Introduction

With ongoing climate change, Climate Services (CS) are increasingly becoming relevant. This has led to increasing investments in CS in many countries especially in the global south (Lam González 2021). Antwi-Agyei et al. (2021) have also noted a surge in the development of national meteorological systems to enhance CS.

Climate Services involve the provision of climate information (i.e., regarding rainfall pattern, wind velocity, and temperature) based on credible scientific data and expertise, to assist in decision-making by smallholders and stakeholders. CS are not analogous to early warning systems which mainly deal with hazardous weather, but can provide seasonal weather forecasts. CS must respond to user needs, and require appropriate engagement between the users and providers. However, previous CS initiatives have been developed using top-down approaches (i.e., with little involvement of end users) which showed limited effectiveness in terms of uptake as they often fail to take into consideration, the needs, expectations, and readiness of end users, especially in developing countries such as Benin (Born et al. 2021; Lourenço et al. 2016; Singh et al. 2018).

While there is a growing consensus on the need to have CS for farmers, it is imperative to first understand information need and use by smallholder farmers, which is lacking in the country of Benin where the concept of climate service is just emerging (Meinke et al. 2006; Vincent et al. 2018). The primary challenge for farmers in Benin is not the long-term change in climate, but rather the day-to-day or seasonal change of weather conditions, known as climate variability.

This paper presents a cross-sectional study that analyzes smallholder farmers’ awareness of CS, measured as scores of Knowledge, Attitude, and Practice (KAP). The paper analyzed the socio-economic factors that explain the KAP scores and contributes to empirical literature in two ways. First, most of the existing work on knowledge and use for CS relied only on farmers’ willingness to pay (e.g., Amegnaglo et al. 2017; Yegbemey et al. 2014a, b). In this study, however, we used a more comprehensive perspective that involves what farmers know/understand (the knowledge), believe (the attitude), and actually do (the practice) with respect to weather information in general and CS in particular. Second, the study highlights the socio-economic drivers of the KAP scores and provides insights into possible intervention areas that can be considered by policy makers.

Increasing climate variability in Benin as a result of global climate change has created the need for smallholder farmers to make significant modifications to their production systems (Wilkinson et al. 2015). Previous studies reported a number of climate change manifestations in West Africa, including some delays/precocities of rains, abnormal variations in rainfall levels, and other extended weather events such as more pronounced droughts, floods, cyclones, frosts, and high heat (Dugué et al. 2021; Kabore et al. 2019; Vodounou and Onibon Doubogan 2016). The negative effects of climate change are also expected to be more serious in developing countries in general, and particularly, in rural areas where agriculture is mostly rainfed and stands as the main source of livelihood for farm households (Amjath-Babu et al. 2016; Banque Mondiale 2021; Byerlee et al. 2008; Pachauri et al. 2014).

Literature on climate change/variability adaptation among smallholder farmers has grown rapidly and suggests that farmers have developed response strategies (e.g., changes in planting dates, adoption of short cycle varieties) that are mostly endogenous (Kabore et al. 2019). Whereas these strategies are well documented, most of them seem to be restrictive or result in higher production costs to farmers (Yegbemey et al. 2014a, b). Indeed, studies show that climate variability is one of the major sources of inefficiency in current production systems (Mushagalusa et al. 2021; Yegbemey et al. 2021). However, it is yet unclear how well the currently reported climate variability responses support the adaptation and continued resilience of smallholder farmers.

Access to robust climate information is vital for anticipating climate-related risks and adapting to climate change (Wilkinson et al. 2015). For example, meteorological records clearly show a change in precipitation patterns globally over the past decade (Behailu et al. 2021). However, studies have suggested that smallholder farmers have limited or no access to such customized and location-specific weather information, but are willing to engage with CS. This offers the opportunity to have a solid weather forecast system, especially for Benin in West Africa where the consequences of climate change are dire. While CS is acknowledged to be very important to help farmers to better respond to climate variability, end users will need to have a level of ownership to be able to access meteorological information for decision-making (Brest and Yann Quilcaille 2015).

Furthermore, a number of studies have found that farmers have limited or no access to customized and location-specific weather information in West Africa, but are willing to engage with CS (Amegnaglo et al. 2017; Yegbemey et al. 2020). However, in terms of awareness and access, Buckland and Campbell (2021) found that demographic, socio-economic, and environmental factors contribute to farmer awareness and access to Climate Services, isolating gender, age, access to extension service, participation in groups/organizations, climate change perceptions, non-farm income, farm size, and agronomic conditions as being significant influencers of farmer awareness, access to, and use of Climate Services.

Nsengiyumva et al. (2022) also found that gender, headship, and wealth status influence responses to climate information and decision-making tools in Rwanda, and in so doing highlight important implications for the design of Climate Services and similar interventions. Friedman et al. (2022) acknowledge that accessibility and use of CS may differ dramatically between groups of farmers, highlighting the need to understand the different paths of information exchange within communities in Papua New Guinea. They noted key gender differences, with women demonstrating stronger information connectivity with family and friends, while men relied on their networks to potentially bridge the gap between external information sources like media, community leaders, church groups, and friends and family.

Despite the importance of CS, the literature review shows the paucity of information on smallholders’ knowledge, attitude, and use of CS for improved farm operations. Against this background, this paper analyzes farmers’ awareness of CS in Benin within the framework of their KAP.

Material and methods

Study area and sampling

The study was conducted in five municipal areas across Benin: Tchaourou, Natitingou, Malanville, Lokossa, and Kétou (see Fig. 1). These areas were identified as agricultural production areas vulnerableFootnote 1 to the effects of climate change according to the Programme d’Action Nationale d’Adaptation au changement climatique and the Ministère du Cadre de Vie et du Développement Durable (MCVDD 2019). The main crops produced in the selected areas include maize, rice, soybeans, and cassava. Within each municipal area, villages (Table 1) were identified in consultation with representatives of farmer organizations and the local agricultural extension officers based on the relative importance of agricultural production and the perceived level of vulnerability to climate variability. Within each village, the sampling units were households, represented by the household heads. We relied on the Central Limit Theorem to select randomly, up to 36 households in each village, implying a total of 540 households. The random selection was done by using the table of random numbers and a list of households generated through a rapid census survey.

Fig. 1
figure 1

Map of the study area

Full size image
Table 1 Sample distribution for the household (HH) survey
Full size table

Data collection and modelling

Primary data was collected through a household survey using a structured questionnaire administered individually to each respondent through a Computer-Assisted Personal Interviews (CAPI) system (i.e., Open Data Kit, ODK). The main data collected were socio-economic characteristics of farmers including age, gender, level of education, socio-cultural group, income, main and secondary activity, household size, number of agricultural workers, area, membership of a farmer organization, farming experience, access to credit, and possession of Information and Communication Technologies (ICT) (i.e., telephone, radio, or television). We measured the score of knowledge about weather information through dummy (yes = 1 and no = 0) responses to a set of 10 questions. The score of attitude with respect to weather information was measured as a dummy variable (yes = 1 and no = 0) in response to a set of 14 questions. The score of practices in relation to weather information was also measured as a dummy variable (yes = 1 and no = 0) set of 16 questions.Footnote 2

In line with Nguyen et al. (2019), this study used the KAP approach to assess farmers’ CS knowledge, attitudes, and practices. Knowledge is defined in context as the comprehension of information. A situation’s attitude can be positive or negative. An ongoing activity that is shaped by societal norms and ideas is practiced. The learning theory (Bandura 1976) and the diffusion of innovation theory (Roger 1995) are the models’ sources for KAP.

Numerous connections between knowledge, attitudes, and practices have been found in earlier research (e.g., Valente et al. 1998). The Knowledge, Attitudes, and Practices (KAP) model was created and is currently regarded as one of the most well-liked survey tools in the field of social research. According to Andrade et al. (2020), the KAP model is a systematic, standardized questionnaire that a target population must complete in order to quantify and analyze what is known (knowledge), believed (attitudes), and practiced (practices) with reference to a topic of interest. As a result, the KAP model data can assist in identifying knowledge gaps, attitudinal obstacles, and practice patterns that may facilitate comprehension and action with respect to a particular issue.

In practice, we designed a KAP model (see Online Appendix) that we implemented in the survey. The significance of the model is on three fronts, as it captures the knowledge levels of the smallholder farmers about CS, their attitude towards CS, and the level of practice of information received via Climate Services. The KAP scores were computed following a number of previous studies (e.g., Roy et al. 2020). Each score was equal to the arithmetic sum of the dummy responses to its related questions. We used descriptive analysis through means (averages) and standard deviations, to present the socio-economic profile of the respondents and the KAP scores. Additionally, based on the assumption that the KAP items are likely to be correlated, we specified a Seemingly Unrelated Regression (SUR) model to analyze the socio-economic determinants of the scores (S) as follows:

$${S}_{ij} = f\left({X}_{i}\right)$$
(1)

With the index j for the knowledge score, the attitude score, or the practice score (j = 1, 2, and 3). Sij is the KAP score j of farmer i; Xi represents the socio-economic and demographic factors linked to farmer i and explains the different levels of indicator type score j of the respondent farmer. We translated Eq. 1 in a set of equations as follows:

$$\{ {S}_{1i}={\alpha }_{1}+\sum\limits_{n}{\beta }_{1n}{X}_{ni}+{\mu }_{1i} {S}_{2i}={\alpha }_{2}+ \sum\limits_{n}{\beta }_{2n}{X}_{ni}+{\mu }_{2i} {S}_{3i}={\alpha }_{3}+\sum\limits_{n}{\beta }_{3n}{X}_{ni}+{\mu }_{3i}$$
(2)

S1, S2, and S3 are the scores of KAP of Climate Services, respectively; α represents the constants; β stands for the coefficients of the socio-economic and demographic factors X; and µ are the random factors (errors). Equation 2 was estimated in Stata 15 software and β is interpreted to analyze the factors that explain the KAP scores.

We selected the following socio-economic variables to run our model specification based on our knowledge of the field and the existing literature:

  • Age: is a quantitative variable measured in nominal years. Ouédraogo and Dakouo (2017) showed that age does not have a significant effect on knowledge and practices of agricultural innovations. Thus, we hypothesized that this variable would not have an effect on the KAP scores.

  • Gender: is a qualitative variable that takes the value 1 if the household head is a man and 0 if it is a woman. Several studies illustrated the possibilities of gender gaps between male and female farmers. For instance, Atreya (2007) reported that men have a very good level of KAP on good pesticide use practices, unlike women in Nepal. Yegbemey et al. (2013) showed that men are more likely to explore new ways to improve agricultural production compared with their women counterparts in Benin. We anticipated a positive correlation between gender (i.e., being male) and the KAP scores.

  • Household size: is a quantitative variable that indicates the number of people in the household. We anticipated that larger households are more likely to have higher KAP scores as they could be more exposed to different sources of information and also need to have a relatively higher production level.

  • Membership in an organization: is a qualitative variable taking the value 1 if the farmer belonged to an organization and 0 if otherwise. According to Zossou et al. (2021), membership in an organization is strongly correlated with access to certain services such as training and information. We hypothesized a positive relationship between this variable and the KAP scores.

  • Formal education: is a qualitative variable that takes the value 1 if the farmer is formally educated, and 0 if otherwise. Education is often associated with knowledge. For instance, Moffo et al. (2020) reported a strong positive correlation is observed between the level of education and KAP score on antimicrobial practices widely used in poultry farms in Benin. Çakmur et al. (2015) showed a statistically significant positive correlation between the high level of education and the knowledge of breeders on zoonotic diseases in Turkey. It was anticipated that formal education would have a positive effect on the KAP scores.

  • Access to ICT: access to ICT in the context of this work comes down to the possession of at least one communication tool (telephone, radio, or television). It is a qualitative variable that takes a value 1 if the farmer has at least one communication tool and 0 if otherwise. Farmer’s possession of communication tools is strongly related to the level of KAP (Rausser et al. 2010). We anticipated that access to ICTs would have a positive effect on the KAP scores.

  • Access to credit: is a qualitative variable taking the value 1 if the farmer has access to credit, and 0 if otherwise. Farmers with access to agricultural credit will tend to increase their knowledge, develop positive attitudes, and adjust their production practices to CS. We hypothesized a positive relationship between access to credit and the KAP scores.

  • Cotton and rice production: These are qualitative variables taking the value 1 if the farmer produces the crop, and 0 if otherwise. We hypothesized that those who produce cotton or rice will likely have higher KAP scores as these are important cash crops in the study area.

Results

Socio-economic characteristics of respondents

Table 2 shows some descriptive statistics of the socio-economic characteristics of the respondents.

Table 2 Descriptive statistics of the socio-economic characteristics of the respondents
Full size table

The majority of the farmers who were randomly sampled are males with an average age of 42 years, and having up to 18 years of experience in farming. Households are relatively large with an average membership of 10 individuals. About 60% of the household members are involved in agricultural production activities. While access to credit is low, ownership of a mobile phone and willingness to engage with free or paid CS were high. We also noted that the majority of farmers had information and communication technology tools.

Respondents’ knowledge, attitude, and practice in relation to weather information and Climate Services

Table 3 presents descriptive statistics of the KAP scores.

Table 3 Descriptive statistics of the Knowledge, Attitude, and Practice (KAP) scores
Full size table

Farmers in Benin had below-average scores in knowledge and practice with weather and CS. On the contrary, the attitude scores of farmers were above average, indicating that farmers have limited knowledge and poor practice but a positive attitude towards weather information and CS. This can be explained by the fact that CS and weather information are not yet available to most of the farmers but with the increasing climate variability, they are important tools that can build farmers’ resilience.

Detailed statistics of the items involved in each KAP score suggest that in terms of knowledge, about half of the respondents have heard of weather/climate information or weather/climate forecasting, can define somehow the concept of weather/climate information or weather/climate forecasting, know that weather or climate information or weather forecasts can be seasonal, and also know that weather or climate information can be realized or not realized. The main knowledge gaps are related to the understanding of CS as the timely production of weather/climate information or data as a function of a climate service and the translation of climate information into use.

Looking at the attitudes, more than 70% of the respondents believed that weather or climate information can be reliable and useful for better planning of production activities, as well as being able to provide economic benefits. Also, more than 70% of the respondents were not sure that weather or climate information could provide specific economic benefits such as the rational use of labor or pesticides.

On the practice towards weather information, around 50% of the respondents claimed that they received or accessed weather forecast information in the past. This percentage drops to 25% if we look at farmers who claimed that they have ever accessed daily forecast information in 2020. It further drops to 6% when the respondents are asked if the information was specific to their village or if they used the information to make decisions about farming activities.

Determinants of the Knowledge, Attitude, and Practice scores

Table 4 presents the results of the regression model to analyze the drivers of the KAP scores. It is worth noting that the design of the study is limited, and did not allow for the analysis of subgroups such as young male/female and adult male/female, which could have explored the heterogeneity of the drivers of KAP scores across different categories of the sample.

Table 4 Result of the Seemingly Unrelated Regression (SUR) model
Full size table

The regression model suggests that KAP scores are influenced by age, gender, formal education, ownership of television and phone by the household, professional training, access to credit, cotton, and rice production. We observed consistent directions of the effects of these variables regardless of the KAP scores.

Age and professional training have positive effects on knowledge score. Formal education and ownership of a television have positive effects on Knowledge, Attitude, and Practice scores. Phone ownership has a positive effect on practice score. Gender and access to credit have positive effects on knowledge and attitude scores. Cotton production and rice production have positive and negative effects, respectively, on practice score.

Discussion

CS provides weather information that is essential for improving the resilience of farmers to climate change (Onyegbula and Oladeji 2017). Knowledge of weather information can allow farmers not only to decide on the dates of plowing and sowing but also to better plan the application of fertilizers and pesticides (Traoré et al. 2017). Yet, as a general pattern, the knowledge and practice scores are low while the attitude score indicates a willingness to engage with weather information and CS. This is a positive signal and indicates key entry points for interventions to promote weather information and CS. According to Gnanglè et al. (2011) and Vodounou and Onibon Doubogan (2016), farmers in Benin adapt to climate variability by developing their own endogenous adaptation strategies. Weather services in Benin provide state or municipal area-level information only, which might not be applicable at the village level.

The results affirm that the farmers were well organized in groups but with little access to financial services. When farmers are in organizations, they could have easy access to certain services such as training, information, grant support, and even access to financial services to support farm enterprises. Access by the majority of the farmers to mobile phones also meant that they could have enhanced access to information. Zossou et al. (2021) noted that organizations in general and particularly farmer organizations create an incentive framework for learning and facilitate access to useful information for agricultural activities. Access to credit is limited. These statistics compare well with some figures from the National Agricultural Census (NAC) survey (Recensement National de l’Agriculture, RNA 2019). For instance, respondents’ age is about 42 years old on average in our sample compared to 43 years old in the NAC. Our sample has about 83% of male farmers against 84% in the NAC, and the household size is about 9 people on average in our sample against 7 in the NAC.

The results highlight that age positively affects the score of knowledge about weather information and CS, corroborating with Buckland and Campbell (2021). Farmers learn from experience, develop skills, and better understand how to improve agricultural production as they grow older. All this is likely to increase their propensity to learn or seek information on weather information and CS. We found positive relationships between formal education and each of the KAP scores. Educated farmers are indeed more likely to be exposed to diverse sources of information such as radio and newsletters. According to Ayedegue et al. (2020) and Zakaria et al. (2020), education contributes to agricultural human capital development and could improve farmers’ ability to choose adaptive strategies to climate variability.

Ownership of ICT tools improved the KAP scores. TV and radio are among other channels where weather forecasts are broadcasted even if they are not necessarily in a format that could be understood or specific/relevant to the location of each farmer. According to Azumah et al. (2018), ICTs promote access to information and are important tools for agricultural growth in developing countries and a vector for improving the living conditions of farmers.

Gender was found to influence the knowledge and attitude scores. Male farmers were more likely to have a better knowledge and a more positive attitude. In the settings of the study area, male farmers are most of the time the household heads. As such, the primary responsibility of caring for the family rests with the male farmers who are likely to actively seek information on how to adapt to challenges such as climate variability (Yegbemey et al. 2013).

Azumah et al. (2022) and Yegbemey et al. (2014a, b) found that farmers who accessed credit are more likely to adopt improved farming strategies, especially those that require additional investments. Our results in this study reveal that access to financial services can improve knowledge and attitude scores. By accessing finance, farmers can invest in their human capital, educate themselves, and be able to invest in new technologies for improved farm outcomes. Therefore, rolling out CS with village savings and loan programs (to support smallholder farmers’ financial needs) could enhance their awareness and use of CS.

Cotton production promotes the practice level of weather information. Cotton is a priority sector benefiting from the particular attention of public authorities and private sectors. In Benin, cotton production is of paramount importance in the economic fabric of the country, thus contributing to a share of 7% of PIB (Banque Mondiale 2016). This indicates that in Benin, the cotton sector is structurally and technically well organized and the farmers are trained and monitored. The yield in cotton production in recent years has dropped drastically from more than 1200 kg/ha in 2002 to 1080 kg/ha in 2019 (Institut National de la Statistique et de l'Analyse Economique, INSAE 2019). Aware of the key role of climate variability in this situation, farmers are trained and informed on practices that are resilient to the effects of climate variability. Household size, membership in a farmer group, and ownership of a radio did not, however, affect levels of knowledge, attitude, and practice of CS but could be given some credence in a wider program design to tackle the effects of climate change in Benin.

By way of limitation, the design of the study and sampling strategy did not allow for the running of a robust subgroup analysis to look at the dynamics within young male/female and adult male/female subgroups. Therefore, future research could account for that to explore the heterogeneity of the drivers of KAP scores across subsets of the sample. The choice of household headship as the study unit meant that more males than females were sampled for the study because household headship in Benin is largely male-dominated. Also, because we used indices (aggregated responses from multiple questions), it is impossible to tell if certain questions influenced the KAP scores more than others. A future study could look at the relationships between the KAP variables and not just what factors influence each score. A study on how knowledge and attitudes influence practice could also be conducted.

Conclusions and policy implication

This study establishes that Benin’s smallholder farmers are aware of the negative effect of climate variability on farm output, but have little knowledge of Climate Services. Nevertheless, they showed a positive attitude and are ready to use Climate Services to enhance their farm operations. Development agencies should consider integrating formal education in programming and promoting televising as key factors in the design of comprehensive Climate Services (CS) models for farmers. These efforts would enhance farmers’ knowledge, attitude, and use of CS, thereby improving resilience against climate change and variability.

The agricultural ministry of Benin should prioritize targeted professional training for farmers of all age groups to enhance their understanding of CS and how to use weather information. Recognizing that age may not influence attitude and practice, investing in knowledge-focused training initiatives will empower farmers to embrace effective CS strategies. There is also the need to implement initiatives that address gender disparities and facilitate access to credit, with a specific focus of enhancing farmers’ knowledge and fostering positive attitudes towards CS in Benin. While these factors may not directly affect practice, they play a crucial role in shaping the foundation for the use of CS. Despite the limited influence of telephone ownership and rice/cotton cultivation on knowledge and attitude, they are strategic to boost farmers' use of CS.