Vol.3, No.6, 871-888 (2012) Agricultural Sciences
Analysis of vulnerability and resilience to climate
change induced shocks in North Shewa, Ethiopia
Gutu Tesso1*, Bezabih Emana2, Mengistu Ketema1
1College of Agriculture, School of Agricultural Economics and Agricultural Business, Haramaya University, Dire Dawa, Ethiopia;
*Corresponding Author: gutessoo@yahoo.com
2HEZBEZ Business and Consultancy PLC, Addis Ababa, Ethiopia
Received 4 August 2012; revised 30 August 2012; accepted 10 September 2012
This article analyzes the vulnerability and resil-
ience levels of farm households in North Shewa,
Ethiopia, using a survey of 452 households.
Agro ecological based classification was done
to analyze vulnerability to climate change in-
duced shocks. Integrated vulnerability analysis
approach was employed to develop indexes for
socioeconomic and biophysical indicators. The
indicators have been classified into adaptive
capacity, exposure and sensitivity to climate
change impact. Then Principal Component Ana-
lysis was used to compute vulnerability index of
each agro ecological zone. The result shows
that farmers living in the highland areas were
very much vulnerable to natural shocks com-
pared to those living in the lowland area. In or-
der to identify and analyse the determinants of
resilience to climate change impacts, ordered
probit model was used. Households were clas-
sified into three categories based on the time
they take to bounce back after natural shocks.
The model outputs indicate that farmers with
better investment on natural resource manage-
ment, access to market, better social network,
access to credit, preparedness, saving liquid
assets, access to irrigation and better level of
education exhibited greater level of resilience
during and af ter climate change induc ed shocks.
Keywords: Climate Change; Vulnerability;
Resilience; Princi pa l Com ponent Analysis; Ord ered
Probit; Ethiopia
Climate Change and its impact on the developed as
well as developing countries are becoming the greatest
worries of life and livelihoods. The impacts of climate
change are heterogeneous across a diverse range of geo-
political scales. For instance, the risk is generally be-
lieved to be more acute in developing countries because
they rely heavily on climate-sensitive sectors, such as
agriculture and fisheries, and have low gross domestic
products, high level of poverty, low level of education,
and limited human, institutional, economic, technical,
and financial capacities as cited in [1-3]. Vulnerability of
countries and societies to the effects of climate change
depends not only on the magnitude of climatic stress but
also on the sensitivity and capacity of affected societies
to adapt to or cope with such stress. Therefore, vulner-
ability is the degree to which a system is susceptible or
unable to cope with the adverse effects of climate change,
including climate variability and extremes. In this regard,
vulnerability is a function of the character, magnitude,
and rate of climate variation to which a system is ex-
posed, its sensitivity, and its adaptive capacity [4].
A number of climate change impact studies have been
conducted in many countries on specific sectors such as
water resources, agriculture, health, coastal zones, and
forestry by using impact models and to a lesser extent
socioeconomic analyses [5-7]. Global recommendation
for Africa calls for an integrated assessment approach for
vulnerability study, at national scale and local level [8].
From the perspective of rural farm households, an analy-
sis of vulnerability to climate change is needed at the
level that would specifically address specific geographic
location so that the smallholders will get a lesson to
tackle climate change problems with the precision that is
necessary [9].
On the other hand, the resilience of households to cli-
mate change impact is another important issue in main-
taining sustainable livelihood. According to DFID, resil-
ience at community level is explained as the ability of
countries, communities and households to manage change,
by maintaining or transforming living standards in the
face of shocks or stresses—such as earthquakes, drought
or violent conflict—without compromising their long-
term prospects [10]. Similarly, resilience is the ability of
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888
a social or ecological system to absorb disturbances
while retaining the same basic structure and ways of
functioning, the capacity for self-organization, and the
capacity to adapt to stress and change. This is a meas-
urement of community’s capacity to absorb external
shocks. In the aftermath of occurrence of climate change
induced shocks, how do farmer bounce back to normal
livelihood is about the resilience level of farming com-
munity. A resilient community is able to respond to
changes or stress in a positive way, and is able to main-
tain its core functions as a community despite those
stresses [11]. An important issue would be what enables
a particular community to easily or hardly bounce back.
It is against this background that this research sets out
to determine quantitatively the magnitudes and patterns
of rural households’ vulnerability to climate change and
then identifies the important determinants for resilience
at household level in North Shewa zone of Ethiopia. The
findings of the research can assist in identifying specific
areas that are most vulnerable to climate change and
guide policymakers and development actors in deter-
mining where investments in reducing vulnerability and
building household’s resilience may be most effective
against adverse effects of climate change.
2.1. The Study Area
The study area is North Shewa Zone of Oromia national
regional state. North Shewa Zone is found in north-west
direction of Addis Ababa. Fiche town which is located at
147 km away from Addis Ababa is the capital of the
zone. The zone has 13 rural districts with a total land
area of 10,323 km2. It is situated between 9˚30"N and
38˚40"E. The zone is boarded by Amahara region in the
north and the east, West Shewa zone in the west and Ad-
dis Ababa in the south. The topography of the area is
mountainous in the highland and midland, while it is
plain in the lowland areas. The altitude of the area ranges
between 1300 - 2500 meters above sea level. It is divided
into three agro-ecologies, namely, 15% highland (>2500
meter above sea level), 40% midland (1500 - 2500 meter
above sea level) and 45% lowland (500 - 1500 meter
above sea level) [12]. The area gets rainfall during both
Belg (February to April) and Meher (June to September)
seasons. The average annual rainfall of the area ranges
from less than 800 mm to 1600 mm while the mean an-
nual temperature varies between 15˚C and 19˚C.
The population of the zone is estimated to be
1,431,305 with population density of 138.7 persons per
km2 and average of 4.6 persons per household. The com-
munity practices mixed farming of cereal crops, pulses
and oil crops. Livestock production also constitutes an
important part of agricultural activities of the zone. The
average land holding is 1.1 hectare per household. Due to
the continuous reduction of farmland to degradation by
frequent flooding and drought, farming intruded into
steep sloping areas, forest lands and expanded to mar-
ginal lands and communal lands covering 81% of the
total area of the zone. Only 3% of the total land is put
under grazing, 3.7% forest land, 11.33% degraded and
bare land and 0.65% is other form of land. The crops,
livestock and other livelihoods of the community are
subjected to damage to climate change induced hazards.
This coupled with the continually decreasing farm size
have serious impact threatening farmers adaptive capac-
ity and livelihood improvements [12].
2.2. Data and Analytical Tools
The data for the research was obtained from a survey of
452 farm households in three districts of the Zone in
2011/2012. The districts include Yaya Gullele, Hidha
Abote and Derra. The specific study sites within the dis-
tricts were selected based on a multi stage random sam-
pling procedure. Consequently, 19 Kebeles were selected
from which the sample households were selected ran-
domly proportional to population size. A structured ques-
tionnaire was used to interview the farmers. Data col-
lected from the farmers include household character-
istics, landholding, crops and livestock production, dis-
aster occurrence, perception level (on precipitation, tem-
perature, soil moisture, air moisture and wind direction),
adaptation strategies pursued, different coping strategies
pursued, level of resilience, and other relevant informa-
In addition, secondary data relevant for this analysis
was obtained from the National Meteorological Service
Agency (NMSA), Central Statistical Authority (CSA),
and Zonal and district agricultural offices. In order to
understand the research questions at community level,
qualitative data were collected through focused group
discussion using checklist prepared for the purpose.
2.3. Method of Anal ysis
2.3.1. Conceptual Framework of Vulnerability
Vulnerability analysis involves various approaches;
the first one is called the socioeconomic vulnerability
assessment approach which focuses on the socioeco-
nomic and political status of individuals or social groups.
Individuals in a community often vary in terms of educa-
tion, gender, wealth, health status, access to credit, ac-
cess to information and technology, formal and informal
(social) capital, political power, and so on, which are
responsible for the variations in vulnerability levels [5,
13]. Consequently, vulnerability is considered to be a
starting point or a state that exists within a system before
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888 873
it encounters a hazard event [14]. In this regard, vulner-
ability is constructed by society as a result of institutional
and economic changes. This explains why the socioeco-
nomic approach focuses on identifying the adaptive ca-
pacity of individuals or communities based on their in-
ternal characteristics. One major limitation of the socio-
economic approach is that it focuses only on variations
within society, but in reality, societies vary not only due
to sociopolitical factors but also due to environmental
factors. It does not also account for the availability of
natural resource bases to potentially counteract the nega-
tive impacts of these environmental shocks. For example,
areas with easily accessible underground water can better
cope with drought by utilizing this resource [5].
The second approach is called the biophysical ap-
proach that attempts to assess the level of damage that a
given environmental stress causes on both social and
biological systems. It is sometimes known as an impact
assessment. The emphasis is on the vulnerability or deg-
radation of biophysical conditions [15]. It is a dominant
approach employed in studies of vulnerability to natural
hazards and climate change [16]. Füssel identified this
approach as a risk-hazard approach. The biophysical
approach, although very informative, has its limitations
[13]. A major limitation is that the assessment of bio-
physical factors is not a sufficient condition for under-
standing the complex dynamics of vulnerability. It also
neglects structural factors and human agency both in
producing vulnerability and in coping or adapting to it.
The approach overemphasizes extreme events while ne-
glecting root causes and everyday social processes that
influence differential vulnerability [15,17].
The third approach is called integrated assessment ap-
proach which combines both socioeconomic and bio-
physical approaches to determine vulnerability. The
IPCC definition—which conceptualizes vulnerability to
climate as a function of adaptive capacity, sensitivity,
and exposure—accommodates the integrated approach to
vulnerability analysis [4,13,18]. According to Füssel and
Klein, the risk-hazard framework (biophysical approach)
corresponds most closely to sensitivity in the IPCC ter-
minology while the adaptive capacity (broader social
development) is largely consistent with the socioeco-
nomic approach [18]. Furthermore, in the IPCC frame-
work, exposure has an external dimension, whereas both
sensitivity and adaptive capacity have an internal dimen-
sion, which is implicitly assumed in the integrated vul-
nerability assessment framework [13].
Although the integrated assessment approach corrects
the weaknesses of the other approaches, it also has some
limitations. The main limitation is that there is no stan-
dard method for combining the biophysical and socio-
economic indicators. This approach uses different data-
sets, ranging from socioeconomic datasets to biophysical
factors. These datasets certainly have different yet un-
known weights [19]. The other weakness of this ap-
proach is that it does not account for the dynamism in
vulnerability. Despite its weaknesses, the approach has
much to offer in terms of policy decisions [5]. Vulner-
ability in this context is a physical risk and a social re-
sponse within a defined geographic territory [19].
In order to solve the challenges of standards for com-
bining the different variables different methods have
been suggested. The first is assuming that all indicators
of vulnerability have equal importance and thus giving
them equal weights [19]. The second method is assigning
different weights to avoid the uncertainty of equal
weighting given the diversity of indicators used. In line
with the second method, many methodological approaches
have been suggested to make up for the weight differ-
ences of indicators. Some of these approaches include
use of expert judgment [5,20,21], principal component
analysis [22,23], correlation with past disaster events
[24], and use of fuzzy logic [25]. Even though there are
attempts in giving weights, their appropriateness is still
dubious; because there is no standard weighting method
against which each method is tested for precision [26].
Annex 1 show different indicators and the scales at
which they could be used and indicators added based on
the context of the study area.
For the analysis of vulnerability in the study area both
physical and social vulnerability perspectives have been
integrated. Fusel and Klein have summarized the frame-
work for vulnerability analysis to include the three com-
ponents; adaptive capacity, sensitivity and exposure. In
the framework, exposure to climate change and variabil-
ity will lead to vulnerability based on the sensitivity level
of the communities’ lives and livelihood. Moreover,
when the capacity to withstand the negative conesquences
of exposure and sensitivity become very low, the vul-
nerability of climate change impact will be very much
higher. In the framework, capacity is generated from the
implementation of adaptation and mitigation intervene-
tions [18].
With this background, the first stage of analysis of
vulnerability in the study area involved descriptive ana-
lysis of the socioeconomic and environmental character-
istics particularly adaptive capacity, sensitivity, and ex-
posure to CC marked by red and green color in the above
framework. Second, the vulnerability indices were ob-
tained by applying Principal Component Analysis on the
adaptive, sensitivity, and exposure variables following
the works of Deressa, Hassan, and Ringler, Fussel and
Ignatius [1,5,13]. Principal component analysis is fre-
quently used in research that constructs indices for which
there are no well-defined weights, such as assetbased
indices used for the measurements of wealth across dif-
ferent social groups. The argument here is that, as with
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888
the asset-based indices for wealth comparison, there are
no well-defined weights assigned to the vulnerability
indices. Therefore, principal component analysis gener-
ated the weights, based on the assumption that there is a
common factor that explains the variance in the vulner-
ability. Instinctively, the first principal component of a
set of variables is the linear index of all the variables that
captures the largest amount of information common to
all the variables. Accordingly, the first component scores
from the principal component analysis measured the
weighted sum of score of all variables.
The model specification is given as
Vulnerabilityadaptive capacity
sensitivity exposure
In this case vulnerability is the difference between
adaptive capacity of a household and its sensitivity and
exposure to climate change induced hazards. When adap-
tive capacity of the area exceeds that of sensitivity and
exposure, the area becomes less vulnerable to climate
change impacts. As explained above, each set (adaptive
capacity, sensitivity and exposure) are composed of dif-
ferent variables. The model specification is as follows:
112 2
jj nnj
nj n jnnnj
 
where V1 is vulnerability index, while Xs, are elements of
adaptive capacity, and Ys are exposure and sensitivity.
The values of X and Y is obtained by normalization us-
ing their mean and standard errors. For instance;
 , where 1
is the mean of x1j across
the different agro ecological zones, 1
is its standard
deviation. In this regard, the first principal component of
a set of variables is the linear index of all the variables
that captures the largest amount of information common
to all the variables. The whole matrix of X1j appears as
1112211 121
12 12
ij ij
mmmn mmmn
 
 
The i and j in the above notation implies the number
of rows (in this case agro ecological zone) and the
number of columns (in this case variables of adaptive
capacity, exposure and sensitivity) respectively. In
Eq.4, the As, are the first component score of each
variable computed using Principal Component Analy-
sis in STATA. Finally, the vulnerability index of each
location is obtained using Eq.4:
111 111
 
 
In calculating the direction of relationship in vul-
nerability indicators (i.e., their sign), a negative value
was assigned to both exposure and sensitivity. The
justification is that areas that are highly exposed to
climate shocks are more sensitive to damage, assume-
ing constant adaptive capacity. The implication is that
a higher net value indicates lesser vulnerability and
vice versa. However, in creating the indices, the scale
of analysis is important. As coated Deressa, Hassan,
and Ringler. [5], vulnerability analysis ranges from
local or household [11] level to the global level [24].
The choice of scale is dictated by the objectives,
methodologies, and data availability. For this study,
the scale of analysis was local level. This is because,
all the earlier studies using aggregated regional and
national levels data has overlooked local variations
which is important for household level analysis.
2.3.2. Determinant s of Resilience
Ordered probit regression model was used to iden-
tify and analyze the determinants of households’ re-
silience to climate change induced shocks. In this
analysis, the level of resilience was classified into
three categories: 1) households that were fast in
bouncing back; which means households that have
gone back to their normal agricultural operation in the
following production season; 2) moderate in bouncing
back; which means households which took one to two
agricultural seasons to get back to normal operation as
before the event; and 3) slow in bouncing back; which
means households which were unable to bounce back
within one to two agricultural seasons to their normal
livelihood activities. In this research, a farmer is said
to have fully bounced back, when it begins its lively-
hood operation as time before the shock. The speed of
bouncing back was measured by number of agricul-
tural seasons taken to bounce back to their livelihood
without external intervention by government or non-
governmental organization. And then comparison was
done based on certain defined characteristics. Thus,
resilience in this measurment involved ordered out-
come. This is with the basic hypothesis that a given
natural shock will have differencial impact on house-
holds’ resilience.
0 if 0YY
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888 875
1 if 01YY
2 if 12YY
Y* is level of resilience and involves ordered outcome,
that is Y = 0 was given to households taking more than
two years to bounce back, Y = 1 was given households
taking greater than one year and less than or equals to
two years; and Y = 2, was given to households taking less
than or equals to one year. The Xij are the explanatory
variables determining the time taken to bounce back. The
independent variables included in the model were avail-
ability of food stock (dummy), income diversification
(number of enterprises), number of plots, number of de-
pendent family members, age of household head (years),
access to credit (dummy), social capital (number of in-
stitutional involvement), area under perennial crops (ha),
preparedness (dummy), propensity to invest on natural
resources (percentage of area under conservation), pro-
pensity to save (percentage of saving), access to irriga-
tion (ha), geographic locations (dummy), etc. βs are pa-
rameters estimated and Uij is the disturbance term.
3.1. Vulnerability Analysis
3.1.1. Vulnerability to Climate Change Induced
In Ethiopia in general and the study area in particular,
small-scale farmers bear largely the brunt of the negative
impacts of climate change, which include increased pov-
erty, water scarcity, and food insecurity. People who are
already poor and marginalized are struggling to cope
with the added burden of increasingly unpredictable wea-
ther, which is triggered by climate change. Families and
communities are getting harder and harder to bounce
back from ever-changing, inconsistent weather affecting
their livelihoods, and many have been forced to sell live-
stock or remove children from school—coping mecha-
nisms that only increase the cycle of vulnerability.
Women headed households, families with high de-
pendency ratio, farmers operating on less fertile and
steeply sloping farms and less diversified enterprises, in
particular, are disproportionately affected by climate
variability. In times of crisis, this categories of commu-
nity tend to move away to look for alternative means of
survival. These households have fewer options to find
other ways of making a living, especially since literacy
rate is very low engaging in alternative coping mecha-
nisms through wage employment. Women are also not
empowered to make household decisions and are fre-
quently without cash savings or assets to sell or to buy
food and other basic items. This vulnerability can be
further classified into social, economic and environ-
mental in the context of agriculture based community. Social Vulnerability
Social vulnerability can be loosely defined as the pre-
disposition of people, organizations, and societies to im-
pacts from natural and man-made disasters. Quantitative
description of the overall social vulnerability of an area
or a region to shocks is measured based on such vari-
ables as proportion of elderly and children, rural housing
density, gender, marital status, age, health status, educa-
tional level of household heads, etc. in the context of
rural household’s social vulnerability to climate change:
it is vulnerability due to the low social profile. Farmers
with high institutional participation, many relatives in a
community, family size with working potential, and par-
ticipation in different social meetings usually have high
social power to withstand adverse effects. Table 1 pre-
sents farmers position in terms of their social status in
the community based on the data from the household
From Table 1 , it is clearly observed that literacy rate
of the community is extreemly low, dependancy ratio of
household members with more than four dependents is
very high, which implies the proportion of dependent
household member with less than 18 and greater than 60
is significant, participation in different institutions is also
low. Thus, it is easier to observe that vulnerability level
of community members to the frequently occuring natu-
ral shocks from their social capital endowment perspec-
tive is high. Economic Vulnerability
The economic vulnerability assessment approach mainly
focuses on the economic status of individuals or social
groups. Individuals in a community often vary in terms
of wealth, health status, access to credit, access to infor-
mation and technology and so on. These variations are
responsible for the variations in vulnerability levels. In
this case, vulnerability is considered to be a starting
point or a state (i.e., a variable describing the internal
state of a system) that exists within a system before it
encounters a hazard event [14,27]. Thus, vulnerability is
considered to be constructed by the society as a result of
economic changes. In general, the economic approach
focuses on identifying the adaptive capacity of individu-
als or communities based on their internal characteristics.
In the study area, climate vulnerability weakens the
different economic capacities. It is historically known for
the socio-economic setbacks and agricultural failures
caused by dry spells and droughts associated with defi-
cits in political-institutional capacities. Economic vari-
ables were selected to be applied for the study area based
on the concept of vulnerability, which is primarly a func-
tion of adaptive capacity. In this context of adaptive
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888
Copyright © 2012 SciRes. OPEN ACCESS
Table 1. Social, economic and environmental vulnerability indicators for the study area.
I. Social Vulnerability Variables Percentage
Sex: Female headed 15.9
Education: illiterate and less than grade 2 86.1
Marital status: Single (including divorce and widow) 14.2
No. of relatives: relative to less than 5 households 38.3
No. institutions: Participation in less than 2.35 institutions 57.1
Dependency: High dependency of 4 person and more 86.3
Farm to farm ext: No access to farmer to farmer extension 31.6
Year Ag. Experience: Lack of farm experience if <3 years 7.3
Access to indigenous early warning information: Having no access 43.8
II. Economic Vulnerability Variables Percentage
Livestock ownership: Own less than 2 tropical livestock unit 35.6
Access to information: Having no access to 73.9
Ownership of perennial crops: no area under perennial crops 87.2
Land size: own less than 0.5 ha of land 36.1
Land fragmentation: own only one plots 74.6
Non-farm income: Have no non-farm income 82.7
Soil and water conservation structures: More than 50% is not conserved 32.3
Income level: Having less than minimum requirement 74.2
Consumption expenditure: Spending less than minimum requirement 62.4
Crop diversity: less than 50% of the 8 major crops grown in the area 70.7
Land under irrigation: no access to irrigation at all 64.2
Land under improved seed: area not covered with improved seed (average of high yielding, drought tolerant, early
maturing) 64
Land under commercial fertilizer: Having no access to fertilizer at all 38.5
Cash reserve: Having no cash saving at all 92
Food reserve: Having no food reserve for next year 71
Credit: Having no access to credit at all 44.5
III. Environmental Vulnerability Va riables (Measures of Sensitivity and Exposure) Amount (%)
Land topography: Slope greater than 15% and 0% slope 49.1
Fertility: Poor fertility and cannot produce without heavy fertilizer use 31.6
Vegetation cover: Bare land 96.3
Frequency of hazards: People facing more than two natural hazards in a year 84.3
Rainfall: Receiving below average 46.2
Temperature: Experiencing above average 95.4
Change in wind direction: Encountering change in wind direction than usual 91.4
ources: Computed from HH survey 2011/12 and Districts report.
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888 877
capacity, farmers’ economic condition palys significant
role in reducing vulnerability. In this connection, the eco-
nomic charcterstics of farmers in the study area shows
that, large majority of the hosueholds operate on less
diversified livelihoods, low non-farm engagement, low
access to credit and market, small landholding, low
holding of perenial crops, small or no area under irriga-
tion, etc. This indicates a high level of economic vulner-
ability of farmers to shocks. Table 1 displays the eco-
nomic characteristics of farmers as related to vulner-
ability to climate change impact.
Clearly the economic status of the households in the
zone to withstand climate change induced hazards can be
judged from the above statistics to be very low. Appar-
ently, large majority of the farmers are economically
vulnerable to the impact of climate change. This can be
evidenced by the fact that 38.5% of the population (or 5
out of 13 rural districts of the zone) are recurrent benefi-
ciaries of safety net program from year to year. Environmental and Physical Vulnerability
There are many environmental challenges that derive
from being a smallholder. The disadvantages include: a
narrow range of resources, which leads to high level of
economic specialization, high population densities, which
can lead to degradation and depletion of limited natural
resources; small watersheds and vulnerable water sup-
plies; and thus easy susceptibility of the farms to cli-
mate change impacts. In this connection, increasingly
different indicators have been used to assess vulnerabil-
ity, both at the national and local scales. At different lev-
els, indicators have been embraced for empirically as-
sessing biophysical vulnerability. These exist on a loca-
tion or geographic specific basis for vulnerability [21,22,
28-30], as well as specifically for climate change [24,
However, the range and extent of indicators varies
from study to study. Complex analyses incorporating
multiple stressors have been carried out at the local level
in various locations [34,35]. The relationship between
natural capital and vulnerability to climate change is ar-
guably one of the least contested. The greater the level of
reliance of a household on natural resources, such as
farming, fishing, or forestry, the greater will be their
vulnerability to climate change. This is because the avai-
lability of such natural resources is dependent on cli-
matic variables such as rainfall, which are projected to
change under climate change. It is likely that the level of
dependence on natural resources will vary from house-
hold to household: for some households farming consti-
tutes the main base of their livelihood; for others it is an
equal or lesser contributor alongside other economic ac-
tivities; and several households do not participate in
farming at all. In this study area, however, almost all
households directly depend on farming activities. Thus
the variable measuring environmental vulnerability con-
siders most vulnerable households (with total depend-
ence on agriculture) as compared to the medium vulner-
ability (partial dependence) and low vulnerability (no
dependence on agriculture).
According to Deressa, Hassan, and Ringler [5], Fussel
[36] and Nhemachena, Benhin, and Glwadys [37], indi-
cators for environmental vulnerability includes but not
limited to slope of the land, soil fertility, rainfall, tem-
perature, frequency of hazards (drought, flooding, forest
fire, disease outbreaks, etc.), vegetation cover, and others.
In the overall vulnerability analysis model, these are
variables for the measurement of sensitivity and expo-
sure. From Table 1, the environmental vulnerability of
the community to climate change can be easily observed.
The undulating and steeply sloping farmlands, low fertil-
ity level due to frequent degradation to soil erosion, ex-
tremely low vegetation cover, frequently occurring cli-
mate change induced shocks (at least 5 in a year), below
average rain and mounting temperature have signifi-
cantly contributing to the vulnerability level of small-
holder farmers.
3.1.2. Measuring Vulnerablity Level by
The above method of measuring vulnerability using
certain social, economic and environmental variables is
usually called the indicator method. The indicator method
of quantifying vulnerability is based on selecting some
indicators from the whole set of potential indicators and
then systematically combining the selected indicators to
indicate the levels of vulnerability as indicated under the
above section; farmer’s vulnerability to climate change.
These levels of vulnerability may be analyzed at local
[11,38,39], national [40], regional [28,41], and global [37]
scales. For the purpose of this study, however, a local
level analysis is proposed based on the recommendations
given by various researchers, who have done at macro
For the analysis of vulnerability condition in the study
area, a household level varaibles were used to make
comparisons between communities residing in different
agro-ecological zones using the principal component
analysis. The variables used in the analysis are listed
under the social, economic and environmental vulner-
abilities in Table 1 above. The variables under I and II in
Table 1 measure adaptive capacity while the variables in
under section III measure the sensitivity and exposure to
climate change impacts. Based on STATA output, the
findings of the study for the agro ecology based classifi-
cation on vulnerability indicators revealed two compo-
nents with eigenvalues greater than 1. These two com-
ponents explain 99 percent of the total variation in the
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888
dataset. The first principal component explained most of
the variation (57.2 percent), and the second principal
component explained 42.8 percent. Based on the level of
variation explained in constructing indices, the first prin-
cipal component was taken, which explained majority of
the variation in the dataset. Table 2 presents the principal
component analysis result for factor scores.
From Table 2, it is observed that the result of the
principal component analysis for factor score was posi-
tively associated with majority of the indicators identi-
fied under adaptive capacity and negatively associated
with all the indicators categorized under exposure and
sensitivity. Therefore, in order to construct vulnerability
indices, indicators of adaptive capacity, which are posi-
tively associated with the first principal component
analysis, and indicators of sensitivity and exposure; which
are negatively associated with the principal component
analysis were taken. In total 22 indices were considered.
Higher values of the vulnerability index show less vul-
nerability and vice versa. This is because, adaptive ca-
pacity is considered as positively contributing to the
reduction of vulnerability, while exposure and sensitiv-
ity are negatively contributing in vulnerability reduc-
Based on the formula under Eq.4, the vulner- ability
index of each agro ecological zone is calculated. In the
calculation of vulnerability index for each agro ecology,
normalized value of each variable (using its mean and
standard deviation) as shown in Annex 2 is used. Ac-
cordingly the vulnerability index of the three agro ecol-
ogy is shown in Figure 1.
The different social, economic and environmental
variables used to generate the 22 indices were tabulated
for each agro-ecology. Annex 2 shows how much values
of each variable deviate from the mean of the total ob-
servation for each agro-ecology. Biggest positive devi-
ance for a given variable indicates that the agro-ecology
has better measurement in that specific factor. On the
other hand, biggest negative deviance implies that the
agro-ecology has lower level of measurement in the spe-
cific factor. For instance, the values for educational level
of household head indicate –0.094 and 0.075 for high-
land and lowland respectively. This implies that the av-
erage education level of farmers in the lowland is 7.5%
higher than the mean of educational level of the total
respondents and while it was 9.4% lower for the high-
The Figure 1 shows the net effect of adaptation, ex-
posure, and sensitivity computed from principal com-
ponent analysis results. It is apparent from the figure that
the net value is only positive for community living in the
lowland areas; while it is negative for those living in
midland and highland agro ecologies. The most vulner-
able agro-ecology is the highland; this is due to small
Table 2. Factor score for the first principal component analysis.
A Social Vulnerability Variables Factor Score
Gender (sex of the household head) 0.9997
Age of the household head above 60
and less than 18 0.8051
Educational level 0.9119
Marital status 0.8974
Number of relatives 0.6145
Institutional participation 0.5787
HH size 0.6673
Farmer to farmer extension 0.8263
Agric. extension 0.9109
Year of agr. experience 0.6531
Indigenous early warning system 0.5761
B Economic Vulnerability Variables
Livestock ownership 0.962
Access to information 0.3305
Ownership of perennial crops 0.6115
Size of land cultivable 0.5004
Number of farm plots 0.9801
Non-farm income 0.9805
level of land conservation 0.2864
Crop diversity 0.6352
Irrigation usage 0.9805
Improved seed usage 0.3983
Commercial fertilizer usage 0.7527
Credit Access 0.9206
C Environmental Vulnerability Va riables
Slope of farmland 0.2386
Fertility of farmland 0.5564
Vegetation cover 0.8129
Natural hazards 0.8459
Rainfall 0.8264
Temperature 0.9578
Wind direction change 0.7006
Sources: STATA output of principal component analysis from data of
2011/2012 household survey.
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888 879
Figure 1. Household’s vulnerability to climate change impacts
in North Shewa. Sources: Own computaiton (lowland 45%,
midland 40% and highland 15%)
land size, highly fragmented farm, low productivity of
land due to fertility lose, high degradation of farmlands
due to steep sloping, lower level of asset building like
livestock and perennial crops, and generally lower level
of experience to adapt to climate change impacts.
In the context of the study are, the midland was less
vulnerable as compared to the highland areas, This less
vulnerability is attributed to lower level of prevalence of
pest and diseases, potential to grow diversity of crops,
relatively gentle sloping of farmlands, moderate rainfall
and low frequency of natural hazards.
Contrary to the expectations the lowland area was not
vulnerable when compared with the midland and high-
land. From the above variables considered in vulnerabil-
ity analysis, the lowland was not vulnerable because of
better experience of operating agricultural activities un-
der stressful conditions, relatively larger farm size with
optimal number of farm plots, moderate slope of farm
lands, better fertility level of farmlands, better size of
land under irrigation, better adaptation to changing cli-
matic conditions and access to early warning informa-
3.2. Household Resilience and Its Indicators
3.2.1. Indicators of Resilience to Climate Change
Due to the frequency of shocks, traditional coping
mechanisms adopted by vulnerable communities are
eroding. During previous drought, floods, disease out-
breaks, landslides and shocks episodes, households have
been able to draw on kinship support network, barter
animals or other assets for food, and/or migrate to areas
with more plentiful natural resources. However, due to a
variety of factors-including continual population growth,
environmental degradations, and the increasing severity
and frequency of climate change induced shocks, com-
munities are less able to provide informal social safety
nets for the neediest households. Similarly, responses
formulated to cope with the periodic occurrence of shocks
have been further challenged by dramatic increase in
food price, and the growing prevalence of conflict on
communal resources along districts and zonal boarders.
As a result, many households affected by the growing
frequencies of shocks are forced to adopt adverse coping
mechanisms such as charcoal production, overgrazing of
reserve and dispute over water and grasses.
Furthermore, combination of severe shortfalls and dis-
ruption of rainfall pattern, depletion of natural resources,
ongoing conflicts and the lack of viable livelihood alter-
natives are increasing which in turn challenge the resil-
ience of vulnerable households. However, some house-
hold still exhibit characteristics of resilience and are al-
ways been able to overcome extreme shocks and sustain
their livelihood and lives. And, what is important under
this section of the study is what characteristics determine
to be resilient to changing climatic conditions.
Resilience is more than an “adaptive capacity”—that
is, society’s capability to draw upon its individual, col-
lective and institutional resources & competencies to
cope with, adapt to & develop from the demands, chal-
lenges and changes encountered before, during and after
a disaster. Much of the literature on resilience from the
perspective of hazards and disasters falls within the do-
main of hazard mitigation planning.
Households that take adverse coping mechanisms of-
ten fail to bounce back after the shock. For instance
households that engage in the sales of liquid assets and
sales of productive assets lacks the capacity to continue
their livelihood operation after disaster shock is over, this
is because they have already deteriorated their opera-
tional capacity. The most damaging form of household
coping strategy (prior to total destitution) is the liquida-
tion of household productive assets such as seeds, tools,
large animals, and land. This category could also include
taking on significant levels of formal or informal debt
from financial institutions or village/neighborhood money
lenders. Such households continue to suffer even after
the stressful seasons are over and consequently fail to be
resilient. Therefore, it is important to consider the level
of such asset maintenance during climate shocks to
measure the level of household’s resilience.
Thus it is important to measure these indicators and
link this information to how resilient a community is
currently. That is which of the personal, community and
institutional factors are strongly affecting resilience in
that community and see resilience as function of the in-
terdependencies between these factors. This also means
that intervention to improve resilience to climate change
must be directed at all factors. They cannot be treated
separately. However, intervention may not be required
for a given predictor if its assessment reveals that it is
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888
Copyright © 2012 SciRes. OPEN ACCESS
present at high levels. Table 3 presents the statistical
measure of the different variables of resilience in the
study area.
From the statistical analysis result, the time taken to
bounce back after climate change induced shocks ranges
from 1 agricultural year to more than 5 years. That is if the
climate change induced shocks seriously affects the crop
and livestock production system during the current year,
only 13.9% can get back to normal operation during the
next production year, where as large majority of the
households as 56.9% needs more than 3 years to bounce
back if no other shock hits them again. The average
number of seasons required for normal resilience is esti-
mated to be 3 years. This is in agreement with the above
analysis of vulnerability, where households residing in the
highland and midland are very much vulnerable to climate
change induced shocks. When disaggregated, household’s
residing in the highland areas take 3.7 years to bounce
back, while those residing in the midland take 3 years to
fully bounce back. The households in the lowland take
less than 1.5 years on average to bounce back.
Even though, the topography of farm lands in the study
area is characterized by steep sloping and ragged terrain,
the investment made by households on conservation of
their natural resources endowment is very low. A large
majority of the households (58%) have worked conser-
vation structure on less than 50% of their farmland and the
annual loss of fertile soil to erosion was high; which in
turn exacerbates the vulnerability of the households and
reduces easy bounce back after natural hazard. In this area,
the dependent family members is high, the family size of
the sample households ranges from 1 to 10; while only
one or two members of household work and cover live-
lihood needs from agricultural activities. In this case,
improving the resilience of a family by being dependent
only on the head of household would be very difficult.
Access to financial services in time of crises is an im-
portant factor to recover from the impact of natural hazard
in the area. Even though 55.5% the households have in-
dicated that they do have access to credit, so far only
41.4% was able to access credit for agricultural operation
to recover from the impact of natural hazards. In spite of
the significant importance of financial and food saving,
the culture of financial saving and keeping food stock
Table 3. Statistical values of factors of resilience to climate change induced shocks.
Variables Mean Maximum Minimum St. Deviation
Time taken to bounce back (Agr. seasons) 3 5 1 1.3898
Diversity of income sources
(Type of crops + type of Livestock + Types non-farm) 7.193 20 1 3.1666
Investment on land conservation
(4 is 100%, 3 is 75%, 2 is 50%, 1 is 25% and 0 is 0%) 1.998 4 0 1.342
Saving (% annual earning) 0.83 0.1 2.39 0.171
Agr. Extension visit (frequency per year) 2.659 7 0 1.6896
Food reserve (% of food total harvest) 0.044 0.1 0.86 0.0709
Preparedness (1 yes, 0 otherwise) 0.946 1 0 0.2244
Educational level of HH head (Year of Schooling) 0.9735 15 0 2.149
Dependency (number of dependents) 3.12 7 1 2.45
Farm size (Ha) 1.13 6.87 0 7.349
Credit access (1 is yes, 0 otherwise) 0.555 1 0 0.497
Distance between plots (hours) 0.59 4 0 0.617
Irrigation (Area irrigated in Ha) 0.053 2.25 0 1.121
Adaptation level (Likert scale ranging between 0 and 1) 0.347 1 0 0.2334
Last year production (1 good, 0 bad) 0.389 1 0 0.488
Asset not liquidated during disaster time (1 yes, 0 otherwise) 0.365 1 0 0.48198
Experience of natural shock (Number/year) 1.694 5 1 0.9625
Years of farming Experience 26.65 80 1 15.4
ource: Own computation from household survey of 2011/2012.
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888 881
over a period is very low. Even during normal years, the
balance between earning and expenditure shows negative
for significant proportion of households (28.9%). That is
large proportion of people usually seek loan from friends
and relatives to sustain their lives until the next production
season. Similarly, 10.8% needs external food support in
addition to own production. These are all tide to prepar-
edness for the coming season’s possible natural shock,
which plays significant role to bounce back. In general,
large majority of the households (86.3%) do not have
preparedness plan either at household or community
level. This makes households to encounter natural shocks
as a surprise. Using chi square test between the prepared
and unprepared there was significant difference in terms
of the time taken to bounce back after a natural shock.
The level of involvement in local institutions in the
area shows that on average, a household participate in
2.35 institutions with a maximum of 6 and minimum of 0.
Participation in existing local institutions and having
relatives in the area were used as a measure of house-
hold’s social capital. Moreover, diversification of income
sources is an important strategy to minimize risk. From
the statistical analysis result, the average number of en-
terprises taken up by farmers is 7.2. Some households
have engaged in the production of only one enterprise
(say production of single crop), while others have en-
gaged in the production of even more than 20 different
crops and livestock enterprises.
3.2.2. Econometric Results: Determinants of
Household Resilience Dependent Variable: Resilience; Time
Taken to Bounce Back to Normal
The frequency distribution of time taken to bounce
back indicates that 57.1% of the respondents were given
value of 0, as it takes them greater than 2 years to bounce
back (years > 2), 29% was given 1, as it takes them more
than one year and less than or equals to 2 years (1 <
Years 2) and 13.9% was given value 2, as it takes them
less than or equals to one year (year 1) to bounce back
to their normal farm oparation as a time before the
shocks. Test of significance using a t-test was done to
make sure that there is a statistically significance diffe-
rence between the categories falling above and below the
cut points for the independent variables. Table 4 presents
the regression coefficient and marginal effect of each
factor on the time taken to bounce back. The Propensity to Invest on Natural
Resources, Maintain Soil Fertility and
Access to Irrigation
Local community have a variety of techniques at their
disposal to enhance the sustainability of the natural re-
sources, which will have significant impact on commu-
nity’s resilience during climate change induced distur-
bances. Some of the practices in the study area include
construction of soil and water conservation structures,
undertaking agro forestry practices, planting of trees
around their farmlands, crop rotation, fallow years and
use of natural fertilizer. In fact, the practice of such
natural resource conservation in turn depends up on
households’ farm size, farm locations, alternative income
and others. The hypothesis here was that households that
have higher proportion of their land conserved as an in-
vestment on their natural resources management will
have better level of resilience. From the t-test result in
terms of resilience level between those having better
investment on their land and those not having, the test
shows significant difference at 1% level of significance.
Similarly, from the econometric result (Table 4), the
regression coefficient for the marginal effect shows 0.062,
which is significant at 1% indicate that increment of
farmland conserved by 25% (out of their total land-
holding), will increase the probability to move to the next
category for bouncing back faster by 6.2%.
In addition to investment made to protect the natural
environment, the natural fertility level of farm plots is
important determinant for speed bounce back and pro-
duce from the land in the following season. Farmers have
already exhausted their farmland and even expanded
agriculture to steep sloping areas, marginal lands and
forest areas. This is due to the continual decrease in the
productivity of their farmland that has come to hardly
sustain household’s food need. Still households with
better fertile soil have better production level and hence
better capacity to bounce back after the natural shock is
over. The above result indicates that the a coefficient of
0.030 for the marginal effect, which implies, households
having better fertile land have a 3% likelihood to bounce
back faster as compared to those with unfertile farmlands.
Moreover, households having area under irrigation ex-
perience better level of resilience. The marginal effect
for area under irrigation is 0.115 indicating a 1 ha in-
crease in area under irrigation would lead to an 11.5%
probability to move from lower category to higher one
for bounce back faster. Prepar edness
In building resilience level of households’, prepared-
ness for the next season’s possible natural shock plays
vital role. Community members that are well prepared
were found to have better level of resilience as compared
to those unprepared. Preparedness in the econometric
model was measured using dummy, where those that
have preparedness were assigned a value of 1 and 0 oth-
erwise. The coefficient of the marginal effect equals
0.196 implies that households with preparedness have
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888
Table 4. Ordered probit model output for time taken to bounce back after natural shock.
Regression Marginal Effect
Va ri abl es
Coefficient St. Error Coefficient St. Error
Propensity to invest on land 0.156*** 0.051 0.062*** 0.020
Propensity to save 0.245 0.463 0.097 0.183
Agr. extension 0.060 0.039 0.024 0.015
Availability of Food reserve 0.087 1.062 0.034 0.419
Preparedness 0.499** 0.274 0.196** 0.104
Educational level HH head 0.073** 0.030 0.029** 0.012
Number of HH’s dependant 0.013 0.055 0.005 0.022
Farm size 0.009 0.010 0.004 0.004
Access to credit 0.422*** 0.131 0.166*** 0.051
Farm plot distance 0.070 0.114 0.028 0.045
Irrigation 0.291*** 0.083 0.115*** 0.033
Saving of productive asset from liquidation 0.451*** 0.136 0.178*** 0.053
Experience of natural shock 0.043 0.069 0.017 0.027
Age HH head 0.003 0.005 0.001 0.002
Agro-ecology: highland 0.054 0.111 0.021 0.044
Midland 0.047 0.012 0.010 0.0078
Lowland 0.0103* 0.166 0.0045 0.007
Sex of HH head 0.060 0.176 0.024 0.070
Social capital: No insti. participated in 0.106* 0.058 0.042* 0.042
Perennial crops ownership 0.118 0.123 0.046 0.049
Access to input/output Market 0.218** 0.094 0.086** 0.037
Diversity income sources (livelihood diversification) 0.035* 0.023 0.014* 0.009
Adaptation level 0.035 0.101 0.014 0.040
No. relatives 0.001 0.001 0.000 0.000
No. farm plots 0.014 0.038 0.005 0.015
Level of soil fertility 0.076*** 0.027 0.030*** 0.011
Log likelihood 345.74
Number of observation 397
LR chi2 (24) 79.54
Prob > chi2 0.000
Pseudo R2 0.4032
*, ** and * indicates significance at 1%, 5% and 10% probability levels respectively.
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888 883
19.6% probability to move to the next category for bounce
back faster as compared to those who do not have pre-
paredness for next year. Educational Level
From range of households’ characteristics, the educa-
tional level of household head was found to be signifi-
cant determinants of resilience to climate change induced
shocks. Heads with higher level of education have better
level of planning, access and understanding of early
warning information, better decision making skills dur-
ing natural shocks, alter agricultural operation, adopt
extension packages and more. Thus education is one of
the key factors in building the resilience level of house-
holds to climate change impacts. The analytical result
shows that this variable is significant at 5% level. An
increase in a year of schooling by one increases the
probability to move to next better category for bouncing
back faster by 2.9%. Access to Credit
One of the most challenging factors in the study area
for smallholder farmers to be resilient to climate change
impact is access to cash needs in times of crises. The
available micro finances institutions in the area are not as
such willing to advance loan during crises. Consequently,
farmers resort to borrowing from local lenders at exorbi-
tantly high interest rates. And cash constraints during
period of natural shocks lead farmers to fall in short of
access to early maturing varieties, drought tolerant varie-
ties and fertilizer. In the model result, access to credit
was significant determinant of resilience at 5% probabil-
ity level. The marginal effect of access to credit shows
that farmers who have access to credit have a 16.6%
probability to move to the next category for bounce back
faster as compared to those who do not have access. Saving of Productive Assets from
Subsequently, and often concurrently with household
short-term strategies, asset divestment (sales) strategies
are employed. Of these, less damaging are divestments
of “liquid” assets such as small animals and household
possessions. Strategies where resources from relatives or
extended family are tapped (e.g., informal loans of food
or money from relatives) also are included in this cate-
gory. The most damaging form of household coping
strategy (prior to total destitution) is the liquidation of
household productive assets such as seeds, tools, large
animals, and land. This category could also include tak-
ing on significant levels of formal or informal debt from
financial institutions or village/neighborhood money-
lenders. Significant percentage of those who sold out
their liquid asset to survive natural shocks has hardly
bounced back. That was because they have already lost
their predictive capacity. And saving of productive asset
during time of shocks was a good determinant of resil-
ience as evidenced by the regression coefficient which
was significant at all conventional probability levels. The
marginal effect of 0.178 indicates that those who have
not liquidated their productive asset has a 17.8% likely-
hood to move to the next better category over those that
have liquidate their assets to bounce back faster after the
climate change induced shock. Social Capital: Involvement in Local
Social networks build a sense of community that con-
tributes to the resilience of individuals and groups. In the
study area types of networks that are important include
families, friends and community organizations. These
groups provide strong bonds within a social group; a
sense of belonging, identity and social support; and
strong linkages to other outside groups that can bring in
additional social, financial or political resources. Suc-
cessful and enduring local institutions create relation-
ships with a common purpose and promote shared inter-
ests, but can also have adaptable and flexible functions.
They can provide emotional and practical support, in-
formation and resource sharing. They stay open, inclu-
sive and diverse, and build community members capital
to mitigate and respond to any natural and manmade
hazards. These local institutions include, Idir, Mahiber,
Iqub, Senebte, Debo, etc. The participation in local insti-
tutions is a strong determinant of household’s resilience
to climate change impact. The marginal effect of 0.042
indicates that involvement in one additional local institu-
tion fosters the likelihood to move to the next category
for bounce back faster by 4.2%. Access to Input/Output Market
Households’ getting easy access to market have a
chance of getting access to input, sale their product, ex-
change information, and diversify their livelihood by
even engaging in small scale trade. The availability of
market in the area benefits households by enabling them
to immediately sale their perishable agricultural com-
modities like vegetables, fruits, and livestock products in
a market to survive from lose that may come due to
change in weather conditions. Moreover, access to mar-
ket or being proximity to market is an important meas-
urement in climate change to bounce back or even to
adapt to the changing condition, presumably because
market serves as a means of exchanging information
with other farmers. In this connection, this study hy-
pothesizes that there is positive relationship between
access to output and input markets and households’ re-
silience to climate change induced shocks. From the
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888
econometric result of Table 4, the regression coefficient
for the marginal effect was 0.086, which implies an
increase in one hour travel away from the market will
decrease the probability to move to lower category for
bounce back by 8.6%. Diversity of Income Sources (Livelihood
The diversity of livelihood sources plays vital role in
that in an event one of the livelihood means is damaged
by climate change induced shocks, households would
survive on the other alternatives. In various climate
change impact literatures, diversifying income sources
stands as the primary measure of household vulnerability
and resilience. The more the household rely on multiple
source of income, the less it is affected by shocks. In this
analysis, livelihood diversity was measured by counting
the different types of crops, livestock and non-farm a
family produce during a year. As a determinant of resil-
ience to climate change, the coefficient marginal effect
for income diversity, 0.014 imply if the household in-
creases its enterprises by 1, the probability to bounce
back faster than normal will increase by 1.4%.
The vulnerability of rural farm households is largely
determined by variety of factors that include social, eco-
nomic, and natural factors. Households living in different
agro ecological location exhibit vulnerability to different
types of hazards. The effect of location in terms of agro
ecology also determines households’ susceptibility to the
risks; where people living in the highland areas are rela-
tively much vulnerable to risks of climate change as
compared to lowlanders, in the context of the study area.
This basically emanates from the topography of farm-
lands, frequency of natural shocks, low experience of
people to adopt to climate change impacts, degradation
of farmlands to erosion and more. Social factors like low
level of literacy or lack of awareness on hazard related
issues have been another exacerbating factor in the dis-
tricts for vulnerability. On the other hand, households
living in the lowland areas were vulnerable to drought,
disease outbreaks and alien weeds. However, when com-
parison is made between three agro-ecological zones in
the study area, lowland was not vulnerable because of
better experience of operating agricultural activities un-
der stressful conditions, relatively bigger size of farm
land with optimal number of farm plots, better access to
credit, moderate slope of farm lands, better fertility level
of farmlands, better adaptation to changing climatic con-
ditions and relatively access to early warning informa-
The resilience levels of farm households living in the
same area differ based on certain socio-economic and
natural factors attributable to lives and livelihood of the
farmers. The capacity to bounce back during and after
climate change induced shocks depends on a number of
households’ characteristics, institutional arrangements,
social networks, economic capacity and natural setting.
Maintaining productive assets from deterioration during
shocks, accessing to irrigation, investing on farmland,
improving the fertility level of farms through usage of
organic processes, having preparedness, diversifying
income sources and participating in local institutions are
some of the households’ action that can build their resil-
ience to climate change impacts. Organizational res-
ponses from government and development actors th-
rough creation of access to market, access to farm loans,
improving educational level, and increased access to
early warning information can be considered as an inter-
vention to build the resilience of community in the study
[1] Ignatius, A. (2012) Spatial vulnerability of rural house-
holds to CC in Nigeria: Implications for internal security.
[2] United Nations Framework Convention on CC (2007) CC:
Impacts, vulnerabilities and adaptation in developing
countries. CC Secretariat, Bonn.
[3] Preston, B.L., Smith, T., Brooke, C., Gorddard, R., Mea-
sham, T. and Withycombe, G. (2008). Mapping CC vul-
nerability in the Sydney Coastal Councils Group. Sydney
Coastal Councils Group by the CSIRO Climate Adapta-
tion Flagship Canberra, Sydney.
[4] Intergovernmental Panel on CC (2001) CC 2001: Impacts,
adaptation and vulnerability. 3rd Assessment Report of the
Intergovernmental Panel on CC, Adaptation to CC in the
Context of Sustainable Development and Equity, Chapter
18, Cambridge University Press, Cambridge, 877-912.
[5] Deressa, T., Hassan, R.M. and Ringler, C. (2008) Meas-
uring Ethiopian farmers’ vulnerability to Climate Change
across regional states. IFPRI discussion paper No. 806,
[6] Pearson, L. and Langridge, J. (2008). CC vulnerability
assessment: Review of agricultural productivity. CSIRO
Climate Adaptation Flagship Working Paper No. 1.
[7] Odjugo, P.A. (2010) General overview of CC impacts in
Nigeria. Journal of Human Ecology, 29, 47-55.
[8] United Nations Framework Convention on CC (2006)
Background paper on impacts, vulnerability and adapta-
tion to CC in Africa. African Workshop on Adaptation
Implementation of Decision 1/CP.10 of the UNFCCC
Convention, Accra, 21-23 September 2006.
[9] Klein, R.J.T. (2004) Approaches, methods and tools for
cc impact, vulnerability and adaptation assessment. In-
Session Workshop on Impacts of, and Vulnerability and
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888 885
Adaptation to CC, 21st Session of the UNFCCC Subsidi-
ary Body for Scientific and Technical Advice, Buenos Ai-
res, 8 December 2004
[10] DFID (2011) Defining disaster resilience. A DFID Ap-
proach Paper.
[11] Adger, W.N. and Kelly, P.M. (1999) Social vulnerability
and the architecture of entitlements. Mitigation and Ad-
aptation Strategies for Global Change, 4, 253-266.
[12] Central Statistical Authority (2007) Population census.
Addis Ababa.
[13] Fussel, H. (2007). Vulnerability: A generally applicable
conceptual framework for CC research. Global Environ-
mental Change, 17, 155-167.
[14] Kelly, P.M. and Adger, W.N. (1999) Theory and practice
in assessing vulnerability to CC and facilitation adapta-
tion. Climatic Change, 47, 925-1352.
[15] Liverman, D.M. (1990) Vulnerability to Global environ-
mental change. In: Kasperson, R.E., et al., Understanding
Global Environmental Change, 27-44, Center for Tech-
nology, Environment, and Development, Clark University,
[16] Hewitt, K. (1995) Excluded perspectives in the social
construction of disaster. International Journal of Mass
Emergencies and Disasters, 13, 317-319.
[17] Pulwarty, R.S. and Riebsame, W.E. (1997) The political
ecology of vulnerability to hurricane related hazards.
Hurricanes: Climate and Socio-Economic Impacts.
[18] Füssel, H.-M. and Klein, R.J.T. (2006) CC vulnerability
assessments: An evolution of conceptual lthinking. Cli-
matic Change, 75, 301-329.
[19] Cutter, S.L., Mitchell, J.T. and Scott, M.S. (2000) Re-
vealing the vulnerability of people and places: A case
study of Georgetown county, South Carolina. Annals of
the Association of American Geographers, 90, 713-737.
[20] Kaly, U. and Pratt, C. (2000) Environmental vulnerability
index: Development and provisional indices and profiles
for Fiji, Samoa, Tuvalu and Vanuatu. Phase II Report for
NZODA. SOPAC Technical Report 306.
[21] Kaly, U., Briguglio, L., McLeod, H., Schmall, S., Pratt, C.
and Pal, R. (1999) Environmental Vulnerability Index
(EVI) to summarize national environmental vulnerability
profiles. SOPAC Technical Report 275, 66 p.
[22] Easter, C. (1999) Small states development: A Common-
wealth Vulnerability Index. The Round Table, 351, 403-
422. doi:10.1080/003585399107947
[23] Cutter, S.L., Boruff, B.J. and Shirley, W.L. (2003) Social
vulnerability to environmental hazards. Social Sciences
Quarterly, 84, 243-261. doi:10.1111/1540-6237.8402002
[24] Brooks, N., Adger, W.N. and Kelly, P.M. (2005) The de-
terminants of vulnerability and adaptive capacity at the
national level and the implications for adaptation. Global
Environmental Change, 15, 151-163.
[25] Eakin, H. and Tapia, B. (2008) Insights into the composi-
tion of household vulnerability from multi criteria deci-
sion analysis. Global Environmental Change, 18, 112-
127. doi:10.1016/j.gloenvcha.2007.09.001
[26] Luers, A.L., Lobell, D.B., Skar, L.S., Addams, C.L. and
Matson, P.A. (2003) A method for quantifying vulnerabil-
ity, applied to the agricultural system of the Yaqui Valley,
Mexico. Global Environmental Change, 13, 255-267.
[27] Allen, K. (2003) Vulnerability reduction and the commu-
nity-based approach: A Philippines study. In: Pelling, M.,
Ed., Natural Disasters and Development in a Globalizing
World, Routledge, New York, 170-184.
[28] Vincent, K. (2004) Creating an index of social vulnerabil-
ity for Africa. Working Paper 56, Tyndall Centre for CC
Research, University of East Anglia, Norwich.
[29] Crowards, T. (1999) An Economic Vulnerability Index for
developing countries, with special reference to the Car-
ibbean: Alternative methodologies and provisional results
(draft). Caribbean Development Bank, Barbados.
[30] Briguglio, L. (1995) Small island developing states and
their economic vulnerabilities. World Development, 23,
1615-1632. doi:10.1016/0305-750X(95)00065-K
[31] Yohe, G. and Tol, R.S.J. (2002) Indicators for social and
economic coping capacity: Moving toward a working de-
finition of adaptive capacity. Global Environmental Change,
12, 25-40. doi:10.1016/S0959-3780(01)00026-7
[32] Downing, T.E., Butterfield, R., Cohen, S., Huq, S., Moss,
R., Rahman, A., Sokona Y. and Stephen, L. (2001) Vul-
nerability indices: CC impacts and adaptation. UNEP
Policy Series, UNEP, Nairobi.
[33] Moss, R., Brenkert, A. and Malone, E. (2001) Vulnerabil-
ity to CC, A quantitative approach.
[34] Bankoff, G., Frerks, G. and Hilhorst, D. (2004) Mapping
vulnerability: Disasters, development and people. Earth-
scan, London.
[35] Cutter, S.L. (1996) Societal responses to environmental
hazards. International Social Science Journal, 48, 525-
536. doi:10.1111/1468-2451.00053
[36] Fussel, H. (2009) Review and quantitative analysis of
indices of CC exposure, adaptive capacity, sensitivity, and
impacts. Potsdam Institute for Climate Impact Research
(PIK), Potsdam.
[37] Nhemachena, C., Benhin, J. and Glwadys, G. (2006)
Vulnerability to CC and adaptive capacity in South Afri-
can agriculture. Food and Water Security Under Global
Change: Developing Adaptive Capacity with a Focus on
Rural Africa, IFPRI Workshop, Washington.
[38] Leon-Vasquez, M., West, T. and Finan, J. (2003) A com-
parative assessment of climate vulnerability: Agriculture
and ranching on both sides of the US-Mexico border.
Global Environmental Change, 13, 159-173.
[39] Morrow, B.H. (1999) Identifying and mapping commu-
nity vulnerability. Disasters, 23, 1-8.
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888
Copyright © 2012 SciRes. OPEN ACCESS
[40] O’Brien, K., Eriksen, S., Schjolden, A. and Nygaard, L.P.
(2004) What’s in a word? Conflicting interpretations of
vulnerability in CC research. www.cice ro. uio. no
[41] Leichenko, R.M. and O’Brien, K.L. (2001) The dynamics
of rural vulnerability to global change: The case of
Southern Africa. Mitigation and Adaptation Strategies for
Global Change, 7, 1-18. doi:10.1023/A:1015860421954
[42] Nyong, A.O., Adepetu, A.A., Ihemegbulem, V.C. and
Dabi, D.D. (2003) Vulnerability and Adaptive capacity of
Rural Populations in the Sahel to CC and climate vari-
ability: A case study of semi-arid Northern Nigeria. EMF
Snowmass WorkshopCC Impacts and Integrated As-
sessment”, 28 July-6 August 2003, Snowmass.
[43] Paavola, J. (2004) Livelihoods, vulnerability and adapta-
tion to CC in the Morogoro region, Tanzania. Working
Paper EDM 04-12, Centre for Social and Economic Re-
search on the Global Environment, University of East
Anglia, Norwich.
[44] Næss, L.O., Norland, I.T., Lafferty, W.M. and Aall, C.
(2006) Data and processes linking vulnerability assess-
ment to adaptation decision-making on climate change in
Norway. Global Environmental.
[45] Kuhl, J.J. (2004) Disaggregating household vulnerability:
Analysing fluctuations in consumption using a simulation
approach. Growth, Poverty Reduction, and Human De-
velopment in Africa, Centre for the Study of African
Economies and Global Poverty Research Group, Oxford.
[46] Eakin, H. (2002) Rural responses to climatic variability
and institutional change in central Mexico. Working Pa-
per 10/2002, Center for US-Mexican Studies, University
of California, San Diego.
[47] Ford, D.J, Barry, S. and Wandel, J. (2006) Vulnerability
to CC in the Arctic: A case study from Arctic Bay, Canada.
Global Environmental Change, 16, 145-160.
[48] Haan, N., Farmer, G. and Wheeler, R. (2001) Chronic
vulnerability to food insecurity in Kenya—2001. A World
Food Program Pilot Study for Improving Vulnerability
enya _va m/i nde x.a sp.
[49] Aandahi, G. and O’Brien, K. (2001) Vulnerability to CCs
and economic changes in India agriculture. Waters of
Hope: The Role of Water in South Asian Development,
Biannual Conference of the Nordic Association for South
Asian Studies, Voss.
[50] Bantilan, C.S. and Anupama, K.V. (2002) Vulnerability
and adaptation in dry land agriculture in India’s SAT:
Experiences from ICRISAT village level studies. Adapta-
tion to CC for Agricultural Productivity, South Asia Ex-
pert Workshop, New Delhi, 1-3 May.
[51] Canadian International Development Agency (2003)
Methodology used in the CIDA study: An example of In-
dia. AIACC Asia-Pacific Region Open Meeting.
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888 887
Annex 1. Indicators and proxy variables for vulnerability analysis.
Categories of Indicator Indicator Level of Analysis Authors
1) HH characteristics Household size HH [11,42]
Female headed households HH, District [40,42,43]
Labour unit HH
Age of HH head HH [44]
Educational level of HH head HH, District and National [24,42,45]
Asset ownership (land, houses, farm
equipments and other assets) HH, District [24]
Non-farm engagement HH, District [11,42,46-48]
Food stock HH, District and National [42]
Drinking water HH [43,49]
Additional variables included for study area
Marital status HH
Access to EWS HH
Experience of Agr. Activity HH
2) Economic characteristics Farm income level HH [48]
% of HH below poverty line District [43]
Expenditure on food HH [40,48]
Infrastructure HH, District and National [40,48]
Additional variables included for the study area
Ownership of radio HH
Ownership of perennial crops HH
Number of farm plots HH
Food reserve HH
Cash reserve HH
3) Institutional characteristics Social network HH [42,47]
Institutional arrangements District and National [40,47]
Additional Variables for the study area
Access to credit HH
Access to Agr. extension HH
4) Farm charters tics livestock ownership HH [43]
Cropping system HH [40,50]
Fertilizer applications HH [49,50]
Irrigation usage (rate or sources) HH, District [40,49]
Variables added for the study area
Area under improved technology HH
5) Environmental (biophysical) characteristics Soil conditions HH, District and National [40]
Climatic conditions HH, District and National [40]
Vegetation District and National [48]
CC induced shocks(drought and flood)District and National [40,51]
Variables added for the study area
Soil and water conservation HH
Topography of the farmlands HH
Source: Adopted with modification from Nhemachena, Benhin, and Glwadys [37] also coated by Deressa, Hassan, and Ringler [5].
Copyright © 2012 SciRes. OPEN ACCESS
G. Te sso et al. / Agricultural Sciences 3 (2012) 871-888
Copyright © 2012 SciRes.