Applied Water Science (2018) 8:193

https://doi.org/10.1007/s13201-018-0840-5

ORIGINAL ARTICLE

Sensitive analysis of ground recharge estimation model, for semiarid

areas
Salah Eddine Ali Rahmani1 · Brahim Chibane1 · Abdelkader Boucefiane2

Received: 24 January 2017 / Accepted: 1 October 2018

© The Author(s) 2018

Abstract
The rapid demographic evolution wields a big pressure on water resources potentialities; in addition, the climate change
effects have influenced the natural renovation rate of water resources around the world. Starting from a previous published
model, a sensitive analysis was established after introducing three parameters as watershed topography, stream network den-
sity and geology index to illustrate the impact of each of these parameters on the rate of groundwater recharge. This modified
model has been improved by in-site experiment (soil chloride mass balance) tested in some selected sites in Sidi Slimane
River located in the region of Djelfa at 300 km from Algiers (Algeria). The results obtained show that the parameters like
geology, topography and the amount of rainfall and the drainage density influence the velocity of groundwater recharge. By
introducing the chloride concentration in soil, the model becomes most independent and gives good results that can be used
in water resources planning in semiarid land.

Keywords Water resources potentialities · Climate change · Groundwater recharge · Chloride mass balance · Sidi Slimane
River

Introduction Regions of Africa and the Middle East are the most
areas influenced by these changes (Rochdane et al. 2012).
Water resources management becomes a serious priority for Groundwater resources have a high importance in these
local authorities and the international community. In area areas because it secures more than 70% of water demand of
where the availability of water depends on the climate condi- all active sectors like agricultural, industrial and domestic
tions, people are happy in wet year and obviously sad in the activities (Foster and Chilton 2003).
dry hydrological year. Changes in recharge are controlled A strategic plan must be implemented to protect aquifers
by the global warming influences on the dynamic of water from a probable depletion and to solve problems of water
cycle around the world (Alcamo et al. 2007). Reduction in unavailability in the short and medium range. Many efforts
rainfall due to climate change in semiarid areas impacts should be made to maintain and optimize the utilization
these effects were summarized by the intergovernmental of groundwater resources in semiarid areas (Loucif Seiad
panel on climate changes (IPCC) (Arnel and Lieu 2001). 2003).
Increasing the temperature of earth will influence the oce- The good quantification of the rate of groundwater
anic thermodynamic in large scale; brutal changes will effect recharge is based on two principal ideas:
the ocean–atmosphere exchange which influences the global
atmospheric circulation. • The availability of sufficient data with good quality.
• The relationship between the used technique and the
physical reality of the watershed, for example, methods
* Salah Eddine Ali Rahmani used to quantify recharge in shallow groundwater aqui-
[email protected] fers cannot be used for karstic aquifers where the velocity
of groundwater flow is more important than the other
1
Geo‑Environment Laboratory FSTGAT​, USTHB, Bo. 32, El hydrogeological systems.
Alia, Bab Ezzouar, 16111 Algiers, Algeria
2
Earth Sciences Department, University of Djilali Bounaama,
Khemis Miliana, Algeria

13
Vol.:(0123456789)
193 Page 2 of 10 Applied Water Science (2018) 8:193

In this paper, we present a previous published model The new model and aquifers’ artificial recharge
(Ali Rahmani et al. 2016) to quantify groundwater recharge and recovery
by introducing some parameters to make the model more
useful and dynamic for all semiarid sites. It can be used in Artificial recharge of aquifers is a needed strategy for water
geographic information systems (GIS) to establish map of resources reinforcement in semiarid regions. The new model
spatial variation of groundwater recharge and map of sensi- will be of interest to select the favorable area for aquifer
tive areas to different sources of pollution. recharge based on five input layers (temperature, rainfall,
slope, stream network density and the geology of the basin).

Groundwater recharge process

Materials and methods
The infiltration rate of rainfall depends on the soils’ texture,
the geologic setting of the recharge areas, the topographical Study area
configuration of the watershed and essentially the hydrocli-
matic conditions (precipitation, temperature, wind speed and The Oued Djelfa-Melah watershed (Fig. 2) was located
direction, the air moisture). at 290 km south to the capital of Algeria (Algiers). It was
In summer, the high amount of rainfall results from under Mediterranean climate (semiarid type), with a cold
aggressive storms that cannot reach the aquifer, but it and rainy winter and a hot, dry summer, with annual rainfall
returns the atmosphere by the bias of the evapotranspira- of around 336 mm. The inter-annual average temperature is
tion; in this case, the evapotranspiration intensity is greater 14.6 °C. The calculation of the water balance by the method
than the recharge process. The recharge process consists of of Thornthwaite revealed that evapotranspiration gives a
three forms: localized recharge, direct recharge and indirect value of 804.66 mm and an ETR of 278.12 mm (Chibane
recharge (inter-flow) (Fig. 1). 2010) (Fig. 3).

Fig. 1  Groundwater recharge mechanism (De Vries and Simmers 2002) modified

13
Applied Water Science (2018) 8:193 Page 3 of 10 193

Fig. 2  Study area (watershed of Djelfa-Melah River)

The region of Djelfa is formed by different geologi- The Albian Struck flushes all around the syncline, and it
cal deposits (Fig. 4); the major deposits are described in consists of massive sandstone fine. There are also of gray
chronological order as the following (Chibane and Ali- clays. The lower Albian forms a productive aquifer and its
Rahmani 2015). thickness is about 400 m. The upper Albian is about 200 m
The Triassic appears in the form of diapirs; the outcrop thick, consisting of limestone and marl alternating; this part
of this type of deposits is known in the Djelfa region by represents a waterproof part.
the “Salt Rock,” located at 20 km north of Djelfa. The Turonian formation is composed of fractured lime-
The Neocomian is impermeable bedrock; it forms the stone with alternations of marls in its uppermost part. Its
reliefs of El-Kef-Haouas and Djebel Aissa, with a thick- thickness is around 450 m, and it forms a karstic aquifer.
ness that reaches 800 m. It is mainly formed by series of The Mio-Plio-Quaternary is mainly sandy, silty and con-
clay sandstone. glomeratic and is surmounted by a calcareous crust; due to
The Barremian formation is essentially consisting of its low permeability, the Mio-Plio Quaternary is operated by
alternating sandstones and sandy clays intensely cracked, small wells; its thickness is variable (250–300 m).
the Barremian forms the major aquifer capacity, and its
thickness ranges from 1500 to 2000 m. Model context
The Aptian forms a waterproof substratum (Aquitard)
of the Albian and shows a continuous narrow strip that Based on a recent model published by Ali Rahmani et al.
appears in the northern and southern flanks, discontinuous, (2016), a sensitive study of the physical context of this hydro-
sometimes hidden by quaternary deposits. The thickness logical-based model was established by introducing two fac-
of this formation is about 200 m to the west and 100 m to tors: topography of the basin and stream network density to
the east.

13
193 Page 4 of 10 Applied Water Science (2018) 8:193

Fig. 3  Sidi Slimane River (March 2016)

extend the model efficiency; the calibration process was done where Gw is the groundwater recharge estimated by Ali
by using soil chloride mass balance (Allison 1988). Rahmani et al. (2016) model, Gf is the geological factor
The general form of the equation of this model is given by that describes the permeability of soil, Tf is the topography
Eq. (1), and the detailed description is given in Table 1: factor that describes the basin inclination, and Ddf is the
drainage density expressed as a part of average runoff.
Gw = 𝛼 × [P − (P − 1) × 𝛽] × 𝛾, (1) Based on theoretical data about characteristics of geologi-
where α and β are the temperature-dependent coefficients cal deposits, topographical settings and drainage density, a
and γ is the adjustment coefficient. weighed rate was given for each of these parameters as the
following:
Modified groundwater recharge (Gwm)
Geological factor (Gf)
Equation (1) is modified by introducing another three param-
eters to make the model more representative of the hydrologi- Groundwater recharge rate depends not only on the amount
cal process; these three parameters are topography coefficient, of rainfall, but also on the geological settings of the water-
geology index and stream network density. The efficiency shed which are important in the recharge process. Perme-
of the model reduces the over-estimation of groundwater able deposits permit rapid infiltration against impermeable
recharge and makes our water resources planning strategy (Castany 1982).
more reasonable. Here, the geological factor was introduced in the modified
By the combination of these factors, the new modified model as a fraction of the total effective precipitation (Ban-
model can be used to draw a potential recharge map and ton and Bangoy 1997; Duda et al. 2011). These coefficients
area with high and low vulnerability of aquifers to pollution are given in Table 2.
sources.
The general form of the modified model equation is given Topography of the watershed (Tf)
by Eq. (2):
Gwm = Gw × Gf × Tf + Ddf, (2) Topography is an essential parameter in the hydrologi-
cal process (runoff, sediments transport and groundwater
recharge).

13
Applied Water Science (2018) 8:193 Page 5 of 10 193

Fig. 4  Geological map of the study area

Table 1  Parameters of the Abbreviation Description Unit Equation

model
Gw Average annual rate of groundwater mm Gw = 𝛼 × [P − (P − 1) × 𝛽] × 𝛾
recharge
P Average annual rainfall mm –
α Temperature coefficient – 𝛼= T 2 −1
2
√ −12T
T
β Temperature coefficient – 𝛽= T 2 −1
T
γ Adjustment coefficient – 𝛾=T 0.21

Table 2  Geological settings factor value for permeable, semipermeable and impermeable layers
Geology of basin Gf Description

Impermeable deposits 0.1 (less than 1%) Very low permeability does not mean there is low recharge, but the
velocity of recharge is very low; recharge in these deposits can be
done by a phenomena called delayed recharge (silt, gypsum, clay
rocks)
Medium permeability 0.5 (less than 4%) Medium permeability (sand, marl, sandy loam, etc.)
High permeability 0.9 (less than 6%) High permeability capacity (gravels, alluvium)

13
193 Page 6 of 10 Applied Water Science (2018) 8:193

Table 3  Topographical factor (Tf) values

Topography (slope) Tf Description

High 0 Zone with high slope (more than 20°) does not promote infiltration, because the inclination of the layer gives kinetic
energy to the precipitations and decreases the time of concentration of water which arises its velocity; in this case,
the rainfall will transform to surface runoff
Medium 0.5 Medium slope (less than 15°) participates with medium range in the groundwater recharge process; it lets infiltration
of only 50% of the possible infiltration volume
Low 1 Low slope helps in groundwater recharge by reducing the velocity of surface runoff and lets the direct infiltration of
rainfall by increasing the time of concentration of precipitation in a specific area (gravels, alluvium)

To give the model more aptitude to adapt with physical Results and discussion
context of the watershed, a topographical weight factor was
assigned to the new modified model. The selection of slope The main results are discussed in this section of the paper,
coefficient is uniformized according to the digital elevation starting by calculating the average recharge for normal con-
model. ditions in relation with basin topography, geological mor-
The description of these weights is given in Table 3. phology and drainage density factors. The results obtained
were compared with in-site experiments.
Drainage density The novelty of this modified model in this paper is the
three parameters introduced in the main model developed
The stream network along the basin describes the stream by Ali Rahmani et al. (2016). Introduced parameters take in
flow and surface runoff in function of the length of each interest the physical form of the watershed, the geology and
stream. The influence of drainage density factor was intro- the stream networks in the entire basin (Table 4).
duced in the model as a fraction of average annual runoff The model becomes more dynamic and elastic, and it can
that supplies the aquifer by the bias of the inter-flow phe- give a very good result to quantify groundwater recharge rate
nomena (McClain et al. 2014). in semiarid area.
Topography has great effects on groundwater recharge
Limit of application of the model process. The slope controls the velocity of the runoff and
the groundwater hydraulic which means every time the slope
To apply the model, the following conditions must be veri- is high; the precipitations will transform to direct runoff.
fied to get good results: Topography factor was devised to three classes: The first
class has 0 weight for slope more than 20°; in this case, the
• The average annual temperature must be great than or topography favors the runoff and the groundwater recharge
equal to 12.8 °C. may approach zero.
• The hydrological system must be closed, and exchange The second class has 0.5 weight, for less than 15°; in this
between watersheds must be assumed negligible. situation, a small part of precipitation will infiltrate and the
other part will transform to direct runoff.
The last class has 1 weight, when the slope is less than
8°; velocity of runoff is low, and precipitation becomes more
concentrated in the land to promote an effective infiltration
of rainwater.

Table 4  Drainage density class

Drainage density (Dd) Dd Description

Low 0 Low drainage density means that there is a little aquifer–river exchange, and the aquifer supply was done
only by effective
rainfall
Medium Dd1% Medium density participates with medium range in the groundwater recharge process; it lets infiltration
of 1% of the surface runoff via ephemeral rivers
High Dd10% High Dd helps in groundwater recharge by increasing the infiltration via streams and rivers (inter-flow
phenomena)

(0.01×P)2.4 (0.01×P)2.4
With Dd1% = 100
and Dd10% = 10
, P is the average annual runoff

13
Applied Water Science (2018) 8:193 Page 7 of 10 193

Chloride mass balance where qCl is the chloride flux at the land surface, CCl is the
chloride concentration in soil water, q is the water flux, and
Theoretical background Dh is the hydrodynamic dispersion coefficient.
The term Dh zCl is assumed negligible for some semiarid
c

Several researchers have used chloride profiles to calculate areas (Bromley et al. 1997).
water fluxes in the unsaturated zone (Allison and Hughes This method helps us to calculate recharge by taking into
1983; Bromley et al. 1997; Scanlon 1991; Scanlon and Gold- account the chloride balance between the atmospheric dray
smith 1997). and wet deposit and the chloride concentration in soils. This
Chloride is a useful geochemical tracer of water move- tracer-based method is useful in semiarid and arid areas.
ment in the unsaturated zone because it is chemically con- Thirteen soil profiles were realized in the Sidi Slimane
servative and generally is not a favorite element to vegetation River in the region of Djelfa, we have analyzed the chloride
and plants. The one-dimensional chloride transport in the concentration in soil at different depths, the standard depth
unsaturated zone under steady-state conditions is described used for all layers is equal to 100 cm, and the details about
by the following equation (Scanlon and Goldsmith 1997): the method are reported by Ali Rahmani et al. (2017).
cCl q Figure 5 shows the infiltration rate via the unsaturated
qCl = CCl q − Dh q = Cl ,
z CCl layers along the two banks of the river.
Groundwater recharge variation is shown in Fig. 5, it was
calculated using the chloride mass balance, and the total
atmospheric flux of chloride was measured to be 15 mg/l

Fig. 5  Chloride concentration Recharge[mm/year] Chloride concentration [mg/l]

and groundwater recharge cal- 0.00 9000

Chloride concentration mg/l

culated for the 13 profiles (Ali 8000
Infiltration rate [mm/year]

5.00
Rahmani et al. 2017) 7000
10.00
6000
15.00
5000
20.00
4000
25.00
3000
30.00 2000
35.00 1000
40.00 0
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13
Soil Profile

Fig. 6  Saline bedrock shown

in the 2D contour plot for the -20
profiles P9–P12 C l-[mg/l]
< 1500
-30 1500 – 3000
3000 – 4500
-40 4500 – 6000
6000 – 7500
> 7500
-50

-60

-70

-80

-90

-100
1 2 3 4
Position

13
193 Page 8 of 10 Applied Water Science (2018) 8:193

(Moulla and Guendouz 1996) for a mean average annual parameters let the water resources manager to establish a
rainfall of 320 mm (Fig. 6). map of potential recharge.
The recharge rate calculated along the river border ranges Figure 8 shows the map recharge obtained from this
from 0.10 to 37.35 mm/year with an average of 12.39 mm/ model, in the region of study. The map shown in Fig. 8
year. shows the spatial variation of groundwater recharge. It
The chloride concentration in soil profiles was var- ranges from 2.7 to 24 mm. The mean is 12 mm/year which
ied between 21.42 and 8150.14 mg/l with a mean of approaches the values founded in Sidi Slimane River
509.87 mg/l with a coefficient of variation of 253% which (12.38 mm/year).
explains the heterogynous of soil profiles in the horizon-
tal and vertical axes along the river. T test was established
to be the value of average estimated groundwater recharge Recommendations
by the soil profile methods; the T test criteria are defined
if the average of recharge is less than 15 mm/year, and the This new experiment lets us to conclude that ground recharge
results of this test are shown in Fig. 7. process is not dependent only on two entries (precipitations
This test lets us to calculate the corrected average and temperature); by introducing another three parameters,
annual recharge with a confident bound of 90%; the value Ali Rahmani et al. model becomes more dynamic to be used
lies between 10.489 and 14.286 mm/year. for the complicated semiarid areas where recharge is limited
The model of Ali Rahmani et al. gives for the same by topographical, geological and drainage density factors.
precipitation 13 mm/year which lets us to conclude that However, the chloride mass balance experiment gives us
the model efficiency is good. In the same way, the idea a good understanding of the recharge process across the
presented in this paper is the dependence of the ground- unsaturated zone, where the climate change and soil layers
water recharge process on the climatic conditions, on the are the major factors that determine which quantity of water
geological settings and on the topographical factors and will reach the saturated zone.
drainage density of the basin. The advantages given by this model are related to five
However, the experimental study lets us to confirm the principal parameters (precipitation, temperature, geol-
result given by the new modified model; the introduced ogy, topography and drainage density of the basin) which

Statistics
0 0,05 0,1 > 0,5 Sample size 65
Yes No Mean 12,388
90% CI (10,489; 14,286)
P = 0,012 Standard deviation 9,1727
The mean of GWRcmb is significantly less than the target (p Target 15
< 0,05).

Where are the data relative to the target?

15 Comments

-- Test: You can conclude that the mean is less than 15 at

the 0,05 level of significance.
-- CI: Quantifies the uncertainty associated with estimating
the mean from sample data. You can be 90% confident
that the true mean is between 10,489 and 14,286.
-- Distribution of Data: Compare the location of the data to
the target. Look for unusual data before interpreting the
test results.

0 8 16 24 32

Fig. 7  Sample t test for the mean of GWRcmb

13
Applied Water Science (2018) 8:193 Page 9 of 10 193

Fig. 8  Groundwater recharge map

make the estimation of groundwater recharge rate with low the aquifer response in function of two variables: the rapid
uncertainty. demographic growth and climate change scenario.

Open Access This article is distributed under the terms of the Crea-
tive Commons Attribution 4.0 International License (http://creat​iveco​
Conclusion mmons​.org/licen​ses/by/4.0/), which permits unrestricted use, distribu-
tion, and reproduction in any medium, provided you give appropriate
Recharge estimation is effectuated by introducing a modi- credit to the original author(s) and the source, provide a link to the
fied previous model. This model is the combination of five Creative Commons license, and indicate if changes were made.
input parameters: rainfall, temperature, geology, topography
and drainage density. The calibration of model was done by
using a chloride mass balance designated to the unsaturated
zone. The in-site experiment shows that the groundwater References
recharge rate varies from 0.10 to 37.35 mm/year with an
Alcamo J, Florke M, Marker M (2007) Future long-term changes in
average of 12.39 mm/year. The model gives an average of global water resources driven by socio-economic and climatic
13 mm/year. This approach will help water resources man- changes. Hydrol Sci J 52(2):247–275. https​://doi.org/10.1623/
ager in semiarid areas to evaluate the annual rate of aqui- hysj.52.2.247
fer recharge to control the abstraction of water and manage Ali Rahmani SE, Chibane B, Boucefiène A (2016) Groundwa-
ter recharge estimation in semi-arid zone: a study case from
water more rationally. The importance of this model is to the region of Djelfa (Algeria). Appl Water Sci. https​: //doi.
establish a map of recharge that helps us in the delimitation org/10.1007/s1320​1-016-0399-y
of areas with high water potentiality. Our actual research Ali Rahmani SE, Chibane B, Hallouz F, Boutelba A, Aiachi N
are focused to perform this model to use it in simulation of (2017) New modified green-AMPT models to calculate water

13
193 Page 10 of 10 Applied Water Science (2018) 8:193

flux via unsaturated layers, case of Sidi Slimane river (Algeria). text explanations (in Polish with English summary). Ministerstwo
LARHYSS J 32:1112–3680 Œrodowiska, Warszawa
Allison GB (1988) A review of some of the physical, chemical and iso- Foster SSD, Chilton PJ (2003) Groundwater: the processes and global
topic techniques available for estimation ground-water recharge. significance of aquifer degradation. Philos Trans R Soc Lond B
In: Simmers I (ed) Estimation of natural groundwater recharge. 358:1957–1972
Reidel Publishing Co., Dordrecht, pp 49–72 Loucif Seiad N (2003) Les ressources en eau et leurs utilisations dans
Allison GB, Hughes MW (1983) The use of natural tracers as indica- le secteur Agricole en Algérie. Revue H.T.E. No. 125 mars 2003
tors of soil–water movement in a temperate semi-arid region. J McClain ME, Subalusky AL, Anderson EP, Dessu SB, Melesse AM,
Hydrol 60:157–173 Ndomba PM, Mtamba JOD, Tamatamah RA, Mligo C (2014)
Arnel N, Lieu C (2001) Hydrology and water resources. In: Climate Comparing flow regime, channel hydraulics and biological com-
change, 2001, impact, adaptation, and vulnerability. Intergovern- munities to infer flow–ecology relationships in the Mara River of
mental panel on climate change. Cambridge University Press, Kenya and Tanzania. Hydrol Sci J 59(3–4):801–819
Cambridge Moulla AS, Guendouz A (1996) A survey of rainwater chemical and
Banton O, Bangoy LM (1997) Hydrogéologie—multiscience envi- isotopic composition in North Algeria. Séminaire International
ronnementale des eaux souterraines. Aupelf Uref. Presses De sur l’Application des Techniques Isotopiques dans le Domaine des
l’Université Du Québec, p 460 Ressources en Eau et en Sol. Mahdia, Tunisia, 14–18 June 1996
Bromley J, Edmunds WM, Fellman E, Brouwer J, Gaze SR, Sudlow J, Rochdane S, Reichert B, Messouli M, Babqiqi A, Khebiza MY (2012)
Taupin JD (1997) Estimation of rainfall inputs and direct recharge Climate change impacts on water supply and demand in Rheraya
to the deep unsaturated zone of southern Niger using the chloride Watershed (Morocco), with potential adaptation strategies. Water
profile method. J Hydrol 188–189:139–154 4:28–44. https​://doi.org/10.3390/w4010​028
Castany G (1982) Principes et méthodes de l’hydrogéologie. Dunod, Scanlon BR (1991) Evaluation of moisture flux from chloride data in
Paris, France desert soils. J Hydrol 128:137–156
Chibane B (2010) Hydrogeological and hydrogeochemical study in Scanlon BR, Goldsmith RS (1997) Field study of spatial variability in
semi-arid area: case study of the Djelfa region. PhD thesis in unsaturated flow beneath and adjacent to playas. Water Resour
hydrogeology, FSTGAT, USTHB, Algeria Res 33:2239–2252
Chibane B, Ali-Rahmani SE (2015) Hydrological based model to esti-
mate groundwater recharge, real evapotranspiration and runoff in Publisher’s Note Springer Nature remains neutral with regard to
semi-arid area. LARHYSS J 23(23):1112–3680 jurisdictional claims in published maps and institutional affiliations.
De Vries JJ, Simmers I (2002) Groundwater recharge: an overview
of processes and challenges. Hydrol J 10:5–17. https​: //doi.
org/10.1007/s1004​0-001-0171-7
Duda R, Witczak S, Zurek A (2011) Map of ground water vulnerability
to pollution in Poland, a scale of 1:500,000—methodology and

13