!

I ScienceDirect'

Heliyon

Volume 8, Issue 12, December 2022, el2000

Resea re h article

Comparing surface water supply index and streamflow

drought index for hydrological drought analysis in Ethiopia
Kassa Abera Tareke 0
b ~ 181 , Admasu Gebeyehu Awoke b

a Department of Hydraulic Engineering, Wallo University, KioT, Ethiopia

h Addis Ababa University, Addis Ababa Institute ofTechnology, Ethiopia

Received 29 June 2022, Revised 21 October 2022, Accepted 23 November 2022, Available on line 30 November 2022, Version of Record
3 December 2022.

(±) What do these dates mean?

[!ii} Check for updates ,

Showless A

::: Outline [ < Share " Cite

https:f/doi.org/10.10l6/j.heliyon.2022.el2000 71
Get rights and content 71

Under a Creative Commons license 71 open access

Abstract
Recently, floods and drought have become common natural hydroclimatic hazards in several countries. Consequently,
the identification of an appropriate drought index is now a challenging task for researchers. It is obvious that there is
not a single best drought index; rather a comparison of indices will give a relative option. The objective of this study
was to compare two hydrological drought indices; the modified surface water supply index (MlSWSI) and stteamflow
drought index (SDI) over eight river basjns, in Ethiopia. The MlSWSI and SDI value was computed from 1973 to 2014
using 34 streamflow stations, 42 rainfall gauge stations, and 3 lake-level data. The two indices results showed that the
1980s were the most severe drought years for all river basins. But for the case of Genale Dawa and Wabishebele basins,
the drought severity increased from 2000 to 2014. Hydrological drought analysis using SDI has more drought
occurrence frequency than Ml SWSI. In all river basins from 1973 to 2014, there were a total of 18 severe drought events
when using M1SWSI, but there were a total of 39 severe and 12 extreme drought events when using SDI. This implied
that Ml SWSI reduced the occurrence probabm~ of severe drought by 53.85% and extreme drought by 100%. It is known
that Ethiopia is stricken by extreme droughts in the last few decades. But MlSWSI doesn't detect those invidious
drought events. In this study, SDI is found to be a better hydrological drought index. Therefore, policy and strategic
planners, master plan developers, and decision-makers can use SDI to analyze historical and future hydrological
drought trends to develop effective drought mitigation measures.

~ Previous Next 0
Keywords
Hydrological drought; Modified surface water supply index; Streamflow drought index

1. Introduction
Floods and droughts are natural hydroclimatic hazards affecting several countries in the world. Globally, flood studies
have good concern than drought due to their fast impact and short duration [ 1 ]. Flood and drought disasters become a
bottleneck for the economic development of many countries. However, drought is the most complex and widespread
!ro!rQ!Qgical extreme than flood [2 ]. Drought has a devastating negative imoact on water suooly, irrigation, hydrooower,
and all kinds of water resource projects [3 ]. As a consequence, recent drought analysis and forecasting studies become
more interesting to develop effective drought mitigation measures [4 ].

The definition of drought is more subjective due to its complex nature and scholars defined it from a different
perspective [5]. However, drought can commonly be classified into four types (a) meteorological drought associated
with scarcity of precipitation for long periods below normal situations [6, 7, 8 ], (b) hydrological drought related to the
low water level in~ and subsurface water resources such as a lake, reservoir, streamflow and groundwater [6 ], [9,
10, 11, 12, 13, 14, 15], (c) agricultural drought related to lack of soil moisture to attain the minimum crop water required
in the soil and distracts agricultural productivity [16, 17, 18, 19] and (d) the fourth one is socio-economic drought which
is the overall welfare crisis of the society caused by severe drought [5, 20, 21, .22, 23 ]. Meteorological drought highly
affects the agricultural systems by aggravating food insecurity, especially in developing countries due to crop failures
before harvesting season while hydrological and agricultural drought causes low production of industries because of
shortage of water sug,g]y to irrigation, municipals, and industries, hydropower generation [24]. The cumulative effects
of meteorological and hydrological droughts lead to socioeconomic drought, which disturbs the entire ecosystem and
badly affects and even loses the lives of humans and animals [8 ]. In addition to this extreme hydrological events has
high influence on water quality and it needs a wide concern [25].

Studies revealed that Ethiopia has faced several severe and extreme drought events in the last few decades as a result of
erratic rainfall and climate chang [26, 27, 28, 29 ]. Northern Tigray region, some parts of Amhara region such as South
Wollo, North Wollo, South Goodar, and Afar region, most parts of Somalia region, and eastern parts of Oromia region
(Borena Zone) have frequently been affected by severe drought in Ethiopia [26, 30, 31. 32, 33, 34]. According to [35],
drought in Ethiopia occurs at a recurrence interval of three to~~. Even though this frequent recurrence is well
recognized there still lacks any firmly-established drought mitigation measure for these disasters. In Ethiopia, drought
response efforts are provided in the form of food aid when food suoplies have decreased significantly due to extended
drought for a short-term recovery.

Relatively, meteorological drought analysis gets more focused than other types of drought in Ethiopia [35, 36, 37, 38, 39,
40, 41 ]. There are few drought indices comparison studies across Ethiopia and specifically Upper Blue Nile in the Abbay
river basin. respectively [42, 43). Bayissa (2018) (43) tried to compare six drought indices in Upper Blue Nile; three
meteorological indices, two agricultural indices, and one hydrological index. Since the number of input data and the
purpose of the index developed affects the selection of the appropriate drought index for a specific study area. Tsige
etal. (2019) [42] also compared two indices (meteorological and agricultural indices) in which the finding were
relatively good compared to Bayissa's finding. However, both scholars were focused on meteorological drought indices
than hydrological drought indices. Hydrological drought analysis is not well studied in Ethiopia as a national level
which has a great influence on national economic development.

Hydrological drought is related to surface and subsurface water shorta~ in terms of volume from the long-term normal
condition. Hydrological drought always lags from meteorological drought and it ~gation orocess from
meteorological to hydrological drought is important to develop effective drought ~ly warning~ [44 ]. For long-
time drought events, meteorological and agricultural droughts have been propagated to hydrological drought which
caused streamflow, reservoir and lake level, and groundwater level reduction. As a result, water supply, irrigation, and
hydropower generation have been directly affected. Because the development of hydrological drought is directly related
to the transformation of precipitation into effective streamflow generation [45). Ethiopia has many perennial rivers and
lakes, but still, those resources are not well utilized [46]. While hydrological and agricultural drought analysis and
monitoring have not been thoroughly explored, meteorological drought analysis has been regularly studied in various
parts of the country [47 ].

Many drought indices are region sensitive and developed in different countries [48]. Most of the indicators are
developed for meteorological and agricultural drought analysis. However, there are also drought indicators used for
hydrological drought analysis such as the palmer hydrological drought index (PHDI) [49], streamflow drought index
(SDI) [50], and surface water supply index (SWSI) [12]. SDI is simple and required a single input data (streamflow) and is
suitable for drought study at the site [47] whereas, SWSI was developed at Colorado University by Shafer and Dezman
(1982) and is used for regional or basin-level hydrological drought analysis by considering the available surface water
from different sources [12]. It requires all forms of surface water sources such as streamflow, precipitation, reservoir
level, and groundwater level. Hydrological drought analysis gives good information for policymakers and sectors related
to water. Because drought related to transboundary rivers has crucial importance for special treaonem during drought
events and it is important to develop a good water resource management strategy before and after drought has
occurred. Many Ethiopian rivers are transboundary rivers and need effective assessment and management. Therefore,
this study was aimed to analyze hydrological drought conditions in Ethiopia by comparing two hydrological drought
indicators (SWSI and SDI). The severity condition of drought obtained by these indicators was also compared using
Pearson's correlation coefficients.

2. Methods and materials

2.1. Study area

Geographically, Ethiopia is located in the eastern horn of Africa at 3° N to 15° N latitude and 33° E to 48° E longitudes.
Approximately the total area of Ethiopia is 1.13 million km 2 [ 51 ]. From this, the study covered about 87.3% of the
country. The rate of population growth increases rapidly and causes deforestation and expansion of urban areas which
indirectly increases hydrodimatic variability and extremes. Figurel shows the major river basins in Ethiopia and the
soatial location of hydrometeorological stations. Of the existing 12 major river basins (Figure 1 ), three are dry and one
river basin has low flow and low recorded data [47]. Therefore, comparing of two hydrological drought indices (SDI and
SWSI) was conducted only on eight major river basins.
 JS'O'O"t: 40'0'0"t:
Ethiopiam Major !Rh•er Basin r-
e
• Str eam_now_Station ~
N
e Lnke_Locntion

A -
• Meteorological_Stntio
Ethio_Main_ Rivers l
None_Study_Basins
D Ethio Basin

0 75 150

JS'O O"E 45'00"E

Download: Download fllgh-res Image (S9SKll)

Download: Download futklze lmoige

Figure 1. Location of hydrometeorological stations.

Ethiopia is the tropical 7.one laying between the equator and the tropic of cancer. It has three different climate zones
according to elevation. These are: Kolla (Tropical 7.one) below 1830 m in elevation and the annual average temperature
and rainfall are 27 •c and 510 mm respectively. The second is Woina Dega (Subtropical 7.one ). which is composed of
~between 1830 and 2440 min elevation and has an average annual temperature of nearly 22 •c and
510-1530 mm of rainfall The third is Dega (Cool zone1 which is located above 2440 min elevation and has an annual
rainfall range of 1270-1280 mm with an average yearly temperature of about 16 •c. The country has four major seasons:
Summer 'Kiremef Oune-August; JJA); autumn 'Tibe' (September-November; SON); winter 'Bega• (December-February;
DJF) and spring 'Belg' (March-May; MAMi However. the coldest month is not always in 'Bega' and the hottest month is
not always in 'Kiremet'. Ethiopia lies near the equator where maximum heat from the sun is received. The length ofdays
and nights is almost the same in most regions.

Precipitation is the main source of streamflow generation for river basins in Ethiopia. F-Or the eight studied basins, most
of them have received high rainfall during the summer season WA). But for the case of Cenale Dawa and W.abishebele,
they receive high rainfall during the $,Pring season (MAM) and low rainfall during swnmer OJA), and the variation of
streamflow level is also corresponding to the precipitation time variance (Yigure2).

~ Abbay ---Tekeze ..,._Awash - Baro

-llE- Omo ...-R.Valley __. Genale - Wabishebele
400

'?300
s
=
<S
200
.5
~ 100

0
Jan Feb Mar Apr May J un Jul A ug Sep Oct Nov Dec

Download: OOWlllaod higlHes image (4941<8}

Download: Dowliload fulklze Image

Flgure2. Mean monthly rainfall ofeight river basins, in Ethiopia.

The spatial distribution of rainfall over the country highly varies; Abbay, Baro Akobo and Omo Clbe River basins receive
high rainfall whereas the remaining five river basins receive medium m low rainfall. As indicated In Flgure3, semi-arid
and arid basins such as Cenale Dawa, Wablshebele, and Ogaden receives below 1000 mm annually. Tekeze, Awash, and
Rift Valley have an annual rainfall range between 770to1100 mm.

N Legend

A CJ E1hio_ Bosi11
)1r:11;11 Annual Jlainfall
(mm)
- 522.74 - 772.91
- 772.92 - 871.04
c::::J 872.05 - ?66.44
L !>6Ci.4s- c.us9.11
- t.089.18- 1.235.5
- 1..235.5 1 - l..l77.I
- 1..177.11 - 1.490..19
- 1.490.4 - 1.726.4

0 80 160 320 480 6-10.

Km

Downlood: Dowrtlood high-res lmoge (S89KB)

Download: Dowlllood fulklze Image

Figure3. Spatial variation of mean annual rainfall in Ethiopia.

3. Methods

3.1. Data preparation

Precipitation data were collected from the Ethiopian Meteorology Institute whereas streamflow and laJre level data
were collected from the Ministry of Water and Energy (MoWE). Based on data record length. quality, and availability:
42, 34, and 3 precipitation gauge stations, streamflow. and lake level stations were collected respectively over eight
major river basins in the country. But for all those stations, the data record length is not the same. Especially,
streamflow data length ls varying from basin to basin. However, the dominant n!Cord range was decided from 1973 to
2014. Missing data for precipitation was completed using the areal ratio method and for streamflow, XISfAT was used.
XL.STAT ls a statistical analysis tool lntx!grated with Microsoft excel used to fill in missing data, analyze hydroJoglc.al tests
such as consistency and homogeneity, and visualize and plot bydrodimatic data (52).

Hydrometeorological data can be affected by natural hazards or by human intervention. For example, the gauge station
may be collapsed, break, or miss reading and data collec:tocs may collect wrong data. Due to those and other reasons,
the information generated from these poor-quality data is affecting the decision of researchers, stakeholders, and
planners. Therefore, before any data is used as model input. data quality analysis is mandatory. Jn any water resource
management activities. the hydrodimadc data should be as much as possible stationary, homogeneous, consistent.
stable variance and mean (53, 54]. So. in this study, consistency and homogeneity tests were checked using double mass
curves and non-dimensional ratio methods. respectively for all selected meu:orological stations [55, 56).
3.2. Original surface water supply index
Shafer and Dezman (1982) [57], in Colorado, developed SWSI to supplement the Palmer Drought Severitv Index by
taking streamflow, reservoir stora~ and snQWPack into account. The steps to calculate the SWSI for a specific basin are
as follows: monthly data are gathered and added for all the reservoir inflow, streamflow monitoring stations, and
precipitation stations throughout the basin. A long-term mean is used to normalize each component's sum. Each
element is given a weight based on how frequently it contributes to the surface water in that basin [ 12 ]. The large
portion of available water resources in Ethiopia is surface water (streamflow) which meteorological drought indexes
such as SfL ROI, and PDSI do not explicitly include.

The main inputs for SWSI are streamflow, precipitation, reservoir storage, and grnundwater level (optional). But
reservoir storage is directly related to the inflow stream condition of the basin and groundwater level is important for
groundwater drought analysis. Therefore, for this research streamflow, precipitation, and lake level instead of reservoir
storage were used as input for SWSI analysis. The equation is given below in Eq. (1):

SWSI = aPNstrm+l>PNprec+cl'Nlal-50 (1)

12

where: SWSI = Surface Water Supply Index, PNstrm, PNprec, and PNlal are a percentage of non-exceedance (%)of
monthly streamflow, precipitation, and lake level respectively, and a, b and c = weight for each hydrologic component in
which; a+ b + c = 1. Subtracting 50 centers the SWSI values around zero, and dividing by 12 compresses the range of
values between -4.2 to +4.2. The non-exceedance orobabilities are taken from probability distributions fitted to each
hydrologic component

3.3. Modified surface water supply index (MlSWSI)

This method was modified by [48] using forecasted streamflow and reservoir storage volume. But it does not explicitly
analyze different hydrological components and this masks important information about the behavior of each
hydrological component The equation is given by Eq. (2):

SWSI = (P-5o) (2)

12

where: P =the non-exceedance probability(%) of a long-term mean.

The surface water supply index (SWSI) is one of the hydrological drought indicators which gives a wide range of
drought characterization than SDI. It is applicable for basin-level drought analysis and it was developed based on PDSI
algorism [58, 59]. For this work without altering the algorism, the equation is modified to make it easy to compare with
the SDI value. SWSI was developed to incorporate multiple hydrologic/meteorological components into a single
objectively derived index value for each river basin [60].

However, still, SWSI is more subjective, and compared with PSDI which is meteorological drought indices, the
hydrological components considered need explicit analysis. So, it is important to modify the equation to make it
comparable with a hydrological index such as SDI by reducing the compressed range from -4.2 to +4.2 into -2.1 to +2.1.
This reduction of the range is done without altering the algorithm of the model and simply reduced the range by
increasing the denominator. Now it can be comparable with the SDI value for a given basin and the equation is given by
Eq. (3):

MlSWSI = aPNatrm+bPNprec+cPNlal-50 (3)

24

where: Mt SWSI is modified SWSI. and all other terms are described in Eq. (1 ).

The value of weighted factors a, b, and c were more subjective in the original SWSI development even though
eliminated by the revised one [ 15 ]. But it is important to make it objective and give a sense of the art of science.
Therefore, the value of these weighted factors is formulated below in Eq. (4).

Pa = Pb = Pc = __.!fi_
X max
(4)

where: Pa Pb and Pc are the proportional value of monthly or annual streamflow, precipitation, and lake level
respectively whereas Xi and Xmax are the observed monthly or annual value and maximum values, respectively for all
components. Now the weighted factors can be determined as follow in Eq. (5).

a-Pa b-Pb
-pt,-pt,-pt
c-Pc (5)

where: Pa, Pb, and Pc are as described earlier in Eq. (4) and Pt is the total proportionality of each component (Pt= Pa+
Pb+ Pc); ab and care weighted factors of each surface water component (streamflow, precipitation and lake level)
respectively.

The probability of non-exceedance for each component was determined using Eq. (6) as shown below developed by
Weibull [61 ].

PN = 1- n7.i (6)

where: PN is the non-exceedance probability of each component, m is the rank and n is the total number of data
considered in the time series.

Table 1 shows the drought criteria originally classified by Shafer and Dezman in 1982 and the modified classification is
downscaled by half from the original.

Tablet. Original SWSI (Shafer and Dezman, 1982) and modified MlSWSI values.

Origin.al SWSI ~ MtSWSI range Description

~+4 ~+2 Abundant supply

3.99to 1.99 1.99 to0.99 Wet

2 tv-0.99 1 to-0.49 Normal

-1 to-1.99 -0.5 to -0.99 Incipient drought

-2 to-2.99 -1 to-1.49 Moderate drought

-3 to-3.99 -1.5 to -1.99 Severe drought

S-4 S-2 Extreme drought

Note: M1 SWSI is a modified surface water supply index.

3.4. Streamflow drought index analysis

The Streamflow Drought Index (SDI) was developed by Nalbantis and Tsakiris (2009) [62], which is used to characterize
the hydrological drought situation of a study area. Since its calculation is similar to SPI's, it has the same efficiency and
simplicity. The SDI gives the advantage of controlling hydrological drought or the availability of water in the short,
medium, and long term because it is based on monthly observed streamflow amounts at various time scales. The SDI is
defined as follows for each reference period k of the ith hydrological year based on the cumulative streamflow volumes
Vi,k:

SDI = Vi,k-Vkm (7)

Sic

where: i - 1, 2; ...• and k - 1, 2, 3, 4.

Since these are calculated over a long period, Vkm and Skare the mean and standard deviation of the cumulative
streamflow volumes of the reference period k, respectively. The SDI runs from -2 to +2 in terms of wetness and dryness.
Below - 2 and above +2, respectively, are the values that are exceedingly dry and wet. The SDI criteria for identifying the
worst and most intense drought occurrences are shown in Table2.

Table2. Drought classification according to the SDI values (Nalbantis and Tsakiris, 2009).

SDI value

Extremely wet
SDI value category

1.5-1.99 Severely wet

1-1A9 Moderately wet

0.5-0.99 Slightly wet

-0.49 to +0.49 Normal

-0.5 to -0.99 Mild drought

-1 to-1.49 Moderately drought

-1.5 to -1.99 Severely drought

:S-2 Extremely drought

SWSI was developed based on the PDSI algorithm. But PDSI is more important for meteorological and agricultural
drought analysis than hydrological drought [63]. Therefore, comparing SWSI with PDSI is subjective and it needs some
modification of the range of SWSI results and weighted factors. After compressing the value of SWSI from -4.1 to +4.1
into -2.1 and +2.1; it is possible to compare the result with SDI for a given river basin. Based on this, the correlation of
the two hydrological drought indices was computed.

4. Results and discussion

4.1. Consistency and homogeneity analysis

The analysis of consistency and homogeneity was conducted using the double mass curve method and non-dimensional
ratio approach method respectively. For the case of Awash and Omo Gibe River basins, two stations were found
inconsistent and nonhomogeneous (see Figures4 and 5 ). However, the remaining 40-gauge stations within the
respective river basins showed good homogeneity and consistency result As shown in Figure4, one gauging station
from the Awash and Omo Gibe River basins (Melkawerer and Bonga), respectively was not consistent with other
stations from the basins. Therefore, the Melkawerer station from the Awash river basin and the Bonga station from the
Omo Gibe River basin was not considered for further drought analysis in this study.
 Tekcze Basin

h-::L
~~20·~
?
:ii~
g ~~

o __ ; ££4J~C
i* I 11 -
:·:-
-
__ ,, ••• =·-
100000 200000 300000 400000 500000
~ 1 0 20000 4000() 60000 80000 100000 120000 140000
Comil:111i\'e rainfall ofall stacioos. (mm) (,.I Cutnulali\•Crainfall of all st:uions{mm)
- -oatiir l)sr - chagni - -G')f1d:u - Oab:n _... H agcr~lam _... Md:ele • Ambagi~is
--.GUlldewoin - -Dcbcro Tabor --.Dangila
OmoG;bc
Awash Basin

'"""'!,I!!!! 5!!!~!!!!!Y
150000 200000 250000
Cumulali't-e rairlfull ofaJI s1a1i(Jn$ (n1m)

50000 100()00 150000 200000 250000 300000 350000

Cumulative rainfall of all st.alions (mm)
- - Awa.sh7kilo GcnaJe Dawa
--. A4:Hs Ababa --. Hom~lc --.Qodcr

Baro Basin
~~60000
SO
OOOL ~
jI 40000
20000 40000 60000 80000 100000

~ ·~ 20000 ---- ---

Cumul;Mi\'C rninfiill ohll $&ilOOt1$(mm0
- -oinir - Coro - -Negde
j ~ 0 =:::.:..__~----~----~----~
% 0 100000 150000 200000

~ Commullltive Rainfall <1f all st.itioos (mm)

---oambela

Wobishebele Basin

Download: Dowllload high-res lma911 (lMB)

Download: Download fulklze Image

Figure4. Consistency test for all river basin using double mass cwve.
 j 30 I A~bbay
Basin
30
Awash Basin

!20
~ 10 . 2010 1 ~
·~ O ~'!' :.' ~' O~ , :t,
o ~~~<~~....,~~ozo a~~5.~§:s~e-9~g
C -~~<~--,~~OzO
'*- -+- Bah ir Dar -+-Chagni -+- Melkawerer -+- Mojo
-+-Gondar - -Makisegnit -+-Awash7kiio - Kombolcha
-+- Haik -+- Metehara
-+- Gundewoin -+- Debere Tabor -+-Addis Ababa -+- Hombele
-+- Dangila -+-Guder

Tekeze Basin

-+- Dabat -+- Hageresclam

-+-Alemteferi -+- Metu -+-Gambel -+- Mekeie Ambagiorgis

Omo Gibe Basin Genale Dawa Basin

Jan Mar May Jul Sep Nov

-+-Assendabo -+-Gojeb
-+- Wolkite ....... Abelti
-+- Bonga -+- Ginir -+-Goro - - Negele

f::i~
Wabishebele Basin
20
15
10
g
·~ Jan Mar May Jui Sep Nov 5
"E 0 +-=...;:=.~~~~~~~~~~!.,
9 -+-Shashcmcne -+- Konso
~-+-Arsi Adamitulu
-+- Dedessa -+- Hosana -+- Endeto -+- Bisidimo
-+- Arbaminch -+- Girawa Jijiga

Download: Download high-I-es lma91' (lMll)

Download: Downlood futklze imgige

Figures. Precipitation data homogeneity test

As shown in Figure5, almost in all river basins the stations are relatively homogeneous. But. in the Awash, Omo Gibe.
and Rift Valley River basins, some stations are non-homogeneous. However, the variation is not significant. therefore
except for Guder station (Awash River basin) and Bonga station (Omo Gibe River basin) all the stations were considered
for further analysis (see Figures).

4.2. Analysis of hydrological drought using SWSI

Even though SWSI analysis requires more input data, this analysis was computed using three input data for Abbay and
Rift Valley River basins (streamflow, precipitation, and lake level), whereas for the remaining six river basins, two input
data were used (streamflow and precipitation) due to data scarcity of lake level All the input data v.rere summed
separately and their non,exceedance probabilitvwas computed using Eq. (6). The value ofcomponent weights (a. b, c)
for each basin is different and it was estimated for seasonal and annual values (Table3 ). There are four seasons in
Ethiopia, Autumn (Tibe or Meher), Winter (Bega), Spring (Tsedey or Belg), and Summer (Kiremt). The value of component
weights was computed for all seasons but the rainiest seasons in Spring and Summer were selected for discussion. The
weightings a, b, and c in Table3 indicated the component potential impact on surface water available in the basin. As
observed from Table3, the most surface water source for all basins in Ethiopia is precipitation. In Abbay and Rift Valley
River basins, the lake level was considered and it shows that the seasonal variation of lake level is insignificant when
compared to the variation of streamflow and precipitation (Table3). The one fact is that the value of component weight
for each basin is not always constant as stated in Table3, rather it will change as the number of gauging stations
increases or decreases. But the variation is not that significant

Table3. Ml SWSI seasonal and annual component weights summary.

Basin Smson Weights

a (1} b(2) c(l}

Abbay Spring 0.21 0.32 0.47

Summer 0.29 0.40 031

Annual 0.29 0.40 031

Awash Spring 0.47 0.53

Summer 0.38 0.62

Annual 0.42 0.58

Baro Spring 031 0.69

Summer 0.33 0.67

Annual 0.45 0.55

Gena.le Da.wa. Spring 038 0.62

Summer 0.66 0.34

Annual 0.44 0.56

Omo Gibe Spring 0.36 0.64

Summer 0.49 0.51

Annual 0.45 0.55

Rift Valley Spring 0.22 0.45 0.32

Summer 0.19 0.44 0.37

Annual 0.27 038 035

Tekeze Spring 0.19 0.81

Summer 0.45 0.55

Annual 0.39 0.61

Spring 0.12 0.88

Wa.bishebele Summer 0.44 0.56

Annual 0.27 0.73

Where 1, 2, and 3 indicate: (1) Streamflow (2) Precipitation (3) Lake level, and(-) implies no data..

In all river basins, the 1980s were the driest years, according to the hydrological drought analysis produced by Ml SWSI
and SDI (Tables4 and 5). But the severe drought has regularly affected the Abbay and Awash River basins (Table4).
MlSWSI gives the drought information over a large area; as a result. it compressed the magnitude of hydrological
drought impact in a specific area. Previous studies revealed that Ethiopia is recurrently affected by severe and extreme
drought events. But here, the result of Ml SWSI indicated that there was no extreme drought event in the last three
decades (Table4). As shown in Table4, the magnitude ofMlSWSI is almost near the moderate drought range category
except for Awash in 2001, Baro in 1985, and Genale Dawa in 2003. Genale Dawa, Wabishebele, and Rift Valley river
basins are located in lower parts of Ethiopia and highly exposed to prolonged drought in the last decades [64, 65 ]. But
the hydrological drought analysis using M1SWSI minimized the frequency and magnitude of severe drought events in
those areas. This indicates that hydrological drought analysis using multiple hydroclimate data may hide the
information and it will directly affect the water resource management system.

Table4. Summary of Severe drought years and magnitude in each river basin in Ethiopia using Ml SWSI.

Rinr B;asln Severe drought YNl5 Moagoltude

Abbay 1982, 1983 -1.52, -1.58

Awash 1979, 1986, 1987, 2001 -1.56, -1.56, -1.72, -1.91

Baro 1985,2002 -1.73, -1.71

GenaleDawa 2003 -1.73

Omo Gibe 1981, 1994, 2016 -1.59, -1.56, -1.73

Rift Valley 1985 -1.64

Tekeze 1996,2008 -1.6,-1.67

Wabishebele 2011 -1.57

Tables. Severe and extreme bydrologll;al drought years and magnitude in Ethiopian river basins using SDI.

Rinr B;asln Severe drought YN?S Magnitude Extreme Magnitude

drougbt,ean

Abbay 1978, 1983, 1984, 1986, 1994,2010 -1.84, -1.66, -1.91, -1.69, -1.69, -1.98 1983, 1984 -2.21, -2.29

Awash 1986, 1987, 2001 1986, 1987, 2002 -2.16, -2.4, -2.38

Baro 1982,1984, 1985,2002 -1.57, -1.88, -1.62, -1.81 2002,2004 -2.01, -2.09

Genale 1996,2002,2003,2010 -1.52, -1.51, -1.76, -1.9

Dawa

Omo Gibe 1980, 1981, 1986 -1.64, -1.77, -1.69

Rift Valley 1980, 1983, 1984, 1985, 1987, 1990, -1.52, -1.54, -1.5, -1.55, - 1.64, -1.74, 1984, 1985, -2.2, -2.1, -2.14,
1999,2002,2004,2010,2012 -1.52, -1.86, -1.91, -1.91, -1.92 2003,2011 -2.21

Tekeze 1996 -1.51

Wabishebele 1990, 2001, 2002, 2004, 2005, 2011 -1.9, -1.86, -1.76, -1.69, -1.7, -1.7 2002 -2.34

4.3. Analysis of hydrological drought using SDI

To eradicate the flow variation of stations across time and space, the daily streamflow data were compiled monthly and
standardized to zero mean and unit standard deviation. Then, SDI was calculated by applying Eq. (7) to the seasonal
base for the three-month length (SDI3) and the yearly base for the twelve-month duration (SD112). The result implies
that the 3-month time scale analysis has a high frequency of drought than the long-term time scale, 12-month. But for
the long-term time scale, drought duration is maximum and the severity is increased. The result is supported by
previous studies related to hydrological and meteorological drought analysis [6, 8, 59]. The most intense and severe
drought years across many river basins were recognized in this study. Tables reveals that during the past three decades,
there have been 11 severe and 4 catastrophic droughts that have had an impact on the Rift Valley River Basin. Besides,
the Wabishebele and Abbay river basins experienced severe drought in the 1980s and 2000s, respectively. The result of
hydrological drought analysis using SDI agrees with previous meteorological drought study findings in Ethiopia [66, 67,
68, 69]. But the result from MlSWSI contrasts with SDI results.

4.4. Comparing SDI and Ml SWSI

SDI value indicated that the Rift Valley river basin is recurrently affected by severe and extreme drought events
(Table5). But the MlSWSI result indicates Rift Valley river basin is less affected by drought compared tn other basins
(Table4). A previous study indicates the Rift Valley river basin is frequently affected by worth drought [70]. The result of
this study agreed with recent findings. However, the Ml SWSI result deviated from the historical trend not only Rift
Valley river basin but also in others. As shown in Table6, except Rift Valley River basin, SDI and MlSWSI have a good
correlation in all river basins. The reason for the Rift Valley River basin is that the streamflow value is very minimum
and directly joined to a lake which increases the lake level regularly. Therefore, the available surface water in the basin
is dominated by lakes and precipitation than streamflow. As a result, the cumulative drought analysis technique using
MlSWSI gives wet season frequently and the overall result of the index can be dominated by the available lake water.

Table6. SDI and MlSWSI correlations.

Basio Colftlatioo Basin Conelation

Abbay 0.70 Omo Gibe 0.64

Awash 0.51 Rift Valley 0.19

Baro 0.72 Tekeze 0.68

GenaleDawa 0.53 Wabishebele 0.62

Table 7 shows that the occurrence of severe and extreme drought frequency is higher for SDI than for Ml SWSI. This is
because SDI gives a site or point drought condition of a single river from the basin but SWSI results are based on the
cumulative contribution of different surface water sources such as streamflow, precipitation, and lake level. It is also
understood that the probability of occurrence of extreme drought using MlSWSI was insignificant (Table7). Except for
the Awash and Tekeze river basins, the frequency of severe drought in all basins has reduced when drought is analyzed
by SWSI. The average occurrence of severe and extreme drought using SWSI was reduced by 53.85% and 100%
respectively when compared with SDI. This implies that SWSI will hide the impact of local drought and affects the
drmlght management program by reducing the severity of drought over the region. Therefore, for water resource
planning and infrastructure development in a river, SDI gives good information about the historical frequency of
drought conditions than Ml SWSI.

Table7. Number of droughts that occurred in Ethiopia from 1973-2014 using SDI and MlSWSI.

Basio Number of Severe Drought Oa:urred Number of Extreme Drought Oa:urred

SDI SWSl % of reduction SDI SWSI % of reduction

Abbay 6 4 33.33 2 0 100

Awmlt 3 4 25 3 0 100

Baro 4 2 50 2 0 100

GenaleDawa 4 75 0 0 100

Omo Gibe 3 3 0 0 0 100

Rift Valley 13 92.3 4 0 100

Tulceze J 2 50 0 0 100

Wabishebele 5 80 1 0 100
The annual time series of SDl and Mt SWSl for all river basins in Ethiopia is shown below in F"igures6, 7, 8, 9, 10, 11, 12,
and 13. All Figuresfrom 6 to 13 imply that the probability of severe and extreme drought oc:currence was high for SDI
than M1SWSI. Because M1 SWSI has been obtained from a combination of many hydroclimatl! variables and the result is
more dominated by wet events than dry events. This is due to the combination of different hydrological components for
a single basin drought analysis. However, the two indexes have good correlations for all river basins except the Rift:
Valley River basin.

~~ r -.-sm -.- MISWSI

o~ -, ~·
.. V'"" •
~
(/) -3
1973 1977 1981 1985 1989 1993 1997 2001 2005 2009 20 13
Years

Download: Dow!llood high-res image (225KB)

Download: Download fulklze Image

Figure&. Comparison between SDl and Ml SWSI in the Abbay river basin.

3
iil
3
~I

1973 1977 1981 1985 1989 1993 1997 2001 2005 2009 20 13
Years

Download: Download high-res lmaee (24SKB)

Download: Dowlllood fulk?ze Image

Figure7. Comparison between SDI and M1SWSl in the Awash river basin.

5
;;; -.-so1 12 -.- MISWSI
3 3
(/)

"O
1L1
0VJ -3
1973 1977 1981 1985 1989 1993 1997 200 1 2005 2009 20 13
Years

Download: Dowlilood lilgh-res Image (194KB)

Download: Download futklze lmaiee

Figures. Comparison between SDl and Mt SWSl in the Baro river basin.

3
r;;
:s:
V>
2
~ 0
1L2
"'
Ci -3
V> 1984 1987 1990 1993 1996 1999 2002 2005 2008 2011 2014
Years

Download: Download high-res lmalJe (2171<8)

Download: Dowliload fulklze Image

Flgure9. Comparison between SDl and Mt SWSl In the Genale Dawa river basin.

3 -+- SDl l2 -+- MISWSI

r;;
;;:: 2
C/l

~ 0
-0

~ -2
Cl
C/l -3
1980 1984 1988 1992 1996 2000 2004 2008 20 12 2016
Years

ol)W!llood: oawnlood fllgh-tts Image (220ICB)

Download: Dowl'llood fulklze lmQ99

Figure 10. Comparison betw"een SDI and MtSWSI in the Omo Gibe river basin.

ll~ 1980 1983 1986 1989 1992 1995 1998 2001

Years
2004 2007 2010 2013

Download: Dowrllood high-res Image (203KB)

Download: Download fulklze Image

Figure1t Comparison between SDI and M1 SWSI in the Rift Valley river basin.

~ ~ k -+- SDl l2 -+- M ISWSI

~-:..
0 -3
~~ ·
C/l 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014
Years

Download: Download lllgh-res Image (lBJICB)

Download : Download fulklze Image

Figure12. Comparison between SDI and M1SWSI in the Tekeze river basin.
 3
VJ
?; 2
VJ

2 0 "ff--79::..,.-..,/1-+.'....._,,.,PY:""w-...,....,,,.._-r-7'1'--r--+--R--~r--i-if--~
"O
~ -2
0
VJ -3
1987 1990 1993 1996 1999 2002 2005 2008 2011 2014
Years

Download: Download lllgh-res Image (209KB}

Download: Downlaod fulklze Image

Figure13. Comparison between SDI and Ml SWSI in the Wabishebele river basin.

Jn the case of the Abbay river basin, the hydrological drought frequency was high for SDI from 1973 to 2000. But from
2000 to 2014, the frequency of drought was obtained high by MlSWSI (Figure6). The result revealed by some
resean:hers such as (71] indicates that the basin was under severe drought from 1980 to 1990.

SDJ and MlSWSI in Awash River have less correlation compared to Abbay and Baro river basins (Table6). Similarly. SDI
has a high frequency from 1973 to 2000 but it becomes wetter from 2001 to 2014 compared to MlSWSL The more
severe probability of the basin is obtained by SDI than MlSWSI (Figure7) (72. 73). also stared that the Awash river is
freciuently affecred by drought during the last five decades.

A drought study is not conducted in the Baro river basin before. The analysis of this study shows, the two hydrological
drought indices (SDI and MtSWSI) have a good correlation for the Baro river basin (Table6). Baro river is the wettest
river basin compared to all basins and the occurrence of drought frequency is less. However, the basin was severely
affected by drought between 1985 and 2002 (Figure8~

Jn Cenale Dawa. the two indices have high variation and from 1984 to 2007 moderare to severe droughts were more
freciuent for MtSWSI than SDI. It is because the Cenale Dawa river basin ls located in arid and semiarid areas in which
the contribution of rainfall for swface water ls minimal (74). But the source of streamflow Is from the highlands of the
central parts of the cowttty and depends mainly on a good dimam zone. As a result, SDI ls relatively wetter than
M1SWSI during the analysis time (Flgure9). The severe drought event occurred ln the basin period in 2003 for both SDI
and M1SWSI. Seven moderate drought events occurred in the Cenale Dawa river basin from 1984 to 2012. 43% of the
drought occurred during 1990, 1996, and 2007 (MlSWSI), 43% were in 1998, 1999, and 2001 (by SDI and M1SWSI), and
14% Is during 2008 (SDI). lbis result is the same as previous drought studies in the basin by (5, 55, 56. 57, 75, 58).

Omo Clbe River basin is located in a good dlma~ zone, which received high rainfall in two seasons (Spring and
summer). As a result, the analysis of drought by different indices (SDI and MtSWSI) relatively gives the same result.
Figure10 shows that from 1980 tD 1993, SDI results in moderate ID severe drought whereas from 1994 to 2016 the
moderam and severe drought events were dominated by MlSWSI.

The Rift Valley River basin is highly dominated by lakes and received minimum rainfall. Many streamflows join into
different small lakes and the rivers flow restricted from flowing a distance. The basin is located in a depression area and
the precipitation variation is insignificant (76]. The fluctuation of lake level in this river basin is constant as stated by
(36). The drought analysis for this basin considered three input parameters such as streamflow, precipitation, and lake
level The result shov.rs that, MlSWSI is influenced by Jake level and which results in the basin being in normal to wetter
conditions (Figure11 ~ But SDI result implied that the area is highly affected by frequent drought and the correlation
between SDI and Ml SWSI is poor (Table6~ This implies that SDI is more suitable for hydrological drought analysis in
this basin than MlSWSI. Because other studies also show that (36, 77] the area is frequently affected by severe drought
but MtSWSI minimized the impact of drought in the area due to the combination of different surface water sources.

SDI and MlSWSI in the Tekeze river basin have a good correlation (Table&). As shown in Figure12, the drought time
series of the two indices have a similar fashion. But relatively, MlSWSI results in more drought frequency compared to
SDI. Due to the construction of the Teke.ze dam, the variation of streamflow in the Tekeze river is balanced (78). SDI
value ls dependent on streamtlow data: therefore, the result also depends on the fluctuation offtow. Hovrever, historical
drought studies imply that the area is frequently affected by the severe drought [31, 79]. But in this study, the annual
SDI12 value of the Tekeze river basin is in the reverses of previous studies. Therefore, the future drought condition of
the Tekeze basin needs a detailed study using different additional streamflow stations located in the tributaries of the
Tekeze river.

Wabishebele River basin is one of the arid basins and is commonly affected by severe drought compared to other river
basins (Awass., 2009). 1990 and 2011 are the most severe drought years whereas 2002 was the extreme drought year in
this river basin. From those drought events, only 2011 was the same for SDI and MlSWSI and the remaining drought
events were obtained by SDI (Figure13).

5. Conclusion
This study has compared two hydrological drought indices (SDI and MlSWSI) and the result showed that the 1980s
were the most prolonged hydrological drought event years in Ethiopia. SDI resulted in more frequently severe and
extreme drought events occurrence over the country than the MlSWSI. MlSWSI uses multi-hydrological components
cumulatively as input and which results in a small magnitude of drought severity and its occurrence frequency was
decreased due to the aggregation of components. The number of severe drought events obtained using Ml SWSI in all
basins was less than 53.85% of the SDI result. Because some hydrological components were dominating the scarce data
and which will affect the overall analysis. The result of SDI values agreed with previous historical drought events.
Therefore, SDI is the best hydrological drought index compared to MlSWSI for all basins in Ethiopia and this index gives
good information for a single river as well as basin-level drought conditions. So, water resource managers and
infrastructure development sectors can use this index for decision-making for the best utilization of the available water
resource within the basin. Water supply, irrigation, and hydropower projects are more dependent on streamflow and
need hydrological drought monitoring.system development. Therefore, decision-makers, policy, and strategic planners,
and master plan developers can use SDI for historical and future hydrological drought analysis to develop effective
drought miti1J9.tion measures in Ethiopia.

Declarations

Author contribution statement

Kassa Abera Tareke, Msc; Admasu Gebeyehu Awoke: Conceived and designed the experiments; Performed the
experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the
paper.

Funding statement
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit
sectors.

Data availability statement

Data will be made available on request.

Declaration of interest's statement

The authors declare no competing interests.

Additional information
No additional information is available for this paper.

Special issue articles Recam mended articles

References
[1] S. Budhakooncharoen
Floods and droughts : sustainable prevention and management
Transac. Ecol. Environ., 64 (2003), p.10

Google Scholer 71

[2] S. Prabnakorn
Integrated Flood and Drought Mitigation Measures and Strategies : Case Study : the Mun River Basin
, Thailand Case Study: the Mun River Basin, Thailand Saowanit Prabnakom, No. May
CRC Press (2020)

Google Scholer 71

[3) A. Araya, L. Stroosnijder

Agricultural and Forest Meteorology Assessing drought risk and irrigation need in northern Ethiopia
Agric. For. MeteoroL, 151 (4) (2011), pp. 425-436

~ View PDF View article View in Scopus 71 Google Scholar 71

[4] A. Hamdy
Drought planning and drought mitigation measures in the Mediterranean region
Drought Manag., 239 (2008), pp. 235-239
Zaragoza : CIHEAM
Google Scholer 71

[SJ G.A. Mera

Drought and its impacts in Ethiopia
Weather Clim. Extrem., 22 (2018), pp. 24-35

~ View PDF View article View in Scopus 71 Google Scholar 71

[6] P. Alegre, E. T. De Jesus, R. junqueira, and M. R Viola, "Meteorological and Hydrological Drought from 1987 to
2017 in Doce River Basin , Southeastern Brazil," Scientific{fechnical Article • RBRH 25 • 2020. pp. 1-10.
Google Scholer 71

[7] Y. Mohammed, A. Yi mom

"Analysis of meteorological droughts in the Lake's Region of Ethiopian Rift Valley using
reconnaissance drought index (RDI),"
Geoenviron. Disasters, 8 (l) {2021), pp. l-16

View in Scopus 71 Google Sc:holar 71

[8] B. Boudad, H. Sahbi, I. Manssouri

Analysis of meteorological and hydrological drought based in SPI and SDI index in the lnaouen Basin
(Northern Morocco)
J. Mater. Environ. Sci., 9 (1) (2018), pp. 219-227

View artide 71 CrossRef 71 View in Scopus 71 Google Scholer 71

[9) X. Liu, X. Xu, M. Yu, J. Lu

Hydrological drought forecasting and assessment based on the standardized stream index in the
Southwest China
Proc:edia Eng., 154 (2016), pp. 733-737

'ibl View PDF View article View in Scopus 71 Google Scholar 71

[10] H.A.J. Van Lanen, N. Wanders, L.M. Tallaksen, A.F. Van Loon, C.H.A.J. Van Lanen
Hydrological drought across the world : impact of climate and physical catchment structure
HydroL Earth Syst. Sci., 17 (2013), pp.1715-1732

View article 71 CrossRef 71 View in Scopus 71 Google Scholer 71

[11] Y. Liu, et al
Evolution of hydrological drought in human disturbed areas : a case study in the laohahe catchment
, northern China
Adv. Meteorol., 2016 (2016)
Google Scholar 71

[12] H. Kwon, S. Kim

Evaluation of semi-distributed hydrological drought using SWSI (surface water supply index)
J. Korean Soc. Agricul. Engineers, 48 (2) (2006), pp. 37-43

View article 71 CrossRef 71 Google Scholar 71

[13] A. F. Van Loon, G. Laaha

Hydrological drought severity explained by climate and catchment characteristics
J. Hydro!., 526 (2015), pp. 3-14
~ View PDF View article View in Scopus 71 Google Scholar 71

[14] C. Sur, D. Kim,J. Lee, M.M. Iqbal, M. Choi

Hydrological drought assessment of energy-based water deficit index (EWDI) at different

geographical regions
Adv. Meteorol., 2019 (2019)
Google Scholar 71

[15] H. Kwon, S. Kim

Assessment of distributed hydrological drought based on hydrological unit map using SWSI drought
index in South Korea
KSCE J. Civ. Eng., 14 (2010), pp. 923-929
View article 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[16] S. Ho, L. Tian, M. Disse, Y. Tuo

Anew approach to quantify propagation time from meteorological to hydrological drought
J. Hydro!., 603 (PB) (2021), Article 127056
'OCJ View PDF View article View in Scopus 71 Google Scholar 71

[17] K. Winkler, U. Gessner, V. Hochschild

Identifying droughts affecting agriculture in Africa based on remote sensing time series between
2000 - 2016 : rainfall anomalies and vegetation condition in the context of ENSO
Remote Sensing, 9 (8) (2017), p. 831
View article 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[18] C. Bhuiyan
Various drought indices for monitoring drought condition in aravalli terrain oflndia
Proceedings of the 20th ISPRS Congress (2004)
Istanbul
Google Scholar 71

[19] B.S. Beyene, H.A.J. Van Lanen, P.j.J.F. Torfa

Drought classification , propagation analysis , and prediction of anomalies
Conference: Technical Report No. 37 {2015)

Google Scholar 71

[20] T.T. Zeleke

Trend and periodicity of drought over Ethiopia
Int. J. Climatol., 37 {13} (2017), pp. 4733-4748
View article 71 CrossRef 71 View in Scopus 71 Google Scholar 71
[21] E. Viste, D. Korecha, A. Sorteberg
Recent drought and precipitation tendencies in Ethiopia recent drought and precipitation tendencies
in Ethiopia ellen viste , diriba korecha , Asgeir Sorteberg
Theoret. Appl. Climatol., 112 (3-4) (May 2013), pp. 535-551
View article 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[22] A. Teshome
Increase of extreme drought over Ethiopia under
Adv. Meteorol. (2019)
Google Scholar 71

[23] A.Ali
Economics of Resilience to Drought - Somalia Analysis
Us Aid (2017), pp. 1-45
Google Scholar 71

[24] A.F. Van Loon

Hydrological drought explained
Wiely Intersciplinory Reviews: Water, 2 (4) (2015)

Google Scholar 71

[25) J. Wu, H. Yoo, X. Yu on, B. Lin

Dissolved organic carbon response to hydrological drought characteristics: based on long-term
measurements of headwater streams
Water Res., 215 (2022), Article 118252
i!J View PDF View article View in Scopus 71 Google Scholar 71

[26] S. Philip, et al
Attribution analysis of the Ethiopian drought of 2015
J. Clim., 31 (6) (2018), pp. 2465-2486
~ View PDF 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[27] I. Masih, 5. Maskey, P. Trambauer

Areview of droughts in the African continent: a geospatial and long-term perspective
Hydrol. Earth Syst. Sci. Discuss., 11 (2014), pp. 2679-2718

View article 71 CrossRef 71 Goo9le Scholar 71

[28] A. Teshome, J. Zhan9

Increase of extreme drought over Ethiopia under climate warming
Adv. Meteorol., 2019 (2019)
Google Scholer 71

(29) T.T. Zeleke, F. Giorgi, G.T. Diro, B.F. Zaitchik

Trend and periodicity of drought over Ethiopia
Int. J. Climatol., 37 {13) {2017), pp. 4733-4748
View article 71 CrossRef 71 View in Scopus 71 Google Scholer 71

[30] D. Chemeda, E. Mukand, S. Babel

Drought analysis in the Awash river basin , Ethiopia drought analysis in the Awash river basin ,
Ethiopia
Water Resour. Manag., 24 (2010), pp.1441-1460

Google Scholar 71

[31) T. Gebrehiwot, A. van derVeen, B. Maothuis

 Spatial and temporal assessment of drought in the Northern highlands of Ethiopia
Int. J. Appl. Earth Obs. Geoinf., 13 (3) (2011), pp. 309-321
~ View PDF View article View in Scopus 71 Google Scholar 71

[32] T. Gissila, E. Black, D.I.F. Grimes, J.M. Slingo

Seasonal forecasting of the Ethiopian summer rains
Int. J. Climatol., 24 (11) (2004), pp. 1345-1358
View in Scopus 71 Google Scholar 71

[33] A.A. Awass

Hydrological Drought Analysis - Occurrence , Severity , Risks : the Case of Wabi Shebele River Basin
Thesis, Dissertation
(2009)
Google Scholar 71

[34] E.B. Van Lanen Hoj

Assessment of the impact of climate change on hydrological drought in lake tana catchment, Blue
nile basin, Ethiopia
J. Geol. Geosci., 3 (6) (2014)
Google Scholar 71

[35] Y. Mohammed, F. Yimer, M. Tadesse, K. Tesfoye

Meteorological drought assessment in north east highlands of Ethiopia
Int. J. Clim. Chang. Strateg. Ma nag., 10 (1) (2018), pp. 142-160
View article 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[36] Y. Mohammed
Analysis of meteorological droughts in the Lake 's Region of Ethiopian Rift Valley using
reconnaissance drought index (RDI)
Geoenvironmentol Disasters, 8 (2021), pp. 1-16
View in Scopus 71 Google Sr:holar 71

[37] E. Lemma, S. Upadhyaya, R. Ramsankaran

Meteorological drought monitoring across the main river basins of Ethiopia using satellite rainfall
product
Environ. Syst. Res. (2022)
Google Scholar 71

[38] A.A. Mekonen, A.B. Berlie, M.B. Ferede

Spatial and temporal drought incidence analysis in the northeastern highlands of Ethiopia
Geonevironmental Disasters, 7 (1) (2020)
Google Scholar 71

[39] B. Gebrehiwot, B. Gessesse, F. Melgani

Characterizing the spatiotemporal distribution of meteorological drought as a response to climate
variability : the case of rift valley lakes basin of Ethiopia
Weather Clim. Extrem., 26 (2019), Article 100237
Google Scholar 71

[40] H.P. Zone, A. Hirko, G. Mergia, A. Nigussie, T. Dandesa

Seasonal and annual meteorological drought frequency: case study east hararge province (zone)
Int. J. Res. Environ. Sci., 7 (l) (2021), pp. 10-19
Google Scholar 71
[41] Getachew Mehabie Mulualem, Yuei-An Liou
Application of artificial neural networks in forecasting a standardized precipitation
evapotranspiration index for the upper Blue basin
J.fWater, 12 (3) (2020)

Google Scholar 71

[42] D.T. Tsige, V. Uddameri, F. Forghanparast, E.A. Hernandez, S. Ekwaro-Osire

Comparison of meteorological and agrirulture-related drought indicators across Ethiopia
Water, 11 (11) (2019), pp. 1-24

Google Scholar 71

[43] Y. Bayissa, et al.

Comparison of the performance of six drought indices in characterizing historical drought for the
upper Blue Nile Basin, Ethiopia
Geosci., 8 (3) (2018)

Google Scholar 71

[44] J. Wu, X. Chen, H. Yao, D. Zhang

Multi-timescale assessment of propagation thresholds from meteorological to hydrological drought
Sci. Total Environ., 765 (2021), Article 144232

1!J View PDF View article View in Scopus 71 Google Scholar 71

[45] J. Wu, X. Chen, X. Yuan, H. Yao, Y. Zhao, A. AghaKouchak

The interactions between hydrological drought evolution and precipitation-streamflow relationship
J. Hydrol., 597 (2021), Article 126210
~ View PDF View article View in Scopus 71 Google Scholar 71

[46] G.G. Halle

Irrigation in Ethiopia, a review
J. Environ. Earth Sci., 3 (10) (2015), pp. 264-269
Google Scholar 71

[47] K.A. Toreke, A.G. Awoke

Hydrological drought analysis using streamflow drought index (SDI) in Ethiopia
Adv. Meteorol., 2022 (2022)

Google Scholar 71

[48] D. Tigkas, H. Vangelis, G. Tsakiris

DrinC: a software for drought analysis based on drought indices
Earth Sci. Informatics, B (3) (2015), pp. 697-709

View article 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[49] J.F. Agwata

Areview of some indices used for drought monitoring
Civ. Environ. Res., 6 (2) (2021), pp. 14-21

Google Scholar 71

[so] T. Akkurt Erogluer, H. Apaydin

Estimation of drought by streamflow drought index (SDI) and artificial neural networks (ANNs) in
ankara-nallihan region
Turkish]. Agric. - Food Sci. Technol., 8 (2) (2020), p. 348

View ortide 71 CrossRef 71 Google Scholar 71

[51] B. Berhanu, Y. Seleshi, S.S. Demisse, A.M. Melesse

 Bias correction and characterization of climate forecast system re-analysis daily precipitation in
Ethiopia using fuzzy overlay
Meteorol. Appl., 23 (2) (2016)
Google Scholar 71

[52] N.P. Vidal, C.F. Manful, T.H. Pham, P. Stewart, D. Keough, R. Thomas
MethodsX the use of XLSTAT in conducting principal component analysis (PCA) when evaluating the
relationships between sensory and quality attributes in grilled foods
MethodsX, 7 (2020), Artide 100835
'ibJ View PDF View article View in Scopus 71 Google Scholar 71

[53] Q. Zhao, Y. Zhu, D. Wan, Y. Yu, X. Cheng

Research on the data-driven quality control method of hydrological time series data
Water (Switzerland), 10 (12) (2018)
Google Scholar 71

[54] N. Branisavijevi, D. Prodanovi, M. Arsi, Z. Simi,]. Borota

Hydro-meteorological data quality assurance and improvement
Journal of the Serbian Society for Computational Mechanics/, 3 (1) (2009), pp. 228-249
https://g rofa r.grf. bg .ac.rs/hon dle/123456789/223 71
Google Scholar 71

[55] T. Kocsis
Homogeneity tests and non-parametric analyses of tendencies in precipitation time series in
Keszthely , Western Hungary
Theor. Appl. Climatol., 139 (2) (2020), pp. 849-859
'i!J View PDF 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[56] J. Sillmonn
Consistency of temperature and precipitation extremes across various global gridded in situ and
reanalysis datasets
J. Clim., 27 (13) Ouly 2014), pp. 5019-5035
Google Scholar 71

[57] N. J. Doesken and T. B. Mckee, "Development of a Surface Water Supply index for the Western United States,"
Colorado Climate Center Department of Atmospheric Science Colorado State University Fort Collins, ColoradoBiz
80523 (303) 491-8545, Climatology Report #91-3, October," 1991.
Google Scholar 71

[58] R. MuliWambua
Analysis of drought and wet-events using SWSI-based severity-duration-frequency (SDF) curves for
the upper tana river basin, Kenya
Hydrology, 6 (2) (2018), p. 43
View article 71 CrossRef 71 Google Scholar 71

[59] R.M. Wambua

Hydrological drought forecasting using modified surface water supply index (SWSI) and streamflow
drought index (SDI) in conjunction with artificial neural networks (ANNs)
Int.]. Serv. Sci. Monag. Eng. Technol., 10 (4) (2019), pp. 39-57
View article 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[60] D.C. Goren, D. Ph

The Surface Water Supply Index : Formulation and Issues Snow Survey and Water Supply
Forecasting Program
 (2011), pp. 1-31
View in Scopus 71 Google Scholar 71

[61] 0.B. Adeboye, M.0. Alatlse

Performance of probability distributions and plotting positions in estimating the flood of river osun
at apoje sub-basin , Nigeria
CIGR E-J., 9 (2007), pp. 1-21
https:flhdl.handle.net/1813/10689 71
Google Scholar 71

[62] I.N.G. Tsakiris

Assessment of Hydrological Drought Revisited
23, Water Resources Management (April 2007), pp. 881-897
Google Scholar 71

[63] W. Wan, J. Zhao, H. Li, A. Mishra

Hydrological drought in the anthropocene: impacts of local water extraction and reservoir
regulation in the U.S
J. Geophys. Res. Atmos., 122 (2017), pp.11,313-11,328
Google Scholar 71

[64] A.T.G.A. Kidane Giorgis

Agricultural Development in Drought-Prone Areas of ethiopia: Interventions, Challenges and Good
Practice
AKLPD Assessment Report, September (2018)
Google Scholar 71

[65] B.N. Wolteji, S.T. Bedhadha, S.L. Gebre, E. Alemayehu, D.O. Gemeda
Multiple indices based agricultural drought assessment in the Rift valley region of Ethiopia
Environ. Challenges, 7 (2022), Article 100488
i!J View PDF View article View in Scopus 71 Google Scholar 71

[66] Y.A. Bayissa

Developing an Impact-Based Combined Drought Index for Monitoring Crop Yield Anomalies in the
Upper Blue Nile Basin, Ethiopia
CRC Press (2018), p. 168
Google Scholar 71

[67] FeyisaSebokaTura
Climate variability and drought in the past 30 Years in central Rift Valley of Ethiopia
J. Nat. Sci. Res., 7 (9) (2017), pp.18-26
[Online]. Available: ISSN 2224-3186 (Paper) ISSN 2225-0921 (Online
Google Scholar 71

[68] J. Nasir, E. Assefc, T. Zeleke, E. Gidey

ELSEVIER

All content on this :sill!: Copyright© 2024 Elsll!!Yi'l!!r B.V.1 its litl!!!nsors, end contributors. All rights ore reserved. induding those far tut ond dote mining, Al training, and similorachnologil!s. For oll
open access content, the creative commons licensing tonns apply.
&ffELXrN

View in Scopus 71 Google Scholar 71

[70] A.W. Teshale, M.D. Olsedo
Temporal drought assessment of lake hawassa sub-basin, central main rift V alley basin of Ethiopian
J. Earth Sci. Climatic Change, 12 (l} (2021), pp. l-3
View article 71 CrossRef 71 Google Scholar 71

[71] Y.A. Bayissa, et al.

Developing a satellite-based combined drought indicator to monitor agricultural drought : a case
study for Ethiopia
GIScience Remote Sens. (2018), pp. 1-31
View article 71 CrossRef 71 Google Scholar 71

[72] D.C. Edosso, M.S. Babel, A. Das Gupta

Drought analysis in the Awash river basin, Ethiopia
Water Resour. Manag., 24 (7) (2010), pp.1441-1460
View article 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[73] A. Belayneh, J. Adamowski, B. Khalil, B. Ozga-zielinski

Long-term SPI drought forecasting in the Awash River Basin in Ethiopia using wavelet neural
network and wavelet support vector regression models
J. Hydro!., 508 (2014), pp. 418-429
1!J View PDF View article View in Scopus 71 Google Scholar 71

[74] N. Kassahun, M. Mohamed

Groundwater potential assessment and characterization of Genale-Dawa river basin
Open J. Mod. Hydrol., 8 (4) (2018), pp. 126-144
View article 71 CrossRef 71 Google Scholar 71

[75] C. David
Garen . Member . ASCE . "Revised surface-water supply index for western United States
J. Water Res. Plann. Manog., 119 (4) (1993}, pp. 437-454
Google Scholar 71

[76] M.D. Belete, B. Diekkri.iger,J. Roehrig

Characterization of water level variability of the main Ethiopian Rift Valley lakes
Hydrology, 3 (l) (2016), pp. 1-14
View in Scopus 71 Google Sc:holor 71

[77] B. Yisehak, K. Ad hen a, H. Shiferaw, H. Hagos, H. Abrha, T. Bezabh

Characteristics of hydrological extremes in kulfo river of southern Ethiopian Rift Valley basin
SN Appl. Sd., 2 (7)(2020}, pp. 1-12
Google Scholar 71

[78] F. Fentaw, D. Hailu, A. Nigussie, A.M. Melesse

Climate change impact on the hydrology ofTekeze basin, Ethiopia: projection ofrainfall-runoff for
future water resources planning
Water Conserv. Sci. Eng., 3 (4) (2018), pp. 267-278
View article 71 CrossRef 71 View in Scopus 71 Google Scholar 71

[79] T. Gebrehiwot, A. Von Der Veen

Assessing the evidence of climate variability in the northern part of Ethiopia
J. Dev. Agric. Econ., 5 (3) (2013), pp.104-119
View article 71 CrossRef 71 Google Scholar 71
Cited by (4)

Deep learning versus hybrid regularized extreme learning machine for multi-month drought forecasting:
A comparative study and trend analysis in tropical region
2024, Hellyon

Show abstract v

Hydrological Drought Severity in Different Return Periods in Rivers of Ardabil Province, Iran 71
2023, Sustainability (Switzerland)

Evaluation of Drought Indexes in the Hydrological Monitoring of Strategic Reservoirs in Ceara, Brazil 71
2023, Revisto Brosileiro de Meteorologic

Characterization of hydro-meteorological droughts based on dynamic future scenarios and effective

rainfall over Central Malawi 71
2023, Theoretic;al and Applied Climatology

© 2022 The Author(s). Published by Elsevier Ltd.