VULNERABILITY AND DETERMINANTS OF ADAPTIVE STRATEGIES OF PASTORALISTS TO CLIMATE CHANGE: THE CASE OF SHINILE ZONE OF SOMALI REGIONAL STATE, ETHIOPIA

VULNERABILITY AND DETERMINANTS OF ADAPTIVE STRATEGIESOF PASTORALISTS TO CLIMATE CHANGE: THE CASE OF SHINILE

ZONE OF SOMALI REGIONAL STATE, ETHIOPIA

Melese Wondimu and Dr.M.Senapathy

ABSTRACTClimate variability and change which are mainly attributable to recurrent droughtshave been a major challenge in the Ethiopian lowlands. Pastoral communities in Shinilezone have been adapting to changing environmental conditions for centuries throughtraditional social structures and resource management systems. The magnitude andrate of current climate change, combined with additional environmental, social andpolitical issues are making many traditional adaptive strategies ineffective and/orunsustainable, amplifying environmental degradation and food insecurity that areforcing them to rapidly find new livelihood strategies. As a result, this study investigatesthe vulnerability and determinants of adaptive strategies of pastoralists to climatechange in Shinile Zone of Somali Regional State. Primary data were collected from 130randomly selected pastoralists using interview schedule which were supplemented bypersonal observation and focus group discussions. Secondary data from differentsources were also included in the study. Both descriptive statistics and MultinomialLogit model were employed to assess the vulnerability of pastoralists to climate changeand the factors influencing pastoralists’ decision on adaptive strategies to climatechange respectively. This study identified the extent of exposure of the study areas toclimate related hazards and the damaging effects of these hazards on livelihood assetsand activities that are highly sensitive, as the way of life of the communities largelydepends on natural resources. In addition, the study pointed out that the adaptivecapacity of the communities is ineffective and unsustainable due to interrelated man-made and natural factors that increase their vulnerability. The adaptive strategies usedin Multinomial Logit model analysis were abnormal migration, livestock diversification,engaging in alternative income activities and receiving food aid. Pastoralists’ choice ofincreased abnormal migration was influenced by age, sex, total annual income,livestock holdings, extension visit, access to credit, market distance and climateinformation. The choice of livestock diversification was also determined by sex,education, total annual income, livestock holdings, extension visit, access to credit,market distance, and climate information. Similarly, age, education, livestock holdings,total annual income, extension visit, access to credit and climate information influenced

1

the decision to engage in alternative income activities. Government policies, integratedsolutions, and strategic investment plans that support improved access to climateforecasting, research into the development of and information about appropriate locallevel climate adaptation technologies, access to credit, education, health,transportation, employment and market development in pastoral areas are importantto reduce the impact of climate change.

Keywords: Climate variability, Pastoral Communities, AdaptiveStrategies, Resource Management Systems, EnvironmentalDegradation, Livelihood Assets.

Introduction1.1. Background of the Study

Climate change is the dominant challenge ever world faced as itaffects the earth surface widely from tropical to arctic regionsand from sea to land and atmosphere (IPCC, 2007). Ethiopia is themost vulnerable to the impacts of climate variability and changedue to high dependence on natural resources and climate sensitivelivelihoods coupled with the existence of rampant poverty andvariable weather events (NMA, 2007). In Ethiopia, the temperaturehas been increasing annually at the rate of 0.2°C over the pastfive decades. This has already led to a decline in agriculturalproduction, and cereal production is expected to decline stillfurther (by 12%) under moderate global warming (Ringer, 2008).Moreover, it has led to a decline in biodiversity, shortage offood and increases in human and livestock health problems, rural-urban migration and dependency on external support. Factorscompounding the impact of climate change in Ethiopia are rapidpopulation growth, land degradation, widespread poverty,dependency on rained agriculture, lack of awareness by policy anddecision-makers about climate change and lack of appropriatepolicies and legislation (Wondwossen, 2008).

2

If the trend of warming and drying continues in Ethiopia,vulnerability among the agricultural farming and pastorallivelihoods is expected to increase. The key climate-inducedvulnerabilities of households in the lowlands are shortage andvariability of rainfall, decline in crop production, chronicwater shortages, floods, livestock and human diseases, conflictsover pasture and water, and livestock and crop price fluctuations(SC-UK/DPPA, 2008).

Accordingly, in recent times, a significant number of people inEthiopia are being affected chronically by drought and/orflooding, leading to deaths and loss of assets and to an appealfor international support. The problem is very serious in thearid and semi-arid areas, especially among the pastoralists(Daniel, 2008).

Ethiopian pastoralists suffered from a history of repeateddrought for nearly one century which has undermined theirlivelihoods in particular ways. The climatic conditions haveworsened and the productivity of the land is deteriorating,leading to conflict over resources. The ‘scarcity-driven’mobility in search for moist areas still prevails in conditionwhere the land area endowed with such conditions is shrinking dueto climate change and its intensification by local practices. Thecompetition for access to shrinking resources has resulted inconflicts (Melat, 2008), that in severe cases have turned intoarmed conflicts among the pastoralist population. As pastoralistmobility cuts across national boundaries, the lack of peace andsecurity within and between Ethiopia’s neighboring countries alsoaffects pastoralist modalities of conflict resolution (Melat,2008).

3

1.2. Statement of the Problem

Pastoralists who live in extreme and precarious climatic Zones,in particular, have developed highly adaptive livelihood systemsthat have enabled them to cope with change. Recurrent drought hasbeen a major issue throughout history in the Ethiopian lowlands,and strategies to cope with, and adapt to these droughts areembedded in these communities’ traditional social structures andresource management systems. However, the rate of this changetogether with a lack of support to their livelihood systems isthreatening to overwhelm their adaptive livelihood strategies andability to cope unless concerted action is taken by alldevelopment actors in consultation with communities, who have anacute sense of the threats and potential for their livelihoods(World Bank, 2011).

As the report of Famine Early Warning System Network (FEWS NET)(2009) showed that the semi-arid and water-limited eastern partsof Ethiopia have experienced reduced rainfall since 1997, as wellas poor rains in 2002 and 2004. These regions consider beinghigh-risk areas, since a combination of climate, health, andpopulation factors lead toward increasing vulnerability. In agro-pastoral and pastoral midland and lowland areas, rainfall totalsmay drop below levels sufficient to support livestock.

According to the report of United Nations Office for theCoordination of Humanitarian Affairs -Pastoralist CommunicationInitiative (UNOCHA-PCI) 2005, livelihoods in rural Somali Regionare subject to numerous shocks and stresses, including recurrentdrought, livestock marketing constraints and restricted mobilitydue to conflict and insecurity. Pastoralists and agro-pastoralists face higher risks of premature mortality than urbanhouseholds. Health and education services are nonexistent in mostrural communities, where immunization rates are low and

4

illiteracy is over 90%. Responding to these multiple sources ofvulnerability with food aid is inadequate and inappropriate.Improved governance is needed at all levels, as is betterprovision of basic services, and support to livestock marketing(UNOCHA-PCI, 2005).

Life is more difficult now for most households in Somali Regionthan it was ten years ago. The impact of livelihood shocks onhousehold viability is also clearly visible illustrating thevolatility, unpredictability and hence extreme vulnerability oflivelihoods in semi-arid regions. Pastoralists in Somali NationalRegional State have faced the natural covariant risk of drought;human illness and livestock diseases, which can turn into amultiple risk in case of an epidemic; the economic risk ofexclusion from markets; and the social risk of violent conflictover increasingly scarce resources, which can turn into the riskof civil strife. This latter risk is amplified by the politicalrisk of marginalization and the environmental risk of pasturedegradation (Devereux, 2006).

Despite strong sense of determination, Borana and Somalipastoralists’ ability to adapt is constrained by many factorsincluding increasing land degradation; conflicts over scarceresources; limited access to information (including on weather,climate change, markets, as well as pest and disease outbreaks);a low level of education, skills and access to financial servicesand markets required to diversity their livelihoods; inadequategovernment policies, capacities and coordination; demographicpressures; and social and gender inequalities andmarginalization, which reduce the voice and adaptive capacity ofthe most vulnerable (Béatrice and Anne, 2009).

The combination of climate related and manmade hazards havechallenged the key components of the pastoralist livelihood and

5

these have led to high risk of vulnerability to climate changeand deterioration of pastoral livelihoods. Adaptation responsesto climate change, however, have to be tailored to localconditions and needs, since the nature of risks and the affectedlivelihood groups vary from one ecosystem to another. Moreover,local specific vulnerability and determinants of adaptivestrategies of pastoralists to climate change have not beenstudied and addressed well yet in the study area. A betterunderstanding of the current situations with regard tovulnerability and determinants of adaptive strategies ofpastoralist to climate change will pave the way to reduce thenegative impact of climate change in the pastoralist areas.Therefore, this study can help to fill these research gaps byproviding important and reliable information about the problem.

1.3. Objectives of the Study

The general objective of the study is to assess the vulnerabilityand explore factors affecting the choice of adaptive strategiesof pastoralists to the impact of climate change in the studyareas. The specific objectives are:

to assess the vulnerability of the livelihood of

pastoralists to climate change; and

to identify factors affecting the choices of adaptive

strategies of pastoralists to climate change.

1.4. Significance of the Study

Pastoralists’ risk and vulnerability to climate change aredetermined by the rate and the scale of climate change,conditions of livelihood assets, compatibility of adaptivestrategies used, enabling conditions and constraints. And if

6

their adaptive strategies are already failing to move with thetimes, then climate change is likely to increase that failure,with huge social and environmental consequences. Properunderstanding of these characteristics and conditions isessential to the formulation of policies, designing appropriatestrategies and practical steps to reduce climate change impactsand promote sustainable development in pastoral communities.Therefore, the results of the study can help as source ofinformation to direct various efforts to reduce the severe impactof climate change in the livelihood of pastoralists.

1.5. Operational Definitions

Adaptation: any modification in natural or human systems inresponse to actual or expected climatic stimuli or their effects,which moderates harm or exploits beneficial opportunities.Climate change: refers to any change in climate over time,whether due to natural variability or as a result of humanactivity. Climate variability: refers to variations in the mean state andother statistics (such as standard deviations, statistics ofextremes, etc.) of the climate on all temporal and spatial scalesbeyond that of individual weather events due to natural internalprocesses within the climate system (internal variability), or tovariations in natural or anthropogenic (manmade) external forcing(external variability). Drought: The phenomenon that exists when precipitation issignificantly below normal recorded levels, causing serioushydrological imbalances that often adversely affect landresources and production systems.Hazards: A dangerous phenomenon, substance, human activity orcondition that may cause loss of life, injury or other healthimpacts, property damage, loss of livelihoods and services,social and economic disruption, or environmental damage.

7

Kebele: Small locality or village or the smallest administrativeunit that includes many households. Pastoralists: the nomads who are mainly livestock producers tosale or exchange of animals and their products to obtain living. Vulnerability: is the degree to which a system is susceptible toand unable to cope up with, adverse effects of climate change,including climate variability and extremes.Woreda: District or the administrative unit under Zone whichincludes many Kebeles.Zone: the administrative unit that includes many Woredas.

1.6. Description of the Study Area

1.6.1. Location

Shinile is one of the nine Zones of Somali region that is locatedat the north-western point of the Somali Region. The Zone borderswith Djibouti in the North, Somalia (Somaliland) in theNortheast, Oromia Regional State in the South and Jijiga Zone andDire Dawa Council in the southeast and Afar Regional State in theWest. Shinile Zone is divided into six Woredas namely, Mieso,Afdem, Erer, Shinile, Aysha and Dembel.

1.6.2. Climate and topography

This region experiences a relatively similar ecological andclimatic system stretches from North Eastern Kenya to SoutheastEthiopia and Northwest Somalia. The Shinile Zone receives anaverage of 500 to 700 mm of rain annually and the annual meantemperature ranges between 22.5 and 32.50C, depending on thelocation within the Zone. There are two rainy seasons, namely theDiraa’ or Gu (short rains) from mid-March to mid-May and the Karan(long rains) from mid-July to mid-October. The dry seasons are

8

Hagaa, from mid-May to mid-July, and Jilaal, from mid-October tomid-March. There is an also short rain (usually 2-5 rainy days)that is locally called hais. It occurs between December andJanuary, but unreliable. The altitude in the Shinile Zone rangesfrom 950 to 1350 meters above sea level (SCUK and Care, 2009).

1.6.3. Livelihoods profileAccording to the CSA (2008), Shinile Zone has an estimatedpopulation of 456,434, of which 245,928 are males and 210,506 arefemales. About 14% (64,122) of the population is classified asurban while the rest 86% (392,312) is rural residents. Theinhabitants of Shinile Zone are Somali peoples (Issa, Gurgura,Gadabursi and Hawiye tribes). The Issa are mainly pastoralistswhile the other groups are mainly agro-pastoralists. A smallgroup of each of these groups is involved in commercialactivities in urban and market centre. Islam is the dominantreligion and most people speak Somali language (Devereux, 2006).

N

Figure 1. Map of the livelihood Zones in Shinile AdministrativeZone

9

Source: SCUK and DPPA (2008) RF 1:250,000

In Shinile administrative Zone there are two major LivelihoodZones (LZs). The largest is the pastoral LZ, which occupies mostof the central and northern parts of the Zone. Pastoralists makeup about 75‐85% of all the households in Ayshia, Dembel, Shinile,Erer, Afdem, and Meisso Woredas of the Zone. Agro-pastoralistsalso found in the southern foothills of the Zone in mostdistricts (Dembel, Shinile, Erer, and Meisso) and form the mostof the remaining population of about 15‐25%. The urban populationis estimated at less than 5% Pastoralists of Shinile Zone rearthe four major livestock species found in the Horn of Africaincluding sheep, goats, cattle and camels (SCUK and DPPA, 2008).

1.7. Research Design

A cross-sectional study employing both qualitative andquantitative method was used for the study.

1.7.1. Sources and Methods of Data Collection

Both primary and secondary data from different sources werecollected. Primary data were collected mainly from the selected130 sampled pastoralists from six Kebeles selected from Erer andShinile districts of Shinile Zone by using interview schedule,personal observation and two focused group discussions. Inaddition to primary data, relevant secondary data were collectedfrom Shinile Zone Administration, Health Department, SomaliNational Regional State Line Bureaus and NGOs operating in theZone. Published and unpublished documents were extensivelyreviewed to secure pertinent secondary information.

1.7.2. Sample Size and Method of Sampling

10

Three stage sampling procedure were employed in selectingprocess. In the first stage two pastoral Woredas (Erer andShinile) were selected among six Woredas because of theirrepresentativeness and accessibility. In the second stage threepastoral Kebeles per Woreda were selected by simple randomsampling method. The numbers of pastoral Kebeles in Erer andShinile districts are six and ten respectively. At last stageafter having the total number of households in each of thepastoral Kebele, only 130 households were selected based onprobability proportional to size. Accordingly, the selected 130sample households were interviewed by using interview schedule.This study applied a simplified formula to determine the requiredsample size at 95% confidence level, degree of variability=0.5and level of precision= 9% (0.09).

n=

N1+N (e )2

Where n is the sample size, N is the population size (totalhousehold heads size), and e is the level of precision. The aboveformula required a minimum of 120 responses but this study wascarried out on 130 respondents.

1.7.3. Data Analysis

Both descriptive statistics and econometric model were used basedon their importance for analyzing the quantitative data thatcollected from primary and secondary sources in the study areas.

1.7.3.1. Descriptive statistics

Descriptive method of data analysis refers to the use of ratios,percentages, means, variances and standard deviations and it wasemployed in the process of examining and describing the

11

vulnerability and adaptive strategies of pastoralists to climatechange. The vulnerability framework was used as a template toanalyze the vulnerability of livelihood and adaptive strategiesof pastoralists to climate change.

1.7.3.2. Analytical econometric model

This study used a MNL Logit model to analyze the determinants ofpastoralists’ choice of adaptation strategies to climate changein the study area because of the draw backs of MNP. The advantageof the MNL is that it permits the analysis of decisions acrossmore than two categories, allowing the determination of choiceprobabilities for different categories and it is alsocomputationally simple (Wooldridge, 2002).

To describe the MNL model, let y denote a random variable takingon the values {1, 2, . . ., J} for J, a positive integer, and letx denote a set of conditioning variables. In this case, y denotesadaptation options or categories and x contain differenthousehold characteristics, economic variables, and institutionalfactors, social capital and climatic elements.

The question is how cetirus paribus changes in the elements of xaffect the response probabilities (P(y = j/x), j = 1, 2, . . .,J. Since the probabilities must sum to unity, P(y = j/x) isdetermined once we know the probabilities for j = 2, . . ., J.

Let x be a 1 × K vector with first element unity. The MNL modelhas response probabilities:

P (y=j /X )=Exp(xβj )

1+∑h=1

jExp(xβh),j=1,...,J

(1)

12

Where βj is K× 1, j = 1,…, JUnbiased and consistent parameter estimates of the MNL model inEq. (1) require the assumption of Independence of Irrelevantalternatives (IIA) to hold. More specifically, the IIA assumptionrequires that the probability of using a certain adaptationmethod by a given household needs to be independent from theprobability of choosing another adaptation method (that is, Pj/Pk

is independent of the remaining probabilities).

The parameter estimates of the MNL model provide only thedirection of the effect of the independent variables on thedependent variable, but estimates do not represent either theactual magnitude of change in or probabilities.

Differentiating Eq. (1) with respect to the explanatory variablesprovides marginal effects of the explanatory variables given as:

∂pj∂xk

=Pj(βjk−∑j=1

j−1Pjβjk)

(2)

The marginal effects or marginal probabilities are functions ofthe probability itself and measure the expected change inprobability of a particular choice being made with respect to aunit change in an independent variable from the mean. In sum,interpreting a parameter estimate for an explanatory variable interms of its effects on event probability in the multiple outcomemodel demonstrate clearly that a variable has not just one but J-1 independent effects on J-1 event probabilities (Green, 2000).

1.7.3.3. Definition of variables and working hypothesis

13

The choice of adaptive strategies of pastoralist to climatechange can be determined by different variables. The majorvariables that are expected to have influence on the decision ofpastoralists’ to choose adaptive strategies are presented andexplained below.

The dependent variables of the model

For this study, the dependent variables in the empiricalestimation are the choice of an adaptation options that werechosen based on the available studies conducted in other placesand observation of the area that the pastoralists dominantlyused. The choice of adaptation options is assumed to be doneamong completely mutually exclusive alternatives based on theassumption of MNL model. They were represented in the model by 1,2, 3, and 4 for pastoralist households or individuals. These are:

1. Increased abnormal migration to search for water and betterpastures. This is abnormal mobility because people travelfarther and migrate for longer periods of time.

2. Livestock diversification and use of more drought tolerantlivestock species,

3. Engaging in alternative income generating activities to getadditional income,

4. Receiving food aid.

The independent variables of the model

The independent variables that are expected (hypothesized) tohave association with the available choices of adaptivestrategies of pastoralists, were selected based on availableliterature and own experience in the study area. The explanatoryvariables of the study described on table 1.

Results and Discussion

14

This chapter is mainly concerned with the presentation andinterpretations of major findings of the study including theanalysis of climate change vulnerability, differentcharacteristics of the sampled household and factors influencingchoice decision on adaptive strategies of pastoralists to climatechange based on the output of Multinomial Logit model. 1.8. Analysis of Climate Change Vulnerability

Vulnerability is understood in terms of climate change exposure,sensitivity, and adaptive capacity. Understanding thesecomponents of vulnerability relies on a combination ofquantitative and qualitative information which allows for a morenuanced and dynamic understanding of vulnerability (IPCC, 2007).

1.8.1. Exposure of the study areas

Exposure can be interpreted as the direct danger (i.e., thestressor), and the nature and extent of changes to a region’sclimate variables (temperature, precipitation, extreme weatherevents) and the experiences of the areas to climate relatedhazards.

1.8.1.1. Climate variability and change

According to the information from the sampled respondents theareas have been experiencing the climate variability and changesince 5 years a minimum 10, 15, 20 and 25 to the maximum as 3%,34%, 45%, 17%, and 1% of them reported respectively.

The participants of FGDs in both Erer and Shinile Woreda alsoconfirmed that they have been observing a shortening of theGu/Diraa and Karan rainy seasons during the last decade as well asrains tend to start later and end earlier than before. Inaddition, they said that rainfall amount, frequency and

15

distributions seem to have decreased since the last ten years.This can lead to scanty or no pasture growth, increased waterscarcity and depletion of resources. Similarly, temperatures haveincreased significantly since 2006, with the year 2009 beingcited as the hottest year ever, both during day and night, whilethe winter period (Jillal) has also become warmer.

An attempt has been made to decipher the long-term climaticpatterns based on available rainfall and temperature data fromHurso station for Erer district. Accordingly, the average annualmaximum and minimum temperature is 32.3 and 15.8oC at Hursostation. The maximum temperature distribution in the study areais characterized by a general increase (positive slope of thelinear fit line) and certain inter-annual variability (Figure 2).

16

1965 1970 1980 1990 2000 2010

0

200

400

600

800

1000

Y = 4.2148X - 7836.9421R 2 = 0.0702

H urso (1965 - 2007)

Annual Rainfall (mm)

Figure 2: Patterns of annual rainfall and temperaturedistribution in the study areaSource: Data from National Meteorological Agency

17

1981 1982 1984 1986 1988 1990

12

16

20

24

28

32

Y = -0.2270X + 466.2156R 2 = 0.2246

Y = 0.0118X + 8.9351R 2 = 0.0081

H urso (1981 - 1989)

Tempe

rature (

0 C)

Rainfall distribution in the study area is generallycharacterized by high degree of inter-annual variability over thepast years. The average annual rainfall in the area ranges from40 to 947 mm. Analysis of the linear trend of annual rainfallfrom the best linear fit line to the data points (the equationshown) indicates a slight increase. This is indicated by thepositive slope of the linear fit.

1.8.1.2. Weather extremes

A drought is defined as a period of months or sometimes yearswhen an area or a region experiences a deficiency in its watersupply due to consistently below average precipitation. Althougha natural phenomenon, its effects can be worsened byanthropogenic factors. Usually droughts can persist for severalyears although even a short, intense drought can causesignificant damage. Drought is part of the normal cycle of lifein Shinile Zone, where rainfall is low at the best of times andabnormally low every few years.

The two dominant weather extremes that show the climate change inthe study area include drought in both Woredas and flood in ErerWoreda. The respondents reported that they have experiencedfrequent drought especially in the recent decade (2000-2009).The participants of the FGDs also similarly reported that thefrequency of drought has increased from once in seven years toonce in two years.

1.8.1.3. Climate related hazards

When respondents are asked to rank the main (climatic and non-climatic) hazards affecting their livelihoods, all of them rankeddrought as the number one hazard. While extreme heat, bushencroachment, and land degradation were selected as the second

18

most important hazard by 49.23%, 23.08%, and 10.77% of therespondents respectively. Third ranked most important hazards areland degradation and flood especially in Aydora and Sedeti by38.46% and 26.15% respectively (Table 2).

1.8.2. Sensitivity of livelihoods to climate change

Community sensitivity to climate change is the degree to which acommunity is adversely or beneficially affected by climate-related stimuli. It largely depends on the main livelihoodactivities of the community especially its dependence onlivestock production, its key livelihood resources, and theimpacts of climate hazards on these key resources.

However, the predominant belief among the communities in thestudy area is that these changes are due to “God punishment” forthe “sins and transgressions” of humankind, climate change isalready happening with its impacts being felt by most of thepeople in the study area, particularly pastoral communities, whoare more vulnerable to its effect as they rely on livestockproduction that is highly weather sensitive.

1.8.2.1. Impacts of climate related hazards on livelihoodassets

According to the respondents who were asked to rate the level ofthe impact of climate related hazards reported that 79.2 % ofthem said the natural assets are significantly affected. Thephysical assets also significantly impacted as 47.7% of themreported. The financial, human, and social resources also damagedsignificantly as 54.6%, 55.4% and 44.6% of them describedrespectively (Table 3).

19

The participants of FGDs in Shinile and Erer also said that thesignificant negative impacts of drought and extreme heat eventson natural resources such as pasture, water sources, and treesincluded the declining availability, productivity and quality ofpastures and water sources. The grazing potential of the land isgoing down, due to shorter rainy seasons, droughts, intenserainfall events, and overgrazing. The cattle eat grass beforeseeds are ready, and non-palatable plants are replacing palatablegrasses.

In addition the FGDs explained that the biodiversity of the areawas degraded. For example they mentioned different grass and treespecies that were extincted such as grass types: ‘ daramo, hadhigo,baraley, darif, madahbur, harfo and arar’ and tree types: ‘harkun, qoh andashaado’. Deforestation was also an important issue in bothdistricts visited, and is all at once a contributor, anamplifier, and a consequence of changing weather patterns. Inaddition, drought decreases tree growth as well as pod and leafproduction. Also, adaptive strategies undertaken by communitiesin times of drought (such as firewood and charcoal selling) leadto increased deforestation rates.

The disappearance of rangeland species gave room to bushencroachment and invasive species. Invasive species in theirecological functions are seen as invasive when they lead to adecline of biodiversity in an area and in their economic impactswhen they do not contribute to land user’s incomes.

The main financial resources upon which Shinile and Erercommunities dominantly depended include livestock and livestockproducts, savings, credit and cash. Because most of these aredirectly dependent on climate sensitive natural resources, thefinancial resources are strongly impacted by drought and extreme

20

temperature. The decreased pasture availability and increasedwater evaporation lead to water shortages, emaciation and deathof livestock, decreased livestock disease resistance andproductivity (in terms of milk and meat). More livestock beingsold on the market and can lower livestock prices, therebyleading to weaker terms of trade and decreased household incomes.As stated by the participants of FGDs, income generated throughlivestock sales is no longer sufficient, compounding poverty,food insecurity, and pastoralist drop-outs.

According to the participant of FGDs in both Woredas, drought hasalso been affecting human resources including education, health,human labor and various abilities/capabilities. Increasedmigration resulting from droughts leads to increased schooldropouts. In addition, the decrease in food (mainly meat andmilk) in times of drought affects human health (especially amongchildren under 5 years, pregnant women and old people), reduceshuman disease resistance, human labor productivity, and humancapability to undertake different activities such as learning inschools, bush clearing, blacksmithing, and business management.

1.8.2.2. The most affected social groups

Even though the exposures to climate change have been the samefor pastoralists of same area, the impact on livelihood ofdifferent wealth categories and social groups has not the same.According to 40%, 30.77%, 16.15% and 13.08% of the sampledpastoralists the most affected social group was children, women,elders and poor respectively. The participants of the FGDs alsoreported that women, children, elders and generally the poor werethe main social groups affected by climate related hazards asthey have poor adaptive capacity due to various economic and

21

social factors. The vulnerability of women and children to theexisting climate related hazards can be determined by factorsincluding lack of ownership over productive assets, low socialstatus, overburdened in raising and caring for children and theelderly under the condition of drought when men migrate to nearbycities to escape the effect of drought and search for betterfuture. Drought induced migration therefore cause gender and agespecific experiences where women, children and elderly tend tosuffer more as a result.

The participants of the FGDs in Shinile and Erer also explainedthat there are various traditional social protection mechanismsinclude: Zakat - religious alms from wealthierrelatives/clan/sub-clan members usually provided to disadvantagedgroups; Irmaansi - households providing milking animals tohouseholds with no lactating stocks across seasons; Maal -sharing of livestock during milking period; Rai- children frompoor pastoralists herd for richer relatives and receive food andother benefits as payment; Keyd - adopting livestock offspringfrom richer households, sometimes on credit; and Dhowrto - surplusmilk and butter is stored for distribution amongst poorhouseholds with no milking animals, especially during dry season(Jilaal). These traditional systems demonstrate that dealing withhazards is not a new challenge in these communities. However,with droughts becoming more and more severe and frequent, theseredistribution systems are becoming impracticable, as the numberof people needing social support is increasing every year.

The effectiveness and sustainability of traditional resourcemanagement systems were also threatened by increased droughtfrequency, due to abnormal mobility and exacerbated resources.Generally in normal times men were more involved into far-distance movements with their herds, while some family members,especially women and children, are left behind with the more

22

vulnerable parts of the herds, like lactating animals, etc.However, most pastoralists react to the decline of naturalresources with an increased mobility. This is actually the bettercase, since also the income base for the family is better, and ingeneral families who can employ this type of response strategieshave higher livestock numbers and better incomes.

1.8.3. Adaptive Capacity

The adaptive capacity of a community is its ability to adjust toclimate change, to moderate or cope with the impacts, and to takeadvantage of the opportunities. Adaptive capacity is often adetermined by a range of factors, processes and structures suchas income, literacy, institutional capacity, social networks, aswell as access to information, markets, technology, and services.Because the availability of these resources and services islimited in many developing countries their adaptive capacity inthe face of climate change is correspondingly low compared todeveloped countries (IPCC, 2007).

1.8.3.1. Adaptation Strategies

The pastoral communities in Shinile and Erer have already beenundertaking local adaptive strategies to address the impacts ofhistoric and current climate variability and extreme weatherevents as well as other climate-related hazards. According to thesampled pastoralists, the most commonly used strategies were:

Receiving food aid (40 observations) Increased abnormal migration to search for water and

better pastures (32 observations) Engaging in alternative income generating activities, such

as, petty trade and sale of charcoal and firewood (31observations)

Livestock diversification and use of more drought tolerant

23

livestock species (27 observations) The participants of FGDs also identified the various adaptationmechanisms in addition to commonly used. These include:

Conserving water for dry periods Digging deeper wells and boreholes Collecting wild foods Boiling and eating chat, which also decreases appetite Purchasing food and feed Borrowing livestock and money from relatives Feeding animals with tree branches and relief food Removing hides from dead animals for sale and home use Bleeding livestock and drinking blood as a source of food Using water supplied by NGOs and government Using traditional asset redistribution mechanisms Early selling of weak and old animals while livestock

prices are still good… etc

The sustainability of different adaptive strategies also dependson the intensity, duration and frequency of hazards. For example,traditional adaptive strategies such as charcoal and firewoodselling, food rationing, and traditional asset redistributionmechanisms, might be efficient adaptive strategies when there isa major drought every 6-8 years, as there used to be. However, ifthere is a major drought almost every year, charcoal and firewoodselling leads to massive deforestation, making this strategyobsolete in the long run, and leading to intensification ofclimate change impacts; continuous food rationing leads tomalnutrition, decreased disease resistance and humancapabilities, and sometimes even death; and traditional assetredistribution mechanisms become obsolete strategies if there aretoo many losses and too many people in need every year.

1.8.3.2. Constraints of Pastoralists

24

Pastoralists in Shinile and Erer Woreda have faced variousconstraints that can make the adaptation strategies ineffectiveat the local level. The sampled pastoralists reported that theyhave various interrelated constraints that can make their livevery difficult in the presence of frequent drought and climaterelated hazards (Table 4).

The participants of FDGs also identified the underlying driversof vulnerability oftentimes mutually-reinforcing, environmental,social, economic and political issues that have implications fortheir vulnerability and adaptive capacity. These were: scarcityof basic infrastructure, resources and services. Many pastoralcommunities in sampled Kebeles have also very limited access toeducational, agricultural, health, electrical energy and watersupply services as well as access to markets, financial resources(such as credit and savings), reliable information (on weatherprojections, climate change and markets), technology andservices. This is due lack of infrastructures related to roads,schools, health posts and centers, borehole, animal health post,market centers, meteorological centers, financial institutions,agricultural offices, etc.

In addition to the problem of scarcity the communities were alsodissatisfied about the quality of animal and human health servicefor it does not have qualified health personnel and limitedsupply of drugs. Moreover, all sampled Kebeles have noelectricity supply and almost no agricultural extension services.

The accessibility of a given place is dominantly determined byavailability of means of transportation. The sampled kebeles inShinile Woreda have connected by railway with Dire Dawa andDjibouti. However, the railway is under construction torehabilitate and to renew as the system is aged. As a result, at

25

the time being it is not functional. It is an option fordiversifying livelihood strategy such as trading. The communityused to trade commodities from and to Djibouti through therailway system. The remaining option is road that is almostabsent as there is no public transportation and even the roadfrom Shinile to Djibouti. In the case of Erer, the sampledKebeles have no access to neighborhood cities through road. Thishinders access to social services and infrastructure, such ashealth facilities, veterinary centers, schools and markets. Withfew resources at their disposal and limited access to qualitysocial services, pastoral communities tend to be very vulnerablewhen hazards occur. For example, during the onset of drought,pastoralists are often unable to quickly transport livestock tomajor distant markets for sale. They are also unable to accesspasture and water due to limited mobility. When livestock andhuman diseases break out, the appropriate medical attention andmedication are not readily accessible. This results in reducedquality of livestock and increased mortality in which some couldhave been prevented. The problem of natural resource degradation due to increasingdeforestation rates, overgrazing and desertification aggravatedthe pastoralist vulnerability in Shinile and Erer Woredas. Thesearid and semi arid lands have been degraded over several decades.Recurrent drought, increased felling of trees for firewood,charcoal production and livestock feed (especially duringdrought), overgrazing, and inappropriate traditional managementpractices have led to increased soil erosion, lower quality ofpasture and bush encroachment, and reduced carrying capacity forlivestock and humans per unit area.

In addition to lack of alternative employment opportunities,pastoral communities in Erer and Shinile have limited education,and skills to engage in alternative and sustainable off-farm

26

income generating activities. There is also no industry apartfrom the local government branch offices and bureaus that provideemployment option for a very few people. There is limitedopportunity for paid job. The lack of industry not only diddeprive the community from diversifying their livelihoodstrategy, it also contributed to migration of educated andskilled human power among the community to other places in searchfor job.

Social and gender inequalities manifested in the communities asmost resources are exclusively controlled by older men / maleelders. Women and youth rarely have a say in decision-making overresources, which limits their capacity to implement adaptationstrategies. In addition, the only source of cash available towomen in many communities is the cash arising from women groups’income generating activities. This creates poor intra-householdincome distribution, disfavoring women. Also, women’s direct andindirect contribution towards livestock production and householdincomes is seldom recognized.

In addition to dwindling functions of customary institutions,modern governance systems poorly recognized and integrated withthe role of traditional social institutions even if they aresignificant in implementation of adaptation strategies.Furthermore, traditional institutions like Kaalmo, Jiisin, and Zkaatin the Somali culture are being stretched and undermined by morepoverty and drought which are increasing the number of communitymembers requiring assistance. In addition, there seems to be alack of capacity, coordination and information sharing betweendifferent governmental agencies on weather and climate changeissues. As well as they have the problem of lack of access toreliable meteorological and climate change information andknowledge.

27

1.9. Demographic Characteristics of Sampled Households

Among the total sampled household heads 87.69% were male and therest 12.31% were female with regards to sex composition in thestudy area. The empirical result shows that the educationalstatus of pastoralists in the study area was considerably low asthe higher level is primary school(1-4 grade) (6.92%), read andwrite only (16.92%), religious education (30.77%) and 45.38% ofthem were illiterate.

The survey results show that the average family size was 8.12with the standard deviation of 2.40 for sampled households duringthe survey year. The maximum and minimum family size was 15 and 4person, respectively. The sampled household head age ranged from28 to 80 years with the mean age of 52.75 years. The householdheads estimated number of relatives they have in the area aminimum of 20 and 75 people at a maximum.

1.10. Farm Characteristics of the Sampled Households

The household head pastoralists in the study area have largerange of farming experience from ten to forty five years.Livestock holding size is one of the indicators of wealth statusof the households in the study area. To assess the livestockholding of each household in terms of total livestock unit (TLU),the TLU per household was calculated. According to the surveyresult, the pastoralist households own 35.07 TLU on average. Thelivestock ownership ranges from a minimum of 5.79 to a maximum of63.11 TLU.

The survey results revealed that livestock production is the mainsource of household income in the study area. The average on-farmincomes of the households for the year 2002 E.C were 7212.77 ETB,

28

with standard deviation of 2610.20, the minimum of 2500 and themaximum of 14000. While the non farm income of the respondent’sranges from 0.00 to 7200 Birr with the standard deviation andmean of 1684.85 and 840.77 Birr respectively. The total annualincomes also found between the maximum of 15600 and the minimumof 3000 Birr. The average and standard deviation of the incomeare 8051.23 and 2850.56 Birr. In addition to animal husbandry13.1 of them engaged in fire wood and charcoal selling and10% ofpetty trading especially chat trading while 76.9% of the have noadditional income sources. Absences of alternative income sourcesin the study area increase their vulnerability in the presence offrequent drought and famine.

1.11. Institutional Support

The information obtained and the knowledge and skills gainedthrough extension message and contents accelerates pastoralist’sdecision on adaptive strategies to climate change and landmanagement. However, only 43.85% of the respondents have reportedthat they obtained extension services at least once a year. Asthe study areas are remote and have no access to publictransportation it difficult to extension workers to deliver theservices easily.

Credit is used to improve the ability of households at criticaltimes of the year to buy necessary goods, services, and todiversify livestock production and income sources. Only about49.23 % of them reported that they had received credit in thepast year from their relative and friends. The received creditused to buy food and water (51 household heads), to diversifylivestock (7) and to create income sources like petty trading(6). There was almost no institution that served as the source ofcredit in the study areas. This also aggravates theirvulnerability as their livestock production is highly sensitive

29

to climate change and variability, and damaged by recurrentdrought.

Access to reliable information about seasonal forecast of theweather condition and climate variability is necessary to predictthe coming weather condition and take measures and reduce thedamage. The pastoralists highly relied on the elders of thecommunity to get information about the climate as 66.92% of therespondents reported that they have access to climate relatedinformation. In addition, it impossible to get meteorologicaldata as there is no center and services in the study area.

The nature and development of markets for inputs and out puts canplay a major role in determining patterns of livestock productionand adaptive strategies of pastoralists. The main markets for thepastoral community in Shinile and Erer Woredas are: Erer, Bikie,Adigala, Shinile, Dire Dawa and Djibouti. Accordingly, theaverage, minimum and the maximum market distances the respondentstraveled were 37, 15 and 50 Kilometers respectively. The absenceof near market center and transportation services increased theirvulnerability as their main source of income is selling oflivestock and its products in which its quality and prices areeasily affected.

1.12. Determinants of the Choice of Adaptive Strategies ofPastoralists to Climate Change

1.12.1. Selection of variables for the Multinomial Logit Model

A Multinomial Logit model was fitted to estimate the effects ofthe hypothesized explanatory variables on the probabilities ofusing adaptive strategies. Prior to the estimation of the modelparameters, it is crucial to look into the problem ofmulticollinearity or association among the potential candidate

30

variables. To this end, the variance inflation factor (VIF) wasused to test the degree of multicollinearity among continuous anddummy variables. The values of VIF for variables were found to besmall (i.e. VIF values less than 10). Based on the VIF result, thedata have no serious problem of multicollinearity. As a result,all the explanatory variables were retained and entered into MNLanalysis.

Similarly, the contingency coefficients, which measure theassociation between various dummy variables based on the chi-square, were computed in order to check the degree of associationamong the discrete variables. The values of contingencycoefficient ranges between 0 and 1, in which zero indicates noassociation between the variables and values close to 1 indicatesa high degree of association. Accordingly, there was no seriousproblem of association among discrete explanatory variables asthe contingency coefficients did not exceed 0.75 which is oftentaken as a cut-off point. Hence, all the seven dummy variableswere entered into MNL analysis.

Econometrics results of the Multinomial Logit Model

This section deals with factors which affect the decision ofpastoralists on adaptive strategies to the climate variabilityand change. The dependent variable is choice decision of adaptivestrategies of climate change which was categorized with anexpected value of 1- indicating pastoralists who could increasedabnormal migration to search water and better pastures (N= 32),2- indicating pastoralists who diversified and used droughttolerant livestock (N=27), 3-indicating pastoralists who engagedin alternative income source activities to generate additionalincome (N=31) and 4-indicating pastoralists who relied on foodaids as adaptive method (N=40). As well as receiving food aidstrategy is a base category. The maximum likelihood method of

31

estimation was employed to estimate the parameter estimates ofthe multinomial logit model and statistically significantvariables were identified in order to measure their relativeimportance on the choice decision of the pastoralists on adaptivestrategies. The results of the maximum likelihood estimates aredepicted in the Table 5.

Age of the household head (AGE)

As can be observed from the results, as compared to the basecategory (receiving food aids) an increase age by 1 yeardecreases the probability of engaging in alternative incomeactivities by 0.003 at 5% significant level. And also an increasein age decreases the probability of engaging in increasedabnormal migration at 5% probability level as compared to thebase category. Age was also negatively correlated with the choicedecision of livestock diversification strategy but notsignificantly. This is may be as age increases the ability tocontrol resources, to move and involve in various incomegenerating activities become difficult due physical unfitness andpreferring receiving food aid as viable adaptive strategy becomea common practice in the study areas.

Sex of the household (SEX)

As compared to receiving food aids being male-headed householdincreases the probability of abnormal migration by 0.42 andlivestock diversification by 0.24 at 10% significant level.

Educational level of household head (EDU)

As compared to illiterate being in religious category increasedthe probability of livestock diversification and engaging inalternative income activities than receiving food aid by 0.0 23and 0.13 at 10% and 5% significant level respectively. This may

32

explain that the availability of educational opportunity can helpto decide to diversify livestock and to involve in availablealternative income activities as viable adaptive strategy.

Livestock ownership (LIVOWN)

Ownership of livestock increases the probability of increasedabnormal migration, livestock diversification, and engaging inalternative income activities of adaptive options as compared toreceiving food aid. A unit increases in the TLU would result inincrease in the probability of choosing increased abnormalmigration, livestock diversification and alternative incomeactivities as adaptive options by 0.007, 0.006 and 0.012respectively at 1% significant level. This may explained by thefact that livestock generally considered being an asset thatcould be used either to diversify livestock production or beexchanged for cash to create income sources. Livestock isconsidered as a measure of wealth and increased availability ofcapital which can increase the abnormal migration of thepastoralists to safeguard their asset as they are a base fortheir means of survival with in harsh climatic condition anddegraded natural resources.

Household income (INCM)

Total annual income of the households promotes switching to useincreased abnormal migration, livestock diversification, andengaging in alternative income activities rather than toreceiving food aids as adaptive strategies. As compared toreceiving food aids a 1 Birr increases in total annual incomeincreases the probabilities of abnormal migration, livestockdiversification and alternative income activities at 5%, 5% and1% significant level respectively. This may be explained by thefact that the availability of more income generating sources that

33

can help to diversify livestock and select drought resistancespecies to increase production and increase their adaptivecapacity to climate change and variability.

Annual extension visit (EXTVST)

Results showed that as compared to receiving food aids anincrease in number of extension visit reduces the probability ofabnormal migration, livestock diversification and engaging inalternative income activities by 0.001, 0.034 and 0.041 at 5%,10% and 5% significant level respectively. This may be explainedby the fact that the role of extension agents in theimplementation of Safety net especially distribution of food forwork may be dominant than other professional advisory in thestudy areas.

Credit used (ASCRDT)

As compared to receiving food aids credit used reduces theprobability of abnormal migration, livestock diversification andengaging in alternative income activities by 0.081, 0.009 and0.12 at 10%, 5% and 5% significant level respectively. This ismight be the fact that the majority of households who receivedthe credit used it to buy basic consumption goods rather than todiversify livestock or to create other income sources that havelimited opportunities in the study areas.

Distance from market centers (MARDIS)

As compared to receiving food aids an increase in market distanceby 1km increases the probability of increased abnormal migrationand livestock diversification by 0.009 and 0.006 at 5%significant level. It is also positively related with engaging inalternative income activities but not significantly. This result

34

may be explained that as pastoralist way of production systemmainly depends on the availability of natural resourcesremoteness from market centers that can serve as a source ofsupply of basic goods and demand of their products can evenaggravate their mobility for searching the resources. And themarket centers may not be the main sources of alternative incomeas they lacked basic knowledge and skills to deal with other jobopportunities rather than involving in chat trading and charcoalselling.

Information on climate change (CLMINF)

As compared to receiving food aids getting information aboutseasonal forecasts and climate change increases the probabilityof abnormal migration, livestock diversification and engaging inalternative income activities by 0.134, 0.083 and 0.17 at 1%, 1%and 5% significant levels respectively. Accordingly, the resultcan be explained that as their production system specifically andlivelihood in general determined by the availability of weatherforecasts and climate information they have been relying on thetraditional information sources and prediction methods for theirway of migration and diversification of livestock and incomesources whether it is reliable or not.

CONCLUSIONS AND RECOMMENDATIONS

1.13. Conclusions

The descriptive analysis that dealt with vulnerability analysisshowed that the communities have been observing the climatevariability and change since five years a minimum and 25 years tothe maximum. The meteorological data also confirmed that themaximum temperature distribution has shown a general increase andcertain inter-annual variability. The rainfall distribution alsogenerally showed high degree of inter-annual variability over the

35

past years. The number one identified hazard affecting thelivelihood of the pastoralists was drought. Extreme heat, bushencroachment, and land degradation were also selected as thesecond most important hazards. Third ranked most important hazardwas flood especially in Aydora and Sedeti. The climate relatedhazards have been significantly damaging the livelihood resources(natural, physical, financial, human and social) and livelihoodactivities mainly livestock production.

The significant impacts of drought and extreme heat events onnatural resources were declining availability, productivity andquality of pastures and water sources. The financial assets alsodegraded because most of them are directly dependent on highlydamaged natural resources. The impacts on human resources wereincreased school dropouts due to frequent migration, reducedhuman disease resistance, human labor productivity, and humancapability to undertake different activities like bush clearing,blacksmithing, masonry and business management due to highprevalence of diseases and food shortage. The various traditionalsocial protection mechanisms have affected and becomingimpracticable as the number of people needing social support isincreasing every year with droughts becoming more severe andfrequent. The effectiveness and sustainability of traditionalresource management systems were also threatened by increaseddrought frequency due to abnormal mobility and exacerbatedresources.

The impacts of climate change on livelihood of different wealthcategories and social groups are not the same. As a result women,children, elders and generally the poor were the main socialgroups affected by climate related hazards due to variouseconomic and social factors that increased their vulnerability.Particularly women and children were the main victims due to

36

lack of ownership over productive assets, low social status,overburdened in raising and caring for children and the elderlyunder the condition of drought when men migrate to nearby citiesto escape the effect of drought and search for better future.Drought induced migration therefore cause gender and age specificexperiences where women, children and elderly tend to suffer moreas a result.

Limited availability of base natural resources, income, literacy,institutional capacity, social networks and access toinformation, markets, technology, services as well as efficientprocesses and structures challenged the effectiveness andsustainability of their adaptive capacity. Moreover, mutuallyreinforcing, environmental, social, economic and political issueshave implicated their vulnerability and adaptive capacity.

In addition to that the results of multinomial logit analysisshowed that sex, livestock owned, annual income, market distanceand climate information have a significant positive impact onincreased abnormal migration while age, extension visits andaccess to credit have a significant negative effect. Livestockdiversification is also positively affected by religiouseducation, sex, livestock owned, annual income, market distanceand climate information significantly, but access to credit andextension visits are negatively influenced significantly. Inaddition, livestock owned, annual income, climate information andreligious education positively affect engaging in alternativeincome activities significantly, whereas access to credit, ageand extension visits have a significant negative effect.

The age of the household head increases the probability ofpreferring receiving food aid (the base) to practicing abnormalmigration and engaging in alternative income activities

37

increases. Male-headed households are more likely to adapt byswitching from receiving food aid to increased abnormal migrationand livestock diversification than female-headed households. Ascompared to illiterate being in religious category increased theprobability of livestock diversification and engaging inalternative income activities than receiving food aid. Theownership of livestock tends to be associated with shifting awayfrom receiving food aid to increased abnormal migration,livestock diversification, and engaging in alternative incomeactivities of adaptive options. The total annual income of thehouseholds promotes switching to use increased abnormalmigration, livestock diversification, and engaging in alternativeincome activities rather than receiving food aids as adaptivestrategies. Results show that better visited pastoralists tend torely on food aids than abnormal migration, livestockdiversification and engaging in alternative income activities.Remoteness from the markets tends to favor the higher probabilityof increased abnormal migration, livestock diversification andengaging in alternative income activities than receiving the foodaids. More access to reliable seasonal forecasts and climateinformation tends to increase the likelihood of abnormalmigration, livestock diversification and engaging in alternativeincome activities than receiving food aids.

1.14. Recommendations

The occurrences of drought without giving pastoralists time torecover from previous experience deteriorate their livelihoodresources, adaptive capacity, and adaptive strategies andincrease their vulnerability that challenges their survival.Generally integrated, concrete, and action orientedinterventions, programs and plans by the stack holders are neededto change their lives.

38

Specifically providing feasible helps by the government and NGOsto increase livestock diversity and adjusting herd compositiontowards fewer grazers (cattle and sheep) and more browsers anddrought-tolerant species (such as camels and goats) throughaffordable credit facilities as a means to implement them arenecessary. Focusing on to improve the quality of the livestockand reducing the number of livestock to simplify managementduring times of drought can reduce their vulnerability.

In addition, as their livelihood mainly depends on the naturalresources capacitating and re-strengthening of traditional rangeland management practices for example scaling up pastureenclosures to feed weak animals, lactating cows and calves duringdry seasons, reforestation and bush clearing/thinningmaintenance, rehabilitation and construction of waterinfrastructure through community participation in theconstruction of various water points like boreholes, ponds,cisterns with support from government agencies and NGOs, canbuild their capacity and bring more effective, efficient andparticipatory management of available natural resources. Although humanitarian food assistance may not be a long-termsolution to the underlining causes of household vulnerability, itseems imperative to continue the relief handout for some time tokeep alive those who have no access either to produce or buyfood. But, the link with the employment generating schemes wouldhelp both in reducing dependency syndrome and contributing tolocal development.

Moreover, the role of improved access to reliable climateinformation and weather forecasting and better productiontechniques in enhancing pastoralists’ awareness are crucial for

39

adaptation decision making and planning. Information pathwaysshould be enhanced, so that local communities can access seasonalweather forecasts and early warnings for climate hazards as earlyas possible and governments can be promptly informed of poor rainconditions and food insecurity issues.

Promoting local level adaptations need to emphasize the criticalrole of providing information by provision of education throughformal, mobile and Kuranic schools in collaboration with donorsand NGOs.

In the collaboration with donors and NGOs provision of communitybased animal and human health service delivery system withsufficient manpower and sustainable drug supply by the governmentis important for improvement of livestock and human productivityinstead of constructing a huge infrastructure which does not gowith the mobility nature of pastoralists.

Establishment of stable livestock marketing activities which areavailable asset based where pastoralists can benefit as sellerand trader in livestock and employment opportunities should getattention and be promoted by the government and NGOs.

40

REFERENCES

Béatrice Riché and Anne Hammill, 2009. Policy Brief Climate-related vulnerability and adaptive-capacity in Ethiopia’s Boranaand Somali communities. Tamara Plush and Demeke Eshetu (eds).Save the Children UK and CARE International. Addis Abeba.

CSA (Central Statistics Authority), 2008. Summary and statisticalreport of the 2007 population and housing census result. FederalDemocratic Republic of Ethiopia Central Statistic Agency. AddisAbaba, Ethiopia.

Daniel, K., 2008. Impacts of climate change on Ethiopia: a reviewof the literature. pp. 9–35. In: Green Forum (ed.), Climatechange – a burning issue for Ethiopia: proceedings of the2ndGreen Forum Conference held in Addis Ababa, 31 October–2November 2007, Green Forum.

Devereux, S., 2006. Vulnerable livelihoods in Somali Region,Ethiopia. IDS Research Report 57. Brighton: Institute ofDevelopment Studies.

FEWS NET (Famine Early Warning System Network). 2009. Recentrainfall and food aid tendencies in Ethiopia:http://www.fews.net/pages/country.aspx?gb=et&l=en, Accessed onJan 30, 2011

Green, W.H., 2000. Econometric Analysis, 4th ed. Prentice-Hall,Upper Saddle River, NJ.

IPCC (Intergovernmental Panel on Climate change), 2007. Climatechange 2007: Impacts, adaptation and vulnerability. Contributionof Working Group II to the Fourth Assessment Report of theIntergovernmental Panel on Climate Change, M.L Parry, O.F.

41

Canziani, J.P. Palutikof, P.J. van der Linden and C.E. Hanson(eds.). Cambridge University Press, Cambridge, UK, 976p.

Melat Gezahegn, 2008. The politics of pastoralist vulnerability:An intersectional perspective. An MSc Thesis Presented to theGraduate School of Development Studies ISS institute of socialstudies. Hague, Netherlands. 76p.

NMA (National Meteorological Agency), 2007. Climate changenational adaptation program of action of Ethiopia (NAPA). Finaldraft report. National Meteorological Agency Addis Ababa,Ethiopia.

Ringer, C., 2008. Food and water under global change: developingadaptive capacity with a focus on rural Africa. Paper presentedin the workshop “How can African agriculture adapt to climatechange? 11–13 December 2008, Nazareth, Ethiopia.

SCUK (Save the Children United Kingdom) and Care International,2009. Climate‐related vulnerability and adaptive‐capacity inEthiopia’s Borana and Somali communities. Final assessmentreport, Addis Ababa, Ethiopia.

SCUK and DPPA (Save the Children United Kingdom and DisasterPrevention and Preparedness Agency), 2008. Livelihoods andVulnerabilities - An understanding of livelihoods in SomaliRegional State, Ethiopia.

UNOCHA- PCI (United Nations Office for the Coordination ofHumanitarian Affairs -Pastoralist Communication Initiative),2005. Vulnerable livelihood in Somali Region, Ethiopia. Issuespaper.

Wondwossen Sintayehu, 2008. Climate change: global and nationalresponse. pp. 37-69. In: Green Forum (ed.). Climate change – aburning issue for Ethiopia: proceedings of the 2nd Green ForumConference held in Addis Ababa, 31 October–2 November 2007, GreenForum.

42

Wooldridge, J.M., 2002. Econometric Analysis of Cross Section andPanel Data. MIT Press, Cambridge, UK.

World Bank, 2011. Costing Adaptation through Local InstitutionsVillage Survey Results: Ethiopia. Washington, DC 20433 U.S.A.www.worldbank.org/sdcc. Accessed on June 9, 2011.

The Editor can include these tables in the paper if he iscomfortable. Otherwise it is not necessary.

43

Table 1 Variables and their measurement included in the models

Explanatory Variables Description

Type of Variables Measurement

AGEAge of the household head Continuous

measured age in year

SEXSex of the household head Dummy

1 if male 0 otherwise

FSIZE Household size Continuousnumber of family members

NRELTVS

The number of relatives of the household in the near by area Continuous

number of relatives

EDUEducational levelof household head

ordinal category

1 for illiterate,2 for read and writes only, 3 for religious education and 4 for primary education (grade1-4).

INCM Household income Continuous

total annual income In Ethiopia Birr (ETB)

LIVOWNLivestock ownership Continuous

Tropical livestock unit (TLU)

EXTVSTAnnual extension visit Continuous

number of visit by agents annually

44

CLMINFInformation on climate change Dummy

1 if yes 0 other wise

ASCRDT Credit used Dummy1 if yes 0 other wise

MARDISDistance from market centers Continuous

distance in kilometers

Table 2. Climatic and non climatic hazards

Hazards Woredas Total

PercentLevel 1 Erer Shinlile Total Percent Drought 58 72 130 100 Level 2 Erer Shinlile Total PercentBush encroachment 12 18 30 23.08Extreme heat 26 38 64 49.23Flood 8 0 8 6.15Human disease 1 1 2 1.54Land degradation 8 6 14 10.77Livestock disease 3 8 11 8.46Pest infestation 0 1 1 0.77Total 58 72 130 100Level 3 Erer Shinile Total PercentBush encroachment 7 25 32 24.62Extreme heat 3 4 7 5.38Flood 34 0 34 26.15Human disease 0 1 1 0.77Land degradation 13 37 50 38.46Livestock disease 1 4 5 3.85

45

Pest infestation 0 1 1 0.77Total 58 72 130 100Source: Own compilation, 2011

Table 3. The Impact level of Climate related hazards on keylivelihood assets

KEY LIVELIHOOD ASSETS

Natural Physical Financial Human

Social

Impact level 2* 3* 1*

2* 3* 1*

2* 3*

2*

3* 1*

2*

3*

Aydora 6 21 012 15 0

10 17

15

12 0

10

17

Berdher 5 18 010 13 0

11 12

15 8 0

11

12

Goad 2 19 011 10 0 4 17 9

12 1

13 7

Millo 8 20 017 11 0

15 13 6

12 0

16

12

Sedeti 6 9 0 7 8 0 8 7 7 8 0 8 7

46

Urdhela 0 16 110 5 1

10 5 6

10 5 8 3

Total 27 103 167 62 1

58 71

58

72 6

66

58

Percent20.8

79.2

0.8

52

47.7

0.8

45

54.6

45

55

4.6

51

45

1*, 2* and 3* are rates of hazards that show low, medium andsignificant impact respectivelySource: Own computation, 2011

Table 4. The Constraints of Pastoralists

Problems

WoredasTotal

Percent

Erer

Shinile

Financial capital 8 9 17 13.08Credit availability 4 5 9 6.92Market access and information 10 11 21 16.15Skills to engage in new income sources 7 8 15 11.54

Inefficient traditional rangeland management 4 7 11 8.46

Seasonal forecasts and climate information 6 8 14 10.77

Limited water sources and reserves 7 13 20 15.38Lack of basic infrastructure 12 11 23 17.69Total 58 72 130 100.00Source: Own survey result (2011)

47

Table 5. Parameter estimates of the Multinomial Logit for climate change adaptationmodel.Variables

Increased abnormal migration Livestock diversification

Engaging in alternative income activities

Coef.

Std.err.

P value

Marginal effect

Coef.

Std.err.

P value

Marginal effect

Coef.

Std.err.

P value

Marginal effect

AGE

-0.212 0.098

0.031

-0.000**

-0.158 0.1

0.115 -0.003

-0.214 0.1

0.032

-0.003**

SEX16.045 8.581

0.062 0.416*

14.593 8.783

0.097 0.236*

-3.864 2.708

0.154 -0.653

FSIZE0.703 0.528

0.183 0.019

0.576 0.539

0.285 0.005

0.461 0.539

0.392 0.023

NRELTV0.062 0.059

0.296 0.003

0.063 0.061

0.301 0.002

0.019 0.057

0.734 0.005

LIVOWN 0.61 0.166 00.007***

0.636 0.167 0

0.006***

0.513 0.162

0.002

0.012***

INCM0.001 0

0.018

0.000**

0.001 0

0.049

0.000**

0.001 0

0.002

0.000***

EXTVST

-2.991 1.291 0.02

-0.001**

-2.412 1.295

0.063

-0.034*

-3.024 1.286

0.019

-0.041**

ASCRDT

-3.584 1.911

0.061

-0.081*

-4.328 1.947

0.026

-0.009**

-4.628 1.907

0.015

-0.116**

48

MARDIS0.187 0.079

0.018

0.009**

0.189 0.081

0.019

0.006**

0.065 0.075

0.382 0.015

CLMINF6.323 2.214

0.004

0.134***

6.142 2.251

0.006

0.083***

4.172 2.087

0.046 0.169**

Read and write 0.76 2.826

0.788 0.026

1.778 2.848

0.532 0.125

0.295 2.859

0.918 0.15

Religious

3.056 2.149

0.155 0.094

3.937 2.172 0.07 0.023*

4.515 2.213

0.041 0.128**

Primary school

-0.131 9.11

0.989 -0.061

2.376 9.098

0.794 0.177

0.766 9.064

0.933 -0.115

Constant

-45.426 . .

-46.652 . .

-15.019 6.91 0.03

*, **, *** Significant at p<0.1, p<0.05 and P< 0.01 level of significance, respectively.Source: Model output (2011)Multinomial Logistic RegressionMaximum Likelihood EstimatesDependent Variable Adaptive strategy Number of observation 130LR chi2 (39) 201.80 Pseudo R2

0.5641Prob > chi2 0.0000 Base category Receivingfood aid strategyLog likelihood -77.97908

49

50