Smallholders’ Conservation Agriculture Adoption Decision in Arba Minch and Derashe Districts of Southwestern Ethiopia

,


Introduction
Agriculture remains one of the most critical sectors for African economies [1], and its success is heavily infuenced by soil and water management, among other factors. In many parts of Africa, inappropriate land use, poor land management, and lack of appropriate inputs have led to soil erosion, biodiversity loss, declining productivity, and food insecurity [2,3]. Like many other African countries, Ethiopia's economy depends mainly on subsistence rainfed agriculture. Traditional cultivation practice is an essential livelihood source for over 80% of the population, contributing about 42% of the gross domestic product (GDP) and more than 80% of its exports [4]. Using a small parcel of land, smallholders produce over 90% of the main crops for the nation [5]. Despite its role, with an average population growth rate of 2.6% [6], most smallholders sufer from food insecurity and poverty due to excessive reliance on traditional, nature-dependent, and low-productive agriculture. Eradicating extreme poverty without adequately addressing land degradation is highly unlikely [7].
Te Ethiopian government has carried out various land management practices on cultivated land, including contour (level) bunds, hillside terraces with aforestation, and hillside closures over the past fve decades to restore land productivity [8]. Sasakawa Global 2000, the Sustainable Land Management Programme, and other related projects have increasingly promoted soil and water management through conservation agriculture (CA) in Ethiopia [9]. CA is defned as soil management practices that minimize disturbance to the soil structure, composition, and biodiversity [10][11][12]. As used in this article, CA refers to soil management combined with zero-till, plants mulched using an indigenous technique, and intercropping practices on the undisturbed soil within the partitioned land. Te crop is planted directly into the preceding year's crop residue on the soil surface for the entire season. In Nepal, the Terai plains lie at the lowest altitude (<1,000 m.a.s.l.) and support 20% of the agricultural land, which is conducive to growing up to three crops, rice (Oryza sativa L.)-wheat (Triticum aestivum L.)-rice, ricewheat, and rice-maize (Zea mays L.), a year if irrigation facilities are present [13].
In Ethiopia, many smallholders use conventional tillage (CT) since the decision to adopt CA is infuenced by socioeconomic, institutional, and biophysical factors [9,14]. Te formal education of household heads, family size, age, frequent contact with development agents, and smallholders' average income have all been major infuencing factors for practicing CA [15]. In addition, livestock ownership has an impact on CA, as farmers with easy access to traction animals are less inclined to switch to minimum tillage [16]. Furthermore, the massive demand for animal fodder may reduce their interest in practicing CA [17].
Derashe and Arba Minch Zuriya are neighboring districts with contrasting tillage and soil management practices; however, both dominantly cultivate maize (Zea mays) and sorghum (Sorghum bicolor) in common. In the early years, Derashe district residents sufered from soil erosion, rainwater shortages, and soil moisture stress [18]. As a result of erratic rainfall as well as smallholders' traditional tillage practices lacking appropriate soil management practices, the situation became critical. Tese factors led to the invention of CA, using an indigenous moisture conservation technique known as Targa-na-Potayta. In this technique, own farm input is used for mulching and decomposed crop residue is used as a nutrient supplement [18][19][20]. In this land management system, only the cob of maize or the head of the sorghum harvest is taken and leaves the stalk in the feld to minimize nutrient mining from biomass removal. Uprooted maize/sorghum stalks layered in such a way serve as moisture-conserving mulch from which nutrients can be released for the next crop. After practicing the technology, the problem of soil moisture stress and soil erosion was signifcantly reduced.
Smallholders in the neighboring Arba Minch Zuriya district practice CT using mineral fertilizer and overirrigating their felds with fooding [21]. Even though chemical fertilizer use has enhanced crop productivity and soil fertility, excessive use has resulted in a decline in the soil organic matter (SOM) content and a consequent decline in soil quality [22]. Besides, in the Arba Minch area, overirrigation using lake water leaches nutrients from the soil. Te state has increased the rate of salinization and thus removed more banana felds from production [21]. Despite all the benefts of CA, smallholders have had a low adoption rate in the Arba Minch Zuriya district.
A large body of research has focused on the importance of CA for soil fertility from own farm input use, sustainable crop productivity, and household income [17,23,24]. Smallholders' CA practice decisions are less studied than their acquired benefts, and there are few universal factors that can explain their decisions [25]. Many studies attribute CA acceptance to several empirically derived farmer characteristics without considering the farmers' decision process (rationally weighing the economic constraints, the information difusion process, and the utility of adopting the new technology). Terefore, the current study attempted to understand factors that limit CA adoption decisions among the neighboring Derashe and Arba Minch Zuriya districts.

Theories of CA Practicing Decisions
Te current study focuses on two paradigms or conceptual models employed to explain the decision of smallholders to accept and practice novel technology: (i) the innovation difusion model and (ii) the economic constraints model. Te innovation difusion model, also called transfer of technology (TOT), follows the initial work of Rogers [26].
Innovation difusion occurs when innovation is communicated over time through specifc channels among the members of a social system [26]. Tis model sees change as a linear process. In this model, innovations generated by smallholders and/or agricultural research are passed down to farmers through extension agents and/or respected smallholders (these are farmers who implement a given technology and become far more proftable with comparing the majority; in most areas, they are named as model farmers) in the area. Tus, extension agents and/or respected smallholders act as modes of communication, and most farmers are recipients of innovation. Te difusion of innovations is composed of four components: (i) innovation, (ii) the communication channel, (iii) time, and (iv) a social system. However, the model's weakness is that it defnes individuals according to their behavior without considering factors that infuence their behavior. In addition, the assumption is that innovative technology should be used unless it is hindered by the lack of efective communication [27].
Te economic constraints paradigm postulates that farmers aim to maximize utility and that uneven resource endowments lead to observed adoption patterns of novel technology [27,28]. Te model emphasizes the role of economic factors at the individual level in determining practicing decisions, and the lack of access to these resources could afect practicing the technology [24]. However, this model allows only strictly rational and informed behavior and fails to capture the efects of individual perceptions of innovation. Tis perception is determined by personal characteristics (human values, education, and experience), physical factors of the land, and institutional factors, e.g., raising awareness through extension [29].
Overall, the technology-practicing decision illustrates that the practice of innovation encompasses a set of phases or levels, such as cognitive, normative, and action-oriented, that govern it. Recognizing these levels will aid in understanding the prerequisites for successive steps in the conservation agriculture technology practicing decisions in Derashe and Arbba Minch Zuriya districts ( Figure 1).

Conceptual Framework of the Study.
Practicing CA technologies is viewed as a variable representing farmers 'behavioral changes in accepting technologies [31]. However, the uptake of the latest technology is a process and can be infuenced by socioeconomic, institutional, and physical factors [14,15,34]. Characteristics of the household head such as age, formal education level, farming experience, total land holding, annual income from farming activities, and total livestock holdings, farming experience, annual net income, lack of access to small-scale irrigation, and extension service are selected variables in the current study expected to afect CA practicing decisions ( Figure 2).

Independent Variables and Teir Expected Outcomes.
Te independent variables that were used in the study are given in Table 1. It shows the independent (explanatory) variables, their description, and the expected outcome of the dependent variable.

Description of the Study Sites.
Te study was carried out in the Arba Minch Zuriya and Derashe districts in the southwestern region of Ethiopia ( Figure 3). Arba Minch Zuriya is one of the 14 Gamo Zone districts with 18 Kebeles (Kebele is the smallest administrative unit of Ethiopia, similar to a neighborhood or a localized and delimited group of people). It is about 505 km south of Addis Ababa, the capital city of Ethiopia. Te district lies between 37°23′51.37″ Eastern longitude and 5°55′16.24″ Northern latitude. Its elevation ranges from 1250 to 2600 m.a.s.l. [34]. In the last ten years, the district received a total annual rainfall of 892 mm, with maximum and minimum temperatures of 30.4 and 15.6°C, respectively [21]. Among the total Kebeles, the study was conducted in two neighboring Kebeles, namely, Zeyse Eligo and Zeyse Wezeqa.
Derashe district is 525 km away from Addis Ababa and geographically located between 5°35′25″ Northern latitude and 37°12′41.00″ Eastern longitude (USGS, 2020). Te elevation ranges from 1250 to 2300 m.a.s.l. [34]. Te area receives a total annual rainfall of 952.1 mm, about 261.2 mm in the long rainy season (June-September) and about 413.8 mm in the short rainy season (March-May) [35]. Most parts of the Southern region are considered short-rainy season (Belg (crops harvested between Megabit (March) and Nehase (August) are considered part of the Belg season crop [36])) growing areas with rainfall reaching its peak during April/May and daytime temperature near its maximum. Te Belg growing season is also marked by erratic rainfall patterns (in terms of amount, distribution, onset, and cessation). From the total of 19 Kebeles, Walessa and Holte Kebeles were purposively selected based on an intensive land management culture and their adjacency to the selected Kebeles of Arba Minch Zuriya district.
Agriculture is the mainstay of the community in the selected Kebeles. Te major types of crops grown in Holte and Walessa Kebeles are maize (Zea mays), sorghum (Sorghum bicolor), tef (Eragrostis tef ), and haricot beans (Phaseolus vulgaris) in integration with Moringa stenopetala trees in an agrisilviculture system. Te local communities of Derashe are hard-working people in Ethiopia where all the household members spend their days in the feld, ensuring they have successful crops [37]. Given their commitment to the harsh environment, they may be regarded as among the most capable managers of the arid ecosystem. In Zeyis Eligo and Zeyis Wezequa Kebeles of Arba Minch Zuriya district, maize, tef, sweet potato (Ipomoea batatas), haricot beans (Phaseolus vulgaris), cassava (Manihot esculenta), and vegetables mixed with mango (Mangifera indica), and banana fruit trees (Musa Sps.) are commonly practiced as fruit treebased agroforestry.

Description of Conservation Agriculture in the Study Area.
Derashe is located in southwest Ethiopia with unpredictable and inadequate rainfall; hence, crop production was threatened by chronic soil moisture stress [18]. Short rainwater quickly evaporates due to soil-water evaporation exacerbated by the high temperature at low altitudes. Runof coming from the nearby higher altitude sloping and undulating terrain also causes soil erosion, reduces soil fertility and productivity, and increases the risk of drought and food insecurity for a long time. To relieve the problem, hundred plus years ago, after many trials, they invented a novel soil moisture conservation structure using crop residues as mulch, locally known as Targa-na-Potayta ( Figure 4).
Targa-na-Potayta is a mulch arrangement technique where maize and sorghum residues are mulched in the feld in a rectangle shape, roughly 60 cm × 80 cm, with continuing patterns [37]. It is arranged so that the more extended parallels are known as "Targa" and the shorter transverse residue stems form a rectangular partition, "Potayta." One large Targa-na-Potayta can use 0.17 ha of land and is known as "Apha." Tey separate each "Apha" using a ditch/trench, and a total of 6 Apha can be found in one hectare of land. When water ponds on the surface of one Potayta follow a high rainfall, they open the rectangular crop residue and pass it to the next Potayta, and if one Apha (consisting of one long structure of Targa-na-Potayta) is ponded, then they pass it into the trench/ ditch that connects the next Apha ( Figure 4).
Te structure helps to (i) Improve rainwater use efciency (ii) Reduce soil erosion and improve soil fertility (iii) Reduce soil compaction (iv) Moderate soil temperature (v) Give higher and stable yields (vi) Save inputs and reduce cost of cultivation [17] Farmers use a metal stick called a "Totale" to plant by counting their steps with minimal tillage. Weeds are continuously controlled, cleared by hoeing, and used as residue. Tey do not allow a single weed in their crops.

Sampling Technique.
Derashe and Arba Minch Zuriya districts are neighboring districts with distinct cultures and languages and fall under diferent administration zones. In all Kebeles of Derashe district, Moringa stenopetala trees are cultivated with maize, sorghum, and tef with various intercropping intensities, and crop rotation is widely practiced. Te Derashe people use the leaves of the M. stenopetala tree as a staple and nutritious vegetable in the semiarid rift valley, adding to the food system for the dietary requirements of children and mature people. Te leaves are harvested every afternoon, added to the three meals, and mixed with sorghum or maize four when cooking the local food"kurkufa." Meanwhile, the residents of the bordering Arba Minch Zuriya district practice contrasting farming systems where banana and maize crops are integrated into separate plots but located close to each other using CT. Te nearby Wezeqa River partly irrigates this cropping system. However, moisture stress also occurs in the dry season when all farmers struggle to irrigate their crop felds simultaneously. Despite the moisture stress during the dry spell, farmers in the Arba Minch Zuriya district do not tend to learn and practice the adjacent Derashe district's moisture conservation and nutrient management skills. In the current study, consultation with the district agricultural ofce experts was made before selecting appropriate Kebeles.
Using a structured household survey, the study collected production and income data for signifcant perennial and annual crops for three consecutive cultivation seasons. Te sampling procedure was performed using a multistage sampling technique. Te study populations were all the households in each study site that cultivate banana + maize/ tef in Arba Minch Zuriya and Moringa stenopetala + maize/  tef/sorghum in Derashe districts. We calculated the total household of the four Kebeles [38] and developed an equation to yield a representative sample for proportions: Tis is valid, where n 0 is the sample size, Z 2 is the abscissa of the normal curve that cuts of an area α at the tails (1 − α) equals the desired confdence level, e is the desired level of precession, p is the estimated proportion of an attribute that is present in the population, and q is (1 − p). Based on the formula, the calculated minimum sample size was 384: When the population size is known (fnite), the formula is as follows: where n 0 is Cochran's sample size computed using the formula for the ideal sample size and N is the size of the population.
n � 384 Terefore, the data were collected from 392 proportionally allocated household heads (HHHs) in the four districts stratifed based on tillage practices.

Methods of Data Collection.
Te primary data were collected using structured questionnaires administered to farm households, key informant interviews, and feld observations. An organized questionnaire was developed to capture information about farmers' socioeconomic, physical, and institutional characteristics. Four key informant interviews were conducted, one in each study Kebele, with carefully selected elderly farmers who have relevant experience and opinions about the topic. Field observations were also used to cross-check the information gathered by the questionnaire survey and key informant interviews. Field observations and informal discussions with farmers and key informants were continuous processes during the entire duration of the research.

Methods of Data Analysis.
A logistic regression model was applied to analyze the determinants of farmers' CA practice decisions in the two study districts. Te model helps to estimate the probability of farm households practicing CA using various independent variables. STATA version 14 was used to analyze the collected data from all surveyed households.

Econometrics Model
(1) Determinants of Farmers' Decision to Practice CA. A binary logistic regression model was applied to answer the factors that infuence CA practice using the Targa-na-Potayta mulching technique. Te model describes the relationship between a dependent variable and a set of independent variables. Te dependent variable was binary or dichotomous and had only two groups: practicing and nonpracticing, whereas the explanatory variables could be continuous, categorical, or dummy [39]. Te logistic function was selected since it approximates the cumulative normal distribution [40]. Most technology-practicing studies have used the logistic regression model to analyze dichotomous practicing decisions in which the dependent variable is binary: 1, if the household head is practicing; otherwise, it is 0 [41]. As a result, the probability of practicing CA is as follows: Applied and Environmental Soil Science  Applied and Environmental Soil Science where, in the notation, P i represents the probability that an individual will make a certain choice in this study, whether the i th farmer practice CA or not, e denotes the base of natural logarithms which is approximated at 2.718, and Z i is a function of explanatory/independent variables (X i ) and expressed as follows: If P i is the probability of the i th farmer to adopt CA, given by (equation (5)), then ( − P i ) is the probability of the i th farmer to not adopt the traditional CA: Dividing (5) by (7), we obtain where P i /1 − P i is simply the odds ratio (likelihood) in favor of the i th farmer to practice to the probability to not practice it.
Taking the natural logarithm of the odds ratio in both sides of (8) will result in what is known as the logit model as indicated below: Te model included all responses (392 from each site, including CA adopters and nonadopters). Te explanatory variables (X i ) included in the model were the age of the household head (AGE), level of formal education (EDU), family size (FAMSIZE), size of total landholding in hectares (TLANHOLD), size of livestock owned in tropical livestock unit (TLU), farming experience (EXPER), net income from annual + perennial crops (NETINCO), provided extension service by development agents (EXTE), and lack of access to small-scale irrigation (IRRI). Logistic analysis was conducted to determine if the farmer practices CA if practicing group 1 and 0 otherwise. Given the above explanatory variables, the general form of equation (10) was rewritten to represent the likelihood of practicing CA in the two study districts: In this study, the above econometric model was used to identify factors that infuence the decision to practice CA.

Household and Socioeconomic Characteristics
4.1.1. Household Characteristics. Among the total sampled respondents, 97.45 percent were male and 2.55 percent were female household heads, of which 96.7 percent were married. In Ethiopia, men are the head of the house and considered representatives of their families. With an average age of 45, CA adopters and nonadopters were each 44.8 and 45.2 years old, respectively (Table 1). CA practicing and nonpracticing groups have a formal educational level of grade 6. Furthermore, 3.32% of CA and 15.81% of CT participants completed high school. However, studies have shown that higher education levels increase the chances of practicing CA because educated farmers are more likely to ease understanding and be receptive to advancing technology or innovations [42]. Even though the average age, farming experience, and formal education of CT practitioners were substantial, they still used traditional tillage by maresha to till their land. Tis indicates the need for more understanding/perception about the comparative benefts of CA among CT practicing groups.
Te average family size of CA adopters and nonadopters was 6.88 and 6.94 per household (Table 1). Study fndings by Sileshi et al. [43] suggest that having a bigger family size increases smallholders' interest in practicing soil and water conservation technologies because they gain additional labor. Households with more members are more likely to adopt it since larger families have more labor available for farm operations, such as weeding, which is critical in CA [24]. However, this assumption only sometimes holds since labor availability depends on how large a household is and the ages and types of persons in that household. In this study, though the total household size was higher for the CT practicing groups, the proportion of economically active members was slightly higher for CA practicing (0.67) compared to CT (0.66). In the Arba Minch Zuriya district, most banana fruit producers use something other than family labor. Instead, they recruit laborers during major agricultural activities because they get massive income from banana fruit marketing. Te lack of social, economic, and environmental benefts of CA-trapped CT practicing smallholders in traditional tillage.
Te total land holding (which consists of homestead land + land for woodlot + main farmland + grazing land) of CA adopters was higher than that of CT practicing groups. However, CT adopters' main farmland (the land use to cultivate Moringa stenopetala + sorghum/maize/tef in Derashe and banana fruit tree + maize/tef in Arba Minch Zuriya district) was signifcantly enlarged compared to CA. Ntshangase et al. [24] found that farmers with larger pieces of land were less likely to adopt CA technology than those with smaller pieces. Te CTgroups in the Arba Minch Zuriya district planted bananas on a large plot of land. Even some farmers extend their land to the coast of Lake Chamo. CT 8 Applied and Environmental Soil Science practicing groups had a mean higher net income of 157,980.60 ETB compared to 49,672.86 ETB from annual-+ perennial crops (Table 2). Tis was due to the huge income from banana fruit production and marketing. It was assumed that a higher income meant the farmer could buy farming inputs and engage in CA. Despite this income, most groups that practice CT continue to practice traditional tillage. Tis is probably because they needed to fully understand the benefts of CA and/or perceived it as exhausting and difcult. Of the total sampled households, 77.4 percent was native to the area and the rest, 22.6 percent, was not native ( Table 2). Being native to a particular area increases smallholders' interest in trying novel technologies.

Determinant Factors for Practicing Conservation
Agriculture. Seven continuous and two dummy explanatory variables were identifed to explain factors infuencing practicing CA in Arba Minch Zuriya and Derashe districts. Te efects of the independent variables on the log odds of practicing CA are reported as the odds ratio alongside the parameter estimates (β). For an independent variable, the odds ratio (e β ) represents the amount by which the odds favoring the decision to adopt CA (adopter � 1) change for a change in that independent variable (Table 3).
(1) AGE: Te direction of infuence of the age variable for practicing CA was predicted to be positive. However, as shown in the model summary table, age infuences the practicing CA negatively and significantly at a 5% probability level. Te negative coefcient and the odds ratio indicated that other factors held constant; the likelihood of a household in favor of practicing CA decreased by a factor of 0.89 for an increase in age by one year. Te endpoint of a 95% confdence interval (CI) of the odds ratio is (0.81, 0.99), suggesting that CA adoption could be decreased as be as little as 0.81 times to as much as 0.99 times as an increase in age by one year. Te negative relation of age with practicing CA was due to the considerable labor and energy requirement in structuring Targa-na-Potayta. Similarly, older farmers want to continue their traditional farming practices rather than try more modern and latest technologies. However, young farmers could be more inclined to try innovations with a lower risk averse and a longer planning horizon. Tis would justify investments in technologies whose benefts are realized over time. Te fnding was with [44], which showed that older farmers were accustomed to conventional farming methods and were unlikely to change. However, the current fnding was contrary to [24], as they revealed that practicing the new technology by older farmers may be attributable due to older access to better resources (for example, land ownership), coupled with experience and knowledge that had been gained over time compared to younger farmers.
(2) EDU: Te direction of infuence of the education variable on the adoption of CA was predicted to be positive. However, as shown in the model summary table, education was found to infuence CA adoption negatively and signifcantly at a 5% probability level. Based on the negative coefcients and odds ratios, the increase in the educational level by one year decreased the favor of adopting CA by 0.77, and the 95% confdence interval (CI) of the odds ratio is (0.65, 0.93). Tis suggests that practicing CA among smallholders could be as little as 0.65 times to as much as 0.93 times lower as the level of education increases by a year. However, some studies have shown that higher education levels increase the chances of practicing CA because educated farmers are more likely to quickly understand and be receptive to novel technology or innovations [24,25,42]. (3) TLU: Te total number of cattle in the tropical livestock unit was statistically signifcant at a 5% probability level and negatively associated with implementing CA. Keeping other factors constant, the odds ratio in favor of practicing CA decreased by 0.85 as livestock ownership increased by 1 TLU. Te negative relation of the livestock size with practicing CA was smallholders choosing to use crop residue as animal fodder, where benefts from livestock are realized over a short period. Jat Sahrawat et al. [17] stated that, in the tropics and subtropics, smallholders prioritize the use of crop residues as cattle fodder due to the high economic and cultural importance of livestock for smallholders. Also, ownership, particularly oxen ownership, reduced the practice of minimum tillage with mulching probably because easy access to traction animals makes it less attractive for farmers to switch to minimum tillage [16]. (4) NETINCO: As hypothesized, the net income from annual + perennial plants is signifcant at a 1% probability level and negatively afects practicing CA. Keeping other factors constant, the odds ratio indicated that, as the increase in the net annual income in one ETB decreases, the favor of adoption by a factor of 0.99, with the endpoints of a 95% confdence interval (CI) of the odds ratio is (−0.000037, −0.000016). Tis may be due to the increased net income by practicing CT causing farmers' to lack interest in CA since many smallholders desire shortterm returns overlooking sustainable land management. However, as the productive potential and annual income from the farmland are reduced, farmers will be more likely to practice CA. Te current fnding was with Nyanga [44], as the most resource-rich farmers are less likely to practice CA since they did not see any need to practice it. Tis could be because the relatively wealthy farmers are comfortable with their production levels and do not Applied and Environmental Soil Science 9 see any need to allocate their resources toward CA. However, the fnding was against Marenya et al. [16], as poor farmers were less likely to practice minimum tillage with mulching as compared to more assets owned by smallholders.
(5) EXT: As hypothesized, getting extension service is signifcant at a 1% probability level and positively correlates with practicing CA. Keeping other factors constant, the odds ratio indicated that, as smallholders get information about CA, the favor of practicing will increase by a factor of 19.01, with endpoints of a 95% confdence interval (CI) of the odds ratio (1.86, 4.03). Tis is due to frequent outreach explaining the contributions of technology to sustainable soil fertility. Crop production will enhance farmers' perception of the benefts of CA and the decision to practice. According to [24,43], increasing contact with farmers can positively infuence CA practice. A study on socioeconomic and institutional factors infuencing the practice of CA in Zimbabwe also revealed a positive and signifcant relationship between extension services in the practice of CA [45].
(6) IRRIGA: As hypothesized, the lack of access to smallscale irrigation was signifcant at a 1% probability level and positively correlated with adopting CA.
Keeping other factors constant, the odds ratio indicated that, as smallholders lack access to smallscale irrigations, the favor of adopting CA increased by 144.43, with endpoints of a 95% confdence interval (CI) of the odds ratio (3.4, 6.54). As a result of a lack of small-scale irrigations or poor moisture conditions, CA adopters as innovators used an indigenous mulching technique known as Targa-na-Potayta, invented almost a century ago. However, CT practitioners in the Arba Mich Zuriya district have easy access to small-scale irrigation systems. Due to this, smallholders in the area are reluctant to practice CA with indigenous mulching, Targa-na-Potayta. Teir CA practicing decision would likely increase if they did not have access to an irrigable water source. During the key informant interview, the Derashe area elderly smallholders stated as follows: "the problem of short and erratic rainfall coupled with lack of irrigable water was a critical challenge for the survival of Derashe people for many years. Using our progenitors' Targa-na-Potayta technology, which they invented with dedication and a creative mind, greatly reduced the demand for rainwater. Te current generation also appreciates them since the technology enhances soil moisture and increases soil fertility and crop productivity." Tese disclosed smallholders of the Derashe area use  the Targa-na-Potayta mulching technique that inherently innovated to relieve the problem of soil moisture stress.
Smallholders' new ideas, practices, and technology choices were theorized to be infuenced by the innovation difusion process and/or economic constraints. Referring to various studies, three socioeconomic, institutional, and physical factors were identifed, and nine explanatory variables were selected as factors that infuence CA adoption decisions. Te results showed that age, year of schooling, total livestock holdings, and net per annum income from annual + perennial crops afects CT practitioners' decision to adopt CA. Furthermore, increased extension agent visits and a lack of access to small-scale irrigation forced CA groups to accept the technology from their progenitors and apply it extensively.
In the innovation difusion model, Rogers [46] emphasized four key variables, namely, innovation, time, communication channels, and the social system. CA using the Targa-na-Potayta mulching technique is a unique feature in the Derashe area. Smallholders in the Derashe area adopted this innovation because of its benefts. However, those in the adjacent Arba Minch Zuriya district were not interested in adopting it, even though they were aware of it. Rogers [46] defned innovation as an idea/practice/object perceived as a novel by an individual or other unit of a working group. Terefore, CA with Targa-na-Potayta is considered an innovation for the Arba Minch Zuriya district smallholders, since apart from hearing and observing it, they did not practice it on their farmland. CA practice by the neighboring Arba Minch Zuriya district smallholders is slow, and they are still in the frst stage of the practicing process, i.e., the awareness stage. According to this study, time alone is not enough to make people accept and embrace novel practices; however, extension agents must play a key role in communication channels. Because their progenitors faced an outwardly overwhelming problem, the Derashe smallholders understood the benefts of CA. Again, families and extension agents also help smallholders practice CA as intensively as their ancestors did.
Rogers [46] also stated the efectiveness of innovation when difused through a given social system considering their cultural similarities. It is generally accepted that innovation difusion has a strong location-specifc component, with geographical proximity and spatial accessibility as important factors [47]. However, people primarily practice innovation based on subjective values and social norms difused through interpersonal networks [26]. Innovation can easily be difused to diferent areas if the technology is considered appropriate by a certain society. According to Assefa and Gezahegn [48], one of the ways that farm-level productivity can be increased is through the introduction of technology from other areas and transferring it to the farmers. It is possible if and only if practicing groups get appropriate information and the risk-taking behavior of farmers is known in advance. Furthermore, Lanckriet et al. [49] stated in their study on CA that spatial factors are relevant and interpersonal relations play a role. As indicated in their study fndings, social networks between diferent villages play an essential role in innovation, and innovations are derived from kinship, friendship, and marriage, and 39.2% of the CT practicing groups cited these factors as reasons for their knowledge of CA.
Most respondents in the Arba Minch Zuriya district stated that designing/implementing Targa-na-Potayta in the plot seemed tedious. However, if there was a problem with access to small-scale irrigation, they indicated their interest in practicing CA. Despite the importance of a social system for easing innovation difusion, an extension system can scale out innovation outside of a social network. Agricultural research/smallholders are considered the source of innovation, extension agents act as modes of communication, and farmers are recipients of the invention. Farmers' rationality is judged either by adopting or rejecting innovations that are seen as the outcome of an innovation-decision process [50]. In the Arba Minch Zuriya district, the lack of scientifcally supported investigation on the importance of CA with Targa-na-Potayta makes extension agents hesitant to understand and difuse the neighboring indigenous practice throughout the community. Tis slows the difusion process and keeps farmers trapped in CT.
Te study fndings showed that CT practicing groups have higher age, schooling, and net per annum income than CA practicing groups. Te economic constraints model assumes that individuals strive for proft or utility maximization, but observed adoption patterns are determined by the asymmetrical distribution of resource endowments among individuals [27,28]. A farmer chooses to adopt a technology if they perceive increased utility from its adoption. Tat is, adoption will occur if U1 > U0, where U1 is the utility if technology is adopted and U0 is the utility without adoption, and the economic constraints model recognizes the importance of proftability and fnancial constraints (access to capital, learning costs associated with innovation, or risk). Observing the comparative benefts, CT practicing groups obtained, on average, a higher net per annum income from sole banana fruit production than CA practicing groups' income from annual + perennial crops. Te high monthly income from banana fruit production makes smallholders inconsiderate of the sustainable benefts obtained from CA with Targa-na-Potayta mulching for soil fertility and productivity.

. Conclusions
Data from primary and secondary sources were used to assess CA practice decisions in southwestern Ethiopia, Derashe, and Arba Minch Zuriya districts. Te study was conducted in 4 adjacent Kebeles (2 from each district) that engage in diferent tillage and soil management activities. Te logistic regression model was used to understand the various determinant factors that limit smallholders' CA practice decisions. In the selected districts, nine independent variables were examined that were supposed to infuence CA practice decisions.
Te logistic regression fndings suggested that households with older age, education, total livestock holdings, and net per annual income were less likely to adopt CA. Increasing extension services and a lack of access to small-scale irrigation systems infuenced the decision to practice CA. One of the principal factors preventing Arba Minch Zuriya district smallholder farmers from adopting CA was receiving the most net income per year from banana fruit production. Te economic constraint assumes that any rational person intends to maximize proft (utility). Similarly, the primary objective of the smallholders in the area was to enhance income using CT at the expense of the soil. Smallholders focused on the momentary benefts of banana fruit, overlooking sustainable productivity because of the high demand nationwide. Lack of an irrigable water source, soil moisture stress, rainwater shortage, and soil erosion make smallholders in the Dershe district practice CA more intensively with the Targa-na-Potayta mulching technique. In the area, the culture of practicing CA and a structured extension system inform smallholders about the importance of CA for sustainable soil and crop productivity. If the government pays careful attention to the importance of soil and water conservation practices, the probability of CA adoption by Arba Minch Zuriya district farmers will increase.

Data Availability
All the data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that they have no conficts of interest.