HESSHydrology and Earth System SciencesHESSHydrol. Earth Syst. Sci.1607-7938Copernicus PublicationsGöttingen, Germany10.5194/hess-20-1355-2016Recent changes in climate, hydrology and sediment load in the Wadi Abd, Algeria (1970–2010)AchiteMohammedOuillonSylvainsylvain.ouillon@ird.frhttps://orcid.org/0000-0001-7964-7787Laboratoire Eau-Environnement, Université Hassiba Ben Bouali, BP 151, Hay Es-Salem, Chlef, AlgeriaLEGOS, Université de Toulouse, IRD, CNRS, CNES, UPS, 14 avenue E. Belin, 31400 Toulouse, FranceUniversity of Science and Technology of Hanoi, Dept. Water-Environment-Oceanography, 18 Hoang Quoc Viet, Cau Giay, Hanoi, VietnamSylvain Ouillon (sylvain.ouillon@ird.fr)6April20162041355137220September201514October20157March201619March2016This work is licensed under a Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/This article is available from https://hess.copernicus.org/articles/20/1355/2016/hess-20-1355-2016.htmlThe full text article is available as a PDF file from https://hess.copernicus.org/articles/20/1355/2016/hess-20-1355-2016.pdf
Here we investigate the changes of temperature, precipitation, river runoff
and sediment transport in the Wadi Abd in northwest Algeria over a time series of
40 hydrological years (1970–2010). Temperature increased and precipitation
decreased with the reduction in rainfall being relatively higher during the
rainy season. A shift towards an earlier onset of first rains during summer
was also found with cascading effects on hydrology (hydrological regimes,
vegetation, etc.) and thus on erosion and sediment yield. During the 1980s,
the flow regime shifted from perennial to intermittent with an amplification
of the variations of discharge and a modification of the sediment regime
with higher and more irregular suspended particulate flux. Sediment flux was
shown to almost double every decade from the 1970s to the 2000s. The sediment regime
shifted from two equivalent seasons of sediment yield (spring and fall) to
a single major season regime. In the 2000s, autumn produced over 4 times more
sediment than spring. The enhanced scatter of the C–Q pairs denotes an
increase of hysteresis phenomena in the Wadi Abd that is probably related to
the change in the hydrologic regime. At the end of the period, due to
irregularity of the discharge, the ability of a rating curve to derive
suspended sediment concentration from river discharge was poor.
Introduction
Fluvial and estuarine suspended sediment fluxes are changing dramatically
under the combined effects of anthropogenic activities and climate change.
On a global scale, recent changes showed a trend towards increasing land
erosion and decreasing fluxes to coastal waters (Walling and Fang, 2003;
Vörösmarty et al., 2003; Wang et al., 2006). The sediment flux
trapped in regulated basins with reservoirs is higher than 50 %
(Vörösmarty et al., 2003). Locally, it can reach more than 60 %
after the impoundment of one single dam, like on the Red River (Vinh et al.,
2014), and more than 80 % on rivers with many dams (86 % on the Yellow
River, Wang et al., 2007; > 95 % on the Ebro River, Durand et
al., 2002). Other engineering activities (meander cutoffs, river-training
structures, bank revetments, soil erosion controls) also significantly
affect sediment fluxes and can participate in the shift from a
transport-limited system to a supply-limited system, like on the
Missouri–Mississipi River system (Meade and Moody, 2010).
Climate change, through increasing temperatures and evaporation, tends to
accelerate the water cycle and modify hydrologic regimes (Bates et al.,
2008). Precipitation intensities and the frequency of extreme events are
projected to increase under climate change, leading to more frequent flood
events of higher magnitude that will, in turn, affect patterns of erosion
and deposition within river basins (Tucker and Slingerland, 1997; Pruski and
Nearing, 2002; Tockner and Stanford, 2002; Coulthard et al., 2012). Recent
studies focused on the impact of climate change on sediment transport
(e.g. Gomez et al., 2009; Hancock, 2009; Walling, 2009; Hancock and Coulthard,
2011; Knight and Harrison, 2013; Lu et al., 2013). Syvitski (2003) showed on
an example that sediment transport may increase due to the increasing
discharge or decrease because of the enhanced temperature. Studies have
compared trends in hydrological and sediment time series to land use changes
(Wang et al., 2007; Memariam et al., 2012; Gao et al., 2012). Climate
projections are consistent regarding warming and the acceleration of the
water cycle (IPCC, 2013), however, they remain to be defined on sediment
transport where projections show a high uncertainty (Shrestha et al., 2013;
Lu et al., 2013). This is in part due to the fact that climate affects many
factors controlling sediment yield, such as surface moisture availability,
weathering processes and rates, and the nature of riparian vegetation
(Nanson et al., 2002).
While sediment transport is well documented in perennial rivers in humid or
temperate climates, its study in semi-arid areas is still fragmentary due to
the difficulty of sampling during flashfloods. Amongst the factors favouring
erosion (slope, nature of rocks, relief, climate, human activities), climate
is recognized to be the main factor in semi-arid Mediterranean areas of
Algeria which experience short and intense rain episodes, high evaporating
power of wind, prolonged droughts and freezing and thawing cycles (Touaibia,
2010; Houyou et al., 2014). Erosion is extremely active and the average
concentration is at least 1 order of magnitude higher than the global
average (Achite and Ouillon, 2007). One of the main impacts of this high
erosion is the rapid silting up of reservoirs (up to 2–5 % per year,
Kassoul et al., 1997; Remini et al., 2009; Touaibia, 2010) with important
consequences on water resource management in a region where 85 % of rain
evaporates (Benhamiche et al., 2014). The high temporal variability and
recent changes in forcings mean that it is necessary to study sediment
dynamics in such environments over time periods of several decades in order
to document and understand the changes in sediment regime.
Location of the Wadi Abd sub-basin within the Mina and Cheliff
basins, and the other main basins of Algeria.
The Wadi Abd catchment area. (a) displays rain and hydrometric stations
including HS1 at Takhmaret and HS2 at Ain Hamara, (b) geology, (c) slopes
from the digital elevation model of north Algeria, (d) vegetation cover from
Landsat ETM+ data of 2009.
Linear erosion forms in the Wadi Abd basin. (a) and
(e) display gullying (depth: 30–50 cm, width > 1 m), (c) and (d) gully erosion
(depth: 50–200 cm), (b) and (f) interill and rill erosion.
Achite and Ouillon (2007 hereafter referred as AO2007) analysed sediment
transport changes in the Wadi Abd, an Algerian wadi over a 22-year period
(1973–1995). Here we extend this analysis of sediment transport changes to
cover a 40-year period (1970–2010). The hydrologic gauging station is
located upstream from a dam and is not affected by any major management.
This river sub-basin is also particularly suitable for such study because
its hydrologic regime was shown to have drastically changed between the
1970s and the 1980s. Precipitation decreased and became more irregular and
the flow regime shifted from perennial to intermittent with 26 % dry days,
on average, in 1990–1995. Variations of discharge were amplified, and a
modified sediment regime occurred with a higher and more irregular suspended
particulate flux, that was 4.7 times higher over the period 1985–1995 than
over 1973–1985. AO2007, showing the advantage of working with over 22 years of
measurement, however, stressed the difficulty of defining a reference
period, and the need to extend the study over a longer period of time. The
objectives of this additional study are to (1) describe precipitation,
discharge and sediment flux variability of the Wadi Abd basin over a 40-year
period; (2) detect the shift, if any, in temperature, runoff and sediment
yield; (3) determine the relationship between sediment load and runoff over
the last 40 years; (4) detect when a shift occurred in the runoff–sediment
load relationship; (5) analyse the possible causes of the change in flow
regime and its consequences on suspended sediment discharge; and (6) assess the
use of rating curves and the physical signification of their parameters when a
river is experiencing a transition and shifts from a perennial regime to an
intermittent regime.
Study area: the Wadi AbdGeneral information
The Wadi Abd, located in northwest Algeria, is a tributary of the Wadi
Cheliff, the principal river of Algeria (Fig. 1). The length of the Wadi
Abd's main stream is 118 km, the basin area is 2480 km2 and the
drainage density is 3.70 km km-2 (Fig. 2a). The Wadi Abd supplies the
downstream Sidi Mohamed Benaouda (SMB) reservoir which has a basin area of
4900 km2. The Wadi Abd catchment area is formed of erodible sedimentary
rocks from the Upper Jurassic (45.9 % of its surface), Middle Jurassic
(20.2 %) and Pliocene (7.4 %) (Fig. 2b). Soft bottom sedimentary
deposits from the Quaternary cover 13 % of the basin along the wadi
(Tecsult International, 2004).
The climate is Mediterranean and is characterized by a dry season from April
to August/September, and a wet season from September to March. The hydraulic
deficit is very high. Annual precipitation is 264 mm, on average, while the
mean potential evapotranspiration over the SMB basin is 1525 mm (Tecsult
International, 2004).
The watershed mainly consists of steep slopes (Fig. 2c) with very sparse
vegetation or bare soil (Fig. 2d). The main land use is natural environment
(73 %; 17 % of forests +56 % of scrub and bare steppe soils),
cultivated lands cover about 26 % and cities 0.4 %. Seven hill
reservoirs were built in the Wadi Abd basin from 1986 to 2004 for
agriculture (irrigation and livestock watering) or for fire-fighting
measures. Their total cumulated capacity is 0.88 hm3, representing
2.3 % of the yearly averaged discharge at Ain Hamara station. These small
reservoirs are now silted up to 70 % of their volume.
In 2008, 123 000 inhabitants were living in the Wadi Abd basin (average
density: 49 inhabitants km-2), 44 % of them living in the city of
Takhmaret. The Wadi Abd is thus little influenced by human activities, in
view of its extensive surface that is subject to severe natural erosion.
In the plain, sheet (interrill) and rill erosion dominates (Fig. 3b and f).
Gully erosion is mainly restricted to the mountainous regions of Frenda and
Tiaret in the north (Figs. 3c, d and 2c), while some mid-slope areas
are gullying (Fig. 3a and e).
Data
Long-term series of temperature measured at three stations in Algeria were
extracted from CRUTEM4 (Jones et al., 2012; Osborn and Jones, 2014). These
stations are located at Chlef (36.20∘ N, 1.30∘ E – 1951–2011),
Miliana (36.30∘ N, 2.20∘ E – 1922–2011) and
Dar El Beida (36.70∘ N, 3.30∘ E – 1856–2011). Annual
average temperatures were calculated for each station from the 12 monthly
averages. The 20 missing monthly data (out of 480) at Chlef, the nearest
station from the Wadi Abd, were extrapolated from the monthly temperatures
measured at Miliana and Dar El Beida using the relationships between the
monthly average temperatures at Chlef and Miliana, and Chlef and Dar El
Beida (Fig. 4). The resulting estimates of temperature at Chlef on seasonal
and yearly scales allowed us to estimate changes by decade over the period 1970–2010.
Relationships between mean annual temperatures at the three
stations of Dar El Beida, Miliana and Chlef (from CRUTEM4).
Rainfall and hydrometric records were provided by the National Agency of
Hydraulic Resources (ANRH). Time series of rainfall data are available at
six stations within the basin (see Fig. 2a): S1 Ain Kermes (altitude: 1162 m),
S2 Rosfa (960 m), S3 Sidi Youcef (1100 m), S4 Tiricine (1070 m),
S5 Takhmaret (655 m) and S6 Ain Hamara (288 m). There were 9076 coincident instantaneous
measurements of water discharge (namely Q, in m3 s-1) and
suspended sediment concentrations (C, in g L-1) recorded at the
Ain Hamara gauging station between September 1970 and August 2010. Water
depths were measured continuously and a calibration between water level and
discharge was regularly performed from velocity profiles. During flow
measurements, water was manually sampled once or twice using a 1 L dip at
the edge of the wadi. The number of samples was adapted to the flow regime.
During baseflow samples were collected every other day, whereas during
floods samples were collected at higher rates (up to one every 30 min).
Water samples were filtered on preweighed Whatman glass fiber filters (GFF),
oven dried at 105 ∘C for 24 h, and weighed again to
determine the concentration. This method, used by ANRH at all hydrologic
stations in Algeria, underestimates the suspended load as compared to its
value averaged over the cross section under low turbulence (i.e. at low
flow) since water is sampling near the surface (Touat, 1989). During floods,
which transport most of the sediment load, turbulence is high enough to
homogenize suspension load. While this underestimation may slightly affect
the budget, it does not severely affect the time variability of suspended
matter which is analysed in this paper. From 9076 coincident instantaneous
data measured during 1213 days, average arithmetic values were calculated
per day so as to obtain 1213 pairs of “mean daily” (C, Q) values. The
resulting “mean daily Q” differs from the (true) daily discharge obtained
from the averaging of 24 h of continuous instant Q.
The Atlantic Multidecadal Oscillation (AMO) index is an index of North
Atlantic temperatures. The monthly unsmoothed values used in this study were
calculated by NOAA, Earth System Research Laboratory, Physical Sciences
Division/ESRL/PSD1 (http://www.esrl.noaa.gov/psd/data/time series/AMO/).
Models and methodsTrends
The analysis of trends was conducted following a method fully described by
Stahl et al. (2010) and Déry et al. (2005) for river runoff. The
Kendall–Theil robust line furnishes a linear equation from a time series of
n measurements such as
y=mt+b
where t is time (year), y denotes the hydrological parameter (precipitation,
river discharge and sediment discharge) and m is the magnitude of the trend
over this period. m is calculated as the median of all slopes mk of
consecutive pairs of values:
mk=yj-yitj-ti,
where k= [1, n(n- 1)/2], i= [1, n- 1], j= [2, n].
This slope is often referred to as the Sen slope (Sen, 1968). The significance
of this trend at a level p was calculated following Ziegler et al. (2003).
Rating curves
C and Q measurements were used to define rating curves that estimate C from
measured values of Q, according to a common approach (e.g. Walling, 1977a;
Asselman, 2000; El Mahi et al., 2012; Tebbi et al., 2012; Louamri et al.,
2013). The most suitable model is a power law of the type C=aQb for
which the coefficients (a, b) determined empirically account for the
effectiveness of erosion and transport. The rating curve established from
the 1213 daily averages of C and Q data enabled the estimation of C then
Qs (Qs=C×Q) for the whole period 1970–2010
from the measured daily Q values.
Comparison between estimates of Qs obtained from Q and the
global rating curve, and measured Qs.
Considering the change in hydrologic regime during the study period, we
wondered if the estimate of C and Qs per subperiods such as decades
would be more appropriate. We therefore applied the four rating curves
established for the 4 decades to the time series of daily Q to obtain daily C
and then daily Qs. This method (B) enabled us to compare the
estimated solid discharge with the value provided by the global relationship
established from 40 years of data (method A). The average error for daily
Qs values was 51 % using method A and 42.1 % using method B.
However, the cumulative flux of suspended matter over the 1213 days for
which daily data are available was overestimated by 3.1 % using method A
while it was underestimated by 5 % using method B. A comparison of the
estimates by these two methods showed that method B is not reliable at high
discharge during the last decade because of an increase in scattering of the
C, Q pairs. The relationship obtained over the last decade (2000–2010) led to
an underestimation of Qs of 23 % over the 314 days for which
daily C and Q are known. In contrast, the global algorithm from method A led to an
underestimation of the same cumulated Qs by only 3.5 % over the same
period. The relationship established over 40 years was therefore used for this study.
It should be noted that although method A provides some daily solid
discharges from the 1213 daily Q values with a high error (the average error
being 51 %), it enabled the reconstruction of good trends of Qs values
over more than 7 orders of magnitude (Fig. 5). The temporal
variability of the coefficients a and b of the rating curves calculated over
years or decades will be discussed in light of the variability of the
forcings and their consequences on sediment transport in order to better
understand their physical meaning.
Average loads
In order to analyse the temporal variability of suspended sediment flux, we
use the average concentration resulting from the ratio between the solid and
the liquid flow rate, denoted as SPM*, which can be defined for any
integration period (day, month, season, year).
Study of breaks: double-mass curve
Double-mass curves were used to determine long-term trends and changes in
the hydrosedimentary regime (Searcy and Hardison, 1960; Walling, 1977b, 2006).
General statistics on the yearly averages of hydrologic
parameters from the Wadi Abd at Ain Hamara gauging station over the period
1970–2010 (note: T at Chlef was estimated from measurements at Dar El Beida
and Miliana for 20 months over 480 months).
StatisticT (Chlef)PQQwMSPM*∘Cmm yr-1m3 s-1m3 s-1103 t yr-1g L-1Mean19.092641.181.2956412.3Min17.521650.370.4633.12.56(Year)1971–19721999–20001992–19931983–19841992–19931975–1976Max20.325062.192.98326650.25(Year)1989–19901995–19961994–19951994–19952007–20082007–2008SD0.6971.20.520.5969610.6CV (%)27.044.445.6123.386.0
Interannual variations of mean yearly temperature (calculated
from September to August monthly temperatures) at three stations in northern
Algeria: Dar El Beida, Miliana and Chlef (from measurements of CRUTEM4 only,
extrapolated values are not shown).
Interannual variations of temperature, precipitation, river discharge and flow regime
The statistics of hydrological parameters at Ain Hamara gauging station over
the period 1970–2010 are reported in Table 1.
Temperature
Temperature in northern Algeria at the three stations of Chlef, Miliana and
Dar El Beida increased from the 1970s onwards (Fig. 6), with higher values
at Chlef than at Dar El Beida and Miliana. On average, temperature at Chlef
increased by 1.17 ∘C from the 1970s to the 2000s (Table 2). The
increase was 0.87 ∘C between the 1970s and the 1980s which is more
than 4 times the difference between the 1980s and the 1990s
(+0.18 ∘C) and the 1990s and the 2000s (+0.12 ∘C).
As has been shown on a global scale, the decade of the 2000s was the warmest (IPCC, 2013).
Precipitation
Annual precipitation at Ain Hamara station was highly irregular, varying
between 165 and 506 mm yr-1 (Table 1, Fig. 7). Mean annual
precipitation (P) was 264 mm, with a coefficient of variation (CV) of 27 %
between 1970–1971 and 2009–2010. The interannual variations of P (Fig. 7)
showed trends towards a decrease of rainfall (-1.86 mm yr-1, on
average, over 40 years, p< 0.05). P decreased by 15 % (from 310 to 264 mm)
between the 1970s and 2000s (Table 2). A more precise analysis shows that
rainfall greatly decreased from the 1970s to the next decade (from 310 to
231 mm, -25 %), then slightly increased in the 2 following decades (see Table 2).
The average precipitation over the six rainfall gauging stations within the
basin was 273 mm yr-1, with consistent variations as compared to the Ain
Hamara station. Five out of six stations show a decrease in precipitation
between 1970–1985 and 1985–2010, the average deficit being equal to 3.7 %.
Interannual variations of annual precipitation, water discharge
and sediment yield at Ain Hamara station.
General statistics on the hydrologic parameters (averages)
from the Wadi Abd at Ain Hamara gauging station per decade and for the
entire period from 1970 to 2010 (note: T at Chlef was estimated from
measurements at Dar El Beida and/or Miliana for 20 months with missing
values over 480 months).
Italic numbers refer either to the full 40-year period or to the two subperiods
of perennial regime (1970–1985) or intermittent regime (1985–2010), while
nonitalic numbers refer to the average values per decade.
PeriodT atP, yearly NDD,Q, yearly Qw, yearly Qs, yearly Q98, average of SSY,SPM*Rating curve parameters Chlefprecipitation averagedischarge discharge of wet sediment load yearly values averagedays AverageAverageCVyearlyAverageCVAverageCVAverageCVAverageCVspecificAverageCVabR2N(∘C)(mm)(%)number(m3 s-1)(%)(m3 s-1)(%)(103 t yr-1)(%)(m3 s-1)(%)sed. yield(g L-1)(%)of dry(t km-2 yr-1)days(Q= 0)1970–201019.09264.1027.028.31.1844.41.2945.7564123.39.1878.6227.612.386.02.2700.6470.43112131970–198018.32310.5319.41.21.1632.91.1632.918078.84.3766.972.74.5447.91.0210.8900.5732401980–199019.19231.2316.824.10.9836.81.0741.533491.77.3968.0134.59.9357.02.0490.6490.4493161990–200019.37250.4240.559.91.1355.11.3455.261498.311.0388.5247.514.3669.22.7530.6590.4183432000–201019.49264.2219.728.11.4543.31.5742.2113090.313.9444.5455.620.5568.74.4400.4120.3843241970–198518.51284.3423.10.81.0237.81.0238.215978.94.1358.864.25.1658.91.2130.8180.5193461985–201019.47251.9629.044.81.2845.11.4543.780897.012.2161.1325.616.6567.42.9740.5760.415867River discharge and flow regime
The annual discharge was 1.18 m3 s-1, on average, and exhibited a
higher interannual variability (CV = 44.4 %) than annual precipitation
(Table 1). Yearly values showed a trend towards an increase of river flow
(+11.3 L s-1 yr-1, on average, over 40 years, p< 0.01;
Fig. 7), with decreasing decadal values between the 1970s and the 1980s,
then increasing values afterwards, similar to P (Table 2).
Detailed analysis of daily river discharge shows that the river was
perennial in the 1970s and then became intermittent during the 1980s (Fig. 8).
The driest year occurred in 1993–1994 with 117 days of the fully dry river. In
Fig. 8, very low river discharges (around 0.01 m3 s-1) were not
considered as days of dry river.
From the 1970s to 2000s, when Q averaged over 10 years increased by 25 %, the
wet discharge Qw (i.e. the yearly average discharge of the days of
running river) increased by more than 35 % (Table 2). Two indicators of
intra-annual discharge variability are shown in Fig. 7: Q98, the 98th
percentile of annual flows calculated from daily discharge and the standard
deviation of daily discharge within each year (σQ). Q98
increased by a factor 3.2 between 1970–1980 and 2000–2010 (Table 2). Q98
is a good indicator of changes in sediment transport as it occurs during the
highest flood events that occur each year.
Variation of hydrological regime with annual percentage of time of
flowing water, Q98 (amongst daily discharges, per year) and annual
standard deviation of daily river discharge.
Interannual variation of sediment loadRating curve
The rating curve obtained from 1213 pairs of daily averages gave the following:
C=2.270Q0.647r2=0.431.
Of the variations of C, 43 % are explained by those of Q. The rating curve
obtained between Q and Qs shows a much higher coefficient of
determination (r2= 0.831) but is biased since Qs=C×Q.
Nevertheless, both relationships give estimates of Qs values
from Q with less than 1 % difference which is less than the uncertainty of Qs.
Yearly sediment fluxes and concentrationsDecadal variability of Qs
Qs increased from 180 to 1130 × 103 t per year between the 1970s
and the 2000s (Table 2). The increase from one decade to the next is
remarkably regular: +85 % between the 1970s and the 1980s, +84 %
between the 1980s and the 1990s, +84 % between the 1990s and the 2000s and is
statistically significant (+19.7 × 103 t yr-1, on average,
p< 0.05). Specific sediment yield follows the same trend (Table 2).
Variability of mean annual load SPM*
The average value of SPM* over the period 1970–2010 is 12.3 g L-1, with
annual values comprised between 2.5 and 50.2 g L-1 (Tables 1 and 2).
Their interannual variation was smaller than that of solid discharge
because annual SPM* is the ratio of the annual Qs to the
annual Q (which increased less than Qs).
Analysis of break points
The double-mass plot enabled us to identify changes in the sediment response
of the stream (Fig. 9). A major break occurred in 1985–1986. A secondary break
was noticed in 1991–1992, but the entire period 1985–1986/2009–2010 may be
considered as a single period (with the relationship “cum Qs” = 0.021 × “cum Q” - 9.417, r2= 0.989). The period
1985–1986/1991–1992 may thus be considered as a transient event towards a new regime.
The response of sediment flow to various constraints differs clearly from
that of discharge from the year 1985–1986 onwards. This break corresponds to
the first year of dry river over a long period in summer (49 days). This
initiates a phase of intermittent flow regime. The averaged parameters for
the two periods 1970–1985 and 1985–2010 were added to the tables, in
addition to average values throughout the full study period and values for
decades to illustrate the dynamics of the hydrological and hydrosedimentary change.
High dependency of the solid discharge on Q variability
The variability of Q and Qs or SPM* at different timescales were
compared. AO2007 showed that, over 22 years, 71 % of the variance of the
annual SPM* values was accounted for by annual discharge and 73 % by the
95th percentile of daily discharge within the given year Q95. This
means that SPM* was mainly driven by the 10–15 highest daily discharges
in a year, suggesting a strong correlation between yearly Qs and the
discharge variability. Finally, they showed a remarkable linearity between
SPM* and the standard deviation of the daily discharge per year (σQ).
Yearly SPM* and yearly σQ still showed a strong linearity over
40-year period (r2= 0.956, Fig. 10a). A higher correlation was
obtained between yearly Qs or SSY, the specific sediment yield, and
yearly σQ (r2= 0.991, Fig. 10b). In conclusion, for this
river, the yearly solid discharge is more closely dependent on the discharge
variability than on discharge values.
Double-mass plot of sediment yield versus water flow.
Variation of the seasonality of climatic and hydrological parameters
The yearly values of temperature at Chlef generally increased but the
monthly averages showed high discrepancies. Temperature from March to
November increased with a maximum of increase in June (+3.30 ∘C,
on average, between the 1970s and 2000s), it remained quite constant in
December and February and decreased by 0.98 ∘C in January over the
same period. Considering the average values per season, winter values
(December–February) decreased by 0.33 ∘C between the 1970s and the 2000s,
while spring values (March–May) increased by 1.66 ∘C, summer values
(June–August) by 2.22 ∘C and fall values (September–November) by
1.29 ∘C. In summary, annual temperature differences increased with minimum
temperatures down slightly and maximum temperatures rising sharply. The
increase was most marked in July–August.
Variation of precipitation, water discharge and sediment
yield averaged per season over each decade.
Precipitation (mm) Water discharge (m3 s-1) Sediment yield (103 t) AutumnWinterSpringSummerAutumnWinterSpringSummerAutumnWinterSpringSummer1970–198068.5102.6128.211.21.261.381.410.5862.243.766.18.41980–199056.094.470.710.11.151.291.080.40128.861.097.246.81990–200067.081.186.915.51.860.991.110.54279.157.8130.9146.02000–201078.698.477.79.53.041.351.060.35804.994.4195.335.4Precipitation (%) Water discharge (%) Sediment yield (%) AutumnWinterSpringSummerAutumnWinterSpringSummerAutumnWinterSpringSummer1970–198022.133.041.33.627.329.830.312.634.524.236.64.71980–199024.240.830.64.429.432.827.710.138.618.329.114.01990–200026.732.434.76.241.322.124.612.045.59.421.323.82000–201029.737.329.43.652.523.218.26.171.28.417.33.1
Yearly average of related sediment load parameters vs. intra-annual
variability of daily river discharge, characterized by their annual standard
deviation. (a) displays SPM*, (b) specific sediment yield.
Trends of the seasonal indexes of precipitation (a), discharge (b)
and sediment discharge (c) in the Wadi Abd basin.
Monthly values of precipitation (a), Q(b) and
Qs(c) averaged over decades in the Wadi Abd basin.
Averaged seasonal values of P, Q and Qs for each decade are given in
absolute values and in percent of the yearly values in Table 3. The seasonal
relative contribution of P, Q and Qs centred and averaged over
9 consecutive years are presented in Fig. 11. The monthly values of P, Q and
Qs per decade over 40 years also clearly illustrate the absolute
changes in intensity and in seasonality of the river regime (Fig. 12). The
main conclusions of the analysis of P, Q and Qs seasonal variations are
the following:
Rainfall decreased in spring and increased in autumn. Precipitation in
autumn increased from 22 to 30 % at the expense of spring rains
(decreasing from 41 to 29 %). For the decade 2000–2010 precipitation
was the same in autumn and in spring (78 mm) while for the decade 1970–1980
spring rainfall was 87 % higher than in fall (see Table 3 and Fig. 11a).
Average monthly rainfall from six weather stations in the river basin for
1970–1985 and 1985–2010 (Fig. 13) illustrates the changes. Two marked
seasons typical of a Mediterranean climate are present (a dry season and a
rainy season) but the following changes are observable: (1) differences
between seasons decrease, as indicated by the CV of monthly rainfall from
57.3 in 1970–1985 to 45.9 % in 1985–2000 (there is a decrease of
spring rains (March–May) and at the beginning of the cold season
(November–December),
as well as the strengthening of rain in the warm season (July–October) and in winter
(January–February)); (2) advancement of the rainy season as evidenced by
precipitation in September and October; (3) spreading of the rainy season
over 9 months (September–May) for 1985–2010 from previously 7 or 8 months (from
October or November onwards); (4) increased regularity of rainy season precipitation.
Proportionally, flow decreased from winter to summer and increased
dramatically in autumn from just over a quarter (27.3 %) of the flow
delivered during 1970–1980 to more than one half (52.5 %) during 2000–2010
(Table 3 and Fig. 11b). Flow decreased in summer and the river became dry
for much of the summer. Over the last decade, it is striking to see the
difference between the average flow rates in fall and spring: the fall rate
is almost 3 times that of spring with almost the same rainfall. This
trend is evident over the 40-year period (Fig. 11b).
These results point towards a change in runoff as defined by the ratio Q/P.
Considering the whole basin area, the river discharge at the Ain Hamara station
averaged over 40 years corresponds to a water depth of 15 mm yr-1,
while the average precipitation is 264 mm yr-1. For comparison, on
average, 85 % of rain in this region evaporates and the remaining 15 %
runs into surface waters or infiltrates underground storage (Sari,
2009, quoted by Benhamiche et al., 2014). On the Wadi Abd, Q/P averages
5.7 %. We calculated the value of Q/P averaged over 3 consecutive years
and over 3 consecutive months (centred) and then took the average per
decade (Fig. 14). It appears that the Q/P ratio remains constant during the
months from December to April (around 4.4 %, on average), it increased
slightly in November and May during the decade 2000–2010 and it increased
significantly from September to November. In other words, runoff increased,
rain decreased slightly and the temperature (and therefore ETP) increased.
As a consequence, infiltration will decrease and the water level in the
aquifers will be lowered.
In absolute values, solid discharge has been increasing in all seasons over
4 decades, but more so in the fall than in the other seasons (Table 3 and
Fig. 12c). During autumn, it more than doubled from one decade to another.
During the other seasons, it doubled or tripled between the 1970s and 2000s
(see Table 2). While during the 1970s the Wadi Abd had two major periods of
roughly equivalent sediment discharge in the fall and spring, suspended
sediment loads were greater in the autumn during the 2000s (> 70 %
of the yearly discharge). The Wadi shifted from a regime with two
equivalent seasons of sediment production to a regime with one dominant
season in the 2000s. Autumn produced over 4 times more sediment than spring
in the 2000s (Table 3, Fig. 11c). This phenomenon does not seem to be due to
some exceptional floods because the trend is observable over 4 consecutive
decades (Fig. 11c).
Monthly values of precipitation averaged over six stations, for
the two periods 1970–1985 and 1985–2010.
DiscussionInterannual variationsHydrology and climate change over 40 years
Temperature increased rapidly between the 1970s and 1980s
(+0.87 ∘C, on average, at Chlef). The increases were lower during
the 3 following decades. An increase in temperature of 1.6 ∘C
between 1977–1979 and 2000–2006 was noted by Dahmani and Meddi (2009) for
the Wadi Fekan basin in west Algeria and Bakreti et al. (2013) also showed a
significant trend of increasing temperature in spring by 0.0183 ∘C
per year in the Tafna basin in west Algeria over the same period. However,
temperature has not increased as rapidly over the 20th century (Fig. 6) and
as mentioned by IPCC (2013), “trends based on short records are very
sensitive to the beginning and end dates and do not in general reflect
long-term climate trends.” The longest available time series of temperature
in Algeria was measured at Dar El Beida near Algiers. At this station,
average temperature increased by 0.62 ∘C between 1850–1900
(29 yearly values available) and 2003–2012 (Fig. 6), while it increased between
1880 and 2012 by 0.85 ∘C globally (IPCC, 2013).
A global trend towards increasing temperatures and increasing dryness in
Algeria from the 1970s onwards has already been described (Meddi and Meddi,
2009). Over the period 1923–2006, north Algeria experienced an alternation of
wet periods (1923–1939, 1947–1973) and dry periods (1939–1946 and from 1974
onwards) (Benhamiche et al., 2014). Over 70 years in the Wadi Fekan, Dahmani
and Meddi (2009) showed that the period 1943–1960 was rather wet, that
1960–1975 was average, and that the period 1975 onwards (up to the end of
their data set in 2004) was dry and of an exceptional long duration. Using
three different statistical tests (Pettitt, Lee–Heghinian and Hubert), Meddi
and Meddi (2007) shown that a shift was observed between 1973 and 1980 over
most of the rain gauges in Algeria. In northwest Algeria, a shift was
noticed in 1973 in winter rainfall and between 1974 and 1980 in spring
rainfall, both of these shifts being responsible for the yearly rainfall deficit
(Meddi and Talia, 2008). From the rainfall data set at the Ain Hamara station
between 1968 and 2007, Hallouz et al. (2013) showed that the break in annual
rainfall occurred in 1976 and calculated a deficit of 19 % between
1968–1976 (304 mm yr-1) and 1976–2007 (247 mm yr-1). At the
stations Ponteba and Rechaiga, near to the Abd basin, the trends of
decreasing total precipitation and of increasing mean length of dry spells
were amongst the five highest in the Maghreb area over the 22 stations
considered by Tramblay et al. (2013, see their Fig. 8).
As a consequence of the decrease of rainfall after the 1970s break which was
observed in most basins of western Algeria, river discharges were generally
seen to decrease as well. Meddi and Hubert (2003) showed that the decrease
in river discharge varied between -37 and -70 % from eastern Algeria
to western Algeria. In the Mecta basin in northwest Algeria, runoff was
estimated to be 28–36 % lower in 1976–2002 as compared to 1949–1976 (Meddi
et al., 2009). In the Tafna basin, also in northwest Algeria, Ghenim and
Megnounif (2013a, b) showed that the decrease in precipitation after the
break point was, on average, 29 % over the whole basin (especially in
winter and spring) and was accompanied by a decrease of 60 % in river flow.
Monthly values of the ratio Q/P averaged over decades.
In contrast, the Wadi Abd behaved differently in that river discharge
increased. The counterintuitive increase of runoff with decreasing rainfall
has also been observed in the Sahel and is referred to as “the Sahelian
paradox” (see Mahé and Paturel, 2009; Mahé et al., 2012). A closer
look at the seasonal variations of the different parameters shows that Q
decreased in winter and spring but that Q/P increased in autumn when
rainfall increased. Overall Q increased. The decrease of rainfall in spring
and its low level in summer may have led to a change in vegetation cover
which would in turn decrease infiltration. However, although studying the
vegetation dynamics of the basin goes beyond the scope of this study, this
aspect could be investigated in the future using satellite data, for example.
What is the influence of large-scale circulation indices?
Changes in precipitation are derived from atmospheric–oceanic signals
(Milliman et al., 2008; Giuntoli et al., 2013). Low-frequency fluctuations
related to climate change are modulated with higher-frequency interannual
fluctuations, such as ENSO (El Niño–Southern Oscillation), NAO (North
Atlantic Oscillation), AMO (Atlantic Multidecadal Oscillation) or MO
(Mediterranean Oscillation). Tramblay et al. (2013) showed that the
precipitation amounts and the number of dry days over the Maghreb were
significantly correlated with the MO and NAO patterns. MO and NAO showed
positive trends from the 1970s onwards which could explain the trend towards
decreasing frontal conditions over the Mediterranean basin and thus
increasing droughts.
Interannual influence by the Austral oscillation ENSO over Algeria was shown
to be higher in northwest Algeria on the highest discharges than on the
average discharge. The maximum Q seems to be smaller during El Niño and
higher during La Niña in northwest Algeria (Ward et al., 2014). The
frequency of extreme rainfall events shows the highest correlation with the
Mediterranean Oscillation index in Algiers and with the Southern Oscillation
index in Oran (Taibi et al., 2014).
In this study, no significant correlation was established between a series
of hydrological parameters in the Wadi Abd and the Southern Oscillation
index. The average of AMO per hydrologic year was calculated from its
monthly values. AMO has increased from 1970s to the 2000s, with negative
values until 1993–1994, then positive values thereafter (except in 1996–1997).
Its decadal average was -0.25 in the 1970s, -0.12 in the 1980s, 0.0 in the
1990s and 0.18 in the 2000s. AMO and the discharge variability of the Wadi
Abd within the year increased coincidently. The yearly AMO values have a
coefficient of determination of 0.226 when correlated with the standard
deviation of daily river discharges within the year, a proxy for the
variability of daily discharge. However, this information does not allow us
to conclude that the Atlantic Multidecadal Oscillation is responsible for
hydrological changes in the Wadi Abd basin.
Break point in 1985–1986: change of flow regime
The several weeks of dry river for the first time in 1985–1986 (49 days) can
be considered as a threshold effect, which marks the start of a new flow
regime. The appearance of a dry regime is a break, a fully nonlinear
phenomenon. It has strong consequences for water infiltration and
groundwater recharge, on the seasonality, intensity and type of floods, and
in turn, on erosion and sediment transport. The year 1985 is also a pivotal year for
recent climate change as evidenced by the rapid increase in global mean air
temperature anomaly from that year until 1993 (Fig. 1 in Lockwood and
Fröhlich, 2007). The hypothesis of a temporary warming caused by dust
emitted during the eruption of Mount Pinatubo had been advanced to explain
the warming since 1985, but climate scientists later recognized that the
temperature anomaly has been increasing since 1993, reaching about
0.6 ∘C by 2007 compared to the global average temperature
calculated for the period 1951–1980 (Lockwood and Fröhlich, 2007).
This threshold is coincident with hydrological shifts in the Tafna basin in
northwest Algeria. Bakreti et al. (2013) analysed the baseflow and baseflow
index of five of the Tafna's sub-basins between 1976 and 2006 and found ruptures of
the baseflow index between 1984 and 1990 depending on the sub-basin, in
1984, 1985 and 1990 in the mountains, and in 1985 and 1986 in the plain.
These changes in flow regimes in the Tafna basin were likely caused by
shifts in rainfall in the late 1970s in the Mounts of Tlemcen and early 1980s in
the plains (Ghenim and Megnounif, 2013a).
Shift of the onset of the first summer flood
The analysis of the time series of daily flows enables the determination of
the start of the first summer flood. The average daily flow per decade
suddenly increases the day at which the first summer flood occurred, at
least once in the decade. By observing these decadal averaged daily flows,
there is no ambiguity on the start of the earliest flood by decade:
In 1970–1980, the first flood starts on 6 September with an average
4-day discharge (6–9 September) of 1.59 m3 s-1, while it was, on
average, 0.58 m3 s-1 over the 4 previous days.
In 2000–2010, the first flood of summer starts on 8 August with an average
4-day discharge (8–11 August) of 2.03 m3 s-1, while it was, on
average, 0.03 m3 s-1 from 4 to 7 August.
During the 2000s, the first flood in summer started close to 1 month
before that of the 1970s and the magnitude was 27 % higher. It can be
asked if this trend was observable over the 40-year period or only between
2 specific decades. The analysis of mean dates and discharges of the first
flood in the late dry season gave the following results for the intermediate decades:
In 1980–1990, the first flood started, on average, on 31 August with a 4-day
average discharge (31 August–3 September) of 2.69 m3 s-1, while the average
rate over the 4 previous days was 0.13 m3 s-1.
In 1990–2000, the first flood started, on average, on 22 August with a 4-day
average discharge (22–25 August) of 7.67 m3 s-1, while the average
rate over the 4 previous days was 0.33 m3 s-1. The existence of
a precursor peak on 17 August, which was not observed in previous decades,
was also observed.
It therefore appears that the date of the first flood advanced by about 10
days each decade over the previous 40 years. The shift in the onset of the
first flood in summer probably has important consequences on flow and
erosion rates.
Relationships between several parameters and sediment yieldTemperature and sediment yield
The curve showing annual suspended load versus the global air temperature
anomaly (base period 1951–1980) calculated by hydrological year from monthly
data provided by NOAA (Hansen et al., 2010; GISTEMP Team, 2015) shows a
correlation between the sediment yield and ongoing climate change
(r2= 0.388, Fig. 15).
Variations of SPM* against the global mean temperature anomaly.
Precipitation and sediment yield
Many authors studied the variations of sediment load per unit of catchment
area against annual rainfall (e.g. Summerfield and Hulton, 1994) or
effective rainfall (e.g. Langbein and Schumm, 1958). On the Wadi Abd, annual
rainfall fell sharply between the 1970s and the 1980s, then slightly
increased over the following decades. Meanwhile, yearly sediment
concentration and suspended sediment discharge have increased. The
comparison of their respective variations shows a lack of correlation
between precipitation and annual sediment yield (r2< 0.1,
regardless of the type of regression considered). Regarding the relationship
between precipitation and erosion, if there are correlations between their
spatial variations reported in the literature (though with a strong scatter,
see Riebe et al., 2001), our study shows that the temporal variations of
precipitation and sediment yield are not correlated in the Wadi Abd. This
may be due to the change of flow regime within the study period.
Runoff and sediment yield
Although runoff was noted to have a limited impact on the distribution of
sediment yield at regional or global scales by Aalto et al. (2006), Syvitski
and Milliman (2007) and Vanmaercke et al. (2014), the temporal variability in
precipitation, runoff (or discharge) and consecutive vegetation cover was
shown to be locally the main impact on fluvial sediment load (see Vanmaercke
et al., 2014, p. 360). On the Wadi Abd, the yearly suspended sediment load
was highly correlated with discharge (Q mean or its highest percentiles) and
to its intra-annual fluctuation (Fig. 10). Although the river regime shift
clearly impacted several parameters, the relationship between yearly
sediment load and discharge variability did not change over the 40-year study period.
On the use of double-mass curves to determine the climate change and anthropogenic influences
Double-mass curves are often used to determine the impact of developments
such as dams on sediment discharge (e.g. Lu et al., 2013). Our findings warn
about extrapolations that could be wrongly made to quantify the impact of a
development by extending the double-mass curves. Indeed, this study shows
that the double-mass curve can change its slope (here increasing) when the
flow regime change is driven by seasonal temporal variation in precipitation
and runoff that is not linked to any specific anthropogenic activity (such as
a dam impoundment) within the basin.
Physical meaning of rating parameters a and bInterannual variation of (a, b)
Since C=aQb, with b≠ 0, then C(1) =a. a thus represents the
sediment concentration when the river discharge is 1 m3 s-1, and
b reflects the sensitivity of concentration to discharge variation. The
general formula lnC=ln(aQb) provides:
dC/C=bdQ/Qb=dC/dQ⋅Q/C=1/a⋅dC/dQ⋅Q(1-b)
thus b varies almost like 1/a (Asselman, 2000). Many papers discuss the
physical meaning of the rating parameters a and b (see AO2007) and try to
connect their values to physiographical characteristics, vegetation cover or
hydrometeorological forcing.
The river's regime change is accompanied by a change in the (a, b) pairs of
rating curves defined for multi-year periods such that a increases and b
decreases (Table 2), following:
b=-0.294lna+0.912r2=0.582lnb=-0.188a+0.042r2=0.649.
Equation (5a) is very similar to that presented by Iadanza and Napolitano (2006)
for the Tiber River after the construction of a dam
(b=- 0.3815 lna+ 0.7794, r2= 0.992). Before the construction of this dam, another
relationship corresponded to more than 3 times higher sediment yields.
Asselman (2000) has suggested interpreting the regression lines in a lna-b
graph as different sediment transport regimes.
On the Wadi Abd, the change in sediment transport regime is not evident from
the yearly (a, b) values but it becomes clearly observable when considering
a and b values averaged over moving periods of several years. The best
correlations were obtained for running averages over 15 years named a15
and b15 (N= 25, from 1970–1985 to 1995–2010, see Fig. 16). The
available data set does not allow us to determine if results obtained from
averaging over longer periods would perform best.
The time evolution of the moving average pair (a15, b15) clearly
shows a first relationship with the values dominated by the pre-1985 regime
(8 values from 1970–1985 to 1977–1991), another one for the values
predominantly after 1990 (12 values from 1983–1997 to 1995–2010), both with
a15 increasing and b15 decreasing, and a transitional regime
centred on the period 1985–1990 (Fig. 16). During the transition period
centred over 1985–1990, b15 was almost constant (between 0.72 and 0.74)
while a15 was increasing from 2.01 to 2.34. During the period
1985–1991, the yearly values of b varied very little (between 0.653 and 0.672)
while yearly a increased significantly from 1.81 in 1985–1986 to
3.23 in 1990–1991. Higher a and lower b values are in the literature typical of
highly arid river basins, such as the ephemeral Nahal Eshtemoa in Israel,
where a= 16.98 and b= 0.43 (Alexandrov et al., 2003).
Relationship between the rating curves parameters averaged over 15 years.
As the break points were coincident, it is possible to analyse the change
(of a15 and b15) in terms of shift of hydrological regime. However, if
the new hydrological regime was immediate from 1985 onwards, the change in
the C–Q relationship was only evidenced in the Wadi Abd at midterm,
considering 15-year average values.
Parameters that explain a (or b)
The coefficient of determination between a and SSY
is low at the annual scale but higher when we consider the moving averages
of a and SSY over 15 years. The SSY explained 95.2 %
of the variance in the interannual scale (Fig. 17), much more than the
average river flow did (r2= 0.839). b15 showed a lower
correlation with the SSY (r2= 0.853) than a15 did.
In summary, the moving average of a is strongly correlated to specific
sediment yield over the same moving period of 15 years, and the moving
average of b can be deduced from a using the relationship which is given in
Fig. 16 as a function of flow regime, either perennial or intermittent.
Validity range of rating curves
The estimation of sediment yield from flow measurements and a rating curve
is still acceptable throughout the study period (Fig. 5). However, the pairs
(C, Q) become increasingly scattered with time around the best-fit curve, as
attested by the decrease of the coefficient of determination from one decade
to another (Table 2).
Relationship between the rating curve parameter a averaged
over 15 years and the averaged values of specific sediment yield over 15 years.
Intermittent flows induce a stronger dependency of river behaviour on
antecedent wetness (Beven, 2002) and antecedent weathering, i.e. a strong
dependency on memory through threshold and hysteresis effects. With
increasing memory effects, coincident values of C and Q become less
dependent on each other and the rating curves less suitable to model their
relation. The study of sediment dynamics in the Wadi Abd in the future will thus likely
require a more appropriate method than rating curves, such as
the study of each individual flood, like Megnounif et al. (2013) did in the
Wadi Sebdou. This finding may have consequences on water management as well.
When dealing with rating curves, water discharge must be recorded at
frequent intervals, although measurements of concentration can be sparser.
When rating curves cannot be applied, river discharge and sediment
concentration should be both frequently and simultaneously measured.
Conclusions
In response to climate change which resulted in an increase in temperature
of around 1.1 ∘C between the 1970s and 2000s at Chlef,
rainfall moved forward during the late warm season and the watershed of Wadi
Abd experienced a significant change in flow regime and an increased
variability at both the inter-annual and intra-annual levels. These changes
ultimately led to a dramatic and continuous increase in sediment load over
4 decades (on average, 84 % more every decade as compared to the previous one).
The main result of our analysis is the shift of the onset of the first
summer flood that occurred 1 month earlier in the 2000s than in the 1970s.
This shift is likely responsible for the cascading effects on the
hydrological regime of the Wadi Abd. In particular, earlier floods during
the warmer season have higher evaporation which limits the groundwater
storage. A parallel study of seasonal changes in vegetation cover is needed
to provide additional information.
The increase in erosion of the watershed (coefficient a) is accompanied by a
decrease in the coefficient b. The traditional rating curves approach, which
was applicable when the river was perennial, is now less adapted to model the
behaviour of the river (Table 2). This could be explained by a more
pronounced hysteresis phenomenon, which is consistent with the change of
hydrological regime in the dry season thereby limiting the utility of rating
curves to model C–Q relationships.
The rapid change in sediment regime which is instantaneously driven by the
changing flow regime should be distinguished from the slow change in the
concentration–flow relationship. The change in flow regime can be precisely
dated in May–July 1986 (with 49 consecutive dry days), while the change in
the C–Q relationship requires averaging over several years of a, b and
specific sediment yield to become evident. Such inertial effect may be
attributed to the time for the basin soil properties (such as humidity) or
vegetation to adapt to the new climate conditions. It likely depends,
amongst other factors, on underground water storage, and thus on basin
lithology and infiltration history. Over the Wadi Abd basin, the time needed
for the flow regime to change after the dryness settlement in early 1970s
(see Fig. 6) is estimated as being around 15 years in this study.
The present analysis only includes hydrological parameters. Management
programs that were conducted to fight erosion in Algeria from the 1960s until
the 1990s by reforesting and setting up banks over cultivated marl and clay
areas proved to be little or not efficient (Touaibia, 2010). Human
activities may have influenced the hydrological regime change and increased
erosion, in particular through firewood cutting during economically
difficult periods (1990s), however the shift was shown to occur earlier. The
lack of data on land use and land cover changes over 40 years does not allow
us to isolate the factors directly related to climate change from those
related to other anthropogenic activities. However, the small population,
the low coverage of pasture (see Fig. 2d), cultivated areas and
vegetation (43 %) in the basin and the small volume of reservoirs
(nominally 2.3 % of the annual discharge, and silted up to 70 %) make us
think that in this system the effects of climate change dominate
anthropogenic effects. The quantification of forcing changes on sediment
sources (raindrop erosion, sheet erosion, rill erosion, gully erosion and
stream channel erosion) may be investigated in situ (e.g. Poesen et al.,
2003) and/or estimated using a numerical model of the hydrologic and
sedimentological functioning of the basin, such as WEPP (Nearing et al.,
1989), EUROSEM (Morgan et al., 1998) or SWAT (Neitsch et al., 2011). Such a
model could help us to test hypotheses and quantify or at least estimate the
effects of different forcing changes (temperature, runoff, vegetation, etc.)
in future studies.
It is important to emphasize that it is impossible to define long-term
hydrological averages in the context of a changing flow regime. The example
of the Wadi Abd shows that the difficulty is challenging with regard to
sediment transport in suspension, since changes of flux cannot be counted as
a fraction but can reach an order of magnitude.
Changes in flow regime in relation to climate change can be investigated
using climate models. Das et al. (2013), using 16 climate projections, showed
that more intense floods of a return period 2–50 years should occur in
the Sierra Nevada, regardless of the rainfall variation. The recent changes
in the Wadi Abd show that extreme events with increasing variability already
occur in the basin. Over Algeria, an increase of 1–2 ∘C in
temperature could induce a reduction of 10 % in precipitation before the
end of the 21st century (Benhamiche et al., 2014) with unknown
consequences on erosion and sediment transport. Lu et al. (2013) calculated
the impact on sediment loads of every 1 % change in precipitation or river
discharge in large Chinese rivers. Such a calculation has no meaning in our
basin since the rainfall and discharge were not monotonic (severe decrease
in the 1970s then slight increase during 30 years) while the sediment loads
have always increased. The difficulty of forecasting climate change-driven
impacts on sediment yield due to nonlinear effects has been underlined by
geomorphologists (see Goudie, 2006; Jerolmack and Paola, 2010; Coulthard et
al., 2012; Knight and Harrison, 2013). The present study illustrates that
the change in flow regime induced a fully nonlinear effect between river
discharge and sediment yield. This needs be considered in forecasts
especially in small river basins in semi-arid areas.
Changes in erosion and sediment transport under new climate constraints will
induce changes from the middle to long term that decision-makers must
integrate into water resources management, habitat status, agricultural
adaptation (O'Neal et al., 2005), landscape evolution (Temme and Veldkamp,
2009) as well as in many other environmental adaptations (Ouillon, 1998). We
thus encourage the local adaptation of sampling strategies and measurements
to take into account changing in flow regimes. Furthermore, due to the
uncertainty of water resources and erosion in the Maghreb (Taabni and El
Jihad, 2012) and in the Mediterranean basin (Nunes et al., 2008), we also
encourage the development of studies on long-term sediment transport in
north African basins, in connection with changes in forcing factors.
Acknowledgements
The authors warmly thank Abda Leila from the Hydrometry office, Ould Lamara Arezki
from the Climatology Office of the Agence Nationale des
Ressources Hydrauliques (ANRH) in Alger, Boudalia Mohamed from the
Hydrometry Office of the ANRH in Oran, and all the staff of ANRH who
participated to the field measurements at Ain Hamara station. They also
acknowledge Abderrezak Kamel Toubal for his help in drawing Figs. 1 and 2.
Two anonymous reviewers are warmly thanked for their reviews and comments
on previous versions of this paper. The editor, Efrat Morin, is gratefully
acknowledged. Emma Rochelle-Newall, a native English speaker, is warmly
thanked for the English corrections.
Edited by: E. Morin
References
Aalto, R., Dunne, T., and Guyot, J. L.: Geomorphic controls on Andean
denudation rates, J. Geol., 114, 85–99, 2006.
Achite, M. and Ouillon, S.: Suspended sediment transport in a semiarid
watershed, Wadi Abd, Algeria (1973–1995), J. Hydrol., 343, 187–202, 2007.
Alexandrov, Y., Laronne, J. B., and Reid, I.: Suspended sediment
concentration and its variation with water discharge in a dryland ephemeral
channel, northern Negev, Israel, J. Arid Environ., 53, 73–84, 2003.
Asselman, N. E. M.: Fitting and interpretation of sediment rating curves, J.
Hydrol., 234, 228–248, 2000.
Bakreti, A., Braud, I., Leblois, E., and Benali, A.: Analyse conjointe des
régimes pluviométriques et hydrologiques dans le basin de la Tafna
(Algérie Occidentale), Hydrolog. Sci. J., 58, 133–151, 2013.
Bates, B. C., Kundzewicz, Z. W., Wu, S., and Palutikof, J. P. (Eds.): Climate
Change and Water. Technical Paper of the Intergovernmental Panel on Climate
Change, IPCC Secretariat, Geneva, 210 pp., 2008.Benhamiche, N., Madani, K., and Laignel, B.: Chapter 12 Impact of climate
change on water resources in Algeria, in: Vulnerability of agriculture, water
and fisheries to climate change: toward sustainable adaptation strategies,
edited by: Behnassi, M., Syomiti Mutengé, M., Ramachandran, G., and Shelat, K. N., Springer,
Dordrecht, 193–205, 10.1007/978-94-017-8962-2_12, 2014.
Beven, K.: Runoff generation in semi-arid areas, in: Dryland rivers: hydrology
and geomorphology of semi-arid channels, edited by: Bull, L. J. and Kirkby, M. J.,
Wiley, London, 57–106, 2002.Coulthard, T. J., Ramirez, J., Fowler, H. J., and Glenis, V.: Using the UKCP09
probabilistic scenarios to model the amplified impact of climate change on d
rainage basin sediment yield, Hydrol. Earth Syst. Sci., 16, 4401–4416,
10.5194/hess-16-4401-2012, 2012.
Dahmani, A. and Meddi, M.: Climate variability and its impact on water
resources in the catchment area of the Wadi Fekan Wilaya of Mascara (West
Algeria), Eur. J. Sci. Res., 36, 458–472, 2009.
Das, T., Maurer, E. P., Pierce, D. W., Dettinger, M. D., and Cayan, D. R.:
Increases in flood magnitudes in California under warming climates, J.
Hydrol., 501, 101–110, 2013.
Déry, S. J., Stieglitz, M., McKenna, E. C., and Wood, E. F.: Characteristics
and trends of river discharge into Hudson, James, and Ungava bays, 1964–2000,
J. Climate, 18, 2540–2557, 2005.
Durand, N., Fiandrino, A., Fraunie, P., Ouillon, S., Forget, P., and Naudin,
J. J.: Suspended matter dispersion in the Ebro ROFI: an integrated approach,
Cont. Shelf Res., 22, 267–284, 2002.
El Mahi, A., Meddi, M., and Bravard, J. P.: Analyse du transport solide en
suspension dans le bassin versant de l'Oued El Hammam (Algérie du Nord),
Hydrolog. Sci. J., 57, 1642–1661, 2012.Gao, Z. L., Fu, Y. L., Li, Y. H., Liu, J. X., Chen, N., and Zhang, X. P.:
Trends of streamflow, sediment load and their dynamic relation for the catchments
in the middle reaches of the Yellow River over the past five decades, Hydrol.
Earth Syst. Sci., 16, 3219–3231, 10.5194/hess-16-3219-2012, 2012.
Ghenim, A. N. and Megnounif, A.: Analyse des précipitations dans le
Nord-Ouest algérien, Sécheresse, 24, 107–114, 2013a.
Ghenim, A. N. and Megnounif, A.: Ampleur de la sécheresse dans le basin
d'alimentation du barrage Meffrouche (Nord-Ouest de l'Algérie),
Physio-Géo, 7, 35–49, 2013b.GISTEMP Team: GISS Surface Temperature Analysis (GISTEMP), NASA Goddard
Institute for Space Studies, http://data.giss.nasa.gov/gistemp/, last
access: 8 December 2015.
Giuntoli, I., Renard, B., Vidal, J. P., and Bard, A.: Low flows in France and
their relationship to large-scale climate indices, J. Hydrol., 482, 105–118, 2013.
Gomez, B., Cui, Y., Kettner, A. J., Peacock, D. H., and Syvitski, J. P. M.:
Simulating changes to the sediment transport regime of the Waipaoa River,
New Zealand, driven by climate change in the twenty-first century, Global
Planet. Change, 67, 153–166, 2009.
Goudie, A. S.: Global warming and fluvial geomorphology, Geomorphology, 79, 384–394, 2006.
Hallouz, F., Meddi, M., and Mahé, G.: Modification du régime
hydroclimatique dans le basin de l'oued Mina (Nord-Ouest d'Algérie),
Revue Sci. Eau, 26, 33–38, 2013.
Hancock, G. R.: A catchment scale assessment of increased rainfall and storm
intensity on erosion and sediment transport for Northern Australia,
Geoderma, 152, 350–360, 2009.
Hancock, G. R. and Coulthard, T. J.: Channel movement and erosion response to
rainfall variability in southeast Australia, Hydrol. Process., 26, 663–673, 2011.Hansen, J., Ruedy, R., Sato, M., and Lo, K.: Global surface temperature
change, Rev. Geophys., 48, RG4004, 10.1029/2010RG000345, 2010.Houyou, Z., Bielders, C. L., Benhorma, H. A., Dellal, A., and Boutemdjet, A.:
Evidence of strong land degradation by wind erosion as a result of rainfed
cropping in the Algerian steppe: a case study at Laghouat, Land Degrad.
Develop., 10.1002/ldr.2295, in press, 2014.
Iadanza, C. and Napolitano, F.: Sediment transport time series in the Tiber
River, Phys. Chem. Earth, 31, 1212–1227, 2006.
IPCC: Climate Change 2013: The Physical Science Basis, in: Contribution of
Working Group I to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change, edited by: Stocker, T. F., Qin, D., Plattner, G.-K., Tignor,
M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P.
M., Cambridge University Press, Cambridge, UK and New York, NY, USA, 1535 pp., 2013.Jerolmack, D. J. and Paola, C.: Shredding of environmental signals by
sediment transport, Geophys. Res. Lett., 37, L19401, 10.1029/2010GL044638, 2010.Jones, P. D., Lister, D. H., Osborn, T. J., Harpham, C., Salmon, M., and
Morice, C. P.: Hemispheric and large-scale land surface air temperature
variations: an extensive revision and an update to 2010, J. Geophys. Res.,
117, D05127, 10.1029/2011JD017139, 2012.
Kassoul, M., Abdelgader, A., and Belorgey, M.: Caractérisation de la
sedimentation des barrages en Algérie, Rev. Sci. Eau, 3, 339–358, 1997.
Knight, J. and Harrison, S.: The impacts of climate change on terrestrial
Earth surface systems, Nat. Clim. Change, 3, 24–29, 2013.
Langbein, W. B. and Schumm, S. A.: Yield of sediment in relation to mean
annual precipitation, T. Am. Geophys. Un., 39, 1076–1084, 1958.Lockwood, M. and Fröhlich, C.: Recent oppositely directed trends in
solar climate forcings and the global mean surface air temperature, P. Roy.
Soc. A, 463, 2447–2460, 10.1098/rspa.2007.1880, 2007.
Louamri, A., Mebarki, A., and Laignel, B.: Variabilité interannuelle et
intra-annuelle des transports solides de l'Oued Bouhamdane, à l'amont du
barrage Hammam Debagh (Algérie orientale), Hydrolog. Sci. J., 58, 1559–1572, 2013.
Lu, X. X., Ran, L. S., Liu, S., Jiang, T., Zhang, S. R., and Wang, J. J.:
Sediment loads response to climate change: A preliminary study of eight
large Chinese rivers, Int. J. Sed. Res., 28, 1–14, 2013.
Mahé, G. and Paturel, J. E.: 1896–2006 Sahelian annual rainfall
variability and runoff increase of Sahelain Rivers, C. R. Geoscience, 314, 538–546, 2009.
Mahé, G., Lienou, G., Descroix, L., Bamba, F., Paturel, J. E., Laraque,
A., Meddi, M., Habaieb, H., Adeaga, O., Dieulin, C., Chahnez Kotti, F., and
Khomsi, K.: The rivers of Africa: witness of climate change and human impact
on the environment, Hydrol. Process., 27, 2105–2114, 2012.
Meade, R. H. and Moody, J.A.: Causes for the decline of suspended-sediment
discharge in the Mississipi River system, 1940–2007, Hydrol. Process., 24, 35–49, 2010.
Meddi, M. and Hubert, P.: Impact de la modification du régime
pluviométrique sur les ressources en eau du nord-ouest de l'Algérie,
IAHS Publ. no. 278, in: Hydrology of the Mediterranean and Semiarid Regions,
edited by: Servat, E., Najem, W., Leduc, C., and Shakeel, A., IAHS Press,
Wallingford, UK, 229–235, 2003.
Meddi, H. and Meddi, M.: Variabilité spatiale et temporelle des
précipitations du Nord-Ouest de l'Algérie, Geographica Technica
no. 2/2007, 49–55, 2007.
Meddi, H. and Meddi, M.: Variabilité des précipitations annuelles
du Nord-Ouest de l'Algérie, Sécheresse, 20, 57–65, 2009.
Meddi, M. and Talia, A.: Pluviometric regime evolution in the North of
Algeria, Arab. Gulf J. Sci. Res., 26, 152–162, 2008.
Meddi, M., Talia, A., and Martin, C.: Evolution récente des conditions
climatiques et des écoulements sur le basin versant de la Macta
(nord-ouest de l'Algérie), Physio-Géo, 3, 61–84, 2009.
Megnounif, A., Terfous, A., and Ouillon, S.: A graphical method to study
suspended sediment dynamics during flood events in the Wadi Sebdou, NW Algeria
(1973–2004), J. Hydrol., 497, 24–36, 2013.
Memariam, H., Balasundram, S. K., Talib, J. B., Sood, A. M., and Abbaspour,
K. C.: Trend analysis of water discharge and sediment load during the past
three decades of development in the Langat basin, Malaysia, Hydrolog. Sci. J.,
57, 1207–1222, 2012.
Milliman, J. D., Farnsworth, K. L., Jones, P. D., Xu, K. H., and Smith, L. C.:
Climatic and anthropogenic factors affecting river discharge to the global
ocean, 1951–2000, Global Planet. Change, 62, 187–194, 2008.
Morgan, R. P. C., Quinton, J. N., Smith, R. E., Govers, G., Poesen, J. W. A.,
Auerswald, K., Chisci, G., Torri, D., and Styczen, M. E.: The European Soil
Erosion Model (EUROSEM): A dynamic approach for predicting sediment
transport from fields and small catchments, Earth Surf. Proc. Land., 23, 527–544, 1998.
Nanson, G. C., Tooth, S., and Knighton, A. D.: A global perspective on dryland
rivers: perceptions, misconceptions and distinctions, in: Dryland rivers: hydrology
and geomorphology of semi-arid channels, edited by: Bull, L. J. and Kirkby,
M. J., Wiley, London, 17–54, 2002.
Nearing, M. A., Foster, G. R., Lane, L. J., and Finkner, S. C.: A process-based
soil erosion model for USDA-Water Erosion Prediction Project technology, T. ASAE,
32, 1587–1593, 1989.
Neitsch, S. L., Arnold, J. G., Kiniry, J. R., and Williams, J. R.: Soil and Water
Assessment Tool – Theoretical Documentation version 2009, Texas Water
Inst. Techn. Report no. 406, Texas A & M University, College Station, 618 pp., 2011.
Nunes, J. P., Seixas, J., and Pacheco, N. R.: Vulnerability of water
resources, vegetation productivity and soil erosion to climate change in
Mediterranean watersheds, Hydrol. Process., 22, 3115–3134, 2008.
O'Neal, M. R., Nearing, M. A., Vining, R. C., Southworth, J., and Pfeifer,
R. A.: Climate change impacts on soil erosion in Midwest United States with
changes in crop management, Catena, 61, 165–184, 2005.Osborn, T. J. and Jones, P. D.: The CRUTEM4 land-surface air temperature data
set: construction, previous versions and dissemination via Google Earth,
Earth Syst. Sci. Data, 6, 61–68, 10.5194/essd-6-61-2014, 2014.
Ouillon, S.: Erosion and sediment transport: width and stakes, La Houille
Blanche, 53, 52–58, 1998.
Poesen, J., Nachtergaele, J., Verstraeten, G., and Valentin, C.: Gully
erosion and environmental change: importance and research needs, Catena, 50, 91–133, 2003.Pruski, F. F. and Nearing, M. A.: Climate-induced changes in erosion during
the 21st century for eight U.S. locations, Water Resour. Res., 38, 1298,
10.1029/2001WR000493, 2002.
Remini, B., Leduc, C., and Hallouche, W.: Evolution des grands barrages en
régions arides: quelques exemples algériens, Sécheresse, 20, 96–103, 2009.
Riebe, C. S., Kirchner, J. W., Granger, D. E., and Finkel, R. C.: Minimal
climatic control on erosion rates in the Sierra Nevada, California, Geology,
29, 447–450, 2001.
Sari, A.: Hydrologie de surface, cours d'initiation, Editions Distribution
Houma, Houma, Alger, 155 pp., 2009.
Searcy, J. K. and Hardison, C. H.: Double-mass curves, USGS Water-Supply
Paper 1541-B, USGS, Washington, D.C., 1960.
Sen, P. K.: Estimates of regression coefficients based on Kendall's tau, J.
Am. Stat. Assoc., 63, 1379–1389, 1968.Shrestha, B., Babel, M. S., Maskey, S., van Griensven, A., Uhlenbrook, S.,
Green, A., and Akkharath, I.: Impact of climate change on sediment yield in the
Mekong River basin: a case study of the Nam Ou basin, Lao PDR, Hydrol. Earth
Syst. Sci., 17, 1–20, 10.5194/hess-17-1-2013, 2013.Stahl, K., Hisdal, H., Hannaford, J., Tallaksen, L. M., van Lanen, H. A. J.,
Sauquet, E., Demuth, S., Fendekova, M., and Jódar, J.: Streamflow trends in
Europe: evidence from a dataset of near-natural catchments, Hydrol. Earth Syst.
Sci., 14, 2367–2382, 10.5194/hess-14-2367-2010, 2010.
Summerfield, M. A. and Hulton, N. J.: Natural controls of fluvial denudation
rates in major world drainage basins, J. Geophys. Res., 99, 13871–13883, 1994.Syvitski, J. P. M.: Supply and flux of sediment along hydrological pathways:
research for the 21st century, Global Planet. Change, 39, 1–11, 2003.
Syvitski, J. P. M. and Milliman, J. D.: Geology, geography, and humans battle
for dominance over the delivery of fluvial sediment to the coastal ocean,
J. Geol., 115, 1–19, 2007.
Taabni, M. and El Jihad, M. D.: Eau et changement climatique au Maghreb:
quelles stratégies d'adaptation?, Les Cahiers d'Outre-Mer, 260, 493–518, 2012.
Taibi, S., Meddi, M., Mahé, G., and Assani, A.: Variability of annual
and extreme rainfall over Northern Algeria and relationship with
teleconnections patterns, Proc. of the Mediterranean meeting on “Monitoring, modeling and early warning of
extreme events triggered by heavy rainfalls”, 26–28 June 2004, Univ. of
Calabria, Cosenza, Italy, 2014.
Tebbi, F. Z., Dridi, H., and Morris, G. L.: Optimization of cumulative trapped
sediment curve for an arid zone reservoir: Foum El Kherza (Biskra, Algeria),
Hydrolog. Sci. J., 57, 1368–1377, 2012.
Temme, A. J. A. M. and Veldkamp, A.: Multi-process Late Quaternary landscape
evolution modelling reveals lags in climate response over small spatial
scales, Earth Surf. Proc. Land., 34, 573–589, 2009.
Tecsult International: Etude de la protection des bassins versants de
l'ensemble des barrages en études d'avant projet détaillé (lots
no. 1 à 4) et des bassins versants des barrages en exploitation
(lots no. 5 et 6), Phase I description du milieu, lot no. 5: barrages
Gargar et S. M. Ben-Aouda, Ministère des Ressources en Eau, Algérie, 2004.
Tockner, K. and Stanford, J. A.: Riverine flood plains: present state and
future trends, Environ. Conserv., 29, 308–330, 2002.Touaibia, B.: Problématique de l'érosion et du transport solide en
Algérie septentrionnale, Sécheresse, 21, 1–6, 2010.
Touat, S.: Contrôle de la représentativité de l'échantillon
des transports solides en suspension, Revue Eaux et Sols d'Algérie, 3, 48–54, 1989.Tramblay, Y., El Adlouni, S., and Servat, E.: Trends and variability in extreme
precipitation indices over Maghreb countries, Nat. Hazards Earth Syst. Sci., 13,
3235–3248, 10.5194/nhess-13-3235-2013, 2013.
Tucker, G. E. and Slingerland, R.: Drainage basin responses to climate
change, Water Resour. Res., 33, 2031–2047, 1997.
Vanmaercke, M., Poersen, J., Broeckx, J., and Nyssen, J.: Sediment yield in
Africa, Earth-Sci. Rev., 136, 350–368, 2014.Vinh, V. D., Ouillon, S., Thanh, T. D., and Chu, L. V.: Impact of the Hoa Binh
dam (Vietnam) on water and sediment budgets in the Red River basin and delta,
Hydrol. Earth Syst. Sci., 18, 3987–4005, 10.5194/hess-18-3987-2014, 2014.
Vörösmarty, C. J., Meybeck, M., Fekete, B., Sharma, K., Green, P.,
and Syvitski, J. P. M.: Anthropogenic sediment retention: major global impact
from registered river impoundments, Global Planet. Change, 39, 169–190, 2003.
Walling, D. E.: Assessing the accuracy of suspended sediment rating curves
for a small basin, Water Resour. Res., 13, 531–538, 1977a.
Walling, D. E.: The response of sediment yields to environmental change,
IAHS Publ. 245, in: Human impact on erosion and sedimentation, edited by:
Walling, D. E. and Probst, J. L., IAHS Press, Wallingford, UK, 77–89, 1977b.
Walling, D. E.: Human impact on land-ocean sediment transfer by the world's
rivers, Geomorphology, 79, 192–216, 2006.
Walling, D. E.: The impact of Global Change on Erosion and Sediment Transport
by Rivers: Current progress and Future Challenges, The United Nations World
Water Development Report 3, UNESCO, Paris, p. 26, 2009.
Walling, D. E. and Fang, D.: Recent trends in the suspended sediment loads
of the world's rivers, Global Planet. Change, 39, 111–126, 2003.
Wang, H., Yang, Z., Saito, Y., Liu, J., and Sun, X.: Interannual and
seasonal variation of the Huanghe (Yellow River) water discharge over the
past 50 years: connections to impacts from ENSO events and dams, Global
Planet. Change, 50, 212–225, 2006.
Wang, H. J., Yang, Z. S., Saito, Y., Liu, J. P., Sun, X. X., and Wang, Y.:
Stepwise decreases of the Huanghe (Yellow River) sediment load (1950–2005):
Impacts of climate change and human activities, Global Planet. Change, 57, 331–354, 2007.Ward, P. J., Eisner, S., Flörke, M., Dettinger, M. D., and Kummu, M.: Annual
flood sensitivities to El Niño-Southern Oscillation at the global scale, Hydrol.
Earth Syst. Sci., 18, 47–66, 10.5194/hess-18-47-2014, 2014.
Ziegler, A. D., Sheffield, J., Maurer, E. P., Nijssen, B., Wood, E. F., and
Lettenmaier, D. P.: Detection of intensification in global- and
continental-scale hydrological cycles: Temporal scale of evaluation, J.
Climate, 16, 535–547, 2003.