Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938

https://doi.org/10.1007/s12524-021-01365-y(0123456789().,-volV)
(0123456789().,-volV)

RESEARCH ARTICLE

High-resolution Alos Palsar for the Characterization of Water Storage

at the Fountaine Des Gazelles Dam in Biskra, Eastern Algeria
Abdelhalim Bendib1

Received: 9 November 2020 / Accepted: 1 April 2021 / Published online: 15 April 2021
Ó Indian Society of Remote Sensing 2021

Abstract
This study estimates water volume variations in the Fontaine des Gazelles Dam in the province of Biskra using high-,
medium-, and low-resolution digital elevation models (DEMs). The water surface was extracted using NDWI, and the
volume of storage was obtained using the water elevation and spatial analysis. The results obtained in comparison with
Peng equation show that there exists a significant change in the water surface. This represents a decrease of more than 29
hm3 between 2005 and 2020, on which precipitations were found to be the main factor of these changes. For validation
purposes, NRMSE and MAPE were used. Respectively, with values of 1.30% and 1.50% for Alos Palsar, 8.10% and 8.70%
for SRTM1, and 8.20% and 10.10% for SRTM3, it is clear that Alos Palsar performs much better results. This leads us to
conclude that Alos Palsar is the most reliable to be adopted in studies related to water volume variations.

Keywords Alos Palsar  Fontaine des Gazelles  Biskra  Precipitations  Water volume

Introduction operation across the country. They were built for providing
an annual volume of more than 20 billion m3 of water for
Algeria is classified as one of the poorest countries in the human consumption, industry, and irrigation (Benfetta &
world in terms of water resources because 90% of its land Ouadja, 2017). According to Bessenasse et al. (2012), the
is desert. More than 2 million sq km are characterized not total capacity of the Algerian dams is estimated to be 6.2
only by high temperatures that can exceed 45 °C but also billion m3, with a considerable quantity estimated at 40
by scarce rainfall. According to Sekkoum et al. (2012), the hm3, decreasing each year for many reasons (leakage,
annual average rainfall is estimated at 100 billion m3 sedimentation, and evaporation). However, the exploitation
which is very irregular (200 mm in the north and 12 mm in of water resources in arid and semiarid zones poses many
the south). issues. The exploitation of surface water using surface
Today, with 40 million inhabitants, the country requires dams is heavily exposed to evaporation in the dry season
a significant mobilization of water resources to accom- (Ouerdachi et al., 2012), the problem of leaks in earth dams
modate its socioeconomic development. For this purpose, (Ratiat et al., 2020), and more recently surface water pol-
the national water strategy focuses on making the most of lution (Guergazi & Achour, 2011). According to Stratfor
existing water resources through desalination for the Worldview (2016) report, Algeria relies heavily on
coastal cities, dams to irrigate the inland mountains and expensive water management solutions, and without gov-
high plateau, and water transfer projects to water-scarce ernment investment, the availability and quality of the
parts of the country (Drouiche et al., 2012). The People’s water supply will continue to decline.
Democratic Republic of Algeria PDRA (2011) reported Artificial surface water is an essential part of engineer-
that to address this situation, a total of 63 dams are in ing structures for water resource management and the main
source for providing water supplies (Guntner et al., 2004).
The majority of these structures are built to store water
& Abdelhalim Bendib
[email protected] during the humid periods, to be used during the dry peri-
ods. The structures also regulate river flows, reduce
1
Department of Geography and Spatial Planning, University catastrophic impacts, control floods, and help with
of Oran 2, Oran, Algeria

123
1928 Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938

irrigation, hydropower generation, and other water uses Description of Study Area
(Gao et al., 2012). Various studies that have examined the
management of dams continue to be of interest to This study concerns an artificial wetland in the munici-
researchers, in areas such as siltation phenomena (Alahiane pality of El-Outaya, which is 35 km northeast of the city of
et al., 2016), control of dam evaporation (Saggai and Bachi Biskra, and more than 450 km from the capital, Algiers.
2018), seepage phenomena (Amnyattalab & Rezaie, 2018), Situated between 35°70 7.200 ’ and 35°80 48.930 ’ North lati-
and seismic hazards (Zacchei et al., 2017). Several studies tude, and 5°340 4.360 ’ and 5°360 56.460 ’ East longitude,
have looked at water storage estimates, and different Fontaine des Gazelles Dam (Fig. 1) was built with a total
approaches have been proposed for a wide range of capacity of 55.49 hm3 (Athmani et al. 2018) to regulate
domains. These approaches include the following: analyt- flows and for the irrigation of more than 1100 ha of agri-
ical methods (AM) and the ensemble Kalman filter devel- cultural land and palm trees in the municipality of El-
oped by Deng et al. (2015), an adaptive network-based Outaya (ANAT, 2003). The dam was inaugurated in 2000
fuzzy inference system proposed by Unes et al. (2019), and went into full operation in 2006 after five years of the
unmanned aerial vehicles techniques (Ridolfi & Manciola, progressive filling (Tebbi, 2014). A quota of 4 hm3 was
2018), a mathematical programming model (Sun et al., designed for irrigation of agricultural land by emptying the
2011), and the ENVISAT Radar Altimeter (RA-2) and dam, and an annual allocation of 10 hm3 was also designed
Advanced Synthetic Aperture Radar (ASAR) images for irrigation by pipeline. This dam is the second in the
(Medina et al., 2010). Data are also obtained through a region, after Foum El Gherza Dam at 42 km, which is
combination of different satellite altimetry databases such intended to irrigate more than 300,000 date palms in the
as Global Reservoir and Lake Monitoring, River Lake oases and to irrigate the perimeters of Sidi Okba, Seriana,
Hydrology, Hydroweb, and ICESat-GLAS level 2 Global Thouda, and Garta, as well as to produce electricity for the
Land Surface Altimetry (Duan & Bastiaanssen, 2013). entire region of Biskra (Remini et al. 2018).
In an era in which significant technological progress is Situated between 347 and 2134 m altitudes, the dam is
occurring, remote sensing in combination with satellite fed by waters flowing from Oued El-Hai, through Oued
data has become a powerful tool for assessing climate Tamtam, and Oued El Melah, with an average annual
change (Yang et al., 2013), and knowledge of variations in supply of more than 20 hm3. The watershed is character-
water volumes in lakes and reservoirs is essential for ized by a surface area of more than 175,871 ha crossing
studying water balance and developing strategies to two climatic zones; a semiarid zone in the province of
maintain it (Duan & Bastiaanssen, 2013). Numerical data Batna with a total area of 143,131.86 ha and an arid zone in
of water volumes in the Fontaine des Gazelles Dam are the province of Biskra with an area of 32,739.71 ha. This
limited because of bureaucratic practices and security entire area is characterized by low slopes, 74.25% of them
reasons. In this paper, an alternative is to use remote do not exceed 12° while 25.75% of the watershed has
sensing techniques and geographic information systems slopes above 12°.
(GIS) in the diachronic estimation of water volumes in the According to the data obtained from WorldClim (http://
arid zone. Through the activities of many earth observation worldclim.org), the study area is characterized by a semi-
satellites, several data with different spatial resolutions and arid climate with large seasonal variations in temperature
coverage can be used to estimate the volumes of water. In and precipitation. Indeed, the annual average temperature
this study, we also look for precision by orienting our is about 20.05 °C and ranges from 3.75 °C in December to
choice toward a comparison of the following three 41.54 °C in July. Also, the area has an annual average
numerical elevation models: Alos Palsar (12.5 m) and two rainfall of 307.69 mm, ranging from 6.81 mm in July to
Shuttle Radar and Topography Mission products (SRTM3 36.82 mm in November.
90 m and SRTM1 30 m).
The study is organized into three main sections. The first
section gives a background to the study area. The second Materials and Methods
section provides highlights on the data collected and the
methodology used in extracting water body surface and The objective of this study is to find the water volume of
dam water storage through an empirical equation. The third the Fontaine des Gazelles Dam using high-, medium-, and
section compares and discusses the results obtained using low-resolution DEMs, as well as to analyze the effect of
three main parameters: the correlation coefficient (R2), the variations in annual precipitation and temperatures on the
normalized root-mean-square error (NRMSE), and the stored water level. In this respect, the present work is based
mean absolute percentage error (MAPE). on the use of numerous satellite images (Table 1) obtained
free of charge from United States Geological Survey

123
Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938 1929

Fig. 1 Study area, watershed limits (top), and north Algeria (bottom) Palsar

(USGS) Earth Explorer portal (https://earthexplorer.usgs. Several DEMs are freely available on the web (Aster
gov/), and (https://glovis.usgs.gov). GDEM, GMTED2010, SRTM products, Alos palsar, etc.).
This paper selects eight scene images of Landsat series In this study, we chose to compare very high-resolution
satellites (including four scenes of Landsat 5 TM for the Alos palsar (12.5 m) with two Shuttle Radar and Topog-
period 2005–2011 and four scenes of Landsat 8 LDCM for raphy Mission products (SRTM3 and SRTM1). This choice
the period 2013–2020) to extract the water body of the of SRTM products is based on the results of a study con-
dam. Furthermore, to get more representative results, the ducted by Rexer and Hirt (2014), in which a comparison
imaging time of the selected images is concentrated mostly between GDEM and SRTM DEM shows that SRTM
in June. According to Lai et al. (2019), at this time, the products are mostly superior and show the best vertical
temperature rises quickly, the reservoir has no ice, the accuracy. The high-resolution Alos Palsar DEM was
weather is fine, and the effect of clouds is less, so the image downloaded from the Alaska Satellite Facility (https://
quality is good. search.asf.alaska.edu) and the low-resolution USGS
SRTM3 (90 m) is obtained too from the official NASA site

123
1930 Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938

Table 1 Details of data used in

Satellite Bands Resolution Path/Row Date of acquisition
the study
Landsat 5 TM 4, 3, 2 30 194/36 7/5/2005
Landsat 5 TM 4, 3, 2 30 194/36 14/6/2007
Landsat 5 TM 4, 3, 2 30 194/36 6/6/2010
Landsat 5 TM 4, 3, 2 30 194/36 9/6/2011
Landsat 8 LDCM 5, 4, 3 30 194/36 14/6/2013
Landsat 8 LDCM 5, 4, 3 30 194/36 20/6/2015
Landsat 8 LDCM 5, 4, 3 30 194/36 28/6/2018
Landsat 8 LDCM 5, 4, 3 30 194/36 30/4/2020
USGS/NASA SRTM1 / 30 / 11/2/2000
USGS/NASA SRTM3 / 90 / 11/2/2000
Alos palsar data / 12.5 656 17/6/2008

USGS earth explorer. The medium-resolution USGS unsupervised classification of images, which highly
SRTM1 (30 m) is obtained from the GPS visualizer site depends on human knowledge of the study area (Li et al.,
(https://www.gpsvisualizer.com/elevation). These models 2016); Automated Water Extraction Index (Feyisa et al.,
were used to extract and compare the level of the water 2014), Modified Normalized Difference Water Index (Xu,
surface, which allowed us to extract the volume of the dam 2006), Revised Normalized Different Water Index (Cao
for different dates. et al., 2008), Index of Water Surfaces (Hassani et al.,
The various satellite data were processed in a multi- 2014), General Water Index (Yang & Xu, 1998), Enhanced
software platform. For this purpose, we used the following water Index EWI (Yang & Du, 2017), and AWEIsh-su-
software: ArcGIS10.3 for data management, spatial anal- perimposed MNDWI (Lai et al., 2019).
ysis, and cartography of the results, ENVI4.7 to extract Taking into account the role of the water surface in
water body, and Global mapper 13 is used in particular to determining the volume of the dam, it is important to
find the exact water elevation. The main steps are sum- choose the index capable of estimating the water surface
marized in Fig. 2. with great precision. In this paper, we used the Normalized
Difference Water Index (NDWI) to extract the water body
Water Body Extraction area. Based on many studies in the field (Acharya et al.,
2018; Peng et al., 2006), we conclude that the NDWI seems
In this study, water body extraction is a decisive step. most able to extract water area with accuracy.
Various quickly and easily methods and index have been Proposed by Mcfeeters (1996), the NDWI enhances
developed through the combination of multi-bands to avoid water spectral signals by contrasting the reflectance
the complications and errors of the supervised/ between different wavelengths and removing a large

Fig. 2 The main flowchart of

retrieving water storage of
Fontaine des Gazelles Dam

123
Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938 1931

portion of noise components in different wavelength empirical water storage estimation, and (3) the MAPE,
regions (Gao, 1996; Gao et al., 2016). which provides an estimate of the overall agreement
between the remote sensing and empirical water storage
Water Body Elevation Extraction estimations.

After extraction of the water surface area using the NDWI,

the problem that arises is to find the exact water elevation Results and Discussion
while tiny shifts between the satellite image Landsat and
booth DEMs (Alos Palsar and SRTM3) can give erroneous In the absence of numerical data, finding alternative means
results. To solve this problem, the polyline layers of the to estimate water storage in dams is becoming increasingly
water surface obtained by applying NDWI were superim- important. The main objective of this study was to estimate
posed on that of the numerical elevation model, and the the volume of water in the Fontaine des Gazelles Dam
average altitude of the water surface was obtained by (province of Biskra) using a method based on remote
extracting the altitudes for each vertex of the perimeter. For sensing and geographic information systems. The process
more accuracy, an important number of vertex (1249) were involved comparing a few DEMs and mapping the dia-
extracted for the selected years; 186 vertex in 2005, 184 chronic variations of the dam’s water surface.
vertex in 2007, 132 vertex in 2010, 169 vertex in 2011, 149 By superimposing the perimeter of the water surface
vertex in 2013, 157 vertex in 2015, 132 vertex in 2018, and with the set of DEMs used in this study, the water elevation
140 vertex in 2020. The obtained altitude allowed us to find of the dam was accurately determined. Moreover, a spatial
the water surface of the dam with high precision. analysis was used to obtain the water volumes for each
In the second step, the obtained water elevation was model for the period 2005–2020 (Table 2). Figure 3 sum-
used as an input to find the water storage. The process is marizes the changes in the dam’s water surface and water
done with the aid of the surface volume tool in ArcGIS volumes over 15 years through topographical profiles.
10.3 software. Analysis of Fig. 3 clarifies that the dam underwent
remarkable changes in its surface, and consequently, in its
Water Storage Estimation and Validation water levels. According to Liebe (2002), this result can be
explained in more than one way: (1) inter-annual rainfall
After extraction of the volume of water by a method based variability, which accounts for a relative maximum water
on remote sensing and spatial analysis (surface volume), it level, (2) time of year/season (intra-annual variations), (3)
is important to assess the results empirically by estimating water withdrawal (for irrigation, gardening, cattle rearing,
the volume of water stored for the selected periods. household uses, construction uses, etc.), (4) storage
Several formulas for several areas and types of water reduction through evaporation, seepage, and percolation,
surface have been developed (Gao et al., 2012; Guntner and (5) the age of the dam.
et al., 2004; Hayashi & Van Der Kamp, 2000; Minke et al., In this study, an empirical method was called to validate
2010; Mohammadzadeh-Habili et al., 2009). In this study, the findings. For this purpose, Peng et al. (2006) formula
we focus on the use of the equation proposed by Peng et al. was chosen. Moreover, before comparing and interpreting
(2006). By its simplicity and the availability of relative the results, it is important to check the compatibility of the
data for each term, this equation is a good choice. present formula. Based on the water elevations and surface
The equation is given as follows: areas, we derived the elevation–area relationship, which
 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi can be used to estimate dam storage. Linear regression was
Viþ1 ¼ Ai þ Aiþ1 þ Ai  Aiþ1 DZ=3Þ used to validate the relationship between the surface ele-
DZ ¼ ðZiþ1  Zi Þ vation and surface area. The estimated correlation of
99.65% (R2 = 0.9932) indicates a strong positive correla-
where Viþ1 is the water storage between two consecutive tion between the surface elevation and surface area
elevations. Ai and Aiþ1 are water surface area at the water (Fig. 4). Thus, the equation proposed was adequate for
elevation of Zi and Ziþ1 , respectively. estimating the volume of water. The volume of water
To evaluate the results and determine the optimal model, stored was obtained and is shown in Table 3.
the following three parameters were considered to quantify The formula was compared with the remote sensing
the accuracy of the water storage estimations: (1) the cor- findings. Based on Fig. 5 and Table 2, it can be concluded
relation coefficient (R2), which defines the weight of the that the results obtained using the Alos Palsar DEM
relationship between the remote sensing and empirical (12.5 m) and the empirical method of Peng et al. (2006)
water storage estimations. (2) The NRMSE, which calcu- demonstrated good agreement. Moreover, the visual com-
lates the residual value between the remote sensing and parison among these methods showed that the medium-

123
1932 Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938

Table 2 Water storage (WS)

Year 2005 2007 2010 2011 2013 2015 2018 2020
estimation by remote sensing
(values in hm3) Alos Palsar (12.5 m) WS 52.18 46.32 43.27 45.34 29.74 17.43 18.16 22.63
SRTM1 (30 m) WS 54.91 50.11 45.56 47.90 31.04 20.04 20.10 24.88
SRTM3 (90 m) WS 49.30 43.18 39.68 42.79 27.41 15.59 16.50 20.60

Fig. 3 Water body extraction and water levels for the period 2005–2020

Fig. 4 Relationship between surface elevation and surface area

Table 3 Water storage (WS)

Year 2005 2007 2010 2011 2013 2015 2018 2020
estimation using the empirical
method of Peng et al. (2006) Water elevation (m) 418.16 417.06 416.27 416.95 413.25 409.30 409.61 411.11
Water storage (hm3) 51.88 46.29 42.40 45.77 29.59 16.69 17.57 22.05

123
Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938 1933

Fig. 5 Time series of water storage estimations by remote sensing and empirical methods (values in hm3)

resolution (30 m) and low-resolution (90 m) SRTM ele- provided a more relative measure of the error, enabling a
vation model differs from the above models. These vari- better assessment of the model’s accuracy. Statistically, the
ances were significant and ongoing over the entire study model featuring Alos Palsar DEM (12.5 m) had a lower
period (2005–2020), raising questions about the suitability error (NRMSE = 1.3%). The highest NRMSE was
of using SRTM products to predict dam water levels. This obtained using a low-resolution SRTM3 model
difference can be explained only by the effects of DEM (NRMSE = 8.20%). The moderate values were obtained by
spatial resolution on the accuracy of the results. On the using the SRTM1 model (resolution of 30 m) with the
other hand, the same results reveal another important NRMSE being equal to 8.10%.
observation. We found that the finest DEM resolution does The MAPE indicates the extent of errors in percentage
not always provide the best accuracy. For instance, the (Ridolfi & Manciola, 2018). As can be seen in Table 4, the
high-resolution (12.5 m) DEM provided better results than extent errors between the three models were distinctive.
the low-resolution (90 m), at the same time, the 90 m With its high resolution, Alos Palsar was characterized by a
resolution DEM achieved better results than medium-res- low extent error (1.50%); this was in contrast to the med-
olution (30 m). This result is approved by many studies ium-resolution SRTM1 DEM, which had an extent error of
(Chen et al., 2020; Kariminejad et al., 2019). 8.70%. The low-resolution SRTM3 DEM had an extent
After comparing the values obtained from remote error of 8.10%, this is an important finding, prompting
sensing with the empirical values (Peng et al. 2006) a high questions about the validity of using this model in esti-
correlation emerged and the values were almost the same. mating dam levels.
More specifically, the results from the high-resolution As a result, the linear regressions for the three models
elevation model (Fig. 6a) yielded a correlation equal to were close to one another. They were also higher, which
99.984% (R2 = 0.9997), whereas the model with a mod- means the correlation coefficient did not provide satisfac-
erate resolution (Fig. 6b) was slightly more precise with a tory and decisive answers with which to judge the appli-
correlation equal to 99.989% (R2 = 0.9998). The low-res- cability of the present approach. Further, the results
olution SRTM3 DEM (Fig. 6c) revealed a correlation of revealed that NRMSE and MAPE facilitated comparison
99.849% (R2 = 0.997). Based on Fig. 7, the NRMSE between models, and it was clear that high-resolution DEM

Fig. 6 Remote sensing and Peng et al. 2006 empirical water storage estimation relationship; a Alos Palsar DEM (12.5 m), b SRTM1 DEM
(30 m), and c SRTM3 DEM (90 m), values in hm3

123
1934 Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938

Fig. 7 Remote sensing and Peng et al. 2006 empirical water storage estimation error; a Alos Palsar DEM (12.5 m), b SRTM1 DEM (30 m), and
c SRTM3 DEM (90 m), values in hm3

Table 4 Correlation coefficient

R2 (%) NRMSE (%) MAPE (%)
(R2), NRMSE, and MAPE for
remotely sensing empirical High-resolution Alos Palsar DEM (12.5 m) 99.984 1.30 1.50
water storage estimations
Medium-resolution SRTM1 DEM (30 m) 99.989 8.10 8.70
Low-resolution SRTM3 DEM (90 m) 99.849 8.20 8.10

yielded the best results. This study concludes that Alos reference year, the dam reached its estimated storage of
Palsar, in this study, is the only valid and reliable model for 52.18 hm3 (418.16 m) after a few years of the progressive
studies related to dam storage levels, especially for those filling (2001–2004). This was followed by five years of
investigating arid areas. remarkable decreases in the volume, estimated at 46.32
The volume of water stored was determined at the end hm3 (417.06 m) and 43.27 hm3 (416.27 m) in 2007 and
of this study using three numerical elevation models (low, 2010, respectively. In 2011, the dam recorded an increase
moderate, and high resolution) from 2005 to 2020. Peng in volume at 45.34 hm3 (413.25 m), this was followed by a
et al. 2006 formula and three validation indices (R2, second wave of stock decline at 29.74 hm3 (413.25 m) in
NRMSE, and MAPE) were also used. The study concluded 2013 and 17.43 hm3 (409.31 m) in 2015.
that the precise volume of the dam could be determined The last two years saw an increase in the water volume
using the high-resolution Alos Palsar model (12.5 m). at a low rate, estimated at 18.16 hm3 (409.61 m) in 2018
Based on Table 5, the results clearly showed that the water and 22.63 hm3 (411.11 m) in 2020.
surface in the dam decreased significantly, from 5.601 In general, from 15 years of data and the models used to
million m2 in 2005 to 5.100 million m2 in 2007 and 4.755 achieve the reported statistics, it can be concluded that
million m2 in 2010. Also, a small increase was noted in there has been a significant change in the water volume of
2011, for a total surface of 4.986 million m2. A consider- the dam. It represents a decrease of more than 29 hm3 (a
able decrease in the dam’s water surface was recorded in decrease of more than 2.5 million m2). According to Chai
2013 and 2015, at 3.667 million m2 and 2.598 million m2, et al. (2018), water level variations are affected by many
respectively. The last two years showed a significant factors, including climate variability, reservoir operation,
increase from 2.672 million m2 in 2018 to 3.069 million m2 backwater effect, irrigation and living water, and human
in 2020. Indeed, this change in the water surface was sand excavation. In this study, it is not easy to assess the
logically followed by changes in the water elevation, and contribution of each factor to water level variations. For
as a result, the storage capacity of the dam. In 2005, the this purpose, this study focused on the effects of climatic

Table 5 Fontaine des Gazelles Dam water storage as obtained by Alos Palsar DEM (12.5 m)
2005 2007 2010 2011 2013 2015 2018 2020

Surface elevation (m) 418.16 417.06 416.27 416.95 413.25 409.31 409.61 411.11
Surface area (million m2) 5.601 5.100 4.755 4.986 3.667 2.598 2.672 3.069
Water storage (hm3) 52.18 46.32 43.27 45.34 29.74 17.43 18.16 22.63

123
Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938 1935

elements (especially precipitation and temperatures). The approximately 28% (90.16 hm3), a considerable part, was
study downloaded the climatic data free of charge from evaporated. This result is approved by a study conducted
(https://globalweather.tamu.edu/). The site contains records by Ouerdachi et al. (2012) where an evaporation loss of
made by the Fontaine des Gazelles monitoring station for 9106 m3 was measured in 2012.
the period 1979–2014. On the other hand, the watershed of the dam undergoes
According to Fig. 8, with a correlation coefficient of specific degradation exceeding 800 t/km2/year with high
more than 95% (R2 = 0.914), there was a strong positive suspended material (turbidity exceeding 160 g/l), which
correlation between precipitation and the volume of water amounts to a loss of enormous capacity of the dam
being stored. Indeed, comparing the annual changes in (Ouerdachi et al., 2012). As mentioned previously, 74.25%
precipitation with the volume of water stored, it seems that of areas do not exceed 12°. This dominance of low slopes
precipitation plays the most important role in water level was reflected by a siltation rate estimated at 0.25 hm3/year
changes, especially when it is an important watershed with in 2017. Indeed, this phenomenon of sedimentation is
an area of more than 175,871 ha crossing two climatic observed, also, in the Dam of Foum El Gherza (at 42 km)
zones. A semiarid zone in the province of Batna (total area and which shares the same area and the same conditions
of 143,131.86 ha), and an arid zone in the province of (climate, soil, etc.). Over 64 years (1950–2014), the dam
Biskra (an area of more than 32,739 ha). experienced an estimated decrease in volume of 32 hm3
Conversely, temperatures exhibited a limited role in the (76%). A study conducted by Ratiat et al. (2020) shows
annual variations of the storage level with a correlation that the dam is subject to leaks in its right bank. The
equal to 68.78% (R2 = 0.4724). However, compared to problem is due to underground seepage caused by a fault
other factors (irrigation, emptying, and discharge), the below the dam foundation. With a flow of leaks corre-
temperature has a significant effect on the variations in sponds to 2.5 l/s to the level of 364 m springs on the toe of
water storage after precipitation. According to Tebbi the dam and in the middle of the spillway flow channel,
(2014) during 2002–2011, the dam received a total con- appear.
tribution of 322 hm3, of which 39% was discharged, 12% In summary, although the Alos Palsar model has its
was drained, 21% was destined for irrigation, and advantages, it has a major disadvantage, as well as the

Fig. 8 Effect of climate parameters on the change in the volume of water stored in the Fontaine des Gazelles Dam, precipitation (high), and
temperatures (low)

123
1936 Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938

weak coverage of Algeria. In 2020, barely 7.5% of the References

national territory has been covered, which makes the study
of other more important dams, such as the Beni Haroune Acharya, T. D., Subedi, A., & Lee, D. H. (2018). Evaluation of water
Dam (the largest in Algeria) still impossible. Another indices for surface water extraction in a landsat 8 scene of nepal.
Sensors. https://doi.org/10.3390/s18082580.
disadvantage is that the Alos Polsar DEM provides a Alahiane, N., Elmouden, A., Aitlhaj, A., & Boutaleb, S. (2016). Small
slightly higher altitude value than the SRTM DEM which dam reservoir siltation in the atlas mountains of central morocco:
requires special attention when using this model. Analysis of factors impacting sediment yield. Environment and
However, the Alos Palsar DEM has proven its abilities Earth Science. https://doi.org/10.1007/s12665-016-5795-y.
Amnyattalab, J., & Rezaie, H. (2018). Study of the effect of seepage
in several fields, such as extracting drainage networks through the body of earth dam on its stability by predicting the
(Niipele & Chen, 2019), and floodplain monitoring (Ar- affecting hydraulic factors using models of Brooks-Corey and
nesen et al., 2013), the results obtained in this study show van Genuchten (Case study of Nazluchay and Shahrchay earth
that Alos Palsar is also capable of retrieving, with accu- dams). . International Journal of Environmental science and
Technology. https://doi.org/10.1007/s13762-017-1549-y.
racy, water storage in an arid area. Arnesen, A. S., Silva, T. S. F., Hess, L. L., Novo, E. M. L. M.,
Rudorff, C. M., Chapman, B. D., & McDonald, K. C. (2013).
Monitoring flood extent in the lower Amazon River floodplain
Conclusion using ALOS/PALSAR ScanSAR images. Remote Sensing of
Environment, 130(3), 51–61.
Athmani, H., Boulahdid, M., Bouchahm, N., & Hocini, N. (2018).
This study aims to estimate the water level of an Algerian Géochimie des métaux traces dans les sédiments du barrage de la
dam whose records are often absent or difficult to obtain. In fontaine des gazelles, Algérie [Geochemistry of trace metals in
this case, Fontaine des Gazelles Dam, located in an arid area, sediment of gazelles fountain dam, Algeria]. Currier du savoir,
26, 481–494.
was chosen. The dam was commissioned in 2005 which made Benfetta, H., & Ouadja, A. (2017). The surface water storage problem
it suitable for this study especially to determine the mor- in arid regions: a case study of the Gargar dam, Algeria. Journal
phology of the dam. This allowed us to estimate the volume of Fundamental and Applied Sciences. http://www.jfas.info
of water with great precision. In this respect, eight satellite Bessenasse, M., Paquier, A., & Moulla, A. S. (2012). A contribution
to the numerical modelling of dam reservoir siltation cycle.
images were chosen to extract the water surface for the period International Water Technology Journal, 2(3), 236–249.
2005–2020. Then, three numerical models of different reso- Cao, R. L., Li, C. J., Liu, L. Y., Wang, J. H., & Yan, G. G. (2008).
lutions were also adopted. These are Alos Palsar (12.5 m) Extracting Miyun Reservoir’s water area and monitoring its
very high resolution, SRTM1 (30 m) medium resolution, and change based on a revised normalized different water index.
Science of Surveying and Mapping, 33(2), 158–160.
SRTM3 (90 m) low resolution. A geospatial information Chai, Y., Li, Y., Yang, Y., Li, S., Zhang, W., Ren, J., & Xiong, J.
process was followed to extract the volume of water. First, the (2018). Water level variation characteristics under the impacts
extraction of the water surface using the famous NDWI of extreme drought and the operation of the Three Gorges Dam. .
allowed us to extract the diachronic variations of the dam. Earth Sci. https://doi.org/10.1007/s11707-018-0739-3.
Chen, Z., Ye, F., Fu, W., Ke, Y., & Hong, H. (2020). The influence of
Then, using the three DEMs, the volume of water was DEM spatial resolution on landslide susceptibility mapping in
determined with great precision. This volume was validated the Baxie River basin. . Natural Hazards. https://doi.org/10.1007/
using a mathematical equation of which Peng et al. 2006 s11069-020-03899-9.
appear to be more efficient. Deng, C., Liu, P., Guo, S., Wang, H., & Wang, D. (2015). Estimation
of nonfluctuating reservoir inflow from water level observations
For validation purposes, three parameters were applied using methods based on flow continuity. Journal of Hydrology,
R2, NRMSE, and MAPE. Respectively, with values of 529(10), 1198–1210.
99.97%, 1.30%, and 1.50% for Alos Palsar, 99.98%, Drouiche, N., Ghaffour, N., Naceur, M. W., Lounici, H., & Drouiche,
8.10%, and 8.70% for SRTM1, and 99.70%, 8.20%, and M. (2012). Towards sustainable water management in Algeria.
Desalination and Water Treatment, 50(1–3), 272–284.
10.10% for SRTM3, it is clear that Alos Palsar performs Duan, Z., & Bastiaanssen, W. G. M. (2013). Estimating water volume
much better results. This leads us to conclude that Alos variations in lakes and reservoirs from four operational satellite
Palsar is the most reliable model to be adopted in studies altimetry databases and satellite imagery data. Remote Sensing of
related to water volume variations, especially in arid areas. Environment. https://doi.org/10.1016/j.rse.2013.03.010.
Agence Natianale d’Aménagement du Territoire ANAT. (2003).
Also, this study found that the finest DEM resolution does Etude schéma directeur des ressources en eau, wilaya de Biskra
not always provide the best accuracy. For instance, the [Water resources master plan study, wilaya of Biskra]. Rapport.
high-resolution DEM provided better results than the low- 60p.
resolution, at the same time, the 90 m resolution DEM Feyisa, G. L., Meilby, H., Fensholt, R., & Proud, S. R. (2014).
Automated water extraction index: A new technique for surface
achieved better results than medium-resolution. water mapping using landsat imagery. Remote Sensing of
Environment. https://doi.org/10.1016/j.rse.2013.08.029.
Gao, B. C. (1996). NDWI, a normalized difference water Index for
Funding The present study was not funded by any company or remote sensing of vegetation liquid water from space. Remote
person.

123
Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938 1937

Sensing of Environment. https://doi.org/10.1016/S0034- Ouerdachi, L., Boutaghane, H., Hafsi, R., Boulmaiz, T. T., &
4257(96)00067-3. Bouzahar, F. (2012). Modeling of underground dams Applica-
Gao, H., Birkett, C., & Lettenmaier, D. P. (2012). Global monitoring tion to planning in the semi-arid areas (Biskra, Algeria). Energy
of large reservoir storage from satellite remote sensing. Water Procedia, 18, 426–437.
resources research. https://doi.org/10.1029/2012WR012063. People’s Democratic Republic of Algeria PDRA. (2011). The algerian
Gao, H., Wang, L., Jing, L., & Xu, J. (2016). An effective modified report- Algeria 2011. Oxford Business Group edition. 66p.
water extraction method for Landsat-8 OLI imagery of moun- Peng, D., Guo, S., Liu, P., & Liu, T. (2006). Reservoir Storage Curve
tainous plateau regions. 9th Symposium of the International Estimation Based on Remote Sensing Data. Journal of Hydro-
Society for Digital Earth (ISDE), IOP Conf. Series: Earth and logic Engineering, 11(2), 165–172.
Environmental Science. Doi:https://doi.org/10.1088/1755-1315/ Ratiat, A., Khettal, T., & Meddi, M. (2020). The piezometric and
34/1/012010 isotopic analysis of leaks in earth dams: the case of the fountain
Guergazi, S., & Achour, S. (2011). Effet de la désinfection par le of Gazelle dam. . Environmental Earth Sciences. https://doi.org/
chlore sur la pollution des eaux de surface de la région de Biskra 10.1007/s12665-020-8886-8.
[Effect of chlorine disinfection on the surface water pollution of Remini, B., Merzoug, H., & Rais, M. A. (2018). Le barrage de foum
the Biskra region]. Larhyss Journal, 9, 131–138. el gherza (Algerie): quand l’eau coule dans le massif karstique
Guntner, A., Krol, M. S., De Araujo, J. C., & Bronsret, A. (2004). [The foum el gherza dam (Algeria): when water flows in the
Simple water balance modeling of surface reservoir system in a karstic massif]. Larhyss Journal, 36, 179–198.
large data-scarce semiarid region. Hydrological sciences journal, Rexer, M., & Hirt, C. (2014). Comparison of free high resolution
49(5), 901–918. digital elevation data sets (ASTER GDEM2, SRTM v2.1/ v4.1)
Hassani, M., Chabou, M. C., Hamoudi, M., & Guettouche, M. S. and validation against accurate heights from the Australian
(2014). Index of extraction of water surfaces from Landsat 7 National Gravity Database. Australian Journal of Earth
ETM? images. Arabian Journal of Geosciences , 8, 3381–3389. Sciences: An International Geoscience Journal of the Geological
Hayashi, M., & Van Der Kamp, G. (2000). Simple equations to Society of Australia, 61(2), 213–226.
represent the volume–area–depth relations of shallow wetlands Ridolfi, E., & Manciola, P. (2018). Water level measurements from
in small topographic depressions. Journal of Hydrology, drones: A pilot case study at a dam site. Water, 10(3), 1–10.
237(1–2), 74–85. Saggaı̈, S., & Bachi, O. E. K. (2018). Evaporation reduction from
Kariminejad, N., Hosseinalizadeh, M., Pourghasemi, H. R., Ownegh, water reservoirs in arid lands using monolayers: Algerian
M., Rossi, M., & Tiefenbacher, J. P. (2019). Optimizing experience. Water Resources. https://doi.org/10.1134/
collapsed pipes mapping: Effects of DEM spatial resolution. S009780781802015X.
CATENA. https://doi.org/10.1016/j.catena.2019.104344. Sekkoum, K., Talhi, M.F., Cheriti, A., Bourmita, Y., Belboukhari, N.,
Lai, Y., Zhang, J., Song, Y., & Cao, Y. (2019). Comparative analysis Boulenouar, N., & Taleb, S. (2012). Water in Algerian Sahara:
of different methods for extracting water body area of miyun Environmental and Health impact. In: R.Y. Ning (ed.), Intech
reservoir and driving forces for nearly 40 years. Journal of the Open. Advancing Desalination. (pp. 197–216).
Indian Society of Remote Sensing, 48(2), 451–463. Stratfor Worldview. (2016). Algeria’s Expensive Water Problem.
Li, W., Qin, Y., Sun, Y., Huang, H., Ling, F., Tian, L., & Ding, Y. https://worldview.stratfor.com/article/algerias-expensive-water-
(2016). Estimating the relationship between dam water level and problem. Accessed 26 Jan 2021.
surface water area for the danjiangkou reservoir using landsat Sun, S., Yan, X., Cui, P., & Feng, J. (2011). A four-step method for
remote sensing images. Remote Sensing Letters, 7(2), 121–130. optimising the normal water level of reservoirs based on a
Liebe., J. (2002). Estimation of water storage capacity and evapora- mathematical programming model- a case study for the
tion losses of small reservoirs in the upper east region of Ghana. Songyuan backwater dam in Jilin Province, China. International
Diploma thesis. Geographische institute der rheinischen frie- Journal of Environmental Research and Public Health, 8(4),
drich-wilhelms-universität bonn, 19p. https://www.zef.de/filead 1049–1060.
min/template/Glowa/Downloads/thesis_liebe.pdf. Accessed 22 Tebbi, F.Z. (2014). Modélisation de la régularisation des barrages
August 2020. dans la région des Aurès, Doctoral thesis, University of Batna.
McFeeters, S. K. (1996). The use of the Normalized Difference Water http://eprints.univ-batna2.dz/759/. Accessed 13 Sept 2020.
Index (NDWI) in the delineation of open water features. Üneş, F., Demirci, M., Taşar, B., Kaya, Y. Z., & Varçin, H. (2019).
International Journal of Remote Sensing. https://doi.org/10. Estimating dam reservoir level fluctuations using data-driven
1080/01431169608948714. techniques. Polish Journal of Environmental Studies, 28(5),
Medina, C., Gomez-Enri, J., Alonso, J. J., & Villares, P. (2010). 3451–3462.
Water volume variations in Lake Izabal (Guatemala) from in situ Xu, H. Q. (2006). Modification of normalized difference water index
measurements and ENVISAT Radar Altimeter (RA-2) and (NDWI) to enhance open water features in remotely sensed
Advanced Synthetic Aperture Radar (ASAR) data products. imagery. International Journal of Remote Sensing. https://doi.
Journal of Hydrology. https://doi.org/10.1016/j.jhydrol.2009.12. org/10.1080/01431160600589179.
016. Yang, C., & Xu, M. (1998). Study on the water-body extraction
Minke, A. G., Westbrook, C. J., & Van der Kamp, G. (2010). methods of remote sensing information mechanism. Geographic
Simplified volume-area-depth method for estimating water Research, 17, 86–89.
storage of prairie potholes. Wetlands, 30, 541–551. https://doi. Yang, J., & Du, X. (2017). An enhanced water index in extracting
org/10.1007/s13157-010-0044-8. water bodies from Landsat TM imagery. Annals of GIS, 23(3),
Mohammadzadeh-Habili, J., Heidarpour, M., Mousavi, S. F., & 141–148.
Haghiabi, A. H. (2009). Derivation of reservoir’s area-capacity Yang, J., Gong, P., Fu, R., Zhang, M., Chen, J., Liang, S., Xu, B., Shi,
equations. Journal of Hydrologic Engineering, 14(9), J., & Dickinson, R. (2013). The role of satellite remote sensing in
1017–1023. climate change studies. Nature Climate Change. https://doi.org/
Niipele, J.N., & Chen, J. (2019). The usefulness of Alos-Palsar Dem 10.1038/nclimate1908.
data for drainage extraction in semi-arid environments in The Zacchei, E., Molina, J. L., & Reyolando, M. L. R. F. (2017). Seismic
Iishana sub-basin. Journal of Hydrology: Regional Studies, hazard and structural analysis of the concrete arch dam (Rules
https://doi.org/10.1016/j.ejrh.2018.11.003.

123
1938 Journal of the Indian Society of Remote Sensing (August 2021) 49(8):1927–1938

Dam on Guadalfeo River). Procedia Engineering. https://doi. Publisher’s Note Springer Nature remains neutral with regard to
org/10.1016/j.proeng.2017.09.334. jurisdictional claims in published maps and institutional affiliations.

123