Sorghum farmers’ perceptions of climate change, its effects, temperature and precipitation trends, and determinants of adaptation strategies in the central plateau zone of Rwanda

Abstract

Climate change is causing significant damage to crop production in the central plateau zone of Rwanda, particularly affecting sorghum, food, and the incomes of smallholder farmers. Understanding farmers’ perceptions and the factors impacting their responses is crucial for improving sorghum production policies and programs. Therefore, a study was conducted to assess sorghum farmers’ perceptions of climate change and the factors determining their adaptation strategies. A multistage sampling method and a cluster random selection were utilized to select 345 respondents from five districts of the study area. The data were analyzed using descriptive statistics and a multivariate probit model. The results showed that 98.8% of farmers were aware of climate change, with deforestation being the main anthropogenic activity causing it. Consequently, 95.7% and 84.3% of farmers experienced grain yield reductions, and over 20 sorghum varieties disappeared. To address these impacts, farmers adopted five adaptation strategies: early maturing sorghum varieties (67%), adjusting planting dates (50.1%), drought-tolerant varieties (46.7%), soil conservation practices (38.3%), and crop diversification (32.8%). The multivariate probit model results showed the age and literacy level of the household head, access to extension services, access to information, access to credit, farming experience, and land size as the important factors influencing at least one of the climate change adaptation strategies. The study concluded that sorghum farmers are aware of the impacts of climate change and are acting to address its negative effects. The results suggest that the government and stakeholders should support farmers in strengthening their adaptation strategies for sustainable sorghum production.

Keywords:

• Adaptation strategies
• baseline survey
• multivariate prohibit model
• smallholder farmers
• sorghum bicolor
• Rwanda
• climate change

Reviewing Editor:

• Manuel Tejada, Universidad de Sevilla, Spain

Subject:

• Agriculture & Environmental Sciences
• Environmental Change & Pollution
• Biodiversity & Conservation

1. Introduction

Climate change is negatively impacting agriculture worldwide by reducing crop yields and the quality of harvests (Portner et al., 2022). It is also affecting crop productivity in Africa, where agriculture relies mainly on the existing climate and weather conditions (Hundera et al., 2019). In Rwanda, climate change is also impacting crops since the agriculture sector is particularly vulnerable to climate and weather related consequences like prolonged droughts, floods, hailstorms, and erratic rainfall (WB & CIAT, 2015). Drought stress induced by climate change is the most severe abiotic constraint that threatens food security in Rwanda (Gasheja & Gatemberezi, 2017).

Among the crops affected by climate change, sorghum is an important crop for smallholder farmers (Niyibigira, 2012), who lack the resources to adapt to the impacts of climate change (NBDF, 2014). Rwanda is among nine countries in eastern and southern Africa where water deficit is the most important constraint in sorghum production, causing more than 2 million kg/year of yield loss (Wormann et al., 2006). This water shortage is worsened by drought caused by climate change (WBG, 2021). Climate change affects sorghum through temperature fluctuations, changes in precipitation patterns, seasonal variations, and drought stress. Climate change induced high temperatures disrupt sorghum crop growth and development, causing reduced yield due to disturbances in cell division, metabolism, photosynthesis, and nutrient uptake (Ndlovu et al., 2021; Prasad et al., 2021). Also, extremely low temperatures delay flowering, induce sterility in pollen development, affect pollination, decrease seed per panicle, and decrease grain yield (Emendack et al., 2021). Too much precipitation damages sorghum crops by saturating the soil and removing air, which is important for plant growth, thus reducing grain yield. Additionally, strong winds and high humidity affect sorghum by causing lodging and fungal infections, respectively (Sharma et al., 2019). Drought negatively affects germination, emergence, plant height, leaf area, and photosynthesis. It also causes premature sorghum plant death, lodging, a reduction in seed size, and yield losses (Assefa et al., 2010; Derese et al., 2018).

Sorghum is one of the key food crops in Rwanda. For example, in 2022, the area under sorghum production was 177,261 ha, making it the fourth most important staple crop after common bean (634,365 ha), cassava (338,674 ha), and maize (301,022 ha) among seasonal crops (NISR, 2022). It is commonly used for health reasons (Ezeanya-Esiobu et al., 2018). Babies are fed with sorghum porridge, believed to be rich in various vitamins. In rural areas, casual laborers and school children take sorghum drinks (porridge, non-alcoholic or alcoholic) early in the morning as their breakfast, and a great number of them can spend the whole day working without eating anything else until their evening meal (Ezeanya-Esiobu et al., 2018). Sorghum is also a source of income for smallholder farmers, and it helps them to cover their daily needs (Niyibigira, 2012).

Despite the numerous uses and benefits of sorghum in Rwanda, the production of sorghum is below its potential level. At the farmers’ level, sorghum grain yield is around 1006 kg. ha⁻¹ (FAO, 2021), whereas potentially it yields about 3600 kg. ha⁻¹ at research stations in Rwanda (OAF, 2016). The low yield is a consequence of numerous biotic and abiotic problems induced by the current global climate change (Gasheja & Gatemberezi, 2017; Niyibigira et al., 2013). For instance, the total sorghum production decreased from 234,763 MT in 2013 (NISR, 2015) to 192,427 MT in 2022 (NISR, 2022).

To address these effects of climate change on sorghum, policymakers and intervention agencies should promote adaptation strategies to be adopted by farmers (Kumar et al., 2020). According to (Dossou-Aminon et al., 2015; Teshome et al., 2021; Likinaw et al., 2022), the adaptation strategies frequently used by smallholder farmers comprise early maturing varieties, crop diversification, adjusting planting dates, soil and water conservation practices, drought-tolerant varieties, crop rotation, and intercropping. These climate change adaptation strategies are occurring in dynamic socioeconomic, political, and biophysical situations that change through time and space. Therefore, smallholder farmers’ responses to these climate change adaptations are influenced by socioeconomic, demographic, and institutional factors (Mihiretu et al., 2020). Thus, identifying factors that determine the farmers’ climate change adaptation strategies could improve the adoption level of these strategies by removing all the barriers that impede the use of suitable strategies. In the past, many scholars (Fosu-Mensah et al., 2012; Mbwambo et al., 2022; Teshome et al., 2021) worked on this subject. They used different econometric models to identify different demographic, institutional, and socioeconomic factors that determine the adoption of climate change adaptation strategies. Their findings highlighted access to information, farming experience, educational level, farm income, contact with extension services, training, and access to credit as the main determinants of the adoption of climate change adaptation strategies. They also reported off farm income, association membership, age, gender, and household size. Then again, few researchers reported household food security status (Ngetich et al., 2022), market information, and distance to the village market (Esayas et al., 2019) as other factors affecting the adoption of climate change adaptation strategies.

On the other hand, farmers need to know the reality of climate change to adopt those adaptation strategies (Meldrum et al., 2017). In this respect, farmers’ perceptions of climate change influence the adoption of existing adaptation strategies as well as their usual agricultural activities (de Matos Carlos et al., 2018; Meldrum et al., 2017). Thus, the first condition for the successful implementation and improvement of climate change adaptation policies is to understand farmers’ perceptions of climate change, the existing adaptation strategies, and the socioeconomic, demographic, and institutional factors influencing their adoption (de Matos Carlos et al., 2018; Fierros-González & López-Feldman, 2021). In Rwanda, few studies on farmers’ perceptions and factors influencing adaptation strategies were conducted, but all of them focused on major crops. Butera et al., (2022), Byishimo (2017), Dusingizimana (2018), and Nyirandorimana et al. (2020) reported the negative effects of climate change on coffee, rice, common beans, maize, and potatoes. Although these researchers assessed farmers’ perceptions and determinants of adaptation strategies, no study has been conducted on sorghum, specifically in this study area of the central plateau zone of the country. Additionally, no one among those studies used a multivariate model that considers interdependence among adaptation strategies.

Additionally, several studies have been conducted on the effects of climate change on sorghum in various parts of the world and Africa. Knox et al., (2012) reported that there is evidence that climate change impacts sorghum production in Africa and Asia. Due to drought, an increase in warm temperatures, and a rise in night temperatures, sorghum will experience a significant yield decrease in India (Sandeep et al., 2018). In Africa, sorghum will also experience yield losses of around 15%. In semi-arid regions of Cameroun, climate change induced drought, poor rainfall distribution, rising temperatures, a decrease in the number of rainy days, and an increase in extreme weather events affect annual sorghum productivity (Tene, 2022). In the northeastern part of Benin, sorghum is seriously affected by climate change, and as a result, there is low productivity, an increase in insects, and biodiversity loss (Dossou-Aminon et al., 2015). According to (Ali et al., 2023), climate change caused 50% of sorghum total production loss in Somalia, while in Somaliland, a climate change induced drought caused a decrease in sorghum production and yield, reducing sorghum farmers’ income and raising food prices. In Ethiopia, the extreme and low rainfall resulting from climate change affects sorghum yield negatively in the highlands of the country (Eggen et al., 2019). Although the effects of climate change on sorghum are global and continental issues, the severity and adverse effects on sorghum differ significantly across regions, countries, and specific climate conditions (Mihiretu et al., 2020). Therefore, it is essential for Rwandan policymakers to identify the perceived changes at the zonal level of the central plateau zone of Rwanda to allow farmers to know how the climate is changing and make the best adaptation measures.

So, there is a lack of information on the farmers’ perceptions, the effects of climate change on sorghum production, adaptation strategies, and factors influencing these strategies in Rwanda. Consequently, it is very difficult to formulate improved adaptation strategies to help sorghum farmers cope with and adapt to the current, continuous climate change. Thus, this study was conducted to assess the farmers’ perceptions of climate change, climate change effects on sorghum production, farmers’ adaptation strategies, and factors determining those adaptation strategies in the central plateau zone of Rwanda. The results of this work could be helpful to policymakers and partners by setting appropriate adaptation strategies and increasing the adoption rate of the existing strategies, as well as to farmers by increasing sorghum yield.

1.1. Determinants of farmers’ adaptation strategies

The capacity of farmers to adapt to various climate change adaptation strategies is influenced by many factors, including demographic and socioeconomic factors (age, sex, and education level), farm characteristics (land size of the farmer), and institutional factors (access to extension services, access to climate information) (Dang et al., 2019). Age was reported as one of the main factors that influence climate change adaptation strategies. Teshome et al. (2021) reported that age positively influenced almost all the adaptation strategies among maize farmers in the eastern parts of Ethiopia. The old age of the farmers is associated with knowledge and wisdom, which enable them to practice different adaptation strategies (Tazeze et al., 2012). It is believed that older farmers accumulate much farming experience through observation and long term exposure to weather conditions, and they are more likely to understand the necessity and significance of applying different adaptation strategies (Anum et al., 2022). On the other hand, it was reported that older farmers tend to be traditional, and it appears that they find it challenging to adopt new technologies. For instance, it was found that adopting improved soil conservation practices was negatively influenced by age (Dang et al., 2019; Gebre et al., 2023).

Sex is also one of the main factors that influence adaptation strategies. The sex of farmers was found to positively influence farmers’ adaptation decisions in the Northern West of Ethiopia (Asrat & Simane, 2018) and the adoption of crop diversification and drought tolerant crops in Kenya (Gebre et al., 2023). According to (Dang et al., 2019), males and females make dissimilar adaptation choices because of their different ways of thinking. Male farmers are more willing to take risks to adopt new technologies, while female farmers, on the other hand, favour safe and conventional practices.

The education level of the farmers positively influenced climate change adaptation strategies in the northern parts of Ghana (Fagariba et al., 2018). Also in Pakistan, it was found that adapting to climate change was positively correlated with the increase in years of schooling (Abid et al., 2015). Educated farmers are more likely to have a high level of reasoning, have more access to new technologies, and tend to better recognize the risks associated with climate change, hence causing them to adopt (Asrat & Simane, 2018). Additionally, educated farmers can better interpret climate information than non educated ones (Atchikpa et al., 2017). Higher education levels among household members enhance the required skills and knowledge to respond to the different effects of climate change (Gebru et al., 2020).

Another determinant of climate change adaptation strategies is the size of the family (Anum et al., 2022). found that family size positively determines the adoption of different climate change adaptation strategies among smallholder farmers in southern Ghana. The same factor was reported in Kenya, where the size of the family was found to be positively correlated with the use of drought-tolerant varieties and crop diversification (Gebre et al., 2023). An increase in productive family members positively influences the adoption of adaptation strategies through enhancements in labour resources, which are necessary for new farming practices that require intensive labour (Belay et al., 2022). On the other hand, family size was reported to negatively affect the adoption of agroforestry, improved agronomic practices, and resilient varieties in the Sahelian region of Niger (Zakari et al., 2022).

Farming experience is also one of the main factors that influence adaptation strategies. Farming experience in Pakistan was reported to determine the adoption of adaptation strategies by increasing the probability of adjusting the planting date and using new crop varieties (Abid et al., 2015). According to (Alam, 2015), the more farmers spend many years in agricultural activities, the more they have a better understanding of the weather and the climate, which helps them adjust to the associated risks.

The land size determines the approach of the farmers towards climate change adaptation strategies in Pakistan, especially crop and variety diversification (Abid et al., 2015), while in Kenya, the same factor influences positively the adoption of early maturing crop varieties and crop diversification (Gebre et al., 2023). Also (Gebre et al., 2023), reported that farmers with large land sizes have more financial capability to invest in various climate change adaptation strategies. In contrast to this, the farm size of the farmers was found to negatively influence the adoption of climate resilient varieties using amended agronomic practices (Zakari et al., 2022).

Off farm income was also reported as one of the determinants of the adoption of climate change adaptation strategies. According to (Atchikpa et al., 2017), climate change adaptation strategies were positively influenced by off farm activities among farmers in Benin and Nigeria. Off farm income was also found to be positively associated with different climate change adaptation strategies such as changing planting dates, irrigation, and planting trees (Dang et al., 2019).

Access to climate information was found to enhance the probability of the adoption of adaptation strategies by 5.7% in the northwest of Ethiopia (Asrat & Simane, 2018). Access to weather information was also reported to be positively associated with the adoption of different climate-smart agriculture practices, such as agroforestry and irrigation. For instance, farmers who had access to climate information were 33.8% more likely to do so than those who did not embrace agroforestry (Kifle et al., 2022). Having access to weather information helps farmers plan ahead and enhance their preparedness for diverse climate change shocks and consequences (Fagariba et al., 2018).

Another important determinant of the adoption of adaptation strategies is contact with extension agents. According to (Asrat & Simane, 2018), respondents’ choice of climate change adaptation strategies was strongly determined by extension agents’ guidance in Ethiopia. Also, farmers’ ability to adapt to climate change was found to be positively correlated with access to extension services. Asrat and Simane (2018) and Fagariba et al. (2018) reported that access to this extension agent’s information boosts behaviour and investment in adaptation strategies.

The positive influence of cooperative membership on the adoption of different climate change adaptation strategies was also reported in Ethiopia and Bangladesh (Aryal et al., 2020; Gebru et al., 2020). Also, knowledge and experience sharing among members of a cooperative influence positively the behaviour of farmers towards climate change adaptation strategies (Swe et al., 2015).

Credit access was reported to be one of the most important determinants of adaptation strategies in different parts of the world (Dang et al., 2019; Fosu-Mensah et al., 2012). For instance, in Benin and Nigeria, farmers’ adaptation strategies to climate change were positively influenced by access to credit (Atchikpa et al., 2017). Having access to credit helps with the adoption of adaptation strategies, as farmers can afford to pay for the cost of different improved technologies (Zakari et al., 2022). This positive association between access to credit and adaptation strategies is contrasted by the report of (Ngetich et al., 2022), who found that access to credit was negatively correlated with staggering planting dates and changing crop varieties in Kenya.

2. Materials and methods

2.1. Study area

The study was conducted in the central plateau zone of Rwanda in a major sorghum growing area covering Nyanza, Ruhango, Gisagara, Kamonyi, and Huye districts, especially in Musha, Rusatira, Kibilizi, Kinazi, and Mugina sectors. The central plateau zone where the study was carried out is located between 1°50′–2°50′S latitude and 29°35′–30°0′E longitude (Figure 1).

Figure 1. Map showing the study area.

This zone is situated in the centre of the country, with an average annual rainfall of 1200–1400 mm (Baligira, 2008). It is characterized by mixed farming agriculture with the main production of banana, sorghum, common beans, cassava, maize, and sweet potato (Pautrizel, 2014). Like in the rest of the country, this area has four climatic seasons: a short dry season (January–February), a long rainy season (Mid March–May), a long dry season (June–August), and a short rainy season (September–December) (Umuhoza et al., 2021).

2.2. Sampling procedures and data collection

The study used a multistage sampling procedure with both purposive and random selection procedures to identify the sample households. The first stage was the selection of the central plateau zone of the country purposively based on the popularity of sorghum production. The second stage was the selection of five districts in the central plateau zone of Rwanda, purposively again based on the popularity of the sorghum crop. The third stage was the selection of five sectors in the five districts, again based on the popularity of the sorghum crop, with the assistance of district and sector agricultural officers. Then, two cells were also selected from each sector, once again focusing on cells where sorghum is more popular. The cell is Rwanda’s smallest local administration entity, with a paid public staff in charge of agricultural activities.

The sample size consisted of 345 farmers who were selected using a cluster random selection from 2535 households (provided by the selected sector offices) of sorghum farmers, considering each district, sector, and cell as a cluster. The households of sorghum farmers were used as a sampling frame, and a list frame was utilised to select the households randomly. To have the list, the focus was on the farmers who grew sorghum annually, excluding those who did not grow it regularly. The sample size was determined using the formula provided in Equation (1)(1) $$\mathrm{n}=\frac{\mathrm{N}}{1+\mathrm{N}\left({\mathrm{e}}^2\right)}$$(1) (Yamane, 1967): (1) $$\mathrm{n}=\frac{\mathrm{N}}{1+\mathrm{N}\left({\mathrm{e}}^2\right)}$$(1) where n = sample size; N = population size; and e = level of precision at 0.05.

Finally, the sample size of each cell was defined using the method provided in Equation (2)(2) $$\left(\mathrm{A}/\mathrm{B}\right)\times \text{C}$$(2) (Cochran, 1977): (2) $$\left(\mathrm{A}/\mathrm{B}\right)\times \text{C}$$(2) where A is the total number of households growing sorghum in the cell, B is the total number of households producing sorghum in the ten cells, and C is the sample size determined in the study area (Table 1).

Table 1. Distribution of sample households by cells.

Sample Districts	Sample Sectors	Sample cells	TNHGS	Sample Size
Ruhango	Kinazi	Rubona	169	23
		Gisali	125	17
Gisagara	Musha	Cyimana	142	19
		Gatovu	196	27
Kamonyi	Mugina	Mugina	277	38
		Nteko	289	39
Huye	Rusatira	Kiruhura	340	46
		Gafumba	292	40
Nyanza	Kibirizi	Mututu	381	52
		Cyeru	324	44
Total			2535	345

TNHGS: Total Number of Households Producing Sorghum.

Using a structured questionnaire associated with the objectives, data were collected by trained and local language speaking enumerators who have adequate knowledge of crop production in order to collect relevant data from the respondents. Before the actual survey, a pre-test survey was conducted in each of the 10 selected cells (one household per cell) with the aim of testing questionnaire design, data collection, survey feasibility, and data processing. The pre-test survey results presented acceptable Cronbach’s alpha values, as presented in Table 2, indicating the internal consistency of the questionnaire. Also, based on these pre-test results, questionnaire modifications were performed to confirm its validity by increasing its accuracy, correctness, meaningfulness and its usefulness. (Kimberlin & Winterstein, 2008) suggested that it is important to know that validity and reliability are not separate qualities since they are connected in their purpose. They also argue that a measurement cannot be valid unless it is reliable, and it must be both reliable and valid to be an accurate representation of a concept or attribute. This data collection was undertaken in December 2019.

Table 2. Reliability coefficient values of different climate change components.

Climate change components	Reliability coefficient value
Awareness and causes of climate change	0.91
Temperature and rainfall trends	0.84
Shocks and coping strategies	0.86
Effects on sorghum	0.74
Adaptation strategies	0.71

Data on farmers’ characteristics, climate change awareness, climate change causes, farmers’ perceptions on temperature and rainfall trends over 30 years, climate change shocks affecting sorghum and coping strategies, climate change effects on sorghum, and farmers’ adaptation strategies were collected. Additionally, a literature review was done, and temperature and rainfall data in the study area over 30 years (1989–2018) were collected from the Rwanda Meteorological Agency for comparison between farmers’ perceptions and observed climate data. During the data collection, farmers were informed about the study’s purpose and importance, asked to participate voluntarily, kept their responses confidential, and had no wrong answers. They were also guaranteed privacy by omitting their names on questionnaires, and no vulnerable groups like children or mentally disabled farmers were interviewed.

2.3. Data analysis

2.3.1. Descriptive analysis

Descriptive statistics such as frequency, percentage, mean, maximum, minimum, and standard deviation were used to analyze the data. To monitor the outliers greater or less than a threshold value and set them to the limit corresponding to ± 1.5 × IQR, the Tukey fence was utilized (Ngongondo et al., 2011).

Also, Von Neumann ratio (VNR) tests (Neumann, 1941), the Buishand range test (BRT) (Buishand, 1982), Alexandersson’s Standard Normal Homogeneity Test (SNHT) (Alexandersson, 1986), and the Pettitt test (Pettitt, 1979) were employed to inspect the data for homogeneity. These tests were carried out under two hypotheses: Ha (alternative hypothesis), there exists a day at which there is a change in the data, and H0 (null hypothesis), the data are similar. Again, the autocorrelation function (r₁) was utilized to check the time-series data for independence and randomness, as given in Equation (3)(3) $$r_1=\frac{{\sum }_{i=1}^{n-1}\left(x_i-\overline{x}\right)\left(x_{i+1}-\overline{x}\right)}{{\sum }_{i=1}^n{\left(x_{i-}\overline{x}\right)}^2}$$(3) (Storch, 1995): (3) $$r_1=\frac{{\sum }_{i=1}^{n-1}\left(x_i-\overline{x}\right)\left(x_{i+1}-\overline{x}\right)}{{\sum }_{i=1}^n{\left(x_{i-}\overline{x}\right)}^2}$$(3) with x representing the mean of the time series, $x_i$ an observation, $x_{i+1}$ the following observation, and n the number of data. Utilizing a two-tailed test at a 95% confidence interval, the autocorrelation coefficient value of r₁ was checked against the null hypothesis, as provided in Equation (4)(4) $$r_1=\frac{-1\pm 1.96\sqrt{\left(n-2\right)}}{n-1}$$(4) . (4) $$r_1=\frac{-1\pm 1.96\sqrt{\left(n-2\right)}}{n-1}$$(4)

In the method described by (Partal & Kahya, 2006), a pre-whitened approach was utilized every time a significant correlation arose in the data series, and that was acquired as explained in Equation (5)(5) $$\left(x_2-r_1{x}_1,x_3-r_1{x}_2,\dots ,{\text{x}}_{\text{n}}-r_1{x}_{n-1}\right),$$(5) : (5) $$\left(x_2-r_1{x}_1,x_3-r_1{x}_2,\dots ,{\text{x}}_{\text{n}}-r_1{x}_{n-1}\right),$$(5) where x₁, x₂, x₃ … x_n are data points of the series. Once more, by utilizing Microsoft Excel, the autocorrelation, homogenization, and outlier discovery were accomplished. Finally, the Mann–Kendall trend analysis (Mann, 1945; Kendall, 1975) was performed to evaluate the trend in rainfall data, while the estimation of the magnitude of trends in the data time series was performed following the method described by (Sen, 2013).

2.3.2. Econometric analysis

Pearson’s correlations were computed to determine the association among explanatory variables. Then, a multivariate probit (MVP) regression model, which is a commonly used statistical approach in this kind of studies (Bahadur & Akhter, 2018; Mbwambo et al., 2022; Sileshi et al., 2019), was utilized to investigate factors affecting the adoption of farmers’ adaptation strategies to minimize the effects of climate change on their sorghum yields. Furthermore, the multivariate probit model is very important in examining factors guiding farmers’ choices when there are more than two sets of response variables that are measured at the nominal level while the explanatory variables are dummy and continuous (Mbwambo et al., 2022). The analysis using the multivariate probit model is based on the expected maximum likelihood of decision making, which explains that if a farmer has a chance to choose among several adaptation strategies, he or she will select the strategy that contributes more in terms of climate adaptability. In that aspect, the farmer adopts a specific adaptation strategy (U*_ij) if the expected utility is greater than the expected benefit from any other existing strategies in the region (i.e. not adopting any strategy) (U_ij), i.e. y*_1i = U*_Ij - U_iJ >0; where y*_ji is the net utility (latent variable) that the farmer can get from endorsing the j_th strategy (Sileshi et al., 2019). Generally, MVP has numerous response variables (y_ji) and numerous hidden variables (y*_ji) but in this specific study, MVP involves five binary choice equations (early maturing varieties, adjusting sowing dates, drought tolerant varieties, soil conservation practices, and crop diversification). Thus, the model presumes that each binary observed takes a value of 1 if, and only if, the continuous hidden variable >0 as given in Equations (6)(6) $$y_{ji}^{\mathrm{*}}=X_{\mathit{\text{ji}}}{\beta }_{\mathit{\text{ji}}}+v_{\mathit{\text{ji}}}$$(6) and (7)(7) $$y_{\mathit{\text{ji}}}=\left\{{}_{0\hspace{0.17em}\mathit{\text{\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}otherwise}}}^{1\hspace{0.17em}\hspace{0.17em}\text{if\hspace{0.33em}\hspace{1em}}{y}_{ji}^{\text{*}}>0\left(\text{j}\hspace{0.17em}=\hspace{0.17em}\text{E},A,D,S,C\right)}\right.$$(7) : (6) $$y_{ji}^{\mathrm{*}}=X_{\mathit{\text{ji}}}{\beta }_{\mathit{\text{ji}}}+v_{\mathit{\text{ji}}}$$(6) (7) $$y_{\mathit{\text{ji}}}=\left\{{}_{0\hspace{0.17em}\mathit{\text{\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}otherwise}}}^{1\hspace{0.17em}\hspace{0.17em}\text{if\hspace{0.33em}\hspace{1em}}{y}_{ji}^{\text{*}}>0\left(\text{j}\hspace{0.17em}=\hspace{0.17em}\text{E},A,D,S,C\right)}\right.$$(7)

Where y_ji is the response variable, $y_{ji}^{\text{*}}$ is the hidden variable that apprehends the observed choice related to five adaptation strategies and is determined by observed characteristics (X) and an observed characteristics apprehended by stochastic error term (Y_ij); $\beta$ ij is a vector of parameters to be estimated. The error terms V_im m = 1, 2, 3, 4, 5 are distributed multivariate normal with a mean of 0 and a variance covariance matrix as presented below with correlation ρ_kj = ρ_kj as off diagonal components and values of 1 on the diagonal as explianed in Equation (8)(8) $$\left[\begin{array}{l}{v}_{1i}\\ v_{2i}\\ v_{3i}\\ v_{4i}\\ {\nu }_{5i}\end{array}\right]\sim n\left(\begin{array}{l}0\\ 0\\ 0\\ 0\\ 0\end{array}\right)\left(\begin{array}{l}1\hspace{0.17em}\rho 12\hspace{0.17em}\rho 13\hspace{0.17em}\rho 14\hspace{0.17em}\rho 15\\ \rho 21\hspace{0.17em}\hspace{0.17em}1\hspace{0.17em}\hspace{0.17em}\rho 23\hspace{0.17em}\hspace{0.17em}\rho 24\hspace{0.17em}\hspace{0.17em}\rho 25\\ \rho 31\hspace{0.17em}\hspace{0.17em}\rho 32\hspace{0.17em}\hspace{0.17em}1\hspace{0.17em}\hspace{0.17em}\rho P34\hspace{0.17em}\hspace{0.17em}\rho 35\\ \rho \text{\hspace{0.05em}}41\hspace{0.17em}\hspace{0.17em}\rho 42\hspace{0.17em}\hspace{0.17em}\rho 43\hspace{0.17em}\hspace{0.17em}1\hspace{0.17em}\hspace{0.17em}\rho 45\\ \rho 51\hspace{0.17em}\hspace{0.17em}\rho 52\hspace{0.17em}\hspace{0.17em}\rho 53\hspace{0.17em}\hspace{0.17em}\rho 54\hspace{0.17em}\hspace{0.17em}1\end{array}\right)$$(8) . (8) $$\left[\begin{array}{l}{v}_{1i}\\ v_{2i}\\ v_{3i}\\ v_{4i}\\ {\nu }_{5i}\end{array}\right]\sim n\left(\begin{array}{l}0\\ 0\\ 0\\ 0\\ 0\end{array}\right)\left(\begin{array}{l}1\hspace{0.17em}\rho 12\hspace{0.17em}\rho 13\hspace{0.17em}\rho 14\hspace{0.17em}\rho 15\\ \rho 21\hspace{0.17em}\hspace{0.17em}1\hspace{0.17em}\hspace{0.17em}\rho 23\hspace{0.17em}\hspace{0.17em}\rho 24\hspace{0.17em}\hspace{0.17em}\rho 25\\ \rho 31\hspace{0.17em}\hspace{0.17em}\rho 32\hspace{0.17em}\hspace{0.17em}1\hspace{0.17em}\hspace{0.17em}\rho P34\hspace{0.17em}\hspace{0.17em}\rho 35\\ \rho \text{\hspace{0.05em}}41\hspace{0.17em}\hspace{0.17em}\rho 42\hspace{0.17em}\hspace{0.17em}\rho 43\hspace{0.17em}\hspace{0.17em}1\hspace{0.17em}\hspace{0.17em}\rho 45\\ \rho 51\hspace{0.17em}\hspace{0.17em}\rho 52\hspace{0.17em}\hspace{0.17em}\rho 53\hspace{0.17em}\hspace{0.17em}\rho 54\hspace{0.17em}\hspace{0.17em}1\end{array}\right)$$(8)

The off diagonal components reveal various sets of climate change adaptation strategies as well as unobserved features that influence the adoption of adaptation strategies (Ahmed et al., 2017; Mbwambo et al., 2022; Sileshi et al., 2019; Teklewold, 2016). The collected data were entered into Excel spreadsheets and analyzed using Stata 13 software.

The description of the independent variables that were hypothesized to determine the adoption of various climate change adaptation strategies is presented in Table 3. These independent variables were utilised based on various findings reported by many researchers (Abera & Tesema, 2019; Belay et al., 2022; Mbwambo et al., 2022; Ngetich et al., 2022; Sileshi et al., 2019; Teshome et al., 2021; Traoré et al., 2021).

Table 3. Description of independent variables used in the model.

Variable	Description
Age	Age of the household head in years
Sex	Dummy of sex of household head (Yes = 1)
Education	Dummy of education of household head (Yes = 1)
Land size	Total area of the plot in Ha
Family size	Household size
Farming experience	Time spent in farming activities in years
Extension contact	Dummy for being visited by extension agent (Yes = 1)
Cooperative member	Dummy for belonging in to farm cooperative (Yes =1)
Information access	Dummy for having a radio, TV, and phone (Yes= 1)
Credit access	Dummy for accessing credit (Yes = 1)
Off farm income	Dummy for having off-farm activities (Yes =1)

3. Results and discussion

3.1. Results

3.1.1. Demographic and socio-economic characteristics of the respondents

The results of the study revealed that the overall mean age of the respondents was 52, ranging from 23 to 90 years, while a significant number of them (47%) had more than 30 years of farming experience (Table 4). The majority of these respondents were between 46 and 60 years old (39.1%), followed by farmers aged between 31 and 45 (30.7%), then those over 60 (26.4%), and finally younger farmers with ages ranging from 16 to 30 (3.8%).

Table 4. Demographic and socioeconomic characteristics of respondents (continues variables).

Variable	Mean	Std. Dev	Min	Max
Age	52.09	13.1240	23	90
Land size (ha)	0.81	0.8222	0.02	5.5
Family size (no.)	5.2	2.1434	1	14
Farming experience	31.59	13.3660	2	70

The average land size owned by the respondents was 0.81 ha, with a minimum of 0.02 ha and a maximum of 5.5 ha (Table 4). The results of this study revealed that the average family size was 5.2 persons, with a minimum of one person and a maximum of 14 persons per household (Table 4). Concerning sex, most respondents were men (58.0%) (Table 5). The findings of the education level revealed that the majority of the respondents (82.3%) were literate, while 17.7% were illiterate (Table 5).

Table 5. Demographic and socioeconomic characteristics of respondents (discrete variables).

Variable	Frequency (n = 345)	Percentage (%)
Sex	200	58.0
Education level	284	82.3
Extension contact	196	56.8
Cooperative member	220	63.8
Information access	236	68.4
Credit access	155	44.9
Off farm income	29	8.4

The results of the study revealed that extension agents were contacting a significant number of sorghum farmers (56.8%), and the majority of those farmers (63.8%) in the study area belonged to at least one farm related cooperative (Table 5). Also, the findings of this study revealed that 44.9% of farmers had access to credit (Table 5).

3.1.2. Farmers’ awareness on climate change and sources of information

The perception and awareness of a given community about climate change determine how its members respond to the negative effects of climate change (Hundera et al., 2019). From the total number of 345 respondents, 341 (98.8%) reported that they heard about climate change. Those farmers who claimed to have heard about climate change got this information from radio (86.4%), extension agents (53.9%), local organization meetings (39.4%), village neighbours (38.6%), places of work (26.1%), and television (23.5%). Only 12 farmers (3.5%) know about climate change through their own observation or indigenous knowledge.

3.1.3. Farmers’ perceptions on causes of climate change

The climate is an interactive, complex system defined as the average weather components over a period of time ranging from months to many millions of years, with a standard of 30 years. The climate system changes gradually over time due to human activities and natural factors (Le Treu et al., 2007). In the past, the climate has been changing, and these changes were mainly caused by natural factors. However, in recent years, climate change has been largely caused by human activities (The Royal Academy, 2012), which resulted in an increased emission of radiatively active gases called greenhouse gases (Pathak et al., 2012). These gases trap the outgoing infrared radiations from the earth’s surface and consequently increase the temperature of the atmosphere, in what is commonly known as global warming (Pathak et al., 2012).

Many interviewed farmers (44.6%) believed that the current climate change is caused by both human actions and natural processes. Out of the total of 154 (44.6%) respondents who perceived both human actions and natural process as the cause of climate change, 93 (60.4%) were men, while the remaining 61 (39.6%) were women. Of these 154 farmers who perceived both human actions and natural processes as drivers of climate change, the Huye and Kamonyi districts recorded the largest number of males and females, having 33 (35.5%) men and 22 (36.1%) women, respectively. On the other hand, 30.1% and 14.2% of farmers believed that climate change was solely the consequence of human actions and natural phenomena, respectively. Out of the total of 104 (30.1%) who perceived the real cause of current climate change, 63 (60.6%) were men and 41 (39.4%) were women. Compared to other districts, Nyanza district registered the largest number of males and females of the 104 farmers, with 29 (27.9%) men and 15 (14.4%) women. Contrarily, 39 (11.1%) of the respondents did not know the real cause of climate change. Concerning human actions, the majority of farmers (72.2%) perceived deforestation as the key anthropogenic activity causing climate change (Table 6). Out of the total of 249 (72.2%) farmers who perceived deforestation as the cause of climate change, 153 (61.4%) were men and 96 (38.6%) were women. Of these 249 farmers, the Nyanza and Kamonyi districts registered the largest number of males and females, with 50 (20.1%) men and 31 (12.4%) women, respectively (Table 6). For five out of seven causes of climate change (deforestation, industry, soil erosion, air pollution, and agricultural activities), the percentages of women who perceived human actions as the main cause of climate change were higher than those of men in the Ruhango district (Table 6).

Table 6. Farmers’ perceptions on human actions’ causes of climate change.

Human actions causes of climate change	% of all households	% of gender per district
		Huye		Gisagara		Nyanza		Ruhango		Kamonyi
		♂	♀	♂	♀	♂	♀	♂	♀	♂	♀
Deforestation	72.2	18.1	8.0	10.4	3.2	20.1	9.2	2.0	5.6	10.8	12.4
Industry	47.0	18.5	11.7	9.9	4.3	25.9	8.6	3.7	4.9	8.6	3.7
Soil erosion	35.7	7.3	4.1	11.4	3.3	25.2	12.2	2.4	5.7	11.4	17.1
Air pollution/transport	30.4	14.3	6.7	11.4	3.8	26.7	13.3	4.8	5.7	3.8	9.5
Agricultural activities	29.0	13.0	4.0	11.0	3.0	26.0	15.0	2.0	3.0	13.0	10.0
Population increase	25.8	5.6	4.5	15.7	4.5	34.8	15.7	2.2	2.2	5.6	9.0
Livestock activities	23.2	6.3	6.3	7.5	1.3	28.8	17.5	1.3	0.0	18.8	12.5

♂: male; ♀: Female. The sum of the percentage for all households does not add up to 100% because some farmers provided multiple responses. This applies to other tables presented with multiple responses.

In the case of natural causes of climate change, out of the total of 345 respondents, 127 (37.1%), 95 (27.5%), and 52 (15.1%) perceived earthquakes, volcanic eruptions, and variations in sun radiation as the main natural causes of climate change, respectively.

3.1.4. Farmers’ perceptions and trends of temperature and rainfall in the period of 30 years

The effects of climate change are being felt all around the world at various levels and magnitudes, depending on available resources. Its effects are experienced through changes in average humidity, rainfall, and temperature (Hundera et al., 2019).

The results of this study revealed that out of the total of 345 sorghum farmers, 332 (96.2%) of them perceived a change in temperature and rainfall trends, 239 (69.3%) highlighted the increased trend of temperature, and 195 (56.5%) perceived a decreasing trend of rainfall. Of the 239 farmers who perceived an increase in temperature, 139 (58.2%) were men and 100 (41.8%) were women. Also, out of the total of 239 farmers, the Nyanza, Huye, and Kamonyi districts recorded the majority of male farmers, having 43 (18.0%), 34 (14.2%), and 29 (12.1%) men, respectively. Concerning females, the Kamonyi, Nyanza, and Huye districts registered the majority of them, with 30 (12.6%), 24 (10.0%), and 21 (8.8%) women, respectively (Table 7).

Table 7. Farmers’ perceptions on the trends of temperature and rainfall over the period of 30 years.

Trends of temperature and rainfall	% of all households	% of gender per district
		Huye		Gisagara		Nyanza		Ruhango		Kamonyi
		♂	♀	♂	♀	♂	♀	♂	♀	♂	♀
Increase in temperature	69.3	14.2	8.8	9.2	3.3	18.0	10.0	4.6	7.1	12.1	12.6
Decrease in rainfall	56.5	22.1	8.7	7.7	3.6	14.9	10.8	5.1	7.2	10.3	9.7
Changes in times of raining	31.6	11.9	10.1	7.3	5.5	17.4	10.1	5.5	5.5	11.0	15.6
Increase in rainfall	12.2	16.7	7.1	14.3	14.3	7.1	2.4	4.8	11.9	7.1	14.3
Increase in frequency of drought	9.5	12.1	12.1	9.1	3.0	15.2	9.1	6.1	6.1	15.2	12.1
Decrease in temperature	2.6	22.2	0.0	0.0	11.1	11.1	0.0	33.3	11.1	11.1	0.0

The study also revealed that, in contrast to other districts, a significant number of men (33.3%) in the Ruhango district perceived a decrease in temperature.

The analysis of 30 years of temperature data showed that farmers’ perceptions are well aligned with the observation of historical temperature data, revealing a positive trend in minimum temperature. In contrast with farmers’ perceptions, a negative trend was observed in maximum temperature except for the Huye district (Figure 2).

Figure 2. Mean annual maximum temperature (Tmax) and minimum temperature (Tmin) trends of Ruhango, Gisagara, Kamonyi, Huye, and Nyanza districts in the central plateau zone of Rwanda (1989–2018).

Concerning rainfall data, farmers’ beliefs also coincide with historical rainfall records of the study area except in the Gisagara and Kamonyi districts for short rain and long dry seasons, respectively, which showed increasing trends (Table 8).

Table 8. Sen’s slope values for seasonal and annual rainfall in central plateau zone of Rwanda (1989–2018).

	Seasons
Stations	Short dry season	Long rain season	Long dry season	Short rain season	Annual
Kinazi	−2.29	−4.00	−0.42	−1.20	−7.00
Gikonko	−1.59	−1.60	−0.22	2.00	−1.60
Mugina	−4.44	−4.00	0.73	−4.15	−7.00
Rubona	−1.83	−1.91	−0.26	−0.33	−4.16
Busoro	−1.29	−3.30	0.00	−1.35	−4.25

3.1.5. Farmers’ perceptions on the effects of climate change induced shocks on sorghum

Climate shocks are natural disasters that surpass the capacity of a given community to cope with them (Anderson, 2001). These shocks are extreme weather events, including droughts, floods, hurricanes, and landslides, among others (Fernando et al., 2021). The results of this study (Table 9) revealed that out of the total of 345 farmers, 300 (87%) and 101 (29.3%) of them experienced drought and too much rain in a short period as the major climate change shocks on the sorghum crop. Of the 300 farmers who perceived drought as a major climate shock, 177 (59.0%) were men, while 123 (41%) were women. Out of this total number of 300 farmers, the Nyanza and Huye districts registered the largest number of males, with 55 (18.3%) and 53 (17.7%) men, respectively. For females, the same Nyanza district registered the majority of them, having 34 (11.3%) women. Compared to other districts, only women from the Ruhango district outnumbered men to perceive the drought. Concerning the amount of rain in a short time, 56 (55.4%) out of 101 farmers were men, while 45 (44.5%) were women. Of these 101 farmers, the Huye and Kamonyi districts showed the largest number of males and females, having 20 (19.8%) men and 12 (11.9%) women, respectively.

Table 9. Farmers’ perceptions on climate change shocks, shocks’ effects, and coping strategies.

Farmers’ perceptions	% of all households	% of gender per district
		Huye		Gisagara		Nyanza		Ruhango		Kamonyi
		♂	♀	♂	♀	♂	♀	♂	♀	♂	♀
Farmers’ perceptions on climate change induced shocks
Drought	87.0	17.7	10.0	7.7	3.3	18.3	11.3	4.7	6.0	10.7	10.3
Too much rain	29.3	19.8	8.9	9.9	6.9	9.9	6.9	10.9	9.9	5.0	11.9
Floods	12.8	13.6	4.5	22.7	9.1	9.1	2.3	11.4	6.8	6.8	13.6
Erratic rainfall	27.8	25.0	13.5	6.3	2.1	26.0	12.5	2.1	7.3	3.1	2.1
Hailstorm	21.4	44.6	18.9	13.5	4.1	2.7	4.1	4.1	4.1	0.0	4.1
Landslide	6.7	17.4	8.7	4.3	0.0	0.0	4.3	26.1	26.1	4.3	8.7
Effects of climate change shocks on sorghum
Decline in crop yield	72.2	20.5	12.4	7.6	3.6	19.3	9.2	2.8	4.4	8.4	11.6
Food shortage	37.4	21.7	11.6	7.0	2.3	27.9	17.8	2.3	3.9	2.3	3.1
Food price increase	32.5	23.2	9.8	14.3	6.3	25.0	16.1	0.9	2.7	1.8	0.0
Loss of entire crop	12.8	4.5	0.0	9.1	2.3	15.9	25.0	9.1	4.5	25.0	4.5
Loss of income	12.2	23.8	19.0	7.1	4.8	7.1	11.9	2.4	9.5	9.5	4.8
Sorghum farmers’ coping strategies for climate change induced shocks
Sell livestock	29.0	20.0	8.0	13.0	5.0	22.0	15.0	0.0	1.0	11.0	5.0
Eat less	24.9	16.3	11.6	4.7	2.3	32.6	16.3	3.5	8.1	3.5	1.2
Borrow money	22.9	26.6	8.9	7.6	2.5	19.0	13.9	1.3	1.3	8.9	10.1
Alternative food	19.1	10.6	12.1	3.0	1.5	34.8	18.2	6.1	9.1	1.5	3.0
No coping measure	17.1	11.9	11.9	5.1	5.1	10.2	3.4	13.6	13.6	5.1	20.3
Off-farm employment	11.3	20.5	20.5	0.0	0.0	15.4	12.8	0.0	7.7	10.3	12.8
Different foods	9.6	9.1	6.1	9.1	3.0	42.4	15.2	3.0	0.0	6.1	6.1

Concerning the effects of climate change shocks on sorghum, out of the total of 345 sorghum farmers, 249 (72.2%) experienced a decline in the yield, while 44 (12.8%) perceived a total crop failure, which resulted in food shortages and food price increases (Table 9). Of the 249 farmers who perceived a decline in sorghum yield, the Huye district registered the largest number of males, with 51 (20.5%) men. On the other hand, out of the total of 44 farmers who experienced a total crop failure, the Nyanza and Kamonyi districts equally registered the largest number of females and males, having 11 (25%) women and 11 (25%) men, respectively (Table 9).

During this study, farmers reported more than 15 sorghum varieties that were mostly affected by climate change shocks. Among them, varieties Nyiragikori, Rudasakwa, Cyamwiha, Gihove, and Amabanda were reported by 17.0%, 14.9%, 8.2%, 7.3%, and 6.1% of farmers, respectively, to have been more affected by these climate change shocks over the period of 30 years.

3.1.6. Sorghum farmers’ climate change coping strategies

It is known that total crop failure and other negative effects on crop yield have bad impacts on farmers’ livelihoods. Therefore, farmers practice different coping strategies to deal with the effects of climate change shocks on sorghum. Out of the total of 345 sorghum farmers, 100 (29.0%) of them preferred selling livestock as their main coping strategy (Table 9), while in contrast, 59 (17.1%) of them did not take any coping measures.

Of the 100 sorghum farmers who opted to sell livestock as a coping strategy, the Nyanza district registered the largest number of males and females, with 22 (22.0%) men and 15 (15.0%) women. The Nyanza district was followed by the Huye district, which showed 20 (20.0%) men and 8 (8.0%) women (Table 9). The results of this study also showed that there was no male respondent in the Ruhango district who opted to sell livestock to cope with climate change shocks. Furthermore, the findings of this study revealed that the Ruhango district recorded the majority of farmers who did not cope with climate change shocks, equally showing 8 (13.6%) men and 8 (13.6%) women (Table 9).

3.1.7. Perceptions of farmers on climate change effects on sorghum crop

Climate change is expressed in various ways because the augmentation of extreme weather conditions such as droughts, cyclones, heat waves, and floods has negative effects on agriculture production (Pathak et al., 2012). In the current study, farmers observed various negative effects of climate change on their sorghum crop. Out of the total 345 farmers, the majority of them (95.7%) acknowledged the reduction in sorghum grain yield as the major effect of climate change on sorghum and the disappearance of sorghum varieties in the study area (84.3%) as the second main effect (Table 10). Among those 330 (95.7%) farmers who perceived the reduction in grain yield, the Nyanza and Kamonyi districts recorded the largest number of males and females, with 59 (17.89%) men and 38 (11.5%) women, respectively. Of the 291 (84.3%) who experienced variety disappearance, the Nyanza district registered the majority of both males and females, with 58 (19.9%) men and 36 (12.4%) women. Also, the results showed Ruhango as the only district where females surpassed males to perceive variety disappearance, with 15 (5.2%) women against 10 (3.4%) men (Table 10). The results of the study also disclosed the names of more than 20 sorghum varieties that were no longer seen in the study area, such as Amabanda, Cyamwiha, Nyiragikori, Rudasakwa, and Kebo, which were reported by 13.5%, 12.9%, 11.5%, 7.4%, and 7.2% of respondents, respectively.

Table 10. Farmers’ perceptions on the effects of climate change on sorghum.

Climate change effects on sorghum	% of all households	% of gender per district
		Huye		Gisagara		Nyanza		Ruhango		Kamonyi
		♂	♀	♂	♀	♂	♀	♂	♀	♂	♀
Grain yied reduction	95.7	15.8	9.1	9.4	3.6	17.9	10.9	4.5	6.1	11.2	11.5
Varieties disappearance	84.3	14.8	8.2	9.6	4.1	19.9	12.4	3.4	5.2	11.7	10.7
Presence of fall armyworms	78.8	15.4	9.9	8.5	3.3	18.8	12.5	2.6	4.4	12.1	12.5
Presence of honeydew/grain	77.1	11.7	9.8	10.2	4.1	20.7	13.2	4.9	6.4	10.9	8.3
Presence of stem borer	63.2	10.6	5.0	9.2	4.1	20.2	13.3	6.0	7.3	10.6	13.8
Lodging of sorghum plants	55.1	11.1	6.3	7.4	3.2	23.2	14.7	4.7	5.8	11.1	12.6
Sorghum plants senescence	52.5	5.0	2.8	10.5	5.0	26.0	18.2	3.9	3.9	16.0	8.8
Presence of aphids	50.7	7.4	5.1	10.3	4.6	29.7	20.0	2.9	2.3	12.0	5.7
Presence of rust	40.3	12.9	3.6	11.5	4.3	34.5	20.9	2.9	5.0	0.7	3.6
Biomass decrease	39.4	5.1	4.4	9.6	2.2	27.9	22.8	1.5	2.2	14.7	9.6

3.1.8. Sorghum farmers’ adaptations strategies and factors determining their adoption

3.1.8.1. Sorghum farmers’ adaptation strategies

Adoption of different adaptation measures reduces the impact of climate change and farmers’ vulnerability (Atchikpa et al., 2017). The potential measures of climate change adaptation proposed to deal with its effects include changing crop management practices, crop diversification, weather forecasting, developing and disseminating new varieties tolerant and resistant to different climate change consequences, improving pest and disease management, initiating crop insurance, and supporting local technical knowledge (Pathak et al., 2012).

As sorghum has economic and cultural importance in the central plateau zone of Rwanda, farmers also used various adaptation measures to deal with the long term effects of climate change. Out of the total of 345 farmers, 289 (83.8%) adopted different adaptation strategies, while 56 (16.2%) did not. For specific adaptation strategies, 231 (67%) sorghum farmers adopted early maturing sorghum varieties, 173 (50.1%) adjusted planting dates, and 113 (32.8%) used crop diversification (Table 11).

Table 11. Sorghum farmers’ adaptation strategies.

Adaptation strategies	% of all households	% of gender per district
		Huye		Gisagara		Nyanza		Ruhango		Kamonyi
		♂	♀	♂	♀	♂	♀	♂	♀	♂	♀
Early maturing variety	67.0	17.3	10.8	12.1	4.3	22.1	12.6	2.6	3.5	8.7	6.1
Adjusting planting date	50.1	23.7	9.2	11.6	4.6	24.9	13.9	2.3	0.6	4.6	4.6
Drought tolerant variety	46.7	11.8	5.0	12.4	5.0	21.1	14.3	3.7	5.0	9.9	11.8
Soil conservation practices	38.3	11.4	9.1	10.6	5.3	28.0	12.9	0.8	2.3	12.1	7.6
Crop diversification practices	32.8	16.8	11.5	10.6	2.7	21.2	17.7	0.9	0.9	10.6	7.1

Of the 231 farmers who adopted early maturing varieties, the Nyanza, Huye, and Ruhango districts registered the largest number of males, with 51 (22.1%), 40 (17.3%), and 28 (12.1%) men, respectively. As for females who opted to use early maturing varieties, the Nyanza and Ruhango districts registered the majority and the minority of them, with 29 (12.6%) men and 8 (3.5%) women, respectively. Out of the total of 173 respondents who chose to adjust their planting dates, the Nyanza recorded the largest number for both sex, having 43 (24.9%) men and 24 (13.9%) women (Table 11). Similar to adjusting the planting date, of the 113 sorghum farmers who adopted crop diversification, the Nyanza district registered the greatest number for both sex, with 24 (21.2%) men and 20 (17.7%) women. On the other hand, the smallest number of farmers who diversified their crops was recorded in the Ruhango district for both sex, equally having one man and one woman.

3.1.8.2. Factors affecting the adoption of climate change adaptation strategies

The statistical results of the variables involved in the multivariate probit model that were believed to affect the adoption of climate change adaptation strategies among sorghum farmers were presented in Tables 4 and 5.

The results of the multivariate probit model are presented in Table 12. As described above, 83.8% of respondents in the study area utilized at least one adaptation strategy to deal with the effects of climate change and variability on sorghum. In the multivariate probit model, it is necessary to find out the validity of the model and the interdependence and dependence variables before explaining the results. The model fitted the data quite well looking at the values of Wald Chi-squared = 531.19 and p < .001. Hence, the hypothesis stating that all coefficients in each equation are equal to zero was rejected. Contrarily, the Chi-square test (chi2 (10) = 81.0655 or p < .000) confirmed that the adoption choices among the five adaptation strategies were not mutually independent. This implies that the coefficient estimates obtained from joint estimation are asymptotically more efficient than the coefficient found from a single equation when the binary outcome variables are correlated (Sileshi et al., 2019; Young et al., 2006). In our case, the joint estimation was supported by the significance of at least a 10% probability level of all possible pairs of correlations between the error terms. All the adaptation strategies were positively correlated (with soil conservation practices and crop diversification, showing the highest correlation coefficient of 0.4988 in Table 12), which indicated that there was a positive or complementary relationship between each adaptation strategy.

Table 12. Results of multivariate probit model for adoption of climate change adaptation strategies.

	Early maturing varieties		Adjusting sowing date		Drought tolerant varieties		Soil conservation practices		Crop diversification
Variables	Coef	z	Coef	Z	Coef	z	Coef	z	Coef	z
Age	0.0057	0.37	0.0551	2.43**	0.0354	2.04**	−0.0154	−0.79	0.0021	0.12
Sex	0.1245	0.79	0.2574	1.17	−0.2491	−1.4	−0.0061	−0.04	0.0364	0.23
Education level	1.6255	6.83***	−0.4924	−1.59	−0.0887	−0.41	0.0987	0.43	−0.3217	−1.53
Land size	−0.2023	−1.77*	−0.1439	−1.1	−0.0558	−0.54	0.1503	1.62	−0.0989	−1.07
Family size	−0.034	−0.92	−0.0677	−1.23	0.0378	0.96	−0.0405	−1.05	0.0191	0.55
Farming experience	−0.0012	−0.08	−0.032	−1.51	−0.0205	−1.19	0.0364	1.9*	0.0133	0.79
Extension agents contact	0.8986	5.48***	3.4516	12.38***	0.4779	2.64***	0.6334	3.6	0.4066	2.56**
Cooperative membership	0.0553	0.31	0.1267	0.56	2.3432	9.47	0.1669	0.93	−0.2533	−1.53
Information access	−0.1741	−0.95	0.085	0.35	−0.1744	−0.86	1.7035	6.44***	0.2154	1.2
Credit access	0.4987	2.67***	0.5829	2.68***	0.3878	2.17**	0.5714	3.18***	0.5443	3.35***
Off farm income	0.0457	0.14	0.1986	0.53	0.2749	0.77	0.3207	1.14	0.2657	0.99
rho21	0.2942	2.38**		Multivariate probit (SML, # draws = 5)
rho31	0.3756	3.86***		Log pseudo likelihood = −708.97984
rho41	0.2954	2.99***		Number of obs. = 345
rho51	0.3758	3.97***		Wald chi2(60) = 531.19
rho32	0.2094	1.77*		Prob > chi^2 = 0.0000
rho42	0.3205	2.45**
rho52	0.2673	2.73***
rho43	0.2577	2.77***
rho53	0.3565	4.19***
rho54	0.4988	6.04***

Likelihood ratio test of rho21 = rho31 = rho41 = rho51 = rho32 = rho42 = rho52 = rho43 =rho53 = rho54 = 0: chi² (10) = 81.0655.

Prob > chi² = 0.0000. rho shows correlation between dependent variables.

^***,**,* significant at the 1, 5 and 10 percent probability levels, respectively.

Additionally, the multivariate probit model results disclosed that some of the selected independent variables significantly affected the sorghum farmers’ choice to adopt adaptation strategies (Table 12). The results from the multivariate probit model revealed that the age of the farmers positively influences the adoption of adjusting planting dates and the use of drought tolerant varieties.

Farmers’ adoption of early maturing varieties increased with their education status, and the interrelation was significant at the 1% level.

Farming experience was also shown to positively affect the use of soil conservation practices with a 10% level of significance.

The visit of extension agents was revealed as a key factor among all explanatory variables, as it affects positively the adoption of all adaptation strategies except soil conservation practices (p < .001 and 0.001 <p < .05).

The result of the multivariate probit model revealed that access to information through radio, TV, and phone has significantly (p < .01) influenced the adoption of soil conservation practices.

The results of the multivariate probit model also indicated that the respondents’ access to credit affected the adoption of climate change adaptation strategies in the study area. Access to credit was positively related to all dependent variables (p < .01) except drought tolerant varieties (.01<p < .05), suggesting that sorghum farmers who have access to credit are more likely to adopt early maturing varieties, drought tolerant varieties, practice intercropping, and use soil conservation practices.

Surprisingly, the results revealed that the likelihood of farmers adopting early maturing varieties decreased with the increase in land size, as they were negatively associated at the 10% level of significance.

3.2. Discussions

3.2.1. Demographic and socio-economic characteristics of the respondents

Farming experience was mainly associated with the age of the respondents, as many of them (74.2%) reported to having started farming at the age of 16. This is in line with what (Teshome et al., 2021) reported, where they found that farming experience was generally associated with the age of the maize farmers in Eastern Ethiopia. According to (Agbo, 2013), this time of experience is enough to have reliable information on climate change and related impacts.

The average land size in the study area was found to be higher than the average size at the national level of 0.4 ha per household (NISR, 2020). This could be explained by the fact that there are many farmers in Rwanda who rent agricultural plots from neighbours with the aim of enhancing their own small land (NISR, 2020).

In the study area, the average family size was found to be greater than the national average which is 4.0 persons per family (NISR, 2023). This may be attributed to the fact that family size is higher in rural areas compared to urban areas, implying a greater need for food in the study area. Consistent with this suggestion (Agidew & Singh, 2018), reported that the increase in family size goes together with the increase in food dependence rate, which may lead to food insecurity. Teshome et al. (2021) also stated that the increase in family size leads to an increase in the number of dependents in a household, which may have a negative effect on food security.

In Rwandan culture, men are heads of the families, and they always have to speak or respond before women in every event. This may be the reason for the high percentage of men among respondents in this study. In line with this finding (Majoro et al., 2020), also found a majority number of men (54.67% among respondents) in their study in Rwanda. On the other hand, the significant number of women (42%) among the respondents may be related to the fact that there is an important number of women who are heads of households as of 2022; 28.9% of all Rwandan households were headed by women. Additionally, women-headed households are more prevalent in rural areas (30%) than in urban areas (26%) (NISR, 2023) probably due to men’s urban emigration for job seeking.

In the Rwandan context, literacy is one of the essential knowledge and skills people learn in schools, including reading, writing, and numeracy (MINEDUC, 2014). The results of this study are almost the same as the average known at the national level for educated people (77.7%) (NISR, 2023).

The fact that a significant number of farmers had access to extension services may be attributed to the current Twigire Muhinzi extension model, which brings together all frontline extension agents in order to work together for the benefit of the farmers. These frontline extension agents involve farmer promoters, farmer field school facilitators, cellar social and economic development officers, agro-dealers, and sector and district agronomists (MINAGRI, 2022).

The high percentage of farmers in cooperatives could be linked with the Twigire Muhinzi extension model, which sensitises farmers to form cooperatives and associations for easy access to extension services. This is confirmed by a recent annual report on the status of cooperatives in Rwanda, which indicates that agricultural cooperatives accounted for 45.8% of all cooperatives existing in the country (RCA, 2022).

Farmers (44.9%) who have access to credit was found to be superior to the national average percentage (11%) of farmers accessing credit (IPAR & AFR, 2018), and this may be justified by the recent expansion of informal tontines (Ibimina), which help many Rwandans to have access to credit without relying on formal financial institutions.

3.2.2. Farmers’ awareness on climate change and sources of information

The percentage of farmers who got climate change information from the radio was almost the same as the average mean known at the national level of 81.3% (NISR, 2023). In contrast, the percentage of farmers who got information from television is higher than the national average of 12.3% for the population who own television (NISR, 2023). This is explained by the fact that all farmers may not own their own television, but after farm activities, many men spend their afternoons outside their homes where they may watch television even if they do not possess it.

The findings of this study are in agreement with (Hundera et al., 2019; Raghuvanshi, 2017), who reported that 100% and 90.3% of farmers, respectively, were aware of climate change and variability in India and Ethiopia. The results of this study also concur with the findings of (Eneji et al., 2020; Mehmood et al., 2022), who found that 67% and 75% of surveyed farmers were aware of climate change, respectively. Regarding sources of information, the findings of this study confirm the results obtained by (Popoola et al., 2020) who highlighted that television and radio are the main sources of farmers’ climate change information. Similar results were also reported in studies conducted by (Adebisi-Adelani & Oyesola, 2014; Okoro et al., 2016; Tessema et al., 2013) emphasizing that the main sources of information were extension agents, village members, and personal observation.

3.2.3. Farmers’ perceptions on causes of climate change

The observed results on human causes of climate change are in accordance with (Teshome et al., 2021; Yamba et al., 2019) who reported that deforestation was perceived to be the main driver of climate change. Similar findings were reported by (Tesfahunegn et al., 2016), who revealed that deforestation (56%), soil poverty due to erosion (50%), air pollution (44%), agricultural activities (42%), livestock activities (40%), and economic development like urbanization (35%) were acknowledged by farmers as the main causes of climate change. In Zambia, farmers also believed that climate change was caused by deforestation, industries, and modernization like the construction of roads and big buildings (Nyanga et al., 2011), and this confirms the results of the present study.

Based on the results of this study, the farmers’ perceptions of current climate change causes were still not well in line with scientific concepts because only 30.1% of the interviewed farmers perceived its real cause, which are anthropogenic activities (The Royal Academy, 2012). Those 30.1% of farmers also missed fossil fuel combustion, the principal cause of climate change emitting 86% of human CO₂ emissions as reported by (GNASL, 2021).

3.2.4. Farmers’ perceptions on trends of temperature and rainfall in the period of 30 years

Farmers’ perceptions on trends of temperature and rainfall in the period of 30 years confirmed the findings of (Msaki et al., 2015; Warner et al., 2015), who reported a substantial temperature increase of 0.47 °C per decade and frequent rainfall shortages in Rwanda. Also, the results of this study confirmed the findings obtained by (Babatolu & Akinnubi, 2016; Dhanya & Ramachandran, 2016), who reported that 89.9% of farmers had perceived an increasing temperature and decreasing rainfall trends. Additionally, similar results were reported by (Dendir & Simane, 2021; Esayas et al., 2019), who revealed that more than 60% and 69.2% of farmers, respectively, had perceived an increasing temperature and decreasing rainfall in southern Ethiopia.

3.2.5. Farmers’ perceptions on the effects of climate change induced shocks on sorghum

The results of this study are aligned with the findings of (Dossou-Aminon et al., 2015), who revealed drought and floods as the main climate change shock on the sorghum crop. Furthermore, drought was also reported as the main shock for sorghum after panicle initiation and during early grain filling (Eggen et al., 2019). According to (Fernando et al., 2021), five climate shock events (drought, floods, extreme temperature, storms, wild fire) account for 73% of climate change shocks, indicating how farmers perceptions were aligned with the scientific reality of climate change shocks.

Similar effects of climate change shocks were reported by (Babatolu & Akinnubi, 2016), when farmers observed a decline in crop harvests resulting in food price increases due to climate change effects. (Eggen et al., 2019; Nsengiyunva, 2014) also reported that sorghum farmers experienced total crop failure due to climate change related drought in Ethiopia and Uganda, respectively. Larger variations in precipitation and temperature caused by climate change increase the frequency and intensity of floods and droughts. This affects crops’ available water, which leads in turn to a reduction in crop yields (Mor, 2017).

3.2.6. Sorghum farmers’ climate change coping strategies

Similar to the findings of this study, (Bawa et al., 2015) found that community members’ resilience to climate change was low, where in the event of floods they normally depend on family and friends for support, and others are helpless and only looking up to God. Adimassu & Kessler (2016) found that farmers sell livestock, while (Abou et al., 2021) reported diversification of income as a coping strategy to deal with climate change shocks on the sorghum crop. Mbwambo et al. (2022) reported selling livestock, seeking off farm employment, and reducing the number of meals per day as coping strategies. Abid et al. (2020) distinguish between two categories of coping strategies: ex-ante and ex-post strategies. The ex-ante strategy consists of adopting strategies ahead of time to handle climate shocks, while the ex-post strategy chooses strategies in response to climatic shocks. In accordance with this categorization, farmers in this study area chose to address climate shock through ex-post methods, demonstrating the necessity of raising their awareness to consider ex-ante strategies in the future.

3.2.7. Perceptions of farmers on climate change effects on sorghum crop

The decrease in grain yield may be caused by drought, which leads to decreased carbon assimilation, stomatal conductance, and cell turgor, hence slowing down crop growth and development, as reported by (Prasad et al., 2021). The results agreed with the findings of (Fosu-Mensah et al., 2012), who reported that the long drought during the reproductive stage affects the size and weight of the grain and consequently the yield of the concerned crops. He also highlighted that more than 80% of farmers felt that the current increase in pest and disease outbreaks and weed infestations were the consequences of climate change. The results of this study also confirmed the findings of (Dossou-Aminon et al., 2015), who reported the increase of insect damage and low productivity as the main impacts of climate change on sorghum. Moreover, in drought prone agro ecologies of Ethiopia, 71% of interviewed farmers experienced a decrease in sorghum production due to climate change. The same farmers acknowledged the increasing number of sorghum stalk bores as another climate change effect that reduces sorghum production (Assefa et al., 2016), while (Traoré et al., 2021) reported poor grain filling due to climate change.

The disappearance of sorghum varieties in the study area may be attributed to the change in environmental conditions suitable for the development of these varieties, which forces them to be either adaptive or exterminated. In agreement with the perceptions of the farmers in this study, (FAO, 2015) reported an increase in crop variety extinction as one of the major effects of climate change. Accordingly (Thormann & Engels, 2015), disclosed how climate change enhance the genetic erosion of different threatened landraces, while (Azeez et al., 2018) highlighted how farmers will lose existing varieties to be replaced with improved varieties that are well adapted to climate change. The perceptions of famers are aligned with the scientific facts of the impact of climate change on sorghum. Climate change alters the incidence of pests and diseases in crops by raising host susceptibility, speeding pathogen transmission, and enhancing vector development. It also causes premature death of stalks and leaves, lodging, and, as a consequence, the seeds’ weight decreases. As well, climate change affects different components of sorghum grain yield, such as the number of grains per panicle and seed size (Edema & Amoding, 2015; Pathak et al., 2012; Yahaya et al., 2023).

Seeing that many sorghum varieties disappeared due to climate change, the Rwanda National Genebank is recommended to carry out the exploration of these lost varieties in their other growing areas in order to collect and conserve them for current and future utilization.

3.2.8. Sorghum farmers’ adaptations strategies and factors determining their adoption

3.2.8.1. Sorghum farmers’ adaptation strategies

Raghuvanshi (2017) reported that the level of climate change awareness influences farmers’ behaviour towards adaptation measures. Hence, the great number of farmers who endorsed the use of climate change adaptation practices could have information on climate change. This is supported by the fact that 98.8% of farmers stated that they were aware of climate change. Also, looking at the percentage of women engaged in agricultural activities at the national level (77%), there is no doubt that women have played a key role in the adoption of these adaptation strategies, as they are more involved in the agricultural sector than men (58.3%) (NISR, 2023).

Many studies carried out in different parts of the world found similar types of farmers’ adaptation strategies. In the North Eastern region of Benin, the majority of sorghum farmers (73.7%) reported the use of drought tolerant sorghum varieties, early sorghum sowing, and crop rotation as the best climate change adaptation strategies. (Atchikpa et al., 2017; Dossou-Aminon et al., 2015; Likinaw et al., 2022; Ngetich et al., 2022) also reported changes in the planting date, use of early maturing varieties, and on farm crop diversification as farmers’ climate change adaptation strategies. (Belay et al., 2022) also reported the use of soil conservation methods as strategies of farmers in the southern part of Ethiopia. Some of the adaptation strategies used by farmers in the central plateau zone are in accordance with what (IPCC, 2023) recommends, like crop diversification and improved varieties. In spite of that, they also lacked important ones like soil moisture conservation, on-farm water management, water storage, and agroforestry, indicating the need to help them improve their existing adaptation strategies.

Though the majority of the farmers in the study area use various adaptation measures, there were 16.2% of them who needed to be sensitized by extension agents on the role of using adaptation measures to tackle the effects of climate change.

3.2.8.2. Factors affecting the adoption of climate change adaptation strategies

The age of the farmers that positively influences the adoption of adjusting planting dates and the use of drought tolerant varieties implies that older sorghum farmers are more expected to adopt adjusting planting dates and drought tolerant varieties than younger farmers. This may be attributed to the fact that older farmers may have more information on the existence of those varieties than younger farmers. The positive influence of age on the adoption of technologies was also reported by (Belay et al., 2022; Mehmood et al., 2022; Zakari et al., 2022).

Literate respondents were more likely to use early maturing varieties as they are able to get information from different sources in comparison to illiterate farmers. Similar to these results (Teshome et al., 2021), reported that the illiteracy of the maize farmers negatively affected the adoption of mineral fertilizers as one of the climate change adaptation strategies in Eastern Ethiopia. The results of the current study demonstrate well the importance of knowledge in combating the current challenges of climate change.

The farming experience that positively affects the use of soil conservation practices indicates that the more the respondents gain experience in farming activities, the more they will be in a good position to practice different soil conservation activities to deal with climate change. This was in line with the findings of (Mbwambo et al., 2022; Ndambiri et al., 2013), who reported that farming experience had a positive and significant impact on the adoption of climate change adaptation strategies.

The visit of extension agents that positively affects the adoption of all adaptation strategies (except soil conservation practices) suggests that information and training, most likely provided by extension agents to sorghum farmers, assisted them in increasing their awareness on the benefits of these strategies, thereby determining their adoptions. Similar findings were reported earlier by (Asrat & Simane, 2018; Mihiretu et al., 2020; Sileshi et al., 2019), who found that the frequency of extension contacts positively influenced the adoption of adaptation strategies as farmers acquired new knowledge and skills that allowed them to implement these adaptation measures.

The access to information that positively influences the adoption of soil conservation practices indicates that the more farmers got information, the more they tended to adopt soil conservation practices in the study area. These results were in line with the findings of (Esayas et al., 2019; Mihiretu et al., 2020) in Ethiopia, who reported that farmers who had better access to information on climate change had a higher probability of adapting to it.

Generally, farmers who get credit are regarded as rich in rural areas, implying that those farmers are in a good position to afford the existing technologies. Access to credit is very important for smallholder farmers, enabling them to invest in acquiring adaptation technologies (Sileshi et al., 2019). This positive correlation between credit access and adaptation strategies was also reported by (Belay et al., 2022; Bakare et al., 2023; Zakari et al., 2022).

The land size that negatively affects the adoption of early maturing varieties could be related to the perceived wealth of large landowners, given that land scarcity is a major issue in Rwanda. Thus, those perceived as rich farmers may opt for a combination of high-yielding late maturing varieties and other adaptation strategies rather than early maturing varieties, as late maturing varieties out-yield early maturing varieties under normal conditions. According to (Bello et al., 2012), late maturing varieties produce more yield than early maturing varieties. The revealed negative association between land size and the adoption of early maturing varieties was also reported earlier by (Deressa et al., 2011) in the Nile basin of Ethiopia.

5. Conclusions

The information generated from farmers’ perceptions of climate change’s effects on crops helps policymakers understand reality and develop appropriate policies. However, in Rwanda’s central plateau zone, there is a lack of such information on sorghum due to few studies that focused on other crops and agroecological areas. Thus, this study aimed to generate this information by assessing farmers’ perceptions of climate change, its effects on sorghum, adaptation strategies, and factors influencing these strategies. The results of the study revealed that sorghum farmers are aware of climate change, but they do not understand the concept of climate change well. Hence, the mandated institutions have to raise the awareness of the farmers concerning the genuine causes of climate change, as that is the starting point of climate change mitigation and adaptation. The majority of farmers also perceived a decrease in rainfall and an increase in temperature over 30 years. Farmers also noted drought as the main shock of climate change, while the reduction in sorghum grain yield and the disappearance of sorghum varieties were mentioned as the main effects of climate change. In the study area, farmers opted to sell livestock and use of early maturing varieties as the best strategies to deal with climate change shocks. The results of the study also revealed that older age, a better education level, access to extension services, access to information, access to credit, and higher farming experience were positively associated with farmers’ climate change adaptation choices. Although farmers are trying to respond to the negative effects of climate change on sorghum, the gap in their climate change knowledge needs to be addressed by extension services. Due to the disappearance of sorghum varieties, it is recommended that the genebank start the exploration, collection, and conservation of these threatened sorghum varieties for current and future utilization. Also, the results imply that the government and partners who are engaged in both agriculture and the fight against climate change should work together to enable farmers to improve their adaptation strategies for sustainable sorghum production. This study focused only on Rwanda’s central plateau zone, yet other sorghum producing agroecologies and other crops are affected by climate change. Therefore, further studies should explore this topic in other sorghum growing areas and consider other crops.

Acknowledgments

We thank Haramaya University, Ethiopia, through the World Bank and the Africa Center of Excellence for Climate Smart Agriculture and Biodiversity Conservation (SABC) for funding this PhD research. Thanks also to the Rwanda Agriculture and Animal Resources Development Board for their valuable support during data collection.

Disclosure statement

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

Additional information

Funding

This work was supported by Africa Center of Excellence for Climate Smart Agriculture and Biodiversity Conservation, Haramaya University.

Notes on contributors

Theogene Niyibigira

Theogene Niyibigira is a researcher at the Rwanda Agriculture and Animal Resources Development Board (RAB) with an MSc in Genetics and Plant Breeding interested in climate-smart agriculture, sorghum crop, and biodiversity conservation research.

Wassu Mohammed

Professor Wassu Mohammed (PhD) is a notable breeder and member of the academic community at Haramaya University, teaching more than 15 courses at the MSc and PhD levels.

Tamado Tana

Professor Tamado Tana (PhD) is a well-known member of the University of Eswatini with interests in crop ecology, crop systems, agronomy, and crop modelling.

Tesfaye Lemma Tefera

Tesfaye Lemma Tefera (PhD) is an Associate Professor of Rural Development at Haramaya University interested in rural livelihoods, food security, climate change adaptation, and agriculture extension.

Placide Rukundo

Placide Rukundo (PhD) is a known breeder in Rwanda, a former Senior Principal Research Fellow in the Rwanda Agriculture and Animal Resources Development Board (RAB), and currently works at the International Potato Centre (CIP), with an interest in plant breeding, molecular biology, and biotechnology.