The latest census “Recensement général de la Population, de l’Habitat, de l’Agriculture et de l’Elevage (RGPHAE)” carried out in 2013 by the “Agence nationale de la Statistique et de la Démographie” or National Agency for Statistics and Demography [74] constitutes the sampling frame and has enabled a breakdown of the number of vegetable farmers according to the four regions that make up the Niayes zone. A systematic random selection was made to select six (6) of the nine (9) that make up the Niayes zone.

The more vegetable farmers and/or a sufficient number of motor-driven pumps for pumping and/or the more public or private intervention for the development of solar pumping systems, the more likely the department is to be chosen. The same is applied to the choice of communes. A total of 401 producers were sampled and interviewed in the field.

The expectation of higher future profits determines technology adoption. The producer will only adopt the new technology if the anticipated profitability is higher than that of non-adoption [75] [76] . If the anticipated utility of the target stage is higher than that of a reference stage, the producer moves from the reference stage to the target stage of the adoption process. For this modelling, we assume the existence of a latent variable $y_i^{*}$ that depends on a combination of characteristics $x_i$ relating to the producer, the cultivated plot but also the spatial dimension.

Let’s assume that the producer’s unobserved utility varies according to the stage of the adoption process. Here, adoption follows three stages, and at each stage it is assumed that an unobserved utility, deemed superior, makes it possible to move from the reference stage to the target stage.

Let $\left(y_j^{*}\right)$, j = 1, 2, 3 be an unobserved utility, derived from knowledge of the technology $y_l^{*}$, the first test $y_2^{*}$, and their continued use $y_3^{*}$.

The latent utility $y_l^{*}$ depends on a vector $X_1$ of $k_1$ explanatory variables of technology knowledge and $y_2^{*}$ depends on a vector $X_2$ of $k_2$ variables at the origin of the first test and $y_3^{*}$ depends on a vector $X_3$ of $k_3$ explanatory variables of continuous use of solar energy. Each utility is associated with a binary variable $y_i$ taking either the value 0, 1, 2, and 3.

Then,

$Y_J=1,if{Y}_J^{*}>0,j=1,2,3$

$Y_J=0,if{y}_J^{*}<0$

However, the utility $y_j^{*}$ can only be observed in an individual when the individual has knowledge to the technology, i.e. that $y_j^{*}$ > 0.

Thus, the model is specified as follows:

Technology awareness phase: Y_{i awareness} = 1 $\left[{X^{\prime }}_i{}_j{\beta }_{awareness}+{\epsilon }_{awareness}>0\right]$ (1)

Technology test phase Y_{i test} = 1 $\left[{X^{\prime }}_i{}_{j}test+{\epsilon }_l>0\right]$ if Y_{i awareness} = 1

Y_{i test} = 0 Otherwise(2)

Technology adoption phase Y_{i adoption} = 1 $\left[{X^{\prime }}_i{}_{j}adoption+{\epsilon }_l>0\right]$ if Y_{i test} = 1

Y_{i adoption} = 0 Otherwise(3)

where β_j et ε_j are respectively the vectors of coefficients to be estimated and of errors.

The conventional approach for this type of model would be to apply the ordered logistic model [77] . However, these decisions are not only ordered but also sequential, as the attainment of one modality is conditional on the attainment of the preceding modalities. Moreover, there is also an aspect of self-selection for each step. Thus, the estimates of the ordered logistic model would be biased as they do not take into account the self-selection effects included in the model [78] . Consequently, the econometric model adapted for this study is the sequential logit model widely used in the literature [79] .

These statistics are presented in Table 1 . The 401 farmers surveyed, with an average age equal to 38.97 years, are 90% male and 10% female. This is justified by the fact that men are in most cases responsible for the plots or landowners. However, there are more women in the marketing and processing activities not analyzed in this work. Adopters, i.e. those currently using solar pumping systems (PVs), are relatively younger than non-adopters, and this could be a determining factor in their decision to adopt them. The average household size of the producers surveyed, estimated at 9.2 individuals, is on average 9.7 members for adopters versus 9 for non-adopters. This small difference in size is significant at the 1% level. With regard to education, for all the producers surveyed, around 46% had attended Koranic school, 30% had at least primary school in formal education, and 23% had no education at all. Here, education takes into account not only formal education but also Koranic schooling, as most producers know how to keep their accounts in Arabic.

In addition, the results show that adopters are better educated than non-adopters. Around 74% of adopters are educated, compared with 70% of non-adopters. Indeed, as the Niayes zone is very close to the big cities, most of these inhabitants have at least attended school, even if only for a year. Access to credit is more marked among adopters than non-adopters. With regard to training, adopters who have received training are almost as numerous as non-adopters.

Concerning the distance of the plot from the market or the main road, the results show that adopters’ and non-adopters’ plots have the same distances on average.

The method of obtaining land for agricultural activities is essentially dominated by inheritance (66%), followed by sharecropping (20.2%). The predominance of inheritance and sharecropping shows the weaknesses of land tenure security in the area. Indeed, during the focus groups, some producers indicated that the reason they were unable to invest in solar panels was that the land they were cultivating did not belong to them, but to family members. In their view, it’s risky to invest in expensive installations, especially in situations where families are at loggerheads over inheritance.

For the 401 farmers surveyed, the perceived benefits of solar installations are essentially diesel savings for 59% of cases, and improved yields for 22.4%.

With regard to financial constraints, the initial investment is declared to be the main perceived constraint to acquiring solar pumping systems. The high interest rate was the second most important constraint perceived by farmers, accounting for 18.2% of cases, and insufficient subsidy was the third most important constraint, accounting for 12% of cases.

More than half the farmers (56%) identified the fragmentation of farms as the main constraint to installing solar pumping systems, and the presence of saline groundwater as the second constraint (18%). The small area under cultivation was cited in 14% of cases as a constraint to installing solar pumping systems on their plots.

The absence of qualified technicians in the Niayes zone and a lack of knowledge about solar installations were the two main constraints declared by farmers concerning the installation of solar pumping systems.

Constraints related to the operation and maintenance of solar installations are mainly dust on the panels (29% of cases) and lack of technical know-how (22% of cases). However, 35% of farmers indicate that their installations have no maintenance or servicing problems.

The advantage of the sequential model is that it highlights the causes of abandonment, highlights the key stages for effective adoption, and sheds light on projects and programs to promote renewable energies in order to provide levers to boost adoption.

With regard to the first stage of the process, factors such as age, marital status, experience, access to credit, the fact that the farmer has noticed climate change in his area, the fact that the farmer is from the Thiès region, the fact that he owns his own home and seniority in the locality were found to be factors influencing the likelihood of the grower being sensitized to the use of solar pumping systems in vegetable cultivation.

The three variables of age, marital status and experience have the same influence on awareness. In fact, younger farmers are more likely to be aware of the use of solar pumping systems than older farmers. On the one hand, during the focus groups, some farmers asserted that young people have the chance to benefit from awareness-raising activities on renewable energies during their school years, as climate concerns have seen a resurgence of interest in recent decades. On the other hand, these young people are involved in vacation or youth activities organized each year by the ministry in charge of youth concerning climate protection. However, the older producers often limit themselves to awareness-raising activities via producers’ associations, even though such activities are rare in the Niayes zone.

Access to credit was found to have a negative influence, at the 10% threshold, on the probability to be aware of the use of solar pumping systems. The greater the producer’s access to credit, the lower his awareness of solar pumping systems. This surprising result could be explained by the statements made by some producers during the focus groups. It was stated that “producers who have greater access to credit are those who are affiliated to producers’ associations”. However, affiliated producers are dominated by older producers. During the focus groups in Thiès, an elderly producer stated that “young people are often resistant to associations, which they describe as lacking transparency and internal democracy”. This attitude of young people towards associations could translate into difficulties in accessing credit. Following this logic, access to credit could be linked to age. In turn, producers with access to credit are older and less aware. We can always rely on the fact that young people are more aware because they are present on social networks or because they benefit from awareness-raising activities on renewable energies either during their school life, during vacation activities or during the youth activities organized each year by the ministry in charge of youth.

Farmers with some knowledge of changing climate parameters are less likely to be sensitized. This is reflected in the fact that producers with certain skills in terms of understanding climate change may not be too interested in awareness-raising activities.

With regard to the second stage of the process, i.e. the test phase, the results show that age, plot size, distance of the plot from the main road and distance of the plot from the market significantly explain the first use of solar pumping systems. A large area is often associated with large harvests, which could be a source of income for purchasing solar pumping systems. With regard to the distance of the plot from the main road or from the market, the results show that if this distance is small, the probability of testing the solar pumping system increases. The reduction in distance could be an indicator of the management of post-harvest losses. Fewer post-harvest losses means higher yields, making it possible to purchase solar pumping systems. Surprisingly, access to credit does not affect the first use of solar pumping systems. During the focus groups, a good number of farmers stated that many farmers do not have access to credit, and of those who do, few have sufficient funds to make costly investments such as the acquisition of solar pumping systems. However, many farmers with solar pumping systems say they have acquired them thanks to good harvests or through programs to promote solar energy in production systems. Good harvests often explain the acquisition of certain agricultural equipment, but not access to credit.

Continuous use, which is the last stage in the sequential process of adoption, is explained, according to the results, by gender, experience, household size and access to credit.

Being male explains significantly at the 1% threshold the continuous use of solar pumping systems. This result was found in Senegal by [80] , who took adoption as a multi-stage process. This situation could be explained by the fact that men predominate in market gardening activities and hold the land tenure. Indeed, the inheritance system in Senegal confers great advantages on men in terms of plot acquisition compared to women. Given that solar pumping systems are investments that require the availability and security of land, it is easier for a man to install them than a woman.

Older farmers are often described as risk-averse and reluctant to change. Younger farmers, on the other hand, are described as receptive to new technologies and risk-takers. It’s worth remembering that in this article, experience is dissociated from age. So, being experienced is not strongly linked to age, even if in most cases, age is often taken as a proxy for experience. During the focus groups, one farmer said “farmers who have benefited from small-scale experiments in the use of solar pumping systems are most often those who succeed in their large-scale use in their plots”. This statement strongly supports the finding that, the more experienced the farmer, the more consistently they use solar pumping systems.

Adoption (continuous use) is also significantly influenced by household size at the 5% threshold. Information from focus groups shows that most households, due to difficult access to credit, mobilize resources through solidarity contributions from their members. However, the study by [81] using adoption as a sequential process, finds that the size of the horticultural producer’s household has no effect on the continued use of mango fly control techniques in Senegal.

Access to credit does not affect the test of solar pumping systems, but it is a determining factor in their continued use. This is because many producers test just because an opportunity is offered to them by projects or programs promoting solar pumping systems in the agricultural sector. In reality, however, the decision to test is not fully matured or intrinsic to the farmer. The farmer just takes advantage of the opportunity. This explains why some farmers just test and then give up. On the other hand, those who continue to use the technology one or more years after a test are farmers who have matured their orientation towards solar pumping technology, which they often acquire on credit.