Droughts are the most commonly practiced indices (Tsakiris

Droughts are common natural
phenomena that arise from a substantial deficit of precipitation and can take
place around any corner of the world, in different climatic regions (Zargar, et al.,
2011). The deficits adversely affect  both surface and groundwater resources and
lead to decrease in water supply, poor water quality, declined agricultural
production, reduced hydro-power generation, distressed riparian and wetland
habitations, and so on (Mpelasoka,
et al., 2008).
In 1997, drought has been
classified into four types by The American Meteorological Society;
meteorological or climatological, agricultural, hydrological, and socioeconomic
(Mpelasoka,
et al., 2008; R. & Junior, 2002). The impacts of droughts are one of
the most significant among all the natural hazards because it is not only
capable of affecting various economic sectors of a nation, but also affecting
the lives and livelihood of people. The agriculture, food security, industry,
human health, animal health and livelihood security of a region can be severely
damaged by the occurrence of a drought event (Svoboda
& Fuchs, 2017). Considering the
impacts and frequency of drought, it is essential to evaluate drought severity,
but the accurate assessment of drought is not easy (Keyantash
& Dracup, 2002). The extent to which a drought can
affect a particular region depend on a various factors, such as the
socioeconomic condition, geographic location and so on (Svoboda & Fuchs, 2017; R. & Junior, 2002). Hence,
the type of impact relevant to a specific region bears significant
consideration to decide what indices to select.

Drought indices can be defined as
quantifiable measures that evaluates drought by integrating data from one or
several variables (Zargar, et al., 2011). More than 100 drought indices have so far
been proposed which quantifies severity levels and
duration of drought (Zargar, et al., 2011, Keyantash &
Dracup, 2002). Several indices that can be used to evaluate droughts are, the
Standardized Precipitation Index (SPI), Standardized Runoff Index (SRI), Standardized
Precipitation Evapotranspiration Index (SPEI), Soil Moisture Index (SMI),
Reconnaissance Drought Index (RDI) and Palmer Drought Severity Index (PDSI).
Among these, SPI and PDSI are the most commonly practiced indices (Tsakiris & Vangelis, 2005). The drought
indices are generally continuous functions of some hydro-meteorological variables,
such as; rainfall, potential evaporation and temperature (Mpelasoka,
et al., 2008).
These indices correspond to different forms of drought, including
meteorological, agricultural and hydrological drought and aid in a variety of operations, including drought early
warning and monitoring and contingency planning (Zargar, et al., 2011). Thus
drought indices can work as a tool
to track droughts. But the choice of an index on a combination of different
indices depends on the characteristics of the specific region and the impacts of drought to the stakeholders (Svoboda & Fuchs, 2017).

Determination of the best fit is quite a lengthy
process which arises the dilemma while selecting which index to use (Svoboda & Fuchs, 2017). Hence, the
indices that are reviewed in this paper are considered to provide options of
different indices that can be best suited for river basin areas with different
climatic and geographical characteristics. The
purpose of this paper is to look into some of the most widely applied drought indices
that are being used across drought-prone river basins regions around the globe
and to come up with applicable indices. 

Basin Specific Climate, Geography and
Drought-Impacts

The papers that have been
selected to review in this article are based on 10 different river basin
regions of 6 different countries. In this section the climatic and geographical
features of these basin areas are briefly discussed along with the impacts of
drought in the respective regions. This section will help to provide the idea
of climate, geography and impact wise suitability of indices.   

China

The
river basins that are studied from China are located in the subtropical climate
region of the country (Fluixá et al., 2017; Li, Q. et al., 2015). The Jinsha River
basin is a watershed with the area of about 1, 32, 000 sq. km, ranging in
elevation from 263m to 5910 m from sea level (Yang et al., 2013). The average annual rainfall of this basin ranges
from 239mm in dry years to
1154.9mm in wet years (Meng et al.,
2012). The Huaihe river basin expands in the east of the country with the basin
area of 270,000 sq. km and has an average annual precipitation of
883mm (Mingkai & Kai, 2017). In China, droughts are
considered to be the most severe natural threat for socioeconomic growth and
ecosystems because, a wide range of areas are affected by droughts that results
in great agricultural losses (Fluixá et al., 2017). In the southern
part of China a number of significant drought event took place during the last
decade and resulted in crop failure, shortage of pure drinking water,
degradation of ecosystem, health issues, and even deaths (Fluixá et al.,
2017; Li, Q. et al., 2015).

India

The Ken and
Ghataprabha river basins of India consist of a geographical area of 28,692 sq.
km and 7231 sq. km respectively (Pathak et al., 2016;
Jain et al., 2012).  The ken river flows within the highest
altitude of 550m and the lowest to be 87m above the sea level (Jain et al., 2012). The subtropical monsoon
climate is the climatic feature of the Ghataprabha river basin, which has 4
monsoon months followed by 8 dry months. The average annual rainfall is around
1248mm (Pathak et al., 2016). On the other hand, the Ken river basin is located in a
semi-arid to sub-humid climatic zone of India, where the mean annual rainfall
ranges from 800mm to 1250mm (Jain et al.,
2012). The major effects of drought around these regions are:
threat to the food security and increase rate of human mortality (Nath et al.,
2017).

Thailand

The Sakae Krang and the Chi river basins
of Thailand are located in a tropical savanna climatic region. The area of Sakae
Krang river basin and Chi river basin is 5191 sq. km and 49,476 sq. km
respectively (Homdee et al., 2016; Wichitarapongsakun et
al., 2016).  The average
precipitation over the considerable small area of Sakae Krang river basin is
around 1000 mm and Chi River basin is approximately 1150 mm per year
(Homdee et al., 2016; Wichitarapongsakun et
al., 2016). Drought hits the basin areas annually resulting in
severe impact on crop production, human health, environment and socioeconomic
conditions (Wichitarapongsakun et al., 2016)

Ethiopia

The Awash river basin of Ethiopia have a
temperate climate and consist of an area about 701,000 sq. km. (Anon., 1965). The highest point
of this basin area is 3000m above the sea level whereas the lowest is at 250m (Anon., 1965). This river basin
has a mean rainfall of 710mm per year (Edossa et al., 2009). In 1888 a historic
deadly famine took place as an impact of drought when 30% of the population
died and 90% of the animal perished (Edossa et al., 2009). Moreover, loss of assets
in the form of crops, livestock, and other productive capitals takes place as
impacts of drought (Edossa et al.,
2009).

Zambia

The Kafue river catchment is located under
the humid-subtropical region of Zambia where the average annual precipitation
ranges from 800mm to 1300mm (Anon.,
2007).
The area of this river basin is 155,000 sq. km (Anon.,
2007).
This basin is the most
developed region of Zambia’s where 50% of the total population and a variety of
animal species live (Lweendo et al., 2017).
Majority of the industrial, municipal water supply, agriculture and mining
activities takes place around this river basin (Lweendo et al., 2017). An event of a severe drought can affect
the whole system of these aspects of this basin.

Bangladesh

Only
7% of the whole GBM basin lies inside Bangladesh and about two-third of the
country is comprised by this basin (Banglapedia, 2014). The climatic zone of
this river basin area falls under is subtropical climate. The total area of
this basin is 1,761,300 sq. km among which 123,291 sq. km falls under the
territory of Bangladesh (Kattelus et al.,
2015). The northern region of the country faces frequent incidents of drought
as a result of which crop production gets affected, malnutrition takes place
(Rakib et al., 2015).

Indices Used for River Basin Areas

This
section covers the different indices that have been applied to the considered
river basins. The purpose is to try finding out which index or indices have
been the best fit under what conditions and why.

About the Indices

The use of Standardized
Precipitation Index (SPI) has been reviewed to be predominant as a drought
index for river basin areas (Fluixá
et al., 2017; Lweendo et al.,
2017; Nury
& Hasan, 2016; Homdee et al.,
2016; Wichitarapongsakun et al., 2016; Jain
et al., 2015; Li, Q. et al., 2015; Rafiuddin et al., 2011; Edossa et al., 2009). The SPI is an
index for drought severity level analysis. Its calculation is based on rainfall
data at various periods of 1 month to 24 months or longer, over a period of 30
or more years of time (Svoboda & Fuchs,
2017). In a couple of study based on two different river basins of
Thailand; the Krang river basin and Chi river basin; the drought severity level
was determined by SPI (Homdee et al.,
2016; Wichitarapongsakun et al., 2016). The drought conditions of Awash River Basin of Ethiopia and Upper Kafue
River Basin of Zambia have also been analyzed, primarily by the use of SPI (Lweendo et al.,
2017; Edossa et al., 2009). A
couple of studies have been conducted for drought detection and assessment on
the Jinsha river basin and Huaihe river basin of China (Fluixá et al., 2017; Li, Q. et al.,
2015).  Both the studies used SPI drought index on
the purpose of formulating new indices; MDI, and indicators; ODE and ODI (Fluixá et al., 2017; Li, Q. et al.,
2015).
Ganges-Brahmaputra river basin of Bangladesh has also been brought under
consideration to diagnose the drought scenario across the country (Nury &
Hasan, 2016; Rafiuddin, M. et al., 2011). According to the outcomes, SPI turned out to
be a consistent drought indicator in the context of Bangladesh (Nury &
Hasan, 2016; Rafiuddin, M. et al., 2011). No other indices were found to be popularly
used to study drought conditions in Bangladesh. A comparison study of different
drought indices was conducted based on Ken river basin of central India where
SPI was one of the considered indices (Jain
et al., 2015).

In
the study of the Upper Kafue River Basin of Zambia a number of standardized
drought indices were selected to evaluate different types of droughts along
with SPI (Lweendo et al., 2017). The
technique that is used to compute SPI, was extended to calculate SRI and SMI
where both of the indices use the process of fitting a probability distribution
to data and converting it to normal distribution (Lweendo et al., 2017). Similar to SPI, they are also quantified on multiple
time scales. SRI uses runoff time series as input data and both the input data
for SRI and SMI computation was derived from the hydrological SWAT model (Shukla & Wood, 2008). The agricultural
drought index, SMI was compared to historical droughts to validate its
efficiency and its performance turned out to be fair in terms of capturing the
major agricultural droughts but it was less sensitive than the other indices (Lweendo
et al., 2017). It is evident that all drought indices demonstrated
comparable and consistent results, even though they were calculated based on
different input parameters. The paper also drew the conclusion
that, all of the indices could successfully detect temporal variability of
major drought events in this humid subtropical basin of Upper Kafue River (Lweendo
et al., 2017).

In
India, a couple of studies were conducted based on the application of several
drought indices in Ken river basin and Ghataprabha river basin to measure their
applicability (Pathak et al., 2016; Jain et al.,
2015). Both the studies were focused on drawing comparison among a
number of drought indices including SPI, SDI, SRI, EDI, CZI, RD, and RDDI (Pathak et al.,
2016; Jain et al., 2015). In
the study of Ghataprabha basin the suitability of SDI and SRI to assess
hydrological droughts was compared and both indices were proven to be
appropriate for the area (Pathak et al., 2016). According to the
study on Ken basin, the comparatively new index called Effective Drought Index (EDI)
was found to be more responsive and showed a better performance (Jain et al.,
2015).

SPEI
was applied to study the scenario of hydrological drought in the studies of
Upper Kafue river basin of Zambia and Chi river basin of Thailand (Lweendo et al., 2017; Homdee et al., 2016). The SPEI considers the
difference between precipitation and potential evapotranspiration (PET) (Homdee
et al., 2016).

Formulation
of new indicators and a drought index took place in the studies based on the
Jinsha river basin and Huaihe river basin of China (Fluixá et al., 2017; Li, Q. et al.,
2015).
A new Multivariate Drought Index (MDI) was developed in the study of Huaihe
river basin and was being compared to SPI and PDSI; as a result of which, MDI
turned out to have certain benefits over SPI and PDSI in drought evolution
monitoring (Li, Q. et al., 2015).
On the other hand, two new indicators; Overall Drought Extension (ODE) and
Overall Drought Intensity; were developed during the study of Jinsha river
basin using the outputs from several meteorological drought indices, such as,
SPI, RAI, PN and DEC (Fluixá
et al., 2017).
The newly developed indicators and index showed a decent agreement with the
records of historical drought events in both the river basins of China (Fluixá et al., 2017; Li, Q. et al.,
2015).

Whys and Wherefores

SPI has been most commonly used by the aforementioned
studies on the purpose of understanding the drought condition of river basins. The simplicity and versatility of this index has made
it more popular among the researchers. One of the benefits of SPI is that it requires
only one input variable that is precipitation data (Svoboda & Fuchs, 2017).
On the purpose of analyzing meteorological droughts, SPI has been recommended
by the World Meteorological Organization to be used all across the world
(Lweendo et al., 2017). SPI is considered
to be a versatile index because its calculation can be done on any timescale
and it is appropriate for both agricultural and hydrological applications. The
studies chose to work with this index because of its application to monitor the
dynamics of a drought, including its development and decline (Lweendo et al.,
2017; Nury & Hasan, 2016; Wichitarapongsakun
et al., 2016; Rafiuddin et al., 2011; Edossa et al., 2009). Moreover,
SPI is not negatively affected by topography and can work as an input to water
management planning, risk management and decision analysis. However, the
problem with using SPI is that, it requires continuous data of at least 30
years and cannot accommodate missing data (Jain
et al., 2015).

A
couple of the abovementioned studies have applied other indices apart from SPI.
SPEI, SRI, and SMI has been successfully used in the study based on Upper Kafue
River Basin of Zambia. Newly developed index MDI turned out to be a great fit
for the study of drought condition in river basins as well. That means, along
with SPI; SPEI, SMI, SRI and MDI can be considered to be well suited to
evaluate drought condition around river basin areas. In one of the studies, the
PDSI index failed to monitor a number of typical drought events (Li, Q. et al., 2015). As a result, this index can be
considered to be less applicable for river basins. Even though PDSI is a widely
used index, its nine-month time scale results into a lag in determining drought
conditions (Svoboda & Fuchs, 2017).
Because of this lag, rapidly emerging droughts can be left unidentified which
makes it less suitable for particular regions. 

 

SPEI
is suitable for drought assessment around river basins because it  offers versatility as SPI and uses monthly
precipitation and temperature data as input which allows this index to
acknowledge the impact of temperature on a drought event (Svoboda & Fuchs, 2017). The issue that makes it less popular that SPI is its
requirement of a serially complete record of data without any missing months. Moreover,
this index failed to represent
extreme drought events in the Chi river basin of Thailand during 1993–1994 and
could only detect a slightly dry period (Homdee et al., 2016).

The
MDI performed the best because it evaluates droughts by using a simple tactic
which considers multiple easily obtainable hydrological parameters, such as;
precipitation, evapotranspiration, soil moisture and runoff (Li, Q. et al.,
2015). The feasibility of MDI to be applied on river basin drought
condition is up to the mark because, the calculation is done based on local
climate and the each parameter is weighed by the hydrological conditions of the
study area (Li, Q. et al., 2015). Among the considered variables, the
calculation of evapotranspiration is considered to be a difficult one. This
variable along with soil moisture and runoff, were simulated using the
Xin’anjiang model (Li, Q. et al., 2015)

The study of Ken river basin successfully used China
Z-Index (CZI) and Rainfall Departure (RD) and suggested these as simple
applicable indicators of drought conditions for a given time over specific
regions (Jain et al., 2015). A straight
forward measure of rainfall deviation is used by this index from its long-term
mean or median based on the regional weather conditions. In general, this
technique is appropriate for defining local weather condition for rainfall deficiencies
of smaller duration (Jain et al.,
2015). It has been reported that CZI and SPI provide similar results, hence,
CZI can be preferred over SPI for its ease of use (Jain et al., 2015).  

EDI
can sometimes perform better than other indices such as, SPI, RD etc. because
it is free from the time scale problem, effective for both long and short
drought and could identify drought earlier than the other considered indices, (Jain et al.,
2015). On the other hand, RDDI showed the poorest performance compared
to others (Jain et al., 2015).