See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/391140800

SWAT-MODFLOW Model for Groundwater Recharge Variation Assessment:

Lower Zab River Basin, Northeastern Iraq

Article in International Journal of Computational and Experimental Science and Engineering · April 2025
DOI: 10.22399/ijcesen.1674

CITATIONS READS

0 116

2 authors:

Mohammed Jallel Al- Janabi Ruqayah Kadhim Mohammed

University of Babylon University of Babylon, Hilla, Iraq and Salford University, Manchester, UK
2 PUBLICATIONS 0 CITATIONS 59 PUBLICATIONS 446 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Mohammed Jallel Al- Janabi on 09 May 2025.

The user has requested enhancement of the downloaded file.

 International Journal of Computational and Experimental
Science and ENgineering
(IJCESEN)
Vol. 11-No.2 (2025) pp. 2494-2504
http://www.ijcesen.com
Copyright © IJCESEN ISSN: 2149-9144
Research Article

SWAT-MODFLOW Model for Groundwater Recharge Variation Assessment:

Lower Zab River Basin, Northeastern Iraq
Mohammed Jalel Rahem1*, Ruqayah Mohammed2
1
Civil Engineering Department, Faculty of Engineering, The University of Babylon, Hilla, Iraq
* Corresponding Author Email: [email protected] - ORCID: 0000-0002-5247-7850
2
Civil Engineering Department, Faculty of Engineering, The University of Babylon, Hilla, Iraq
Email: [email protected] - ORCID: 0000-0002-5247-7850

Article Info: Abstract:

DOI: 10.22399/ijcesen.1674 For surface and groundwater simulations, recharge is a fundamental water balance unit.
Received : 21 February 2025 Yet, measuring its regional spatiotemporal variation poses a substantial challenge.
Accepted : 05 April 2025 Mathematical and empirical simulation are the most frequently used approaches at the
basin scales. However, the accuracy and dependability of integrated models may be
Keywords : limited due to the limitless number of unknowns and uncertainties they contain. In the
Lesser Zab River Basin, A QSWATMOD variant was suggested for evaluating
Basin management, groundwater recharge in northeastern Iraq. The paired version was calibrated using a
Lesser Zab Basin,
Land use/land cover,
hydraulic head and daily circulating drift. In comparison to the SWAT version, the
Sustainable Management, QSWATMOD version performed well throughout the flow drift simulation. The results
Variation detection. of the study verified that during the predicted period, the watershed saw significant
fluctuations in groundwater recharge. The wet season is when a significant amount of
recharge takes place. It makes a significant contribution to the region's average yearly
precipitation. The water stability components were assessed locally, and the direct waft
additives (lateral and surface) demonstrated significant contributions. All things
considered, the QSWATMOD model predicted groundwater recharge in the study area
with a respectable degree of accuracy.

1. Introduction method to measure groundwater recharge, which

entails examining aquifer and surface conditions or
Because of the increasing demand for water both. [18,19].
supplies, climate change, and poor resource The MODFLOW model is commonly used to
management, sustainable development and resource analyze unconfined groundwater flow, whereas
management are essential in watersheds with both surface water flow is often simulated using the
agricultural and urban land use [1,2,3,4]. To SWAT model. SWAT is widely used across various
preserve and oversee the water resources within the catchment scales to assess surface and groundwater
watershed, a comprehensive Understanding The quality and quantity and to forecast the impacts of
hydrological cycle and the system that goes along specific catchments on land use/cover and climate
with it are extremely significant. [5,6]. Watershed change.
recharge-discharge, water quality, and water supply While SWAT focuses on surface processes, its
are all impacted [6,7,8]. Accurate estimation of aggregated nature limits its ability to model
water balance components is also crucial [5,6,9]. groundwater flow. Several tactics have been carried
Therefore, Assessing groundwater recharge in the out to cope with the original SWAT groundwater
reservoir under study is the primary goal of this module's shortcomings. Alternatively, the
research project. Many methods, including physical MODFLOW model, with its distributed and
[10,11,12], tracer [13,14,15], and numerical physically primarily based method, is a good option
approaches [16,17,9], were used in earlier studies to for modeling groundwater flow considering various
assess groundwater recharge. When evaluating boundary situations. However, it's far inadequate
recharging, each approach has drawbacks [11]. for modeling floor water float. The coupled
More and more researchers are using the numerical QSWATMOD approach is essential for resolving
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

the issues of those fashions and offering a greater parametrization tools. in recent hydrological
correct depiction of the hydrological device. investigations [31, 32]. In the northern Danish
The integration of SWAT and MODFLOW Ogerby River Basin, [31] employed the approach of
fashions was first of all delivered through a standardization technique for a coupled model to
Sophocleous et al. [20] and has considering that assess how water flow responded to groundwater
been enhanced through version overall performance withdrawal. An important benefit of QSWATMOD
comparisons with the SWAT version [18,21]. is its ability to simulate coupled surface-
Bailey et al.' integrated model framework uses the groundwater scenarios. And has demonstrated high
SWAT model to model surface hydrological precision in streamflow evaluation [25,31,22].
approaches. [22] developed. Newton-Raphson However, like other numerical models,
formulation for MODFLOW, or MODFLOW- QSWATMOD is subject to uncertainty [33],
NWT [23], which also can address unconfined particularly in regional-scale environmental
aquifer subsurface water flow problems, is utilized models, which frequently experience a great deal of
to simulate groundwater flow processes. Previous uncertainty because of many unknown factors, with
studies using the QSWATMOD model, as water table fluctuation (WTF) being one of the
described by Wang and Chen [24], are divided into most direct, precise, and uncomplicated approaches
two categories: situational and non-situational [11,34,35,36]. This approach depends on aquifer
simulations. Although QSWATMOD situational parameters, which can be ascertained in a variety of
scenario simulation studies consider how ways, and observed groundwater level data series
groundwater abstraction affects surface water [11,35,37]. Enhancements have been made to boost
resources [25], Simulations of water flow, efficiency and decrease subjectivity [38,39].
groundwater discharge, surface-regional water Chung et al. [40] investigated the recharge of the
balance, and groundwater interaction are the main groundwater reservoir in South Korea. One
areas of interest for non-situational investigations. technique for integrated modeling is to connect a
[18,21,22]. According to Ntona et al. [26], hydrological model, such as SWAT, with the WTF
QSWATMOD has been widely used as a model for method as a model in Jeju Island. Using the SWAT
analyzing surface–groundwater interactions model in combination with TWTFM. Since it
worldwide from 1992 to 2020 and remains essential guarantees accurate results, it is advised to consider
in current research. In Africa's Limpopo River an alternate method for predicting groundwater
Basin, for example, Mosase et al. [27] used recharge [35,36,41]. We use the QSWATMOD
QSWATMOD to estimate the water table level, model to evaluate groundwater recharge in our
surface and subsurface water interaction, and the study area. To the best of our knowledge, no
spatiotemporal distribution of recharge. previous study has evaluated groundwater recharge
Additionally, Sophocleous and Perkins [28] using a coupled model; this study will be the first to
employed QSWATMOD to explore the impacts of do so.
water abstraction on water table levels and Such as parameterization choices, input databases,
streamflow in three basins in Kansas, USA, while and model structure [24]. For example, during the
also analyzing various scenarios, including how wet season, a big recharge value is transferred to
water resources are affected by climate change. MODFLOW by the surface water model (SWAT),
Gao et al. [21] used an integrated model to assess according to Guevara-Ochoa et al. [34]. Due to the
the spatiotemporal variations of surface water absence of a module in QSWATMOD to account
resources in a U.S. River catchment. Additionally, for the interaction between groundwater and soil
Taye Samiromi and Koush [18] investigated the saturation. While various techniques are used for
relationship between surface and groundwater in estimating recharge, the water Another strategy that
Iranian agricultural watersheds using the integrated is suggested is the table fluctuation method.
model highlighting the fact that the primary cause Estimating groundwater recharge involves various
of water scarcity in the study area is excessive methods,
groundwater use, not just climate change. Chunn et The present study aims to put forward quantitative
al. [29] evaluated the effects of groundwater and observational techniques for assessing
abstractions and climate change on the groundwater replenishment's spatial and temporal
groundwater-surface water relationship in western pattern. Using the QSWATMOD model, the
Canada using the same model. Highlighting the research basin assesses the temporal and spatial
important role that groundwater misuse plays in distribution of groundwater recharge. This study
lowering groundwater levels and river flow, as also presents the seasonal fluctuations in
mentioned by [18,30]. To lower the uncertainty groundwater recharge in the area. Additionally, it
associated with parameter calibration, The acquires and discusses the water balance
QSWATMOD model has been subjected to components.
2495
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

2. Materials and methods area, such as elevation accurate to 30 meters and

the specific SRTM type, were acquired from the
2.1. Basin description USGS (US Geological Survey)
(http://earthexplorer.usgs.gov). Subsequently, using
Between latitudes 43° 21′ 41" and 46º 17′ 55" N the ARC-GIS application, the surrounding areas
and 35° 1′ 29" and 36º 54′ 41" E is where the were eliminated, and the final representation
Lower Zab Basin is located. According to Figure 1, validating the watershed's geometry was created, as
the Lower Zab River's smaller portion is in Iran depicted in Figure 1. The Lesser Zab River serves
while its greater portion is in northeastern Iraq. The as the primary tributary of the Tigris River. In the
entire basin extent is 19700.845 Km2, 74.77% is Iraqi portion of the basin, the largest dam is the
placed inside Iraq which is 14729.690 km2 and Dokan Dam (35°57'14" N and 44°57'10'' E), while
25.23 % is placed inside Iran which is 4970.310 Iran is currently constructing one dam and planning
km2. The topographic characteristics of the study to build two more [42].

Figure 1. Description of the study area: flow network, digital elevation model (DEM), flow gauges, weather station,
and rainfall station

The primary water sources for the stream include Zagros region. With annual average air
rainwater, snowmelt, and various springs, resulting temperatures ranging from 14.5°C to 18.5°C,
in great discharge rates in the spring period and average temperatures show a similar gradient.
lower rates in the summer. The weather within the Numerous unique lithostratigraphic units, ranging
basin differs, ranging from semi-arid conditions in from the Precambrian to the present, can be found
the northern and northeastern regions to arid in the LZB, including sedimentary, metamorphic,
climates in the southern and southwestern areas. igneous, and Quaternary deposits. The landforms in
The LZR exhibits the lowest, average, and highest the catchment show significant topographic
discharges of 6, 227, and 3,420 m³/s, in that order variances that affect the moisture content and soil
[43]. The basin traverses a range of geological, chemistry. The land use and land cover within the
environmental, and climatic areas. The upper and LZB mainly consist of bare lands and flooded
lower regions of the catchment are predominantly vegetation, with bare lands accounting for
categorized by a mix of different rock types with approximately 70% of the total area and flooded
various lithological characteristics and more easily vegetation comprising about 15%. The remaining
eroded clastic rocks, respectively. Annual rainfall 15% includes crops, water bodies, trees, and areas
along the river differs considerably variation, designated for agriculture. The SWAT model
ranging from less than 20 cm at its Iraqi Tigris classified the land use into eight distinct categories,
River confluence to more than 1000 mm in Iran's as demonstrated in Figure 2.
2496
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

2.2. SWAT-MODFLOW model layers, rows, and columns. Models simulate

groundwater heads down to the grid cell level.
Due to its aggregated nature, the SWAT model is Once QSWATMOD has been configured, for
limited in its ability to address groundwater flow, recharging scenarios, the MODFLOW groundwater
and MODFLOW finds it challenging to estimate flow equation will be utilized and the groundwater
surface water conditions (Figure 2). SWAT module will be turned off.

∂⁄∂x∙ (K_xx∙ ∂h⁄∂x) + ∂⁄∂y∙ (K_yy∙ ∂h⁄∂y) +

∂⁄∂z∙ (K_zz∙ ∂h⁄∂z)-W=S_S ∙ ∂⁄∂t (2)

The hydraulic head (h) is measured in meters. The

hydraulic conductivity along the x, y, and z
directions is denoted by Kxx, Kyy, and Kzz,
respectively, and is measured in meters per day. Ss
represents the dimensionless specific storage of the
aquifer, t (measured in days) denotes time, and W
(measured in days) indicates the source or sink. A
positive W value signifies. In contrast, a negative
value denotes abstraction. Recharge into the
aquifer.
Figure 2. The lower Zab River Basin land use and land
cover (LULC) classification 2.3. Configuring and linking models

Because the groundwater model does not have a The SWAT model
land surface hydrology model, the lower Zab River The research area was constructed using SWAT
Basin land use and land cover (LULC) (version 2012). The SWAT model operates on a
classification must be used. To get around these daily time series and is semi-distributed.
restrictions, the integrated QSWATMOD model Watersheds and sub-watersheds (sub-basins) were
substitutes MODFLOW for the SWAT model established through the use of a digital elevation
module for groundwater [24,29]. Data from other model (DEM). Hydrological Response Units
models must be transferred for the QSWATMOD (HRUs) are conceptual units that have been used to
model to calculate groundwater recharge divide the landscape for this model. An area with
accurately. It has been possible to transfer data comparable soil, land use, slope, and management
between individual models by using a mapping features is designated as a piece HRU. The SWAT
scheme Model-simulated deep percolation (HRUs) model uses conceptual units (HRUs) to simulate the
are used as recharge to the detailed HRUs amount and quality of water resources [44]. A
(DHRUs) in MODFLOW for each grid cell, crucial first step is determining the appropriate
whereas MODFLOW is used to obtain groundwater number of HRUs to modify the model size without
discharge in each sub-basin in the SWAT model. omitting critical information [45]. The SWAT
[22]. The SWAT model models groundwater model's various HRU definition options allow users
recharge, or deep percolation, using Equation 1. to depict the study watershed's heterogeneity in
LULC, soil types, and topography. QSWAT is a
𝜔rchrg,i = (1 − exp [−1⁄𝛿 ]) ∙ 𝜔𝑠𝑒𝑒𝑝 + graphical user interface (QSWAT3, version 1.7.1)
𝑔𝑤
utilized to build and execute the study watershed's
exp [−1⁄𝛿 ] ∙ 𝜔𝑟𝑐ℎ𝑟𝑔,𝑖−1 (1) SWAT model. With a digital elevation model
𝑔𝑤
(DEM) of 30 m resolution, the area was contoured
On the day I, the total amount of water that leaves and separated into 37 sub-basins (Figure 3). To
the soil profile is represented by ɷrchrg,i (mm), and replicate the water network of the research area, a
the total amount of water that recharges the aquifers network of streams was created. After that, 74
is represented by ω_seep (mm). The amount of HRUs with various HRU choices were created. For
water that recharges the aquifers on day I-1 is this study, a station gauge called Dokan (Figure 1)
represented by rchrg,i-1 (mm), The groundwater was used to calibrate streamflow. The model was
materials' transit time is depicted by δ_gw (days). updated with imported weather data. From 1997 to
Using Equation (2), MODFLOW represents 2020, The first three years (1998–2000) of the
groundwater flow utilizing a finite difference simulation of a SWAT model were known as the
method. MODFLOW discretizes the aquifer into "warm-up period," during which the model was
unable to produce data that could be interpreted.
2498
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

conductivity. As expected, the first simulation

showed differences between calculated and
observed head values. To address this, the
parameter estimation (PEST) algorithm was
employed to calibrate the conceptual model. It
meant establishing additional pilot sites where there
was little data. The minimum and maximum
hydraulic conductivity are given identical values.
During the calibration procedure, the pilot points
were determined. Following the steady-state
simulation, the computed and observed heads
showed results that made sense, and the transient-
state simulation was set up. Before starting the
process of coupling SWAT and MODFLOW, the
transient simulation was set up with specific yield
Figure 3. The lower Zab River Basin hydrological soil
types
and specific storage.

MODFLOW model SWAT-MODFLOW model integration

The GMS software MODFLOW, version 10.4.64, Bailey et al. [22] presented the connecting code, a
was utilized to develop the conceptual model of the monthly time-step combination of the MODFLOW
study region. A finite difference method is used in and SWAT models. A crucial step in integrating
MODFLOW to analyze the groundwater flow in the Decomposed Hydrological Response Units
aquifer in three dimensions [46]. The MODFLOW- (DHRUs) is created by converting SWAT
NWT (Newtonian Solver) engine was employed in Hydrological Response Units (DHRUs). When
this study to account for variations in the grid cells' combined with a MODFLOW grid shapefile, gives
confinement and wetness over time. In an the model reliable spatial information. The meshing
unconfined aquifer, MODFLOWNWT can replicate strategy can be useful because it enables models to
Without rendering the cell hydraulic heads neutral, exchange their calculated outcomes. Wanting to
drying them out., which makes it superior to other import or alter the information entered. Park et al.’s
MODFLOW versions [24]. We also looked at an procedure indicates the full process of creating and
upstream weighting (UPW) package that gives dry transferring existing mesh files and model
cells zero magnitudes to compute the head from the integration. [44]. QSWATMOD, a graphical user
intake to the dry cells and to flow out. All things interface built on top of QGIS, was used in this
considered; the MODFLOW-NWT run provides a study. For the coupled model before, during, and
more accurate estimate of the watershed's after processing, Its Python code makes it ideal for
subsurface water flow [24]. 3259 active grid cells creating linkage files between MODFLOW and
have been identified within the Lower Zab River SWAT.
watershed, which, at a lateral resolution of 2959 m Bailey and associates. [22] describe the basic
by 2317 m, spans 19700.4 km2. After importing processes. QSWATMOD requires a few key inputs
the stream network established during the SWAT to integrate models.
delineation procedure as a shapefile into These include the SWAT output file, shapefiles for
MODFLOW, the driver package was used to create subbasins, hydrological response units (HRUs), and
293 stream cells for the watershed. The DEM raster the network of rivers.
file and SWAT model provide conceptual top Additionally, a MODFLOW model folder in native
elevation data. About 300 meters below the surface text format is necessary as part of the preprocessing
was the lowest elevation. phase.
Sources and sinks, beginning head, river The MODFLOW and SWAT models are two
conductance, and hydraulic conductivity are options for stream networks;
important factors in the modeling of groundwater the SWAT stream was utilized. MODFLOW grid
flow in a steady state. The SWAT percolation cells were paired with the appropriate hydrological
output was used to calculate the recharge rate, response units (DHRUs) to link the models.
producing a single MODFLOW number. Thirty Before the coupling process, the QSWATMOD
monitoring wells data sets from the research region integration was calibrated using a zonal polygon for
were interpolated to get the initial head values. the MODFLOW model's aquifer parameters.
Similarly, an interpolation method was used to A single working directory contained all the
determine the initial values for hydraulic integrated model files in an orderly fashion.

2499
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

2.4. Calibration and assessment of model observed data. The degree to which the observed
performance and simulated data match is assessed using the
regression coefficient (R2). NSE was chosen as the
Calibration of SWAT study's objective function to determine the
Observed streamflow data can be used to assess the simulation through the SWATCUP program's
SWAT model's suitability for streamflow calibration and validation phase. Two stream
simulation. It is necessary to adjust the parameters gauges (Dokan and Dibbis) are part of the study
that affect the simulated streamflow until they watershed. as indicated in Table 1. Due to missing
match the real streamflow values. All sources of data from the other stream gauge, which makes it
uncertainty regarding the model, parameters, and challenging to choose, calibration was done at
input data were taken into account by the Dokan. [49].
Sequential Uncertainty Fitting program (SUFI-2)
algorithm. Was used in conjunction with the Calibration of SWAT-MODFLOW models
SWAT-CUP program to accomplish Blöschl and Model parameters are changed during the
Sivapalan [47]. Jafari et al. [48] provided a QSWATMOD model calibration process to
thorough explanation of SUFI-2 and other guarantee that the simulated outcomes closely
algorithms. This study used a 20-year simulation of match the observed data. This procedure is essential
the SWAT model., with 2001–2005 serving as the because it improves the model's predictions for
calibration period and 2006–2008 as the validation surface and groundwater flows in terms of accuracy
period. The monthly time series in our and dependability. A thorough simulation of the
investigation provided both a validation and hydrological cycle is made possible by the
calibration for the streamflow that was observed. combination of MODFLOW (Modular Finite-
Sensitive parameters were selected from several Difference Groundwater Flow Model) and SWAT
SWAT hydrological parameters in order to (Soil and Water Assessment Tool). SWAT
calibrate the model by offering a value within an primarily deals with surface water flow, while
acceptable range. Evaporation, groundwater, MODFLOW focuses on groundwater flow. The
surface runoff, and simulated soil water are all study overcomes the drawbacks of applying each
controlled by certain factors. Figure 4 provides model independently by integrating them.
comprehensive explanations of a few chosen During calibration, specific parameters related to
factors. The model's performance during calibration land use, soil properties, and aquifer characteristics
can be assessed using statistical measures like the are adjusted. Statistical measures like the
regression coefficient (R2) Nash-Sutcliffe correlation coefficient, root mean square error, and
efficiency (NSE) and its values Nash-Sutcliffe efficiency are frequently used to
(https://mountainscholar.org/bitstream/ ). The NSE evaluate the objective, which is to minimize the
value should be higher than 0.4–0.6 for the difference between simulated and observed data.
calibration procedure. The closer the NSE is to The successful calibration of the coupled
one, the more the model outputs resemble the QSWATMOD model results in improved

Table 1. Fit and range value details for the SWAT model calibration parameters
Calibrate
Range
Factors Explanation d
value
Value
Wet condition II initial
r_CN2.mgt SCS runoff curve 0.154 -0.2 − 0.2
number
v_ALPHA.B Alpha factor of baseflow
0.39 0−1
F.gw (days)
v_GW_DEL Groundwater delay
2.47 1 − 45
AY.gw (days)
Water depth threshold
v_GWQMN.
for "revap" to occur in 0.061 -0.1 − 0.1
gw
the shallow aquifer (mm)
v_GW_REV Groundwater "revap" 0.01 −
0.05
AP.gw coefficient 0.09
Soil evaporation
v_ESCO.hru 0.975 0.0 − 1
compensation factor (-)
The soil layer's available
r_SOL_AW
water capacity 0.126 -0.2 − 0.2
C.sol
(mm mm-1 )

2500
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

performance compared to using the SWAT model enhance streamflow simulation. The aquifer's HK,
alone. Understanding water balance and managing SS, and Sy parameters were taken into account
water resources in the watershed depend on more during the QSWATMOD calibration procedures.
precise estimates of groundwater recharge, which Table 2 lists the calibration parameters that were
this improved model offers. Five additional SWAT selected and the ranges of values that correspond to
parameters were chosen for recalibration to them.

Table 2. Descriptions and value ranges of the calibration parameters chosen by SWAT-MODFLOW.
Value
Factor Explanation
Range
ALPHA_BF 0.01-0.2 SWAT, Baseflow alpha factor (days)
CH_k2 1-50 Effective hydraulic conductivity of the main channel (mm/h), SWAT
CN2 0.01-1 First SCS runoff curve number for SWAT and moisture condition II
EPCO 0.01-1 SWAT and the plant uptake compensation factor (-)
Compensation factor for soil evaporation (-), SWAT aquifer hydraulic conductivity (m/day),
ESCO 0.01-1
MODFLOW specific yield (-), MODFLOW specific storage (m-1), and MODFLOW
Hk 0.1-100 SWAT, Baseflow alpha factor (days)
0.000001- Effective hydraulic conductivity of the main channel (mm/h), SWAT
Ss
0.005
Sy 0.0001-0.4 First SCS runoff curve number for SWAT and moisture condition II

3. Results and discussion Table 3. The final SWAT-MODFLOW calibrated

parameter value
Factor Value Explanation
3.1. Calibrated parameters Range
SWAT, Baseflow alpha factor
ALPHA_BF 0.05
In the following section, the final automated (days)
parameters used in the integrated QSWATMOD Effective hydraulic conductivity
model are explained. The values of the calibration CH_k2 20 of the main channel (mm/h),
SWAT
parameters in Table 1 are substantially greater than First SCS runoff curve number
those in Table 3. This difference can be explained CN2 0.1 for SWAT and moisture
by the modification made to account for the condition II
simulated streamflow's lower base flow and peak SWAT and the plant uptake
EPCO 0.02
flow conditions. To fine-tune aquifer compensation factor (-)
Compensation factor for soil
characteristics before the coupling process, zonal evaporation (-), SWAT aquifer
polygons were created based on the hydraulic head hydraulic conductivity (m/day),
ESCO 0.02
distribution [49,50]. A visual depiction of the final MODFLOW specific yield (-),
automated aquifer parameter value ranges in the MODFLOW specific storage (m-
study area is shown in Figure 5. The range of the 1), and MODFLOW
calibrated hydraulic conductivity values is 0.54 to
15.2 meters per day.
Specific yields and storage values, respectively,
range from 0.00076 to 0.33 and 0.000001 m-1 to
0.0037 m-1 for the study. Figure 5 shows the
apparent relationship between the different aquifer
characteristics. Interestingly, after calibrating the
study watershed's hydraulic head, it shows an
inverse relationship between specific yield and
specific storage. A comparison of monthly
streamflow for the SWAT and integrated models
using simulated and measured data is shown in
Figure 5. The streamflow was calibrated between
2001 and 2005 and then validated between 2006
and 2008. Table 4 summarizes model performance
data for the SWAT and QSWATMOD models for
NSE and R2 between streamflow calculations and
measurements. Figure 4. Distribution of automated aquifer parameters
in the Lower Zab River Basin.

2500
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

3.2. Performance of the model during validation the wet season, which ran from January to March
and calibration (Figure 6.c). The average recharge during the rainy
season varies from 0.303 to 54.23 mm/month. In
Table 4 shows that the integrated model contrast, the dry season (June to September)
outperformed the SWAT model in terms of experiences the lowest recharge values of the year,
streamflow performance statistics. The accuracy with the maximum reaching only 11.161
with which the QSWATMOD model captured the mm/month (Figure 6.b). Recharge rates are lower in
observed streamflow under low base flow and peak higher-elevation areas. The recharge accounts for
flow conditions is shown in Figure 5. In the 15–40% of the average yearly precipitation in the
calibration and validation stages, both models met Lower Zab River watershed, which includes our
the evaluation criteria and accurately represented study area [52]. Furthermore, wintertime is when
streamflow [51]. the recharge is highest and most noticeable, with
February showing the highest recharge value.
According to this study, the average annual
precipitation in the area is caused by groundwater
recharge to the tune of 0.16%.

Figure 5. Observed and simulated flow by (a) SWAT

model; and (b) SWAT-MODFLOW, and precipitation at
the stream station in the years 2000–2008

Table 4. SWAT and SWAT-MODFLOW performance Figure 6 (a) Monthly recharge; and (b) dry season
metrics during calibration and verification. recharge; and (c) wet season recharge in the Lower Zab
Calibration Validation River Basin.
Name of the model
R2* NSE** R2* NSE**
4. Conclusions
SWAT 0.83 0.79 0.80 0.72
To assess the groundwater recharge of the Zab
SWAT-MODFLOW 0.89 0.77 0.90 0.74 River basin, a numerical model approach was
employed. QSWATMOD was used to determine
* Regression coefficient; ** Nash-Sutcliffe efficiency the spatiotemporal distribution of groundwater
recharge in the study region. For the linked model,
3.3. Groundwater recharge observations of groundwater level and streamflow
were calibrated. Streamflow measurements were
Figure 6 shows the groundwater recharge conducted between 2001 and 2005. and validated
distribution on average from 2001 to 2020. from 2006–2008. The model's performance has
Significant changes in groundwater recharge over been evaluated using the statistical parameter
time and space are shown in the figure for the entire functions R2 and NSE. The integrated model
watershed. These patterns are evident in the performed more effectively throughout the
monthly average recharge distribution, which falls calibration and validation phases when modeling
between 0.215 and 25.052 mm/month (Figure 6.a). the streamflow. In addition to aquifer parameter
The complexity of groundwater recharge calibration, recalibrating specific SWAT
distribution in this study makes it challenging to parameters was a demonstrated factor in improved
correlate with the region's physical topography, streamflow simulation for integrated models. The
although the southwest exhibits the highest QSWATMOD model displayed spatiotemporal
recharge levels. Groundwater recharge in the study distribution changes over the year and determined
area showed a pattern of heavy precipitation during the region's average monthly groundwater recharge.
2501
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

There was a noticeable difference between the dry Extensive Stormwater Infiltration. J. Hydrol. 544,
and wet seasons, and the model correctly 524–537.
represented the seasonal recharge in the area. Even [4] Barron, O.V., Donn, M.J., Barr, A.D., (2013).
though there are many urban areas in the area, 16% Urbanisation and Shallow Groundwater: Predicting
Changes in Catchment Hydrological Responses.
of the yearly average precipitation comes from
Water Resour. Manag., 27, 95–115.
groundwater recharge. Significant contributions [5] Fleckenstein, J.H., Krause, S., Hannah, D.M.,
were made by surface and lateral direct flow Boano, F., (2010). Groundwater-SurfaceWater
components, which are connected to the study Interactions: New Methods and Models to Improve
area's geological characteristics. All things Understanding of Processes and Dynamics. Adv.
considered, the present study proved the usefulness Water Resour., 33, 1291–1295.
and application of the QSWATMOD model for [6] Sophocleous, M., (2002). Interactions between
groundwater recharge analysis. Swat-Modflow Groundwater and Surface Water: The State of
displays the study watershed's groundwater Science. Hydrogeol. J., 10, 52–67.
recharge's temporal and spatial distribution. As [7] Sharp, J.M., (2010). The Impacts of Urbanization
on Groundwater Systems and Recharge.
such, it provides a more accurate depiction of the
Aquamundi, 1, 51–56.
region's groundwater and surface water. [8] Wang, Y., Zheng, C., Ma, R., (2018). Review: Safe
and Sustainable Groundwater Supply in China.
Author Statements: Hydrogeol. J., 26, 1301–1324.
[9] Yifru, B.A., Chung, I.M., Kim, M.G., Chang, S.W.,
 Ethical approval: The conducted research is (2021). Assessing the Effect of Land/Use Land
not related to either human or animal use. Cover and Climate Change on Water Yield and
Groundwater Recharge in East African Rift Valley
 Conflict of interest: The authors declare that
Using Integrated Model. J. Hydrol. Reg. Stud., 37,
they have no known competing financial 100926.
interests or personal relationships that could [10] Lee, C.H., Yeh, H.F., Chen, J.F., (2008).
have appeared to influence the work reported in Estimation of Groundwater Recharge Using the
this paper Soil Moisture Budget Method and the Base-Flow
 Acknowledgement: The authors declare that Model. Environ. Geol., 54, 1787–1797.
they have nobody or no-company to [11] Scanlon, B.R., Healy, R.W., Cook, P.G., (2002).
acknowledge. Choosing Appropriate Techniques for Quantifying
Groundwater Recharge. Hydrogeol. J., 10, 18−39.
 Author contributions: The authors declare that
[12] Healy, R., Cook, P.G., (2002). Using Groundwater
they have equal right on this paper. Levels to Estimate Recharge. Springer, 10, 91–109.
 Funding information: The authors declare that [13] Chand, R., Chandra, S., Rao, V.A., Singh, V.S.,
there is no funding to be acknowledged. Jain, S.C., (2004). Estimation of Natural Recharge
 Data availability statement: The data that and Its Dependency on Sub-Surface Geoelectric
support the findings of this study are available Parameters. J. Hydrol., 299, 67–83.
on request from the corresponding author. The [14] Rangarajan, R., Muralidharan, D., Deshmukh, S.D.,
data are not publicly available due to privacy or Hodlur, G.K., Gangadhara Rao, T., (2005).
ethical restrictions. Redemarcation of Recharge Area of Stressed
Confined Aquifers of Neyveli Groundwater Basin,
India, through Tritium Tracer Studies. Environ.
References Geol., 48, 37–48.
[15] Shahul Hameed, A., Resmi, T.R., Suraj, S.,
[1] Arshad, A., Mirchi, A., Samimi, M., Ahmad, B., Warrier, C.U., Sudheesh, M., Deshpande, R.D.,
(2022). Combining Downscaled-GRACE Data with (2015). Isotopic Characterization and Mass Balance
SWAT to Improve the Estimation of Groundwater Reveals Groundwater Recharge Pattern in Chaliyar
Storage and Depletion Variations in the Irrigated River Basin, Kerala, India. J. Hydrol. Reg. Stud., 4,
Indus Basin (IIB). Sci. Total Environ., 838, 156044. 48–58.
[2] Aslam, R.A., Shrestha, S., Usman, M.N., Khan, [16] Aslam, R.A., Shrestha, S., Usman, M.N., Khan,
S.N., Ali, S., Sharif, M.S., Sarwar, M.W., S.N., Ali, S., Sharif, M.S., Sarwar, M.W.,
Saddique, N., Sarwar, A., Ali, M.U., et al., (2022). Saddique, N., Sarwar, A., Ali, M.U., et al., (2022).
Integrated SWAT-MODFLOW Modeling-Based Integrated SWAT-MODFLOW Modeling-Based
Groundwater Adaptation Policy Guidelines for Groundwater Adaptation Policy Guidelines for
Lahore, Pakistan under Projected Climate Change, Lahore, Pakistan under Projected Climate Change,
and Human Development Scenarios. Atmosphere, and Human Development Scenarios. Atmosphere,
13, 13, 2001.
[3] Locatelli, L., Mark, O., Mikkelsen, P.S., Arnbjerg- [17] Cheema, M.J.M., Immerzeel, W.W., Bastiaanssen,
Nielsen, K., Deletic, A., Roldin, M., Binning, P.J. W.G.M., (2014). Spatial Quantification of
(2017), Hydrologic Impact of Urbanization with Groundwater Abstraction in the Irrigated Indus
Basin. Groundwater, 52, 25–36.
2502
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

[18] Taie Semiromi, M., Koch, M., (2019). Analysis of Jeppesen, E., Jensen, J.S., Jensen, J.B., et al.,
Spatio-Temporal Variability of Surface– (2020). Quantifying the Streamflow Response to
Groundwater Interactions in the Gharehsoo River Groundwater Abstractions for Irrigation or
Basin, Iran, Using a Coupled SWAT-MODFLOW Drinking Water at Catchment Scale Using SWAT
Model. Environ. Earth Sci., 78, 201. and SWAT–MODFLOW. Environ. Sci. Eur., 32,
[19] Dowlatabadi, S., Ali Zomorodian, S.M., (2015). 113.
Conjunctive Simulation of Surface Water and [32] Zambrano-Bigiarini, M., Rojas, R., (2013). A
Groundwater Using SWAT and MODFLOW in Model-Independent Particle Swarm Optimisation
FiroozabadWatershed. KSCE J. Civ. Eng., 20, 485– Software for Model Calibration. Environ. Model.
496. Softw., 43, 5−25.
[20] Sophocleous, M.A., Koelliker, J.K., Govindaraju, [33] Yuan, L., Sinshaw, T., Forshay, K.J., (2020).
R.S., Birdie, T., Ramireddygari, S.R., Perkins, S.P., Review of Watershed-Scale Water Quality and
(1999) Integrated Numerical Modeling for Basin- Nonpoint Source Pollution Models. Geosciences,
Wide Water Management: The Case of the 10, 25.
Rattlesnake Creek Basin in South-Central Kansas. [34] Guevara-Ochoa, C., Medina-Sierra, A., Vives, L.,
J. Hydrol., 214, 179–196. (2020). Spatio-Temporal Effect of Climate Change
[21] Gao, F., Feng, G., Han, M., Dash, P., Jenkins, J., on Water Balance and Interactions between
Liu, C., (2019) Assessment of Surface Water Groundwater and Surface Water in Plains. Sci.
Resources in the Big Sunflower River Watershed Total Environ., 722, 137886.
Using Coupled SWAT–MODFLOW Model. [35] Healy, R., Cook, P.G., (2002). Using Groundwater
Water, 11, 528. Levels to Estimate Recharge. Springer, 10, 91–109.
[22] Bailey, R.T., Wible, T.C., Arabi, M., Records, [36] Hung Vu, V., Merkel, B.J., (2019). Estimating
R.M., Ditty, J., (2016). Assessing Regional-Scale Groundwater Recharge for Hanoi, Vietnam. Sci.
Spatio-Temporal Patterns of Groundwater–Surface Total Environ., 651, 1047–1057.
Water Interactions Using a Coupled SWAT- [37] Sim¸sek, C., Demirkesen, A.C., Baba, A.,
MODFLOW Model. Hydrol. Process., 30, Kumanlıo˘ glu, A., Durukan, S., Aksoy, N.,
4420−4433. Demirkıran, Z., Hasözbek, A., Murathan, A.,
[23] Techniques, M., (2005). MODFLOW-NWT, a Tayfur, G., (2020). Estimation Groundwater Total
Newton Formulation for MODFLOW-2005. US Recharge and Discharge Using GIS-Integrated
Geol. Surv. Tech. Methods, 6, 44. Water Level Fluctuation Method: A Case Study
[24] Wang, Y., Chen, N., (2021). Recent Progress in from the Ala¸sehir Alluvial AquiferWestern
Coupled Surface–Ground Water Models and Their Anatolia, Turkey. Arab. J. Geosci., 13, 143.
Potential in Watershed Hydro-Biogeochemical [38] Delottier, H., Pryet, A., Lemieux, J.M., Dupuy, A.,
Studies: A Review. Watershed Ecol. Environ., 3, (2018). Estimating Groundwater Recharge
17−29. Uncertainty from Joint Application of an Aquifer
[25] Yifru, B.A., Chung, I.M., Kim, M.G., Chang, S.W., Test and the Water-Table Fluctuation Method.
(2020). Assessment of Groundwater Recharge in Hydrogeol. J., 26, 2495−2505.
Agro-Urban Watersheds Using Integrated SWAT- [39] Park, S., (2018). Enhancement of Coupled
MODFLOW Model. Sustainability, 12, 6593. Surface/Subsurface Flow Models in Watersheds:
[26] Ntona, M.M., Busico, G., Mastrocicco, M., Analysis, Model Development, Optimization, and
Kazakis, N., (2022). Modeling Groundwater and User Accessibility by Using SWAT-MODFLOW
SurfaceWater Interaction: An Overview of Current Simulation. Available
Status and Future Challenges. Sci. Total Environ., online:https://mountainscholar.org/bitstream/
846, 157355. handle/10217/193164/Park_colostate_0053A_1519
[27] Mosase, E., Ahiablame, L., Park, S., Bailey, R., 9.pdf?sequence=1 (accessed on 11 September
(2019). Modeling Potential Groundwater Recharge 2024).
in the Limpopo River Basin with SWAT- [40] Chung, I.M., Kim, Y.J., Kim, N.W., (2021).
MODFLOW. Ground. Sustain. Dev., 9, 100260. Estimating the Temporal Distribution of
[28] Sophocleous, M., Perkins, S.P., (2000). Groundwater Recharge by Using the Transient
Methodology and Application of Combined Water Table Fluctuation Method and Watershed
Watershed and Ground-Water Models in Kansas. J. Hydrologic Model. Appl. Eng. Agric., 37, 95−104.
Hydrol., 236, 185–201. [41] Jie, Z., van Heyden, J., Bendel, D., Barthel, R.,
[29] Chunn, D., Faramarzi, M., Smerdon, B., Alessi, D., (2011). Combination of Soil-Water Balance Models
(2019). Application of an Integrated SWAT– and Water-Table Fluctuation Methods for
MODFLOW Model to Evaluate Potential Impacts Evaluation and Improvement of Groundwater
of Climate Change and Water Withdrawals on Recharge Calculations. Hydrogeol. J., 19,
Groundwater–Surface Water Interactions in West- 1487−1502.
Central Alberta. Water, 11, 110. [42] Mohammed, R., and Scholz, M., (2019). Climate
[30] Pisinaras, V., (2016). Assessment of Future change and water resources in arid regions:
Climate Change Impacts in a Mediterranean uncertainty of the baseline period. Theor. Appl.
Aquifer. Glob. NEST J., 18, 119–130. Climatol., 137(1), 1365−1376.
[31] Liu, W., Park, S., Bailey, R.T., Molina-Navarro, E., https://doi.org/10.1007/s00704-018-2671-6
Andersen, H.E., Thodsen, H., Nielsen, A.,
2503
 Mohammed Jalel Rahem, Ruqayah Mohammed / IJCESEN 11-2(2025)2494-2504

[43] Mohammed, R., Scholz, M., (2017). Impact of

Evapotranspiration Formulations at Various
Elevations on the Reconnaissance Drought Index.
Water Resour. Manage., 31, 531-548.
[44] Park, E., (2012). Delineation of Recharge Rate
from a HybridWater Table Fluctuation Method.
Water Resour. Res., 48.
[45] Faramarzi, M., Srinivasan, R., Iravani, M., Bladon,
K.D., Abbaspour, K.C., Zehnder, A.J.B., Goss,
G.G., (2015). Setting up a Hydrological Model of
Alberta: Data Discrimination Analyses Before
Calibration. Environ. Model. Softw., 74, 48–65.
[46] Abbaspour, K.C., (2010). SWAT-CUP SWAT
Calibration and Uncertainty Programs. Arch.
Orthop. Trauma Surg., 130, 965–970.
[47] Blöschl, G., Sivapalan, M., (1995). Scale Issues in
Hydrological Modelling: A Review. Hydrol.
Process., 9, 251–290.
[48] Jafari, T., Kiem, A.S., Javadi, S., Nakamura, T.,
Nishida, K., (2021). Fully Integrated Numerical
Simulation of Surface Water-Groundwater
Interactions Using SWAT-MODFLOW with an
Improved Calibration Tool. J. Hydrol. Reg. Stud.,
35, 100822.
[49] Abbaspour, K.C., Rouholahnejad, E. Vaghefi, S.,
Srinivasan, R., Yang, H., Kløve, B.A., (2015)
Continental-Scale Hydrology and Water Quality
Model for Europe: Calibration and Uncertainty of a
High-Resolution Large-Scale SWAT Model. J.
Hydrol., 524, 733–752.
[50] Ali, I. M., Naje, Ahmed Samir, Nasr, M. S., (2020).
Eco-friendly chopped tire rubber as reinforcements
in fly ash based geopolymer concrete. Global Nest
Journal, 22(3), 342 – 347.
[51] Mohammed, Ibrahim, Hashim Al-Khalaf, Safaa K.,
Alwan, Husam H., Naje, Ahmed Samir, (2022).
Environmental assessment of Karbala water
treatment plant using water quality index (WQI).
Materials Today: Proceedings, 60,1554 – 1560.
[52] Hassan, Ali A., Hadi, Raid T., Rashid, Adil H.,
Naje, Ahmed Samir, 2020. Chemical modification
of castor oil as adsorbent material for oil content
removal from oilfield produced water. Pollution
Research, 39 (4),892 – 900.

2504

View publication stats