Gender perspective on climate change adaption strategies in livestock farming in Gandaki Province, Nepal

Abstract

Background:

Climate change has a significant impact on livestock farming around the globe. Farmers have adopted different strategies to mitigate the adverse impact of climate change. Females in developing countries are more vulnerable to climate change impacts and have lower adaptive capacity and they bear additional roles and responsibilities in livestock rearing compared to their male counterparts.

Aim:

The main aim of this study is to examine the gender perspective on climate change adoption strategies in livestock farming in Gandaki province, Nepal.

Methods:

A multistage random sampling technique was employed to select 1,158 households from five districts in Gandaki province, western Nepal. A household head or household member who was 45 years or older resided in that area for at least 15 years and owned at least one primary livestock at the time of the survey was selected as the ultimate respondent from each selected household. Both structured and unstructured questionnaires were prepared. A structured questionnaire was used for the household survey, while a checklist (guideline) was prepared for focus group discussions. Data were collected through face-to-face interviews, and both descriptive and inferential statistics were used for data analysis.

Results:

The results revealed that buffalo was the primary livestock among farmers. More than half of farmers, both men and women were aware of the impact of climate change on livestock. While this study did not find significant gender-based differences in adaptation strategies, the odds of adoption are higher among males than females. Jobs other than agriculture and livestock, as well as access to credit, emerged as key determining factors associated with adaptation strategies among farmers in Gandaki province.

Conclusion:

There is no significant gender-based difference in adaptation strategies; however, employment outside agriculture and livestock, along with access to credit, are the key determining factors associated with adaptation strategies.

Introduction

The latest Intergovernmental Panel on Climate Change (IPCC-AR6, 2023) synthesis report reveals that the average global surface temperature increased by 1.08°C (0.95–1.20) during the period 2011–2020. Other visible evidences reported by IPCC, such as extreme weather events, sea level rise, ocean acidification, and changes in biodiversity, also confirm that climate change has occurred globally. The temperature in Nepal has been found to increase at an annual rate of 0.054°C. The maximum and minimum temperatures are increased at the rate of 0.05°C and 0.03°C annually, respectively. A national-level study reveals similar temperature trends between 1971 and 2014 (DHM, 2017).

Livestock is an integral part of agriculture globally and it contributes to global calorie and protein supplies such as meat, milk, and eggs (FAO, 2009; Thornton, 2010; Thornton and Gerber, 2010). Nepal is an agriculture-dominated country, with livestock being an essential part of the mixed agriculture system. The agriculture and forestry sectors contribute 25.8% to the gross domestic product (GDP) and provide employment opportunities to 73.9% of the total employed population. Similarly, livestock alone contributes 27% of agriculture’s GDP and about 13% of the national GDP (CBS, 2019; MOF, 2021).

Nepal, recognized as “one of the most disaster-prone countries” in terms of climate-related hazards, sees a particularly strong impact on its farmers. Nepalese smallholder farmers are vulnerable to climate change because of high dependence on rain-fed agriculture, limited access to resources, subsistence farming, lack of information (Gentle and Maraseni, 2012; Budhathoki and Zander, 2020), and fragile mountain ecosystem (Wong, 2020). Climate change affects the livestock sector through different channels, including competition for natural resources, quality and quantity of feeds, the emergence of new kinds of diseases, and heat and cold stresses (Rojas-Downing et al., 2017). Thus, farmers need to cope with the effect of climate change through adoption-led mitigating practices (Aryal et al., 2020).

Adaptation in the human system is “the process of adjustment to the actual or expected climate and its effects to moderate harm or exploit beneficial opportunity” (IPCC, 2023). In other words, adaptations are the adjustments made by people in their context to mitigate the negative impact of climate change. Therefore, adaptation is the response to environmental stimuli that affect the systems. Adaptation is supposed to effectively combat climate change, particularly for smallholder and subsistence farmers. Adopting any practice depends on the system’s adaptive capacity, regions, and communities and is primarily influenced by socioeconomic status (Smit et al., 2001). The different means of adaptation at the household level may include the use of adaptive species of livestock, nutritional management strategies, risk management practices, and institutional approaches, i.e., increased accessibility to veterinary services, provision of subsidies, reduction of taxes, and insurance of livestock (Deressa et al., 2008; Rojas-Downing et al., 2017).

Globally, about 43.0% of females are engaged in the agricultural sector indicating the significant role of women in agricultural production, livestock, fishery, and forestry (Akter et al., 2017). Females’ participation in the agricultural sector has been increasing in recent decades in developed countries too (Statistics Canada, 2017; Ball, 2020).

Climate change harms the overall well-being of men and women; however, the impact is perceived to be different among women (Singh et al., 2010; Jianjun et al., 2015). Furthermore, females in developing countries are more vulnerable to climate change impacts and have lower adaptive capacity (Denton, 2002; Goh, 2012). The study of adaptation status to mitigate the effect of climate change should consider the gender dimension because the farming livelihood of small-scale farmers is affected by the use and control over the resources, social practices, socioeconomic status, level of education, access to information, labor division, decision-making process, and ownership of land, which vary among males and females (Bayard et al., 2007; Terry, 2009; Vincent et al., 2014; Jerneck, 2017; Rahman et al., 2023).

Research indicates that class, gender, and culture (Nielsen and Reenberg, 2010; Sultana, 2014;); job other than agriculture and livestock or income from other sectors (McKune et al. 2015); remittance (Bhattarai, 2015); and caste and ethnicity (Ranjitkar, 2020) significantly influence adaptation strategies for both men and women. In addition, education (Diiro et al. 2015), age, land access, farm size (Addison, 2018; Shahid et al. 2021), access to credit, and household size (Gaya et al. 2017) are important factors of adaptation strategies of men and women in agriculture and livestock sectors.

Despite the women’s major responsibilities and significant participation and contribution to rearing livestock in Nepal, their role in livestock production has been underestimated, underrated, and unnoticed. Nepal is a traditional paternal society where men are superior to women. Therefore, the adaptation strategies are shaped by the norms and culture of patriarchal society (Bhattarai, 2020). Furthermore, women’s attitudes and behavior are also conditioned by their culture and society (Upadhay, 2003). A number of studies have been conducted on adaptation practices against the impact of climate change on the agriculture and livestock sector in Nepal. However, there are still inadequate studies on gender differentials in adaptation practices against climate change’s impact on livestock in Nepal. Moreover, gender issues have been seldom discussed in policy-making processes at the national and local levels. Therefore, this study examines the gender differential in adaptation strategies to climate change impact on livestock in western Nepal.

Materials and Methods

Study area

This study was conducted in the Gandaki province of western Nepal. The total area of the province is 21,974 km² and a population of 2,479,745 (1,180,460 men and 1,299,285 women) with a sex ratio of 90.85 (CBS, 2021). The province has a diverse topography, so the temperature varies from subtropical in the Terai (plain) region to an alpine climate in the Himalayan region. There are 11 districts with 85 local governments (one metropolitan city, 26 municipalities, and 58 rural municipalities) (Provincial Policy and Planning Commission of Gandaki Province [PPPC], 2019). The average annual maximum temperature varies between 0.1°C in Manang and 30°C in Nawalpur. Similarly, the average annual minimum temperature varies between −5°C in Manang and Mustang and 30°C in Nawalpur districts. The average rainfall varies between 250 and 2,500 mm per annum. The Himalayan district of Mustang records the minimum rainfall, while Lumle of Kaski district records the maximum rainfall in the province.

Gandaki province was purposively selected for the study because it comprises holdings of major livestock types (cattle, buffalo, Yak/Nac/Chauri (Himali Cow), sheep, goat, and pig). Out of the 11 districts of the province, only five districts (Tanahu, Kaski, Gorkha, Mustang, and Nawalpur) were chosen purposively based on the holding of significant livestock species of Nepal. Tanahu was chosen based on the maximum number of cattle, goats, and pigs, followed by the maximum number of buffalo in Kaski and the maximum number of sheep in the Gorkha district. The remaining two districts (Mustang and Nawalpur) were chosen to represent the province’s local farmers from the Himalayan and Terai (plain) regions. Thus, these five districts covered the entire topography (mountainous Himalayan region, Hill (middle region, and plain Terai region) of the province, from the Himalayas to the Terai.

Sample and population

Only 1,158 households from five districts were interviewed using a multistage random sampling technique. The population of this study was 47,915 households in the sample districts. Since there was no relevant information on farmer’s perception toward climate change impact and adaptation strategies to the livestock in the study area, the value of proportions p and q was assumed to be equal to 0.50 at a 5.0% level of significance with a margin of error of 3.5, resulting a sample of 772. The average value of the design effect 1.5 is used to adjust for the loss of sampling efficiency due to multistage sampling, resulting in a minimum sample size of 1,158 (772 × 1.5). The distribution of the population and sample is presented in Table 1.

Table 1. Distribution of sample across the selected sites of sampled districts.

Study sites	Households	Sample wards	Population	Weights based on HHs	Allocated samples
Mustang					23
Lomanthang rural municipality	532	5	172	1.11	13
Dalome rural municipality	404	5	36	0.84	10
Nawalpur					335
Hupsekot rural municipality	4,588	6	552	9.58	111
Devchuli municipality|	9,256	17	1290	19.32	224
Tanahu					398
Gathering rural municipality	4,137	5	767	8.63	100
Bandipur rural municipality	4,853	6	533	10.13	117
Bhimad municipality	7,488	9	667	15.63	181
Gorkha					163
Ajirkot rural municipality	3,618	5	899	7.55	87
Dharche rural municipality	3,128	7	450	6.53	76
Kaski					239
Machhapuchhre rural municipality	5,512	9	370	11.50	133
Pokhara Metropolitan City (Mauja)	1,027	20	1027	2.14	25
Kalilake	2,610	19	2610	5.42	63
Rupa municipality (Hansapur)	762 |	7	762	1.59	18
Total	47,915		N = 10135	100.00	N = 1158

Source: CBS,915 and self-calculation.

Sampling design

Initially, Gandaki province was selected purposively. Three districts, Kaski, Gorkha, and Tanahu, were chosen based on the higher number of livestock (buffaloes, sheep, goats, and pigs), and Mustang and Nawalpur were chosen to represent the mountain and Terai regions of the province at the second stage. In the third stage, municipalities and rural municipalities were randomly selected, and then, wards were selected randomly. A complete list of households in the wards was obtained from the 2011 population census (CBS, 2011). Households from the selected wards were chosen using a random route sampling technique. Finally, a household head or household member who was 45 years or older had lived in the area for at least 15 years and owned at least one primary livestock at the time of the survey was selected as the ultimate respondent. In case of the unavailability of the household head, a responsible household member who was 45 years and above was chosen for the survey. The minimum age of the respondents was 45 years because the impact of climate change and its impacts can only be perceived during a longer period of time (at least 30 years) so younger respondents were not included in the survey.

Survey instruments

Both structured and unstructured questionnaires were prepared. A structured questionnaire was used for the household survey, while a checklist (guideline) was prepared for focus group discussions (FGDs) to understand the hidden realities that cannot be obtained through survey interviews (Parker and Tritter, 2006). For the validity of the survey instrument, a rigorous review was conducted at the initial stage, followed by a discussion with subject experts about the content of the survey instruments. For reliability, the draft questionnaire was pre-tested in a similar setting to check for any inconsistencies. After pre-testing (10% of the total sample), required modification was made for unclear and poorly constructed statements. In addition, Cronbach Alpha (α) was more significant (greater than 0.92) for the structured questionnaires. Initially, the questionnaire was prepared in English, and back-to-back translation was made into Nepali language. Again, a round table discussion with the experts was held to assess the overall quality of the survey instrument.

The survey questionnaire included three sections. The first section of the questionnaire was about background information, including the availability of land for agriculture and livestock. The second section of the questionnaire consisted of statements related to local farmers’ perception of climate change and its impact on livestock. Forty-four questions were included to assess the farmers’ perceptions toward climate change and its impact on livestock in the questionnaire. Climate change was defined as the perceived change in climatic parameters (average minimum and maximum temperature and rainfall) over 15 years. All perception-related questions were prepared based on a five-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree).

The third section of the questionnaire focused on adaptation strategies that had particular potential in the context of Nepal. Together, 56 questions related to the different dimensions of adaptation strategies were included in the questionnaire. These questions were further divided into seven groups: adapted improved nutritional strategies, changed income strategies, received veterinary, insurance services and subsidies adapted improved feeding strategies, modified shed and changed feeding strategies, minimized risk, diversified the livestock breeding, and adapted stress tolerant breeds. All adaptation-related questions were also prepared using a five-point Likert scale (as in perception). However, the last section of the questionnaire contained issues related to adaptation barriers and gender involvement in livestock rearing.

Statistical analysis

Categorical variables were described using frequency distribution (number and percentage), while quantitative variables were described using descriptive statistics (mean and standard deviation). An independent t-test assesses gender differences in adaptation. Factor analysis was used to decrease the number of variables into a small number of factors for the adaptation strategies. A multiple linear regression model was used to identify the determinants of Adaptation strategies among males and females.

Statistical models

Independent t-test

An independent t-test (to test the significant difference between men and women) was run to assess the mean difference of adaptation strategies between men and women. This test is used to compare the mean between men and women if they differ. The mathematical form of the model is

$$\mathrm{t}=\frac{{\stackrel{─}{\mathit{x}}}_1-{\stackrel{─}{\mathit{x}}}_2}{\sqrt{(\frac{\mathit{s}_1{}^2}{{\mathit{n}}_1}-\frac{\mathit{s}_2{}^2}{{\mathit{n}}_2})}}$$ (1)

Exploratory factor analysis

Exploratory factor analysis was used to reduce the large set of variables into manageable factors. At the initial stage, a Kaiser–Meyer–Olkin measure of sampling adequacy was conducted, and the result was found significant (p < 0.001), indicating that the sample for the factor analysis was adequate. Bartlett’s test of sphericity was also significant (p < 0.001), which permitted running the factor analysis. Only factors with total initial eigenvalues greater than one were considered for further analysis. Out of 56 responses, only 31 were retained during factor analysis. These thirty-one responses were divided into eight components having total initial Eigenvalues greater than 1, explaining 70.41% of the total variance. Principal component analysis was used to extract the factors.

Out of the total components extracted by factor analysis, the first component received veterinary, insurance service, and subsidies, which showed about a 27.0% variation. The second component adapted heat-tolerant breeds and explained 11.42% variation, followed by 9.18% by the modified shed and changed feeding strategies by the third component. The fourth component is livestock diversification strategies, with 6.89%. Similarly, the fifth component is changed income strategies, with a 4.59% variation followed by 4.38% for improved feeding strategies. Furthermore, the seventh component, “risk management strategies,” covers 3.37%, and the eighth component is about improved nutritional strategies. Table 2 shows the eigenvalues and percent of the variation explained by the components of factor analysis.

Table 2. Rotated component matrix showing percent of variation, eigenvalues, and Cronbach Alpha.

Variables	Component
	1	2	3	4	5	6	7	8
Feeding livestock with proteins, vitamins, and minerals								0.850
Feeding livestock with iron, calcium, and iodine								0.849
Participated in livestock insurance schemes								0.604
Involved in credit schemes for livestock					0.687
Adapted income diversification					0.716
Involved in income generation activities					0.820
Involved in cost-reducing activities					0.703
Received veterinary services	0.630
Benefited from decreased tax provision in livestock	0.779
Benefited from subsidy provision in livestock	0.792
Received livestock insurance subsidies	0.802
Received training in the production and conservation of feed	0.661
Use of new species of grass						0.642
Use of a new type of animal feed						0.735
Use of improved animal health technology						0.648
Production of energy-giving crops						0.591
Fed hot Kudo in winter and cold Kudo in summer			0.687
Use of modified roof			0.826
Increment in the height of the roof			0.757
Use of improved shed			0.662
Increased water intake during periods of high heat			0.630
Reduced age to slaughter cattle							0.762
Practiced the extended lactation period							0.770
Minimized disease outbreaks of livestock							0.717
Adapted more fecund genotype				0.807
Adapted faster growing genotype				0.837
Adapted more weighting genotype				0.728
Adapted heat-tolerant breeds		0.643
Adapted cold-tolerant breeds		0.819
Adapted disease-tolerant breeds		0.805
Developed local breeds that can adapt to climate change		0.803
Percentage of variance	27.457	11.423	9.184	6.898	4.595	4.378	3.366	3.170
Cumulative variance	27.457	38.880	48.064	54.962	59.557	63.934	67.300	70.471
Eigenvalue	8.786	3.655	2.939	2.207	1.470	1.401	1.077	1.015
Cronbach alpha (α)	0.871	0.890	0.821	0.899	0.823	0.811	0.757	0.756
Total alpha (α)	0.913

1. Received veterinary, insurance services, and subsidies; 2. Adapted stress-tolerant breed; 3. Modification of shed and feeding strategies; 4. Livestock diversification strategies; 5. Changed Income strategies; 6. Improved feeding strategies; 7. Risk management strategies; 8. Improved nutritional strategies.

Finally, these components were added to get the total adaptation index. This variable has been treated as a dependent variable to run the multiple linear regression.

Multiple linear regression

Multiple linear regression examined the determinants associated with adaptation practices among males and females. The mathematical form of the multiple linear regression model is as follows:

$$\mathit{y}={\mathit{a}}_0+{\mathit{a}}_1{\mathit{x}}_1+{\mathit{a}}_2{\mathit{x}}_2+\mathrm{..........................}+{\mathit{a}}_{\mathit{n}}{\mathit{x}}_{\mathit{n}}+\varepsilon$$ (2)

where a_1, …, a_n are the regression coefficients; i is the dependent variable (adaptation strategies); x_i is the set of independent variables (i = 1, …, n); and ε is the error term.

Based on an extensive literature review, a set of independent variables was chosen for the regression analysis (Hassan and Nhemachena, 2008; Sahid et al., 2021). The leading independent variables selected for the regression analysis are age, sex, household size, educational level, food availability (in months), availability of land for livestock, jobs other than in agriculture and livestock, family type, caste group, and involvement in credit schemes for livestock.

Results

Background characteristics of farmers

Table 3 shows the description of selected background characteristics of farmers. The mean age of farmers was 54.56 years, with an average household size of 5.19, which is greater than the national average (4.32) (CBS, 2021). The mean educational attainment of the sample farmer was close to the basic level (more than 12.0% were illiterate, while 33.1% had informal education, 29.0% had basic level, and about 25.0% had secondary and higher level education). The average food availability among the farmers was 8.38 months. Likewise, the average land available for agriculture was 0.33 hectares, compared to 0.13 hectares for livestock farming among the farmers (the availability of land among females was less than for males). Less than half of the respondents had agriculture as the main occupation, and more than half of the farmers were from the upper caste and nuclear families.

Table 3. Selected socioeconomic characteristics of farmers.

Characteristics	Male Mean (SD)	Female Mean (SD)	Total Mean (SD)
Age of respondents (years)	55.39 (9.19)	53.11 (8.93)	54.56 (9.16)
Household size	5.38 (2.31)	4.86 (2.04)	5.19 (2.23)
Education (1 = illiterate, 2 = literate (informal), 3 = basic level (1–8), 4 = secondary (9–12), and 5 = bachelor and above)	2.90 (1.05)	2.45 (1.05)	2.73 (1.07)
Food availability (in months)	8.82 (5.31)	7.64 (3.70)	8.38 (4.69)
Availability of land for agriculture (hectare)	0.36 (0.13)	0.28 (0.21)	0.33 (0.11)
Availability of land for livestock (hectare)	0.15 (0.08)	0.09 (0.05)	0.13 (0.08)
*Jobs in other than agriculture and livestock
No other job (0)	339 (46.1%)	216 (51.2%)	555(47.9%)
Jobs in other areas (1)	397 (53.9%)	206 (48.8%)	603 (52.1%)
*Caste group
Upper caste (coded as 1)	437 (59.4%)	195 (46.2%)	632 (54.6%)
Janajati (coded as 2)	243 (33.0%)	162 (38.4%)	405 (35.0%).
Others (coded as 3)	56 (7.6%)	65 (15.4%)	121 (10.4%)
*Family type
Nuclear (1)	374 (50.8%)	256 (60.6%)	630 (54.4%)
Extended (2)	362 (49.2%)	166 (39.3%)	528 (45.6%)
Total sample (n)	736	422	1158

^*values are numbers and percentages in parentheses.

An independent t-test for the difference in the means of selected background characteristics between males and females showed a significant age difference (t = 4.14; p <0.01), household size (t = 3.90; p < 0.01), education (t = 7.02; p < 0.01), availability of land for agriculture (t = 4.08; p < 0.01), livestock (t = 3.4; p < 0.01), and food availability (t = 3.92; p < 0.01) and gender. In addition, a chi-square test of independence was also run to assess the association between gender and other selected categorical variables (education, jobs other than agriculture, caste group, and family type). This showed that education (chi-square = 56.82; p < 0.01), caste group (chi-square = 26.32; p < 0.01), and family type (chi-square = 10.48; p < 0.01) were significantly associated with the gender of the respondent.

Farmer’s perception of climate change

About two-thirds (66.8%) of the farmers agree that the annual average maximum temperature has increased. Similarly, 71.5%, followed by 66.7%, report that the timing and amount of rainfall have changed in the last 15 years. Farmers also perceive a significant increase in the frequency of drought, floods, and incidence of diseases in the agriculture and livestock sector in the last 15 years. The chi-square test of independence assesses the gender differential in perception toward climate change. The analysis reveals that there is a significant association between the decrease in annual average minimum temperature (chi-square = 15.84; p < 0.01), increase in drought incidence (15.31; p < 0.01), and increase in the incidence of livestock diseases (chi-square = 13.87; p < 0.01). This indicates that males and females perceive climate change differently.

Farmer’s perception of the impact of climate change on livestock

Figures 1 and 2 show males’ and females’ perceptions about the impact of climate change on livestock, respectively. More than 41.0% of males perceive that climate change is causing a decrease in livestock weight, productivity, and number. Nearly 63% and 60.5% of males report an increase in the incidence of diseases and morbidity (Fig. 1).

Fig. 1. Male perception about the impact of climate change on livestock.

Fig. 2. Female perception of the impact of climate change on livestock.

Similarly, more than half of females also perceive that climate change impacts livestock’s decrease in weight and the chance of morbidity. As in males, about one-fourth of females are neutral and state that the impact of climate change on livestock neither agrees nor disagrees.

Gender difference in adaptation strategies

Factor analysis has extracted eight components as the significant strategies of adaptation. These components have been considered as dependent variables and gender as independent variables. An independent sample t-test was conducted to examine the significant difference in adaptation strategies by gender. Table 4 shows the t-values, their degrees of freedom, and p-value. The analysis reveals that there is no significant difference in adaptation strategies by gender. It indicates that gender does not play a role in adaptation strategies. This result is also supported by the finding of FGDs where the participants report that the decision of adaptation is jointly made in the family.

Table 4. Independent sample t-test showing the gender differences in adaptation strategies among farmers.

Adaptation strategies	t-statistic	DF	Significance (p)
Received veterinary, insurance services, and subsidies	0.705	1156	>0.05
Adapted stress-tolerant breed	0.308	1156	>0.05
Modification of shed and feeding strategies	-1.801	1156	>0.05
Livestock diversification strategies	1.131	1156	>.05
Changed income strategies	0.552	1156	>0.05
Improved feeding strategies	1.584	1156	>0.05
Risk management strategies	0.607	1156	>0.05
Improved nutritional strategies	0.095	1156	>0.05
Overall adaptation	1.682	1156	>0.05

Determinants of adaptation strategies

Multiple linear regression was conducted to identify the determinants of adaptation practices among males, females, and total. Table 4 shows the coefficients and p-values of regression analysis of adaptation practices among males, females, and farmers. All models’ overall goodness of fit (F-values) is statistically significant, with an acceptable coefficient of multiple determination (more than 20.0% for all models). Socio-economic and demographic factors are the determinants of adaptation strategies among farmers.

As shown in model 1 of Table 4, the coefficients of age, years of schooling, availability of foods, availability of land for livestock, jobs other than agriculture and livestock, caste group, and access to credit of livestock are statistically significant.

Table 5 also shows the determinants of adaptation strategies to climate change on livestock by gender (males and females). Model 2 (Table 5) illustrates that age, years of schooling, availability of land for livestock, jobs other than agriculture and livestock, caste group, family type, household size, and access to credit on livestock are the significant determinants of adaptation strategies among males.

Table 5. Determinants of adaptation to climate change on livestock among male, female, and both.

Predictors	Model 1 (both)		Model 2 (male)		Model 3 (female)
	Standardized coefficients	p-value	Standardized coefficients	p-value	Standardized coefficients	p-value
Age	0.057	<0.05	0.108	<0.01	−0.032	>0.05
Gender	0.048	>0.05
Years of schooling	0.098	<0.01	0.132	<0.01	0.068	>0.05
Availability of food (months)	−0.029	<0.05	−0.033	>0.05	−0.094	<0.05
Availability of land for livestock (months)	0.126	<0.01	0.164	<0.01	0.049	>0.05
Job other than agriculture and livestock	0.130	<0.01	0.114	<0.01	0.141	<0.01
Family type	0.060	>0.05	0.112	<0.01	−0.003	>0.05
Household size	−0.056	>0.05	−0.101	<0.05	0.023	>0.05
Caste group	0.091	<0.01	0.147	<0.01	−0.044	>0.05
Access to credit on livestock	0.420	<0.01	0.403	<0.01	0.410	<0.01
R square	0.237		0.289		0.248
F statistic	30.991	<0.01	23.099	<0.01	12.489	<0.01
Sample size (n)	1158		736		422

Furthermore, model 3 (Table 5) presents the determinants of adaptation strategies to climate change impact on livestock by females. Among the predictors included in the model, only availability of food, jobs other than agriculture and livestock, and access to credit on livestock are the significant predictors.

Discussion

Livestock is one of the major sources of livelihood for smallholder farmers in developing countries such as Nepal. Although females have a significant contribution in the livestock sector, they are less acknowledged in Nepal (Upadhyay, 2003). The impact of climate change is a serious threat to livestock production and productivity. Males and females face the impact of climate change differently (Singh et al. 2010). The role of gender in adaptation has also been changing in the context of climate change (Jianjun et al., 2015).

This study shows significant differences in selected background characteristics by gender. These findings are also consistent with the results of the studies by Akter et al. (2016) and Jianjun et al. (2015). Factor analysis identified eight significant strategies adapted by the farmers and it was noted that gender did not show any significant differences in adaptation strategies. A Nigerian study by Olutegbe (2019) also claims that gender does not matter in the adaptation strategies of agriculture farmers. However, other studies reveal that gender is a significant factor in adaptation strategies (Deressa, 2009).

A multiple linear regression model was used to identify the significant determinants of adaptation strategies. Different socio-economic factors have been identified as the key determinants of adaptation strategies of males and females. The age of farmers positively impacts adaptation strategies, indicating that older farmers are likelier to adopt different strategies than younger ones. The result of FGDs is also consistent with the findings of quantitative analysis. One possible reason may be that older farmers might have better experience handling the climatic and nonclimatic hazards in the livestock sector than younger farmers and older farmers may have known better options of indigenous knowledge to handle the worst situation. A Ghanaian study by Guodaar and Asante (2018) supports this finding. However, this result is opposed by the findings of a Nepalese study conducted by Khanal and Wilson (2018). As in the t-test, gender does not show a significant influence on adaptation strategies in multiple linear regression analysis. Conversely, it is a significant predictor of adaptation intensity among smallholder rice producers in Chitwan of Nepal (Upendram et al., 2023) and Habtemariam et al. (2020) study of adaptation decisions among Ethiopian smallholder agricultural farmers. Farmers with higher years of education are more likely to adopt different strategies. It may be because of the reason that education enhances the farmers’ knowledge of adaptation strategies and their benefits. According to Alauddin and Sarker (2014) and Deressa et al. (2009), the increased level of education enhances the adaptive capacity of the farmers.

This study indicates that the availability of food is negatively related to adaptation strategies. The analysis of qualitative findings also supports this finding. Participants of FGDs having enough food availability were not eager to adapt any strategies due to food security. The availability of land for livestock also has a positive and significant impact on adaptation strategies among farmers in the study area. Only a few farmers in the study area were engaged in other jobs than agriculture and livestock. The finding is clear that jobs other than agriculture and livestock sectors significantly have a significant positive impact on adaptation strategies. The findings of FGDs also support this claim. One of the participants reported that farmers with jobs other than agriculture and livestock (alternate income) had capacity to buy new technology and other materials for adaptation. This finding is also consistent with the findings of other studies (Alauddin and Sarker, 2014; Mersha and Laerhoven, 2015).

Different caste groups were involved in livestock farming in Gandaki province. The caste group [upper caste (Brahmin, Chhetri, Thakuri, and Giri), Janajati (Gurung, Magar, Tamang, and Magar), and others (Dalits and other minorities)] is also a significant determinant of adaptation strategies. Furthermore, access to credit for livestock motivates farmers toward adaptation and it has shown a considerable positive impact on adaptation strategies.

Table 5 also shows that males from extended families are more likely to adopt strategies compared to nuclear families and household size is negatively associated with adaptation strategies among males. This result is also consistent with qualitative findings. Most male participants in FGDs from extended families report that they are not serious about adaptation because they rely on other family members already involved in income-generating activities, such as nonfarm work. However, household size is positively related to adoption of vaccination practices among Vietnamese poultry farmers (Qui et al. 2024).

Adaptation is likely to decrease as there is an increase in the availability of food among females. Females who have jobs other than agriculture and livestock are also more likely to adopt different strategies. This finding is consistent with the male and overall model. This result is also parallel with the findings of qualitative analysis. Access to credit for livestock is positively related to adaptation strategies among female farmers. The probable reason may be that access to credit could enhance their financial capacity to adopt new technologies and strategies.

Conclusion

Adaptation to climate change is a multifaceted process affected by both climatic and nonclimatic factors. Gender-based differences in adaptation strategies are observed due to the social construction of gender, which determines the roles and responsibilities of males and females. Commonly, there are significant differences in access to resources and power among males and females, which shape adaptation strategies differently. Although Nepal is a patriarchal society where the role of males is assumed superior to females and there is a significant difference in some selected background characteristics of respondents of the province, the findings of this study do not show significant differences in adaptation strategies by gender (males and females) but the odds of adoption is higher among males than females. This may be due to the role of situational factors that were not considered in the study.

Age, years of schooling, availability of food and land for livestock, caste group, and access to subsidies on livestock are the determining factors of adaptation strategies among farmers. This study highlights that some of the determinants of adaptation strategies differ between males and females. Age, years of schooling, availability of food, availability of land for livestock, family type, household size, job other than agriculture and livestock, caste group, and access to subsidies on livestock are significant determinants of adaptation strategies among males. Meanwhile, the availability of food, jobs other than agriculture and livestock, and access to credit on livestock are determining factors of adaptation strategies among females. These results reveal that some determinants of adaptation strategies among males and females are different. Therefore, the provincial and local governments should consider gender dimensions while formulating policies and strategies at the provincial level.

Jobs other than agriculture and livestock (proxy measure of income) and access to credit on livestock are the key determinants of adaptation strategies among farmers (males and females). This study points out that jobs other than agriculture and livestock enhance the adaptive capacity of males and females. Creating additional job opportunities other than agriculture and livestock may be another option for improving the adaptive capacity of local people. Access to credit on livestock is another key determinant of adaptation options among Nepalese farmers. The accessibility of the credit scheme is not favorable to smallholder farmers in Nepal because of administrative hurdles. Therefore, the concerned authority of provincial and local governments should pay attention to formulate effective policy measures toward the accessibility of credit for smallholder farmers.

Conflict of interest

All the authors have no conflict of interest.

Funding

Research Directorate, Tribhuvan University, under the national priority project (Award No.: TU-NPAR-077/78-ERG-08).

Authors’ contribution

Vikash Kumar KC designed the concept, prepared the entire draft, and finalized (introduction, methodology, results and discussion, and conclusion) the manuscript. Ananta Raj Dhungana analyzed the data and Purna Bahadur Khand collected the relevant materials.

Data availability

Data are provided in the manuscript.