FARMERS’ RESPONSE AND WILLINGNESS TO PARTICIPATE IN

WATER HARVESTING PRACTICES: A CASE STUDY IN DEJEN

DISTRICT / EAST GOJAM ZONE

M.Sc. Thesis

MOLLA TAFERE YEHZBALEM

April 2005

Alemaya University
 FARMERS’ RESPONSE AND WILLINGNESS TO PARTICIPATE IN WATER
HARVESTING PRACTICES: A CASE STUDY IN DEJEN DISTRICT / EAST GOJAM
ZONE

A Thesis Submitted to the Faculty of Agriculture

Department of Agricultural Economics, School of Graduate Studies
ALEMAYA UNIVERSITY

In partial Fulfillment of the Requirements for the degree of

MASTER OF SCIENCE IN AGRICULTURE (AGRICULTURAL ECONOMICS)

By:
Molla Tafere

April 2005

Alemaya University
 SCHOOL OF GRADUATE STUDIES
ALEMAYA UNIVERSITY

As members of the examining Board of the Final M.Sc. Open Defense, we certify that we
have read and evaluated the thesis prepared by MOLLA TAFERE entitled FARMERS’
RESPONSE AND WILLINGNESS TO PARTICIPATE IN WATER HARVESTING
PRACTICES: A CASE STUDY IN DEJEN DISTRICT / EAST GOJAM ZONE and
recommended that it be accepted as fulfilling the thesis requirement for the degree of Master
of Science in Agriculture (agricultural economics).

------------------------------------------------ ---------------------- -----------------------------

Name of chairman Signature Date
------------------------------------------------ ---------------------- -----------------------------
Name of Major Advisor Signature Date
------------------------------------------------ ---------------------- -----------------------------
Name of Co-advisor Signature Date
------------------------------------------------ ---------------------- -----------------------------
Name of Internal Examiner Signature Date
------------------------------------------------ ---------------------- -----------------------------
Name of External Examiner Signature Date

Final approval and acceptance of the thesis is contingent upon the submission of the final copy
of the thesis to the Council of Graduate Studies (CGS) through the Departmental Graduate
Committee (DGC) of the candidate’s major department.

I hereby certify that I have read this thesis prepared under my direction and recommend that it
be accepted as fulfilling the thesis requirement.
-------------------------------------------------- ---------------------- ---------------------------
Name of Major Advisor Signature Date

ii
 DEDICATION

I dedicate this thesis to my father TAFERE HEHZBALEM , and my mother YITAKTU

KELEME as well as to all my brothers and sisters for nursing me with affections and love
and their dedicated partnership in the success of my life.

iii
 STATEMENT OF AUTHOR

I declare that this thesis is my work and all sources of materials used for the thesis have been
genuinely acknowledged. This thesis has been submitted in partial fulfillment of the
requirements for M.Sc degree of Alemaya University and is deposited at the University
Library to be made available to borrowers under rules of the Library. I solemnly declare that
this thesis is not submitted to any other institution anywhere for the award of any academic
degree, diploma, or certificate.

Brief quotations from the thesis are allo wable without special permission provided that
accurate acknowledgement of sources is made. Requests for permission for extended quotation
from or reproduction of this manuscript in whole or in part may be granted by the head of the
major department or the Dean of the School of Graduate Studies when in his or her judgment
the proposed use of scholarship. In all other instances, however, permission must be obtained
from the author.

Name: ..............
Place: Alemaya University, Alemaya
Date of Submission:.....................
Signature:.................

iv
 ABBREVIATIONS

ACSI Amhara Credit and Savings Association

ANRS Amhara National Regional State
BoFED Bureau of Finance and Economic Development
BoA Bureau of Agriculture
CSA Central Statistical Authority
DA Development Agent
DTU Development Technology Unit
FAO Food and Agricultural Organization
FTC Farmers Training Center
IFAD International Fund for Agricultural Development
LEISA Low external Input Sustainable Agriculture
LPM Linear Probability Model
NGO Non-Governmental Organization
PA Peasant Association
RWH Rainwater Harvesting
SPSS Statistical Measures for Social Sciences
VIF Variance Inflation Factor
WH Water Harvesting
WTP Willingness to Participate

v
 BIBLOGRAPHY

The author was born in Wemberma district, West Gojam zone, Amhara National Regional
State. He attended his elementary education in Wogedad elementary school. He completed his
secondary education in Shikudad Senior Secondary School, Bure town. The author joined the
then Alemaya University of Agriculture (AUA) in 1983 E.C. and graduated with B.Sc. degree
in Agricultural Economics in July 1986.

After his graduation, he was employed in Beneshangul- Gumuz National Regional State
Investment Bureau and worked there as project planning and control junior expert. Then he
was transferred to the Amhara National Regional State and worked in east Gojam zone
agriculture office in different districts of the zone.

vi
 ACKNOWLEDGEMENT

It is God who is to be first acknowledged. He gave me health, strength and patience to

withstand the inconveniencies, which I came across through all the processes of education and
thesis preparation.

My heartfelt thanks are due to my major advisor, Dr. Wegayehu Bekele, for his willingness to
advice and guide me with understanding throughout the course of the study. I want to due
further thanks to him for his charitable time devotion and correcting the thesis on time.

I am indebted to thank my parents particularly my mother and father as well as all my brothers
and sisters who are the sources of my special strength towards the successful completion of
the study and all my life.

The staff of Bibugn district Agriculture and Rural Development Office and all government
employs of the district deserve heartfelt thanks for their encouragement and support.

Finally I would like to express my sincere thanks to all individuals and organizations who in
one way or another support for the success of my research undertaking.

vii
 TABLE OF CONTENT

STATEMENT OF AUTHOR iv

ABBREVIATIONS v

BIBLOGRAPHY vi

ACKNOWLEDGEMENT vii

TABLE OF CONTENT viii

LIST OF TABLES xi

LIST OF FIGURES xii

LIST OF TABLES IN THE APPENDIX xiii

ABSTRACT xiv

1. INTRODUCTION 1

1.1. Background 1
1.2. Statement of the Problem 4
1.3. Objectives of the Study 5
1.4. Scope and Limitations of the Study 6
1.5. Significance of the Study 6

2. LITERATURE REVIEW 8

2.1. Concepts and Definitions 8

2.2. Historical Development of Water Harvesting Practices 9
2.3. Advantages, Opportunities and Weaknesses of Water Harvesting 10
2.3.1. Advantages 10
2.3.2. Weaknesses 12
2.3.3. Opportunities: 13
2.4. Classification of Water Storage Technologies and Systems 14
2.4.1. Underground tanks 14
2.4.2. Aboveground tanks 15
Contd…

viii
 (Table of content contd…)

2.4.3. Surface ponds 15

2.5. Behavioral Studies 16
2.6. Decision Behavior of Farmers about Water Harvesting 17

3. METHODOLOGY 23

3.1. Description of the Study Area 23

3.1.1. Brief description of the Amhara National Regional State 23
3.1.1.1. Population 24
3.1.1.2. Agriculture 24
3.1.1.3. Transport and other infrastructure features of ANRS 25
3.1.1.4. Water resources and irrigation potential of the region: 25
3.1.2. Brief description of Dejen district 27
3.1.2.1. Geography and location 27
3.1.2.2. Population characteristics 27
3.1.2.3. Climate 28
3.1.2.4. Topography and soil 28
3.1.2.5. The farming system 28
3.1.2.6. Agricultural extension 29
3.1.2.7. Livestock production 29
3.1.2.8. Infrastructure 30
3.2. Method of Data Collection and Analysis 32
3.2.1. Data sources and data collection procedure 32
3.2.1.1. Data sources 32
3.2.1.2. Sampling procedure 32
3.2.1.3. Sample size 33
3.2.2. Methods of data analysis 34
3.2.2.1. Binary logistic model 34
3.2.2.2. Preference decision model (multinomial logistic regression) 37
3.3. Definitions of Variables and Working Hypothesis 40
3.3.1. Participation decision study: 40
3.3.1.1. Dependent variable for the participation decisio n model 40
3.3.1.2. The independent variables 41
3.3.1.2.2. Hypothesis 43
3.3.1.2.2.1. Household characteristics 43
3.3.1.2.2.2. Economic factors 44
3.3.2. Study for preference among water storage structure groups 48
3.3.2.1. Dependent variable 49
3.3.2.2. Independent variables for the choice model 50
3.3.2.2.1. Variable definitions for the multinomial logit model 50
3.3.2.2.2. Hypotheses 51

Contd…

ix
 (Table of content contd…)

4. RESULTS AND DISCUSSION 55

4.1. Results and Discussion on Participation Decision 55

4.1.1. Results of descriptive statistics 55
4.1.1.1. Household characteristics 55
4.1.1.2. Economic situations 57
4.1.1.3. Farmers' institutional environment 58
4.1.1.4. Training, visiting and perception measures 59
4.1.1.5. Tests of the mean and frequency differences of variables 62
4.1.1.6. Household's willingness to participate (WTP) 64
4.1.2. Econometric results for the binary logistic regression model 65
4.1.3. Sensitivity analysis for participation decision 72
4.2.1.1. Mean and frequency comparisons for multinomial logit analysis 75
4.2.1.2. Preference among water storage technology groups 77
4.2.2. Econometric results of the multinomial logit model 78

5. CONCLUSIONS AND RECOMMENDATION 84

5.1. Conclusion 84
5.2. Policy Implications 87

6.REFERENCES 89

7. APPENDIX 94

7.1. Appendix Tables 94

7.2. Survey Questionnaire 97

x
 LIST OF TABLES

Table Page

1: Definition of explanatory variables to explain participation in water harvesting ................42

2: Definition of variables for multinomial logit .......................................................................50
3: Level of education of sample farmers ..................................................................................56
4: Responsibilities of head of the household ............................................................................57
5: Money shortages ...................................................................................................................58
6: Access to credit.....................................................................................................................59
7: Participation and training......................................................................................................60
8: The importance of water harvesting technology ..................................................................60
9: Irrigation use. ........................................................................................................................61
10: Food availability status of sample households. ..................................................................61
11: T-test for mean difference of continuous variables ............................................................62
12: Chi-square test for frequency difference ............................................................................63
13: Parameter Estimates for binary logit ..................................................................................67
14: Sensitivity analysis .............................................................................................................73
15: Multiple Comparisons ........................................................................................................75
16: Results of chi-square tests ..................................................................................................76
17: Parameter Estimates of the Multinomial Logistic Regression ...........................................79

xi
 LIST OF FIGURES

Figure:1 Map of the study area...............................................................................................31

xii
 LIST OF TABLES IN THE APPENDIX

Appendix Tables Pages

1: Conversion factors used to compute adult equivalent (AE) 94

2: Conversion factors that used to estimate tropical livestock unit 94
3: Values of VIF for continuous variables, which are used in binary logit 95
4: Contingency coefficients of discrete variables in binary logit 95
5: Values of VIF for continuous variables of the multinomial logit 96
6: Contingent coefficients of discrete variables for multinomial model 96

xiii
 FARMERS’ RESPONSE AND WILLINGNESS TO PARTICIPATE IN WATER
HARVESTING PRACTICES: A CASE STUDY IN DEJEN DISTRICT / EAST GOJAM
ZONE

ABSTRACT

Despite the significant contributions of agriculture in the overall economy of Ethiopia, its
productivity and yield per unit area is very low. Many factors contributed to bring up the
problem among which moisture stress is to be cited. The response to cope up with drought and
hence to improve the status of productivity of the agriculture is to adopt some form of
irrigation practice. However, developing large scale irrigation is a costly alternative, which
may require large quantity of capital resource. This is a difficult job to practice for the small
scale resource poor Ethiopian farmers. The alternative, which may compromise the need for
expanding irrigation and the capital shortage, may be promoting small scale irrigation
schemes through practicing water harvesting. Therefore, the major concern of this study is to
identify determinants of farmers’ response and willingness to participate in water harvesting
practices and the choice decision among alternative water storage structure technology
groups. The study is aiming at proposing or indicating policy measures for promotion and
adoption of water harvesting activities. The study was conducted in Dejen district, Amhara
National Regional State. A total of 120 farmers had been interviewed in the study area in
October and November 2004 to generate primary data for the study.

A binary logit model for the willingness/ to participate decision study and multinomial logit
model for the choice decision among alternative water harvesting storage technology groups
were employed. A total of 14 explanatory variables for the binary logit model and 10 for the
multinomial logit models were used out of which 7 variables were significant to affect the
willingness of farmers in water harvesting practices. These are education level of head of the
household, total tropical livestock unit owned, training, financial constraint of the household,
distance of development center from homestead, attitude towards water harvesting technology
and labour availability.

xiv
Taking the specific characteristics of farmers into account in planning water harvesting
projects may help policy makers to come up with projects that can win acceptance by farmers.
The findings of this study indicate that any effort in promotion of water harvesting activity
should recognize the socio-economic, household and technological characteristics.

xv
 1. INTRODUCTION

1.1. Background

Agriculture occupies a key position in the Ethiopian economy and more than 90 percent of the
agricultural productions are generated from the peasant sector. Agriculture provides livelihood
to more than 85 percent of the population and more than 87 percent of the economically active
labor force is engaged in agriculture. Nearly, 90 percent of the export earnings and more than
50 percent of the country’s GDP usually come from the agricultural sector, (CSA, 1997).
Moreover, agriculture is the single most important source of food for the nation.

Given this state of affairs, the creation of conducive atmosphere for the better performance of
the agricultural sector is not only an important and realistic option to increase food availability
and food accessibility but it is also a means to improve the standard of living of the quasi-
totality of the population in the country (Degnet, 1999).

Available evidences, however, indicate that the agricultural sector in Ethiopia has always
remained dominated by small-scale resource poor peasant farms whose production and
productivity are very low. The production systems are traditional and subsistence in nature.
High rate of population growth, unfavorable natural and policy environment have also been
the causes of the inefficiency of agriculture in the country.

More over, Ethiopia currently depends on rain fed agriculture with limited use of irrigation.
The rain fed agriculture is declining in productivity as caused by variable rainfall, frequent
floods and droughts. The inconsistency in the amount and seasonal pattern of rainfall and its
inter annual variation constitute a major cause for frequent failures of crops and scarcity of
livestock feed. The annual rainfall distribution in most parts of Ethiopia, including the
highlands, is not only uneven but also highly unpredictable in its inter annual variations
(Habtamu, 1999).
The frequency of drought phenomena has increased tremendously in recent decades,
particularly since 1950’s. The analysis of the rainfall data in this period indicated the
occurrence of drought in every two years, which was once in seven years for earlier times
(Berhanu, 1999).

As a result of the variable rainfall, increase in the frequency and severity of drought, Ethiopia
has experienced repeated food shortages and famine which has caused a great deal of human
suffering and migration as well as considerable losses in human and livestock lives during the
worst famines like the ones that occurred in the years 1974-75 and 1984 (Ngigi, 2003). This
condition added with other social and political problems has made the country a regular
recipient of food aid in the past two decades.

On the other hand, Ethiopia has ample amount of irrigation potential, which is yet little
exploited. The irrigation potential of Ethiopia including large, medium and small scale
schemes is estimated to be 3.7 million hectares out of which only about 3% is exploited. Some
literatures indicate that out of all potentially irrigable land, about 400,000 hectares are
estimated to have potential for small-scale irrigation schemes. In addition, the rainfall that
occurs over the extent of the country gives rise to an estimated 110 billion cubic meter of
water as mean annual flow (FAO, 1993).

This immense amount of water resource can be used for irrigation using different alternatives.
The most emphasized and most realistic option in Ethiopia is applying all possible efforts to
exploit the aforementioned irrigation potential of the country. A number of factors led to this
conclusion, the most obvious of which is that irrigation increases the potential for producing
more food more consistently in drought prone areas. This remains the central hypothesis for
these activities and investments. Another factor favoring the adoption of irrigation was that
irrigation was seen as a window of opportunity during the mid-1980s (Catterson et al., 1999).
Although, irrigation may be the most obvious response to drought, large-scale irrigation
schemes have proved to be costly and can only benefit a fortunate few who could not afford to
invest (Critchley, 1991). Considering this reality, many experts and other NGOs forwarded
their advice to focus on small-scale irrigation development schemes. Small-scale irrigation

2
development has got vital role in assisting the development of sustainable agriculture in
Ethiopia.

In addition to using rivers and springs for small-scale irrigation, other sources of water for
agricultural production are required to meet the increasing demands of the growing population
of the country. Water harvesting could be one of the coping mechanisms to obtain sufficient
water for farming. Water harvesting can play a significant role, and can make the people self-
sufficient in agricultural products, combat problems of environmental deterioration and
provide base for sustainable agriculture. The development of small- irrigation schemes through
water harvesting techniques will help to distribute the runoff potential from time of excess
rainfall to the short fall season so that moisture deficit can be alleviated, there by reducing
crop failures.

People in some parts of the world where water shortage exists have successfully utilized
water-harvesting systems. The application of water harvesting techniques, although potentially
high, is still low in Ethiopia. Even though, the technology is politically accepted and
recognized, it is new to most development workers and government personnel, hence there are
inadequate strategies, human resources and policies for its promotion (Ngigi, 2003).

This study was undertaken in Dejen district, Amhara National Regional State (ANRS). This
district is one of the districts in the region where rainwater harvesting technologies are being
promoted. This study was done to identify the socio-economic and cultural factors that can
affect farmers’ willingness to participate in water harvesting practices. The aim of the study is
to generate information for policy makers and executive officials for intervention.

3
1.2. Statement of the Problem

The major constraint to successful crop production in the rainfed agriculture is moisture stress
in dry seasons. Hence, problems associated with dependence on rainfed agricultural systems
are common in Ethiopia: repeated famine, repeated crop failure, and human as well as live
stock life loss are some to be mentioned. Therefore, one alternative to improve agricultural
production in the rain fed agriculture is to develop water- harvesting techniques and then use
the limited water efficiently (CTA, 2002).

This led the concerned officials and experts to take the issue seriously and seek long lasting
solutions. Among the means designed to cope with the problem, water harvesting got the top
priority (RDPE, 2002). Hence, a considerable amount of money, time and efforts are being
spent on different water harvesting practices in the past three years.

There are many water harvesting/ runoff farming systems practiced and evaluated at ma ny
different places in the world. Some of the systems have been outstanding successes, while
others were complete failures. Although different rainwater harvesting technologies have
worked under similar or diverse climatic and geographical conditions, it should not be
assumed that a particular technology would be viable in another area with similar conditions,
especially due to socio-cultural factors. Nevertheless, a technology that has failed in one
locality may be a solution to a different area.

It is evident that there are disparities in the rate of adoption and adaptation, and types of
technologies in different countries and even regions within a country. There are various
promising aspects of rainwater harvesting technologies, geared towards improving subsistence
food production, being adopted by rural land users in the Greater Horn of Africa. However,
despite the success of a number of land user initiated systems, the rate of adoption is still low
making their impacts marginal.

This is often because the technologies and designs were not suitable for either the environment
or the cultural habits of the beneficiaries, or because operation and maintenance of the
schemes turned out to be either too costly and/or too time consuming (Ngigi, 2003).

4
In some instances installations with high capital costs can only be justified if there is a strong
communal commitment usually in the form of labor. For a water- harvesting project to be
successful, the society must possess a high degree of individual commitment (FAO, 1994).

In line with this, what is the response of Ethiopian farmers? Are they committed and willing to
participate in any form of water harvesting? What would the probable reasons be for their
positive or negative responses? I think these and other related issues are still questions to the
government, policy makers and for concerned experts. In Amhara region, still there is no such
well-organized and documented research result, which can reveal the status of promotion and
adoption of water harvesting works in the region. Of course, there are reports prepared by
different governmental and non-governmental organizations based on short visit observations.
These report results, however, are not based on systematic analysis and are not sufficient to
point out the wide-ranging factors that affect the promotion and adoption of water harvesting
practices in the region. Hence, the issue requires to be studied intensively and empirically.

Therefore, the problem that should be addressed in this study is to understand whether the
people know the concepts and roles of water harvesting, and how much they are willing to
participate in water harvesting works in general and in different technology options in
particular. Undertaking a study on these aspects is of paramount importance and timely vis a
vis the current rural development strategy of the country.

1.3. Objectives of the Study

The specific objectives of this study are:

1. To analyze factors affecting farmers' willingness to participate in water harvesting
activities; and
2. To identify factors affecting farmers' preferences among alternative water harvesting
storage structure groups.

5
1.4. Scope and Limitations of the Study

The principal concern of this study is to identify and quantify key social, economic, technical
and institutional factors that significantly affect the willingness of farmers to participate in
water harvesting practices. The study is confined in Dejen district, East Gojam Zone, Amhara
National Regional State. Due to constraints aris ing from shortage of financial resources, time,
vehicles and other logistics related problems, it was found necessary to consider only one
district, Dejen. Hence, the research results are primarily based on data collected from a
restricted but randomly selected sample of 120 farm household heads, from five peasant
associations in the district.

However, recommendations and policy implications drawn out of this study could be used in
other locations having comparable or similar conditions.

1.5. Significance of the Study

In the present times, the major concern of Ethiopia is her inability to provide its expanding
population with adequate food and clothing for sustenance and comfort. Agriculture, which is
the single most important sector of the Ethiopian econo my, is mainly dependent on rainfed
farming. In the face of repeated drought phenomena in Ethiopia during recent decades,
dependence on rainfed agriculture has increased vulnerability of farmers to food shortage,
malnutrition and hunger. Repeated famine and hunger caused by drought and land degradation
have become a well known image of Ethiopia in recent years.

Doing all possible efforts to alleviate the ever increasing agricultural problems is, therefore,
not an assignment to be left for tomorrow. In this endeavor, expanding irrigation has been
taken as one important strategy to relieve agriculture from dependence on nature and thus
improve its productivity.

There is ample potential to expand small-scale irrigation in the country so far as the country is
rich in its water resources. Spring development, river diversion and water harvesting are

6
among the possible alternatives. However, choosing the least cost, and socio-economically
acceptable technology demands a conscious research and planning in the field concerned.

Water harvesting is taken as one of the was to expand small-scale irrigation and to alleviate
moisture stress and thereby to improve the productivity of agriculture farmers. Therefore,
identifying and quantifying the socio-economic, technical and institutional factors affecting
farmers' willingness to participate in rainwater harvesting at micro level are firstly expected to
be useful for policy makers and donors who are involved in the promotion of water harvesting
activities in the region and at national level. Secondly, the outcome of this study is expected to
be useful for agricultural researchers to develop and design a technology that can better fit
farmers' socio-economic and agro-ecological conditions. Thirdly, the study is expected to
indicate the weaknesses and strengths of the strategies, which are being practiced for water
harvesting.

7
 2. LITERATURE REVIEW

This chapter discusses on definitions, historical perspectives, advantages and disadvantages of

water harvesting. It also comprises classification of water storage structure groups, briefs
issues to be considered for a proper promotion and adoption of water harvesting activities
across potential users.

2.1. Concepts and Definitions

Ngigi (2003), defined water harvesting as the collection of the rainfall/ runoff for domestic
consumption or food production purposes, which will otherwise cause erosion.

According to Reij et al. (1993), water harvesting is usually employed as an umbrella term
describing a whole range of methods of collecting and concentrating various forms of runoff
(roof top runoff, overland flow, stream flow, etc.) from various sources (precipitation, dew,
etc.) and for various purposes (agricultural, livestock, domestic and other purposes).

Water harvesting is the collection of runoff for productive purposes. Instead of runoff being
left to cause erosion, it is harvested and utilized. In the semi-arid drought prone areas where it
is already practiced, water harvesting is a directly productive form of soil and water
conservation. Both yields and reliability of production can be significantly improved with this
method (FAO, 1995).

According to Boers and Ben-Asher (1982), rainwater harvesting is defined as a method for
inducing, collecting, storing and conserving local surface runoff for agriculture in arid and
semi-arid regions.

Water harvesting can be considered as a rudimentary form of irrigation. The difference is that
with water harvesting (WH) the farmer (or more usually, the agro pastorals) has no control
over timing. Runoff can only be harvested when it rains. In regions where crops are entirely
rainfed, a reduction of 50 percent in the seasonal rainfall, for example, may result in a total

8
crop failure. If, however, the available rain can be concentrated on a smaller area, reasonable
yields will still be received. Of course, in a year of severe drought there may be no runoff to
collect, but an efficient water harvesting system will improve plant growth in most parts of the
years (CTA, 2002).

Water harvesting involves collecting rainwater or runoff from a surface and storing it for
domestic use, livestock and crop production. The surface can be a roof, rock or simply ground
surface. Storage options include tanks, ponds, reservoirs, and in the soil itself (FAO, 2000).

2.2. Historical Development of Water Harvesting Practices

It is a common practice for most of the households in Ethiopia to employ ‘informal’ rainwater
harvesting by placing bowels, jugs, or any convenient material at hand under eaves or eve n
trees during rainfall. Despite the above fact, there is no readily available document, which can
indicate the actual date as to when water harvesting has been started in Ethiopia.

However, some available physical evidences show that in Ethiopia ancient churches,
monasteries and castles used to collect rain water from rooftops; and the history of rain water
harvesting by the Axumite Kingdom dates back as early as 560 BC (Habtamu, 1999). During
this period, rainwater was harvested and stored in ponds for agriculture and drinking water
supply purposes. This is confirmed from literature and visual observations on the remains of
ponds that were once used for irrigation during that period (Ngigi, 2003). The documents also
provide additional evidence for roof water harvesting set up in the remains of one of the oldest
palaces in Axum. Other evidences include the remains of one of the old castles in Gondar,
constructed in the 15-16 century which has a pond built for rainwater harvesting for drinking
and religious rituals by the Kings. Also during the rule of King Lalibela, ponds and
underground water storage tanks were used both for drinking and religious rituals within the
system of rock- hewn churches that have existed up to now (Ibid).

The first water harvesting facility, however, was in all likelihood nothing more than a
depression in a rock surface that tapped rainwater. The collected water served as a drinking
water supply for man and animals. These water depression storages are still found in many

9
parts of the world and serve as a source of drinking water supply for many forms of wildlife. It
is highly probable that the first constructed water harvesting facility was simply an excavated
pit or other water storage container placed at the outfall of a rocky ledge to catch runoff water
during a rainstorm. The next evolution might have been to construct a rock diversion wall or
gutter to provide a larger collection area. Researchers have found signs of early water
harvesting structures to have been constructed over 9000 years ago in the Edom mountains in
southern Jourdan (FAO, 1994). There is evidence in Iraq that simple forms of water harvesting
were practiced in the Ur (Iraque) area in 4500 Bc. Along desert roads, from the Arabian Gulf
to Mecca, there still exis t water harvesting systems that were constructed to supply water for
trade caravans (FAO, 1994).

As for the recent development of water harvesting, a growing awareness of the potential of
water harvesting for improved crop production arose in the 1970s, with the widespread
drought in Africa leaving a trail of crop failures. The stimulus was the well-documented works
on water harvesting in the Negev Desert of Israel (Critchley, 1991).

Though water harvesting, in its traditional form, has long been practic ed in Ethiopia,
promotion and application of rainwater harvesting techniques as alternative interventions to
address water scarcity were started through government initiated soil and water conservation
programmes. It was started as a response to the 1971-74 droughts with the introduction of food
for work programmes, which were intended to generate employment opportunities to the
people affected by the drought. Water harvesting during this time included, among others,
construction of ponds, micro dams, bunds, and terraces in most parts of the drought affected
areas in Tigray, Wollo and Hararghe regions (Ngigi, 2003).

2.3. Advantages, Opportunities and Weaknesses of Water Harvesting

2.3.1. Advantages

Agricultural development based on water conservation and irrigation is often considered a

promising venue for poverty alleviation in rural areas. For example, the availability of water
for a small domestic garden plot, usually managed by women, can make a significant

10
difference to household nutrition and thus contribute to improved livelihoods. Water
harvesting may make this possible (FAO, 2003).

Water harvesting can be used for flood control mainly in two ways. The first is by maximum
infiltration and holding of rainwater in the soil where it falls and the second is by building
dams and reservoirs to store peak floods (Hatibu and Mahoo, 2000). In addition, adoption of
rainwater harvesting goes hand in hand with horticultural production. It also reduces the
livestock walking time to look for water, which may otherwise reduce the grazing time and
hence reduce growth and productivity.

Boers and Ben-Asher (1982) pointed out that water harvesting would lead to substantial yield
increases by reducing the water constraint to plant in arid and semi-arid regions. It coincides
with plant nutrient harvesting because organic matters and other plant nutrient materials will
not be washed away. Water harvesting systems are relatively small-scale operations in terms
of catchments area, volume of storage, and capital investme nt so that small scale farmers may
prefer to adopt from other forms of irrigation.

Rainwater harvesting is an active anti-drought way that has a significant economic benefit.
Poverty reduction pace has been hastened in the province of Gansu, China, by adopting
rainwater harvesting. Yields have shown to increase substantially when crop water supply is
stabilized by the use of combinations of small scale supplementary irrigation schemes and
runoff collection (Perrier, 1988). The system in addition, helps to promote eco-environment
improvements. Gansu province is using water cisterns to greatly develop forestry to improve
eco-environment and control soil erosion for better ecosystem (Weide, 2003).

Water harvesting is a simple, cheap, and environmentally frie ndly technology. It minimizes
some of the problems associated with irrigation such as competition for water between various
uses and users, low water use efficiency, and environmental degradation. The technology can
also be integrated with many land use systems; hence it is appropriate for local socio-cultural,
economic and biophysical conditions (LEISA, 1998).

11
 Rees (2000) mentioned the following advantages: water harvesting gives very convenient
supply of water as per requirement; it is largely independent of outside organization for
construction and maintenance, and also gives fairly high water quality, which may be further
increased by simple means. It is also believed to reduce the area of rotting iron. Because of
fewer overhangs, water harvesting reduces run-off erosion, especially in townships and urban
areas.

DTU (2002) suggested that collecting rainwater is not only water conserving, it is also energy
conserving. Furthermore, rainwater is soft and can significantly reduce the quantity of
detergents and soaps needed for cleaning as compared to typical municipal tap water. The
introduction of more formal rainwater harvesting will normally be accompanied by three
benefits. The most obvious is a reduction in the time spent carrying water from point sources-a
reduction more or less proportional to the volume of water no longer carried. The second is an
increase in household water consumption wherever it was previously constrained by the effort
of collection. The third is a common, although not invariable, increase in water quality.

According to Mintesinot (2002), rainwater harvesting is successful where other sources are not
available or when full potential of other sources cannot be achieved. According to him, water
harvesting is considered as a source of superior quality water in some areas with low
operational and maintenance costs and easy to manage by users. It can be used as a buffer
security, and also avoids conflict among users when system is owned individually. Rainwater
harvesting is a simple technology, which has the potential to dramatically improve water
security.

2.3.2. Weaknesses

The disadvantages in runoff irrigation especially in micro-catchments system are that the
terrain should preferably be even and the success of this method of irrigation depends very
much on rainfall and soil type (FAO, 1994). In addition, water harvesting system for
supplementary irrigation is small scale to make significant change at macro level within a
short time period (FAO, 2003). The catchments area and storage capacity of a system are also

12
relatively small as well as rainwater storage tanks may take up valuable spaces (Lo, 2003).
Rainwater harvesting may also cause incidence of disease such as malaria FAO, (2003).

Rees (2000) suggested that water harvesting involves high cost per household unless cheaper
storage methods can be found. It also discriminates against the poor who have grass roofs and
insufficient funds. Storage structures may be vulnerable to earth tremors to which the area is
subject. Moreover, water harvesting is vulnerable to drought years.

Mintesinot (2002) also suggested that water harvesting involves the risk of flat test and easier
contamination of open tanks, and high dependency on rainfall. Users usually suffered from
lack of knowledge on the technology and have poor awareness. Water harvesting demands
high per capita cost and difficulties in replicating, insufficient catchments areas. It is incapable
for providing water throughout the year as single source, and difficult for thatched roofs.

2.3.3. Opportunities:

Mintesinot (2002) suggested that water harvesting will have good opportunity when problems
like ground water depletions/salinity /high cost of water treatment is an issue. If well designed,
managed and promoted, water ha rvesting has user acceptance and users are committed to
participate as well it is easy to mobilize rural communities. Its opportunity may also be seen
when other water sources are at a distance.

Under bimodal high rainfall situations, chance of success for water harvesting system is very
high. It also helps ensure certainty of water supply (even if the duration may be short). Water
harvesting is relatively easy to demonstrate the advantages to water vendors (water trading,
relief to women, increase awareness among stakeholder, help to reach communities on
demand responsive approaches) (Rees, 2000).

13
2.4. Classification of Water Storage Technologies and Systems

Because of the intermittent nature of precipitation runoff, water storage must be an integral
part of any water harvesting system (FAO 1994). Any container capable of holding water is a
potential water storage facility. The storage generally means confinement in excavated pits,
cisterns, ponds, bags or tanks.

There is almost unlimited number of options for storing water. The type of storage selected for
use in a water harvesting system depends upon many factors such as the ultimate use of the
water, availability of construction materials, availability and skill level of labour, and the site
topography. The water storage facilities can vary in shape, size, functions, uses of water, and
in local names. The choice of system therefore will depend on a number of technical and
economic considerations such as space availability, local traditions for water storage, cost of
purchasing new tank, cost of materials and labour for construction, materials and skill
available locally, and ground conditions.

2.4.1. Underground tanks

Excavated tanks, dugouts or cisterns serve the purpose of storing surface runoff for longer
time, which can be used for irrigation, livestock or domestic water supply. Below ground
tanks have the greatest potential for cost savings, particularly if built in a stable soil, which can
be relied on to take all or part of the load. Walls can be much weaker and materials that have
good waterproofing properties but suffer from a low strength can be used (Rees, 2000;
Martinson et al., 2001). According to the same authors, this type of structures is generally
cheaper due to lower material requirements, but more difficult to empty by leaving tap on.
Underground structures require little or no space aboveground and unobtrusive. Surrounding
ground gives support allowing lower wall thickness and thus lowers costs. The cons of the
underground water harvesting structure group are that it is problematic for water extraction
and often requiring a pump. Leaks or failure are more difficult to detect, and tree roots can
damage the structure. There is danger to children and small animals if tank is left uncovered
(Rees, 2000).

14
 Dom shaped and cylindrical structures are the main types of underground structure types,
which are being practiced in the study area. The cost calculation for both types is too difficult
as it depends on the soil type, slope of the plot, availability of local materials and other similar
concerns. However, based on some available documents and standards the cost of Dom shape
is estimated to be 5420 birr. Similarly, the cost of cylindrical type structure is assumed to be
6210 birr.

2.4.2. Aboveground tanks

Thanks aboveground are usually meant for roof top rainwater collection. Collecting water
from roofs covered by corrugated, iron sheet, thatch or tiles and requires storage facilities
aboveground like oil drums or tanks built from concrete.

Aboveground designs are generally more popular than below ground types. However, the cost
is often higher as the tank must now cope with the full force of the water pressure acting on it.
The principle of functional separation allows some scope for cost reduction by using an
inexpensive material for structure while waterproofing can be done by either a mortar or liner
(Martison et al., 2001). Rees (2000) suggested that aboveground structure groups are easy to
inspect for cracks or leakages. Many designs exist to choose, and can be manufactured from a
wide variety of materials. It is easy to construct from local materials, as well it is easy to
extract water. He also pointed out that they are more expensive. In addition, and also easily
damaged, as it is prone to attack from weather. Failure may be dangerous.

The typical example is Ferro-cemented which can be constructed from concrete and other
materials. It can be constructed in various sizes, which may require relatively a higher cost.
The cost is estimated to be about 7230 birr.

2.4.3. Surface ponds

Ponds provide important sources of water for crop production, livestock, washing, basing, and
for drinking when better water is not available. If the pond and its watershed are protected
from human and livestock contamination, these sources can provide relatively safe water,

15
especially for washing and basing. On the contrary, when they are not protected from
contamination, they can be a source of disease.

Ponds for water harvesting is very appropriate to the local situation (Borsa, 2001). It can be
sustainable and replicable provided that the management system is reinforced. It provides a
viable option for providing water to most population and their animals. Trapezoidal and geo-
membrane ponds, which are introduced in the district, are typical examples. Geo- membrane
ponds are constructed with about 3200-4000 birr.

2.5. Behavioral Studies

According to Rogers and Shomaker (1971), adoption is defined as the decision making
process in which an individual passes from first hearing about an innovation to final adoption.
The decision of whether or not to adopt a new technology hinges upon a careful evaluation of
a large number of technical, economical and social factors.

Yapa and Mayfield (1978) suggested that the adoption of an entrepreneurial innovation by an
individual requires at least four conditions. These are (a) the availability of sufficient
information (b) the existence of a favorable attitude towards the innovation (c) the possessio n
of the economic means to acquire the innovation (d) the physical availability of the
innovation.

Degnet (1999) had summarized different empirical studies on the association between
adoption decision and the factors, which influence adoption, particularly in less developed
countries into the following groups. 1) Personal characteristics such as age, education, gender,
farming experience and social status of the head of the family, 2) farm characteristics-farm
size, availability of livestock, oxen availability, off farm income, and availability of cash, 3)
supply and institutional factors such as farmers` access to credit, agricultural extension
services, market and price of products, access to and availability of inputs.

16
2.6. Decision Behavior of Farmers about Water Harvesting

The success or failure of any rainwater harvesting technology will ultimately depend on the
degree of acceptance by the land users. It is essential that the needs and aspirations of the land
users are clearly understood and fully provided for in the planning, designing and
implementation process. It should give sense in terms of productivity of resources used. It is
difficult to judge if the resource is used efficiently in case where the water harvested is
essential for survival. Efficient resource utilization can be judged very easily where water
harvesting is a means of profit (Martison et al., 2001).

Many researchers and experts in the field of natural resources conservation and water
harvesting forwarded their reasons about different factors that affect the willingness of farmers
to participate and efficiently use water- harvesting works. As mentioned in CTA (2000)
widespread adoption of water harvesting techniques by the local population depends on cost
and simplicity of the technology for implementation and maintenance.

Another consideration would be whether success in rainwater harvesting promotion and

adoption is facilitated by integrating different forms of rainwater harvesting systems, and
whether the introduction of storage systems would be facilitated in situations when no off
farming and in-situ water conservation are already practiced.

Adoption of rainwater harvesting technologies despite their technical benefits will depend on
knowledge of socio-economic and cultural dynamics, on the part of the technology developer,
and on the farmers/ community perceptions. A comparison of promotion approaches of the
same technology in different environments, either by the same development agency or
different actors, reveals the importance of participatory, land user friendly approaches, and
due consideration of socio-economic and cultural backgrounds (Ngiggi, 2003).

Abdulkarim (2002) enumerated conditions that should be considered during water harvesting
system planning and design for the technology to be more acceptable by the users. These are
a) Socio-economic aspects which may include community acceptance and participation,

17
understanding of needs and aspirations, prioritization of needs, appropriate technology, proper
planning and analysis, pilot scheme approach, technical services, cost and benefit
consideration of investment and operation, application, b) Agronomic aspects which includes:
appropriate agro climatic zones, crop selection, soil condition, crop management, c)
Environmental aspect which includes ecological impact, downstream effect, introduction of
disease, d) Institutional/legal aspect which includes land and water use right, management and
organization, services.

FAO (1993) put remarks on the socio-economic factors to be considered for adoption of water
harvesting technologies. According to this document, water harvesting/ run off farming
techniques are technically sound methods of water supply for most parts of the world. What is
not often realized is that these techniques are also relatively expensive. They require
considerable efforts to install and must have periodic maintenance for continued success and
satisfactory operation. There have many water harvesting/runoff farming systems constructed
and evaluated at ma ny different places in the world; some of the systems have been
outstanding successes, while others were complete failures. Some of the systems failed,
despite proper material and designed for extensive labor with high capital costs. In some
instances, installation with high capital costs can only be justified if there is a strong
communal commitment usually in the form of labour. For a communal water harvesting
project to be successful, the local users must possess a high degree of individual commitment
to group projects.

There is a strong interaction of the physical water harvesting system and social structure of the
water user. If the physical system is changed or altered then consideration must be given, to
the impact of the changes on the social structure of the water users, small scale projects may
be more appropriate for communities without any sort of social organization tradition. In many
run off farming systems the micro catchments design is a preferred approach. It is easily
installed and maintained on a family scale as well as suitable to larger communal enterprises.

Understanding the social aspects of a local community or group of potential water users is
important in designing the appropriate system. There is a positive albeit complex, link

18
between water services for irrigation and other far use, poverty alleviation and food security
(IFAD, 2001). The availability of water confers opportunities to individuals and communities
to boost food production, in both quantity and diversity, to satisfy their own needs and also to
generate income from surpluses. Irrigation has a land augmenting effect and can therefore
mean the difference between extreme poverty and the satisfaction of the households basic
needs.

Throughout the arid and semi-arid regions of the world there is a wide range of indigenous
water harvesting techniques known, but experience has shown that they do not always transfer
well from one set of conditions to another. Apart from basic technical prerequisites, the
selected method must be compatible with local life-styles, social systems and the willingness
of the beneficiaries to adopt it. People naturally do not like to put a lot of efforts into schemes
which have a low success rate nor if their land tenure is insecure.

FAO (1995) stated similar criteria to be considered while selecting water harvesting
techniques. To mention, firstly, before selecting a specific technique due consideration must
be given to the social and cultural aspects prevailing in the area of concern as they are
important and will affect success or failure of the project. Second, it is becoming more widely
accepted that unless people are actively involved in development projects, which are aimed to
help them, the projects are doomed to fail. It is important that the beneficiaries participate in
every stages of the project.

As a concluding remark, it is essential to mention and summarize focusing on very influential

factors that determine farmers’ willingness to participate in water harvesting practices. Hence,
the following points were commonly cited in many literatures, as determinates of farmers’
decision to participate in water harvesting activities.

Awareness and sensitization: surprisingly, despite the fact that rainwater harvesting have been
around for hundreds of years, it has never been given sufficient attention as viable solutions to
our food and environmental problems. If it had been given sufficient attention, like other types

19
of agricultural technologies, with the accompanying services, equipment and personnel, the
situation would have been radically different (Critchley, 1991; FAO, 1994).

Poverty and economic status: for much of rural Africa, it has been argued that inefficiency in
agricultural production often reflects poverty rather than ignorance or mismanagement. In the
long term, one of the solutions to poverty could be rainwater harvesting. Thus, simple
rainwater harvesting technologies for users are necessary while involving the beneficiaries on
the choice of the technologies (Ngiggi, 2003).

Technology development and dissemination: rainwater harvesting is relatively new to most

people and only few institutions and trained personnel, in particular at community level, have
acquired adequate knowledge. Sometimes, it has been found that the requirements of the
technologies are not at part with the resource level of the target implementers, especially due
to poverty. Land users want to invest in simple technologies, which they can replicate in their
own time without demand for skill from outside (FAO, 1994; Martison et al., 2001).

Legal, policy and institutional issues: despite the centrality and potential of rainwater
harvesting in alleviating food scarcities, it is surprising to find no comprehensive policy
guiding it in the Greater Ho rn of Africa countries (Ngiggi, 2003).

Socio-cultural perspectives: since the technologies were addressing a basic felt need, that of
access to food, there was found no conflict with the socio-cultural aspects of all the
communities analyzed. The vario us aspects of the socio-cultural settings of the communities
are the results of their long history of social existence and relationship with the environment
(Rees, 2000; Ngigg, 2003).

Gender aspects: We should put more effort on women participation in rainwater harvesting
activities, especially, through various systems of empowerment. Women are essentially closer
to the rainwater harvesting technologies because their work entrails a close relationship with
them. They relate directly and are in constant interaction with the technologies in their daily

20
activities for subsistence needs especially food production, and the search for water (Critchley,
1991; Ngiggi, 2003).

Environmental and health issues: in general, rainwater harvesting systems are environmentally
friendly, but there are some minor concerns such as increased mosquito, snails, poor water
quality and so on. The main problem with the on farm storage technology is breading of
mosquitoes, which cause malaria. However, this problem is not pronounced. Rainwater
harvesting technologies have been reported to reduce soil erosion by capturing sediments and
hence improving soil fertility and reducing situation of water reservoirs (Hatibu and Mahoo,
2000).

Extension services, actors’ collaboration and ne tworking: Making use of the available
extension services within government departments and equipping them with the necessary
skills and material support would enhance adoption and replication. Land users
exchange/exposure visits and stakeholders’ collaboration and networking would suffice in
disseminating the technologies (Mamdouh, 1999; Ngiggi, 2003).

Water related conflicts: rainwater harvesting systems are meant to reduce water scarcity
related conflicts especially competition over limited water for various uses. Downstream-
upstream conflicts could also be reduced if upstream users harvest rainwater, hence allowing
adequate water to reach down stream users (Christopher and Silva-Ocha, 2001; Martinson et
al., 2001).

Peoples’ priorities: if the objective of rainwater harvesting projects is to assist resource poor
farmers to improve their production systems, it is important that the farmer's/ agro pastoralist's
priorities are being fulfilled, at least in part. Otherwise, success is unlikely (Critchley, 1991;
FAO, 1994). It is becoming more widely accepted that unless people are actively involved in
the development projects, which aim to help them, the projects are dimmed to failure. It is
important that the beneficiaries participate in every stage of the project. When the project is
being planned, the people should be consulted and their priorities and needs assessed. During
the construction phase the people again should be involved supplying labor but also helping

21
with field layouts after being trained with simple surveying instruments (Hatibu and Mahoo,
2000; Christopher and Silva-Ocha, 2001).

Area differences: it is tempting to assume that a system which works in one area will also
work in another, superficially similar zone. However, there may be technical dissimilarities,
such as availability of stone or intensity of rainfall, and distinct socio-economic differences
also (Critchley, 1991).

Equity: land tenure issues can have a variety of influences on water harvesting projects. On
the one hand, it may be that lack of tenure means that people are reluctant to invest in water
harvesting structures on land, which they do not formally own. Where land ownership and
rights of use are complex, it may be difficult to persuade the cultivation to improve land that
someone else may use later. On the other hand, there are examples of situations where the
opposite is the case in some areas farmers like to construct bunds because it implies a more
definite right of ownership (Chritchley, 1991).

Village land use management: degraded land in and around villages can only be improved if
the communities themselves face land use management issues. One of the techniques, which
can assist in rehabilitation of degraded land, is water harvesting, but it is only one tool among
several others and cannot be effective in isolation (Ngigi, 2003).

22
 3. METHODOLOGY

In this chapter description of the study area, methods of data collection and analysis as well as
definition of variables and formulation of hypothesis are discussed in detail.

3.1. Description of the Study Area

Description of the study area is one of the essential components of any scientific report.
Hence, the socio-economic, demographic and biophysical features of the Amhara National
Regional State and Dejen district are briefly described.

3.1.1. Brief description of the Amhara National Regional State

The Amhara National Regional State (ANRS) is one of the constituent states of the Federal
Democratic Republic of Ethiopia. It is located in the northwestern part of the country between
80 45’ and 130 45’ North latitude and 350 46’ and 400 25’ East longitude. The boundaries of
ANRS adjoin Tigray Region in the North, Oromiya Regional State in the South, Afar regional
State in the East, Beneshangul Gumuz in the Southwest and Sudan in the Northwest. The
region is divided into 11 administrative zones, including the special zone, Bahir Dar. The
other 10 administrative zones are East Gojam, West Gojam, Awi, North Gondar, South
Gondar, North Wollo, South Wollo, North Sheo, Wag Himra, and Oromya. The region
consists of 106 districts and more than 5,300 peasant associations.

The total area of the region is estimated to be about 170,150 square kilometers. The
topography is divided mainly into plains, mountains, valleys and undulating lands. The high
and mid altitude areas (about 65 percent of the total area) are characterized by chains of
mountains and a central plateau. The lowland part, constituting 33% of the total area, covers
the western and eastern parts of the region. These are mainly plains that are large river
drainage basins. Of the total area of the region, 27.3% is under cultivation, 30% is under
grazing and browsing, 14.7% is covered by forest, bush, and herbs, and 18.9% is currently not
used for productive purposes. The remaining 9.1% represents settlement sites, swampy areas,

23
and lakes. The highland areas are rugged and mountainous with peak rising up to 4620 meters
at Rasdashen (the highest peak in Ethiopia).

The recorded mean annual temperature of the region ranges from 12.40 in Mehal Meda (high
altitude) to 27.80 C in Metemma arid kolla. The mean annual rainfall recorded for the region is
in the range of 598.3 mm in Lalibela and 1692mm at Chagni. The southern plateau and central
parts of the region receive about 1000 mm of annual rainfall (CSA, 2003). The Northwestern
and Northeastern parts of the region along the boundary with the Sudan, Tigray and Afar
regions receive the lower amount of rainfall, which is less than 700 mm. The highest rainfall
occurs in "Meher" season, which starts in Mid-June and ends in Early September.

3.1.1.1. Population

Based on the 1994 population and housing census of Ethiopia, the total population of the
ANRS is predicted for the 2004 to be 18654799. Of the total, the numbers of male and female
population were estimated to be 9348508 and 9306291, respectively. Moreover, about
16564096 or (88.79%) people are living in rural areas and the rest 2090703 or (11.21%) are
urban dwellers.

3.1.1.2. Agriculture

The agricultural sample survey result of the statistical authority shows that the total area
cultivated under major crops (temporary crops, cereals, pulses and oil seeds) is estimated to be
3,081,220 hectares. Of these, 80.73 percent is under cereals, 12.53 percent is under pulses and
the rest 6.74 percent is under oil seeds.

Based on 2003 BoFED annual report, total supply or available food grains in the region was
21330330 quintals and the per capita availability is estimated to be about 11.8 quintals. During
the same year 28511860 hectares of land were covered with different crops and the total
production was reported to be 92746585 quintals. Manually operating traditional tools are
dominantly in use without modification. Moreover, smallholder farmers that contribute for the
less produc tivity of the sector dominate the agricultural sector in the region. Regardless of its

24
low productivity, agriculture is the source of employment for more than 87 percent of the
residents of the region and serves as main source of income.

3.1.1.3. Transport and other infrastructure features of ANRS

Regarding road coverage, there were 606 kms asphalt, 2384 kms all weathered, 2809kms dry
weathered roads. Road density per '000 square kilometer was about 35.8 kilometers.The
numbers of functional telephone lines were 53983 and 135 satellite stations, which comprise
27 automatic, 60 semi automatic, and 2 manual stations were on service. Telephone service
coverage per 100,000 persons was 304.8. There were 5 branch, 13 post level, 25 regular, 97
agents' post offices and 2 postman.

3.1.1.4. Water resources and irrigation potential of the region:

The major water sources of the region are the Abay, Tekeze and Awash river basins and the
Afar drainage basin. The Abay river basin, the largest of all basins in the count ry, has about
198,812 km2 area, of this 89,857 km2 (46%) falls in the region and it drains more than 80
percent of the land surface (BoFED, 2003). Some of the surface water resources including lake
Tana (the largest inland lake in the country), and the other creator lakes like Zengena, Tirba,
Hike and etc are found in the Nile river basin.

Tekeze and Awash River basins drain the remaining 20 percent of the land surface and a very
small portion of the region is drained by Afar drainage basin. Tekeze has about 88,800 km2
areas and the upstream part of the catchments area of the river is within the region. The Awash
river, which has Borkena, Robit, Upper Mile, Upper Kesem and Girana sub basin tributaries,
drains the South and Southeast plateau. Eastern lowlands drain itself in lake Abe. The Afar
drainage basin, on the other hand, serves only as drainage network for small portion of the
eastern lowlands including the Raya Kobo valley, which are within the lowest rainfall area in
the region.

The overall surface water flow by the tributaries of these river basins was estimated to be
about 35 billion m3 per annum, which is about 29 percent of the country's potential (122

25
billion m3 ). The hydropower potential of the region was estimated to be more than 6,000 mega
watts per year, out of which not more than 0.2 percent is yet developed. Moreover, studies also
indicated that from these river basins and tributaries about 22,600 tones of fish could be
harvested per year, out of which only less than 0.5 percent (20 tones) utilized. The ground
water potential of the country as a whole was not studied. But some minor practical work to
utilize the ground water potential for irrigation had been started in the Kobo Girana Valley
Development in the region.

Irrigation potential of the region is also enormous as indicated by the primary surveys and
master plan studies conducted in the region. Based on this survey, about 700,000 hectares of
land are suitable for conventional gravity irrigation. Out of this total, 420,000 hectares are for
medium to large irrigation potential in the Nile river basin. Another 30,000 hectares were
estimated to be potentially irrigable within the Awash river and Afar drainage basin. The
remaining 250,000 hectares were the potentials for small-scale irrigation in the region.

Despite the aforementioned irrigation potential of the region, rain fed agriculture is the
predominant means to produce different crops that are used for consumption and sale. The
contribution of the irrigation agriculture was about 8 percent of the regional crop production,
which was about 52.5 million quintals. As reported by BoA (1999), of the total irrigated land
96 percent was covered by traditional irrigation and 4 percent was by modern irrigation
schemes. Out of the total irrigated crop production 95 and 5 percents were produced by
traditional and modern irrigation schemes, respectively.

Nowadays, it is crucial and highly important to increase the rate of development of irrigation
agriculture to make the agriculture production sustainable so as to bring about sustainable
economic growth. One means designed to expand small scale irrigation is promoting rainwater
harvesting program. It is believed to solve the food insecurity problem of the region.

26
3.1.2. Brief description of Dejen district

Assessment of the some socio-economic and biophysical backgrounds is essential before some
study is carried upon at a specified region.

3.1.2.1. Geography and location

Dejen district is one of the 14 districts in East Gojam zone with an area of 571 square kms. It
is known to be the second smallest district in the zone with respect to its area coverage. It
borders one National Regional State, Oromiya, and other 4 districts. In the south, it borders
with Oromiya National Regional State, in the Southeast it shares border with Sheble Berenta
and Enemay districts, in the North with Debay Tilate and in the west with Awable districts.
The capital town of Dejen district is Dejen town and located at about 230 k.m. in the northeast
direction from Addis Ababa, the capital city of Ethiopia. The district has 19 peasant
associations and two- urban kebele administrative.

The study was conducted in five peasant associations called koncher Sasaberay, Borebor
Shencha, Werk Amba Berch, Yetnora Tik and Kurar. They in sum comprise a total population
of 30575.

3.1.2.2. Population characteristics

Based on 1994 population and house census results, about 132 609 people are expected to live
in Dejene district in 2004. Among those, 67142 were predicted to be male and 65450 to be
female. In addition, the prediction result indicated that 128,930 people are expected to live in
rural areas and the remaining 3,672 people in towns. The population density of the district is
about 232 people per square kilometer. The economically active population (15-64) years of
age accounts 47% of the total population.

27
3.1.2.3. Climate

Climatically, the district can be categorized mainly into two zones. These are mid-altitude,
which comprises about 68 percent of the total land mass and 13 peasant associations, and the
lowland, which comprises about 32% of the total land mass and 6 peasant associations (PAs).
The low land PAs experience frequent drought, and hence moisture stress is the binding
problem for their agricultural production. The mid-altitude PAs on the other hand, receive
relatively sufficient amount of rainfall and are more productive.

The temperature also shows a wide range of variation between PAs in the gorge of the Blue
Nile and those that are located in mid-altitude areas. Very high temperature is recorded in low
land PAs and from moderate to cold in mid-altitude PAs.

3.1.2.4. Topography and soil

Undulating and plain are the two major topographical features of Dejen district. Those PAs in
the valley of the Blue Nile have highly undulated topography consisting of hilly lands and
ridges. On the other hand, about 9 PAs have flat type of topography.

The type and texture of the soil varies to a great extent. Those PAs that are located at mid-
altitude mostly have heavy black soil. On the other hand, lowland PAs have gray, red whitish
soil.

3.1.2.5. The farming system

The farming system in the region is characterized by mixed farming. The agro-climatic
condition of the region is favorable for growing diversified types of crops and rear different
species of animals. Tef, maize, wheat and barely are the dominant crops frequently grown in
the district. Production is undertaken mainly by waiting the rainy season that is once per year.
If rain is not sufficient in amount and do not keep its normal cycle, farmers in the area often
face hazards of drought and hence food shortage.

28
The land use system is mainly influenced by the land tenure. Regarding the land tenure
system, all farm households in the study area as in the rest of the country, had been entitled to
use their holdings but not allowed mortgaging or selling it. The land size varies from one PA
to another due to the differences in the available land resource and the population size among
the PAs.

Farmers in the study area use their land mainly to produce cereal crops, and to some extent to
graze their animals. They never leave their lands as fallows, because there is an obvious trend
of shrinking land-holding from time to time, which is already below the optimum holding size
to provide a living with the available technology being used by the community.

3.1.2.6. Agricultural extension

In Amhara region, the bureau of Agriculture and Rural Development is the principal
responsible to run extension of agricultural technologies that are developed and released by
research centers. At present the extension approach is undergoing a transition from one DA in
each PA or village to 3 specialized diplomas graduate DAs in each farmers training centers
(FTC). The main task of the DAs in the former system/approach was to demonstrate,
popularize and disseminate agricultural technologies and modern farming systems. On the
other hand, the task of DAs in the new system is a bit different. It mainly focuses on training
farmers at FTC for longer time.

3.1.2.7. Livestock production

Livestock play a significant role in the mixed farming system of the area. Their main
contribution is in providing draft power, cash generation, food (example milk), and as a status
symbol. Livestock types kept by the farmers include cattle, sheep and goats, equines and
poultry. Oxen are kept to provide draft power, cows to provide farm households with milk and
butter for consumption and sale, donkeys for transporting goods, whilst sheep, goats and
poultry are mainly kept for sale as well as for their meat. The feed sources commonly used for
livestock, include natural grazing, crop residues and in rare cases byproducts from industries.
The contribution of natural pasture as sources of feed is very limited due to the extensive

29
coverage of the land by crops. Consequently, natural grazing for cattle in particular is limited
to farm boundaries and the lower slopes of the hillsides. Goats and sheep are, however,
entirely fed from the natural vegetation in the bushes and hillsides.

At present, livestock based farming is becoming reduced and has got lesser emphasis in the
study area. Two contradicting forces play their role for the reduction in livestock population.
On the one hand, due to the ever- increasing trend of population growth, even marginal lands
are becoming under cultivation, leaving minimum and even no land for grazing and animal
production. The other cause for the reduction of animal population in the area is that farmers
use traditional and extensive system of animal production that cannot cope up with the
prevailing shortage of grazing land. Therefore, the regional and district agricultural experts
and other concerned parties have to do a lot to improve the existing problem.

3.1.2.8. Infrastructure

Presence of infrastructure is an important vehicle for the transformation of a rural economy.

Roads and communication net works, health and educational infrastructures, and potable water
supply, availability and access to input and output markets are some of the infrastructure
components that are necessary to improve the production and productivity of the rural poor. In
Dejen district, there are about 21 elementary and one high school. With respect to health
infrastructure, there are 1 health center and 6 health posts. Only about 23 percent of the people
in the district are provided with potable water. When we see the road coverage, about 40
kilometers asphalt, 10 kilometers all weathered and 23 kilometers dry weathered roads are
available.

There are two major marketing places in the district called Dejen and Hagereselam market
places. The Dejen market is larger in its volume of transaction than the Hagerseselam. In
addition to the above two marketing places, there are others such as Yetnora, Kurar and
Enajima which farmers use to exchange industrial and farm products. Pack animals are the
main means of transportation and larger share of farm produce is transported using them.
Human labor is another means to transport farm produce to the local market.

30
Map of East Gojjam administrative zone
Amhara Amhara national regional state
National
Regional
State

Dejen district

31
 3.2. Method of Data Collection and Analysis

Here procedures and principles of data collection for this study are discussed. Sample size
determination and interviewing procedures are also parts of the main concerns for this part. As
a natural follow up of data collection is analysis and interpretation of the collected data using
appropriate statistical tools. Hence, theoretical econometric models, which are used for the
study, are discussed.

3.2.1. Data sources and data collection procedure

3.2.1.1. Data sources

Both primary and secondary data sources were used for this study. The primary data were
collected from October to December 2004, using structured questionnaires. The data were
collected on socio-economic aspects of the household such as farmers age, education level,
plot size, soil type, livestock number, marketing, production and other biophysical and
institutional factors which were believed to explain farmers decision behavior regarding
rainwater harvesting participation.

Secondary data were collected from relevant sources such as the Amhara National Regional
State Finance and Economic Development Bureau, Agriculture and Rural Development
Bureau, Dejen district Agriculture and Rural Development Office, and other related bureaus,
offices and officials. The data were collected from reports, statistical documents as well as
published and unpublished documents.

3.2.1.2. Sampling procedure

Many researchers and experts agreed upon that, at least theoretically, there are two different
methods of sample selection. The first is probability sampling, which requires known non- zero
probabilities of selection (Pi > 0) for all elements (i = 1, 2, ---, N) listed in the specified frame
population. The second is non-probability sampling that fails to attach equal chance of being
included in the sample for each of the members of the population. Though, each method

32
carries some strengths and weakness of their own, the second method is blamed more as it
depends on subjective judgments. Hence, in this study, the former method was used to select
sample farm households.

Sample farm households were selected in such a way that first, 5 peasant associations (PAs)
were selected using random sampling method. The PAs were, Koncher Sasaberay, Borebor
Shencha, Werkamba Berch, Yetnora Tik and Kurar. Following the selection of PAs, 120
sample farm households were selected. Here again, proportional probability sampling
technique was employed to select the required number of farm households from each PA.
Rosters, which consist of recent lists of PAs and their members, were obtained from Dejen
district Finance Office. Extension agents and PA committee members were consulted to
include all farm household heads if there are some individuals who were not in the list and
exclude those whose names were in the list but were not in the area.

Prior to the final administration of the questionnaires, several steps were passed through. First,
enumerators were given classroom training and briefings on the objectives and contents of the
questionnaire and were also acquainted with the basic techniques of socio-economic data
gathering and interviewing techniques. Secondly, the questionnaires were tested at the farm
level on 15 randomly selected farm households. Thirdly, some amendments on the
questionnaire were made following the results of the pretest.

3.2.1.3. Sample size

Principally, the sample size required depends on the required precision, the variance of
variables among the total population, and the sampling technique. In practical terms, however,
the sample size is often restricted by the available fund, time and other related reasons. It is
often not feasible to study the entire population because of the following reasons; the physical
impossibility of checking all items in the population, the cost of studying all the items in the
population, the adequacy of sample results, to contact the whole population would be time
consuming and the destructive nature of certain tests (Lind et al., 1994). Therefore,
considering financ ial constraints, time shortages, lack of transportation and other infrastructure
facilities, 120 households were included in the model.

33
3.2.2. Methods of data analysis

Both descriptive statistics and econometric models were employed to study the relationship
between the dependent and explanatory variables. Using descriptive statistics the mean, range,
minimum as well as maximum values of variables were indicated. The result obtained is used
as an indicator of the relationship between explanatory variables and the dependent variable.

In addition, econometric models were used to study the relationship between variables
empirically. Thus, the binary logit model was used to analyze the willingness of farmers to
participate in water harvesting activities. The multinomial logit on the other hand, was used to
analyze the choice decision of households among alternative water harvesting storage
structure groups.

3.2.2.1. Binary logistic model

In participation decision studies, responses to a question such as whether farmers are willing
to participate in a given technology could be 'yes' or 'no', a typical case of dichotomous
variable. A variety of statistical models can be used to establish a relationship between the
household characteristics and the willingness for participation. Conventionally, linear
regression analysis is widely used in most economic and social investigations. This is because,
it has some desirable properties for specific type of enquiry and data and is widely available in
computer packages (Green, 1991). Moreover, it is easy to interpret and it is a reasonable
procedure even if some of the assumptions underlying it are not met in the data (ibid).
However, the same source further stated that while estimates derived from linear regression
analysis may be robust in the face of errors in some assumptions, other assumptions are
critical and their failure will lead to quite unreasonable estimates. To mention some weakness,
the linear probability Model (LPM) may generate predicted values outside the 0-1 intervals,
which violates the basic tenets of probability. The other problem with LPM is that the variance
of the disturbance term is heteroschedastic. Furthermore, the assumption of normality in the
disturbance term is no longer tenable.

34
The inadequacy of the linear probability model suggest that a non- linear specification may be
more appropriate and the candidate for this will be an S-shaped curve bounded in the interval
of 0 and 1 (Amemiya, 1981; Maddala, 1983). These authors suggested the S-shaped curves
satisfying the probability model as those represented by the cumulative logistic function (logit)
and cumulative normal distribution function (probit).

The choice between these two models revolves around practical concerns such as the
availability and flexibility of computer program, personal preference, experience and other
facilities. In fact it represents a close approximation to the cumulative normal distribution.
Hosmer and Lemshew (1989), pointed out that a logistic regression has got adva ntage over
others in the analysis of dichotomous outcome variables. There are two primary reasons for
choosing the logistic distribution. These are 1) from a mechanical point of view, it is an
extremely flexible and easily used function, and 2) it lends itself to a meaningful
interpretation. The logit model is simpler in estimation than the probit model. Therefore, a
binary logistic regression model was used to study the decision behavior of sampled
households (Pindyck and Rubinfeld, 1981).

Following Hosmer and Lemshew (1989), the logistic distribution function for identification of
the willing and non-willing farmers can be defined as:

1
Pi = .........................................................................................…....................1
1 + e− Ζi

Where Pi is the probability of being willing to participate for the ith farmer and Zi is a function
of m explanatory variables (Xi), and expressed as:

Zi = ßo + ß1X1 + ß2X2 +-----+ ßmX m .................................................….........2

35
Where ßo is the intercept and ßi are the slope parameters in the model. The slope tells how the
log-odds in favor of being willing to participate in water harvesting practices change as
independent variables change.

Since the conditional distribution of the outcome variable follows a binomial distribution with
a probability given by the conditional mean Pi, interpretation of the coefficient will be
understandable if the logistic model can be rewritten in terms of the odds and log of the odds,
(Gujarati, 1995). The odds to be used can be defined as the ratio of the probability that a
farmer will practice (Pi) to the probability that he/she will not (1-Pi).

But

1
1-Pi = ................................................................….........................3
1 + eΖi

Therefore,

Pi 1 + e Ζi
= = eZi ..........................................................................................4
1 − Pi 1 + e− Ζi

And

Pi 1 + e Ζi βo + ∑
i =1
βixi
= = e ...................….....................…................5
1 − Pi 1 + e− Ζi

Taking the natural logarithm of the odds ratio of equation (5) will result in what is known as
the logit model as indicated below:

 P  βo + ∑ βixi
ln  i  = ln e i =1 = Zi ...........…......................................................6
1 − Pi 

If the disturbance term Ui is taken into account the logit model becomes

m
Zi = ßo + ∑β x + u
i i i ..........................................................................7
i =1

36
Hence, the above econometric model was used in this part of the study to identify variables
that affect the willingness to participate in water harvesting practices.

3.2.2.2. Preference decision model (multinomial logistic regression)

In choosing water harvesting storage structure groups, there is no natural ordering in the
alternatives. In such conditions unordered choice models can be motivated by a random utility
model (Green, 2000). Therefore, a choice has to be made between multinomial probit and logit
models. A multinomial probit model is less restrictive than the multinomial logit model.
However, the multinomial probit model is gained at considerable computational expenses. It is
also difficult to get accurate estimates of the covariance matrix of the ßi (Judge et al., 1985).

Therefore, multinomial logit is used for studying choice of farmers among water harvesting
structure groups.

Following Green (2000), the multinomial logit was used to determine factors affecting
farmers' preference among alternative water harvesting storage structure groups. The model is
specified as follows:

e β oxi
P(Y=j) =
eβ 1xi +... + β jxi

Where Y = 0 No preference among the alternatives

Y = 1 prefer underground alternatives
Y = 2 prefer aboveground alternatives
Y = 3 prefer surface pond structure alternative

The estimated equations provide a set of probabilities for the j+1 choices for a decision maker.
Before proceeding, we must remove indeterminacy in the model. If we define ßj = ßj + q for
any vector q, then the identical set of probabilities result because the terms involving q will
drop out. A convenient normalization that solves the problem is to assume that ßo = 0 (Green,
2000).

37
 β jXj
e
Therefore, the probabilities are Prob(Y=j) = ; for j = 1…j
1+ ∑e
βjXj

1
Prob(Y=0) =
∑e β jXj

Before running the models, it would be necessary to check whether there is multicollinearity
among the candidate variables and verify the degree of association among discrete variables.
The reason is that the existence of multicollinearity will affect seriously the parameter
estimates. If multicollinearity turns out to be significant, the simultaneous presence of the two
variables reinforces the individual effect of the variables. However, omitting significantly
interacting terms incorrectly will lead to a specification bias. To this end, the coefficients of
the interaction of the variables indicate whether or not one of the two associated variables
should be eliminated from the analysis.

According to Gujarati (1995), there are various indicators of multicollinearity and no single
diagnostic will give us a complete handle over the collinearity problem. Of various indicators
of multicollinearity, the variance inflation factor (VIF) is used in this study to check whether
there is multicollinearity or not among the continuous explanatory variables. Where each
continuous explanatory variable is regressed on all the other continuous explanatory variables
and coefficients of determination for each axillary or subsidiary regression will be computed.
Moreover, Gujarati (1995), stated that a high R2 obtained could only be a surface indicator of
multicollinearity. Therefore, a measure of multicollinearity associated with the variance
inflation factors is defined as:

1
VIF (Xj) =
1 − R2
Where R2 is the coefficient of determination when the variable Xj is regressed on the other
explanatory variables. A VIF value greater than 10 is used as a signal for a strong
multicollinearity (Gujarati, 1995).

38
Likewise, there may also be interaction between two qualitative variables, which can lead to
the problem of multicollinearity or association. To detect this problem, coefficients of
contingency were computed from the survey data. The contingency coefficients are computed
as follows:

X2
C=
N + X2

Where:

C = coefficients of contingency,

X2 = Chi-square random variable,

and N = total sample size.

The parameters of the model were estimated using the iterative maximum likelihood
estimation procedure. This yields unbiased and asymptotically efficient and consistent
parameter estimates.

In reality, the significant explanatory variables do not all have the same level of impact on the
decision of farmers. The relatively important explanatory variables in decision can be
measured by examining variable elasticities defined as the percentage change in probabilities
that would result from a percentage change in the value of these variables. To compute the
elasticity, one needs to select a variable of interest, compute the associated Pi, vary the Xm of
interest by some small amount and recompute the Pi, and then measure the rate of change as
dPi/dXi, where dXi and dPi stand for percentage changes in Xi and Pi, respectively. When dXi
is very small, this rate of change is simply the derivative of Pi with respect to Xm and it is
expressed as follows (Aldrich and Nelson, 1990).

dPi exp( Zi ) 1
= * β̂ = P(1-P) β̂
dX i 1 + exp( Zi ) 1 + exp( Zi )

39
The impact of each significant explanatory variable on the probability of willingness is
calculated by keeping the continuous variables at their mean values and the dummy variables
at their most frequent values.

3.3. Definitions of Variables and Working Hypothesis

Once the analytical procedures and their requirements are known, it is necessary to identify
the potential explanatory variables and describe their measurements. Therefore, this part was
treated in two divisions.

3.3.1. Participation decision study:

In this division, identifying and defining dependent and explanatory variables for the
participation decision study was done based on theoretical frameworks and empirical findings.

3.3.1.1. Dependent variable for the participation decision model

The dependent variable for the binary logistic analysis has dichotomous nature representing
the preferred status of the farmer to participate in water harvesting practices. The variable
takes the value 1 if the farmer is willing to participate in rainwater harvesting and 0 otherwise.

It has been noted that one of the gauges of a community’s acceptance of a new idea is the
willingness to participate and cost share (Ngiggi, 2003). Thus, in this study a farmer is said to
be willing if he expresses his willingness to construct water harvesting structures by covering
at least half of the costs of the structure.

40
3.3.1.2. The independent variables

The independent variables of the participation decision study are those, which are expected
(hypothesized) to have association with the participation decision behavior of farmers in water
harvesting practices. Different researchers came up with different results as to what factors can
influence farmers' decision to participate or to have water harvesting structures. Some mention
the socio-economic factors as the leading. Sambrook and Akhter (2001) identified a marked
association between socio-economic characteristics of the household and the practice of
domestic rainwater harvesting. Others list out biophysical and climatic factors such as rainfall,
soil type etc. Critchley (1991), and others still consider a combination of socio-economic and
biophysical factors. Despite their differences in relative weighting of factors, most researchers
(FAO, 1994; Niggi, 2003) considered the socio-economic, cultural and technological
characteristics as the decisive factors for promotion and adoption of rainwater harvesting
practices across potential users. To mention some, household characteristics (age, education
level, farming experience, sex and so on), farm characteristics (labour availability, farm size,
etc), and institutional arrangements (land tenure, frequency of development agent contact,
availability and adequacy of credit facility, etc) are to be considered.

Based on theoretical background and empirical results of different studies on water harvesting
practices, and other related natural resources conservation decision carried out else where as
well as considering the information from the informal survey, the following variables are
hypothesized to influence farmers' decision to participate in water harvesting practices.

41
Table 1: Definition of explanatory variables to explain participation in water harvesting

Variables Code Definition

1 AGE Age of the head of the household since birth in years.

2. TTLU Total tropical livestock unit owned per household in numbers with values 0.1, 0.2, ...

3. LABORAVA Labour availability in adult equivalent per household with the value of all rational numbers 0.1, 0.2, ..

4. CULTAREA Total cultivated land size owned in hectares with the values of all rational numbers

5. DISTEXTC Time required in minutes to reach (by walking) the nearest extension center

6. DISTMARK Walking distance in time measures (minutes) to reach nearest market center

7. EDUC Education level of head of the household 0 = illiterate, 1 = 1-4 years formal education and read and write, 2 = 5 and

above years of formal education

8. RESPONSI If the household head has assumed any administrative responsibility in his PA, which assume values 1= yes and 0 = no

9. TRAINING Training and visit in different water harvesting practices (workshops, seminars, etc) which was 1= yes and 0 = no

10. FINANCON Financial constraints faced by the respondent 1= yes and 0 = no

11. IRRIGUSE Whether the respondent has owned irrigated plot 1= yes and 0 = no

12. FOODSHOR Whether the sample household has faced food shortage in 5 years time 1= yes 0 = no

13. ADQOLO The respondents access to adequate credit facility 1= yes and 0 = no

14. ATITECHN General attitude towards the importance of water harvesting technology 1= if considered important and 0 = otherwise

42
3.3.1.2.2. Hypothesis

3.3.1.2.2.1. Household characteristics

Age of the household head (AGE ): Age matters in any occupation. Older farmers are more
likely to reject new agricultural technologies. They tend to be too traditional and security
conscious to take the risk of adopting an innovation. On the contrary, younger farmers are
often expected to take risk due to their longer planning horizon. Therefore, adopters are
relatively younger and middle-aged farmers (Dasgupta, 1989). DTU (2002) has reported that
households headed by elderly people in Ethiopia and Sri Lanka have neither thought of nor are
interested to participate in domestic rainwater harvesting. Bekele and Holden (1998) reported
similar negative relationship between age and adoption of land conservation practices in the
Ethiopian highlands. Contrary to this, Lapar and Pandey (1999) has found positive
relationships between age and decision of farmers to adopt soil conservation practices.
However, in this study, it is hypothesized that increased age would have a negative impact on
the participation decision in water harvesting practices.

Responsibility (REPONSIB ): This variable defines whether the respondent has assumed any
type of administrative responsibility in his village or PA level during the time of the survey. If
the household heads assumed any type of responsibility, the chance of becoming accessible to
information may increase and hence become willing easily and quickly than those who do not
assumed any type of responsibility. Participating in formal and informal organizations within
and outside the village as members and office bearers facilitates willingness for new
agricultural technologies (Dasgupta, 1989). Therefore, this variable is expected to be
associated positively with the willingness of farmers to participate in water harvesting works.

Education level (EDUC): The variable refers to the number of years of formal education the
household head has attended. It is assumed to increase a farmer's ability to obtain, process, and
use information relevant to the uses of water harvesting practices. A value 0 was assigned if
the farmer was illiterate. The value 1 represented those respondents who have 1-4 years of
formal education. This group also comprises those sample household heads that learned in

43
religious schools and adult education programs. Similarly, the value 2 was assigned to
represent those farm household heads having 5 and above years of formal education.

Willing farmers tend to be literate or to have a higher level of formal education (Dasgupta,
1989). Empirical evidences indicate that the higher the level of education, the greater is the
possibility for farmers to become aware of the uses of water harvesting practices for securing
food self sufficiency (Sambrook and Akhter, 2001). Paulos (2002) has also identified strong
positive relationship between willingness to participate in soil conservation practices and
education level of the respondents. It is expected that those farmers with increased formal
education be disposed to decide to participate in water harvesting works. Thus, education level
is hypothesized to influence decision of farming households for participation positively and
significantly.

3.3.1.2.2.2. Economic factors

Total tropical livestock unit (TTLU): This variable represents the total number of livestock
in tropical livestock units owned per individual sample household. Animal rearing is one
component of the farming systems of the study area. Sample farmers rear livestock for
various purposes including meat, milk, eggs production, draft power and others. Most farmers
in the study area practice rearing livestock, which constituted cattle, small ruminants, and pack
animals. The number of livestock owned by a farmer was hypothesized to be positively related
to the willingness of a farmer to participate in water harvesting practices. Sambrook and
Akhter (2001) identified positive relationship between number of cattle owned and willingness
to participate in water harvesting works. Tesfaye et al. (2001) have obtained similar
relationship between tropical livestock units owned and fertilizer adoption decision. As a
farmer owns more livestock units, he requires large amount of water in his vicinity to provide
his livestock population with water. On the other hand, more livestock unit means more asset
and more asset possession leads to investment decision. Therefore, it is expected to affect the
willingness of farmers positively.

44
Total size of cultivated land (CULTAREA): It represents the total owned cultivated land by a
household. It is an indication for the wealth status of a household. Adopters of new
agricultural technologies have larger units of farm size holding (Rogers et al., 1971; Dasgupta,
1989). Bekele and Holden (1998) found significant and positive relationship between adoption
of land conservation technologies and per-capita cultivable land. Mulugeta (2000) reported
similar positive and significant relationship between size of cultivated area owned and
adoption of physical soil and water conservation practices. As land ownership is equated with
asset ownership, a farmer with large cultivable land is considered to be wealthy. In addition, a
farmer who owned a large size of cultivated land is expected to have enough land to adopt the
technology. Therefore, land size holding is expected to be associated positively with the
willingness of farmers to adopt the technology.

Labour availability (LABORAVA ): Water harvesting is a labour demanding practice.

Constructing water harvesting structures, watering the plants, protecting the water from
evaporation and the like practices all require labour. This variable represents availability of
labour in adult equivalent ratios. Hence, a farmer who has larger labour unit is expected to be
willing easily and quickly than those farmers who faced labour shortages. FAO (1994) and
Ngiggi (2003) have reported availability of labour as first criterion to participate in water
harvesting works. Thus, this variable is hypothesized to correlate positively with the
willingness of farmers to participate in water harvesting activities.

Financial constraints (FINANCON): In addition to its huge labor requirements, water

harvesting practices require a relatively large amount of money. It may be a requirement or
essential to buy local and industrial materials to construct water storage structures. For this,
availability of cash at hand is a necessity. The variable takes a value 1 if the household head
faced shortage of cash during the past cropping season and 0 if otherwise. Shortage of money
may discourage a farmer from practicing water harvesting activities. DTU (2002) confirmed
that those Ethiopian farmers who were not interested to participate in domestic rainwater
harvesting, has reported financial shortages as the main cause for their decisions. Hence, this
particular variable is expected to correlate negatively with the willingness of farmers to
practice water harvesting.

45
Shortage of food (FOODSHOR ): A household faced with food shortage may not have cash at
his hand to adopt water harvesting activities. However, in any possible alternative he has to
overcome the food insecurity problem. One means may be participating in water harvesting
practices. To adopt water harvesting practices on the other hand, the farmer may face cash
shortages. He may solve this by taking loan or with assistance from the government. Ngiggi
(2003) mentioned water harvesting as the main means to alleviate poverty and food shortages
in the long run. Similarly, FAO (2000) pointed out that water harvesting is a proven
technology to increase food security in drought prone areas. This variable is hypothesized to
correlate positively with the willingness of farmers to participate in water harvesting works.

3.3.1.2.2.3. Institutional factors

Distance to nearest development center (DISTEXTC): This variable refers to the time a
farmer may need to walk to get his extension agent. The farther an extension office is located
from farmers' homes, the less likely it is that farmers will have access to information. Earlier
adopters have more change agent contact than later adopters (Rogers and Shomaker., 1971;
Dasgupta, 1989). Chilot (1994) has found significant negative relationships between distance
to an extension office from homestead and adoption of wheat technologies. As a farmer is
nearby his development agent, there may be a chance to have more contact hours than the
farmer who live at a farther distance. More contact may help a household to be convinced by
the development agent to participate in water harvesting works. Therefore, distance from the
nearest development center in this context is expected to be negatively correlated with the
willingness of farmers to participate in water harvesting works.

Distance to market (DISTMARK): This variable represents the walking time required to
reach the nearest market center, which can be expressed in minutes. This variable has key
importance in decision making to participate in water harvesting works. The longer the
number of walking hours required, the less probable for being willing to participate in water
harvesting activities. This hypothesis was formulated due to the fact that vegetables and fruits,
which will be produced using harvested water, are perishable and can be rotten before

46
reaching market places and final consumers. Lapar and Pandey (1999) came up with a
negative relationship between adoption decision of farmers for soil conservation practices and
distance to market center. Therefore, this variable is hypothesized to affect the willingness of
farmers negatively.

Access to credit (ADEQOLO ): It is a dummy variable taking a value 1 if the farmer

responded as he has access to credit and 0 otherwise. Access to credit for agricultural purposes
can relax farmers' financial constraints and expected to make farm households willing to
participate in water harvesting activities. In the study area, the Amhara Credit and Savings
Institute (ACSI) is the main source of credit for farm households. In this study, access to
credit was expected to increase the probability of being willing to participate in water
harvesting practices. Sambrook and Akhter (2001) reported that those farmers who enjoy
credit facility were willing to participate in water harvesting works. Tesfaye et al. (2001) have
found a positive relationship between access to credit and wheat adoption decision. Therefore,
this variable is anticipated to have positive relations with the decision of farmers to participate
in water harvesting activities.

3.3.1.2.2.4. Attitude and training measures

Training and visiting (TRAINING): Training and visit in areas of rainwater harvesting are
preconditions for sustained efforts in water harvesting sector (Ngiggi, 2003). If we want a
steadily increasing number of farmers to undertake water harvesting, the implementation of
substantial training program should get a high priority (Reij et al., 1993). A typical training
course is likely to include demonstration, practice survey, visits to fields where water
harvesting structur es have previously been built and used, discussion and field visits,
participating in planning and so on.

Training and visiting will enable farmers to increase their knowledge and improve their skills.
Sambrook and Akhter (2001) found a strong positive relationship between training users in
different water harvesting matters and willingness to participate in water harvesting activities.
Tesfaye et al. (2001) has arrived at similar results in wheat adoption study in Amhara region.

47
Therefore, training and visiting in different water harvesting activities are expected to be
correlated positively and significantly with the willingness of a farmer to participate in water
harvesting activities.

Irrigation (IRRIGUSE): This represents whether or not a household owns irrigated plot/s at
the time of survey. A farmer who was practicing irrigation on his own plot may not be
interested to have additional. This was practically seen during the time of the informal survey.
The reason was that farmers have limited land size and other resources to construct another
irrigation structure. Of course, the positive contribution or influence of irrigation experience
for willingness should not be denied. However, the influence of experience is expected to be
minimum when compared with the above mentioned problems. Thus, this variable is
hypothesized to correlate negatively with the willingness of farmers to participate.

Attitude towards the importance of water harvesting technology (ATITECHN): Another

measure of the attitude of a given respondent to adopt a technology is his perception about the
importance of water harvesting technology. The existence of a favorable attitude towards new
agricultural technologies facilitates adoption (Yapa and Mayfield, 1978). FAO (1994)
considered technological appropriateness as a key determinant factor for the adoption and
promotion of water harvesting practices across potential users. Bekele and Holden (1998)
found positive relationship between attitude towards new land conservation technologies and
adoption by users. Sometimes, water harvesting technologies are blamed to be beyond
farmers' know- how and ability to construct, maintain and manage the system. The users will
not easily adopt any technology that is too complex. Therefore, farmers' attitude towards the
uses and convenience of water harvesting technology is expected to be related positively to the
willingness of farmers to accept the technology.

3.3.2. Study for preference among water storage structure groups

Consumers make choices so as to maximize their utility, that is, to make them as happy as
possible. Thus, preferences were defined in terms of utility: to say a bundle (x1 , x2 ) was
preferred to a bundle (y1 , y2 ) meant that the x-bundle had a higher utility than the y-bundle.

48
The preferences of the consumers are the fundamental description useful for analyzing choice
and utility is simply a way of describing preferences.

Thus, farmers are expected to choose that water storage technology group which can generate
a higher benefit or utility among the available three water storage technology groups. The
choice decision may depend on the available budget and other characteristics of the
technology as well as the plot. Therefore, analysis of the choice decision among potential
water storage technology groups was made at a plot level. This was done because decision
behavior of users varies depending on the plot character such as soil type, slope, distance, plot
area and other related plot characters. As done in the participation decision sub part, in this sub
part too, definition of variables and formulation of hypothesis is performed.

3.3.2.1. Dependent variable

The dependent variable for this analysis was structure groups among which users decided to
construct. Hence, a value 1 represents underground structure groups (Dom shape etc). A value
2 was assigned to represent aboveground structure groups (Ferro cemented, etc). Similarly, the
value 3 represented surface pond structure such as geo- membrane lined, trapezoidal pond and
other pond types.

49
 3.3.2.2. Independent variables for the choice model

3.3.2.2.1. Variable definitions for the multinomial logit model

Table 2: Definition of variables for multinomial logit

Variables Code Definition

1. TTLU As defined in table 1

2. LABORAVA As defined in table 1

3. PLOAREA Area of a specific plot in hectares with values of all rational numbers

4. DISTPLO Distance of the plot from homestead in kms

5. FERTILIT Perceived fertility status of the specific plot taking values 1= fertile 2 = medium 3 = low fertile

6. OWNPLO Ownership of the plot 1= owned 0 = hired

7. EDUC As defined in table 1

8. FINANSTA As defined in table 1

9. SOILTYP Soil type of the plot 1= black 2 = gray 3 = red

10. SLOP Slope of the plot 1= flat 2 = medium 3 = steep

50
3.3.2.2.2. Hypotheses

Education level (EDUC): Education level of the head of the household help farmers to easily
get, analyze, and interpret information (Dasgupta, 1989) about water harvesting storage
structure groups based on different characteristics of each group. In addition to their
indigenous knowledge, this further assist educated farmers to choose the technology groups,
which are most appropriate to their socio-economic and other household characteristics. As
mentioned in the literatures review section, water storage groups have their own pros and
cones. Some are convenient; others can be built at a relatively lower cost. Thus, educated
farmers can compare the pros and cones of each group and can choose one based on some
criteria. Therefore, it is difficult to determine a priori the impacts of education on choice
decision for water storage technology groups.

Area (PLOAREA): This variable refers to the area of a particular plot, which was selected for
constructing water storage structures by a sample household. As mentioned before, the
analysis of the choice decision between technology groups is made at a plot level. Rees (2000)
mentioned aboveground and surface pond structure group s are generally demanding a large
area for their construction. Therefore, this variable may serve as an input to choose between
water storage technology groups. As aboveground and surface pond structure groups require
relatively large area, the area of the plot on which water storage structure has to be built is
expected to relate positively with these structures. However, plot area is expected to correlate
negatively with the underground type of structure group.

Distance of the plot from residential hous e (DISTPLO): Distance of a plot from homestead
is expected to highly affect the technology types to be selected. Those technology groups,
which need day to day follow up, should necessarily be built near home (Martinson et al.,
2001). In addition, aboveground structures are usually built near residence to collect water
from roofs. Hence, distance matters for choosing water storage structures. Therefore, it is
expected that distance of the plot may relate negatively with aboveground and underground
structure groups and positively with surface pond type structure groups.

51
Ownership of the plot (OWNPLO): Ownership of the plot may affect the choice decision of
farmers for alternative water storage technology groups. Rees (2000) remarked that one of the
pros of aboveground structures is that it can be constructed from lighter materials so that they
can be carried from place to place. Those farmers who hire others land may need to construct
easily movable aboveground structure groups. On the other hand, farmers working on their
plot are expected to construct stable and long lasting structures. These may be underground
and surface pond structure groups. Thus, ownership of the plot is expected to have positive
impact on the choice decision of farmers for underground and surface pond structure groups.

Fertility of the plot (FERTILIT): This variable refers to the fertility status of a particular plot.
It assumes a value 1 if the plot is highly fertile, a value 2 represents medium fertility status and
a value 3 was assigned to represent low fertility level of the plot. Most underground structure
groups are usually built for the purpose of crop production but aboveground and surface pond
structure groups can in addition serve for storing water for drinking purpose for human beings
and animals respectively. However, this does not mean that underground structure groups can
never serve the purpose of storing water for drinking purposes.

Thus, fertility status of a plot is expected to correlate negatively with the underground
structure groups and positively with the other two groups.

Total Tropical Livestock Unit (TTLU): As a household owns large livestock unit, it may be
interested to invest in convenient but costly projects, such as aboveground structure groups.
Martinson et al. (2001) categorize aboveground structures as generally convenient and costly
than underground and surface pond structure groups. Therefore, livestock ownership is
expected to positively affect the choice decision of the aboveground technology groups and
negatively correlated with the underground and surface pond technology groups.

Labour Availability (LABORAVA ): Water harvesting is a labour demanding practice.

Constructing water storage structures, watering the plants, protecting the water from
evaporation and the like practices require labour. This variable represents availability of
labour in adult equivalent ratios. Hence, a farmer who has large labour units is expected to

52
prefer those technology groups, which require larger amount of labour, such as surface pond
and underground technology groups (Martinson et al., 2001). Therefore, it is hypothesized that
availability of labour may be associated positively with the choice decision of farmers for
underground and surface pond structure groups and negatively with aboveground technology
groups.

Financial constraints of the household (FINANCON ): In addition to its larger labor

requirements, water harvesting practices require huge amount of money. Some technology
groups require relatively huge amount of financial outlay than others, with a typical example
being aboveground structure groups (Rees, 2000; Martinson et al., 2001). Hence, shortage of
money is anticipated to go positively related to the choice decision for underground and
surface pond water storage technology groups. However, a farmer with critical shortage of
money may not desire to decide to construct costly technology groups such as aboveground
technology groups. Thus, shortage of money is expected to affect negatively the choice
decision for aboveground technology groups. It is expected to affect the choice decision for
underground and surface pond structure groups positively.

Slope (SLOP): Water harvesting is not recommended for areas where slopes are greater than
5% due to uneven distribution of runoff and large quantities of earthwork required which is
not economical (FAO, 1994). Hence, the steepness or flatness of a plot may affect the water
storage types to be selected. Due to the aforementioned reasons, flat type of topography is
suitable to build underground and surface pond structures, while there is no so much
consideration about the slopes to construct aboveground structures. Thus, slope can be taken
as a variable to conduct empirical studies about the decision of farmers to choose among water
storage technology groups. The value 1 represents flat topography, the value 2 represents
medium and 3 was assigned to represent very steep topography. Hence, slope is hypothesized
to influence the choice decision for underground and surface pond technology groups
negatively. However, it is difficult to decide a priori the relationship between the slope of the
plot and the choice decision for aboveground structure groups.

53
Soil type (SOILTYP ): For this particular variable, the value 1was assigned to represent black
Soil category, while 2 and 3 represent gray and red soil categories, respectively. The type of
soil is expected to affect the type of structure to be selected for construction in a particular
area, DTU (2000). Surface pond structures are usually constructed in black soils. On the other
hand, red type of soil cannot be easily ruptured (Brady and Well, 2002) so that it is more
suitable to construct underground water harvesting structure types. Hence, the type of soil is
expected to affect negatively the choice decisions for surface pond and aboveground structure
groups. It was hypothesized to positively affect the choice decision for underground structure
groups.

54
 4. RESULTS AND DISCUSSION

This part is mainly concerned with the description and interpretation of the findings. As
already noted, a structured questionnaire was administered to 120 sample households in Dejen
district. The main aim was investigating determinants of farmers' response and willingness to
participate in water harvesting practices and their preferences among water storage technology
groups. The questionnaire was designed in such a way that it enables to collect data on
personal and socio-economic characteristics of farm households as well as on the farmers'
willingness to participate in rainwater harvesting practices.

4.1. Results and Discussion on Participation Decision

In this section the results of both descriptive statistics and econometric models for
participation decision are discussed in detail.

4.1.1. Results of descriptive statistics

Influential factors affecting willingness of farmers to participate in water harvesting works are
described in this subdivision. The description was made using mean, minimum as well as
maximum values, range and standard deviations. In addition, mean difference for continuous
variables and frequencies of discrete variables were tested using T-test and chi-square test
respectively.

4.1.1.1. Household characteristics

Out of the total sample farmers the maximum and minimum ages registered were 70 and 22
years respectively, with a standard deviation of 11.12 years. The average age of the sample
farmers was 42.23 years and the average age figure is 42.46 and 41.92 years for the willing
and non-willing farmers respectively.

From the total sample household heads, 88.3% were able to read and write. To put this in
grade level groups, 57 farm household heads were between 1-4 years of formal education.

55
This group included those sample farmers who get education from regular schools, adult
education programs, and from religious institutions. The survey result shows that in general,
92.6% of the willing farmers were literate, and the remaining 7.4% were illiterate. On the
other hand, relatively a large number of the non-willing farmers (17.3%) were illiterate. This
shows that as a farmer becomes literate the probability for being willing increases.

Table 3: Level of education of sample farmers

Willing Non-willing Total

Attributes Number % Number % Number %

5 and above years of formal 6 8.8 3 5.77 9 7.5

education

1-4 years of formal education 57 83.8 40 76.92 97 80.8

Unable to read and write 5 7.4 9 17.31 14 11.7

Source: results of descriptive statistics

Of the total sample household heads, 64 or 53.3% were assumed some responsibility at their
village or PA level. The figure was 54.4% and 51.1% for the willing and non-willing farmers
respectively. The higher figure for the willing respondents when compared with the non
willing may indicate that as the head of the household assumed some responsibility, the
chance of getting information and hence understanding about the uses of water harvesting
increases. This contributes to decide to construct some form of water harvesting.

56
Table 4: Responsibilities of head of the household

Willing Non-willing Total

Attributes Number % Number % Number %

Yes 37 54.4 27 51.9 64 53.3

No 31 45.6 25 48.1 56 46.7

Source: results of descriptive statistics

4.1.1.2. Economic situations

The land size holding of the sample farmers ranges from 0.375 to 4.125 hectares. The average
land holding is known to be 1.82105 hectares with a standard deviation of 0.945225 hectares.
The survey result indicated that about 28% of the respondents had a farm size of 1 hectare or
less. Willing farmers owned on the average 1.98735 hectares of cultivated land. The
corresponding figure for the non-willing farmers was 1.6036 hectares. About 72.5% of the
respondents cultivated others land through short term rent or through share cropping
arrangements. The figure will be 74.3 and 71.4 percent for the willing and non-willing
farmers, respectively.

On average, the sample households kept about 3.83 total tropical livestock unit: the minimum
and maximum livestock kept being 0.130 and 8.72 tropical livestock units respectively with a
standard deviation of 1.96. Willing farmers own on average 4.02 livestock units and the non-
willing have reported average 3.58 livestock units per household. The larger average tropical
livestock unit owned in the willing farming household indicated that larger livestock
ownership leads farmers to decide for participation in water harvesting activities. The sample
survey indicated that 60% of the respondents did not face shortage of oxen for their farming
practices. The figure was reported as 57.7 percent for the willing and 63% for the non-willing
farming households.

57
 The mean available labour in adult equivalent was 3.01 with the maximum and minimum
being 1 and 6 respectively. Willing farmers have relatively larger labour units in adult
equivalent when compared with non-willing sample households which may indicate that
labour availability is a key component to be considered for participation decision. This means
that to undertake water harvesting works households need to have sufficient labor availability.
Shortage of labor supply may lead a household not to participate in water harvesting practices.

Shortage of money may discourage farmers from participating in newly released agricultural
technologies. Accordingly, 38.2% of the willing and 84.6% of the non-willing farmers faced
money shortages during the past cropping season.

Table 5: Money shortages

Willing Non-willing Total

Attributes Number % Number % Number %

Reported shortage of money 26 38.2 44 84.6 70 58.3

Reported as no shortage of many 42 61.8 8 15.4 50 41.7

Source: results of descriptive statistics

4.1.1.3. Farmers' institutional environment

Farmers’ institutional environment has important bearing on the preferred status of the farmers
with respect to willingness to participate in water harvesting technologies. The important
institutional concerns, considered in this study, are credit facility, market accessibility, and
agricultural extension (distance to the nearest development center).

The minimum and maximum time required to arrive at the nearest development center were 1
minute and 2 hours respectively. The average time required by samp le households is about 18
minutes with a standard deviation of 22.1 minutes. On average, willing farmers walk for 14.2

58
minutes while the figure was 20.9 minutes for the non-willing farmers. This indicates that
distance from development center may matter to make decision for participation.

Similarly, the maximum time required to arrive at the nearest marketing center is 3 hours
while the minimum time was found to be only 10 minutes with a standard deviation of 39
minutes. On average, sample households were required to walk for 58.04 minutes to arrive at
the nearest local market center. Similarly, willing and non-willing farmers walk for 50.51 and
67.88 minutes respectively.

In this study, access to credit was found to affect the probability of being willing to participate
in water harvesting practices. The study found that about 34.2% of the respondents have faced
problems in getting adequate loan facilities. 28.8% of the non-willing and 38.2% of the willing
farmers suffered the same problem. This result indicated that shortage of money might
discourage users from participating in water harvesting activities.

Table 6: Access to credit

Willing Non-willing Total

Attributes Number % Number % Number %

I can get adequate lo an 42 61.8 37 71.2 79 65.8

I cannot get adequate loan 26 38.2 15 28.8 41 34.2

Source: results of descriptive statistics

4.1.1.4. Training, visiting and perception measures

Training users and visits to create awareness are preconditions for they to make decisions to
participate in water harvesting sector. As indicated in Table 7 below, 67 sample household
heads or 55.83% have participated in training and other water harvesting related matters. The
figure is 85.3% and 17.31% for the willing and non-willing farming households, respectively.
The gap between the two groups is too large which implies that training and participating

59
users in water harvesting related matters is an important factor to promote the technology
quickly.

Table 7: Participation and training

Willing Non-willing Total

Attributes Number % Number % Number %

Participated in WH matters 58 85.3 9 17.31 67 55.83

Not participated 10 14.7 43 82.69 53 44.17

Source: results of descriptive statistics

The attitude of users towards the importance of water harvesting practice is another factor,
which was considered to explain the willingness of farmers to participate in water harvesting
practices. Accordingly, 52.5% from the total respondents have responded that water harvesting
technologies are not as such complex to implement. They reported water harvesting practice as
economically important. As can be seen from Table 8, 80.9% of the willing and 15.4% of the
non-willing farmers shared the same idea. This variable was seen as having very large
difference between the willing and non willing sample household groups. This indicates that
those farmers who considered the available technology groups as simple as they can
implement and maintain with their resources and ability are quick to adopt the technology.

Table 8: The importance of water harvesting technology

Willing Non-willing Total

Attributes Number % Number % Number %

Water harvesting is important 55 80.9 8 15.4 63 52.5

WH is not as such important 13 19.1 44 84.6 57 47.5

Source: results of descriptive statistics

60
Few farmers in the study area, practice small scale irrigation schemes. Of all the total
respondents, 27.5% have reported that they have constructed small scale irrigation of their
own. The figure is slightly smaller for the willing compared to the non-willing farm
households indicating that those farmers who have none before are more interested to have in
the future.

Table 9: Irrigation use.

Willing Non-willing Total

Attributes Number % Number % Number %

Practiced irrigation before 18 26.5 15 28.8 33 27.5

Not practiced irrigation before 50 73.5 37 71.2 87 72.5

Source: results of descriptive statistics

Considering all sample households 60% responded that they have faced food shortages in the
past 5 years. The figure is reported to be 66.2% for the willing and 51.9% for the non-willing
farming households.

Table 10: Food availability status of sample households.

Willing Non-willing Total

Attributes Number % Number % Number %

Yes 45 66.2 27 51.9 72 60

No 23 33.8 25 48.1 48 40

Source: results of descriptive statistics

61
4.1.1.5. Tests of the mean and frequency differences of variables

The mean values of the continuous variables in both willing and non-willing groups were
compared using t-test. The test is used to indicate the mean difference between groups. That is
why the test was used to identify the mean difference between willing and non-willing
respondents. The t-values of 6 continuous variables were computed and out of these variables
the two groups were found to differ significantly in five of them.

Table 11: T-test for mean difference of continuous variables

Continuous Variable Willing Non-willing Total T-value

Mean Mean Mean St.D.

AGE 42.46 41.92 42.23 11.12 .259

LABORAVA 3.51 3 3.28 3.82 2.18*
TTLU 4.02 3.58 3.83 1.96 2.77*
DISTEXTC (in minute) 14.2 20.9 18 22.18 1.855*
DISTMARK (in minute) 50.51 67.88 58.04 39 2.487*
CULTAREA 1.987 1.604 1.821 .945 2.24*

Source: results of T-test

* Significant at 5% probability level

Accordingly, the mean difference of the variables labour availability (LABORAVA), total
tropical livestock unit (TTLU), distance from extension center (DISTEXTC ), distance from
nearest market center (DISTMARK) and total cultivated area owned (CULTAREA) are found to
be significant at 5% probability level.

Indeed, the two groups ma y not only differ in terms of quantitative variables, but also in terms
of qualitative variables. In this respect, a chi-square test was used to examine the existence of
statistically significant differences between the two groups. Accordingly, 7 discrete variables

62
were considered and the two groups were found to be different in terms of 3 of the 7 variables
(Table 12). More specifically, the chi-square test reveals that 3 discrete variables showed
statistically significant differences between the two groups at 1% probability level.

Table 12: Chi-square test for frequency difference

Dichotomous Sc Chi-
variables ore Willing Non-willing Total square

Number % Number % Number %

0 6 8.8 3 5.77 9 7.5

EDUC 1 57 83.8 40 76.92 97 80.8 2.833
2 5 7.4 9 17.31 14 11.7

1 37 54.4 27 51.9 64 53.3

RESPONSI 0 31 45.6 25 48.1 56 46.7 .073

1 58 85.3 9 17.31 67 55.83

TRAINING 0 10 14.7 43 82.69 53 44.17 58.19**

1 26 38.2 44 84.6 70 58.3

FINANCON 0 42 61.8 8 15.4 50 41.7 30.04**

1 18 26.5 15 28.8 33 27.5

IRRIGUSE 0 50 73.5 37 71.2 87 72.5 .083

1 45 66.2 27 51.9 72 60
FOODSHOR 0 23 33.8 25 48.1 48 40 2.494

1 42 61.8 37 71.2 79 65.8

ADQOLO 0 26 38.2 15 28.8 41 34.2 1.155

1 55 80.9 8 15.4 63 52.5

ATITECHN 0 13 19.1 44 84.6 57 47.5 33.75**

Source: results of Chi-square test

**Significant at 1% probability level

63
The significant variables are training (TRINING), financial constraint (FINANCON) and
attitude towards water harvesting techno logy (ATITECHN) level.

4.1.1.6. Household's willingness to participate (WTP)

As it has been mentioned one of the gauges of a community’s acceptance of a new idea is the
willingness to participate and cost share. It was encouraging to note that in all the situations
where the idea of cost sharing was introduced for the rainwater harvesting technologies, it was
readily accepted provided the part to be contributed by the community members was within
their reach (Ngiggi, 2003).

Thus, the meaning of willingness to participate in this study carries some commitment. Hence,
the questionnaire was prepared in three different stages. The first type of question puts the
costs totally at the shoulder of the respondent. The question informed the respondents that no
financial supports are to be expected from the government or other donor agencies. It tells the
farmer that he and only he himself should cover all types of costs. Accordingly, 15 % of the
respondents were willing to have some structures. The second type of question demands the
respondents to cover half of the costs to construct and run the water harvesting structures.
Here, 68 farmers or 56.67 percent of the sample households were willing to have some water
harvesting structures. The third question inquires the farmer to respond whether he is willing
or not to have some water harvesting structures irrespective of the issue of costs. The content
of the question in here leads the respondent to show his willingness to construct any type of
water harvesting structure in his plot, though the costs were to be covered by government or
other NGOs. Accordingly, 89.2 percent of the sample farmers showed their desire to have
some type of structure.

Therefore, in the context of this study, a respondent is said to be willing if he/she falls in
categories two or three. They are considered as real demanders of the technology. This three
stage questioning was believed to reveal the willingness to pay for the technology. Those who
are willing to participate are also willing to pay for the technology.

64
Adding the number of farmers who fall under groups two and three yield the total number of
willing respondents. Hence, 68 farmers or 56.67% of the total respondents were considered to
be the willing farmers. On the other hand, 52 farmers or 43.33 % of the respondents were
considered to be non-willing farmers. Those farmers who desire but not willing to pay were
asked, as to why they are not willing to contribute at least half of the costs of the technology,
which they desire. Most responded that they are unable to afford at least half of the costs of
the technology. According to this group of respondents, the costs of the technologies, which
are being practiced, were beyond their ability to pay. They also pointed out that they have
critical shortages of labour to contribute. Still there are other respondents within this group
who demands only governments or other NGOs contributions for the total costs of the
structures.

However, fewer in number, there are farmers (13) or 10.8% of the total respondents who
neither desire nor are willing to participate. Most of them have no information about the uses
of water harvesting structures.

4.1.2. Econometric results for the binary logistic regression model

In the preceding parts of this thesis the descriptive analysis of important explanatory variables
that were expected to have impact on the decision of a given farmer to participate in water
harvesting works were presented. In this section, the selected explanatory variables were used
to estimate the binary logistic regression model to analyze the determinants of household's
willingness to participate. A binary logistic regression model was fitted to estimate the effect
of hypothesized explanatory variables on the probabilities of being willing or not. SPSS for
windows was used for the econometric analysis.

Prior to the estimation of the model parameters, it is crucial to look into the problem of
multicollinearity or association among the potential candidate variables. To this end, the
variance inflation factor (VIF) was used to test the degree of multicollinearity among the
continuous variables.

65
 The values of the VIF for continuous variables were found to be small (i.e VIF values less
than 10). To avoid serious problems of multicollinearity, it is quite essential to omit the
variable with value 10 and more from the logit analysis. The data have no serious problem of
multicollinearity (Appendix Table 3). As a result, all the 6 continuous explanatory variables
were retained and entered into the binary logistics analysis.

Similarly, the contingency coefficients, which measure the association between various
discrete variables based on the Chi-square, were computed in order to check the degree of
association among the discrete variables. The values of contingency coefficients ranges
between 0 and 1, with zero indicating no association between the variables and the values
close to 1 indicating a high degree of association. Accordingly, the results of the computation
reveal that there was no serious problem of association among the discrete explanatory
variables (Appendix Table 4).

Eventually, a set of 14 explanatory variables (6 continuous and 8 discrete) were included in

the model and used in the logistic analysis. These variables were selected based on theoretical
explanations and the results of various empirical studies. To determine the best subset of
explanatory variables that are good predictors of the dependent variable, the logistic regression
was estimated using enter method of Maximum Likelihood Estimation, which is available in
statistical software program (in this case SPSS version 11). In this method, all the above
mentioned variables were entered in a single step. For estimation of the logistic regression
model, some of the exp lanatory variables that are expected to improve the model fitness were
selected and included in the model analysis. The result of the analysis is presented in Table 13
below.

66
Table 13: Parameter Estimates for binary logit

Variable B Odds ratio Wald Sig.

AGE 0.078 1.081 .951 0.329

EDUC 2.669 14.421 4.435 0.035**
TTLU 0.531 1.701 4.095 0.043**
TRAINING 2.797 16.403 15.887 0.000***
FINANCON -2.691 0.068 8.759 0.003***
IRRIGUSE -0.712 0.491 0.745 0.388
CULTAREA 0.123 1.131 1.234 0.267
DISTEXTC -0.030 0.970 3.115 0.078*
ADEQOLON 0.013 1.013 0.000 0.987
ATITECHN 2.837 17.064 12.595 0.000***
LABORAVA 0.814 2.257 6.057 0.014**
DISTMARK -0.005 0.995 0.183 0.669
RESPONSI 0.037 1.038 0.002 0.963
FOODSHOR 0.451 1.570 0.316 0.574
Constant -5.778 0.003 3.185 0.074

-2 log Likelihood Ratio 53.412

Pearsons Chi-Square (X2) 110.804
a
Correctly predicted (count R2) 91
b
Sensitivity 91
c
Specificity 92
***, ** and * shows significance at 1%, 5% and 10% probability levels, respectively
a
Based on a 50-50 probability classification scheme
b
Correctly predicted willing groups based on a 50-50 probability classification scheme
c
Correctly predicted willing groups based on a 50-50 probability classification scheme

Source: results of binary logit analysis

A Variable(s) entered on step 1: AGE, EDUC, FARMEXPE, TTLU, TRAINING, FINANCON,

IRRIGUSE, CULTAREA, DISTEXTC, ADEQOLON, ATITECHN, FAMILSIZ, LABORAVA, DISTMARK,
REPONSI, FOODSHOR.

67
The various goodness of fit measures validate that the model fits the data well. The likelihood
ratio test statistics exceeds the Chi-square critical value at less than 1% probability level
indicating the hypothesis that all coefficients except the intercept are equal to zero is rejected.
The value of Pearson Chi-square test shows the overall goodness of fit of the model at less
than 1% probability level.

Other summary statistics for the goodness of fit, which are not based directly on the distance
between the observed and fitted values, are the various measures of classification accuracy.
An intuitively appealing way to summarize the results of a fitted logistic model is via a
classification Table. This classification is the result of cross-classifying the outcome variable,
y, with a dichotomous variable whose values are derived from the estimated logistic
probabilities. In this approach, estimated probabilities are used to predict group membership.
Presumably, if the model predicts group membership accurately according to some criterion,
then this is thought to provide evidence that the model fits. The model results show that the
logistic regression model correctly predicted 91% of the total sample farmers: 91% willing
and 92% non-willing.

Among the 14 explanatory variables included in the analysis, 3 variables were found to have a
significant impact on determining the willingness of a given farmer with less than 5%
probability level. These variables include education level of head of the household (EDUC),
labor availability (LABORAVA) and total tropical livestock unit owned (TTLU). Three
variables were considered significant at less than 1% probability level. These variables were
training and visit of the head of the household in different water harvesting matters
(TRAINING), financial constraints of the household (FINANCON), and general attitude towards
the importance of water harvesting technology (ATITECHN). Distance to extension center
(DISTEXTC) affects the willingness at 10% probability levels. The other 8 variables were not
significant at less than 10% probability level.

Education level of the head of the household (EDUC): The variable is significant at (P<
0.05) and has a positive association with the willingness of a respondent to participate in water
harvesting works. This variable as hypothesized affects the willingness in such a way that as

68
the education level of household heads get higher, they can easily obtain, process and use
information relevant to the uses of water harvesting practices. The positive effect of this
variable indicates the importance of education in influencing farmers’ decision to participate.
This result is consistent with the ideas stated in Dasgupta (1989), Roger and Shomaker (1971)
and Sambrook and Akhter (2001).

The odds ratio tells us that keeping the influences of other factors constant, as the household
head becomes literate, the decision to be willing increases by a factor of 14.42.

Total Tropical livestock unit owned (TTLU): It is found significant at (P< 0.05) and affects
the decision to participate positively. This means that as farmers own large livestock units, the
chance of becoming willing to participate gets increasing. This is due to two main reasons.
The first is that more livestock ownership is considered as more asset possession, which in
turn leads to investment decision. The second is that as farmers own large livestock
population, they need to have water available at their vicinity to provide their livestock
population with drinking water. Dasgupta (1989) and Rogers and Shomaker (1971) reported
similar result.

The odds ratio of 1.701 for tropical livestock unit indicates that with the assumption of ceteris
paribus, the probability of being willing increases by a factor of 1.701 as livestock ownership
increased by 1 tropical livestock unit.

Labour availability (LABORAVA ): It was found significant at (P<0.05) and positively

related with the willingness of farmers indicating that an increase in labour allows a farmer to
achieve large labour force to participate in water harvesting works. This result was consistent
with many other research results, which were conducted, in earlier times as well as agrees with
the ideas mentioned in the hypothesis part of this thesis. Niggi (2003) and FAO (1994)
mentioned availability of labour as an important element for the promotion of water harvesting
works.

69
With the assumption of constant influences of other factors, the likelihood ratio indicates that
the probability of being willing increases by a factor of 2.257 as labour availability increases
by one adult equivalent unit.

Training and Visit (TRAINING): As mentioned in the variable definitions part, training in
the context of this study includes participating in water harvesting training program, video
show related to water harvesting, participating farmers' field day, visits of water harvesting
works in another PA and village, participation in meetings and other water harvesting related
issues which were conducted in his/her PA or village. The model result indicates that it affects
the decisions of farmers to participate in water harvesting practices positively and very
significantly (P< 0.01). This means that as the head of the household participate in training
and other related water harvesting programs, he may become aware and can acquire sufficient
knowledge about the uses of water harvesting. This may lead the farmer to decide for
participation. The result obtained in this study coincides with other research results obtained in
this particular subject. For example, (FAO, 1994; Sambrook and Akhter, 2001) confirmed that
awareness and sensitization to be highly and positively correlated with the willingness of users
to participate. Concerning this variable, there is no economic theory or empirical result that
stated in contrary to the result obtained in the study.

The odds ratio of 16.403 for training implies that the effects of other factors kept aside, the
odds ratio in favor of being willing increases by a factor of 16.403 as the head of the
household trained and gets participated in different water harvesting related programs.

Attitude towards water harvesting technologies (ATITECHN ): In many literatures such as

Ngiggi (2003) and FAO (1994) the issue of technological convenience and acceptance by the
users for a proper promotion and adoption of water harvesting technology have been stressed.
The ease of implementation may be affected by many factors such as availability of local
construction materials, ease of maintenance and construction, and so on. The attitude of the
users towards the available water harvesting technologies convenience and uses is an
important variable for participation decision. It is found, as expected, to be highly significant
at (P< 0.01) and positively related with the decision of a fa rmer for participation. The result of

70
this study coincides with the ideas in the aforementioned literatures. The essence of the
question was that 'is the available water harvesting technologies easy to implement and have
economic importance'? Majority of the respondents who responded by saying yes were willing
to participate in water harvesting works. This was the main reason for the slope to become
positive.

The odds ratio 17.064 indicate that keeping the effects of other factors constant, willingness of
a user increases by a factor of 17.04 as he/she develops positive attitude towards the uses of
water harvesting technology.

Financial constraints of the household (FINANCON ): Availability of money is a

prerequisite for a technology to be adopted and promoted properly. Shortage of money was
significant at (P< 0.01) but negatively related to the willingness of farmers. This result agreed
with what was expected earlier in the hypothesis and with the logical framework that a farmer
with shortage of money withdraws from adopting a new technology. The negative relationship
indicates that the odds ratio in favor of the probability of being willing decreases with an
increase in shortage of money. Niggi (2003) reported similar results.

The odds ratio of 0.068 for money shortages implies that, other things being constant, the odds
ratio in favor of being willing decreases by a factor of 0.068 as money shortage increases by a
unit. To express the concept conversely, the decision to participate increases by a factor of
0.068 as availability of money increased by one unit.

Distance from extension center (DISTEXTC): This variable is significant at (P< 0.1) and
related negatively with the willingness of farmers to participate in water harvesting works. The
result is cons istent with the idea in the hypothesis, which means those farmers who are nearby
the development center may have more contact ours with their development workers. On the
contrary, non-willing farmers are located at a farther distance from development center.

71
The odds ratio 0.970 indicate that with the assumption of ceteris paribus, the willingness of a
farmer gets decreasing by a factor of 0.970 as distance of homestead from development center
increases by 1 minutes waking time.

4.1.3. Sensitivity analysis for participation decision

All significant variables do not have the same level of impact on the farmer's participation
decision. The relative importance of the qualitative explanatory variables can be seen by
examining the changes in probabilities tha t would result from changes in values of these
variables. To rank these factors a "typical farmer" is defined by the most frequent values of the
qualitative variables included in the model. Thus a typical farmer is one who has good attitude
and perception about the uses of water harvesting technology (52%), who participated in
training and visit program (55.83%), has faced financial shortages during the last cropping
season (58.3%), can able to read and write (92.5%). Thus, the probability that the typical
farmer will show interest to participate in water harvesting activities was estimated to be
0.981. The effect of significant dummy variables was calculated by changing their values
keeping all the continuous variables at their mean values and the dummy variables at their
most frequent values (Table 14). The predicted probabilities show how the likelihood of
willingness to participate in water harvesting practices was affected by changes in the
significant dummy variables. Hence, the probability of willingness to participate in water
harvesting activities decreases by 22.5% for a typical farmer who did not have good attitude
towards water harvesting. Similarly, the probability of willingness to participate in water
harvesting decreases by 21.74% for a typical farmer who did not take training. The probability
of willingness to participate in water harvesting decreases by 19.52% as a typical farmer is
illiterate.

Likewise, the relative importance of the significant continuous explanatory variables on

farmers' willingness to participate in water harvesting activities can be seen by explaining
variable elasticity, defined as the percentage change in the value of these variables.
Accordingly, a 10% increment in adult equivalent labour availability would decrease the
probability of willingness to participate in water harvesting practices by 0.43 percent.

72
Similarly, the probability of willingness would decrease by 0.33 percent as livestock
availability increases by 10% tropical livestock unit.

Table 14: Sensitivity analysis

Probabil Change in Percentage

Description ity probability change in
probability

Typical farmer 0.981

Typical farmer but did not take training 0.768 0.2134954 21.74
Typical farmer but do not have good attitude about WH 0.761 0.2206885 22.50
Typical farmer but did not face financial constraint 0.998 -0.016783 1.71
Typical farmer but illiterate 0.790 0.1915073 19.52
A 10% increase in TTLU 0.985 -0.003250 0.33
A 10% increase in distance from the development center 0.980 0.0009984 0.102
A 10% increase in labour availability 0.986 -0.004165 0.43

Source: From survey

4.2. Results of Preference Analysis for Water Storage Structure Groups.

In this part the results of both descriptive statistics and multinomial logit model are discussed
in brief to explain the choice decision of farm households among alternative water storage
technology groups.

4.2.1. Results of descriptive statistics and mean test

Results of descriptive statistics for variables, which were taken in both models, are described
in the binary logistics part. Thus, neglecting the commonly used variables in the two models,
in this section the results of the descriptive statistics of variables, which were taken only in
multinomial logit, are discussed.

73
Of all the total 355 plots, 180 or 50.7% were black soil categories while 78 plots or 22% and
97 or 27.3% were gray and red soil plots respectively. Red soil categories are more suitable to
construct underground structure groups while ponds are usually built in black soil categories.

Slope of plots was categorized into three. Accordingly, 198 or 55.8% of the total plots have
flat slope. On the other hand, 146 plots or 40.8% and 11 or 3.4% of the plots were categorized
as medium and steep sloped plots respectively.

Fertility level varies among plots. Taking the evaluation of land distribution and DAs
classification of plots, of all sample plots, 131 or 36.90% were registered to be fertile plots.
Similarly, 150 or 42.3% and 74 or 20.845% were classified as medium and low fertile plots,
respectively.

The maximum plot area was 1.5 hectares while the minimum was 0.01 hectare. The mean plot
size is known to be 0.39 hectare. The distance of a plot from homestead is estimated to range
from a minimum of zero to a maximum of 5 kilometers with a standard deviation of 2.1407. It
is measured in kilometers.

Plot ownership shows the status of ownership or holding of a specific plot. From the total 355
sample plots, sample household farmers owned 296 plots or 83.38% and the rest 59 plots or
16.62% were obtained by hiring from other farmers.

74
4.2.1.1. Mean and frequency comparisons for multinomial logit analysis

Table 15: Multiple Comparisons

Dependent (I) (J) Mean Std. Error Sig.

Variable STRUCT STRUCTU Difference
U (I-J)

1 2 -0.219 0.126 0.084

PLOAREA
1 3 0.107 0.117 0.361
2 3 0.325* 0.137 0.018

DISTPLO 1 2 -1.029* 0.283 0.000

1 3 -8.26E-02 0.261 0.752
2 3 0.946* 0.308 0.002

TTLU 1 2 -0.506* 0.254 0.047

1 3 -1.196* 0.234 0.000
2 3 -0.690* 0.276 0.013

MANEQIVP 1 2 0.21 0.16 0.179

1 3 -0.11 0.15 0.438
2 3 -0.33 0.17 0.059

Source: results of mean comparison

*The mean difference is significant at the .05 levels.

The mean difference with reference to each technology group is seen above in Table 15. The
mean difference between aboveground and surface pond structure groups with respect to a
variable area of the plot is significant at 5% probability level. Similarly, the mean difference in
underground and aboveground as well as aboveground and surface pond structure groups are
significant at 1% with reference to a variable distance of the plot.

75
Preference decisions for the three groups of water storage technolo gy groups are not only
dependent on mean differences of continuous variables. It is also highly affected by frequency
differences of discrete variables. For this frequency differences of discrete variables were
tested using chi-square test. The result is presented below in Table 16.

Table 16: Results of chi-square tests

Variables Score No Under Above Surface Total Chi-

Choice Ground Ground Pond Square
1 81 32 11 7 131
FERTIL 2 120 11 5 14 150 16.82**
3 270 25 20 29 74

1 135 7 12 26 180
SOILTYP 2 67 3 6 2 78 58.86***
3 68 26 2 1 97

1 127 29 16 26 198
SLOP 2 130 4 6 3 146 54.04***
3 10 - 1 - 11

OWNPLO 1 214 35 19 28 296 12.031**

2 56 1 1 1 59

Source: results of Chi-square test

** Significant at 5% probability level
*** Significant at 1% probability level

The results of the Chi-square test indicate that all variables are significant. Soil and slope
categories of a plot are significant at 1% probability level. Similarly, variables ownership and
fertility leve l of the plot are significant at 5% probability levels.

76
4.2.1.2. Preference among water storage technology groups

Experts and researchers in the field of rainwater harvesting and natural resources conservation
in general categorize different types of water harvesting storage structures into some groups
based on specific criteria. For this study, storage structures were grouped mainly into three.
These were aboveground, underground, and surface pond structure groups. These categories at
the moment are being practiced in the study area. This type of classification are commonly
practiced by many researchers and organizations, for example Rees (2000) and Martinson et
al. (2001) used the same classification as did in this study.

Accordingly, willing respondents were asked to reveal their preferences among water storage
technology groups. Of the entire total willing farmers, 16 respondents or 23.53 % preferred
aboveground structure groups. 32 farmers or 47.1 percent of the total willing respondents
wanted to have underground structures. Still 20 farmers or 29.4 % of the willing respondents
preferred to have surface pond structure groups. Sample farmers were also asked the reasons
for preferring a specific group. Most farmers who preferred aboveground type of structures
forwarded their reasons as, the aboveground structures are convenient to manage. Some others
within this group put their reasons, as their plot is stony and cannot be dug deep into the
ground to construct underground type of structures. Some farme rs within this group have also
a fear that if they build underground type of structures, at some day their children or their
animals may fall into the water harvesting structures.
Those who prefer the underground structure group on the other hand reasoned their
preferences as underground structures prevent evaporation. Some farmers within this group
also added other reasons that underground groups are easy to construct.

Surface pond structure demanders on the other hand, have their own reasons for their
preferences. Most surface pond preferring farmers said that this structure group is the most
suiting structure type for their plot because as their plot has black soil, it will be costly if they
want to build underground structures. In addition, surface pond structures can vary in size as
the ability to pay of a given farmer. They can hold large or small quantity of water as per
requirement and ability.

77
4.2.2. Econometric results of the multinomial logit model

As done in the binary logistics part, various tests of multicollinearity were conducted and
hence variables were found free from the problem of multicollinearity. See appendix table 5
for variance inflation factor and appendix table 6 for contingency coefficient.

The various goodness of fit measures were checked and validate that the model fits the data.
The likelihood ratio test statistics exceeds the Chi-square critical value at less than 1%
probability level. This implies that the hypothesis, which says all coefficients except the
intercept is zero, was rejected. The value of Pearson Chi-square test shows the overall
goodness of fit of the model at less than 1% probability level.

As can be seen from Table 17 below, of all the 10 explanatory variables considered as
determinants for the choice decision of sample households among different water storage
structure groups, 4 and 2 variables were found to have significant impact on the choice
decision for underground and surface pond technology groups respectively at (P< 0.01). 1
variable for the aboveground and 2 variables for the surface pond structure groups were
significant at (P< 0.05). Similarly, 2 variables for aboveground, 1 variable for surface pond
and 1 variable for underground structure groups were significant at (P< 0.1) to affect the
choice decision. The effect of some significant variables is not similar for the three structure
groups. Some may be highly significant to affect the choice decision for a particular group and
may be insignificant for the other groups.

78
Table 17: Parameter Estimates of the Multinomial Logistic Regression

Structure B Odds ratio Wald Sig.

Intercept 5.274 5.409 0.020

PLOAREA -0.154 0.860 0.250 0.617
DISTPLO -0.603 0.547 6.796 0.009***
SOILTYP 1.118 3.060 15.826 0.000***
Under OWNPLO 2.018 7.523 3.316 0.069*
Ground FERTILIT -0.909 0.403 8.002 0.005***
SLOP -1.620 0.198 10.680 0.001***
EDUC 0.928 2.529 1.014 0.314
TTLU -0.143 0.867 1.400 0.237
FINANCON .160 1.174 0.105 0.746
LABORAVA 0.846 2.33 0.821 0.365

Intercept 5.533 6.266 0.012

PLOAREA 0.323 1.380 1.100 0.294
DISTPLO -0.514 0.598 5.876 0.015**
SOILTYP -0.517 0.597 2.327 0.127
OWNPLO 1.552 4.721 2.133 0.144
FERTILIT 0.569 1.77 2.256 0.133
Above SLOP -0.394 0.674 1.785 0.182
Ground EDUC 1.344 3.834 2.779 0.095*
TTLU 0.138 1.148 0.801 0.371
FINANCON -0.976 0.377 2.949 0.086*
LABORAVA -0.450 0.637 0.169 0.681

Surface Intercept 3.230 2.399 0.121

pond PLOAREA 0.523 1.690 3.690 0.055*
DISTPLO 0.180 1.197 1.275 0.259

79
 SOILTYP -1.386 0.250 7.338 0.007***
OWNPLO 1.693 5.436 4.018 0.045**
FERTILIT 0.415 1.514 1.761 0.184
SLOP -1.899 0.150 7.358 0.007***
EDUC 1.312 3.714 2.487 0.115
TTLU -0.118 0.889 0.869 0.351
FINANCON 4.222E-02 1.043 0.006 0.936
LABORAVA 1.390 4.016 3.908 0.048**

Source: results Multinomial analysis

*** Significant at 1% probability level -2 Log Likelihood = 393.214

** Significant at 5% probability level
*Significant at 10% probability level

Interpretation of empirical results:

Soil type of the plot (SOILTYP ): This variable is highly significant at (P< 0.01) to affect the
choice decision for underground and surface pond structure groups but insignificant to affect
the choice decision for aboveground structure groups. The positive sign of the coefficient in
underground case and the negative coefficients in the surface pond and aboveground structure
groups is found to match exactly with the idea in the hypothesis. This means that red soils are
more preferable to construct underground structure groups but surface ponds are usually
constructed in black soils. This is the usual practice in the study area and true in the theoretical
framework, which is stated in (Brady and Well, 2002).

The odds ratio 3.060 in underground structure group indicate that with the assumption of
ceteris paribus, the choice decision in favor of underground structure group increases by a
factor of 3.060 per unit increase in the soil category. Similarly, the decision in favor of surface
pond structure group decreases by a factor of 0.250 per unit increase in the soil category.

Slope of the plot (SLOP): It affects the choice decision for underground and surface pond
structure groups significantly at (P< 0.01) but insignificant to affect the choice decision for
aboveground structure group. The coefficients of this variable in all the three structure groups

80
capture a negative sign. This result is consistent with the idea in the hypothesis and the general
theoretical framework, which is mentioned in (FAO, 1994). Users usually prefer to construct
underground and surface pond structure groups in plots having flat topography mainly to
avoid or to minimize the large earth works required to construct them.

The odds ratio 0.198 in the underground structure group case indicate that considering the
influences of other factors constant, the choice decision in favor of the underground structure
group decreases by a factor of 0.198 as the steepness of the slope increases by one category
unit. In a similar fashion, the choice decision in favor of surface pond structure group gets
decreasing by a factor of 0.150 as slope category increases by a unit.

Distance of the plot from homestead (DISTPLO): This variable is significant at (P< 0.01)
and (P< 0.05) to affect the choice decision for underground and aboveground structure groups
respectively but insignificant for surface pond structure group. The parameter in all the three
groups also took the expected sign, which is negative in underground and aboveground but
positive in surface pond structure groups. The result agreed with the idea in the hypothesis.
The negative sign of the coefficients was as anticipated indicating that as the distance of a plot
from homestead is large, users are not interested to construct above and underground structure
groups. Theoretically, it was true for both cases.

The odds ratios 0.547 for the underground and 0.598 for the aboveground structure groups
indicate that keeping the influences of other factors constant, the choice for above and
underground structure groups decrease by the rate of the above mentioned factors as distance
of the plot increases by 1 kilometer.

Area of the plot (PLOAREA ): The variable affects the choice decision for surface pond
structure group significantly at (P< 0.1). The positive coefficients of this variable in the
aboveground and surface pond structure groups but negative in the underground group case
are found to be consistent with the hypothesis. It indicates that the first two structure groups
require relatively larger area than the aboveground group for their construction.

81
The odds ratio 1.690 for this variable in surface pond structure group indicate that keeping the
influences of other factors constant, the decision in favor of the surface pond structure group
increases by a factor of 1.690 as plot size increases by one unit.

Labour availability (LABORAVA ): The variable is significant at (P<0.05) to affect the

choice decision for the surface pond structure group. The coefficients in the three structure
group cases also capture the anticipated sign, which was negative for the aboveground and
positive for the underground and surface pond cases. This agrees with the theory that
underground and surface pond structure groups in general demand large labour units than the
aboveground ones (Martinson et al., 2001). On the other hand, aboveground structure groups
require relatively large amount of financial resources.

The odds ratio 4.016 for the choice of surface pond structure group indicates that with the
assumption of ceteris paribus, the choice for surface pond structure group increases with a
factor of 4.016 per unit increase in adult equivalent labour.

Ownership of the plot (OWNPLO): This variable significantly at (P< 0.05) and (P< 0.1)
affects the decision behavior of farmers in the choice of surface pond and underground
structure groups, respectively. The sign of the coefficients in all the three groups is seen to be
positive. The coefficient in the aboveground case is found to be opposite to the idea in the
hypothesis. This means that a farmer prefers to construct all structure groups in his own plot.
Many literatures repeatedly underscore the land tenure as the determinant factor for the
adoption of water harvesting practices.

The odds ratio 7.52 indicates that the decision to choose underground structure groups
increases by a factor of 7.52 as ownership of a plot increases by a unit. Similarly, choice
decision in favor of surface pond structure group increase by a factor of 5.436 with per unit
shift in the status of ownership of a plot.

Fertility of the plot (FERTILIT): It refers to the fertility status of the plot on which water
harvesting structure is to be constructed. This was significant at (P< 0.01) for underground
structure groups but not for the other two. The coefficients of the variable in all the three

82
structure groups also capture the anticipated sign. This result is consistent with the hypothesis,
which argues that aboveground and surface pond structure groups are built not only for the
purpose of crop production but also for drinking water supply. Therefore, aboveground and
surface pond types can be built in both fertile and infertile plots. That is why fertility of plots
is not so much concerns of respondents to construct the two structure groups.

The odds ratio 0.403 show that as the plot becomes fertile, the decision to construct
underground structure groups increases by a factor of 0.403. Conversely, the decision in favor
of the underground structure groups gets decreasing by a factor of 0.403 as fertility level
decreases by a category unit.

Education level of head of the household (EDUC): Education level is significant to affect the
decision for aboveground structure groups at (P< 0.1) but insignificant to affect the choice
decision for underground and surface pond structure groups. The coefficients in all the three
groups were found to be positive. The result shows that increase in the level of education of
the head of the household enables him to identify and to select convenient structure group.

The odds ratio 3.834 for the aboveground structure group indicate that keeping the influences
of other factors constant, the decision to choose aboveground structure groups gets increasing
as the education level increased by one category unit.

Financial constraints of the household (FINANCON): Financial constraint is significant at

(P< 0.1) to affect the choice decision for aboveground structure group. The coefficients
capture the hypothesized sign for all structure groups. The result in all the three groups is
consistent with the idea in the hypothesis and with the theoretical framework, which says
aboveground structure groups generally require large amount of money when compared with
the other two groups (Rees, 2001).

The odds ratio 0.377 indicate that leaving the influences of other factors aside, the decision to
choose aboveground structure groups decreases by a factor of 0.377 per unit increase in
money shortage.

83
 5. CONCLUSIONS AND RECOMMENDATION

5.1. Conclusion

In spite of being the dominant sector of the Ethiopian economy, agriculture in the country, was
and still is characterized by low productivity in general and low yield per unit area in
particular. Many people attribute the problem to the growth of population at a rate faster than
what the means of sustenance would guarantee, drought, environmental degradation, poor
cultural practices, limited accessibility and use of improved technology, insufficient
infrastructure and ill-thought-out policies. This low performance of the agricultural sector
coupled with rapid population growth, has had an impact on household food insecurity and
profound effect on per capita food production.

There is, therefore, a desperate need for increasing productivity of Ethiopian agriculture, since
its effect is reflected in the overall economy of the country. Among others, this requires
overcoming the moisture stress problem, which is believed to play a pivotal role in the
agricultural development of the country.

More recently, Ethiopia is trying to expand small scale irrigation schemes, especial those that
can be constructed and managed individual resource poor farm household levels. Among the
means to accomplish this objective is promoting water harvesting practices at individual farm
household levels. However, there is no study as such which can indicate the status of
promotion and adoption of water harvesting works across potential users.

This study has tried to look into the socio-economic, physical and other related factors, which
can affect the willingness of farming households to participate in water harvesting works and
preferences among water storage technology groups. For this, data were collected from 120
farm households drown randomly from Dejen district. The primary data were collected using
structured questionnaire. Secondary data were collected to supplement the data obtained from
the survey.

84
Fourteen variables hypothesized to explain farmers' willingness to participate in water
harvesting activities were used to study the participation decision of farm households in water
harvesting activities. Similarly, ten hypothesized explanatory variables were used to identify
the preferences of farm households among water storage technology groups. Evidences from
the descriptive analysis indicate that willing farmers have better education standards, own
more livestock units, have relatively large adult equivalent labour force, have better access to
credit, and most of them did not face financial constraint in the past cropping season.
Moreover, willing farmers perceive water harvesting as the main means of alleviating
moisture stress and believed that the available water harvesting technologies, even if not all
can be implemented and sustained at individual farm household levels. Most willing farmers
have exposures to training, field visits, and other water harvesting related matters, which made
them aware of the means and uses of water harvesting. Non-willing farmers on the other hand,
are located relatively at a farther distance from extension and market centers. This means they
do have lesser development agent contact and have less access to market to sale their produce.
Most of the non-willing farmers own irrigated plots. Majority of them faced financial and
labour shortages.

The result of the binary logit analysis indicated that three variables at (P< 0.01) level and three
variables at (P< 0.05) were found to be significant to affect the willingness of users to
participate in water harvesting activities. Training was found to have a positive and significant
impact on farmers' willingness to participate in water harvesting practices at (P< 0.01) level of
significance implying that trained farmers are more opt in understanding the advantages of
water harvesting and could easily decided to take part willingly in water harvesting practices.
Attitude towards the importance of water harvesting activities for food security is another
highly significant and positively related variable to affect the willingness of sample
households to participate in water harvesting works. This means that favorable attitude
towards the importance of water harvesting technologies is an important input to decide for
participation. Financial constraint is negatively and significantly correlated to the willingness
of users to participate in water harvesting practices. The result revealed the truth in such a way
that a farmer faced with financial constraints cannot purchase local and industrial materials
needed for constructing water storage structures. Labour availability is the other key variable

85
affecting the willingness of farmers to participate in water harvesting activities. It is significant
at less than 0.05 percent significance level and positively related with the willingness of users
for participation, which indicated that availability of labour is an essential element to practice
water harvesting activities. Livestock availability is significant at (P < 0.05) and positively
related with the willingness to participate in water harvesting works. Education level of head
of the household is also significant at (P< 0.05), which implies that as the education level of
head of the household increases the probability of participation willingly is also get increasing.

The result of the multinomial logit analysis revealed that slope of the plot on which water
harvesting structure is to be constructed matters for choice decision making among alternative
water harvesting structure groups. Thus, it affects the choice decision for underground and
surface pond structure groups at (P< 0.01) levels. Those water storage structure groups, which
need large earthwork, need not be built in steep plots. Soil type of the plot significantly affects
the choice decision for underground and surface pond structure groups at (P< 0.01) and (P<
0.01). This means that red soil plots are usually preferable to build underground structure
groups while surface pond structure groups are preferably be built in black soil plots. Distance
of the plot from homestead is another key variable that serve as an input to choose among
water storage structure groups. The variable was significant at (P< 0.05) probability levels to
affect the choice decision for aboveground and underground structure groups and negatively
related with the choice decision for the two groups. The result indicated that as the distance of
a plot from homestead increases users are not interested to build the two structure groups.
Availability of labour affects the choice decision for surface pond structure groups
significantly at (P< 0.05) probability level. Similarly, ownership of a plot affects the choice
decision for surface pond structure group at (P< 0.05) probability levels and affects the choice
decision for underground structure group at (P< 0.1) probability levels. Fertility level of the
plot was also found to affect the choice decision for underground structure groups at (P< 0.01)
probability levels. This is true because underground structure groups are usually built for the
purpose of crop production, which may take fertility level of the plot as main criterion. Area of
the plot is an important variable influencing the choice decision among water storage structure
groups. The result of the multinomial analysis indicated that the area of a plot was found to

86
have significant and positive influence to affect the choice decision for surface pond structure
groups.

5.2. Policy Implications

Some implications of this study were found to be relevant. The importance of expanding small
scale irrigation schemes through water harvesting practices for sustained development of the
agricultural sector becomes growing especially at the concerned experts and political leaders
level. However, the awareness of the users and development workers is seen to be very low.
Based on the main findings of the study, the following recommendations are made.

1. Participating users in training, farmer's field day, field visits, visits to other villages and
other water harvesting related issues is an essential element to promote water harvesting in
small holder, resource poor farm household level. To accomplish this responsibility, the
government has to first equip the pertinent experts who are working specially at PA, district,
and regional levels with the necessary knowledge about the uses and means of implementing
water harvesting technologies. In addition to providing informal trainings, expanding formal
education to the farming household enhances promotion and adoption of water harvesting
works in various dimensions. Furthermore, inclusion of topics on water management into the
formal education curricula can help to improve awareness level.

2. The farmers' willingness to participate in water harvesting works was also found to be
highly associated with their understanding and knowledge level of water harvesting, whether
the technologies can be implemented and managed at their know how and resource levels. To
make users accept and implement water harvesting technologies, in addition to training and
extension work, government and other development cooperative organizations should provide
technologies appropriate to the specific socio-economic, cultural and bio-physical
circumstances. Moreover, the individuals, groups as well as policy makers should work
towards making the people aware of the uses of water harvesting activities to fight against
drought and food shortages.

87
3. Financial status of the farming households is another key factor explaining the decision
behavior of farmers for participation in water harvesting works. Those households facing
financial constraints were not willing to participate in water harvesting activities. Working to
alleviate the financial constraints of users is, therefore, essential for policy makers and other
NGOs to promote water-harvesting practices in the long run. This can be carried out using
various means, one of which is provision of adequate loan with possible minimum interest
rates. The other means may be creating favorable marketing policy and organizing users to
find their potential markets and improving market infrastructure.

4. Making more contact hours between farmers and development workers can also enhance
willingness of users for participation in water harvesting works. This was seen from the
econometrics result, which indicated that those farm household heads that were nearby the
development center were found to be more willing. The result was brought by more contacts
made by the development worker and farmer. Therefore, policy makers should rather work to
assign sufficient number of DAs in the rural area.

5. Finally, it would be necessary to indicate the preference decision behavior of farm

households among alternative water harvesting technology groups based on the characteristics
of a plot. As plots vary in various biophysical factors, the preference decision of users among
water harvesting technology groups also vary following the variation in characteristics of
plots. Therefore, experts and policy makers should consider developing and promoting water
storage structure groups focusing on the characteristics of a particular area and locality.

88
 6.REFERENCES

Abdulkarim Guleid, 2002. Water harvesting in the Somalia National Regional State of
Ethiopia: Workshop on the experiences of water harvesting in Ethiopia: principles and
practices. December 28-30. Mekele, Ethiopia. 34-37p.

Aldrich, J. and D. Nelson, 1984. Linear Probability, Logit and Probit Model: Quantitative
Applications in the Social Science. Sera Miller McCun, Sage Pub.Inc. University of Minisota
and Iowa, London. 47p.

Amemiya, T., 1981. "Q ualitative Response Models: A Survey". Journal of Economic
Literature. (19): 48-53.

Bekele Shiferaw and T.Holden, 1998. Resource degradation and adoption of land
conservation technologies in the Ethipian highlands: a case study in Andit Tid, North Shewa.
The Journal of the International Association of Agricultural Economists. 18( 3): 244.

Berhanu Adugna, 1999. Drought in Ethiopia. Ethiopian Journal of Water Science and
Technology. Arbaminch, Ethiopia. 3( 1): 132-137.

BoA, 1999. Annual report for Bureau of Agriculture. Amhara National Regional State, Bahir
Dar, Ethiopia. 12-16p.

BoFED, 2003. Annual report for Bureau of Finance and Economic Development. Amhara
National Regional State, Ethiopia. (2): 15-23.

Boers, T. and J. Ben-Asher, 1982. A review of rainwater harvesting. Agricultural water

management. .5(4): 145-158.

Borsa Natea, 2001. Workshop on the experiences of water harvesting in the dry lands of
Ethiopia: Principles and practices. December 28-30, Mekele, Ethiopia. 25-26p.

Brady, c. and R.Well, 2002. The Nature and Properties of Soil. Thirteenth Edition.
Prenticehole, New Jersey, USA. 165-171p.

Catterson, T., Moges Worku, Messel Endalew, Carmela G., Frank B., Abebaw Wolde
Amannuel and Kibru Mamusha, 1999. Programmatic environmental assessment of small- scale
irrigation in Ethiopia. Catolic Relief Services, U.S. Catolic Conference, Baltimore, Maryland.
U.S.A. 3-4p.

Chilot Yirga, 1994. Factors influencing adoption of new wheat technologies in the Wolmera
and Addis Alem areas of Ethiopia. An M.Sc Thesis presented to the School of Graduate
Studies of Alemaya University. 63p.

89
Christopher, A. and P. Silva-Ochoa, 2001. Collective action for water harvesting irrigation in
the Lerma-Chapala Basin, CAPRI working paper. Mexico. (20): 15.

Critchley, W. and C. Sieger, 1991. A manual for the design and construction of water
harvesting schemes for plant production. FAO, Rome. (17): 91.

CSA, 1997. Agricultural sample survey. Report on land utilization. Addis Ababa, Ethiopia.
(4): 209.

CSA, 2003. Ethiopian agricultural sample enumeration results for Amhara Region. Statistical
report on socioeconomic characteristics of the population in agricultural households and land
use. Addis Ababa, Ethiopia. (1): 43-45.

CTA, 2000. Managing dry land resources. A manual for Eastern and Southern Africa. 53p.

CTA, 2002. Managing dry land resources. A manual for Eastern and Southern Africa. 46p.

Dasgupta, S., 1989. Diffusion of Agricultural Innovations in Village India. Department of

Sociology and Anthropology, University of Prince Edward Island. Canada. 74-86p

Degenet Abebaw, 1999. Determinants of adoption of high yielding varieties of maize in

Jimma Zone: The case of Mana and Kersa woreda. An M.Sc Thesis Presented to the School of
Graduate Studies of Alemaya University. 1-2p.

DTU, 2000. Very low cost domestic roof water harvesting in the humid tropics: existing
differences, DFID KAR contract R7833. School of Engineering University of Warwick.
Report A3. 27p.

DTU, 2002. Very low cost domestic roof water harvesting in the humid tropics: existing
differences, DFID KAR contract R7833. School of Engineering University of Warwick. (1):
23p.

FAO, 1993. Irrigation in the near east region in figures. Proceedings of the FAO expert
consultation. Rome, Italy. 56-78p.

FAO, 1994. Water harvesting for improved crop production. Proceedings of the FAO
consultation. Cairo, Egypt. 31-40p.

FAO, 1995. Irrigation water development models. Proceedings of the FAO expert
consultation. Rome, Italy. 112-115p.

90
FAO, 2000. The state of food and agriculture. Proceedings of the FAO experts consultation.
Rome, Italy.133p.

FAO, 2002. Unlocking the water potentials of agriculture. Proceedings of the FAO
consultation team. Rome, Italy.37p.

FAO, 2003. Unlocking the water potentials of agriculture. Proceedings of the FAO
consultation team. Rome, Italy.18p.

Green, H. W., 1991. Econometric Analysis. New York University. MacMillan Publishing
Company. 697-701p.

Green, H.W., 2000. Econometric Analysis. Fourth Edition. New York University. MacMillan
Publishing Company. 857- 871p.

Gujarati, N., D., 1995. Basic Econometrics. Third Edition. McGraw-Hill,Inc. New york.

Habtamu, 1999. Rainwater harvesting concepts and issues, proceedings of the annual meeting
on soil and water conservation, Addis Ababa, Ethipia. (2). 24p.

Hatibu, N. and F. Mahoo, 2000. Rainwater harvesting for natural resources management: A
planning guide for Tanzania, Technical handbook, RELMA. (22): 39p.

Hosmer, D., and S. Lemeshew, 1989. Applied Logistic Regression. A Wiley-Inter science
Publication. Newyork.

IFAD, 2001. Irrigation scheduling: From theory to practice. Center for development
cooperation services. Free University, Amsterdam. 39p.

Jdge, G., .E.Griffiths, .R.Carter, H. Lutkepohl and T. Lee, 1985. The Theory and Practice of
Econometrics. Wiley Series in Probability and Mathematical Statistics. Newyork. 756-778p.

Lapar, A. and S. Pandey, 1999. Adoption of soil conservation: the case of the
Philippine uplands. The Journal of the International Association of Agricultural Economists,
(21). 3.

LEISA. 1998. Challenging Water Scarcity. ILEIA News letter for the Low External Input and
Sustainable Agriculture. (14). 1.

Lind, A. Douglass and R. D.Masn, 1994. Basic Statistics for Business and Economics.
University of Toledo, USA. 128-131p.

Lo, K., Andrew, 2003. Rainwater Harvesting and Utilization. Water Resources Bureau of
Gansu Province. International Rainwater Catchements System Association, Gansu Research
Institute for Water Conservancy. Gansu, China. 151-153; 173-178p.

91
Maddala, G.S., 1983. Limited Dependent and Qualitative Variables in Econometrics.
Economic Society Monographs. Cambridge University Press, USA. 67p.

Martinson, D. Ranatunga and A. Gunarante, 2002. Reducing rainwater harvesting system

cost, Sustainable environmental sanitation and water services, 28th WEDC conference.
Calcutta, India. 43p.

Mintesnot Behailu, 2001. The experience of water harvesting in the drylands of Ethiopia:
Principles and practices. Workshop on water harvesting in Ethipia. December 28-30. Mekele,
Ethiopia. 23p.

Momdouh, N., 1999. Assessing desertification and water harvesting in the Middle East
and North Africa, ZEF discussion papers on development policy. Bonn, Germany. (10). 57p.

Mulugeta Enki, 2000. Determinants of adoption of soil conservation practices in central

highlands of Ethiopia: The case of three weredas of Selale. An M.Sc. Thesis Presented to the
School of Graduate Studies. Alemaya University. 67p.
Ngigi, N., Stephen, 2003. Rainwater Harvesting for Improved Food Security. Nairobi, Kenya.
89-94, 233-244p.

Paulose Asrat, 2002. Determinants of farmers' willingness to participate in soil conservation

practices in the highlands of Bale: the case of Dinsho farming area. An M.Sc. Thesis Presented
to the School of Graduate Studies, Alemaya University. 98p.

Perrier, 1988. Opportunities for the productive use of rainfall normally lost to cropping for
temporal or spatial reasons. In: drought research priorities for the dry land Tropics. Bidinger,
F.R. and Johnson, C. (eds). Inc. RISAT, Patancheru, India: 113-131p.

Pindyck, R., and C. Rubinfeld, 1981. Econometric Models and Econometric Forecasts. Second
Edition. McGraw-Hill book Co. New York.

RDPE, 2002. The Federal Democratic Republic of Ethiopia, Rural Development Policy.
Ministry of Press and Information. 129p.

Rees, D., 2000. Partially Below Ground (PBG) tank for rain water storage. Instructions for
Manufacture. Development Technology Unit, School of Engineering, University of Warwick.
TR-RWH01. 5-7p.

Reij, C., P. Mulder and L. Begemann, 1993. Water Harvesting for Plant Production. A
Comprehensive Review of the Literature. Third Printing. The World Bank.Washington, D.C.
1(91). 3-13p.

92
Rogers, M. and F. Shomaker, 1971. Diffusion of Innovations. Second editition. Free Press of
Glencoe. NewYork. 185-187p.

Sambrook, B. and S. Akhter, 2001. Social and gender analysis: Findings from the inception
phase. Domestic Roofwater Harvesting Research Program, RN-RWH02. Hambantota,
Srilanka. 34-36p.

Storck, H., Bezabih Emana, Berhanu Adnew, A.Borowiecki and Shimeles W/Hawariate, 1991.
Farming systems and farm management practices of small holders in the hararghe highlands:
Farming systems and resources economics in the tropics. Wssenshaftsver lag vauk, Kiel,
Germany. 11. 41-48p.

Tesfaye Zegeye, Girma Taye, D. Tanner, H. Verkuijl, Aklilu Agdie and M. Wangi, 2001.
Adoption of improved bread wheat varieties and inorganic fertilizer by small-scale farmers in
Yelmana Densa and Farta Districts of Northwestern Ethiopia. Ethiopian Agricultural Research
Organization. Addis Ababa, Ethiopia. 4p.

Yapa, L. and C.Mayfeld, 1978. Non-adoption of innovations: evidence from discriminate

analysis. Economic Geography. 54(2). 145-156p.

Weide, S., 2003. Innovations adopted on integrated management of rainwater, groundwater

and surface water for sustainable utilization of water resources in Gansu, China. Water
Resource Bureau of Gansu Province. International Rainwater Catchement System Association.
Gansu Research Institute for Water Conservancy. China. 87p.

93
 7. APPENDIX

7.1. Appendix Tables

Appendix Table 1: Conversion factors used to compute adult equivalent (AE)

Age group (years) Male Female

Less than 10 0.6 0.6
10-13 0.9 0.8
14-16 1 0.75
17-50 1 0.75
Greater than 50 1 0.75

Source: Stork et al., 1991

Appendix Table 2: Conversion factors that used to estimate tropical livestock unit

Animal Category TTLU Animal Category TTLU

Calf 0.25 Donkey (young) 0.35
Weaned Calf 0.34 Camel 1.25
Heifer 0.75 Sheep & Goats 0.13
(adult)
Cow and Ox 1 Sheep & Goats 0.06
(young)
Horse 1.10 Chicken 0.013
Donkey (adult) 0.70

Source: Stork, et al, 1991.

94
Appendix Table 3: Values of VIF for continuous variables, which are used in binary logit

Continuous Variables R2 Value VIF

AGE .195 1.242

TTLU .105 1.117

LABORAVA .298 1.425

CULTAREA .437 1.1777

DISTEXTE .119 1.136

DISTMARK .202 1.253

Appendix Table 4: Contingency coefficients of discrete variables in binary logit

EDU FINANC IRRIGU ADEQO ATITEC ATITEC FOODS RESPO

EDUC 1

FINANC 0.164 1

IRRIGUS 0.140 0.292 1

ADEQOL 0.012 0.138 0.219 1

ATITECH 0.07 0.370 0.301 0.052 1

TRAININ 0.088 0.391 0.359 0.051 0.458 1

FOODSH 0.126 0.027 0.034 0.143 0.057 0.020 1

RESPONS 0.076 0.149 0.011 0.170 0.144 0.137 0.181 1

Source: survey data

95
Appendix Table 5: Values of VIF for continuous variables of the multinomial logit

Continuous Variables R2 value VIF

PLOAREA .099 1.109

DISTPLO .086 1.094

TTLU .042 1.044

Source: From survey data

Appendix Table 6 : Contingent coefficients of discrete variables for multinomial model

Discrete variables SOILTYP FERTILIT OWNPLO EDUC FINANCON SLOP

SOILTYP 1

FERTILIT .173 1

OWNPLO .140 .088 1

EDUC .121 .084 .003 1

FINANCON .082 .139 .008 .076 1

SLOP .259 .141 .046 .162 .109 1

Source: From survey data

96
7.2. Survey Questionnaire

General Information

Name of the woreda _____________________

Name of peasant association _____________________
Name of the village
Name of the respondent
Name of the interviewers
1. Personal or household characteristics
No Name Age Sex Formal schoolings Religion Relationship to the
completed head of the household
1
2
3
4
5
6
7

1.1. Marital status of the respondent. 1) married 2) single 3) divorced

1.2. Number of years that the respondent has been living in this region___________
1.3. Did you have some social position in the community so far? 1) yes 2) no
1.4. if yes, type of responsibility.................................................
1.5. When did you start your own farming? _____________ years ago.
1.6. Do you have any vocational training in agriculture? 1) yes 2) no
1.7. If yes, type of training.....................
When trained...................
Title of training.................
1.8. Would you tell me the type of houses you have?
1) Grass roofed 2) Corrugated tin roofed
2. Farm characteristics
2.1.How much is your total farm land in hectares? _____________
2.2. Would you tell the allotment of your land?
Plot plot area in distance soil slope crops ownership fertility irrigated
number name timad from type grown owned/hired or not
home(km)

97
2.3. Have you taken some land on lease or some other terms from other farmers for
cultivation? 1) yes 2) no
2.4. If yes amount of rent..................................................
2.5. Have you given some land on lease or some other terms to other farmers?
1) yes 2) no
2.6. Amount of rent...........................................
2.7. What crops did you grow during the last crop season?
Crops grown Area in Total yield in Remark
hectares hectares
Tef
Maize
Sorghum
Chickpeas
Beans
Wheat
Barley
Lin seed
Others
Total

2.8. How was the yield of your crop last cropping season season?
1)high 2) medium 3) low
2.9. From your total plot, how many not irrigated?...............................
2.10. have you faced food shortages in the past 5 years? 1. yes 2. no
2.11. If yes, how did you overcome? 1. buying 2. borrowing 3. aid 4. others
2.12. mention your agricultural problems in increasing orders? 1. Minimum farm land 2.
Pests 3.infertility of plot 4. animal disease 5. drought 6. shortage of input
7. labour shortage 8.lack of technical support 9. others
2.13. What do you suggest for a solution? 1. using irrigation 2. using credit
3. resettlement 4 . Others
2.14. Would you please tell me the type and number of livestock you have? _______
Live stock Sold during the year Died during Available
type the year during the year
No Birr
Oxen

Bulls
Cows
Calves

98
 Sheep
Goats

Donkeys
Mules

Horse

Chicken
Others

2.15.Mention problems faced in livestock keeping?......................................

2.16. Did you face labour shortage in your farming? 1. yes 2. no
2.17. If yes, for which farming activities?................................
2.18. If how did you overcome? ...........................................
2.19. Did you face shortage of oxen? 1. yes 2. no
2.20. If yes, how did you overcome?..............................
3. Institutional factors
3.1. How many minutes do you travel to reach the nearest:
1. Market center _____________ 2. Extension office ---------------
3. Main road _____________ 4. District town_____________
5. School_____________ 6. Church__________7. Clinic

3.2. Mention main sources of your income 1. .........2.............3..........

3.1. Have you ever faced money shortages to run your farming?
1) yes 2) no
3.3. If yes, how do you solve such a problem?
1) by borrowing money from friends 2) by borrowing money from merchants
3) by getting inputs on credit 4) others specify
3.4.The sources from where you usually borrow money?
1) banks 2) merchants 3) friends 4) others, specify
3.5. For what purposes do you use the credit?
1) to buy inputs 2) to by oxen 3) others, specify
3.6. Have you paid back your loan? 1) yes 2) no
3.7. Why did not you paid full? 1) due to insufficient return
2) lenders do not collect on time
3) others, specify
3.8. Is the credit facility adequate to your need? 1) yes 2) no
3.9. Have you faced problem of getting a loan? 1) yes 2) no
3.10. If yes, which problem? 1) Administrative problem
2) Collateral
3) Others, specify
3.11. How many times per month (or other) do you usually discuss agricultural matters with
the extension staff? 1) monthly 2) quarterly

99
 3) other time 4) on discussion
3.12. From whom do you get most frequent farm advice?
1) DA 3) woreda experts
2) local leaders 4) others, specify
4. Media exposure and communication
4.1.Did you have a radio/tape? 1) yes 2) no
4.2. How often do you listen? 1. daily 2. weekly 3. other time, mention
4.3.Have you any access and able to read written agricultural materials?
1. yes 2. no
4.4. Mention agricultural technologies you read or listen...............................
5. Information on marketing and consumption activities
5.1. How much of the crops produced on farm are sold? 1. half 2. Quarter
3. no sold 4. other amount
5.2. When do you usually sell your crops? 1. At harvest time 2. At summer
3. As per requirement 4. Other mention
5.3. In your opinion, what do you say about the time of selling of the products?
1. Convenient 2. Not convenient
5.4. If you think that you sell at the wrong time what makes you to do so?
1. To pay for input credit
2. To meet cash requirements to meet other payments
3. If other me ntion
5.5 To whom do you usually sell crop products? 1. Merchants 2. Other farmers
3. Retailers 4. Other, mention
5.6 . Did you face any marketing problem to your farm products? 1. yes 2. no
5.7 . If yes, mention the problem........................................................
6. Information on participation
6.1. Have you ever used any form of irrigation? 2. yes 2. no
6.2. If yes, how many hectares do you cultivate annually with irrigation? ----------
6.3. Which crops/vegetables/fruits you used to grow?------------------
6.4. Have you come across a neighbour or any other farmer who is using irrigation?
1 yes 2. no
6.5.In general what is your idea about the benefits of practicing irrigation?
1. very good 2. good 3. not necessary to use irrigation
7. Information on water harvesting
7.1. Have you heard/ read about water harvesting? 1. yes 2. no
7.2. From whom and which source/s have you got information about water harvesting?
1. DA 2. from local leaders 3. from radio 4. from experts 5. others specify
7.3. When did you get first information? ____________
7.4.How many structure types do you know? Name/mention all types you know

1. ____________ 2. ____________ 3. ____________ 4. _________

5. ____________ 6. ____________ 7. ____________ 8. _________

7.5.. Have you taken any training about water harvesting? 1. yes 2. no
7.6. When did you take training?...................................................
7.7. mention the advantages of the training?......................................
7.8. Have you ever initiated to participate in meeting, workshop, seminar and other

100
 related water harvesting issues by MOA or other organizations? 1. yes 2. no
7.9. If no, why? 1. Not educated 2. No opportunity
3. Not consulted 4. Others, specify
7.10. Can you openly give suggestions on water harvesting? 1. yes 2. no
7.11. Have you participated in any of the following water harvesting matters organized by
NGO, government, etc?
Forms of participation NGO Government Others
Water harvesting training program
Demonstration of water harvesting practices
Farmers` field day or other experience sharing
Water harvesting video show
Water harvesting tour in another village
Developing water harvesting program
Other participation, mention

7.12. Are you practicing in any form of water harvesting? ____________

7.13.If yes, which ones on which plot? ___________
7.14. what was your reason to build this structure on the mentioned plot?.........
7.15. who initiate to build?..............
7.16. did they make you convince or enforce?...............................
7.17. How did you judge the benefits of water harvesting after participation?...........
7.18. Is there someone who enforce you to build without your willingness?.............
7.19. Have you come across a neighbor who built without his willingness?.........
7.20. Who initiate / enforce you to build?......................
7.21. If not, why do you not in need of participating? 1. I have no information 2. due to
shortage of cash 3. water harvesting is not useful 4. if other, mention
7.22. Have you come across a neighbor who successfully build and used water harvesting? 1.
yes 2. no
7.23. which type?......................................................
7.24. If yes, Who enforce you to build? 1. your interest 2. government
7.25. If you come across successfully built and used, tell us the advantages and disadvantages
of WH? advantage..............disadvantage.....................
7.26. Due you believe that the costs of water harvesting can be covered by individual
household level? 1. yes 2. no
7.27. If difficult to cover at individual level, what do you suggest for a solution?..............
7.28. Are you willing to participate/ to have any form of water harvesting practice/s covering
all costs whic h are required to construct the structure you prefer?
1. yes 2. no
7.29.If yes which type.............. on plot number (referring 2.2)................................
7.30. Are you willing to construct WH structures covering half of the costs of the structure you
prefer? 1. yes 2. no
7.31. If yes, which type.............on plot number (referring 2.2)..........................
7.32. Are you willing to construct WH structure on your plot, if government is to cover all
costs required to construct? 1. yes 2. no
7.33. If yes, which type..............on plot number (referring 2.2)........................
7.34. If yes, which structure type/s do you want to have, in which plot?

101
Type of technology Sandy soil plot Loam soil plot Clay soil plot
1 Pond
2. Hemispherical
3. Bottle Shape
4. Dom cap tank
5. Hand dug well
6. cylindrical
7. cone cap
9. trench
10. spring
11. others. specify

7.35. Which structure groups you prefer most? 1. underground 2. aboveground 3. surface
pond
7.36. What are your reasons to prefer such groups?...........................................
7.37. If not willing to construct any, why do you not in need of participating?
1. I have no information 2. due to shortage of cash 3. water harvesting is not useful
4. the technology is not appropriate 5. if other, mention
7.38. Put the following water harvesting technologies in order of difficulty to implement in
terms of technological complexity, labor requirement and cost. (From easy to complex)
Type of technology Technological Construction Labor Generally
complexity Cost requirement Preferred to
implement
1 .Pond
2. Hemispherical
3. Bottle Shape
4. Dom cap tank
5. Hand dug well
6. cylindrical
7. cone cap
9. trench
10. spring
11. others. specify

7.39. What potential problems one will face in participating in water harvesting practices?
1. financial problem
1. Technological problem
2. Know how problem
3. Others
7.40. In your view how do you judge the advantages of water harvesting?
1. Very good 2. good 3. not necessary
7.41. If you are willing to have some form of water harvesting structure, which crops you
planned to grow? 1. vegetables 2. fruits 3. cereal crops 4. others, specify
7.42. What are your reasons. ------------------------------------------------------------------------------
-----------------------------------------------------------------------------------------------------

102
7.43. Have you received any of the following support from government or other
concerned?
1. technical information/ assistance
2. cost sharing/ financial assistance 4. casual labor for construction of WH structures
3. Implements/ inputs 5. Others specify
7.44. Give your opinion on the following points
No Description Agree Disagree Not clear
1 WH generate benefit beyond costs
2 Drought becomes increasing
3 Rain fed agriculture is insufficient to feed your family
4 Each family has to participate in WH to secure food
5 WH fosters development
6 WH structures are convenient to use
7 WH structures are easy to implement
8 WH structures are not costly
9 Farmers know the existing difference among WH
structures.
10 Farmers participate from their own willingness
11 WH structures share a large area
12 WH cause malaria

8. Awareness, attitudes, perceptions and benefits of water harvesting.

8.1. Do you know the existing differences among water harvesting practices?
1. yes 2. no
8.2. If yes, mention some of them.
------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------
8.3. In your opinion how do you evaluate the importance of water harvesting in securing
food supply and generating income to your family? 1. Necessary 2. unnecessary
8.4. If your choice is "necessary" could you mention some of the specific advantages that
could be derived by practicing water harvesting? ------------------------------------------------------
-------------------------------------------------------------------------------------------------
8.5. Could you have a point to mention about the disadvantages of water harvesting?------------
--------------------------------------------------------------------------------------------------
8.6. Which of the available technology groups/ practices most suit your interest. List in
order of importance. ----------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------
8.7. If the farmer decides to use the technologies / practices? (Mark if answer is given)
1. Observed the benefits other farmers obtained from using the technology
2. Persuaded by other farmers 3. Persuaded by change agents
4. Persuaded by others specify
8.8.If the farmer uses water harvesting practices? (Mark if answer is given)
1. to produce crops
2. as source of drinking water supply for the family
3. as source of drinking water supply for animals

103
4. others specify

104
105