Extreme Drought Assessment in Sumatra-Indonesia Using SPI and EDI

Abstract

Even though Sumatra is very vulnerable to the effects of drought, research identifying drought on this island is very limited, especially extreme drought. This research was conducted to identify extreme drought in Sumatra using the Standardized Precipitation Index (SPI) and the Effective Drought Index (EDI). This study uses precipitation data from the Indonesian Meteorology, Climatology and Geophysics Agency (BMKG) and Global Precipitation Climatology Center (GPCC) reanalysis data. The composite indices were conducted to discover some phenomena that cause the drought based on El Niño and positive Indian Ocean Dipole (IOD) events. The results showed that the El Niño and positive IOD phenomena were more likely to influence extreme droughts. However, the droughts in 2014 and 2008 tended to be influenced by the negative Sea Surface Temperature anomaly (SSTA). The spatial analysis results show that the areas that experience extreme drought more often are the west coast of Sumatra (except Aceh province), especially based on SPI12 and EDI. The composite indices results show that the drought that occurred in Aceh province was more influenced by the El Niño phenomenon in the December-January-February (DJF) period and the positive IOD in the June-July-August (JJA) – September-October-November (SON) period. In addition, Aceh province is an area that is not affected by El Niño-positive IOD, especially during the SON-DJF period. These results can be used to mitigate drought, especially when El Niño-positive IOD phenomena cause it.

Full text

Science and Technology Indonesia
e-ISSN:2580-4391 p-ISSN:2580-4405
Vol. 8, No. 4, October 2023
Research Paper
Extreme Drought Assessment in Sumatra-Indonesia Using SPI and EDI
Suhadi1, Iskhaq Iskandar2*, Supari3, Muhammad Irfan2, Hamdi Akhsan1
1Graduate School of Sciences, Faculty of Mathematics and Natural Sciences, Sriwijaya University, 30862, Indonesia
2Department of Physics, Faculty of Mathematics and Natural Sciences, Sriwijaya University, 30862, Indonesia
3Indonesia Agency for Meteorology, Climatology and Geophysics BMKG, Jl.Angkasa I No 2 Kemayoran, Jakarta, 3540, Indonesia
*Corresponding author: iskhaq@mipa.unsri.ac.id
Abstract
Even though Sumatra is very vulnerable to the effects of drought, research identifying drought on this island is very limited, especially
extreme drought. This research was conducted to identify extreme drought in Sumatra using the Standardized Precipitation Index
(SPI) and the Effective Drought Index (EDI). This study uses precipitation data from the Indonesian Meteorology, Climatology
and Geophysics Agency (BMKG) and Global Precipitation Climatology Center (GPCC) reanalysis data. The composite indices were
conducted to discover some phenomena that cause the drought based on El Niño and positive Indian Ocean Dipole (IOD) events.
The results showed that the El Niño and positive IOD phenomena were more likely to influence extreme droughts. However, the
droughts in 2014 and 2008 tended to be influenced by the negative Sea Surface Temperature anomaly (SSTA). The spatial analysis
results show that the areas that experience extreme drought more often are the west coast of Sumatra (except Aceh province),
especially based on SPI12 and EDI. The composite indices results show that the drought that occurred in Aceh province was more
influenced by the El Niño phenomenon in the December-January-February (DJF) period and the positive IOD in the June-July-August
(JJA) – September-October-November (SON) period. In addition, Aceh province is an area that is not affected by El Niño-positive
IOD, especially during the SON-DJF period. These results can be used to mitigate drought, especially when El Niño-positive IOD
phenomena cause it.
Keywords
Extreme Drought, Sumatra, Indonesia, Standardized Precipitation Index, Effective Drought Index, El Niño, Positive IOD
Received: 11 May 2023, Accepted: 5 September 2023
https://doi.org/10.26554/sti.2023.8.4.691-700
1. INTRODUCTION
As an annual natural disaster that impacts various sectors of
life (WMO,2016), drought is a condition with lower precipi-
tation than usual in certain areas and periods (WMO,2012).
Referring to the impact of drought, the four types of drought
dened by the World Meteorological Organization (WMO) are
Meteorological Drought, Agricultural Drought, Hydrological
Drought, and Socio-Economic Drought (Svoboda and Fuchs,
2016;WMO,2016). The four droughts usually occur sequen-
tially, depending on how long the drought lasts. Meteorological
drought is inuenced by precipitation as the main indicator
(Abbas et al.,2021;Bougara et al.,2021;Eslamian et al.,2017;
Salehnia et al.,2017) is the drought that occurs in the rst stage.
This drought then continues as an agricultural drought (Adisa
et al.,2021), marked by reduced soil moisture and causing
damage to agriculture. The third stage of this drought is a
hydrological drought marked by the diminishing water supply
in the river. The third stage of this drought is hydrological
drought, marked by the diminishing supply of water in the
river (Eslamian et al.,2017;Zargar et al.,2011). In the end,
this drought will impact the socio-economic sector. Not only
in the socio-economic sector but in some serious cases, such
as in Africa, the impact of this drought has resulted in death
(Adisa et al.,2021). One of the eorts that can be made to
mitigate the impact of the drought is a drought assessment.
In conducting a drought assessment, there are at least 23
meteorological drought indices with diculty levels of use,
namely easy, medium, and dicult (Svoboda and Fuchs,2016;
WMO,2016). The ease of use of these various indices is usu-
ally determined based on the input parameters used. Apart
from this level of convenience, of course, this use is due to the
advantages and disadvantages of the indices, which usually refer
to the topographical conditions of an area (Adisa et al.,2021) .
Several studies have been conducted to analyze a suitable index
for assessing a drought that has occurred. For example, the Per-
cent Normal Index (PN) and Rainfall Anomaly Index (RAI)
detect more droughts occurring in the Tafna-Northwestern
Algeria watershed than the Standardized Precipitation Index

Suhadi et. al. Science and Technology Indonesia, 8 (2023) 691-700
(SPI) and Decile Index (DI). Bougara et al. (2021) show that
PN and RAI detect more than droughts detected by SPI. Mean-
while, EDI is more sensitive than SPI in detecting drought in
Bangladesh (Kamruzzaman et al.,2019) . However, SPI is an
index the WMO recommends to detect drought (Svoboda and
Fuchs,2016;WMO,2016).
Apart from being recommended by the WMO, SPI is a
widely used method (Chandrasekara et al.,2021) , which McKee
et al. (1993) introduced to detect meteorological drought in
Colorado. However, this index is weak because it only uses
precipitation to indicate drought. Another method that uses
precipitation as calculation input is EDI. Byun and Wilhite
(1999) introduced EDI to overcome the lack of SPI. EDI
uses a precipitation time reduction function, claimed as an
evapotranspiration calculation not used in SPI calculations.
As a new method, EDI has been used by several researchers
in detecting drought. In South Africa, EDI (and SPI) have
been used to detect drought (Adeola et al.,2021;Adisa et al.,
2021). In a recent study, the EDI was modied to determine
the impact of social drought on the Korean Peninsula (Park
et al.,2022) .
Sumatra is a region often aected by drought, especially in
the environmental sector (Davies and Unam,1999;Nurdiati
et al.,2021;Prinz,2009;Sarmiasih and Pratama,2019) in
the form of forest and land res. Even Supari et al. (2016)
show that Sumatra is drought-prone, especially in July-October.
In 1997/98, several areas of Sumatra (namely, South Suma-
tra, Riau, and Jambi) experienced drought-related forest res
(Prinz,2009). This re then causes an eect in the form of
an increase in air pollutants such as SO
2
, CO, CH
4
and CO
2
(Davies and Unam,1999). Apart from 1997/98, the impact of
a more severe drought occurred in 2015/16. The drought that
happened that year caused forest res covering an area of 2.6
million ha and was a wider re than the res in 2019, which
caused forest res covering an area of 328,722 ha (Sarmiasih
and Pratama,2019). The extent of this burned area is also
indicated based on the increase in hotspots in South Sumatra
(Nurdiati et al.,2021). Even though the impact of this drought
is very signicant, research related to drought, especially the
extreme drought in Sumatra, is still limited. Therefore, car-
rying out a drought assessment as a disaster mitigation eort
is necessary. So far, most of the drought identication has
been carried out in the central to eastern parts of Indonesia
(Kuswanto et al.,2019;Nurdiati et al.,2021;Pramudya et al.,
2019), particularly in Java and Kalimantan. Therefore, this
research used the SPI and EDI methods to identify drought,
especially extreme drought in Sumatra.
2. EXPERIMENTAL SECTION
2.1 Location
The research area is Sumatra, as shown in Figure 1. In the west-
ern part of Indonesia, this island is an area with precipitation
variability triggered by the Asia-Australia monsoon (Iskandar
et al.,2011). Apart from Monsoon, ENSO - IOD is also a
phenomenon that aects the precipitation on this island during
Figure 1. The Area of Interest of Research. The Numbers
Indicated in Yellow are Rainfall Stations
Figure 2. Monthly Climatology in 9 Rainfall Station
interannual periods (Aldrian and Dwi Susanto,2003;Avia and
Soati,2018;Saji and Yamagata,2003;Iskandar et al.,2017a;
Supari et al.,2016). In addition to the precipitation variabil-
ity, several unique precipitation patterns occur on this island,
especially in the northern region (Darlan et al.,2020). This
uniqueness changes the opposite pattern from other regions;
when the Southern region enters the wet season, this part ex-
periences drought and vice versa (Iskandar et al.,2017b). In
contrast, this uniqueness occurs in the SON period of El Niño
(Qian,2020). Topographically, contrasting elevation dier-
ences along the western and eastern coasts are also visible on
this island. In the part of the west of the island lies the Bukit
Barisan, which stretches from north to south. While in the
east, there are many rivers and swamps.
2.2 Materials
In this study, we used precipitation station data from the BMKG
website, namely http://dataonline.bmkg.go.id. At this site, 33
©2023 The Authors. Page 692 of 700

Suhadi et. al. Science and Technology Indonesia, 8 (2023) 691-700
Figure 3. Drought Indices in the Sultan Iskandar Muda,
Malikussaleh, and Deli Serdang Rain Gauges. The Horizontal
Green, Yellow, and Red Lines are Moderate, Severe, and
Extreme Drought Limits
rainfall stations spread along Sumatra. However, due to the
availability of data at each station and the need for data for
at least 30 years, this study only used nine stations, as shown
in Table 1. The period of this data is from January 1990 to
December 2019. Precipitation data from these BMKG stations
will be used to identify extreme drought time series in Sumatra
from January 1990 to December 2019.
In addition to using BMKG precipitation data, this research
uses Global Precipitation Climatology Center (GPCC) reanal-
ysis data. This reanalysis data has various resolutions, namely
0.25°, 0.5°, 1°and 2.5°. However, this study uses data with
a resolution of 0.25°from January 1990 to December 2019.
This reanalysis data will be used to see the frequency distribu-
tion of extreme droughts over 30 years.
2.3 Methods
2.3.1 SPI
SPI is a method that determines precipitation conditions based
on the gamma distribution function (Edwards and McKee,
1997;McKee et al.,1993). Calculation of precipitation condi-
Figure 4. Same with Figure 2, but for the FL Tobing,
Minangkabau, and Japura Rain Gauges
tions is done through several steps. The rst step in calculating
the SPI is to t the precipitation data (at least 30 years) to the
cumulative gamma distribution function (Adisa et al.,2021;
Huang et al.,2016).
G(Pj)=1
𝛽𝛼Γ(𝛼)∫pj
0
P𝛼−1
je−Pj/𝛽dPj(1)
where
𝛼
and
𝛽
are the form factor and scale factor respec-
tively, P
j
is the accumulated precipitation for j-month, and the
gamma function is marked with
Γ
(
𝛼
). One characteristic of the
gamma function is that it cannot be calculated when P
j
=0. To
overcome this, the cumulative distribution function G(P
j
) is
changed to
H(pj)=q+ (1−q)G(Pj)(2)
where q is the probability of zero precipitation, after the
cumulative distribution function H(s) is standardized, the nal
step is to convert this cumulative distribution function into
an SPI value using the following approximation technique
(Abramowitz and Stegun,1964) .
©2023 The Authors. Page 693 of 700

Suhadi et. al. Science and Technology Indonesia, 8 (2023) 691-700
Figure 5. Same with Figure 2, but for the Sultan Thaha,
Palembang, and Raden Intan Rain Gauges
SP I =−t−c0+c1t+c2t2
1+d1t+d2t2+d3t3;
t=ln 1
H(Pj)2for 0<H(Pj)<0.5 (3)
SP I =+t−c0+c1t+c2t2
1+d1t+d2t2+d3t3;
t=ln 1
(1−H(Pj))2for 0.5 <H(Pj)<1(4)
with c
0
= 2.515517, c
1
= 0.802853, c
2
= 0.010328, d
1
=
1.432788, d
2
= 0.189269, d
3
= 0.001308. SPI can be calcu-
lated using accumulated precipitation at various time scales,
namely 1, 3, 6, 9, 12, and 24 months. However, in this study,
the accumulated precipitation used is 3, 6, and 12 months to
analyze drought on a 3-month, 6-month, and 12-month scale.
Figure 6. (a) Sea Surface Temperature, and (b) Precipitation
Anomaly in 2008
Figure 7. Extreme Drought Frequency of SPI3, SPI6, SPI12,
and EDI
2.4 EDI
EDI was introduced by Byun and Wilhite (1999) to identify
precipitation conditions based on eective precipitation (EP).
EPi=
N

m=1©«
m

j=1
Pjª®¬/j(5)
The EDI value is calculated based on the ratio between the
precipitation needed to return to normal (Precipitation Return
to Normal - PRN) and its standard deviation.
EDIi=PRNi
𝜎(PRNi)(6)
𝜎
(PRN
i
) is the standard deviation of PRN in each calen-
dar month (standard deviation of PRN climatology). PRN is
calculated based on Jain et al.,2015;Raude et al.,2018.
©2023 The Authors. Page 694 of 700

Suhadi et. al. Science and Technology Indonesia, 8 (2023) 691-700
Table 1. Rainfall Stations
Rainfall Stations Lat Lon
Sultan Iskandar Muda 5.52 95.42
Malikussaleh 5.23 96.95
Deli Serdang 3.62 98.71
FL Tobing 1.55 98.88
Minangkabau -0.79 100.29
Japura -0.33 102.32
Sultan Thaha -1.63 103.64
Palembang -2.98 104.77
Raden Inten -5.16 105.11
PRNi=DEPi/©«
N

j=1
1
jª®¬(7)
DEPi=EPi−¯
EPl(8)
DEP and
¯
EPl
are the deviation of the eective precipitation
and the average precipitation in each calendar month (clima-
tological average), respectively. Precipitation conditions (dry
or wet) identied using SPI and EDI are determined based on
the scale shown in Table 2.
Table 2. SPI and EDI Scales
Conditions SPI/EDI
Extremely Drought ≦-2.00
Severely Drought -1.5 to -1.99
Moderately Drought -1.0 to -1.49
Normal -0.99 to 0.99
Moderately Wet 1.0 to 1.49
Severely Wet 1.5 to 1.99
Extremely Wet ≧2.00
As is known, several factors related to drought are El Niño
and positive IOD. To nd out the relationship between this
phenomenon and the drought, we used the period of El Niño
positive and IOD from 1990 to 2020. We obtained data for
positive IOD and El Niño by calculating the Sea Surface Tem-
perature (SST) in the IOD and Niño3.4 blocks. This SST
data is ERA5 reanalysis data with a resolution of 0.25°. After
calculating, the positive IOD and El Niño events obtained are
shown in Table 3. El Niño and positive IOD conditions are
determined based on the Standard Deviation of Sea Surface
Temperature Anomaly (SSTA). We then use the period (sea-
son) that shows positive El Niño and IOD events as the basis
for composite calculations of SPI3, SPI6, SPI12, and EDI.
We only include months showing El Niño and positive IOD
phenomena in one season to ensure accurate composite results.
For example, El Niño in 2009 in SON started in October-
November, so the composite index (SPI3, SPI6, SPI12, and
EDI) was only carried out in October-November.
3. RESULTS AND DISCUSSION
3.1 Precipitation Climatology
The monthly precipitation climatology of the nine stations is
shown in Figure 2. These results indicate that in the period
from January to August (JFMAMJJA), precipitation at the Sul-
tan Iskandar Muda, Malikukksasaleh and Deli Serdang stations
was less than 200 mm. This condition is also shown by Qian
(2020) , which states that the northern region of Sumatra tends
to experience precipitation of more than 200mm during the
SON period. Apart from being due to the west monsoon winds,
which bring a mass of water from Asia to Australia, this is also
due to the more dominant westerly winds that blow from the
Indian Ocean Qian (2020) . Monthly precipitation recorded at
200mm/month-550mm/month at FL Tobing and Minangk-
abau stations indicates that these two stations experience higher
precipitation than the other seven stations. This higher precipi-
tation is because the Tobing and Minangkabau FL station areas
directly face the Indian Ocean, so the IOD phenomenon aects
this region (Iskandar et al.,2011;Lee,2015;Mori et al.,2004).
The highest precipitation at this station occurs in November,
and another peak in March. Precipitation patterns at Japura,
Sultan Thaha, and Palembang stations are the same as the
previous stations (FL Tobing and Minangkabau). These three
stations’ precipitation patterns also show peaks in March and
November.
Nonetheless, the precipitation is lower than at the FL To-
bing and Minangkabau stations. This condition is due to the
obstruction of the mass of seawater (brought by the westerly
wind) by the Bukit Barisan, which stretches along the western
side of Sumatra. Unlike the others, the peak of precipitation
at the Raden Intan station occurs in January (DJF season). As
previously explained, this condition was also shown by Qian
(2020) that the northern part of Sumatra tends to experience
higher precipitation at SON while the southern region at DJF.
If sorted by the amount of monthly precipitation, it can be
seen that the high precipitation respectively is TM (FL Tobing-
Minangkabau), JSPR (Japura-Sultan Thaha-Palembang-Raden
Intan), and IMD (Sultan Iskandar Muda-Malikussaleh-Deli
Serdang).
3.2 Temporal Analysis
The results of drought identication based on SPI3, SPI6,
SPI12, and EDI are temporally shown in Figure 3-5. Figure
3for Sultan Iskandar Muda, Malikussaleh and Deli Serdang
stations, Figure 4for FL Tobing, Minangkabau and Japura
stations, and Figure 5for Sultan Thaha, Palembang and Raden
Intan stations. As it is known that, generally, the drought
that occurs, especially in Indonesia, is caused by a positive
El Niño - IOD. For example, the droughts of 1997/98 and
2015/16. The El Niño - positive IOD that occurs simultane-
ously causes these droughts. In 1997/98, the drought reached
©2023 The Authors. Page 695 of 700

Suhadi et. al. Science and Technology Indonesia, 8 (2023) 691-700
Table 3. El Niño and Positive IOD Event. (A) for August, (ON) for Oct and November, (MA) for March and April, (M) for
March, (..M) for May, (JA) for July and August, (SO) for September and October, (D) for December, and (J) for January
Phenomena Periods Years
El Niño JJA 2002(A)
SON 1991(ON); 2002; 2009(ON)
DJF 1991/92; 2002/03(J); 2009/10
MAM 1992(MA); 2010(M)
Positive IOD JJA 2011(A); 2012(JA); 2019
SON 2011(SO); 2012(SO); 2019
El Niño and JJA 1994(JA); 1997; 2006(A); 2015
Positive IOD SON 1994; 1997; 2006; 2015
DJF 1994/95( J); 1997/98; 2006(D); 2015/16
MAM 1998(MA); 2015(..M); 2016(MA)
extreme levels at Malikussaleh stations (SPI3, SPI6), Deli Ser-
dang (SPI3, SPI6, SPI12), FL Tobing (SPI3, SPI6, SPI12,
EDI), Minangkabau (SPI3, SPI6, SPI12, EDI), Japura (SPI3,
SPI6, SPI12, EDI), Sultan Thaha (SPI3, SPI6, SPI12, EDI),
Palembang (SPI3, SPI6, SPI12, EDI), Raden Intan (SPI3,
SPI6, SPI12, EDI). Whereas in 2015/16, extreme drought
was detected at Sultan Iskandar Muda (SPI3), Malikussaleh
(SPI3), Deli Serdang (SPI6, EDI), Minangkabau (SPI3, SPI6),
Japura (SPI3, SPI6, EDI), Sultan Thaha (SPI3, SPI6, SPI12,
EDI), Palembang (SPI3, SPI6), and Raden Intan (SPI3). El
Niño in 2009/10 caused extreme drought to be detected at
Deli Serdang station (SPI6, SPI12), Japura (SPI3). Whereas
the positive IOD in 2012 caused extreme drought, only de-
tected at the Raden Intan station (SPI3). In addition to the
extreme drought caused by the positive El Niño-positive IOD
phenomena, several droughts unrelated to these phenomena
also occurred in 2008 at the Sultan Iskandar Muda station
(SPI3, SPI6), Malikussaleh (SPI6, EDI). In addition to the
extreme drought in 2008, the 2014 drought was not caused
by the El Niño-positive IOD phenomena but by the negative
SSTA (Iskandar et al.,2017a) .
In Figure 6a, it can be seen that a strong negative sea surface
temperature anomaly occurred in the May-June-July period.
This anomaly caused a reduction in the mass convection activity
of seawater into the atmosphere in that period (Figure 6b),
which in turn led to reduced precipitation, especially on the
island of Sumatra. This negative SSTA was also the cause of
the drought that occurred on the island of Java in 2008 (Nita
et al.,2020) .
3.3 Spatial Analysis
The results of identifying extreme droughts are spatially shown
based on the frequency of extreme droughts, as shown in Figure
7. This frequency indicates the number of extreme droughts
that have occurred in 30 years (1990-2019). Based on SPI3
calculations which describe precipitation conditions for three
months, extreme droughts ranging from 10-22 times occurred
almost along the west coast of Sumatra, more precisely in the
south of North Sumatra, West Sumatra, Bengkulu, and most
Figure 8. Composite of Indices in El Niño Events
of South Sumatra.
Drought analysis based on SPI6 shows that the frequency
of extreme droughts is as much as 10-25 times more visible in
the provinces of North Sumatra, West Sumatra, Bengkulu, to a
small part of South Sumatra. Most of the drought frequency in
this region has been detected based on SPI3, except for the east
coast of South Sumatra and Lampung, which were detected to
show a decrease in the frequency of extreme drought. SPI12
detects the number of extreme droughts around 20-25 times
occurring in the southern provinces of North Sumatra and
West Sumatra. The frequency distribution of extreme droughts
shown by EDI is the same as that shown by SPI12, and it is
just that the number of extreme droughts detected by EDI is
©2023 The Authors. Page 696 of 700

Suhadi et. al. Science and Technology Indonesia, 8 (2023) 691-700
Figure 9. Sea Surface Temperature Anomaly in El Niño
less than SPI12. EDI detects the number of extreme droughts
in the south of North Sumatra province about ten times and
in West Sumatra about 17 times. The lowest extreme drought
frequency detected by EDI occurred in Aceh Province. It was
also shown by Supari et al. (2023) .
As previously explained, SPI3, SPI6, SPI12, and EDI de-
scribe precipitation conditions (dry/wet) on a 3-month, 6-
month, and 12-month scale. Based on the frequency of ex-
treme droughts, it can be seen that the distribution of extreme
droughts on a seasonal scale is wider than on inter-seasonal
and annual scales. The areas that often experience drought
(3-month, 6-month, and 12-month scale) are West Sumatra
to the southern part of North Sumatra. This condition is re-
lated to research by Daoed et al. (2018), which shows that West
Sumatra tends to be vulnerable to drought, especially in river
basins. The climatological results in section 3.2 show that this
area (Tobing and Minangkabau FL stations) has the highest
monthly precipitation. However, this area has tended to experi-
ence more extreme drought over the past 30 years. This result
shows that identifying drought with SPI and EDI is local, which
refers to regional precipitation. Regions with low precipitation
throughout the year do not always refer to drought because
drought refers to the condition of below-normal precipitation
in a certain area and time. In addition, these results also show
that SPI and EDI, in particular, cannot detect drought spatially.
Figure 10. Composite of Indices in Positive IOD Events
3.4 Drought Correlation with El Niño and Positive IOD
The composite of SPI3, SPI6, SPI12, and EDI results at the
time of El Niño, positive IOD, and El Niño-positive IOD, re-
spectively, are shown in Figure 8, Figure 10, and Figure 11. El
Niño is more dominant on SON, mainly based on SPI3, SPI6,
and EDI composites. During the JJA period, the inuence
of El Niño (SPI3 composite) tends to occur in the northern
provinces of Aceh, North Sumatra, West Sumatra, and parts of
Riau to Bengkulu. In this period (JJA), the SPI12 composite
showed a more dominant El Niño eect in the province of
West Sumatra. In the DJF period, it can be seen that El Niño
is more dominant in the north, while in the MAM period, El
Niño is more dominant on the east coast of Sumatra. When
El Niño occurs, almost the entire area of Sumatra experiences
drought, especially during the SON to MAM seasons. How-
ever, dry-normal conditions were exchanged in the provinces
of Riau and Bengkulu during the DJF and MAM seasons.
During the DJF season, Riau province tends to be in nor-
mal conditions, while Bengkulu province tends to be dry, and
vice versa during the MAM season. This normal condition
is because, during the DJF season, there is an increase in sea
surface temperature, especially on the coast of Riau province
(Figure 9). In addition, this DJF period is the beginning of
the monsoon cycle, and The dominant Inter-Tropical Conver-
gence Zone (ITCZ) is at the equator Qian (2020) . It causes
an increase in the convection of water mass activity in the air,
and this area experiences an increase in precipitation, so Riau
province tends to be in normal conditions. During the MAM
season, the sea surface temperature around this area is colder
©2023 The Authors. Page 697 of 700

Suhadi et. al. Science and Technology Indonesia, 8 (2023) 691-700
Figure 11. Composite of Indices in El Niño and Positive IOD
Events
again, like in the JJA and SON seasons.
The composite drought indices during the positive IOD
period are shown in Figure 10. At the beginning of the positive
IOD (JJA), the positive IOD inuence was more dominant
in the provinces of Aceh, West Sumatra-Riau, and Bengkulu-
Lampung. The eect of this positive IOD is getting stronger
in the SON period. Based on the SPI3 composite, the areas
that are dominantly aected by positive IOD are the northern
part of North Sumatra, West Sumatra, and parts of Jambi to
Lampung provinces. The inuence of IOD on the 6-month
(SPI6) and 12-month (SPI12 and EDI) scales is more than
dominant in the areas of Aceh, Riau, to Lampung.
The eect of El Niño-positive IOD based on the SPI3,
SPI6, SPI12, and EDI composites (Figure 11) is the same
as when an El Niño occurs. However, when these two phe-
nomena occur simultaneously, the eect of drought is stronger.
Based on the SPI3 and SPI6 composites, the inuence of these
two phenomena is more dominant in the JJA to DJF seasons.
Although based on the SPI6 composite, the inuence of this
phenomenon is slightly weakened in the JJA season, it has
strengthened in the MAM season. The inuence of these two
phenomena was also shown based on the EDI composite. How-
ever, during the JJA season, the EDI composite showed the El
Niño-positive IOD eect stronger than the SPI6 composite
but weaker than the SPI3 composite. The SPI12 composite
shows a more dominant El Niño-positive IOD eect in the
SON to MAM season.
4. CONCLUSION
Based on the monthly precipitation amount, the west coast
of Sumatra receives the highest precipitation throughout the
year (TM stations). In contrast, the lowest is on the northern
coast of northern Sumatra (IMD stations). In addition, the
peak of precipitation in the northern part of Sumatra occurs
during the SON period, while the southern part experiences
a lack of precipitation. Extreme drought on a 3-month scale
tends to be spread across almost all of Sumatra except for Aceh
province, which experiences the least extreme drought. The
frequency of drought on a 12-month scale is most frequent
in the provinces of West Sumatra and southern North Suma-
tra. In addition, the results of this study indicate that SPI12
detects more extreme drought than EDI. However, the distri-
bution of extreme drought frequencies detected based on EDI
is comparable to SPI12. Huang et al. (2016) suggest a strong
correlation between SPI12 and EDI. Based on the results of
the temporal analysis, it appears that some extreme droughts
are not aected by El Niño-positive IOD phenomena, such as
the extreme droughts in 2014 and 2008, which negative SSTA
more inuenced, so some droughts that are not related to these
phenomena are thought to be caused by negative SSTA. El
Niño is more than dominant in northern Sumatra, especially
during the DJF period. Positive IOD is more inuential in this
region during JJA-SON, as shown in the study by Kurniadi
et al. (2021) . The southern part of Sumatra is also aected by
positive IOD, especially during the SON period. When an El
Niño-positive IOD occurs simultaneously, North Sumatra to
Lampung is more dominantly aected by this phenomenon,
especially during the SON-DJF period. It was also shown by
Kurniadi et al. (2021) and Nurdiati et al. (2022) .
5. ACKNOWLEDGMENT
This study is part of the rst author’s dissertation supported
by the Ministry of Education, Culture, Research, and Tech-
nology through Hibah Penelitian Disertasi Doktor 2022, No.
142/E5/PG.02.00.PT/2022.
REFERENCES
Abbas, A., M. Waseem, W. Ullah, C. Zhao, and J. Zhu (2021).
Spatiotemporal Analysis of Meteorological and Hydrological
Droughts and their Propagations. Water,13(16); 2237
Abramowitz, M. and I. A. Stegun (1964). Handbook of Math-
ematical Functions with Formulas, Graphs, and Mathematical
Tables, volume 55. US Government Printing Oce
Adeola, O. M., M. Masinde, J. O. Botai, A. M. Adeola, and
C. M. Botai (2021). An Analysis of Precipitation Extreme
Events based on the SPI and Edi Values in the Free State
Province, South Africa. Water,13(21); 3058
©2023 The Authors. Page 698 of 700

Suhadi et. al. Science and Technology Indonesia, 8 (2023) 691-700
Adisa, O. M., M. Masinde, and J. O. Botai (2021). Assessment
of the Dissimilarities of EDI and SPI Measures for Drought
Determination in South Africa. Water,13(1); 82
Aldrian, E. and R. Dwi Susanto (2003). Identication of Three
Dominant Rainfall Regions within Indonesia and their Rela-
tionship to Sea Surface Temperature. International Journal
of Climatology: A Journal of the Royal Meteorological Society,
23(12); 1435–1452
Avia, L. and I. Soati (2018). Analysis of El Niño and IOD Phe-
nomenon 2015/2016 and Their Impact on Rainfall Variabil-
ity in Indonesia. Conference Series: Earth and Environmental
Science,166(1); 012034
Bougara, H., K. Baba Hamed, C. Borgemeister, B. Tischbein,
and N. Kumar (2021). A Comparative Assessment of Meteo-
rological Drought in the Tafna Basin, Northwestern Algeria.
Journal of Water and Land Development, (51); 78–93
Byun, H. R. and D. A. Wilhite (1999). Objective Quantica-
tion of Drought Severity and Duration. Journal of Climate,
12(9); 2747–2756
Chandrasekara, S. S., H.-H. Kwon, M. Vithanage, J. Obey-
sekera, and T. W. Kim (2021). Drought in South Asia:
A Review of Drought Assessment and Prediction in South
Asian Countries. Atmosphere,12(3); 369
Daoed, D., B. Rusman, B. Istijono, A. Hakam, and M. Syukur
(2018). Evaluation of Drought Vulnerability on Watersheds
in West Sumatera Province by using Cropwat-8 and GIS.
International Journal on Advanced Science, Engineering and In-
formation Technology,8(6); 2443–2449
Darlan, N. H., S. S. Arif, P. Sudira, and B. D. A. Nugroho
(2020). Spatial and Temporal Analysis of Seasonal Rainfall
on the East Coast of North Sumatra, Indonesia. Indonesian
Journal of Geography,52(3); 360–367
Davies, S. J. and L. Unam (1999). Smoke-haze from the
1997 Indonesian Forest Fires: Eects on Pollution Levels,
Local Climate, Atmospheric CO2 Concentrations, and Tree
Photosynthesis. Forest Ecology and Management,124(2-3);
137–144
Edwards, D. C. and T. B. McKee (1997). Characteristics of 20
th Century Drought in the United States at Multiple Time Scales,
volume 97. Colorado State University Fort Collins
Eslamian, S., K. Ostad Ali Askari, V. P. Singh, N. R. Dalezios,
M. Ghane, Y. Yihdego, and M. Matouq (2017). A Review of
Drought Indices. International Journal of Constructive Research
in Civil Engineering,3; 48–66
Huang, Y. F., J. T. Ang, Y. J. Tiong, M. Mirzaei, and M. Z. M.
Amin (2016). Drought Forecasting using SPI and EDI under
RCP-8.5 Climate Change Scenarios for Langat River Basin,
Malaysia. Procedia Engineering,154; 710–717
Iskandar, I., M. Irfan, F. Syamsuddin, A. Johan, and P. Poer-
wono (2011). Trend in Precipitation Over Sumatera Under
the Warming Earth. International Journal of Remote Sensing
and Earth Sciences (IJReSES),8(1)
Iskandar, I., W. Mardiansyah, D. Setiabudidaya, P. Poerwono,
I. Yusyian, and Z. Dahlan (2017a). What Did Drive Extreme
Drought Event in Indonesia During Boreal Summer/fall
2014? Journal of Physics: Conference Series,817(1); 012073
Iskandar, I., P. A. Utari, D. O. Lestari, Q. W. Sari, D. Se-
tiabudidaya, M. Khakim, I. Yustian, and Z. Dahlan (2017b).
Evolution of 2015/2016 El Niño and its Impact on Indone-
sia. 1857(1)
Jain, V. K., R. P. Pandey, M. K. Jain, and H. R. Byun (2015).
Comparison of Drought Indices for Appraisal of Drought
Characteristics in the Ken River Basin. Weather and Climate
Extremes,8; 1–11
Kamruzzaman, M., S. Hwang, J. Cho, M. W. Jang, and
H. Jeong (2019). Evaluating the Spatiotemporal Character-
istics of Agricultural Drought in Bangladesh using Eective
Drought Index. Water,11(12); 2437
Kurniadi, A., E. Weller, S. K. Min, and M. G. Seong (2021).
Independent ENSO and IOD Impacts on Rainfall Extremes
Over Indonesia. International Journal of Climatology,41(6);
3640–3656
Kuswanto, H., F. Hibatullah, and E. S. Soedjono (2019). Per-
ception of Weather and Seasonal Drought Forecasts and
its Impact on Livelihood in East Nusa Tenggara, Indonesia.
Heliyon,5(8)
Lee, H. S. (2015). General Rainfall Patterns in Indonesia and
the Potential Impacts of Local Seas On Rainfall Intensity.
Water,7(4); 1751–1768
McKee, T. B., N. J. Doesken, and J. Kleist (1993). The Rela-
tionship of Drought Frequency and Duration to Time Scales.
Proceedings of Eighth Conference on Applied Climatology,17(22);
179–183
Mori, S., H. Jun Ichi, Y. I. Tauhid, M. D. Yamanaka,
N. Okamoto, F. Murata, N. Sakurai, H. Hashiguchi, and
T. Sribimawati (2004). Diurnal Land–sea Rainfall Peak
Migration Over Sumatera Island, Indonesian Maritime Con-
tinent, Observed by TRMM Satellite and Intensive Rawin-
sonde Soundings. Monthly Weather Review,132(8); 2021–
2039
Nita, I., A. N. Putra, and A. Fibrianingtyas (2020). Analysis of
Drought Hazards in Agricultural Land in Pacitan Regency,
Indonesia. SAINS TANAH-Journal of Soil Science and Agro-
climatology,17(1); 7–15
Nurdiati, S., F. Bukhari, M. T. Julianto, A. Sopaheluwakan,
M. Aprilia, I. Fajar, P. Septiawan, and M. K. Najib (2022).
The Impact of El Niño Southern Oscillation and Indian
Ocean Dipole on the Burned Area in Indonesia. Terrestrial,
Atmospheric and Oceanic Sciences,33(1); 16
Nurdiati, S., A. Sopaheluwakan, and P. Septiawan (2021). Spa-
tial and Temporal Analysis of El Niño Impact on Land and
Forest Fire in Kalimantan and Sumatra. Agromet,35(1);
1–10
Park, C. K., J. Kam, H. R. Byun, and D. W. Kim (2022). A
self-calibrating Eective Drought Index (scEDI): Evaluation
Against Social Drought Impact Records Over the Korean
Peninsula (1777–2020). Journal of Hydrology,613; 128357
Pramudya, Y., T. Onishi, M. Senge, K. Hiramatsu, and P. M.
Nur (2019). Evaluation of Recent Drought Conditions by
Standardized Precipitation Index and Potential Evapotran-
©2023 The Authors. Page 699 of 700

Suhadi et. al. Science and Technology Indonesia, 8 (2023) 691-700
spiration Over Indonesia. Paddy and Water Environment,17;
331–338
Prinz, D. (2009). Contributor and Victim-Indonesia’s Role in
Global Climate Change with Special Reference to Kaliman-
tan. Jurnal Sains & Teknologi Lingkungan,1(2); 139–153
Qian, J. H. (2020). Multi-scale Climate Processes and Rainfall
Variability in Sumatra and Malay Peninsula Associated with
ENSO in Boreal Fall and Winter. International Journal of
Climatology,40(9); 4171–4188
Raude, J. M., R. M. Wambua, and B. M. Mutua (2018). De-
tection of Spatial, Temporal and Trend of Meteorological
Drought Using Standardized Precipitation Index (SPI) and
Eective Drought Index (EDI) in the Upper Tana River
Basin. Open Journal of Modern Hydrology,8; 83–100
Saji, N. and T. Yamagata (2003). Possible Impacts of Indian
Ocean Dipole Mode Events on Global Climate. Climate
Research,25(2); 151–169
Salehnia, N., A. Alizadeh, H. Sanaeinejad, M. Bannayan,
A. Zarrin, and G. Hoogenboom (2017). Estimation of Me-
teorological Drought Indices based on AgMERRA Precipita-
tion Data and Station-observed Precipitation Data. Journal
of Arid Land,9; 797–809
Sarmiasih, M. and P. Y. Pratama (2019). The Problematics
Mitigation of Forest and Land Fire District (Kerhutla) in
Policy Perspective (A Case Study: Kalimantan and Sumatra
in Period 2015-2019). Journal of Governance and Public Policy,
6(3); 270–292
Supari, S., I. Iskandar, M. Irfan, and H. Akhsan (2023).
Drought Assessment in Aceh and North Sumatra Using Ef-
fective Drought Index. Science and Technology Indonesia,8(2)
Supari, S., R. Muharsyah, and A. Sopaheluwakan (2016). Map-
ping Drought Risk in Indonesia Related to El-Niño Hazard.
AIP Conference Proceedings,1730(1)
Svoboda, M. and B. Fuchs (2016). Handbook of Drought
Indicators and Indices. Drought and Water Crises: Integrating
Science, Management, and Policy; 155–208
WMO (2012). Standardized Precipitation Index User Guide.
World Meteorological Organization,1090
WMO (2016). Handbook of Drought Indicators and Indices.
In British Medical Journal; 2366
Zargar, A., R. Sadiq, B. Naser, and F. I. Khan (2011). A
Review of Drought Indices. Environmental Reviews,19(NA);
333–349
©2023 The Authors. Page 700 of 700