ArticlePDF Available

Assessing structural uncertainty caused by different weighting methods on the Standardized Drought Vulnerability Index (SDVI)

February 2019
Stochastic Environmental Research and Risk Assessment 33(2)

DOI:10.1007/s00477-019-01648-4

Authors:

Demetrios E Tsesmelis

Hellenic Open University

Panagiotis D. Oikonomou

University of Vermont

Constantina G. Vasilakou

Nikolaos A. Skondras

Agricultural University of Athens

Show all 8 authorsHide

Indices are used for representing complex phenomena; however, concerns usually arise regarding their objectivity and reliability, particularly dealing with their uncertainties during the development process. The current overarching objective is to reveal the significance of employing different weighting techniques in the application of the Standardized Drought Vulnerability Index (SDVI) and demarcate any pertinent implications that may emerge in drought decision making. Greece, as it is very often facing the catastrophic effects of droughts, presents an almost ideal case for the SDVI testing. SDVI outcomes were tested utilizing five weighting techniques deriving from four weighting methods. The analyses indicated that the use of complex weighting models may not be necessary in all cases and that the simple equal weighting method seems more effective to estimate drought vulnerability. It also seems more important to address the search for valid, reliable and relevant individual indicators forming the complex index as well as appropriate index development processes that would measure performance of water bodies, systems and schemes, monitor the process of equitable sharing, and provide mechanisms for monitoring the state and changes in interdependent water systems.

sualization of four sub-indices: demand, supply, impacts and infrastructure for August 2000

…

The basic statistics of precipitation in Sterea Hellas

…

Advantages and disadvantages of various weighting methods

…

Possible four scenarios in Aliartos areaStochastic Environmental Research and Risk Assessment

…

Descriptive statistics of the five weighting techniques in Sterea Hellas based on real conditions

…

Figures - uploaded by Demetrios E Tsesmelis

Content may be subject to copyright.

Content uploaded by Demetrios E Tsesmelis

Content may be subject to copyright.

Content uploaded by Panagiotis D. Oikonomou

Content may be subject to copyright.

1 23

Stochastic Environmental Research

and Risk Assessment

ISSN 1436-3240

Stoch Environ Res Risk Assess

DOI 10.1007/s00477-019-01648-4

Assessing structural uncertainty caused

by different weighting methods on the

Standardized Drought Vulnerability Index

(SDVI)

Demetrios E.Tsesmelis, Panagiotis

D.Oikonomou, Constantina

G.Vasilakou, Nikolaos A.Skondras,

Vassilia Fassouli, et al.

1 23

Your article is protected by copyright and

all rights are held exclusively by Springer-

Verlag GmbH Germany, part of Springer

Nature. This e-offprint is for personal use only

and shall not be self-archived in electronic

repositories. If you wish to self-archive your

article, please use the accepted manuscript

version for posting on your own website. You

may further deposit the accepted manuscript

version in any repository, provided it is only

made publicly available 12 months after

official publication or later and provided

acknowledgement is given to the original

source of publication and a link is inserted

to the published article on Springer's

website. The link must be accompanied by

the following text: "The final publication is

available at link.springer.com”.

ORIGINAL PAPER

Assessing structural uncertainty caused by different weighting

methods on the Standardized Drought Vulnerability Index (SDVI)

Demetrios E. Tsesmelis

·Panagiotis D. Oikonomou

·Constantina G. Vasilakou

·Nikolaos A. Skondras

Vassilia Fassouli

·Stavros G. Alexandris

·Neil S. Grigg

·Christos A. Karavitis

1,4

©Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract

Indices are used for representing complex phenomena; however, concerns usually arise regarding their objectivity and

reliability, particularly dealing with their uncertainties during the development process. The current overarching objective

is to reveal the signiﬁcance of employing different weighting techniques in the application of the Standardized Drought

Vulnerability Index (SDVI) and demarcate any pertinent implications that may emerge in drought decision making.

Greece, as it is very often facing the catastrophic effects of droughts, presents an almost ideal case for the SDVI testing.

SDVI outcomes were tested utilizing ﬁve weighting techniques deriving from four weighting methods. The analyses

indicated that the use of complex weighting models may not be necessary in all cases and that the simple equal weighting

method seems more effective to estimate drought vulnerability. It also seems more important to address the search for

valid, reliable and relevant individual indicators forming the complex index as well as appropriate index development

processes that would measure performance of water bodies, systems and schemes, monitor the process of equitable sharing,

and provide mechanisms for monitoring the state and changes in interdependent water systems.

Keywords Drought · Drought vulnerability · Drought Vulnerability Index · Indices · Indicators · Structural uncertainty ·

Water resources management

1 Introduction

Decision making may generally be deﬁned as the art and

science behind the identiﬁcation and selection of the

actions that best fulﬁll the overall objectives of a stated

issue (Corvala

´n et al. 2000; Bianco 2006; Adair 2010).

However, in many cases, such a deﬁnition may seem both

simple and elusive. Part of that resides in systems com-

plexity and the resulting difﬁculty of the emerging man-

agement problems (Daellenbach 1994; Qudrat-Ullah et al.

2007). The various sets of indicators and indices are usu-

ally employed for the simpliﬁcation, quantiﬁcation and

provision of pertinent information concerning complex

events and phenomena (Breier et al. 2012; Dahl 2012;

Becker et al. 2017). They usually derive from an appro-

priate selection and elaboration of primary or secondary

data (Segnestam 2002). On a greater scale, they may also

be used to evaluate and rank the national/regional perfor-

mance of various countries towards achieving speciﬁed

goals (Bandura 2005,2008; Rogge 2012; Landerretche

et al. 2017). In the pertinent literature, their overall

objective is to facilitate the communication among the

respective authorities and foster decision and policy mak-

ing (European Environment Agency 1999,2005; Rogge

2012). As tools, their useful value increases in cases where

the lack of direct knowledge sets them as the only means

&Demetrios E. Tsesmelis

tsesmelis@aua.gr

Department of Natural Resources Development and

Agricultural Engineering, Agricultural University of Athens,

75 Iera Odos, 11855 Athens, Greece

Colorado Water Institute, Colorado State University, 1033

Campus Delivery, Fort Collins, CO 80523-1033, USA

Department of Civil and Environmental Engineering,

Colorado State University, 1372 Campus Delivery,

Fort Collins, CO 80523-1372, USA

Department of Civil and Environmental Engineering, Faculty

Afﬁliate, Colorado State University, 1372 Campus Delivery,

Fort Collins, CO 80523-1372, USA

123

Stochastic Environmental Research and Risk Assessment

https://doi.org/10.1007/s00477-019-01648-4

Author's personal copy

towards obtaining a fracture of the required information

(Haines-Young et al. 2012). Climatic, social, and ecologi-

cal indicators and indices such as Human poverty index,

population density, crop type, rainfall etc. play an essential

role in water resources management adapted to the com-

plexity of such systems mostly through social and envi-

ronmental learning.

Such systems’ complexity leads also to an extremely

important obstacle of the decision making environment:

uncertainty. Decision makers are often forced to deal with

several uncertainty generating sources including the vari-

ous factors embodied in the decision making process itself.

Such sources are mainly convoluted around the quality and

validity of the input data and the outcomes of the selected

decisions (French 1995; Briggs and Peat 2000; Burgman

2005; Norton 2005; Sage 2007; Williams 2011). Further-

more, lack of pertinent data and information are crucial

issues that may also constitute another impediment in the

whole process (Morc¸ӧl2006; Adam 2008). However,

indices have caused many debates regarding their objec-

tivity and reliability towards dealing with uncertainty

(Nardo et al. 2005; Saisana et al. 2005; Singh et al. 2009;

Rogge 2012). In this essence, Hilborn (1987) and Peterson

et al. (1997) identify three types of uncertainty that pertains

in composite indicators:

●Statistical Uncertainty—referring to the uncertainty on

the state of each variable (the components of the

composite indicators).

●Model Uncertainty—signifying the relation among the

variables, in other words, the structure of the composite

indicators.

●Fundamental Uncertainty—oncerning the novel situa-

tions which cannot be described by the existing

models/composite indicators.

Moreover, the greatest number of the disputes and criti-

cisms mainly focuses on the development process of such

indicators. Especially, when a particular process is obscure

and not characterized by clarity (OECD 2008). To some

extent, such debates may be justiﬁed since the development

process involves some steps in which the subjectivity may

be great and often unavoidable (Cherchye et al. 2007;

OECD 2008). In those steps, due to the lack of a univer-

sally accepted formulation procedure, the respective

developer is usually compelled to mostly arbitrarily select

or develop the various components in order to carry out the

index formulation. Consequently, these steps “bear the

signature” of the respective development procedures and

they are characterized by their reliability. Similarly to

many other customized tools and practices, the transfer-

ability of a successfully applied index structure may not be

enabled without speciﬁc considerations (Welte et al. 2004;

Peterson et al. 2007). All in all, the phases that mainly

concentrate the majority of the emerged criticism concern

the weighting, normalization, and aggregation processes

(OECD and European Commission 2008; Dobbie and Dail

2013; Gan et al. 2017). Each of those phases may be

achieved by a variety of methods and tools, but the

employment of simple approaches is mostly facilitating

both comprehension and calculation of an index (Booysen

2002).

Decision making in the ﬁeld of water resources man-

agement, most of the times, includes thorny issues that are

part of a complex environment. It usually incorporates

conﬂicting natural and anthropogenic interests and objec-

tives, multiple water uses, sensitive social and ecological

equilibria, recurring natural hazards. Frameworks like

integrated water resources management and sustainable

development are implement within the environmental

decision making process in order to mitigate adverse

effects and achieve consensus among different water user

groups (Margerum and Born 1995; Grigg 1996; Loucks

et al. 2005; Grigg 2008; Karavitis et al. 2015b). Thus,

regarding the implementation of any comprehensive

drought management scheme, the crucial obstacles to be

surmounted are having as epicenters drought’s onset, areal

extent, and severity. In order to overcome them drought

index methods are coming to an aid. Such an effort por-

trays drought characteristics based on a speciﬁc index,

which index may serve at the same time as a drought alert

mechanism (Oikonomou et al. 2018). This research need

was early on addressed in the literature (Palmer 1965;

National Research Council 1986; Dracup 1991; Karavitis

1992; Grigg 1996; Karavitis 1999; Wilhite et al. 2000).

Thus, a drought index, particularly a drought vulnerability

related one, should support decision making and convey

necessary information in demarcating drought’s beginning,

severity ﬂuctuations, and termination.

One of the attempts to ﬁll this gap was the development

of the Standardized Drought Vulnerability Index (SDVI)

that was initially designed to use Equal Weighting (Kar-

avitis et al. 2012b,2013,2014). Thus, in order to test if its

applicability is ampliﬁed in a weighted version, the current

approach was launched. The comparison between statisti-

cal techniques for weighting indicators has already been

extensively studied in the scientiﬁc literature (Saisana et al.

2005; OECD and European Commission 2008; Zhou et al.

2010). In the present attempt, SDVI is examined under ﬁve

weighting techniques that derive from four different

weighting methods. The SDVI was initially developed

using Equal Weighting and thus, three additional weighting

methods are tested, namely; the Principal Component

Analysis, the Multiple Correspondence Analysis, and the

Analytical Hierarchy Process. Their resulting values are

evaluated through statistical analyses. More speciﬁcally,

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

the statistical signiﬁcance of their differences in terms of

produced values is examined through correlations, struc-

tural uncertainty deriving from weights and input values.

The Sterea Hellas region in central Greece was selected as

an area of interest/application of the pertinent procedure

being inﬂicted very seriously by the 1989–1990, 1992–

1993, 2000 and 2007–2008 droughts (Tsesmelis 2010;

Karavitis et al. 2011a). The central objective of the present

effort is to reveal the signiﬁcance of the differences among

the examined weighting techniques and pinpoint any

potential implications that may emerge in decision making

from the application of the SDVI. Furthermore, a potential

contribution may be a framework of spatially assessing the

effects of different weighting methods in developing indi-

ces. The different models are compared employing three

different testing approaches. The ﬁrst part of the effort

analyzes the occurred drought event of August 2000. The

produced vulnerability maps are representing only the 2000

drought event. As the scope of the work was to test

structural uncertainty during the calculation of the SDVI,

the 2000 drought also serves as a case study event at the

region. In the second part, a higher spatial variability of

drought vulnerability is simulated, where the contributing

characteristics to the SDVI were randomly generated. The

scope of such an effort is to escape the local characteristics

and test the structural uncertainty of the SDVI in a hypo-

thetical region with greater variability. Finally, different

scenarios are employed for a speciﬁc location in the region

to also focus on the local effects of the different random

generated SDVI values.

2 Study area

Greece is located in south-eastern Europe and presents the

typical climatic characteristics of the northern Mediter-

ranean climate. The Sterea Hellas region (Fig. 1) located

almost at the center of Greece, occupies an area of 23,818.3

and constitutes of Attica (sector GR 06) and Eastern

Sterea Hellas (sector GR 07), two of the 14 water sectors in

the country. Sterea Hellas’ climate is generally temperate

along the coastline and colder in the interior, while its

average annual precipitation ranges from about 340 to

1050 mm/year (Fig. 1and Table 1). Table 1shows the 24

rain stations used with available rainfall data records from

1981 to 2010. It has to be pointed out that the stations

present signiﬁcant variations due to exposure and altitude.

Overall, Greece has an average annual rainfall ranging

from 250 to 2150 mm/year (decreasing from northwest to

southeast), with a country average of 760 mm/year

exhibiting at the same time, a very strong seasonal distri-

bution with cool rainy winters and prolonged hot dry

summers (Karavitis et al. 2011a,2014; Tsesmelis et al.

2019). Sterea Hellas region exhibits also the same trends

(Fig. 1). Hence, the precipitation’s extreme seasonal dis-

tribution makes water storage systems an absolute neces-

sity, a fact even more pronounced as Sterea Hellas is the

epicenter of such issues, since its total population surpasses

4.8 million inhabitants, making it the most populous region

of the country. Attica sector in particularly, where also the

capital city of Athens is located, displays the greatest

population density of the whole country with 987.9

inhabitants/Km

(HSA 2011). Attica’s water supply comes

from the dams and reservoirs of Marathon, Mornos, and

Euenos as well as from the natural lake of Hylice. The

water bodies of Marathon, Mornos, and Hylice, are within

the study area. Mornos reservoir and Hylice Lake lay west

of Attica in the prefectures of Phocis and Boeotia respec-

tively, supplying Attica including Metropolitan Athens,

with 200 km long aqueducts yielding an average water

discharge of 530910

/year. Thus the area, as well as

the whole of Greece, is heavily dependent on the annual

precipitation, since water supply of the major cities as well

as the large irrigation schemes to the center of it are relying

on such reservoir storage operated mostly on an annual

basis (Karavitis et al. 2011a,b). During periods of drought,

the reservoir stored water is usually reduced and hence,

increasing the vulnerability to pertinent drought impacts

affecting almost half of the whole country’s population.

All in all, any precipitation deﬁciency and moreover a

drought are producing signiﬁcant impacts, as the droughts

of 1989–1990, 1992–1993, 2000 and 2007–2008 have

clearly demonstrated (Karavitis 1998; Tsesmelis 2010;

Karavitis et al. 2012a). In this context, Greece and partic-

ularly Sterea Hellas are presenting almost an ideal envi-

ronment for the SDVI testing in an effort to assess its

structural uncertainty caused by different weighting

methods.

Additionally, in order to test SDVI from a regional to a

local scale, scenarios for a speciﬁc local were applied at the

Aliartos area. Aliartos area is a rural area in Sterea Hellas

(Fig. 1) comprises for its most part from the Boeotikos

Cephissus—Lake Hykice River basin. The hydrologic

complex of Boeoticos Cephissus and Lakes Hylice and

Paralimne is one of the most important hydrological sys-

tems of Sterea Hellas, supplying as pointed out, industrial,

agricultural and urban water. Morphologically, the region

in its biggest part is an extensive plain, composed of loose

sediments with slopes of less than 5%. The main part of the

plain is at an altitude between 100 and 200 m asl. Hence,

for such testing, the presented in the following methodol-

ogy was applied on the whole Sterea Hellas for regional

and in Aliartos area for local scale.

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

3 Methods and tools

3.1 The standardized Drought Vulnerability

Index

The SDVI is an index developed in the context of the

Drought Management Centre (DMCSEE) Project by the

Agricultural University of Athens (AUA) research team

(Karavitis et al. 2011b,2012b,2014). One of the parame-

ters needed to calculate SDVI is the Standardized Precip-

itation Index (SPI). The principal reason that SPI was

chosen in relation to other indices is that it can be easily

calculated from precipitation data time series. Such an

advantage may be proven valuable especially in cases

where there is lack of other climatic data. Furthermore,

according to Guttman (1998) and Hayes et al. (1999), the

SPI central beneﬁt is that through the different time scales

of precipitation irregularities, it can describe short term and

long term drought events and impacts. Additionally, it is

statistically consistent and has a great performance in

drought risk analysis (Guttman 1999).

SDVI aims at integrating the various manifestations of

drought (Meteorological, Agricultural, Hydrological, and

Socioeconomic) and the concept of vulnerability into a

single value. It comprises of six components cSPI-6 (re-

classiﬁed SPI-6), cSPI-12, Supply, Demand, Impacts, and

Infrastructure (Karavitis et al. 2014; Oikonomou et al.

2019). These six components are classiﬁed into four vul-

nerability categories (0–3 scale) according to their perfor-

mance as presented in Table 2. To derive such a

categorization extensive pertinent literature review and

expert knowledge was applied for the percentages scaling

of the various components (Karavitis et al. 2012b,2014;

Tsesmelis 2017). The vulnerability value is calculated by

the average scaled value of the components following

Eq. (1). In this regard, the SDVI adopts the mechanics of

the Environmental Vulnerability Index that was developed

by the South Paciﬁc Geoscience Commission (Kaly et al.

2004; Skondras et al. 2011).

SDVI ¼1

i¼1

Scaled Values of the Components ð1Þ

The SDVI application procedure is describe in detail by

Karavitis et al. (2014), and brieﬂy in the following:

1. Precipitation patterns usually affect the surface waters

and the recharge of the aquifers. SDVI employs the

SPI that is an index based on precipitation time series.

For the calculation of the index, the ﬁrst step is to

identify the probability density function, which better

describes the distribution of the time series over the

different time scales (6 and 12 months) and then the

current process is operated for every month of the

precipitation time series. Urban water supply and

Fig. 1 Sterea Hellas elevation (left) and annual average precipitation (1981–2010) (right). The black line portrays the counties boundaries and the

black dot on the upper right map depicts the location of Aliartos area

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

irrigation are usually rely signiﬁcantly on annual

reservoir and/or aquifer storage. Hence, the 12-month

SPI depicting long-term anomalies becomes pivotal for

the SDVI estimation, since it may exemplify the

overall water availability (irrigation, hydropower,

industry domestic and tourism). SPI-6 may denote

the non-irrigation water availability for rain-fed crops,

as it may effectively show the rainfall patterns in

different seasons, pointing out medium-term trends. It

is has presented that their mutual incorporation

improves interconnections, operability and may lead

to more sound results (Karavitis et al. 2014). The

values of cSPI-6 and cSPI-12 are calculated on local

(meteorological station) scale. Overall, SPI does not

impose any restrictions, except that it does not have

application in the deserts or at the earth poles.

2. Supply and Demand, which describe the deﬁcits in the

corresponding capacities (including network losses),

based on pertinent data. They depend on the available

water quantity, existing water consumption, reported

uses, and demographic, social and technical develop-

ment patterns.

3. Impacts that express the losses in monetary units.

Usually, impacts mitigation measures on the demand

side of the supply-minus-demand economic drought

equation, have the basic objective to trim the water use

Table 1 The basic statistics of precipitation in Sterea Hellas

Station UTM X UTM Y Average annual precipitation (1981–

2010)

Elevation in m

asl.

Standard

deviation

Minimum Maximum

Agia Triada 405136 4244800 986.56 400 220.27 534.3 1382.7

Amﬁssa 358886 4265827 681.94 180 161.66 328.4 1025.7

Gravia 363497 4280548 822.31 450 143.23 431.3 1048.9

Davlia 389166 4248703 770.83 380 153.20 463.5 1062.1

Distomo 383406 4253888 605.78 450 121.66 329.4 821.7

Eptalofos 367725 4273077 972.64 830 221.66 612.1 1420.1

Zilefto 349557 4310404 394.19 120 166.89 96.9 835.7

Thisvi 409381 4233654 458.00 174 149.83 185.6 874.9

Itea 363056 4254654 377.45 20 119.48 187.7 660.9

Kalithea 451708 4238840 468.85 333 157.18 321.3 753.6

Kaloskopi 354830 4282551 879.01 1000 235.07 568.0 1740.1

K.Tirothea 388071 4274616 655.60 170 129.32 388.8 991.9

Lamia 361050 4306493 538.38 144 115.08 328.0 870.5

Livadia 400880 4254100 737.21 200 156.38 414.1 1037.8

Lilea 369237 4276752 848.55 330 335.74 465.2 1272.4

Pavlos 421355 4264972 479.07 200 160.67 82.5 741.5

Trilofo 345367 4317887 598.79 580 126.42 373.0 826.5

Tymfristos 319174 4309189 1022.99 850 382.17 141.0 1713.3

Elliniko 475537 4194336 363.63 15 96.39 155.4 546.6

Asteroskopio 475089 4202597 397.06 110 138.72 150.6 896.0

Elefsina 440916 4212390 344.17 31 91.06 117.4 468.3

Tatoi 480783 4218209 435.09 235 153.46 169.6 812.7

Filladelphia 477479 4210345 452.17 138 177.52 180.5 1015.5

Ypati 346524 4303061 790.74 286 293.18 327.6 1431.0

Table 2 SDVI components

vulnerability scale classiﬁcation Vulnerability level Scales categorization based on SPI classes and experts’ knowledge

cSPI-6 and cSPI-12 Supply, demand, impact and infrastructure

Non vulnerable 0 Wet ≥1.50 0 No deﬁcits

Vulnerable 1 Quite wet 0–1.49 1 15% Deﬁcits

Highly vulnerable 2 quite dry 0 to −1.49 2 16–50% Deﬁcits

Extremely Vulnerable 3 Dry ≤−1.50 3 [50% Deﬁcits

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

of the least unit impact, provided that legal consent

permits it. Then, the resulting impacts are mainly

focused on economic costs and production losses.

Environmental impacts are analyzed only if they can

have a direct monetary quantiﬁcation. Drought social

impacts assessments are very rare and usually present

qualitative criteria and social stress indicators with the

latter ones being very difﬁcult to quantify. The current

effort has not endeavor in such an estimation. It is

pointed out that the developed Drought Vulnerability

Index is portraying above all the magnitude of the

societal exposure and disruption to drought and its

output is also a societal Index of drought pressures. All

in all, it may be suggested that for the last two

categories of impacts more specialized analyses should

be applied in order to quantify them in an another

pertinent research approach. Such approach may use

the SDVI as an initial indication for more specialized

research tools.

4. Infrastructure that imports to the index the status of the

current operational water infrastructure in terms of its

deﬁciency magnitude. The components of infrastruc-

ture and supply should be cautionary treated, as on one

hand they may easily spark some misperceptions (e.g.

a 15% of an infrastructure deﬁciency may in some

cases, because a proportional supply deﬁcit). On the

other hand, an infrastructure may be in excellent

operational condition, and a supply deﬁcit still to be

induced e.g. in agriculture as it may be decided to

divert irrigation water during a drought in order to

fully supply an urban population. Hence, the infras-

tructure component incorporates its condition and

degree of efﬁciency in conveying water timely and

appropriately.

In this regard, the SDVI is calculated on a monthly step. In

a given area (basin/county) the total supply data and

demand deﬁcits, as well as the resulting impacts are rep-

resented as ratios (proportions) over the highest value of

the corresponding parameters, while the infrastructure sub

index is portrayed by a weighted average value (ratio) of

the infrastructure performance (Karavitis et al. 2014). The

SDVI may be expanded to other regions with similar or

different climates since the SDVI index is calculated using

average values and deviations for the supply, demand and

impact sub-indices (Karavitis et al. 2015a; Oikonomou

2017; Oikonomou and Waskom 2018; Oikonomou et al.

2019; Tsesmelis 2017). The SDVI results are classiﬁed into

six vulnerability categories ranging from 0 to 3 as shown in

Table 3(Karavitis et al. 2014). Finally, all the obtained

data as well as the SDVI results, were visualized on a GIS

environment using geostatistical techniques (Karavitis

et al. 2012a,2014).

3.2 Weighting methods and structural

uncertainty analysis

Weighting methods are a step during the development of an

index that usually attracts criticism because different

weighting methods could have different index values. In

this context, the equal weights technique is used mainly

because in most cases there is no objective way to deter-

mine the relative importance of each component (Garriga

and Foguet 2010). In similar approaches incorporating

social, economic, and environmental aspects in composite

indicator development for e.g. land exposure to natural

events as the: Human Development Index and the Envi-

ronmentally Sensitive Areas (ESA) index, the equal

weights usage achieves effective results (Sagar and Najam

1998; Kosmas et al. 2006; Becker et al. 2017). Thus, the

equal weights approach by giving the same importance on

every index component, and especially for vulnerability

related indices is the most common technique employed

(Eakin and Bojo

´rquez-Tapia 2008). Other methods used in

developing indices may be the Principal Component

Analysis, the Multiple Correspondence Analysis, the

Analytical Hierarchy Process, the Conjoint Analysis, and

the Beneﬁt of the Doubt (OECD and European Commis-

sion 2008; Rogge 2012; Kosmas et al. 2014; Li et al. 2015).

Each of those simpler methods is characterized by its

respective advantages and disadvantages, while providing

different sets of weights and, consequently, different sets of

index values. Table 4explicitly portrays such advantages

and disadvantages stating relevant comparisons among the

pertinent methods as well as corresponding references.

Table 3 SDVI scaled values (Karavitis et al. 2014)

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

Table 4 Advantages and disadvantages of various weighting methods

Weighting

technique

Advantages Disadvantages Source

Equal weighting

(EW)

Removes the overweighting

and underweighting

features

Easier to construct

May deviate from reality

Disguises the absence of a statistical or

an empirical basis

OECD and European Commission (2008)

Principal

component

analysis (PCA)

Sustains the maximum

variance within the data

Greater weights are assigned

to variables with greater

variability

Depends on the data

Sensitive to outliers

Requires correlation among the

variables

Minimization of the variables

contribution with different behavior

OECD and European Commission (2008)

Multiple

correspondence

analysis (MCA)

Can be applied with both

quantitative and nominal

variables/indicators

Delicate and quite difﬁcult to apply if

necessary data are not available

Abdi and Valentin (2007)

Analytical

hierarchy

process (AHP)

Decomposes a decision

problem into its

components

Captures both subjective and

objective evaluation

measures

Controls inconsistencies

Supports group decision-

making

Develops scales where

measures ordinarily do not

exist

Ranking irregularities can occur

Compensation between good scores on

some criteria and bad scores on other

criteria can occur

The number of pairwise comparisons to

be made, may become very large

Difﬁcult to distinguish among the

applied scales

Kasperczyk and Knickel (1996), Zahir (1999),

Ramanathan (2001), Millet and Wedley

(2002), Macharis et al. (2004)

Conjoint analysis

(CA)

Breaks the task into a series

of choices

The results may be used for

model development

Calculates utilities at the

individual level

Straightforward experimental

designs;

Easily used in hybrid

methodologies

Designing conjoint studies can be

complex

Poorly designed studies may over-value

preference variables and undervalue

concrete variables

Desarbo et al. (1995), Orme (2009)

Beneﬁt of the

doubt (BD)

Useful in policy making

Helps to deﬁne trade-offs

The weights are determined

by the observed

performances

The indices are on a linear

combination of observed

best performances

In some cases the optimal set of

weights may be undetermined

Higher weights are assigned to higher

scores

In some cases, the best performer will

not see its progress reﬂected in the

index

Cherchye et al. (2004), OECD (2008), Rogge

(2012)

Unobserved

components

model (UCM)

Decomposes the problem at

hand

Provides powerful models

Easily interpreted results

Data dependent

Sufﬁcient data dependent

The indicators should not be correlated

Kaufmann et al. (1999,2003), Fomby (2008),

OECD and European Commission (2008)

Budget allocation

process (BAP)

Transparent

Relatively straightforward

Short duration

Requires well-deﬁned framework

Optimal for a maximum of 10–12

indicators

The selected experts should be

specialists on the given sub-index

Munda (2005,2007), OECD and European

Commission (2008)

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

Additionally, according to the OECD and European

Commission (2008), in the case of similar results with the

various weighting methods the simplest method should be

chosen. Furthermore, other complex weighting methods

have been emerged that apply fuzzy logic (Sodhi and

Prabhakar 2012) or even combine spatial and temporal

weighting (Wang et al. 2005; Huang et al. 2010; Harris

et al. 2010). In such a setting, the unavoidable cus-

tomization increases the inherent subjectivity of the pro-

cess (Singh et al. 2009), while, at the same time, the

uncertainty towards the formulation of decisions may reach

critical levels. The latter methods have not been considered

and thus they are not further examined. In this rationale,

simpler waiting methods have been chosen in this effort,

despite the nascent complexity of any vulnerability illus-

tration, following the premise of accepting a reasonable

approximation of such illustration, which may be very

valuable for applicable drought management decision-

making.

In the current effort, the original weighting method

(equal weighting-EW) of computing the SDVI index is

compared to four weighting techniques that derive from

three different weighting methods employed. Apart from

the EW method, the next examined one, Principal Com-

ponent Analysis (PCA), is a multivariate statistical proce-

dure that uses orthogonal transformation to identify the

principal components that explain most of the variance in

the original data set. The index components are weighted

with the proportion of variance in the original set of vari-

ables and it could be argued that it is a relatively more

objective option for weight selection (Booysen 2002).

Multiple Correspondence Analysis (MCA) is also a sta-

tistical method which identiﬁes the relationship patterns

between categorical variables and it could be seen as

analogous to Principal Component Analysis (Abdi and

Valentin 2007). Last, the Analytic Hierarchy Process

(AHP) is a multi-criterion decision making method which

through expert opinion and pairwise comparison the

weights of the SDVI models are calculated (Saaty

1980,1988; Sadiq and Tesfamariam 2009). Thus, the main

objective of comparing the SDVI results deriving from

different weighting methods is to help gain a better

understanding of their impacts on the SDVI results. In

other words, it is a means to test the robustness of the

SDVI.

In order to assess the structural uncertainty caused by

the selected four weighting methods, ﬁve techniques

(PCA1, PCA2, EW, AHP, and MCA) were applied. Then,

three distinct approaches in the area were used to supply

different data sets to the models for the structural uncer-

tainty analysis of SDVI. The ﬁrst approach comprises of

the existing data for the recorded conditions. The vulner-

ability to drought values derived by the ﬁve techniques

applied in Sterea Hellas and they were spatially represented

in GIS environment. Index values of each produced map

were extracted in order to create a sample of data (N =

23,429, the number of raster tiles) according to the existing

conditions of the study area and to examine the similarity

of the results. The second approach explores the situation

where the study area exhibits large variability in the SDVI

components, as the existing natural and anthropogenic

conditions in Sterea Hellas are rather homogeneous. In

order to achieve this, ten thousand random component class

values (0–3), from each component, were generated from a

uniform distribution using R (R Core Team 2016), so as to

simulate an area that could have greater variability.

Finally, the third approach was to test local effects of the

different weighting techniques. The Aliartos area serves as

case study. Four scenarios (Α, B, C and D; presented in

detail on Tables 5and 11) were generated, each one rep-

resenting a feasible and plausible set of SDVI components’

state. For the scenarios creation, SDVI is divided in three

ﬁxed components and three varied ones. The ﬁxed com-

ponents are cSPI-6, cSPI-12, and infrastructure, whereas

varying ones are considered to be supply, demand and

impacts. Thus, scenario A depicts the real recorded con-

ditions for the selected area. Scenarios B, C, and D reﬂect

future plausible combinations of conditions in all the

varying components (increasing/declining demand,

increasing/declining supply, etc.).

All in all, it is believed that the above methodology

enhances the understanding of the examined weighting

techniques’ effects on SDVI values in various conditions

(real and hypothetical). This multi-approach analysis may

contribute in drought vulnerability assessment, by trying to

quantify the uncertainties in the various index weighing

methods.

3.3 Data sets for the SDVI

In the present attempt, the SDVI is applied to picture the

vulnerability conditions that occurred in Sterea Hellas on

August 2000 a dry hydrological year (Karavitis et al.

2012a,b). The gathered data sets reﬂect the conditions

occurred in August 2000 and are presented in Fig. 2.

More speciﬁcally:

Table 5 Possible four scenarios in Aliartos area

Scenario Fixed Varied

cSPI-6 cSPI-12 Infr. Supply Demand Impact

A331033

B33112 2

C33121 1

D331112

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

●Precipitation data from 1981 to 2010 for 24 stations in

the area (Table 1), were collected from the National

Meteorological Service of Greece (HNMS 2016), the

Special Secretariat for Water (SSW 2016) and the

Public Power Corporation (PPC S.A. 2016). Data gaps

in all 24 stations were minimal and raw data seem more

appropriate to be inserted so as to represent the natural

drought conditions (extreme minimum values), hence,

no attempt made to apply “corrective” homogeneity in

the data sets (Karavitis et al. 2011a). The implicit sta-

tionary assumptions of the SPI were tested with all

stations exhibiting stationary process (Karavitis et al.

2015b). The SPI then was calculated on monthly basis

both with the 6 month (SPI6) and 12 month (SPI12)

step.

●The Demand component was calculated by taking into

account urban, industrial and agricultural water demand

on county scale. Attica region has the greatest industrial

concentration of the whole Greece with an industrial

water consumption of more than 50910

year

−1

(Karavitis et al. 2001; Barraque

´et al. 2008). The urban

water supply of Attica region, including Metropolitan

Athens, has an average water consumption of 365 9

/year. The remaining urban water consumption

in the area of Sterea Hellas reaches 70 910

/year

(SSW 2013). In this case.

i. For the calculation of agricultural demand, both

crop demands and water losses were estimated

using the crop water requirements as a function

of Crop evapotranspiration and Actual

evapotranspiration (Suat et al, 2003). In other

words, Crop Water requirement equals Crop

evapotranspiration minus Actual evapotranspira-

tion (Doorenbos, J. & Pruitt, W.O., 1977; Allen

et al, 1999). Actual evapotranspiration Et

,is

calculated on a plot scale, however pertinent data

are not available on the Case Study Area (Allen

et al, 1999; Jensen & Allen, 2016). It is also

pointed out that during a drought (without

irrigation), actual evapotransipation (Et

) reaches

asymptotically to zero (as by deﬁnition refers to

the condition of existing water supply), since

drought progresses and water supply diminishes.

Hence, Et

cannot alone represent the water

demand (or Crop Water requirements). Crop

evapotranspiration ET

may be estimated by the

equation: ET

*ET

, where Kc is the crop

coefﬁcient according to the growth rate and ET

is the reference evapotranspiration (Allen et al,

1999; Jensen & Allen, 2016). Then, the equation

used for sub-index agricultural water demand (in

other words crop water requirement) was:

SubindexWater Demand Aug 2000

¼ETcAug 2000

ETcAug 1



100;

Which becomes by substituting from the previ-

ous equation:

Fig. 2 Collected datasets for Sterea Hellas region

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

SubindexWater Demand Aug2000

¼EToAug2000

EToAug 1



100;

Therefore, it is only needed to estimate ET

ii. ET

can be estimated by several methods which

include models based on pan evaporation, radi-

ation, temperature, and mass transfer (Alexandris

et al. 2006; Ambas and Baltas 2012; Nikam et al.

2014; Chatzithomas and Alexandris 2015). For

the calculation of SDVI the temperature based

method of Hargreaves-Samani (1985) was chosen

(Eq. 2). The method was selected as it is

recommended in cases where reliable data are

lacking (Droogers and Allen 2002; Allen et al.,

1999); furthermore, it usually performs better

than other temperature based estimation methods

(Valipour 2014; Valipour and Eslamian 2014).

ETo¼0:0023

Tmax Tmin

ðÞ

0:5Tmax þTmin

2þ17:8



Ra

ð2Þ

where Tmax and Tmin and are maximum and

minimum temperature respectively (°C) and Rais

the extraterrestrial radiation (mm/day).

iii. The data used for the calculation of urban and

industrial water demand were retrieved from the

River Basins Management Plans of Greece (SSW

2013).

Thus, it would seem that the water demand

estimation strengthens the index testing process

by applying it to the complete spectrum of water

uses (agricultural, domestic and industrial

demand).

●The Supply component was calculated on a county

scale, the average size of the counties is approximately

3200 km

, according to the registered delivered ﬂow

(including losses) in the major water supply networks.

The counties boundary lines are presented in Fig. 1. The

corresponding data were collected from the pertinent

authorities (SSW 2013).

●For the calculation of the economic and environmental

impacts component (as analyzed in Sect. 3.1) the

reduction of mainly the agricultural production was

used on county scale. More speciﬁcally, for that

particular component, the difference (%) in agricultural

production between August of 2000 and the average

agricultural production of the same month for previous

5 years was calculated. The corresponding data were

collected from the Hellenic Statistical Agency (HSA

2011).

●The calculation of the Infrastructure component was

performed according to the difference between the

designed and the actual capacity of the reservoirs

(dams) and other pertinent infrastructure. The corre-

sponding data were collected on county scale from the

River Basins Management Plans by the Special Secre-

tariat for Water (SSW 2013).

The w1 to w6 symbols in Fig. 2represent the weights

assigned to each component for the calculation of the SDVI

values.

4 Application, results and discussion

Based on the gathered data and the previously presented

methodology, the SDVI has been calculated with each of

the following weighting techniques:

●The equal weighting technique (EW) The initial form of

the SDVI employs that technique which is described by

Eq. (1).

●The principal component analysis (PCA) In order to

have a large database for a complete as possible

provision of principal components, the application of

the PCA was performed on the raw data gathered from

the seven of the initial DMCSEE countries (Karavitis

et al. 2013) namely Slovenia, Hungary, Serbia, Mon-

tenegro, Bulgaria, FYROM and Greece, but of course

excluding data sets of Sterea Hellas, the case study area.

The region encloses 224 precipitation stations (more

than 700,000 values) and provided six principal

Table 6 The weights from the

ﬁve weighting techniques Methods cSPI-6 cSPI-12 Supply Demand Infrastructure Impacts

EW 0.166 0.166 0.166 0.166 0.166 0.166

PCA1 0.100 0.050 0.320 0.320 0.150 0.060

PCA2 0.170 0.170 0.200 0.200 0.100 0.160

AHP 0.201 0.193 0.154 0.154 0.125 0.173

MCA 0.230 0.250 0.100 0.100 0.130 0.190

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

components. However, only the ﬁrst two components

have been selected for the calculation of the SDVI

weights. Two different approaches were tested. In the

ﬁrst (PCA1), the indicators’ weights have been deter-

mined by the squared loadings of the ﬁrst principal

component (Explained variance: 46.40%). In the second

approach (PCA2), the squared standardized values of

both the ﬁrst and second principal components (that

were selected out of six and explain 74.12% of the data

variance) were multiplied to their respective standard-

ized percentage of explained variance and were added

to each other. The result of the latter process was a

weighted averaged set of weights.

●The analytic hierarchy process (AHP) The AHP

application was built on the weighting preference of

15 water resources management experts (University

Professors, Researchers from Greece, Slovenia, Serbia

and Portugal, and Greek Ministries Employees). It is

worth stating that all the experts provided zero valued

consistency ratios, meaning that their responses were

not inconsistent due to confusion or other related

reasons (Saaty 1980). That was quite expected consid-

ering the small number of components and the expe-

rience of the correspondents. The ﬁnal weights are

calculated as the mean value derived by the experts’

weighting preference. In this technique, the SDVI

weights were calculated on individual level (per expert)

and the ﬁnal weights derived from the mean value of

the 15 sets of weight.

●The multiple correspondence analysis (MCA) Com-

pared to PCA—which was used on the raw collected

data—the MCA was used on the scaled (ordinal) values

of the SDVI components. The weights of this technique

derived from the ﬁrst six dimensions (Explained

variance: 73.57%).

The weights which derived from the two PCA, AHP

and MCA approaches, as well as, from the EW process

were inserted into ﬁve linearly aggregated models

(Table 6). All the six index sub-indices have been visual-

ized in a GIS environment and are displayed in Figs. 3and

Fig. 3 SPI 6 and 12 monthly step and classiﬁed SPI 6 and 12 monthly step (cSPI6 and cSPI12) visualization for August 2000

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

4. August 2000 was chosen for the SPI visualization, as

Greece was severely stricken by the 2000 drought. Con-

secutively, the SDVI calculations were made for August

2000. According to Fig. 3and based on SPI-6 (upper left

quarter), the area of interest displays “severe” to “extreme”

drought conditions in a uniform pattern. Conversely,

according to SPI-12 (lower left quarter), the area presents

almost “normal” conditions with some areas being classi-

ﬁed as “mildly” to “very wet” while others classiﬁed as

being “mildly” to “severely dry”. However, the data

transformation from SPI to cSPI involved a decreased

number of classes and produced maps which display quite

uniform drought vulnerability conditions. More speciﬁ-

cally, cSPI-6 presents uniform conditions of extreme vul-

nerability across the area of interest, except on three areas

with “highly vulnerable” classiﬁcation. The cSPI-12 the

uniformity of high drought vulnerability is mildly disrupted

by some areas classiﬁed as “less vulnerable” and “vulner-

able” or even “extremely vulnerable”.

Figure 4presents the scaled values of the remaining

SDVI components calculated on county scale, based on the

classiﬁcation of Table 2. In the case of Demand compo-

nent, the area of interest displays almost uniform condi-

tions of “no vulnerability” with scattered areas classiﬁed as

“vulnerable” to “extremely vulnerable”. Those areas are

either highly urbanized locations or areas with signiﬁcant

agricultural production. Continuing, the Supply sub-index

in Fig. 4reﬂects conditions of high to extreme vulnera-

bility. The produced result was expected since ﬂows pre-

dominantly for agriculture, in the major water supply

networks were dramatically reduced during August of

2000. The Impacts component pinpoints the loss of agri-

cultural production reﬂecting “vulnerable” to “extreme

vulnerable” conditions. The result may seem rough enough

but data on ﬁner spatial resolution were not available.

Finally, the Infrastructure component reﬂects the difference

between the designed and the actual capacity of water

supply infrastructure. It has to be mentioned that apart from

two areas that display conditions of “less vulnerability”, the

Fig. 4 Visualization of four sub-indices: demand, supply, impacts and infrastructure for August 2000

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

remaining area of interest is classiﬁed as being “vulnera-

ble” to “highly vulnerable” to drought. On one hand, by

referring also in combination with Fig. 1, the northwestern

part that displays lesser vulnerability is a highly

mountainous area with minimal anthropogenic activities

and, consequently, it cannot be affected by the examined

infrastructure deﬁciencies. On the other hand, the high

vulnerability region in Fig. 4coincides with the Athens

Fig. 5 August 2000 SDVI results for the ﬁve weighting techniques

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

Metropolitan and industrial area, where all the water is

supplied by the relevant infrastructure and if there are

deﬁciencies are directly associated to high impacts.

According to the gathered data and the produced results,

the summer season of 2000 (August in particular) was a

difﬁcult period for Sterea Hellas (and the country in gen-

eral) in terms of drought conditions. During that period

quite a portion of the population was affected in terms of

limited water supply and high agricultural losses. Drawing

on the produced material (Fig. 5), the whole region of

Central Greece may be classiﬁed as being “Highly Vul-

nerable” by four out of the ﬁve (EW, AHP, and PCA2 and

MCA) weighting approaches and “Medium Vulnerable” by

the remaining one (PCA1). Looking at Table 5, PCA1 and

2 prioritize the overall supply and demand and MCA

boosts the meteorological component through SPI and by it

predominantly the natural water supply. Such facts are

reﬂected in Fig. 5, which displays the SDVI results derived

from the ﬁve weighting techniques that were tested.

Regarding PCA, Fig. 5correctly displays with lower

Table 7 Descriptive statistics of

the ﬁve weighting techniques in

Sterea Hellas based on real

conditions

SDVI EW SDVI PCA1 SDVI PCA2 SDVI AHP SDVI MCA

N 23,429 23,429 23,429 23,429 23,429

Mean 1.541 1.443 1.584 1.599 1.632

Std. error of mean 0.002 0.002 0.002 0.002 0.002

Median 1.494 1.430 1.530 1.548 1.570

Std. deviation 0.311 0.359 0.322 0.315 0.314

Variance 0.097 0.129 0.103 0.099 0.099

Range 1.826 1.850 1.900 1.875 1.860

Minimum 0.664 0.800 0.730 0.682 0.620

Maximum 2.490 2.650 2.630 2.557 2.480

Table 8 Correlation matrix

between the ﬁve weighting

techniques based on real

conditions in Sterea Hellas

SDVI EW SDVI PCA1 SDVI PCA2 SDVI AHP SDVI MCA

SDVI EW 1 0.879 0.985 0.988 0.507

SDVI PCA1 1 0.912 0.849 0.419

SDVI PCA2 1 0.989 0.501

SDVI AHP 1 0.509

SDVI MCA 1

Table 9 Descriptive statistics

for Sterea Hellas results of the

ﬁve weighting techniques with

random input values

SDVI EW SDVI PCA1 SDVI PCA2 SDVI AHP SDVI MCA

N 10,000 10,000 10,000 10,000 10,000

Mean 1.498 1.498 1.499 1.499 1.500

Std. error of mean 0.005 0.006 0.005 0.005 0.005

Median 1.500 1.510 1.500 1.502 1.500

Std. deviation 0.461 0.556 0.471 0.468 0.491

Variance 0.213 0.309 0.222 0.219 0.241

Range 2.830 2.950 2.830 2.850 2.900

Minimum \0.001 \0.001 \0.001 \0.001 \0.001

Maximum 2.830 2.950 2.830 2.850 2.900

Table 10 Correlation matrix of

the ﬁve weighting techniques

with random input values

SDVI EW SDVI PCA1 SDVI PCA2 SDVI AHP SDVI MCA

SDVI EW 1 0.830 0.981 0.988 0.942

SDVI PCA1 1 0.868 0.775 0.614

SDVI PCA2 1 0.982 0.915

SDVI AHP 1 0.973

SDVI MCA 1

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

vulnerability the Athens metropolitan area, where infras-

tructure conveys the urban water supply, than the central

agricultural areas where irrigation water supply was inter-

rupted in favor of the cities. Additionally, it should be

mentioned that in the PCA1 approach the produced weight

for the impact component is rather low resulting in

underestimating drought vulnerability, when there are

signiﬁcant impacts. It should be also mentioned that the

increase in ET

caused an increase in the demand in the

irrigated areas, which was not supplied due to a manage-

ment decision to divert water to the cities, despite the fact

that the infrastructure operational condition was at the

normal levels, and thus in these areas impacts were more

intense.

The results of the weighting techniques uncertainty

analysis are described below with the order mentioned in

Sect. 3. Table 7shows the descriptive statistics of the

techniques based on the real conditions (Fig. 5) in the study

area. It should be underlined that the area has a relative

small variability. Table 8displays the similarity metric

(correlation) between the techniques results (real condi-

tions). All the techniques exhibit a high correlation apart

from the MCA method which has the lowest correlation

with all the other ones. Commenting on the MCA, as it

prioritizes the meteorological dimension, natural water

deﬁciency dictates a high vulnerability value, which was

true for the central agricultural areas, but not true for the

Athens area, as Fig. 5portrays, where the water deﬁcit is

consistently and efﬁciently supplied by the greatest water

infrastructure works in the country. However, since the

supplied water comes from surface water reservoirs in the

western part of Sterea Hellas, a meteorological drought in

the reservoirs’ watersheds plays an extremely important

role for the future years’ water supply and thus the index

results should initiate drought contingency measures.

The following tables (Tables 9,10) display the

descriptive statistics and correlations produced by random

components class values for the ﬁve weighing techniques

and the corresponding statistics were derived. It is only

noted that the pertinent values of median, mean and stan-

dard deviation fall within the expected limits of the

generation approach. The values produced with the random

mode (Table 8) have a greater standard deviation than the

real condition’s one (Table 7), in an effort to overcome the

region’s homogeneity and thus to be able to test conditions

with higher variability of drought vulnerability. The tech-

niques with random values are correlated rather well and

MCA is strongly correlated with all the other ones, with the

PCA1 technique being the one having the lowest correla-

tion with the rest of the techniques (Table 10), which can

be attributed to the low weight assigned to the impact

component, in other words as explained it prioritizes sup-

ply and demand.

The results of the two previous approaches on examin-

ing the similarity of the techniques with real values and

second with randomly generated ones for cases that there is

a larger variability in an area are based on spatial data and

the overall performance was calculated. For the third

approach to test local effects of the different weighting

techniques the SDVI components’ selected for the sce-

narios are shown in Table 11. The results point out that the

SDVI values for the Aliartos area with different weighting

techniques can have signiﬁcant differences among them-

selves, depending on the scenario, resulting in different

drought vulnerability classiﬁcation. Especially, when the

PCA1 and MCA techniques are used for when impacts are

signiﬁcant, the SDVI values tend to be classiﬁed one level

lower or higher, respectively, than with another weighting

approach. This is attributed to the fact that the PCA1 pri-

oritize Supply and Demand whereas MCA boosts the

meteorological dimension as it has been also noted in the

previous approaches.

5 Conclusions

Decision makers usually count on a great number of tools,

methods and techniques as well as on simpliﬁed heuristics,

rules of thumb and frames, which are usually intertwined to

provide a representation of the complex reality so as to lead

to appropriate courses of action. In this context, indices are

used as tools to represent complex systems and they are

Table 11 Possible scenarios in Aliartos area and the computed SDVI values for each weighting technique

SCENARIO Fixed Varied Weighting techniques results

cSPI-6 cSPI-12 INFR. SUPPLY DEMAND IMPACT EW PCA1 PCA2 AHP MCA

A 3 3 1 0 3 3 2.16 1.74 2.20 2.29 2.44

B 3 3 1 1 2 2 1.99 1.68 2.04 2.12 2.25

C 3 3 1 2 1 1 1.83 1.62 1.88 1.94 2.06

D 3 3 1 1 1 2 1.83 1.36 1.84 1.96 2.15

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

built around such elements. Subjectivity and uncertainty

are constantly present in the various index development

phases and affect their reliability and effectiveness towards

informed decision making. Hence, testing of weighting

methods for indices becomes crucial in any effort to

examine drought vulnerability.

The scope of this effort was to investigate the effect of

various weighting methods and techniques upon the values

of such an index, as the Standardized Drought Vulnera-

bility Index (SDVI), and to pinpoint any potential inter-

connection and implication towards decision making

efforts. Four weighting methods were examined, namely;

the Principal Component Analysis (PCA), the Analytic

Hierarchy Process (AHP), the Multiple Correspondence

Analysis (MCA) and the Equal Weighting (EW). In the

case of PCA, two different component based techniques

were tested. The comparison of the examined techniques

was achieved via a framework that incorporates three dis-

tinct testing approaches. Firstly, the SDVI outcomes from

the ﬁve weighting techniques were correlated under a real

drought case in the region of Sterea Hellas, Greece. Then,

random SDVI values were generated to test the weighting

effects in the case that a region may exhibit higher spatial

component variability. Finally, it was examined if there

were any signiﬁcant differences among the examined

weighting techniques on a local scale, by developing sce-

narios for an agricultural area (Aliartos), a part of Sterea

Hellas region.

The analysis of the applied techniques with the various

tested approaches (actual values, random values and sce-

narios) indicated that in all cases the Equal Weighting

Technique (EW), the Principal Component Analysis

(PCA2) and the Analytic Hierarchy Process (AHP) might

equally serve as tools for the decision makers, since they

all perform in a similar manner for the vulnerability

characterization. At the same time, PCA1, presented high

correlations with the other methods, but during extreme

drought events falls short to depict the state of the system

as it is capturing lower system drought impacts by design.

Contrariwise, the MCA method does not always follow the

other ones, especially in cases where the sample depends

on speciﬁc conditions as the applied scenarios depicted.

The most important ﬁnding of the present effort may refer

to the statistically signiﬁcant differences that were detected

among the actual drought case, the random SDVI values,

both on regional scale, and the scenario-based analysis on

local level. However, such differences may not be easily

detectible in real conditions applications and the presence

of these differences dictates that decision making efforts

may be at risk in cases where the various indices are treated

as fully reliable tools or without further reservations. In this

context, the index classiﬁcation pattern/scale, which is

usually less ﬂexible compared to the respective index

values, becomes more important to decision makers and it

may affect their decision.

All in all, the use of complex weighting techniques

seems not always necessary and the simple equal weighting

method may be effective in representing drought vulnera-

bility. However, the inherent complexity and uncertainty in

drought vulnerability assessment propagates in using sim-

ple weighting methods and does not necessarily mean that

the simple equal weighting method is more effective than

other complex ones. SDVI performed well in all approa-

ches and characterized well enough the drought vulnera-

bility conditions to support drought management options.

Nonetheless, research using a greater number of indicators,

various regions, alternative index structural steps, different

map scales, decision making simulations and above all,

reliable data is required for a more comprehensive appli-

cation. It is important to address the search for valid,

reliable and relevant indicators that would measure per-

formance of water bodies and environmental conditions,

monitor the process of equitable water sharing, and provide

mechanisms for monitoring state and changes in interde-

pendent water systems.

References

Abdi H, Valentin D (2007) Multiple correspondence analysis. In:

Salkind N (ed) Encyclopedia of measurement and statistics.

SAGE Publications Inc, Thousand Oaks, pp 651–657

Adair J (2010) Decision making and problem solving strategies, 1st

edn. Kogan Page, London

Adam F (2008) Encyclopedia of decision making and decision

support technologies. Information Science Reference, Hershey

Alexandris S, Kerkides P, Liakatas A (2006) Daily reference

evapotranspiration estimates by the “Copais” approach. Agric

Water Manag 82:371–386. https://doi.org/10.1016/j.agwat.2005.

08.001

Allen RG, Pereira LS, Raes D, Smith M (1999) Crop evapotranspi-

ration. Guidelines for computing crop water requirements. FAO

irrigation and drainage paper no. 56. Rome, Italy: United Nations

—FAO, p 300

Ambas VT, Baltas E (2012) Sensitivity analysis of different

evapotranspiration methods using a new sensitivity coefﬁcient.

Glob Nest J 14:335–343

Bandura R (2005) Measuring country performance and state behavior:

a survey of composite indices. Prepared for the book project the

new public ﬁnance: responding to global challenges. Ofﬁce of

Development Studies United Nations Development Programme,

New York

Bandura R (2008) A survey of composite indices measuring country

performance: 2008 update. Ofﬁce of Development Studies

United Nations Development Programme, New York

Barraque

´B, Karavitis CA, Katsiardi P (2008) The range of existing

circumstances in the water strategy man case studies. In:

Koundouri P (ed) Coping with water deﬁciency. Springer,

Amsterdam, pp 45–112

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

Becker W, Saisana M, Paruolo P, Vandecasteele I (2017) Weights and

importance in composite indicators: closing the gap. Ecol Ind

80:12–22. https://doi.org/10.1016/j.ecolind.2017.03.056

Bianco D (2006) Decision making. In: Helms MM (ed) Encyclopedia

of management, 5th edn. Thomson/Gale, Detroit, pp 160–164

Booysen F (2002) An overview and evaluation of composite indices

of development. Soc Indic Res 59:115–151. https://doi.org/10.

1023/A:1016275505152

Breier GP, de Paula I, ten Caten C, et al (2012) A review of value

tools used in sustainability assessment. In: 18th International

conference on industrial engineering and operations manage-

ment, July 9–11, 2012, Guimaraes, Portugal

Briggs J, Peat FD (2000) Seven life lessons of chaos: spiritual wisdom

from the science of change. Harper Perennial, New York

Burgman M (2005) Risks and decisions for conservation and

environmental management. Cambridge University Press,

Cambridge

Chatzithomas CD, Alexandris SG (2015) Solar radiation and relative

humidity based, empirical method, to estimate hourly reference

evapotranspiration. Agric Water Manag 152:188–197. https://

doi.org/10.1016/j.agwat.2015.01.019

Cherchye L, Moesen W, Puyenbroeck TV (2004) Legitimately

diverse, yet comparable: on synthesizing social inclusion

performance in the EU. J Common Mark Stud 42:919–955.

https://doi.org/10.1111/j.0021-9886.2004.00535.x

Cherchye L, Moesen W, Rogge N et al (2007) Creating composite

indicators with DEA and robustness analysis: the case of the

Technology Achievement Index. J Oper Res Soc 59:239–251.

https://doi.org/10.1057/palgrave.jors.2602445

Core Team R (2016) R: A language and environment for statistical

computing. R Foundation for Statistical Computing, Vienna

Corvala

´n C, Briggs D, Zielhuis G (2000) Decision-making in

environmental health: from evidence to action. World Health

Organization, New York

Daellenbach HG (1994) Systems and decision making: a management

science approach, 1st edn. Wiley, Chichester

Dahl AL (2012) Achievements and gaps in indicators for sustain-

ability. Ecol Ind 17:14–19. https://doi.org/10.1016/j.ecolind.

2011.04.032

Desarbo WS, Ramaswamy V, Cohen SH (1995) Market segmentation

with choice-based conjoint analysis. Market Lett 6:137–147.

https://doi.org/10.1007/BF00994929

Dobbie MJ, Dail D (2013) Robustness and sensitivity of weighting

and aggregation in constructing composite indices. Ecol Ind

29:270–277. https://doi.org/10.1016/j.ecolind.2012.12.025

Doorenbos J, Pruitt WO (1977) Crop water requirements. FAO

irrigation and drainage paper 24, FAO, Rome, p 144

Dracup JA (1991) Drought monitoring. Stoch Hydrol Hydraul 5:261–

266. https://doi.org/10.1007/BF01543134

Droogers P, Allen RG (2002) Estimating reference evapotranspiration

under inaccurate data conditions. Irrigat Drain Syst 16:33–45.

https://doi.org/10.1023/A:1015508322413

Eakin H, Bojo

´rquez-Tapia LA (2008) Insights into the composition of

household vulnerability from multicriteria decision analysis.

Glob Environ Change 18:112–127. https://doi.org/10.1016/j.

gloenvcha.2007.09.001

European Environment Agency (1999) Environmental indicators:

typology and overview. Copenhagen

European Environment Agency (2005) EEA core set of indicators:

guide. Publications Ofﬁce, Luxembourg

Fomby T (2008) The unobservable components model. http://faculty.

smu.edu/tfomby/eco5375/data/notes/The%20Unobservable%

20Components%20Model.pdf. Accessed 25 Mar 2018

French S (1995) Uncertainty and imprecision: modelling and analysis.

J Oper Res Soc 46:70–79. https://doi.org/10.1057/jors.1995.8

Gan X, Fernandez IC, Guo J et al (2017) When to use what: methods

for weighting and aggregating sustainability indicators. Ecol Ind

81:491–502. https://doi.org/10.1016/j.ecolind.2017.05.068

Garriga RG, Foguet AP (2010) Improved method to calculate a water

poverty index at local scale. J Environ Eng 136:1287–1298.

https://doi.org/10.1061/(ASCE)EE.1943-7870.0000255

Grigg NS (1996) Water resources management: principles, regula-

tions, and cases, 1st edn. McGraw-Hill Professional, New York

Grigg N (2008) Total water management: practices for a sustainable

future. American Waterworks Association, Denver

Guttman NB (1998) Comparing the palmer drought index and the

standardized precipitation index 1. J Am Water Resour Assoc

34:113–121. https://doi.org/10.1111/j.1752-1688.1998.tb05964.

Guttman NB (1999) Accepting the standardized precipitation index: a

calculation algorithm 1. J Am Water Resour Assoc 35:311–322.

https://doi.org/10.1111/j.1752-1688.1999.tb03592.x

Haines-Young R, Potschin M, Kienast F (2012) Indicators of

ecosystem service potential at European scales: mapping

marginal changes and trade-offs. Ecol Ind 21:39–53. https://

doi.org/10.1016/j.ecolind.2011.09.004

Hargreaves GH, Samani ZA (1985) Reference crop evapotranspira-

tion from temperature. Appl Eng Agric 1:96–99. https://doi.org/

10.13031/2013.26773

Harris P, Charlton M, Fotheringham AS (2010) Moving window

kriging with geographically weighted variograms. Stoch Environ

Res Risk Assess 24:1193–1209. https://doi.org/10.1007/s00477-

010-0391-2

Hayes MJ, Svoboda MD, Wilhite DA, Vanyarkho OV (1999)

Monitoring the 1996 drought using the standardized precipitation

index. Bull Am Meteorol Soc 80:429–438. https://doi.org/10.

1175/1520-0477(1999)080%3c0429:MTDUTS%3e2.0.CO;2

Hellenic Statistical Authority (HSA) (2011) 2011 Population-housing

census. Hellenic Statistical Authority

Hilborn R (1987) Living with uncertainty in resource management.

North Am J Fish Manag 7:1–5. https://doi.org/10.1577/1548-

8659(1987)7%3c1:LWUIRM%3e2.0.CO;2

Huang B, Wu B, Barry M (2010) Geographically and temporally

weighted regression for modeling spatio-temporal variation in

house prices. Int J Geogr Inf Sci 24:383–401. https://doi.org/10.

1080/13658810802672469

Jensen ME, Richard GA (eds) (2016) Evaporation, evapotranspira-

tion, and irrigation water requirements, 2016. Task Committee

on Revision of Manual 70

Kaly U, Pratt C, Mitchell J (2004) The environmental vulnerability

index 2004

Karavitis CA (1992) Drought management strategies for urban water

supplies: the case of metropolitan Athens. Ph.D. Dissertation,

Colorado State University

Karavitis CA (1998) Drought and urban water supplies: the case of

metropolitan Athens. Water Policy 1:505–524. https://doi.org/

10.1016/S1366-7017(99)00009-4

Karavitis CA (1999) Decision support systems for drought manage-

ment strategies in metropolitan Athens. Water Int 24:10–21.

https://doi.org/10.1080/02508069908692129

Karavitis C, Bosdogianni A, Vlachos E (2001) Environmental

management approaches and water resources in the stressed

region of Thriassion, Greece. Glob NEST J 3:131–144

Karavitis CA, Alexandris S, Tsesmelis DE, Athanasopoulos G

(2011a) Application of the standardized precipitation index

(SPI) in Greece. Water 3:787–805. https://doi.org/10.3390/

w3030787

Karavitis CA, Alexandris SG, Fassouli VP et al (2011b) Vulnerability

assessment, task 4.2.5, DMCSEE project. In: 5th DMCSEE

consortium meeting and training, 28 June–1 July 201, Lasko,

Slovenia

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

Karavitis CA, Chortaria C, Alexandris SG et al (2012a) Development

of the standardised precipitation index for Greece. Urban Water J

9:401–417. https://doi.org/10.1080/1573062X.2012.690431

Karavitis CA, Skondras NA, Tsesmelis DE et al (2012b) Drought

impacts archive and drought vulnerability index. In: Gregoric

ˇG

(ed) Drought management centre for south-east Europe—

DMCSEE. Summary of the result of the project, co-ﬁnanced

by the South east Europe transnational Cooperation programme

(contract no. See/a/091/2.2/X). Slovenian Environmental

Agency, pp 33–37

Karavitis CA, Alexandris SG, Fassouli VP et al (2013) Assessing

drought vulnerability under alternative water demand deﬁcits

scenarios in South-Eastern Europe. Porto, Portugal

Karavitis CA, Tsesmelis DE, Skondras NA et al (2014) Linking

drought characteristics to impacts on a spatial and temporal

scale. Water Policy 16:1172–1197. https://doi.org/10.2166/wp.

2014.205

Karavitis CA, Oikonomou PD, Waskom RM, et al (2015a) Applica-

tion of the Standardized Drought Vulnerability Index in the

lower south Platte Basin, Colorado. In: 2015 AWRA annual

water resources conference, 16–19 November 2015, Denver

Karavitis CA, Vasilakou CG, Tsesmelis DE et al (2015b) Short-term

drought forecasting combining stochastic and geo-statistical

approaches. Eur Water 49:43–63

Kasperczyk N, Knickel K (1996) Analytic Hierarchy Process (AHP).

http://www.ivm.vu.nl/en/Images/MCA3_tcm234-161529.pdf.

Accessed 15 June 2018

Kaufmann D, Kraay A, Zoido-Lobato

´n P (1999) Aggregating

governance indicators. World Bank, New York

Kaufmann D, Kraay A, Mastruzzi M (2003) Governance matters III:

governance indicators for 1996–2002. World Bank, NewYork

Kosmas C, Tsara M, Moustakas N et al (2006) Environmentally

sensitive areas and indicators of desertiﬁcation. In: Kepner WG,

Rubio JL, Mouat DA, Pedrazzini F (eds) Desertiﬁcation in the

mediterranean region. A security issue. Springer, New York,

pp 525–547

Kosmas C, Kairis O, Karavitis C, Ritsema C, Salvati L, Acikalin S,

Alcala M, Alfama P, Atlhopheng J, Barrera J, Belgacem A, Sole

Benet A, Brito J, Chaker M, Chanda R, Coelho C, Darkoh M,

Diamantis I, Ermolaeva O, Fassouli V, Fei W, Feng J, Fernandez

F, Ferreira A, Gokceoglu C, Gonzalez D, Gungor H, Hessel R,

Juying J, Khatteli H, Khitrov N, Kounalaki A, Laouina A,

Lollino P, Lopes M, Magole L, Medina L, Mendoza M, Morais

P, Mulale K, Ocakoglu F, Ouessar M, Ovalle C, Perez C, Perkins

J, Pliakas F, Polemio M, Pozo A, Prat C, Qinke Y, Ramos A,

Ramos J, Riquelme J, Romanenkov V, Rui L, Santaloia F,

Sebego R, Sghaier M, Silva N, Sizemskaya M, Soares J, Sonmez

H, Taamallah H, Tezcan L, Torri D, Ungaro F, Valente S, de

Vente J, Zagal E, Zeiliguer A, Zhonging W, Ziogas A (2014)

Evaluation and selection of indicators for land degradation and

desertiﬁcation monitoring: methodological approach. Environ

Manag 54:951–970. https://doi.org/10.1007/s00267-013-0109-6

Landerretche O, Leiva B, Vivanco D, Lo

´pez I (2017) Welcoming

uncertainty: a probabilistic approach to measure sustainability.

Ecol Ind 72:586–596. https://doi.org/10.1016/j.ecolind.2016.08.

050

Li Q, Li P, Li H, Yu M (2015) Drought assessment using a

multivariate drought index in the Luanhe River basin of

Northern China. Stoch Environ Res Risk Assess 29:1509–

1520. https://doi.org/10.1007/s00477-014-0982-4

Loucks DP, Beek EV, Stedinger JR et al (2005) Water resources

systems planning and management: an introduction to methods,

models and applications. UNESCO, Paris

Macharis C, Springael J, De Brucker K, Verbeke A (2004)

PROMETHEE and AHP: the design of operational synergies

in multicriteria analysis: strengthening PROMETHEE with ideas

of AHP. Eur J Oper Res 153:307–317. https://doi.org/10.1016/

S0377-2217(03)00153-X

Margerum RD, Born SM (1995) Integrated environmental manage-

ment: moving from theory to practice. J Environ Plan Manag

38:371–392. https://doi.org/10.1080/09640569512922

Millet I, Wedley WC (2002) Modelling risk and uncertainty with the

analytic hierarchy process. J Multi-Crit Decis Anal 11:97–107.

https://doi.org/10.1002/mcda.319

Morc¸ӧl G (2006) Handbook of decision making. CRC Press, Boca

Raton

Munda G (2005) Multiple criteria decision analysis and sustainable

development. In: Multiple criteria decision analysis: state of the

art surveys. International Series in Operations Research &

Management Science, vol 78. Springer, New York, pp 953–

986. https://doi.org/10.1007/0-387-23081-5_23

Munda G (2007) Social multi-criteria evaluation (SMCE). Springer,

New York

Nardo M, Saisana M, Saltelli A et al (2005) Handbook on

constructing composite indicators. OECD Statistics Working

Paper. STD/DOC(2005)3

National Meteorological Service of Greece (HNMS) (2016) Meteo-

rological data of the Sterea Hellas Stations. Hellinikon, Athens

National Research Council (1986) Drought management and its

impact on public water systems: report on a colloquium

sponsored by the water science and technology board. National

Academy Press, Washington

Nikam BR, Kumar P, Garg V et al (2014) Comparative evaluation of

different potential evapotranspiration estimation approaches. Int

J Res Eng Technol 3:544–552

Norton BG (2005) Sustainability. University of Chicago Press,

Chicago

OECD, European Commission (ed) (2008) Handbook on constructing

composite indicators: methodology and user guide. Organisation

for Economic Co-operation and Development Publishing, Paris

Oikonomou PD (2017) Methodologies for transforming data to

information and advancing the understanding of water resources

systems towards integrated water resources management. Ph.D.

Dissertation, Colorado State University, Department of Civil &

Environmental Engineering, Fort Collins, Colorado, USA

Oikonomou PD, Waskom RM (2018) Assessing drought vulnerability

in Northeast Colorado. 2018 Fall Meeting of the American

Geophysical Union, Washington, D.C., 10–14 December

Oikonomou PD, Karavitis CA, Kolokytha E (2018) Multi-index

drought assessment in Europe. Proceedings of 3rd International

Electronic Conference on Water Sciences (ECWS-3). https://doi.

org/10.3390/ECWS-3-05822

Oikonomou PD, Tsesmelis DE, Waskom RM, Grigg NS, Karavitis

CA (2019) Enhancing the standardized drought vulnerability

index by integrating spatiotemporal information from satellite

and In Situ data. J Hydrol 569:265–277. https://doi.org/10.1016/

j.jhydrol.2018.11.058

Orme BK (2009) Which conjoint method should i use? Sawtooth

software research paper series, article originally published in

sawtooth solutions, 1996. Sawtooth Software Inc., Sequim, WA

Palmer WC (1965) Meteorological drought. US Department of

Commerce, Weather Bureau Washington, Washington

Peterson G, De Leo GA, Hellmann JJ et al (1997) Uncertainty,

climate change, and adaptive management. Conserv Ecol 1:4

Peterson AT, Papes¸ M, Eaton M (2007) Transferability and model

evaluation in ecological niche modeling: a comparison of GARP

and Maxent. Ecography 30:550–560. https://doi.org/10.1111/j.

0906-7590.2007.05102.x

Public Power Corporation (PPC S.A.) (2016) Precipitation and

temperature data for the Sterea Hellas Stations, Athens, Greece

Qudrat-Ullah H, Spector JM, Davidsen PI (eds) (2007) Complex

decision making. Springer, Berlin

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

Ramanathan R (2001) A note on the use of the analytic hierarchy

process for environmental impact assessment. J Environ Manag

63:27–35. https://doi.org/10.1006/jema.2001.0455

Rogge N (2012) Undesirable specialization in the construction of

composite policy indicators: the environmental performance

index. Ecol Ind 23:143–154. https://doi.org/10.1016/j.ecolind.

2012.03.020

Saaty TL (1980) The analytic hierarchy process: planning, priority

setting, resource allocation. McGraw-Hill, New York

Saaty TL (1988) What is the analytic hierarchy process? In: Mitra G,

Greenberg HJ, Lootsma FA et al (eds) Mathematical models for

decision support. Springer, Berlin, pp 109–121

Sadiq R, Tesfamariam S (2009) Environmental decision-making

under uncertainty using intuitionistic fuzzy analytic hierarchy

process (IF-AHP). Stoch Environ Res Risk Assess 23:75–91.

https://doi.org/10.1007/s00477-007-0197-z

Sagar AD, Najam A (1998) The human development index: a critical

review. Ecol Econ 25:249–264. https://doi.org/10.1016/S0921-

8009(97)00168-7

Sage A (2007) Decision theory. In: McGraw-Hill encyclopedia of

science and technology, 10 edn. McGraw-Hill Education,

pp 302–308. https://doi.org/10.1036/1097-8542.182500

Saisana M, Saltelli A, Tarantola S (2005) Uncertainty and sensitivity

analysis techniques as tools for the quality assessment of

composite indicators. J R Stat Soc Ser A (Statistics in Society)

168:307–323. https://doi.org/10.1111/j.1467-985X.2005.00350.x

Segnestam L (2002) Indicators of environment and sustainable

development: theories and practical experience. World Bank,

New York

Singh RK, Murty HR, Gupta SK, Dikshit AK (2009) An overview of

sustainability assessment methodologies. Ecol Ind 9:189–212.

https://doi.org/10.1016/j.ecolind.2008.05.011

Skondras NA, Karavitis CA, Gkotsis II et al (2011) Application and

assessment of the Environmental Vulnerability Index in Greece.

Ecol Ind 11:1699–1706. https://doi.org/10.1016/j.ecolind.2011.

04.010

Sodhi B, Prabhakar TV (2012) A simpliﬁed description of fuzzy

TOPSIS. arXiv:12055098[cs]

Special Secretariat for Water (SSW) (2013) Management plans of the

River Basins in Greece. Ministry of Environment, Energy and

Climate Change (MEECG), Athens

Special Secretariat for Water (SSW) (2016) Precipitation and

temperature data for the Sterea Hellas Stations, Hydroskopion.

Ministry of Environment, Energy and Climate Change

(MEECG), Athens

Tsesmelis DE (2010) SPI Application in Greece for Integrated

Drought Management. Master’s Thesis, Department of Natural

Resources Development and Agricultural Engineering, Agricul-

tural University of Athens, Athens, Greece (in Greek)

Tsesmelis DE (2017) Development, implementation and evaluation of

drought and desertiﬁcation risk indicators for the integrated

management of water resources. Ph.D. Dissertation, Department

of Natural Resources Development and Agricultural Engineer-

ing, Agricultural University of Athens, Athens, Greece (in

Greek)

Tsesmelis DE, Karavitis CA, Oikonomou PD, Alexandris S, Kosmas

C (2019) Assessment of the vulnerability to drought and

desertiﬁcation characteristics using the standardized drought

vulnerability index (SDVI) and the environmentally sensitive

areas index (ESAI). Resources 8:6. https://doi.org/10.3390/

resources8010006

Valipour M (2014) Temperature analysis of reference evapotranspi-

ration models. Met Appl 22:385–394. https://doi.org/10.1002/

met.1465

Valipour M, Eslamian S (2014) Analysis of potential evapotranspi-

ration using 11 modiﬁed temperature-based models. Int J Hydrol

Sci Technol 4:192–207. https://doi.org/10.1504/IJHST.2014.

067733

Wang Q, Ni J, Tenhunen J (2005) Application of a geographically-

weighted regression analysis to estimate net primary production

of Chinese forest ecosystems. Glob Ecol Biogeogr 14:379–393.

https://doi.org/10.1111/j.1466-822X.2005.00153.x

Welte DR, Feenstra T, Jager H, Leidl R (2004) A decision chart for

assessing and improving the transferability of economic evalu-

ation results between countries. Pharm Econ 22:857–876. https://

doi.org/10.2165/00019053-200422130-00004

Wilhite DA, Hayes MJ, Svoboda MD (2000) Drought monitoring and

assessment: status and trends in the United States. In: Vogt J.V.,

Somma F. (eds) Drought and drought mitigation in Europe.

Advances in Natural and Technological Hazards Research, vol

14. Springer, Dordrecht, pp 149–160. https://doi.org/10.1007/

978-94-015-9472-1_11

Williams BK (2011) Adaptive management of natural resources—

framework and issues. J Environ Manag 92:1346–1353. https://

doi.org/10.1016/j.jenvman.2010.10.041

Zahir S (1999) Clusters in a group: decision making in the vector

space formulation of the analytic hierarchy process. Eur J Oper

Res 112:620–634. https://doi.org/10.1016/S0377-2217(98)

00021-6

Zhou P, Ang BW, Zhou DQ (2010) Weighting and aggregation in

composite indicator construction: a multiplicative optimization

approach. Soc Indic Res 96:169–181. https://doi.org/10.1007/

s11205-009-9472-3

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in

published maps and institutional afﬁliations.

Stochastic Environmental Research and Risk Assessment

123

Author's personal copy

A preview of this full-text is provided by Springer Nature.
Learn more

Preview content only

Content available from Stochastic Environmental Research and Risk Assessment

This content is subject to copyright. Terms and conditions apply.

Drought Risk Evaluation in Iran by Using Geospatial Technologies

Article

Full-text available

Jun 2022

A drought risk map has been developed at the national scale by using remote-sensing satellite data over Iran by combining output layers resulting from three main components of a risk-evaluation procedure including Hazard Quantification (HQ), Vulnerability Assessment (VA) and Identification of Elements at Risk (IER) in a GIS environment. In this respect, Drought Severity (DS) was calculated by using the monthly Normalized Difference Vegetation Index (NDVI) (over 31 years from 1986–2016). Iran landcover classification and a slope map, population density maps, and irrigated farm percentages at the provincial scale were utilized within the drought risk evaluation (DRE) process. The final risk map reveals that the northwest of the country, with a climate similar to the central European weather conditions, is exposed to the maximum drought risk. In contrast, the areas with an arid climate, mainly located in the middle of Iran, exhibits minimum risk against drought. Based on the risk map, the southern part of the Caspian Sea shows very low drought risk due to the moderate and subtropical climate in this region. The outputs of this research will provide advice and warnings to help decision makers reduce drought risk consequences after prioritizing risk areas at the administrative scale.

Greece on a sustainable future: reviewing constraints and practices regarding forest and water resources management, flora and fauna biodiversity

Article

Jan 2022

The forest policy in Greece and the current regulatory framework is not efficient in supporting the implementation of sustainability at a satisfactory level. The main scope of this study is to review and present constrains and practices across the sectors of forest and water resources management, flora and fauna biodiversity. The hypothesis is that common practices in the forest field combined with inefficient and obsolete legislation are responsible for delays in the implementation of a national forest policy, which will promote sustainability. A systematic reviewing methodology was applied so to ensure a rigorous and repeatable method of sustainability constraints identification and evaluation. The identification of the constraints can promote the improvement of legislation, the revision of common practices concerning the forest sector and finally can help the forest managers to better understand how to work effectively within legal, regulatory and operational environments deriving from forest policy.

Enhancing drought monitoring and prediction in diverse climates by using composite drought indices

Article

Full-text available

Nov 2023
STOCH ENV RES RISK A

In order to improve the reliable monitoring and prediction of drought behavior, it is crucial to comprehensively consider composite indices. Initially, four univariate drought indices were evaluated: the Standardized Precipitation Index, the Standardized Soil Moisture Index of the top two layers (SSI1 and SSI2), and the Standardized Precipitation Evapotranspiration Index. Subsequently, this paper introduces five composite drought indices: the Multivariate Standardized Drought Index, the modified Aggregate Drought Index, the Joint Drought Deficit Index, the Machine Learning (ML-based) Drought Index (SVMs), and the Artificial Intelligence (AI-based) Drought Index (ANF-PSOs) models based on precipitation (P), soil moisture at two layers (SM1 and SM2), and reference evapotranspiration as input variables. These drought indices are formulated using P-ETref inputs, P-SM1 inputs, and P-SM2 inputs. The study covers 30 main basins classified into six climate zones, including coastal wet, mountain, semi mountain, semi desert, desert, and coastal desert across Iran from 1979 to 2021. The performance of the studied models was evaluated using the correlation coefficient and root mean square error in comparison with SPI at the same time scales as the target model. Drought characteristics, including the number of drought events, duration, frequency, and intensity, were determined for each model at monthly, seasonal, and yearly time scales. The results revealed that the recommended P-SM1 inputs for SVMs and ANF-PSOs models significantly outperformed the MSDIs, modified ADIs, and JDIs models. The results indicated that the introduced composite models effectively captured the comprehensive SM situation without being heavily influenced by individual parameters. The behavioral patterns of these indices remained consistent, except for the specific performance of ETref, which caused some inconsistencies. Moreover, a comparison among different climates revealed that ETref played a prominent role in the discrepancies observed in the output of the composite models, resulting in a strong relationship between P and ETref. Consequently, when constructing composite indices, the information conveyed by ETref was more readily disregarded by JDI and ADI but retained in ANF-PSO. This research sheds light on the mechanisms of these ML-based and AI-based composite approaches in integrating different drought features. It offers valuable insights into the performance of composite drought models and provides benchmarks, particularly in dry climates, to enhance drought monitoring methods.

Weakened economic impacts with future intensifying drought in Chinese mainland

Article

Oct 2023
J CLEAN PROD

Drought vulnerability range assessment: A dynamic and impact-driven method for multiple vulnerable systems

Article

Apr 2023

Current approaches to identify vulnerability to drought often lack ground-truthing and transparent methodology; they are also often biased by expert subjectivity. Furthermore, most drought vulnerability assessments are seldom system-specific but of general social relevance. Consequently, resultant vulnerability maps are typically a general snapshot of case-specific decisions, lacking teeth in their statistical soundness and system-specific relevance. Thus, this study presents an impact-based method for a transparent selection of effective vulnerability factors for six vulnerable systems (social, economic, agricultural, ecological, water resources, and energy-industry) The study is based on a backward multivariate linear regression (MLR) model, which links vulnerability factors to historical drought impacts for the case of Iran. The drought vulnerability index (DVI) and the comprehensive drought vulnerability index (CDVI) are calculated by aggregating the vulnerability of the six systems. Four different weighting methods are then applied to combine driving vulnerability factors to assess the vulnerability of each system for three timesteps (2005, 2010, and 2015). The vulnerability values are presented as ranges by combining the results of four weighting methods. Furthermore, identified driving vulnerability factors are classified by their manageability. A range of 154 vulnerability factors were tested as possible drivers, from which 44 were identified as effective factors. Overall, the results especially highlight that the social and economic systems are vulnerable to drought. From a systemic perspective, socioeconomic factors, such as infrastructure development, health conditions, and industrialization, might play a crucial role in reducing vulnerability. Thus, they should be prioritized in vulnerability reduction programs.

Proposing an ensemble machine learning based drought vulnerability index using M5P, dagging, random sub-space and rotation forest models

Article

Full-text available

Mar 2023
STOCH ENV RES RISK A

Drought is one of the major barriers to the socio-economic development of a region. To manage and reduce the impact of drought, drought vulnerability modelling is important. The use of an ensemble machine learning technique i.e. M5P, M5P -Dagging, M5P-Random SubSpace (RSS) and M5P-rotation forest (RTF) to assess the drought vulnerability maps (DVMs) for the state of Odisha in India was proposed for the first time. A total of 248 drought-prone villages (samples) and 53 drought vulnerability indicators (DVIs) under exposure (28), sensitivity (15) and adaptive capacity (10) were used to produce the DVMs. Out of the total samples, 70% were used for training the models and 30% were used for validating the models. Finally, the DVMs were authenticated by the area under curve (AUC) of receiver operating characteristics, precision, mean-absolute-error, root-mean-square-error, K-index and Friedman and Wilcoxon rank test. Nearly 37.9% of the research region exhibited a very high to high vulnerability to drought. All the models had the capability to model the drought vulnerability. As per the Friedman and Wilcoxon rank test, significant differences occurred among the output of the ensemble models. The accuracy of the M5P base classifier improved after ensemble with RSS and RTF meta classifiers but reduced with Dagging. According to the validation statistics, M5P-RFT model achieved the highest accuracy in modelling the drought vulnerability with an AUC of 0.901. The prepared model would help planners and decision-makers to formulate strategies for reducing the damage of drought.

Spatio-Temporal Changes and Influencing Factors of Meteorological Dry-Wet in Northern China during 1960–2019

Article

Full-text available

Jan 2023

In northern China, precipitation fluctuates greatly and drought occurs frequently, which mark some of the important threats to agricultural and animal husbandry production. Understanding the meteorological dry-wet change and the evolution law of drought events in northern China has guiding significance for regional disaster prevention and mitigation. Based on the standardized precipitation index (SPI), this paper explored the spatio-temporal evolution of meteorological dry-wet in northern China. Our results showed that arid area (AA) and semi-arid area (SAA) in the west showed a trend of wetting at inter-annual and seasonal scales, while humid area (HA) and semi-humid area (SHA) in the east showed a different dry-wet changing trend at different seasons under the background of inter-annual drying. AA and HA showed obvious “reverse fluctuation” characteristics in summer. The drought frequency (DF) and drought intensity (DI) were high in the east and low in the west, and there was no significant difference in drought duration (DD) and drought severity (DS) between east and west. The DD, DS and DI of AA and SAA showed a decreasing trend, while the DD and DS of HA and SHA showed a slight increasing trend, and the DS decreased. In summer and autumn, the main influencing factors of drying in the east and wetting in the west were PNA, WP, PDO and TP1, and the fluctuations of NAO-SOI, NAO-AMO and PNA-NINO3.4 jointly determined the characteristics of SPI3 reverse fluctuations of HA and AA in summer.

High Resolution Future Projections of Drought Characteristics in Greece Based on SPI and SPEI Indices

Article

Full-text available

Sep 2022

Future changes in drought characteristics in Greece were investigated using dynamically downscaled high-resolution simulations of 5 km. The Weather Research and Forecasting model simulations were driven by EC-EARTH output for historical and future periods, under Representative Concentration Pathways 4.5 and 8.5. For the drought analysis, the standardized precipitation index (SPI) and the standardized precipitation-evapotranspiration index (SPEI) were calculated. This work contributed to achieve an improved characterization of the expected high-resolution changes of drought in Greece. Overall, the results indicate that Greece will face severe drought conditions in the upcoming years, particularly under RCP8.5, up to 8/5 y of severity change signal. The results of 6-month timescale indices suggest that more severe and prolonged drought events are expected with an increase of 4 months/5 y, particularly in areas of central and eastern part of the country in near future, and areas of the western parts in far future. The indices obtained in a 12-month timescale for the period 2075–2099 and under RCP8.5 have shown an increase in the mean duration of drought events along the entire country. Drought conditions will be more severe in lowland areas of agricultural interest (e.g., Thessaly and Crete).

Construction and application of comprehensive drought monitoring model considering the influence of terrain factors: a case study of southwest Yunnan, China

Article

Full-text available

May 2022
ENVIRON SCI POLLUT R

Droughts in winter and spring are one of the most prominent natural disasters in the Yunnan Province in China. They occur frequently, with long durations and have a wide range of damage, which has a serious impact on social and economic development, as well as agricultural production and, therefore, strongly impacts the lives of the people living in the region. The traditional drought monitoring model does not take terrain into consideration, thereby affecting the comparative nature of results, as baseline conditions are not the same. Therefore, this study proposed a comprehensive drought monitoring model considering the influence of terrain factors to improve the evaluation effect. Firstly, based on NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS) and Tropical Rainfall Measurement Mission (TRMM 3B43) data, vegetation condition index (VCI), temperature condition index (TCI), precipitation condition index (TRCI), and three terrain factors ground elevation (DEM), slope (SLOPE), aspect (ASPECT) were selected as model parameters. Then, a comprehensive drought monitoring model without considering terrain factors (Model A) and a comprehensive drought monitoring model of considering terrain factors (Model B) were constructed by using multiple linear regression models. Finally, the effects of the two models were evaluated by using standardized precipitation evapotranspiration index (SPEI) in southwest Yunnan Province, China, and model B was used to analyze the drought in winter and spring in the study area from 2008 to 2019. The results showed that (1) the correlation coefficient of model B was higher than that of model A in winter and spring and the standard error of model B was lower than that of model A. (2) The grade consistency rate of Model A and SPEI was 0.92 in winter and 0.33 in spring; the grade consistency between model B and SPEI was 0.83 in winter and 0.75 in spring, and therefore the monitoring effect of model B was more stable. (3) There were periodic droughts during the study period, and the degree of drought in spring was less than in winter. Medium and severe droughts were observed in winter. Thus, this study concluded that the effect of terrain has an important influence on the evaluation of droughts. The comprehensive drought monitoring model which considers topographic factors can effectively identify the occurrence of drought, and therefore provide significant input with regards to disaster prevention and mitigation policies in southwest Yunnan.

Evaluating the Degradation of Natural Resources in the Mediterranean Environment Using the Water and Land Resources Degradation Index, the Case of Crete Island

Article

Full-text available

Jan 2022

Natural resources degradation poses multiple challenges, particularly to environmental and economic processes. It is usually difficult to identify the degree of degradation and the critical vulnerability values in the affected systems. Thus, among other tools, indices (composite indicators) may also describe these complex systems or phenomena. In this approach, the Water and Land Resources Degradation Index was applied to the fifth largest Mediterranean island, Crete, for the 1999–2014 period. The Water and Land Resources Degradation Index uses 11 water and soil resources related indicators: Aridity Index, Water Demand, Drought Impacts, Drought Resistance Water Resources Infrastructure, Land Use Intensity, Soil Parent Material, Plant Cover, Rainfall, Slope, and Soil Texture. The aim is to identify the sensitive areas to degradation due to anthropogenic interventions and natural processes, as well as their vulnerability status. The results for Crete Island indicate that prolonged water resources shortages due to low average precipitation values or high water demand (especially in the agricultural sector), may significantly affect Water and Land degradation processes. Hence, Water and Land Resources Degradation Index could serve as an extra tool to assist policymakers to improve their decisions to combat Natural Resources degradation.

Assessment of the Vulnerability to Drought and Desertification Characteristics Using the Standardized Drought Vulnerability Index (SDVI) and the Environmentally Sensitive Areas Index (ESAI)

Article

Full-text available

Mar 2019

The degradation of natural resources at an intense rate creates serious problems in the environmental systems particularly with the compounding effects of climatic vagaries and changes. On the one hand, desertification is a crucial universal, mostly an anthropogenic environmental issue affecting soils all over the world. On the other hand, drought is a natural phenomenon in direct association with reduced rainfall in various spatial and temporal frames. Vulnerabilities to drought and desertification are complex processes caused by environmental, ecological, social, economic and anthropogenic factors. Particularly for the Mediterranean semi-arid conditions, where the physical and structural systems are more vulnerable, the abuse and overuse of the natural resources lead to their degradation and ultimately, if the current trends continue, to their marginalization. The scope of the current effort is trying to find any common drivers for the pressures of both processes. Thus, the vulnerabilities to drought and desertification are comparing by using the Standardized Drought Vulnerability Index (SDVI) and the Environmentally Sensitive Areas Index (ESAI). The indices are calculated from October 1983 to September 1996 in Greece. Greece is prone to desertification and it is often experiencing intense droughts, thus it presents an almost ideal case study area. The results may indicate that the most important factor for such procedures is the deficits in water resources, either due to lower than usually expected rainfall or to higher societal water demand.

Assessing Drought Vulnerability in Northeast Colorado

Poster

Full-text available

Dec 2018

Multi-Index Drought Assessment in Europe

Conference Paper

Full-text available

Nov 2018

Any attempt for the application of integrated drought management, requires identifying and characterizing the event per se. The questions of scale, boundary, and of geographic areal extend are of central concern for any efforts of drought assessment, impacts identification, and thus of drought mitigation implementation mechanisms. The use of drought indices, such as Standardized Precipitation Index (SPI) and the Standardized Precipitation Evapotranspiration Index (SPEI), has often lead to pragmatic realization of drought duration, magnitude and spatial extend. The current effort presents the implementation of SPI and SPEI on a Pan-European scale and it is evaluated using existing precipitation and temperature data. The E-OBS gridded dataset for precipitation, minimum temperature, and maximum temperature used covered the period 1969 – 2018. The two indices estimated for time steps of 6, and 12 months. The results for the application period of recurrent droughts indicate the potential that both indices offer for an improvement on drought critical areas identification, threshold definitions and comparability, towards contingency planning leading to better mitigation efforts.

Development, implementation and evaluation of drought and desertification risk indicators for the Integrated Management of Water Resources

Thesis

Full-text available

Feb 2017

Demetrios E Tsesmelis

Η παρούσα διδακτορική εργασία έχει ως τομέα μελέτης την Ολοκληρωμένη Διαχείριση των Υδατικών Πόρων (ΟΔΥΠ), βάσει των αρχών της οποίας δημιουργήθηκε ένα εργαλείο (δείκτης) για την αναγνώριση των ξηρασιών και των τρωτών περιοχών, καθώς και τη σύνδεση της τρωτότητας στηv ξηρασία και την ερημοποίηση. Η Ελλάδα παρουσιάζει συχνά φαινόμενα ξηρασίας, τυχαία και περιοδικά. Αυτό οφείλεται στη χωροχρονική κατανομή των κατακρημνισμάτων, δηλαδή στη διαδοχή των υγρών (Οκτώβριος-Μάρτιος) και των ξηρών περιόδων (Απρίλιος-Σεπτέμβριος). Αν δεν σημειωθούν βροχοπτώσεις κατά τη διάρκεια των υγρών περιόδων τότε ενδέχεται να δημιουργηθούν προβλήματα στη διαθεσιμότητα των υδατικών πόρων κατά τη διάρκεια των ξηρών περιόδων. Για την αντιμετώπιση αυτού του προβλήματος απαιτείται τακτική παρατήρηση και επεξεργασία των μετεωρολογικών δεδομένων, αλλά και ταυτόχρονη εφαρμογή του προληπτικού σχεδιασμού. Ένας τρόπος πρόληψης είναι ο υπολογισμός απλών και σύνθετων δεικτών ξηρασίας μιας περιοχής ή μιας χώρας για την έγκαιρη αναγνώριση του σχετικού προβλήματος. Σε αυτό το πλαίσιο, παρουσιάζεται η ανάπτυξη και η δημιουργία ενός σύνθετου δείκτη που αποτελείται από έξι υπο-δείκτες και συγκεκριμένα από τους υπο-δείκτες κατηγοριοποιημένα SPI6 (cSPI6 και cSPI12), Ζήτηση (Demand), Εφοδιασμός (Supply), Επιπτώσεις (Impacts) και Υποδομές στους Υδατικούς πόρους (Infrastructure), έτσι οι υπο-δείκτες του Standardized Drought Vulnerability Index (SDVI) σχετίζονται με όλες τις πτυχές μιας ξηρασίας (μετεωρολογική, αγροτική, υδρολογική και κοινωνικοοικονομική). Συνυπολογίζοντας όλα τα παραπάνω, εκτιμάται η τρωτότητα σε ένα σημείο ή μια περιοχή. Έπειτα, εξετάστηκε η σχέση των υπο-δεικτών και αναπτύχθηκαν συντελεστές βαρύτητας με στατιστικές και εμπειρικές μεθόδους. Ο σύνθετος δείκτης εφαρμόστηκες σε διάφορες συνθήκες (σενάρια, πραγματικές και υποθετικές τιμές) και διαφορετικές περιοχές (Ελλάδα και Νοτιοανατολική Ευρώπη με τη χρήση δεδομένων από 324 μετεωρολογικούς σταθμούς και περίοδο τουλάχιστον 30 ετών). Η αρχική υπόθεση του SDVI με ίδια βάρη και η μέθοδος των κύριων συνιστωσών δείχνουν ότι προσεγγίζουν καλύτερα την τρωτότητα. Στην συνέχεια, εξετάζεται εάν υπάρχει σύνδεση των δύο διαδικασιών τρωτότητας στην ξηρασία και τρωτότητας στην ερημοποίηση. Για τη συσχέτιση αυτή χρειάστηκε να γίνει όμοια κατηγοριοποίηση και των σύνθετων δεικτών σε 3 κλάσεις (χαμηλή, μέτρια και υψηλή). Με βάση αυτή την παραδοχή, εμφανίστηκε μία σχέση μεταξύ τους, της οποίας η εφαρμογή θα ήταν δόκιμο να εξεταστεί και σε άλλες περιοχές. Τέλος, αναλύθηκαν οι δείκτες και των δύο διαδικασιών και δημιουργήθηκε ένας σύνθετος δείκτης που εμφανίζει συνδυαστικά την υδατική και εδαφική υποβάθμιση μιας περιοχής έχοντας, επίσης, τη δυνατότητα να τις εμφανίσει και ξεχωριστά. Οι δείκτες που περιλαμβάνονται στην τελική εξίσωση είναι αυτοί της Ξηρότητας (Aridity), της Ζήτησης (Drought Demand), των Επιπτώσεων από την ξηρασία (Drought Impacts), της Αντοχής στην ξηρασία (Drought Resistance), των Υποδομών στους υδατικούς πόρους (Infrastructure), της Έντασης Χρήσης Γης (Land use intensity), του Μητρικού Υλικού (Parent Material), της Φυτοκάλυψης (Plant Cover), των Βροχοπτώσεων (Rainfall), της Κλίσης (Slope) και της Σύστασης του Εδάφους (Soil Texture). ENG The present work refers to the implementation of the integrated water resources management methodology for the development of a tool in the form of an index for the recognition of droughts and pertinent vulnerable areas. Such an attempt is further expanded in an effort to produce the connection of vulnerability to drought with that to desertification. Greece is sensitive to drought phenomena (random and periodical). This is a consequence of predominant spatial and temporal distribution of precipitation in the area, expressed with the existing climatic succession of wet (October-March) and dry periods (April-September). Specifically, lower precipitation values during the wet periods could cause serious issues in the water resources availability. Regular observation and processing of meteorological data as well as concurrent application of contingency planning are necessary in order to deal with such issues. Early recognition and possible prevention could be achieved through the estimation of simple and composite drought indices of an area. In this context, the development of a composite index was presented, that included six indicators, namely categorized SPI (cSPI6 and cSPI12), Demand, Supply, Impacts and Infrastructures. The Standardized Drought Vulnerability Index is attempting to relate with all the aspects of drought (meteorological, agricultural, hydrological and socio-economic) and provide an estimation procedure. Thereafter, the relation between the indicators was examined and weights with statistical and empirical methods were developed. The composite index was implemented in various conditions (existing, real and artificial generated data for scenarios testing) and in different areas (Greece and South Eastern Europe with data from 324 meteorological stations for a period of more than 30 years). The initial hypothesis of equal weighting and the method of principal components expressed the vulnerability estimation more effectively. Furthermore, it was analyzed whether there was a correlation between drought and desertification vulnerability. To associate these attempts, a similar classification of these composite indices into three classes (low, medium and high) was necessary. Based on this assumption a relationship between drought and desertification vulnerability was surfaced. Finally, the indicators of both procedures were analyzed and a composite index was created, which shows the water resources and soil degradation of a region. The indicators that have occurred in the final equation are Αridity, Water Demand, Drought Impacts, Drought Resistance, Water Resources Infrastructure, Land Use Intensity, Parent Material, Rainfall, Slope and Soil Texture. All in all, it would be appropriate for further research to implement this methodology in other regions worldwide.

Decision-Making in Environmental Health

Book

Apr 2000

Enhancing the Standardized Drought Vulnerability Index by Integrating Spatiotemporal Information from Satellite and In Situ Data

Article

Feb 2019
J HYDROL

Drought is a complex natural hazard with its adverse multifaceted impacts cascading in every physical and human system. The vulnerability magnitude of various areas to drought mostly depends on their exposure to water deficiency, the existing water management policy framework and its implementation. The Standardized Drought Vulnerability Index (SDVI) is an integrated attempt towards characterizing drought vulnerability based on a comparative classification system, incorporating precipitation patterns, the supply and demand trends, and the socioeconomic background as the most crucial contributors to drought vulnerability. This work attempts to evolve the SDVI by presenting a more rigorous method of index parameters estimation and argues that the combination of in-situ and satellite data improve the index results in an effort to further minimize the paucity of drought related information. At the same time, it helps to surpass previous limitations in temporal and spatial propagation of the vulnerability concept. The new framework is applied in the South Platte Basin, within Colorado, on the 2012 summer drought (July-September). The proposed index modification may convey drought information in a more holistic manner to decision makers. SDVI could aid in advancing the understanding of each component contribution through in situ and remote sensing data integration and in avoiding existing practices of broken linkages and fragmentation of the reported impacts. Thus, it is believed that the SDVI could serve as an additional tool to guide decisions and target mitigation and adaptation actions, allowing for a more integrated management approach.

Application and assessment of the standardized drought vulnerability index in the lower South Platte basin, Colorado, USA

Presentation

Nov 2015

R: A Language and Environment for Statistical Computing

Book

Jan 2015

Core R Team

When to use what: Methods for weighting and aggregating sustainability indicators

Article

Oct 2017

Sustainability indices (SIs) have become increasingly important to sustainability research and practice. However, while the validity of SIs is heavily dependent on how their components are weighted and aggregated, the typology and applicability of the existing weighting and aggregation methods remain poorly understood. To close the knowledge gap regarding when to use which weighting and aggregation methods for constructing SIs, we review the most commonly used methods for weighting and aggregating SIs, discuss their benefits and drawbacks, and suggest a process-oriented approach for choosing appropriate weighting and aggregation methods depending on research objectives. Our review synthesis was based on peer-reviewed journal articles, books, and reports by international organizations, governmental agencies, and research institutions. After carefully examining their principles, characteristics, and applications, we selected and classified the frequently used methods for indicator weighting and aggregation. We systematically discuss the benefits and drawbacks of nine weighting methods and three aggregation methods. We propose a four-step process for choosing the most suitable weighting and aggregation methods based on: research purposes, spatial and temporal scales, and sustainability perspectives. In this research, we chose the most commonly used methods for weighting and aggregating SIs and analyzed the characteristics, strengths and weaknesses of each method. We found that choosing appropriate weighting and aggregation methods for a specific sustainability assessment project is an extremely important and challenging task. To meet this challenge, we propose a process-oriented approach for properly selecting methods according to the purpose, scale and sustainability concept. This approach can facilitate the proper selection of these methods in sustainability research and practice.

Evaporation, evapotranspiration, and irrigation water requirements

Book

Jan 2016

Sponsored by the Committee on Evapotranspiration in Irrigation and Hydrology of the Irrigation and Drainage Council of the Environmental and Water Resources Institute of ASCE Evaporation, Evapotranspiration, and Irrigation Water Requirements is a comprehensive reference to estimating the water quantities needed for irrigation of crops based upon the physics of evaporation and evapotranspiration (ET). This new edition of MOP 70, which updates and expands the 1990 original, provides improved and standardized methods to estimate evaporation and ET and to apply and evaluate calculations. The methods, which have both a physical and practical basis, improve clarity, accuracy, and consistency for systems design and management. They are also useful in legal applications involving water agreements, water disputes, and water rights. As competition for water increases and water resources are depleted, this book provides the critical tools to accurately quantify amounts of water consumed by irrigated agriculture. The first part covers basic concepts and physical principles, such as evaporation and ET processes, soil-water-plant systems, energy balance, surface energy-air mass interactions, and evaporation from water surfaces. The second part on ET from land surfaces explains components of measurement and details on estimation; reference crop ET; evaporation from soil; the crop coefficient method; the Penman-Monteith and energy-balance equations; and regional estimates. A third part discusses estimation of irrigation water requirements (IWR) and streamflow depletions. Thirteen appendixes provide tables of mean and basal crop coefficients and background information on concepts and equations used throughout the manual. MOP 70 is intended for use by consulting engineers and scientists working on water issues and by instructors of agricultural and civil engineering, hydrology, and environmental and agricultural sciences.

Assessing structural uncertainty caused by different weighting methods on the Standardized Drought Vulnerability Index (SDVI)

Abstract and Figures

Recommended publications

Spatial and temporal analysis of drought in Greece using the Standardized Precipitation Index (SPI)

A comparison of the degree of implementation of marine biodiversity indicators by European countries...

Spatiotemporal Characteristics of Meteorological Drought for the Island of Crete

An Objective Assessment of Hyperspectral Indicators for the Detection of Buried Archaeological Relic...