Study area

This study was conducted around Nikki, Sakabansi, and Fombawi APDs in the Municipality of Nikki (9°56′2″N and 3°12′16″E), located in the Department of Borgou in northeastern Benin. The Municipality of Nikki holds 20 APDs. The three dams were selected based on previous studies [14, 17] that revealed intensive dry season vegetable production activities around the dams. Furthermore, the three APDs vary in terms of the type of dyke, the capacity and importance of the watershed, the market access and urbanization level, and the number of vegetable producers involved. Nikki is the district capital of the Municipality of Nikki and is considered as a more urbanized area compared to Sakabansi and Fombawi villages, which are rural settings. The features of the APDs [17] are presented as follows: (1) located within the boundary of Nikki town, the Nikki agro-pastoral dam was constructed in 1972 and renovated in 1996. It has a capacity of 257,000 m³ and a watershed area of 120 km² [14]. The Nikki APD is 1 km from the town. The vegetable plots are located upstream of the dam. There are two vegetable producer associations in Nikki: Ansouroukoua, created in 1990 with 20 members and now consisting of 150 members (130 women and 20 men), and Donmarou, created in 2007 and consisting of 50 members (40 women and 10 men) [19]. The members have access to markets in Nikki town and in Chikandou (Nigeria, 22 km away). (2) Constructed in 1985 and covering 1 ha, the Sakabansi agro-pastoral dam has a capacity of 200,000 m³ surrounded by a watershed of 20 km² [14]. The dam is 3 km from the village of Sakabansi, which has a local market. The Sakabansi vegetable producer association, named Ankua mon, was created in 1998 and now has 30 members (29 women and 1 man) [19]. The association was restructured in 2004 to help members to obtain support for seeds and training in farming practices. The vegetable production plots are located upstream of the Sakabansi dam. (3) Constructed in 1989, the Fombawi agro-pastoral dam has a capacity of 17,000 m³ and a watershed of 2.4 km² [14]. It is 500 m from the village of Fombawi, which has a local market. Thirty-three women, working in the association Angara debu, farmed around the dam, with plots located downstream. Because of the perennial destruction of the plots by local livestock, vegetable farmers in Fombawi did not cultivate during the data collection phase of this study.

Data collection

The research was carried out in two phases: an exploratory survey and an in-depth survey. The exploratory survey aimed to obtain an overview of the organization of vegetable producers operating around each dam, identify stakeholders and key informants and investigate vegetable producers’ perceptions of the main source of dams’ siltation and factors that affect water quality. Focus group discussions with representatives of the vegetable producer associations around each dam and participant observation were used to collect information in this phase. During the focus group discussions, participants were asked to list and rank the main source of dam’s siltation in each village. In the in-depth phase, semi-structured individual interviews were conducted with 70 respondents using a questionnaire. Participants were selected using a purposive sampling [20] of 37 (18 males and 19 females) out of 200 producers in Nikki, 15 out of 32 females in Fombawi (men are not involved in vegetable production there), and 18 (17 females and the only 1 male) out of 30 producers in Sakabansi. To be included in the sample, the respondent has to be a vegetable producer around the APD, belong to a vegetable producer association and be willing to be involved in the study. Key informants and association leaders were also interviewed to obtain follow-up information and for triangulation. The farm plot size of 55 producers was measured in Nikki and Sakabansi. Fertilizers and pesticides used by vegetable producers were recorded in the three villages. The main subjects of the questionnaires included socio-economic characteristics of respondents, cropping systems with emphasis on diversity of vegetables cultivated, soil fertility management and irrigation practices, pest and disease management practices, constraints that hindered their production, organization of the producers around the dams and respondents’ perception of how vegetable production could affect APD quality. The respondents were asked to prioritize by scoring and ranking their constraints on a 10-point scale (from very high importance = 1 to little importance = 10). The perceived sources of silting of each APD were also prioritized by producers using the ranking method. In this regard, each producer was asked to list the perceived main sources of the dam’s siltation, rank them and indicate whether his/her activity could affect the dam’s water quality and siltation and other APD ecosystem services (e.g. habitat suitability for crocodiles and fish).

Data analysis

We used descriptive statistics to characterize vegetable species diversity and farming practices patterns around the three APDs. To compare the total land area between female and male vegetable farmers and among surveyed villages, we performed a two-sample Wilcoxon test after checking for normality and homoscedasticity assumptions (which were not met) [21]. To test whether farming practices used by producers were independent of gender and villages where they were operating, we used Fischer’s exact test instead of a Chi-square test since some cells of the contingency table had frequencies less than five, and these cells represented more than 20% of the total number of cells in the contingency table [21, 22]. The same approach was used to test whether the vegetable species cultivated by a producer was independent of the producer’s gender and the village in which he/she was operating. We used a generalized linear model with the Poisson error structure and a log link function [22] to test whether the number of vegetables cultivated varied between women and men producers and among villages. To understand how difference in access to land affects the diversity of vegetables cultivated, we ran Pearson correlation test between the number of vegetables cultivated and the total land size. We computed the Kendall’s coefficient of concordance (W) [23] to test whether there was agreement among vegetable producers regarding the ranking of the constraints hindering their activity around APDs. We used simple logistic regression to test whether the probability that a vegetable producer perceived that his/her activity (vegetable production) affected APD water quality and crocodiles and contributed to dam siltation (dependent variable) was affected by respondent’s age, gender, experience in vegetable production around the dam, position related to the dams (upstream vs. downstream), vegetables species richness (number of vegetables cultivated), and village in which he/she operates (each as independent variable). For the logistic regression, we used a generalized linear model with a binomial error structure and a logit link function [22]. We used multinomial logistic regression [24] to test whether the probability that a given source of dam siltation (perceived by vegetable producers) is ranked as the first source (dependent variable) is affected by participant’s age, gender, experience in vegetable production around the dam, position related to the dams, vegetables species richness, and village where he/she operates (independent variables). We assigned water erosion and runoff (one of the dependent variable categories) as a baseline against which all other categories were compared [24]. All statistical analyses were performed using R software version 3.0.2 [25].

Land access and diversity of vegetables cultivated around agro-pastoral dams

Vegetable production around the APDs was a part-time activity dominated by women (73% of respondents). More than 90% of the respondents indicated that they were part-time vegetable farmers. Only one man was involved in vegetable production around Sakabansi’s APD, while no men were involved in Fombawi. It was around Nikki’s APD that men were actively involved in this activity, representing about 49% of the respondents in that area. The respondents’ ages varied from 16 to 78 years, with an average age of 44 ± 13.51 years. They have been involved in vegetable production around APDs since on average 8 ± 5.25 years (ranging from 1 to 25 years). There was a large variation in cultivated land size, which ranged from 92 to 2500 m². The overall average cultivated land size was small (412.6 ± 429.85 m²). However, the land size was significantly affected by gender (Wilcoxon test W = 142.5, p < 0.001, Fig. 1a) and varied significantly between APDs (Wilcoxon test W = 534.5, p < 0.001, Fig. 1b). Men had larger plot sizes (669.2 ± 590.1 m²) compared to women (277.2 ± 226.7 m²). The average land size around the Nikki dam (523.5 ± 484.4 m²) was higher than in Sakabansi (184.56 ± 90.1 m²). To better capture how the large variation in land size was reflected through gender difference, we categorized producers into three groups based on their land sizes (Fig. 2): large vegetable producers (more than 1000 m² of land), medium vegetable producers (plot size between 500 and 1000 m²), and small vegetable producers (plot size less than 500 m²). Again, there was a significant gender-dependence in land size for vegetable production around APDs (Fisher’s exact test; p = 0.0014, Fig. 2). Up to 77% of producers who had less than 500 m² were women, while farmers with more than 1000 m² were exclusively men. Fifty-five per cent of medium size farmers were also men. Additional file 1: Table S1 shows a summary of the descriptive statistics of the socio-demographic and biophysical characteristics of the respondents.

Overall, 16 vegetables species were cultivated around the three APDs, from which about 63% were African indigenous vegetables (Table 1). Eleven vegetables were among the most cultivated (see Fig. 3) with amaranth (Amaranthus cruentus L.), black benniseed (Sesamum radiatum Schumach. & Thonn.) and okra (Abelmoschus esculentus (L.) Moench) as the top three vegetables cultivated around the APDs. They were cultivated respectively by 91, 70 and 56% of the respondents. The number of vegetables cultivated per producer varied from one to eight, with an average of 4.37 ± 1.84 vegetables per producer. In Nikki and Sakabansi, there was a significant correlation between land size and the number of cultivated vegetables (r = 0.48, p = 0.0001), indicating that producers with large plot sizes seemed to cultivate the highest number of vegetables. There was no significant difference between men and women (Z = − 1.291, p = 0.197) and among villages (Z = 0.341, p = 0.733) for the number of cultivated vegetables. However, the type of vegetables cultivated was dependent on gender (Fisher’s exact test; p = 0.0077) and village (Fisher’s exact test; p < 0.0001). The most cultivated vegetables by women producers were amaranth, black benniseed, okra, jute mallow (Corchorus olitorius L.) and Roselle (Hibiscus sabdariffa L.). False sesame (Ceratotheca sesamoides Endl.) was exclusively cultivated by women. The most cultivated vegetables by men were pepper (Capsicum frutescens L.), amaranth and okra. In Fombawi and Sakabansi, the most cultivated vegetables were similar and included amaranth and black benniseed, while in Nikki, they included amaranth and pepper. Furthermore, false sesame was exclusively cultivated in Sakabansi, while African eggplant (Solanum macrocarpon L.) was cultivated only in Nikki. Additionally, cabbage (Brassica oleracea L.) was mainly cultivated by producers around Nikki APD. Tomato (Lycopersicon esculentum Mill.) and Roselle were mainly cultivated in Nikki and Sakabansi.

Common names	Scientific names + authorship	Family	Type/origins
Amaranth	Amaranthus cruentus L.	Amaranthaceae	Indigenous/traditional
Black benniseed	Sesamum radiatum Schumach. & Thonn.	Pedaliaceae	Indigenous/traditional
Okra	Abelmoschus esculentus (L.) Moench	Malvaceae	Indigenous/traditional
Pepper	Capsicum frutenscens L.	Solanaceae	Indigenous/traditional
Jute mallow	Corchorus olitorius L.	Tiliaceae	Indigenous/traditional
Lettuce	Lactuca sativa L.	Asteraceae	Exotic
Tomato	Lycopersicon esculentum Mill.	Solanaceae	Exotic
Roselle	Hibiscus sabdariffa L.	Malvaceae	Indigenous/traditional
African eggplant	Solanum macrocarpon L.	Solanaceae	Indigenous/traditional
Cabbage	Brassica oleracea L.	Brassicaceae	Exotic
False sesame	Ceratotheca sesamoides Endl.	Pedaliaceae	Indigenous/traditional
Carrot	Daucus carota L.	Apiaceae	Exotic
Lagos spinach	Celosia argentea L.	Amaranthaceae	Indigenous/traditional
Green beans	Phaseolus vulgaris L.	Leguminosae	Exotic
Egusi	Cucumeropsis mannii Naud.	Cucurbitaceae	Indigenous/traditional
Onion	Allium cepa L.	Alliaceae	Exotic

Cropping systems and farming practices around agro-pastoral dams

The cropping system patterns were gender- and village-dependent (Fisher’s exact test; p = 0.0454 and Fisher’s exact test; p = 0.004 respectively). Monocropping and intercropping were practised mainly by women, while the majority of men practised crop rotation. In Nikki, more producers were involved in crop rotation than in Fombawi and Sakabansi. However, more producers were involved in monocropping and intercropping in Fombawi and Sakabansi than in Nikki. Similarly, soil fertility management practices were gender- and village-dependent (Fisher’s exact test; p < 0.0001 and Fisher’s exact test; p < 0.0001, respectively). Men used mainly chemical fertilizers for soil fertility management, while women relied mainly on organic fertilizers. In cases where women used chemical fertilizers, it was usually in combination with an organic one. Producers in Fombawi and Sakabansi used mostly organic fertilizers for managing the fertility of their soil, while producers in Nikki relied on both organic and chemical fertilizers. Organic fertilizers used around the APDs included cattle and small ruminant manure, household waste, and occasionally poultry manure. Chemical fertilizers included urea and NPK. Pest and disease management practices were also gender-dependent (Fisher’s exact test; p < 0.0001) and village-dependent (Fisher’s exact test; p < 0.0001). Women applied neem (Azadirachta indica A. Juss) leaves and seed extracts, wood ashes, and prophylactic practices, while men relied more on chemical insecticides for pest and disease management. Chemical insecticides and neem extracts were mainly used by producers in Nikki, while producers in Fombawi and Sakabansi applied wood ashes and prophylactic practices. Regarding irrigation, 75% of our respondents used buckets associated with cans with holes and 25% of them used watering cans. In Fombawi and Sakabansi, producers used water directly from the dams to water their crops, while in Nikki, all the producers used water from small wells and boreholes they had created in the dams’ watercourse.

Constraints faced by vegetable producers around agro-pastoral dams

Vegetable producer organizations

Vegetable producers around the ADPs were organized in associations. More than 91% of the respondents were members of a vegetable farmers’ group (Additional file 1: Table S1). There were two associations working around Nikki’s APD, while in Fombawi and Sakabansi, one association worked around the dam in each village. Although almost all producers belonged to an association, each producer worked individually on his plot. The motivation underpinning the creation of the associations was the facilitation in access to credit and technical assistance from local banks and NGOs, extension services and development agencies. Each association was led by an executive committee whose president was usually the oldest member; the secretary and the treasurer were selected from among the members with basic educational backgrounds. To become a member of an association, the applicant should contact the executive committee who would inform the other members of the association. Based on internal discussions regarding the applicant’s characters and his ability to respect the group’s rules, the executive committee would later give approval or not. Once the applicant was approved, he/she would pay the membership fee, which is 500 F CFA for each of the two associations working around Nikki’s dam and 1000 F CFA for each association in Fombawi and Sakabansi, respectively. In Nikki, members paid an annual membership fee (which varied from 1200 to 3200 F CFA) to both associations. The money collected was used for soya bean trading. Soya beans were bought and stored when the price was low and sold when it became high. The profit generated was used to assist members facing financial challenges and special events (such as marriage, baptism, and death). In Sakabansi and Fombawi, members did not pay an annual membership fee.

Vegetable producers’ perception and knowledge of dam siltation and water quality

According to the vegetable producers, several factors contributed to dam siltation and also affected the quality of dam water and the living organisms in the dam, including crocodiles (Fig. 4). Across the three areas, these factors included water erosion and runoff, livestock, vegetable production and farming around the dams. Water erosion and runoff was cited by 81% of the respondents and approximately 50% of the respondents ranked it as the first cause of threat to the dams. They explained that due to water erosion and runoff, a considerable amount of sediment was transported into the waterhole during the rainy season, contributing to the siltation. Livestock was considered as the second most important threat to the dam siltation and quality (36% of responses) and was ranked as the first cause by approximately 23% of the respondents. The respondents explained that livestock accessed the waterholes from all sides, and this contributed to the siltation. Additionally, when drinking water, cattle defecate and urinate directly into the dams, and this can affect the water quality and subsequently the crocodiles living in the dams. Vegetable production and farming around the dams were also identified as sources of siltation and threats to the dam’s water quality (31 and 16% respectively) and were ranked as the first cause by 23% and 4% of the respondents, respectively. In fact, more than 75% of the vegetable plots were located upstream of the dams (Additional file 1: Table S1). With frequent tillage, soils become more vulnerable to erosion and easily transported into the waterholes by runoff water. Furthermore, inputs used (pesticides and fertilizers) can affect the dam’s water quality and subsequently the living organisms including the crocodiles. However, the likelihood that a producer perceived that his/her activity affected the dam’s water quality and crocodiles and contributed to dam siltation was not affected by the participant’s age (Z = − 0.176, p > 0.05), gender (Z = 1.001, p > 0.05), experience in vegetable production around the dam (Z = 0.630, p > 0.05), position related to the dams (Z = 0.087, p > 0.05), vegetables species richness (Z = − 0.015, p > 0.05), or the village in which he/she worked (Z = − 0.147, p > 0.05, Additional file 1: Table S3).

Gender, vegetable farming systems and diversity of vegetable species around agro-pastoral dams

Vegetable production around the APDs is less intensive and characterized by low use of external inputs and rudimentary or small-scale tools and equipment. Vegetable species are dominated by African indigenous species, and the use of improved seeds is low. It is a part-time activity for the majority of people involved and represents a typical example of traditional vegetable farming systems common in West Africa [26]. Indeed, in traditional vegetable farming systems, vegetable production and marketing are undertaken mainly by women [2, 26, 27]. This is the case in Fombawi and Sakabansi, the rural areas, where this activity is undertaken exclusively by women. However, men become more involved as soon as income generation potential increases and the activity takes on a commercial orientation [27, 28]. This can explain why more men were involved in vegetable production and mainly cultivated species with high commercial or economic values in Nikki, the more urban area, with improved roads and market access.

Our results revealed that there was a strong gender difference in cropping systems and farming practices around the APDs investigated, except irrigation practices. These findings are consistent with previous studies on gender dimensions in agriculture in sub-Saharan Africa [29]. The gendered difference observed in the cropping systems may be due to differences in access to land. In fact, men had larger plot sizes compared to women, as it was also reported in vegetable production systems in Buea, Cameroon [26] and in southwestern Nigeria [30]. This supports the general gender-differentiated resource base of African farming systems [29, 31]. Thus, the fact that monocropping and intercropping were predominantly practised by women could be seen as a strategy to cope with the challenge of limited plot sizes to maximize the utilization of the small plot of land they have.

Like Reyes-García et al. [32], we found that men and women producers used different soil fertility, and pest and disease management practices. Organic fertilizers and traditional pest management practices were predominantly used by women. This may be due to difference in knowledge of farming practices and in access to information and the ability to act upon it [26, 31]. In fact, female farmers are less likely to access agricultural information and knowledge through extension services, or receive very limited access to quality services and agricultural technology information compared to their male counterparts [30, 33]. Differences in access to inputs may also play a role in the gendered nature of vegetable production around APDs [26, 29]. Indeed, chemical inputs used around the APDs are mainly those intended for cotton cultivation. In the study areas, only cotton farmers, who are exclusively men, can have access to those inputs. Furthermore, difference in access to labour can explain the gender difference in vegetable production systems around APDs. Indeed, because of the gender-differentiated nature of rural household labour shortages or the gendered nature of labour-hiring practices in African farming systems [31], women usually have less access to labour for labour-intensive farming activities.

Vegetable species cultivated around APDs were mainly leafy vegetables. Our findings also indicated that species cultivated were predominantly African indigenous vegetable species (63% of vegetables cultivated). This highlights the potential role of vegetable production around APDs towards improving food and nutritional security in the study area [3–5, 8, 34]. Indeed, proteins, carbohydrates, dietary fibres, potassium, calcium, magnesium, phosphorus, Vitamin A, Vitamin C and Vitamin E have been shown to be significantly higher in indigenous vegetables than their exotic counterparts [35]. Our findings that women cultivated mainly traditional vegetables in comparison with men corroborates previous results by Ndenga et al. [36] that indicated women vegetable producers in the Rift Valley and Central provinces in Kenya allocated more land for the production of traditional vegetables than men.

Challenges to vegetable production around agro-pastoral dams

Some of the key challenges faced by vegetable producers around APDs, namely, lack of equipment, access to inputs and credit or capital, and access to and fluctuation of market opportunities are not specific to working around agro-pastoral dams. These constraints have been reported previously in vegetable production systems in Benin, Cameroon, Nigeria, Senegal and South Africa [8, 9, 26, 30, 37]. Although we did not assess how gender affects the perception of the constraints (e.g. Ngome and Foeken [26] and Alao et al. [30]), we found a strong and significant agreement on the ranking of the most important constraints among the respondents, indicating that gender did not considerably affect the perception of the constraints in our study. In addition, contrary to previous investigations in southern Benin and elsewhere in West Africa that reported access to land as a major constraint for vegetable production, especially in urban areas [9, 11, 12, 37, 38], we found that land access is not a major challenge for vegetable producers around APDs, even in the more urbanized study area. Indeed, challenges related to land access have been ranked as second to last (out of ten) around almost all the three APDs (see Table 2). This is inconsistent with results by Ogunjimi and Adekalu [37] and Cissé et al. [39], which reported strong competition over lands around the major perennial streams in Nigeria and around APDs in Ouagadougou, Burkina Faso, respectively.

However, conflicts with livestock keepers (due to recurrent destruction of vegetable plots and seedlings by livestock) were challenges specific to vegetable production around APDs, especially in the rural study areas. It is very likely to be the case around other APDs in northern Benin where vegetable production occurs. Indeed, the APDs were constructed initially to provide additional drinking water for livestock production after the severe drought experienced in Benin in the 1970s, which had seriously affected the national livestock population [14]. Consequently, up to now, some livestock herders believe that the APDs are for them and they have more use-rights than other stakeholder groups, though there is a dam management committee (CoGes) for each APD [16, 17].

Another key challenge for producers around the APDs is access to water, especially at the peak of the dry season. This constraint has also been reported by small-scale irrigation farmers along major rivers in Nigeria [37]. Unfortunately, since the majority of the producers are located upstream of the dams, the strategies they have developed to cope with this challenge, which include digging wells and boreholes in the dams’ watercourse, potentially contribute to dam siltation. In addition, with frequent tillage, soils become more vulnerable to erosion and easily transported into the waterholes by runoff [40]. Some of the producers are perfectly aware of this since they have identified farming and vegetable production around the dams as causes of dam siltation in the study areas. However, very few respondents ranked vegetable production as the first cause of dam siltation, while more experienced producers around APDs and those with high vegetable species richness were more likely to rank farming as the first cause of dam siltation.

Although pest and disease management was reported to be a very key challenge for vegetable farmers in southern Benin and in other regions of Africa [41, 42], it was not perceived as a major constraint around APDs. It was only in urban Nikki that pest and disease pressures emerged as a serious constraint. This can be explained by the fact that urbanization process modifies farmer practices in response to increased food demand, changes dietary patterns towards fruits and vegetable consumption, and increases competitiveness of fruits and vegetable production in the urban and peri-urban areas over staple crops [41]. Pest and disease pressures combined with the challenge of limited access to pesticides have led men producers in Nikki to have recourse to non-recommended pesticides (namely, those intended for cotton cultivation). This practice is not specific to vegetable production around APDs and has actually been reported in the fruit and vegetable production systems of small farmers in sub-Saharan Africa [41]. This raises the issues of the quality of vegetables produced and the health risks not only for consumers and producers but also living organisms in the dams such as crocodiles and fishes [38, 41, 42]. Moreover, utilization of APD water can also represent significant health risks due to bacteriological and microbial contaminations [39, 43]. As explained by respondents, livestock defecate and urinate directly into the dams when drinking water, contributing to water contamination. In addition, given the inappropriate waste disposal and poor waste water management system, especially in the town of Nikki [17], it is very likely that runoffs, streams and effluents that feed the dams carry pathogens contaminating dam water [43].

Implications and research avenues for sustainable vegetable production and integrated afro-pastoral dam management

Our results revealed a strong gender gap in vegetable production systems around agro-pastoral dams in northern Benin. A tailored and gender-specific training programme on sustainable production practices targeted to women will certainly contribute to filling this gap and improving vegetable production around APDs. Indeed, women’s training on improved vegetable production practices can dramatically increase production of leafy vegetables, profitability and income, thus improving women’s livelihoods [44, 45]. This capacity building programme should be supported by increased agricultural extension service support to vegetable farmers around APDs, which could improve technical efficiency of producers [46]. Environmentally friendly solutions and innovations developed by women producers to manage pests and diseases (e.g. botanical extracts) need to be strengthened to reduce health risks for consumers and ensure healthy ecosystems.

Our results also indicated that vegetable producers around APDs were organized into farmer groups or associations, and almost all the respondents belonged to one. However, facilitation to access external support was the main motivation behind group membership, with little concerns about cooperative activities such as mutual support for labour, group marketing strategies, and regular meetings. Promotion of vegetable production around APDs by addressing production bottlenecks and training on sustainable cropping practices could certainly be a driving force in the resurgence of vegetable farmer groups.

Currently, the management of APDs is subject to contestation, and the use of the dams results in recurrent conflicts not only among human stakeholders (e.g. between herders and vegetable producers) but also between humans and wildlife (e.g. human-crocodile conflicts) [17]. Indeed, prior to the 2005 decentralization reforms in Benin, APDs and their management were under the control of the Centre Communal de Promotion Agricole (the Communal Centre for Agriculture Promotion under the Ministry of Agriculture, Livestock production and Fisheries) and their maintenance by the Direction Générale de l’Eau (the General Directorate of Water at the Ministry of Environment, Housing and Urban Planning). Following the decentralization reforms, the control and management of water resources have been transferred to the local government (i.e. the Municipal Assembly or Council) [14, 16, 17, 47]. As a result, the 20 APDs in the Municipality of Nikki are now under the control of the Municipal Council of Nikki [17]. For each APD, a CoGes is put into place by the Municipal Council to ensure the monitoring and maintenance of the APD and to manage and control the access to and use of the APDs’ water and surrounding land so as to prevent siltation and water pollution [14, 17]. However, the ability of CoGes to effectively and sustainably manage the APDs is debatable [14, 16, 17]. Many of the CoGes were inactive or have collapsed [14, 16]. The most recent inventory and assessment of the existing 250 geo-referenced APDs in Benin showed that many of the APDs were not regularly maintained and 44% of them needed urgent rehabilitation of their dyke [47]. Five of the 250 APDs are abandoned and no longer in use [47]. Furthermore, according to many stakeholder groups, in addition to the recurrent conflict between farmers and herders, the siltation and shrinking storage over time have become key impediments to the use of the APDs [17].

In this complex and conflicting context, developing scenarios for the future of this common resource may be helpful. Scenarios are a useful tool for grappling with the uncertainty and complexity of social-ecological challenges [48]. Rather than predicting what will happen in the future, scenarios are ‘what if?’ stories, which can assist in decision-making and strategic planning [49]. Scenarios can help identify a desired future state for APDs to which to work towards or help prepare for a range of possible futures [49]. We suggest the implementation of the Transformative Scenario Planning approach [48] to stimulate social learning by enabling all stakeholders and users of APDs, including the municipal council, to engage and discuss options for coping with uncertainty through collaborative actions. The process will help discuss and find a common solution for the recurrent destruction of the vegetable plots, for instance by defining corridors, accepted by all parties, for livestock. The conditions and prerequisites (e.g. availability of irrigation facilities, improved accessibility to the sites through creation of roads) for moving vegetable producers to the downstream sites will also be discussed for sustainable use of the common resource. Additionally, it would be interesting to conduct the transformative scenario planning process per gender to allow a comparative analysis of women and men perspectives for the future of APDs.

In this study, we assessed perception and knowledge of only vegetable producers. It would be interesting to investigate the perception of other users (fishermen, herders, domestic users, crop farmers, etc.) on APD siltation and water quality. This will provide holistic insights on the different stakeholders’ knowledge and perception of the common resource, which will be crucial in the scenario planning process. Further investigation is also needed to empirically quantify the contribution of vegetable production to APD siltation, eutrophication and proliferation of aquatic invasive species, which are becoming a challenge to APDs [17]. This assessment will provide information to inform decision-making regarding whether to move farming activities to downstream sites.

A water quality assessment will also help ascertain dam water quality for the various uses (e.g. agricultural and domestic uses, suitability for crocodiles) and inform policy decision-making at the district level. Contamination level assessments (bacteriological, microbial, and heavy metals) of vegetables produced and fishes from the dams are also needed to evaluate risks for consumers.