Determinants of adoption of land management practices among farmers in Western Lake Tana and Beles River watersheds (Ethiopia) as a climate change adaptation strategy

Abstract

This study analyzes farmers’ decisions to use land management adaptation practices in the face of climate change. It also looks at the socio-economic elements that influence adaptation practices. To collect primary data, a multistage and simple random sampling procedure was used to choose a sample of 338 farmers. The data was analyzed using a multivariate probit model. The results reveal that a farmer’s likelihood of adopting a specific land management adaptation measure is linked to and defined by a set of complementary adaptation measures. Grade bund terracing (97.1%), contour plowing (96.9%), adequate waterway (88.4%), compost preparation (84.3%), hedge planting (73.4%), and mulching (66.6%) were the most commonly chosen land management adaptation measures, followed by afforestation (61.2%), gulley rehabilitation (57.8%), woodlot implementation (55.5%), and area closure (55.4%). This shows that grade bund terracing and contour plowing are the most likely chosen land management adaptation practices while area closure is the less likely choice of practice. Land management adaptation strategies by farmers had a combined chance of success of 96 %, compared to a combined probability of failure. The implication is that farmers will utilize a combination of land management strategies to address climate change variables. The results also demonstrate that the sex, family size, farm experience, frequency of extension visits, and farmers’ level of education have a substantial impact on the common underlying socioeconomic component for choosing land management adaption techniques. As a result, developing a package of land management adaptation methods is critical for possible climate change treatments and strategies.

Keywords:

• climate change
• land management
• adaptation practices
• resilient agriculture
• Ethiopia

1. Background of the Study

Climate change, according to FAO (2008), is defined as long-term changes in average meteorological conditions. The type of shift in temperature, rainfall, and the incidence of extremes is referred to as climatic variability (UNFCCC, 2007). According to several authors, there has been broad acceptance and proof that global climate change is a reality (IPCC, 2007; World Bank, 2010). Nyboer et al. (2019) also concluded that the 20th century was most likely the warmest of the millennium, with an increase in the average global surface temperature of roughly 0.72°C to 0.85°C

According to Esham and Garforth (2013), farmers’ susceptibility to climate change and variability has been rising in underdeveloped and developing nations. Due to their low capacity for adaptation and lack of access to sources of income, adaptive skills, and resources, developing nations like South Asia and East Africa are particularly susceptible to the effects of climate change (heating trends and precipitation fluctuations; Fahad & Jing, 2018). Experts say, the record one killing crops and cattle across Ethiopia, Kenya, and Somalia have underscored the increasing frequency of drought due to climate change. High population growth rates, high and rising levels of poverty, low per capita incomes, high levels of inequality, and declining GDP growth rates, among other issues, are likely to exacerbate the effects (Mubiru et al., 2018).

Ethiopia is currently going through climate change. According to the UNDP, Ethiopia’s mean annual temperature increased by 1.3°C between 1960 and 2006, with a 0.28°C increase every decade on average (UNDP, 2010). Every ten years, Ethiopia’s average lowest temperature has risen by around 0.37°C, while seasonal mean temperatures have risen in several parts of the country. Since 1900, the highest daily temperature has risen by 1.5 degrees Celsius (NMA (National Meteorological Agency), 2007; IPCC, 2014a; Belay et al., 2013). Furthermore, rainfall distribution across the country was both temporally and spatially varied.

Since a large portion of Ethiopia’s economy depends on climate-sensitive agricultural production, the effects of climate change on farmers who depend on crop and livestock production have been severe and well-documented in the literature (Agrrarwal et al., 2019; Rahut & Ali, 2018), and the effect of climate change on food security is due to yield loss (Aniah et al., 2019; Zhao et al., 2018).

According to NMA (National Meteorological Agency; 2007), Ethiopia’s ability to adapt to climate change is hampered by the country’s poor socioeconomic development, inadequate infrastructure, high population growth, institutional capacity, and reliance on climate-sensitive activities at risk. As a result, climate change is expected to have major impacts on future land use, agriculture, and food security. National GDP growth is also projected to decline by 3–10% annually depending on these effects (World Bank, 2010).

Climate change and unpredictability can exacerbate land degradation by making current land management approaches unsustainable, According to Worku and Mekonnen (2012) and Berry (2003), poor land-use practices and population pressure were among the main sources of land degradation in Ethiopia. In Ethiopia, the yearly cost of land degradation due to changes in land use and the cover was estimated to be over $4.3 billion (Samuel Gebreselassie et al., 2015).

Climate change adaptation provides an opportunity for a multidisciplinary and integrated strategy to address climate-related hazards, which are expected to severely impede agricultural productivity. Smallholder farmers employed diverse coping, on-farm, and off-farm adaptation techniques to decrease the negative impacts of climatic and ecological changes on their livelihood. According to Hadgu, et al. (2014), site-specific adaptation methods can assist lessen the impact of climate change while also maximizing the benefits.

The Amhara region is one of the nine regional states in Ethiopia which is being affected by recurrent drought because of its arid nature in the northern and northeastern parts of the region emanated from climate change and variability. The study area namely Lake Tana (the largest lake in Ethiopia) and the upper Beles watershed are among the ecologically sensitive areas in west Gojjam of the Amhara region. These watersheds drain towards lake Tana from the east side and towards the Beles River from the west side. In the study area, farmers use different land management adaptation measures. In addition, farmers also use diversification to non-farm livelihood practices that enable farm households to have better income, enhance food security, and increase agricultural production.

The majority of climate change studies have focused on the impact of climate change on crop and livestock production, soil and nutrients, climate change adaptation and mitigation, and coping mechanisms for immediate climate variations and adaptation. Many more studies on climate change perception, adaptation, and soil and water conservation strategies have also been conducted by Gezie (2019; Teklewold et al. (2019). However, this research did not address how farm households make decisions about implementing adaptation technology. Furthermore, studies on the topic of land management adaptation approaches to climate change at the farm or household level, particularly in the Lake Tana and Beles River watersheds, are frequently inadequate. What variables did farmers examine when determining which land management and climate change adaptation practices to implement? Are their decisions synchronized with or in opposition to their actions? What factors influence land management adaptability in terms of socio-economic factors? Despite their efforts, none of them had been adequately investigated or reviewed. This is especially true in this watershed, which is particularly vulnerable. Furthermore, the study contributes to the existing literature on land management adaptation techniques in the study area such as Addisu et al. (2016), Amare and Simane (2017), Gezie (2019), Jeetndra Prakash et al. (2021), in aiding the development of appropriate land management interventions and specific policy recommendations to facilitate the adoption of regional adaptation packages. Farmers’ practices in selecting land management adaptation techniques to reduce climate change are the subject of the study.

As a result, the study will investigate smallholder farmers’ decisions to adopt alternative land management strategies in response to climate change. It also looks at how various land management adaptation options interact, as well as whether farmers should use complementary or substitutable agricultural adaptation practices. It also explains the key socio-economic elements that influenced farmers’ decisions to implement land resource management strategies in the study area.

2. Methodology

2.1. Description of the study area

The survey, which was conducted west of Lake Tana between the lake and the Beles River basin, covered three kebeles (Charma Dusuman, Wombera Eyesus, and Kunzella Zuria.) as shown in Figure 1. In total, there are 2703 families in three of the villages, where 2384 of them are headed by men and 319 by women. The study only reflects the highland (northwest) part of Ethiopia and does not cover the entire country. In Charma Dusman Kebele, there are 4,153 residents, where 2,237 of them are men and 1,916 of whom are women. The Woreda agriculture department reports that the populations of Wombera Eyesus and Kunzela Zuria kebeles are 8,286 and 6,045, respectively, Out of these, 4322 and 3198 are males, and 3964 and 2847 are females in Wombera Eyesus and Kunzela Zuria kebeles, respectively (WAO, 2017).

Figure 1. Location map of the Study Area.

The massive Lake Tana, with a catchment area of 15,054 km2, and the Beles River, which feeds into the Blue Nile, a highly valued water resource, are both located on the border of the research area. The Beles River Basin is situated in northwest Ethiopia. Nearly 550 kilometers separate it from Addis Abeba. Dangila (a town located 490 kilometers from Addis Abeba), and Durbete are the two towns from where the upper part of the Watershed are reached (which is 510 km from Addis Abeba). The characteristics of the watershed are explained in Table 1.

Table 1. Characteristics of the watersheds

Watersheds	Characteristics of the watersheds	Source
Lake Tana and Belese watershed	a) The lake Tana Basin located at 11°36’ N & 23’ E with°37 4 average natural altitudes of 1786 m asl. The average annual rainfall is 1248 mm per year (mm yr−1) Atmospheric temperature ranged from 13 to 22 °C covers an area of 15,114 km2 &tropical Lake with surface area of 3111 Km2 It is the largest freshwater resource of Ethiopia (50%) & the source of the Blue Nile(Abay) River. b) Belles river basin Located in the southwestern direction of Lake Tana(10° 56’ to 12° North latitude and 35°12’ to 37° East longitude. The total area of the basin is about 14,200 km 2. The topography of the area is mainly flat with altitude variables from 458 to 2729 m above sea level. Beles is one of the most important basins in Ethiopia and one of the major sub-basins of the Upper Blue Nile Basin	Ashebir, 2017, (Goshu & Aynalem, 2017

The farming system includes crop and livestock production. Aside from it, there are lake and river-based irrigation systems. Agriculture is the main source of income for the locals, but petty trade is also prevalent. Teff, maize, and finger millet are three of the most prevalent crops farmed in the area. Cattle, donkeys, mules, sheep, and goats are among the animals raised in the area. The area has hydropower potential, and the government is currently working on a scheme to start irrigation-based sugarcane production.

2.2. Sampling techniques and sample size

A multistage sampling design was used in this investigation. First, the zone¹ and woreda were carefully chosen for their ecological significance as well as their proximity to the research site. Three kebeles were chosen at random (using the lottery technique) and in proportion to the total size of the kebeles out of the 27. The kebele² agricultural office provided a sample frame, and a simple random sampling technique was used to choose the target farmers from each group. Using Yamane’s (1967), a sample size of 338 (12.5 %) was chosen from the entire population of households:

(1) $n=\frac{N}{1+N(e^2)}=\frac{1419}{1+1419({0.04}^2)}=338$(1)

Where n = the sample size; N = population size; e = error of margin

2.3. Method of data collection

A semi-structured questionnaire was used to obtain primary data. The questionnaire asked about the households’ socioeconomic characteristics, animal and crop production, farmers’ perception, climate factors including rainfall distribution and temperature, and land resource adaptation measures such as area closure, forest protection, water resource development and mulching, handcraft, small trade, wage worker, and other non-farm adaption techniques are examples. Before the actual data collection, the questionnaires were pretested by distributing them to 30 farmers, and adjustments and corrections were made. The woreda office of agriculture provided information on the population size of the kebeles, the farming system, and the area or location. In addition, six key-informant interviews and four focus group discussions (each with eight persons) were conducted.

2.4. Model specification

2.4.1. Dependent and independent variables

The dependent variable “land management adaption practices” was regressed against independent variables in this study. The dependent variables were selected from different literature and farmers’ practices on the ground. The most common climate variability and adaptation strategies in rural Ethiopia include, contour plowing, graded bund terracing, mulching, compost preparation, installation of a wood lot, forestation, hedge planting, area closure, suitable waterway, and gully rehabilitation were all examined as dependent variables (Elasha et al., 2006; Addisu et al., 2016; Wolka & Zeleke, 2017; Cherinet, 2017; Gezie, 2019). The socioeconomic elements that influence farmers’ land management adaptation methods to climate change are the independent variables. Table 2 lists the description, measurement, expected sign, and literature sources.

Table 2. Socioeconomic variables affecting land management adaptation practices

I. Dependent variables	Description of the climate adaptation practices
$C_P$	Contour Ploughing
$G_b$	Grade Bund Terracing
$M_p$	Mulching Practice
$C_p$	Compost Preparation
$W_l$	Wood Lot Implementation
$F_p$	Forestation Practice
$H_p$	Hedge Planting
$A_c$	Area Closure
$p_w$	Proper waterway
$G_r$	Gulley Rehabilitation
II. Independent Variables	Measurement	Hypothesis	Supporting References
1. Age of household head	Number	Positive/ negative	Yesuf et al., 2008; Temesegen et al., 2009, Acquah 2011 &Quayum & Ali, 2012, Uddin et al., 2014
2. Sex of household head	Dummy (1 = Male, 0, otherwise)	Positive or negative	Asfaw & Admassie, 2004; Amare & Simane, 2017; Dolisca et al., 2006; Hemachena and Hassan, 2007
3. Education level	Level category	Positive	Uddin et al., 2014.
4. Family size	Number	Positive/negative	Temesegen et al., 2009, Menberu&Yohannes,2004 Uddin et al., 2014
5. Farm experience	Years	Positive	Franklin Nantui Mabe et al., .
6. Farm size	Hectare	Positive/negative	Temesegen et al., 2009; Temesgen et al., 2010. Etsay et al., 2019
7. Distance to market	Kilometer	Negative	Amare & Simane, 2017
8. Frequency of extension visit	Number	Positive	AkhterAli & OlafErenstein, 2017.
9. Crop income	Birr	Positive	Temesegen et al., 2009, T. T. Deressa et al. (2009) and Uddin et al., 2014
10. Oxen ownership	Number	Positive	Barungi & Maonga, 2011
11. Access to credit	Dummy(yes = 1, 0 otherwise)	Positive/ negative	Uddin et al., 2014 and AkhterAli & OlafErenstein, 2017

2.4.2. A multivariate probit model

The decision of a farmer to apply a certain adaptation strategy is binary or discrete, necessitating qualitative choice models. Because a farmer may choose mixed adaptation technologies, the decision to use one practice may be influenced by the decisions made for the other practices, the Binary Probit/Logit, and Multinomial Probit/Logit models do not account for possible interrelationships between the various adaptation practices. As a result, the multivariate probit model takes into account interdependencies and multiple technology selection decisions (Arinloye et al., 2015; Degye et al., 2013; Yu et al., 2008). To examine the socio-economic elements that influence climate adaption options, a multivariate probit model (MVP) is used.

The multivariate probit model expressed as the selection of climate change strategy i by farmer j is Y*ij. This is defined as the choice of farmer j to adapt (mitigate) the technology in climate adaptation strategy i, and it is expressed as follows:

Y*ij is the multivariate probit model stated as farmer j’s choice of climate change approach i. This is described as farmer j’s decision to adapt (mitigate) technology in climate adaptation strategy i and it is written as follows:

(2) ${{\mathrm{Y}}^{\ast }}_{\mathrm{i}\mathrm{j}}=X_i{\beta }_j+U_{ij}$(2)

$y_{ij}=1({y^{\ast }}_{ij}>o)$

$U_i=(U_{i1}\dots \dots \dots U_{im})$

Where: Y*ij is the dependent variable (contour plowing, graded bund terracing, mulching, compost preparation, implementation of wood lot, forestation, hedge planting, area closure, proper waterway, and gully rehabilitation).

X_j is a vector of explanatory (socioeconomic) variables.

β_ij is a vector of parameters to be estimated and

U_ij is a disturbance term.

Farmers can pick from a variety of adaption strategies in multivariate models. As a result, the error terms (ei) follow a multivariate normal distribution with zero conditional means (μ; = 0), variance normalized to unity (σ²= 1), and the symmetric covariance matrix Ψ is:

(3) $\psi =\left(\begin{array}{ccccccccccc}1& C_a& G_b& M_p& C_p& W_l& F_p& H_p& A_c& P_w& G_r\\ C_a& 1& {\rho }_{C_a{G}_b}& {\rho }_{C_a{M}_p}& {\rho }_{C_a{C}_p}& {\rho }_{C_a{W}_l}& {\rho }_{C_a{F}_p}& {\rho }_{C_a{H}_p}& {\rho }_{C_a{A}_c}& {\rho }_{C_a{P}_w}& {\rho }_{C_a{G}_r}\\ G_b& {\rho }_{G_b{C}_a}& 1& {\rho }_{G_b{M}_p}& {\rho }_{G_b{C}_p}& {\rho }_{G_b{W}_l}& {\rho }_{G_b{F}_p}& {\rho }_{G_b{H}_p}& {\rho }_{G_b{A}_c}& {\rho }_{G_b{P}_w}& {\rho }_{G_b{G}_r}\\ M_p& {\rho }_{M_p{C}_a}& {\rho }_{M_p{G}_b}& 1& {\rho }_{M_p{C}_p}& {\rho }_{M_p{W}_l}& {\rho }_{M_p{F}_p}& {\rho }_{M_p{H}_p}& {\rho }_{M_p{A}_c}& {\rho }_{M_p{P}_w}& {\rho }_{M_p{G}_r}\\ C_p& {\rho }_{C_p{C}_a}& {\rho }_{C_p{G}_b}& {\rho }_{C_p{M}_p}& 1& {\rho }_{C_p{W}_l}& {\rho }_{C_p{F}_p}& {\rho }_{C_p{H}_p}& {\rho }_{C_p{A}_c}& {\rho }_{C_p{P}_w}& {\rho }_{C_p{G}_r}\\ W_l& {\rho }_{W_l{C}_a}& {\rho }_{W_l{G}_b}& {\rho }_{W_l{M}_p}& {\rho }_{W_l{C}_p}& 1& {\rho }_{W_l{F}_p}& {\rho }_{W_l{H}_p}& {\rho }_{W_l{A}_c}& {\rho }_{W_l{P}_w}& {\rho }_{W_l{G}_r}\\ F_p& {\rho }_{F_p{C}_a}& {\rho }_{F_p{G}_b}& {\rho }_{F_p{M}_p}& {\rho }_{F_p{C}_p}& {\rho }_{F_p{W}_l}& 1& {\rho }_{F_p{H}_p}& {\rho }_{F_p{A}_c}& {\rho }_{F_p{P}_w}& {\rho }_{F_p{G}_r}\\ H_p& {\rho }_{H_p{C}_a}& {\rho }_{H_p{G}_b}& {\rho }_{H_p{M}_p}& {\rho }_{H_p{C}_p}& {\rho }_{H_p{W}_l}& {\rho }_{H_p{F}_p}& 1& {\rho }_{H_p{A}_c}& {\rho }_{H_p{P}_w}& {\rho }_{H_p{G}_r}\\ A_c& {\rho }_{A_p{C}_a}& {\rho }_{A_p{G}_b}& {\rho }_{A_p{M}_p}& {\rho }_{A_p{C}_p}& {\rho }_{A_p{W}_l}& {\rho }_{A_p{F}_p}& {\rho }_{A_p{H}_p}& 1& {\rho }_{A_p{P}_w}& {\rho }_{A_p{G}_r}\\ P_w& {\rho }_{P_p{C}_a}& {\rho }_{P_p{G}_b}& {\rho }_{P_p{M}_p}& {\rho }_{P_p{C}_p}& {\rho }_{P_p{W}_l}& {\rho }_{P_p{F}_p}& {\rho }_{P_p{H}_p}& {\rho }_{P_p{A}_c}& 1& {\rho }_{P_p{G}_r}\\ G_r& {\rho }_{G_r{C}_a}& {\rho }_{G_r{G}_b}& {\rho }_{G_r{M}_p}& {\rho }_{G_r{C}_p}& {\rho }_{G_r{W}_l}& {\rho }_{G_r{F}_p}& {\rho }_{G_r{H}_p}& {\rho }_{G_r{A}_c}& {\rho }_{G_r{P}_w}& 1\end{array}\right)$(3)

Where ${\rho }_{ij}$ denotes the relationship between several forms of climate change adaptation strategies. It is conceivable to use the multivariate probit model independently if these correlations in the off-diagonal elements of the covariance matrix become non-zero. Otherwise, we apply the multinomial probit model for adaptational practices. The statistical analyses are considered under three levels of significance. These are at 1%, 5%, and 10% levels of significance.

3. Results of the study

3.1. Demographic and socio-economic characteristics of the participants

According to the survey result, approximately 81 % of the sampled homes are between the ages of 31 and 60, with approximately 10% of the sampled households being over the age of 60. The average age of the household is 45 years old. Male-headed households accounted for 94% of the respondents, while female-headed households accounted for 6%. In the study, 50% of the households were illiterate (unable to read and write), 35% could only read and write, and 10% were in grades one through four. Around 30% of homes contain six to seven persons, while 31% have more than seven. The average family size is about 7, which is over the national family size of the country.

According to the research, 8% of households own less than 0.5 hectares of land, while 30% own between 0.5 and 1 hectare. Between 1.11 and 1.55 hectares of land are owned by around 21% of families. The result indicated that the respondents have on average three oxen. According to Melaku (2011), 34% of Ethiopian farmers have no oxen and 29 % have two oxen. The average land size of the respondents was about 1.5 hectares. Farmers had an average of 20 to 30 years of farming experience. This accounts for roughly 34% of farm households. Similarly, the sampled households were asked if they had received extension services in the previous month, and the frequency of extension contact was reported to be between 3 and 4 times a month (33 %) for households, and 1 to 2 times (5 %) for responses (Table 3).

Table 3. Demographic and socio-economic characteristics of smallholder farmers

variables	N	Minimum	Maximum	Mean	Std. Deviation
1. Age of hh	337	24.00	73.00	45.75	9.11
2. Family size of hh	336	2.00	9.50	6.57	2.42
3. Oxen number	329	1.00	5.50	2.59	0.97
4. Land size	317	0.25	3.25	1.55	0.99
5. Farm experience	336	5.00	45.00	25.95	9.98
6. Market Distance	338	3.00	13.00	4.86	3.21
7. Extention frequency	338	1.50	9.00	4.31	2.33
8. Crop income	335	1,000.00	25,000.00	14,602.41	9,721.63
9. Total income of HH	337	5,000.00	35,000.00	22,715.23	11,735.87
10. Education level of Household head					Freq (%)
Illiterate					170 (50.3%)
Read and write					119 (35.2%)
1–4 years of education					36 (10.7%)
5–8 years of education					10 (3%)
9–10 years of education					3 (0.89%)
Total					338
11. Do households get credit services?
No					134 (39.6%)
Yes					197 (58.3%)
Total					331

3.2. Farmers’ land management adaptation practices to climate change

Farmers used ten key land management practices out of 15 alternative practices used in the area. Grade bund terracing (97 %), contour plowing (96.4%), cut-off drains (96.4%), proper waterway (88%), and compost preparation (85%) were the top five practices used to manage their land management practices. Area Closure and Gully rehabilitation on the other hand is the least land management strategy that farmers adopt to adapt to climate change (55.6 %). However, depending on their socioeconomic variables, farmers implemented multiple adaptation strategies (jointly) on their plots of land (Figure 2).

Figure 2. Farmers’ land management adaptation practices to climate.

3.3. Model result of land management adaptation practices

Farmers’ decisions about which adaptation strategy mechanism to use had a major impact on the climate change adaptation strategies that they chose (Table 4). The model proved significant at the 1% level (Wald chi-square (100) = 1271.30), and the factors included in the model have sufficient explanatory power. The likelihood ratio test of the null hypothesis of independence (ij = 0) (χ2 (45) = 323.225, Prob. > χ 2 = 0.0000) is significant at 1%, indicating that the model’s goodness of fit and farmers’ decisions to choose different adaptation practices were interdependent or correlate to one another.

Table 4. Model fitness, probabilities, and correlation matrix of land adaptation choices

Variables	contour plowing (pr) (1)	Grade bund Terracing (2)	Mulching (3)	Compost Preparation (4)	Wood lot Implementation (5)	Forestation (6)	Hedge Planting (7)	Area Closure (8)	Proper waterway (9)		Gulley Rehabilitation (10)
Predicted probability	0.969	0.971	0.666	0.843	0.555	0.612	0.734	0.554	0.884		0.578
Joint probability of success	0.096653
Joint probability of failure	9.25E-07
Estimated correlation of land management practices (Pair-wise correlation coefficients)
ρ 42	0.287*		ρ 108	0.414***
ρ 43	0.431**		ρ 109	0.361**
ρ 54	0.587***		ρ 51	−0.420*
ρ 64	0.227*		ρ 61	−0.53***
ρ 65	0.721***		ρ 63	−0.226**
ρ 75	0.178**		ρ 72	−0.201*
ρ 76	0.404***		ρ 73	−0.228**
ρ 87	0.319***		ρ 74	−0.256**
ρ 92	0.281**		ρ 82	−0.250*
ρ 96	0.207*		ρ 84	−0.406***
ρ 97	0.273**		ρ 104	−0.232**
ρ 98	315**
ρ 106	0.214**
ρ 107	0.319**
Likelihood ratio test of ρ21 = ρ31 = ρ41 = ρ 51 = ρ61 = ρ71 = ρ 81 = ρ91 = ρ101 = ρ32 = ρ42 = ρ52 = ρ62 = ρ72 = ρ82 = ρ92 = ρ102 = ρ43 = ρ53 = ρ63 = ρ73 = ρ83 = ρ93 = ρ103 = ρ54 = ρ64 = ρ74 = ρ84 = ρ94 = ρ104 = ρ 65 = ρ75 = ρ85 = ρ 95 = ρ105 = ρ76 = ρ86 = ρ 96 = ρ106 = ρ 87 = ρ97 = ρ107 = ρ98 = ρ108 = ρ109 = 0: chi2(45) = 323.225 Prob.> chi2 = 0.0000 Numbers of draws (SML, #draws) = (5) Numbers of observation = 312 Log pseudolikelihood = −1296.4571 Wald chi2(100) = 1271.30, Prob.> chi2 = 0.0000

***, **, * are statistically significant at 1%, 5%, and 10%, level of significance, respectively.

The land management adaption practices’ correlation coefficient (ij) was determined, as well as their complementarity and substitutability. The results showed that eight combined practices, namely ${\rho }_{61}$ (forestation and contour plowing), ${\rho }_{72}$ (hedge and bund terrace), ${\rho }_{84}$ (area closure and bund terracing), ${\rho }_{63}$ (forestation and mulching), ${\rho }_{73}$ (hedge planting and mulching), ${\rho }_{74}$ (hedge planting and compost),${\rho }_{84}$ (area closure and compost), and ${\rho }_{104}$ (gully rehabilitation and compost), were substitutable to land management adaptation practices to climate change, with their coefficient signs negative. The rest of the combinations of land adaptation strategies, on the other hand, have proved complementary, and farmers have chosen to apply them together. The findings revealed that adaption strategies such as ${\rho }_{65}$ (forestation and woodlot), ${\rho }_{54}$ (woodlot and compost), and ${\rho }_{76}$ (Hedge planting and forestation) are highly complementary. Furthermore, forestation and contour plowing, as well as area closure and compost preparation, have all been found to be highly substitutable land management adaptation practices.

Table 4, further showed that the chance of farmers adopting climate change practices was assessed. As a result, the likelihood of implementing counter plowing, grade bund terracing, compost preparation, and proper waterway implementation is relatively high (96.9%, 97.1%, 84.3%, and 88.4%) respectively, compared to the likelihood of implementing other practices such as mulching (66.67 %), wood lot implementation (55.56 %), forestation (61.20 %), hedge planting (73.4 %), area closure (55.4 %), and gully rehabilitation (57.80 %).

The joint probabilities of the ten agricultural technologies’ climate change adaptation strategies indicate that households are more likely to adopt land management practices together. In comparison to their failure to adopt the ten technologies, households are 9.6% more likely to adopt them jointly (i.e. 0.096). This suggests that the chances of selecting the right mix of climate change adaptation practices are high. According to this evidence, distinct criteria for each adaptation method will decide the best mix of land adaptation practices to climate change. The predicted probability indicated that Grade bund terracing, and contour plowing, were the most commonly chosen land management adaptation measures used by farmers, while area closure, Wood lot Implementation, and Gully Rehabilitation are the less likely choice practice.

3.4. Determinants of smallholder farmer’s land management adaptation practices

Farmers’ choice of adaption practices is influenced by the age of the household head. However, except for grade bund terracing practice, which is statistically favorable (positive) at a 10% level of significance, the age of the household head had no significant effect on adaption practices. On the contrary, sex plays a substantial impact in farmers’ decision-making regarding climate change adaptation methods. At the 1% level of significance, being a male farmer has a positive and highly significant effect on contour plowing and graded bund terracing practices compared to female farmers. Hedge planting, forestation, and area closure, on the other hand, have a negative and severe impact on farmers’ adaptation strategies (Table 5).

Table 5. Statistical analysis results for the socioeconomic factors influencing land management practice adopted by smallholders in Western Lake Tana and Beles River watersheds (Ethiopia) as an adaptation strategy to climate change

VARIABLES	Contour plowing (1)	Graded bund Terracing (2)	Mulching (3)	Compost Preparation (4)	Wood lot Implementation (5)	Forestation (6)	Hedge Planting (7)	Area Closure (8)	Proper waterway (9)	Gulley Rehabilitation (10)
Age of HH	−0.157(0.542)	0.627* (0.377)	0.303 (0.225)	−0.028 (0.274)	−0.275 (0.218)	−0.193(0.213)	−0.013(0.210)	−0.215 (0.207)	−0.007(0.216)	−0.166 (0.214)
Sex of HH	3.051***(0.459)	3.335*** (0.470)	0.236 (0.469)	0.525 (0.564)	0.084 (0.333)	−0.783*(0.404)	−1.141***(0.397)	−0.687*(0.384)	0.028 (0.387)	−0.115(0.402)
Education level	0.519** (0.242)	0.332* (0.181)	0.036 (0.094)	0.155 (0.113)	0.052 (0.096)	−0.103 (0.099)	−0.028 (0.09)	−0.199**(0.087)	0.267*(0.137)	−0.013(0.089)
Family size	−0.465** (0.203)	−0.425***(0.159)	−0.126 (0.093)	0.218** (0.107)	0.186**(0.085)	0.207**(0.092)	0.122 (0.081)	0.016 (0.083)	0.134 (0.114)	0.008(0.091)
Farm experience	0.000 (0.168)	0.040 (0.222)	0.170* (0.099)	0.265** (0.104)	0.192**(0.093)	0.162*(0.095)	−0.092 (0.105)	0.071 (0.093)	0.122 (0.106)	0.201**(0.096)
Land Size	0.016 (0.117)	−0.050 (0.104)	−0.140** (0.070)	−0.0551(0.087)	−0.142**(0.072)	−0.086 (0.066)	−0.043 (0.068)	−0.065 (0.070)	−0.029(0.084)	−0.019 (0.067)
Crop income	0.183 (0.136)	0.126 (0.084)	−0.118* (0.068)	−0.108 (0.082)	0.046 (0.066)	0.083 (0.072)	0.138**(0.061)	−0.063 (0.067)	−0.058(0.074)	0.049 (0.066)
Distance to market	0.223 (0.201)	−0.122 (0.235)	−0.010(0.123)	−0.311**(0.134)	0.093 (0.122)	0.038 (0.130)	0.128 (0.119)	0.401***(0.125)	−0.112(0.134)	0.085 (0.117)
Frequency of extension	0.415 (0.285)	0.154 (0.116)	0.236***(0.067)	0.214** (0.096)	0.069 (0.070)	0.015 (0.065)	0.128**(0.064)	0.204***(0.063)	0.027(0.086)	0.135**(0.065)
Oxen ownership	−0.383 (0.237)	−0.304 (0.195)	0.197 (0.145)	−0.039 (0.187)	−0.299**(0.143)	−0.072 (0.147)	−0.112(0.136)	0.215 (0.137)	−0.355**(0.164)	0.012 (0.141)
Constant	−1.088 (1.360)	−1.760 (1.123)	−0.702 (0.860)	−0.529 (1.037)	−0.088 (0.741)	0.533 (0.738)	1.154 (0.763)	0.303 (0.759)	1.143(0.931)	−0.574 (0.81)
Number of obs. = 312 Wald chi2(100) = 1271.3, (SML, # draws) = (5) Log pseudolikelihood = −1296.4571, Prob.> chi2 = 0.0000 Likelihood ratio test of rho21 = rho31 = … . … . = rho109 = 0: chi2(45) = 323.225. Prob.> chi2 = 0.0000

Note: The dependent variable is land adaptation choice; ***, **, * are significant at 1%, 5%, and 10%, level respectively * Standard errors are in parenthesis.

Farmers with higher education have a favorable and significant impact on contour plowing, bund terracing, and proper waterway techniques, according to the study. On the contrary, it has a negative and large impact on a region that is more closely associated with climate adaptation strategies. According to Tizale (2007), household size is a proxy for labor availability, allowing farmers to implement labor-intensive adaptation methods. The adaption technique of counter-plowing and graded bund terracing has been considerably and negatively influenced by the size of a household’s family. However, as a climate change adaptation strategy, it has a beneficial and considerable impact on compost preparation, wood lot implementation, and forestation.

On the one hand, the sample households’ farm experiences have a favorable and significant impact on farmers’ climate adaptation techniques, such as compost preparation, wood-lot implementation, and gully rehabilitation. Mulching and forestation adaptation tactics, on the other hand, benefit from it. The likelihood of adopting climate change adaptation techniques is unaffected by oxen ownership. Oxen ownership, on the other hand, has had a detrimental and severe impact on wood lot plantations.

The land resource adaption mechanism of climate change has been found to be adversely related to land size. Mulching and wood lot practices were shown to be negatively and significantly connected to climate adaptation mechanisms in the study. Crop farm income, meanwhile, is a measure of financial capacity that aids in the adoption of agricultural methods or technology. At a 5% level of significance, hedge planting adaptability is positively and significantly associated with farmers’ household income. However, at a 5% level of significance, crop income has a negative impact on implementing mulching as a climate adaptation strategy.

Farmers’ decisions in the use of land management adaption strategies are substantially influenced by the residents’ proximity to the local market. The longer distance between a farmer’s home and the market has had a negative and considerable impact on compost preparation techniques. However, as a strategy of climate adaptation mechanisms, there is a favorable and strong relationship between market distance and area closure practices.

Farmers’ decisions on all land management climate change adaptation choices were found to be influenced by extension contact frequency in this study. At a 5% level of significance, mulching and area closure have a significant impact on the likelihood of employing climate change adaptation techniques. Similarly, climate change adaptation strategies have been positively and considerably influenced by compost preparation, hedge planting, and gully rehabilitation.

4. Discussion of results

Working-age households make up the majority of the population. When compared to younger farmers, older farmers are more likely to utilize land management adaptation practices to combat climate change. The model also revealed that, of all the land adaptation strategies, only age was a relevant determinant, especially when employing the grade bund terracing practices. As a climate adaptation method, the majority of the farmers (97 %) adopted grade bund terracing practices. In comparison to young farmers, elderly farmers are more inclined to change land management climate adaptation strategies. To use grade bund terracing as a climate adaptation method, you’ll require technical capabilities. A similar finding was reported by Acquah (2011) and Quayum & Ali, 2012).

Male-headed households make up the majority of the households (97.5 %). When compared to other climate adaptation strategies like contour plowing, graded bund terracing, hedge planting, forestation, and area closure, sex has a substantial impact on the utilization of land management adaptation practices. In comparison to female-headed families, male-headed households are more likely to adopt contour plowing and graded bund terracing as climate change adaptation strategies. Female-headed households use hedge planting, forestation, and area closure activities more than male-headed households. This could be because male respondents have higher technical abilities and more access to information than female respondents. Asfaw and Admassie (2004), came to the same conclusion. Amare and Simane (2017) found that being a female farmer was negatively and significantly connected to the use of agronomic methods and livelihood diversification measures in a comparable study. This shows that agricultural women are more vulnerable and less able to cope with shocks than males, especially when climate variability is significant (Food and Agriculture Organization of the United Nations (FAO), 2013). Households headed by men use more adaption techniques than households headed by women (Temesegen et al., 2009).

The average household has four to five members, which is similar to the national average (i.e. 4.6 individuals per household). Climate change adaptation practices are influenced by the size of the household’s family. Family size, on the other hand, could be a source of labor for climate adaptation. Contour plowing and graded bund terracing were also found to be significantly and negatively linked with farm household land management adaption methods in the model. Compost preparation, woodlot implementation, and forestation practices, on the other hand, have a favorable and significant relationship. The reason for this is that a higher family size usually means more laborers who are willing to work. Uddin et al. (2014), who researched factors affecting adaption strategies in Bangladesh, came to the same conclusion.

Family size, on the other hand, has a large and negative impact on climate change adaptation measures, notably for farmers’ terracing and contour plowing practices. This is because if there are enough options for off-farm labor, household liquidity grows faster than on-farm operations. This lowers the quality of a household’s labor endowment and lowers actual internal labor availability (labor opportunity cost; Uddin et al., 2014). According to Teshome and Aberra (2014), the greatest statistically significant determinant of land management methods adoption is household size.

Farmers’ experiences have an impact on climate change adaptation and land management techniques. In the research area, around 30% of the farmers had 20 to 30 years of farming experience. Compost preparation, woodlot (tree planting), gully rehabilitation, and mulching adaption methods were all found to be significantly and favorably associated with farmers’ agricultural experiences in the model. Farmers’ adoption of particular adaptation strategies is likely to be influenced favorably by their farming experience in the research location. The reason for this is that if you have more agricultural experience, you’re more likely to notice and adjust to climate change and variability (Franklin Nantui Mabe et al.,).

Another factor that determines land management adaptation to climate change is land. A land size of 0.5 to 1.5 hectares is owned by around half of the households. Only the size of a farmer’s land has had a negative and significant impact on mulching and woodlot implementation procedures. This could be because implementing these methods becomes more difficult as the land size grows, as it necessitates more labor and costs.

Farm income, derived from crop and livestock production, has an impact on farmers’ climate change adaptation methods. The farmers’ annual average crop revenue was 14,602.41 ETB,³ which accounts for roughly 64% of their total income. The model also revealed that while farm income increases the likelihood of utilizing hedge planting as a climate change adaptation technique, it has a negative impact on the usage of mulching. Uddin et al. (2014) discovered that a farmer with more wealth in Bangladesh is more likely to use mulching as a climate change adaptation method. The findings are likewise similar to those of T. T. Deressa et al. (2009), who investigated Ethiopian farmers’ perceptions of climate change and adaptive measures. Farmers want to divert their money to other purposes, such as off-farm activities or alternative investments. This means that higher-income farmers are more likely to adopt adaptive methods than lower-income farmers (Uddin et al., 2014).

According to the survey results, 89 % of farmers traveled fewer than five kilometers from their homes. Farmers’ land management adaption strategies to climate change have been influenced as a result of this. The conclusion implies that in the research area, households living away from markets were more likely to employ area closure as a climate change strategy. A study conducted by Amare and Simane (2017) in the Muger Sub-basin of Ethiopia’s Upper Blue Nile basin found that distance to the marketplace and field size had a detrimental impact on the implementation of agronomic methods. On the contrary, it was associated with area closure procedures and land management activities in a positive and significant way.

Farmers adopted a variety of land management adaptation strategies, both in terms of quantity and type. They used a variety of or many adaptation practices at the same time to either replace or supplement current procedures. According to the report, there is a significant correlation between farmers’ adaptation techniques and climate change. Farmers’ decisions to adopt land adaptation management strategies to climate change are more complementary than replacement approaches. Farmers’adopted a variety of land adaptation strategies to mitigate the effects of climate change. The findings revealed that the cross-correlation among the adaptation practices implied that adaption techniques are highly complementary. Farmers in the Nile basin of Ethiopia used a variety of climate change adaptation tactics, according to Teklewold et al., 2019).

5. Conclusion and recommendations

We explored various land management adaptation options and discovered the factors that influence them. The results may be extrapolated to similar areas, and the actual significance is important for agricultural production and the development of climate adaptation strategies in the study area. The findings demonstrated a strong relationship between adaptation practices and climate change, indicating that complementarity, rather than supplementary adaptation practices, influences farmers’ decisions.

The probability of farmers applying specific land management adaptation decision was found to be correlated. Several socioeconomic factors influenced farmers’ decisions to choose different land management adaptation practices, found to be as sex, family size, agricultural experience, frequency of extension services, and household education level. When compared to the probability of employing other practices, the likelihood of using contour plowing (96.9%), grade bund terracing (97.1%), compost preparation (84.3%), and establishing proper waterways (88.4%), compost preparation (84.3%) is substantially greater compared to other land management adaptational practice like mulching (66.6%), forestation (61.2%), Gulley Rehabilitation (57.8%), wood lot implementation (55.5%) and Area Closure (55.4%).

The joint probabilities of land management adaptation practices to climate change practices show that households are more likely to adopt these practices together. When comparing the likelihood of farmers adopting the behaviors together versus their failure to do so, the likelihood of households adopting them together is 9.6%. As a result, developing packages of land management adaptation methods as a climate change approach is critical. Furthermore, policy and development interventions to minimize climate change and variability should promote different land management adaption methods. Finally, all socio-economic variables that influence land management adaptation strategies in response to climate change must be considered. Future research may focus on crop management practices such as changing agricultural calendars, introducing seed varieties which are tolerant to climate change, and enhancing land management adaptation practices to climate change as a way forward.

Availability of data and materials

We declare that whatever data have been used in the manuscript will be kept intact. These data can be made available to anyone who desires to see them from the corresponding author on request.

Acknowledgements

We highly acknowledge Bahir Dar University for its financial support, and developmental agents and farmers who participated in the field survey.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was financially supported by Bahir Dar University, College of Agriculture and Environmental Sciences, Bahir Dar, Ethiopia.

Notes on contributors

Astewel Takele

Astewel Takele Department of Agricultural Economics, College of Agriculture and Environmental Sciences, Bahir Dar University, Bahir Dar, Ethiopia

Assefa Abelieneh

Assefa Abelieneh Department of Rural Development and Agricultural Extension, College of Agriculture and Environmental Sciences, Bahir Dar University, Bahir Dar, Ethiopia

Beneberu A. Wondimagegnhu

Beneberu A. Wondimagegnhu Ethiopian Policy Studies Institute, Economic Policy Studies Sector

Notes

1. Zones and Woredas are geographical administrative units

2. Kebele is the smallest administrative unit

3. ETB = Ethiopian birr; 1USD = 53.73 ETB