Application of Numerical Hydrodynamic Modeling for Flood Mapping and Simulation of the Euphrates River

Floods are ranked among the most destructive natural hazards globally, imposing severe threats to human lives and infrastructure [1]. In arid and semi-arid regions, such as the Euphrates River basin, one of the most important water resources in Iraq, these risks are exacerbated by hydrological instability and the scarcity of historical data for accurate forecasting [2]. Climate change has increased hydrological extremes, including droughts and floods, placing significant pressure on the operation of the main control structure of the Euphrates River, Haditha Dam [3–5]. Sometimes, droughts can last for up to three years, followed by a rainy year that may lead to floods, causing disasters, significant economic losses, and the displacement of residents from their homes [2, 6]. For example, the city of Al-Khalidiya in Anbar Province was flooded as a result of heavy rains on March 26, 2026, after seasons of drought (Figure 1).

Although dams such as Haditha Dam are vital for managing water resources, the possibility of dam failure or emergency releases presents an existential risk to downstream communities [7, 8]. Reservoir operations and changing climatic flow regimes play a significant role in the dynamics of the Euphrates River and need to be integrated into complex hydraulic models [9, 10].

The development of hydrodynamic numerical models has evolved significantly over the past decades. Early approaches relied on simplified analytical equations and field observations, which were limited in representing complex flood dynamics. With the advancement of computational techniques in the 1970s and 1980s, one-dimensional (1D) numerical models based on the Saint-Venant equations were introduced, enabling improved simulation of river flow. Subsequently, the 1990s and 2000s witnessed the emergence of two-dimensional (2D) models such as Hydrologic Engineering Center- River Analysis System (HEC-RAS) and MIKE21, which allowed a more realistic representation of floodplain processes, including flow depth, velocity, and spatial inundation extent. The integration of these models with geographic information systems (GIS) further enhanced their capability to produce accurate flood hazard maps and assess risks to populated areas.

These models are among the most useful tools for simulating unsteady flow and studying dam-failure scenarios or unexpected discharges [11, 12]. Therefore, many researchers have used these models and techniques to analyze river floods and generate flood maps to support decision-making, emergency planning, and risk analysis.

In recent years, the field has rapidly progressed toward high-resolution modeling and the integration of artificial intelligence techniques. Modern approaches combine physics-based hydrodynamic models with data-driven methods, significantly improving computational efficiency and predictive accuracy. The growing importance of these models is driven by increasing climate variability, rapid urbanization in flood-prone areas, and the urgent need for effective flood risk management and early warning systems. Their ability to simulate multiple scenarios, integrate spatial data, and support decision-making has made them indispensable tools for researchers and practitioners worldwide.

Flood hazard assessment has also evolved substantially through the integration of hydrodynamic numerical models and GIS. Recent studies, such as Mustafa et al. [13], demonstrated the effectiveness of coupling HEC-RAS 2D models with GIS datasets, including digital elevation models (DEMs), land-use data, and soil maps, to delineate flood-prone urban areas in Iraq. Similarly, hydrodynamic simulations conducted along the Euphrates River in Al-Anbar showed high predictive accuracy in estimating flow depth and velocity under various discharge scenarios [14]. Furthermore, dam-break simulations highlighted the severe flood risks associated with potential failure events, emphasizing the need for reliable flood inundation mapping [15]. Globally, the field has shifted toward integrating artificial intelligence with physics-based hydrodynamic models, as seen in recent frameworks such as FloodCast [16] and DL Hydro-FRAN [14], which significantly enhance prediction speed and accuracy.

Based on the above review, despite these advancements, limited studies have focused on high-resolution flood mapping along the Euphrates River, particularly in populated areas, indicating a critical research gap that this study aims to address. There is a clear lack of recent studies after 2023 specifically on the Euphrates River in Anbar Governorate. In addition, few studies have linked flooding to urban exposure, and the application of high-resolution 2D models along the entire river remains limited.

In this research, the research methodology used is ‎integrated research methodology which incorporates ‎descriptive, analytical and quantitative research ‎methodologies to evaluate and determine the risks of the ‎flood along Euphrates River in the study area which is ‎between Haditha Dam and ‎Heet City. Such a process requires accurate data gathering, ‎analysis, and interpretation of digital and descriptive data to ‎manage flood risks. The Input data to the HEC-‎RAS model includes the longitudinal river section ‎‎(geometric data), ‎which is drawn with (Ras‎ Mapper) option and based on DEM as a ‎precise topographical foundation. Then the coordinates of ‎the river cross sections projected on the longitudinal river section ‎along the river. ‎After ‎completing the geometric data, the boundary conditions were entered, ‎where ‎the discharges were entered in U/S and the normal depth in D/S based on observed data. After that, ‎the ‎model calibration begins. The flood wave ‎propagation will be analyzed with the help of HEC-RAS 2D ‎‎(version 6.6 or higher), which will calculate inundation areas ‎and create flood hydrographs to find out the water surface ‎elevation, depth, and velocity of the flood-affected areas. ‎The obtained results will be then combined with Geographic ‎Information Systems (ArcGIS) (version 10.5 or more) with ‎the help of such tools as Geo-RAS to create in-depth flood ‎maps and identify and classify high-risk zones and create a ‎spatial reference database to help manage flood risks and ‎sustainable planning in the area. Figure 3 shows the ‎approach of the current research.‎

The geometric input data were prepared using a DEM with a spatial resolution of 12.5 m within a GIS environment. Longitudinal and cross-sectional profiles were extracted using Hydrologic Engineering Center- Geographic River Analysis System (HEC-GeoRAS). Boundary conditions were obtained from the operational records of Haditha Dam and the gauging station in Hit. Manning’s roughness coefficients were adopted from previous studies and further refined through calibration using observed data from the Water Resources Directorate in Al-Anbar. This approach ensures accurate flow representation and reliable flood mapping results and Table 1 shows input data for HEC-RAS 2D model for flood mapping.

Figure 3. Methodology of the present study

Table 1. Input data for HEC-RAS 2D model for flood mapping

Note: Hydrologic Engineering Center- River Analysis System (HEC-RAS); Hydrologic Engineering Center- Geographic River Analysis System (HEC-GeoRAS); Geographic Information Systems (GIS); digital elevation model (DEM)

The course of the Wadi Hawran is narrowed by the river's longitudinal cross-section, while its discharges are considered as lateral flow in boundary conditions. Data provided by the Iraqi Ministry of Water Resources regarding the Wadi Hawran's discharges were used.

Calibration:

It is the process for adjusting input parameters to minimize the difference between ‎model simulations and observed data, ensuring the model accurately represents a real-world system. It ‎involves comparing numerical results against observed data, often focusing on parameter refinement to ‎improve predictive accuracy, in present study the calibration is achieved for determination of Manning coefficient for the Euphrates River in studied area.

Several previous studies indicate that the Manning roughness index in rivers, such as the Euphrates, ranges between [values to be inserted here]. Based on available observational data obtained from the Anbar Governorate Water Resources Directorate, the official government body responsible for managing the Euphrates River, these values were tested for their consistency with the observed data in a section of the river downstream from the city of Hit. Table 2 shows the results of the calibration process, indicating that the roughness index value closest to the observed data is (0.032). ‎

Table 2. Results of calibration

Validation:

It is the process of evaluation the model’s predictive capabilities against independent data, proving the ‎calibrated model holds true in a new scenario.‎ In numerical modeling used in hydraulic and water resources engineering, validation is among the most ‎important stages of model evaluation before using it for forecasting or decision support.‎ In present study based on the available data, validation or verifying the accuracy of the numerical model (HEC-RAS) is achieved by comparing the model's ‎results with actual or observed data. ‎Statistical indicators are used to measure model accuracy using Nash-Sutcliffe Efficiency (NSE) formula or (Eq. (1)) and Table 3 represent the Nash-Sutcliffe Efficiency Rating.

$N S E=1-\frac{\sum\left(Q_{o b s}-Q_{s i m}\right)^2}{\sum\left(Q_{o b s}-\ddot{Q}_{o b s}\right)^2}$         (1)

where, Q_obs = observed value of (discharge or water level), Q_sim = simulated value of (discharge or water level), $\ddot{Q}_{o b s}$ = average observed value of (discharge or water level).

The equation is applied to measure accuracy for determination both discharges (Q) and water level (H) (see Table 4). The accuracy or Efficiency for discharge is (0.99), and for water level (0.83). These results indicate a high degree of accuracy in representing the numerical model of both the discharge and the water level.

Table 3. Nash-Sutcliffe efficiency

Table 4. Observed and simulated values of Q and H

The results demonstrate a nonlinear increase in inundation extent with return period. Specifically, inundated areas expand as the return periods increase from 5 to 50 and 100 years, indicating the high sensitivity of the Euphrates floodplain to extreme discharge events. The limited channel conveyance capacity of the Euphrates River further contributes to this pattern, as the main channel is insufficient to convey high-return-period discharges, resulting in overbank flooding and expansion into adjacent urban areas.

The study also confirms the superiority of unsteady flow modeling, which provides a more realistic representation of flood dynamics, including water depth, velocity, and peak arrival time, compared with the steady flow model. In addition, the findings highlight the significant impact of climate change, suggesting that intensified hydrological extremes have increased the likelihood and severity of flood events.

Urban areas show high vulnerability, as a considerable portion of the urban land located within the floodplain is highly susceptible to inundation, particularly under high-return-period scenarios. The influence of topography and river morphology is also evident, with the low longitudinal slope and wide cross-sections contributing significantly to lateral flood expansion and larger inundation areas.

Furthermore, the discharge–inundation relationship is strongly nonlinear, meaning that higher discharges cause disproportionately larger flooded areas. The integrated hydraulic–GIS approach proved effective in generating accurate flood hazard maps and identifying high-risk zones. Finally, the study emphasizes the need for periodic flood hazard updates to ensure that flood maps reflect changing climatic conditions and land-use dynamics.