Professional Documents
Culture Documents
Multi Criteria Decision Analysis Case Studies in Engineering and The Environment 2Nd Edition Igor Linkov Online Ebook Texxtbook Full Chapter PDF
Multi Criteria Decision Analysis Case Studies in Engineering and The Environment 2Nd Edition Igor Linkov Online Ebook Texxtbook Full Chapter PDF
Multi Criteria Decision Analysis Case Studies in Engineering and The Environment 2Nd Edition Igor Linkov Online Ebook Texxtbook Full Chapter PDF
https://ebookmeta.com/product/multi-criteria-decision-making-
methods-with-bipolar-fuzzy-sets-muhammad-akram-shumaiza-jose-
carlos-rodriguez-alcantud/
https://ebookmeta.com/product/many-criteria-optimization-and-
decision-analysis-state-of-the-art-present-challenges-and-future-
perspectives-dimo-brockhoff-editor/
https://ebookmeta.com/product/modelling-turbulence-in-
engineering-and-the-environment-2nd-edition-kemal-hanjalic/
https://ebookmeta.com/product/machine-learning-for-decision-
sciences-with-case-studies-in-python-1st-edition-s-sumathi/
Data Science and Multiple Criteria Decision Making
Approaches in Finance: Applications and Methods Gökhan
Silahtaro■lu
https://ebookmeta.com/product/data-science-and-multiple-criteria-
decision-making-approaches-in-finance-applications-and-methods-
gokhan-silahtaroglu/
https://ebookmeta.com/product/formal-analysis-by-abstract-
interpretation-case-studies-in-modern-protocols-benjamin-aziz/
https://ebookmeta.com/product/case-studies-in-infectious-
disease-2nd-edition-peter-lydyard/
https://ebookmeta.com/product/philosophy-on-fieldwork-case-
studies-in-anthropological-analysis-1st-edition-nils-bubandt-
editor/
https://ebookmeta.com/product/case-studies-in-advanced-
engineering-design-proceedings-of-the-1st-international-
symposium-1st-edition-c-spitas/
Multi-Criteria Decision
Analysis
Environmental Assessment and
Management
Series Editor:
Dr. Glenn W. Suter II
U.S. Environmental Protection Agency
Cincinnati, OH, USA
PUBLISHED TITLES
Multi-Criteria Decision Analysis
Case Studies in Engineering and the Environment, Second Edition
Igor Linkov, Emily Moberg, Benjamin D. Trump, Boris Yatsalo, and
Jeffrey M. Keisler
Ecological Causal Assessment
Susan B. Norton, Susan M. Cormier, and Glenn W. Suter II
Second Edition
Reasonable efforts have been made to publish reliable data and information, but the author and publisher
cannot assume responsibility for the validity of all materials or the consequences of their use. The authors
and publishers have attempted to trace the copyright holders of all material reproduced in this publica-
tion and apologize to copyright holders if permission to publish in this form has not been obtained. If any
copyright material has not been acknowledged please write and let us know so we may rectify in any future
reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, trans-
mitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter
invented, including photocopying, microfilming, and recording, or in any information storage or retrieval
system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, access www.copyright.com or
contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-
8400. For works that are not available on CCC please contact mpkbookspermissions@tandf.co.uk
Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used
only for identification and explanation without intent to infringe.
Typeset in Palatino
by codeMantra
Emily Moberg would like to dedicate this edition of the book to her
vii
viii Contents
Dredging .........................................................................................................34
Impacts of Dredging...................................................................................... 36
Environmental and Ecological Effects ........................................................ 37
Socioeconomic Impacts ................................................................................. 38
Dangers of Contaminated Sediments ......................................................... 38
Alternatives Selection.................................................................................... 41
Criteria Selection ............................................................................................ 41
Criteria Identification ....................................................................................44
Formulation of Alternatives ......................................................................... 45
References ....................................................................................................... 46
Background................................................................................................... 117
Importance ............................................................................................... 117
Uncertainty and Risk .................................................................................. 118
Moving Forward .......................................................................................... 120
References ..................................................................................................... 121
xvii
xviii Foreword
Lynn Scarlett,
Chief External Affairs Officer, The Nature Conservancy, and
former Deputy Secretary, US Department of the Interior
Preface
As humanity continues into the 21st century, we are faced with an increas-
ingly complex set of environmental stressors and a burgeoning population
affected by these factors. Fortunately, we are in the process of developing more
technical and innovative ways of combating these problems. Environmental
issues have become complex to the point of engendering specialties within
the field, with many projects requiring experts from different fields within
the general category of environmental science. Beyond this, many envi-
ronmental projects and problems span much broader disciplines, bringing
together experts in finances, urban space allocation, and other interdisciplin-
ary efforts. With myriad decision alternatives and criteria which span many
disciplines, decision-making is a complex process. Structured approaches
for decision-making have been proposed, rigorously examined, and utilized,
yet despite the inherent applicability of such methodology, such approaches
are rarely seen within environmental management. Multi-Criteria Decision
Analysis (MCDA) is one such method that can not only aggregate the varie-
gated views of conflicting stakeholders but also weigh criteria with different
units, scales, and meanings against each other effectively.
The impetus to explore this topic with an educational book of case studies
came about as we saw opportunities to implement MCDA methodologies
in order to solve problems within the environmental realm passed up time
and again. Upon receiving feedback that a lack of knowledge and experience
in navigating the many options within MCDA could be a major roadblock,
we sought to aggregate much of the literature and work that had been done
in environmental remediation and planning with MCDA into a common
place. Since the goal was inherently to teach an applied science, we chose
a case study-based approach which focuses on common or general types
of environmental issues so that students are exposed to problems they will
likely face and current practitioners will be exposed to problems they can
find immediately applicable.
As such, we have structured this book in four parts. The first is a general
introduction to MCDA methods and application and subsequently delves fur-
ther into its specific applicability to environmental science. By the end of this
section, the reader is given an overview of the types of MCDA available and
has a conceptual framework of how it is applied. The second part of the book
takes a more in-depth look at one well-analyzed case—that of sediment man-
agement. Chapters in this section explore different steps of MCDA processes
(such as Problem Formulation, MCDA Model Development, Criteria weight-
ing, and Alternative Scoring). This problem is then explored using various
MCDA methods, including Analytical Hierarchy Process (AHP), outrank-
ing, and multi-attribute utility theory. The reader can thus see clearly how
xxi
xxii Preface
will prepare themselves for a future which will undoubtedly demand such
rigorous decision-making processes, while professionals and academics can
take the lead in transforming environmental management by using these
rigorous methodologies to confront the complex, multidisciplinary problems
that characterize their fields. We hope that this book proves useful both in
the classroom and in the field, as we are sure it will!
xxv
Authors
Dr. Igor Linkov is the Risk and Decision Science Focus Area Lead with the
US Army Engineer Research and Development Center and Adjunct Professor
with Carnegie Mellon University. Dr. Linkov has managed multiple risk and
resilience assessments and management projects utilizing MCDA in many
application domains, including critical infrastructure, environment, trans-
portation, energy, homeland security and defense, advanced materials and
biotechnology, supply chain, and cybersecurity. He has published widely on
environmental policy, environmental modeling, and risk analysis, includ-
ing 25 books and more than 400 peer-reviewed papers and book chapters
in top journals. He is Elected Fellow with the American Association for the
Advancement of Science (AAAS) and Society for Risk Analysis.
xxvii
xxviii Authors
Multi-Criteria Decision
Analysis: Methods
and Applications
1
Introduction to Multi-Criteria Methods
Background
Think about the most recent time you encountered an “environmental issue”
in the news. In the first edition of this book, we highlighted natural gas drill-
ing controversies, the aftermath of the Tohoku tsunami in Japan, and the
battle over the US Environmental Protection Agency’s potential regulation of
carbon emissions as examples of the issues you might have seen. In the inter-
vening years, the environmental issues have not stopped; climate change
(in)action, pesticide regulation, air quality issues from burning crops, algal
blooms, and taxes on plastic bag or straw usage have all been major news
stories across the world. These stories are often longer than a single article;
coverage of events can last months. The articles will quote politicians, scien-
tists, health care professionals, business leaders, and affected citizens.
The multiple voices and viewpoints showcased in these articles highlight
one of the hallmarks of environmental decisions—they often involve com-
plex science, many stakeholders, and potential solutions which need to be
judged against many different criteria in order to be fully compared.
Unfortunately, our basic human tendency is to simplify such complicated
decisions and situations until they are more manageable, which results in the
loss of information about the problem, loss of information about other view-
points, and loss of information about uncertainty; in essence, we struggle to
incorporate all the available information to make a fully informed choice.
The result can be controversies and issues that rage on for months and years
without satisfactory resolution.
Unaided, we are quite bad at making complex decisions (McDaniels et al.
1999), and environmental decisions often fall into this complex category. In
environmental problems, the expected inputs include modeling or moni-
toring data, risk analysis, cost or cost–benefit analysis, and stakeholders’
preferences; integrating this information is a major challenge (Linkov and
Ramadan 2004). This integration process can be opaque and seen as unfair,
especially when stakeholder preferences are not dealt with in a manner that
is perceived as objective.
3
4 Multi-Criteria Decision Analysis
MCDA Methods
One type of method that has been used increasingly in many fields, includ-
ing environmental decision-making, is Multi-Criteria Decision Analysis
(MCDA). The method allows for preferences and performance about dif-
ferent management alternatives to be assessed in a clear, formal way that
is both mathematically rigorous and transparent to stakeholders. The basic
outline for an MCDA which will be followed in this book is:
MAUT
Multi-Attribute Utility Theory or Multi-Attribute Value Theory (MAUT
and MAVT respectively; they are often used interchangeably) resolves the
disparate units (cost, environmental impact, etc.) of our criteria into a util-
ity or value so comparison among alternatives’ scores can occur. We also
elicit information about how important each criterion is relative to the
others. With these two pieces of information, we can now combine the
preferences (weights of criteria) with the scores (which have been trans-
formed into a single type of unit, the value or utility) to make a coherent
decision.
Methodologically, the critical step is eliciting or defining the value func-
tions. From the decision maker, we elicit a “utility (or value) function” for
each criterion; this describes how much utility she derives from different
values of the criterion. A simple linear value function could have 0 util-
ity at the lowest potential value and 1 at the highest potential value; other
potential shapes could include exponential, logistic, and step functions.
The shape of the function is important. If I am running in hot weather
and need a drink, the utility of water may be very high; once I have a few
glasses, I may derive no additional utility from more water (this describes a
saturating function). Figure 1.1 shows a few possible shapes for these value
functions.
In Chapter 5, we explore in more detail how we can elicit these value
functions.
Once we have a value function for each criterion, we can transform the
performance of the alternatives (which are measured typically in the units
of the criteria) into this value/utility unit and choose the alternative with the
greatest overall value. Because we have already done the hard work of get-
ting our units to match, choosing the best alternative is a matter of choosing
that with the highest overall value much in the way we would add the cost of
individual items into a total cost.
More formally, if Vi ( ai ) is the performance of alternative a on criterion i,
n
the overall value for alternative a is V ( a ) = ∑ wi Vi ( ai ), where wi is the weight
i=1
assigned to a criterion i. The weights thus do as their name suggest and
weigh how much the performance on a criterion counts towards the total
by how much we value it relative to other criteria. The best alternative
(of alternatives a, b, c … etc.) is then whichever total value (e.g. V (b)) is
largest.
Utility functions are very powerful, but they also assume a rational deci-
sion maker who prefers more value to less and who is willing to make the
trade-offs implied via their value functions (Linkov et al. 2004). As we will
explore in later chapters, this is not always the best assumption!
6 Multi-Criteria Decision Analysis
FIGURE 1.1
Potential value function shapes. Top left: a linear function. This implies that an increase in
the value of the criterion of size x always produces the same increase in utility, no matter the
starting value of the criterion. Note that different low/high values can change the steepness
of the line. Top right: saturating utility function. Unlike in the linear case, we get much larger
gains in utility from initial changes and much less as we approach the highest potential values.
Bottom left: an exponential curve. Here, we do not get much utility until the criterion is almost
to its maximum value, at which point it changes rapidly. Bottom right: a step-wise function.
In this case, the criterion values are binned—we only get more utility once we cross into the
next bin.
AHP
The Analytical Hierarchy Process (AHP) is a method developed by Thomas
Saaty in the 1980s. AHP is immediately distinctive because instead of
using direct weights or value functions, it uses pairwise comparisons;
Introduction to Multi-Criteria Methods 7
FIGURE 1.2
Analytic hierarchy process data schematic. The left-hand figure shows how the “weighting”
occurs—the importance of criteria relative to each other is assessed. The right-hand side of the
figure shows how for each criterion, the alternatives are assessed against each other for per-
formance on that criterion. The bolded cells in the bottom left show how Alternative 2 might
perform better than Alternative 1 on one criterion but worse than it on another.
8 Multi-Criteria Decision Analysis
Outranking
Outranking attempts to order the alternatives by finding the ones which out-
perform or dominate. The idea is that within the algorithm, pairwise com-
parison occurs, but this is the basis for how ordering occurs, rather than
how data are entered (as in AHP). The following explanation of how one of
the most prominent outranking method, Preference Ranking Organization
Method for Enrichment Evaluations (PROMETHEE), works has been adapted
from Yatsalo et al. (2007) and Brans and Vincke (1985):
First, we should introduce the ideas of preference and indifference.
Alternative a would be preferable to Alternative b if z(a) > z(b), where z is the
criterion by which they are being compared. Indifference is when z(a) = z(b);
that is, the performance of these alternatives is equal to the decision maker
(Brans and Vincke 1985).
Each criterion, i, has a weight (wi) for that criterion; the sum of the weights
over all criteria is equal to one.
We then calculate a preference for the alternatives a’s and b’s performance
against this criterion. This preference is expressed
f i zi ( a ) − zi ( b ) if zi ( a ) ≥ zi ( b )
Pi ( a, b ) =
0 otherwise
P ( a, b ) = ∑ w P ( a, b )
i i
TABLE 1.1
Preference Function Types
Name Functional Form: f i = Description
Usual 1 if zi ( a ) ≥ zi ( b ) The decision maker strictly prefers a if
its performance exceeds that of b at all
0 otherwise
Quasi 1 if zi ( a ) ≥ zi ( b ) + l The decision maker strictly prefers a if
its performance exceeds that of b by a
0 otherwise margin l; otherwise she is indifferent
Linear 1 if zi ( a ) ≥ zi ( b ) + l The decision maker strictly prefers a if
preference its performance exceeds that of b by a
zi ( a ) − zi ( b ) margin l; otherwise their preference
otherwise
l grows larger as the difference in
performance grows
Level 1 if zi ( a ) ≥ zi ( b ) + l + m The decision maker strictly prefers a if
its performance exceeds that of b by a
0 if zi ( a ) + l ≤ zi ( b ) margin l + m. The decision maker
is indifferent until a exceed b by
1/2 otherwise margin l. In that middle space, there
is weak preference
Linear with The decision maker strictly prefers a if
indifference 1 if zi ( a ) ≥ zi ( b ) + l + m its performance exceeds that of b by a
margin l + m. The decision maker is
0 if zi ( a ) + l ≥ zi ( b ) indifferent until a exceeds b by margin
l. In that middle space, there is a
z i ( a ) − zi ( b ) − l linearly increasing preference
otherwise
m
Gaussian ( zi ( a)− zi (b ))2 The decision maker begins to prefer a
−
1− e 2σ 2 if zi ( a ) ≥ zi ( b ) as soon as its performance exceeds
b’s, but the preference grows
0 otherwise according to a Gaussian (an S curve)
a outranks b if Q + ( a ) ≥ Q + ( b ) and Q − ( a ) ≤ Q − ( b );
a is indifferent to b if Q + ( a ) = Q + ( b ) and Q − ( a ) = Q − ( b );
10 Multi-Criteria Decision Analysis
Together, all these rules allow us to compare the performance of two alterna-
tives first within a criterion (the preference function), then across criteria (the
preference index), then across alternatives (the positive and negative flows,
then the rules to compare them).
Outranking does not optimize, but rather makes comparisons, making it a
flexible, useful model (Linkov et al. 2004). Its usage will be further explored
in Chapter 6.
Uncertainty in Models
These models use discrete values as weights or scores, but it is easy to imag-
ine cases where these values are not strictly known. One way to deal with
this uncertainty is to use Stochastic Multi-Criteria Analyses, which explore
the weight space and find out how the system behaves under different sets of
weights (Lahdelma and Salminen 2001). These methods either use complex
math to evaluate multidimensional integrals or use numerical methods like
Monte Carlo simulations (which essentially try out combinations of different
possible values and see how the overall distribution of outcomes behaves)
(Tervonen and Lahdelma 2007).
This book uses a method for probabilistic multi-criteria acceptability anal-
ysis ProMAA (Yatsalo et al. 2010), which allows specification of both criteria
values and weight coefficients as probability distributions and probabilistic
methods are used for treatment of uncertainties and assessment of probabil-
ity distributions in an integrated scale (e.g. in a utility scale) for the alterna-
tives under consideration. We will further explore the effect of uncertainty
throughout this book. As we will see later, being able to use distributions
and incorporate uncertainty is very useful! Of course, there are many other
books and papers focused on model, parameter, and scenario uncertainties
as they relate to making policy decisions (e.g., Morgan and Henrion 1990;
Linkov and Burmistrov 2004).
This book uses case studies to explore how MCDA methods are used
in practice. These case studies have been implemented in the software
DECERNS (Decision Evaluation in Complex Risk Network Systems), which
can accompany this text. DECERNS offers a simple user interface which
allows users to implement models easily and with the same method of data
entry and analysis. Figure 1.3 shows the architecture for the DECERNS
program.
DECERNS implements many different MCDA approaches, including those
discussed earlier in this chapter. They are listed fully below:
FIGURE 1.3
DECERNS architecture. The gray shading indicates the Graphic User Interface which is how
you access the MCDA models and exploit their functionality. The Application Programming
Interface combines the Geographic Information System (GIS) subsystem, the Decision Support
Tools, and the models. The Decision Support Tools are also highlighted as this is the subsystem
we are using in this book.
12 Multi-Criteria Decision Analysis
• Advanced MCDA
MAUT
ProMAA
Fuzzy-MAVT
Fuzzy-PROMETHEE
Within DECERNS, the effects of using different models (AHP vs. MAUT,
etc.) can be explored with ease by implementing the model with multiple
methods. MAVT, AHP, TOPSIS, and PROMETHEE have an option to conduct
sensitivity analyses to changes in weights (Yatsalo et al. 2010). MAUT and
ProMAA compute functions of random values instead of using Monte Carlo
methods to deal with uncertainty; the “fuzzy methods” are also computed
as functions of fuzzy variables.
The case studies that are illustrated in this book can be loaded into DECERNS
so that further exploration can be conducted or alterations can be made.
For more information about how to operate DECERNS, the reader should
read through the Software Appendix.
Concluding Remarks
MCDA methods are a powerful set of tools that combine information about a
decision maker’s preferences and about the performance of different alterna-
tives to reach a decision and are an important component of environmental
modeling toolbox (Keisler and Linkov 2014). The rest of this book explores
how these factors are measured and used in practice.
References
Belton, V. and T. J. Stewart (2002). Multiple Criteria Decision Analysis: An Integrated
Approach. Dordrecht, Kluwer Academic Publishers.
Brans, J. P. and P. Vincke (1985). “A Preference Ranking Organisation Method.”
Management Science 31(6): 647–656.
Keisler, J. and Linkov, I. (2014). “Models, Decisions and Environment.” Environment,
Systems, Decision 4: 369–372.
Lahdelma, R. and P. Salminen (2001). “SMAA-2: Stochastic Multicriteria Acceptability
Analysis for Group Decision Making.” Operations Research 49(3): 444–454.
Linkov, I. and Burmistrov, D. (2003). “Model Uncertainty and Choices Made by
Modelers: Lessons Learned from the International Atomic Energy Agency
Model Intercomparisons.” Risk Analysis, 23: 1335–1346.
Introduction to Multi-Criteria Methods 13
1 This chapter is based on Huang et al. (2011), on Cegan, et al. (2017), and on Kurth et al. (2017).
15
16 Multi-Criteria Decision Analysis
TABLE 2.1
Summary of Search Terms
MCDA Keywords Environmental Phrases Subject Area
MCDA Contamin* or remediat* Environmental sciences
MCDM or multi-criteria Ecosystem Environmental studies
decision-making
AHP Land Engineering, environmental
Outranking Nano* Social sciences,
mathematical methods
MAUT Site select* Management science
MAVT Sustainab* Operations research and
management sciences
ELECTRE Waste
ANP Water or coastal
Swing weight* Natural resource*
Expected utility Risk and environ*
TOPSIS Aquatic or terrestrial
SMAA or stochastic multi- Energy
criteria acceptability analysis
PROMETHEE Emission or atmosph*
Note: Search terms are indicated here by subject areas.
* Indicates a wildcard, so all words including the letters prior to it were queried.
Watwa-Dorf, Urundi.
Der gewöhnliche Gruss eines Höheren besteht bei allen Wasinja,
von Ukerewe bis Ussui, im Niederknien und Händeklatschen. Der
Gegrüsste erwidert darauf nicht.
Die Wasinja sind zwar weniger unternehmungslustig als die
Wanyamwesi, durch ihre Intelligenz und ihre Geschicklichkeit, die
sich besonders in Schmiedearbeiten äussert, aber doch sicher
berufen eine Rolle zu spielen.
Während die Wasinja schon von verschiedenen Reisenden, von
Speke bis auf die neueste Zeit besucht und besonders von
Stuhlmann vorzüglich beschrieben wurden, gelangen wir westlich
von ihnen zu einem Bantustamm, von dem kaum mehr als der Name
bekannt war: den Wa r u n d i. Mit den ihnen nahe verwandten Waha
und Wanyaruanda bewohnen sie ein weites Gebiet, von Ussui bis
zum Russisi, von Unyamwesi bis zum Tanganyika und reichen
nördlich bis nahe an den Albert Edward-See. Ueberall stehen sie als
»Wahutu« (Unterworfene) dem Adel der Watussi gegenüber.
Die Warundi sind zweifellos sehr alte Ansiedler der von ihnen
bewohnten Gebiete; irgend welche Tradition über Einwanderung
besteht, soviel ich erfahren konnte, nicht. Dennoch sind die Warundi
wahrscheinlich keine Urbevölkerung, sondern eine solche haben wir
in den Wa tw a zu sehen, welche überall im Lande verstreut leben.
Der Name Watwa (oder Batwa) ist bekanntlich ein weit verbreiteter
und wird hauptsächlich den Pygmäenvölkern in den südlichen
Kongowäldern beigelegt. In den schwach bewohnten Urwäldern
konnten die Watwa sich begreiflicherweise reiner erhalten als in dem
offenen Urundi, inmitten einer dichten Ackerbaubevölkerung. Der
Blutmischung waren hier die Wege geebnet und thatsächlich finden
wir, dass die Watwa Urundi's durchschnittlich nicht kleiner sind als
die umwohnenden Warundi.
Sie leben in kleinen Niederlassungen mit sehr schlechten
Grashütten, benutzen im Gegensatz zu den Warundi, die stets auch
Speere führen, ausschliesslich Bogen und Pfeile und lebten
ursprünglich von der Jagd. Mit der Zunahme der Bevölkerung nahm
jedoch das Erträgniss derselben ab, doch wandten sich die Watwa
keineswegs dem Ackerbau, sondern der Töpferei zu. Mit einem
Stück Kalebasse als einzigem Geräth und einem Schnurende zum
Anbringen der Ornamente fertigen sie ungemein geschmackvolle
Töpfe und Krüge an, welche sie an die Ackerbauer verkaufen.
Sie werden sehr verachtet und gelten als Pariastamm. Kein
Mrundi würde aus demselben Gefäss wie ein Mtwa trinken, auch
sollen Heirathen nicht vorkommen. Dennoch ist, wie gesagt, die
Blutmischung unverkennbar und zwar nicht nur bei den Watwa,
sondern auch bei den Warundi. Denn unter den vielfach
hochgewachsenen Warundi trifft man, besonders im Norden,
auffallend häufig Leute von etwa 1,35 m Höhe mit kurzem Hals,
röthlichen Lippen und gedrungener Gestalt, auch erwachsene,
auffallend lichtfarbige Weiber mit dem Kinde auf dem Rücken bei
einer Höhe von 1,20 m.
Offenbar hat man es hier mit Fällen von Atavie zu thun, bei
welchen der Typus einer Watwa-Urbevölkerung zu Tage tritt, welche
in den Warundi aufgegangen ist. Die heutigen Watwa dagegen
stellen nur einen von der herrschenden Rasse durch die
Lebensweise unterschiedenen Pariastamm dar. Der Uebergang
zwischen ihnen und den Kongo-Watwa bilden die Watwa der Berge
westlich vom Tanganyika, die ebenfalls von Jagd und Töpferei leben,
nach den Märkten der Eingeborenen kommen, jedoch bereits als
Zwerge bekannt sind. Alle Watwa sollen eine eigene Sprache
besitzen, doch konnte ich trotz vieler Bemühung nur Kirundi-Wörter
von ihnen erhalten. Sie scheinen sehr stumpfsinnig, »tu wayovu«
(Wir sind Elephantenjäger) ist das einzige, was sie auf alle Fragen
antworten. Auch ihre Geräthschaften, mit Ausnahme jener für
Töpferei, haben nichts charakteristisches und gleichen jenen der
Warundi.
Wenn also auch in gewissen Distrikten eine Aufnahme von
Watwa-Elementen bei den Warundi wahrscheinlich ist, so haben sie
sich im Allgemeinen doch sehr rein und besonders von hamitischen
(Watussi) Mischungen ziemlich frei erhalten.
Die Wa r u n d i sind ein kräftiger, mittelgrosser Stamm;
hochgewachsene und herkulisch gebaute Leute sind nicht selten.
Die Gesichtszüge sind rein negerhaft, die Hautfarbe dunkelbraun,
bei der geringen Reinlichkeit oft fast schwarz erscheinend. Die
Busen junger Weiber sind wohlgeformt und nicht zitzenförmig.