The study was conducted in the Department of Kolda, located in southwestern Senegal, between 12˚52' north latitude and 14˚57' west longitude ( Figure 1 ). Covering an area of 13,721 km², approximately 7% of the national territory [24] , the department had an estimated population of 306,591 in 2023 [25] . It is bordered to the north by the Gambia, to the east by the Tambacounda region, to the west by the Sedhiou region, and to the south by Guinea-Bissau and Guinea-Conakry [24] . The area is subject to a Sudano-Guinean climate characterized by an average rainfall of 1087.055 mm, with maximum temperatures ranging from 30˚C to 40˚C between March and September and lowest temperatures from 25˚C to 30˚C from December to January [26] .

Agriculture is a dominant activity, practiced by 70.07% of households. It relies mainly on rain-fed agriculture, which makes production highly dependent on climatic conditions. The main crops are cereals, legumes, and vegetables. After millet and fonio, maize is the most widely grown cereal on family farms in the Kolda region. Maize is one of the most important crops, with an area of 10,820 ha sown in 2022 [26] . Generally cultivated around concessions, maize has a crop cycle of 75 to 90 days, from sowing to maturity [27] .

Since, at the time of the study, there was no database on the size of the maize-growing population in the study area, the sample size was determined using the formula of [28] , described by [29] :

$no=\frac{Z^{2}pqN}{e^2\left(N-1\right)+Z^{2}pq}$ (1)

with: no: sample size; z: the standard value of 1.96 at a confidence interval of 95% and a precision of ±5%; p: the estimated share of the population = 0.1; q: 1 − p; N: total number of households in the study area (N = 5831); e: the desired level of precision (e = 0.032).

The value of p = 0.1 was chosen based on preliminary information and expert knowledge indicating that only a small share of maize producers was expected to present the characteristic of interest. Using this estimate rather than the conservative p = 0.5 prevented an unrealistically large sample size while remaining consistent with local expectations. The precision level e = 0.032 (±3.2%) was selected as a compromise between the standard 5% margin and field constraints, ensuring reliable estimates while keeping the required sample size feasible.

The resulting sample size is as follows:

$no=\frac{{\left(1.96\right)}^2\ast 0.1\ast 0.9\ast 5831}{\left[{\left(0.035\right)}^2\ast \left(5831-1\right)\right]+\left[{1.96}^2\ast 0.9\ast 0.1\right]}=319.5$

Although the initial target was 319 respondents, according to the value predicted by the sampling method, the survey successfully covered 300 maize growers, representing a robust and diverse sample despite the limited availability of some participants during the data collection phase.

Data collection was based on a questionnaire developed from an in-depth review of the scientific literature to ensure methodological rigor and the relevance of the information gathered. To adapt the tool to the local context and validate the clarity of the questions, a pilot survey was conducted in the study area. The questionnaire, comprising 118 questions divided into seven main sections: 1) socio-demographic characteristics of respondents, 2) specific features of the maize cultivation system in the area, 3) producers’ perception of climate change risk, 4) climate-smart agricultural practices, 5) motivations for adopting climate-smart agricultural practices, 6) factors influencing adoption, and 7) constrains encountered during adoption. The questionnaire was designed using the KoboToolbox platform (version 2.024.05) and then administered via the KoboCollect application (version 2023.2.4), which enabled efficient and secure data collection between October 28 to November 23, 2024.

The survey was conducted in the 13 villages where maize production was highest.

Data collection was carried out using the snowball method, where selected initial participants helped to identify other potential participants. In the interests of a higher response rate and contextual observation, this method was combined with the door-to-door method, which involves going directly to individuals’ homes to collect data.

Basic descriptive statistics, such as mean, frequency, and percentage, were used to present the results. To assess climate change risk perception, the climate-smart agricultural practices adopted and the associated promoting or hindering factors of their adoption in the study area, scoring criteria were defined, scores calculated, and a ranking made.

Maize growers’ perception levels of climate risk were ranked using the Climate Change Risk Perception Score (CCRPS). The respondents were asked to share their views on eight climate change risks to gain a thorough understanding of their perceptions regarding these specific climate change risks. For the estimation of CCRPS, a list of eight questions was asked with a 4-point Likert scale [30] : 0 = “no risk perception”; 1 = “low risk perception”; 2 = “moderate risk perception”; 3 = “high risk perception”. Low perception was assigned to households that expressed little concern for climate change and had a low perceived exposure to its effects. In contrast, high risk perception was assigned to households that expressed high concern and high perceived exposure to its impacts. In moderate risk perception, more measured statements about the severity and urgency of climate change were included.

The CCRPS was calculated by using Equation (2):

$\text{CCRPS}=\text{CCRPn}\ast 0+\text{CCRPl}\ast 1+\text{CRPm}\ast 2+\text{CCRPh}\ast 3$ (2)

where: CCRPn: number of households with no perception of risk, CCRPl: number of households with a low-risk perception, CCRPm: number of households with medium risk perception, and CCRPh: number of households with a high-risk perception.

The maximum possible CCRPS value is reached when all 300 maize growers fall into the “high risk perception” category, giving a maximum of 300 × 3 = 900. This reference value was then used in the calculation of the Standardized Climate Change Risk Perception Index (SCCRPI) (Equation (3)), so that the results could be interpreted more easily:

$\text{SCCRPI}=\frac{\text{CCRPS}}{\text{Maximum}\text{CCRPS}\text{value}}\ast 100$ (3)

Similar methods have been employed in previous studies by [31] and [32] to assess perceptions of climate change risk in Bangladesh and Ethiopia, respectively.

The climate-smart agricultural practices used were identified and ranked using the Relative Importance Index (RII). In addition to assessing the level of use of these practices, the RII also helps prioritize knowledge and practices in order of frequency of adoption, so that adaptation actions or support can be targeted at the right practice(s). This method has been used in previous research by [33] and [34] to examine the most popular agroecological practices among small-scale farmers in northern and eastern Ghana, respectively. Formula (4) was used to calculate the RII:

$\text{RII}=\sum \frac{W}{A\times N}$ (4)

where, W is the weighting of a particular response on a Likert scale [30] with 4 response modalities: never used = 1, rarely used = 2, often used = 3, and used every year = 4. A is the highest response (4), and N is the number of respondents taken into account (300).

WAI is a statistical analysis method that determines the mean of an outcome by multiplying the weight given to a particular event by the associated quantitative result, then adding up all the results. This is very useful for determining a theoretically expected outcome where each event has different probabilities of occurring. WAI is therefore crucial in determining the factors motivating producers to adopt climate-smart agricultural practices. The WAI was estimated based on a 4-point Likert scale [30] : 1 = “disagree”; 2 = “more or less agree”; 3 = “agree”; 4 = “strongly agree”.

Several studies have used the WAI to evaluate climate change adaptation measures [35] [36] . The WAI is mathematically given by Formula (5) [37] [38] :

$\text{WAI}=\frac{\sum F_i{W}_i}{\sum F_i}$ (5)

where: F denotes frequency, W represents weight or importance, and it shows the score of each factor and opportunity related to climate-smart agricultural practices, or the score of each factor that can influence the adoption of climate-smart agricultural practices.

To classify the barriers to the adoption of climate-smart agricultural practices, the Problem Confrontation Index (PCI), given by Formula (6), was used. This index has been used in previous studies, e.g., [35] [36] [39] , and [40] , to identify and rank perceived barriers that prevent farmers from adopting climate-smart agricultural practices. For the estimation of PCI, a list of questions was asked with a 4-point Likert scale [30] to classify the obstacles that hinder the adoption of climate-smart agricultural practices: 0 = “disagree”; 1 = “more or less agree”; 2 = “agree”; 3 = “strongly agree”.

$\text{PCI}=Pn\times 0+Pl\times 1+Pm\times 2+Ph\times 3$ (6)

where Pn is the number of maize growers who rated the obstacle as no problem, Pl is the number of maize growers who rated the obstacle as low-level, Pm is the number of maize growers who rated the obstacle as moderate-level, and Ph is the number of maize growers who rated the obstacle as high-level.

Table 1 shows the socio-demographic characteristics of maize growers in the study area. Analysis shows that men make up the majority in the maize growing activity, representing 66.67% of the sample, compared to women who only represent 33.33%. Looking at the age range, we see that a large proportion of these growers (47%) are adults, aged between 35 and 50, while those who are over 50 and less than 35 are most poorly represented (19%). Regarding ethnic groups, the majority of respondents (94.6%) are Fulani, while other groups have low proportions: Mandingo 4.4%, Diola (0.6%), and others 0.3%. Considering the marital status of the respondents, the majority of them are monogamous (50%) or polygamous (45.67%), while those who are widowed and singles represent a marginal share with 2.33% and 2% respectively. In terms of education level, a large proportion of the growers surveyed have no formal education (32%), unlike the others who have varying types and levels of education, dominated by Qur’anic (28%) and Basic school (26%). The analysis also showed that people with Middle school, High school, and Tertiary education were less represented, with respective frequencies of 8%, 3.67%, and 2.33%. Nearly 81% of growers are members of a farmers’ organization. In terms of household size, the average was 13 members. It was also noted that producers have an average of 27.9 years’ experience in maize production. The average area dedicated to maize growing is 3.19 hectares, with an average yield of 2179.19 kg/ha in 2023. In the same year, the total volume of maize produced was 8629.59 kg, divided between domestic consumption (740.61 kg) and sales (1330.98 kg). Maize sales generate an average income of 331,316 Fcfa. Furthermore, the labor force devoted to the cultivation system averages 2.2 workers per farm.

The results presented in Table 2 shed light on how farmers in the study area have perceived climate change risk over the past ten years. The analysis highlights that the most strongly perceived hazards are rising temperatures SCCRPI = 92.11), recurrent droughts (SCCRPI = 87.33), and decreasing rainfall (SCCRPI = 73.67). Other climate phenomena, such as delayed rainfall, shorter rainy seasons, and stronger winds, were also frequently reported by farmers during the same period.

Note: 0: No risk perception; 1: Low risk perception; 2: Medium risk perception; 3: High risk perception.

Table 3 shows the relative importance of climate-smart agricultural practices adopted by maize growers. The analysis reveals that among the 15 identified climate-smart agricultural practices, drill seeding, crop rotation, use of recommended chemical fertilizers, assisted natural regeneration, and early sowing were ranked first, second, third, fourth, and fifth, with RII values of 0.998, 0.864, 0.827, 0.75, and 0.703, respectively. The least adopted climate-smart agricultural practices are planting living hedges or windbreaks (RII = 0.412), the application of water conservation techniques (RII = 0.278), and the use of crop irrigation (RII = 0.272).

Note: RII: Relative Importance Index.

Table 4 represents the factors that motivate growers to adopt climate-smart agricultural practices. The results show that improving food security, increasing yields, and improving incomes are, in order of importance, the key reasons that motivate producers to adopt climate-smart agricultural practices. However, wind erosion control, soil protection, and maintaining soil fertility are less motivating for the adoption of climate-smart agricultural practices, according to the respondents.

Table 5 represents the factors favoring the adoption of climate-smart agricultural practices. The results indicate that understanding the effects of climate change, access to weather and climate information, access to secure land, and support from producer organizations are the main factors promoting the adoption of climate-smart agricultural practices, with respective WAI values of 3.91, 3.84, 3.84, and 2.96. And the factors that have less influence on the adoption of the climate-smart agricultural practices are access to sustainable agricultural technologies, government support, and access to agricultural finance to implement those practices.

The results in Table 6 show that difficult access to agricultural credit (PCI = 738), difficulties in marketing agricultural products (PCI = 718), weak government support (PCI = 628), high cost of agricultural inputs (PCI = 582), lack of capacity building for producers (PCI = 464) are ranked in this order as the most important obstacles that can block the adoption of climate-smart agricultural practices by maize growers. Factors such as unavailability of improved seeds (PCI = 455), lack of manpower (PCI = 427), poor soil fertility (PCI = 394), lack of access to land (PCI = 250), and limited access to weather information (PCI = 190) also appear as limitations blocking the adoption of cli-mate-smart agricultural practices, but to a lesser extent, compared with the first obstacles cited.

Notes: WAI: Weighted Average Index.

Notes: PCI: Problem Confronting Index.

This study shows that, in the study area, men are more involved in maize production than women. This predominance of the male gender in this production sector can be explained on the one hand by the fact that men have much greater access to land, and on the other by the fact that women are much more concentrated in households or constitute the labor force. These results concur with those of [41] , [32] , and [42] , where men are predominantly represented in agricultural activities. Most producers are adults in the 35 - 50 age bracket. This can be explained by the fact that most producers are heads of household. These results are in line with those of [32] [43] and [42] , who stated in their studies that the majority of producers are adults. Of the farmers who attended school, most stopped at elementary school. This could be explained by the fact that farmers drop out of school early to focus on agricultural activities as their main source of income. These findings are in line with those of [42] , who revealed that the majority of farmers in Africa drop out of school very early. The average number of years of farming experience for producers is almost 28 years. This could be explained by the fact that the majority of active producers are adults who have been active in the agricultural sector for a long time. The majority of farmers are Fulani. This may be because the area is largely dominated by this ethnic group.

It was noted that rising temperatures, recurrent droughts, and reduced rainfall were mainly identified as the greatest climatic risks perceived by maize growers in the study area. This can be explained by the fact that these changes are more perceptible because of their direct and depressive influence on agricultural production. These findings corroborate those of [31] and [32] , who reported a greater perception of droughts, delayed onset of rainfall, and short duration of the rainy season in Bangladesh and Ethiopia, respectively.

The results show that drill seeding, crop rotation, assisted natural regeneration, and early sowing are the practices most widely used by growers in the study area. This could be explained by the fact that this practice is better suited to crop management. In fact, in Senegal, growers cultivate mostly with bovine, equine, or asine traction. In addition, State-owned companies supporting producers, such as ANCAR (National Rural Agricultural Consultancy Agency), supply agricultural equipment such as seed drills, hoes, etc., making it easier for growers to practice drill seeding. In fact, according to [43] , line seeding facilitates crop management. These three practices have been considered as CSA by previous studies such as [39] and [43] . State subsidies for chemical fertilizers and their availability on the market make it easier for growers to access these types of inputs. In addition, the training of growers in the use of chemical fertilizers by projects or NGOs also facilitates their use. At the same time, trees play a decisive role in agricultural systems, providing people with a means of subsistence. Growers aware of the importance of trees in their plots tend to assist natural regeneration. Moreover, early sowing is adopted in the study because of the early onset of the rainy season. And as soon as the first rain falls, growers sow the maize so that it can complete their production cycle. The study carried out by [44] in the Kolda region showed the occurrence of early rainfall episodes. These results are in line with those of [39] in Ghana, [45] in Kenya, [46] in Vietnam, and [47] in Guatemala, all of whom reported the use of this practice by growers to improve the structure and fertility of their soils.

However, irrigation and water management techniques remain largely unadopted by producers, largely because maize cultivation in the Kolda region relies primarily on rainfall and covers vast areas, making irrigation difficult, if not impossible. In addition, certain practices such as installing hedges or windbreaks around fields, planting trees within plots, or soil conservation techniques are among the least used. This low adoption rate could be explained by a lack of awareness or outreach about these practices among farmers.

The results of this study reveal that, according to producers’ perceptions, the application of climate-smart agricultural practices (CSAPs) improves household food security, increases yields, and improves incomes. These are important motivating factors as they regulate the decisions of maize growers in the study area to adopt CSAP. This is linked to the importance attached by small-scale farmers to food security and household income sources. These findings are close to those of several studies in Sub-Saharan Africa where smallholder farmers adopted climate-smart agricultural practices based on their ability to increase household security and improve agricultural yields and on-farm income [48] - [51] .

The results reveal that understanding the effects of climate change, access to secure land and climate information, support from producer organizations, and access to agricultural insurance are factors that facilitate the adoption of climate-smart farming practices. These findings are in line with those of [39] , who identified secure land tenure, understanding of the effects of climate change, and access to sustainable agricultural technologies as the main factors promoting the adoption of climate-smart agricultural practices. For instance, [49] observed that households and individuals with access to secured land tenure invested in agroforestry practices as an adaptation practice to improve food and livelihood security. This is linked to the fact that most growers in the study area have access to climatic information. Secure land ownership encourages growers to invest in or implement intelligent farming practices in their crop plots, unlike those without titled plots. What’s more, the organizations that support producers often provide training and financing. In addition, agricultural insurance, which compensates for economic losses due to damage caused by unpredictable weather events, encourages producers to develop more resilient farming practices in a context of climate change. It enables farmers to prepare for the risks associated with climatic shocks and to limit the possible consequences in the event of an incident.

This study reveals that limited access to agricultural credit is the main obstacle to the adoption of climate-smart agricultural practices. This constraint can be explained in particular by the cumbersome financing procedure, which is often lengthy and unresponsive, as well as by the geographical distance between financial institutions and the areas where producers live. These findings are in line with those of [52] and [46] , who asserted that access to credit has a positive correlation with the adoption of climate-smart agricultural practices. As far as the marketing of agricultural produce is concerned, growers are finding it extremely difficult to sell their harvests. This can be explained, on the one hand, by the fact that producers do not have the means to store their produce after harvest and, on the other hand, by the low selling price of produce found on the market. These results are in line with those of [53] , who found that marketing constitutes a problem to the intensification of maize production for producers. Among other obstacles, we have low government support, high cost of agricultural inputs, lack of training for producers, unavailability of improved maize seeds, etc. Indeed, the high cost of fertilizers and climate-smart farming interventions requires appropriate access to credit and funds to purchase the necessary inputs.