Assessing The Impacts of Climate Variability on Rural Households in Agricultural Land Through The Application of Livelihood Vulnerability Index

Climate variability adversely affects rural households in Ethiopia as they depend on rain-fed agriculture, which is highly vulnerable to climate fluctuations and severe events such as drought and pests. In view of this, we have assessed the impacts of climate variability on rural household’s livelihoods in agricultural land in Tarcha zuria district of Dawuro Zone. A total of 270 samples of household heads were selected using a multistage sampling technique with sample size allocation procedures of the simple random sampling method. Simple linear regression, the standard precipitation index, the coefficient of variance, and descriptive statistics were used to analyze climatic data such as rainfall and temperature. Two livelihood vulnerability analysis approaches, such as composite index and Livelihood Vulnerability Index-Intergovernmental Panel on Climate Change (LVI-IPCC) approaches, were used to analyze indices for socioeconomic and biophysical indicators. The study revealed that the variability patterns of rainfall and increasing temperatures had been detrimental effects on rural households' livelihoods. The result showed households of overall standardized, average scores of Wara Gesa (0.60) had high livelihood vulnerability with dominant major components of natural, physical, social capital, and livelihood strategies to climate-induced natural hazards than Mela Gelda (0.56). The LVI-IPCC analysis results also revealed that the rural households in Mela Gelda were more exposed to climate variability than Wara Gesa and slightly sensitive to climate variability, considering the health and knowledge and skills, natural capitals, and financial capitals of the households. Therefore, interventions including road infrastructure construction, integrated with watershed management, early warning information system, providing training, livelihood diversification, and SWC measures' practices should be a better response to climate variability-induced natural hazards.


Introduction
The detrimental effects of climate change and variability have become an environmental and socioeconomic problem that is rapidly causing climate-driven hazards for people around the world (Adu et al., 2018). Globally, climate-related hazards are seen to have a huge impact on young, elderly, poor and marginalized populations such as households headed by women and people with limited access to resources (IPCC, 2014;Tanner et al., 2015;Paul et al., 2019).
Climate-related hazards have many indirect impacts on the livelihoods, health, water, agricultural production and socioeconomic welfare of systems (Gezie, 2019;Masuda et al., 2019;Endalew & Sen, 2020). Climate variability is predicted to increase the frequency and severity of certain severe weather events (IPCC, 2018), and disasters such as floods of agricultural lands, droughts, storms, and cyclones (Ullah et al., 2018). Also, Africa is the utmost vulnerable continent to climate variabilitywith 350-600 million Africans facing increased water stress by the 2050s (Hahn et al., 2009).
Climate change and variability are adversely affecting smallholder farming households in Africa because their activity depends on climate-regulated water resources with low adaptive capacity (Adu et al., 2019). Similarly, dependence on agriculture, pastoralism and lack of irrigation means that African farmers are especially vulnerable to climate hazards (Hahn et al., 2009;Araro et al., 2019). Indeed, rural households' livelihood is considered to be highly vulnerable to climate change and variability (Turpie & Visser, 2013). This livelihood vulnerability of rural farmers in Africa is triggered by exposure to climate change and variability and by combining social, economic, and environmental factors that interact with it, including Sub-Saharan Africa (Ofoegbu et al., 2017). The agricultural sector in Sub-Saharan Africa is extremely susceptible to potential climate changes and variability (Turpie & Visser, 2013).
Food insecurity is one of the major drivers that determine development dynamics in East Africa, especially in Ethiopia; due to these the country faces drought and poverty in different periods due to climate changes and variability that was directly affecting the agricultural output (Few et al., 2015;Ademe et al., 2020;Ketema & Negeso, 2020). Ethiopia is an agro-based economy where agriculture contributes 45% to the gross domestic product (GDP). The agriculture sector is a source of livelihood for more than 80% of the population (Dendir & Simane, 2019). In fact, rain-fed agriculture in the country is more vulnerable to the adverse effects of climate variability (Gezie, 2019) and extreme events like drought and pests (Endalew 98 Ginjo Gitima et al. / Geosfera Indonesia 6 (1), 2021, 96-126 & Sen, 2020. Even if productivity grew, climate variability would still dramatically impact incountry (Teshome & Baye, 2018).
In addition, climate change projected in Ethiopia is expected to result in decreased precipitation variability and an increase in temperature (1.1 to 3.1°C by 2060 and 1.5 to 5.1°C by 2090) with a rise in the frequency and intensity of extreme events such as flood and drought (National Meteorological Agency, 2007). Other studies indicate an increase of temperature in all seasons of 1.4°C to 2.9°C by the 2050s (Conway & Schipper, 2011). Besides, rainfall and temperature patterns show large regional differences (Gezie, 2019). Such trends of increasing temperature, the high variability of precipitation, and the rising frequency of extreme events are expected to continue in the country (Dendir & Simane, 2019).
Vulnerability assessment approaches tend to be inextricably related to the vulnerability concept and interpretation. In line with, the outcome of vulnerability and its conceptual meanings, Dessai & Hulme (2004) highlight the different approaches that the two concepts take (without explicitly referring to them) to inform climate adaptation policy. Physical vulnerability concepts prefer to adopt a top-down approach to assessing the strategy of climate adaptation, while vulnerability of contextual concepts focus on socio-economic vulnerability that follow a bottom-up approach (Young et al., 2009). A top-down approach usually starts with international climate forecasts, which can then be rationalized and used to determine climate change's regional effects.An essential feature of bottom-up approaches is primarily the participation of the stakeholders and population of the scheme in classifying climate-change stresses, influences and adaptive strategies (Fellmann, 2012). According to Neupane et al. (2013) socioeconomic parameters such as access to essential resources like forest, land, and water should also be reflected in the vulnerability analysis. Moreover, the importance of incorporating socioeconomic systems with biophysical systems (integrated approach) at varied spatial and social scales in the vulnerability assessment. An integrated approach is effective and may adequately capture all possible dimensions of vulnerability when one integrates both the biophysical (sensitivity and exposure) and the socioeconomic (adaptive capacity) aspects of vulnerability (Endalew & Sen, 2020).
Studies suggest that poor households' livelihood in rural areas of Ethiopia are the most vulnerable to climate change and variability (Deressa et al., 2009). Similarly, current climate shocks and stresses already have an overwhelming impact on the vulnerability of farmers, particularly in rural communities (Sujakhu et al., 2019). Likewise, climate variability vulnerability is understood to be the result of the interaction between the biophysical drivers (include climatic exposure) and the function of the system's sensitivity and adaptive capacity.
The exposure constituents entail individuals, biological systems, ecological capacities, services, assets, infrastructure, financial, or social resources in places and settings that could be unfavorably influenced by climate change and variability (Ademe et al., 2020). Sensitivity is the degree to which the rural household is adversely affected by exposure to climatic variables' variations (Teshome, 2017). The adaptive capacity constituent the capacity of systems or people ability, establishments, people, and different ecosystems to conform to potential harm, exploit openings, or react to varied consequences (Amuzu et al., 2018).
Different scholars have been conducted to study the vulnerability of Ethiopian households to climate-related extreme events. For instance, a study conducted by Dercon et al. (2005) using panel data set. However, most of these studies are very general and the results are aggregated at national or regional levels. These studies have also been limited concerned about rural livelihoods vulnerability to climatic-hazards on district and context-specific nature at a local level. In addition, aggregated national results do not capture the complex state of vulnerability at the local level, while they are important to understand development priorities (Simane et al., 2014;Narayanan & Sahu, 2016). Moreover, the context-specific essence of risk and interventions did not examine the degree to which rural livelihoods in agricultural land are vulnerable to climatic-related extreme events (Ford et al., 2010;Azene et al., 2018).
Hence, our study focuses on livelihood vulnerability to climate variability at contextspecific nature in Tarchazuria district of Dawuro zone. Also, Dendir & Simane (2019) suggested that stakeholders plan context-specific intervention is important than the national level to reduce rural farmers' vulnerability to climate variability and strengthen farm households' adaptive capacity. Tarchazuria district faced climate-related natural hazards and no study has examined in our study area in local detail. The rural farm households in the district are predominantly rain-fed and hence are prone to risks of climate variability. Due to frequent climatic events like drought, floods, and rainfall irregularities, there are the main problems on indirect costs, crop failure, death of livestock, water shortage, and loss of biodiversity. Moreover, climate variability has also direct and indirect impacts on the prevalence and spread of diseases and pests in the study area. Therefore, this study aimed to assess the impacts of climate variability on rural households 100 Ginjo Gitima et al. / Geosfera Indonesia 6 (1), 2021, 96-126 in agricultural land through the application of the Livelihood Vulnerability Index in the Tarchazuria district of Dawuro Zone.

Biophysical Setting of The Study Area
This study was conducted at Tercha Zuria district in the Dawuro zone of Southwest Ethiopia. Geographically, the study area located between 7°05'00" to 7°15'00"N latitude and 36°45'00'' to 37°20'00''E longitude (Figure 1).The study area is located at 510 Km in Southwest of Addis Ababa the capital city of Ethiopia. The district shares borders in the North with Maraka and Tocha district, in the South and Southwest Gojeb river, in the East and Northeast Gena district and in the West Konta special district. The district covers a total area of 588 square kilometers. The physiographic setting of the study area is a dissected and rugged landscape, having well-drained and moderately weathered brown soil (Nitisols) and Orthic Acrisols. Thus, soil erosion and floods in the area is mainly attributed to the dissected and rugged topography. The geology of the study area is abundant with rhyolites and trachy basalts mainly overlying in the Precambrian basement and tertiary volcanism (Bore & Bedadi, 2015;Gitima & Legesse, 2019).
The elevation ranges lie between 918 m to 2170 m above sea level. The dominant agro-ecology in the districtis tropical (kola) and sub-tropical (Woina-dega) agro-climate. The average annual minimum and maximum temperatures of 13 years were 14.65℃ to 16.12℃ and 26.4℃ to 29.3℃, respectively. The 13 years (2007-2019) of mean annual rainfall was 1398.8 mm, and the mean monthly rainfall ranges between 18.6 mm and 323 mm (National Meteorological Agency, 2019).
The rainfall is a bimodal type in the study area: the short rainy season is between March and May, and the long rainy season between June and September (Bore & Bedadi, 2015).
Agriculture is mainly composed of crop production and animal husbandry and it is the main source of livelihood of the population in the district. The dominant activities under land use pattern in the study area include the cultivation of perennial crops such as enset (Enseteventricosum), banana, coffee, mango, avocado and etc. Whereas the annual food crops, including cereals (maize, sorghum, teff), pulses (beans, peas), (maize and teff are largest produced), and root crops like potatoes, yams, sweat potatoes and cassavas. Generally, mixed agriculture is the major economic activity in the study area (Gitima & Legesse, 2019). However, the watershed has ample potential for cultivations, its farm productivity is very low because farmers use traditional means of production. Besides, crop production is mainly rain-fed coupled with poor market access makes the livelihood of farming households extremely stagnant (Abebe, 2014).

Data Sources and Collection Tools
The data required for the current study is obtained from both primary and secondary sources and also these necessary data were of both qualitative and quantitative in nature. The primary data were collected through the questionnaire, key informant interviews, FGDs, and field observations. Questionnaire was used to collect information from the sampled rural households. Prior to the survey, the enumerators were trained how to interview and fill the questions. Close-ended and open-ended format questions were prepared to the selected sample rural household heads and administered through face-to-face interview to get information about the impacts of climate variability on rural household livelihoods. Also, two focus group discussions, the discussion among a small group of six to seven members of the farmers were carried out in the district. In addition, key informant interviews were held with respondents from different sections of the community such as three development agents, two from non-government organizations, four model farmers, and three elderly farmers. Moreover, secondary data were collected from published and unpublished documents. Furthermore, time series climatic data such as temperature and rainfall were obtained from the regional meteorological agency (Hawassa) to predict the trend and variability over time. The reference periods for the climatic data were between 2007 and 2019. This range was chosen based on the concept of climate variability and its resulting effects on the rural livelihoods in agricultural land.

Research Design and Sampling Procedure
This study employed a cross-sectional survey research design and longitudinal time series meteorological data were used records over the period of 2007-2019. In selecting representative sample households, multistage sampling techniques were carried out to select sample household heads for the study from the district. The first stage, Tarchazuria district, was selected using purposive sampling techniques among the ten districts of Dawuro zone because in the district rural farmers' livelihoods affected by climate variability like drought and extreme events, and climate data availability and meteorological station in the area. Secondly, two kebeles were purposively selected using on the above district selection technique i.e., : Mela Gelda (372 household heads) and Wara Gesa (464 household heads).Finally, simple random sampling procedure was applied to select 270 representative farm household heads for the study.

Methods of Data Analysis
The unit of analysis of this study focused on rural farm household heads. Qualitative data were analyzed by using thematic analysis of categorization; the data were gathered through observation, interview and focus group discussions. Quantitative data were analyzed by descriptive statistics such as percentage, mean, ratio, maximum, and minimum by using Microsoft Excel. Metrological data such as rainfall was analyzed by using standardized precipitation index and coefficient of variation (CV), whereas, temperature was analyzed by means of simple linear regression and standardized temperature anomalies. Household Exposure (HE) and household Sensitivity (HS) indices complemented with basic household information of farmers were analyzed using descriptive statistics.

Simple Linear Regression
It is the mainly used to analyze the association between one quantitative result and a single quantitative explanatory indicator. The method is important to detect and characterize the long-term trend and variability of temperature and rainfall values at the annual/monthly time scale. The parametric test takes into account random variable Y on time X in a simple linear regression. The regression line slope coefficient was interpolated that computed from the data is a coefficient of the regression or the Pearson correlation coefficient (Teshome, 2017). It can be calculated with eq. 1: Where: refers natural disasters (rainfall and temperature variability) during the period; α is constant of regression; represents slope of the regression equation; refers to number of years from 2007 to 2019.

Standardized Precipitation Index (SPI) Standardized Precipitation Index (SPI) developed by the (World Meteorological
Organization, 2012). The number of cold nights and warm days per month was calculated using the monthly observation of minimum and maximum temperature, respectively. The SPI was used to identify droughts across the years from 2007 to 2019. It is a statistical measure indicating how unusual an event is, making it possible to determine how often droughts of certain strength are likely to occur. The practical implication of SPI-defined drought, the deviation from the normal amount of precipitation, would vary from one year to another. It can be calculated with eq. 2: where; SPI= anomaly of rainfall (irregularity) in different time period; xi is yearly rainfall in the study period; ̅ is the long-term average yearly rainfall; and is the standard deviation of rainfall in observed time period (Teshome, 2017). Accordingly, the drought severity classes are: extreme drought (SPI <-1.65), moderate drought (-0.84 >SPI > -1.28), severe drought (-1.28 > SPI > -1.65) and no drought (SPI >-0.84) (World Meteorological Organization, 2012).

Constructing Livelihood Vulnerability Index
Vulnerability is one factor determining whether people have risks to their livelihoods in agricultural land or not (Suryanto & Rahman, 2019). Thus, the index is used for comparison among the communities. In addition, the Sustainable Livelihood Framework (SLF) where vulnerability context is the major determinant of sustainability of livelihood assets as it directly influences livelihood strategies, institutional process, and livelihood outcomes of the community.   (Hahn et al., 2009). We used both equal and unequal weights in this study, then used an integrated method to compute composite vulnerability indices using weighting average systems.
According to Adu et al. (2018), a single component is consisting several subcomponents (indicators), each of these indicators is calculated on a different scale, such as percentages or ratios and etc., therefore, it was necessary to the data into indices using either eq. (3) or eq. (4). (3) which can be also expanded as: where; LVI-IPCCh indicates the LVI for household h represented using the IPCC vulnerability framework, e is the households' exposure result, a is households' the capacity of adapative result, and s is the household's sensitivity result (weighted mean score of the health, knowledge, skills, natural capital and financial major components) which ranged from (-1) the least vulnerable to (+1) the most vulnerable on the LVI-IPCC scale (Adu et al., 2018).

Maximum And Minimum Temperatures Over The Last 13 Years
The average temperature hurts agricultural output and significantly reduces agricultural output. A one percent increase in average temperature would reduce agricultural output by 2.5% in the long run. The long-run elasticity of agricultural output concerning average temperature is -2.5 indicating that agricultural output is most sensitive to an average temperature increase in the long run. A decrease in agricultural productivity is likely as a result of increased temperature variability. This may be due to the fact that high temperatures deplete soil nutrients, making livestock and agricultural productivity difficult (Ketema & Negeso, 2020). Climate variability causes the frequency and severity of weather events.
Accordingly, an analysis of the climate variability in the study area over the last 13 years ( Similarly, the study made by Kedir & Tekalign (2016) in the pastoral community of the Karrayu people in the Oromia region reported that the mean maximum monthly temperature indicates an increasing trend except for July and August. Rainfall in Ethiopia is a major input in determining output due to this the country is named as rain-fed agriculture, where rainfall play an important role (Ketema &Negeso, 2020). As shown in figure 4 the analysis of metrological data of rainfall indicates the annual temporal variations. The annual rainfall variability from 2007 through 2019 can be detected from the CV value. The result showed that the study area's annual temporal CV was 19.5 percent, indicating a low variability in rainfall. According to Asfaw et al. (2018), CV below 20% implies less variability and hence annual rainfall experienced less variability. However, key informant interviewers indicated that climate variability has become unpredictable and associated with erratic rainfall. They also claimed that rainfall's erratic nature brings indescribable hardship to study communities as most of them expressed unhappiness to the current irregular, and unstable nature of rainfall currently experienced. Similar findings have been found by Araro et al. (2019) in Konso district of Southern Ethiopia, unexpected rain followed by heavy flood and drought. These variations in rainfall pattern have a direct impact on crop yields, livestock production and price fluctuation from the agricultural perspective.
Also, FGDs discussants reported there is a high variability of rainfall and rainy seasons could either delay when farmers predict a fall of rains when they least expected them in the district.
Therefore, FGDs discussants suggested livelihood diversification strategies, and water harvesting methods during the rainy seasons should be the best options to adapt to existing rain variability and extreme weather events. Likewise, Kedir & Tekalign (2016) suggested that proper use of water harvesting technology should be devised to use and manage the intense rainfall of July and August in their study in central Ethiopia. Moreover, early warning systems and integrated watershed and environmental management measures are required to minimize/avoid disaster and design possible remedial actions.
The rainfall anomaly also witnessed for the presence of annual variability and the trends being below the long-term average. As shown in figure 4, the SPI (rainfall anomalyvariability and irregularity) can identify and monitor droughts. The evaluation of SPI at a certain location is based on a series of accumulated rainfall for a different monthly time scale in a year. The rainfall series is fitted to probability distributions that are subsequently transformed into normal distributions. It follows that the average SPI for the target location and the chosen period is zero. Negative SPI numbers specify less than median or long-term average rainfall, whereas positive SPI values indicate greater than median rainfall (Mohammed & Scholz, 2019).  The standard deviation is one way of summarizing the spread of a probability distribution; it directly related with the degree of uncertainty allied thru predicting the value of a random variables. High values indicate more uncertainty than low values (Teshome, 2016). Accordingly, May (129.6), April (79.5), and October (77.8) had the highest standard deviation indicates more uncertainty in the district (Table 3) December (43.7). It has been observed from the study that rainfall is generally at its peak among August, July, and September, receiving more than three fourth of the amount of rainfall in these months. Practically, assessment of livelihood vulnerability is too complicated and difficult to be covered all because there are many aspects, dimensions and factors that relating to livelihood vulnerability, e.g., economic, political, demography, etc., and it was certainly mentioned in some reports (Can et al., 2013). This study only focuses on some major components that influence rural livelihoods in agricultural lands of households due to climate variability in the Tercha District of Dawuro zone.
The results of LVI standardized average scores of all 13 indexed major components calculated from 45 subcomponents or indicators commune are presented collectively in Table   4. In addition, when the total standardized weighted scores of the indicators of forest resources showed that Mela Gelda (0.53) was less vulnerable than Wara Gesa (0.73). These were because of the large percentage of households depending on forest resources recorded in Wara Gesa (73) than Mela Gelda (54). In comparison, the highest percentage of households reported that about a change of tree cover and severe damage to common forests in Mela Gelda than Wara Gesa. The key informant interviewee realized the farmers located near the main roads and close to the market place clear forests because charcoal is their income source.Wara Gesa (0.74) showed a slightly higher vulnerability standardized score in terms of water resources than Mela Gelda (0.70) on this aggregated major component. The indicators of water resources were more vulnerable to climate-induced natural hazards due to a high percentage of households reporting water conflict in past years and households to utilize water from unprotected sources.

Financial Capital Vulnerability
As indicated in When indicators reviewed the major components networks and relationships, Wara Gesa was the most vulnerable in terms of households' heads reported that a high percentage of household heads not associated with any organization/cooperative in Wara Gesa (75.3) than Mela Gelda (37.5), and a higher percentage of household heads had loose ties to relatives/neighbors in Wara Gesa (23) than Mela Gelda (12). By organization affiliation on households' vulnerability to climate variability, the results show that levels of vulnerability in WaraGesa (0.38) was highest vulnerable to climate-induced natural hazards than Mela Gelda (0.20), this was because of a high percentage of households not a member of the organization like idir and ikub, etc.

Livelihood Strategies Vulnerability
The indexed livelihood strategies component /profile consisted of four subcomponents/indicators. Considering the percentage of households dependent exclusively on agriculture as a source of income as an indicator a higher vulnerable in Mela Gelda (83) than Wara Gesa (62.4), and average inverse agricultural livelihood diversification index a higher vulnerable in Wara Gesa (0.685) than Mela Gelda (0.50). Wara Gesa (54%) shows a slightly greater vulnerability to climate variability based on the percentage of households unable to save crops for contingency than Mela Gelda (52%). Wara Gesa also showed greater Ginjo Gitima et al. / Geosfera Indonesia 6 (1), 2021, 96-126 vulnerability (77.4 %) on the percentage of households categorized themselves poor than Mela Gelda (63%). Based on similar indicators that calculate their respective methods of the LVI-IPCC contributing factors were computed by grouping exposure, sensitivity, and adaptive capacity into three groups (Table 5). The LVI-IPCC contributing factors in the study area showed households for Mela Gelada (0.64) have a higher standardized average score than Wara Gesa (0.57). According to the IPCC classification of vulnerability exposure to natural hazards caused by climate variability was a high contributing factor for rural households. Yet, Wara Gesa households (0.55) have a greater capacity for adaptation than MelaGelda (0.47). The sensitivity contributing factor value for Wara Gesa (0.60) is slightly lesser than that of the Mela Gelda (0.62) indicating that Mela Gelda was more sensitive than Wara Gesa. The standardized weighted result of the overall LVI-IPCC score was for Mela Gelda (0.105) and for Wara Gesa (0.012), indicating that the showing of the incidence of great vulnerable conditions of rural households to climate variability-induced natural hazards in the district which is a similar result to that of the LVI standardized weighted scores.  Figure 6 also shows the vulnerability triangle that plots scores of contributing factors for adaptive capacity, exposure, and sensitivity. The vulnerability triangle reveals that the livelihoods in agricultural land of rural households in Wara Gesa were more vulnerable in terms of household adaptations' capacity considering the major components of the sociodemographic profile, livelihood strategies, and social networks. The rural livelihoods in agricultural land of households in Mela Gelda were more exposed than Wara Gesa to climate variability and slightly sensitive to climate variability, taking into consideration of the health, and knowledge and skills, natural capitals, and financial capitals of the households in the study area.

Conclusion
Rural households in Mela Gelda were a higher vulnerable than those in Wara Gesa in terms of indexed major components such as health, skill, and knowledge, socio-demographic profile, income and wealth, policy and leadership services. In comparison, farm households in Wara Gesa were more vulnerable in terms of land resources, forest resources, water resources, networks and relationships, organizational affiliation, and livelihood strategies.
The livelihoods in agricultural land of rural households in Wara Gesa were more vulnerable in terms of the capacity for household adaptations considering socio-demographic profile, livelihood strategies, and social networks. The rural households in Mela Gelda also more exposed than Wara Gesa to climate variability and slightly sensitive to climate variability, considering the health, knowledge and skills, natural capitals, and financial capitals of the households in the study area. Hence, interventions including road infrastructure construction, integrated with watershed management, specific area early warning information system, livelihood diversification, afforestation/reforestation, and land degradations rehabilitation should be a better response to climate variability-induced natural hazards in the study area.

Conflict of Interest
The authors declare that there is no conflict of interest.

Acknowledgments
The authors would like to thank the Tercha district agricultural offices experts for their support in providing the necessary data for the study. In addition, we have enormously benefited from the study communities, and they shared for us their knowledge and experiences with patience without the feeling of tiredness. We also wish to thanks the