Journal of Environmental Management 227 (2018) 107–116

Contents lists available at ScienceDirect

Journal of Environmental Management

journal homepage: www.elsevier.com/locate/jenvman

Research article

Assessment of urbanization impact on groundwater resources in Hanoi, T

Vietnam
Vu Thanh Tama, Tran Thi Viet Ngab,∗
a
National Center for Water Resources Planning & Investigation of Vietnam, 93/95 Vu Xuan Thieu, Sai Dong, Hanoi, Viet Nam
b
National University of Civil Engineering, 55 Giai Phong Road, Hai Ba Trung District, Hanoi, Viet Nam

A R T I C LE I N FO A B S T R A C T

Keywords: In this study, the impact of urban development on groundwater resources of Hanoi City was assessed in terms of:
Urbanization (1) change of land use practice, and (2) increasing groundwater abstraction due to urban population growth. To
Land use change this end, a coupled hydrological rainfall-runoff and groundwater flow simulation with WetSpa and MODFLOW
Groundwater recharge codes was carried out with a set of spatial and temporal data of meteo-hydrology, land use, groundwater ab-
Numerical modeling
straction, population growth, and losses from the city water supply and sewerage system. The results of the
Hanoi City
simulation indicate that infiltration from rainfall contributes with 53.6% to the recharge of the groundwater
system in Hanoi City, followed by seepage from rivers and lakes (31%). The remaining 15.4% was attributed to
leakage from the municipal water supply and sewerage networks. This study also suggests that the main cause of
groundwater level's decline in the city is the extensive groundwater abstraction, while an increase of the urban
impervious areas due to urbanization only causes a slight decrease of groundwater recharge.

1. Introduction deep percolation depends on the characteristics of the unsaturated

In many cities in the world, urbanization changes the land use
practice, usually by reducing the natural land area and increasing the
impervious cover area (Murakami et al., 2005; Xu et al., 2007; Batisani
and Yarnal, 2009; Mohan et al., 2011; Hassan and Nazem, 2016). The
urban hard surfaces such as rooftops, driveways, streets, swimming
pools, and patios don't allow water to penetrate the soil. In addition,
streets act as streams, collecting storm water and channeling it into
waterways. This leads to the decrease in the volume of rainfall that
infiltrates through the ground and the resulting increase in the volume
of surface water (Foster et al., 1993; Garcia-Fresca and Sharp, 2005;
Mao and Cherkauer, 2009; Sajikumar and Remya, 2015). However, the
reduction of direct infiltration can be counterbalanced by the reduction
in evapotranspiration, and new sources of recharge may arise in the
urban environment, such as losses in the sewerage system or in the
water distribution system (Kim et al., 2001; Wolf et al., 2004;
Blackwood et al., 2005).
On the other hand, the increase of groundwater abstraction to meet
the water supply demand of a growing urban population growth is one
of the causes of groundwater depletion in megacities (Lerner and
Barrett, 1996; Foster et al., 1998; Hoque et al., 2007; Mpamba and
Hussen, 2008; Naik and Tambe, 2008). This decline of groundwater
level can lead to an increase of the groundwater recharge rate as the
zone, mainly the hydraulic conductivity, root depth, water content of
the soil, and the gradient of hydraulic potential between the un-
saturated and saturated zone (Bouwer, 1978; Hillel, 1982).
Thus, the impact of urbanization on groundwater resources can be
assessed by evaluating the two most important variables: recharge rate
and groundwater level. However, evaluating the impact of recharge on
groundwater resources in an urban environment is different from
evaluating the impact on natural systems for three reasons: (1) recharge
sources are radically different; (2) lack of detailed information on the
sewerage system and water distribution system as well as the position,
duration and volume of water leaked from these systems; and (3) re-
duction of recharge and decline of groundwater level due to urbani-
zation are inter-related as mentioned above, which cannot be in-
dependently quantified by a single simulation. Several methods have
been developed to distinguish sources and to quantify the recharge rate
of an urban groundwater system based on hydro-geochemistry and
isotope tracer analysis of groundwater water samples (Kim et al., 2001;
Choi et al., 2005; Morris et al., 2006; Vázquez-Suñé et al., 2010). In
essence, these methods only provide a “snapshot” of recharge around
the time of the sampling; thus they are not suitable for assessing the
impact of the urbanization process on the variation of groundwater
recharge and resources. In addition, these methods do not show the
impact of increasing impervious cover area and declining groundwater

∗
Corresponding author.
E-mail address: ngattv@nuce.edu.vn (T.T.V. Nga).

https://doi.org/10.1016/j.jenvman.2018.08.087
Received 31 December 2017; Received in revised form 16 August 2018; Accepted 23 August 2018
Available online 30 August 2018
0301-4797/ © 2018 Elsevier Ltd. All rights reserved.
V.T. Tam, T.T.V. Nga
level to groundwater recharge during the urban development. A com-
prehensive evaluation of all recharge sources can be achieved through
flow and transport simulation when previous data on groundwater head
and quality are available (Yang et al., 1999; Vázquez-Suñé and
Sánchez-Vila, 1999; Trowsdale and Lerner, 2007).
A recent study demonstrated that the main contributors to total
recharge of Barcelona city aquifers are: rainfall, surface runoff and
water bodies (48%), followed by losses from the sewerage network
(30%), and losses from the water supply network (22%) (Vázquez-Suñé
et al., 2010). For tropical cities where annual precipitation is 2.5–4.5
times higher than in Barcelona, contribution from rainfall and the re-
sulting overland flow and surface runoff to total recharge can be fore-
seen as a major portion. As land use change is one of the factors in-
fluencing the distribution and contribution of overland flow and surface
runoff to ground water recharge, land use maps at different time per-
iods can be used as a variable to assess the impact of urbanization
process to the groundwater resource.
Hanoi is considered the second most rapidly developing city in
Vietnam in the last two decades. Also in this period, the Quaternary
aquifers, the most important source for the city water supply are re-
ported as being degraded: the annual lowering of groundwater level
0.3–0.5 m causes land subsidence of up to 41 mm/year, additionally,
the aquifer are exposed to saltwater intrusion from southeast (On,
2012). The report raised the two main causes for the degradation: re-
duction of groundwater recharges and increase of groundwater ab-
straction. This research analyses the impact of urbanization develop-
ment to groundwater resources of Hanoi City, based on land use change
and groundwater abstraction. Using these variables as input, a coupled
hydrological-groundwater flow modeling is used to simulate the tem-
poral changes of groundwater recharge and level. Based on this simu-
lated result, a detailed analysis of the relationship between rainfall,
river discharge, components of recharge, and groundwater level is
carried out to reveal the factors that have the most significant influence
on the groundwater resources in the study area. This study provides
insights into the impact of urbanization on groundwater budget in de-
veloping cities where groundwater is the main source for urban water
supply.
2. Site description
2.1. Hanoi's hydrogeology
Hanoi, the second largest city in Vietnam with a current area of
about 3300 km2 and 7.6 million of inhabitants, is located at the heart of
the Red River Delta plain (Fig. 1). Three quarters of the city's area is of
flat terrain gradually steeping down in the North-South and West-East
directions with an averaged elevation 5–20 m above mean sea level
(AMSL). The remaining area is covered by hills and mountains with an
elevation of 50–750 m AMSL located in the West and North-West
(Fig. 1). Flowing through Hanoi city, the Red River is the second largest
in Vietnam with a mean flowrate of 1054 and 4228 m³/s for dry and
rainy season respectively. The mean annual precipitation in Hanoi City
is about 1760 mm, out of which 78% falls in the hot rainy season from
May till October and the remaining 22% falls in the cool dry season
from November till April every year.
Hydrogeologically, the groundwater of Hanoi City resides in porous
media of Quaternary formations distributed in the flat terrain and in
fault/fracture zones of pre-quaternary hard rock formations exposed on
the hills and mountains. The Quaternary formations are strati-
graphically divided into two aquifers intercalated with two aquitards
from top to down as follows: (1) the upper aquitard (UA), with aver-
aged thickness of 18.5 m and hydraulic conductivity 0.05 m/day, ex-
poses on the ground surface in the southern area of Hanoi; (2) the
Holocene aquifer (QH), with average thickness of 18.9 m and hydraulic
conductivity 12.5 m/day, underlies directly the UA and outcrops
mainly in the North and as narrow bands along rivers; (3) the lower
108
Journal of Environmental Management 227 (2018) 107–116
aquitard (LA), with average thickness of 16.5 m and hydraulic con-
ductivity 0.02 m/day, is completely covered by the UA and QH for-
mations; and (4) the Pleistocene aquifer (QP), of average thickness
59.5 m and hydraulic conductivity 31.9 m/day, overlays directly on the
partly weathered or denuded surface of the hard rock formations. More
information on the lithology, thickness and hydraulic conductivity of
these aquitards and aquifers is provided in Table 1.
A three-dimensional representation of the hydrogeological strati-
graphy for Hanoi City built on the basis of 238 stratigraphic well-logs is
shown in Fig. 2. This diagram shows the existence of many “vertical
hydrogeological windows” where the LA aquitard is absent and the QH
and QP aquifers are interconnected. The QH aquifer is recharged from
rainfall and surface water percolating vertically through the UA aqui-
tard or with river water in the southern part of Hanoi, where the Red
River system cuts into the aquifer. The QP aquifer is only replenished
with rain and surface water in the Northern and Western part of Hanoi
City where it outcrops. Based on the pumping tests conducted by Quyen
(2016), the QP aquifer has been estimated to receive water discharging
from weathered and fractured zones at the rate of approximately 2.76 l/
s/km along the Western contact margin. At the bottom, this aquifer is
lined with layers of siltstone, shale and schist of hard rock formations,
and therefore, it was assumed there is no flow through this layer.
2.2. The development of Hanoi City
The development of Hanoi City started in the 11th century as the
capital of Vietnam, and before the year 1975 remained in the style of a
town with no high-rise buildings. Only when the country's economy
burst into bloom in the last decade of the 20th century has Hanoi was
growing rapidly to become a modern crowded city with an urban po-
pulation increasing from 1.44 million in 1990 to 3.41 million in 2015
(Table 2). A sudden development in city's infrastructure accompanied
this growth, meaning urbanized land with new roads and buildings
replaced agricultural and natural (A&N) land; more sewerage and water
supply networks were constructed; and many small lakes and ponds
were partly or fully filled up with soil to be transformed into resident
areas or industrial parks (Table 2).
2.3. Water supply and wastewater generation in Hanoi City
Since the beginning of the nineteenth century, groundwater has
been the principal water supply source for Hanoi City. Before 1990s, a
volume of about 350,000 m3/day of groundwater was exploited from
the QP aquifer and about 200,000 m3/day from the QH aquifer. Water
supply for the urban area consist of mainly deep large diameter wells
drilled in the QP aquifer over 16 well fields and an underground dis-
tribution system hundreds kilometers in length. By 2010, the ground-
water abstraction volume increased to about 950,000 m3/day from the
QP aquifer and 350,000 m3/day from the QH aquifer. This increased in
the abstraction caused a declined rate of the groundwater levels of
0.5–0.8 m/year for the QP aquifer during the period 1990–2010. The
groundwater level decline stopped after 2010 when a volume of
250,000–350,000 m3/day from a surface water treatment plant located
80 km upstream of the Red River was added to meet the increasing
demand of newly built residential areas and industrial parks.
Concomitant with urban population growth, the sewerage and
water networks in the urban area have been also expanded and up-
graded. However, leakage and losses from these systems are still con-
siderable. By comparing the water volume produced by water treatment
plants and the volume gauged at the water meter of consumers, it was
estimated that the losses through the distribution system were about
36% (in 1990s) to 27% (in 2010s) (VWSA, 2015). The leakage from the
urban sewerage system was also estimated at 25% (in 1990s) and 21%
(in 2010s) by subtracting the sewage volume gauged at the urban outlet
from the total water supply volume gauged at consumer's water meters.
It was noticed that the calculated leakage value may underestimate the
V.T. Tam, T.T.V. Nga Journal of Environmental Management 227 (2018) 107–116

Fig. 1. Location of study area.

Table 1 WetSpa and MODFLOW codes is used. This coupled model takes to-
Summary information of the aquifers and aquitards based on 127 pumping tests pography, soil properties, and temporally-variable land use and rainfall
in Hanoi City (Tran, 1993). as inputs to calculate seasonal variation of recharge and groundwater
Water-bearing Mean Mean hydraulic Mean specific level. The temporal urban land use changes are taken into consideration
formation thickness, (m) conductivity (m/d) yield (%) by the WetSpa model while the temporal groundwater abstraction
variation is simulated by the MODFLOW model. Because of hydro-
Upper aquitard (UA) 18.5 0.05
logical distributed property of WetSpa modeling, respective recharge
Holocene aquifer 18.9 12.5 10.7
(QH) from rainfall in urban land as well as in A & N land can be easily cal-
Lower aquitard (LA) 16.5 0.02 culated. Recharge from the Red River is calculated with the RIVER
Pleistocene aquifer 59.5 31.9 17.6 package embedded in MODFLOW code while recharge from other open
(QP) surface water bodies such as small rivers, channels and lakes is esti-
Weathered - 60 0.04
fractured hard
mated as the product of their distribution area and leakage flux value
rocks determined from seepage experiments. Recharge due to leakage from
the supply network and sewage network is also estimated in the
MODFLOW modeling. All these components are summarized as total
actual value as the gauged sewage volume also includes the overland recharge of the groundwater system under study. The calculated mean
flow during storm events. annual and seasonal precipitation, total and component recharge, and
groundwater level values are analyzed in relation to changes in land use
and increasing groundwater abstraction to reveal the impact of urba-
3. Methodology
nization development on the groundwater system concerned.

This research assesses the impact of urbanization on groundwater

resources by analyzing the temporal variation of recharge and 3.1. Simulation of time variation of groundwater recharge by WetSpa
groundwater levels under changes of land use and groundwater ab- hydrologic model
straction volume for the municipal water supply. For this purpose, a
coupled rainfall-runoff and groundwater flow modeling based on WetSpa, a grid-based rainfall-runoff hydrologic model for water and

109
V.T. Tam, T.T.V. Nga Journal of Environmental Management 227 (2018) 107–116

Fig. 2. Three dimensional diagram of the hydrogeological stratigraphy of Hanoi City.

energy transfer between the soil, plants and atmosphere was used in storage, which then contributes to groundwater discharge forming the
this study to simulate time variation of groundwater recharge in ac- base flow of a stream hydrograph (Liu et al., 2004). Recharge was es-
cordance with land use changes resulted from Hanoi City urbanization. timated using Equation (2) based on the relationship between hydraulic
Each grid cell includes four top-down layers representing vegetation conductivity and effective saturation (Brooks and Corey, 1966):
zone, root zone, transmission zone, and saturated zone. The hydrologic
(2 + 3B )/ B
processes considered in the model are precipitation, interception, de- θ − θr ⎞
R = Ks ⎛
⎜ ⎟

pression, surface runoff, infiltration, evapotranspiration, percolation, ⎝ s − θr ⎠

interflow, groundwater flow and water balance in the root zone and the
saturated zone (Liu et al., 2003). The water balance of the root zone
was considered as a key factor controlling the runoff, interflow and
groundwater recharge and was calculated for each grid cell (Equation
(1)):
Δθ
D =P−I−S−E−F−R
Δt
where D is the root zone depth; Δθ is the change in soil moisture
content; Δt is the time interval; P is the precipitation; I is the initial
abstraction (interception and depression losses); S is the surface runoff;
E is the actual evapotranspiration; F is the interflow; and R is the
percolation out of the root zone.
The percolation out of the root zone increases the groundwater
θ (2)
where R is the recharge or percolation; Ks is the saturated soil hydraulic
conductivity; θ is the soil moisture content; θr is the residual moisture
content; θs is the soil porosity; B is the pore size distribution index.
Baseflow simulation was performed at the sub-catchment scale
using a lumped linear reservoir approach (Equation (3)):
SG
(1) Qg = Cg
1000 (3)
where Qg is the base flow; Cg is the groundwater flow recession con-
stant, and SG is the groundwater storage.
The model's input data include: (1) temporal variables such as daily
rainfall, temperature and evapotranspiration time series together with
five maps in raster format representing land use status of Hanoi City

Table 2
Change of population, land use practice, sewage discharge and water consumption through the urbanization process of Hanoi City (summarized from JICA, 2007 and
GSO, 2016).
Parameters 1991–1995 1996–2000 2001–2005 2006–2010 2011–2015

Urban population (106 people) 1.44 1.99 2.61 3.04 3.41

Non-urban population (106 people) 2.98 3.20 3.55 3.89 4.15
Urbanized land area (km2) 181.32 253.2 325.76 417.64 476.25
A&N land area (km2) 2930.13 2878.86 2827.91 2760.26 2734.34
Open surface water area (km2)a 177.44 156.83 135.23 110.99 98.1
Average sewage dischargeb (103 m3/day) 240 300 375 470 580
Water consumption for urban area (m3/day) 320c 390c 480c 600d 730d

a
Not including the Red River.
b
Measured at the urban drainage outlet.
c
Not counting individual production wells.
d
Also includes exploitation of surface water from the Red River.

110
V.T. Tam, T.T.V. Nga
during the periods 1991–1995, 1996–2000, 2000–2005, 2006–2010
and 2001–2015; (2) soil classification and digital elevation model
(DEM) raster. The above-mentioned daily meteorological data are col-
lected from the National Lang Monitoring Station, the daily Red River
discharge data are from the Long Bien Gauge (Fig. 1), and the land use,
soil classification and DEM maps are provided by the Vietnam National
Institute of Agricultural Planning and Projection (NIAPP). Initial
groundwater storage, which was estimated as the product of effective
porosity of the QH aquifer and groundwater head monitored in January
1991, was used as the simulation start value for the simulation. The
model's parameters such as Manning's roughness coefficient for chan-
nels, potential runoff coefficient, depression storage capacity, and
groundwater flow recession coefficient was calibrated for the
1991–1995 period. This was done to ensure that the simulated
groundwater recharge value was lower than the saturated hydraulic
conductivity values determined by double ring infiltrometer experi-
ments in the field. In the subsequent simulation periods, these cali-
brated parameters remained unchanged, but the initial groundwater
storage for each period was calculated with the groundwater head of
the previous simulation period, and the land use map was replaced with
that of the corresponding period. The WetSpa model was operated on a
daily computation basis and produced a series of raster maps of
monthly groundwater recharge from January 1991 till December 2015
for use in the groundwater flow model. For lake cells, the data collected
from previous seepage experiments representing the mean leakage flux
value through bed sediment into groundwater system was used instead
of the recharge value computed by the WetSpa model.
3.2. Simulation of groundwater flow using MODFLOW
The groundwater flow modeling aimed to assess the temporal
change of the groundwater level resulting from the variation of
groundwater recharge and increasing groundwater abstraction. The
model was based on the MODFLOW code comprising four layers, re-
presenting the whole Quaternary formations as described in Table 1. On
the surface, the model received the simulated monthly rasters of
groundwater recharge as time-variable input. To use of these rasters,
the entire modeled area was divided into 11 recharge zones, each in-
cluding cells of relatively similar characteristics of topography, land use
Fig. 3. (a) Groundwater recharge raster computed by WetSpa model for January 1991, (b) Division of recharge zones with assigned recharge value for use in the
groundwater flow model.
111
Journal of Environmental Management 227 (2018) 107–116
and soil, and each region was assigned an averaged recharge value of all
cells located within that zone (Fig. 3). In the urban zone, the estimated
leakage from the supply network and sewage network were the inputs
to the model as an additional source of groundwater recharge. The
value of these recharge sources were spatially uniform over the urban
zone, but temporally changed on the basis of the above-mentioned loss
rate and monthly gauged volumes of sewage discharge and total water
supply. The monthly aggregated groundwater abstraction volume of the
16 centralized water-supply well fields (each consisting of 8–12 pro-
duction wells) were the inputs to the model for the simulation of the
impact of abstraction on groundwater level variation. The monthly
variation of the Red River water level was also considered as an input to
the model for simulation of hydraulic interaction between the river and
the groundwater aquifers. All these temporally variable values were
applied at the beginning of each simulated stress period of one month
and thereafter they remained constant during that period.
In the West where the Quaternary formations is in contact with the
hard rock formation, the model is assigned a specified constant flux
boundary with the flux value 2.76 l/s/km determined by Quyen (2016)
along this contact. The remaining model's hydraulic boundaries are of
specified head type with a value assigned to each section of the
boundary estimated on the basis of the groundwater level observed in
the boreholes of the groundwater monitoring network (Fig. 1).
Hydrogeological properties of the aquifer such as hydraulic con-
ductivity and storage coefficients were selected from pumping tests as
summarized in Table 1 and entered into model as constant parameters.
The model's parameter calibration only focuses on the hydraulic con-
ductivity and storage coefficients of the aquitards as there were no
pumping test conducted for these formations. For this calibration was
the groundwater level observed in 36 boreholes of the monitoring
network was used (Fig. 4a). The model validated for the period January
1995–December 2000, showing a good match between simulated and
observed groundwater head of QP aquifer in the monitoring well Q64A
at Hanoi Center (Figs. 1 and 4b).
V.T. Tam, T.T.V. Nga Journal of Environmental Management 227 (2018) 107–116

Fig. 4. (a) Plot of computed versus observed groundwater head in 36 monitoring wells after model calibration (b) Simulated and observed groundwater head of QP
aquifer at Hanoi Center during the model validation period.

4. Results and discussions decades the mean annual rainfall has slightly increased, with predic-
tions that by the year 2030 the total precipitation in the dry season will
4.1. Growing pressure on groundwater resources as increasing in urban decrease by 10.3% while in the wet season will increase by 20.9% in
water demand and decreasing of precipitation comparison with the base line period 1981–2000 (MONRE, 2016).
Because of increasing water use in upstream regions, no compliance
The urbanization process of Hanoi City accompanies with growth of between variation of rainfall and the Red River discharge was observed
urban population and city area, and increase in groundwater ex- for Hanoi City: the annual mean discharge decreases by 9.1% with
ploitation. In the period 1991–2015, the urban area and groundwater decrease of 2.4% in the dry season and 11.9% in the rainy season
abstraction linearly increased with the city's population growth (Table 3).
(Table 2). It is also observed that the annual rainfall is also lightly in-
creasing about 4.4% per year with total rainfall in dry season de- 4.2. Estimation of groundwater recharge components
creasing 17% while in the rainy season increases 12.8% per year
(Table 3). This complies with the latest report on climate change and The annual total recharge to the groundwater system was summed
sea level rise scenarios for Vietnam which states that in the last three up from 6 components: (1) recharge from rain on A & N land, (2) re-
charge from rain on urbanized land, (3) recharge from surface water
area (i.e. lakes, channels, small rivers), (4) recharge from the Red River,
Table 3 (5) leakage from city water supply system, and (6) leakage from city's
Annual and season variation of rainfall and the Red River discharge during sewerage system (Fig. 5a and b). The annual total recharge varies in the
period 1991–2015.
same way as the annual total rainfall but in a descending trend in
Year Annual Annual Dry season Rainy season contradicting to the increasing trend of annual total rainfall (Fig. 5a).
total river This is explained by the notable decrease of recharge from the Red
rainfall discharge Total Monthly Total Monthly
River, which accounts for 29.2% of the annual total recharge (Fig. 5b
(mm) (103 m3/s) rainfall average rainfall average
(mm) discharge (mm) discharge and Table 4). It is noted that both the value and the percentile of most
(103 m3/s) (103 m3/s) of the recharge components are perfectly matching the urbanization
development, i.e. increase in city groundwater exploitation and sewage
1991 1536.50 3380.83 304.5 1408 1232 5460
discharge, and reduction in open surface water area (Fig. 5b, Tables 2
1992 1371.30 2654.50 328.6 1298 1102.3 4162
1993 1442.40 2906.08 259.5 1167 1226.9 4545 and 4). Exception is the reduction of A & N land and the increase of
1994 2536.00 3644.67 217 1226 2249.9 5794 urban land, where corresponding recharge components are in opposite
1995 1220.30 3840.83 208.6 2103 1040.9 5633 way. There, the impact of the increasing rainfall to the groundwater
1996 1596.80 4040.25 317.1 2093 1040 5978
recharge is larger than that of the urbanization development. Due to
1997 2037.51 3553.33 700.5 1758 1581.81 5333
1998 1590.60 3560.83 330.9 2035 1076.9 5502
reduction of recharge, the groundwater level of QH aquifer has declined
1999 1556.60 3455.96 355.9 963 1266.5 5372 (Fig. 6a). Declination of QP aquifer's groundwater level depends less
2000 1278.10 2868.09 398.1 1817 1054.5 4396 likely on the decreasing recharge but rather on the increasing
2001 2264.70 3722.95 272.9 1279 1930.6 5834 groundwater abstraction (Fig. 6b and Table 2). However, because of the
2002 1264.80 3679.90 86 1579 1130.8 6066
combined influence of many factors as previously mentioned, it is very
2003 1582.50 2711.67 260.6 1417 1427.2 4115
2004 970.70 2946.69 245.1 1158 679.4 4552 hard to see synchronic variation of rainfall, recharge, abstraction vo-
2005 1771.80 3084.17 159.2 1560 1535.6 4493 lume and groundwater level.
2006 1240.90 2833.33 203.8 1587 1049 4248
2007 1659.70 3280.83 272.7 1403 1479 5072
4.3. Impact of urbanization on groundwater resources
2008 2267.10 4085.00 207.7 1597 1814.7 5782
2009 1612.10 3078.33 406.4 2532 1471 4603
2010 1239.20 2542.17 155.2 1133 1076.6 3845 The impact of the urbanization process to groundwater resources is
2011 1795.20 2318.33 186.8 1413 1537.3 3287 better envisaged with seasonal analysis. Fig. 5a, c and e show that
2012 1920.40 3185.00 238.1 1403 1700.8 4898 during period from 1991 to 2015 the recharge in both dry and rainy
2013 2676.40 3360.00 228 1643 2336.4 5072
season is reduced yearly about by 1.5% although the rainfall only by
2014 1662.00 2726.67 432.8 1753 1357 3660
2015 1395.00 2670.83 95 1772 1263 3413 about 1.1% per year in the dry season. The Red River and rainfall on A
& N land are the two main sources of decreasing recharge (Fig. 5b).

112
V.T. Tam, T.T.V. Nga
Fig. 5. Variation in annual average, dry season and rainy season for total groundwater recharge, components recharge and rainfall in the period 1991–2015.
Moreover, the results of the simulation indicate that the recharge var-
iation in the dry season is rougher than in the rainy season (Fig. 5d and
f). This can be explained by the rougher recharge from the Red River
and especially by the more rigorous change of rainfall in the dry season
because the recharge to groundwater during that time depends greatly
on soil moisture status and rain intensity. In the rainy season, when the
soil nearly reaches its moisture capacity, the recharge quantity depends
more on the increase in the infiltration area than on the rain intensity
itself. Therefore, the decrease of 6.7% of A & N land area during period
1991–2015 is considered to be the main cause of the reduction of re-
charge from that land although the rainfall increased by 1.5% per year.
The analysis of seasonal variation of groundwater level also reveals
several aspects of the impact of urbanization to groundwater resources.
First, the synchronic variation of groundwater level of QH aquifer
shows the dry season has a stronger influence on recharge compared to
the rainy season (Fig. 6c and e). Secondly, the descending variation of
groundwater level of both aquifers is interpreted as the consequence of
increasing the abstraction from the QP aquifer (due to population
113
Journal of Environmental Management 227 (2018) 107–116
growth) and decrease of recharge (due to the reduction of A & N land).
However, for the QP aquifer the slightly more descending groundwater
level shows more influence of recharge in the dry season in comparison
with the rainy season (Fig. 6d and f).
For further impact assessment of descending recharge (due to the
increase of urbanized area) and increasing groundwater abstraction
(due to urban population growth), two scenarios are simulated and
analyzed: 1 - groundwater abstraction volume was unchanged since
1991 but land use varies during period from 1991 to 2015, and 2 - land
use was unchanged since 1991 but groundwater abstraction volume
varies during the period 1991–2015. The simulated results show that
the scenario 2 causes a more decline of groundwater level than the
scenario 1 (Table 5).
Thus, it can be concluded that the urbanization process has been
impacting on the groundwater resources of Hanoi City in different
ways. The increase of urban population leads to increase of leakage and
seepage from the city's water supply pipe network and sewage channel,
but doesn't impact much on the groundwater resources because this
V.T. Tam, T.T.V. Nga Journal of Environmental Management 227 (2018) 107–116

Table 4
Percentile of each groundwater recharge component for Hanoi City.
Year Total recharge from all sources (106 m3) Component recharge (calculated in % of total recharge)

A & N land Urban land Open surface water Red river City water supply system City sewerage system

1991 611.1 48.6 1.8 2.3 39.1 4.7 3.5

1992 598.2 54.5 1.7 2.3 32.7 5.1 3.6
1993 605.6 50.3 2.1 2.2 36.3 5.6 3.6
1994 785.5 55.5 2.5 1.7 32.8 4.8 2.7
1995 612.2 45.0 2.3 2.1 40.3 6.8 3.5
1996 623.9 43.7 2.9 2.0 41.4 5.9 4.2
1997 676.8 48.9 3.1 1.8 36.6 5.8 3.8
1998 663.5 51.7 4.1 1.8 32.3 6.3 3.9
1999 613.7 45.7 3.6 1.9 37.5 7.1 4.2
2000 546.2 42.9 3.4 2.0 38.8 8.2 4.7
2001 730.1 51.1 4.7 1.5 32.5 5.8 4.4
2002 535.3 36.3 3.9 1.9 43.7 8.3 6.0
2003 535.7 44.1 3.9 1.9 35.4 8.7 6.0
2004 390.2 32.2 2.1 2.5 42.6 12.5 8.2
2005 548.9 50.5 4.8 1.7 28.4 8.8 5.8
2006 451.7 44.7 5.7 2.0 28.6 10.4 8.6
2007 471.6 43.5 6.4 1.9 29.7 10.3 8.2
2008 607.5 49.3 7.2 1.4 27.4 8.4 6.4
2009 425.2 49.7 3.8 1.8 23.2 12.3 9.1
2010 375.1 49.0 8.1 2.0 16.1 14.5 10.3
2011 407.8 53.7 9.6 1.7 11.2 12.2 11.5
2012 549.8 58.0 11.0 1.3 11.9 9.4 8.5
2013 535.8 57.4 11.6 1.3 11.1 9.9 8.7
2014 390.2 50.7 10.5 1.7 10.8 14.2 12.0
2015 382.7 50.4 11.1 1.7 9.5 15.1 12.2
Average 547.0 48.3 5.3 1.8 29.2 8.8 6.6

contributes only 15.6% of total recharge to the groundwater system.
The most important recharge sources are from the Red River and from
precipitation, which contribute with 29.2% and 53.6% to the total re-
charge, respectively. The decrease of A & N land due to urbanization
process results in only a slight decrease of recharge from precipitation,
while the decrease of the Red River discharge and water level is the
main cause of reduced recharge to the groundwater system. This re-
duced recharge influences mostly the QH aquifer, causing the decline of
groundwater level of this aquifer. The biggest impact of the urbaniza-
tion process on groundwater resources is given by the increase of
groundwater withdrawal, which causes severe depletion of the
groundwater level of the QP aquifer.
Comparing with other typical urban hydrological models like ILSAX
(O'Loughlin, 1993), AQUACYCLE (Mitchell et al., 2001) or SWMM
(Rossman, 2015), WetSpa model doesn't comprehensively represent the
spatial and temporal dynamics of the rainfall-runoff response as well as
the characteristics of each components of the urban water hydrological
system i.e. water supply and reuse, wastewater and storm water.
However, when being coupled with MOFLOW the WetSpa model can
provide more understanding about the variation of groundwater re-
charge (which also depends on groundwater level) due to the temporal
and spatial change of urban land use. In terms of assessment of
groundwater recharge for Hanoi City where urbanized land area (which
is considered non-permeable or least permeable) makes up only a small
portion (6.2% in 1990s to 15.3% in 2010s) of total city land area, the
spatial and temporal dynamics of the rainfall-runoff response in the
whole city area is more important than in the urbanized land area.
Therefore, the WetSpa model is more suitable than the above-men-
tioned urban hydrological models, especially for the purpose of asses-
sing the impact of urbanization on groundwater resources in Hanoi. It is
noted that only the WetSpa model's parameters were calibrated using
double ring infiltrometer experiments as described earlier but the
computed groundwater recharge hasn't been validated. However, the
averaged recharge equal to 16.42% of mean annual rainfall computed
with the WetSpa model is comparable to the value of 15%–20% of
rainfall estimated by rainfall infiltration breakthrough (RIB) method
carried out for the nearby province (Roi, 2014). Furthermore, the
results of the present study are consistent with a mass balance analysis
using stable isotopes (δ18O, δD of water) and chloride, showing that the
infiltration from evaporation-affected surface water bodies contributes
more to groundwater recharge than infiltration from the Red River in
the suburban areas of Hanoi City (Kuroda et al., 2017). Nevertheless,
the purpose of this study is not to determine very accurately the re-
charge value, but to assess the impact of the urbanization process to the
groundwater resources of Hanoi City as a whole. In this sense, we de-
monstrated that the coupled WetSpa – MODFLOW modeling gives an
insight to the contribution of each recharge component as well as an in-
depth explanation on the volumetric change of groundwater budget
during the urbanization development.
Compared with studies based on groundwater hydrochemistry and
isotopic analysis of water samples, the coupled hydrological-ground-
water flow modeling approach in this study shows some advantages,
especially for a complicated surface-groundwater interaction system as
in Hanoi City. The numerical modeling approach in this study can
evaluate distinctively and quantitatively the impact of different urba-
nization development factors on groundwater resources, which is very
hard to achieve by single water balance modeling, infiltration mod-
eling, or groundwater hydrochemistry and isotopic analysis alone.
5. Practical application and future research perspectives
From the research results, it is revealed that many aspects of ur-
banization can impact on groundwater resources: land use changes, the
increase of paved and concrete surfaces; increase of groundwater ab-
straction, increase of leakage from the extension of the water supply
and sewerage networks. It is recommended that these aspects should be
respectively evaluated for an overall assessment of impacts of urbani-
zation on groundwater resources. When assessing the impact of urba-
nization on groundwater resources, it is important to quantify the re-
charge components resulting from land use changes. The application of
a distributed hydrological rainfall-runoff modeling is an appropriate
approach for this purpose. The increase of groundwater abstraction
must be considered in the assessment of impacts of urbanization on
groundwater resources for cities where groundwater is the main source

114
V.T. Tam, T.T.V. Nga Journal of Environmental Management 227 (2018) 107–116

Fig. 6. Variation in annual average, dry season and rainy season for total groundwater recharge, groundwater level in QH and QP during the period 1991–2015.

Table 5 for municipal water supply. In this case, the application of coupled
Averaged groundwater level according to simulation of two scenarios: Scenario hydrological-groundwater flow modeling is appropriate.
1 only land use variation and Scenario 2 only groundwater abstraction varia- Groundwater will remain an important source for water supply in
tion. Hanoi and many cities in Vietnam because the surface water system is
Period Average groundwater level of QH Average groundwater level of QP contaminated and the cost of surface water treatment and distribution
aquifer (m) aquifer (m) are very high. It highlights the need of a holistic approach to study and
evaluate the stability and potential risks of urban groundwater system,
Scenario 1 Scenario 2 Scenario 1 Scenario 2
and the coupled hydrological-groundwater flow modeling could served
1991–1995 5.89 5.84 −6.79 −7.03 as a useful tool for the development of management policy or
1996–2000 5.78 5.71 −7.34 −7.96 groundwater resources.
2001–2005 5.64 5.56 −9.34 −9.99
2005–2010 5.56 5.51 −13.42 −13.89
2011–2015 5.40 5.34 −14.42 −15.02
6. Conclusions

A coupled hydrological rainfall–runoff and groundwater flow si-

mulation was carried out to assess the impacts of urbanization on the

115
V.T. Tam, T.T.V. Nga
groundwater resources of Hanoi City. The results of the simulation
suggest that extensive groundwater abstraction due to rapid urban
population growth is the main cause of groundwater level decline in the
city, followed by the descending recharge due to decrease of the Red
River's discharge and water level. Change of land use practice resulting
from urbanization, namely increase of urbanized area and decrease of
agricultural & natural land only causes a slight decrease of groundwater
recharge.
Acknowledgements
This study was carried out within the framework of research project
NCCB-DHUD.2012-G/02, funded by the National Foundation for
Science and Technology Development of Vietnam (NAFOSTED). The
authors are grateful to all the project staffs for their active contribu-
tions. Special thanks to Mr. Tran Viet Hoan from NAWAPI for his en-
thusiastic collaboration in the numerical modeling work of this study.
References
Batisani, N., Yarnal, B., 2009. Urban expansion in Centre County, Pennsylvania: spatial
dynamics and landscape transformations. Appl. Geogr. 29, 235–249. https://doi.org/
10.1016/j.apgeog.2008.08.007.
Blackwood, D., Ellis, J., Revitt, D., Gilmour, D., 2005. Factors influencing exfiltration
processes in sewers. Water Sci. Technol. 51 (2), 147–154.
Bouwer, H., 1978. Groundwater Hydrology. McGraw-Hill, New York.
Brooks, R.H., Corey, A.T., 1966. Properties of porous media affecting fluid flow. J. Irrig.
Drain. Div. ASCE IR2 61–90.
Choi, B., Yun, S., Yu, S., Lee, P., Park, S., Chae, G., Mayer, B., 2005. Hydrochemistry of
urban groundwater in Seoul, South Korea: effects of land-use and pollutant recharge.
Environ. Geol. 48 (8), 979–990. https://doi.org/10.1007/s00254-004-1205-y.
Foster, S.S.D., Lawrence, A.R., Morris, B.M., 1998. Groundwater in Urban Development.
Paper no 390. World Bank Technical, Washington DC, USA.
Foster, S.S.D., Morris, B.L., Lawrence, A.R., 1993. Effects of urbanization on groundwater
recharge. In: Wilkinson, W.B. (Ed.), (1993): Groundwater Problems in Urban Areas.
Proceeding of Institution of Civil Engineers, London, pp. 43–63. June 1993. https://
doi.org/10.1680/gpiua.19744.0005.
Garcia-Fresca, B., Sharp, J.M., 2005. Hydrogeologic considerations of urban develop-
ment: urban-induced recharge. Rev. Eng. Geol. 16, 123–136. https://doi.org/10.
1130/2005.4016(11.
GSO, 2016. Statistical Yearbook of Hanoi 2015. (in Vietnamese).
Hassan, M.M., Nazem, M.N.I., 2016. Examination of land use/land cover changes, urban
growth dynamics, and environmental sustainability in Chittagong city, Bangladesh. J.
Environ. Dev. Sustain. 18 (3), 697–716. https://doi.org/10.1007/s10668-015-
9672-8.
Hillel, D., 1982. Introduction to Soil Physics, first ed. Academic Press, USA.
Hoque, M.A., Hoque, M.M., Ahmed, K.M., 2007. Declining groundwater level and aquifer
dewatering in Dhaka metropolitan area, Bangladesh: causes and quantification.
Hydrogeol. J. 15 (8), 1523–1534. https://doi.org/10.1007/s10040-007-0226-5.
JICA, 2007. The Comprehensive Urban Development Programme in Hanoi Capital City of
the Socialist Republic of Vietnam (HAIDEP) - Final Report. Sections 1, 4. pp. 9–11.
Kim, Y., Lee, K., Sung, I., 2001. Urbanization and the groundwater budget in metropolitan
Seoul area, Korea. Hydrogeol. J. 9 (4), 401–412. https://doi.org/10.1007/
s100400100139.
Kuroda, K., Hayashi, T., Do, A.T., Canh, V.D., Nga, T.T.V., Funabiki, A., Takizawa, S.,
2017. Groundwater recharge in suburban areas of Hanoi, Vietnam: effect of de-
creasing surface-water bodies and land-use change. Hydrogeol. J. 25 (3), 727–742.
https://doi.org/10.1007/s10040-016-1528-2.
Lerner, D.N., Barrett, M.H., 1996. Urban groundwater issues in the UK. Hydrogeol. J. 4
(1), 80–89. https://doi.org/10.1007/s100400050096.
Liu, Y.B., De Smedt, F., Hoffmann, L., Pfister, L., 2004. Assessing land use impact on flood
processes in complex terrain by using GIS and modeling approach. Environ. Model.
Assess. 9 (4), 227–235. https://doi.org/10.1007/s10666-005-0306-7.
116
Journal of Environmental Management 227 (2018) 107–116
Liu, Y.B., Gebremeskel, S., De Smedt, F., Hoffmann, L., Pfister, L., 2003. A diffusive
transport approach for flow routing in GIS-based flood modeling. Hydrogeol. J. 283
(1–4), 91–106. https://doi.org/10.1016/S0022-1694(03)00242-7.
Mao, D., Cherkauer, K.A., 2009. Impacts of land-use change on hydrologic responses in
the Great Lakes region. J. Hydrol. 374 (1–2), 71–82. https://doi.org/10.1016/j.
jhydrol.2009.06.016.
Mitchell, V.G., Mein, R.G., McMahon, T.A., 2001. Modeling the urban water cycle. J.
Environ. Model. Softw. 16 (7), 615–629. https://doi.org/10.1016/S1364-8152(01)
00029-9.
Mohan, M., Pathan, S.K., Narendrareddy, K., Kandya, A., Pandey, S., 2011. Dynamics of
urbanization and its impact on land-use/land-cover: a case study of Megacity Delhi. J.
Environ. Protect. 2, 1274–1283. https://doi.org/10.4236/jep.2011.29147.
MONRE, 2016. Climate Change and Sea Level Rise Scenario for Vietnam. Vietnam Natural
Resources and Environment Publisher (in Vietnamese).
Morris, B.L., Darling, W.G., Cronin, A.A., Rueedi, J., Whitehead, E.J., Gooddy, C., 2006.
Assessing the impact of modern recharge on a sandstone aquifer beneath a suburb
Doncaster, UK. J. Hydrol. 14 (6), 979–997. https://doi.org/10.1007/s10040-006-
0028-1.
Mpamba, N.H., Hussen, A., 2008. Evidence and implications of groundwater mining in
the Lusaka urban aquifers. Phys. Chem. Earth 33 (8–13), 648–654. https://doi.org/
10.1016/j.pce.2008.06.015.
Murakami, A., Zain, M.A., Takeuchi, K., Tsunekawa, A., Yokota, S., 2005. Trends in ur-
banization and patterns of land use in the Asian mega cities Jakarta, Bangkok, and
Metro Manila. Landsc. Urban Plann. 70 (3–4), 251–259. https://doi.org/10.1016/j.
landurbplan.2003.10.021.
Naik, P.K., Tambe, J.A., 2008. Impact of urbanization on the groundwater regime in a fast
growing city in central India. Environ. Monit. Assess. 146 (1–3), 339–373. https://
doi.org/10.1007/s10661-007-0084-6.
On, D.H., 2012. Groundwater exploitable reserves of Hanoi and related issues to be
solved. In: Tuc, N.V. (Ed.), Proceedings of Workshop on Groundwater Exploitation
Reserve of Hanoi City. VUSTA, pp. 48–53 Hanoi March 2012. (in Vietnamese).
O'Loughlin, G., 1993. The ILSAX Program for Urban Storm Water Drainage and Analysis -
User's Manual for Microcomputer. School of Civil Engineering, University of
Technology, Sydney, Australia 2.13.
Quyen, P.H., 2016. Preliminary Report on Aquifer Tests in Hanoi - Governmental Project
of Groundwater Protection in Megacities in Vietnam. Data and Document Archives,
Northern Division for Water Resources Planning and Investigation. (In Vietnamese).
Roi, N.D., 2014. Estimation of groundwater recharge of the Holocen aquifer from rainfall
by RIB method for Hung Yên province. VNU J. Sci. Earth Environ. Sci. 30 (4), 49–63
ISSN 2588-1094.
Rossman, L.A., 2015. Storm Water Management Model - User's Manual Version 5.1. U.S.
EPA.
Sajikumar, N., Remya, R.S., 2015. Impact of land cover and land use change on runoff
characteristics. J. Environ. Manag. 161, 460–468. https://doi.org/10.1016/j.
jenvman.2014.12.041.
Tran, M., 1993. Report on Hydrogeological - Geological Engineering Mapping for Hanoi
City, Scale 1:50.000. Geological Archives, General Department of Geology and
Mineral Resources of Viet Nam (In Vietnamese).
Trowsdale, S.A., Lerner, D.N., 2007. A modeling approach to determine the origin of
urban ground water. J. Contam. Hydrol. 91 (1–2), 171–183. https://doi.org/10.
1016/j.jconhyd.2006.08.011.
Vázquez-Suñé, E., Sánchez-Vila, X., 1999. Groundwater modeling in urban areas as a tool
for local authority management: Barcelona case study (Spain). In: Ellis, J.B. (Ed.),
Impacts of Urban Growth on Surface Water and Groundwater Quality, pp. 65–72
IAHS no. 259.
Vázquez-Suñé, E., Carrera, J., Tubau, I., Sánchez-Vila, X., Soler, A., 2010. An approach to
identify urban groundwater recharge. Hydrol. Earth Syst. Sci. 14, 2085–2097.
https://doi.org/10.5194/hess-14-2085-2010.
VWSA, 2015. Vietnam Water Supply and Sewerage Utility Directory. Vietnam Water
Supply Association.
Wolf, L., Held, I., Eiswirth, M., Hotzl, H., 2004. Impact of leaky sewers on groundwater
quality. Acta Hydrochim. Hydrobiol. 32 (4–5), 361–373. https://doi.org/10.1002/
aheh.200400538.
Xu, J.G., Liao, B.G., Shen, Q., Zhang, F., Mei, A.X., 2007. Urban spatial restructuring in
transitional economy - changing land use pattern in Shanghai. Chin. Geogr. Sci. 17,
19–27.
Yang, Y., Lerner, D., Barrett, M., Tellam, J., 1999. Quantification of groundwater recharge
in the city of Nottingham, UK. Environ. Geol. 38 (3), 183–198.