1 Introduction

Climate-smart agriculture (CSA) is a comprehensive management strategy for crops, livestock, forests, and fisheries that tackles the interconnected issues of food security and climate change [1]. It aims to increase productivity sustainably, adapt to climate change and build resilience to it, and reduce or eliminate greenhouse gas emissions [2]. Although agriculture is one of the most important economic sectors in Africa that employs about 70% of the population and accounts for about 14% of the Gross domestic product (GDP) in Sub-Saharan Africa [3], it is a highly vulnerable sector to the impact of climate change [4]. Similarly, in Ethiopia, agriculture contributes 37.64% to the GDP, a source of employment for about 67% of the population, and more than 79% of foreign exchange earnings [5]. However, its performance has remained unsatisfactory in feeding its ever-growing population [6], which has risen to a projected figure of more than 126 million in the year 2023, which is a 2.55% increase from the year 2022 [7]. Therefore, this rapidly growing population needs greater economic performance than before to at least guarantee food security. Moreover, the contribution of the agricultural sector to Ethiopian GDP has been declining over time due to its dependence on rain-fed and traditional farming systems, low or no adoption of improved agricultural technologies, vulnerability to climate change stresses, and farmers’ intrinsic inefficiency in utilizing the available agricultural resources [6, 8,9,10,11,12]. Furthermore, the country has only an area of about 1.1 million square kilometers. Thus, given this scarce area of land along with the rapidly growing population, intensification, which requires investments in CSATs is a better choice that is used to increase agricultural productivity, resilience, and income [12,13,14].

Climate-smart improved wheat variety refers to the type of varieties that are resistant to disease, can withstand drought, and/or yield more [2, 15,16,17,18]. Irrigation technology encompasses effective water harvesting and management, whether through large-scale systems like canals and pumps or smaller-scale micro-irrigation strategies like pond construction, water wall construction, and utilizing all available water for wheat production [2, 4, 16, 17, 19, 20]. It also includes any technology that moves water from its sources to previously unreachable areas [21]. Similarly, integrated soil fertility management includes the careful application of chemical fertilizers together with integrated sustainable practices like crop rotation, intercropping, and the integrated use of organic fertilizers like farmyard manure, compost, and careful application of chemical fertilizers [2, 4, 17, 19, 22,23,24,25]. Hence, since the adoption of improved wheat varieties, irrigation, and integrated soil fertility management raise resilience, lower greenhouse gas emissions, and increase agricultural productivity, they are climate-smart agricultural technologies [2, 4, 15,16,17,18,19].

As the principal food crop in terms of area covered and production volume, cereals make up over 88% of Ethiopia’s grain crop production [26]. However, given the unsatisfactory performance of cereal production in feeding its ever-growing population and the rise in vulnerability to climate change on the other sides, increasing cereal production and productivity is a matter of existence [10, 27]. Thus, it is crucial to create and spread CSATs that can increase household resilience to food insecurity in the face of climate change [28,29,30].

Ethiopia is the leading producer of wheat in Sub-Saharan Africa (SSA) followed by South Africa, Sudan, and Kenya [31]. Among cereal crops, in Ethiopia, wheat ranks third in terms of total production next to maize and teff with 1.86 million ha of land and productivity of 31.11 qt/ha [26]. Similarly, in the Oromia region, wheat production took the third rank 1.95 million numbers of holders, 1.03 million hectares of land, and productivity of 33.64 qt/ha [26]. In the study area, Arsi zone, wheat production had the second rank in terms of production next to maize with 393,316 numbers of holders, 256,172.2 hectares of land, and productivity of 35.30 qt/ha [ibid]. Thus, wheat plays a major role in the socioeconomic growth of smallholder farmers in the research area and remains a strategic crop to meet the growing demand for food [10]. However, although the study area’s average productivity (33.1 qt/ha) is higher than the national (29.8 qt/ ha) and regional averages (31.9 qt/ ha), compared to 51.6 qt/ha of the world average [26, 31], wheat productivity is still very low. Governments and other development programs aimed at boosting agricultural productivity and enhancing resilience to climate change recommend the adoption of climate-smart agricultural technologies [32]. Moreover, the government of Ethiopia through its recent policies has been advocating for increasing adoption of CSATs to improve resilience to climate change and food security of smallholder farmers.

In Ethiopia, [8, 33] and [34,35,36,37,38,39,40] conducted their studies on the determinants of adoption of improved technologies in wheat production. However, they considered only one technology, that is improved wheat seed varieties, and ignored the reality that farmers may adopt multiple technologies as complements, and/or substitutes. Besides, although studies by [41, 42] and [30] examine the determinants of the adoption of agricultural technologies in different parts of the country, the selected technology package is not climate-smart and looked at other crops (maize, rice, groundnut, and teff), and overlooked wheat crop. Furthermore, studies by [11, 12, 14, 42,43,44], and [21, 45, 46] captured determinants of the adoption of climate-smart agricultural technologies as an index which is suffering from aggregation problem as well as the possibility that different technologies might be affected by different factors. Likewise, they didn’t conduct their studies at plot level and hence, ignore the reality that wheat farmers may adopt certain technologies on some of their plots but not on others.

Hence, given the widespread food insecurity situations in the nation overall and the study area specifically, which are primarily caused by changing climatic conditions manifested by erratic rainfall, floods, and disease outbreaks, more research is required to better understand the causes of low adoption rates. To the best of our knowledge, no study has yet identified factors that triggered the adoption of climate-smart agricultural technologies by wheat farmers in Ethiopia in general and in the study area in particular. Likewise, climate-resilient crop varieties, integrated soil fertility management, and irrigation are climate-smart agricultural technologies [1, 2, 18, 21, 23]. Hence, this study is carried out to identify factors that affect the adoption of climate-smart agricultural technologies (i.e. improved wheat varieties, integrated soil fertility management, and irrigation) in wheat production in the Arsi zone, Oromia region of Ethiopia.

2 Research methodology

2.1 Description of the study area

Arsi zone is situated within a total area of 21,008 km2 in the Oromia National Regional State of Ethiopia. It borders the Bale zone to the south, the West Arsi zone to the southwest, the East Shewa zone to the northwest, the Afar Region to the north, and the West Hararghe zone to the east. Assela town, the zone’s capital, is situated 75 km south of Adama town and 175 km from Addis Ababa.

The Zone has four agro-climatic zones namely; lowland, midland, highland, and cool that account for about 20.12%, 50.72%, 27.32%, and 1.84%, respectively. It is also distinguished by four seasons namely; summer, autumn, winter, and spring. The altitude of the zone ranges from 805 to 4195 m above sea level. The Zone’s average annual temperature ranges from 10° to 25 °C with an average annual rainfall of 1020 mm [47]. A relatively short rainy season (Belg), which lasts from February to March, and a lengthy rainy season (Meher), which lasts from June to September, characterize the rainfall pattern.

There are 2,637,657 people living in the Zone, of whom 1,323,424 are men and 1,314,233 of them are women. It has a population density of 133.05. In the Zone, there are a total of 541,959 households, with an average of 4.87 persons per household [48]. Among the total districts of the zone, Hetosa, Digalu-tijo, Tiyo, Lode Hetosa, Limuna Bilbilo, Munesa, and Dodota districts have a greater potential in the production of wheat out of 24 districts of the Zone (Fig. 1).

Fig. 1
figure 1

Location of the study area

Full size image

2.2 Data types, sources, and methods of collEctions

Primary and secondary data were used for this study. Primary data were collected by a team of skilled enumerators with previous experience in data collection, familiarity with the research area, and proficiency in the local language. A semi-structured questionnaire was used to gather the primary data that was prepared after conducting a pre-test to refine some of the questions based on the feedback obtained from this exercise. As a result, certain changes were made to the questionnaire based on the results of the pre-test before the actual survey was conducted. Additionally, secondary data were collected from published and unpolished documents, Arsi Zone Office of Agriculture and Rural Development, Dodota, Lode-Hetosa, and Digalu-Tijo districts offices of agriculture and natural resources for augmenting primary data.

2.3 Sampling technique and sample size

To select a representative sample of households for an interview, a multistage sampling technique was used. In the first stage, the Arsi zone was selected purposively due to its high potential in the production of wheat in the region and the country. In the second stage, three districts were purposively selected based on wheat production potentials and the presence of climate-smart wheat technology interventions. These are, Digalu Tijo from the highland, Lode Hetosa from the midland, and Dodota district from the lowland were selected. In the third stage, three kebelesFootnote 1 from Dodota, four kebeles from Lode Hetosa, and five kebeles from Digalu Tijo districts, a total of 12 kebeles were selected randomly based on probability proportional to the number of kebeles in each district. In the final stage, 422 wheat-producing households were selected randomly based on probability proportional to population size. However, the study was carried out at the plot level because farmers might use a particular technology on some of their plots but not on others. In light of this, data at the plot level were gathered from 628 plots handled by 422 randomly selected wheat farmers. The formula provided by [49] was used to determine the sample size:

$$n = \frac{{Z^{2} pqN}}{{e^{2} (N - 1) + Z^{2} pq}} = \frac{{(1.96)^{2} (0.5)(0.5)(2,637,657)}}{{(0.05)^{2} (2,637,657) + (1.96)^{2} (0.5)(0.5)}}{\kern 1pt}$$
(1)

where n is the sample size; Z (1.96) is the standard cumulative distribution at 95% confidence level; e is the level of precision; p (which is = 0.5) is the estimated proportion of an attribute in the population, q = 1-p; and N is the size of the total population from which the sample is drawn. Hence, based on the formula stated above, the required sample size is 384. However, considering a non-response rate of 10%, the data were collected from 422 households.

2.4 Method of data analysis

To identify the major determinants of climate-smart agricultural technologies adoption by smallholder wheat farmers (Improved wheat, integrated soil fertility management, and irrigation) Multivariate Probit (MVP) model was employed. The MVP model for the three climate-smart agricultural technologies is specified as:

$${{\text{Y}}}_{{\text{ij}}}^{*}={\upbeta }_{{\text{j}}}{{\text{X}}}_{{\text{ij}}}+{\upvarepsilon }_{{\text{ij}}} ({\text{j}}=1,\dots 3)$$
(2)

where, \({{\text{Y}}}_{{\text{ij}}}^{*}\) is a latent variable associated with technology \(j\) and individual \(i\) that can be translated into binary outcomes such that:

$${Y}_{ij}=1 if {{\text{Y}}}_{{\text{ij}}}^{*}>0\,and\,0, ot{\text{herwise}}$$
(3)

The \({{{\varepsilon }}_{{{ij}}}}\)represents error terms distributed as multivariate normal, each with zero conditional mean, and variance–covariance matrix Ω, where Ω, has values of 1 on the leading diagonal and correlations \(\rho_{jk} = \rho_{kj}\) as off-diagonal elements as presented below.

$$\Omega =\left[\begin{array}{c}1 {\uprho }_{12} {\uprho }_{13 }\\ {\uprho }_{21 1 } {\uprho }_{23}\\ {\uprho }_{31} {\uprho }_{32} 1\end{array}\right].$$
(4)
$$E\left( {{\varepsilon \mathord{\left/ {\vphantom {\varepsilon x}} \right. \kern-0pt} x}} \right) = 0$$
$$Var\left( {{\varepsilon \mathord{\left/ {\vphantom {\varepsilon x}} \right. \kern-0pt} x}} \right) = 1$$
(5)
$$Cov\left( {{\varepsilon \mathord{\left/ {\vphantom {\varepsilon x}} \right. \kern-0pt} x}} \right) = \rho$$

The off-diagonal elements of the variance–covariance matrix \({\rho }_{jk}\) represent unobserved correlations between the disturbance terms related to the jth and kth types of CSAT adoptions namely; improved wheat varieties, integrated soil fertility management, and irrigation \(( j=1, 2\dots 3;k=\mathrm{1,2}\dots 3;j\ne k)\).

2.5 Definition of variables and working hypothesis

The dependent variables of the model were the adoption of improved wheat varieties, integrated soil fertility management, and irrigation. All are dummy variables that took a value of 1 if the framer adopted the technology and 0, otherwise. Similarly, the definition and measurement of the independent variables for climate-smart agricultural technologies(CSATs) are given by the following Table 1.

Table 1 Definitions of independent variables hypothesized to influence CSATs adoptions
Full size table

3 Results and discussion

3.1 Household socio-economic characteristics (continuous variables)

The average educational level of the sample household heads was grade 4, with a range of 0 to 12. Similarly, it was discovered that the sample household heads had an average of 27.3 years of farming experience, with minimum and highest values of 3 and 62 years, respectively. The average family size among the sample respondents was 4.30 persons per household in adult equivalent, which is lower than the national average of 4.6 persons per household [26]. In addition, the average number of contacts a farm household had with extension workers was found to be two per year, with minimum and highest values of 0 and 10 times, respectively. The mean livestock holding for the sample respondents was determined to be 4.40 in the Tropical Livestock Unit (TLU), with a minimum and maximum of 0 and 11 TLU, respectively. The average total income of the sample households was determined to be 85,978.00 Ethiopian birr (ETB). The results also indicate that the sample respondents had to travel 5.10 km on average to the closest market center. Likewise, the average distance that the household traveled to the nearest wheat plot is 1.2 km (Table 2).

Table 2 Summary statistics of continuous variables used in the adoption of CSATs
Full size table

3.2 Household socio-economic characteristics (dummy variables)

The result reveals the sex distribution of the sampled households is quite skewed towards the male, which accounts for 88.4% of respondents. Besides, only 47.9% of households participated in off/non-farm income-generating activities. Similarly, 62.6% of the sample households had training related to the use of CSATs. Besides, only 40.8% of the sample households have access to credit for purchasing necessary inputs, 93.8% are members of social organizations (Edir and Equp), and 70.4% are members of cooperatives. It further shows that, of the sample households, 73.9% have access to market information on input and output price, while the remaining 26.1% do not. Similarly, 65% of the wheat plots of the sample households were flat, while the remaining 35% were medium and gentel slope (Table 3).

Table 3 Summary statistics of dummy variables used in the adoption of CSATs
Full size table

3.3 The adoption status of CSATs

The results show that improved wheat varieties were adopted on 95.22% of the plots whereas integrated soil fertility management was used on 85.19% of the plots. Meanwhile, irrigation was adopted only on 19.43% of the plots (Table 4).

Table 4 Plot level adoption status of CSATs
Full size table

3.4 Adoption of CSATs by agro-ecologies

The relationships between selected districts and the adoption status of the sample households were compared by the following Table 5. The result revealed that all three technologies were highly adopted in highland agroecology accounting for about 43.1%, 37.9%, and 8.3% of improved wheat varieties, integrated soil fertility management, and irrigation, respectively. However, the smallest rate of adoption of the technologies occurs in lowland agroecology which accounts for about 18.5%, 15.9%, and 5% for improved wheat varieties, integrated soil fertility management, and irrigation, respectively. The result of the chi-square test revealed that there were statistically significant mean differences among agroecologies in the adoption of improved wheat varieties and integrated soil fertility management at a 5% and 10% level of significance.

Table 5 Adoption of CSATs and Agro-ecologies
Full size table

3.5 Linkages among the technologies

The multivariate probit model (MVP) is suitable because the correlation coefficients of the MVP error terms are statistically different from zero in one of the three scenarios (Table 6). After the influence of the observed factors has been taken into consideration, the correlation coefficients measure the pair-wise correlation between decisions to adopt any one of the three technologies with the other two technologies [50]. The study’s findings also show that households had a 95.25%, 85.27%, and 19.52% probability of adopting improved wheat varieties, integrated soil fertility management, and irrigation technologies, respectively. The least likelihood of adoption of irrigation could be associated with problems of accessibility and financial constraints. It also demonstrates that the proportion of farmers using all three technologies jointly is 15%, implying that about 85% use either 1 or 2 of the technologies.

Table 6 The Multivariate Probit model’s correlation matrix of CSATs
Full size table

The correlation coefficients for the error terms show complementarities in technologies (positive correlation) and substitutabilities in technologies (negative correlation) [51]. The findings of the simulated maximum likelihood estimation show a positive and significant association between integrated soil fertility management and household decisions to adopt improved wheat varieties. Improved wheat variety along with the use of integrated soil fertility management can raise resilience, lower greenhouse gas emissions, and increase agricultural productivity. Hence, they are a complement to one another. Moreover, the outcome further reveals that there is no distinct line between different technologies and hence, supports the interdependence assumption across the technologies.

3.6 Determinants of adoption of climate-smart agricultural technologies (CSATs)

Even though farmers employ several technologies in combination, there are a lot of factors that can affect their decision to select one over another. Multivariate Probit Model (MVP) to identify the key factors that impede the adoption of CSATs from a survey of prior research’s literature. A statistically significant chi-square test result shows that the independent variables included in the model are pertinent in explaining the differences in the adoption of the three technologies under consideration in the study (Wald chi2 (51) = 114.36, p = 0.000). As evidenced by the likelihood ratio test’s rejection of the independent assumption on the adoption of CSATs, multivariate regression gives results that are more reliable than individual univariate (likelihood ratio test of Rij = 0: chi2(3) = 2.00, p = 0.045).

Furthermore, before model estimation, various tests were made to know whether or not the model surpasses various assumptions. Therefore, the Variance Inflation Factor (VIF) was used to perform multicollinearity tests among all of the independent variables. There were no significant issues with multicollinearity among the explanatory factors since the values of VIF for all variables included in the model were all below 10 (Appendix). Likewise, the Breusch-Pagan test result showed that the model's heteroskedasticity (appendix) problem did not exist. The model goodness of fit also indicated that the model was able to fit the data(appendix). However, endogeneity was tested for climate-smart technology adoptions using the Durbin-Wu-Hausman test. The test result revealed that there was a problem of endogeneity (χ2 = 0.652, P-value = 0.023). Hence, to overcome the problem of endogeneity, two instrumental variables namely; access to credit, and the distance to the wheat plot were selected as instrumental variables, and the adoption of climate-smart technologies was regressed on these variables including other variables in the structural model. Then, the predicted value of climate-smart technology adoption was taken and included in the original model (multivariate probit) model.

The result of the MVP model revealed that among the 17 variables included in the analysis, six of them, namely, level of education, family size, farm size, credit, distance to the nearest market, and market information, were found to have a statistically significant effect on the decision to adopt improved wheat varieties whereas education level, farming experience, family size, and credit, were found to significantly affect the choice to use integrated soil fertility management. The outcomes of the model further indicated that family size, farm size, frequency of extension contact, and total annual income were the significant factors that influenced the decision to use irrigation (Table 7). However, the sign for farm size in irrigation was not as expected. Hence, the discussions about each significant variable are presented as follows:

Table 7 Results of the Multivariate Probit Model
Full size table

The adoption of improved wheat varieties and integrated soil fertility management are positively and significantly influenced by the household head’s education level at 10% and 1% levels of significance, respectively. As a result, households are more likely to adopt improved wheat varieties and integrated soil fertility management as their years of schooling increase. In terms of how to interpret marginal effects, the marginal effect of 0.004 for the level of education in improved wheat varieties indicates that for each additional year in education, the farm households are 0.4% more likely to adopt improved wheat varieties. Similarly, the marginal effect of 0.018 of the level of education for integrated soil fertility management indicates that for each additional year in education, the farm households are 1.8% more likely to adopt integrated soil fertility management. This may be because farmers with a higher level of education can increase awareness and digest information about the advantages of utilizing climate-smart crop technologies more than farmers with fewer years of schooling. Hence, education plays a crucial role in stimulating adoption-related decisions. This result concurs with the findings of [52, 53,54,55], and [44].

Farming experience was also positively and significantly related to integrated soil fertility management at a 1% level of significance. The marginal effect of 0.007 for farming experience implies that an increase in years of farming by one year would increase the probability of adopting integrated soil fertility management by 0.7%. This is because farmers would gain knowledge of the benefits of various advanced technologies and methods as their years of farming experience increased. So, the likelihood of implementing integrated soil fertility management increases with experience. This finding concurs with that of [46].

Family size was negatively related to improved wheat varieties and positively related to integrated soil fertility management at a 5% level of significance. It is also related to irrigation at a 10% level of significance. The marginal effect of -0.008 for improved wheat varieties implies that each additional number of families in adult equivalent would decrease the likelihood of adopting improved wheat varieties by 0.8%. However, the marginal effect of 0.028 for integrated soil fertility management and irrigation indicates that each additional number of families in adult equivalent would increase the probability of adopting integrated soil fertility management and irrigation by 2.8%. According to the findings of the study, although households with a smaller number of families are more likely to adopt improved wheat varieties, they are less likely to adopt integrated soil fertility management. The negative link between improved wheat varieties and family size is due to the greater expense of food consumed at home, which would make it more difficult to afford commercial inputs like improved wheat varieties. This outcome agrees with the findings of [56]. Its positive effect on integrated soil fertility management is plausible because, in the study area, the use of integrated soil fertility management is a labor-intensive technology.

Although the farm size has a positive and statistically significant effect on the adoption of improved wheat varieties at a 10% level of significance, it has a negative and significant effect on the adoption of irrigation at a 1% level of significance. If improved wheat varieties are adopted, the marginal effect for farm size is 0.011, which suggests that every 1-hectare increase in farm size would increase the probability of adoption of improved wheat varieties by 1.1%. Similarly, the marginal effect of − 0.058 for farm size under irrigation implies that every 1-hectare increase in cultivated land size would be 5.8% less likely to adopt irrigation technology than those having small size of farms. Therefore, as farm size increases, the possibility of utilizing improved wheat varieties increases, whereas the likelihood of using irrigation falls. Farmers who own more land are more likely to produce more, which results in greater financial resources that they may use to buy improved inputs, leading to a positive effect of farm size on the adoption of improved wheat varieties. This outcome is in line with the results of related investigations [46, 57,58,59,60,61]. However, the negative effect of farm size on irrigation is due to the fact that, in the context of limited land, farm households are forced to intensify farming to meet the rising demand for food among their members and outside [32, 62]. Hence, adopting irrigation technology to increase output from agriculture might be more likely if the amount of cultivated land is small. Previous adoption studies found that households with lower land holdings would be more likely to apply sustainable intensification practices [44, 54, 63, 64].

The frequency of extension contact was positively and significantly associated with the likelihood of adopting irrigation at a 10% level of significance. The marginal effect of 0.025 for the frequency of extension contact implies that as the frequency of contact with development agents increases, the probability of adopting irrigation increases by 2.5%. As a result, farmers who have frequent interaction with development agents are more inclined to use irrigation. This can be explained rationally by the fact that development agents actively encourage the use of irrigation technology by balancing the drawbacks of not employing it and thereby encouraging adoption. Similarly, development agents support input distribution, create awareness, improve access to agricultural credit, and convey agricultural research and development outcomes to smallholder farmers. This finding is therefore in line with those of [12, 46, 59, 61, 65,66,67], who identified extension services as a key driver of the adoption of various CSA techniques.

At 5% and 10% levels of significance, access to credit positively and significantly influences the likelihood of adopting improved wheat varieties and integrated soil fertility management, respectively. Farm households with access to credit are 2.6% more likely to adopt improved wheat varieties than those who do not, as the marginal effect for access to credit for improved wheat varieties is 0.026. Similarly, the marginal effect of 0.047 for access to credit for integrated soil fertility management implies that farm households who have access to credit are 4.7% more likely to adopt integrated soil fertility management than those who have no access to it. This is because having access to credit makes it possible for households to purchase modern agricultural technologies without experiencing income problems, which opens the door for the timely execution of modern farm inputs. This result is similar to the findings of [11, 44 ,68], and [69, 70].

Distance to the nearest market was negatively and significantly related to improved wheat varieties adoption at a 1% level of significance. The marginal effect of − 0.010 for distance to the nearest market center indicates that for every 1 km increase in the distance to the nearest market, they are 1.0% less likely to adopt improved wheat varieties. As a result, households that are close to the nearest market have a greater probability of adopting improved wheat varieties than those that are far away. This makes sense given that market distance raises transportation and transaction costs and restricts access to information about new technology, which makes it more unlikely for distant locations to utilize inputs that have been purchased from the market. This result is consistent with the findings of [32, 44, 64, 71, 72], and [69].

At a 10% level of significance, access to market information positively and significantly determines the adoption of improved wheat varieties. The marginal effect of 0.034 for market information implies that farm households that have access to market information are 3.4% more likely to adopt improved wheat varieties than farm households without such access. This is because access to market information enables farm households to easily obtain necessary information about the availability and price of inputs and outputs. This result is concurring with the finding of [73].

At a 1% level of significance, the log of total annual income has a positive and significant effect on the likelihood of using irrigation. The marginal effect of 0.090 for total income indicates that an increase in household total income by one birr would lead to an increase in the probability of adopting irrigation by 9.0%. Due to the capital-intensive nature of irrigation technology, a household with a higher total income cannot be financially limited and prevented from making timely use of irrigation technology. This outcome concurs with the findings of [74,75,76].

4 Conclusion and policy implications

4.1 Conclusion

Climate-resilient agricultural technology adoptions are essential for sustainably increasing productivity, building resilience to climate change, and reducing greenhouse gas emissions. Similarly, as the opportunity to expand cultivated land is exhausted in Ethiopia due to growth in population and urbanization, the necessity to implement climate-smart agricultural technology in the agricultural sector has become critical. Despite this, the county’s agriculture sector is known for using traditional techniques and technologies. The success of the sector in terms of increasing its contribution to the overall growth of the economy and ensuring food self-sufficiency depends on the development and implementation of climate-smart technologies. So, research on the factors that hinder the adoption of climate-smart technologies is crucial to enhancing agricultural productivity, building resilience, and reducing poverty in the study area in particular and the country in general. Therefore, this study analyzed factors that affect the adoption of climate-smart agricultural technologies among wheat farmers using plot and household-level data collected from a sample of 422 wheat-producer households in the Arsi zone.

The study used a multivariate probit (MVP) model since it was found that farmers were more inclined to use a variety of technologies rather than only one. The three climate-smart agricultural technologies that were taken into consideration for this study were improved wheat varieties, integrated soil fertility management, and irrigation. The multivariate correlation coefficients’ result shows that there was a significant and positive relationship between the adoption of integrated soil fertility management and improved wheat varieties. The MVP model’s findings showed that the factors influencing farmers’ decision to adopt a technology vary between technologies. Therefore, the level of education, family size, farm size, credit, distance to the nearest market, and market information, were found to have a statistically significant effect on the decision to adopt improved wheat varieties whereas level of education, farming experience, family size, and credit, were found to significantly affect the choice to employ integrated soil fertility management. It is also revealed that family size, farm size, frequency of extension contact, and total income were significant factors that determine the decision to use irrigation.

4.2 Policy implications

The following suggestions were made in light of the study’s findings:

The adoption of improved wheat varieties and integrated soil fertility management were both positively and significantly influenced by the level of education. Hence, the government must train farmers by strengthening and establishing both formal and informal types of farmers’ education programs, farmers' training centers, and technical and professional schools. Besides, it has been discovered that having access to credit significantly increases the likelihood that climate-smart technologies will be adopted. Hence, facilitating access to credit services might ease the financial burden of acquiring and owning innovative agricultural technologies. Policymakers must make sure that farmers have access to credit to hasten the adoption of climate-smart technologies. To do this, they must expand the number of microfinance institutions in rural areas that offer credit services to smallholder farmers.

Moreover, distance to the closest market center is another significant factor that influences the adoption of climate-smart technologies, according to the study. Hence, to ensure accessibility and availability of climate-smart technologies the market must be established near the home of farm households. Furthermore, to increase access to market information, the quantity and quality of extension contact, and total annual income, the government, non-governmental organizations, and donor agencies should work cooperatively with farm households.