Determinants of farmers’ adoption of soil and water conservation practices in northeastern Ethiopia Authors: Dessale Wasie Published: 2026 ABSTRACT AbstractSoil erosion continues to undermine the agricultural productivity and rural livelihoods in the Ethiopian highlands, yet the determinants of sustained adoption of soil and water conservation (SWC) practices remain insufficiently understood. This study analyzed the intensity and determinants of farmers’ adoption of SWC practices in the Golina watershed of Northeastern Ethiopia. A cross-sectional survey of 364 households was conducted using a three-stage sampling technique, complemented by field observations, focus group discussions, and key informant interviews. Descriptive statistics and a Poisson regression model were employed to examine factors influencing the number of SWC practices adopted per household. The results show that 99.2% of farmers had previously participated in SWC activities, while 66.8% were currently maintaining at least one SWC practice. Contour farming (100% in Woleh; 95.7% in Wofela) and soil bunds (90.4% in Wondie Ashewa) were the most widely adopted measures. The regression analysis indicated that farming experience, total income, membership in local organizations, land slope, knowledge of SWC, and positive perception of SWC benefits significantly increased adoption intensity, while greater land distance and better land quality reduced it. Overall, adoption intensity is strongly shaped by socioeconomic conditions, institutional engagement, land characteristics, and farmers’ perceptions. The findings highlight the need to strengthen knowledge dissemination, enhance community-based institutions, and promote context-specific interventions to improve the sustainability and intensity of SWC adoption for sustainable land management and rural livelihood improvement.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-63287-4.Keywords: Adoption intensity, Farmers’ perception, Golina watershed, Poisson regression, Soil and water conservation, Watershed managementSubject terms: Ecology, Ecology, Environmental sciences, Environmental social sciences, Environmental studies Introduction Land degradation in the form of soil erosion remains among the most pressing environmental and socioeconomic challenges affecting agricultural production worldwide, particularly in developing countries1. In Sub-Saharan Africa, rapid population growth, land scarcity, and limited access to conservation resources have intensified pressure on fragile ecosystems, resulting in accelerated soil degradation and declining farm productivity2. Ethiopia is one of the most severely affected by soil erosion, especially in its highland areas, where steep slopes, intense rainfall, and unsustainable land-use practices exacerbate the problem3. In Ethiopia, sheet and rill erosion by water are the dominant forms of land degradation, significantly reducing soil fertility and agricultural productivity4. Estimates indicate that average annual soil loss from cultivated land reaches approximately 42 tons ha⁻¹ year⁻¹, and exceeds 300 tons ha⁻¹ year⁻¹ on steep and poorly vegetated slopes4. These rates far surpass the commonly accepted tolerable soil loss threshold of 11 tons ha⁻¹ year⁻¹5. The consequences are both on-site, such as nutrient depletion and yield reduction, and off-site, including sedimentation of water bodies, damage to hydraulic infrastructure, and broader environmental degradation6. Nationally, the country is estimated to lose about 1.3 billion metric tons of fertile soil annually, resulting in substantial economic losses and contributing to food insecurity and rural poverty. The associated economic loss is estimated at $139 million, accounting for approximately 3–4% of the country’s agricultural GDP7. These losses have far-reaching consequences, including declining agricultural productivity, food insecurity, and rural poverty8. To combat these challenges, the Ethiopian government, in collaboration with development partners, has implemented extensive soil and water conservation (SWC) programs since the 1970s3. These initiatives have included physical measures such as soil bunds, terraces, check dams, cutoff drains, and waterways, often integrated with biological measures such as area closures and afforestation9. In recent years, nationwide mass mobilization campaigns have further promoted watershed-based conservation efforts aimed at enhancing sustainable land management, improving food security, and supporting socioeconomic development2. Despite these substantial investments, the success of SWC interventions has been uneven in many parts of Ethiopia. In many cases, conservation measures have been implemented through top-down approaches with limited community participation, leading to poor maintenance and low long-term sustainability2. In some areas, poorly designed or improperly maintained structures have even aggravated erosion problems. Moreover, many watershed projects phase out within a few years, raising concerns about the durability and continued adoption of introduced practices10. The Golina watershed in Northeastern Ethiopia represents a typical highland area severely affected by soil erosion and land degradation. In response, various integrated SWC practices, both physical and biological have been introduced by the government, research institutions, and development organizations11. While these interventions have been widely implemented, soil erosion continues to threaten agricultural productivity and rural livelihoods in the area. Evidence suggests that technical effectiveness alone does not guarantee sustainability; rather, farmers’ decisions to adopt, maintain, or abandon conservation practices are influenced by a range of socioeconomic, institutional, and perceptual factors. Previous studies in similar Ethiopian highland watersheds have predominantly focused on the biophysical performance and short-term impacts of SWC measures. However, limited attention has been given to the determinants of farmers’ adoption and the post-project sustainability of these practices at the watershed scale3. In particular, there is a lack of comprehensive analysis examining how household characteristics, access to resources, institutional support, and farmers’ perceptions interact to influence the intensity and continuity of SWC adoption. Understanding the determinants of farmers’ adoption of soil and water conservation practices is therefore crucial for designing effective, context-specific, and sustainable land management strategies. Without such evidence, future interventions risk repeating past shortcomings and failing to achieve lasting impacts. This study contributes to the existing literature in three important ways. First, it moves beyond evaluating the biophysical performance of SWC measures by examining the intensity and sustainability of farmers’ adoption behavior. Second, it integrates socioeconomic, institutional, biophysical, and attitudinal factors within a single analytical framework to explain adoption intensity. Third, it provides empirical evidence from the Golina watershed of Northeastern Ethiopia, where the long-term maintenance and sustainability of introduced SWC practices have received limited research attention. Therefore, the study offers valuable insights for designing more effective and context-specific conservation interventions in erosion-prone highland areas. Accordingly, this study aims to investigate the determinants of farmers’ adoption of soil and water conservation practices in the Golina watershed of Northeastern Ethiopia. Specifically, the study examines farmers’ perceptions and analyzes the socioeconomic and institutional factors influencing the intensity of adoption at the household level. Materials and methods Description of the study area The study was conducted in the Golina watershed, which is located in Raya Kobo Woreda, North Wollo Zone, Amhara Region, Ethiopia. Geographically, the watershed lies between 11° 55’0’’ − 12° 10’0’’ N latitude and 39° 25’0’’ − 39° 50’0’’ E longitude (Fig. 1). It covers an estimated area of 821.72 km² and is drained by three main rivers, Golina, Hormat, and Amidwuha, which converge at Megenagna and exit the watershed to the east. The Golina watershed is characterized by semi-arid climate conditions. Annual mean maximum temperatures range from 18 °C to 30 °C, while mean minimum temperatures vary between 5 °C and 15 °C. Annual precipitation ranges from 764 mm to 1092 mm, with spatial variability influenced by topography and elevation. Map of the Golina watershed, Northeastern Ethiopia, prepared by the authors using ArcGIS Desktop 10.8.2 (Esri Inc., Redlands, CA, USA; https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview). According to the Central Statistical Agency (2012), the national census, this district has a total population of 221,958, of which 111,605 are male, and 110,353 are female. A total of 54,466 households were counted in this district, resulting in an average of 4.08 persons per household. Types and sources of data The study was carried out through the use of both quantitative and qualitative data collection techniques. For this study, both Primary and secondary input data sources were used. Primary data were obtained through field observation, transect walks, direct measurements, household survey, focus group discussions (FGDs), and key informant interviews (KIIs). Secondary data, including socio-economic, demographic, vegetation, physiographic, soil, hydrologic, and farming systems, were collected from sources such as the Central Statistical Agency (CSA) of Ethiopia, Raya Kobo Woreda Administration, and Agricultural Development offices. Additional secondary sources, such as watershed base map, land use and land cover, soil, slope, development map, population, and other necessary information, were collected from the three kebele agricultural offices. Shapefiles of sub-watersheds and other necessary GIS and RS files were collected from the Raya Kobo Woreda office of agriculture, Natural Resources Management Case Team. Other secondary sources of information used for this study included published materials such as books, journals, annual reports, plans, official records, census records, project reports, research papers, and web pages. Sampling method and sample size determination To get the representative sub-watersheds and households, a three-stage sampling technique was employed. The Golina watershed was purposively selected in the first stage due to its severe soil erosion problems, long history of soil and water conservation interventions, and unsatisfactory adoption levels. In the second stage, representative sub-watersheds were selected purposively from major watersheds based on erosion sensitivity, prior soil and water conservation (SWC) interventions, and project phase-out history. The selection was carried out in consultation with the Woreda natural resources management case team. In the third stage, sample households were selected from the chosen sub-watersheds. The total number of households residing in the selected sub-watershed areas was 4,085. From the total households of the selected sub-watersheds, the size of the sample respondents that had to be interviewed for the data collection was determined by using Yamane’s formula (1967) as shown in the following Eq. (1). The reasons for deciding on this formula are the homogeneity of the households in their physical-environmental, social, and cultural aspects, as well as resource endowments. Where: nh = number of sample size; Nh = population size (total household); and e = the level of precision (5%, 7%, and 10%), but for this study margin of error is set at 5%. Based on the above sample size determination formula, a total of 364 household farmers were selected for the questionnaire data collection. Then, the figure was distributed to each sub-watershed proportionally by using the proportionality formula12. Where, nh= sample size of the stratum; Nh= total household head in each stratum; N=total population (total head of the household in the study area); and n = total sample size of the study. The total number of households and the sampled households for the study in each respective catchment are presented below in Table 1. Table 1The number of total HHs and the number of sampled HHs in each respective part of the watershed area.Sub watershedsTotal number of householdsNumber of sample householdsWofela1297116Wondie-Ashewa1170104Woleh1618144Total4,085364 The number of total HHs and the number of sampled HHs in each respective part of the watershed area. Respondents were selected using systematic random sampling techniques on lists of households obtained from the respective local agricultural offices. Thus, every 11th household was selected after randomly choosing the first household within the first 11 units. Because the household population was homogeneous in terms of the livelihood conservation technologies implemented, sample household respondents were selected randomly. In addition, fifteen key informants knowledgeable about local circumstances were purposefully selected. The identified elders (above the age of 50 years) who participated under different SWCs for many years, Kebele leaders, Woreda SWC experts, project coordinators, development agents (DAs), and female house heads were included as key informants. Similarly, 10 participants in the focus group discussion (FGD) were purposefully selected in each sub-watershed (3 FGD) because they were considered to provide adequate data on the issue under study. The selection of participants was carried out purposefully with elderly farmers, village leaders, and model farmers, based on experience and involvement in SWC practices. Methods of survey data collection The household survey, focus group discussion (FGD), key informant interview (KII), and field observation were employed to collect relevant information about the status of soil and water conservation practices and farmers’ perceptions of these measures in the study area. A formal household survey was undertaken with the selected sample household respondents to gather information on demographic and socioeconomic characteristics, types of SWC practices implemented, farmers’ perceptions, and general conditions of the study area. The process involved the preparation of questionnaires, recruitment, and training of field assistants, pre-testing questionnaires, refinement based on feedback, and actual data collection. A semi-structured questionnaire, containing both open and closed-ended questions, was prepared. It comprised three main sections. The first section of the questionnaire addresses the socioeconomic and demographic characteristics of respondents, the second section describes the types of soil and water conservation practices implemented, and the final section deals with respondents’ perceptions towards SWC practices (Appendix A). For easy understanding of the questionnaires of the respondents, the questionnaire was translated from English to ‘Amharic’, the local language of the study area. An in-depth interview was conducted because it allows for a comprehensive exploration of participants’ thoughts, experiences, and perspectives. To triangulate and supplement the survey data, FGDs and KIIs were conducted. These were used to enrich the information gathered through the questionnaire survey and to capture any details that may have been overlooked in the questionnaire. FGDs were used to clarify areas of ambiguity and gather detailed qualitative data on perceptions of soil and water conservation practices. A checklist with open-ended questions, prepared in Amharic (Appendix B), guided the discussions. Detailed notes were taken to capture participants’ views on the effectiveness of SWC measures in reducing soil erosion. Key informant (KI) interviews were conducted with knowledgeable individuals to collect additional information on soil erosion, historical trends, changes, and personal observations of SWC practices. Open-ended questions were used to allow informants to freely share their knowledge and experiences (Appendix C). Finally, field observations were conducted using a well-prepared observation checklist to validate information gathered from other sources, such as survey and interview data (Appendix D). Observations focused on the physical condition, design, and maintenance of SWC structures across the watershed. This provided an overall view of the various sites throughout the research process. Statistical analysis The collected data were systematically organized, summarized, tallied, and coded using Microsoft Excel 2021, and then analyzed using SPSS version 29. SPSS was selected over alternatives such as STATA because of its user-friendly interface, wide applicability in social science research, and strong support for both descriptive and inferential statistical procedures. In addition, SPSS provides integrated reliability testing tools (Cronbach’s alpha) and flexible handling of Likert-scale data, which were central to this study. The data gathered from the household survey were analyzed using the Econometric model and descriptive and inferential statistics. Descriptive statistics such as frequencies, mean, SD, percentages, and graphs were used for some socioeconomic variables. Inferential statistics, such as the t-test, One-way ANOVA, and chi-square test, were employed. The t-test was used to compare the technical quality and current status of SCW practices implemented across the subwatersheds. The one-way ANOVA was used to compare the mean differences of socioeconomic and demographic characteristics among sub-watersheds. The chi-square test was used to assess the association between categorical variables and the adoption rate of SWC practices. Farmers’ perceptions toward SWC practices were measured using a five-point Likert scale, with responses coded numerically as follows: 1 = Strongly Disagree, 2 = Disagree, 3 = Neutral, 4 = Agree, and 5 = Strongly Agree. Prior to analysis, the internal consistency and reliability of the Likert-scale statements were tested using Cronbach’s alpha coefficient, with values equal to or greater than 0.70 considered acceptable for reliability. The perception score for each respondent was computed by summing the scores of individual statements. In addition to mean score analysis, the Relative Importance Index (RII) was employed to rank perception statements and key factors influencing SWC adoption. The RII was calculated to provide a clearer and more objective prioritization of factors based on farmers’ responses, enabling comparison of the relative importance of different perceptions and constraints. Where: W = weight assigned to each response (1–5 based on Likert scale); A = highest possible weight (5 in this case); N = total number of respondents. This approach enhances interpretability beyond simple Likert-scale averages. Furthermore, an econometric (Poisson regression) model was used to identify factors influencing SWC adoption, integrating demographic, economic, institutional, physical, and attitudinal variables. Qualitative data from FGDs and KIIs were transcribed and analyzed thematically. Model specification In this study, the dependent variable was adoption intensity, defined as the total number of soil and water conservation (SWC) practices adopted and maintained by each household. The variable takes non-negative integer values (0, 1, 2, …, n), representing the extent of adoption rather than a simple adoption/non-adoption decision. Because the dependent variable is count data, a Poisson regression model was employed to identify factors influencing adoption intensity. The Poisson model is appropriate for analyzing discrete count outcomes and enables assessment of how household, institutional, physical, and attitudinal factors influence the number of SWC practices adopted. To analyze the determinants of adoption intensity, a Poisson regression model was employed. The Poisson model is particularly suitable for dependent variables measured as count data, such as the number of SWC practices adopted by a household. Unlike binary models (like probit or logit), which are restricted to yes/no outcomes, the Poisson regression allows for modeling discrete, non-negative integers and provides insights into the factors influencing the extent of adoption rather than just the decision to adopt13. The Poisson regression can be specified as follows: Where Pr(yi/xi) is the probability of farmer i adopting y techniques at a time and λi is the Poisson parameter for farmer i. Given that the expected number of SWC techniques adopted per period is E(y|xi), then the mean parameter as a function of the regressors, xi and a parameter vector β is given by: Where, The parameter λ is assumed to be log-linearly related to regressors xi. Therefore, The log-likelihood function is given by: Where y is the number of technologies adopted, β is a 1 × k vector of parameters, and x is ak × 1 vector with the values of k independent variables in the ith observation and n is thenumber of observations. To facilitate interpretation of the Poisson regression results, Incidence Rate Ratios (IRRs) were computed by exponentiating the estimated coefficients. The IRRs indicate the multiplicative effect of a one-unit change in an explanatory variable on the expected number of SWC practices adopted. Model adequacy was assessed using the log-likelihood value, likelihood ratio chi-square statistic, and pseudo-R². Robust standard errors were employed to reduce potential bias arising from heteroscedasticity. Description of variables and hypothesis The dependent variable in this study is the intensity of adoption, defined as the number of SWC practices adopted by each household. This variable takes values of 0, 1, 2, … up to the maximum number of practices observed. Farmers were considered adopters if they constructed and maintained at least one of the introduced SWC structures, either in its recommended form or with slight modifications, as defined by14. A range of explanatory variables hypothesized to influence adoption decisions were included, encompassing demographic, socioeconomic, institutional, physical, and attitudinal factors. The explanatory variables that are expected to influence the adoption of SWC, along with their units of measurement and expected effect, were summarized in Table 2. Table 2Definition of explanatory variables and their hypothesis.VariableDescriptionMeasurementExpected effectDependent variable SWC practicesCountNumber of SWCT adoptedExplanatory variables AgeContinuousYears+/- GenderDummy1, for males & 0, for female+/- Family sizeContinuousNumber+/- EducationCategorical0, Illiterate, 1, Informal Edu., 2, Elementary, & 3, Secondary school+ Dependency ratioContinuousThe ratio of non-working age to working age- Farming experienceContinuousYears+ Livestock ownershipContinuousIn TLU+/- Non-farm activitiesDummy1 if a farmer has off-farm and 0 otherwise+ Total incomeContinuousEthiopian birr+ TenureDummy1 if the farmer feels that the land belongs to him/her at least in his/her lifetime; 0 otherwise+ Extension contactContinuousNumber of contacts per year+ TrainingDummy1, for yes & 0, for no+ Credit UtilizationDummy1 if he/she accessed 0 otherwise+ MembershipDummy1 if a farmer is a member and 0 otherwise+ Land OwnershipDummy1 if a farmer owns and 0 otherwise+ Landholding sizeContinuousHectare+ Land distanceContinuousKilometre- Land qualityCategorical1, High, 2, Medium, &3, Low+ Land slopeCategorical1, Not slopy, 2, Steeply, 3, moderate, & 4, rugged+ Potential of SWCDummy1 if exists, 0 otherwise Erosion perceptionDummy1 if perceived as a problem, 0 otherwise+ Knowledge of SWCDummy1 if the respondent has awareness, 0 otherwise Perception on SWCDummy1, if the soil erosion problem is resolved & 0 otherwise+ Definition of explanatory variables and their hypothesis. The demographic factor includes the age, gender, education level, family size, working-age members, and dependency ratio of the household head. The socio-economic factor of the respondents includes farming experience, ownership (measured in Tropical Livestock Units), asset ownership, engagement in non-farm activities, and annual household income. The institutional factors include land tenure security, extension contact, access to training, credit availability, cooperative membership, and ownership rights. The physical factors include farm size, distance of farmland from homesteads, land quality, slope characteristics, and availability of SWC construction materials. Attitudinal factors include farmers’ awareness of soil erosion problems, knowledge about SWC practices, and their perceptions regarding the effectiveness of SWC practices. Categorical explanatory variables were carefully evaluated according to their measurement scales before inclusion in the regression model. Education level, land quality, and land slope were treated as ordinal variables because their categories represent ordered levels. Education was coded from illiterate to secondary school and above, land quality from low to high, and land slope from flat to steep/rugged conditions. Binary variables such as gender, training participation, membership in organizations, and land tenure were coded as dummy variables. Ethical approval and informed consent All methods used in this study were carried out in accordance with relevant ethical guidelines and regulations for research involving human participants. The study protocol was reviewed and approved by the appropriate institutional body of Woldia University. Prior to data collection, informed consent was obtained from all respondents involved in the household survey, focus group discussions, and key informant interviews after clearly explaining the purpose of the study. Participation was voluntary, and respondents were assured of the confidentiality and anonymity of the information they provided. Results and discussion Demographic and socio-economic characteristics of the respondents This study examined various demographic and socio-economic characteristics as crucial factors influencing farmers’ participation in SWC practices. These include age, gender, family size, education level, dependency ratio, farming experience, livestock ownership, participation in non-farm activities, total income, land tenure, access to extension services, training participation, credit availability, membership in organizations, land ownership, landholding size, distance to farmland, land quality, and land slope. Continuous variables One-way ANOVA results revealed statistically significant differences (p < 0.01) among sub-watersheds for key continuous variables, including age, farming experience, landholding size, and income levels (Table 3). These differences have direct implications for SWC adoption. For instance, farmers in Wondie Ashewa were older and more experienced, who often perceive the long-term benefits of SWC, and enhance awareness of land degradation and commitment to SWC practices implementation efforts. Table 3Demographic and Socioeconomic characteristics of sample households (continuous variables).VariablesSub-watersheds: Mean ± SDOverallStd. ErrorWolehWofelaWondie AshewaAge51.31 ± 8.8445.72 ± 9.6956.87 ± 15.9351.12 ± 12.30.101*Family size6.1 ± 1.214.84 ± 2.055.65 ± 1.695.57 ± 1.730.09*Dependency ratio0.246 ± 0.160.328 ± 0.210.252 ± 0.150.274 ± 0.180.00965**Farming experience28.3 ± 9.6821.18 ± 11.4835.61 ± 15.1328.12 ± 13.230.694*TLU1.99 ± 1.32.28 ± 1.683.08 ± 1.652.39 ± 1.590.08375*Total income16,597 ± 829297,153 ± 41,90152,121 ± 61,16150,934 ± 52,8252768.824*Landholding size2.3 ± 0.792.625 ± 1.174.98 ± 2.473.17 ± 1.930.10153*Distance to farmland0.74 ± 0.322.08 ± 0.72.375 ± 0.731.639 ± 0.940.04942**&**are significant at 1% & 5% level respectively. Demographic and Socioeconomic characteristics of sample households (continuous variables). *&**are significant at 1% & 5% level respectively. Prior studies confirm that age and farming experience positively influence conservation behavior, due to the accumulation of knowledge and land stewardship values15. Conversely, Wofela households were relatively younger with smaller landholdings, yet they reported higher income levels, suggesting stronger market integration and a higher capacity to invest in labor- and capital-intensive SWC measures, such as stone bunds, check dams, etc16. Family size showed significant differences; Woleh had the largest households and Wofela the smallest. While larger families can contribute more labor to SWC, they also experience higher consumption needs, which may lead to land scarcity. The dependency ratio was highest in Wofela, indicating a higher proportion of non-working members, thereby reducing the effective labor force available for conservation activities. This, combined with a smaller household size, may further constrain labor availability and must be considered in the planning and mobilization of community-based SWC programs17. The higher Tropical Livestock Units (TLU) and larger landholdings were observed in Wondie Ashewa. These may provide greater potential for organic soil improvement through manure use, allocating space for conservation structures, and diversified livelihoods. However, these advantages also come with risks, such as increased grazing pressure, which can exacerbate land degradation if unmanaged18. In contrast, higher income levels in Wofela may enhance the ability of households to invest in labor-intensive or capital-intensive SWC measures. Higher income levels enhance a household’s ability to invest in SWC technologies and structures, such as stone bunds or check dams16. On the other hand, economic limitations in Woleh may restrict such investment, despite other favorable conditions. Conversely, the distance from homesteads to farmland showed an opposite pattern. Woleh households had the shortest average distance (0.74 km), while in Wofela and Wondie Ashewa, average distances exceeded 2 km. Proximity to farmland is an important motivator for SWC adoption, as it allows for frequent monitoring and timely maintenance of conservation structures. This proximity may compensate for economic limitations by encouraging more consistent engagement in SWC practices15. Categorical and dummy variables Chi-square (χ²) test results indicated statistically significant variation in categorical socio-economic characteristics (Table 4). These differences may influence the adoption and implementation of SWC practices. Gender composition varies significantly across sub-watersheds, with male-headed households dominating, particularly in Woleh. While Wondie Ashewa showed a higher proportion of female-headed households, this reflects a relatively higher involvement of women in land management. These gender dynamics influence access to labor, resources, and decision-making power, with female-headed households often facing constraints that affect SWC participation14. Table 4Demographic and Socioeconomic characteristics of sample households (categorical and dummy variables).VariablesCategorySub-watersheds: Frequency (%)Overallχ2WolehWofelaWondie AshewaGenderMale130 (90.3)93 (80.2)66 (63.5)289 (79.4)26.6**Female14 (9.7)23 (19.8)38 (36.5)75 (20.6)Educational statusIlliterate72 (50.0)12 (10.3)26 (25.0)110 (30.2)97.5**Informal Edu.44 (30.6)41 (35.3)64 (61.5)149 (40.9)Elementary28 (19.4)47 (40.5)8 (7.7)83 (22.8)Secondary school0 (0.0)16 (13.8)6 (5.8)22 (6.0)Land tenureYes129 (89.6)104 (89.7)104 (100.0)337 (92.6)11.6*No15 (10.4)12 (10.3)0 (0.0)27 (7.4)Training participationYes144 (100)87 (75.0)98 (94.2)329 (90.3)254**No0 (0.0)29 (25.0)6 (5.8)35 (9.7)Credit availabilityYes0 (0.0)0 (0.0)0 (0.0)0 (0.0)No144 (100)116 (100)104 (100)144 (100)Membership in organizationsYes85 (59.0)93 (80.2)96 (92.3)274 (75.3)No59 (41.0)23 (19.8)8 (7.7)90 (24.7)Land ownershipYes129 (89.6)104 (89.7)104 (100.7)337 (92.6)11.6*No15 (10.4)12 (10.3)0 (0.0)27 (7.4)Land qualityHigh0 (0.0)0 (0.0)2 (1.9)2 (0.5)21**Medium129 (89.6)116 (100)99 (95.2)344 (94.5)Low15 (10.4)0 (0.0)3 (2.9)18 (5.0)Land slopeNot slopy15 (10.4)69 (59.5)54 (51.9)138 (37.9)104**Steeply29 (20.1)0 (0.0)3 (2.9)32 (8.8)Moderate100 (69.4)47 (40.5)44 (42.3)191 (52.5)Rugged0 (0.0)0 (0.0)3 (2.9)3 (0.8)Extension servicesYes144 (100)116 (100)104 (100)144 (100)No0 (0.0)0 (0.0)0 (0.0)0 (0.0)** & * is significant at 1% & 5% level respectively. Demographic and Socioeconomic characteristics of sample households (categorical and dummy variables). ** & * is significant at 1% & 5% level respectively. Educational status also shows statistically significant variation across sub-watersheds. Wofela had the highest proportion of respondents with elementary and secondary education, while Woleh had the highest illiteracy rate. Education enhances understanding of conservation benefits and increases willingness to adopt innovations, as literate farmers are more likely to access information, understand technical training, and implement new technologies19. Thus, farmers in Wofela may have a comparative advantage in knowledge-driven SWC interventions. There were statistically significant differences in land tenure security, with full ownership in Wondie Ashewa compared to slightly lower percentages in Woleh and Wofela. Land ownership is a key motivator for long-term land improvement investments. Secure tenure encourages households to implement sustainable practices with confidence in future benefits20. Participation in SWC-related training showed highly significant differences. All respondents in Woleh received training, compared to lower levels in Wofela. Access to training enhances farmers’ technical skills, awareness, and motivation, directly influencing the adoption and proper implementation of SWC technologies15. Credit availability was uniformly absent across all sub-watersheds. This is a critical constraint, as access to financial services can support investment in SWC infrastructure and labor. Meanwhile, membership in organizations was highest in Wondie Ashewa and the lowest in Woleh. Participation in cooperatives and local associations enhances collective action, access to shared knowledge, access to information, and external support, thus influencing SWC participation positively21. Land quality and land slope show statistically significant differences. Woleh had a higher proportion of steeper slopes and lower-quality land, while Wofela had flatter terrain. These physical characteristics affect erosion risk and determine the type and intensity of SWC practices required. Steeper slopes, for instance, demand labor-intensive measures such as stone bunds and hillside terraces15. Although all respondents reported access to extension services, the quality and frequency of these services may differ across sub-watersheds, influencing the overall outcomes of SWC implementation. Effective extension support is essential for the sustained adoption and maintenance of SWC practices22. Farmers’ perception of erosion and soil and water conservation practices Farmers’ perceptions of soil erosion and conservation practices provide critical insights into the local realities of land degradation and the effectiveness of interventions. These perceptions not only reflect environmental conditions but also shape the willingness and capacity of communities to adopt and sustain soil and water conservation (SWC) measures. Evidence from in-depth interviews and focus group discussions indicated that all respondents across the three sub-watersheds perceived soil erosion as a major problem on their farmlands and expressed a strong belief that erosion can be mitigated through appropriate conservation measures. Perception of soil erosion severity and trends The perceived severity of erosion varied across sub-watersheds (Table 5). The majority of farmers in Wondie Ashewa (72.1%) and Woleh (68.8%) perceived erosion severity as high, while Wofela showed relatively lower concern (57.8%). Overall, the majority of farmers (66.2%) perceived erosion as highly severe, where higher perceptions of severity reflect greater exposure to erosion risks on productivity and land quality. Such findings are consistent with Teshome et al. (2013), who emphasized that farmers’ perceptions of erosion are often closely linked to visible indicators such as gully formation, topsoil loss, and declining yields. These perceptions align with observed topographic, land-use conditions, and past erosion experience, where steeper slopes and deforestation exacerbate erosion risks23. Table 5Farmers’ perception of soil erosion (%).VariablesCategorySub-watershedsOverallWolehWofelaWondie AshewaSeverity level of soil erosionHigh68.857.872.166.2Moderate10.433.625.022.0Low20.88.62.911.8Soil erosion over timeIncreased52.886.287.573.4Decreased10.48.61.97.4Stayed the same36.85.2010.619.2 Farmers’ perception of soil erosion (%). Regarding erosion changes over time, the majority of farmers in Wondie Ashewa (87.5%) and Wofela (86.2%) reported that erosion has “Increased,” compared to 52.8% in Woleh. The overall responses reveal that 73.4% of farmers believe soil erosion has intensified over time, with only 7.4% perceiving a decrease. This trend highlights the pressing need for effective SWC measures to mitigate the impacts of increasing soil degradation, particularly in Wofela and Wondie Ashewa, where the perception of worsening conditions is most pronounced. A study by24 emphasizes that recognizing such farmer-reported trends is vital, as perception-driven responses can determine the extent of local participation and long-term sustainability of SWC. The higher perception of increased erosion in Woleh and Wondie Ashewa may be attributed to steeper slopes, deforestation, or land-use changes that have exacerbated soil loss. In contrast, the relatively lower perception of erosion intensification in Wofela suggests that conservation efforts or land-use stability might have played a role in maintaining soil conditions. The survey findings were supported by FGD participants, who consistently described soil erosion as a major threat to agricultural productivity. The Participant stated, “Every year we observe fertile soil being washed away during the rainy season, which reduces crop yield and increases the need for conservation measures.” Studies have shown that integrating local perceptions into conservation planning fosters greater community engagement and sustainability25. Perceptions of SWC effectiveness Farmers across the three sub-watersheds strongly recognized the effectiveness of SWC practices in controlling runoff, reducing erosion, conserving moisture, improving soil fertility, and enhancing land productivity (Table 6). In Woleh, all respondents strongly agreed that SWC practices effectively reduced surface runoff and soil erosion. Similarly, 58.6% of farmers in Wofela and 78.8% in Wondie Ashewa strongly agreed, with the rest agreeing. These findings align with studies by26, who observed that well-implemented SWC structures, such as terraces, bunds, and vegetative barriers, play a crucial role in minimizing soil loss and controlling water flow, ultimately preventing land degradation. Table 6Farmers’ perceptions of the effectiveness of SWC practices.ItemsSub-watershedsFrequency (%)MeanSDDNASAEffectively reduced surface runoff & soil erosion.Woleh0000144(100)4.26Wofela00048(41.4)68(58.6)Wondie Ashewa00022(21.2)82(78.8)SWCP contributes to moisture conservation.Woleh0000144(100)4.25Wofela00048(41.4)68(58.6)Wondie Ashewa00024(23.1)80(76.9)SWC practices improved the fertility of farmlandWoleh00088(61.1)56(38.9)4.56Wofela00048(41.4)68(58.6)Wondie Ashewa00024(23.1)80(76.9)SWC practices increased the productivity of landWoleh000115(79.9)29(20.1)4.49Wofela00048(41.4)68(58.6)Wondie Ashewa00024(23.1)80(76.9)SWC practices could be Sources of additional benefits.Woleh0000144(100)4.12Wofela00048(41.4)68(58.6)Wondie Ashewa009(8.7)26(25)69(66.3)SWC leads reduction of your scarce farm plotWoleh000100(69.4)44(30.6)4.29Wofela080(69)021(18.1)15(12.9)Wondie Ashewa087(83.7)026(25)69(66.3)Take more time and laborWoleh00086(59.7)58(40.3)4.3Wofela012(10.3)12(10.3)59(50.9)33(28.4)Wondie Ashewa03(2.9)6(5.8)26(25)69(66.3)Difficulty during farm implementation.Woleh0014(9.7)72(50.0)58(40.3)3.1Wofela0116(100)000Wondie Ashewa073(70.2)022(21.2)9(8.7)1 = SD-strongly disagree, 2 = D-disagree, 3 = N-neutral, 4 = A-agree, 5 = SA-strongly agree. Farmers’ perceptions of the effectiveness of SWC practices. 1 = SD-strongly disagree, 2 = D-disagree, 3 = N-neutral, 4 = A-agree, 5 = SA-strongly agree. SWC practices were widely perceived as effective in enhancing moisture conservation, with the unanimous agreement in Woleh (100% strongly agreed), followed by Wofela (58.6% strongly agreed) and Wondie Ashewa (76.9% strongly agreed). Such responses reflect the critical role of SWC practices to improve soil moisture retention by reducing runoff velocity and increasing infiltration, which is crucial for sustaining crop growth in semi-arid regions27. Perceptions regarding soil fertility improvement and land productivity were also positive. In Woleh, 38.9% strongly agreed and 61.1% agreed that SWC improved soil fertility, while in Wofela and Wondie Ashewa, over three-quarters strongly agreed. These results are consistent with the work of28, who found that SWC interventions significantly enhance organic matter retention and nutrient availability, leading to improved soil fertility and crop yields over time. Furthermore, 79.9% of farmers in Woleh agreed that SWC increased land productivity, while 58.6% in Wofela and 76.9% in Wondie Ashewa strongly agreed. These perceptions are consistent with the findings of21 who highlighted the significant impact of SWC measures on soil health and agricultural productivity. Farmers in Woleh unanimously (100%) perceived SWC practices as sources of additional benefits, reflecting strong local endorsement of conservation efforts. In Wofela, 58.6% strongly agreed, and 41.4% agreed; in Wondie Ashewa, 66.3% strongly agreed, 25% agreed, and 8.7% disagreed (D). This demonstrates the broader socioeconomic advantages of SWC measures, such as improved livelihoods and environmental sustainability29. Despite widespread recognition of SWC benefits, farmers identified several challenges associated with SWC implementation. One of the most significant concerns was the reduction of cultivable land due to SWC structures. In Woleh, 69.4% agreed and 30.6% strongly agreed that SWC structures reduce available farm plots. In contrast, 69% in Wofela and 83.7% in Wondie Ashewa disagreed, suggesting that land occupation concerns vary by location and farm size as land scarcity exacerbates farmers’ concerns about losing arable plots30. similarly observed that while SWC may initially reduce cultivable area, long-term gains in soil productivity often outweigh this loss. Another widely reported challenge was the high time and labor demand of SWC practices. SWC practices were perceived as labor-intensive, especially in Wondie Ashewa (66.3% strongly agreed) and Woleh (59.7% agreed, 40.3% strongly agreed). This finding aligns with26, who reported that labor-intensive SWC interventions can pose a barrier to adoption, especially for resource-limited farmers. The difficulty of implementing SWC structures on farmlands also varied across sub-watersheds. In Woleh, 50% of farmers agreed that SWC practices were challenging to implement, while in Wofela, 100% disagreed. Wondie Ashewa showed mixed responses, with 70.2% disagreeing. These differences may reflect terrain complexity, technical support availability, and prior experience with SWC structures29. emphasize that topography, soil type, and farmer experience influence the ease of SWC implementation. Overall, farmers across the three sub-watersheds overwhelmingly recognized the benefits of SWC practice, particularly in reducing surface runoff, conserving soil moisture, improving soil fertility, and enhancing land productivity. Despite these advantages, challenges persist, particularly concerns related to the labor and time required for implementation, the amount of land occupied by SWC structures, and difficulties encountered during the establishment of these measures were also evident, particularly in Wofela and Wondie Ashewa30. highlighted similar issues, emphasizing that the labor-intensive nature of SWC practices can hinder their adoption among smallholder farmers. The internal consistency of the items measuring farmers’ perceptions of soil and water conservation (SWC) practices was assessed using Cronbach’s Alpha reliability coefficient. The computed Cronbach’s Alpha value was 0.706. According to commonly accepted reliability thresholds, Cronbach’s Alpha values are interpreted as follows: excellent (α > 0.90), good (0.70 < α ≤ 0.90), acceptable (0.60 < α ≤ 0.70), poor (0.50 < α ≤ 0.60), and unacceptable (α < 0.50)31. Therefore, the obtained alpha value of 0.706 indicates good internal consistency among the perception items, suggesting that the variables reliably measure the underlying construct of farmers’ perceptions toward SWC practices. This confirms that the scale is sufficiently reliable for subsequent statistical analysis. Farmers across the Golina watershed demonstrated strong recognition of the benefits and effectiveness of soil and water conservation (SWC) practices. Using the Relative Importance Index (RII) to objectively rank perceptions across the study area, the most highly valued benefit of SWC was its effectiveness in reducing surface runoff and controlling soil erosion (RII = 0.962), followed by improvements in soil moisture conservation (RII = 0.960), which is critical for sustaining crop growth in this semi-arid region. Farmers also strongly perceived SWC practices as enhancing soil fertility (RII = 0.912, Rank 3) and increasing land productivity (RII = 0.897, Rank 4), indicating recognition of long-term agronomic and livelihood benefits. Additional advantages, including fodder, fuelwood, and broader environmental services, were acknowledged (RII = 0.87, Rank 5) (Table 7). Table 7Relative importance index (RII) and ranking of farmers’ perceptions of SWC practices.NoPerception statementRIIRank1SWCP had effectively reduced surface runoff & soil erosion.0.96212SWCP contribute for moisture conservation.0.96023SWC practices improved the fertility of farm land.0.91234SWC practices increased productivity of land.0.89745SWC practices could be Sources of additional benefits.0.87056SWC leads reduction of your scarce farm plot.0.60587Take more time and labor0.86268Difficulty during farm implementation.0.6217 Relative importance index (RII) and ranking of farmers’ perceptions of SWC practices. Key informants emphasized that SWC structures have contributed to improved soil moisture retention and reduced runoff. As one development agent explained, ‘Fields protected by terraces and bunds retain more moisture and remain productive even during periods of erratic rainfall. Despite these benefits, farmers identified several challenges: high labor and time requirements (RII = 0.862, Rank 6), reduction of cultivable land due to SWC structures (RII = 0.605, Rank 8), and difficulty during implementation (RII = 0.621, Rank 7). These findings align with previous studies emphasizing the multi-dimensional benefits and challenges of terraces, bunds, check dams, and vegetative barriers26,28,29. While the benefits of SWC strongly encourage adoption, constraints related to labor, land occupation, and technical difficulties remain critical barriers. Addressing these challenges through labor-efficient designs, enhanced technical support, and community-based planning is essential for increasing the intensity and sustainability of SWC adoption in the Golina watershed. FGD participants frequently highlighted labor shortages and loss of cultivable land as barriers to maintaining SWC structures. One farmer noted, ‘Constructing terraces requires substantial labor, and some farmers are reluctant because the structures occupy part of their farmland. Adoption status of soil and water conservation practices Farmers’ participation in soil conservation practices The adoption of soil and water conservation (SWC) practices by farmers is a critical component of sustainable land management. Results from the three sub-watersheds reveal high overall participation rates, though with differences in timing, type, and maintenance of practices (Table 8). Farmers across all sub-watersheds exhibited a high participation rate of SWC practices, with 99.2% of farmers reporting prior engagement in SWC activities. This aligns with studies by3,32, which highlight the increasing adoption of SWC measures in erosion-prone areas due to growing concerns over soil degradation and declining agricultural productivity. Similarly33, observed that farmers in Ethiopia increasingly adopt SWC practices due to government support and training programs. Table 8Farmers’ Participation in Soil Conservation Practices (%).VariablesCategorySub-watershedsOverallWolehWofelaWondie AshewaHave you ever practiced any SWC?Yes144(100)114(98.3)103(99)361(99.2)No02(1.7)1(1)3(0.8)when did you start taking these measures?< 5years00005-10years15(10.4)12(10.5)6(5.8)33(9.14)10-15years29(20.1)24(21.1)17(16.5)70(19.39)> 15years100(69.5)78(68.4)80(77.7)258(71.47)Which types of SWC measures have been implemented before?Traditional015(13.2)17(16.5)32(8.9)Introduced0000Both144(100)99(86.8)86(83.5)329(91.1)Are you constructing and maintains at least one of the introduced SWC technologies currently?Yes82(56.9)80(69)81(77.9)243(66.8)No62(43.1)36(31)23(22.1)121(33.2)Percentages for adoption-related questions were calculated based on households that had practiced SWC (n = 361). Farmers’ Participation in Soil Conservation Practices (%). Percentages for adoption-related questions were calculated based on households that had practiced SWC (n = 361). The timing of SWC adoption shows that the majority of farmers (71.5%) have been practicing SWC for more than 15 years, while 19.4% have done so for 10–15 years, and only 9.1% for 5–10 years. These indicate that the conservation efforts in the study area have a relatively long history, shaped by large-scale government and community mobilizations since the 1980s. Similarly3, emphasized that widespread and sustained adoption requires consistent promotion over decades rather than short-term interventions. When considering the type of SWC measures adopted, 91.1% of farmers practiced both traditional and introduced techniques, while 8.9% had only practiced traditional methods. These findings emphasize the importance of integrating traditional and introduced measures to enhance adoption rates. Similarly, Studies by34 indicate that the combination of traditional and modern SWC practices enhances conservation effectiveness. However, the proportion of farmers actively constructing and maintaining introduced SWC technologies was lower. Overall, 66.8% of farmers currently construct and maintain at least one introduced measure. Adoption was lowest in Woleh (56.9%), moderate in Wofela (69%), and highest in Wondie Ashewa (77.9%). These disparities suggest context-specific barriers such as land scarcity, high labor demand, economic constraints, post-conflict damage, or reduced institutional support18. similarly found that sustained adoption depends on perceived benefits and the availability of technical and material support. Farmers’ adoption level of soil and water conservation practices Adoption levels of specific SWC technologies vary widely across the three sub-watersheds (Table 9), reflecting differences in awareness, resource availability, and site suitability. Contour farming has the highest adoption rate across all sub-watersheds, with full adoption in Woleh (100%), 95.7% in Wofela, and 74.03% in Wondie Ashewa. Its wide acceptance is likely due to simplicity, low labor demand, and visible erosion control benefits33. confirm that contour plowing is one of the most rapidly adopted techniques in Ethiopia due to effective extension promotion. Table 9Farmers’ adoption level of SWC practices.TechnologySub-watershedsFrequency (%)Not knownKnown but not adoptedAdoptedContour farmingWoleh00144(100)Wofela05(4.3)111(95.7)Wondie Ashewa027(25.97)77(74.03)Mixed croppingWoleh074(51.4)70(48.6)Wofela028(24.14)88(75.86)Wondie Ashewa046(44.24)58(55.76)Crop rotationWoleh0118(81.95)26(18.05)Wofela045(38.8)71(61.2)Wondie Ashewa067(64.4)37(35.6)CompostWoleh0132(91.66)12(8.33)Wofela0110(94.8)6(5.2)Wondie Ashewa0100(96.1)4(3.9)Soil bundWoleh042(29.2)102(70.8)Wofela086(74.1)30(25.9)Wondie Ashewa010(9.6)94(90.4)Stone bundWoleh0130(90.3)14(9.7)Wofela0104(89.7)12(10.3)Wondie Ashewa097(93.3)7(6.7)Stone faced soil bundWoleh0104(72.3)40(27.7)Wofela059(50.9)57(49)Wondie Ashewa066(63.5)38(36.5)Hillside terracesWoleh044(30.6)100(69.4)Wofela082(70.7)34(29.3)Wondie Ashewa038(36.5)42(40.38)Check damWoleh0115(79.9)29(20.1)Wofela0110(94.8)6(5)Wondie Ashewa099(95.2)5(4.8)Cutoff drainsWoleh0115(79.9)29(20.1)Wofela063(54.4)53(45.6)Wondie Ashewa082(78.8)22(21.2)WaterwaysWoleh0116(100)0Wofela5(4.3)105(90.5)6(5.2)Wondie Ashewa0104(100)0Farm pondWoleh0144(100)0Wofela0116(100)0Wondie Ashewa0104(100)0HalfmoonWoleh0132(91.66)12(8.33)Wofela5(4.3)105(90.5)6(5.2)Wondie Ashewa1(1)33(31.7)70(67.3)Bund stabilizationWoleh0132(91.66)12(8.33)Wofela5(4.3)105(90.5)6(5.2)Wondie Ashewa1(1)33(31.7)70(67.3) Farmers’ adoption level of SWC practices. Mixed cropping adoption rates are moderate, with the highest uptake in Wofela (75.9%) compared to 48.6% in Woleh and 55.8% in Wondie Ashewa. This practice is valued for its dual benefits of crop diversification and soil erosion control, as noted by Shiferaw et al. (2009). Crop rotation adoption shows substantial variation, with very low adoption in Woleh (18.1%), moderate adoption in Wondie Ashewa (35.6%), and relatively high adoption in Wofela (61.2%). The low adoption in some areas likely reflects land shortages and perceptions that rotation reduces short-term returns35. Similarly noted that crop rotation requires both adequate landholding and awareness of long-term fertility benefits. Compost application is minimal, with only 8.33% in Woleh, 5.2% in Wofela, and 3.9% in Wondie Ashewa, indicating limited awareness and access to organic inputs, as discussed by3. Soil bunds and stone bunds also show an uneven adoption rate. Soil bund is highly adopted in Wondie Ashewa (90.4%), moderate in Woleh (70.8%), and lowest in Wofela (25.9%). While stone bunds show low adoption across all areas, reflecting their high labor and material requirements18. Stone-faced bunds and hillside terraces are moderately practiced, with adoption highest in Woleh. This aligns with36 findings that technical training is essential for adopting such labor-intensive practices. Other water harvesting and flood and drainage management practices, such as check dams, cutoff drains, waterways, and farm ponds, remain underutilized. These practices, despite their effectiveness, are often hindered by labor and material costs37. Adoption of half-moon structures and bund stabilization is higher in Wondie Ashewa (67.3%) compared to Woleh and Wofela, reflecting the effectiveness of integrating traditional and modern practices in degraded land stabilization, as noted by38. The adoption of SWC practices is influenced by factors such as labor demands, costs, and technical support. For instance, simpler, low-cost practices like contour farming and mixed cropping are widely adopted, while labor-intensive or capital-demanding technologies such as bunds, terraces, and farm ponds face adoption barriers. Limited awareness and technical capacity hinder the use of practices like composting and crop rotation. Factors influencing adoption intensity of SWC practices The Poisson regression analysis reveals significant socio-demographic, economic, institutional, and land-related factors influencing farmers’ adoption intensity of soil and water conservation (SWC) practices. The likelihood ratio (LR) chi-square value of 98.20 (p < 0.01) indicates that the model is statistically significant and fits the data well. The pseudo-R² value of 0.0737 is acceptable for cross-sectional count data models (Table 10). Table 10Summary results of the Poisson regression analysis of farmers’ adoption Intensity.Number of soil water conservationCoefficientStd. errorzP>|z|Age0.00908690.00712941.270.202Gender0.17615120.12004111.470.142Family size0.01966550.02277290.860.388Education0.07771890.04719891.65*0.100Dependency ratio-0.35194230.2186655-1.610.108Farming experience0.01358110.0066612.04**0.041Livestock ownership0.00917310.03238990.280.777Non-farm activities-0.0430.1235-0.1210.728Total income0.000160.000142.507**0.0113Tenure0.24658750.20168811.220.221Extension contacts0.01593450.01193871.330.182Training0.19208190.13350241.440.150Credit0.01566330.08776840.180.858Membership0.1838160.07986512.30**0.021Ownership0.27421470.25462761.080.282Landholding size0.05454550.03009321.81*0.070Land distance-0.13910260.0622651-2.23**0.025Land quality-0.47311410.1844615-2.56***0.010Land slope0.36626950.1361482.69***0.007Potential of SWC0.35998950.13124692.74***0.006Erosion perception0.53776470.4924661.090.275Knowledge of SWC0.39063790.11853643.30***0.001Perception of SWC0.70941540.25363352.80***0.005Constant2.5614291.0959832.34**0.019Number of Observations: 364Log-Likelihood: -616.89Likelihood Ratio (LR) Chi-Square: 98.20Prob> chi2 = 0.0000Pseudo R2 = 0.0737*** p < 0.01, ** p < 0.05, * p < 0.10 Summary results of the Poisson regression analysis of farmers’ adoption Intensity. Number of Observations: 364 Log-Likelihood: -616.89 Likelihood Ratio (LR) Chi-Square: 98.20 Prob> chi2 = 0.0000 Pseudo R2 = 0.0737 *** p < 0.01, ** p < 0.05, * p < 0.10 Socio-demographic variables Among socio-demographic variables, farming experience positively and significantly affects adoption intensity (β = 0.0136, p < 0.05). This suggests that farmers with longer farming experience are more likely to implement a greater number of SWC practices, possibly due to accumulated knowledge and greater exposure to erosion problems39. Education level shows a positive effect at the 10% significance level (β = 0.0777, p < 0.10), implying that more educated farmers tend to adopt a higher number of SWC practices. Similar findings were reported by40, who noted that education improves farmers’ capacity to adopt improved SWC structures in Wolaita Zone. However, age, gender, family size, and dependency ratio were not statistically significant, indicating that these variables do not independently explain variation in adoption intensity within the study area. This could reflect differences in labor allocation or priorities within households3. Economic Variables. Total household income positively and significantly influences adoption intensity (β = 0.00016, p < 0.05), suggesting that wealthier households have greater financial capacity to invest in conservation measures. These findings align with41, who argue that wealthier households have better financial capacity to invest in conservation technologies. Similarly, landholding size is positive and weakly significant (β = 0.0545, p < 0.10), implying that farmers with relatively larger land sizes tend to implement more SWC practices. Livestock ownership, non-farm activities, tenure status, and credit access were not statistically significant, indicating that economic diversification and asset ownership do not uniformly translate into higher conservation intensity in the study area. This contrasts with some studies20, which emphasizes the role of asset ownership in adoption decisions. Institutional and support factors Membership in local organizations is positively and significantly associated with adoption intensity (β = 0.1838, p < 0.05). This finding highlights the importance of social capital, collective action, and information exchange in promoting conservation practices3. However, extension contact and training were not statistically significant in explaining adoption intensity, suggesting that mere contact may not be sufficient unless it effectively enhances farmers’ knowledge and perception. According to20 argues that extension effectiveness depends on participatory approaches and context-specific training rather than frequency of contact. Land characteristics Land-related variables show strong and significant effects. Land slope has a positive and highly significant influence (β = 0.3663, p < 0.01), indicating that farmers cultivating steeper lands tend to exhibit higher SWC adoption intensity due to higher perceived erosion risk. Land quality, however, has a negative and significant effect (β = − 0.4731, p < 0.01), suggesting that farmers with relatively fertile or better-quality land are less likely to intensify SWC adoption, possibly because they perceive lower immediate erosion threats. This aligns with Nyssen et al. (2008), who found that farmers prioritize conservation on degraded lands rather than fertile plots. Land distance negatively and significantly affects adoption intensity (β = − 0.1391, p < 0.05), indicating that plots located farther from homesteads receive fewer conservation investments, likely due to monitoring and labor constraints. This consistent with40, who reported that plots farther from homesteads receive fewer conservation investments due to monitoring and labor constraints. Perception and knowledge variables Knowledge of SWC practices is positively and highly significantly associated with adoption intensity (β = 0.3906, p < 0.01), demonstrating that informed farmers are more likely to implement multiple conservation measures. Similarly, farmers’ perception of SWC benefits shows a strong positive and significant effect (β = 0.7094, p < 0.01). This variable has one of the largest coefficients in the model, indicating that favorable attitudes toward SWC substantially increase adoption intensity. The perceived potential or effectiveness of SWC practices also positively affects adoption (β = 0.3600, p < 0.01). In contrast, perception of erosion severity was not statistically significant, suggesting that awareness of erosion alone may not be sufficient unless accompanied by a positive evaluation of SWC effectiveness. These findings are consistent with the theory that awareness and perception of resource degradation drive conservation behaviors38. Similarly20, highlights that farmers’ perception of erosion severity alone is insufficient unless accompanied by confidence in SWC effectiveness. The findings provide important policy insights for sustainable land management interventions in the Ethiopian highlands. In particular, strengthening farmers’ knowledge, enhancing participation in local organizations, and targeting conservation support to vulnerable and erosion-prone areas may improve the long-term adoption and sustainability of SWC practices. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Author contributions Dessale: conceptualization, data collection, analysis, writing the original draft, review, and editing. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.