Acta geographica Slovenica, 62-2, 2022, 23–47
SPATIOTEMPORAL EVOLUTION
OF DROUGHTS AND THEIR
TELECONNECTIONS WITH
LARGE-SCALE CLIMATIC INDICES
IN THE LOWER SEBOU BASIN
IN NORTHWESTERN MOROCCO
Oualid Hakam, Abdennasser Baali, Khalil Azennoud, Touria El Kamel,
Yassine Ait Brahim, Youssra Ahouach
Effects of drought on soils and rainfed crops (cereals).
O
U
A
L
ID
 H
A
K
A
M
62-2_acta49-1.qxd  13.12.2022  12:07  Page 23
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
24
DOI: https://doi.org/10.3986/AGS.10508
UDC: 551.577.38(64)
COBISS: 1.01
Oualid Hakam1, Abdennasser Baali1, Khalil Azennoud1, Touria El Kamel, Yassine Ait Brahim2,
Youssra Ahouach1
Spatiotemporal evolution of droughts and their teleconnections with large-scale
climatic indices in the Lower Sebou Basin in northwestern Morocco
ABSTRACT: The Lower Sebou Basin, placed in a Mediterranean climate, has the particularity of being
exposed to the influence of disturbances from the Atlantic Ocean, making periods of drought and climatic
phenomena variable in space and time. Applying the world’s most recognized drought indices, shows that
the duration, frequency and severity of droughts have increased since the start of the 21st century. These
results revealed and placed in the even wider regional climatic context, including the two dominant atmos-
pheric oscillations such as the North Atlantic Oscillation (NAO) and the Mediterranean Oscillation (MO),
suggest that the significant drought trends determined are correlated with the relative facts of the two oscil-
lations.
KEY WORDS: drought, teleconnection, Lower Sebou Basin, Morocco
Prostorsko–časovni razvoj suš in njihove povezave z obsežnimi podnebnimi indeksi
v porečju Lower Sebou v severozahodnem Maroku
POVZETEK: Kotlina spodnjega Sebouja, kjer je sredozemsko podnebje, ima posebnost, da je izpostavl-
jena vplivom motenj, ki prihajajo iznad Atlantskega oceana. Zaradi tega so obdobja suše in podnebnih
pojavov spremenljiva v prostoru in času. Uporaba uveljavljenih svetovnih indeksov suše kaže, da so se tra-
janje, pogostost in intenzivnost suš od začetka 21. stoletja povečala. Doseženi rezultati, postavljeni tudi
v širši regionalni podnebni kontekst, vključno z dvema prevladujočima atmosferskima nihanjema, kot sta
severnoatlantsko nihanje (NAO) in sredozemsko nihanje (MO), kažejo, da so ugotovljeni pomembni tren-
di suše povezani z relativnimi dejstvi obeh nihanj.
KLJUČNE BESEDE: suša, klimatske povezave na daljavo, kotlina spodnjega Sebouja, Maroko
The article was submitted for publication on December 24th, 2021.
Uredništvo je prejelo prispevek 24. decembra 2021.
1 Sidi Mohamed Ben Abdallah University, Faculty of Sciences Dhar El Mehraz, Laboratory of engineering,
electrochemistry, modeling and environment, Fez, Morocco
oualid.hakam@gmail.com (https://orcid.org/0000-0001-6831-5365), baaliabdennasser@gmail.com,
khalil.azennoud.fst@gmail.com (https://orcid.org/0000-0003-2539-394X), touria.elkamel@gmail.com
(https://orcid.org/0000-0003-4508-2399), youssra.ahouach@gmail.com
2 Mohamed VI Polytechnic University, International Water Research Institue, Ben Guerir, Morocco
aitbrahim.yassine@gmail.com (https://orcid.org/0000-0003-3098-7339)
62-2_acta49-1.qxd  13.12.2022  12:07  Page 24
1 Introduction
Drought is a climatic phenomenon considered to be one of the most loss–making natural disasters in the
world and affects a large number of people each year. An estimated 55 million people globally are affect-
ed by droughts every year (Vatter, Wagnitz and Hernandez 2019; United Nations Office for Disaster Risk
Reduction 2021). It occurs when there is less rainfall in an area for a long period or when there is a poor
distribution of rainfall over some time. The American Meteorological Society has classified droughts into
four types: meteorological drought, hydrological drought, agricultural drought and socio–economic drought
(Yihdego, Vaheddoost and Al-Weshah 2019). They may not occur simultaneously, but meteorological drought
remains the driving force behind the others (Wilhite 2006). This drought can take the form of permanent
droughts, like what is occurring in arid and semi–arid or Mediterranean climates during the summer sea-
son or in subtropical regions in winter. As well as it may turn out to be a natural disaster when it becomes
high frequency with longer duration. Therefore, the assessment and characterization of meteorological
drought are most important to understand these risks. However, the assessment of drought characteris-
tics and its spatial variability is still very difficult (Vicente-Serrano et al. 2020). For this purpose, many
indices have been developed to identify and quantitatively characterize drought events, in order to mon-
itor and predict the onset, termination and impacts of drought (Bayissa et al. 2018). On the other hand, no
single index can represent all aspects of meteorological drought (Yihdego, Vaheddoost and Al-Weshah 2019).
To overcome this constraint, a multi–index approach was adopted for the characterization of drought in
the Lower Sebou Basin (LSB). As a result, two drought indices can be used as a basis for characterizing
meteorological drought in the LSB, the Standardized Precipitation Index (SPI) (McKee, Doesken and Kleist
1993) and the Standardized Precipitation–Evapotranspiration Index (SPEI) (Vicente-Serrano, Beguería
and López-Moreno 2010). Widely used in the world (Vicente-Serrano et al. 2012; Beguería et al. 2014; Šebenik,
Brilly and Šraj 2017) and recommended by the World Meteorological Organization (World Meteorological
Organization 2012), these indices have been recently applied in the LSB due to their performance and region-
al applicability compared to other drought indices (Hakam et al. 2022a). The SPI and SPEI approaches
are similar, but there are obvious differences between the input parameters for the calculation. The SPI
calculation requires rainfall data, and is well suited in time and space, but the SPEI uses the difference between
rainfall and potential evapotranspiration (PET), which can account for the possible effects of tempera-
ture variability and temperature extremes beyond the context of global warming.
In Morocco, as the Mediterranean semi–arid areas, drought is one of the most frequent and widespread
natural disasters and has become a structural element of the country’s climate in recent decades, as shown
by Stigter, Nunes and Pisani (2014) for case studies in south Portugal, Spain and Morocco. According to a recent
study by Woillez (2019), rainfall in Morocco has generally decreased significantly over the period 1951–2010,
between -10 and -25 mm per year in the north of the country, with temperature increases in the order of
+1 °C to +2 °C on average annually between 1901 and 2012. Over the last 30 years, the trend is clearly high-
er than the global average, with + 0.42 °C per decade on average since 1990, compared to + 0.28 °C per decade
on average on all continents (Woillez 2019), resulting in an increase in the frequency and intensity of extreme
events such as drought and flooding. The LSB was not spared in recent decades and experienced heatwaves
in 1995 and 2001 and floods in 1996, 2002 and 2009 (Agence du Bassin Hydraulique de Sebou 2010).
The spatio-temporal evolution of droughts in the LSB is controlled by landscape heterogeneity and
specific climatic events such as extreme weather anomalies and large-scale atmospheric circulations (Lamb
and Peppler 1987; Zamrane, Mahé and Laftouhi 2021; Hakam et al. 2022b). Some studies have demon-
strated an anti–correlation of rainfall in Morocco, especially in the northwestern part, and the North Atlantic
Oscillation (NAO) (Marchane et al. 2016; Driouech et al. 2021). Knowing that the calculation of this oscil-
lation, which makes its effects felt in winter, is based on the difference in atmospheric pressure between
the Icelandic low and the Azores high, neglecting the Mediterranean Oscillation (MO) despite the prox-
imity of the Mediterranean Sea. Indeed, it has been shown that the MO index influences the seasonal
variability of rainfall in the Mediterranean basin, especially in winter (Criado-Aldeanueva and Soto-Navarro
2020). Currently, drought studies in the LSB are scarce, most of them being based on simple statistical
approaches (El Jihad, Peyrusaubes and El Bouzidi 2014; Acharki et al. 2019).
More vulnerable than other basins, facing numerous droughts (Ajlal and Agoutime 2021; Hakam et
al. 2022a) and subject to a drought warning by the Regional Observatory of the Environment and Sustainable
Development in Morocco (2014), the climatic study of the LSB has become paramount. As no regional
Acta geographica Slovenica, 62-2, 2022
25
62-2_acta49-1.qxd  13.12.2022  12:07  Page 25
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
26
60° E30° E
0° 
60
° N
30
° N
30
° N
0°
 
0°
 
60
° E
30
° E
0°
 
30
° E
|
|
|
|
Mechra
Bel Ksiri
Zirara
Mnasra
Lalla
Mimouna
4° W
4° W
5° W
5° W
6° W
6° W
35
° N
35
° N
34
° N
34
° N
33
° N
33
° N
Legend
| Meteorological station
Hydrographic network
Sebou watershed
Lower Sebou Basin
Elevation (m)
0–50
50–100
100–200
200–300
300–600 ±
Content by: Hakam Oualid
Map by: Hakam Oualid]
Source: Faculty of Sciences Dhar El Mehraz
© 2021, (FSDM) Faculty of Sciences Dhar El Mehraz
0 50 10025
kmMedi
terra
nean
 Sea
Atlan
tic 
Ocea
n
Mo
roc
co
Figure 1: Geographical location of the LSB and location of the typical meteorological stations.
Figure 2: Main climatic teleconnections dominating in the LSB. p
62-2_acta49-1.qxd  13.12.2022  12:07  Page 26
Acta geographica Slovenica, 62-2, 2022
27
#
#
"
"
90
° E
60
° E
30
° E
30
° E
0°
 
0°
 
30
° W
30
° W
60
° W
90
° W
30° N
30° N
M
or
oc
co
N
o
rt
h
 A
tl
a
n
ti
c 
O
ce
a
n
M
ed
it
er
ra
n
ea
n
 S
ea
C
on
te
nt
 b
y:
 H
ak
am
 O
ua
lid
M
ap
 b
y:
 H
ak
am
 O
ua
lid
So
ur
ce
: F
ac
ul
ty
 o
f S
ci
en
ce
s D
ha
r E
l M
eh
ra
z
©
 2
02
1,
 (F
SD
M
) F
ac
ul
ty
 o
f S
ci
en
ce
s D
ha
r E
l M
eh
ra
z
Le
ge
nd S
eb
ou
 w
at
er
sh
ed
Lo
w
er
 S
eb
ou
 B
as
in
Te
le
co
nn
ec
tio
ns
#
N
A
O
"
M
O
±
0
50
0
1 
00
0
1 
50
0
2 
00
0
25
0
km
62-2_acta49-1.qxd  13.12.2022  12:07  Page 27
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
study has targeted the meteorological characteristics of the drought and its relationship with large-scale
atmospheric dynamics, such a study in the LSB as a large agricultural plain would help to put into perspective
the investments under the Morocco-Green Plan strategy (38.6 billion dirhams) and the threatened infra-
structure (Ministry of Agriculture and Maritime Fisheries 2020). This strategy aims to develop a modern
and efficient agriculture that meets the requirements of national and international markets (22% of Morocco’s
total exports, i.e. approximately 5.57 billion dollars) (Harbouze et al. 2019). This study also opens doors
for the development of drought early warning systems in North Africa as the dissemination of this data
will allow proactive decisions on water use and agricultural practices.
2 Data used and methodology
2.1 Study area
The Sebou watershed, downstream of which the LSB is located in the North West of Morocco and covers
an area of 40,000 km2. It flows from the Middle Atlas Mountain and ranges over a length of 500 km to its
outlet in the Atlantic Ocean (Figures 1, 2). Indeed, the LSB takes benefits from a sub–humid Mediterranean
climate with an oceanic influence, particularly in its western part. Rainfall exceeds 500 mm per year in
most of the basin and average temperatures oscillate around 19 °C (Table 1). Consequently, the abundance
of rainfall and the importance of the basin’s water and biophysical resources expose the basin to overex-
ploitation of resources and vulnerability to the effects of drought in the future.
2.2 Data used
Meteorological data provided by the Regional Office for Agricultural Development (ROAD) from the four
weather stations cover the period from 1984 to 2016. The distribution and characteristics of these stations
are presented in Figure 1 and Table 1.
Two important large-scale climate anomalies were selected to analyse the main drivers of drought in the
LSB: NAO (North Atlantic Oscillation) and MO (Mediterranean Oscillation) (Zamrane, Mahé and Laftouhi
2021). The NAO is called a meridional dipole of atmospheric pressure with centres of action near the Azores
(37 °N, -25 °E) and Iceland (65 °N, -18 °E) (Figure 2). The Mediterranean Oscillation (MO) is defined as
a dipolar behaviour of the atmosphere between the East–West sub–basins of the Mediterranean. This index
measuring the intensity of the dipole was defined as the normalized 500 hPa height anomalies between Alger
(36.4 °N, 3.1 °E) and Cairo (30.1 °N, 31.4 °E) (Figure 2) (Conte, Giuffrida and Tedesco 1989). The monthly
NAO and MO data were extracted from the National Oceanic and Atmospheric Administration (NOAA).
2.3 Drought indices (DIs)
Drought indices (DIs) help converting the indicator variables such as temperature and rainfall into a numer-
ical value to represent the extent of the water deficit in a way that is easy to understand. DIs used in this
study, namely the Standardized Precipitation Index and the Standardized Precipitation–Evapotranspiration
Index mainly characterize meteorological drought (Table 2).
28
Table 1: Characteristics of the meteorological stations of the LSB.
Stations rainfall Average PET Geographical coordinates
(mm per year) temperature (mm per year)
(°C per year) Long (°W) Lat (°N) Z (m)
Lalla Mimouna 574.7 17.15 1157.5 6.11 34.85 15
Mechra Bel Ksiri 515 19.35 1441.4 5.96 34.57 24
Mnasra 567 16.65 1239.5 6.48 34.46 10
Zirara 387 19.05 1480 5.74 34.24 55
Note. PET: potential evapotranspiration calculated based on the Thornthwaite equation; Long: longitude; Lat: latitude; Z: elevation in meter.
62-2_acta49-1.qxd  13.12.2022  12:07  Page 28
Acta geographica Slovenica, 62-2, 2022
29
Standardized Precipitation Index (SPI) is based on historical rainfall records at a given location to cal-
culate the probability of rainfall at any time-scale between 1 and 48 months.
(1)
With Pi: the rainfall of month or year i; Pm: the average rainfall of the month or year i in the entire peri-
od; S: the standard deviation of the series over the considered time-scale. Data were normalized using the
gamma distribution method to overcome the constraint of rainfall that is not normally distributed over
the year (Stagge et al. 2015).
Standardized Precipitation–Evapotranspiration Index (SPEI) is an index developed by Vicente-
Serrano, Beguería and López-Moreno (2010). The procedure for calculating the SPEI is also similar to that
of the SPI, except that it uses the difference between rainfall and potential evapotranspiration (PET) cal-
culated according to the method of Thornthwaite (1984). In the SPEI calculation, we need the monthly
water balance (Di), which is based on the difference between rainfall and potential evapotranspiration (PET).
This represents a simple climatic water balance calculated at different time-scales (1-, 3-, 6- and 12-months).
The difference between accumulated monthly rainfall and potential evapotranspiration was calculat-
ed as follows:
(2)
Subsequently, the monthly Di values are aggregated over different time-scales as is the case for rainfall in
SPI calculations. A three–parameter Log–logistic distribution of Di values was used to fit the data series
of accumulated monthly difference values (Vicente-Serrano and Beguería 2016). Therefore, negative SPEI
values indicate a dry condition due to lack of rainfall and/or higher PET compared to the historical aver-
age recorded. 
The Di values were summarized on different time-scales:
(3)
where k is the monthly time-scale and n is the number of calculations. Di values are aggregated depend-
ing on the time-scale before being standardized using the Log–logistic distribution with the following
probability density function (Vicente-Serrano, Beguería and López-Moreno 2010; Pyarali et al. 2022):
(4)
where α, β, and γ are scale, shape, and origin parameters, respectively. The probability distribution func-
tion of Di according to the Log–logistic distribution is then given by:
(5)
Table 2: Characteristics of drought indices (DIs).
Drought Name Variables Probability distribution Reference
indices (DIs)
SPI Standardized P Gamma McKee, Doesken and Kleist (1993)
Precipitation Index
SPEI Standardized Precipitation– P and PET Log–Logistic (3 parameters) Vicente-Serrano, Beguería
Evapotranspiration Index and López-Moreno (2010)
Note. P: Monthly–accumulated precipitation; PET: Potential Evapotranspiration.
62-2_acta49-1.qxd  13.12.2022  12:07  Page 29
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
The SPEI can easily be obtained as the standardized values of F(x). For example, following the classical
approximation of Abramowitz and Stegun (1965):
(6)
Where
(7)
p is the probability of exceeding a determined Di value, p= 1−F(x). If p > 0.5, then p is replaced by 1 − p
and the sign of the resultant SPEI is reversed. The category of SPI and SPEI varies from ≥ 2.0 (extremely
wet) to ≤ -2.0 (extremely dry) (Table 3).
30
Table 3: Wet/Dry classification of SPI and SPEI. The colour scale indicates the drought classes in figure 6.
SPI/SPEI Values Class
≥ 2 Extremely wet
1.5 to 1.99 Very wet
1 to 1.49 Moderately wet
–0.99 to 0.99 Near normal
–1 to –1.49 Moderately dry
–1.5 to –1.99 Severely dry
≤ –2 Extremely dry
2.4 Identification and characterization of drought events
The »runs« theory proposed by Yevjevich (1967) has been applied to identify and characterize meteoro-
logical drought events based on DIs. A drought event is defined as a consecutive sequence of months (t)
with a drought index value (Xt) below a chosen threshold (X0); hence each drought event is characterized
by the following parameters (Figure 3):
• drought duration (D) is defined as the time interval between the beginning and the end of a drought
event;
• drought intensity (Is) is defined as the number of months during which the DIs values are lower than –1;
• drought severity (S) is defined as the sum of the monthly DIs values when they are lower than –1 (X0)
during the period considered;
• peak intensity indicates the lowest DIs value during a drought event.
2.5 Trend analysis (Mann–Kendall test)
The Mann Kendall (M–K) trend test helps to find the presence of a monotonic trend in a time series of
hydro–climatic variables (Acharki et al. 2019). In the case of the LSB, the M–K test is applied to analyse
the temporal characteristics of the SPI and SPEI indices.
The test is based on the (S) statistic defined as follows:
(8)
Where S is the number of positive differences minus the number of negative differences, N is the num-
ber of data points, xj and xk are monthly values of months j and k with (j > k).
62-2_acta49-1.qxd  13.12.2022  12:07  Page 30
Acta geographica Slovenica, 62-2, 2022
31
1
–1
0
2
–2
3
–3
4
–4
Time (Months)
Peak
Drought Duration (D)
SP
I/
SP
E
I
Extreme Drought
Severe Drought
Moderate Drought
Moderately Wet
Severely Wet
Extremely Wet
Peak intensity (P)
Drought severity (S)
= P
= Sum of all SPI/SPEI values
in duration when they are lower than -1
Figure 3: Drought characteristics for DIs following the »Run« theory.
Noting 
(9)
When N > 10, the distribution of S is assumed to be normally distributed with mean E (S) = 0 and variance
(var(S)) given by:
(10)
where N is the number of tied (zero difference between compared values) groups and tk the number of
data points in the kth tied group. Then the values of S and var(S) are used to calculate the standardized
statistical test Z under the following formula:
(11)
A positive (negative) value of Z indicates an ascending (descending) trend and its significance is com-
pared to the critical value α or significance level of the test. Absolute values of Z ≥ 1.65, 1.96 and 2.58 are
adopted, respectively, indicating significant levels of (α) 0.1, 0.05 and 0.01.
In order to find a monotonic trend in the time series used, it is necessary to identify the period of the
beginning of the trends. Therefore, the Sequential Mann–Kendall (SQMK) test statistic is particularly use-
ful for such change detection analysis (Gerstengarbe and Werner 1999). The SQMK test procedure is different
from the Z test and consists of the following steps:
• Values of the series x are replaced by their ranks ri, ranked in ascending order;
• Quantities ri (i = 1, 2, …, n) are compared with rj (j = 1, 2, …, i - 1);
62-2_acta49-1.qxd  13.12.2022  12:07  Page 31
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
• Statistic ti is defined as follows:
(12)
• Variance of the statistic var (ti) and the mean E (ti) of the test is calculated as follows:
(13)
and
(14)
• Values of u(ti) statistic can then be calculated as:
(15)
While the forward sequential statistic, u(ti) is estimated using the original time series x1, x2 … xn, values
of the backward sequential statistic, u’(ti) are estimated in the same manner but starting from end of the
series. In estimating u’(ti) the time series is resorted so that last value of the original time series comes first
xn, xn-1…1. The sequential version of Mann–Kendall test statistic allows detection of approximate begin-
ning of a developing trend. When u(ti) and u’(ti) curves are plotted, the intersection of the curves u(ti) and
u’(ti) locates approximate potential trend turning point. If the intersection of u(ti) and u’(ti) occur within
± 1.96 (95% confidence level) of the standardized statistic Z, a detectable change at that point in the time
series can be inferred. Moreover, if at least one value of the reduced variable is greater than a chosen level
of significance of Gaussian distribution the null hypothesis (H0: Sample under investigation shows no begin-
ning of a new trend) is rejected.
2.6 Drought events and large-scale climate anomalies
To search for possible relationships, wavelet coherence analysis was applied to DIs (SPI and SPEI) and large-
scale climate anomaly indices, such as NAO and MO. Wavelet coherence is an approach used for analysing
the degree of coherence of cross wavelet transform in time-frequency space. It has been widely used to analyse
the periodicity of drought and the relationships between hydrological variability and possible teleconnections
(Zamrane, Mahé and Laftouhi 2021).
Following Torrence and Webster (1999), the wavelet coherence coefficient can be defined as follows
(Chang et al. 2019):
(16)
where R2(𝛼, 𝜏) takes values between 0 (no coherency) and 1 (perfect coherency), α is the scale expansion
parameter, τ is the dimensionless time–shift parameter, Wxy(𝛼, 𝜏) is the cross wavelet transform of the two
time series, Wx and Wy are the sums of ranks of observations in samples 𝑥𝑡 and 𝑦𝑡, respectively, and S rep-
resents a smoothing operator, which is defined as:
(17)
where Sscale and Stime represent smoothing along the wavelet scale axis and in time, respectively.
32
62-2_acta49-1.qxd  13.12.2022  12:07  Page 32
3 Results
3.1 Temporal evolution of the SPI and SPEI
The monthly SPI and SPEI values that have been calculated on four time scales (1-, 3-, 6- and 12-months)
can be used for the monitoring and evaluation of different types of droughts and their full application can
allow a comprehensive detection and evaluation of droughts and floods. They were first calculated using
the two monthly climate data observed from four weather stations over the period from 1984 to 2016
(Figure 4). Then, the time series of SPI and SPEI were averaged over the four stations to characterize the
dry or wet conditions in the LSB. Figure 4 shows the process of SPI and SPEI changes at different time-
scales (1-, 3-, 6- and 12-months) during the period 1984–2016. Both indices indicated an extreme severity
of drought after the beginning of the 21st century where the frequency and intensity of drought periods
are increased. Noting that longer the time-scale, the more obvious the severity and duration of the drought.
Typical drought periods or years occurred in 1992–1995, 1999–2002, 2005–2006, 2012 and 2015–2016 by
applying SPI–12 and SPEI–12. However, the indices fluctuated frequently, with a wide range, on time scales
of 1- and 3-months, which suggests the impact of short–term climate change. In addition, the range of
positive and negative fluctuations of SPI was larger than that of SPEI. However, on longer time-scales (6-
and 12-months), SPI and SPEI values had a smooth trend, the volatility has decreased and the character-
istics of inter-annual and inter-decadal changes were evident, hence the duration of the drought increased.
The temporal trend of SPI and SPEI showed a relatively stable evolution and a smaller amplitude differ-
ence at different time-scales, but there are slight differences in the fluctuation value and continuity, explaining
different drought intensities and frequencies. The difference between SPI and SPEI decreased with the increase
of the time-scale, where the values fluctuated between -4 and 4 at short time-scales (1-month) and between
–1 and 1 at long time-scales (12-months). Furthermore, it should also be noted that this difference has
increased in recent years (Figure 5).
3.2 Monthly variations of the SPI and the SPEI
The SPI and SPEI monthly distribution values were quite clear at different time-scales, which can reflect
the variation of dry and wet conditions for each month in the LSB (Figure 6). It was clear that the dura-
tion and intensity of drought in some months has increased considerably and the effect of time-scales on
drought duration and intensity was evident, particularly after the beginning of the 21st century. It is also
observed that the droughts reflected by SPI and SPEI were slightly different in each month and at differ-
ent time-scales, especially on the 1– and 3-months scales. Thus Figure 6 shows that the distribution of drought
expressed by the SPI was random on short scales (1-month and 3-months), especially in the winter sea-
son. The dry months reflected by SPEI–1 and SPEI–3 were concentrated in the winter months before 2004,
and after this year the dry months start to shift towards the summer and into the end of the year (December).
Furthermore, droughts reflected by SPEI were more frequent on short time-scales, especially 1-month,
while droughts reflected by SPI were common on 6-months and 12-months scales; which reflects the effect
of the evapotranspiration parameter included in the SPEI calculation on the variation in drought condi-
tions in the LSB.
3.3 Temporal variability
The evolution of SPEI and SPI values and the annual and seasonal trends, as well as the graphical results
of the SQMK test identified the year when the turning point and the magnitude of the change occurred
in the LSB. Knowing that the identified turning point and the magnitude of change year for each period
represent the beginning of a negative or positive trend (Figure 7).
Acta geographica Slovenica, 62-2, 2022
33
Figure 4: Temporal evolution of SPI and SPEI at time-scales of 1-, 3-, 6- and 12-months in the LSB during 1984–2016. p p. 34
Figure 5: Difference (D) between the SPI and SPEI indices in the different time-scales (1-, 3-, 6- and 12-months) in LSB during 1984–2016. p p. 35
Figure 6: Heat monthly variation in the SPI and SPEI at 1-, 3-, 6- and 12-months’ time-scales during 1984–2016 in the LSB, colour legend indications
are obtained in Table 3. p p. 36
62-2_acta49-1.qxd  13.12.2022  12:07  Page 33
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
34
Year
–5
–4
–3
–2
–1
0
1
2
3
4
S
P
I/
S
P
E
I
SPI – 3SPEI – 3
Year
–3
–2
–1
0
1
2
3
S
P
I/
S
P
E
I
SPI – 6SPEI – 6
Year
–3
–2
–1
0
1
2
3
S
P
I/
S
P
E
I
SPI – 12SPEI – 12
–5
–4
–3
–2
–1
0
1
2
3
4
S
P
I/
S
P
E
I
Year
SPI – 1SPEI – 1
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
62-2_acta49-1.qxd  13.12.2022  12:07  Page 34
Acta geographica Slovenica, 62-2, 2022
35
Year
–4
–3
–2
–1
0
1
2
3
4
D
–
3
Year
–3
–2
–1
0
1
2
3
D
–
6
Year
–2
–1
0
1
2
D
–
1
2
–4
–3
–2
–1
0
1
2
3
4
5
D
–
1
Year
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
62-2_acta49-1.qxd  13.12.2022  12:07  Page 35
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
36
1
1984 1992 2000 2008 2016
2
3
4
5
6
7
8
9
10
11
12
SPI–1
M
o
n
th
s
1
1984 1992 2000 2008 2016
2
3
4
5
6
7
8
9
10
11
12
SPI–3
M
o
n
th
s
1
1984 1992 2000 2008 2016
2
3
4
5
6
7
8
9
10
11
12
SPI–6
M
o
n
th
s
1
1984 1992 2000 2008 2016
2
3
4
5
6
7
8
9
10
11
12
SPI–12
M
o
n
th
s
1
1984 1992 2000 2008 2016
2
3
4
5
6
7
8
9
10
11
12
SPEI–1
M
o
n
th
s
Years
Years
Years
Years
Years
Years
Years
Years
1
1984 1992 2000 2008 2016
2
3
4
5
6
7
8
9
10
11
12
SPEI–3
M
o
n
th
s
1
1984 1992 2000 2008 2016
2
3
4
5
6
7
8
9
10
11
12
SPEI–6
M
o
n
th
s
1
1984 1992 2000 2008 2016
2
3
4
5
6
7
8
9
10
11
12
SPEI–12
M
o
n
th
s
62-2_acta49-1.qxd  13.12.2022  12:08  Page 36
Acta geographica Slovenica, 62-2, 2022
37
The annual and seasonal (winter, spring, summer and autumn) ZSPEI values were –0.85, –0.50, –0.26, –3.51
and 0.54 respectively, and showed a decreasing trend, except for the autumn season (ZSPEI = 0.54). While the
annual and seasonal ZSPI values were 0.39, –0.60, –0.23, –2.25 and 1.07 respectively, in contrast to SPEI, ZSPI–12
showed an increasing trend (Figure 7). According to the SPI and SPEI drought indices, extreme drought events
on an annual scale were observed in 2006 and 2016 (Figures 7). For seasonal droughts, winter was the most
affected by drought followed by spring, however the SPEI considered autumn as the least drought–prone sea-
son. It is worth noting that u(t) and u’(t) curves of SPI at the annual and autumn scales were in very good
agreement with each other, which might suggest that annual rainfall is mainly impacted by autumn rainfall
(Figure 7). In general, the SPI and SPEI series showed high variability, with no significant trends over the
study period, except for the summer when the most severe droughts were recorded only from the year 2000
onwards. As shown in Figure 7, the downward trends in summer were highly significant from 2004 to 2016
at 95% significance level, hence the intersection of the u(t) and u’(t) curves of SPI–3 and SPEI–3 during the
summer season occurred in 1998 and 2003, respectively, and exceeded the confidence limit in 2004 (Fig. 7).
This is clearly indicated in the monthly variation of SPEI–3 and SPI–3 in Figure 6.
3.4 Spatial and inter-annual variability
At the annual scale, the spatial distribution of the drought duration trends of SPI and SPEI was interpo-
lated by the Inverse Distance Weight interpolation method (IDW method) (Figure 8a). The drought duration
reflected by SPI–12 and SPEI–12 was identical with an opposite spatial distribution. In terms of trends,
all stations showed non–significant increasing trends; however, the SPI–12 suggested that the two stations
(Lalla Mimouna and Mnasra) showed a decreasing trend (tendency to be wet). Whereas, the other sta-
tions showed an increasing trend, mainly located in the interior of LSB. The duration of droughts in the
LSB increased from West to East. The SPEI–12 showed that the duration of drought is longer in the west-
ern regions and shorter in the north-eastern part of the basin.
The spatial distribution of drought severity trends in the LSB was different from that of drought dura-
tion, where the Mnasra station recorded the longest period and the least severe at the same time (Figure 8b).
However, the SPI–12 and SPEI–12 indices showed a similar overall behaviour of the drought severity dis-
tribution except for the Zirara station. Droughts were more frequent in the Northeast of the basin, especially
in the Mechra Bel Ksiri station. In terms of trend, all stations showed increasing trends, only Lalla Mimouna
and Mechra Bel Ksiri stations recorded a significant increasing trend for SPEI and SPI respectively, main-
ly in the North West and Northeast of the LSB. In general, droughts occur within the basin showing significant
upward trends.
3.5 Correlations between SPI and SPEI
Pearson correlation plots between SPI and SPEI series were made at different time-scales for each station
(Figure 9). The relationships between this DIs at different time-scales had relatively high correlation coef-
ficients (r). However, there was a marked difference in the correlation coefficient (r) of each station at different
time-scales, especially at the 1-month time-scale. On the short time-scales (1- and 3-months), the corre-
lation coefficients (r) between SPI and SPEI were relatively low in the summer period (r < 0.5). As the
time-scale increased, the low correlation coefficients shifted towards the autumn season on a 6-months’ time-
scale. On an annual scale of 12-months, in all stations there were very strong correlations (0.7 < r < 0.95)
except the Zirara station (0.4 < r < 0.6). The correlation between SPI and SPEI in the western part of the
LSB was generally higher than in the eastern part (Zirara station) since SPI neglected the temperature vari-
ation in the calculation and PET is high at this station (Table 1).
Therefore, the droughts identified by SPI and SPEI may be different in the East of the LSB (Figure 9).
Figure 7: Seasonal and annual variations of M–K trend test and abrupt changes in SPEI (left) and SPI (right) in the LSB during 1984–2016. p p. 38–39
Figure 8: Distribution of trends in drought duration (a) and drought severity (b) for SPEI (left) and SPI (right). Significant positive trend (Z > 1.96),
insignificant trend (–1.95 < Z > 1.96) and significant negative trend (Z < –1.95). p p. 40
Figure 9: The correlations between the SPI and SPEI of typical meteorological stations in the LSB at scales of 1-, 3-, 6- and 12-months. Correlations
significant at the 95% confidence level are indicated by triangles. p p. 41
62-2_acta49-1.qxd  13.12.2022  12:08  Page 37
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
38
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
Winter (DJF) Z 0.50= –
–3
–2
–1
0
1
2
3
S
P
E
I –
3
SPEI U(t) U'(t)Confidence level 95 %
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
–3
–2
–1
0
1
2
3
S
P
E
I –
3
Spring MAM = – 26( ) Z 0.
Annual = –0.85Z
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
–3
–2
–1
0
1
2
3
S
P
E
I –
12
Autumn SON = 0.54( ) Z
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
–3
–2
–1
0
1
2
3
S
P
E
I –
3
Summer JJA = –3 51( ) Z .
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
–3
–2
–1
0
1
2
3
S
P
E
I –
3
62-2_acta49-1.qxd  13.12.2022  12:08  Page 38
Acta geographica Slovenica, 62-2, 2022
39
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
Winter (DJF) Z 0.60= –
–3
–2
–1
0
1
2
3
S
P
E
–
3
SPEI U(t) U'(t)Confidence level 95 %
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
–3
–2
–1
0
1
2
3
S
P
E
–
3
Spring MAM = – 2( ) Z 0. 3
Annual =Z 0.39
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
–3
–2
–1
0
1
2
3
S
P
E
–
12
Autumn SON = .( ) Z 1 07
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
–3
–2
–1
0
1
2
3
S
P
E
–
3
Summer JJA = –( ) Z 2.25
–3
–2
–1
0
1
2
3
1984 1989 1994 1999 2004 2009 2014
M
–
K
 s
t a
t i
s t
ic
s
Year
–3
–2
–1
0
1
2
3
S
P
E
–
3
62-2_acta49-1.qxd  13.12.2022  12:08  Page 39
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
40
Content by: Hakam Oualid
Map by: Hakam Oualid
Source: Faculty of Sciences Dhar El Mehraz
© 2021, (FSDM) Faculty of Dhar El MehrazSciences
–1.96 to 0
1.96 to 2.96
–2.96 to –1.96
0.01 to 1.95
Z value
Legend
a)
Atlantic
Ocean
3
° 
N
4
35
° 
N
6 W°
6 W°
Atlantic
Ocean
3
° 
N
4
35
° 
N
6 W°
6 W°
Atlantic
Ocean
Atlantic
Ocean
b)
3
° 
N
4
35
° 
N
6 W°
6 W°
3
° 
N
4
35
° 
N
6 W°
6 W°
62-2_acta49-1.qxd  13.12.2022  12:08  Page 40
Acta geographica Slovenica, 62-2, 2022
41
0
.0
0
.2
0
.4
0
.60
.8
1
.0
Jan.
Feb.
Mar.
Apr.
May.
Jun.
Jul.
Aug.
Sep.
Oct.
Nov.
Dec.
M
o
n
th
s
L
al
la
 M
im
o
u
n
a
M
ec
h
ra
 B
el
 K
si
ri
M
n
as
ra
Z
ir
ar
a
Correlation coefficient (r)
1
–
m
o
n
th
 t
im
e
sc
al
e
–
0
.0
0
.2
0
.4
0
.60
.8
1
.0
Jan.
Feb.
Mar.
Apr.
May.
Jun.
Jul.
Aug.
Sep.
Oct.
Nov.
Dec.
M
o
n
th
s
Correlation coefficient (r)
3
–
m
o
n
th
 t
im
e
sc
al
e
–
0
.0
0
.2
0
.4
0
.60
.8
1
.0
Jan.
Feb.
Mar.
Apr.
May.
Jun.
Jul.
Aug.
Sep.
Oct.
Nov.
Dec.
M
o
n
th
s
Correlation coefficient (r)
6
–
m
o
n
th
 t
im
e
sc
al
e
–
0
.0
0
.2
0
.4
0
.60
.8
1
.0
Jan.
Feb.
Mar.
Apr.
May.
Jun.
Jul.
Aug.
Sep.
Oct.
Nov.
Dec.
M
o
n
th
s
Correlation coefficient (r)
1
2
–
m
o
n
th
 t
im
e
sc
al
e
–
62-2_acta49-1.qxd  13.12.2022  12:08  Page 41
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
3.6 Coherence between drought indices (DIs) and large-scale climate indices
Wavelet coherence (WC) identifies both the frequency bands and time intervals of co–variations between
SPI/SPEI and large-scale climate indices (i.e. NAO and MO) in the LSB (Figures 10, 11). The WC, which
varies in value from 0 to 1, measures the cross–correlation between the drought and climate indices as
42
Years
Y
e
a
rs
0.5
1
2
4
8
16
32
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
b) Years
Y
e
a
rs
0.5
1
2
4
1
0.7
0.5
0.3
0.1
0
8
16
32
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
a)
Figure 10: Wavelet coherence spectrum between the large-scale climate indices (NAO) and the SPEI (a) and SPI (b) series of the LSB. The colours from
blue to red indicate increasing coherence. The 95% significance level with respect to the red noise is represented by a white outline.
62-2_acta49-1.qxd  13.12.2022  12:08  Page 42
a function of frequency. The coloured shading represents the magnitude of the coherence, as indicated in
the colour bar, indicating the time-scale variability in the correlation between the two-time series. The
white contours represent the significant sections that have a significance level of 95%. Arrows represent
consistency (right–oriented: positive correlation, left–oriented: negative correlation). As shown in Figures
10 and 11, for a shorter periodicity of less than one year, significant coherence between the DIs (SPI and
Acta geographica Slovenica, 62-2, 2022
43
Years
Y
e
a
rs
0.5
1
2
4
8
16
32
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
b) Years
Y
e
a
rs
0.5
1
2
4
1
0.7
0.5
0.3
0.1
0
8
16
32
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
2
0
1
0
2
0
1
1
2
0
1
2
2
0
1
3
2
0
1
4
2
0
1
5
2
0
1
6
a)
Figure 11: Wavelet coherence spectrum between the large-scale climate indices (MO) and the SPEI (a) and SPI (b) series of the LSB. The colours from
blue to red indicate increasing coherence. The 95% significance level with respect to the red noise is represented by a white outline.
62-2_acta49-1.qxd  13.12.2022  12:08  Page 43
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
SPEI) and the climate indices (NAO and MO) was observed intermittently from year to year. Strong neg-
ative consistency is detected around 8–16 years (multi–annual) throughout the period considered
(1984–2016). The NAO had the first–order relationship with drought variation over an 8–16 years’ peri-
od, appearing since 1990 for SPI and during the whole period considered (1984–2016) for SPEI.
Considerable energy is observed during a short period in the mid–1990s and from 2006 to 2015 for the
2–4 years’ group (Figure 10). In the case of the MO, the coherence with the DIs (SPI and SPEI) was almost
similar to that of the NAO but with low energy. The same energies are detected during the same periods
of NAO (Figure 11). NAO has mostly strong negative coherence, while in MO the coherence is partly pos-
itive and partly negative.
4 Discussion and conclusion
The results reveal some differences between DIs in drought monitoring, found in most studies (Liu et al.
2021; Hakam et al. 2022a), which are due to the variation in climatic conditions at the different LSB sta-
tions and especially at the Zirara station (Figure 9). The fact that the calculation of DIs is based on rainfall
and that the variability of rainfall is much higher than that of PET, considered stationary (without tem-
poral trend), the importance of the latter would be negligible and drought conditions are almost controlled
by the temporal variability of rainfall. However, some authors have cautioned against systematically neglect-
ing the importance of the effect of temperature on drought conditions, and numerous empirical studies
have shown that increasing temperature significantly affects drought severity (McGuire et al. 2010; Linares
and Camarero 2012). Due to the consideration of temperature by SPEI, the evolution of annual drought
showed a decreasing trend, while the drought reflected by SPI–12 showed an increasing trend, so the per-
formance of SPI might be insufficient in reinforcing the conclusions on the warming trend of the climate
in Morocco (Figure 7).
At the seasonal scale, both SPI–3 and SPEI–3 showed decreasing trends, except for autumn where SPI–3
recorded a positive value (ZSPI–3 = 1.07). This indicates that the fall season would be marked by a slight
increase in rainfall and a shift in the wettest season of the year, as rainfall is relatively more abundant in
the fall and less abundant in the winter (ZSPI–3 = –0.60). Consequently, the autumn drought that started
in September was increasingly less severe in the LSB and when rains started early with high intensity, a win-
ter drought would be expected (Figure 6). These results further confirm the findings of Acharki et al. (2019)
and Diani et al. (2019). In addition, SPI–3 and SPEI–3 showed that the summer season was the most
drought–prone with a significant downward trend at the 0.05 level, triggered in 2004 (Figures 6, 7). This
triggering of dry months can explain the increase in average temperatures from the year 2004 (Driouech
2010). Thus, the intensification of drought episodes in summer was also related to the nature of the winds
that prevail in this season, namely the warm continental winds from the southeast (Chergui and Sirocco).
These winds led to high evapotranspiration marked by the trend of SPEI–3 to the lowest value of
Z (ZSPEI–3 = –3.52) and the small amounts of rainfall received are mainly of stormy origin (ZSPEI–3 = –2.25).
Concerning the winter season, no month escaped the drought either by SPI–3 or by SPEI–3 and even on
different time-scales. The abnormal dryness of the rainy season months was due to a relative frequency
of random anticyclonic situations. This winter drought seemed to change from month to month, and was
shorter than in the summer season (Figure 6). The reason for this is that winter anticyclonic situations,
generating dry and cold weather, are less persistent than stable and hot summer weather (summer anti-
cyclones), therefore more quickly degraded by rainfall disturbances.
The climate in LSB is subject to the contrasts of the Mediterranean climate concomitant with ocean-
ic influences which give it particular climatic rhythms on the one hand, and on the other hand the north–west
winds, meeting the Iberian chains (Spain) arriving dry in the north–west of Morocco. In addition, there
is a simple succession of two climatic environments (cyclonic and anticyclonic), which have a remarkable
influence on the inter-annual variability of the climate in Morocco (Driouech et al. 2021) and in partic-
ular in the LSB (Hakam et al. 2022b).
The dominant influence of large-scale climate indices on the evolution of drought was slightly man-
ifested in the LSB. Consistency analysis indicates a significant influence of climate indices (NAO and MO)
on drought variability at the annual to multi-year (5-year) scale, particularly during years with drought
conditions (Figure 5). Both NAO and MO were strongly influenced by the North–East Atlantic cyclonic
44
62-2_acta49-1.qxd  13.12.2022  12:08  Page 44
systems that force Mediterranean cyclogenesis (Zeroual, Assani and Meddi 2017), inducing a strong cor-
relation between them (Criado-Aldeanueva, Soto-Navarro and García-Lafuente 2014). The OM can be
defined as an oscillation of sea level pressure anomalies (SLP) in the central and western Mediterranean,
an important source of cyclogenesis. As the occurrence of these cyclones is partly related to the activity
of NAO–governed North Atlantic fronts, a strong correlation can be expected (Criado-Aldeanueva and
Soto-Navarro 2020). For example, Angulo-Martínez and Beguería (2012) found that the erosive force of
precipitation in the Ebro Basin (NE Spain) is strongest during negative phases of the NAO and OM. It is
weaker during positive phases.
The statistical analysis is highly dependent on the observational (in–situ) data, such as the duration
of observation, authentic source, spatial density and critical criteria used for the analysis, trends and peri-
odicity. In addition, the analysis of global and regional climate models could provide physical dynamics
of climate change and improve the reliability of the analysis results.
5 References
Abramowitz, M., Stegun, I. A. 1965: Handbook of mathematical functions. New York.
Acharki, S., Amharref, M., El Halimi, R., Bernoussi, A.-S. 2019: Assessment by statistical approach of cli-
mate change impact on water resources: Application to the Gharb perimeter (Morocco). Journal of Water
Science 32-3. DOI: https://doi.org/10.7202/1067310ar
Agence du Bassin Hydraulique de Sebou. 2010: Evènements hydro-pluviométriques dans le bassin du Sebou
2009–2010. Maroc.
Ajlal, S., Agoutime, Z. 2021: Changements climatiques et leurs impacts dans la région du Gharb, Maroc.
Journal of African Studies and River Nile Basin 3-11.
Angulo-Martínez M., Beguería S. 2012: Do atmospheric teleconnection patterns influence rainfall ero-
sivity? A study of NAO, MO and WeMO in NE Spain, 1955–2006. Journal of Hydrology 450,168. DOI:
https://doi.org/10.1016/j.jhydrol.2012.04.063
Bayissa, Y., Maskey, S., Tadesse, T., Van Andel, S. J., Moges, S., Van Griensven, A., Solomatine, D. 2018:
Comparison of the performance of six drought indices in characterizing historical drought for the upper
Blue Nile basin, Ethiopia. Geosciences 8-3. DOI: https://doi.org/10.3390/geosciences8030081
Beguería, S., Vicente‐Serrano, S. M., Reig, F., Latorre, B. 2014: Standardized precipitation evapotranspira-
tion index (SPEI) revisited: Parameter fitting, evapotranspiration models, tools, datasets and drought
monitoring. International Journal of Climatology 34-10. DOI: https://doi.org/10.1002/joc.3887
Chang, X., Wang, B., Yan, Y., Hao, Y., Zhang, M. 2019: Characterizing effects of monsoons and climate
teleconnections on precipitation in China using wavelet coherence and global coherence. Climate
Dynamics 52. DOI: https://doi.org/10.1007/s00382-018-4439-1
Conte, M., Giuffrida, A., Tedesco, S. 1989: The Mediterranean Oscillation. Impact on precipitation and
hydrology in Italy. Conference on Climate and Water. Helsinki.
Criado-Aldeanueva, F., Soto-Navarro, F. J., García-Lafuente, J. 2014: Large-scale atmospheric forcing influ-
encing the long-term variability of Mediterranean heat and freshwater budgets: Climatic indices. Journal
of Hydrometeorology 15-2. DOI: https://doi.org/10.1175/JHM-D-13-04.1
Criado-Aldeanueva, F., Soto-Navarro, J. 2020: Climatic indices over the Mediterranean Sea: A review. Applied
Sciences 10-17. DOI: https://doi.org/10.3390/app10175790
Diani, K., Kacimi, I., Zemzami, M., Tabyaoui, H., Haghighi, A. T. 2019: Evaluation of meteorological drought
using the Standardized Precipitation Index (SPI) in the High Ziz River basin, Morocco. Limnological
Review 19-3. DOI: https://doi.org/10.2478/limre-2019-0011
Driouech, F. 2010: Distribution des précipitations hivernales sur le Maroc dans le cadre d’un changement
climatique : descente d’échelle et incertitudes. Ph.D. thesis, University of Toulouse. Toulouse.
Driouech, F., Stafi, H., Khouakhi, A., Moutia, S., Badi, W., ElRhaz, K., Chehbouni, A. 2021: Recent observed
country‐wide climate trends in Morocco. International Journal of Climatology 41. DOI: https://doi.org/
10.1002/joc.6734
El Jihad, M. -D., Peyrusaubes, D., El Bouzidi, A. 2014: Sécheresses saisonnières et changement climatique
dans le Gharb (Maroc). Rur@lités 2014-4.
Acta geographica Slovenica, 62-2, 2022
45
62-2_acta49-1.qxd  13.12.2022  12:08  Page 45
O. Hakam, A. Baali, K. Azennoud, T. El Kamel, Y. Ait Brahim, Y. Ahouach, Spatiotemporal evolution of droughts and their …
Gerstengarbe, F. -W., Werner, P. C. 1999: Estimation of the beginning and end of recurrent events within
a climate regime. Climate Research 11-2. DOI: https://doi.org/10.3354/cr011097
Hakam, O., Baali, A., Ait Brahim, Y., El Kamel, T., Azennoud, K. 2022b: Regional and global teleconnections
patterns governing rainfall in the Western Mediterranean: Case of the Lower Sebou Basin, North-West
Morocco. Modeling Earth Systems and Environment. DOI: https://doi.org/10.1007/s40808-022-01425-3
Hakam, O., Baali, A., El Kamel, T., Ahouach, Y., Azennoud, K. 2022a: Comparative evaluation of precip-
itation-temperature based drought indices (DIs): A case study of Moroccan Lower Sebou basin. Kuwait
Journal of Sciences 49-3. DOI: https://doi.org/10.48129/kjs.13911
Harbouze, R., Pellissier, J. -P., Rolland, J. -P., Khechimi, W. 2019: Rapport de synthèse sur l’agriculture au
Maroc. Internet: https://hal.archives-ouvertes.fr/hal-02137637/document (14. 6. 2022).
Lamb, P.J., Peppler, R.A. 1987: North Atlantic Oscillation: Concept and an application. Bulletin of the American
Meteorological Society 68-10. DOI: https://doi.org/10.1175/1520-0477(1987)068<1218:NAOCAA>2.0.CO;2
Linares, J. C., Camarero, J. J. 2012: From pattern to process: Linking intrinsic water‐use efficiency to
drought‐induced forest decline. Global Change Biology 18-3. DOI: https://doi.org/10.1111/j.1365-
2486.2011.02566.x
Liu, C., Yang, C., Yang, Q., Wang, J. 2021: Spatiotemporal drought analysis by the standardized precipita-
tion index (SPI) and standardized precipitation evapotranspiration index (SPEI) in Sichuan Province,
China. Scientific Reports 11. DOI: https://doi.org/10.1038/s41598-020-80527-3
Marchane, A., Jarlan, L., Boudhar, A., Tramblay, Y., Hanich, L. 2016: Linkages between snow cover, tem-
perature and rainfall and the North Atlantic Oscillation over Morocco. Climate Research 69-3. DOI:
https://doi.org/10.3354/cr01409
McGuire, A. D., Ruess, R. W., Lloyd, A., Yarie, J., Clein, J. S., Juday, G. P. 2010: Vulnerability of white spruce
tree growth in interior Alaska in response to climate variability: Dendrochronological, demographic,
and experimental perspectives. Canadian Journal of Forest Research 40-7. DOI: https://doi.org/10.1139/
X09-206
McKee, T. B., Doesken, N. J., Kleist, J. 1993: The relationship of drought frequency and duration to time
scales. Proceedings of the 8th Conference on Applied Climatology. Boston.
Ministry of Agriculture and Maritime Fisheries. 2020: Regional agricultural plan – Region of Gharb – Chrarda –
Beni Hssen. Internet: https://www.ormvag.ma/PDF/Brochures/Plaquette PAR ORMVAG.pdf (14. 6. 2022).
Pyarali, K., Peng, J., Disse, M. Tuo, Y. 2022: Development and application of high resolution SPEI drought
dataset for Central Asia. Scientific Data 9. DOI: https://doi.org/10.1038/s41597-022-01279-5
Regional Observatory of the Environment and Sustainable Development. 2014: Climate change in the region
of Gharb, Chrarda and Beni Hssen. Internet: https://www.4c.ma (24. 12. 2021).
Šebenik, U., Brilly, M., Šraj, M. 2017: Drought analysis using the standardized precipitation index (SPI).
Acta geographica Slovenica 57-1. DOI: https://doi.org/10.3986/AGS.729
Stagge, J. H., Tallaksen, L. M., Gudmundsson, L., Van Loon, A. F., Stahl, K. 2015: Candidate distributions
for climatological drought indices (SPI and SPEI). International Journal of Climatology 35-13. DOI:
https://doi.org/10.1002/joc.4267
Stigter, T. Y., Nunes, J. P., Pisani, B., Fakir, Y., Hugman, R., Li, Y., Tomé, S. et al. 2014: Comparative assess-
ment of climate change and its impacts on three coastal aquifers in the Mediterranean. Regional
Environmental Change 14. DOI: https://doi.org/10.1007/s10113-012-0377-3
Thornthwaite, C. W. 1984: An approach toward a rational classification of climate. Geographical Review
38-1. DOI: https://doi.org/10.2307/210739
Torrence, C., Webster, P. J. 1999: Interdecadal changes in the ENSO–monsoon system. Journal of Climate
12-8. DOI: https://doi.org/10.1175/1520-0442(1999)012<2679:ICITEM>2.0.CO;2
United Nations Office for Disaster Risk Reduction. 2021: Special report on drought 2021. Internet:
https://www.undrr.org/publication/gar-special-report-drought-2021 (14. 6. 2022).
Vatter, J., Wagnitz, P., Hernandez, E. 2019: Drought risk. The global thirst for water in the era of climate
crisis. Internet: https://d2ouvy59p0dg6k.cloudfront.net/downloads/drought_risk__wwf_.pdf (14. 6. 2022).
Vicente‐Serrano, S. M., Beguería, S. 2016: Comment on ‘candidate distributions for climatological drought
indices (SPI and SPEI). International Journal of Climatology 36-4. DOI: https://doi.org/10.1002/joc.4474
Vicente-Serrano, S. M., Beguería, S., López-Moreno, J. I. 2010: A multiscalar drought index sensitive to
global warming: The standardized precipitation evapotranspiration index. Journal of Climate 23-7. DOI:
https://doi.org/10.1175/2009JCLI2909.1
46
62-2_acta49-1.qxd  13.12.2022  12:08  Page 46
Vicente-Serrano, S. M., Beguería, S., Lorenzo-Lacruz, J., Camarero, J. J., López-Moreno, J. I., Azorin-Molina,
C., Revuelto, J. et al. 2012: Performance of drought indices for ecological, agricultural, and hydrological
application. Earth Interactions 16-10. DOI: https://doi.org/10.1175/2012EI000434.1
Vicente-Serrano, S. M., Quiring, S. M., Peña-Gallardo, M., Yuan, S., Dominguez-Castro, F. 2020: A review
of environmental droughts: Increased risk under global warming? Earth-Science Reviews 201. DOI:
https://doi.org/10.1016/j.earscirev.2019.102953
Wilhite, D. 2006: Drought monitoring and early warning: Concepts, progress and future challenges. Internet:
https://public.wmo.int/en/resources/library/drought-monitoring-and-early-warning-concepts-progress-
and-future-challenges (14. 6. 2022).
Woillez, M. - N. 2019: Revue de littérature sur le changement climatique au Maroc: Obsrvations, projec-
tions et impacts. DOI: https://doi.org/10.3917/afd.woill.2019.01.0001
World Meteorological Organization. 2012: Standardized precipitation index. User guide. Internet:
https://library.wmo.int/doc_num.php?explnum_id=7768 (14. 6. 2022).
Yevjevich, V. M. 1967: Objective approach to definitions and investigations of continental hydrologic droughts.
Ph.D. thesis, Colorado State University. Colorado.
Yihdego, Y., Vaheddoost, B., Al-Weshah, R. A. 2019: Drought indices and indicators revisited. Arabian Journal
of Geosciences 12-3. DOI: https://doi.org/10.1007/s12517-019-4237-z
Zamrane, Z., Mahé, G., Laftouhi, N.-E. 2021: Wavelet analysis of rainfall and runoff multidecadal time
series on large river basins in western North Africa. Water 13-22. DOI: https://doi.org/10.3390/w13223243
Zeroual, A., Assani, A. A., Meddi, M. 2017: Combined analysis of temperature and rainfall variability as they
relate to climate indices in northern Algeria over the 1972–2013 period. Hydrology Research 48-2.
DOI: https://doi.org/10.2166/nh.2016.244
Acta geographica Slovenica, 62-2, 2022
47
62-2_acta49-1.qxd  13.12.2022  12:08  Page 47