Ar 867 e
Ar 867 e
Ar 867 e
IN'j)
g
Government of Ethiopia
Water Resources Development Authority
UNDP/FAO
FINAL REPORT
August 1987
Arrangement of Report
VOLUME 1
Main Report
1
Annex G Geology
Annex G
Annex J
VOLUME 4
VOLUME 5
Annex K Implementation
Page Nr
ABBREVIATIONS (iv)
CHAPTER 1 INTRODUCTION
1.1 Background 1
1.2 Objectives 2
1.3 Arrangement of the Report 2
1.4 Acknowledgements 3
4.1 General 19
4.2 Crops and Farming Systems 20
4.3 The Pastoralists and the Environment 25
4.4 Irrigation, Drainage and Flood Protection 27
4.4.1 Irrigation System 27
4.4.2 Drainage and Reclamation 29
4.4.3 Flood Protection 30
(i)
CONTENTS (cont.)
Page Nr
CHAPTER 6 EVALUATION 58
6.1 Methodology 58
6.2 Economic Bene fits and Costs 58
6.3 Results of Economic Analysis 60
6.4 Other Benefits 62
6.5 Finance 63
REFERENCES 66
LIST OF FIGURES
PLATES
Plate 1 Satellite Image of the Project Area 7
Plate 2 Awara Melka from the West 14
DRAWING
(bound in Album 1)
(iv)
ABBREVIATIONS (cont.)
UNITS
All units and their abbreviations are SI (Systeme Internationale), which are
usually equivalent to metric units, except the following:
qt quintal (100 kg)
yr year
day
md man-day
MM man-months
CURRENCY
Unless stated otherwise all financial and economic calculations are made in the
Ethiopian currency unit, the Birr, at 1986 prices. The exchange rate used is the
official rate, namely Birr 2.07 to US$ 1.
(v)
CHAPTER 1
INTRODUCTION
1.1 Background
In the early 1960s a major river basin study was undertaken for the Awash river
(Ref. 1). Among the areas identified for potential irrigation development was
the Kesem-Kebena plain, a relatively flat area of some 20 000 ha on the left
bank of the Awash in its middle course, intersected by the tributary rivers,
Kesem and Kebena, flowing off the western escarpment of the Great Rift valley.
The report on this basin study, issued by FAO in 1965, recommended:
During the 1970s and early 1980s the right bank developments of the Amibara
system, representing Stage 2 of the 1965 report's programme, were constructed.
The completion of the Bolhamo scheme and its linking to the Amibara canals
reached final design stage in 1986. (The location of these schemes relative to
the Kesem-Kebena plain is shown on Drawing Nr 1 in Album 1.) In the early 1970s
some fragmentary basin-wide studies threw doubt on the size of the irrigable
area on the Kesem-Kebena plain, indicating that it might be less than 10 000 ha
gross.
No further field studies were undertaken until this study, but a 1984 report by
the Bulgarian organisation Agrocomplect proposed a run-of-river development for
the Kesem and Kebena, irrigating 1 350 ha gross in the dry season and 6 500 ha
in the rainy season (Ref. 7, p.65).
In 1985 the Ethiopian Water Resources Development Authority (WRDA) issued terms
of reference for this comprehensive feasibility study of the Kesem Irrigation
Project (KIP). In January 1986 the work was entrusted to Sir M. MacDonald &
Partners Ltd. (MMP), under a contract with FAO, which in turn is acting for
UNDP. Work began in January 1986 and field investigations, including geological
and hydrogeological subsurface drilling and testing carried out by the
Government of Ethiopia, continued until February 1987. An Interim Report was
discussed in December 1986 and the Draft Final Report was submitted for
discussion in July 1987. This document is the Final Report.
Some changes have been made to the scope and timetable of the study during the
course of the work. Firstly, the area covered by the soil survey was raised from
17 550 ha (a figure derived from the 1965 report's proposed gross irrigation
area) to nearly 22 000 ha so as to cover the whole plain, the observation
density being maintained at eight observation sites per square kilometre.
Secondly, the geological drilling programme was extended in duration and was
1
complemented by geophysical investigations using a hired seismograph. Thirdly,
extra mapping was undertaken, to extend the 1985 mapping in certain places. The
overall duration of the study was extended by about seven months because of the
extension of the field work.
1.2 Objectives
The objectives of the project are to use all resources to maximise agricultural
potential under a balanced environmental and ecological system, producing cash
and industrial crops for export and/or import substitution, plus food crops at
least for local self-sufficiency (Ref. 14). This emphasis on maximisation rather
than optimisation is significant and will be discussed in later chapters: some
compromises are needed in planning the joint use of the resources of land,
water, human potential and capital.
The objective of this study is to enable WRDA to determine the relative priority
of the Kesem Project within the overall development of the Awash valley (Ref.
15). The study is therefore intended primarily for comparison with other
relevant studies, rather than for isolated appraisal against national or
international criteria, which is significant for the interpretation of the
economic analysis. It will be useful at this point to explain the use of the
word "settlement". Except for very occasional use to mean a place where people
live, like a village, it is used in this report to indicate all aspects related
to the present indigenous occupants of the project area. For historical and
institutional reasons, this usage is retained despite the fact that the semi-
nomad people are not expected to settle in the forseeable future.
An initiative which is closely linked to the Kesem Irrigation Project is the
Ethio-Bulgarian Agricultural Joint Venture of Kesem Kebena (EBJV), which is
described in Section F1.5 in Annex F. Formed in late 1984, it aims to develop
parts of the KIP area by run-of-river irrigation, without a dam. Towards the end
of the study period of this report, the EBJV took over the existing state farm
in the area, and has plans to extend it. The Joint Venture can become an
alternative project if KIP is not implemented, or a management and finance
package for KIP's first stage if it is: in either case the new field data
contained in this report will be valuable for the EBJV's planning and design
process. The link which has been established through EBJV's recent involvement
in KIP's Steering Committee should be actively maintained.
2
1.4 Acknowledgements
Although the Consultant takes final responsibility for this report, the study
has been a cooperative effort by many people. For the fieldwork in 1986 an
integrated team was formed by WRDA and MMP, consisting of about 40 professionals
and a larger number of assistants, technicians and administrators.
3
CHAPTER 2
2.1 General
Previous reports which refer to the Kesem-Kebena plain are:
the 1965 report on the 'Survey of the Awash River Basin', by SOGREAH
for FAO (Ref. 1);
the series of technical reports and assignment notes prepared by
Australian specialists from 1972 to 1974 under the overall title
'Development of the Awash Valley' (Ref. 3);
a prefeasibility study of 1980 for a proposed Ethio-Yemeni joint
venture (Ref. 5);
The first two of these reports give the results of field investigations, while
the other three consist mainly of the re-presentation of selected field data
from the first two, plus updated cost and benefit estimates and some new
proposals for action.
Many of the annexes of this report contain references to, and comments on, these
and other earlier reports. The rest of this chapter is devoted to a brief
general review of the more relevant reports.
4
2.2 The 1965 FAO/SOGREAH Report
The 1965 SOGREAH report (Ref. 1) covered the KIP area only as one part of the
whole Awash basin, but gave considerable attention to the Kesem dam as one of
two suggested major storages (the other being Tendaho on the Lower Awash). The
soil survey of the Kesem-Kebena plain had an observation density of about one
per 200 to 300 ha (compared with one per 12.3 ha for this study), and was
plotted at the relatively large scale of 1 : 100 000 using air photographs at
1 : 40 000. Soils were classified by suitability for irrigation using US Bureau
of Reclamation criteria, and some 18 700 ha on the Kesem-Kebena plain were
classified as Class II, leading to the proposal for irrigation of 17 550 ha
gross. This now turns out to have been an over-estimate of the area of suitable
land by a ratio of around 1 : 1.4. The reasons are complex, but one significant
factor is that the extent of highly saline and sodic soils is much greater than
was recognised in the low-density soil survey for the 1965 report.
For regulation of the flow of the Kesem, the 1965 report identified a potential
reservoir volume of 370 hm3, but this study has shown that the maps used in 1965
were inaccurate and a volume of 500 hm3 can in fact be stored. This was however
offset, in the 1965 report, by a relatively low estimate of the sedimentation
rate (2 to 3 hm3/yr compared with this study's estimate of 5.7 hm3/yr), and a
high estimate of the catchment's water yield (600 hm3/yr compared with this
study's 500 hm3/yr). The reservoir was seen as a seasonal rather than an over-
year storage (initial live storage 48% of average annual yield instead of 100%
in this study), so its regulated output was seen as unrealistically reliable.
The 1965 report accordingly estimated that the reservoir could irrigate 22 500
ha and provide firm power of 4.8 MW, both of which now appear too high. Before
completion of the dam there was to be a run-of-river phase irrigating "at most
2 000 ha of land."
The SOGREAH study identified two damsites within the 1 km long gorge on the
Kesem river. Either site would use the same reservoir. That study included some
drilling and permeability testing at the downstream site, which was preferred to
the upstream site because of the supposed better shape of the gorge. This now
turns out to be due mainly to inaccurate mapping. High permeabilities were
measured at the downstream site (probably related to a nearby fault) and were
countered by proposals for a very extensive, probably excessive, grout curtain
rather than by closer consideration of the upstream site. Both rockfill and
hollow gravity concrete dam types were proposed and the report left the choice
to be determined by later studies, though stating the expectation that rockfill
would probably be favoured.
In summary, the 1965 report identified the Kesem Irrigation Project and gave an
initial estimate of its extent and merits which, though qualitatively correct,
were quantitatively somewhat optimistic. The area of suitable soils and the
regulated yield of the reservoir were the crucial parameters which appear to
have been over-estimated. These comments should not, however, be interpreted as
criticism of the SOGREAH study; the database, especially topographic maps and
soil surveys, was very limited, and KIP was only one part of the 1965 report's
subject matter.
5
potential. Their usefulness is however very varied. One particularly significant
one was Informal Technical Report 23, by H.E. Voelkner, on the Afar pastoralists
and the problems of settling them on irrigation schemes (Annex C, p.C-47).
Voelkner's demographic work was especially valuable. On the subject of soils and
land suitability, Informal Technical Report 14 by D.T. Currey questioned the
1965 estimate of suitable land area, suggesting that the true figure was only
9 150 ha, little more than half of that estimate. The reasoning behind this
assertion was not well founded, however, and Currey did not carry out or have
access to data from any significant new soil survey. The report dealing with
organisational structure (Nr 6, by Donohoe; Annex F pp. F-3 and F-37)
recommended a single organisation dealing with agriculture and engineering,
which was the practice at the time.
Perhaps because of the weak basis of Currey's challenge to the 1965 estimate of
irrigated area, or because of the fragmented nature of the Australian reports,
the three desk studies of the early 1980s tended to go back to the 1965 report
for their base data. In the matter of soil suitability this is unfortunate,
since the time-honoured figure of 17 550 ha for the irrigation area appears
again and again.
6
The earlier-developed parts of the Amibara complex, particularly the Melka Sadi
area, have suffered groundwater and salinisation problems within a few years
after the start of large-scale irrigation. This has led to a pilot drainage
scheme and a master drainage plan, and further investment in drainage works is
expected during the late 1980s.
Feasibility study and design reports on these schemes appeared from 1969 to 1986
(Refs. 2, 4, 6 and 12), while the drainage problems and their proposed solutions
were covered by three reports in 1985 (Refs. 8, 9 and 10). These have all been
used in the course of this study, and have produced valuable information on the
Awash river, construction costs, and some other aspects. There is however one
fundamental difference between the Amibara system and KIP which limits the
transferability of experience between them. The soils in the Amibara system are
mostly basin clays and levee soils related to the Awash, whereas on the Kesem-
Kebena plain the predominant soils are those developed on alluvial fan deposits
related to the rivers and wadis flowing eastward from the rift valley
escarpment. For this reason, most of the conclusions reached in this report on
matters related to soils, particularly land suitability, drainage, and crop
yields, are based on the present Consultant's fieldwork and wider experience and
owe little to the previous reports.
Many other and more specialised reports have been used in the course of this
study and are mentioned in the relevant annexes. Material from the NOMADEP
project, which promoted development among the Afar on the Kesem-Kebena plain,
has been valuable, as have documents of the Ministry of Agriculture's Third
Livestock project and of the International Livestock Centre for Africa
(Annex C). Various government statistical publications, and also the reports and
budgets of the Ministry of State Farm Development and its branches, have been
valuable for agricultural and economic data (Annexes B and N). Agricultural
information was also of course gained from many reports of the Institute of
Agricultural Research. On organisation and management, a Halcrow - ULG report of
1982 was useful (Annex F). Specialist reports on seismology and power planning
are mentioned in Annex J.
In addition to all the reports and documents specific to the Awash valley or to
Ethiopia, a number of international publications have been used, notably the
Soils Bulletins and the Irrigation and Drainage Papers produced by the FAO.
Finally, the written sources have been supplemented by many interviews and
conversations with people having relevant experience, both in government
departments such as WRDA, MOA and MSFD and also in independent or United Nations
agencies such as FAO, NOMADEP and ILCA.
7
CHAPTER 3
The Kesem-Kebena plain covers an area of about 22 000 ha. It is part of the
floor of the Great Rift Valley, which in this area is oriented roughly north-
south and drained by the Awash river flowing northward. The plain lies on the
west bank of the Awash and is about 12 km wide, bounded on the west by the
dissected foothills of the rift valley's western escarpment. Figure 3.1, which
is bound at the back of this volume, shows the main features, and the wider
surroundings are seen on Drawing 1 in Album 1. To the north the plain is bounded
by Mount Dof an, a small volcano joined to the escarpment by a ridge running
east-west: this volcano pushed itself up right on the line of the Awash, causing
the river to deposit sediment upstream of the blockage and eventually to form a
new course round the volcano on its east side. The north-east corner of the
plain is formed of these deposits, characterised by very flat topography and
heavy soils, and tends to resemble the Melka Sadi area, part of the Amibara
Irrigation Project, on the other side of the Awash. The rest of the plain is,
however, mostly made up of alluvial deposits brought down from the escarpment to
the west by various watercourses, notably the Kesem and Kebena rivers and Wadi
T'unfeta. These two rivers are wide and braided where they enter the plain,
becoming narrower between their own levees, and eventually showing a tendency to
meander in the last few kilometres before joining the Awash. The wadis that
drain the foothills immediately to the west, which tend to have catchments
extending only 5 to 30 km westward, only flow briefly after storms and do not
reach the Awash at all. Wadi T'unfeta ends in a pronounced conical fan in the
north-east, while several wadis, end in an ill-defined basin south-west of
Gurmile Hill, separated from the Kesem by that river's left levee. To the south,
some 25 km south of Dof an, the plain is bounded by higher ground that extends
eastwards from the large Fantale Volcano to the Awash and, in the south-eastern
corner, by the Filweha stream at its foot. This stream is unlike any other
watercourse in the area because it rises in a group of hot springs related to
the area's widespread volcanicity, and carries a remarkably steady flow (1 to
2 m3/s) of sediment-free, warm and slightly saline water. In the middle of the
plain is a group of volcanic craters and their side slopes, collectively called
Gurmile Hill. This feature is about 2 km in diameter.
This study included a semi-detailed soil survey and a land suitability classifi-
cation which are described in Annex A and mapped at 1 : 20 000 in Drawings Al
to A4 in Album 1. The fieldwork, done from February to June 1986, covered some
21 820 ha with a total of 1 768 auger holes and pits, and also involved numerous
infiltration and hydraulic conductivity tests. The overall observation density
thus exceeded 8 per km2, and the auger holes were mostly 3 m deep, some 4.5 m.
Use of the 1 : 20 000 air photographs taken in advance of the study (1984)
enabled the large volume of data from these field investigations to be plotted
as soil maps (Drawings Al and A2) which record the very fragmented soils in some
detail.
This soil survey shows that the soils of the plain are very variable, both
laterally and within individual profiles, and that large areas are affected by
problems of salinity and/or sodicity. Most of the soils in the area are
6
developed on a complex pattern of intercalating alluvial fans. Along the banks
of the major watercourses, particularly the Awash and Kesem rivers, a narrow
trip of levee soils may be found. Relict levee soils may also be found along old
dry watercourses throughout the area, but their extent is very limited.
Soil textures vary considerably, reflecting the very complex recent geomor-
phological and hydrological history of the area. Alluvial fans continue to be
developed, as meandering watercourses repeatedly alter their courses. Air
photographs taken over the last 20 years indicate, for instance, that the
outwash fan from the Wadi T'unfeta has gradually moved south away from Dofan
Mountain. Extreme spatial variation in soil textures indicate that this
phenomenon has been common throughout the area for some time. Most of the soils
are predominantly medium textured: typically silty loams, silty clay loam, very
fine sandy loam and very fine sandy clay loams, with potentially moderate to
high water holding capacities. Although some of these soils are well or
moderately well drained, many are imperfectly to very poorly drained. In some
cases this is due to the presence of underlying heavier textured material, but
more often because of the massive nature and inherently poor physical
characteristics of the silty material itself.
The land suitability classification, unlike those in previous studies which used
USBR methods, is based on FAO guidelines published in 1984. The new approach
defines land suitability for irrigation in a manner specific to particular land
uses rather than in a general sense, as well as taking into account non-soil
factors like water resources, social constraints, and economic aspects. This
means that several classifications have been made, one for each of six potential
land uses (details in Annex A, Chapter A5). Since it would be impracticable to
map all six, and to use such maps for designing an irrigation layout, a
generalised classification has also been developed and this is mapped in
Drawings A3 and A4 in the Album. Its seven classes or 'mapping units', were
designed specifically for this project and are not related to any standard or
other classification scheme.
The results of the survey indicate that some 17 200 ha (79% of the area
surveyed and mapped) are at least marginally suitable for some form of irrigated
development, though some of this cannot be commanded for gravity irrigation.
However, for many of the land utilisation types under consideration the areas of
suitable land are significantly smaller, as shown in Table 3.1. It can be seen
that for cotton, the main cash crop and not a particularly demanding one as far
as soils are concerned, only 62% of the gross area, or 13 400 ha, is suitable. A
further 3 BOO ha can be used for irrigated pasture or woodlots but not for high-
value or annual crops, and even this only brings the total gross suitable area
to 17 200 which is considerably less than the 1965 reportes estimate for
undifferentiated irrigable land (18 700 ha).
9
TABLE 3.1
Summary of Areas Suitable for Irrigation by Land Utilisation Type
Adoption of skilled soil and water management practices will be crucial to the
success of irrigation development in the area. In addition to the practices
necessary to combat and prevent excessive sodicity and salinity, measures will
also be needed to overcome the poor physical properties of many of the soils.
Potentially the most serious of these include the tendency of the siltier
textures to form surface crusts (thereby reducing the rates of infiltration); a
general lack of good soil structure, and the inherently low soil permeability of
the heavier textures which are widespread as depositional bands or major soil
layers. These inherent deficiencies would be aggravated by poor soil management,
notably by inopportune timing of field operations, especially those using heavy
farm machinery which is liable to cause compaction and surface smearing in the
silty soils. Significant improvement of soils by good agricultural management is
not possible.
To summarise the position with regard to land resources, the soils of the Kesem-
Kebena plain have been found to be less good than was indicated by the 1965
study. The main problems are:
10
The soils are generally unlike those of the Amibara system, and crop yields
cannot be expected to match those of the Amibara farms. Nevertheless, there is
potential for irrigation, with appropriate cropping patterns and some recla-
mation, of the order of 12 000 to 16 000 gross ha on the plain.
subsurface flow from the volcanic bedrock in the west and south and
from underlying bedrock aquifers;
infiltration in the beds of the rivers and wadis that cross the plain
from the west;
11
Chemically, the higher groundwater has fairly high levels of sodium and
potassium relative to calcium and magnesium, and also high and variable fluoride
levels. In places the water is hot, locally exceeding 50°C, which gives rise to
the localised high soil temperatures noted in the soil maps (see Drawing A2).
All these properties reflect the influence of vulcanism. In the deeper aquifer
however, at depths down to 100 m, water quality is quite good: total dissolved
solids range from 250 to 1 000 mg/I and fluoride is generally below 1 mg/l. For
drinking water supply, wells should be situated near,the Kesem and Kebena rivers
to ensure acceptable fluoride levels. For irrigation, the deep aquifer's water
is usable, with medium to high salinity hazard and low to medium sodium.
There is thus good potential for the use of groundwater as a source of domestic
water, and yields of around 20 1/s can be expected from wells 100 m deep and
200 mm or 250 mm diameter. Use of groundwater for irrigation is also feasible,
but the quantity would only be of the order of 20 hm3 per year. This can of
course be pumped in particular months, to complement run-of-river surface
supplies, and the practicalities of this are discussed in Chapter 5 and in
Annex L in connection with the 'Small Project' development scenario.
The hydrology and sediment transport of the Kesem have been studied in some
detail, as is described in Annex I. There are now over 20 years of flow
measurements at the Awara Melka gorge, whose catchment is very little different
from that of the dam sites. After detailed study and revision of rating curves
as part of this study, they indicate a mean annual flow volume of 500 hm3 for
this period, which includes a sequence of particularly dry years since the mid-
1970s. For the regulated yield of a reservoir, the low flows and the shape of
the recessions are more significant than the mean annual yield, because in wet
years much of the water will spill anyway. Studies of the dam sites in the Kesem
gorge (Annex J) show that it is feasible to build a dam there to form a
reservoir of volume up to 500 hm3.
The 22-year sequence of monthly flows, derived from the Awara Melka gauge
records, was used to simulate the operation of the Kesem Reservoir under
32 different combinations of target draw-off and reservoir sedimentation (see
Chapter 17, Annex I). These showed that, with a 500 hm3 reservoir containing no
sediment, the drawoff could reach about 450 hm3 in most years. It was assumed
that the relative monthly distribution of drawoff volumes would follow the
pattern of the proposed KIP cropping systems, but the absolute volumes would be
maximised. Exploitation to this relatively high level (90% of mean annual yield)
would involve occasional water shortages, some restrictions in about one year in
four, and a severe shortage about once a decade. The impact of shortages would
be easy to minimise because they would be predictable several months in advance
and crop plantings could be modified. Reducing the target drawoff to
350 hm3/year would increase the reliability so much that a water shortage would
be very rare - one in 20 years or less, but average agricultural production
would be much reduced.
The reservoir would, however, not remain free of sediment for long. Estimation
of future sedimentation rates is very difficult, but for this study a thorough
investigation has been made, involving sampling and photograhic analysis at many
12
points in the catchment, with access by helicopter. An intense suspended
sediment sampling programme was carried out by a resident WRDA team at Awara
MeIke during the 1986 wet season, and data from other reservoirs such as Koka
were also used. Because no one method is reliable, nine new estimates were made
and compared with three from previous studies. They gave widely differing
results (0.6 to 11 million tons per year), but the methods are not all of equal
relevance. Taking the average of the four estimates most likely to be correct
for the Kesem, this study arrives at a working estimate of 8.3 million tons,
equivalent to about 5.7 hm3 per year. The proportion made up by bed-load is
relatively high, 30% to 40% according to the estimates that make specific
forecasts for bed-load. These figures compare with the apparent estimate of the
1965 report (Ref. 1) of 5.4 million tons per year.
The effect of sedimentation on the useful regulated yield of the Kesem Reservoir
has been estimated from the 32 cases for which its operation was simulated for
22 years. The simulation covered silting of up to 60% of the reservoir volume,
which would reduce the achievable target regulated yield, at constant relia-
bility level; from 450 hm3 to 300 hm3 per year. If the forecast siltation rate
of 5.7 hm-qyr is correct, this condition would be reached in about 50 years.
With the overall cropping pattern and irrigation efficiency of the proposed KIP
this means that the new reservoir could irrigate around 17 000 ha, while
sedimentation would reduce this to about 13 000 ha in the first 50 years.
The simulation of reservoir operation shows that sequences of dry years are very
significant: unless the reservoir was being worked excessively hard, the 1982 to
1985 dry sequence would have been more severe in its effects than the isolated
very dry year 1972. The reservoir would therefore normally be used as an
overyear rather than merely a seasonal storage, having a live storage
approaching the mean annual flow.
In the reservoir area there live about 200 people, called Soudanis, who are
pastoralists and who also practice irrigated and rainfed agriculture. Many of
the men also take paid employment.
For the purposes of a large irrigation project, these present inhabitants do not
represent the relevant human resources for the development. The human resources,
at least in the short term, are the highlanders who would come down into the
rift valley, as permanent workers with families or as migrant labour, in the way
that they already come to the existing state farms on both sides of the Awash
river.
13
Although they do not usually have a long tradition of irrigation behind them,
they do come from a very well-established agricultural tradition and are well
suited to the work. The population of the highlands is large and expanding fast,
so there is in effect an adequately large source of workers for Middle Awash
projects. But it is not easy for farms in the Middle Awash area to attract
suitably motivated people, either as managers, technical staff or labourers. At
present, on existing state farms in the region, the standard rate of pay for
labourers is Birr 2.00 (about US$ 1) per day, and housing for migrant labourers
is of low standard. Mobilisation of large numbers of labourers for KIP will
require an improvement in one or both of these factors.
The main part of the state farm, at Awara Melka itself, is an old scheme that
has been extended over the decades. It was started in 1905 by a Frenchman whose
name, Saboret, is now borne by the village that overlooks the farm and is almost
synonymous with Awara Melka. For a while it was under private Italian owner-
ship, and in the mid-1970s it was taken over by the government as a state farm.
The farm is described in Annex L (Section L2.4). It draws water from a rudimen-
tary gated headworks structure and partly collapsed weir in the small gorge at
Awara MeIke, from which the flow passes through a short feeder canal to the
scheme's north-west corner and thence eastward in a main canal along the Kesem's
right bank levee. The scheme once covered 1 400 ha, but the effective area is
now less than 800 ha because of salinisation and abandonment of some areas, and
because of water shortages. The rubble bank that is built each year in the gorge
cannot divert all the Kesem's low flow into the intake, and distribution is poor
because there are few gates and no measuring structures. The main crops are
fruit trees (270 ha) and tobacco (300 ha), the latter being planted continuously
on the same land which implies a risk of disease buildup. There is no drainage
system and, over most of the scheme, no watertable problem.
The Yalo unit of the same state farm is much younger, having been started in
1963 and having a present effective area of about 400 ha: over 600 ha was
developed, but some has been abandoned because of salinity and flooding. The
crop is cotton and the yield is low at about 1.5 t/ha. Pumping from the Awash
was considered but never done, and the Kebena's flow dwindles to nothing, or
very little, in most dry seasons.
The main official "settlement scheme" is at Doho, on the eastern end of the
Awara MeIke farm and receiving water from its main canal. It dates back to the
drought of the early 1970s, when a relief settlement was set up at Doho,
involving distribution of food rations. Self-help irrigation was introduced with
the help of NOMADEP, the French aid project serving the Afar, based on family
smallholdings. This was then taken under the control of the government's
settlement and famine relief organisation (RRC) and converted into a collective
farm. The area cultivated, once 356 ha, is now between 60 ha and 80 ha, and the
14
PLATE 1
Satellite Image
of the Project Area
#
. (January1976:false colour)
PLATE 2
The third form of irrigated agriculture is the spontaneous schemes set up by the
Afar. Annex C contains details of five such schemes on the plain, all near the
places where the Kesem and Kebena emerge from the hills (and mostly just outside
the proposed KIP scheme). They are very small, and are significant mainly for
the precedent that they represent for irrigated farming by the Afar. Their
hydraulic arrangements are extremely crude and vulnerable to floods, but with
periodic repair the water usually reaches the fields and cropping is fairly
successful. Most produce is food, especially maize, but some cash crops are
grown. Cultivation is entirely by hand, except where the Soudanis bring in ox-
ploughing.
These Soudanis, 200 of whom live in the proposed reservoir area in an isolated
and ethnically distinct community, practise both rainfed and irrigated
agriculture as well as keeping livestock. They are generally enterprising
people, and have become involved to some extent in Afar irrigation down on the
plain. In one place the Afar pay the Soudanis in cash to plough their land with
oxen. In another (on the Kebena) some 16 Soudani households are farming on Afar
land in co-operation with 7 Afar households, the Afars having, invited the
Soudanis to join the venture because they themselves lack most of the necessary
skills. This is the only, but not the first, such scheme and appears successful
after 2 years. The Soudanis are willing to join it because they are short of
good irrigable land in their own area, where they irrigate 15 ha to 20 ha from
the Kesem, growing maize and other food crops.
In summary, the existing agriculture of the area is characterised by deficient
infrastructure and declining cropped areas, apart from a few recent tiny
schemes. The total cropped area is around 1 300 ha, less than a tenth of the
available land on the plain.
The state farm has recently become part of the Ethio-Bulgarian Joint Venture
described in Annex F (Section F1.5). It proposes to increase the run-of-river
area to 6 800 ha (which implies low cropping intensity) and to prepare for
eventual development of the whole plain, but work has not yet started.
15
to the north there are no all-season tracks, so that Yalo farm tends to be cut
off at times during the rainy season. There is a ford across the Kesem at
Saboret and another one 15 km downstream, not far from the confluence with the
Awash, but both are impassable for days at a time during or after floods. The
downstream ford links with a track from Awash town, which becomes very muddy in
the rains. There is also a third track to the highway, south-eastwards from
Saboret via the Filweha springs, but it passes through the National Park and is
seldom used. As a result of this study there are several rough tracks to and
around the dam site, on both banks.
The main settlement on the plain is Saboret (or Awara Melka), which is on the
slopes of the easternmost ridge of the rift valley escarpment and just south of
the Kesem river. There is an old fort-like structure, some more recent offices,
and a few hundred houses of traditional construction. Further out on the plain
there are the state farm's staff houses, built of blockwork and with corrugated
roofs. The only other modern-style buildings are the few sheds and offices of
the settlement authorities at Doho. Yalo is a small settlement of traditional
houses strung out along the right levee of the lower Kebena river. The Afar have
mobile dwellings of sticks, grass and animal skins.
Public services are almost entirely absent from the area. There is no post
office or telephone. Some of the state farm houses have a simple water supply
but most inhabitants of Saboret take domestic water from the canals, while the
Afar use rivers and a few dug wells. The state farm management, in the absence
of other civil authorities or police, is the effective representative of the
state.
Relative to other services, the provision of health care is slightly better than
one might expect. Detailed studies for this report are described in Annex E. The
state farm runs two clinics (at Saboret and Yalo), while the Ministry of Health
runs two ex-NOMADEP clinics at the same places and a smaller clinic at Doho -
these Ministry clinics are however not open every day. The main diseases treated
are malaria, respiratory infections and dysentery. Maternal and child health is
largely neglected, and the clinics are understaffed. The Afar have traditional
midwives who provide not only services related to childbirth but also general
medical help: some of them were trained by NOMADEP and their skills represent a
base on which part of any future medical service to the Afar can be built.
16
grabens), which the river breaks through in a series of gorges (and which
incidentally make east-west road building impracticable). Further north, in the
catchments of the wadis that disgorge onto the plain, the landscape is dominated
by more normal dendritic watercourse patterns between branching ridges. The
western edge of the plain is interrupted at many places by north-south scarps
and intrusive ridges.
These surrounding areas are important for the project not only as the source of
storm runoff and sediment but also as the wet-season grazing lands of the Afar
and Soudanis, who use some 1 500 km2 to 2 000 km2 for this purpose. To the
south-west their grazing areas meet those of the more aggressive Kerayu tribe.
All these pastoralists take their animals to the wet-season grazing lands as
soon as new grass begins to grow there each year, leaving some people and a few
animals on the plain. They tend to return after the rains in response to
shortage of drinking water for livestock, rather than staying longer and
exhausting the grazing completely.
The Awash National Park, some 20 years old, aims to conserve the fauna and flora
of a sample of the semi-arid environment. It succeeds in doing this in its
central and south-eastern parts, on both sides of the railway and highway that
run from Awash town to Metahara. Parts of its nominal area, which extends north-
west to Fantale volcano and north almost to Saboret, are however not under the
park authorities' effective control. In the west, particularly, the Kerayu both
live and graze their animals within the park's boundaries. In the north the Afar
of the Kesem-Kebena plain tend to bring their animals into the park occasion-
ally, but their dwellings remain outside it.
These incursions reduce the effective area of the National Park, but the
pastoralists' livestock are using grazing that would otherwise mostly go to
waste, as does much of the grass in the effective part of the park.
3.7 Institutions
Government departments already involved in the Kesem-Kebena plain are the
Ministry of State Farm Development (MSFD) as operators (until very recently) of
the state farm, the Ministry of Agriculture (MOA) as successors to RRC in taking
responsibility for the settlement farms, and the Water Resources Development
Authority (WRDA) which operates river gauges and organised this study. Existing
large-scale irrigation schemes on the other side of the Awash bring together the
same three departments (WRDA is part of the Water Resources Commission, which
does not come under any ministry and reports directly to the Council of
Ministers, so that for many practical purposes it is like a ministry). The three
are independent of each other and each has its own local offices, MSFD and MOA
operating these through intermediate regional offices. The situation is
described in Annex F and illustrated diagrammatically there by Figure F1.1.
Until 1981 the water resources, agricultural and settlement functions were all
exercised by a single integrated authority in the Awash valley, but then the
present vertically separated structure was introduced. This present structure
has advantages in that the different objectives of the various functions are
clearly identifiable at all levels and choices on priorities and emphasis can be
made at national level. It also has disadvantages at local level, particularly a
frequent lack of communication, co-ordination and co-operation among the local
offices of the three organisations. These disadvantages can be counteracted, for
instance by the setting-up of a non-executive project-level co-ordinating
committee, and on the Amibara scheme changes are in hand to improve co-
ordination.
17
Towards the end of the information-gathering phase of this study, in late 1986,
the Awara Melka State Farm was handed over formally by MSFD to the Ethio-
Bulgarian Joint Venture. This undertaking, formed in 1984, is described in
Section F1.5 of Annex F. Its objective is to develop the Kesem-Kebena plain for
irrigated agriculture, starting with run-of-river schemes, based on the existing
farms but larger. By the time of writing this report the institutional change
had not yet had practical consequences in the field.
18
CHAPTER 4
4.1 General
This chapter discusses the options that have been analysed in the course of the
study, describing the choices that have been made and the reasons behind them.
It is left to Chapter 5 to describe the project.
In accordance with the Terms of Reference and with the objectives of the
Ethiopian Government, the starting point for the formulation of the project was
the maximisation of the use of the resources described in the previous chapter.
The area covered by the soil survey was extended by almost 25% to take in all of
the Kesem-Kebena plain, some 21 800 ha. The topographic mapping was also
extended to cover the whole reservoir area and the whole plain. The soil survey
and land suitability classification showed that only about 17 000 ha of the
plain is suitable for irrigation, and some of this is out of command. The
maximum practicable net irrigated area is about 14 000 ha, of which only about
10 000 ha is suitable for cash crops. The project was initially formulated to
use all of this. The reservoir formed by the highest practicable Kesem dam was
found able to provide more than enough regulated flow for 14 000 ha of irrigated
land, at least at the beginning of its life before its effectiveness is reduced
by sedimentation. This led to the formulation of the so-called 'Large Project',
which is the development to which most of this report refers.
The maximisation of use of resources does not, however, necessarily lead to
optimum development. Reductions in the size of the dam do not offer useful
economies, but it was found worthwhile to investigate the merits of irrigating
less than the maximum area of land on the Kesem-Kebena plain. This was done by
formulating an alternative scenario involving the irrigation of only the better
lands on the plain, about 9 000 ha. This scenario is referred to as the 'Medium
Project', and is of course only one of any number of intermediate cases that
could be formulated. To investigate the possibility of developing irrigation on
the plain without a dam, a third scenario called the 'Small Project' was also
roughly formulated, but this was outside the scope of the data collection work
of this study so that the formulation is less precise than the other two. All
three were considered in the economic analysis (Annex N).
When the Kesem Reservoir can provide more regulated flow than is needed by the
Kesem Irrigation Project (KIP), it can be used to increase the area irrigated by
gravity or pumping from the Awash river, into which the Kesem and also all KIP's
drainage flows discharge. Although the details of such downstream irrigation are
outside the scope of this study, it is considered in the evaluation of KIP.
The main components of KIP, arising from the approach outlined here, are:
a dam, in a gorge some 10 km west of the plain, for seasonal and over-
year storage of the flow of the Kesem;
an irrigation system leading water from a simple diversion weir on the
Kesem to all the usable land on the plain;
a drainage system to remove excess rain and irrigation water, to
control groundwater levels, and to enable saline and sodic soils to be
leached;
19
flood control works to protect the irrigated lands from the large
rivers and also from the wadis that enter the plain from the foothills
of the rift valley escarpment;
20
TABLE 4.1
Summary of Crops Considered
Crops for Crops for Crops considered
state farms settlement areas but found unsuitable
Citrus Pasture Mangoes
Tobacco Cowpea Bananas
Cotton Groundnuts Sugar
Wheat Sesame Rice
Maize Wheat Kenaf
Woodlots Maize Safflower
Teff
Haricot beans
Tomatoes
Citrus is already being grown in the area and shows the highest returns, of the
order of Birr 10 000 per hectare per year. Oranges and grapefruit are the best
types. Market considerations limit the area to about 400 or 450 ha, and there
are sufficient suitable soils for this: export is not a practicable proposition
because of quality problems and heavy competition on the world market.
The next most attractive crop in terms of gross margin is Virginia tobacco,
which can serve as a substitute for imports from Zimbabwe and Malawi. A small
area is already being grown at Awara Melka, but not on the most suitable soils
and with a continuous farming system which in the long run would be susceptible
to disease problems. On the plain as a whole there are enough soils usable for
tobacco to crop nearly 2 000 ha every year on a 2-year rotation, but they are
fragmented in a way that makes it difficult to design a manageable irrigation
layout for more than 3 000 ha net, i.e. 1 500 ha cropped each year. The market
is also limited, and the estimated long-term maximum marketable from KIP is also
about the production from 1 500 ha annually, Thus, for this crop, the soil and
market constraints coincide and result in the selection of an area of a little
over 3 000 ha, with a 2-year rotation and therefore about 1 500 ha in production
every year. For good utilisation of curing barns and other facilities it is
proposed that planting and harvesting should proceed continuously throughout the
year: with the crop in the ground for 4 months out of the 24, only one-sixth of
the tobacco land would be occupied by tobacco at any one time.
The project's main potential export crop is cotton. It is already grown
extensively in the region, particularly on the other side of the Awash river,
and the market for high-quality hand-picked cotton is effectively unlimited. It
can be grown on the same land every year, and is reasonably tolerant to saline
and sodic conditions. This enables it to be grown on all the area suitable for
annuals that is not chosen for one of the priority crops, citrus or tobacco.
Land preparation and cultivation can and should be extensively mechanised,
because labour in the area is scarce and the cropping calendar leaves limited
time for the preparation. Mechanical picking is possible and has been
experimented with in the area, but it suffers from several disadvantages. It
does not allow the picking to be spread in time, so that early and late-opening
bolls are lost. It requires different varieties which are not available for the
area, and different cropping techniques including the use of defoliants. Ginning
costs tend to be higher due to the unclean seed cotton. Finally it results in
lower quality than hand-picking, and it is the high quality of hand-picked
Ethiopian cotton which the export market particularly values. It is therefore
recommended that cotton should be picked by hand, even though this requires
large numbers of labourers to be brought from other parts of the country.
21
The tobacco crop occupies the land for only 4 months in 24, and the cotton crop
for 6 or 7 months out of 12, although there is also a period needed for land
preparation outside these months. There is therefore scope for growing break
crops within the cropping systems dominated by the two main crops. Figure 4.1
shows the crops chosen and the proposed cropping calendars, which have been
chosen for the reasons set out in the next two paragraphs.
In the case of the cotton system, the time is limited but it is possible to grow
wheat between November and February. Because of the time constraint this will be
limited to half the area, so each field within the cotton-wheat system can grow
two crops of cotton and one of wheat in every two years. Wheat is chosen because
it can grow in the cool season, is valuable as a food crop, and can use the same
soils as cotton.
In the areas devoted primarily to Virginia tobacco there is time for more than
one break crop. The first choice is cotton, because of its high return, its
value as an export crop, and its suitability to all the soils likely to be
chosen for tobacco. It would be grown at its optimum time, that is mid-April to
December, using all the land vacated by tobacco in the 12 months ending each
April: with a 2-year rotation for the tobacco this means half the area every
year. This leaves gaps of several months between tobacco and cotton crops, but
the timing varies. Land vacated by tobacco between May and October inclusive has
a winter gap of 7 to 111 months which is suitable for growing wheat, and this
crop is chosen for the same reasons as in the cotton-wheat system. Land vacated
by tobacco between November and April inclusive would then have a cotton crop,
as stated above, leaving a gap of 81- to 12 months before the next tobacco crop.
Although some of the cotton would be off the land by mid-October, some of it
would only be harvested in December: in theory a part of the land could be
planted to wheat, with the same quick turn-round as is proposed for the cotton-
wheat system, but this would be very difficult to manage in an already
complicated rotation, so the whole of this area is proposed for another crop.
The one chosen is maize, which is a valuable food crop and can be grown between
December and May, leaving enough time to prepare the land for the following
tobacco crop.
The cropping intensity for each of these two cropping systems, which are
proposed for state farms, is 150%. The cotton-wheat lands have two cotton
and one wheat crop per 2 years, and the tobacco system lands have either a
tobacco-wheat-cotton or a tobacco-cotton-maize rotation, thus also three crops
in2 years. The existing irrigated state farms have not as yet practised
multiple cropping on a large scale, but the Ministry of State Farm Development
(MSFD) recognises that such farming will be necessary in the future and is
ready to begin developing the appropriate management and agricultural skills.
An intensity of 150% is considered to be near the limit for practicable
application, and even this level will not be easy to achieve. The calendars on
Figure 4.1 show that all lands would have a fallow period of 4 months or more in
alternate years, which should be sufficient for maintenance of irrigation and
drainage systems.
Table 4.2, taken from Annex B, summarises the estimated financial crop gross
margins of these state farm crops. The state farm areas will inevitably, because
of the fragmentation of soils, include a proportion of land that is not suitable
for any of these crops: mapping unit 4 in the KIF general suitability classifi-
cation. Where they occur in large areas, such soils will not be devoted to state
farms at all, but to settlement areas. The small areas surrounded by better
soils cannot be used in that way. Such patches of land would be bordered by
canals and drains anyway, even if unused, so in accordance with the objective of
maximum use of land resources it is proposed to plant woodlots on them. These
would normally be irrigated, but because the trees would be fairly tolerant of
22
TABLE 4.2
Summary of Gross Margins for State Farms
(Ethiopian Birr at 1986 financial prices)
TABLE 4.3
Summary of Gross Margins for Settlement Areas
(Ethiopian Birr at 1986 financial prices)
23
water shortage, and of relatively low value, their irrigation can be stopped
during times of peak requirements or in seasons of water shortage. Their
inclusion in the scheme therefore increases the scope for flexibility in
management. The fuelwood they produce will play an important role in preserving
the environment in the surrounding areas, by satisfying the demand for fuel for
domestic cooking and also, potentially, for tobacco curing barns. Several tree
species have been identified, including some, like Acacia nilotica, which are
already growing in the area.
Gross margins for the crops considered suitable for settlement areas are given
in Table 4.3. For settlers and part-time pastoralists the return to labour will
be of more interest than the return per unit of land, and crop choice will also
be governed by tastes and habits and by the local market. Cowpea is popular
because of its relatively low labour requirement, ability to grow in almost any
season, and versatility as a food and fodder crop at a range of growth stages.
Groundnuts are favoured in the diet and are presently considered too labour-
intensive, but with higher yields this might be seen as worthwhile. Sesame is
suitable for cultivation for its edible seed. Wheat would probably be grown on a
part of the settlement arable area, but maize, though grown to a small extent by
settlers now, is not attractive in terms either of return to land nor return to
labour. The Afar should begin to use animal power for most land preparation and
cultivation; there is already one camel trained for draft work, and oxen can be
used in the usual way. This will require a small input of training and
development work by outsiders. Some Afar elsewhere are already using tractors,
but until the KIP Afar learn mechanical skills this is not recommended because
of the resulting dependence.
Among the crops not recommended, bananas and mangoes are being grown now on the
Awara Melka State Farm. The area of bananas is however quite small, yields are
not good, and the relevant ministry had already decided to centralise the Awash
Valley's banana production elsewhere. It is however possible that recent changes
in development plans for the Lower Awash will cause a reversal of this decision,
in which case a small area of bananas would be included in KIP. Mangoes on the
other hand may well have some long-term potential in the KIP area, when
sufficient research has been done and markets identified.
Sugar is grown not far to the south on the Ethiopian Sugar Corporation's
Metehara estate. Agronomically, sugar could be grown on KIP, but the distance to
the Metehara factory (35 km and more) is too high for processing there, while
the area of really suitable soils at KIP, and the likely yields, do not justify
a new estate and factory. The world market price is relatively low, and at the
national level a further development at Finchaa, under more suitable conditions,
is planned.
Rice could grow well at KIP, and could use considerable areas of low-lying land
with heavy soils that are not suitable for other annuals. It is excluded from
consideration for this project because of the lack of research information on
rice culture in the Middle Awash region, and the lack of a large internal market
because of dietary habits. The areas it would use can also be used for pasture.
Kenaf could, in principle, be grown as a substitute for imported jute but it is
not included because of the lack of experience of its production and processing.
Teff, the staple food of highland Ethiopians who grow it as a rainfed crop at
higher altitudes, cannot be recommended unless or until a variety is found or
developed which can do well at KIP's altitude and under irrigation. Haricot
beans and tomatoes both show some promise, but research and variety trials are
needed before they can be recommended for large-scale production: the beans can
be exported but the tomatoes might face market limitations. All these crops
24
Figure 4.1
Cropping Calendar .for N r -
100%
F M A M J J A SONO J F M AIM J J A S OHIO
114
I
- -,":/,
WHEAT
tts .
.
......
.....
.
.
.
..
.
.
.
-
t
. . s.....
. . . . .
. .
-
.....
. -
75% 1
. . . . . . .
, .. . 1.. - %
.
.
.
.
.
.. .. -
.
. : -
.
..
.
410,
S,
- .. .... . ...
.
4
. .
-.
WHEAT, .
-
.
1.
sP
41,
:Ito
50 °/0 it
At
0.. #
:4,40,
*St 10%.
t, IA.* '
4.0
WHEAT t
25%
444* *v .
044 . 4 . 44
ttm.
s 4%A0:*?
SSA .44
02,4Vii
.4°14 44444
Ptr 4K,41
Ps," k. ..44
tat' Zbrattedkab,
100% e 00-loo
J
04-
I F
o-,
M A
-
M J J A
.
S
. .... .
O N O J. F
,-,
M
, ..,.
A
,
M
,
J J A S 0
(TOBACO.°0. ../-\'(....\--
.,\--;.---.`,\ ,A7
\- .
00 °o- . -
o
:"- .
COTTON
. .
I-
\-s
../..,../.\,./..\./.
...\, Y,\,--,I,--.',
--_,,
, -I(- MA ZE ,
---;\ ;.(::1-..,-s% r` -
o -
- o
o
o
-,
75%
... .. /" \=',
\ \Ì\ -'\I
i i
.
o °°
o
.0
0
:*4"40 V'
ko
: 'V
.
.
.
. ...
.
.. .. - . -.
. .
"
COTTON
. .
. -. -
.
.
.
. . -
'
.
o o°0 0.
0 00 : 0000:
oo
"4
P ' 1si.
4411.
1OP 4 1
4141
.
.
-.
.
. - .
.. -.... ' l -
. .
.
,
-
.
.
.
.
.
0 0*
°TOBACCO:
,.
*
0
. o' ,.,*o . o
0 *0 * °°
0
°
WHEAT
0,. o 0-0° ,',</
50% ^ s
o'oci0-°.; 0°0 " -
. .
o 0 o.o ,
\'.... :"\:,\:,.........\;;:\/,........
% '7/ \ 7/ \ ,N, 7\ '., .
,,:: 0 o
oo
. o .0* eo .° 10 0 o°1*
7:v7-%`;7'\/,',J.,/,,, 00 e o : 0. 000 0,:),, o - - - - -
\--\ s''. MA.ZEN,, ;\,,- %. 0 *0 ° . 00 0 1 i -: .*
.....
. . . , C 0 TT ON I . .
/..\/......\-,.. --.:-.\-.1-\\- 0 0 00 0 .
.
.
. .
.
.
.
.
.
. .
. t. . . .
.
. . . . : : :
*0 ,o
25%
...,\E.\/i 1/..\<r
,..,/
:' /" .'{* \ Z....
i'''/..\ 1' ....`
. ' *
'
o° 0*0
0 o 0* 0 0 * * e 0
* . *0 °* 0 ° .
. . . .
. .... .
. . .
.
* Th 0
PPRIVP,4""
V.".
* 00 O.
*
%o 0 0 0
* . . .
° . - . . . .
o... 000 o
444ftt r,
. -
/ 4tp, "4 . . . . . .
00/0
A. A*. 4,
Mts..satioliss-4'
i
- " .
-
. -
25
The 12 000 Afar in the KIP area are well aware of this history and are still
suspicious of large-scale irrigation development. However, they have the
experience of some very dry years in the recent past which have convinced many
of them that they cannot continue indefinitely to rely entirely on livestock,
and they have the example of the small initiatives which a few of them have
undertaken in irrigated agriculture, supported by an outside organisation
(NOMADEP) in the 1970s and to some extent by the existing small state farm in
the area. They have in one place called on the Soudanis to help and advise them,
and they are also receiving significant assistance from a few individuals at the
agricultural research station.
The project will also need to concern itself with the wider area used by the
Afar for grazingz which, though not precisely bounded, is estimated to extend to
roughly 2 000 kmL. A range and environment management programme will be needed,
to improve grass species, resist tendencies to deforestation, and protect the
environment from other pressures of increasing population. Simple measures to
improve the retention of rainwater should be used, in conjunction with the
efforts to improve pasture.
26
4.4 Irrigation, Drainage and Flood Protection
4.4.1 Irrigation System
The main factors governing the choices of irrigation method, layout and design
are:
the fact that water is relatively plentiful in comparison with land;
the proposed crops;
the nature and distribution of the soils;
the land slopes and microtopography;
the institutional arrangements;
the need to limit the number of places where irrigation
water is conveyed across the eastward-flowing rivers.
The detailed discussion of the choices proposed for this project, and the
reasoning behind them, is in Annex L. After consideration of sprinkler, trickle,
sub-irrigation and five types of surface irrigation, it is proposed that the
annual crops should be irrigated by sloping furrows. Level furrows or basins
would require too much land levelling, sprinkler or trickle would require
pumping which would increase cost and complexity, and soils are not suitable for
corrugations. Border strips could be used, but furrows are well-proven in the
area and familiar to managers. For citrus it is recommended that furrows should
be used, the trees being grown on beds between widely-spaced furrows. This will
require less labour than the present small-basin method, and will also improve
access and help to avoid diseases associated with standing water in contact with
tree-trunks. For irrigated pasture, to keep capital costs low, a simple
controlled flooding system is proposed: water would be released from small
contour canals and would flow down the slope, any tendency to concentrate canals
being countered by short, level spreader furrows. These will be periodically
renewed or added to in the light of the local behaviour of water flows.
The operation method of an irrigation scheme needs to be well matched to the
design of the physical works, although the latter should wherever possible be
adaptable to different operating methods if future changes make them desirable.
For this project it is proposed that the operation should be as simple as is
consistent with good efficiency. It is proposed to base the layout and the
operation on a basic land unit of standard size 22.5 ha net (25 ha gross), which
will make the rotation of irrigation water supplies relatively simple. The
larger canals will flow continuously and the smaller ones be rotated, so as to
avoid operating canals with very low discharges relative to their design
condition.
Most cross-regulators will be long weirs, which operate automatically without
any complex control gear and thus need no adjustment and little maintenance.
Only the larger cross-regulators will have gates and need to be monitored and
adjusted. Where a smaller canal draws water off a larger one, the head regulator
of the smaller one will be a movable weir, which provides for adjustment and
measurement of flow according to need. This is considered preferable to the use
of constant-head orifice gates. Head regulators of field canals are simple gated
structures, since they only need to pass about 100 1/s.
27
The use of lined canals has been carefully considered, but soil permeabilities
are relatively low and water loss by seepage is not a serious problem because
drains are provided anyway and water will often need to be passed to the Awash
river for downstream schemes, so that drainage water will be re-used. In some
places, canals will need lining over short distances through sandy patches, and
this can best be done with clay imported from drain excavations in suitable
soils. Concrete is used for the first stretch of the North Primary Canal, where
it passes out of the Awara MeIke gorge and on to the plain.
The layout of the irrigation system has involved some compromises, as discussed
in Chapter L3 of Annex L and shown on the drawings. The fragmentation of the
soils results in a very patchy map of suitable land. The horticulture area is
small for market reasons, and the first effective priority was to find about
3 000 ha suitable for tobacco, so as to harvest about 1 500 ha of tobacco each
year in a 2-year rotation with other crops, this being the estimated market
limit. This proved just possible, with three tobacco-system farms of which two
draw water from the main canal system in only one place each. The third tobacco
farm will have to draw water in two places. The remainder of the land suitable
for annual crops was nearly all allocated to cotton and its break crop wheat,
while large blocks of land of suitability mapping unit 4 (good only for pasture
or woodlots) were allocated to the so-called settlement areas for the Afar and
Soudanis: the extent of such areas matches well with the estimates of areas
needed for irrigated pastures and small gardens. Because of the fragmentation of
soils, there are small patches of mapping unit 4 land enclosed within the state
farm areas, and small patches of better land enclosed within the settlement
areas. As mentioned above, the former are assumed to be used for irrigated
woodlots, since they will be adjacent to canals and drains anyway and can thus
be served at minimal cost; fuelwood will be needed for domestic and crop
processing purposes, although partially replaceable by hydro-electric energy
from the dam, and a plentiful supply will be a valuable factor in combatting
deforestation of the surrounding country by the increased population. The latter
areas are expected to be used for the gradual adoption of irrigated agriculture
by the Afar, for whose purposes the fragmentation of the soils will not be as
serious a disadvantage as it would be for the state farms' large-scale
operation.
One objective of the layout design was to permit handover of water to each user
organisation, such as a state farm or one of three settlement offices, at few
places, preferably only one. This is valuable for the establishment of clear-cut
divisions of responsibility, and for measurement of water and collection of
payments. Another objective was to place the settlement areas on the periphery
of the scheme along the north, west and south sides, so that the Afar can move
to and from the surrounding hills without crossing the state farms; this proved
quite easy to achieve.
It is debatable whether it is worthwhile to take a canal across the Wadi
Tiunfeta in the north of the plain: the soils north of it are generally poor and
the north-eastern corner is also subject to flooding and cannot economically be
provided with deep drainage. In the area-maximising scenario, the Large Project,
this northern area of 1 460 ha is used entirely as a settlement area, and is
good for this purpose because it is on the periphery and has a natural border to
the state farm area in the form of the wadi itself and a floodway. When the
objective is selective land use, as in the Medium Project, it is considered
better to terminate the canal system on the south side of the T'unfeta and save
the cost of the inverted siphon. This still leaves the land on the north side
for the Afar, but without formal irrigation. The primary canal must however
terminate in an escape structure, and it would be easy and advantageous to
discharge some water, except at times of water shortage, into the wadi to be
used by local informal canals on its north bank.
28
To the south of the irrigation area is the Filweha Spring, which produces a
remarkably steady flow. The possibility of using this for irrigation was
considered but rejected. The water, though hot, is probably just usable for
irrigation from a chemical point of view, provided care were taken. But the
location of the spring is too low for gravity irrigation of any but a very small
area. Since KIP will normally be passing water into the Awash river anyway, it
is considered better to leave the Filweha Spring water to flow into the Awash
and be diluted by normal surface water, as is already happening, after which it
can be used for downstream irrigation. This will avoid any possibility of
problems caused by any long-term changes in the chemical composition of the
spring water.
This report assumes seven state farms for the Large Project. In practice these
could be separate units of one or two state farms, with negligible difference in
staffing costs.
29
Water drained from the eastern parts of the plain will need to be pumped up into
the Awash or its tributaries at times of high water levels in the Awash, which
persist for days or weeks in most years. The pumping head is minimised by
avoiding deep drainage in some low areas, as just mentioned, and the volume
pumped is minimised by separating the whole drainage network into gravity-
drained networks for the higher western areas and pumped networks for the lower
eastern ones. The higher areas can drain into the Kesem, Kebena or T'unfeta
except briefly during flash floods. The lower areas will not require pumping all
the time, since flap-gated bypasses will allow the water to drain by gravity
whenever the Awash is low enough. The recurrent costs of pumping are therefore
not very great, but the pumpstations add both to the initial cost and the
operational complexity of the project. In the Medium Project case, therefore,
low-lying areas are excluded from the scheme so that no pumped drainage is
needed. The same measure reduces flood protection costs, and it happens that the
areas affected are generally not those with the best soils.
Most areas will either need drainage within the first 5 years, because of low
permeabilities or shallow watertables now, or they will never need deep drainage
at all because of adequate permeabilities and a high or sloping location.
Permeability in the top 3 m is much more important than watertable rise in
determining drainage requirements, in contrast in the Amibara case. The scope
for improving economic indicators by delaying the installation of field drains
until watertables rise is therefore very limited. Nor is there significant scope
for improving permeability by agricultural management. Only in the existing
state farm at Awara Melka is it proposed to provide for possible future drain
installation. That area has good natural drainage: .the Frenchman Saboret must
have been either clever or lucky in his selection of a place for his farm.
Once deep drainage and controlled irrigation have been provided for, the
reclamation of the saline and sodic lands will proceed in the usual way, mainly
by leaching. Some soils will also require the application of gypsum.
30
slopes of the diversion channels need to be carefully chosen. Wadi A is to be
led across a small part of the irrigation scheme and into the Kesem river, but
Wadi F is larger and less conveniently located. After comparative studies of
various options, it is proposed to let it cross the main canal and then spread
out into a large area of useless soils, as it already does. The floodwater
flowing on down the slope will then be collected by a drain and led into the
Kesem by gravity, except for large floods which will be temporarily stored and
discharged into the Kesem later through a flap gate.
31
Leakage through abutments would need to be restricted by a grout curtain at
either site, but the downstream site, because of the topography and the jointing
on the right abutment particularly, is slightly more likely to show unexpected
leakage. Leakage that does not endanger the structure can be tolerated anyway,
because the drawoff for irrigation is 10 km downstream.
With the geology of the two sites so similar, the choice of site is in effect
separate from the choice of dam type. Although the differences are not drastic,
the upstream site is preferred on grounds of geology, seismicity, topography and
leakage. It is therefore recommended for the dam.
Others were considered, including the arch type, but rejected. Some of the rock
layers, especially the red bole, are not strong enough for an arch dam, and
large-scale replacement of these layers with concrete would not be economic.
Suitable materials are available locally for both rockfill and hollow mass
concrete dams. The rockfil dam is definitely feasible, but the design of a
concrete dam would probably encounter difficulties because of a fracture zone
running along the gorge. Relative cost estimates show the rockfill type cheaper
by a small margin (10 to 15%), so it is recommended. The rockfill embankment is
accordingly assumed for estimation and project evaluation and is shown in the
drawings in Album 2, while both are discussed and illustrated in Annex N. The
choice of type should be reviewed, after further site investigation, at final
design stage.
The design of ancillary works, particularly the spillway, is bound up with the
choice of dam type because a concrete dam can have an integral spillway in the
dam wall while an embankment must have a separate spillway, and similarly with
diversion, drawoff, bottom outlet and hydropower works. These have therefore
been included in the cost comparison. The use of one or more emergency spillways
on saddles has been considered and rejected because of ground conditions, the
effects on side valleys, and the economies of scale. With a rockfill dam the
preference is for a single ungated overflow weir spillway discharging into the
gorge via a concrete chute and flip bucket. The small hydropower station, with
bypass for irrigation releases, can conveniently and economically be
accommodated in part of the foundations of the flip bucket. The bottom outlet
can be constructed in the diversion tunnel after the diversion phase is over.
For any useful reservoir level, a saddle dam will be needed in a valley 2.5 km
north of the main dam. By locating it slightly away from the watershed it has
been found possible to keep its length down to about 350 m, and a semi-
homogeneous embankment about 25 m high is appropriate.
For reservoir storage levels in the region of 930 m above sea level, a number of
low bunds would be needed to prevent spillage of water over saddles at levels of
936 to 940 m. They would however be very low (up to 5 m) and only wetted briefly
during major floods. A main dam crest level of 941 m would require six such
bunds totalling 3 km in length but making up only 3% of dam costs. Higher levels
would begin to involve bunds to a much greater extent, and they would be higher
and more critical, while lower levels would lose active storage at the rate of
over 30 hm3 per metre for relatively small cost savings, so the proposed crest
level is 941 m above sea level. With due allowance for freeboard (taking into
32
account spillway optimisation), this corresponds to a regulated storage level of
930 m above sea level and a total storage volume of 500 hm3. This makes the main
dam about 90 m high.
Another choice involved in the dam design at this stage is the type and height
of cofferdam and size of diversion. Economic optimisation shows a preference for
a high cofferdam and small tunnel and the choice also involves practical
considerations like access, timing, and inundation of a borrow area. The
proposed cofferdam is about 30 m high and is subsequently to be incorporated
into the upstream shell of the rockfill dam; the proposed diversion tunnel
diameter is 5 m.
The possibility of a very small dam in the gorge, of the order of 25 m high, has
been very briefly considered. It would fill up with sediment in five or ten
years unless a way were found to flush nearly all the river's sediment load
through the dam, retaining water only after the end of the flood season. In view
of the reservoir shape and the relatively high proportion of coarse bed load,
this is not considered practicable, and such a dam is not recommended.
33
4.6 Services and Infrastructure
Because there are now no public services and very little infrastructure in the
project area, the project package will have to include these. Without them it
would not be possible to build, staff and operate a large irrigation project, so
they are prerequisites or essential components rather than optional extras added
for social reasons.
The components involved, all of which are treated as parts of the project at
least for cost estimating purposes, include the following:
roads;
power supplies;
housing;
domestic water supply;
offices and workshops;
clinics and health services;
schools and education;
police and civil administration;
recreational facilities.
There is already a reasonable all-weather road from the Addis Ababa/Assab
highway to Awara Melka at the south-west corner of the scheme. The project would
improve and maintain it but not replace it. It is recommended that the track
from Awash town to the south-east corner should not be developed as an access
road, and the track through the National Park via Filweha Springs should be
closed at its northern end and not used at all for the project: the purpose of
these measures is to avoid damage to the ecological balance of the National
Park. At the north of the scheme a connecting road should be built to link the
project with the extended Bolhamo irrigation scheme, north of Dofan mountain,
which is expected to be in operation by that time and to have a bridge over the
Awash to the Amibara area. This will provide a second outlet on to the Assab
highway and will facilitate contact with the Amibara system and the research
station that it contains. With the access thus provided at the north end of the
plain and at its south-west corner, the project needs a main spine road from
which feeder roads can branch. The best route for this north-south road is along
the western side of the irrigation scheme, uphill of the main canal. This
location keeps it on well-drained and hard soils, gives the road designer
freedom to choose a good alignment without constraints from numerous canals and
drains, avoids having to cross these, and separates long-distance from local
traffic. Major project roads will branch from this spine road and serve the
irrigation areas and settlements, giving traffic access to the roads that are
proposed alongside all canals and drains. These major roads, and the north-south
spine road, will be raised to avoid wetness and also surfaced with selected
gravel, which is plentiful near the western edge of the plain.
Electric power supplies will be needed during construction and can be provided
from the national interconnected system by installing a 15 kV power line from
the existing substation at Awash Town, which connects both to the existing
132 kV line and to the proposed 220 kV line. It is proposed to extend the 15 kV
line to the dam site for power supply during construction, and also so that it
can afterwards bring power in the other direction, from the Kesem hydropower
plant to the rest of the project area. This plant will not be able to supply
power to the national grid because of its small size and low reliability, but it
can provide cheap local energy for domestic, pumping and crop processing within
the project.
34
Because of the unwillingness of the Afar to undertake manual work in agricul-
ture, and the size of the project's labour demand, large numbers of labourers as
well as staff and managers will have to come from elsewhere. The project there-
therefore needs to provide housing for large numbers of immigrants, generally
from the highlands. The number of employed persons is estimated at slightly less
than one per net irrigated hectare, and with dependants this can mean an
immigrant population of up to 60 000 people. This compares with the present Afar
population of about 12 000. The choices to be made in project planning concern
what sort of housing will be provided for the various categories of immigrants,
it being assumed that the Afar will continue to build and use their traditional
mobile houses.
For senior and intermediate staff it is assumed that permanent houses will be
built by the project of sand-cement blockwork with asbestos-cement or similar
roofs, to the normal standard for government employees, i.e. Type C or Type D
houses according to seniority. The more difficult choice is the type of housing
for agricultural labourers, who are of course much more numerous, constituting
85% of the immigrant employees. It also has to be decided how many of them are
to be permanent residents and have housing for their families. In recent years
the state farms in the Middle Awash have experienced increasing difficulty in
attracting enough labourers for the cotton harvest, which is seasonal work
relying mainly on migrant ('casual') labourers. For KIP there is a slightly less
seasonal demand for labour because of the break crops, but the cotton-wheat
system still needs 1.2 labourers per ha in September, less than 0.2 in March,
and about 0.7 on average. The patterns for all cropping systems have been
analysed (Chapter FS in Annex 5), and the proposal for purposes of analysis in
this study is that the proportions of permanent and casual labour should be such
that the permanent labourers are occupied for 90% of the time, the remaining
10%, or five weeks per year, being left for holidays and sickness. This
assumption results in a pattern where casual labourers do about 10% of the work
on the tobacco system and 22% on the cotton-wheat system, but make up 30%
and 44% respectively of the September workforce. For the project as a whole they
would represent 35% of the peak number. This proportion is important for
estimating housing requirements because a permanent labourer, having his family
with him, needs about four times as much housing space as a migrant casual
labourer. The casual labourers would stay in the area, on average, 3 to
4 months. The proposed proportion of permanent labourers is higher than on
existing monocrop state farms in the area but, as stated above, state farms are
finding it difficult to attract enough labour and improved provision of housing
for families is a factor in easing this problem.
The housing for agricultural labour is a major cost component for the whole
project (15% to 30% depending on housing type), so it is important what standard
of house is chosen. It could be the government standard Type G housing, which
costs about Birr 23 000 (US$ 11 000) per labourer with family. This is more than
the project can possibly afford. At the other extreme it could be a 'site-and-
services' approach whereby the labourers were allocated space to build their own
houses, perhaps being provided with some free materials. As a reasonable
compromise, the assumption for analysis in this study is that a housing type,
or a choice of options, would be developed so as to keep the average cost to
Birr 6 000 per labourer with family, only a quarter of the Type G cost. The
details would be worked out on site, using local materials and some degree of
occupant participation, but this amount could cover a minimal structure of
concrete floor, blockwork walls and metal or asbestos-cement roof, or
alternatively a more extensive dwelling of cheaper materials. On a project of
this size it will be worthwhile to set up an experimental programme to develop
the most economic solution locally by trials and experimentation. It is assumed
that the relatively small number of non-agricultural labourers and junior staff
will have Type G housing.
35
The project package will of course include offices and workshops for the
agricultural, engineering and social services involved, as well as the necessary
vehicles and equipment.
Water supply for the large population will need to be provided, but this is not
a problem since groundwater conditions are quite favourable: proposals are
described in Annex M. Sanitation is also necessary, and is accounted for in the
cost estimates. The Afar will be provided with simple water supplies and stock
watering points to improve health and living standards and to help keep the
livestock away from irrigation canals.
The large new population will require the usual range of state services,
particularly health, education and civil administration (including police, tele-
communications, etc.). Since there are no facilities in the area now, all these
are costed as part of the project for economic analysis, though the costs are
not very high in comparison with the agricultural sector. The project will place
considerable emphasis on preventive health care and on health education. Primary
schools and basic clinics will be distributed throughout the area, normally two
of each per state farm plus three clinics for the Afar, while a single secondary
school will be built at the Project Centre. This central township is proposed to
be built on a presently vacant site half-way along the north-south road and just
to the west of the irrigated area, where there is a patch of good soils that are
too high to be included in the gravity irrigation scheme. Other buildings will
generally be located within the scheme boundaries but on land that is not usable
for agricultural purposes: each state farm will have a headquarters area and one
other housing area ('villaget), both including labour housing with a school and
clinic. Recreational facilities, such as sports fields, will be provided, though
they have not been detailed at this feasibility study stage.
4.7 Institutions
There are decisions to be made about the project's institutional arrangements
which will be important for its performance and success but which do not need
to be made at feasibility-study stage. Annex F contains a discussion of the
options, particularly the choice between integration and separation of
functions, and the choice between national-level centralisation and project-
level autonomy. The two choices are linked and two combinations are singled out
for serious consideration, namely the centralised and functionally separated
structure which operates at comparable projects now (called Model A), and a
possible integrated project structure with a degree of local autonomy (Model C).
Both models have disadvantages, and in each case there are measures that can be
taken to minimise those disadvantages. In an absolute sense and for the long
term, the Consultant would recommend Model C, and it may be worthwhile for the
authorities to consider changing to such a structure for all large irrigation
projects at some future date when there are more of them than at present. But it
is unlikely that a change from the present national policy would be made just
for Kesem, so the proposal on which this study bases its estimates and further
analysis is a modified form of Model A, called Model AX in Annex F. The
modifications, some of which have already been proposed or are in the course of
being introduced on other projects, concern coordination at project and national
level between the main functions (agriculture, water resource control, and
settlement services). These modifications would go a long way towards removing
the disadvantages of the centralised-separated structure.
36
not be regarded as decisive. The Small Project in any case represents a much
smaller undertaking and the regulation of the Kesem's flow has wide implications
for the whole Awash basin.
All three of these EIRR values are far below the levels usually considered for
financing by organisations with a commercial or banking outlook, so that any
decision on project implementation will be made on wider criteria: similarly the
decision on project size cannot be made on economic grounds alone. In keeping
with the underlying policy to maximise the use of all resources, the Large
Project is the subject of most of this report, the Medium Project being analysed
alongside it in relevant places. The rest of the economic evaluation, and the
recommendations, are the subjects of later chapters of this tvlain Report. The
next chapter describes the proposed project, with the dam and reservoir,
distinguishing between the Large Project and the Medium Project where necessary.
Whenever the distinction is not mentioned, the Large Project is described.
Further consideration of a small project, with conjunctive use of groundwater
but no dam, is outside the scope of this study but should receive attention,
beginning with new data collection. The concept is close to that of the Ethio-
Bulgarian joint venture.
38
CHAPTER 5
THE PROJECT
5.1 General
The purpose of this chapter is to describe the proposed project in summary form.
Details can be found in the relevant annexes in Volumes 2 to 5, and in the
drawings in Album 1 (irrigation, drainage and flood protection) and Album 2
(dam). The reasoning behind these feasibility-study-stage designs has been
discussed in the previous chapter, so this one is generally confined to a
concise and quantitative description of the proposed project.
As explained at the beginning of Chapter 4, the project has initially been
designed to make maximum use of the land resources of the Kesem-Kebena plain, in
accordance with government policy. This concept is called the Large Project to
distinguish it from alternative scenarios called the Medium and the Small
Projects, which are based on the alternative strategy of selective land use.
Because much of the land on the plain is of poor quality, the selective approach
turns out to be economically more attractive, so the Medium Project must be
taken seriously. The scheme of that name which is described here, irrigating
8 920 ha net, is one representative of a range of possible schemes that could be
formulated. Depending on the criteria used, the best size might be anywhere
between about 7 000 and 11 000 ha. The Small Project represents the potential
for a scheme with no dam for seasonal storage of river flow, relying only on the
natural flows of the Kesem and the Kebena plus supplementary pumping from
groundwater in the dry season: at 4 315 ha it is probably near the upper size
limit for such a run-of-river scheme with normal cropping intensities.
The dam and hydropower station would be the same for the Large or Medium
Project. Many other project elements would differ little between those two
cases. Unless stated otherwise this report refers to the Large Project, but the
differences are mentioned when they are significant.
39
The net and gross irrigated areas achieved by these two layouts are set out and
compared in Table 5.1. The areas devoted to the higher-value state farm crops,
citrus and tobacco, are limited by market considerations to about 450 ha and
3 000 ha respectively. The rest of the land of mapping units 1 to 3, in the
suitability classification of Annex A, is allocated to the cotton-wheat farming
system, which is the next highest in gross returns and has no effective
marketing limit. In the Large Project most of the land of unit 4, which is
suitable only for irrigated pasture woodlots, is allocated to the 'settlement
areas', i.e. to irrigated pasture and gardens for the Afar and Soudani
populations. Exceptions to this target allocation include considerable areas of
unit 4 land in small patches surrounded by state farm lands, and patches of
better soils within the settlement areas. The former are used for woodlots
within the state farms and the latter are suitable for the gradual adoption of
irrigated agriculture, on up to 20% of the settlement land, by the Afar. The
detailed layout for the Large Project is shown on Drawings Ll and L2 in Album 1.
TABLE 5.1
Net Irrigated Areas
Settlement
(pasture and food crops) 1 260 14 4 180 30
It can be seen that the differences in cropped area between the Large and Medium
Projects are almost entirely in the cotton/wheat area and the settlement area.
This is because the citrus and tobacco system areas, being limited by markets,
can and should be maintained near the market limits.
The tobacco and cotton/wheat areas have a cropping intensity of 150% in each
case, with the cropping patterns of Figure 4.1. This results in the annual
cropped areas set out in Tables 5.2 and 5.3. A comparison between the two tables
shows that the exclusion of the poorer eastern areas from the Medium Project has
drastically reduced the proportion of woodlots in the cotton/wheat state farms.
In fact, when planning for selective rather than maximum land use one could
decide to omit woodlots altogether leaving unused patches in the state farms.
The Medium Project as formulated here, however, keeps about two-thirds of the
Large Project's woodlot area.
40
TABLE 5.2
Farming Systems and Net Cropped Areas - Large Project
Associated
woodlots 690 690
Associated
woodlots 330 330
Settlement System
Pastures with
gradual
introduction
of arable
farming 4 180 4 180
41
TABLE 5.3
Farming Systems and Net Cropped Areas - tvledium Project
Associated
woodlots 600 600
Settlement System
Pastures with
gradual
introduction
of arable
farming 1 260 1 260
Total 8 920 1 580 4 990 2 495 790 430 660 1 260
42
5.2.2 Irrigation System
Irrigation water will be extracted from the Kesem river by a gravity headworks
structure in the Awara MeIke gorge, approximately at the site of the existing
and partially broken weir. The intakes can be combined with the weir abutments,
and it will probably be convenient to combine the necessary road bridge with the
same structure. A simple mass concrete ogee weir Is proposed, as shown on
Drawing L7 in Album 1. For the Large Project the offtakes will be designed for
12 m3/s on the north bank and 6 m3/s on the south bank. They will skim the water
laterally from the river just upstream of the main weir, then pass it through
trash screens and gates and out of the gorge in rectangular concrete flumes.
Some way downstream, when any coarse sediment in the flow has concentrated near
the bottom of the flume, a horizontal slab will skim off the relatively clean
irrigation water and the water from the bottom of the flume will be returned to
the river through a gated scour sluice, except at times of water shortage when
the river water will be carrying little sediment anyway. Once the dam is
operating the river's sediment load will be small, but during the run-of-river
phase it will carry its full load and there will always be some sediment from
the river bed and tributary catchments downstream of the dam.
From the headworks the water will be conveyed by two primary canals, as shown on
Drawings Ll and L2. The north one, 16.4 km long, will begin with a piped reach
to take the water round the corner from the gorge as far as the flood bund, and
thereafter there will be a short concrete-lined reach. Otherwise the canals will
all be unlined, except locally where passing through sandy patches, where an
imported clay lining, using material from drain excavation, will be used. The
North Primary Canal will be aligned across the land slope, crossing two large
wadis (A and F) and the Kebena river, by means of reinforced concrete inverted
siphons (Drawing L9 in Album 1). The southern primary, only 5.5 km long, will go
down the slope close to the right bank of the Kesem, generally following the
route of the existing scheme's main canal: that canal will, however, need
extensive remodelling for the larger discharge, and new drop structures.
Distribution of water to the field units and pasture areas will be by a system
of secondary and tertiary canals with discharges from 0.2 to 2.6 m3/s. Their
total length is 175 km. For state farm areas, they will supply water to field
canals each of which will irrigate an area of about 22.5 ha net. Furrow
irrigation will be used on these field units. The irrigated pasture areas for
the Afar will also be supplied by field canals (not shown on Drawings Ll and
L2), but they will not serve a standard area and will be curved, following the
topography at minimal slope. Water will be released at intervals from these
field canals and will flow across the natural ground surface, being repeatedly
'distributed laterally by contour -furrows or 'spreader ditches', which can be
cheaply added to when necessary. This provides a cheap distribution system for
the low-value land.
Canal cross-regulators will be gated on the primary canals, but secondaries and
tertiaries will have long weirs of duckbill shape to maintain near-constant
upstream water levels with a minimum of operation input. Head regulators on
secondaries and tertiaries will be movable weirs provided with flow measuring
facilities. All primary, secondary and tertiary canals will have escape
structures.
The Large Project layout is such that the central water resources authority will
hand over measured quantities of water to the state farms and settlement offices
at only eleven places in the northern primary's system and ten in the south.
This is the result of careful adaptation of the layout to the soils. No state
farm or settlement unit will have to convey water through its area and hand it
on to another party, and most of them will receive water at only one or two
places.
43
In the area of the existing farms, some of the canals can be retained: at Aware
MeIke the eastward-flowing canals will be used as field canals but new
tertiaries will be built along the lines of the former southbound canals which
are now generally out of use. At Yalo it is not worth while to retain the canals
because the layout can be improved to bring in more of the unit 1 land and
exclude poor land from a tobacco farm.
TABLE 5.4
Land Drainage Treatments
The field drains will be perforated plastic pipes with filter surrounds, laid by
trenching at a depth of about 2.0 m below the surface, the total length of pipe
being about 3 600 km. They will discharge into 240 km of buried plastic pipe
collectors with buried manholes at the junctions (most of the buried collectors
will serve two field units, one on each side). They in turn will discharge into
open drains totalling 182 km.
44
In the Large Project (Figure 5.1) all the drains of the South Block, and the
eastern parts of the other blocks, will lead to drainage pumpstations. For most
of the time the water will discharge by gravity, bypassing the pumps through
flap gates, but for a few weeks each wet season the Awash river's water level
will be too high and the drain water will be pumped, usually against heads of
around 3 to 4 m. The more westerly parts of the Gurmile and North Blocks will,
however, be drained directly into the rivers or the T'unfeta floodway, without
pumping. In the north-east corner the pasture land is low but its value does not
justify a fourth pumpstation so some flooding will be accepted.
In the case of the the Medium Project (Figure 5.2), pumping of drainage water is
avoided altogether. The open drains discharge either into the rivers or onto the
low-lying eastern parts of the plain, where the drainage water will be
beneficial in promoting growth of grass in the dry season for the Afar
livestock.
To protect the irrigated land from floods in the Awash, Kesem and Kebena rivers
and the Wadi Tiunfeta, earthern flood bunds totalling 85 km in the Large Project
case are proposed. The average height is about 2.5 m and the maximum 5 m,
allowing a 0.5 m freeboard over the estimated 1-in-20 year flood levels. Nearly
all of this is avoided in the Medium Project, since the flood banks along the
tributary rivers in the western part of the plain do not need to be high.
About ten ephemeral wadis discharge floodwater onto the plain from the
escarpment foothills to the west. The proposed protection involves six
interceptor channels totalling 19 km and designed for discharges of up to
110 m'is to divert flood flows into the rivers. Their alignments are designed
for adequate slope, and storm runof f from the land between them and the main
canal will be picked up by another drain just west of the main canal. On each
wadi, upstream of the point where it is diverted into an interceptor, there will
be a flow dispersion structure designed to promote deposition of part of the
sediment load. All these works will be designed on the best available
information and then constructed and watched carefully, modifications and
improvements being made after each major flood in the light of the new
information it provides. This approach is more economic than an attempt to
design against any flow and sediment behaviour from the start, which would
inevitably lead to overdesign.
45
light levelling requirement (up to 200 m3/ha): 1 600 ha
medium levelling requirement (200 to 500 m3/ha): 4 200 ha
high levelling requirement (500 to 1 000 m3/ha): 4 000 ha
Laser-controlled elevating scrapers and motor graders are recommended for the
land levelling. The land preparation will be completed by land planing to give
uniform land slopes within about +_ 5 cm. Settlement areas require no land
levelling nor planing.
Cost estimates include the provision of 16 400 siphon pipes for furrow
irrigation, and 1 200 groundwater observation pipes so that the movements of the
watertable can be monitored.
46
The spillway will be an ungated concrete structure passing through the left
abutment, well clear of the main dam, though intersecting its low wing
embankment. The ogee crest at level 930 m will be 100 m long, slightly curved in
plan, and accompanied by a road bridge. Downstream of the weir the chute floor
will drop immediately by about 5 m, then remain near horizontal for a length of
about 60 m while tapering to 70 m wide, then turn down to a steep chute
terminating in a flip bucket at level 866 m. Rock anchors and underdrains will
be provided to prevent uplift problems, and the flip bucket will be located just
short of the fault zone to avoid foundation problems. Much of the rock excavated
to form the spillway will be used in the dam.
The 6 MW hydropower system will incorporate a bypass so that its tunnel can pass
irrigation water even when the turbines are not running. A separate intake on
the left abutment will control the flow of water into a horizontal concrete-
lined low-pressure tunnel at about level 930 m. A little downstream of the dam
axis this will join with a sloping tunnel containing steel penstocks leading to
two horizontal-axis Francis turbines in a small powerhouse located under the
spillway's flip bucket and thus sharing its foundations. The draft tube below
will lead laterally and discharge into the bottom outlet's stilling basin.
Two generators will pass power at 6.6 kV to transformers supplying the 15 kV
transmission line. This line will initially, during the dam construction period,
bring power from the national grid's substation at Awash Town, and will then
take power in the other direction, from the new station to the project. The link
to the grid will remain, being used during dry years when the reservoir is drawn
down too low for hydropower operation.
47
About 80% of the land is suitable only for pasture or woodlots (general
suitability unit 4), so the loss of potential cash crop land is not large. The
remainder, about 800 ha, consists of better soils in small patches which can be
used for growing food and cash crops.
As mentioned in Section 5.2, the pasture areas will be provided with secondary
and tertiary canals supplying water to field canals aligned almost along the
contours rather than parallel and straight. Water will be released from these by
gaps in their banks and will flow downslope, being spread by short contour
furrows. The naturally occurring grasses will be complemented, and partially
replaced, by improved grasses such as Cenchrus ciliaris and possibly Rhodes
grass. The Afar and Soudanis are expected gradually to take up the growing of
food crops such as cowpea, groundnuts, sesame, maize and wheat plus perhaps some
cash crops. Land will be allocated to clans and lineages within the existing
tribal structure, and the involvement of the Afar both in the allocation and the
water management will be maximised.
The Afar will live outside the irrigated area but close to their irrigated
pastures, and will never need to cross the state farms. In the Gurmile Block
their pastures lie to the north and south of the ,vedge of very poor land where
Wadi F now discharges, and this area will provide both access and living space
for the Afar.
In the case of the Medium Project, as can be seen in Figure 5.2, a strip of low-
lying land 2 to 4 km wide and 14 m long, along the eastern edge of the plain,
will be left unused by the formal irrigation scheme but will be watered locally
by the drain outfalls. This is also the area which is naturally flooded by the
rivers, and is thus an important part of the Afar's dry-season grazing area at
present. With this land still available for the Afars, and its dry-season grass
production slightly enhanced by drain discharges, it is considered that the
irrigation scheme, covering about 11 000 ha gross, will only need to include
about 1 300 ha net of irrigated pasture and gardens. The provision in the South
Block is the same as for the Large Project because the state farm area in that
block is hardly reduced at all, but in the other blocks there are only small
areas of irrigated pasture. As well as the eastern strip, the whole of the area
north of Wadi T'unfeta is left for the Afar, without formal irrigation. Other
arrangements concerning the Afar will be the same in the Medium and Large
Projects.
Once the project is in full operation, the overall supervision of the settlement
areas will be from three field offices located near the Afar villages, and a
headquarters office in the project centre. Each field office will include a
health centre. Before and during the construction period there will be an
advance programme to prepare the Afar for the changes brought by the project.
This will begin with the existing spontaneous irrigation schemes operated by
small groups of Afar and Soudanis, and will foster development of irrigated
agriculture, animal power, local participation and responsibility, and improved
health for people and livestock. A modest 8-year programme of outside inputs is
proposed (Annex C), costing about Birr 655 000 which is only about 0.1% of Large
Project costs.
For the surrounding areas, which will continue to be used by the Afar for wet-
season grazing, a small range and environment management programme is proposed.
Its main elements will be water harvesting and improvement of the mix of grass
species, and details will be progressively adapted in the light of experience.
The cost, for the initial 6 years, is estimated at only Birr 600 000 and the
continuation thereafter will be part of the settlement authority's work.
48
The project's south-western and main access road will pass through the Awash
National Park, but near its edge and following the route of the existing road.
That part of the park is not completely effective anyway, so the increase in
road traffic will not represent a serious loss in the conservation work of the
park authorities. The present track from Awash Town to the south-east corner of
the Kesem-Kebena plain will not be developed as a major access road, so as to
minimise effects on the park on that side. The central track, through the most
effective part of the park south of Saboret and Doho, will be closed at the
northern end to prevent project traffic from disturbing the wildlife. The
project's provision of irrigated pasture and woodlots should reduce the
incursions of pastoralists and fuelwood seekers into the fringes of the park.
With all these provisions built into the project, its environmental impact will
be slight and generally positive (a more detailed statement is in Annex D,
Section D3.6). The project will provide on-going monitoring of the environment.
Apart from the details of health services, which are described in Annex E, all
these project elements are described in Annex M. The reasoning underlying the
project provisions has been discussed in Section 4.6 above.
The project will upgrade the existing 32 km long access road between Awara Melka
and the Addis Ababa/Assab highway, and then extend it northward with a 21 km
long main road, 8 m wide and gravelled, along the western side of the irrigation
scheme. From the north-western corner of the scheme there will be a link road
about 16 km long, 4.5 m wide and gravelled, passing over the western flank of
Mount Dof an to link with the roads in the Bolhamo scheme and thus with that
scheme's proposed bridge over the Awash to Amibara. The north-south road will
cross the Kesem by a high-level bridge and the other watercourses by causeways
with culvert pipes (Irish bridges). Within the irrigated area there will be a
minimum network of gravelled through roads with one more crossing of each of the
49
two rivers: these total 56.5 km and will be 7 m wide. They will give access to
the roadways beside all secondary and tertiary canals and drains, so that
wheeled access will be provided to every field unit. They will also link all the
settlements. New access roads will be provided to the dam site on both banks.
Airstrips will be provided, for light aircraft from other parts of the country
and for crop-spraying.
Electricity supplies will be provided to all the settlements. In the early
stages of project development a power line will be built from the grid's
existing major substation at Awash town. This will be at 15 kV, passing north
from Awash along the eastern edge of the National Park, then west to Saboret and
the dam site with a branch to the Project Centre: a total distance of 50 km. If
and when a hydropower station is built at the dam, it will provide power to the
whole project area most of the time, using the same network, but the connection
to the grid will be retained for use when the reservoir water level is too low
to drive the turbines.
Housing is a major cost element. With the assumptions discussed in Section 4.6
above, the numbers of buildings required is as given in Table 5.5. Senior staff,
in accordance with national policy, will have Type C houses (about 58 m2 gross),
intermediate staff Type D (47 m4) and junior staff Type G (40 m2 with separate
communal shower and toilet blocks). To keep costs down, Types C and D units are
built in pairs and Type G units in blocks of four. For agricultural labourers
the non-standard low-cost housing described in Section 4.6 will be used, in
blocks which each accommodate 4 permanent labourers and their families, or
16 'casual' (migrant) labourers without families.
TABLE 5.5
Numbers of Buildings, Large Project
Office 8 7 3 18
Workshop 2 7 9
School 2 14 16
Clinic 1 14 3 18
50
The locations of the settlements are shown on Drawings Ll and L2. The Project
Centre contains the whole water resources staff and the central, project-level
offices of the state farms and settlement authorities, as well as offices and
housing relating to the other services. Each state farm has a headquarters and
one outlying village, each of these sites having a primary school, a clinic, and
about half the housing. These settlements, as well as the three outstations of
the settlement authority, will be on land that cannot be used for irrigation,
while the Project Centre site is on land with good soils but out of command from
the North Primary Canal. The central workshops of the Water Resources Authority,
located at the Project Centre, will maintain the vehicles of the Settlement
Authority and the various service offices, which are not big enough to justify
their own workshops.
Si
5.6 Institutions
As explained in Section 4.7 above, it is assumed for the purposes of this study
that the present policy will be maintained and the main functions will be
exercised by separate project-level offices of the relevant national bodies,
rather than by an integrated project organisation. To avoid the potential
disadvantages of a separated system while retaining its advantages, a project-
level non-executive coordinating committee is proposed (Annex F, p. F-15).
The three main functions would be exercised by the following organisations:
the Kesem Water Resources Office (KWRO), operating the dam and power
station, the larger canals and drains, drainage pumpstations and major
roads;
the Kesem State Farm Office (KSFO), supervising the individual state
farms and liaising with regional and national offices of MSFD;
the Kesem Settlement Office, responsible for the welfare of the Afar
and Soudani populations and operating through three outstations.
Estimated staffing numbers are set out in Annex F (Chapter F5) and used for cost
estimating. Table 5.6 summarises the totals, assuming that all teachers live at
the Project Centre (which they might not). The table excludes the 7 352 perman-
ent and 3 984 casual agricultural labourers.
The project formulation includes provision for a permanent monitoring and
reporting programme. Although this will cover all aspects of the project and its
environmental interactions, it is suggested that it should be coordinated by the
KWRO. The programme, which is also mentioned in Section 03.7 will cover
operation and maintenance work, the state of the canals, drains, flood
protection works, structures and buildings, the wider environment and its
vegetation cover, soil chemistry, water quality, livestock, environmental health
and social conditions.
5.7 Implementation
The implementation programme for KIP is discussed in Annex K, from which the bar
charts are reproduced here in Figure 5.3. These represent the quickest practica-
ble programmes and the start dates may be postponed, for instance due to the
need to await completion of other studies before deciding on implementation of
KIP.
The critical path runs through the arrangement of finance, the design of
priority works and the construction of the dam. With a four-year dam construc-
tion period this means that the regulated irrigation water supply will not be
available earlier than six years hence, in 1993. In the mean time an advance
phase of irrigation development can take place, using run-of-river supplies from
the new Kesem diversion weir and a cheap temporary Kebena offtake combined with
the road crossing. The proposed phasing of irrigation development is sumrnarised
in Table 5.7, from which it can be seen that the advance package is the same for
the two scenarios: the figure shows a two-year construction period so that the
advance areas only precede the rest by two years, though any delay in starting
or completing dam construction would stretch this gap. The next phase is timed
to come into production the year after the dam's completion, and in view of the
need to set up camps and infrastructure a three-year construction programme is
assumed, so that work overlaps with the advance phase's completion. In the
Medium Project the area after the advance phase is commissioned in three
successive years, the second and third phases being built in only two years
because a momentum would by this time have been established. For the Large
52
Figure 5.3
Tentative Project Development Programmes
Relative Timing Project Years: -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13
Suggested Ethiopian (Julian) Calendar: 1979 1980 1981 1962 1983 1984 1985 1986 1987 1988 1989 1990 1991 1292 1993
Absolute
Timing International (Gregorian) Calendar: 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
Lag
1
Packages G4, N2 r,
Package N3
Infrastructure:
Advance work -..:
Remainder
....mi...
gIMIIIIMAINIMIN t MINIM
Dam
Suggested Ethiopian (Julian) Calendar: 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 199111992 1993
Absolute 1
Timing International (Gregorian) Calendar: 1987 1968 1989 1990J1991 1992 1993 1994 1995 1996 1997 199811999 2000
Arrangement of finance
Irrigation areas:
Existing schemes
Advance (run-of-river), Packages 51, Cl
Package 52X
Package G2X r
Infrastructure:
Advance work
Remainder
Dam r
Hydropower 1--
Key:
Preparatory phases Tendering and award Construction Operation
of contracts or implementation
TABLE 5.6
Staff Numbers, Large Project
Senior Intermediate Junior Total
KWRO 8 43 85 136
KSFO (Central) 5 13 12 30
7 state farms 99 379 1 085 1 563
Sub-total, state farms (104) (392) (1 097) (1 593)
KSFO (Central) 3 3 4 10
3 outstations 3 36 114 153
TABLE 5.7
Staged Development of Irrigated Areas
Large Project Medium Project
Year of first State farms Settlement State farms Settlement
production (net ha) (net ha) (net ha) (net ha)
Run-of-river (advance' Phae):
4 1 520 400 1 520 400
Using Reservoir:
6 2 500 1 360 2 670 280
7 2 400 960 1 770 380
8 3 060 800 1 700 200
9 430 660
53
Project the pattern is similar except that a relatively small amount of work is
held back until Year 8 and commissioned for Year 9: this comprises the low-lying
north-east corner of the project and is delayed to reduce the pressure of
construction, particularly large-scale earthworks for drains and flood protec-
tion, in Years 6 and 7. Thus the bulk of the construction is concentrated into
five years.
Other elements of the programme are self-explanatory. The advance programme of
assistance to the Afar, and the range management programme, start relatively
early and the on-going work in those fields is to be handed over to the
settlement office after the project is fully operational. Some of the infra-
structure is needed early eadvance work°, such as access roads, some housing,
basic water and power supplies; the rest is built progressively in step with the
irrigation areas that it serves. The design process is shown spread over four
years, although it could be done more quickly, because it is advantageous to do
the detailed design for irrigation and drainage works during the construction
period (much topographic survey is involved, for land levelling particularly).
The Ethio-Bulgarian Joint Venture (EBJV) has now formally taken over the Awara
Melka State Farm and was preparing to start design work during 1987. It should
be regarded as a financing and implementation package rather than as an
alternative project, and it has the great advantage of being already in place,
while any project arising out of this report will take some time to set up. The
EBJV had not, at the time of completion of this study, issued any detailed
implementation programme, so the project planning in this study has not taken
special account of it. It may prove possible and advantageous for the EBJV to
undertake the advance (run-of-river) phase of the KIP proposed here, but two
difficulties can be foreseen: firstly, the land blocks suitable for run-of-river
irrigation will later be separated by other blocks and will not form a coherent
unit in the long term, and secondly the EBJV is not at present organised or
motivated to include significant provision for the Afars. With close cooperation
it should be possible to overcome these difficulties and integrate the two
initiatives. It might be conv-enient to allocate the whole South Block (4 600 ha
net) to the EBJV, since this is a distinct unit with its own primary canal and
contains most of the existing state farm. No use could then be made of the
Kebena river as a water source, but its dry season flow is anyway much less than
the Kesem (contrary to EBJV estimates).
5.8 Estimated Costs
The initial and recurrent costs of all the proposed project elements have been
estimated, in both financial and economic terms. These are 'best estimates' or,
in statistical terms, expected values: they therefore do not include a
contingency allowance in the sense of a safety margin against underestimation.
Such bias in cost estimating is considered appropriate for budgeting but not for
economic analysis. For budgeting purposes it would be prudent to add a
contingency allowance of 10% or 15% to the financial figures given here. The
estimates do, however, include allowances for miscellaneous or unbilled items,
as well as for design, supervision and administration.
An abstract of the estimates of initial costs is presented here in Tables 5.8
and 5.9, and the Appendix to this volume gives per hectare costs and the
breakdown by years. In this context 'initial' costs include the procurement of
agricultural machinery in the first ten years, although the subdivision in Annex
N is slightly different. Details are in the annexes and a key to their location
is given in Chapter N5 of Annex N, which also summarises all costs.
The relative sizes of the different cost elements, expressed by the percentages
on the right hand side of Tables 5.8 and 5.9, are of particular interest, as is
the distribution of costs between state farms and settlement functions (common
54
TABLE 5.8
Summary of Initial Costs, Large Project
(Million Ethiopian Birr at 1986 financial prices)
55
TABLE 5.9
Summary of Initial Costs, Medium Project
(Million Ethiopian Birr at 1986 financial prices)
Irrigated area:
- irrigation system and land
preparation 78.8 8.5
drainage system 54.1 3.9
flood protection 8.0 0.8
Sub-total 140.9 13.2 154.0 32
56
costs have been divided in the ratio of the water volumes used). The settlement
function represents only 17% of initial costs for the Large Project and 10% for
the Medium Project. Buildings make up a quarter of the initial costs in either
case, which is more than the cost of the dam and hydropower station.
Within the irrigation area, which at 30 to 40% is the largest cost category,
deep drainage accounts for more than a third. A more detailed breakdown of
irrigation, drainage and flood protection costs is given in Annex L. The total
of these categories amounts to some Birr 16 000 per hectare, equivalent to about
US$ 8 000 per hectare which is, by international standards, not unreasonable in
a situation with quite heavy drainage needs and fairly awkward topography.
The distinction between 'design and engineering' and 'supervision and adminis-
tration' is an arbitrary one, and the estimate for the former may be too low,
depending on the arrangements adopted. Some elements (like irrigation systems)
have relatively higher design costs than others (like buildings).
In the annexes, all costs are broken down into local currency, direct foreign
currency, and indirect foreign currency components. Direct foreign currency
accounts for some 53% of initial costs, distributed as detailed in Appendix 1.
The estimates of recurrent costs are summarised in Table 5.10 and breakdowns by
category and function are given in the last two tables of the Appendix. To
enable their significance, relative to initial costs, to be assessed, Table 5.10
gives the discounted Present Values of the total estimated recurrent costs, for
a nominal 50-year period and for different discounting rates. It must be
remembered that these figures for recurrent costs are at full development, i.e.
from year 20 onwards (costs are less in some earlier years), and also that
agricultural inputs and labour wages are excluded (being accounted for in crop
gross margins). The settlement function accounts for only 5 to 8% of recurrent
costs, with common services' costs being divided in the proportion of water use.
TABLE 5.10
Summary of Recurrent Costs
(million Ethiopian Birr at 1986 Financial Prices)
Large Medium
Project Project
These tables exclude the Small Project, the costs of which have been estimated
in less detail and only in economic terms. Most of the discussion of this
scenario is in Section L11.3 of Annex L. The total economic initial cost is Birr
160 million (without agricultural machinery), compared with Birr 477 million for
the Large Project and Birr 380 million for the Medium Project.
57
CHAPTER 6
EVALUATION
6.1 Methodology
The evaluation of the project has been carried out by an economic cost-benefit
analysis using border pricing. The methodology follows the relevant guidelines
published by the Ethiopian Government in 1981 (Ref. 13). It has been discussed
with the responsible department (DPSA, authors of the guidelines) and with the
WRDA/FAO economists: the guidelines are under review in 1987 but no changes had
been issued at the time this analysis was done. The treatment of foreign
exchange is that it is not itself shadow-priced, but all local currency elements
are multiplied by a Standard Conversion Factor of 0,75, which is equivalent to a
1.33 factor on foreign exchange. Unskilled labour is shadow priced by applying a
factor of 0.67 relative to other local costs, giving a Specific Conversion
Factor of 0.50 to convert unskilled labour wages to border prices. The border
prices of traded inputs and outputs have been calculated by allowing for freight
and insurance in the usual way. Details are given in Annex N, particularly
Chapter N4.
To give adequate weight to the dam and other civil engineering works \,vhich have
long economic lives, the analysis was calculated over a period of 75 years, i.e.
for the first 70 years of the dam's operation.
TABLE 6.1
Summary of Economic Crop Gross Margins
Range of economic
gross margins(1)
(Birr/ha)
State farm crops:
citrus 6 400 to 11 700
tobacco 5 400 to 7 000
cotton 2 200
maize 2 200
- wheat 1 700
Smallholder food crops 600 to 1 300
For convenience the Medium and Small Project cases have been evaluated with the
same weighted average yields as the Large Project. Since their selective land
use would enable more of the poorer soils to be avoided, their average yields
would in fact be slightly higher, so this represents a slight bias against them
in the analysis.
The economic costs have been calculated in parallel with the financial costs
throughout this study. Most of the initial costs are broken down into four
types:
unskilled labour;
skilled labour;
imported materials;
imported equipment.
The sub-totals for these types have then been subdivided into direct and
indirect foreign exchange and local currency, and converted to economic costs by
factors laid down by the government's Development Projects Study Agency (DPSA).
Unskilled labour costs are 100% local currency and are converted to economic
costs by the Specific Conversion Factor of 0.50. Skilled labour is also 100%
local currency but subject to the Standard Conversion Factor of 0.75. Imported
materials are estimated to include, on average, 40% direct foreign currency, 31%
indirect foreign currency and 29% local currency, with a weighted average
conversion factor of 0.78 (these costs include cement, which though now
partially imported will in future be locally made, and this is allowed for
here). The corresponding percentages for imported equipment, allowing for local
handling and transport, are 90%, 2% and 8%, and the factor is 0.98. Engineering
recurrent costs were subject to a weighted mean factor of 0.78. With the
exception of fuel costs, all the financial costs exclude taxes, because a
clearly-defined project like KIP can obtain tax exemption. Direct foreign
currency accounts for 51% of initial financial costs (excluding agricultural
machinery).
The economic costs are summarised along with the financial ones in Chapter N5 of
Annex N: for the initial costs they turn out, in total, to be numerically just
over 82% of the financial costs which have been summarised at the end of the
preceding chapter.
A deduction is made in the economic analysis for the expected future returns to
agriculture in the area without the project. The 'without-project' situation is
defined, for this analysis, as comprising the continuation of the existing state
farm with foreseeable changes due to the age of the standing fruit trees. The
cropped areas are 210 net ha of citrus, 300 net ha of tobacco and 410 net ha of
cotton, giving a total economic gross margin of Birr 3.75 million per year in
the long term (details in Annex N, Table N4.8).
59
The classification of costs used in these tables, for instance into initial and
recurrent costs or into agricultural and non-agricultural, is arbitrary and
there are some items, like workshops used to maintain agricultural machines,
which could be classified in different ways. For detailed breakdowns the reader
is referred to the annexes. One matter, however, goes beyond mere classifica-
tion details, namely, the cost of attracting adequate labour to the area. In
this analysis the wages (which are low) appear under crop gross margins and the
housing costs (which are high) appear under initial costs, but in fact there is
a potential trade-off between these two. It might be possible and advantageous
to provide less housing and pay the labourers more, perhaps assisting them to
build their own houses. This would change the balance between the figures in
this analysis, even if there were no net change.
60
TABLE 6.2
Summary of Economic Analysis Results
Net ha Initial EIRR(2)
irrigated economic
cost, M Birr(1)
A. Base Case
Large Project 14 090 477 0.55
Medium Project 8 920 380 2.11
Small Project 4 315 160 2.5
Notes: (1) These initial costs exclude agricultural machinery costs, which
amount to Birr 29 million for the Large Project and Birr 24
million for the Medium Project.
(2) Economic internal rate of return.
61
project costs, are moderately significant and contribute a differential of about
1% to the EIRR. Provision for the Afar, representing only about 9% of initial
costs, contributes about 0.3% to the Large Project's low EIRR. It must be kept
in mind that the irrigated pasture uses low-value land suitable for little else,
and has low development costs per hectare, so the areas, if compared with those
of the state farms, tend to give an exaggerated impression of importance.
The growing of break crops, maize and wheat, contributes nearly 2% to the EIRR
and is thus of considerable economic significance as well as being decisive for
local and regional self-sufficiency in food production. The EIRR is moderately
sensitive to estimates of crop yields and to cost estimates.
According to DPSA criteria, the minimum EIRR for favourable consideration in
national planning and financing is about 10%. It is however not necessarily
right to apply such a criterion uniformly to all kinds of project. This project
develops an area \,vhich is currently almost completely without infrastructure and
services, and these have been fully costed as a prerequisite of the development
and thus a necessary part of the project. This study has taken account of
drainage and salinity problems in the irrigated areas, which has not always been
done in the feasibility studies of other projects. KIP's economic result should
therefore be compared not primarily with a hypothetical, nation-wide and multi-
sector target like 10%, but rather with the results of appraisals of other rural
projects that provide their own infrastructure, carried out in a comparable way.
In the light of criteria beyond the purely economic, it might be found desirable
to implement the project despite its forecast EIRR in the range 1 to 4%.
62
6.5 Finance
This study includes, in Annex N, a financial analysis for the project and for
its two productive subdivisions, the state farms and the settlement authority.
Not surprisingly in view of the results of the economic analysis, the financial
internal rate of return is negative. With the hypothetical assumption of a
'soft' loan of 476 million Birr (of which 14 million is for working capital)
with a grace period of 15 years and interest rate of 2%, the project still has
an accumulated deficit of over 500 million Birr after 30 years. Such a loan
package might be approached by separating the 'settlement' element, which
represents just over 10% of the total initial costs for the Medium Project, from
the state farm element. All these figures are at 1986 prices assuming a foreign
exchange rate of 2.07 Birr per US dollar: with 5% annual inflation the 'price
contingencies' would add 40% to the nominal financial cost.
63
CHAPTER 7
This chapter brings together the main conclusions of the whole study. To enable
the reader to obtain an overview of what is a very complex study, the chapter is
kept extremely brief and details must be sought elsewhere in this volume or in
the annexes.
The Kesem Irrigation Project (KIP) is found to be technically feasible. On the
Kesem-Kebena plain a gravity irrigation scheme can extend to about 14 000 ha
net, though the soils are variable and generally fairly poor and it is
preferable to restrict the formal irrigation to the better areas, the optimum
being probably between 8 000 and 10 000 ha net.
Preferred main crops are tobacco and citrus, which attain good returns but face
limited markets, plus cotton, wheat and maize. Soils usable for these crops are
sufficient for 7 000 to 10 000 ha net, and further areas can be used for
woodlots to produce fuelwood and for irrigated pasture to compensate the 12 000
Afar pastoralists for the loss of their present dry-season grazing area. Maximum
use of the land resources implies the inclusion of about 4 000 ha of such
pasture, but selective land use could leave much of the present grazing area to
the Afar and thus reduce the need for formally irrigated pasture.
The project requires regulation of the Kesem river's highly seasonal flow by
means of a dam about 90 m high, for which a suitable site and materials exist.
The reservoir formed by the dam would be filled with sediment, probably in less
than 100 years, but would provide water for KIP for around 70 years. In earlier
decades, before sediment deposits build up, it could provide regulated flow for
other schemes in the Awash Valley, and later these schemes could be provided
for by other reservoirs, either new or re-allocated.
Use of less than the maximum technically feasible area on the Kesem-Kebena plain
represents a modification to the project's objectives, which have been defined
as aiming to maximise agricultural potential under a balanced environmental and
ecological system. The maximisation should however be viewed for the Awash basin
as a whole, and since there are better soils further down the valley it would be
inappropriate to irrigate KIP's poorest soils. It is therefore recommended that
the plain should be used selectively, irrigating 9 000 to 10 000 ha net as in
the 'Ivledium F.)rojeot.' scenario of this study. At the same time the question of
developing a much smaller area, such as 4 000 ha, with no dam, should be
investigated.
The area now has little infrastructure and the project would open up some
22 000 ha on the plain and a much larger area in the foothills of the escarp-
ment. It would link with the proposed extended Bolhamo scheme to the north.
KIP would bring up to 60 000 people, mostly highlanders, into the area and would
provide housing for these immigrants and work for up to 13 000 of them, as well
as health and education facilities. The environmental impact is small but
positive.
This project is not economically attractive. A careful economic analysis, which
takes account of all the relevant costs including social and infrastructure
components and the need for drainage and reclamation of unfavourable soils,
shows an economic internal rate of return of just over 2%. Under some
circumstances it might rise to 4%. Financing would be difficult and would
probably require separate packaging of the components dealing with the Afar.
64
The Consultant's recommendation is that this project should be compared with
other projects in the same sector, preferably after they also have been studied
in a compatible manner. Projects that can take advantage of pre-existing
infrastructure and services may appear economically preferable to KIP, but the
economic criteria are not the only ones, and in the development of a country
like Ethiopia the opening-up of new areas may be a target in itself. The
provision of housing and livelihood for some 60 000 people from the highlands is
also a factor, the valuation of which is beyond the scope of this study. Such a
comparison might show KIP to be one of the better rural development projects
available.
In the mean time, the Ethio-Bulgarian Joint Venture is already active in the
project area, with aims that are more limited than those of KIP but not
incompatible with them. It is not an alternative or competing project, but a
financing and execution package. As such it has the advantage of being already
in place, whereas any project springing from this report will take at least two
years to reach the stage of action in the field. Integration of the two
initiatives should be pursued as soon as possible, using this report to help
plan the EBJV's first phase as part of a larger project. In particular, new data
should be gathered on relevant matters which were outside the scope of this
project, such as the low flows of the Kebena river.
65
REFERENCES FOR MAIN REPORT
66
APPENDIX 1 TO MAIN REPORT
67
TABLE 1/1
Initial Costs per Hectare, in US Dollars
(Thousand US$ per ha, at 1986 financial prices)
68
E/uscription Year Year Year Year Year Year Yuar Year Year Total Direct
0 1 2 3 4 5 6 7 8 foreign
exchange
Irrigated area: (%)
- irrigation system and land
preparation - 5.96 13.08 7.12 20.42 24.35 13.43 6.48 90.83 51
- deep drainage 4.54 10.70 6.16 18.00 20.62 10.82 3.37 74.22 49
- major canals and diversion weir 1.56 4.32 6.61 3.07 6.26 4.74 2.06 - 28.60 43
large drains and pumping stations - - 1.15 3.27 2.13 5.28 3.52 1.07 0.53 16.94 59
- flood protection vvorks 0.32 1.87 1.55 4.82 5.12 2.35 9.34 25.37 63
Sub-total 1.56 16.29 35.53 20.03 54.78 58.35 29.73 19.72 235.96 51
Major roads:
access to Awara IvIelka 0.88 0.66 0.44 0.22 - - 2.20 63
- Amara Ivielka to project centre 1.60 1.20 0.80 0.40 - - 4.01 60
- access to darn site - 0.57 0.43 0.29 0.14 - - - 1.43 60
- gravelled roads - 2.56 3.77 1.21 1.62 2.25 2.25 13.66 62
other Inajor roads - - - 1.67 1.67 - 3.34 57
Afar, range and environmental 0.12 0.12 0.20 0.15 0.15 0.15 0.15 1.18 59
Supervision and administration 1.34 3.70 5.57 7.34 4.00 2.22 0.62 28.58 52
Total initial costs 0.12 13.21 52.10 84.26 87.77 142.18 118.69 79.43 33.24 611.02 53
Major roads:
- access to Awara Melka - 0.88 0.66 0.44 0.22 - 2.20 63
- Awara Melka to Project Centre - 1.60 1.20 0.80 0.40 - - 4.01 60
0.57 0.43 0.29 0.14 - _ _ 1.43 60
- access to dam site
- gravelled roads - - 2.56 2.56 2.02 4.27 2.25 - 13.66 62
- other major roads - - - 0.50 0.50 - 1.00 57
Atar, range and environmental 0.12 0.12 0.20 0.15 0.15 0.15 0.15 0.15 1.10 59
Sub-total - 0.22 0.16 0.11 5.00 2.43 10.49 7.05 6.89 33.16 90
Supervision and administration 0.64 2.85 4.26 3.74 6.04 3.96 1.07 22.57 52
Total Initial Costs 0.12 11.32 50.35 87.10 82.05 125.78 97.07 25.96 6.89 486.65 53
Agricultural machinery:(2)
- replacement 4.04 4.04
operation 5.22 5.22
Engineering O&M:(3)
- plant and vehicles 2.73 0.57 3.30
building maintenance 0.92 0.08 1.00
ID and FP 1.32 0.28 1.60
- miscellaneous 0.63 0.27 0.90
71
TABLE 1/5
Recurrent Costs, Medium Project
(Million Ethiopian Birr per year at 1986 financial prices,
at full development excluding agricultural inputs and labour)
Agricultural machinery:(2)
- replacement 3.13 3.13
operation 4.10 4.10
Engineering O&M:(3)
plant and vehicles 2.32 0.28 2.60
- building maintenance 0.79 0.06 0.85
ID and FP 1.10 0.10 1.20
miscellaneous 0.64 0.11 0.75
72
Figu. 3.1
The
tri o-
f"-
Thì
f
.38tic,n rchcmc 1
SAD!
Road or track
rs
-
-
11 -7 0
1
i
' t i
-- - "in2oóooN I
Gurmile Hill ..
Nj
i
tA
DOH
I
-
Springs
8
o
Figure 5.1
LEGEND
Canal
BLOCK PART OF
Drain
AMIBARA
North-South access road k IRRIGATION
Alternative dam access routes \ PROJECT
Drainage pump station
Staged development unit (SDU) 16
Contour
FARMING SYSTEMS
Citrus
Tabacco etc.
Cotton/ Wheat YALO
Pasture 21
SCALE
5
, -
/
1 o 1 2 3 4 5 km. 12
GURPAILE
11
Gurmile Hil BLOCK /
17 20
e
SSDDLE DAf 1
14
15
SOUTH
DOC-10
-
BLOCK
o 8
1010000,,
iv DAr1
s
Filweha
Springs
Figure 5.2
wadi
LEGEND
,
Canal I
Noi.31:-: sLockir,' ' PART OF
Drain
North-South access road ,.., g i
,
/1(
--ig-
AM1BARA
IRRIGAT,CX,
-,=-:' 22 ,/ / PROJECT
Alternative dam access routes 4),
....,,,,4".
'I
,/ 23 / 1-:
Y
'.
\ ,
18 'tf.`-_-'
Drainage pump station
Staged development unit (SDU) 16
Contour
-
FARMING SYSTEMS
,
-- ,-,--=,- 11
24x - _.- d?,-;
Citrus
Tobacco etc ,..----
Cotton/ Wheat '(ALOE'21x
Pasture
1
___ r 21x,
5 \
i
1
Ç=/ /I
' (URM iL4-E,
,
-k Gin-mile Hi S L_ 0 C/1,-: /
16x---- ,
vog_i
_J L g
/ / 17x
gigig
1 f
cc
= 15x j
14
15x
C
1 0x ""-'" SOUTH
lc-6g s e /27 .1)01-10 9
- River_ , BLOCK-
/
SPBORET ,
7
8
2
ULM DM ,
kg,/
red
Filweha e
RESEP,VOIn Springs