Impact Assessment of Small‐Scale Pump Irrigation in the

Somali Region of Ethiopia

PLI Policy Project

September 2010
Acknowledgements

The assessment team is grateful to all who participated in the impact assessment and made
a valuable intellectual contribution to the assessment process. These are the participants of
the ABGs from Gode, Kelafo and Mustahil districts; CHF experts namely Mr. Abuboker Haji
Ali, Mr. Abdulahi Mohamed, Mr. Mohamed Ahmed, Mr. Issie Ali, Mr. Heys Farah and Mr.
Samuel Tesfayethe; irrigation experts consulted in Gode town namely Mr. Teyib Sherif of
FAO, and Mr. Dereje Rorisa and Salehu Woldeyes working for the government of Ethiopia on
the South Gode irrigation development project; private pump owners Mr. Yussuf Issie and
Ismail Mahad in Kelafo; elders and ex‐settlers from the Wollo and Gojam area; W/ro Meselu
Abebe, owner of Jignoch Hotel who organised the meeting held with onion traders from
Somalia in Kelafo; and government employed human health experts consulted in Mustahil
town. The team would also like to thank Mr. Solomon Tibebu of CHF Addis Ababa for his
support in the design, planning and organisation of the assessment.
Contents

SUMMARY .......................................................................................................................... I

1. INTRODUCTION ......................................................................................................... 1
1.1 Background to small‐scale irrigation development in Ethiopia ................................ 1
1.2 CHF L‐SAP Project background .................................................................................. 3
1.3 Project implementation............................................................................................. 4

2. ASSESSMENT METHODOLOGY ................................................................................... 5

2.1 Field assessment design ............................................................................................ 5
2.1.1 Assessment areas ......................................................................................... 6
2.1.2 Sampling method and sample sizes .............................................................. 7
2.1.3 Data collection methods ............................................................................... 8
2.1.4 Pre‐testing .................................................................................................... 9
2.1.5 Triangulation ............................................................................................... 10
2.1.6 Data analysis ............................................................................................... 10
2.2 Economic analyses ................................................................................................... 10

3. ASSESSMENT FINDINGS ........................................................................................... 11

3.1 Background information on crop farming along the Wabe Shabelle River............. 11
3.1.1 Traditional land management and livelihood systems ............................... 11
3.1.2 The development of irrigated agriculture in south Gode ........................... 13
3.1.3 Traditional crop farming practice ............................................................... 15
3.2 The L‐SAP irrigation project findings ....................................................................... 16
3.2.1 Operational status of ABGs ......................................................................... 17
3.2.2 Group membership and land status ........................................................... 18
3.2.3 Production histories .................................................................................... 19
a. Gode ..................................................................................................... 20
Top: Failed banana crop at Ilan ............................................................... 23
b. Kelafo ................................................................................................... 23
c. Mustahil ............................................................................................... 24
3.2.4 Trends in production, costs, sales and net benefits ................................... 26
a. Trends in wealth groups, non‐ABG members in Gode ........................ 26
b. Trends among L‐SAP participants ........................................................ 27
3.2.5 Time costs of farming by ABG households ................................................. 29
3.2.6 Zekat ........................................................................................................... 30
3.2.7 Benefits and risks of growing different crop varieties ................................ 30
3.2.8 Perceptions of the benefits of irrigated farming ........................................ 31
3.2.9 Common risks and challenges .................................................................... 32
a. Environmental factors ......................................................................... 32
b. Economic factors ................................................................................. 33
c. Social factors ....................................................................................... 34
3.3 Cost‐benefit analysis of irrigated farming in Gode .................................................. 34

4. DISCUSSION ............................................................................................................. 37
4.1 The importance of pre‐project analysis .................................................................. 37
4.1.1 Participatory approaches ............................................................................ 37
4.1.2 Pre‐project lessons from the literature ...................................................... 38
4.2 Project monitoring ................................................................................................... 40
4.3 Livelihoods issues .................................................................................................... 40
4.3.2
4.3.3
4.3.4

4.3.1
 Livelihood impact and implications ............................................................ 40
Economic issues ‐ markets and trade ......................................................... 41
Environmental and health impacts and implications ................................. 43
Social impact and implications ................................................................... 43
4.4 Future considerations .............................................................................................. 44
4.4.1 Securing access to land ............................................................................... 44
4.4.2 Irrigation investments are development investments ............................... 44
4.4.3 Design issues ............................................................................................... 44
4.4.4 Recommendations ...................................................................................... 45

5. CONCLUSIONS ......................................................................................................... 46

ANNEXES ......................................................................................................................... 48
Annex 1 Inputs provided per ABG ............................................................................. 48
Annex 2 Number of irrigations required per crop .................................................... 49
Annex 3 Water pump maintenance costs incurred by ABGs .................................... 50
Annex 4 Results of ABG crop preference ranking ..................................................... 51
SUMMARY

Gode zone in the Somali National Regional State of Ethiopia is a remote area in the
east of the country, characterized by marked under‐development and frequent
humanitarian crises. In addition, various forms of conflict in and around the zone,
including in neighbouring Somalia, have created a very difficult operational context
for international NGOs.

The L‐SAP project of CHF International was funded by OFDA for one year, in three
woredas in Gode zone, and ended in December 2008. The project aimed to improve
the household income and assets of targeted poor households through establishing
group‐based small scale irrigation schemes along the Wabe Shabelle River for the
production of food and cash crops. The project established 18 ‘Asset‐building
Groups’ (ABGs), each comprising at least 50 households. The ABGs were provided
with water pumps, fuel, seeds and tools, and training. Each group was to be
allocated 25 ha of land, equivalent to 0.5 ha/household for a group of 50
households. The project was assessed in mid 2010, 18 months after the project
ended.

Main findings
1. The project areas had a long history of both private sector and government
project‐based pump irrigation, dating back to the 1960s and in terms of private
activities, influenced by experiences from small‐scale irrigation in neighbouring
Somalia. Many years before the project, communities had evolved locally‐
appropriate irrigation systems based on dhulqeyb (land sharing) or birgerac
(pump renting). These existing systems – developed to suit the natural and
socio‐political environment ‐ seemed to have been overlooked during the design
of the project.
2. An impact assessment of the project was initially viewed as useful due to the
proposed target beneficiaries in Gode woreda being destitute women
pastoralists, and the potential to offer an alternative livelihoods option to these
women via small‐scale irrigation. However, the participants in Gode were all
pre‐established crop producers and some had experience of pump irrigation
dating back to 2001. Similarly in Mustahil and Kelafo woredas, the project
participants all had prior and often long experience of pump irrigation.
3. The above findings only became apparent when the assessment team first
visited Gode, and highlighted a set of difficult project attribution issues.
Attribution was further complicated when in Mustahil and Kelafo, it also
became evident that intended beneficiaries were farming as individuals not
groups, and had mostly reverted to the well‐established dhulqeyb or birgerac
systems. Project pumps in these areas had simply been added to pre‐existing
pumps provided by private owners and previous projects. The measurement of
impact and attribution was further hindered by the lack of verifiable crop
production or sales data from the irrigated farms, either in CHF or among the L‐
SAP project participants.
4. At the time of the assessment in mid 2010, CHF reported that 15 out of 18 of
the original ABGs were still functioning. The assessment sampled 13 of the

i
 remaining 15 ABGs, but found that only three were still functioning as groups
(both in Gode) and one of these had been linked to another NGO. As noted in
point 3 above, all ABGs in Mustahil and Kelafo had reverted to pre‐existing
individual farming arrangements. It seems possible that in many project
locations, local people agreed to form groups mainly to access high‐value and
freely‐distributed project resources such as pumps, spares, fuel, tools and
seeds.
5. Of the 25 ha planned for each ABG to cultivate, in Gode only around 5 to 10 ha
had been cultivated by the groups. In Kelafo and Mustahil, the land allocated to
the ABGs was ‘incorporated’ by the community and in most cases, allowed them
to expand the pre‐existing irrigated area. However, the recent floods in Mustahil
swamped some of the irrigated plots and reduced the area under cultivation in
other former ABG sites.
6. The pump‐based irrigation supported by the project produced substantial gains
in terms of the volume of agricultural production, and income from the sales of
produce, in all three woredas. These trends corresponded with increases in
household consumption of produce, and more expenditure on health and
education. These findings pointed to an overall benefit to those households
which continued to farm, relative to the pre‐project period. However, from the
perspective of aid investment these benefits have to be compared with the
costs of the project, and the increased costs incurred by households (e.g. for
fuel and spare parts) for using pump irrigation.
7. Not all project participants maintained their plots, and, local informants
described the various risks of crop production and marketing in Gode zone, and
associated impacts on harvests. These risks were numerous and included crop
diseases, crop damage by wild animals, flooding, wind storms, salinity problems,
human diseases, fuel price increases, storage problems, market instability, and
conflict.
8. The issues outlined in point 6 above were captured in the cost‐benefit analysis
for the project in Gode zone. When only the L‐SAP project period was
considered, the cost‐benefit ratio was 1:0.4; if a longer time period was
considered, covering three crop plantations, the cost‐benefit ratio was 1:0.7.
Although the benefits increased over time, as more crops were produced and
sold, the costs also increased over time as the water pumps became less
efficient, and required more fuel and spare parts. In addition, it was evident that
water pumps could be acquired privately in Gode at approximately half the cost
of the pumps which were supplied by CHF.
9. The assessment covered various policy and institutional issues related to small‐
scale irrigation. Of particular relevance was the history of riverine irrigation in
Somalia, where minority ethnic groups practicing irrigation were displaced from
their land and suffered extreme impoverishment. This displacement occurred
when land was appropriated by private commercial companies and later,
through the violent actions of more powerful clans. In Ethiopia, although the
Rer Barre are associated with the Hawiye clan, they could be considered a
minority group. If so, their food insecurity not only relates to the risks of a crop‐
based livelihood summarized in point 7, but also to the risk of displacement and
limited security of land tenure.

ii
Main recommendations
1. Support to small‐scale irrigation is a development activity, requiring good initial
analysis of the pre‐existing systems, and related constraints and opportunities.
Constraints are not only technical, but relate to complex social, environmental,
policy and institutional issues. A one‐year project timeframe is not sufficient
either to ensure the success of new irrigation schemes or to assess the results.
2. CHF and OFDA should not directly support small‐scale (i.e. one‐piston pump)
group‐based irrigation schemes targeted at poor households but instead,
consider how best to encourage the strengthening and appropriate expansion
of pre‐existing, privately‐run schemes. Poor households might then access the
private service through a voucher system. As water pumps can be purchased
locally in Gode at half the cost of the pumps provided by CHF, the delivery of
more free pumps through aid projects is difficult to justify.
3. Technical approaches to be tested include spreading the costs of future
schemes by ensuring multiple productive water uses, and, promoting other low‐
input water technologies such as rainwater harvesting, treadle pumps, etc. that
can potentially be used not only for irrigation, but also for rainfed farming.
4. Support and stimulate the private and governmental sectors as appropriate for
the supply of and access to the other inputs necessary for the success of
irrigated and rain‐fed farming, such as agricultural and health extension
support, supply chains for farm inputs and credit, improved post‐harvest
storage, and marketing options.
5. Options in 2 to 4 above require good local analysis with stakeholders. In
addition, future projects need to consider risks and how to minimize them. For
some risks, such as the control of pests or wild animals, further costs will be
incurred which in turn, further question the economic viability of the systems.
6. In the absence of secured land tenure, crop farming remains especially risky for
minority groups. This raises the question of how NGOs should engage regional
government on land tenure issues e.g. through linkages with the Ethiopia Land
Administration Project.

iii

1.
 INTRODUCTION

Under the 12‐month emergency Livelihood Support for Somali Agro‐Pastoralists

program (L‐SAP), funded by the Office for Foreign Disaster Assistance (OFDA), CHF
International (CHF) established small scale irrigation interventions along the Shabelle
river, in three woredas of Gode zone, in Somali National Regional State. The project
set up 18 ‘Asset Building Groups’ (ABGs), with the aim of building household assets
through diversifying farm production and increasing income from irrigated crop
farming. The project helped to organise the allocation of land to the groups and
provided Indian‐made Anil water pumps, fuel, farm tools and seeds and training on
irrigated farming. The project ended in December 2008.

This report describes an impact assessment of the ABG irrigated farming activities,
conducted in July and August 2010. The assessment focused on household income
and asset creation, and the performance and the sustainability of the groups since
the project ended, with a view to learning lessons to inform the scaling up of similar
interventions in future. The assessment also considered the social and
environmental impacts of the intervention. Of the 15 ABGs reported to be still
functioning at the time of the assessment, 13 ABGs were visited.

The report begins with a short background section on the development of small‐
scale irrigation generally in Ethiopia (section 1.1), and then describes the CHF project
intervention (section 1.2) and the assessment methodology (section 2). The report
then gives a brief history of crop farming and the development of irrigated
agriculture in the project area, placing the CHF intervention in an historical and
geographical context, and describes the results of the impact assessment, including a
cost‐benefit analysis (section 3). The report ends with a discussion and set of
recommendations (section 4).

1.1 Background to small‐scale irrigation development in Ethiopia

Large‐scale irrigation programs and related technologies are relatively well known in
Ethiopia, and the government actively promoted these schemes during the 1970s
and 1980s. However, during the 1990s many irrigated state farms were abandoned
and investment in large‐scale and medium‐scale schemes stagnated. At the same
time, there was a corresponding expansion of small‐scale communal irrigation
schemes. Until recently there has been relatively limited government investment in
improving these traditional small‐scale irrigation schemes or in expanding modern
schemes either through providing incentives to encourage private sector
involvement, or through government research and extension programs. The
government’s Water Sector Development Strategy (2002) and Plan for Accelerated
and Sustained Development to End Poverty (PASDEP), 2005/6‐2009/10 emphasises
the importance of irrigation development in stimulating rural economic growth and
development, and ensuring long‐term food security, and sees the further

1
development of small‐scale irrigation and water harvesting, along with agricultural
research, as playing a significant role.1

Estimates of the potential irrigable area in Ethiopia vary. For the Somali Region, an
IWMI (2005) report uses estimates from a 2004 study that suggest a total irrigable
potential of 500,000 ha, of which only around 12,000 ha (2.4%) was under irrigation
of any kind. Of this, the report estimates only 1,800 ha (15%) to be under modern
small‐scale irrigation schemes, whilst around 8,200 ha was under traditional small‐
scale irrigation2. Another report suggests that around 265,000 ha of land in the
Wabi Shabelle basin is ‘very suitable’ for irrigation, whilst around 90,000 ha is
‘moderately suitable’3. Traditional small‐scale irrigation schemes can include spate
irrigation, controlled or uncontrolled flooding, lift irrigation using buckets or gravity‐
fed surface irrigation canals. These schemes are usually organised by groups of
interested households, sometimes with government technical and material support.
The users typically farm small plots of 0.25 to 0.5 ha and produce both food and cash
crops4. In contrast, modern small‐scale schemes may involve the construction of
dams, river diversions, the use of water pumps, or sprinkler or drip irrigation
systems.

Until relatively recently, the modern schemes depended on the support of

government or NGOs, which typically, ‘handed over’ to water user or management
committees once the scheme was completed, sometimes with an expectation of
ongoing technical or material support from the government or NGO. These schemes
were usually designed with an emphasis on cash crops. What the traditional and
modern schemes have in common, and what distinguishes them as ‘small‐scale’ is
that they are generally owned and managed by the farmers themselves, on their
own land5.

In terms of government responsibility for small‐scale irrigation, in the past the

mandate was shared between the Ministry of Agriculture and Rural Development
(MOARD) and the Ministry of Water Resources (MoWR). In 2004, the responsibility
for irrigation at the federal level was handed over from the MoWR to the MoARD.

The potential benefits of small‐scale irrigated agriculture are well known ‐ irrigation
can mitigate against rainfall variability and crop failure, allowing farmers to intensify
their production by producing two or three harvests annually (rather than one), and
increase their incomes through the production of high value cash crops. The
potential impact on livelihoods includes higher consumption and better nutrition

1
The Water Sector Development Program planned to develop an additional 127,138 ha of small‐scale
irrigation during the period 2002‐2016.
2
Awulachew, S.B. et al. (2005), Experiences and opportunities for promoting small‐scale/ micro
irrigation and rainwater harvesting for food security in Ethiopia. International Water Management
Institute (IWMI), Working Paper 98.
3
Ayele Gebre‐Mariam (2005). The Critical Issue of Land Ownership. WP 1 ‘Governance and Conflict
Transformation’, Working Paper No. 2, Bern: NCCR North‐South.
4
5
Awulachew, S.B. et al. (2005), ibid.
Carter, R. and Danert, K. (2006). FARM Africa Ethiopia: Planning for Small‐Scale Irrigation
Interventions. FARM Africa Working Paper No. 4.

2
(through crop diversification), purchase of assets such as livestock, and investment in
children’s education. However, there are also significant challenges and potentially
negative impacts including the high cost of inputs (pumps, fuel, agricultural inputs),
pests and diseases affecting high value horticultural crops, storage problems and the
perishability of some crops, plus transportation and market access problems, and
increased labour demands.

The literature on the development of small‐scale irrigation in Ethiopia highlights a

number of technical, social, management and institutional conditions that need to
be in place for the successful, sustainable development of small‐scale irrigation
schemes. This is discussed further in the light of the assessment findings, at the end
of this report (section 4).

1.2 CHF L‐SAP Project background

In September 2007, CHF was awarded a Cooperative Agreement from OFDA to

implement L‐SAP over a 12‐month period. The project was an emergency response
to the drought of 2006 in agro‐pastoral and pastoral areas of Somali Region, and the
goal was to, ‘To improve the capacity of agro‐pastoralists from selected woredas in
Gode Zone to cope with shocks and achieve sustainable and improving livelihoods’.

Within L‐SAP, the ‘Agriculture and Food Security’ sub‐component of the project
aimed to assist vulnerable agro‐pastoral communities in Gode, Kelafo and Mustahil
woredas in Gode zone ‘to increase and diversify their agricultural production, which
will enhance income generation and enable the accumulation of essential household
assets.’ This was to be done through helping communities to establish small‐scale,
pump‐based irrigation schemes along the Wabe Shabelle river for the production of
maize for home consumption, as well as cash crops – fruit, vegetables and oil crops ‐
for sale. A key element of the intervention was the formation of the ABGs which on a
cooperative basis, were to share the project inputs, manage the scheme and share
the benefits. These groups were to begin a group savings and loans program with the
aim of covering the future costs of pump maintenance and repair, as well to
generate capital for other group and/or individual‐based income generating
activities. In this way, CHF hoped to respond to the emergency and at the same time
support vulnerable households to increase their income and assets and thus mitigate
the effects of future droughts.

The project gained a three‐month cost extension from September 2008, and
officially ended in December 2008. After the project, CHF had minimal contact with
the ABGs. In September 2009 and before an external evaluation of the L‐SAP, CHF
were awarded USAID funding for another small‐scale irrigation project which would
enable them to scale‐up the L‐SAP approach with new ABGs in all five woredas of
Gode zone. An impact assessment was requested as a means of learning lessons
from L‐SAP that could inform the new project and other government and INGO
small‐scale irrigation projects in the zone. It also provided a good opportunity to look
at sustainability issues almost one and a half years after the initial L‐SAP project.

1.3
 Project implementation

Following approval of L‐SAP in late 2007/early 2008, CHF and the Water and
Agriculture Bureaus of Gode zone selected six potential project sites each in Gode,
Kelafo and Mustahil woredas for the establishment of 18 ABGs (six per woreda),
which would establish an irrigation scheme in their communities. At the same time,
CHF negotiated with local clan leaders to allocate 25 hectares (ha) of land close to
the river and suitable for irrigation, for each ABG. Each community then selected 50
vulnerable households to be a member of the ABG, with each household using a plot
of around 0.5 ha. The criteria for the selection of the ABG members included:
 the level of poverty of the community i.e. the ABG members did not own
individual private irrigation plots before and could not afford to buy a pump
by themselves;
 the ABG members live close to the area allocated for irrigation and belong to
the same sub‐clan;
 the ABG members are willing to participate in the project and willing and
able to contribute to the purchase of the water pumps;
 the majority of the participants should be female (specific to Gode woreda).

CHF provided each ABG with two Anil‐Indian water pumps of one piston capacity,
fuel for the first production season, farm tools and 10 quintals of seeds including
maize, sesame, onion and tomato, and banana plants. Details of the inputs provided
are shown in Annex 1. The project also conducted a number of trainings e.g. in
irrigated agriculture, in nursery establishment, savings and credit etc.

Each group was asked to make an initial contribution of ETB 50 per member (total of
ETB 2,500 per ABG) to set up an ‘operations and maintenance’ (O&M) fund, in
addition to a monthly savings contribution of around ETB 3 per member, which was
to cover the salary of the pump operator. However, because of the short project
time frame and CHF internal staffing constraints, the planned savings and credit
programme did not materialise and the ABGs made their own arrangements for
covering these costs6.

6
Information provided by senior CHF management staff.

2. ASSESSMENT METHODOLOGY
The assessment was based on two main activities. First, a field assessment was
designed with CHF to examine the performance of the irrigation schemes and
related issues. Second, economic analyses were conducted which in part, drew on
data collected in the field.

2.1 Field assessment design

The initial design of the study was based on information provided by CHF and an
understanding that the ABGs differed both in their locations and in the participants’
main sources of livelihood prior to the project. Therefore, two main clusters of ABGs
were identified in initial meetings with CHF:
i. Gode ABGs, made up of impoverished ex‐pastoralists, largely women.
ii. Kelafo and Mustahil ABGs, formed predominantly by members of the Jirir
tribe7 who traditionally, depended on crop farming for their livelihoods.
However, during the early stage of the field work the authors noted that the Gode
groups were predominantly agro‐pastoralists, who were permanently settled and
who had conducted rain‐fed crop farming in the area for many years. As a result, the
ABGs were re‐classified based on the three separate locations, as Gode, Kelafo and
Mustahil ABGs.

Based on discussions with CHF senior management and program staff, the study was
designed around six key questions relating to the potential livelihoods impact of the
irrigation intervention and the implications for scaling‐up. Indicators were developed
around each of these questions, leading to the development of a set of participatory
tools to measure the indicators.

Key question Indicators

 What was the household‐level livelihoods  Household income from irrigated farming; uses
impact of the intervention? of income obtained
 How sustainable is the intervention, both  Functioning and performance of the ABG groups
in terms of the sustainability of the ABGs post‐project, without any external support
and the irrigation scheme?  Cost of inputs deployed (by project) and value of
 How did the formal or informal policy produce sales from irrigated farming
and institutional environment assist or  ABG utilisation of project inputs provided
constrain the project, for example in  Land use/access arrangements, changes in land
relation to clan, land access or trade use
issues?  ABG access to markets (internal and cross‐
 What are the main opportunities and border trade);
constraints around marketing of ABG  Documented good practice and key lessons for
produce, for the existing groups and for improved program delivery and policy making.
new groups?
 What are the implications for scaling up
in relation to organisational technical
capacities?

7
In the literature, this group is known as the Rer Barre, farmers of Bantu descent who are associated
with the Hawiye clan in Gode zone.

5
The assessment collected data from two main areas viz. project areas and non‐
project areas.

Project areas ‐ primary data was collected as follows:

 individual household interviews with ABG participants to capture perceptions of
impact at the household level;
 focus group discussions with ABG members to capture information related to
group performance, and the costs and benefits of the irrigation scheme, as well
as to triangulate the data from the individual interviews;
 key informant interviews with,
o elders and religious leaders from the sampled communities to identify
changes in community wealth status;
o irrigation experts involved in the design and implementation of irrigation
programs in Gode zone.
These interviews were also used to triangulate data from the individual and
focus group discussions.

Non‐project areas ‐ secondary data was collected from non‐project sites and from
discussions with key informants in Gode town.

2.1.1 Assessment areas

The CHF intervention area comprised riverine areas of Gode, Kelafo and Mustahil
woredas in Gode zone, Somali Region (Figure 1).

Figure 1. Map of Somali Region, Ethiopia, showing CHF project woredas

Source: UNOCHA Ethiopia

6
2.1.2 Sampling method and sample sizes

L‐SAP reports and CHF staff confirmed that the performance of the ABGs differed by
location and therefore the geographical sampling was stratified to cover all three
woredas – Gode, Kelafo and Mustahil. CHF informed the PIA team that three of the
Gode ABGs were no longer functioning, that one of the Mustahil ABGs was
inaccessible due to flooding, and that one of the ABGs in Kelafo was not secure
enough to visit8. Therefore, 13 operational ABGs (with 650 registered members)
were included in the assessment.
 In Gode, all three of the reported functioning ABGs were selected and
households for individual interview were randomly sampled from these AGBs,
using a list provided by CHF.
 In Kelafo and Mustahil, a random stratified sampling approach was used to
identify both the sample sites for the focus group discussions and the households
for individual interview within each assessment woreda. The sampling frame was
derived from the lists of ABG sites and household participants of the ABGs
established by CHF in each assessment district.

Individual household interviews were held with a total of 104 households (16% of
the 650 registered members). Focus group discussions (FGDs) were held in 8 of the
operational and accessible ABGs at the time of the assessment. Both the household
interviews and the FGDs were pre‐arranged by the CHF field staff and were mostly
held at, or close to, the ABG irrigation sites9.

It should be noted that when first visiting the field, it became evident that the ABGs
in Kelafo and Mustahil were no longer functioning as groups.

Table 1 Distribution of sampled ABGs

Woreda Number of ABGs Number of Number of
Household Focus groups
interviews
Established Reported as Assessed Female Male
2008 functional 2010
2010
Gode 6 3‐4 3 35 0 3
Kelafo 6 6 5 21 15 3
Mustahil 6 6 5 15 18 2
Total 18 15 13 71 33 8

Gode ‐ All of the Gode ABGs, including those classified as ‘non‐operational’ were
visited. Bardo‐quorah was visited for the pre‐test and was found to be partially
functioning. The team visited Dabaldag and Dud‐ade for the collection of secondary
data and found that both of these ABGs disbanded prior to the first harvest.

8
CHF staff had not been able to visit this site since the project started for security reasons.
9
The Qonlawe ABG was not accessible and the interviews were conducted at the Danyere ABG site.

7
Kelafo ‐ In Kelafo, the Luqdhere ABG was excluded because of ongoing security
problems.

Mustahil ‐ In Mustahil, the Ibad‐badi site was excluded since it could not be accessed
because of floodwater remaining from the severe flooding of March and April 2010.
A field researcher visited all of the other sampled sites except Qonlawe in Mustahil,
which was also inaccessible because of flooding.

2.1.3 Data collection methods

The assessment was conducted by Tufts University from July 13 to August 4, 2010.
One member of the assessment team spoke the Somali language.

Individual household interviews ‐ were conducted using semi‐structured interviews

and were structured around a standardized set of exercises using participatory
methods, particularly ‘before’ and ‘after’ proportional piling. These tools were used
to capture perceptions of change in overall household production and the
proportion of total production obtained from farming in 2007 i.e. before the project
established the ABG irrigated plot, and in 2010. The same tool was used to capture
perceptions of change in the relative proportion of farm production that was
consumed (by humans or livestock) and sold, as well as to identify changes in the
uses of the income derived from produce sales, particularly the proportion spent on
farm inputs.

Focus group discussions ‐ were conducted with ABG participants at 8 ABG sites
(Table 1). The focus group discussions were held with ABG members other than
those involved in the individual interviews. The discussions were primarily used to
capture quantitative and qualitative data relating to the ABG’s performance and
achievements since the start of the L‐SAP project, including data on the inputs
received from CHF and the use of these inputs, the costs incurred and benefits
obtained from the irrigation scheme to date and any challenges faced by the group.
A timeline was developed with each group to define the project’s temporal
boundaries and to understand the history of rain‐fed and irrigated farming activities
in their respective areas.

Key informant interviews ‐ were held with elders at the Gode ABG sites. These were
structured around a ‘before’ and ‘after’ proportional piling exercise to capture and
explain any changes in community wealth classes between 2007 (pre‐project) and
2010. Key informant interviews with government and INGO irrigation experts
consulted in Gode town provided useful information on irrigation program
interventions in south Gode by government and aid agencies since the early 1960s.

Secondary data ‐ was derived from the literature and key informant interviews and
community discussions as well as direct observation during field visits to the three
‘non‐operational’ ABG sites in Gode and through site visits and/or informal
discussions with ABG participants in the five ‘non‐FGD’ sites in Kelafo and Mustahil.

Table 2
 Summary of participatory methods
Method Use/issue Sample
Before and after To measure:
proportional piling  relative changes in overall HH production 104 households
 relative changes in HH production derived from
farming
 relative changes in proportion of farm production
consumed and sold
 relative changes in uses of HH income derived
from farm sales
 relative changes in the proportion of time spent
on farm activities
 relative changes in wealth, assessed in terms of
receipt or payment of zakat
Focus group discussions To collect: 8 groups
 qualitative perceptions of the impact of the CHF
inputs
 quantitative data on the number and type of
plantations and estimates of the yield and cash
income obtained since project start
 data on changes in the size and gender
composition of the ABG and the size of land
holdings since project start
 crop preference rankings for irrigated crops
 qualitative information on the challenges
encountered with the project and irrigated
farming activities generally
Key informant To:
interviews  measure relative changes in community wealth Gode ABG site
classes between 2007 and 2010, and reasons
 collect technical information and timelines Various
relating to irrigation program interventions in
south Gode
Semi‐structured Used with all methods to cross‐check information and
interviews clarify responses

2.1.4 Pre‐testing

Field testing of the data collection tools was done on July 14 to 15, 2010. The FGD
pre‐test took place at Bardo‐quorah, one of the sites identified by CHF as non‐
operational. During this visit the team identified that the Gode ABGs were not ‘ex‐
pastoralists’ but were from well‐established, settled communities, with a history of
engaging in rain‐fed and irrigated farming. The pre‐test of the household interviews
was done at the Barsan ABG site. Following the pre‐tests, the research team revised
the tools and finalised the data sheets.

9
2.1.5 Triangulation

Prior to the field work, one of the research team members reviewed the project field
and donor reports and ABG production data provided by CHF. This included data on
the amount and type of crops harvested and area planted by ABG for the first
production season. During the field work, documentation in the Gode, Kelafo and
Mustahil field offices was used to match project service delivery and inputs with the
outputs and impact captured during the assessment. Unfortunately, CHF records
were incomplete and since the ABGs did not have any written records of inputs
received, expenditure or crop production data, the study team relied on oral reports
from the FGDs to cross‐check this data. The CHF production data could not be
verified by these oral reports and did not match with technical data relating to the
capacity of the pumps, the area irrigated, the number of irrigations required per crop
or with crop yield data for the area.

Other resource documents that were used for triangulation were the Save the
Children UK/DPPC Livelihoods Profiles (2008) for Somali Region, particularly the
Shabelle Riverine Livelihood Zone. Focus group discussions and key informant
interviews were also used to triangulate the data collected during the household
interviews. This was done by collecting more qualitative data on project impact, and
comparing the perceptions of the FGD participants with the more quantitative
perceptions captured in the household interviews. Finally, the assessment team held
discussions with CHF management and field staff before and during the assessment,
and during a feedback meeting at the end of the assessment in the Gode field office.

2.1.6 Data analysis

All the quantitative data from the household interviews was analysed using SPSS in
PASW Statistics 18 and Microsoft Office Excel version 2007.

2.2 Economic analyses

This component of the study focused on a cost‐benefit analysis (CBA) of the CHF
irrigation activities under L‐SAP. The analysis used:
 Data provided by CHF on project costs, including materials inputs such as
water pumps, spares, fuel, seeds and farm tools, and operational costs such
as staff time, travel, per diems and organizational overheads.
 Data collected from the field on crop production and sales, and costs incurred
by farmers and agro‐pastoralists.

As explained in later sections of the report, the CBA was restricted to Gode, where
relatively better production and sales data was available, and where project
attribution was less affected by various private sector activities around irrigation,
and other projects.

10

3.
3.1

ASSESSMENT FINDINGS

Background information on crop farming along the Wabe Shabelle River

3.1.1 Traditional land management and livelihood systems

The Wabe Shabelle riverine area passes through Gode zone. Approximately 95% of
the zone is grazing land, while the remaining 5% is arable. Of the arable land, around
20% was under cultivation in mid 2008, mostly using irrigation along the river. The
rural population of the assessment woredas make up around 65% of the rural
population of the zone, with around 52% of people living in the Shabelle riverine
area10 .

Table 3 Population of assessment woredas

Woreda Rural population Population in riverine area

Mustahil 53,227 37,259
Kelafo 88,156 61,709
Gode 95,647 23,912

Gode zone, and particularly Gode woreda is among the most drought‐affected
pastoral areas in Ethiopia. At the same time, the communities located in the low‐
lying areas along the river in Mustahil and Kelafo are also affected by regular
flooding caused by heavy rainfall in the highlands. The floods destroy crops and
livestock, but also provide an opportunity to practice flood recession agriculture.

The people in the rural areas are a mix of crop farmers and pastoralists who make
their livelihoods largely from crop farming and livestock rearing respectively. The
pastoralists tend cattle, camels, small ruminants and donkeys through seasonal
movements between the wet season grazing areas in the interior and the dry season
grazing areas along the Shabelle river. The farmers depend mainly on growing
maize, produced under gravity or pump‐fed surface irrigation along the river banks 11,
and sorghum which is traditionally grown in the two rainy seasons – the main gu
rainy season and the shorter deyr rainy season. These farmers also have small herds
of cattle, small ruminants and donkeys, which they use to transport produce to local
markets and for collecting and transporting firewood.

Along the Shabelle river, farmers also use irrigation to produce cash crops such as
sesame and onion, usually on relatively small plots of land (0.5 to 5 ha) very close to
the river where irrigation is easier. However, because of the hot, dry climate even
farmers practicing irrigated agriculture tend to focus production in the rainy seasons,
either because traditional gravity‐fed irrigation channels are full and it is easy to
irrigate using buckets, or if they are using pumps, because they can save on the cost

1
0 Save the Children UK and Disaster Prevention and Preparedness Agency (2008); Livelihoods and
Vulnerabilities: An understanding of livelihoods in Somali Regional State, Ethiopia, Updated version,
June 2008, Addis Ababa, Ethiopia.
11
Gravity irrigation can be used where the river is embanked; where riverbeds are deep, pumps are
needed; farmers also plant maize and sorghum in riverbeds.

11
of fuel. The dry season (hagay) is also known for strong winds that create dust and
sand storms making farm work difficult and covering crops. In Kelafo and Mustahil
woredas, regular flooding allows farmers to engage in uncontrolled or controlled (i.e.
using channels or pumps) flood‐recession production of staple crops further back
from the river. Although yields can be good, flood‐recession farming is risky as floods
can return before the crops are harvested.

Other economic activities, particularly for poorer households include the sale of
charcoal and firewood and grass for thatching and for livestock feed, as well as farm
labour.

As in other parts of the region, in all of the areas covered by the assessment, land
use and access is collectively managed by the clan and sub‐clan lineages occupying
the area. Clan members have the right to fence land and conduct irrigated farming
and there is no limitation on the amount of land that any household can use.
However, the clan has the right to redistribute the best land e.g. the land closest to
the river that is easier to irrigate, amongst its clan members at any time. This could
include the plots of individuals already involved in irrigated farming at the time of
the redistribution. Non‐clan members who intend to conduct irrigated farming must
get permission from the clan that owns the land.

While Gode woreda is dominated by the Ogadeni clans, Kelafo and Mustahil woredas
are a mix of Ogadeni and Rer Barre groups. In these areas, pastoralists and crop
farmers differ either by tribe or clan and inter‐ethnic conflicts often break out,
affecting the livelihoods of both groups. The farmers are mainly Jirir from the Rer
Barre ethnic group, while the pastoralists in the area are mainly Jile from the Ogaden
and Hawiya ethnic groups12. The traditional dry season grazing areas of the Jile
pastoralists include lands used for irrigated crop farming in the Shabelle riverine
areas. Conflicts arise as pastoralists attempt to access potential dry season grazing
land and livestock water points and damage crops, either accidentally through
livestock escaping to fields close to the roads leading to the water points, or
intentionally by pastoralists releasing livestock into the farmers’ crops. Although
these conflicts are sometimes violent, there are also traditional processes of
cooperation and exchange relations which prevail between the pastoralist and
farming/ agro‐pastoralist clans13. Issues around land ownership and user rights over
land used for irrigated farming also cause conflict between the Ogaden and Rer
Barre and the Hawiya and Rer Barre tribes in areas where both the Jile and the Jirir
conduct irrigated farming.

1
2 Jile and Jirir are Somali names given to Somalis with ‘soft hair’ (from the Ogaden/Hawiya tribes) and
‘hard hair’ (from the Rer Barre tribe) respectively.
13
Ayele Gebre‐Mariam (2005), ibid.

12
3.1.2 The development of irrigated agriculture in south Gode

South Gode refers to the Shabelle riverine area south of Gode town through Kelafo
and Mustahil woredas to the border with Somalia. West Gode refers to the Somali
regional government irrigated farm west of Gode town. Figure 2 summarises
irrigation‐related research and development work conducted in south Gode by the
government and NGOs since the 1960s.

Figure 2 Timeline of major irrigated farming related events in south Gode

Time Key event

1960s  Under Haile Selassie’s government, Alemaya College of Agriculture opened a research
centre in Gode with the main objective of assessing the possibility of using the Wabe
Shabelle River for large‐scale irrigated crop‐farming.
 The impact of the research centre included the initiation of crop farming in the Wabe
Shabelle River basin and the initial design of large‐scale irrigation programs in south
Gode, later developed by the EPRDF and Derg governments.
 Local Somali farmers in Gode produced food crops, mainly maize and sorghum,
through rain‐fed farming and fruit such as papaya and lemon using irrigated farming
methods, with the technical and material support of the Gode research centre.

1970s  Following the 1972‐74 drought, pastoralists were settled on around 1,000 ha of
irrigable land.
 During the Ethio‐Somalia war in 1977, the Somali crop farmers migrated to Somalia
and their land was taken over by the Derg government. The government’s motto was
‘we produce while fighting’.
 Consequently, a gravity‐based irrigation scheme was designed with the aim of building
a dam at west Gode14 and a series of canals to enable the development of a large‐scale
irrigation program for south Gode, Kelafo and Mustahil; this was known as the south
Gode irrigation program.
 In 1979, through its Relief and Rehabilitation Commission, the Derg government
brought settlers from Wollo and Tigray to south Gode. The settlers started producing
food crops through a pump‐based irrigation scheme at selected sites such as Welwel
and Barsan.

1980s  In 1980, the Derg established the south Gode state farm at Welwel, Gelewage, Barsan,
Gudare and Bedled.
 In 1981, the RRC transferred the responsibility of managing the south Gode irrigation
program to the Awash Irsha Limat Corporation, which established the Gode Irsha Limat
project.

1991  The Gode Irsha Limat project hired settlers from Wollo and Tigray and started pump‐
based irrigated production of cotton and fruit such as banana and papaya. This
program ended with the fall of the Derg regime in 1991.
1991/2  Returnees from Somalia claimed their land that had been converted into the state
farm, now occupied by the settlers.
1992/3  The EPRDF transitional government established a committee under the prime
minister’s office to handle issues related to claims on the state farm land by the Somali
returnees and the settlers. The settlers were returned home.
1994  The EPRDF government announced a new development motto ‘self‐sufficiency in food
crops’; the Gode state farm was transferred to the Somali Region Livestock, Crop and

1
4 This site was used for the dam built under the west Gode irrigation project designed for settlers by
the EPRDF government in the early 2000s.

13

1995
 Natural Resources Development Bureau (LCNRDB).
2000  The LCNRDB launched an irrigation project designed for returnees from Somalia with
claims to the previous state farm. Water pumps of 4 piston capacity were installed at
Gelewage (421 ha), Barsan (400 ha), Bedled (596 ha), Gudare (428 ha) and
Iskasheto/Welwel (118 ha) to be used by a total of 2,259 households. Each household
was given up to 1 ha land mainly for maize production on an individual basis, but using
state ‐owned tractors. Project beneficiaries contributed 500 ETB per household per
2001/2 cultivation period for fuel and running costs until the Ogaden drought of 1999.

 As a response to the drought, the government of Somali region decided to cover the
fuel cost for the pumps for the irrigation project; however, the area of land planted to
maize started shrinking, mainly because of the lack of budget for pump and tractor
2006 maintenance costs.
 CCM, an Italian NGO, implemented a food/cash for work program for the construction
2007/8 of canals for a gravity‐based irrigation scheme in Kelafo and Mustahil 15.
 The Pastoral Community Development Project (PCDP) was launched by the federal
2008 government in collaboration with the government of Somali region on a pilot basis in
selected woredas and kebeles; inputs delivered included water pumps of 4 piston
capacity, fuel, seeds, farm tools, and running costs delivered via the concerned
government bureaus.
2008/9  The PCDP phased out and farming activity was interrupted mainly due to problems
related to pump spare parts and maintenance costs.
 The International Committee of the Red Cross (ICRC) distributed pump spare parts to
2009 individuals and groups using pumps in Gode, Kelafo and Mustahil.
 CHF launched the L‐SAP project, to promote small‐scale irrigated farming in Gode,
Kelafo and Mustahil; they organised ABGs in selected villages and provided inputs
including water pumps of 1 piston capacity, fuel and seeds for the first cultivation
2009/10 period and farm tools. The project ended in Dec. 2008.
 The Wabi‐Shebele Development Association (WASDA) implemented a food/cash for
2010 work program for canal construction for a gravity‐based irrigation scheme in the Kelafo
area.
 The Somali Regional government launched a new irrigation project at Barsen and Iftin
for 800 households; 2‐piston capacity water pumps and running costs including fuel
were provided by FAO. The project failed immediately after the first cultivation period
mainly due to management problems.
 Save the Children US established new ABGs in selected villages in Gode woreda where
L‐SAP used to operate; inputs included small water pumps of 1 piston capacity, fuel
and seeds (mainly onion) for the first cultivation period and tools.
 Severe floods destroyed irrigated crops in the Mustahil area.
 Currently, the south Gode irrigation development project provides technical advice to
individuals involved in small‐scale irrigated farming using private pumps of 1 piston
capacity.

In all of the areas targeted by the various irrigation projects over the years, some of
the ex‐project beneficiaries have continued with irrigated farming using privately‐
owned pumps. The pumps are usually brought in as contraband goods from Somali
and are used either on a dhulqeyb (land sharing) or birgerac (pump renting) basis16.
Under dhulqeyb, in exchange for the use of the pump the landowner gives the pump
owner free use of half of the land to be irrigated by the pump. In this case, the pump

1
5 Participants of this scheme indicated that gravity‐based irrigation requires a two to three year
period to become functional, depending on the water levels in the Shabelle River.
16
A one‐piston pump brought from Somalia costs around ETB 9,000 to 10,000. The CHF pumps cost
ETB 19,000, since the CHF head office requires inputs to be purchased from a licensed supplier.

14
owner is responsible for oil and pump maintenance costs, while each party is
responsible for the fuel and labour inputs required for cultivating their respective
plots. This system is limited to the fertile lands located nearer the river that are used
for cash crop production.

Under the birgerac system, the pump users arrange their own fuel and also pay 10‐
12 ETB per hour for the pump during the cultivation period. Under this system, the
pump user pays 50% of the pump running costs per cultivation (ie. fuel and the
hourly payment) to the pump owner after the harvest; the pump owner also receives
land from the land owner for free. On average, it takes a small pump of one‐piston
capacity around 16 hours to irrigate one ha of maize17; since maize normally requires
four rounds of irrigation over a four to five month cultivation period, the pump user
could end up paying around ETB 640 to ETB 960 per ha of maize, plus fuel costs.
Annex 2 shows the number of irrigations normally required per crop per cultivation
period. The farmers form cooperatives, known as ishkasheto to collectively use
water pumps that belong to individual investors.18

The availability of private pumps for rent has been steadily increasing, particularly
since the demise of the government‐run irrigation projects. In Kelafo and Mustahil
particularly, increasing numbers of farmers and traders from Somalia are joining
local farmers in the irrigated production of cash crops, mainly onion, to meet the
increasing market demand in Somalia. This is partly due to the impact of Al‐Shabab
on farming activity in Somalia; many of these farmers left Somalia either because
their land was returned to its original owners by Al‐Shabab, or to escape the
stringent Sheriya system. According to cash crop traders in the area, mainly women
from both Ethiopia and Somalia, one of the reasons for the increased trade is the
improved security imposed by Al‐Shabab, and the removal of checkpoints (isbadho)
in the areas that it controls, including Baladweyn. Truck owners interviewed by the
field researcher said that Al‐Shabab charges a ‘reasonable’ amount of tax and this,
together with removal of the check points, has allowed an increase in the trucking of
cash crops from Kelafo and Mustahil for the Somali market. As a result, the demand
for onion has increased and the price has gone up from 160 ETB/quintal (100 kg) in
2009 to 450 ETB/quintal in July 2010, for a standing crop in the field. Although the
price of onion in the Jijiga and Dhegahbur markets in Ethiopia is 15 ETB/kg (1,500
ETB/quintal), access to these markets has been blocked since around 2005, when the
Somali regional government blocked the road to deny the ONLF access to income
(dheqala) from taxing the Somali trucks on the trade route.

3.1.3 Traditional crop farming practice

As already mentioned, irrigation is generally limited to the lands located close to the
river where farming with either a gravity or pump‐based surface irrigation system is
easier. Otherwise, flood recession and traditional rain‐fed farming are practiced.
Rain‐fed farming is dependent largely on the rainfall from the Hararghe and Bale

1
7 Estimated from manufacturer specifications, and cross‐checked with ABG members and private
pump owners.
18
Ayele Gebre‐Mariam (2005), ibid.

15
mountains to the north and west that determines the level of the Wabe Shabelle
River. Generally, the major crop plantation seasons happen twice per year towards
the end of the deyr and gu rainy seasons. The deyr plantation is dependent on the
short belg rains in the highland areas to the north‐west and particularly on the flood
water which comes from the northern part of the Somali region and adjacent
eastern Hararghe areas, which is used during the jilal dry season. The gu plantation
largely depends on the flood water from the long kremt rains in the Bale mountains,
which is used during the hagay dry season.

Farmers from the area believe that the flood water from the belg rains via the
northern part of Somali region is not good for crops while the flood water from the
Bale Mountains produces healthier crops. This perception coincides with the
relatively higher incidence of crop failure encountered in the jilal season than in the
hagay season, as reported by the ABGs. Although this needs to be further
investigated, the general understanding is that the concentration of salt is higher in
the flood water which comes from the Somali region compared with the flood water
from the Bale Mountains. If true, salinity could contribute to crop failures
encountered by the ABGs mostly with the jilal plantations (see below).

3.2 The L‐SAP irrigation project findings

The project timeline is shown in Figure 3 with other key events in the area.

Figure 3 L‐SAP project timeline

Date Project events General

2006 Poor gu and deyr rains
Sept‐Oct 2007 L‐SAP funding approved by OFDA,
Regional agreement signed
Oct‐Dec 2007 Recruitment and induction of L‐SAP staff; Poor deyr rains, drought
technical training of staff;
Jan‐Feb 2008 Site identification; ABG formation and Adoption of distress coping
beneficiary selection; technical training mechanisms (slaughter destocking);
and signing of MOUs with ABGs; purchase migration to river areas; livestock
of 36 pumps deaths.
Mar‐Apr 2008 Land preparation; Training of ABGs in Poor gu rain; Hyper‐inflation of staple
irrigated farming and nursery food prices
establishment; distribution of seeds and
tools
May‐Jun 2008 Distribution of water pumps (2 per ABG); Poor gu rain; livestock deaths
Distribution of fuel to ABGs for first round
of irrigation
July‐Sep 2008 Refresher training & supervision of ABGs;
3 months no‐cost extension approved
Nov 2008 L‐SAP field staff’s contracts ended Poor deyr rain
Dec 2008 Official end of L‐SAP

16
3.2.1 Operational status of ABGs

The L‐SAP project was implemented in a total of 18 kebeles across the three woredas
i.e. 6 kebeles each in Gode, Kelafo and Mustahil. The number of ABGs established by
CHF in 2008 and their current status are shown below.

Table 4 Operational status of ABGs, mid 2010

Woreda ABGs ABGs reported as Actual status in mid 2010

established functional
2008 2010
Gode 6 3‐4 Three ABGs still functioning on a group basis.
 Barsen and Ilan ABGs are still operating as a
group, but were struggling to cover the
costs of pump repairs and fuel.
 Dud‐Ade and Dabaldeg ABGs were
disbanded in 2008.
 Bardo‐quorah ABG was linked to SCUS.
 Bahktiley ABG members were not farming
as a group.

Kelafo 6 6 No ABGs still functioning as groups ‐ the Kelafo

ABGs were incorporated into various privately
and community‐managed irrigation schemes
using a mix of CHF, private and government
pumps that serve individual plots.

Mustahil 6 6 No ABGs still functioning as groups – as above,

the Mustahil ABGs were incorporated into
various privately and community‐managed
irrigation schemes using a mix of CHF, private
and government pumps that serve individual
plots. Because of recurrent floods, farmers are
practicing flood recession farming rather than
using pumps.

Gode ‐ of the ‘non‐operational’ ABGs in Gode:

 Bardo‐quorah was still functioning, at least in name. Following the breakdown of
the CHF pumps in 2009, this ABG could not afford to pay for spare parts and
instead purchased a subsidised pump from the government. The group said that
they have been operating at a loss and were renting a donkey cart19 to transport
firewood to Gode, which they sold to cover the costs of fuel. They requested
support for fuel and spare parts. One of the reasons originally given for the
disbandment of the group was that since the close of L‐SAP, Save the Children US
(SC US) has established a new irrigation scheme in the area, which included many
of the former Bardo‐quorah ABG members. Although this could not be fully

1
9 Half of the sale proceeds were paid to the cart owner.

17
 ascertained, on visiting the site it appeared that some of the former ABG
member households may be supported by the SC US project.
 Dud‐Ade and Dabaldeg ABGs both disbanded in 2008.
o In the case of Dud‐Ade, the ABG planted maize on 25 ha of land and
received an initial two barrels of fuel (400 litres) from CHF, which they
used for the first round of irrigation. The group were expecting to receive
an additional nine barrels (1,800 litres)20 from CHF but this fuel was
reportedly sold in Gode town before it reached the group, and the ABG
subsequently dissolved. It is not clear who sold the fuel ‐ according to CHF
staff, the ABG participants were farming on an individual rather than a
group basis and therefore it was difficult for them to hold either the ABG
committee, the kebele administration (or CHF) accountable for the loss of
the fuel.
o In the case of the Dabaldeg ABG, the site chosen for the ABG irrigation
scheme was a previously uncultivated site, close to the west Gode
irrigation project. When the assessment team visited, most of the
community members had left the village to become registered as settlers
with the west Gode project and there were no signs of any irrigation
equipment at the site. The group disbanded before the first plantation for
reasons that are unclear. However former ABG members suggested that
the site was not suitable for farming. Although the Tufts team did not
have time to investigate further, there may also have been some misuse
of fuel in Debaldeg21.

Kelafo and Mustahil ‐ in Kelafo and Mustahil all of the ABGs were reported by CHF as
still being operational. However, these communities were already involved in
irrigated farming using private water pumps well before the formation of the ABGs
by CHF. The CHF pumps delivered to the Kelafo ABGs were ‘pooled’ with the other
pumps already being used by the community, and enabled the community to expand
and intensify irrigated farming in the area. The ABG pumps and land have effectively
been incorporated into the various ongoing irrigation schemes in the community,
although the ABG participants, like others in the community, are conducting irrigated
farming on an individual, rather than a group basis. In most of these areas, the CHF
pumps have been handed over to individuals selected from the former ABG
members. In Mustahil, because of the recurring floods, farmers are practicing flood
recession farming, rather than using the pumps.

3.2.2 Group membership and land status

The project was targeted at 50 households per kebele and the CHF list of registered
ABG participants indicated a total of 900 direct beneficiary households. As soon as
2
0 Note that CHF records in the Gode field office showed a total of three barrels of fuel (600 litres)
being delivered to the Gode groups.
21
Note that misuse of fuel was also reported in one of the non‐sampled ABGs in Kelafo.

18
the project started, many of the groups voluntarily included additional, non‐
registered, ABG members who they felt met the selection criteria. All of the Gode
ABG members were female (except the chairman of the Barsan group), as per the
CHF selection criteria. Between the end of the project in December 2008 and the
time of the assessment in July 2010, there were further changes to the group
membership. For example, in some groups (Bakhtiley in Gode, Qufdhaley Kincho in
Kelafo) members left the ABG or the village because of economic problems.
However, because of the changes in most of the groups’ mode of operation from
farming as a group to farming as individuals on a private basis, as explained above, it
was not possible to give accurate figures of current ABG ‘membership’.

With regard to the land cultivated by the ABGs, the project was planned on the basis
of 25 ha per ABG, equivalent to 0.5 ha per household, and CHF negotiated with clan
leaders to allocate land to the ABG on that basis. During the FGDs in Gode, the
research team found that only around 5 to 10 ha had been cultivated by the groups.
This was supported by the groups’ accounts of the amount and types of crops
produced and sold, as well as by direct observation at the sites. In Kelafo and
Mustahil, as mentioned above, the land allocated to the ABGs was incorporated by
the community and in most cases, allowed them to expand the irrigated area,
although this could not be verified by the site visits. However, the recent floods in
Mustahil swamped some of the irrigated areas and reduced the area under
cultivation in other former ABG sites, for example God‐aro. Again, it was not possible
to accurately estimate the area of land currently under cultivation by (former) ABG
members.

3.2.3 Production histories

Although in theory irrigation means that crops can be planted outside of the main
rainy seasons, in practice farmers in the area stick to these seasons since they can
reduce the amount of water that needs to be pumped from the river thus reducing
the amount of fuel they need i.e. they practice supplementary rather than full crop
irrigation. In the dry, windy hagay season (July to September), crops need more
water and farmers are reluctant to plant. The CHF pumps were delivered nine
months into the L‐SAP project and three months before the original project end. This
timeline corresponds with the oral histories told by the FGDs, which indicate that
there have been four potential production seasons for maize since the ABGs were
formed until July 201022.

The timelines below, based on focus group discussions, summarise the history of the
ABGs assessed in Gode, Kelafo and Mustahil. As mentioned above, the CHF
production data for the 2008/09 cultivation season could not be verified, and the
groups themselves did not have any documented production records. Therefore, the
production data is based on the FGD discussions with the groups. While the Gode

22
At the time of the assessment in July 2010 the Tufts team noted that pumps for the USAID‐funded
SHAPE project had just arrived in Gode town, eight to nine months after project start‐up and four
months before the official project end in November 2010 (i.e. missing two potential production
seasons).

19
ABGs were able to provide aggregated production data for the group, in Kelafo and
Mustahil, where farming was individually based, it was not possible to obtain
estimates of the total production of the ABGs.

a. Gode

Information on the histories, activities and impacts of the three ABGs visited during
the assessment is provided below.

Barsen

1991 Barsan village created by returnees from Somalia.

1997 25 women raised funds to start crop farming and received 25 ha of former Derg state
farmland.
They produced vegetables using bucket irrigation, and rain‐fed and flood recession maize.
2001 They purchased a 1 piston capacity pump and started irrigated farming.
‐2007 They did not make a profit because of the high running costs of the pump.

2008 CHF re‐organised them into the ABG together with another 24 women and a man (ABG chair)
from the village. They received 2 Indian Anil water pumps of 1 piston capacity, 3 barrels of
fuel (600 litres)23, 10 quintals of seeds and various farm tools24. They added an additional 6
ha of land.
1st cultivation ‐ ABG members shared 50 quintals of maize between them i.e. 1 quintal per
person. The rest of the produce was sold and saved for fuel and spare parts for the pumps.

2009 Strong wind (dufan) and soqudud affected the first cultivation of maize and the crop failed.
2nd cultivation ‐ nothing was shared to the ABG members.

2009 Fuel price increased from 7.50 to 11 ETB/ litre; they purchased spare parts for >6,500 ETB.
They raised 150 ETB/person for fuel, spare parts and the pump operator’s salary (400
ETB/month).
3rd cultivation ‐ ABG sold 47 quintals maize for 250 ETB/quintal to WFP on credit 25. At the
time of the assessment the group had not yet received the payment

2010 They purchased a 1 piston capacity pump from the government for a subsidised price of
4,500 ETB. The fuel price increased to 17 ETB/ litre.
4th cultivation ‐ they shared 500 ETB/ person from maize by‐products sold as animal feed and
other farm produce. Spare parts have cost 7,000 ETB so far.

2
3 The three ABGs assessed in Gode woreda confirmed that the harvest collected from the first
cultivation period was produced only with the 600 litres of fuel received from CHF i.e. they did not
purchase additional fuel. Assuming a fuel consumption rate of 30 litres/ha (minimum), and that
maize requires four irrigations per four‐month cropping period (see Annex 2) this would be enough to
irrigate around five ha of maize.
24
Seed inputs included maize, sesame, onion, tomato, banana, beans and grass; six different types of
tools were provided. The seeds and tools provided to the ABGs were uniform in all three woredas
(see Annex 1).
25
The main reason why the Barsen ABG sold the maize on credit was due to the UN’s procurement
procedure that requires them to sign an agreement with the regional government to which the
payment should be made. Given this procedure and the fact that payment has been delayed for
around five months, CHF should follow up the payment on the ABG’s behalf, especially since the
women were not even clear whether FAO or WFP collected the maize.

20
Ilan

1999 Ilan village was created by members of the Reer Mahad Assie clan who started rain‐fed
farming on the former state farm on an individual basis.

2003 The government launched the PCDP at Ilan for 400 households on 400 ha land. They received
two water pumps of four‐piston capacity each, 25 kg of maize seeds and 600 litres of fuel for
the cultivation of 1 ha maize per household.

2008 CHF organised 50 women into an ABG and donated two water pumps, fuel (600 litres), seeds
and tools. The group was given 27 ha by the clan.
1st cultivation ‐ the group shared 100 ETB/ person from sale of produce from the first
cultivation

2009 The group produced fodder for own livestock.

2nd cultivation ‐ the first crop of maize failed due to dufan, but the ABG sold sesame and
tomato for 6,000 ETB; spare parts cost 2,750 ETB.

2009 They spent 1,750 ETB on spare parts, and raised additional money from the group for spare
parts and fuel. The members raised 50 ETB per person for costs.
3rd cultivation ‐ the second crop of maize and sorghum was sold as animal feed (for 5,000 ETB
and 1,500 ETB respectively); they also sold tomato (for 2,500 ETB); grass was fed to own
livestock.

2010 The ABG is cultivating maize for animal feed, tomato and water melon but salinity is
becoming a problem on the farm.

Bakhtiley

The village was so old that the ABG participants did not know exactly when it was formed.

1999 The name Bakhtiley was given to the village in connection with the massive livestock
mortality caused by the 1999 Ogaden drought.

2008 CHF organised 50 women into an ABG and gave pumps, fuel (600 litres), seeds and tools.
1st cultivation – the maize crop was lost to warthogs.

2009 The group could not afford to pay for the pump maintenance ‐ spare parts cost 7,950 ETB, so
they gave the pumps to Mr. Nuur Ibrahim the current pump operator who invested his two
cows (sold for 2,300 ETB and 1,100 ETB) to cover the cost26. Mr. Nuur charged 120 ETB per
plot per irrigation round, towards fuel and pump operating costs.

2010 The members began to farm on an individual basis; some established new plots close to the
river.

26
The Bakhtiley ABG called their plot ‘Isnugto’ meaning ‘self‐sucking’ because of the large proportion
of income from the sale of farm produce that has to be used for fuel and pump spare parts. In all of
the areas assessed in Gode woreda, the women ABG participants requested a donkey cart, to be used
partly for transporting firewood to sell in Gode town to raise cash for fuel and spare parts.

21
Upper left:
Upper right:
Lower left:
Lower right:

Wabe Shabelle River at Ilan

Barsan ABG water pump and pump operator
Salinity at Barsan
Irrigated plots at Bardo Qorah

Top left:
Top right:
Upper left:
Upper right:

Baktiley ABG plot

Maize stalk borer damage, Bardo Qorah
Water melon damage caused by porcupines, Ilan
Hagay wind at Dabaldeg

22
Top:

Failed banana crop at Ilan

Lower: Pepper crop at Ilan

b. Kelafo

It was not possible to collect production data for the former ABGs as groups since
farming was done on an individual basis.

Agweyne

1993 Agweyne village was created by ex‐pastoralists interested in rain‐fed and flood recession
farming.

2006 The village lost cattle, small ruminants and donkeys to the 2005/6 drought.

2007 28 members of the village contributed money to rent a water pump from Kelafo town; each
person planted half ha maize. The owner collected the pump at the end of 2007.

2008 CHF organized the ABG and provided two pumps, 12 barrels of fuel (2,400 litres), seeds and
tools.

2010 The ABG rented another pump as the CHF pump was unusable. The ABG purchased an
additional 105 litres of fuel.

Beladamin

1983 The Derg government launched an irrigation project with 2 German DOSS pumps of 4 piston
capacity and the Jirirs started maize farming as a cooperative; the government collected the
cost of fuel in kind after harvest. ‘The maize yield was so good that the life of the community
changed.’

1991 The Said Bare government dissolved and returnees from Somalia came to the Kelafo area.

23

1992
 After the downfall of the Derg, the Ogadenis destroyed the village and took the pumps; 2
months later the community developed canals and restarted irrigated farming through a
gravity system. The Jirir agreed that in future they would purchase a Kalashnikov after each
harvest in case of future attack.

1996 The level of the Shabelle River decreased and the gravity scheme failed to work.

1998 A village member brought a water pump of 4 piston capacity and irrigated
farming restarted through the birgerac (pump rental) system ‐ the pump was rented for 10
ETB/hour; fuel cost 6 ETB/litre. Irrigated farming was conducted with private pumps to 2008.

2008 CHF organized the ABG and provided two pumps, 12 barrels of fuel (2,400 litres), seeds and
tools.

2009 118 households shared the pumps and cultivated land on individual irrigated plots. In all of
the Kelafo ABGs,an additional 5 ha was cultivated for the kebele chair and vice chairpersons
and the pump operators27 .

2010 There were 12 private pumps operating in the village at the time of the assessment in August
2010.

Irkabanan

1983 The Derg government brought water pumps and the community began irrigated farming.

1992 Following the downfall of the Derg, the Ogadenis invaded the village and took the water
pumps. Three private pumps were brought to the village and irrigated farming restarted via
the birgerac system. The pumps were rented for 200 ETB/ha per irrigation round.

2008 CHF organized 50 households into the ABG and provided pumps, fuel (2,400 litres), seeds and
tools. The ABG purchased an additional 200 litres of fuel for the first cultivation.

2009 119 households shared the pumps and cultivated land individually

2010 The ABG purchased an Anil water pump of one‐piston capacity for 7,000 ETB from the
government with income from the sale of livestock 28.

c. Mustahil

A total of five ABGs were assessed in Mustahil woreda. Godaro is close to Mustahil
town, while Kunaso and Kobane ABGs are located to the north and Qonlawe and
Danyere ABGs to the south. Generally, the areas to the north of Mustahil are highly
flood prone and more suitable for livestock than crop production – the groups
reported only one successful cultivation during the last four potential production
seasons, because of floods29.

The Mustahil groups reported that the prevalence of water borne disease
particularly ‘bloody urine’ (schistosomiasis, also called bilharzia) has been increasing

2
7 Five ha represents half of the area that it is possible to irrigate with one pump of one‐piston
capacity.
28

sales after each harvest was to save money for pump maintenance costs.
29
Note that there have been further floods since the time of this assessment.
In Kelafo, the ABG participants indicated that the main purpose of purchasing livestock from crop
24
over the past few years30. This was borne out by the case record at the government
hospital in Mustahil which indicated that the number of cases admitted has
increased from two per month in 2007 to around seven per month in 2010. The
hospital has officially reported the increase to the Regional Health Bureau. A ‘before
and after’ scoring exercise with the Godaro ABG showed an increase in the incidence
of bloody urine, mainly in children, from around three cases in 2008 to seven cases
in 2010 and an increase in malaria from 12 to 15 cases per season over the same
period. The hospital is conducting extension work in areas where communities
engage in irrigated farming to advise farmers how to reduce the risk of infection e.g.
by discouraging children from playing barefoot in irrigation canals.

Kunaso

1990 Mardhis village split into two and Kunaso was created; the community began saving to buy a
water pump, but failed.

2000 CCM brought a DOSS pump of four‐piston capacity and fuel for one cultivation. The pump
was used from 2000 to 2004.

2008 CHF organized the ABG and provided two water pumps, fuel (1,400 litres), seeds and tools.
They purchased an extra 6 barrels of fuel for the 1st cultivation.

2009 There was a minor flood and the ABG was able to cultivate a 2 nd plantation following flood
recession; however, the harvest was nearly nil due to tree‐locust and disease outbreaks.

2010 They purchased eight barrels of fuel for the 3 rd plantation but a severe flood destroyed the
crops. Using the flood recession, the group cultivated a 4th plantation of maize.

Kobana

1998 Kobane village was formed by Jile (Hawiya) and Jirir agro‐pastoralists displaced by flood from
a village east of Mustahil; they started flood recession farming of maize and sorghum.

2005 A good harvest was obtained from maize and sorghum following flood recession (medium
flood).

2007 Sorghum fields were lost to birds.

2008 CHF organized the ABG and provided two pumps, fuel (1,100 litres), seeds and tools. The
ABG bought an extra 6 barrels of fuel and used 40 litres of fuel per plot.

2009 The 2nd plantation was lost to a tree‐locust outbreak.

2010 Severe flooding destroyed the maize and sesame cultivated during the 3 rd plantation
Increasing incidence of ‘bloody urine’, especially among children, reported to the Regional
Health Bureau.

3
0 Schistosomiasis is a parasitic disease transmitted by aquatic snails. The larvae of infected snails
penetrate human skin and over time can cause damage to internal organs and impair growth in
children. The disease is commonly associated with irrigated areas where surface water in canals
carries the infected snails. Risk can be minimised by good design of the scheme, such as
reducing/avoiding potential areas of ponding and allowing parts of the irrigation channels to dry out
periodically.

25
Upper left:
Upper right:
Mid left:
Mid right:
Lower left:
Lower right:

Irrigated plots, Kelafo

Irrigated plots, Kelafo
Sesame harvest, Kelafo
Transporting maize crop, Kelafo
Flooding, Mustahil
Cattle eating maize stove, Kelafo

3.2.4 Trends in production, costs, sales and net benefits

a. Trends in wealth groups, non‐ABG members in Gode

As indicated in preceding sections, in all of the areas covered by the assessment,

community members who were not ABG members were also involved in some kind
of irrigated farming using private pumps. ‘Before and after’ proportional piling of
community wealth classes was conducted with non‐project households in the Gode
ABG areas to identify any trends in wealth status between 2007 and 2010 in the
general population. The exercise was not conducted in Mustahil and Kelafo since the

26
communities did not make a distinction between ABG members and the general
community31. The results were:
 the proportion of ‘better‐off’ households decreased by around 9% (from 33%
to 24% of the population);
 the proportion of ‘poor’ households decreased by around 33% (from 67% to
34% of the population);
 the proportion of ‘middle’ households increased by around 42% (from 0 to
42% of the population).

The reduction in the poorest wealth category and increase in the middle category
was attributed to the increase in irrigated farming activity through water pumps
brought to the area both by government and private individuals. For example,
between 2009 and 2010, the regional government sold around 50 small (1‐piston
capacity) water pumps at subsidised prices to individuals and groups in Gode district
alone. As outlined already, in the Kelafo area, clan members returning from Somalia
carried their pumps with them, and their experience of irrigation in Somalia.

b. Trends among L‐SAP participants

Trends in farm production and related home consumption of produce and sales of
produce are shown in Figure 4, along with trends in expenditure, by type of
expenditure.

When viewing Figure 4 it should be noted that:

 for Gode woreda the assessment covered only the three out of six ABGs
which were reported as still functioning in mid 2010. Therefore the results
are biased towards remaining ABGs and overlook the limited impacts and
changes in the three non‐functioning ABGs.
 for Kelafo and Mustahil, project inputs were combined or absorbed into pre‐
existing irrigation systems by local communities. It was difficult to
disaggregate impact and changes due to project inputs from private
activities. The possibility that the trends shown are attributable mainly to
private sector or pre‐existing community arrangements has to be considered.

In terms of general trends the results showed that:

 Trends in production, consumption, sales and expenditure were similar in the
three woredas;
 Household production from farms was far higher in 2010 than in 2007, being
3.1 times higher in Gode, 2.2 times higher in Kelafo and 2.3 times higher in
Mustahil;
 Marked increases in the home consumption of farm produce were evident;
 Expenditure on farm supplies increased substantially, being 8.5 times higher
in Gode, 3.5 times higher in Kelafo and 3 times higher in Mustahil;

3
1 In Kelafo and Mustahil where all village members participated in irrigated farming in the ‘ABG’
areas , informants insisted that all households were poor in 2007.

27

Figure 4
 Trends in household financial indicators

Notes for Figure 4

1. Data derived from proportional piling with household informants. When viewing the graphs, the x‐
axis scales are arbitrary and are used only for illustrating the trends for each indicator between 2007
and 2010.
2. ‘Home consumption’ refers to all produce used in the home, including tomatoes, onions and
peppers for making sauce for human consumption, and grasses, maize and sorghum leaves used to
feed livestock.
3. Farm expenses for irrigated plots included cash spent on seed, fuel and spare parts for the water
pumps, farm tools and hired labour (for example, the pump operator or farm labourers).
4. Social expenses included weddings and funerals.
5. Statistically significant changes are illustrated using 95% confidence intervals for each indicator.

 increases in farm production and the need to acquire cash to meet rising
farm expenses, led to correspondingly large increases in the sale of farm
produce, being 5 times higher in Gode, 3 times higher in Kelafo and 2.7 times
higher in Mustahil;

28


 moderate increases in health, education, and household consumables were
evident, with education expenses related to both Koranic and formal
education;
expenditure on assets, especially livestock, increased slightly and was most
noticeable in Kelafo.

Table 5 represents the net profit obtained by the ABG households in 2007 and
(current/ former) ABG households in 2010, as measured using proportional piling.

Table 5 ABG household net profit from farming

Woreda Mean score 2007 Mean score 2010

Sales Expenditure Net profit Sales Expenditure Net profit
Gode (n=35) 11 5 6 54 42 12
Kelafo (n=36) 16 9 6 47 33 14
Mustahil (n=33) 18 12 6 48 36 12

There was a high level of agreement across the three areas regarding the net profit
from farming both before and after the L‐SAP. On average, the net profit from
farming doubled between 2007 and 2010 for the ABG households.

3.2.5 Time costs of farming by ABG households

In all areas, the amount of time spent on farming was far higher in 2010 than in 2007
(Figure 5). In Gode and Mustahil, households were spending at least twice as much
time on their farm as in 2007, while in Kelafo households are spending half as much
time again on the farm as in 2007.

Figure 5 Trends in time spent on farming by L‐SAP participants (n=104)

Mustahil

Kelafo
Before
After

Gode

0 20 40 60 80

Mean proportion (%) of time spent on farming

Note
Data derived from ‘before‐ and‐after’ scoring with 100 stones.

29
Reasons for these changes included:
 an increase in the scale and intensity of irrigated farming by the ABG
households;
 an increase in the scale and intensity of irrigated farming by communities
generally through private investment in irrigation, particularly since the
interruption of the large‐scale government irrigation schemes in 2006.

Specific to Mustahil, the floods of April and May 2010 destroyed the first round of
crops and meant that farmers had to cultivate twice already this year, partly
explaining the increase in time spent on farming in that area. In Kelafo, the increase
is relatively smaller probably since farmers in that area were already heavily involved
in farming, while in Gode, the increase is greater because they were less involved in
farming prior to the project, than in the other areas. The extra time was mainly used
for irrigating the plot and for guarding the crops against animal damage, both wild
and domestic (the risk increases during the dry season when there is a shortage of
their natural sources of food). These activities were mostly done during the evenings
and early morning. Therefore, although the workload increased, the additional time
did not necessarily affect the time spent on other, non‐farming activities.

3.2.6 Zekat

The assessment planned to use the payment and receipt of zekat by L‐SAP
households to cross‐check for any change in their income or wealth status. However,
although the number of sampled households receiving zekat declined from 7 in 2007
to 1 in 2010, the zekat payment system did not reflect actual household wealth
status because there are two types of zekat payment rules. Under ‘normal’ zekat a
household paid 10% of its income (in livestock, grain or cash), based on wealth,
whereas under wajib zekat all healthy adults voluntarily gave something to someone
in need in the community, in the form of firewood, water, labour etc.). Since 1977,
religious leaders have stopped collecting and dispatching zekat and it is generally
paid to someone in return for a service, such as Quranic teaching.

In the case of the Gode ABGs, the zekat payment was deducted by the ABG
committee from the group’s produce – 1 quintal for every 10 quintals produced by
the group.

3.2.7 Benefits and risks of growing different crop varieties

Local preferences for different crops are summarized in Table 8, with reasons for the
ranking. Across all areas, maize was ranked first followed by sesame, onion and
tomato (See Annex 4).

30

Table 6
 Crop preferences (n=9 groups)

Crop and median Reasons for ranking

rank (range)
Maize – 1st (1,3)  We can sell both as animal feed and human food
 We feed maize leaf to our livestock
 We can grow using flood recession
 It does not require guarding from birds
 Even when it fails, we can feed the leaf to calves and milking animals at
least
Sesame – 2nd (1,4)  There is relatively less disease problem
 It does not require guarding against wild or domestic animals except
camels
 It requires relatively less water
 It has a relatively good market comparatively
 It can be made into oil for home use

Onions – 3rd (1,4)  It is difficult to store when there is no market

 It is susceptible to disease
 It requires too much watering
 Low price and there is no alternative way of using bulk production
 It kills the soil

Tomatoes ‐ 4th (3,4)  Low price and there is no alternative way of using bulk production
 It is difficult to store as well as transport to distant located markets

The ranking and reasoning agreed with the information from focus groups on the
production history of the Gode ABGs, where all of the groups prioritised the
cultivation of maize (section 3.2.3a). All groups had sold maize stalks or leaves as
fodder in at least one of the four production seasons since L‐SAP started. The Gode
ABGs could not find a market for their tomato and onions in the first 2009 cultivation
and all groups also mentioned the high cost of irrigating onion, since it required
irrigation on average every three days for a six‐month growing period. Also, onions
required the purchase of pesticides needed, and spraying during each watering.

3.2.8 Perceptions of the benefits of irrigated farming

One of the key reasons identified by the study participants for undertaking irrigated
farming was to protect their livestock assets against recurrent drought by growing
crops for fodder. The value of fodder was explained by reference to two trends viz.
the reduction in pasture as traditional pastureland was converted to farmland, and
the damage being caused to traditional pasture due to invasion by Prosopis juliflora.
Informants referred to the former state farms, now abandoned, that were being
encroached by Prosopis (in Somali known as birsoobis, literally ‘when the stem is cut
it sprouts with shoots’)32; this was a particular problem in the Kelafo area. Related to

3
2 Ayele Gebre‐Mariam (2005), ibid.

31
 this trend, informants also referred to the benefit of farming for controlling the
expansion of Prosopis33.

Maize could be sold for fodder after two to three months, either as maize shoots or
at the green stage. This was practiced to reduce the cost of production by avoiding
the third and fourth rounds of irrigation needed to bring the maize to full term.
Selling maize for fodder at the early stage also reduced the risk of crop failure due to
disease and damage by wild animals34. However, the Somali regional government
had a policy of encouraging human food production in the region and government‐
supported farmers e.g. those receiving farm inputs or food aid were not allowed (in
theory) to produce maize for animal feed. Therefore, producers noted the potential
benefits of cashing in the maize crop early against the potential negative
consequences in terms of access to government assistance.

3.2.9 Common risks and challenges

Referring to repeated crop failures, informants highlighted some of the factors, both
natural and manmade, that contributed to the relatively low returns (in their view)
from pump‐based irrigated farming. In areas covered by the assessment, L‐SAP
participants encountered one to three total or partial crop failures between 2008
and 2010 from a potential of four harvests.

a. Environmental factors

Poor harvests – crop failures were attributed to disease35, pests including tree locust
and birds36, and damage from wild animals37. Crop failures were said to be more
likely in jilal season because of the ‘bad’ floodwater (see below) and because
damage from wild herbivores was more common than in the hagay season.
However, the hagay season plantation was also said to be difficult due to the wind
storms (dufan) that cause high water evaporation and bring heavy dust, increasing
the number of rounds of irrigation required and reducing seed germination
respectively.

Water and soil salinity ‐ all of the FGDs reported failures of crops planted in jilal for
at least one type of crop. As described above, these crop failures were at least partly

3
3 USAID already has a program for Prosopis control in Gode zone and there appears to be some
evidence that irrigated farming can help to stop the reinvasion of cleared land.
34
For example, Bakhtiley ABG lost maize planted during the first cultivation in 2008/9 due to damage
by warthogs and told the field researcher that they highly regretted not selling it earlier as fodder.
35
Gududiye ‐ a disease affecting all crops except sesame ‐ is named after the red colour of the roots of
the affected crop and was a commonly‐reported problem in all assessment areas.
36
In maize and sorghum, birds cause problems both by eating maize seeds from the field and by
attacking the matured sorghum. In the case of sorghum, the cost of production increases because
labour has to be hired to protect the sorghum fields. In the case of maize, farmers irrigate the newly
planted maize immediately to protect the seed against birds; this increases the number of irrigations
applied per cultivation from four to five rounds.
37
Warthog and porcupine are common problems in all areas covered by this assessment (see photo),
as well as hyena and fox, particularly affecting maize and watermelon.

32
associated with what was believed to be the relatively saline floodwater from the
highland belg rains and used for irrigation in jilal. Although the risk of increasing
secondary salinity, due to irrigation had been recognized for some time, informants
felt that the problem was increasing.

The assessment team saw evidence of salinity in several of the study sites (see
photograph, page 22).

Some of the factors that may contribute to the salinity problem included:
 the increasing use of small one‐piston water pumps, which has forced farmers to
concentrate on land close to the river where irrigation is easier;
 poor design of the irrigation in that there was no process for leaching off the
saline water (e.g. sub‐surface drainage channels) and preventing the build up of
salt in the soil;
 high temperatures and arid climate which cause high evapotranspiration; the use
of loose soil for the construction of the irrigation canals which means that the
water is absorbed, leaving the salt behind (thus aggravating secondary salinity).

The risk of irrigated farming activity being affected by secondary salinisation is

reportedly higher in the Mustahil area than in Kelafo and Gode. In the former, the
concentration of salt in the soil is relatively high and run‐off water from the hillsides
during the rainy season was reported to aggravate the salinity problem.

Water borne diseases – this problem was limited to Mustahil, where the prevalence
of schistosomiasis was reported to have increased over the last two years as
irrigated farming expanded. As mentioned above, good design of the irrigation
scheme can reduce the risk as can extension efforts aimed at raising awareness
about how to prevent children and adults becoming infected.

b. Economic factors

High costs of production – the increasing price of fuel was seen by informants as one
of the main reasons for the relatively low net benefits from irrigated farming.
Between 2008 and 2010, the fuel price increased from 11 ETB to 17 ETB per litre (a
54% increase). The effect of increasing fuel prices was exacerbated as fuel
consumption of the pumps increased as they aged, and also, by the heavy wind
storms that increased the number of irrigations required (especially in hagay).
Although there was no major change in the prices of spare parts for the pumps,
informants complained about the declining quality of the parts available in the local
markets, which meant that they had to be replaced with increasing frequency.

Absence of reliable markets and low market prices – in Gode the ABGs failed to find
a market for their onion and tomato harvests from the first‐round harvest in 2009. In
all the assessment areas, the ABG participants felt that the prices paid for both cash
and food crops in the local markets were too low. They blamed the regional
government for blocking the road from Gode to Jijiga, effectively cutting them off

33
from the Jijiga market. They also noted the almost blanket injection of poorly
targeted food aid into the area, which they said had depressed farm prices.

c. Social factors

Conflict – in the Kelafo area, pastoralists’ dry season grazing land had been
converted to farmlands and here, informants reported tension and conflict between
camel pastoralists and sesame farmers. Specific examples included cases of camels
being burned alive or knifed by crop farmers, and a pastoralist who held a sesame
farmer at gunpoint while his camels fed on the crop38. In Mustahil, where the crop
farmers were agro‐pastoralists with relatively large livestock herds themselves, a
significant part of the pasture land was fenced off.

Pump management – in the ABGs covered by the assessment, there were no major
problems related to the use or management of the water pumps. In Kelafo and
Mustahil, where farming is conducted on an individual basis, each pump user
connected the fuel tube of the pump directly to their fuel container rather than
putting it into the main tank. In this way, each person could only use the pump for
as long they provided fuel. The communities in these areas also used a system where
they allocated the land close to and far from the river equally, thus sharing the extra
cost of fuel incurred in irrigating plots further from the river.

One exception was the Tundhow ABG in Kelafo, which had not used their pumps
since the first cultivation because of a conflict between the participants and the
former ABG committee members, with the latter blamed for misusing part of the
fuel collected from CHF in 2008. Although the ABG participants agreed to privatize
the pumps, they failed to agree on who should be responsible for pump
management – candidates had been identified from four different sub‐clans of the
six sub‐clans that shared the pumps.

3.3 Cost‐benefit analysis of irrigated farming in Gode

A cost‐benefit analysis (CBA) was conducted using information provided by CHF on

project costs, and information collected during the assessment on benefits, being
mainly the income from produce sales. Farm expenses by project participants were
also considered in the analysis. The CBA was restricted to Gode, as ABGs in this area
were able to provide production data estimates, whereas in Kelafo and Mustahil,
people were farming on an individual basis and representative production data was
not available.

The Gode and Kelafo ABGs collected three harvests from the four plantations
conducted between the hagay 2008 season and the jilal 2010 season. Generally, the
first plantation in 2008 was perceived to be the best in terms of yield, followed by
the third plantation in the second half of 2009. In Mustahil, where the third

38
These types of conflicts took place even before L‐SAP and were not necessarily caused by or
exacerbated by the project.

34
plantation was destroyed by floods in early 2010, farmers had collected two harvests
by August 2010.

It was notable that attempts to conduct the CBA was constrained by absent or
incomplete data on the outputs of irrigated farming during the project. The
production data obtained from the CHF Gode office was assumed to be inaccurate,
as the figures were identical across ABGs and such an outcome was viewed as highly
unlikely by the assessment team. The CHF production data also contradicted some of
the information obtained from the ABGs. For example, the CHF data (albeit provided
for a time period after the end of the L‐SAP project) indicated that all of the ABGs
had harvested four times between July 2008 and July 2010 including those of
Mustahil. This contrasted with the three harvests reported by ABGs during the
assessment. Also, it was not possible for informants to estimate the value of the crop
by‐products fed to their own livestock, or the value of farm produce that was
consumed before being harvested; both were potentially substantial benefits. In
Gode, where irrigated farming was conducted as a group venture, the ABGs had not
kept any records.

With the above issues in mind, two cost‐benefit ratios were calculated as follows:

1. A CBA during the CHF project period only, based on 1 cultivation, harvest and
sale of produce:

2. A CBA based on production during and after the project, assuming 3

cultivations, harvests and sale of produce.

As noted above, the CBAs did not include benefits such as the value of crop by‐
products fed to livestock or direct household consumption of produce. These
omissions would lead to an under‐estimation of the cost‐benefit ratio. However, the
calculations also omitted the benefits of pre‐project farming activities, and the need
for a full CBA to consider only the additional benefits provided by the project,
relative to the pre‐project period. This omission would lead to an over‐estimate of
project benefits. Furthermore, the CBA also omitted the value of labour inputs and
the additional time spend on farming activities relative to the pre‐project period (e.g.
see Figure 5).

35
Type of CBA

Table 7
 Amount (US$)

Cost‐benefit analysis of pump‐based

irrigation in Gode
CBA for one plantation (2008)
Costs
Project costs1
Cost of seed provided ‐ total cost plus transport 5918
Cost of pumps ‐ 12 pumps x $1698/ pump 20385
Cost of fuel provided to ABGs ‐ 20 barrels x $213/ barrel 2 4260
Cost of agricultural tools provided to ABGs 4441
Training for group members 333
CHF technical and admin staff costs 2790
Project vehicle rent and fuel costs 3686
CHF overhead costs 1889
Sub‐total project costs for all 6 ABGs 43702
Community costs (first plantation)
Initial contribution by the ABGs (2,500 ETB per ABG) 1364
Transportation of produce to market 109
Sub‐total community costs for 6 ABGs 1473
Total costs ‐ 6 ABGs 45175
Benefits3
Estimated value of produce sales from 1st plantation, Gode ABGs 4 16394
Cost‐benefit ratio 1:0.4

CBA for 3 plantations (2008‐2010)

Costs
Project costs 43702
5
5818
Total
Community costs: reported total costs for 3 plantations 49520
Benefits
Estimated value of produce sales from 3 plantations 5 34636
Cost‐benefit ratio 1:0.7

Notes for Table 7

1
2

3 Costs for 6 Gode ABGs calculated as one third of total costs for 18 ABGs, except for fuel cost.
4 Actual cost estimated from number of barrels of fuel delivered to Gode ABGs.
5 Values estimated from FGDs and L‐SAP project documents; note 2 ABGs had nil production.
The sales value is equivalent to 82 US$ (902 ETB) per hh assuming 4 ABGs of 50 hh each.
ABGs.
Costs and benefits calculated from Barsan ABG which reported the highest income of the Gode

36

4.

4.1
 DISCUSSION

The importance of pre‐project analysis

4.1.1 Participatory approaches

The physical geography and clan composition of Gode zone means that livelihoods in
the zone have been strongly influenced by trends in neighbouring Somalia, including
trends in the development of irrigation systems on the Somalia side of the border. In
both Somalia and Ethiopia there is a long history of irrigation, dating back to the
1960s, and locally, an extensive period of experimentation and adaptation of
irrigation approaches independently of government or aid projects. There was
already considerable local knowledge on what works and why before the CHF
project, with the development of the dhulqeyb and birgerac systems to suit local
social, environmental and market contexts and opportunities.

The L‐SAP project was funded by OFDA, and best practice for the design of
humanitarian projects includes participatory assessment; the first Common Standard
of the Sphere handbook is ‘participation’39. This standard captures not only the right
of local people to contribute to the design of projects from which they are supposed
to benefit, but also, recognizes that local knowledge is a key intellectual and practical
resource for project design. The L‐SAP project clearly included much local
involvement in project implementation, but what was less clear was the extent to
which the group‐based approach to irrigation and the specific material inputs drew
on local knowledge or an understanding of pre‐existing irrigation systems and
histories. The group‐based ABG approach seemed not to take account of the strong
tendency for well‐established irrigation systems in the areas to be organized on an
individual basis, with pumps hired from private owners. This is an example of a pre‐
existing service which livelihoods‐based approaches should aim to understand and
improve40 or strengthen, rather than replace with an alternative approach. Similarly,
there were pre‐existing, clear and equitable systems for allocating land for irrigation.
For example, the land closest to and furthest from the river was traditionally shared
equally between the community members involved in farming, in order to share the
higher benefit from irrigation close to the river and the higher costs of irrigating plots
further away. In communities such as Kelafo, the land allocated to the ABG was later
shared out in this way. This suggests that in such communities – and contrary to L‐
SAP ‐ it would not be appropriate to target selected poor households with small
pumps for irrigation close to the river.

In Kelafo and Mustahil, the rapid reversion to pre‐project irrigation systems

indicated that communities viewed the ABG approach as unnecessary, but a
relatively acceptable way to acquire the free material inputs of L‐SAP such as pumps,

3
9 The Sphere Project (2005), Humanitarian Charter and Minimum Standards in Disaster Response.
http://www.sphereproject.org
40

and exclude poorer farmers; further research would be needed into exactly how these systems work
and how they could be expanded to benefit more farmers, including poorer farmers.

37
There are concerns that the private systems are disproportionately beneficial to the pump owners
fuel, seeds and tools. In these areas, these project resources were simply absorbed
into the existing systems, usually with private operators taking over control of the
pumps. The planned introduction of a group‐based savings and credit scheme
potentially could strengthen existing systems by providing a source of credit for the
individual purchase of farm inputs. However, although planned, this did not happen
under L‐SAP and could not be considered during the assessment.

These experiences illustrate the importance of understanding and working with

private sector stakeholders from the onset. In terms of scaling‐up the CHF approach,
it might be argued that CHF should base further support on an analysis of constraints
and opportunities with private pump owners. As these stakeholders can purchase
pumps at around the half the cost of CHF (see footnote 16, page 14), OFDA and CHF
will need to consider the economic logic of delivering more pumps free‐of‐charge.

4.1.2 Pre‐project lessons from the literature

The participatory analysis outlined can be strengthened if experiences from previous

small‐scale irrigation are considered. Indeed, many of the issues described in section
3.2.9 are also reflected in the literature on small‐scale irrigation in Ethiopia. Some of
the key lessons from the literature are summarised in Table 8 in relation to the
critical engineering, socio‐economic, environmental and institutional aspects of the
schemes which can lead to their under‐performance. In analysing the opportunities
for expanding small‐scale irrigation in Ethiopia, IWMI (2005) summarise the
difficulties as follows: ‘… even in countries where water resources potential is
relatively well known and known to be substantial, other conditions may not be
conducive for sustainable irrigation development to achieve food security, improve
livelihoods and reduce poverty. Such conditions may vary from attributes such as
topography, soils conditions and rainfall characteristics, to technical and
socioeconomic issues such as lack of physical infrastructure, access to innovations
and information, markets, credits, extension, and other institutional support services
needed to enhance viable irrigation farming. There is ample evidence that most of
these conditions have not been sufficiently met in the expansion of small‐scale
irrigation, micro irrigation and rainwater harvesting in Ethiopia. Thus, the impacts of
these initiatives in most regions of the country have been limited, especially in
addressing the country’s chronic food insecurity problems’.41

Other reviews by FARM Africa (2006) suggest that in any scheme, the analysis of the
potential benefits and costs of introducing small‐scale irrigation needs to be weighed
against the benefits and costs of alternative means of improving agricultural
production, for example through improved post‐harvest practices, better market
linkages, improved seeds and so on, which may provide wider benefits than a small‐
scale irrigation scheme42. This advice is echoed in the IWMI (2005) report which
concluded: ‘Thus, we hypothesize that efforts to achieve food security and reduce
poverty through irrigation will achieve greater impacts, if complemented by
simultaneous efforts to increase productivity in the rainfed sub‐sector, which
4
1 Awulachew, S.B. et al. (2005), ibid.
4 Carter, R. and Danert, K (2006), ibid.
2
38

Table 8
 Lessons from the literature, Ethiopia

Aspect of the scheme Common factors threatening the success of SSI

Engineering/ physical  Over‐emphasis on physical design and infrastructure, and lack of an
integrated approach to the enhancement of livelihoods through the
scheme (a focus on the ‘hardware’ of irrigation)
 Poor physical design due to lack of data or limited technical experience
 Pump capacity not adequate to cover the planned irrigation area; this
can result in a smaller area being irrigated and/or reduced irrigation
frequency, both leading to lower production (and therefore income)
 Focus on single rather than multiple water use systems where domestic,
sanitation/ hygiene, livestock needs as well as agricultural needs are
considered
Economic  Lack of access to inputs, particularly improved inputs where these are
required for the production of cash crops; lack of a source of income for
the purchase of inputs (seed, fertiliser, pesticides, credit)
 Lack of internal markets for the sale of outputs and low farm‐gate prices;
local markets may become saturated when producers in a scheme grow
the same crops leading to low prices, or even the loss of perishable crops
 Lack of planning for operation and maintenance, including identifying
sources of income for operation and maintenance costs; inflation in the
price of fuel and other inputs
Social and gender  Increased burden of labour because of agricultural intensification not
always costed (particularly impact on women farmers)
 Conflicts between use of land for small scale irrigation and other uses
(e.g. grazing)
 Lack of attention to gender issues, for example women’s access to land,
labour and water, and their participation in water user committees
Environment and
 Poor catchment management/ limited soil conservation measures can
environmental health
result in flash floods, silt‐laden soils and debris
 Increased competition for water resources can create conflict between
upstream and downstream water users,
 Soil degradation through salinisation, waterlogging, erosion and soil
nutrient mining are common problems of irrigation schemes
 Year round impoundment of water can lead to an increased incidence of
Institutional malaria and schistosomiasis (especially below 2000m)
 New community management mechanisms introduced without taking
into account existing structures can lead to overlap, confusion and
exclusion – and may not be sustainable
 Lack of local‐level government bodies for the guidance of irrigation
operation, maintenance and water management at community level,
particularly once schemes are ‘handed over’ to farmers
 Weak linkages to the agricultural extension system for follow up and
support with crop selection, the sourcing of inputs, production
techniques, pest and disease management, and storage and marketing,
Policy particularly where new, high value crops are introduced
 Weak regulatory/ legal framework for the resolution of issues related to
Table adapted from: Carter, R. water
and Danert, K., ibid;
user rights Annen,
(e.g. betweenC. (2004). Irrigation
upstream Planning and
and downstream users,
Implementation: Ethiopia Case between
Studies inpastoralists
IWMI (2004) ‘Investment
and in Agricultural
farmers, between Water
traditional Management
irrigators and
in Sub‐Saharan Africa: Diagnosisusers
of Trends and Opportunities’;
of improved schemes) Awulachew, S.B. et al. (2005), ibid.

39
contributes to the overwhelming proportion of agricultural production and nearly all
the staple grain in the country. In fact, this may have a much more direct impact on
food security than irrigation alone, because most of the food crops are rain‐fed.’ 43

A comparison of the results from the L‐SAP project (section 3) with the issues
detailed in Table 8 is sobering, and reinforces the conclusion that the initial project
design was weak, and did not take account of lessons already learned in Ethiopia on
small‐scale irrigation. Specific project issues are discussed in more detail in the
following section.

4.2 Project monitoring

The lack of verifiable project monitoring data on production, and limited

documentation of the costs and benefits of the ABG schemes under L‐SAP limited
the learning from the project. In this regard, the CHF field offices, particularly in
Mustahil and Kelafo need to be supported to establish better documentation
systems.

4.3 Livelihoods issues

4.3.1 Livelihood impact and implications

The assessment showed that there has been a substantial increase in household on‐
farm production since 2007 across all three areas. Although this led to more
household consumption of farm produce, most of the increase in production was
crops for sale. However, the livelihood impacts of these changes were minimal due
to the corresponding increases in expenditure on farm inputs particularly, fuel and
pump spare parts, and the relatively poor market outlets and low market prices for
the crops produced.

Of the cash received from produce sales, the amount spent on farm inputs
accounted for almost three quarters of income, a significantly higher proportion of
income than in 2007. Similarly, the amount spent on non‐farm uses accounted for
around one quarter of total farm income at the time of the assessment, compared to
around 43% in 2007. Household spending on consumables and health and education
increased in all areas, and this increase was significant for spending on education
and health care in Kelafo and Gode. It was notable that the increase in income has
not translated into a significant increase in household investment in key assets such
as livestock, except in Kelafo.

In terms of the net profit from farming, although the proportional piling exercises
suggested that net profit has doubled since 2007 (Table 5) – with similar scores given
across the three sites – this was not supported by data from FGDs, such as one to
three full or partial crop failures during the previous four production seasons. The
groups also described substantial expenditure on spare parts, especially after the

4
3 Awulachew, S.B. et al. (2005), ibid.

40
first production season, with two of the FGDs stating that they could not afford
repairs to the pumps. One group had purchased another pump when the CHF pumps
broke down, while another group shifted to a pump rental system. The groups in
Gode were selling firewood to raise funds to cover the costs of fuel and spare parts,
and all requested a donkey cart to help with this activity.

Although the cost of spare parts was high, the price of fuel was far more important
in terms of impacting on net profits. Since the project started in 2008, the price of
fuel increased by over 50% and this was cited as one of the major factors affecting
the profitability of farms. For an irrigated production area of 5 ha of maize at 2009
prices in Gode (i.e. fuel cost 11 ETB/litre and maize sold for 250 ETB/quintal), we
estimated that an increase in the price of fuel to 17 ETB led to a 34% reduction in net
profit. In comparison, a 50% increase in the cost of spare parts led to only a 3%
reduction in net profit. Market prices were also a key determinant of profitability.
For example, a 20% increase in the price of maize led to a 36% increase in net profit.
Additional calculations using data provided by local informants indicated that the
production of irrigated maize alone was not profitable or sustainable if current
trends in fuel price continued. For example:
 We estimated a net profit of 13,973 ETB from a 5 ha maize plot, based on
four irrigations per cultivation period, an optimum yield of 20 quintals/ha,
and the sale of all produce (at 2009 prices as above).
 However, at an inflated fuel price of 17 ETB, if five irrigations were needed
and the households consumed 30% of the maize production, then the profit
turned to a loss ‐ even if the maximum maize yield was obtained.

This type of analysis helps to explain why farmers preferred to sell maize for fodder
at an early stage of growth, thus reducing irrigation costs as well as the risk of crop
damage. People also placed a high value on maize by‐products, for example, maize
leaves fed to their own livestock.

Although similar economic analyses suggested that the production of cash crops is
relatively more profitable, the high cost of production and susceptibility to disease
increase the risk of crop failure and reduce profitability . For example, onions had to
be irrigated every three days over a growing period of 140 days, making fuel costs
very high. To counter this, communities allocated plots only for onion production
closer to the river.

4.3.2 Economic issues ‐ markets and trade

The inclusion of high value cash crops seems to be essential to the profitability of
small‐scale irrigated farming. It follows that access to markets and a fair market price
for these crops will be a main determinant of profitability. However, access to the
Jijiga market was limited and so farmers depended on local markets. This meant that
some people (particularly in Gode) struggled to find a market for their crops.
Although the number of traders accessing the Somalia markets had increased

41
following the expansion of Al‐Shebab, re‐opening access to the Jijiga market would
avoid forcing producers to depend on Somalia markets alone.

When researching the potential for small‐scale irrigated agriculture, we found few
studies that used the potential yields for different combinations of irrigated crops
and the potential market for these crops, as the starting point for analysis. The use
of simple models to calculate the potential costs and benefits of irrigated agriculture
under different scenarios relating to crop yields, market prices, pump operating and
maintenance costs etc. is needed during the design of future projects.
The impact of aid food in depressing market prices for farm produce and
discouraging local producers was highlighted by informants. Although beyond the
scope of the assessment, this is another factor that needs to be considered when
assessing the potential market for locally‐produced irrigated crops.

Food aid in Mustahil, cited as a cause of low prices for produce from irrigated L‐SAP plots

Future projects also need to consider ways to support producers to identify markets
and secure production agreements in advance of the growing season. For example,
where local sources are used for purchasing aid food, aid agencies could sign
agreements with producers in advance, based on the biannual assessments
conducted by UN agencies to forecast food needs. In this regard, UN agencies need
to review the current stringent aid food procurement procedures whereby payments
for grain purchased from local farmers e.g. in Gode, need to be made via the
regional government in Jijiga. Also, in terms of crop selection, farmers need to have
better market information to enable them to coordinate what they produce to take
advantage of local markets, rather than producing the same type of crops in a given
season and then over‐supplying the market.

42
4.3.3 Environmental and health impacts and implications

The key environmental issue arising from the assessment was the impact of
increasing soil and irrigation water salinity on reducing crop yields. Salt lined the
irrigation channels in project sites, and informants (particularly in Gode) stressed the
problem of soil salinity during the jilal cropping season. It was unclear whether
design features (such as sub‐surface drainage channels) could be used in future to
reduce the salinity problems.

In part, the increase in the incidence of schistosomiasis in Mustahil also related to

design factors. The risk of schistosomiasis can be reduced both through the design of
irrigation channels (to reduce ponding and allow for periodic drying out of the
channel) as well as through raising awareness around the management of the
scheme. This could be addressed in future through a more multi‐disciplinary,
integrated approach to irrigation development that ensures that farmers involved in
new irrigation schemes are linked to the agricultural and health extension systems in
the area.

A wider issue was the general suitability of the area for small‐scale irrigated
agriculture, in terms of environmental characteristics. Although the L‐SAP project
proposal anticipated at least three irrigated cultivations per year, in reality, farmers
did not produce more two cultivations per year, timed to the traditional rain‐fed
cropping seasons. The arid soils, high temperatures and hagay season dust storms
meant that the costs and risks of crop production outside the traditional rain‐fed
cropping seasons were high. In addition, in parts of Mustahil and Kelafo, farmers
faced regular floods which destroyed crops and profits for that season. Within this
environmental context and unless the costs of production can be substantially
reduced, livelihoods based on irrigated crop farming are high‐risk.

4.3.4 Social impact and implications

The assessment highlighted a number of lessons related to the potential social

impacts of aid‐related small‐scale irrigation schemes. First, there is a clear risk of
conflict in the conversion of former pastureland to farmland when land was formerly
used as dry season grazing for livestock. Unfortunately, there have been a number of
cases of conflict between farmers and pastoralists in the L‐SAP areas in Kelafo, not
necessarily related to the project.

In contrast, in those communities with a history of practicing irrigated agriculture

there are already clear and equitable systems for allocating land for irrigation. For
example, the land closest to and furthest from the river is traditionally shared
equally between the community members involved in farming in order to share the
higher benefit from irrigation close to the river, and the higher costs of irrigating
plots further away. In these communities, the land allocated to the ABG was shared
out in this way. This suggests that in such communities, it will not be appropriate to
target selected poor households with small pumps for irrigation close to the river.

43

4.4
Future considerations

4.4.1 Securing access to land

Along the Wabe Shabelle and Juba rivers in Somalia, the history of irrigation includes
appropriation of productive riverine land either by government or more powerful
clans, leading to the impoverishment of small‐scale farmers or agropastoralists from
minority clans who had hitherto occupied these areas.44 In Ethiopia, the Wabe
Shabelle has seen large‐scale commercial irrigation schemes with limited
sustainability (e.g. see Figure 2). In the absence of formalized land tenure
arrangements which secure the use of riverine land for less influential clans, there is
a risk that such groups will be displaced at some point. Furthermore, as groups such
as the Rer Barre are so dependent on crop farming, if displaced they have few
alternative livelihoods options to turn to.

This kind of analysis indicates that although there may be options for technical
support to strengthen small‐scale irrigation in Gode zone, the key issues may be
around land tenure and the complexities of land rights in this part of Ethiopia.
Further analysis of these issues seems to be fundamental to the future of riverine
farmers in the zone, and could involve linkages with the ongoing Ethiopia
Strengthening Land Administration Program. 45

4.4.2 Irrigation investments are development investments

The assessment showed that the establishment of sustainable small‐scale irrigation

schemes requires longer‐term technical support than can be provided by a one‐year
project. The L‐SAP project was designed as an emergency intervention and phased
out immediately after the distribution of inputs. This did not allow CHF either to
learn lessons about the impact of its work, or to provide the groups with the
necessary support to manage the scheme and deal with the many practical and
technical problems they faced (rising fuel costs, finding markets for produce, analysis
of the costs and benefits of producing different crops, access to credit and extension
services etc). Most of the ABG groups were no longer operating as groups, but
people continued to engage in irrigated agriculture on a private, individual
household basis.

4.4.3 Design issues

In terms of the design and scale of irrigation schemes, available evidence indicates
that the risk of salinity increases with the use of land close to a river, and, it is
difficult to limit or control the number of households participating in the scheme. In
contrast, the use of land far from a river requires more powerful water pumps
and/or a gravity‐based irrigation system, which is too costly for individual, small‐
scale irrigation users.

4
4 Besteman, C., Cassanelli, L.V. (eds.)(2003), The Struggle for Land in Southern Somalia: The War
Behind the War. Haan Publications London and Transaction Publications, New Brunswick.
45

44
http://eltap.net/
Larger scale irrigation projects that compete with pastoralism for dry season grazing
areas close to a river, can easily result in land disputes between pastoralists and crop
farmers, and run the risk of causing irreversible damage to these grazing areas.
However, to avoid this problem a scheme has to be large enough to justify the use of
more powerful pumps and a gravity system that enables land to be irrigated further
inland. Such schemes require huge initial investments, and the extent to which they
are sustainable in the environmental, social and governance context of Somali
Region would appear to be under question, given the history of comparable schemes
in the region.

4.4.4 Recommendations

The options for an organization such as CHF include:

 Shift focus away from supporting small‐scale (i.e. one‐piston pump) group‐
based irrigation schemes targeted at a selected number of poor households,
which then become responsible for management and control of fuel, spare
parts and other inputs.
 Support and encourage the continued expansion of privately‐run schemes
(for example using pumps with a higher capacity that can support irrigation
further from the river) and support the participation of poor households by
helping them to cover the costs of fuel and pump hire, for example through a
voucher system46. This would also have the advantage of reducing the risks
of agency‐provided inputs being misappropriated, as happened in a number
of ABGs under L‐SAP.
 Spread the costs of future schemes by ensuring that schemes allow for
multiple productive water uses e.g. for domestic use, for livestock, for
sanitation and hygiene, as well as for farming – this could potentially spread
the costs of the scheme over multiple uses/ users, as well as creating
benefits such as reduced labour for women and children in fetching water.
 Consider the costs and benefits of promoting other water technologies such
as rainwater harvesting, manual water lifting pumps/ treadle pumps, drip
kits, that can potentially be used not only for irrigation, but also for rainfed
farming and for other purposes47; these would avoid the high costs of
collective schemes and be more accessible to individual farmers if supplied
through the private sector.
 Support to and stimulation of the private and governmental sectors as
appropriate for the supply of and access to the other inputs that are
necessary for the success of irrigated and rain‐fed farming, such as

46
More detailed research would be needed into the current modalities, and the relative distribution
of costs and benefits of the existing pump rental schemes in Gode and elsewhere in the region.
47
Ibid. Some of these technologies (e.g. treadle pumps) have already been introduced in Tigray and
Amhara regions.

45
 agricultural and health extension support, supply chains for farm inputs
(improved seeds, pesticides and fertilisers) and credit, improved post‐harvest
storage, producer agreements and marketing options and so on.

5. Conclusions

As evidenced in this project, and supported by the literature, the value of irrigated
farming to a community is dependent on a number of variables, some of which are
within the control of a project and some of which are not. These need to be
thoroughly considered at the design stage of any irrigation intervention. If an
organisation is to set up a new irrigation scheme, it needs to have technical capacity
to be able to assess the potential costs and benefits from all its various aspects – the
irrigable potential of the land to be developed, the multiple needs for water and
potentially competing needs, the social organisation and gender implications of the
scheme, access to input markets (for fuel and spare parts as well as for seeds,
fertilisers and pesticides) and access to and size of markets for the cash crops to be
produced. It also needs to assess the environmental context and environmental
implications of the scheme. The PIA results show that the L‐SAP project had a
positive impact on household income, but not to the extent anticipated in the
project design. A few of the Gode ABGs are still operating on a collective basis, but
the schemes are performing well below their potential and their economic
sustainability is in question. In Kelafo and Mustahil, the ABGs no longer exist and the
schemes are being operated and managed by the community on a private individual
basis. However, again there are economic and environmental factors threatening the
sustainability of these schemes – namely, the limited access to the Jijiga and
Degahabur markets for produce, and consequent reliance on the Somali market, and
in the case of Mustahil, the regular threat of floods which destroy inputs and crops
and make the schemes economically unviable.

The FARM Africa (2006) paper suggests that each scheme needs to be considered on
a case‐by‐case basis, taking into account the potential benefits to a community
versus issues of cost and sustainability. The results from the L‐SAP PIA would support
the idea that the very different environmental, social, economic and political
contexts operating even within Somali region justifies this kind of case‐by‐case
approach in future project design. The use of a simple modelling tool could help to
estimate the potential costs and benefits, and the sensitivity of the potential
profitability of an irrigation intervention to changes in key variables such as the price
of fuel and the market prices of crops.

The PIA results show that farmers in Kelafo and Mustahil who already had
experience of irrigated farming were immediately able to adopt and adapt the CHF
inputs to their own systems of production. In Gode, where there was less experience
of irrigated farming, the groups probably required more support during and after the
project to be able to gain experience and build their capacity in all of the different
aspects of irrigated production. For the former, the way forward may be to support
the continuing expansion and intensification of privately run schemes, supporting
and strengthening the private sector supply chains for both inputs and outputs, and

46
enabling poorer households to participate by covering their costs. For both, there is
a need to investigate ways of bringing down the operating and maintenance costs
through the introduction of low‐input alternative technologies.

For any scheme, there needs to be much stronger linkages to local‐level support and
guidance services such as the health and agricultural extension services, and credit
supply services that are equally as necessary for the success of small‐scale irrigation
schemes as the physical design of the scheme. Although there has to be a balance
between providing adequate support to farmers and the risk of creating dependency
on external agencies, the results of the PIA suggest that a one‐year timeframe is not
enough to ensure the sustainable success of this type of intervention. The lack of
time for evaluation or impact assessment of the L‐SAP has meant that some of the
same mistakes are likely to have been repeated in the subsequent SHAPE project.

There is relatively little documented evidence of the performance of small scale

irrigation in Somali region. It is important that the lessons learned from CHF’s L‐SAP
and SHAPE projects are widely shared to inform the design of future interventions.

47
 ANNEXES

Annex 1 Inputs provided per ABG (based on CHF documents in Gode field
office)

Seed Maize – 350 kg; Sesame – 80 kg; Tomato (large) – 3.25 kg; Onion – 3
kg; Sweet pepper – 3 kg; Sudan grass – 10 kg; Small cow pea – 10 kg;
Banana ‐ ? not listed

Tools 50 spades, 50 sickles, 2 rolls polybags (small), 2 rolls polybags (large),

small rope (100m), 3 wheelbarrows, 50 local axes, 50 Kelafo hoes, 50
frock hoes, 35 Kawawes
1
Water 2 Anil Indian water pumps – handover document in Gode field office
pumps states ‘2 diesel irrigation generators with centrifugal pump and full
accessories’
2
Fuel Gode ‐ Undated documents in the CHF Gode field office record the
handover of 3 barrels of fuel (600 litres) to the Gode ABGs
Kelafo – Undated documents in the CHF Gode field office record the
handover of varying amounts of fuel from 10 barrels (2,000 litres) to
14 barrels (2,800 litres) of fuel to the Kelafo ABGs
Mustahil – No documents could be found. The ABGs assessed
reported receiving from 5.5 barrels (550 litres) to 7 barrels (1,400
litres) of fuel.

Notes:
1
The Anil pumps were 1‐piston capacity pumps and were selected by CHF because
they are water‐cooled pumps (and therefore suitable for the Gode climate), spare
parts are readily available in Gode, and they are small and can easily be moved if the
river floods.
2
The fuel was not delivered to the ABGs in one batch, but in varying amounts prior
to each irrigation round.

48
Annex 2

Number of irrigations required per crop

Crop Frequency of irrigation No. of irrigations needed per

needed cultivation period

Maize for seed/ Every 21 days for 4 to 5 4 normally, 5 if very dry

consumption months
Maize for fodder Every 21 days for 2 to 3 2 or 3 depending on the market
months demand (there is a good market for
maize leaf from small herds dependent
on buying feed)

Sesame Every 21 days for 2 to 3 4

months
Hot and sweet pepper Every 5 days for 1 month 6

Tomato Every 5 days for 1 month 6

Onion Every 3 days for 6 months 60‐70 (NB higher than in highland
areas)

Banana Every 7 days for 9 months 36

49

Annex 3
ABG

Water pump maintenance costs incurred by ABGs (for 2 pumps)

Pump maintenance cost per cultivation (ETB) Remarks

1st 2nd 3rd 4th
Bakhtiley 870 7100 3400 No costs given for 1st
cultivation for Gode groups;
Ilan 2750 1750 pumps possibly working below
Barsen 6500 7000 capacity
Qonlawe 2500 4000 6000
Agweyn 3000 1800 1000
Qudhaley 3000 6000 5000 6000
Beladamin 2500 6500 2500
Average of total cost of 16,700
Irkabanan 5566 5566 5566 ETB for 3 cultivations
Danyare 2000 3000 2000
Pumps not working after 1st
Kunaso 4000 cultivation
Average of total cost of 3,650
Kobane 1825 1825 ETB for 2nd & 3rd cultivations

Average over 3 cultivations is

Average 3224 3881 3974 4700 3,693 ETB for 2 pumps

50

Annex 4
ABG

Results of ABG crop preference ranking

Maize Sesame Onion Tomato

Barsen 1 2 3 4
Ilan 2 1 4 3
Bakhtiley 3 2 1 4
Agweyn 1 4 2 3
Beledamin 1 4 2 3
Irkabanan 1 2 4 3
Kunaso 1 2 3 4
Qonlawe 1 2 4 3
Kobane 1 2 3 4

51