are used as warning
systems. On the other hand, long-term predictions are mostly used for policy analysis
purposes. Furthermore, if the prediction leading time to flood is three days longer than the
confluence time, the prediction is considered to be long-term [37,58]. From this perspective,
in this study, we considered a lead time greater than a week as a long-term prediction. It
was observed that the characteristics of the ML methods used varied significantly
according to the period of prediction. Thus, dividing the survey on the basis of short-term
Here, it is also worth emphasizing that, in this paper, the prediction lead-time was
of time with great destructive power, they can be predicted with either “short-term” or
“long-term” lead times to the actual flood. In fact, this paper is concerned with the lead
times instead of the duration or type of flood. If the lead-time prediction to a flash flood
was short-term, then it was studied as a short-term lead time. However, sometimes flash
floods can be predicted with long lead times. In other words, flash floods might be
predicted one month ahead. In this case, the prediction was considered as long-term.
In this study, the ML methods were reviewed using two classes—single methods and
Single methods
Short-term
Hybrid methods
Single methods
Long-term
Hybrid methods
n=1
Step 1
Qn
Step 4 Step 6
T1 AND T2 AND T3 Yes Yes
Long-term? Hybrid?
No
Acceptable number of
Step 2 No
n+1
results achieved? No
Step 5
No
Single?
Yes
Yes Step 7
Table 3: hybrid
Step 3 Refine the results & Table 2: Single Table 4: Single Table 5: hybrid
method, short-
add to database method, short-term method, long-term method, long-term
term
Step 1 involved running the queries one by one; step 2 involved checking the results
of the search, and initiating the next search; step 3 involved identifying the comparative
studies on ML models of prediction, refining the results and building the database; step 4
involved identifying if it was a single or hybrid method, constructing Table 1, and step 7
involved constructing the other Tables. The four tables provide the list of studies on
different prediction techniques, which entail the organized comprehensive surveys of the
literature.
models, artificial intelligence (AI) methods, and ML, short-term prediction remains a
challenging task [147-152]. This section is divided into two subsections—single and hybrid
methods, i.e., ANNs, MLP, nonlinear autoregressive network with exogenous inputs
(NARX), M5 model trees, DTs, CART, SVR, and RF, followed by a comprehensive
and discussion of these methods follow so as to identify the most suitable methods
Flood Resource
Modeling Technique Reference Prediction Type Region
Variable
Streamflow and
ANN vs. statistical [1] Hourly USA
flash food
Flood frequency
MLP vs. Kohonen NN [154] Long-term China
analysis
Peak flow of
BPANN [155] Daily Canada
flood
Monthly and
BPANN vs. DBPANN [156] Rainfall–runoff China
daily
Rainfall extreme
SVM vs. ANN [120] Daily India
events
Kim and Barros [148] modified an ANN model to improve flood forecasting short-
term lead time through consideration of atmospheric conditions. They used satellite data
from the ISCCP-B3 dataset [172]. This dataset includes hourly rainfall from 160 rain gauges
within the region. The ANN was reported to be considerably more accurate than the
statistical models. In another similar work, Reference [44] developed an ANN forecast
model for hourly lead time. In their study, various datasets were used, consisting of
forecast models showed promising results for 5-h lead time. In another attempt, Danso-
Amoako [1] provided a rapid system for predicting floods with an ANN. They provided a
reliable forecasting tool for rapidly assessing floods. An R2 value of 0.70 for the ANN
model proved that the tool was suitable for predicting flood variables with a high
Panda, Pramanik, and Bala [151] compared the accuracy of ANN with FFANN, and the
results were benchmarked with the physical model of MIKE 11 for short-term water level
prediction. This dataset includes the hourly discharge and water level between 2006 and
2009. The data of the year 2006 was used for testing root-mean-square error (RMSE). The
results indicated that the FFANN performed faster and relatively more accurately than the
ANN model. Here, it is worth mentioning that the overall results indicated that the neural
networks were superior compared to the one-dimensional model MIKE 11. Nevertheless,
11 [8].
Kourgialas, Dokou, and Karatzas [150] created a modeling system for the prediction
of extreme flow based on ANNs 3 h, 12 h, and 19 h ahead of the flood. They analyzed five
years of hourly data to investigate the ANN effectiveness in modeling extreme flood
hydrological models. Lohani, Goel, and Bhatia [4] improved the real-time forecasting of
rainfall–runoff of foods, and the results were compared to the T–S fuzzy model and the
subtractive-clustering-based T–S (TSC-T–S) fuzzy model. They, however, concluded that
the fuzzy model provided more accurate predictions with longer lead time. The hourly
rainfall data from 1989 to 1995 of a gauge site, in addition to the rainfall during a monsoon,
was used. Pereira Filho and dos Santos [153] compared the AR model with an ANN in
simulating forecast stage level and streamflow. The dataset was created from independent
flood events, radar-derived rainfall, and streamflow rain gauges available between 1991
and 1995. The AR and ANN were employed to model short-term flood in an urban area
utilizing streamflow and weather data. They showed that the ANN performed better in its
Ahmad and Simonovic [155] used a BPNN for predicting peak flow utilizing causal
meteorological parameters. This dataset included daily discharge data for 1958–1997 from
gauging stations. BPNN proved to be a fast and accurate approach with the ability of
improve the simulation of daily streamflow using BPNN, Reference [156] used division-
based backpropagation to obtain satisfying results. The raw data of local evaporation and
rainfall gauges of six years were used for the short-term flood prediction of a streamflow
time series. The dataset of one decade from 1988 was used for training and the dataset of
five subsequent years was used for testing. The BPNN model provided promising results;
however, it lacked efficiency in using raw data for the time-series prediction of streamflow.
In addition, Reference [157] showed the application of BPNN for assessing flash floods
using measured data. This dataset included 5-min-frequency water quality data and 15-
min-frequency rainfall data of 20 years from two rain gauge stations. Their experiments
expert knowledge by the user. In addition, their ANN prediction model showed great
ability to deal with a noisy dataset. Ghose [158] predicted the daily runoff using a BPNN
prediction model. The data of daily water level of two years from 2013–2015 were used.
The accurate BPNN model was reported with an efficiency of 96.4% and an R2 of 0.94 for
flood prediction.
Pan, Cheng, and Cai [159] compared the performances of ELM and SVM for short-
However, ELM was suggested as a faster method for parameter selection and learning
loops. Reference [154] also conducted a comparison between fuzzy c-means, ANN, and
MLP using a common dataset of sites to investigate ML method efficiency and accuracy.
The MLP and ANN methods were proposed as the best methods. Chang, Chen, Lu, Huang,
and Chang [160] and Reference [161] modeled multi-step urban flood forecasts using
BPNN and a nonlinear autoregressive network with exogenous input (NARX) for hourly
forecasts. The results demonstrated that NARX worked better in short-term lead-time
prediction compared to BPNN. The NARX network produced an average R2 value of 0.7.
This study suggested that the NARX model was effective in urban flood prediction.
Furthermore, Valipour et al. [24] showed how the accuracy of ANN models could be
Bruen and Yang [162] modeled real-time rainfall–runoff forecasting for different lead
times using FFNN, ARMA, and functional networks. Here, functional networks [173] were
compared with an FFNN model. The models were tested using a storm time-series dataset.
The result was that functional networks allowed quicker training in the prediction of
rainfall–runoff processes with different lead times. The models were able to predict floods
with short lead times. Reference [164] estimated water level–discharge using M5 trees and
ANN. This dataset was collected from the period of 1990 to 1998, and the inputs were
supplied by computing the average mutual information. The ANN and M5 model tree
performed similar in terms of accuracy. Reference [166] tested four DT models, i.e.,
alternating decision trees (ADTs), reduced-error pruning trees (REPTs), logistic model
trees (LMTs), and NBTs, using a dataset of 200 floods. The ADT model was reported to
areas. In other research, Reference [165] compared the performance of an NBT and DT
prediction model, using geomorphological disposition parameters. Both models and their
hybrids were compared in terms of prediction accuracy in a catchment. The advanced DTs
were found to be promising for flood assessment in prone areas. They concluded that an
independent dataset and benchmarking of other ML methods were required for judgment
of the accuracy and efficiency of the method. Reference [171] worked on a dataset
including more than 100 tropical cyclones (TCs) affecting a watershed for the hourly
MLR, and CLIM were compared. The evaluation results showed that MLP and DTs
provided better prediction. Reference [163] applied a dynamic ANN, as well as a Z–R
included three-dimensional radar data of typhoon events and rain gauges from 1990 to
2004, including various typhoons. The results indicated that the ANN performed better.
Aichouri, Hani, Bougherira, Djabri, Chaffai, and Lallahem [167] implemented an MLP
model for flood prediction, and compared the results with the traditional MLR model. The
rainfall–runoff daily data from 1986 to 2003 were used for model building. The results and
comparative study indicated that the MLP approach performed with better yield for river
rainfall–runoff. In a similar research, Reference [98] modeled and predicted the river
rainfall–runoff relationship through training six years of collected daily rainfall data using
MLP and MLR (1990 to 1995). Furthermore, the data of 1996 were used for testing to select
the best performing network model. The R2 values for the ANN and MLR models were
0.888 and 0.917, respectively, showing that the MLP approach gave a much better
prediction than MLR. Reference [169] proposed a number of data-based flood predictions
for daily stream flows models using MLP, WT, MLR, ARIMA, and ANN. This dataset
included two time series of streamflow and a meteorological dataset including records
from 1970 to 2001. The results showed that MLP, WT, and ANN performed generally
better. However, the proposed WT prediction model was evaluated to be not as accurate
as ANN and MLP for a one-week lead time. Reference [170] designed optimal models of
ANN and MLP for the prediction of river level. This study indicated that an optimization
tool for the ANN network can highly improve prediction quality. The candidate inputs
included river levels and mean sea-level pressure (SLP) for the period of 2001–2002. The
MLP was identified as the most accurate model for short-term river flood prediction.
Nayak and Ghosh [120] used SVM and ANN to predict hourly rainfall–runoff using
weather patterns. A model of SVM classifier for rainfall prediction was used and the results
were compared to ANN and another advanced statistical technique. The SVM model
appeared to predict extreme floods better than the ANN. Furthermore, the SVM model
proved to function better in terms of uncertainty. Gizaw and Gan [48] developed SVR and
ANN models for creating RFFA to estimate regional flood quantiles and to assess climate
change impact. This dataset included daily precipitation data obtained from gauges from
1950 to 2016. RMSE and R2 were used for the evaluation of the models. The SVR model
estimated regional flood more accurately than the ANN model. SVR was reported to be a
suitable choice for predicting future flood under the uncertainty of climate change
scenarios [118]. In a similar attempt, Reference [69] provided effective real-time flood
prediction using a rainfall dataset measured by radar. Two models of RF and SVM were
short-term floods, considering the complexity of the algorithm, ease of use, running speed,
accuracy, and input dataset. This table was created based on the revisions that were made
on the articles of Table 1 and also the accuracy analysis of Figure 3, where the values of R2
and RMSE of the single ML methods were considered. The quality of ML model prediction,
in terms of speed, complexity, accuracy, and ease of use, was continuously improved
floods.
Fairly
ELM Fair Fairly high Fair Historical
high
Fairly
ANFIS Fair Fair Fairly high Historical
high
uncertainty, longer lead time, speed, and computation costs, there is an ever increasing
trend in building hybrid ML methods. These hybrid methods are numerous, including
more popular ones, such as ANFIS and WNN, and further novel algorithms, e.g., SVM–
BB, RNN–SVR, RSVRCPSO, MLR–ANN, FFRM–ANN, and EPSs. Table 3 presents these
methods; a revision of the methods and applications follows along with a discussion on
the ML methods.
HEC–HMS–ANN vs.
[179] Rainfall–runoff Hourly Taiwan
HEC–HMS-SVR
SOM–R-NARX vs. R-
[181] Regional flood Hourly Taiwan
NARX
Flash flood:
RNN–SVR, RSVRCPSO [184] Hourly Taiwan
rainfall forecasting
Jimeno-Sáez, et al. [174] modeled flash floods using ANN and ANFIS, applying a
dataset collected from 14 different streamflow gauge stations. RMSE and R2 were used as
evaluation criteria. The results showed that ANFIS demonstrated a considerably superior
ability to estimate real-time flash floods compared to ANN. Chang and Chang [175]
constructed an accurate water level forecasting system based on ANFIS for 1–3 h ahead of
the flood. The ANFIS successfully provided accurate water level prediction. The hourly
water level of five gauges from 1971 to 2001 was used. They concluded that the ANFIS
model could efficiently deal with a big dataset [176] through fast learning and reliable
prediction. A further comparison showed that the ANFIS hybrid model tuned by SVR
provided superior prediction accuracy and good cost-effective computation for nonlinear
and real-time flood prediction. In addition, the model with human interaction could
ANFIS model based on a precipitation dataset, which provided reliable hourly predictions
with an R2 more than 0.85. The results were reported as highly satisfactory for the typhoon
season. Reference [67] used ANFIS for ungauged sites of 151 catchments; the results were
evaluated and compared to the ANN, NLR, NLR-R modes using a Jackknife procedure.
The evaluation showed that the ANFIS model provided higher generalization capability
compared to the NLR and ANN models. The ANFIS model implemented an efficient
mechanism for forecasting the flood region, and providing insight from the data, leading
daily flow prediction using ANN, ANFIS, MLR, and MNLR. Furthermore, the
performances of the models were calculated with RMSE and R2. This dataset included
showed that MNLR models with lower RMSE values had a better performance than the
ANFIS, MLR, and ANN models. Furthermore, MNLR was suggested as a low-cost and
efficient model for the daily prediction of flow. In a similar attempt, Choubin, Darabi et al.
(2018) [133] evaluated the accuracy of ANFIS, considering three common ML modeling
tools—CART, SVM, and MLP. The evaluation suggested that the CART model performed
best. Therefore, CART was strongly suggested as a reliable prediction tool for hydro-
meteorological datasets. Kim and Singh [96] developed three models, namely generalized
NNM), and MLP, for flood prediction. Furthermore, the prediction performance was
GRNNM in forecasting flood discharge. The hybrid models, overall, were shown to
overcome the difficulties when using single ANN models. Reference [178] proposed an
advanced ensemble model through combining FR and SVM to build spatial modeling in
flood prediction. The results were compared with DT. This dataset included an inventory
map of flood prediction in various locations. To build the model, up to 100 flood locations
were used for training and validation. The evaluation results showed a high success rate
for the ensemble model. The results proved the efficiency, accuracy, and speed of the
Young, Liu, and Wu [179] developed a hybrid physical model through integrating the
HEC–HMS model with SVM and ANN for accurate rainfall–runoff modeling during a
capability for hourly prediction. However, the SVR model had much better generalization
and accuracy ability in runoff discharge predictions. It was concluded that the predictions
MP, which is a hybrid of wavelet and season multilayer perceptron for daily rainfall
prediction. The season algorithm is a novel decomposition technique used to improve data
quality. The resulting hybrid model was referred to as the W-SAS–MP model. This dataset
included the daily rainfall data of three decades since 1974. The W-SAS–MP model was
reported as highly efficient for enhancing daily rainfall prediction accuracy and lead time.
Chang, Shen, and Chang [181] developed a hybrid ANN model for real-time
forecasting of regional floods in an urban area. The advanced hybrid model of SOM–R-
NARX was an integration of the NARX network with SOM. Their big dataset included 55
rainfall events of daily rainfall. The evaluation suggested that SOM–R-NARX was accurate
with small values of RMSE and high R2. Furthermore, compared to the cluster-based
hybrid inundation model (CHIM), it provided hourly prediction accuracy. Reference [182]
proposed a model of wavelet-based NARX (WNARX) for the daily forecasting of rainfalls
on a dataset of gauge-based rainfall data for the period from 2000 to 2010. The prediction
performance was further benchmarked with ANN, WANN, ARMAX, and NARX models,
Partal [110] developed a model for the daily prediction of precipitation with ANN
and WNN models. In their case, WNN showed significantly better results with an average
value of 0.79 at various stations. In Reference [60], they compared WNN with ANFIS for
daily rainfall. The results showed that the hybrid algorithm of WNN performed better with
an R2 equal to 0.9 for daily lead time. Reference [105] proposed a hybrid model of wavelet,
bootstrap technique, and ANN, which they called WBANN. It improved the accuracy and
reliability of the ANN model short-term flood prediction. The performance of WBANN
was compared with bootstrap-based ANNs (BANNs) and WNN. The wavelet
resampling produced consistent results. French, Mawdsley, Fujiyama, and Achuthan [183]
proposed a novel hybrid model of ANN and a hydrodynamic model for the accurate short-
term prediction of extreme storm surge water. The ANN–hydrodynamic model generated
realistic flood extents and a great improvement in model accuracy. Reference [184]
optimization algorithm (CPSO). This dataset was obtained from three rain gauges from the
period of 1985 to August 1997, which included the data of nine typhoon events. The results
suggested that the proposed model yielded better performance for rainfall prediction. The
RSVRCPSO model, in comparison with SVRCPSO, resulted in less RMSE learning and
which was a hybrid form of linear regression and a state-space neural network (SSNN).
The performance of the proposed model was benchmarked against the hybrid method of
MLR–ANN. This dataset included the total rain, wind, and humidity measures from 1989–
2008 based on 371 rain gauge stations of six typhoons. The results indicated that the
method was highly robust with a better prediction accuracy in terms of R2, peak discharge,
and total volume. Rajurkar et al. [186] modeled rainfall–runoff by integrating ANN and a
simplified linear model. Furthermore, this dataset included the daily measurements of
rainfall in the period of 1963–1990. The hybrid model was found to be better for providing
Hsu et al. [187] proposed a hybrid model from the integration of a flash-flood routing
model (FFRM) and ANN, called the FFRM–ANN model, to predict hourly river stages.
The ANN algorithms used in this study were the FFNN and FBNN. Data from eight
typhoon events between 2004 and 2005 of rainfall and river stage pairs were selected for
model training. The results indicated that the hybrid model of FFRM–ANN provided an
efficient FFRM for accurate flood forecasting. The comparison of the hybrid method
against each algorithm used in the study proved the effectiveness of the proposed method.
Reference [188] developed a hybrid prediction model by integrating ANN and a nonlinear
against two models of nonlinear perturbation model (LPM), and NLPM integrated with
in the period of 1973–1999. They reported that the NLPM–ANN worked better than the
models of LPM and NLPM–API. The results of the case studies of various watersheds
Through an EPS model, Reference [189] aimed at limiting the range of the
uncertainties in runoff simulations and flood prediction. The classifier ensembles included
MLP, SVM, and RF. Note that the ensemble of MLP was a novel approach in flood
prediction. The proposed EPS presented a number of integrated models and simulation
runs. The model validation was successfully performed using a dataset from various rain
gauges of precipitation data during the 2013–2014 storm season. Using the EPS model
evaluated as reliable and robust in estimating flood duration and destructive power. In
another case, Reference [190] developed an EPS model of six ANNs for daily streamflow
prediction based on daily high-flow data from the storm season of 2013–2014. The
proposed model had a fast development time, which also provided probabilistic forecasts
to deal with uncertainties in prediction. The ensemble prediction system was reported as
generalization, and timeliness are suggested as the basic criteria (Singh 1989). The
timeliness is one of the most important criteria, and it is only achieved through using
robust yet simple models. Furthermore, the performance of the prediction models is often
evaluated through root-mean-square error (RMSE), mean error (ME), mean squared error
(MSE), Nash coefficients (E), and R2, also known as the correlation coefficient (CC). In this
survey, the values of R2 and RMSE were considered for performance evaluation. CC (Eq.1)
∑𝑁
𝑖=1(𝑥𝑖 −𝑥̄ )(𝑦𝑖 −𝑦̄ )
𝐶𝐶 = Eq.1
√[∑𝑁 2 𝑁 2
𝑖=1(𝑥𝑖 −𝑥̄ ) ][∑𝑖=1(𝑦𝑖 −𝑦̄ ) ]
where xi and yi are the observed and predicted values and the i-th residue; x and y are
2
∑𝑛
𝑖=1(𝑋𝑜𝑏𝑠,𝑖 −𝑋𝑚𝑜𝑑𝑒𝑙,𝑖 )
𝑅𝑀𝑆𝐸 = √ Eq.2
𝑛
where Xobs defines observed variables and Xmodel prediction values for year i, where
generally R2 > 0.8 is considered as an acceptable prediction. However, a lower value for
RMSE suggests a better prediction. Overall, forecasting models of floods are reported as
accurate if RMSE values are close to 0, and R2 values are close to 1. The specific intended
purpose, computational cost, and dataset would be our major consideration criteria.
Furthermore, the generalization ability, speed and cost of implementation and operation,
ease of use, low-cost maintenance, robustness, and accuracy of the simulation are other
Here, it is worth mentioning that the value of RMSE can be different across various
studies. In addition, the values of RMSE in some studies were calculated for various sites.
To present a fair evaluation of RMSE, we made sure that the unit of RMSE was the same,
and, for the multiple RMSEs, the average was calculated. We also double-checked for any
possible error. The comparative performance analysis of single and hybrid ML methods
for short-term flood prediction using R2 and RMSE are presented in Figures 7 and 8
respectively.
Comparative performance analysis of single methods for short-term flood
prediction
1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
RMSE R2
1.4
1.2
1
0.8
0.6
0.4
0.2
0
TSC-T–S (Lohani, Goel et al. …
WNN-BB (Kasiviswanathan,…
SAS-MLP (Altunkaynak and…
ANN-BB (Kasiviswanathan,…
RSVRCPSO (Hong 2008)
ANFIS (Kisi, Nia et al. 2012)
Despite performing weakly in a few early studies, especially in the generalization aspect,
better results. In this context, the BPNN and functional networks are suggested as being
difficult to be implemented by the user. However, the models were shown to be reasonably
accurate, efficient, and fast with the ability to deal with noisy datasets. However, the
enhanced through integration with autoregressive models. MLP and DTs provide equally
acceptable prediction yields with ANNs. Among DTs, the ADT model provided the fastest
and most accurate prediction capability in determining floods. Although not as popular as
ANNs, the rotation forest (RF) and M5 model tree (MT) were reported as efficient and
robust. References e.g. [69,136] proposed RF-based models that were as effective as ANNs
runoff modeling and classification with better generalization ability and performance. In
many cases, SVM performed even better, especially for very short lead times [122,125]. In
Nevertheless, the prediction ability decreased for longer lead times. This issue was
addressed using the LS-SVM model, which also showed better generalization ability [127].
predicting hazardous flood quantiles, which revealed the effectiveness of SVM in real-time
flood forecasting.
Overall, the reviewed single prediction models could provide relatively accurate
short-term forecasts. However, for predictions longer than 2 h, hybrid models such as
ANFIS, and WNN performed better. The performance comparisons of the ANFIS model
with BPNN and AR models, with average correlation coefficients higher than 0.80, showed
the superiority of ANFIS in a wide range of short-term flood prediction applications, e.g.,
considerably superior ability for estimating real-time flash flood estimation compared to
most ANN-based models, particularly 1–3 h ahead of flood, providing high accuracy and
reliability. More advanced ANFIS hybrid models tuned by SVR provided even better
prediction accuracy and good cost-effective computation for nonlinear and real-time flood
by increasing the prediction lead time, R2 decreased. For daily flow, MNLR was suggested
with a superior performance over the ANN, ANFIS, and MLR models. In cases where
hydro-meteorological data are readily available, CART was superior to ANFIS, SVM, and
MLP; T–S fuzzy was also a good choice. On the other hand, WNN performed significantly
better than MLP, ANNs, and ANFIS for daily predictions. For accurate longer lead-time
Overall, the novel hybrid models designed using ML, soft computing, and statistical
accuracy and lead time, leading to more realistic flood models with even better
susceptibility assessment. On the other hand, novel ensemble methods not only improved
the accuracy robustness of predictions, but also contributed to limiting the range of
uncertainties in models. Among the EPS methods, the ensembles of ANN, MLP, SVM, and
water resource management potential over longer periods of time, from weekly to monthly
and annual predictions [191]. In the past decades, many notable ML methods, such as
ANN [74], ANFIS [68,192], SVM [193], SVR [193], WNN [51], and bootstrap–ANN [51],
were used for long lead-time predictions with promising results. Recently, in a number of
studies (e.g., References [55,194-198]), the performances of various ML methods for long
lead-time flood predictions were compared. However, it is still not clear which ML method
performs best in long-term flood prediction. In this section, Tables 4 and 5 represent a
a summary of the major single ML methods used in long-term flood prediction, i.e., MLP,
and discussion of these methods follow, identifying the most suitable methods presented
in the literature.
Table 4. Long-term flood prediction using single ML methods.
BPNNs vs.
[200] Reservoir levels Monthly Turkey
BFGSNN
For seasonal flood forecasting, Elsafi [197] proposed numerous ANNs and compared
the results. The water level data from different stations from 1970–1985 were selected for
training, and the data from 1986–1987 were used for verification. The ANNs worked well,
especially where the dataset was not complete, providing a viable choice for accurate
prediction. ANNs provided the possibility of reducing the analytical costs through
reducing the data analysis time that used to face in e.g., [198]. Similarly, reference [87] used
different stations was considered and the ANN model was applied to various stations to
evaluate prediction performance. The authors summarized that the ANN models offered
Reference [202] used an ANN model for stream assessment for long-term floods. This
dataset was collected from more than 100 sites of numerous flood streams. They concluded
that the ANN model, compared to Hilsenhoff’s biotic index (HBI), significantly improved
the prediction ability using geomorphic data. However, the ANN had generalization
problems. Nevertheless, the ANN in this case proved useful to water managers.
Singh [199] used a number of BPNNs to build prediction models of heavy rains and
floods. This dataset included the period of 1871–2010 on a monthly time scale. The results
indicated that the BPNN models were fast and robust with simple networks, which made
them great for forecasting nonlinear floods. Reference [200] aimed to better analyze
nonlinear floods through modeling with BPNN and local linear regression (LLR)-based
models for long-term flood forecasting. This dataset included almost two decades of
rainfall, outflow, inflows, evaporation, and water level since 1988. Their evaluation
concluded that LLR showed better prediction than the Broyden Fletcher Goldfarb Shanno
neural network (BFGSNN) model in terms of performance and accuracy with bigger
values of R2 and lower values of RMSE. However, BPNN outperformed the other methods
with relatively good results. Among the ANN variations, [151] proposed a BPNN model
as the most reliable ANN for long-term flood prediction. Reference [201] also compared
the performances of ANNs with BPNN and MLP in the long-term prediction of flood
discharge. Promising results were obtained when using MLP. However, generalization
remained an issue.
Lin, Cheng, and Chau [203] applied an SVM model for estimating streamflow and
reservoir inflow for a long lead time. To benchmark, they used ANNs and ARMA. The
prediction models were built using monthly river flow discharges from the period of 1974–
1998 for training, and 1999–2003 for testing. Through a comparison of model performance,
SVM was demonstrated as a potential candidate for the prediction of long-term discharges,
model for estimating soil moisture using remote-sensing data, and the results were
compared to predictive models based on BPNN and MLR. Training was performed on the
data of the period of 1998 to 2002, and testing used data from 2003 to 2005. The SVM model
was shown to be more accurate and easier to build compared to BPNN and MLR.
Reference [204], employed RT to model forest flood. Data from 2009–2012 at 50 sites were
used for model building. The prediction of annual forest floods was reported through a
long-term floods considering the complexity of algorithm, ease of use, running speed,
accuracy, and input dataset. This table was created based on revisions that were made on
articles of Table 4, as well as the accuracy analysis in Figure 9, where values of R2 and
RMSE for the single ML methods were considered. The quality of the ML model
floods.
A critical review on the long-term flood prediction using hybrid methods is presented
autoregressive ANN integrated with sigmoid and radial activity functions. The proposed
hybrid method outperformed the conventional statistical methods of ARMA and ARIMA
with lower values of RMSE. They reported that ARIMA was suitable for the prediction of
monthly and annual inflow, while the dynamic autoregressive ANN model with a sigmoid
activity function could be used for even longer lead time. This dataset included monthly
Streamflow
Hybrid WNN vs. M5 model tree [206] Monthly Australia
water level
Weekly and
WNN-BB vs. WNN vs. ANN [50] Streamflow Canada
few days
ANFIS, ANNs vs. SVM vs. LLR [214] Streamflow Short-term Turkey
Flood
NLPM–ANN [215] Yearly China
forecasting
PCA
EEMD–ANN vs. SVM vs. ANFIS [220] Runoff forecast Monthly China
WNN vs. ANN vs. WLGP [51] Streamflow Monthly Iran
Adamowski [25] developed models based on ANN and WNN, and compared their
prediction performances with statistical methods. WNN was proposed as the most
accurate prediction model, as previously confirmed by Cannas et al. (2005) [207] for
monthly rainfall–runoff forecasting, and also for further engineering application [208]. In
a similar work, Reference [209] compared the performances of ANN and WNN for the
prediction of peak flows. They also reported WNN as most reliable for simulating extreme
event streams, whereby decomposition improved the results considerably. Higher levels
of wavelet decomposition further improved the testing results. The statistical performance
Ramana [210] also combined the wavelet technique with ANN for long-term flood
prediction. They considered 74 years of data for the period of 1901 to 1975. A dataset of 44
years was used for calibration, and the remainder was used for validation of the model.
Their results showed a relatively lower performance for ANNs compared WNN models
in modeling rainfall–runoffs. Cannas et al. [207] proposed WNN for monthly rainfall–
attempt, Kasiviswanathan, He, Sudheer, and Tay [50] used WNN and WNN–BB, which is
an ensemble of WNN utilizing the block bootstrap (BB) sampling technique, to identify a
robust modeling approach among ANN and WNN, by assessing accuracy and precision.
This dataset included measurements from 1912 to 2013 at several flow gauge stations. The
results suggested WNN–BB as a robust model for long-term streamflow prediction for
longer lead times of up to one year. Tantanee et al. [211] proposed a hybrid of wavelet and
autoregressive models, called WARM, which performed more effectively for long lead
times. Prasad [206] proposed another similar hybrid model with the integration of WNN
and iterative input selection (IIS). The hybrid model was called IIS–W-ANN, and was
benchmarked with the M5 model tree. Their dataset included streamflow water level
measurements from 40 years. The IIS–W-ANN hybrid model outperformed the M5 tree.
This study advocated that the novel IIS–W-ANN method should be considered as an
excellent flood forecasting model. Nevertheless, the model could be further optimized for
such optimizers can complement IIS–W-ANN for fine-tuning the hidden-layer weights
and biases for better prediction. Mekanik [212] used ANFIS to forecast seasonal rainfall. A
comparison of the performance and accuracy of the ANN model and a physical model
showed promising results for ANFIS. Rainfall measurements of 1900–1999 were used for
training and validation, and the following decade was used for testing. The results showed
that ANFIS outperformed the ANN models in all cases, comparable to Predictive Ocean
Atmosphere Model for Australia (POAMA), and better than climatology. Furthermore, the
addition, the study suggested ANFIS as an alternative tool for long-term predictions.
ANFIS was reported as being easy to implement with low complexity and minimal input
performances of ANFIS, ANNs, and SVM. This dataset included monthly flow data from
1953 to 2004, where the period of 2000–2004 was used for validation. ANFIS and SVM were
evaluated as being better for long-term predictions. References [224,226] compared the
performances of ANFIS, ANNs, and SVM for the monthly prediction of floods. The
comparison results indicated that the ML models provided more accuracy than the
more accuracy vs. SVM. However, for low-flow predictions, the SVM and ANN models
NLPM–ANN to predict wetness and flood. To do so, the seasonal rainfall and wetness data
of various stations were considered. The NLPM–ANN model was reported as being
model for long lead times, and comparedits accuracy with ANN and SVM models. The M-
EMDSVM model was created through modification of EMD–SVM. The evaluation results
showed that the M-EMDSVM model was a better alternative to ANN, SVM, and EMD–
SVM models for long lead-time streamflow prediction. The M-EMDSVM model also
Zhu, Zhou, Ye, and Meng [217] contributed to the integration of ML with time-series
accuracy of a number of models. For that matter, they integrated SVM with discrete
wavelet transform (DWT) and EMD. The hybrid models were called DWT–SVR and EMD–
SVR. The results indicated that decomposition improved the accuracy of streamflow
prediction, yet DWT performed even better. Further comparisons of SVR, EMD–SVR, and
DWT–SVR models showed that EMD and DWT were significantly more accurate than SVR
prediction in real-life cases. He utilized the K-nearest neighbor regression for the purpose
EPS of ANNs, the hybrid model of K-NN was proposed to increase the generalization
ability of neural networks, and was further compared with the results using MLP, MLP–
PLC, and ANN. The hourly water level data of the reservoir from 132 typhoons in the
period of 1971–2001 were used. The proposed EPS had a promising ability of
Bass and Bedient [218] proposed a hybrid model of surrogate–ML for long-term flood
prediction suitable for TCs. The methods used included ANN integrated with principal
component analysis (PCA), Kriging integrated with PC, and Kriging. The models were
reported as efficient and fast to build. The results demonstrated that the methodology had
Ravansalar [51] compared the performances of the prediction models of WNN, ANN,
and a novel hybrid model called wavelet linear genetic programming (WLGP) in dealing
with the long-term prediction of streamflow. The results showed an accuracy of 0.87 for
the WLGP model. The comparison of the performance evaluation showed that WLGP
significantly increased the accuracy for the monthly approximation of peak streamflow.
prediction is presented. Figure 9 represents the values of RMSE and R2 for single methods
of ML, where ANNs, SVMs, and SVRs show better results. Figure 10 represents the values
of RMSE and R2 for hybrid methods of ML, where decomposition and ensemble methods
ANNs are the most widely used ML method due to their accuracy, high fault
tolerance, and powerful parallel processing in dealing with complex flood functions,
especially where datasets are not complete. However, generalization remains an issue with
ANN. In this context, ANFIS, MLP, and SVM performed better than ANNs. However,
wavelet transforms were reported to be useful for decompositions of original time series,
improving the ability of most ML methods by providing insight into datasets on various
0.8
0.6
0.4
0.2
RMSE R2
prediction.
were reported as better than those which were trained using un-decomposed time series.
However, despite the achievement of WNNs, the predictions were not satisfactory for long
lead times. To increase the accuracy of the longer-lead-time predictions up to one year,
novel hybrids such as WARM, which is a hybrid of WNN and an autoregressive model,
and wavelet multi-resolution analysis (WMRA) were proposed. In other cases, it was seen
cleaner inputs. For example, wavelet–neuro-fuzzy models [228] were significantly more
accurate and faster than single ANFIS and ANNs. However, with an increase in the lead
time, the uncertainty in prediction increased. Thus, the evaluation of model precision
autoregressive, DWT, IIS, and EMD, contributed highly to developing hybrid methods for
longer prediction lead time, good stability, great representativeness, and higher accuracy.
These data decomposition methods were integrated with ANNs, SVM, WNN, and FR, and
they are expected to gain more popularity among researchers. The other trend in
generalization. The EPS of ANNs and WNNs, using BB sampling, genetic programming,
simple average, stop training, Bayesian, data fusion, regression, and other soft computing
methods. In ensembles, however, it is noted that human decision as the input variable
provided superior performance than models without this important input. However, the
suitable for monthly prediction. Their performance comparisons with SVM, ANFIS, and
prediction. Here, it is also worth mentioning the importance of further signal processing
techniques (e.g., Reference [228]) for both long-term and short-term floods.
Comparative performance analysis of single methods for long-term
1.2
0.8
0.6
0.4
0.2
RMSE R2
prediction.
This paper suggests that the drawbacks to major ML methods in terms of accuracy,
methods, as well as using an ensemble variation of the ML method. It is expected that this
5. Conclusions
The current state of ML modeling for flood prediction is quite young and in the early
stage of advancement. This paper presents an overview of machine learning models used
in flood prediction, and develops a classification scheme to analyze the existing literature.
The survey represents the performance analysis and investigation of more than 6000
articles. Among them, we identified 180 original and influential articles where the
performance and accuracy of at least two machine learning models were compared. To do
so, the prediction models were classified into two categories according to lead time, and
further divided into categories of hybrid and single methods. The state of the art of these
classes was discussed and analyzed in detail, considering the performance comparison of
the methods available in the literature. The performance of the methods was evaluated in
cost, and speed. Despite the promising results already reported in implementing the most
popular machine learning methods, e.g., ANNs, SVM, SVR, ANFIS, WNN, and DTs, there
was significant research and experimentation for further improvement and advancement.
In this context, there were four major trends reported in the literature for improving the
quality of prediction. The first was novel hybridization, either through the integration of
two or more machine learning methods or the integration of a machine learning method(s)
with more conventional means, and/or soft computing. The second was the use of data
decomposition techniques for the purpose of improving the quality of the dataset, which
highly contributed in improving the accuracy of prediction. The third was the use of an
models and decreased the uncertainty of prediction. The fourth was the use of add-on
optimizer algorithms to improve the quality of machine learning algorithms, e.g., for better
tuning the ANNs to reach optimal neuronal architectures. It is expected that, through these
four key technologies, flood prediction will witness significant improvements for both
short-term and long-term predictions. Surely, the advancement of these novel ML methods
depends highly on the proper usage of soft computing techniques in designing novel
learning algorithms. This fact was discussed in the paper, and the soft computing
the future.
Here, it is also worth mentioning that the multidisciplinary nature of this work was
the most challenging difficulty to overcome in this paper. Having contributions from the
coauthors of both realms of ML and hydrology was the key to success. Furthermore, the
novel search methodology and the creative taxonomy and classification of the ML methods
For future work, conducting a survey on spatial flood prediction using machine
learning models is highly encouraged. This important aspect of flood prediction was
excluded from our paper due to the nature of modeling methodologies and the datasets
machine learning models for spatial flood analysis revolutionized this particular realm of
methodology, supervision, communication, resources, data curation, and project management. A.M.
contributed to the original draft preparation, data collection, formal data analysis, investigation,
critical review, and final revisions, K.C., the hydrology expert, contributed to supervision, validation,
Funding: This research was funded by the Norwegian University of Science and Technology AI Lab
and the European Research Consortium for Informatics and Mathematics (ERCIM).
Acknowledgments: Dr. Amir Mosavi contributed to this research during his tenure through an
ERCIM Alain Bensoussan Fellowship Program under supervision of Prof. Pinar Ozturk.
ML Machine learning
AI Artificial intelligence
DT Decision tree
AR Autoregressive
AR Autoregressive
FR Frequency ratio
RF Rotation forest
WBANN Wavelet–bootstrap–ANN
RT Regression trees
T–S Takagi–Sugeno
E Nash coefficients
References
1. Danso-Amoako, E.; Scholz, M.; Kalimeris, N.; Yang, Q.; Shao, J. Predicting dam failure risk for
sustainable flood retention basins: A generic case study for the wider greater manchester area.
2. Xie, K.; Ozbay, K.; Zhu, Y.; Yang, H. Evacuation zone modeling under climate change: A data-
3. Pitt, M. Learning Lessons from the 2007 Floods; Cabinet Office: London, UK, 2008.
4. Lohani, A.K.; Goel, N.; Bhatia, K. Improving real time flood forecasting using fuzzy inference
5. Mosavi, A.; Bathla, Y.; Varkonyi-Koczy, A. Predicting the Future Using Web Knowledge: State
of the Art Survey. In Recent Advances in Technology Research and Education; Springer: Cham,
6. Zhao, M.; Hendon, H.H. Representation and prediction of the indian ocean dipole in the poama
7. Borah, D.K. Hydrologic procedures of storm event watershed models: A comprehensive review
8. Costabile, P.; Costanzo, C.; Macchione, F. A storm event watershed model for surface runoff
based on 2D fully dynamic wave equations. Hydrol. Process. 2013, 27, 554–569.
models for forecasting rainfall–runoff from precipitation data in urban areas. J. Hydrol. 2010,
382, 88–102.
10. Fernández-Pato, J.; Caviedes-Voullième, D.; García-Navarro, P. Rainfall/runoff simulation with
2D full shallow water equations: Sensitivity analysis and calibration of infiltration parameters.
11. Caviedes-Voullième, D.; García-Navarro, P.; Murillo, J. Influence of mesh structure on 2D full
shallow water equations and SCS curve number simulation of rainfall/runoff events. J. Hydrol.
12. Costabile, P.; Costanzo, C.; Macchione, F. Comparative analysis of overland flow models using
13. Xia, X.; Liang, Q.; Ming, X.; Hou, J. An efficient and stable hydrodynamic model with novel
source term discretization schemes for overland flow and flood simulations. Water Resour. Res.
14. Liang, X.; Lettenmaier, D.P.; Wood, E.F.; Burges, S.J. A simple hydrologically based model of
land surface water and energy fluxes for general circulation models. J. Geophys. Res. Atmos. 1994,
99, 14415–14428.
15. Costabile, P.; Macchione, F. Enhancing river model set-up for 2-D dynamic flood modelling.
16. Nayak, P.; Sudheer, K.; Rangan, D.; Ramasastri, K. Short-term flood forecasting with a
17. Kim, B.; Sanders, B.F.; Famiglietti, J.S.; Guinot, V. Urban flood modeling with porous shallow-
water equations: A case study of model errors in the presence of anisotropic porosity. J. Hydrol.
18. Van den Honert, R.C.; McAneney, J. The 2011 brisbane floods: Causes, impacts and
19. Lee, T.H.; Georgakakos, K.P. Operational rainfall prediction on meso-γ scales for hydrologic
20. Shrestha, D.; Robertson, D.; Wang, Q.; Pagano, T.; Hapuarachchi, H. Evaluation of numerical
21. Bellos, V.; Tsakiris, G. A hybrid method for flood simulation in small catchments combining
22. Bout, B.; Jetten, V. The validity of flow approximations when simulating catchment-integrated
of the 1-D modeling highlighted by the 2-D approach. Nat. Hazards 2015, 77, 181–204.
24. Valipour, M.; Banihabib, M.E.; Behbahani, S.M.R. Parameters estimate of autoregressive
moving average and autoregressive integrated moving average models and compare their
25. Adamowski, J.; Fung Chan, H.; Prasher, S.O.; Ozga‐Zielinski, B.; Sliusarieva, A. Comparison of
multiple linear and nonlinear regression, autoregressive integrated moving average, artificial
neural network, and wavelet artificial neural network methods for urban water demand
26. Valipour, M.; Banihabib, M.E.; Behbahani, S.M.R. Comparison of the ARMA, ARIMA, and the
autoregressive artificial neural network models in forecasting the monthly inflow of Dez dam
27. Chow, V.T.; Maidment, D.R.; Larry, W. Mays. Applied hydrology; International Edition;
28. Aziz, K.; Rahman, A.; Fang, G.; Shrestha, S. Application of artificial neural networks in regional
flood frequency analysis: A case study for australia. Stoch. Environ. Res. Risk Assess. 2014, 28,
541–554.
29. Kroll, C.N.; Vogel, R.M. Probability distribution of low streamflow series in the united states. J.
30. Mackey, B.P.; Krishnamurti, T. Ensemble forecast of a typhoon flood event. Weather Forecast.
31. Haddad, K.; Rahman, A. Regional flood frequency analysis in eastern australia: Bayesian GLS
regression-based methods within fixed region and ROI framework–quantile regression vs.
32. Thompson, S.A. Hydrology for Water Management; CRC Press: Boca Raton, FL, USA, 2017.
33. Kerkhoven, E.; Gan, T.Y. A modified ISBA surface scheme for modeling the hydrology of
Athabasca river basin with GCM-scale data. Adv. Water Resour. 2006, 29, 808–826.
34. Burnash, R.J.; Ferral, R.L.; McGuire, R.A. A Generalized Streamflow Simulation System, Conceptual
Modeling for Digital Computers; Stanford University: Stanford, CA, USA, 1973.
35. Yamazaki, D.; Kanae, S.; Kim, H.; Oki, T. A physically based description of floodplain
inundation dynamics in a global river routing model. Water Resour. Res. 2011, 47,
doi:10.1029/2010WR009726.
36. Fawcett, R.; Stone, R. A comparison of two seasonal rainfall forecasting systems for Australia.
37. Mekanik, F.; Imteaz, M.; Gato-Trinidad, S.; Elmahdi, A. Multiple regression and artificial neural
network for long-term rainfall forecasting using large scale climate modes. J. Hydrol. 2013, 503,
11–21.
38. Mosavi, A.; Rabczuk, T.; Varkonyi-Koczy, A. R. Reviewing the novel machine learning tools for
materials design. In Recent Advances in Technology Research and Education; Springer: Cham,
39. Abbot, J.; Marohasy, J. Input selection and optimisation for monthly rainfall forecasting in
Queensland, Australia, using artificial neural networks. Atmos. Res. 2014, 138, 166–178.
40. Fox, N.I.; Wikle, C.K. A bayesian quantitative precipitation nowcast scheme. Weather Forecast.
41. Merz, B.; Hall, J.; Disse, M.; Schumann, A. Fluvial flood risk management in a changing world.
42. Xu, Z.; Li, J. Short-term inflow forecasting using an artificial neural network model. Hydrol.
43. Ortiz-García, E.; Salcedo-Sanz, S.; Casanova-Mateo, C. Accurate precipitation prediction with
support vector classifiers: A study including novel predictive variables and observational data.
44. Kim, S.; Matsumi, Y.; Pan, S.; Mase, H. A real-time forecast model using artificial neural
network for after-runner storm surges on the Tottori Coast, Japan. Ocean Eng. 2016, 122, 44–53.
45. Mosavi, A.; Edalatifar, M. A Hybrid Neuro-Fuzzy Algorithm for Prediction of Reference
46. Dineva, A.; Várkonyi-Kóczy, A.R.; Tar, J.K. Fuzzy expert system for automatic wavelet
shrinkage procedure selection for noise suppression. In Proceedings of the 2014 IEEE 18th
International Conference on Intelligent Engineering Systems (INES), Tihany, Hungary, 3–5 July
47. Suykens, J.A.; Vandewalle, J. Least squares support vector machine classifiers. Neural Process.
48. Gizaw, M.S.; Gan, T.Y. Regional flood frequency analysis using support vector regression
parameters using extreme learning machine and artificial neural network. Eng. Appl. Comput.
50. Kasiviswanathan, K.; He, J.; Sudheer, K.; Tay, J.-H. Potential application of wavelet neural
network ensemble to forecast streamflow for flood management. J. Hydrol. 2016, 536, 161–173.
51. Ravansalar, M.; Rajaee, T.; Kisi, O. Wavelet-linear genetic programming: A new approach for
52. Mosavi, A.; Rabczuk, T. Learning and intelligent optimization for material design innovation.
In Learning and Intelligent Optimization; Springer: Cham, Switzerland, 2017; pp. 358–363.
53. Dandagala, S.; Reddy, M.S.; Murthy, D.S.; Nagaraj, G. Artificial neural networks applications
in groundwater hydrology—A review. Artif. Intell. Syst. Mach. Learn. 2017, 9, 182–187.
54. Deka, P.C. Support vector machine applications in the field of hydrology: A review. Appl. Soft
55. Fotovatikhah, F.; Herrera, M.; Shamshirband, S.; Chau, K.-W.; Faizollahzadeh Ardabili, S.;
Piran, M.J. Survey of computational intelligence as basis to big flood management: Challenges,
research directions and future work. Eng. Appl. Comput. Fluid Mech. 2018, 12, 411–437.
56. Faizollahzadeh Ardabili, S.; Najafi, B.; Alizamir, M.; Mosavi, A.; Shamshirband, S.; Rabczuk, T.
Using SVM-RSM and ELM-RSM Approaches for Optimizing the Production Process of Methyl
57. Tsai, L.T.; Yang, C.-C. Improving measurement invariance assessments in survey research with
missing data by novel artificial neural networks. Expert Syst. Appl. 2012, 39, 10456–10464.
58. Sivapalan, M.; Blöschl, G.; Merz, R.; Gutknecht, D. Linking flood frequency to long-term water
doi:10.1029/2004WR003439.
59. Maier, H.R.; Jain, A.; Dandy, G.C.; Sudheer, K.P. Methods used for the development of neural
networks for the prediction of water resource variables in river systems: Current status and
60. Lafdani, E.K.; Nia, A.M.; Pahlavanravi, A.; Ahmadi, A.; Jajarmizadeh, M. Research article daily
rainfall-runoff prediction and simulation using ANN, ANFIS and conceptual hydrological
62. Seo, D.-J.; Breidenbach, J. Real-time correction of spatially nonuniform bias in radar rainfall
63. Grecu, M.; Krajewski, W. A large-sample investigation of statistical procedures for radar-based
64. Maddox, R.A.; Zhang, J.; Gourley, J.J.; Howard, K.W. Weather radar coverage over the
65. Campolo, M.; Andreussi, P.; Soldati, A. River flood forecasting with a neural network model.
66. Prakash, O.; Sudheer, K.; Srinivasan, K. Improved higher lead time river flow forecasts using
sequential neural network with error updating. J. Hydrol. Hydromech. 2014, 62, 60–74.
67. Shu, C.; Ouarda, T. Regional flood frequency analysis at ungauged sites using the adaptive
68. Ashrafi, M.; Chua, L.H.C.; Quek, C.; Qin, X. A fully-online neuro-fuzzy model for flow
69. Yu, P.-S.; Yang, T.-C.; Chen, S.-Y.; Kuo, C.-M.; Tseng, H.-W. Comparison of random forests and
support vector machine for real-time radar-derived rainfall forecasting. J. Hydrol. 2017, 552, 92–
104.
70. Nourani, V.; Baghanam, A.H.; Adamowski, J.; Kisi, O. Applications of hybrid wavelet–artificial
71. Zadeh, M.R.; Amin, S.; Khalili, D.; Singh, V.P. Daily outflow prediction by multi layer
perceptron with logistic sigmoid and tangent sigmoid activation functions. Water Resour.
72. Li, L.; Xu, H.; Chen, X.; Simonovic, S. Streamflow forecast and reservoir operation performance
assessment under climate change. Water Resour. Manag. 2010, 24, 83.
73. Wu, C.; Chau, K.-W. Data-driven models for monthly streamflow time series prediction. Eng.
74. Sulaiman, J.; Wahab, S.H. Heavy rainfall forecasting model using artificial neural network for
flood prone area. In It Convergence and Security 2017; Springer: Singapore, 2018; pp. 68–76.
75. Kar, A.K.; Lohani, A.K.; Goel, N.K.; Roy, G.P. Development of flood forecasting system using
statistical and ANN techniques in the downstream catchment of mahanadi basin, india. J. Water
76. Jain, A.; Prasad Indurthy, S. Closure to “comparative analysis of event-based rainfall-runoff
77. Lohani, A.; Kumar, R.; Singh, R. Hydrological time series modeling: A comparison between
adaptive neuro-fuzzy, neural network and autoregressive techniques. J. Hydrol. 2012, 442, 23–
35.
78. Tanty, R.; Desmukh, T.S. Application of artificial neural network in hydrology—A review. Int.
79. Kişi, O. Streamflow forecasting using different artificial neural network algorithms. J. Hydrol.
80. Shamseldin, A.Y. Artificial neural network model for river flow forecasting in a developing
81. Badrzadeh, H.; Sarukkalige, R.; Jayawardena, A. Impact of multi-resolution analysis of artificial
intelligence models inputs on multi-step ahead river flow forecasting. J. Hydrol. 2013, 507, 75–
85.
82. Smith, J.; Eli, R.N. Neural-network models of rainfall-runoff process. J. Water Resour. Plan.
83. Taormina, R.; Chau, K.-W.; Sethi, R. Artificial neural network simulation of hourly
groundwater levels in a coastal aquifer system of the Venice Lagoon. Eng. Appl. Artif. Intell.
84. Thirumalaiah, K.; Deo, M. River stage forecasting using artificial neural networks. J. Hydrol.
85. Panagoulia, D. Artificial neural networks and high and low flows in various climate regimes.
86. Panagoulia, D.; Tsekouras, G.; Kousiouris, G. A multi-stage methodology for selecting input
variables in ann forecasting of river flows. Glob. Nest J. 2017, 19, 49–57.
87. Deo, R.C.; Şahin, M. Application of the artificial neural network model for prediction of
parameters and climate indices in Eastern Australia. Atmos. Res. 2015, 161, 65–81.
88. Coulibaly, P.; Dibike, Y.B.; Anctil, F. Downscaling precipitation and temperature with temporal
89. Schoof, J.T.; Pryor, S. Downscaling temperature and precipitation: A comparison of regression-
based methods and artificial neural networks. Int. J. Climatol. 2001, 21, 773–790.
90. Hassan, Z.; Shamsudin, S.; Harun, S.; Malek, M.A.; Hamidon, N. Suitability of ANN applied as
a hydrological model coupled with statistical downscaling model: A case study in the northern
91. Zhang, J.-S.; Xiao, X.-C. Predicting chaotic time series using recurrent neural network. Chin.
92. Huang, G.-B.; Zhu, Q.-Y.; Siew, C.-K. Extreme learning machine: Theory and applications.
93. Lima, A.R.; Cannon, A.J.; Hsieh, W.W. Forecasting daily streamflow using online sequential
94. Yaseen, Z.M.; Jaafar, O.; Deo, R.C.; Kisi, O.; Adamowski, J.; Quilty, J.; El-Shafie, A. Stream-flow
forecasting using extreme learning machines: A case study in a semi-arid region in Iraq. J.
95. Sahoo, G.; Ray, C. Flow forecasting for a Hawaii stream using rating curves and neural
96. Kim, S.; Singh, V.P. Flood forecasting using neural computing techniques and conceptual class
97. Rumelhart, D.E.; Hinton, G.E.; Williams, R.J. Learning representations by back-propagating
98. Riad, S.; Mania, J.; Bouchaou, L.; Najjar, Y. Rainfall-runoff model using an artificial neural
99. Senthil Kumar, A.; Sudheer, K.; Jain, S.; Agarwal, P. Rainfall-runoff modelling using artificial
neural networks: Comparison of network types. Hydrol. Process. Int. J. 2005, 19, 1277–1291.
100. Zadeh, L.A. Soft computing and fuzzy logic. In Fuzzy Sets, Fuzzy Logic, and Fuzzy Systems:
Selected Papers by Lotfi a Zadeh; World Scientific: Singapore, 1996; pp. 796–804.
101. Choubin, B.; Khalighi-Sigaroodi, S.; Malekian, A.; Ahmad, S.; Attarod, P. Drought forecasting
in a semi-arid watershed using climate signals: A neuro-fuzzy modeling approach. J. Mt. Sci.
layer perceptron network and adaptive neuro-fuzzy inference system for forecasting
precipitation based on large-scale climate signals. Hydrol. Sci. J. 2016, 61, 1001–1009.
103. Bogardi, I.; Duckstein, L. The fuzzy logic paradigm of risk analysis. In Risk-Based Decisionmaking
in Water Resources X; American Society of Civil Engineers: Reston, VA, USA, 2003; pp. 12–22.
104. See, L.; Openshaw, S. A hybrid multi-model approach to river level forecasting. Hydrol. Sci. J.
105. Tiwari, M.K.; Chatterjee, C. Development of an accurate and reliable hourly flood forecasting
model using wavelet–bootstrap–ANN (WBANN) hybrid approach. J. Hydrol. 2010, 394, 458–
470.
106. Guimarães Santos, C.A.; da Silva, G.B.L. Daily streamflow forecasting using a wavelet
transform and artificial neural network hybrid models. Hydrol. Sci. J. 2014, 59, 312–324.
107. Supratid, S.; Aribarg, T.; Supharatid, S. An integration of stationary wavelet transform and
nonlinear autoregressive neural network with exogenous input for baseline and future
108. Shoaib, M.; Shamseldin, A.Y.; Melville, B.W. Comparative study of different wavelet based
neural network models for rainfall–runoff modeling. J. Hydrol. 2014, 515, 47–58.
109. Dubossarsky, E.; Friedman, J.H.; Ormerod, J.T.; Wand, M.P. Wavelet-based gradient boosting.
110. Partal, T. Wavelet regression and wavelet neural network models for forecasting monthly
111. Shafaei, M.; Kisi, O. Predicting river daily flow using wavelet-artificial neural networks based
on regression analyses in comparison with artificial neural networks and support vector
112. Kumar, S.; Tiwari, M.K.; Chatterjee, C.; Mishra, A. Reservoir inflow forecasting using ensemble
models based on neural networks, wavelet analysis and bootstrap method. Water Resour.
113. Seo, Y.; Kim, S.; Kisi, O.; Singh, V.P. Daily water level forecasting using wavelet decomposition
114. Sudhishri, S.; Kumar, A.; Singh. J. K. Comparative Evaluation of Neural Network and
Regression Based Models to Simulate Runoff and Sediment Yield in an Outer Himalayan
116. Vapnik, V.; Mukherjee, S. Support vector method for multivariate density estimation. Adv.
117. Li, S.; Ma, K.; Jin, Z.; Zhu, Y. A new flood forecasting model based on SVM and boosting
118. Dehghani, M.; Saghafian, B.; Nasiri Saleh, F.; Farokhnia, A.; Noori, R. Uncertainty analysis of
streamflow drought forecast using artificial neural networks and Monte‐Carlo simulation. Int.
119. Dibike, Y.B.; Velickov, S.; Solomatine, D.; Abbott, M.B. Model induction with support vector
machines: Introduction and applications. J. Comput. Civ. Eng. 2001, 15, 208–216.
120. Nayak, M.A.; Ghosh, S. Prediction of extreme rainfall event using weather pattern recognition
and support vector machine classifier. Theor. Appl. Climatol. 2013, 114, 583–603.
121. Granata, F.; Gargano, R.; de Marinis, G. Support vector regression for rainfall-runoff modeling
in urban drainage: A comparison with the EPA’s storm water management model. Water 2016,
8, 69.
122. Gong, Y.; Zhang, Y.; Lan, S.; Wang, H. A comparative study of artificial neural networks,
support vector machines and adaptive neuro fuzzy inference system for forecasting
groundwater levels near lake okeechobee, Florida. Water Resour. Manag. 2016, 30, 375–391.
123. Jajarmizadeh, M.; Lafdani, E.K.; Harun, S.; Ahmadi, A. Application of SVM and swat models
for monthly streamflow prediction, a case study in South of Iran. KSCE J. Civ. Eng. 2015, 19,
345–357.
124. Bao, Y.; Xiong, T.; Hu, Z. Multi-step-ahead time series prediction using multiple-output
125. Bray, M.; Han, D. Identification of support vector machines for runoff modelling. J.
126. Tehrany, M.S.; Pradhan, B.; Mansor, S.; Ahmad, N. Flood susceptibility assessment using GIS-
based support vector machine model with different kernel types. Catena 2015, 125, 91–101.
127. Kisi, O.; Parmar, K.S. Application of least square support vector machine and multivariate
adaptive regression spline models in long term prediction of river water pollution. J. Hydrol.
129. Sachindra, D.; Huang, F.; Barton, A.; Perera, B. Least square support vector and multi-linear
130. De’ath, G.; Fabricius, K.E. Classification and regression trees: A powerful yet simple technique
131. Tehrany, M.S.; Pradhan, B.; Jebur, M.N. Spatial prediction of flood susceptible areas using rule
based decision tree (DT) and a novel ensemble bivariate and multivariate statistical models in
132. Dehghani, M.; Saghafian, B.; Rivaz, F.; Khodadadi, A. Evaluation of dynamic regression and
artificial neural networks models for real-time hydrological drought forecasting. Arabian J.
133. Choubin, B.; Zehtabian, G.; Azareh, A.; Rafiei-Sardooi, E.; Sajedi-Hosseini, F.; Kişi, Ö.
comparative study of different approaches. Environ. Earth Sci. 2018, 77, 314.
134. Choubin, B.; Darabi, H.; Rahmati, O.; Sajedi-Hosseini, F.; Kløve, B. River suspended sediment
modelling using the cart model: A comparative study of machine learning techniques. Sci. Total
135. Liaw, A.; Wiener, M. Classification and regression by randomforest. R News 2002, 2, 18–22.
136. Wang, Z.; Lai, C.; Chen, X.; Yang, B.; Zhao, S.; Bai, X. Flood hazard risk assessment model based
137. Tehrany, M.S.; Pradhan, B.; Jebur, M.N. Flood susceptibility mapping using a novel ensemble
weights-of-evidence and support vector machine models in GIS. J. Hydrol. 2014, 512, 332–343.
138. Bui, D.T.; Tuan, T.A.; Klempe, H.; Pradhan, B.; Revhaug, I. Spatial prediction models for
machines, artificial neural networks, kernel logistic regression, and logistic model tree.
139. Etemad-Shahidi, A.; Mahjoobi, J. Comparison between m5′ model tree and neural networks for
prediction of significant wave height in lake superior. Ocean Eng. 2009, 36, 1175–1181.
140. Dietterich, T.G. Ensemble methods in machine learning. In International Workshop on Multiple
A novel machine learning-based approach for the risk assessment of nitrate groundwater
142. Moore, K.J.; Kurt, M.; Eriten, M.; McFarland, D.M.; Bergman, L.A.; Vakakis, A.F. Wavelet-
bounded empirical mode decomposition for measured time series analysis. Mech. Syst. Signal
143. Wang, W.-C.; Chau, K.-W.; Xu, D.-M.; Chen, X.-Y. Improving forecasting accuracy of annual
runoff time series using ARIMA based on EEMD decomposition. Water Resour. Manag. 2015, 29,
2655–2675.
144. Al-Musaylh, M.S.; Deo, R.C.; Li, Y.; Adamowski, J.F. Two-phase particle swarm optimized-
support vector regression hybrid model integrated with improved empirical mode
decomposition with adaptive noise for multiple-horizon electricity demand forecasting. Appl.
145. Ouyang, Q.; Lu, W.; Xin, X.; Zhang, Y.; Cheng, W.; Yu, T. Monthly rainfall forecasting using
EEMD-SVR based on phase-space reconstruction. Water Resour. Manag. 2016, 30, 2311–2325.
146. Zhang, J.; Hou, G.; Ma, B.; Hua, W. Operating characteristic information extraction of flood
discharge structure based on complete ensemble empirical mode decomposition with adaptive
147. Badrzadeh, H.; Sarukkalige, R.; Jayawardena, A. Hourly runoff forecasting for flood risk
1633–1643.
148. Kim, G.; Barros, A.P. Quantitative flood forecasting using multisensor data and neural
149. Saghafian, B.; Haghnegahdar, A.; Dehghani, M. Effect of ENSO on annual maximum floods and
volume over threshold in the southwestern region of Iran. Hydrol. Sci. J. 2017, 62, 1039–1049.
150. Kourgialas, N.N.; Dokou, Z.; Karatzas, G.P. Statistical analysis and ann modeling for predicting
hydrological extremes under climate change scenarios: The example of a small Mediterranean
151. Panda, R.K.; Pramanik, N.; Bala, B. Simulation of river stage using artificial neural network and
coefficient using support vector machine and adaptive neuro-fuzzy inference system
153. Pereira Filho, A.J.; dos Santos, C.C. Modeling a densely urbanized watershed with an artificial
neural network, weather radar and telemetric data. J. Hydrol. 2006, 317, 31–48.
154. Jingyi, Z.; Hall, M.J. Regional flood frequency analysis for the Gan-Ming River basin in China.
155. Ahmad, S.; Simonovic, S.P. An artificial neural network model for generating hydrograph from
156. Ju, Q.; Yu, Z.; Hao, Z.; Ou, G.; Zhao, J.; Liu, D. Division-based rainfall-runoff simulations with
157. Sahoo, G.B.; Ray, C.; De Carlo, E.H. Use of neural network to predict flash flood and attendant
water qualities of a mountainous stream on Oahu, Hawaii. J. Hydrol. 2006, 327, 525–538.
158. Ghose, D.K. Measuring Discharge Using Back-Propagation Neural Network: A Case Study on
159. Pan, H.-X.; Cheng, G.-J.; Cai, L. Comparison of the extreme learning machine with the support
vector machine for reservoir permeability prediction. Comput. Eng. Sci. 2010, 2, 37.
160. Chang, F.-J.; Chen, P.-A.; Lu, Y.-R.; Huang, E.; Chang, K.-Y. Real-time multi-step-ahead water
level forecasting by recurrent neural networks for urban flood control. J. Hydrol. 2014, 517, 836–
846.
161. Shen, H.-Y.; Chang, L.-C. Online multistep-ahead inundation depth forecasts by recurrent
162. Bruen, M.; Yang, J. Functional networks in real-time flood forecasting—A novel application.
163. Chiang, Y.-M.; Chang, F.-J.; Jou, B.J.-D.; Lin, P.-F. Dynamic ANN for precipitation estimation
164. Bhattacharya, B.; Solomatine, D.P. Neural networks and M5 model trees in modelling water
165. Heiser, M.; Scheidl, C.; Eisl, J.; Spangl, B.; Hübl, J. Process type identification in torrential
modeling at haraz watershed, Northern Iran. Sci. Total Environ. 2018, 627, 744–755.
167. Aichouri, I.; Hani, A.; Bougherira, N.; Djabri, L.; Chaffai, H.; Lallahem, S. River flow model
168. Torabi, M.; Hashemi, S.; Saybani, M. R.; Shamshirband, S.; Mosavi, A. A Hybrid clustering and
classification technique for forecasting short‐term energy consumption. Environ. Prog. Sustain.
169. Adamowski, J.F. Development of a short-term river flood forecasting method for snowmelt
driven floods based on wavelet and cross-wavelet analysis. J. Hydrol. 2008, 353, 247–266.
170. Leahy, P.; Kiely, G.; Corcoran, G. Structural optimisation and input selection of an artificial
neural network for river level prediction. J. Hydrol. 2008, 355, 192–201.
171. Wei, C.C. Soft computing techniques in ensemble precipitation nowcast. Appl. Soft Comput. J.
172. R Schiffer, R. A.; Rossow, W. B. The International Satellite Cloud Climatology Project (ISCCP):
The first project of the world climate research programme. Bulletin of the American Meteorological
174. Jimeno-Sáez, P.; Senent-Aparicio, J.; Pérez-Sánchez, J.; Pulido-Velazquez, D.; María Cecilia, J.
Estimation of instantaneous peak flow using machine-learning models and empirical formula
175. Chang, F.-J.; Chang, Y.-T. Adaptive neuro-fuzzy inference system for prediction of water level
176. Mosavi, A.; Lopez, A.; Varkonyi-Koczy, A. R. Industrial Applications of Big Data: State of the
Art Survey. In Recent Advances in Technology Research and Education; Springer: Cham,
177. Rezaeianzadeh, M.; Tabari, H.; Yazdi, A.A.; Isik, S.; Kalin, L. Flood flow forecasting using ANN,
ANFIS and regression models. Neural Comput. Appl. 2014, 25, 25–37.
178. Tehrany, M.S.; Pradhan, B.; Jebur, M.N. Flood susceptibility analysis and its verification using
a novel ensemble support vector machine and frequency ratio method. Stoch. Environ. Res. Risk
for accurate rainfall-runoff modeling during extreme typhoon events. Appl. Soft Comput. J. 2017,
53, 205–216.
180. Altunkaynak, A.; Nigussie, T.A. Prediction of daily rainfall by a hybrid wavelet-season-neuro
181. Chang, L.-C.; Shen, H.-Y.; Chang, F.-J. Regional flood inundation nowcast using hybrid SOM
182. Nanda, T.; Sahoo, B.; Beria, H.; Chatterjee, C. A wavelet-based non-linear autoregressive with
exogenous inputs (WNARX) dynamic neural network model for real-time flood forecasting
183. French, J.; Mawdsley, R.; Fujiyama, T.; Achuthan, K. Combining machine learning with
184. Hong, W.-C. Rainfall forecasting by technological machine learning models. Appl. Math.
185. Pan, T.-Y.; Yang, Y.-T.; Kuo, H.-C.; Tan, Y.-C.; Lai, J.-S.; Chang, T.-J.; Lee, C.-S.; Hsu, K.H.
Improvement of watershed flood forecasting by typhoon rainfall climate model with an ANN-
186. Rajurkar, M.; Kothyari, U.; Chaube, U. Modeling of the daily rainfall-runoff relationship with
187. Hsu, M.-H.; Lin, S.-H.; Fu, J.-C.; Chung, S.-F.; Chen, A.S. Longitudinal stage profiles forecasting
188. Pang, B.; Guo, S.; Xiong, L.; Li, C. A nonlinear perturbation model based on artificial neural
189. Doycheva, K.; Horn, G.; Koch, C.; Schumann, A.; König, M. Assessment and weighting of
application to runoff simulations and flood warning. Adv. Eng. Inform. 2017, 33, 427–439.
190. Fleming, S.W.; Bourdin, D.R.; Campbell, D.; Stull, R.B.; Gardner, T. Development and
Iran, 2016.
192. Kisi, O.; Sanikhani, H. Prediction of long-term monthly precipitation using several soft
computing methods without climatic data. Int. J. Climatol. 2015, 35, 4139–4150.
193. Liang, Z.; Li, Y.; Hu, Y.; Li, B.; Wang, J. A data-driven SVR model for long-term runoff
prediction and uncertainty analysis based on the Bayesian framework. Theor. Appl. Climatol.
194. Han, S.; Coulibaly, P. Bayesian flood forecasting methods: A review. J. Hydrol. 2017, 551, 340–
351.
195. Choubin, B.; Moradi, E.; Golshan, M.; Adamowski, J.; Sajedi-Hosseini, F.; Mosavi, A. An
classification and regression trees, and support vector machines. Elsevier Sci. Total Environ. 2018,
651, 2087–2096
196. Teng, J.; Jakeman, A.; Vaze, J.; Croke, B.F.; Dutta, D.; Kim, S. Flood inundation modelling: A
review of methods, recent advances and uncertainty analysis. Environ. Model. Softw. 2017, 90,
201–216.
197. Elsafi, S.H. Artificial neural networks (ANNs) for flood forecasting at Dongola station in the
198. Mohammadzadeh, D.; Bazaz, J. B.; Yazd, S. V. J.; Alavi, A. H. Deriving an intelligent model for
soil compression index utilizing multi-gene genetic programming. Springer Environ. Earth Sci.
199. Singh, P.; Borah, B. Indian summer monsoon rainfall prediction using artificial neural network.
200. Shamim, M.A.; Hassan, M.; Ahmad, S.; Zeeshan, M. A comparison of artificial neural networks
(ANN) and local linear regression (LLR) techniques for predicting monthly reservoir levels.
201. Rezaeian-Zadeh, M.; Tabari, H.; Abghari, H. Prediction of monthly discharge volume by
different artificial neural network algorithms in semi-arid regions. Arabian J. Geosci. 2013, 6,
2529–2537.
202. Gazendam, E.; Gharabaghi, B.; Ackerman, J.D.; Whiteley, H. Integrative neural networks
models for stream assessment in restoration projects. J. Hydrol. 2016, 536, 339–350.
203. Lin, J.-Y.; Cheng, C.-T.; Chau, K.-W. Using support vector machines for long-term discharge
204. Cunningham, S.C.; Griffioen, P.; White, M.D.; Nally, R.M. Assessment of ecosystems: A system
for rigorous and rapid mapping of floodplain forest condition for Australia’s most important
205. Ahmad, S.; Kalra, A.; Stephen, H. Estimating soil moisture using remote sensing data: A
206. Prasad, R.D.; Ravinesh, C.; Li, Y.; Maraseni, T. Input selection and performance optimization of
ANN-based streamflow forecasts in the drought-prone murray darling basin region using IIS
207. Cannas, B.; Fanni, A.; Sias, G.; Tronci, S.; Zedda, M.K. River flow forecasting using neural
208. Najafi, B. An Intelligent Artificial Neural Network-Response Surface Methodology Method for
Accessing the Optimum Biodiesel and Diesel Fuel Blending Conditions in a Diesel Engine from
the Viewpoint of Exergy and Energy Analysis. Energies 2018, 11, 860.
209. Singh, R.M. Wavelet-ANN model for flood events. In Proceedings of the International
Conference on Soft Computing for Problem Solving (SocProS 2011), Patiala, India, 20–22
210. Ramana, R.V.; Krishna, B.; Kumar, S.R.; Pandey, N.G. Monthly rainfall prediction using wavelet
211. Tantanee, S.; Patamatammakul, S.; Oki, T.; Sriboonlue, V.; Prempree, T. Coupled wavelet-
autoregressive model for annual rainfall prediction. J. Environ. Hydrol. 2005, 13, 124–146.
212. Mekanik, F.; Imteaz, M.A.; Talei, A. Seasonal rainfall forecasting by adaptive network-based
fuzzy inference system (ANFIS) using large scale climate signals. Clim. Dyn. 2016, 46, 3097–
3111.
213. Wang, W.C.; Chau, K.W.; Cheng, C.T.; Qiu, L. A comparison of performance of several artificial
intelligence methods for forecasting monthly discharge time series. J. Hydrol. 2009, 374, 294–
306.
214. Kisi, O.; Nia, A.M.; Gosheh, M.G.; Tajabadi, M.R.J.; Ahmadi, A. Intermittent streamflow
forecasting by using several data driven techniques. Water Resour. Manag. 2012, 26, 457–474.
215. Li, C.; Guo, S.; Zhang, J. Modified NLPM-ANN model and its application. J. Hydrol. 2009, 378,
137–141.
216. Huang, S.; Chang, J.; Huang, Q.; Chen, Y. Monthly streamflow prediction using modified EMD-
217. Zhu, S.; Zhou, J.; Ye, L.; Meng, C. Streamflow estimation by support vector machine coupled
with different methods of time series decomposition in the upper reaches of Yangtze River,
218. Bass, B.; Bedient, P. Surrogate modeling of joint flood risk across coastal watersheds. J. Hydrol.
219. Araghinejad, S.; Azmi, M.; Kholghi, M. Application of artificial neural network ensembles in
220. Tan, Q.-F.; Lei, X.-H.; Wang, X.; Wang, H.; Wen, X.; Ji, Y.; Kang, A.-Q. An adaptive middle and
long-term runoff forecast model using EEMD-ANN hybrid approach. J. Hydrol. 2018,
doi:10.1016/j.jhydrol.2018.01.015.
221. Nosratabadi, S., Mosavi, A., Shamshirband, S., Kazimieras Zavadskas, E., Rakotonirainy, A.
and Chau, K.W., 2019. Sustainable business models: A review. Sustainability, 11(6), p.1663.
222. Høverstad, B.A.; Tidemann, A.; Langseth, H.; Öztürk, P. Short-term load forecasting with
seasonal decomposition using evolution for parameter tuning. IEEE Trans. Smart Grid 2015, 6,
1904–1913.
223. Tantithamthavorn, C.; McIntosh, S.; Hassan, A.E.; Matsumoto, K. Automated parameter
2016 IEEE/ACM 38th International Conference on Software Engineering (ICSE), Austin, TX,
224. Varkonyi-Koczy, A.R. Review on the usage of the multiobjective optimization package of
modefrontier in the energy sector. In Recent Advances in Technology Research and Education;
225. Dineva, A.; Várkonyi-Kóczy, A.R.; Tar, J.K. Anytime fuzzy supervisory system for signal auto-
healing. In Advanced Materials Research; Trans Tech Publications: Tihany, Hungary, 2015; pp.
269–272.
226. Torabi, M.; Mosavi, A.; Ozturk, P.; Varkonyi-Koczy, A.; Istvan, V. A hybrid machine learning
approach for daily prediction of solar radiation. In Recent Advances in Technology Research and
wavelet-artificial neural network and comparison with adaptive neurofuzzy inference system
(case study: Verayneh station, Nahavand). Adv. Civ. Eng. 2014, 2014, 279368.
228. Badrzadeh, H.; Sarukkalige, R.; Jayawardena, A. Improving ann-based short-term and long-
term seasonal river flow forecasting with signal processing techniques. River Res. Appl. 2016,
32, 245–256.