Water storage practices for small-scale irrigation systems in East Hararghe zone , Ethiopia

Agricultural productions of smallholder farmers are vulnerable to rainfall shortage and variability in Ethiopia. Thus, this study was aimed to examine water storage strategies practiced by farmers for small-scale irrigation systems in East Hararghe zone, Ethiopia. Data were collected from 300 irrigation user households using an interview schedule. Collected data were analyzed using descriptive statistics and multivariate probit (MVP) model. The research findings revealed that about 62.67, 54.33, 20.33 and 17.67% of households have used borehole, shallow wells, hand-dug, and farm pond water storage practices for small-scale irrigation, respectively. The finding of the MVP model showed that the choice of using borehole water storage practices affected by education level and frequency of extension contacts. The choice of shallow wells to store water was influenced by livestock size, participation in non-farm activities, and the use of the improved seed. Furthermore, the choice of using hand-dug is affected by household size, farm size, access to training, external support, and weather information. Finally, the size of livestock ownership influenced the choice of ponds for small-scale irrigation. Therefore, various stakeholders should work to enhance the water storage capacity of the farmers by reducing the hindrances and strengthening enabling factors in the study areas.


INTRODUCTION
Agriculture is integral to reduce poverty and ensure adequate food supplies for the growing population (Kaczan et al. ). However, agricultural production of many developing countries highly depends on rain-fed and vulnerable to both intra-annual and inter-annual fluctuations in rainfall (Pachauri et al. ). Rainfall shortage and variability result in the risk of intra-seasonal droughts. Drought aggravates a high level of food insecurity and poverty (Cochrane ).
Climate change is leading to an increase in the frequency of droughts thereby negatively affect the production of crops particularly during dry periods (Field et al. ; Shikuku et al. ; Asravor ). These challenges can be alleviated by storing locally available water from the ground, surface, or rain when and where it is abundant (Nicol et al. ; Rockström & Falkenmark ). They argued that agricultural water storage can relatively increase agricultural and economic growth and enhance people's well-being by supporting crops and/or livestock during dry periods.
Small-scale irrigation has vital importance to raise agricultural production and productivity of smallholder farmers to achieve food self-sufficiency and food security at the According to FAO (), about 97% of Ethiopia's food crops are produced by rain-fed agriculture, whereas only 3% is from irrigated agriculture. While irrigation has the potential to increase cereal yields by up to 40% (Diao et al. ), only 8.70% of the smallholder farmers use irrigation water in Ethiopia (Sheahan & Barrett ). There is a huge gap between the potential and the level of irrigation applied in the country due to exhaustively unidentified factors in specific areas of the country. Hence, strengthening the options available for storing water and reducing the hindrance in irrigation activities need to be made to meet the growing food demands in Ethiopia. Agricultural water storage from underground and surface sources such as ponds, wells, borehole, and natural wetlands/streams are just among a range of possible water storage options (Hagos et al. ).
In areas where the amount and distribution of rainfall is insufficient to sustain crop growth and development, an alternative approach is to store water. East Hararghe zone is one of the zones affected by the disappearance and shortage of water sources, climate variability, and recurrent weather-induced shocks such as drought (Setegn et al. ; Tesfaye & Seifu ). Due to the topography and hydro-geological condition, the zone is a water resourcescarce area ( Jema et al. ). The zone is ranked as the least among all zones of the Oromia region in terms of surface and groundwater potential (Berhane et al. ). It is also among the chronic food insecure area and characterized by a dense population (Mulugeta et al. ; Alemayehu et al. ). Under such variable conditions, the storage of irrigation water during the wet season can increase local water availability during dry periods. This helps to mitigate the negative effects of intra-seasonal dry spells and bridging the dry seasons and wet seasons thereby improving the agricultural productivity of smallholder farmers. Thus, water storage for small-scale irrigation is an essential tool to enhance agricultural production and productivity (Wisser et al. ). Therefore, this study was aimed to examine water storage strategies practiced by smallholder farmers for small-scale irrigation systems; and to identify factors that affect the choice of irrigation water storage for small-scale irrigation in East Hararghe zone, eastern Ethiopia.

Description of the study areas
The study was conducted in Haramaya and Kersa districts.
Both districts are found in East Hararghe zone, Oromia Regional State, eastern Ethiopia ( Figure 1). The capital city of the zone, Harar, is located at 526 km from the capital city of Ethiopia, Addis Ababa. The agro-climatic characteristic of the zone is 7.67% highlands, 24.57% midland, and 67.76% lowland. The zone is bordered by Bale zone to the south, west Hararghe zone to the west, Dire Dawa administrative council to the north, and Somali regional state to the north and east. The altitude of the zone ranges from 1,200 to 3,400 metres above sea level. The annual rainfall of the area ranged from 400 to 1,200 mm and the minimum and maximum temperature is 12 and 32 0 C, respectively. Office, ). According to (CSA ), the zone has a total population of 3,490,222 persons of whom 1,772,964 were males and 1,717,258 were females. The farming system is characterized by mixed crop-livestock farming. Therefore, the crop-livestock production system is the dominant activity of the farmers in the area to generate income and improve their livelihood using both irrigation and rainfall.

Sample size and sampling techniques
Multi-stage sampling procedures were employed to select the study area and sample households. In the first stage, two districts namely Haramaya and Kersa were selected from the East Hararghe zone by considering their high potentiality in water storage capacity and irrigation practices compared to the remaining districts of the zone. In the second stage, three kebeles from each district were selected randomly.
Probability proportional to sample size methods was applied to draw the sample households from each kebeles based on the number of households in sampled kebeles.

Methods of data collection
This study was based on a cross-sectional household survey to collect quantitative data. Quantitative data were collected using interview schedule. Moreover, relevant and necessary secondary information and records for the study were collected from different published and unpublished literature such as journal articles, research conducted by others, reports of government and non-government organizations, books, and conferences.

Methods of data analysis
Quantitative continuous/dummy types of variables were analyzed using descriptive statistics such as frequency distribution, percentage, mean and standard deviation. A multivariate probit (MVP) model was applied to identify Smallholder farmers are more likely to use a combination of water storage practices for irrigation systems to deal with water scarcity risks. This means that there is a probability of using more than one water storage practices for irrigation systems. Hence, the study applied the MVP econometric model to measure the influence of the set of independent variables on each of the different water storage practices for irrigation systems.
The dependent variable in the empirical estimation for this study is the selection of water storage practices for irrigation systems from the set of water sources such as water extracted directly from ponds or natural streams, open shallow wells, and borehole. For this study, an MVP econometric approach was characterized by a set of m dependent variables yhpj such that: where: j ¼ 1, 2 … m ¼ water storage practices for irrigation systems X 0 hpj ¼ vector of independent/explanatory variables β j ¼ vector of parameter to be estimated U hpj ¼ error terms distributed as multivariate normal distribution with zero means Y hpj ¼ whether or not a household using particular water storage strategies.
Because the use of many water storage practices is possible, the error terms in Equation (1) are assumed to jointly follow a multivariate normal distribution with zero conditional mean and variance normalized to unity. The covariance matrix represents the unobserved correlation between the stochastic components of the jth and mth type of water storage strategies. This assumption means that Equation (2) gives a MVP model that jointly represents decisions to use particular water storage strategies.

Dependent variable
Water storage strategies practiced by farmers for irrigation systems is the dependent variable of the study. Six water storage practices were identified. The dependent variables of the study are specified as: where k is the water storages practiced by farmers.
Y k1 is water storage and irrigation using ponds Y k2 is water storage and irrigation using shallow wells (open) Y k3 is water storage and irrigation using boreholes Y k4 is water storage and irrigation using hand-dug Y k5 is water storage and irrigation using flood and stream diversion Y k6 is water storage and irrigation using water harvesting.
Operational definition of independent variables (see Table 2).  Regarding external support related to water storage/irrigation activities from government or non-government, only 4% of the total sample households received support, while the remaining 96% of the households did not receive any support related to water storage or irrigation development strategies.

RESULTS AND DISCUSSION
Out of the total sample households, 43% of them accessed weather-related information, while the remaining 57% did not access weather information. Moreover, of the total sample households, 41% of them had contact with extension workers concerning water storage strategies and utilization practices while 59% have no contact with extension workers related to water storage strategies (Table 3).
The mean distance to the nearest market in a walking  (Table 4).

Irrigation systems practiced in the study areas
Most sample households practiced small-scale irrigation using motor pump technology. According to survey results, about 84.67, 14, and 1.33% of sample households responded that they used motor pumps, natural stream or flood diversion of water to crop field and fetching water from shallow   hand-dug and boreholes were substituting.
As indicated in Table 6

Household size (HHSZ)
Household size has a negative and significant effect at less than 5% significance level on using hand-dug water storage practice. An increase in the household size decreases the  *, **, and *** imply significance at 10, 5 and 1% significance levels, respectively. RSE is Robust Standard Error. Source: Own analysis based on survey data, 2019.
They indicated that the households use the household members as a labour force and easily undertake the irrigation activity than lower household size because of labour intensive nature of small-scale irrigated farming.

Education status (EDULEV)
The likelihood of using hand-dug water storage practices was negatively and significantly affected by level of edu- showed that the positive relationship between NFINCgenerating activities and irrigation practice by households was because the income from non-farm activities was used to cover the water storage and irrigation costs such as inputs and pieces of equipment required. and Nhundu et al. () found that access to TRAWS for irrigation practices positively and significantly influences the engagement in irrigation practice. This implies that farmers who attended more water storage training were found with a higher probability of engagement in irrigation practice than their counterparts.

External support (EXSUP)
Access to external support on water storage strategies for small-scale irrigation has a negative and significant effect on households' choices of hand-dug water storage strategy.
Access to external support would decrease the probability of choosing hand-dug water storage practices for smallscale irrigation. This means that households received external support in kind or cash from various government and non-government organizations to invest in other options of water storage practices such as borehole and well construction than households who did not receive the support. Since the cost of irrigation equipment has increased beyond what most farmers can afford, the support of government and non-government organizations enhances the farmers' involvement in irrigation development activities (Abiyu et al.

Weather information (WRINFO)
Farmers obtain weather-related information through demon- irrigation. This indicates that households' access to adequate weather information focused on various water storage strategies rather than using hand-dug.

Extension contact (EXCON)
The likelihood of choosing shallow wells and hand-dug

Use of improved seed (IMPS)
Irrigation requires the use of modern inputs such as IMPS and fertilizers, which further enhance agricultural productivity (Diao et al. ; Gebregziabher & Holden ). Empirical literature shows that modern inputs are used more in irrigated farming than rain-fed farming and as a result productivity is higher in the former. For instance, Gollin et al. () underlined that the green revolution had the highest impact on irrigated cereals. However, the MVP model results revealed that the use of IMPS has a negative and significant effect on households' choice of shallow wells water storage practices.
Using IMPS would decrease the probability of choosing shallow wells for small-scale irrigation. The plausible explanation is that households who use shallow wells water storage practices for small-scale irrigation are less likely to use IMPS than those who choose other water storage practices.

DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.