- Email: [email protected]
Groundwater recharge estimation under semi arid climate: Case of Northern Gafsa watershed, Tunisia
Accepted Manuscript Groundwater recharge estimation under semi arid climate: Case of Northern Gafsa watershed, Tunisia Melki Achraf, Abdollahi Khodayar, Fatahi Rouhallah, Abida Habib PII:
S1464-343X(17)30164-4
DOI:
10.1016/j.jafrearsci.2017.04.020
Reference:
AES 2888
To appear in:
Journal of African Earth Sciences
Received Date: 6 January 2017 Revised Date:
12 April 2017
Accepted Date: 20 April 2017
Please cite this article as: Achraf, M., Khodayar, A., Rouhallah, F., Habib, A., Groundwater recharge estimation under semi arid climate: Case of Northern Gafsa watershed, Tunisia, Journal of African Earth Sciences (2017), doi: 10.1016/j.jafrearsci.2017.04.020. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
ACCEPTED MANUSCRIPT Groundwater recharge estimation under semi arid climate: case of Northern
2
Gafsa watershed, Tunisia
3
a,*
, Abdollahi Khodayar b, Fatahi Rouhallah c, Abida Habib a
Melki Achraf
5
a
6
Tunisia. [email protected]; [email protected]
7
b
Faculty of Earth Sciences, Shahrekord University, Iran, [email protected]
8
c
Water Resources Research Center Shahrekord, Iran, [email protected]
M AN U
SC
Geo-model Laboratory, Faculty of Sciences, University of Sfax, BP : 1171 - 3000 Sfax,
9
10
15
16
EP
14
AC C
13
TE D
11
12
RI PT
4
17
18
19
1
ACCEPTED MANUSCRIPT 1
Abstract Natural groundwater recharge under semi arid climate, like rainfall, is subjected to large
3
variations in both time and space and is therefore very difficult to predict. Nevertheless, in order
4
to set up any strategy for water resources management in such regions, understanding the
5
groundwater recharge variability is essential. This work is interested in examining the impact of
6
rainfall on the aquifer system recharge in the Northern Gafsa Plain in Tunisia.
RI PT
2
The study is composed of two main parts. The first is interested in the analysis of rainfall
8
spatial and temporal variability in the study basin while the second is devoted to the simulation
9
of groundwater recharge. Rainfall analysis was performed based on annual precipitation data
10
recorded in 6 rainfall stations over a period of 56 years (1960-2015). Potential
11
evapotranspiration data were also collected from 1960 to 2011 (52 years).
M AN U
SC
7
The hydrologic distributed model WetSpass was used for the estimation of groundwater
13
recharge. Model calibration was performed based on an assessment of the agreement between
14
the sum of recharge and runoff values estimated by the WetSpass hydrological model and those
15
obtained by the climatic method. This latter is based on the difference calculated between
16
rainfall and potential evapotranspiration recorded at each rainy day. Groundwater recharge
17
estimation, on monthly scale, showed that average annual precipitation (183.3mm/year) was
18
partitioned to 5, 15.3, 36.8, and 42.8% for interception, runoff, actual evapotranspiration and
19
recharge respectively.
21 22
EP
AC C
20
TE D
12
Key words: Rainfall, Tunisia, Groundwater, Semi arid, Actual evapotranspiration, WetSpass
23 24 25
2
ACCEPTED MANUSCRIPT 1
1. Introduction Groundwater recharge assessment is required for sustainable management of water
3
resources, storm water management and subsurface contaminant transport assessment. Previous
4
estimations showed that more than 1.5 billion people worldwide rely on groundwater for
5
drinking (Clarke et al., 1996). As the world's human population is growing, more people will
6
come to rely on groundwater resources, especially in arid and semi arid areas (Simmers, 1990
7
and Hamed, 2011). In this context, tens of millions of dollars have been invested during the
8
previous decades to estimate groundwater recharge rates (Flint et al., 2001), such as the case for
9
the statewide maps of annual recharge for Texas and Minnesota States, which were produced by
SC
Kees (2005) and Lorenz ( 2007).
M AN U
10
RI PT
2
Spatial variation of recharge rates may occur in both systematic and random ways. This
12
is valid for direct (diffuse) recharge and localized (horizontal water movement) recharge.
13
Systematic trends for groundwater recharge are mainly related to climatic factors while other
14
factors like soil, land use and geology may affect the process. It is therefore important to
15
simulate rainfall and understand its variability patterns in order to evaluate its impact on
16
recharge. However, rainfall in arid and semi-arid region is characterized by its erratic
17
spatiotemporal distribution and oceanic influence (Moulin et al., 1997; Guenni and
18
Hutshinson, 1998; Matari et al., 1999; Kipkorior, 2002; Cassou C., 2004).
EP
TE D
11
Precipitation variability is regularly the most important factor affecting recharge rates
20
and it can be considered as the remainder of the rainfall in the water budget for most watersheds.
21
Temporal variability of precipitation in forms of seasonal variation, year-to-year differences, and
22
longer-term trends are also important. In finer resolutions, frequency, duration, and intensity of
23
individual precipitation events also have an effect on recharge rates (Richard, 2010).
AC C
19
24
In Tunisia, groundwater resources are often overexploited by an ever growing
25
population. Indeed, economic development in the region is usually related to access to fresh
26
water, which is rare and unevenly distributed in both time and space. The southern areas of 3
ACCEPTED MANUSCRIPT Tunisia are the most sensitive to the scarcity of fresh water, among which the Gafsa Region
2
where water resources face a big stress under the effect of climatic variation (Recurrence of
3
drought) and overexploitation to satisfy industrial (Phosphate processing and transformation)
4
and agricultural needs. This clearly shows the need for proper management strategies of this
5
precious resource, based on reliable recharge assessment.
RI PT
1
In this context, this work contributes to the evaluation of the water potential of this
7
region through recharge estimation of the biggest groundwater reservoir (Northern Gafsa Basin).
8
The Northern Gafsa Reservoir is composed by 3 horizons which are clay sand of plio-
9
quaternary, clay sand of Miocene and limestone of late cretaceous. These horizons behave as a
10
single hydrogeological entity because of the inexistence of impermeable layers which separate
11
them (Farhat and Moumni, 1989).
M AN U
SC
6
Previous studies on this aquifer system (1572 km2) showed that recharge is essentially
13
influenced by the contributions of floodwaters of both Sidi Aich and Elkebir intermittent streams
14
(Zammouri and Moumni, 1988). Nevertheless, a big difference is detected between recharge
15
estimated by hydrogeological models and recharge simulated based on the contribution by Sidi
16
Aich Stream. In fact, recharge estimated by hydrogeological models is relatively low and does
17
not exceed 1.23 Mm3/year while infiltration modeling showed that Sidi Aich watercourse has an
18
important contribution (4.7Mm3/year) to recharge of the Northern Gafsa groundwater system
19
(Chaieb, 1992). Exploitation rate increased from 148 to 988 l/s during 7 years (1988 to 1995),
20
resulting thereby in a significant drawdown of the groundwater table that may end up with the
21
complete depletion of groundwater resources (Zammouri and Moumni, 1988). A sharp drop in
22
the groundwater level in the deep aquifer resulted in a downward drainage gradient from the
23
shallow aquifer (Yermani, 2002). The presence of fresh groundwater derived from meteoric
24
sources (less than 50 years) was identified based on the stable isotopes of the water molecule,
25
implying a relatively recent groundwater recharge of the shallow aquifer of Northern Gafsa.
AC C
EP
TE D
12
4
ACCEPTED MANUSCRIPT 1
This groundwater is interpreted as contemporaneous recharge at the high-altitude surrounding
2
mountains (Mokadem et al., 2016). This study is interested in examining the interrelationship between yearly rainfall series
4
recorded over a period of 52 years (1960-2011) and groundwater recharge in the aquifer of
5
Northern Gafsa (South-Western Tunisia). This work is composed by two main parts, the first is
6
intended for the analysis of rainfall data using statistical tests (frequency, deciles distribution
7
and the Standard Deviation Index (SDI)) and the second part is devoted to recharge estimation
8
by the climatic method and WetSpass Model (Water and Energy Transfer between Soil, Plants
9
and Atmosphere under quasi-Steady State).
11
2. Study zone and data base
M AN U
10
SC
RI PT
3
The watershed of Northern Gafsa is located in South-Western Tunisia (Fig. 1). It is
13
geographically located between latitudes 34°05' to 34°45' North and longitudes between 8°03'
14
and 9°35' East. Gafsa Watershed is characterized by semi arid climate, with an annual average
15
temperature of 19.7°C. Annual rainfall varies between 150 and 212 mm/year while annual
16
average potential evapotranspiration rate is 1800 mm/year. The annual average relative humidity
17
is 45%. The watershed is characterized by a high density of its stream network in the western
18
part. There are two major watercourses (Sidi Aich and El kebir) and several secondary streams
19
(Fig. 1).
AC C
EP
TE D
12
20
The geological map (Fig. 2) shows that the lower Cretaceous, which represents the oldest
21
formation, occupies a small area in the north of the watershed. The Upper Cretaceous manifests
22
with narrow manner in the frontier zones. Concerning the Mi-Pliocene series, they are mainly
23
presented in the west of the basin and in the extreme south. As for the Quaternary, It is almost
24
present throughout the entire study basin.
5
ACCEPTED MANUSCRIPT The study of the temporal variation of precipitation is based on the analysis of long time
2
series of rainfall in different stations all over the study basin. Rainfall data were compiled from 6
3
rainfall stations during 56 years (1960-2015), as presented in Table 1. The choice of these
4
stations was not arbitrary, but took into account their distribution over the watershed and the
5
extent of the chronology of observations. Climatic data (temperature, wind, potential
6
evapotranspiration) were collected at the meteorological station Gafsa SM.
RI PT
1
Groundwater depth was monitored by 18 piezometers (Table 2). Thematic maps
8
representing topography, slope, soil type, and land use were extracted from the agricultural map
9
of Gafsa, provided by the Regional Department of Agricultural Development. The data base of
10
rainfall, groundwater depth, temperature, wind and potential evapotranspiration was supplied by
11
the General Directorate of Water Resources of the Tunisian Ministry of Agriculture and the
12
National Institute of Meteorology.
13
3. Methodology
TE D
14
M AN U
SC
7
The relationship between precipitation and groundwater recharge is complex as it
16
involves different factors, which vary in both time and space. The approach consists of four
17
different parts: (i) analysis of rainfall data series (ii) estimation of the sum of infiltration and
18
runoff rates based on a water balance equation (iii) calibration of the WetSpass model (Batelaan
19
et al., 2001), and (iv) estimation of monthly distributed groundwater recharge.
AC C
EP
15
20
The first part is devoted to the analysis of the yearly rainfall series by means of statistical
21
tests, including frequency, deciles distribution and Standard Deviation Index (SDI). The
22
frequency analysis is based on yearly classification. Annual rainfall series are classified in an
23
increasing order and then divided into five classes based on their respective frequencies (Table
24
3).
6
ACCEPTED MANUSCRIPT 1
Deciles distribution, developed by Gibbs and Maher (1967), is another method used in the
2
identification and characterization of drought sequences. This method is based on the
3
arrangement of rainfall events recorded in long-term deciles (Table 4).
4
6
The Standard Deviation Index can be calculated by comparing the average annual rainfall (Pm) to the value of standard deviations (σ), as shown in Table 5. σ = [(1/ (N-1)) ∑(Pi-Pm)2 ] 1/2 (1)
7
N and Pi represent the number of samples and annual rainfall for the year i respectively.
SC
8
RI PT
5
9
In the second part of the study, the climatic method is applied to estimate the sum of
11
infiltration and runoff rates. This method is based on the calculation of the difference between
12
rainfall inputs and losses (potential evapotranspiration) (Flippi, 1990). It gives the sum of
13
infiltration and runoff. This was done using daily rainfall time series and potential
14
evapotranspiration for a duration of 52 years (1960-2011). The difference between rainfall and
15
potential evapotranspiration was calculated for all rainy days.
TE D
M AN U
10
The third part is planned to calibrate WetSpass model (Water and Energy Transfer
17
between Soil, Plants and Atmosphere under quasi-Steady State) and to estimate recharge.
18
WetSpass model determines the long-term average spatially distributed recharge as a spatial
19
variable depending on soil texture, land-use, slope, meteorological conditions, etc. This model is
20
based on the flowing relationship (Batelaan et al., 2001):
21
P= IA + ET + R + I
22
With IA, ET, R and I are interception, actual evapotranspiration, runoff and recharge
23
respectively. According to WetSpass model, the total water balance for a raster cell was split
24
into independent water balances for the vegetated, bare-soil, open-water and impervious parts of
25
each cell. This methodology allows one to account for the non-heterogeneity of the land-use per
26
cell, which is dependent on the resolution of the raster cell. The processes in each part of a cell
AC C
EP
16
(2)
7
ACCEPTED MANUSCRIPT are set in a cascading way. This means that an order of occurrence of the processes, after the
2
precipitation event, is assumed. The water balance components of four land use types, including
3
vegetated area, bare-soil surface, open-water, and impervious area are considered to calculate
4
the total cell-based water as follows:
5
ݒܶܧݒܽ=ݎ݁ݐݏܽݎܶܧ+ܽ ݏܧݏ+ܽܧ+ܽ݅( ݅ܧ3)
6
ܵ ݒܵݒܽ=ݎ݁ݐݏܽݎ+ܽ ݏܵݏ+ܽܵ+ܽ݅ܵ݅ (4)
7
ܴ ݒܴݒܽ=ݎ݁ݐݏܽݎ+ܽ ݏܴݏ+ܴܽ+ܴܽ݅݅ (5)
RI PT
1
Where ETraster, Sraster, Rraster are respectively the actual evapotranspiration, surface runoff, and
9
groundwater recharge of a raster cell, each having a vegetated, bare-soil, open-water and impervious area component indicated by proportions av, as, ao, and ai, respectively.
M AN U
10
SC
8
The WetSpass Model has been used for water balance simulation of catchments under
12
humid climate condition. Therefore, it needs to be calibrated in order to be applied for arid and
13
semi arid climates (Abdollahi et al., 2016). The calibration of WetSpass model is based on
14
changing the parameters a, α, LP and A1, which are related to interception, actual
15
evapotranspiration, runoff and recharge respectively. Sutanto et al. (2012) found ‘a’ to be 4.5
16
mm but the final value for the study area can be obtained by calibration.
EP
17
TE D
11
(6)
LAI is the leaf area index, Pm is monthly precipitation [mm/month] and ID represents a
19
minimum daily threshold. LP, the surface runoff calibration parameter (default is 0.65), is a
20
reducing factor for potential evapotranspiration that depends on the soil moisture condition.
21
Pistocchi et al. (2008) suggested α =1.5 as an average value at monthly scale. This soil moisture
22
condition coefficient was used as a starting value for calibration (Abdollahi, 2015).
AC C
18
23
24
CH = 1 if ETm ≤Pm 8
ACCEPTED MANUSCRIPT 1
CH is a coefficient [-] representing soil moisture condition (Bahremand et al., 2007) and ETm is
2
the potential evapotranspiration [mm/month]. A1 is a calibrated parameter in WetSpass related
3
to the infiltration rate of the soil type.
4
Tv is actual transpiration, W is the available water for transpiration, and Trv is the reference
6
transpiration. This model is based on climatic data, such as precipitation, temperature, wind
7
speed, and potential evapotranspiration and physical parameters such as soil type, slope,
8
topography, land use, and groundwater depth. All these data are introduced into the model in the
9
form of ASCII maps.
SC
RI PT
5
11
M AN U
10
4. Results and discussions
Statistical analyses of annual rainfall data, based on frequency, deciles distribution and
13
index of the number of standard deviation, were performed for the rainfall time series, extending
14
over the period (1960-2015). The results obtained showed 41.1% of recurrence of drought (one
15
year in three is dry), split into 26.8% and 14.3% for dry and very dry years respectively. Normal
16
years represent 25% of all data. This analysis also detected 16.1% and 17.9 % for the frequency
17
of humid and very humid years respectively. An increase of drought is detected between the two
18
periods (1960-1979) and (1980-2015). The frequency of drought is intensified from 30 % during
19
1960-1979 to 47.2% in the period (1980-2015) (Table 6).
AC C
EP
TE D
12
20
The sum of infiltration and runoff was estimated by the climatic method as the difference
21
between rainfall and potential evapotranspiration for each rainfall event. The results obtained
22
varied between 0 and 283 mm/month during the period of 52 years (1960-2011), with a monthly
23
average of about 9.4 mm/month (Fig. 3). The comparison between the two periods (1960-1979)
24
and (1980-2011) shows a decrease of the sum of recharge and runoff from 130.8 in the former to
25
100.8 mm/year in the latter. Furthermore, this comparison expresses an increase of dry year 9
ACCEPTED MANUSCRIPT 1
frequency from 30% to 44%, associated with a reduction of rainfall from 212.9 (1960-1979) to
2
174.4 mm/year (1980-2011). Figure 3 also shows 12 exceptional events, where the values of the
3
sum of infiltration and runoff exceed 50 mm/year. The results obtained by the climatic method and WetSpass Model were compared. The
5
corresponding differences in the estimation of the sum of infiltration and runoff quantities were
6
2.4 ; 2.1 ; 1.4 ; -0.4 and 1 mm/month successively for the very dry, dry, normal, humid and very
7
humid years. Large differences were obtained in the very dry period (28.8 mm/year), dry period
8
(25.2 mm/year) and normal period (16.8 mm/year). However, the very humid and humid periods
9
present small differences of 12 mm/year and -4.8 mm/year respectively. This would imply that
10
default parameters of WetSpass model are appropriate for very humid and humid periods. The
11
study area is characterized by semi arid climate, high evapotranspiration and poor vegetation.
M AN U
SC
RI PT
4
WetSpass Model calibration was essentially based on changing parameters a, alfa (α), LP
13
and A1 (Tables 7 and 8) in order to minimize the interception and increase the actual
14
evapotranspiration. The calibration exercise resulted in new differences of 0.2 ; 0.4 ; 0.1 ; -
15
0.4 and 0.3 mm/month for the very dry, dry, normal, humid and very humid years respectively
16
(Fig. 4). This clearly shows that the results of calibration are adequate for all periods and under
17
all drought conditions.
EP
TE D
12
WetSpass Model simulation resulted in dividing average precipitation into 5, 15.3, 36.8
19
and 42.8% for interception, runoff, actual evapotranspiration and recharge respectively.
20
Groundwater recharge during 52 years (1960-2011) in Northern Gafsa Basin varied from 29.8
21
mm in 1994 to 196.4 mm in 1989. The annual recharge average value is 86.4mm/year.
AC C
18
22
The spatial variability of groundwater recharge is shown to be mainly influenced by
23
precipitation, slope and land use. Figure 5, which represents the spatial distribution of
24
groundwater recharge, shows that recharge varied between 67 to 120 mm/year over most of the
10
ACCEPTED MANUSCRIPT 1
basin. The corresponding zones are characterized by low altitudes (300 to 350 m), slopes less
2
than 3% and domination of the rural pathways (low permeability). Recharge values varying between 120 and 151 mm/year are detected in limited regions
4
which are essentially placed in the South East of the basin. These parts are characterized by
5
altitudes varying from 400 to 470 m, slopes less than 2% and the domination of cereal land use.
6
The highest recharge values (151 to 176 mm/year) are mainly detected in limited zones located
7
in the North. These zones are characterized by high topography (480-550 m), feeble slopes
8
(<1%) and olive tree land use (Fig. 5). The estimated recharge rates in the bordering zone,
9
formed mainly by fractured limestone and sand, were shown to vary from 90 to 151mm/year in
10
the south-eastern part of the basin and between 69 and 90 mm/year in the northern and south-
11
western zones
M AN U
SC
RI PT
3
Simulations showed that the study basin is characterized by an average actual
13
evapotranspiration rate of 74.3 mm/year. Its spatial distribution varies from 0 to 1597 mm/year
14
and the entire basin is characterized by a spatial variability of actual evapotranspiration between
15
62 and 75 mm/year, except for the northern zone, where the highest actual evapotranspiration
16
rates occur (75 to 1597 mm/year) (Fig. 6). This high actual evapotranspiration resulted from
17
high precipitation (212 mm/year) (Table 1), existence the open water (sebkha) and moderate
18
permeability of soil.
EP
Concerning runoff rate, the model shows a variation between 3 and 204 mm/year, with
AC C
19
TE D
12
20
an average of 31mm. Most of the basin is characterized by runoff depths not exceeding 19
21
mm/year (Fig. 7), because of low slopes (<3%). The highest runoff depths (81 to 204 mm/year)
22
were detected in the mountainous region, located in the north-east of study basin. Furthermore,
23
relatively important runoff depths (46 to 81 mm/year) were detected in the north and the south
24
of the study basin. The important runoff depths observed in the north are explained by the
25
existence of an open water depression (sebkha) and the high concentration of the stream
26
network. In the South, high runoff is mainly associated with high urbanization rates. 11
ACCEPTED MANUSCRIPT The estimation of interception shows a low rate (< 16 mm/year) over most of the basin
2
(Fig. 8). These low observed values are explained by the feeble density of the vegetation cover
3
and the dominance of the bare soil. The highest observed values (19 to 22 mm/year) coincide
4
with orchard zones. The cereal, palm tree and the market gardening zones are characterized by
5
interception varying from 16 to 19 mm/year.
RI PT
1
6
7
5. Conclusion
Analysis of rainfall data over a period of 56 years (1960-2015) showed that the Basin of
9
Northern Gafsa is characterized by an erratic climatic variation, where dry years are very
10
frequent. The recurrence of rainfall variations in the study area are 41.1, 25 and 34% for dry,
11
normal and humid years respectively. Drought became very frequent, reaching 30% during the
12
period (1960-1979) and 47.2% in the period (1980-2015).
M AN U
SC
8
Annual recharge rates, simulated by WetSpass model over a period of 52 years, were
14
shown to vary between 29.8 to 196.4 mm, with a mean 86.4mm/year. Annual average
15
evapotranspiration, runoff and interception rates were estimated to be 74.3, 30.9 and 10.2
16
mm/year respectively. This estimation shows that average annual rainfall is divided into 5, 15.3,
17
36.8 and 42.8% for interception, runoff, actual evapotranspiration and recharge respectively.
EP
TE D
13
In this study, the impact of water and soil conservation works on the watershed
19
hydrologic response was not taken into consideration. It is therefore important to consider it in
20
subsequent studies in order to elucidate its effect on runoff and recharge rates.
AC C
18
21
22
23
12
ACCEPTED MANUSCRIPT 1
Figures
2 3
Fig. 1. Situation map of Northern Gafsa Watershed.
4
Fig. 2. Geology and spatial distribution of piezometers and rainfall stations in Northern Gafsa.
6
Watershed (Agricultural map of Gafsa, 2008).
RI PT
5
7
Fig. 3. Sum of infiltration and runoff estimated by climatic method.
SC
8 9
Fig. 4. Average deviation of the sum of infiltration and runoff before and after calibration.
M AN U
10 11 12
Fig. 5. Spatial distribution of recharge in Northern Gafsa Aquifer.
13
17 18 19 20 21
TE D
16
Fig. 7. Spatial distribution of runoff in Northern Gafsa Basin.
EP
15
Fig. 6. Spatial distribution of actual evapotranspiration in Northern Gafsa Basin.
Fig. 8. Spatial distribution of interception in Northern Gafsa Basin.
AC C
14
22 23 24 25 26 13
ACCEPTED MANUSCRIPT 1
Tables
2 3
Table 1 Characteristics of rainfall stations and their corresponding time series.
4
Table 2 Characteristics of piezometers.
RI PT
5 6 7
Table 3 Classification of drought based on frequencies (Smith et al., 1993).
Table 4 Classification of drought based on deciles (Hayes et a.,1999).
M AN U
9
SC
8
10
11
Table 5 Drought severity (Beran and Rodier, 1987).
12
16
17
18
19
TE D
15
Table 7 Values of WetSpass model parameters (a, alfa and LP) after calibration.
EP
14
Table 6 Drought classification in Northern Gafsa Basin over the period (1960-2015).
Table 8 Value of WetSpass model parameter (A1) after calibration.
AC C
13
20
21
22 14
ACCEPTED MANUSCRIPT 1
References
2 3
Abdollahi, K., Bashir., I., Verbeiren, B., Harouna, M R., Griensven, V., Huysmans, M., Batelaan, O. 2017. A distributed monthly water balance model: formulation and application
5
on Black Volta Basin. Environ Earth Sci (2017) 76:198 DOI 10.1007/s12665-017-6512-1.
6
Abdollahi, K., 2015. Basin scale water balance modelling for variable hydrological regimes and
8 9
temporal scales. Doc Thesis, Vrije Universiteit Brussel. 65p.
Agricultural map of Gafsa., 2008. Regional Commissariat for Agricultural Development of
SC
7
RI PT
4
Gafsa.
Ampe, EM., Vanhamel, I., Salvadore, E., Dams, J., Bashir, I., Demarchi, L., Batelaan, O., 2012.
11
Impact of urban land-cover classification on groundwater recharge uncertainty. Selected
12
Topics in Applied Earth Observations and Remote Sensing, IEEE Journal 99:1-9.
M AN U
10
Bahremand, A., De Smedt, F., Corluy, J., Liu, YB., Poorova, J., Velcicka, L., Kunikova, E.
14
,2007. WetSpa model application for assessing reforestation impacts on floods in Margecany-
15
Hornad Watershed, Slovakia. Water Resources Management, 21(8):1373-1391.
18 19 20 21 22 23 24 25
methodology for regional groundwater modelling. IAHS Publication 269:11-18.
EP
17
Batelaan, O., De Smedt, F., 2001. WetSpass: a flexible, GIS based, distributed recharge
Batelaan, O., De Smedt, F., 2007. GIS-based recharge estimation by coupling surface-
AC C
16
TE D
13
subsurface water balances. Journal of Hydrology, 337(3):337-355. Beran, MA., Rodier, JA., 1987. Aspects hydrologiques des sécheresses. Contribution au programme hydrologique international. Rapport UNESCO/OMM, 172p. Cassou, C., 2004: From climate change to time regimes: the North Atlantic oscillation. Meteorology 2004. Decadal trend in the North; 46: 21-32. 33. Hurrell JW. Chaieb, H., 1992. Modeling the contribution of Oued Sidi Aich on groundwater recharge of Gafsa North. General Directorate of Water Resources of the Ministry of Agriculture, Tunisia.
15
ACCEPTED MANUSCRIPT 2 3 4 5 6
Clarke, R., Lawrence, A., Foster, S., 1996. Groundwater: A threatened resource. Nairobi, Kenya: United Nations Environment Programme Environment Library No. 15. Farhat,H., Moumni, L., 1989. Hydrogeological study of northern.
Gafsa basin. General
Directorate of Water Resources of the Ministry of Agriculture, Tunisia. Flint, A. L., Flint, L. E., Kwicklis, E. M., Bodvarsson, G. S., Fabryka-Martin, J., 2001.
RI PT
1
Hydrology of Yucca Mountain, Nevada. Rev. Geophys., 39, 447-470.
Hamed, Y., 2011. The hydrogeochemical characterization of groundwater in Gafsa -
8
Hayes, Michael J.; Svoboda, Mark D.; Wilhite, Donald A.; and Vanyarkho, Olga V., 1999
9
"Monitoring the 1996 Drought Using the Standardized Precipitation Index" (1999). Drought
12 13 14 15 16
M AN U
11
Mitigation Center Faculty Publications. Paper 31.
Keese, K. E., Scanlon, B. R., Reedy, R. C., 2005. Assessing controls on diffuse groundwater recharge using unsaturated flow modeling. Water Resour. Res.,41, 1-12. Lorenz, D. L., Delin, G. N., 2007. A regression model to estimate regional groundwater recharge. Ground Water, 45, 196–208.
TE D
10
SC
7
Matari, A., Kerrouchi, M., Bousid, H., Douguedroit, A., 1999. Drought in western Algeria. Publications of the International Association of Climatology, Volume 12, 1999. Mokadem, N., Demdoum, A., Hamed, Y., Bouri, S., Hadji, R., Boyce, A., Laouar, L., Saad,
18
S. 2016. Hydrogeochemical and stable isotope data of groundwater of a multi-aquifer system:
19
Northern Gafsa basin e Central Tunisia. J. Afr. Earth Sci. 114 (2016) 174-191.
21
AC C
20
EP
17
Moulin ,C., Lambert, CE, Dulac F., Dayan U., 1997 : Control of atmospheric export of dust from north Africa by the North Atlantic oscillation. Nature 1997 ; 387 : 691-4.
22
Pistocchi. A., Bouraoui, F., Bittelli, M., 2008. A simplified parameterization of the monthly
23
topsoil water budget. Water Resources Research 44:W12440, doi:10.1029/2007WR006603.
24
Richard, W. Healy., 2010. Book of Estimating Groundwater Recharge,7p.
25
Sid
26
Boubaker
region
(Southwestern
Tunisia).
http://dx.doi.org/10.1007/s12517-011-0393-5. 16
Arabian
J.
Geosciences.
ACCEPTED MANUSCRIPT 1
Simmers, I., 1990. Aridity, groundwater recharge and water resources management. In
2
Groundwater Recharge, A Guide to Understanding and Estimating Natural Recharge.
3
International Contributions to Hydrogeology Vol. 8, ed. D. N. Lerner, A. S. Isaar and I.
4
Simmers. Hanover: Verlag Heinz Heise, 3-22.
7 8
payments and policy. Drought Network News, 11-12.
RI PT
6
Smith, D.I., Hutchinson, M.F., Arthur. M., 1993. Australian climatic and agricultural drought :
Yermani, M., 2002. Study of hydrodynamic of Northern Gafsa aquifer system (central Tunisia). Doc Thesis, University of Sfax, Tunisia.
SC
5
Zammouri, M., Moumni, L., 1988. Simulation of the behavior of Gafsa Northern groundwater
10
in threshold level. General Directorate of Water Resources of the Ministry of Agriculture,
11
Tunisia.
AC C
EP
TE D
M AN U
9
17
1 2 3 4 5
X UTM 482243.11 Sidi Aich 480132.71 Sidi Boubaker 450786.60 Zannouch 504457.54 Ouled Ahmed Saad 501488.13 Hay amaimia 472587.43 Gafsa SM
Y UTM Altitude(m) Average annual rainfall (mm) 300 151 3808011.15 550 212 3843619.62 500 197 3836643.38 374 161 3814154.28 392 166 3829444.28 469 151 3832032.09
AC C
EP
TE D
6
Station name
M AN U
Number
SC
Table 1
RI PT
ACCEPTED MANUSCRIPT
RI PT
ACCEPTED MANUSCRIPT
Station name
X UTM
Y UTM
M AN U
Number
SC
Table 2
Altitude (m)
Depth (m)
Bir Hammem
514578.58
3819452.45
404
60
2
Bir Hir Fej
502373.52
3837079.98
430
49
3
Med Ezzeddine b Ali
487749.56
3810054.55
337
16
4
Med Salah Bouterâa
495509.89
3813620.43
345
10
5
Oglet Jedida
500152.97
3819409.88
347
10
6
Sidi Aich Pz 1
476820.35
3828957.21
423
70
7
Pz Kef Derbi
463391.23
3830590.41
512
134
8
Pz Ouled Mbarek
9
S. Aich Pz 8
10
Souinia
11
El Ayaiech ben Tahar
12
Hedi b Nasser
501629.83
13
Hamed b Ali Ghouma
490897.69
14
Bir Nouayel 1
498547.87
15
Bir Jedid
16
19
3832275.93
442
113
3834841.20
457
100
487673.69
3830153.90
412
51
510778.49
3824867.93
376
39
3830252.01
387
39
3819198.74
359
13
3826709.84
365
24
477122.39
3826676.39
414
47
Lazher Chrayti 2
489306.56
3810206.64
340
11
Ahmed b Med Salah Bir Hir Guettis
496987.85 465663.07
3809594.52 3823141.92
385 409
43
PZ Mzizra
448213.00
3837057.20
507
30
EP
478701.27
478501.01
AC C
17 18
TE D
1
53
M AN U
SC
Table 3
RI PT
ACCEPTED MANUSCRIPT
Class Very dry Dry Normal Humid Very humid
AC C
EP
TE D
F : Frequency F < 0,15 0,15==0,85
Table 4 Distribution Very much less than 20% Less than 20% Close to 20% More than 20% Very much more than 20%
Classification Vey dry Below normal Near normal Above normal Very humid
AC C
EP
TE D
M AN U
Decile 1-2 3-4 5-6 7-8 9-10
SC
RI PT
ACCEPTED MANUSCRIPT
Type of drought Moderate High
AC C
EP
TE D
Very high
criterion of comparison
M AN U
Table 5
SC
RI PT
ACCEPTED MANUSCRIPT
Pm - σ < Pi < Pm
Pm - 2 σ < Pi < Pm - σ Pi < Pm - 2 σ
ACCEPTED MANUSCRIPT
Table 6 Frequency
Standard deviations
Results Very dry Dry Very humid Humid Humid Normal Dry Normal Dry Very humid Dry Humid Very humid Very humid Normal Very humid Very dry Normal Normal Normal Very dry Dry Normal Dry Humid Normal Dry Very dry Dry Very humid Very humid Normal Normal Dry Very dry Very humid Dry Dry Humid Normal Very dry Dry Very humid Humid Dry Humid Humid Very dry Very humid
TE D
M AN U
SC
RI PT
Deciles distribution
AC C
EP
Year 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
ACCEPTED MANUSCRIPT
Normal Humid Normal Very dry Dry Normal Dry
AC C
EP
TE D
M AN U
Very dry Year Dry year Normal year Very humid year Humid year
SC
Legend
RI PT
2009 2010 2011 2012 2013 2014 2015
ACCEPTED MANUSCRIPT
Table 7 Very humid
RI PT
Very dry Original
years
Dry years
Normal years
Humid years
years
a alfa wslope wlanduse wsoil x LP Mean intensity Beta Contribution
4.5 1.5 0.4 0.3 0.3 0.5 0.85 4 0.75 0.5
2 5 0.4 0.3 0.3 0.5 5 4 0.75 0.5
2 5 0.4 0.3 0.3 0.5 5 4 0.75 0.5
4.5 1.5 0.4 0.3 0.3 0.5 0.85 4 0.75 0.5
4.5 1.5 0.4 0.3 0.3 0.5 0.85 4 0.75 0.5
2 5 0.4 0.3 0.3 0.5 5 4 0.75 0.5
AC C
EP
TE D
M AN U
SC
Parameter
ACCEPTED MANUSCRIPT
RI PT
Table 8 Very humid
Original
Very dry years
Dry years
Normal years
clay silty clay sandy clay clayloam silty clayl sandy clayl silt loam silty loam sandy loam loamy sand Sand
0.21 0.23 0.25 0.27 0.29 0.32 0.35 0.37 0.4 0.44 0.47 0.51
0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
0.06 0.07 0.08 0.08 0.09 0.10 0.11 0.11 0.12 0.13 0.14 0.15
0.11 0.12 0.13 0.14 0.15 0.16 0.18 0.19 0.20 0.22 0.24 0.26
M AN U
TE D EP AC C
Humid years
years
0.21 0.23 0.25 0.27 0.29 0.32 0.35 0.37 0.4 0.44 0.47 0.51
0.17 0.18 0.20 0.22 0.23 0.26 0.28 0.30 0.32 0.35 0.38 0.41
SC
Parameter
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 1.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 2.
EP
TE D
M AN U
SC
May-12 Sep-10 Jan-09 May-07 Sep-05 Jan-04 May-02 Sep-00 Jan-99 May-97 Sep-95 Jan-94 May-92 Sep-90 Jan-89 May-87 Sep-85 Jan-84 May-82 Sep-80 Jan-79 May-77 Sep-75 Jan-74 May-72 Sep-70 Jan-69 May-67 Sep-65 Jan-64 May-62 Sep-60
AC C
RI PT 250
Quantity variation (mm)
ACCEPTED MANUSCRIPT
300
200
150
100
50
0
Month
Fig. 3.
3 2.5 2 Average deviation before calibration
1.5
RI PT
Average deviation(mm/month)
ACCEPTED MANUSCRIPT
1 0.5
Average deviation after calibration
0
-1 Humid years
Very humid years
M AN U
Very dry Dry years Normal years years
SC
-0.5
AC C
EP
TE D
Fig. 4.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 5.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 6.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 7.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 8.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Statistical tests, climatic method and WetSpass model were used in this study: • to characterize the temporal rainfall variation in Northern Gafsa Watershed, Tunisia. • to calibrate WetSpass model in semi arid climate. • to estimate the average recharge of Northern Gafsa groundwater during 52 years (1960-2011).
S1464-343X(17)30164-4
DOI:
10.1016/j.jafrearsci.2017.04.020
Reference:
AES 2888
To appear in:
Journal of African Earth Sciences
Received Date: 6 January 2017 Revised Date:
12 April 2017
Accepted Date: 20 April 2017
Please cite this article as: Achraf, M., Khodayar, A., Rouhallah, F., Habib, A., Groundwater recharge estimation under semi arid climate: Case of Northern Gafsa watershed, Tunisia, Journal of African Earth Sciences (2017), doi: 10.1016/j.jafrearsci.2017.04.020. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
ACCEPTED MANUSCRIPT Groundwater recharge estimation under semi arid climate: case of Northern
2
Gafsa watershed, Tunisia
3
a,*
, Abdollahi Khodayar b, Fatahi Rouhallah c, Abida Habib a
Melki Achraf
5
a
6
Tunisia. [email protected]; [email protected]
7
b
Faculty of Earth Sciences, Shahrekord University, Iran, [email protected]
8
c
Water Resources Research Center Shahrekord, Iran, [email protected]
M AN U
SC
Geo-model Laboratory, Faculty of Sciences, University of Sfax, BP : 1171 - 3000 Sfax,
9
10
15
16
EP
14
AC C
13
TE D
11
12
RI PT
4
17
18
19
1
ACCEPTED MANUSCRIPT 1
Abstract Natural groundwater recharge under semi arid climate, like rainfall, is subjected to large
3
variations in both time and space and is therefore very difficult to predict. Nevertheless, in order
4
to set up any strategy for water resources management in such regions, understanding the
5
groundwater recharge variability is essential. This work is interested in examining the impact of
6
rainfall on the aquifer system recharge in the Northern Gafsa Plain in Tunisia.
RI PT
2
The study is composed of two main parts. The first is interested in the analysis of rainfall
8
spatial and temporal variability in the study basin while the second is devoted to the simulation
9
of groundwater recharge. Rainfall analysis was performed based on annual precipitation data
10
recorded in 6 rainfall stations over a period of 56 years (1960-2015). Potential
11
evapotranspiration data were also collected from 1960 to 2011 (52 years).
M AN U
SC
7
The hydrologic distributed model WetSpass was used for the estimation of groundwater
13
recharge. Model calibration was performed based on an assessment of the agreement between
14
the sum of recharge and runoff values estimated by the WetSpass hydrological model and those
15
obtained by the climatic method. This latter is based on the difference calculated between
16
rainfall and potential evapotranspiration recorded at each rainy day. Groundwater recharge
17
estimation, on monthly scale, showed that average annual precipitation (183.3mm/year) was
18
partitioned to 5, 15.3, 36.8, and 42.8% for interception, runoff, actual evapotranspiration and
19
recharge respectively.
21 22
EP
AC C
20
TE D
12
Key words: Rainfall, Tunisia, Groundwater, Semi arid, Actual evapotranspiration, WetSpass
23 24 25
2
ACCEPTED MANUSCRIPT 1
1. Introduction Groundwater recharge assessment is required for sustainable management of water
3
resources, storm water management and subsurface contaminant transport assessment. Previous
4
estimations showed that more than 1.5 billion people worldwide rely on groundwater for
5
drinking (Clarke et al., 1996). As the world's human population is growing, more people will
6
come to rely on groundwater resources, especially in arid and semi arid areas (Simmers, 1990
7
and Hamed, 2011). In this context, tens of millions of dollars have been invested during the
8
previous decades to estimate groundwater recharge rates (Flint et al., 2001), such as the case for
9
the statewide maps of annual recharge for Texas and Minnesota States, which were produced by
SC
Kees (2005) and Lorenz ( 2007).
M AN U
10
RI PT
2
Spatial variation of recharge rates may occur in both systematic and random ways. This
12
is valid for direct (diffuse) recharge and localized (horizontal water movement) recharge.
13
Systematic trends for groundwater recharge are mainly related to climatic factors while other
14
factors like soil, land use and geology may affect the process. It is therefore important to
15
simulate rainfall and understand its variability patterns in order to evaluate its impact on
16
recharge. However, rainfall in arid and semi-arid region is characterized by its erratic
17
spatiotemporal distribution and oceanic influence (Moulin et al., 1997; Guenni and
18
Hutshinson, 1998; Matari et al., 1999; Kipkorior, 2002; Cassou C., 2004).
EP
TE D
11
Precipitation variability is regularly the most important factor affecting recharge rates
20
and it can be considered as the remainder of the rainfall in the water budget for most watersheds.
21
Temporal variability of precipitation in forms of seasonal variation, year-to-year differences, and
22
longer-term trends are also important. In finer resolutions, frequency, duration, and intensity of
23
individual precipitation events also have an effect on recharge rates (Richard, 2010).
AC C
19
24
In Tunisia, groundwater resources are often overexploited by an ever growing
25
population. Indeed, economic development in the region is usually related to access to fresh
26
water, which is rare and unevenly distributed in both time and space. The southern areas of 3
ACCEPTED MANUSCRIPT Tunisia are the most sensitive to the scarcity of fresh water, among which the Gafsa Region
2
where water resources face a big stress under the effect of climatic variation (Recurrence of
3
drought) and overexploitation to satisfy industrial (Phosphate processing and transformation)
4
and agricultural needs. This clearly shows the need for proper management strategies of this
5
precious resource, based on reliable recharge assessment.
RI PT
1
In this context, this work contributes to the evaluation of the water potential of this
7
region through recharge estimation of the biggest groundwater reservoir (Northern Gafsa Basin).
8
The Northern Gafsa Reservoir is composed by 3 horizons which are clay sand of plio-
9
quaternary, clay sand of Miocene and limestone of late cretaceous. These horizons behave as a
10
single hydrogeological entity because of the inexistence of impermeable layers which separate
11
them (Farhat and Moumni, 1989).
M AN U
SC
6
Previous studies on this aquifer system (1572 km2) showed that recharge is essentially
13
influenced by the contributions of floodwaters of both Sidi Aich and Elkebir intermittent streams
14
(Zammouri and Moumni, 1988). Nevertheless, a big difference is detected between recharge
15
estimated by hydrogeological models and recharge simulated based on the contribution by Sidi
16
Aich Stream. In fact, recharge estimated by hydrogeological models is relatively low and does
17
not exceed 1.23 Mm3/year while infiltration modeling showed that Sidi Aich watercourse has an
18
important contribution (4.7Mm3/year) to recharge of the Northern Gafsa groundwater system
19
(Chaieb, 1992). Exploitation rate increased from 148 to 988 l/s during 7 years (1988 to 1995),
20
resulting thereby in a significant drawdown of the groundwater table that may end up with the
21
complete depletion of groundwater resources (Zammouri and Moumni, 1988). A sharp drop in
22
the groundwater level in the deep aquifer resulted in a downward drainage gradient from the
23
shallow aquifer (Yermani, 2002). The presence of fresh groundwater derived from meteoric
24
sources (less than 50 years) was identified based on the stable isotopes of the water molecule,
25
implying a relatively recent groundwater recharge of the shallow aquifer of Northern Gafsa.
AC C
EP
TE D
12
4
ACCEPTED MANUSCRIPT 1
This groundwater is interpreted as contemporaneous recharge at the high-altitude surrounding
2
mountains (Mokadem et al., 2016). This study is interested in examining the interrelationship between yearly rainfall series
4
recorded over a period of 52 years (1960-2011) and groundwater recharge in the aquifer of
5
Northern Gafsa (South-Western Tunisia). This work is composed by two main parts, the first is
6
intended for the analysis of rainfall data using statistical tests (frequency, deciles distribution
7
and the Standard Deviation Index (SDI)) and the second part is devoted to recharge estimation
8
by the climatic method and WetSpass Model (Water and Energy Transfer between Soil, Plants
9
and Atmosphere under quasi-Steady State).
11
2. Study zone and data base
M AN U
10
SC
RI PT
3
The watershed of Northern Gafsa is located in South-Western Tunisia (Fig. 1). It is
13
geographically located between latitudes 34°05' to 34°45' North and longitudes between 8°03'
14
and 9°35' East. Gafsa Watershed is characterized by semi arid climate, with an annual average
15
temperature of 19.7°C. Annual rainfall varies between 150 and 212 mm/year while annual
16
average potential evapotranspiration rate is 1800 mm/year. The annual average relative humidity
17
is 45%. The watershed is characterized by a high density of its stream network in the western
18
part. There are two major watercourses (Sidi Aich and El kebir) and several secondary streams
19
(Fig. 1).
AC C
EP
TE D
12
20
The geological map (Fig. 2) shows that the lower Cretaceous, which represents the oldest
21
formation, occupies a small area in the north of the watershed. The Upper Cretaceous manifests
22
with narrow manner in the frontier zones. Concerning the Mi-Pliocene series, they are mainly
23
presented in the west of the basin and in the extreme south. As for the Quaternary, It is almost
24
present throughout the entire study basin.
5
ACCEPTED MANUSCRIPT The study of the temporal variation of precipitation is based on the analysis of long time
2
series of rainfall in different stations all over the study basin. Rainfall data were compiled from 6
3
rainfall stations during 56 years (1960-2015), as presented in Table 1. The choice of these
4
stations was not arbitrary, but took into account their distribution over the watershed and the
5
extent of the chronology of observations. Climatic data (temperature, wind, potential
6
evapotranspiration) were collected at the meteorological station Gafsa SM.
RI PT
1
Groundwater depth was monitored by 18 piezometers (Table 2). Thematic maps
8
representing topography, slope, soil type, and land use were extracted from the agricultural map
9
of Gafsa, provided by the Regional Department of Agricultural Development. The data base of
10
rainfall, groundwater depth, temperature, wind and potential evapotranspiration was supplied by
11
the General Directorate of Water Resources of the Tunisian Ministry of Agriculture and the
12
National Institute of Meteorology.
13
3. Methodology
TE D
14
M AN U
SC
7
The relationship between precipitation and groundwater recharge is complex as it
16
involves different factors, which vary in both time and space. The approach consists of four
17
different parts: (i) analysis of rainfall data series (ii) estimation of the sum of infiltration and
18
runoff rates based on a water balance equation (iii) calibration of the WetSpass model (Batelaan
19
et al., 2001), and (iv) estimation of monthly distributed groundwater recharge.
AC C
EP
15
20
The first part is devoted to the analysis of the yearly rainfall series by means of statistical
21
tests, including frequency, deciles distribution and Standard Deviation Index (SDI). The
22
frequency analysis is based on yearly classification. Annual rainfall series are classified in an
23
increasing order and then divided into five classes based on their respective frequencies (Table
24
3).
6
ACCEPTED MANUSCRIPT 1
Deciles distribution, developed by Gibbs and Maher (1967), is another method used in the
2
identification and characterization of drought sequences. This method is based on the
3
arrangement of rainfall events recorded in long-term deciles (Table 4).
4
6
The Standard Deviation Index can be calculated by comparing the average annual rainfall (Pm) to the value of standard deviations (σ), as shown in Table 5. σ = [(1/ (N-1)) ∑(Pi-Pm)2 ] 1/2 (1)
7
N and Pi represent the number of samples and annual rainfall for the year i respectively.
SC
8
RI PT
5
9
In the second part of the study, the climatic method is applied to estimate the sum of
11
infiltration and runoff rates. This method is based on the calculation of the difference between
12
rainfall inputs and losses (potential evapotranspiration) (Flippi, 1990). It gives the sum of
13
infiltration and runoff. This was done using daily rainfall time series and potential
14
evapotranspiration for a duration of 52 years (1960-2011). The difference between rainfall and
15
potential evapotranspiration was calculated for all rainy days.
TE D
M AN U
10
The third part is planned to calibrate WetSpass model (Water and Energy Transfer
17
between Soil, Plants and Atmosphere under quasi-Steady State) and to estimate recharge.
18
WetSpass model determines the long-term average spatially distributed recharge as a spatial
19
variable depending on soil texture, land-use, slope, meteorological conditions, etc. This model is
20
based on the flowing relationship (Batelaan et al., 2001):
21
P= IA + ET + R + I
22
With IA, ET, R and I are interception, actual evapotranspiration, runoff and recharge
23
respectively. According to WetSpass model, the total water balance for a raster cell was split
24
into independent water balances for the vegetated, bare-soil, open-water and impervious parts of
25
each cell. This methodology allows one to account for the non-heterogeneity of the land-use per
26
cell, which is dependent on the resolution of the raster cell. The processes in each part of a cell
AC C
EP
16
(2)
7
ACCEPTED MANUSCRIPT are set in a cascading way. This means that an order of occurrence of the processes, after the
2
precipitation event, is assumed. The water balance components of four land use types, including
3
vegetated area, bare-soil surface, open-water, and impervious area are considered to calculate
4
the total cell-based water as follows:
5
ݒܶܧݒܽ=ݎ݁ݐݏܽݎܶܧ+ܽ ݏܧݏ+ܽܧ+ܽ݅( ݅ܧ3)
6
ܵ ݒܵݒܽ=ݎ݁ݐݏܽݎ+ܽ ݏܵݏ+ܽܵ+ܽ݅ܵ݅ (4)
7
ܴ ݒܴݒܽ=ݎ݁ݐݏܽݎ+ܽ ݏܴݏ+ܴܽ+ܴܽ݅݅ (5)
RI PT
1
Where ETraster, Sraster, Rraster are respectively the actual evapotranspiration, surface runoff, and
9
groundwater recharge of a raster cell, each having a vegetated, bare-soil, open-water and impervious area component indicated by proportions av, as, ao, and ai, respectively.
M AN U
10
SC
8
The WetSpass Model has been used for water balance simulation of catchments under
12
humid climate condition. Therefore, it needs to be calibrated in order to be applied for arid and
13
semi arid climates (Abdollahi et al., 2016). The calibration of WetSpass model is based on
14
changing the parameters a, α, LP and A1, which are related to interception, actual
15
evapotranspiration, runoff and recharge respectively. Sutanto et al. (2012) found ‘a’ to be 4.5
16
mm but the final value for the study area can be obtained by calibration.
EP
17
TE D
11
(6)
LAI is the leaf area index, Pm is monthly precipitation [mm/month] and ID represents a
19
minimum daily threshold. LP, the surface runoff calibration parameter (default is 0.65), is a
20
reducing factor for potential evapotranspiration that depends on the soil moisture condition.
21
Pistocchi et al. (2008) suggested α =1.5 as an average value at monthly scale. This soil moisture
22
condition coefficient was used as a starting value for calibration (Abdollahi, 2015).
AC C
18
23
24
CH = 1 if ETm ≤Pm 8
ACCEPTED MANUSCRIPT 1
CH is a coefficient [-] representing soil moisture condition (Bahremand et al., 2007) and ETm is
2
the potential evapotranspiration [mm/month]. A1 is a calibrated parameter in WetSpass related
3
to the infiltration rate of the soil type.
4
Tv is actual transpiration, W is the available water for transpiration, and Trv is the reference
6
transpiration. This model is based on climatic data, such as precipitation, temperature, wind
7
speed, and potential evapotranspiration and physical parameters such as soil type, slope,
8
topography, land use, and groundwater depth. All these data are introduced into the model in the
9
form of ASCII maps.
SC
RI PT
5
11
M AN U
10
4. Results and discussions
Statistical analyses of annual rainfall data, based on frequency, deciles distribution and
13
index of the number of standard deviation, were performed for the rainfall time series, extending
14
over the period (1960-2015). The results obtained showed 41.1% of recurrence of drought (one
15
year in three is dry), split into 26.8% and 14.3% for dry and very dry years respectively. Normal
16
years represent 25% of all data. This analysis also detected 16.1% and 17.9 % for the frequency
17
of humid and very humid years respectively. An increase of drought is detected between the two
18
periods (1960-1979) and (1980-2015). The frequency of drought is intensified from 30 % during
19
1960-1979 to 47.2% in the period (1980-2015) (Table 6).
AC C
EP
TE D
12
20
The sum of infiltration and runoff was estimated by the climatic method as the difference
21
between rainfall and potential evapotranspiration for each rainfall event. The results obtained
22
varied between 0 and 283 mm/month during the period of 52 years (1960-2011), with a monthly
23
average of about 9.4 mm/month (Fig. 3). The comparison between the two periods (1960-1979)
24
and (1980-2011) shows a decrease of the sum of recharge and runoff from 130.8 in the former to
25
100.8 mm/year in the latter. Furthermore, this comparison expresses an increase of dry year 9
ACCEPTED MANUSCRIPT 1
frequency from 30% to 44%, associated with a reduction of rainfall from 212.9 (1960-1979) to
2
174.4 mm/year (1980-2011). Figure 3 also shows 12 exceptional events, where the values of the
3
sum of infiltration and runoff exceed 50 mm/year. The results obtained by the climatic method and WetSpass Model were compared. The
5
corresponding differences in the estimation of the sum of infiltration and runoff quantities were
6
2.4 ; 2.1 ; 1.4 ; -0.4 and 1 mm/month successively for the very dry, dry, normal, humid and very
7
humid years. Large differences were obtained in the very dry period (28.8 mm/year), dry period
8
(25.2 mm/year) and normal period (16.8 mm/year). However, the very humid and humid periods
9
present small differences of 12 mm/year and -4.8 mm/year respectively. This would imply that
10
default parameters of WetSpass model are appropriate for very humid and humid periods. The
11
study area is characterized by semi arid climate, high evapotranspiration and poor vegetation.
M AN U
SC
RI PT
4
WetSpass Model calibration was essentially based on changing parameters a, alfa (α), LP
13
and A1 (Tables 7 and 8) in order to minimize the interception and increase the actual
14
evapotranspiration. The calibration exercise resulted in new differences of 0.2 ; 0.4 ; 0.1 ; -
15
0.4 and 0.3 mm/month for the very dry, dry, normal, humid and very humid years respectively
16
(Fig. 4). This clearly shows that the results of calibration are adequate for all periods and under
17
all drought conditions.
EP
TE D
12
WetSpass Model simulation resulted in dividing average precipitation into 5, 15.3, 36.8
19
and 42.8% for interception, runoff, actual evapotranspiration and recharge respectively.
20
Groundwater recharge during 52 years (1960-2011) in Northern Gafsa Basin varied from 29.8
21
mm in 1994 to 196.4 mm in 1989. The annual recharge average value is 86.4mm/year.
AC C
18
22
The spatial variability of groundwater recharge is shown to be mainly influenced by
23
precipitation, slope and land use. Figure 5, which represents the spatial distribution of
24
groundwater recharge, shows that recharge varied between 67 to 120 mm/year over most of the
10
ACCEPTED MANUSCRIPT 1
basin. The corresponding zones are characterized by low altitudes (300 to 350 m), slopes less
2
than 3% and domination of the rural pathways (low permeability). Recharge values varying between 120 and 151 mm/year are detected in limited regions
4
which are essentially placed in the South East of the basin. These parts are characterized by
5
altitudes varying from 400 to 470 m, slopes less than 2% and the domination of cereal land use.
6
The highest recharge values (151 to 176 mm/year) are mainly detected in limited zones located
7
in the North. These zones are characterized by high topography (480-550 m), feeble slopes
8
(<1%) and olive tree land use (Fig. 5). The estimated recharge rates in the bordering zone,
9
formed mainly by fractured limestone and sand, were shown to vary from 90 to 151mm/year in
10
the south-eastern part of the basin and between 69 and 90 mm/year in the northern and south-
11
western zones
M AN U
SC
RI PT
3
Simulations showed that the study basin is characterized by an average actual
13
evapotranspiration rate of 74.3 mm/year. Its spatial distribution varies from 0 to 1597 mm/year
14
and the entire basin is characterized by a spatial variability of actual evapotranspiration between
15
62 and 75 mm/year, except for the northern zone, where the highest actual evapotranspiration
16
rates occur (75 to 1597 mm/year) (Fig. 6). This high actual evapotranspiration resulted from
17
high precipitation (212 mm/year) (Table 1), existence the open water (sebkha) and moderate
18
permeability of soil.
EP
Concerning runoff rate, the model shows a variation between 3 and 204 mm/year, with
AC C
19
TE D
12
20
an average of 31mm. Most of the basin is characterized by runoff depths not exceeding 19
21
mm/year (Fig. 7), because of low slopes (<3%). The highest runoff depths (81 to 204 mm/year)
22
were detected in the mountainous region, located in the north-east of study basin. Furthermore,
23
relatively important runoff depths (46 to 81 mm/year) were detected in the north and the south
24
of the study basin. The important runoff depths observed in the north are explained by the
25
existence of an open water depression (sebkha) and the high concentration of the stream
26
network. In the South, high runoff is mainly associated with high urbanization rates. 11
ACCEPTED MANUSCRIPT The estimation of interception shows a low rate (< 16 mm/year) over most of the basin
2
(Fig. 8). These low observed values are explained by the feeble density of the vegetation cover
3
and the dominance of the bare soil. The highest observed values (19 to 22 mm/year) coincide
4
with orchard zones. The cereal, palm tree and the market gardening zones are characterized by
5
interception varying from 16 to 19 mm/year.
RI PT
1
6
7
5. Conclusion
Analysis of rainfall data over a period of 56 years (1960-2015) showed that the Basin of
9
Northern Gafsa is characterized by an erratic climatic variation, where dry years are very
10
frequent. The recurrence of rainfall variations in the study area are 41.1, 25 and 34% for dry,
11
normal and humid years respectively. Drought became very frequent, reaching 30% during the
12
period (1960-1979) and 47.2% in the period (1980-2015).
M AN U
SC
8
Annual recharge rates, simulated by WetSpass model over a period of 52 years, were
14
shown to vary between 29.8 to 196.4 mm, with a mean 86.4mm/year. Annual average
15
evapotranspiration, runoff and interception rates were estimated to be 74.3, 30.9 and 10.2
16
mm/year respectively. This estimation shows that average annual rainfall is divided into 5, 15.3,
17
36.8 and 42.8% for interception, runoff, actual evapotranspiration and recharge respectively.
EP
TE D
13
In this study, the impact of water and soil conservation works on the watershed
19
hydrologic response was not taken into consideration. It is therefore important to consider it in
20
subsequent studies in order to elucidate its effect on runoff and recharge rates.
AC C
18
21
22
23
12
ACCEPTED MANUSCRIPT 1
Figures
2 3
Fig. 1. Situation map of Northern Gafsa Watershed.
4
Fig. 2. Geology and spatial distribution of piezometers and rainfall stations in Northern Gafsa.
6
Watershed (Agricultural map of Gafsa, 2008).
RI PT
5
7
Fig. 3. Sum of infiltration and runoff estimated by climatic method.
SC
8 9
Fig. 4. Average deviation of the sum of infiltration and runoff before and after calibration.
M AN U
10 11 12
Fig. 5. Spatial distribution of recharge in Northern Gafsa Aquifer.
13
17 18 19 20 21
TE D
16
Fig. 7. Spatial distribution of runoff in Northern Gafsa Basin.
EP
15
Fig. 6. Spatial distribution of actual evapotranspiration in Northern Gafsa Basin.
Fig. 8. Spatial distribution of interception in Northern Gafsa Basin.
AC C
14
22 23 24 25 26 13
ACCEPTED MANUSCRIPT 1
Tables
2 3
Table 1 Characteristics of rainfall stations and their corresponding time series.
4
Table 2 Characteristics of piezometers.
RI PT
5 6 7
Table 3 Classification of drought based on frequencies (Smith et al., 1993).
Table 4 Classification of drought based on deciles (Hayes et a.,1999).
M AN U
9
SC
8
10
11
Table 5 Drought severity (Beran and Rodier, 1987).
12
16
17
18
19
TE D
15
Table 7 Values of WetSpass model parameters (a, alfa and LP) after calibration.
EP
14
Table 6 Drought classification in Northern Gafsa Basin over the period (1960-2015).
Table 8 Value of WetSpass model parameter (A1) after calibration.
AC C
13
20
21
22 14
ACCEPTED MANUSCRIPT 1
References
2 3
Abdollahi, K., Bashir., I., Verbeiren, B., Harouna, M R., Griensven, V., Huysmans, M., Batelaan, O. 2017. A distributed monthly water balance model: formulation and application
5
on Black Volta Basin. Environ Earth Sci (2017) 76:198 DOI 10.1007/s12665-017-6512-1.
6
Abdollahi, K., 2015. Basin scale water balance modelling for variable hydrological regimes and
8 9
temporal scales. Doc Thesis, Vrije Universiteit Brussel. 65p.
Agricultural map of Gafsa., 2008. Regional Commissariat for Agricultural Development of
SC
7
RI PT
4
Gafsa.
Ampe, EM., Vanhamel, I., Salvadore, E., Dams, J., Bashir, I., Demarchi, L., Batelaan, O., 2012.
11
Impact of urban land-cover classification on groundwater recharge uncertainty. Selected
12
Topics in Applied Earth Observations and Remote Sensing, IEEE Journal 99:1-9.
M AN U
10
Bahremand, A., De Smedt, F., Corluy, J., Liu, YB., Poorova, J., Velcicka, L., Kunikova, E.
14
,2007. WetSpa model application for assessing reforestation impacts on floods in Margecany-
15
Hornad Watershed, Slovakia. Water Resources Management, 21(8):1373-1391.
18 19 20 21 22 23 24 25
methodology for regional groundwater modelling. IAHS Publication 269:11-18.
EP
17
Batelaan, O., De Smedt, F., 2001. WetSpass: a flexible, GIS based, distributed recharge
Batelaan, O., De Smedt, F., 2007. GIS-based recharge estimation by coupling surface-
AC C
16
TE D
13
subsurface water balances. Journal of Hydrology, 337(3):337-355. Beran, MA., Rodier, JA., 1987. Aspects hydrologiques des sécheresses. Contribution au programme hydrologique international. Rapport UNESCO/OMM, 172p. Cassou, C., 2004: From climate change to time regimes: the North Atlantic oscillation. Meteorology 2004. Decadal trend in the North; 46: 21-32. 33. Hurrell JW. Chaieb, H., 1992. Modeling the contribution of Oued Sidi Aich on groundwater recharge of Gafsa North. General Directorate of Water Resources of the Ministry of Agriculture, Tunisia.
15
ACCEPTED MANUSCRIPT 2 3 4 5 6
Clarke, R., Lawrence, A., Foster, S., 1996. Groundwater: A threatened resource. Nairobi, Kenya: United Nations Environment Programme Environment Library No. 15. Farhat,H., Moumni, L., 1989. Hydrogeological study of northern.
Gafsa basin. General
Directorate of Water Resources of the Ministry of Agriculture, Tunisia. Flint, A. L., Flint, L. E., Kwicklis, E. M., Bodvarsson, G. S., Fabryka-Martin, J., 2001.
RI PT
1
Hydrology of Yucca Mountain, Nevada. Rev. Geophys., 39, 447-470.
Hamed, Y., 2011. The hydrogeochemical characterization of groundwater in Gafsa -
8
Hayes, Michael J.; Svoboda, Mark D.; Wilhite, Donald A.; and Vanyarkho, Olga V., 1999
9
"Monitoring the 1996 Drought Using the Standardized Precipitation Index" (1999). Drought
12 13 14 15 16
M AN U
11
Mitigation Center Faculty Publications. Paper 31.
Keese, K. E., Scanlon, B. R., Reedy, R. C., 2005. Assessing controls on diffuse groundwater recharge using unsaturated flow modeling. Water Resour. Res.,41, 1-12. Lorenz, D. L., Delin, G. N., 2007. A regression model to estimate regional groundwater recharge. Ground Water, 45, 196–208.
TE D
10
SC
7
Matari, A., Kerrouchi, M., Bousid, H., Douguedroit, A., 1999. Drought in western Algeria. Publications of the International Association of Climatology, Volume 12, 1999. Mokadem, N., Demdoum, A., Hamed, Y., Bouri, S., Hadji, R., Boyce, A., Laouar, L., Saad,
18
S. 2016. Hydrogeochemical and stable isotope data of groundwater of a multi-aquifer system:
19
Northern Gafsa basin e Central Tunisia. J. Afr. Earth Sci. 114 (2016) 174-191.
21
AC C
20
EP
17
Moulin ,C., Lambert, CE, Dulac F., Dayan U., 1997 : Control of atmospheric export of dust from north Africa by the North Atlantic oscillation. Nature 1997 ; 387 : 691-4.
22
Pistocchi. A., Bouraoui, F., Bittelli, M., 2008. A simplified parameterization of the monthly
23
topsoil water budget. Water Resources Research 44:W12440, doi:10.1029/2007WR006603.
24
Richard, W. Healy., 2010. Book of Estimating Groundwater Recharge,7p.
25
Sid
26
Boubaker
region
(Southwestern
Tunisia).
http://dx.doi.org/10.1007/s12517-011-0393-5. 16
Arabian
J.
Geosciences.
ACCEPTED MANUSCRIPT 1
Simmers, I., 1990. Aridity, groundwater recharge and water resources management. In
2
Groundwater Recharge, A Guide to Understanding and Estimating Natural Recharge.
3
International Contributions to Hydrogeology Vol. 8, ed. D. N. Lerner, A. S. Isaar and I.
4
Simmers. Hanover: Verlag Heinz Heise, 3-22.
7 8
payments and policy. Drought Network News, 11-12.
RI PT
6
Smith, D.I., Hutchinson, M.F., Arthur. M., 1993. Australian climatic and agricultural drought :
Yermani, M., 2002. Study of hydrodynamic of Northern Gafsa aquifer system (central Tunisia). Doc Thesis, University of Sfax, Tunisia.
SC
5
Zammouri, M., Moumni, L., 1988. Simulation of the behavior of Gafsa Northern groundwater
10
in threshold level. General Directorate of Water Resources of the Ministry of Agriculture,
11
Tunisia.
AC C
EP
TE D
M AN U
9
17
1 2 3 4 5
X UTM 482243.11 Sidi Aich 480132.71 Sidi Boubaker 450786.60 Zannouch 504457.54 Ouled Ahmed Saad 501488.13 Hay amaimia 472587.43 Gafsa SM
Y UTM Altitude(m) Average annual rainfall (mm) 300 151 3808011.15 550 212 3843619.62 500 197 3836643.38 374 161 3814154.28 392 166 3829444.28 469 151 3832032.09
AC C
EP
TE D
6
Station name
M AN U
Number
SC
Table 1
RI PT
ACCEPTED MANUSCRIPT
RI PT
ACCEPTED MANUSCRIPT
Station name
X UTM
Y UTM
M AN U
Number
SC
Table 2
Altitude (m)
Depth (m)
Bir Hammem
514578.58
3819452.45
404
60
2
Bir Hir Fej
502373.52
3837079.98
430
49
3
Med Ezzeddine b Ali
487749.56
3810054.55
337
16
4
Med Salah Bouterâa
495509.89
3813620.43
345
10
5
Oglet Jedida
500152.97
3819409.88
347
10
6
Sidi Aich Pz 1
476820.35
3828957.21
423
70
7
Pz Kef Derbi
463391.23
3830590.41
512
134
8
Pz Ouled Mbarek
9
S. Aich Pz 8
10
Souinia
11
El Ayaiech ben Tahar
12
Hedi b Nasser
501629.83
13
Hamed b Ali Ghouma
490897.69
14
Bir Nouayel 1
498547.87
15
Bir Jedid
16
19
3832275.93
442
113
3834841.20
457
100
487673.69
3830153.90
412
51
510778.49
3824867.93
376
39
3830252.01
387
39
3819198.74
359
13
3826709.84
365
24
477122.39
3826676.39
414
47
Lazher Chrayti 2
489306.56
3810206.64
340
11
Ahmed b Med Salah Bir Hir Guettis
496987.85 465663.07
3809594.52 3823141.92
385 409
43
PZ Mzizra
448213.00
3837057.20
507
30
EP
478701.27
478501.01
AC C
17 18
TE D
1
53
M AN U
SC
Table 3
RI PT
ACCEPTED MANUSCRIPT
Class Very dry Dry Normal Humid Very humid
AC C
EP
TE D
F : Frequency F < 0,15 0,15=
Table 4 Distribution Very much less than 20% Less than 20% Close to 20% More than 20% Very much more than 20%
Classification Vey dry Below normal Near normal Above normal Very humid
AC C
EP
TE D
M AN U
Decile 1-2 3-4 5-6 7-8 9-10
SC
RI PT
ACCEPTED MANUSCRIPT
Type of drought Moderate High
AC C
EP
TE D
Very high
criterion of comparison
M AN U
Table 5
SC
RI PT
ACCEPTED MANUSCRIPT
Pm - σ < Pi < Pm
Pm - 2 σ < Pi < Pm - σ Pi < Pm - 2 σ
ACCEPTED MANUSCRIPT
Table 6 Frequency
Standard deviations
Results Very dry Dry Very humid Humid Humid Normal Dry Normal Dry Very humid Dry Humid Very humid Very humid Normal Very humid Very dry Normal Normal Normal Very dry Dry Normal Dry Humid Normal Dry Very dry Dry Very humid Very humid Normal Normal Dry Very dry Very humid Dry Dry Humid Normal Very dry Dry Very humid Humid Dry Humid Humid Very dry Very humid
TE D
M AN U
SC
RI PT
Deciles distribution
AC C
EP
Year 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
ACCEPTED MANUSCRIPT
Normal Humid Normal Very dry Dry Normal Dry
AC C
EP
TE D
M AN U
Very dry Year Dry year Normal year Very humid year Humid year
SC
Legend
RI PT
2009 2010 2011 2012 2013 2014 2015
ACCEPTED MANUSCRIPT
Table 7 Very humid
RI PT
Very dry Original
years
Dry years
Normal years
Humid years
years
a alfa wslope wlanduse wsoil x LP Mean intensity Beta Contribution
4.5 1.5 0.4 0.3 0.3 0.5 0.85 4 0.75 0.5
2 5 0.4 0.3 0.3 0.5 5 4 0.75 0.5
2 5 0.4 0.3 0.3 0.5 5 4 0.75 0.5
4.5 1.5 0.4 0.3 0.3 0.5 0.85 4 0.75 0.5
4.5 1.5 0.4 0.3 0.3 0.5 0.85 4 0.75 0.5
2 5 0.4 0.3 0.3 0.5 5 4 0.75 0.5
AC C
EP
TE D
M AN U
SC
Parameter
ACCEPTED MANUSCRIPT
RI PT
Table 8 Very humid
Original
Very dry years
Dry years
Normal years
clay silty clay sandy clay clayloam silty clayl sandy clayl silt loam silty loam sandy loam loamy sand Sand
0.21 0.23 0.25 0.27 0.29 0.32 0.35 0.37 0.4 0.44 0.47 0.51
0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
0.06 0.07 0.08 0.08 0.09 0.10 0.11 0.11 0.12 0.13 0.14 0.15
0.11 0.12 0.13 0.14 0.15 0.16 0.18 0.19 0.20 0.22 0.24 0.26
M AN U
TE D EP AC C
Humid years
years
0.21 0.23 0.25 0.27 0.29 0.32 0.35 0.37 0.4 0.44 0.47 0.51
0.17 0.18 0.20 0.22 0.23 0.26 0.28 0.30 0.32 0.35 0.38 0.41
SC
Parameter
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 1.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 2.
EP
TE D
M AN U
SC
May-12 Sep-10 Jan-09 May-07 Sep-05 Jan-04 May-02 Sep-00 Jan-99 May-97 Sep-95 Jan-94 May-92 Sep-90 Jan-89 May-87 Sep-85 Jan-84 May-82 Sep-80 Jan-79 May-77 Sep-75 Jan-74 May-72 Sep-70 Jan-69 May-67 Sep-65 Jan-64 May-62 Sep-60
AC C
RI PT 250
Quantity variation (mm)
ACCEPTED MANUSCRIPT
300
200
150
100
50
0
Month
Fig. 3.
3 2.5 2 Average deviation before calibration
1.5
RI PT
Average deviation(mm/month)
ACCEPTED MANUSCRIPT
1 0.5
Average deviation after calibration
0
-1 Humid years
Very humid years
M AN U
Very dry Dry years Normal years years
SC
-0.5
AC C
EP
TE D
Fig. 4.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 5.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 6.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 7.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 8.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Statistical tests, climatic method and WetSpass model were used in this study: • to characterize the temporal rainfall variation in Northern Gafsa Watershed, Tunisia. • to calibrate WetSpass model in semi arid climate. • to estimate the average recharge of Northern Gafsa groundwater during 52 years (1960-2011).