The contribution of trees and palms to a balanced diet in three rural villages of the Fatick Province, Senegal

Abstract

Improving the quality of people’s diets represents a major challenge for developing countries, particularly in Sahelian African countries. Foods from trees, shrubs, and palms are present in many meals and may improve dietary quality, especially for rural communities but, as their contributions have rarely been quantified, investigating the link between the intake of tree foods and the nutritional composition of diet is important. This study assesses the contribution of tree and palm foods to dietary intake in three rural communities in Senegal, using three household food consumption surveys to quantify the dietary intake of meals. The consumption of tree and palm foods in meals was frequent, mentioned by 93–99% of the families depending on the time of year. Products of 11 tree and palm species were used in meals, with Adansonia digitata (baobab) leaves the most frequently mentioned. The energy contribution of tree and palm foods was low, but their contribution of certain micronutrients was high, with 56% of the daily household food consumption of vitamin A, 45% of vitamin C, 17% of vitamin B6 and 21% of iron (Fe). Products from 18 tree and palm species were consumed as snacks or beverages between meals, and consumption was frequent in February and June, but rare in October. We discuss the possibilities of improving nutrition in rural communities from an increased use of trees and palms.

Keywords:

• Agroforestry
• fruits
• vitamins
• minerals
• nutrition

Introduction

Food security in West Africa has been improving considerably in the last 25 years, both in absolute and in relative terms, but still it is estimated that 10% of the population is undernourished (FAO 2015). The region is facing widespread chronic undernutrition and associated nutrient deficiencies that dramatically affect the health and well-being of both adults and children. Improving dietary quality represents a major challenge for the sub-region, and increasing the variety of foods is important to enhance this quality (IARSAF 2007).

Trees have significant potential to complement agricultural production by providing better and more nutritionally balanced diets (Vinceti et al. 2013). Vegetal resources make up most of what people eat, particularly in rural areas, and it is estimated that 50% of all fruits consumed by humans originate from trees (Powell et al. 2013). Food from tree species in sub-Saharan Africa traditionally acts to build resilience into the farming system in terms of food security (Bourou et al. 2012), and has been reported to play a pivotal role in balancing diets with additional ingredients and snacks (Lamien et al. 2009; Thiombiano et al. 2013). Studies from other parts of the world have found that wild foods are important sources of micronutrients but with low energy, although their contributions to diets vary (Samson & Pretty 2006; McMichael et al. 2007; Termote et al. 2012). Thus, trees outside forests, forests, and woodlands worldwide supply rural communities with a wide range of foods, and contribute to food security and nutrition directly and indirectly by providing fruits, seeds, and leaves (Arnold et al. 2011; Thiombiano et al. 2012; de Foresta et al. 2013).

Tree foods are often rich sources of vitamins, minerals, proteins, fats, and other nutrients (FAO 1992; Ho et al. 2012). A recent review of African fruit trees by Stadlmayr et al. (2013) showed their high nutritional value but also highlighted the huge variability and low quality of some of the data reported in the literature. In many parts of West Africa, trees are frequently found in fields, and farmers have easy access to their products. Tree products in certain regions are consumed in the majority of meals (Lykke et al. 2002). Hence, it is often mentioned in the agroforestry literature that food tree products are important for balancing the nutrition of rural people in the Sahel (Lamien et al. 2009; Ræbild 2012), yet how they contribute has seldom been well described. Likewise, increasing the consumption of food products from trees which have high contents of micronutrients, vitamins, and minerals is seen to be a way to improve nutrient quality and diet (Stephenson et al. 2010; Kehlenbeck et al. 2013). However, there are few studies of the actual consumption of food tree products (Vinceti et al. 2008), and rarely have the contributions of food tree products to balanced nutrition been quantified (Powell et al. 2013).

We hypothesized that food products from trees, shrubs, and palms (from here on called tree and palm foods) play an important role in the diets of rural communities and represent an important source of nutrients. Our aim was to quantify the contribution of tree and palm foods to the dietary intake of rural communities.

Materials and methods

Study area

The study was carried out in the three rural villages of Boly Serere (14°04′28″N, 16°18′49″W), Poukham Tock (14°20′57″N, 16°21′04″W), and Samba Dia (14°08′00″N, 16°43′35″W), all located in Fatick Province, Senegal. The three villages were selected based on their proximity (0.8–1.5 km) to forest reserves (forest reserve of Welor, forest reserve of Mahecor, and the biosphere reserve of Samba Dia, respectively, in Boly Serere, Poukham Tock, and Samba Dia sites) and diverse agroforestry systems (including agroforestry parklands and fallows) which could be an important source of food products. According to data from the rural community census, the numbers of inhabitants in 2013 in Boly Serere, Poukham Tock and Samba Dia were 328, 1321, and 3530, respectively. The communities are heavily dependent upon agriculture, animal husbandry, and collection of food products. Other incomes for the households are small businesses and remittances from relatives in cities. The climate is semi-arid with an average annual precipitation of 600–700 mm. The main cultivated species are millet, groundnuts, maize, sorghum, cowpea, rice, and bissap (Hibiscus sabdariffa). Most of the agricultural crops are grown in the rainy season, which typically lasts from July to September, but there is limited agricultural activity in the long dry season from October to June.

Sampling

Data were collected in 2013 using a convenient sample and a two-stage approach. As there are no rural census lists available, the participating households were chosen by randomly selecting four different streets in each village. Along the streets, a systematic sampling was undertaken by asking every second household on both sides of the street or path whether they would participate in the surveys. Twenty-four, 50, and 40 households (114 in total) representing 68, 43, and 13% of the households were selected in Boly Serere, Poukham Tock, and Samba Dia, respectively. A household is here defined as the set of people eating from the same bowls, including guests. Absent members of the household who were not participating in meals were thus excluded. Three household food consumption surveys were carried out to represent the different periods and concomitant food availability: February (post-harvest), June (start of the food shortage period), and October (harvest). February corresponds to the post-harvest period with an abundance of food products, whereas June and October represent different stages of what is known as the hunger season or the food scarcity period when stocks of staple crops may be exhausted. During the survey in October, crops were ripe but still not dry in the fields, and because of a lack of food, people had started using the new crops by artificially drying the cereals. Interviews were carried out in the same households for all surveys, with the exception of 10% which had to be replaced because the person to be interviewed was absent. If an interviewee was absent, the next household in the street was asked to replace it.

Interviews

Women were interviewed first as they were in charge of all meals eaten at home. Four steps were investigated: (1) general non-dietary information related to household size (age and sex); (2) a quick list of foods and beverages consumed by the household members during meals, starting from the most recent to the earliest meal during the last 24 h; (3) records of time and occasions at which foods were consumed, and (4) a detailed list of cooked/uncooked food items for the meals, including amounts, origins (market, cultivated, or collected), and cooking methods. Amounts were estimated using local units, which were then converted into weight, or using the price per unit to estimate the amount of purchased food consumed by households. These prices were obtained by simultaneous surveys of the market closest to each village to record prices of food products and beverages for use in the calculations. For the liquid products, a volume–weight conversion was done using the volumetric mass table of these products. No reports of leftovers occurred.

Finally, the other members of the household who were present (including men and children) were interviewed to get information about the snacks, beverages, and food consumed between meals and outside the home. The quantity of consumed snacks could not be determined, as respondents frequently found it difficult to quantify the amounts they consumed.

Calculation of nutrient compositions

The net weight of food was calculated using a table of waste percentage (Dop et al. 2003), and the nutrient contents (energy, protein, lipids and fat, carbohydrate, vitamin A, vitamin B6, vitamin B12, vitamin C, vitamin D, vitamin E, iron, and zinc) of food consumed were calculated based on how food was consumed (raw or cooked) using the West African food composition table (FAO 2012). Food items were grouped into 12 groups: (1) cereals; (2) roots and tubers; (3) legumes; (4) vegetables; (5) tree and palm foods; (6) meat and fish; (7) oil and fat; (8) bouillon cubes, spices and salt; (9) milk; (10) nuts and seeds (only groundnuts); (11) sugar, and (12) beverages. Tree and palm foods were defined as all food products from woody and palm species (including native as well as exotic species) used as food items, ingredients, or condiments to cook meals. The contribution of each food item and food group to nutrient supply was calculated.

Since people in the households eat from the same bowls, we could not estimate the individual intake. Instead, the average nutrient intake per person was calculated based on adult male equivalents, using reference nutrient intakes (RNIs) for vitamins and minerals (FAO/WHO 2004). For each nutrient, the number of adult male equivalents per household was computed using age, sex, lactation, and pregnancy of each member of the household (Weisell & Dop 2012). The daily nutrient intake of the household was then divided by the total adult male equivalents to obtain the daily nutrient intake per adult male equivalent. These values were then compared to the daily recommended intake per adult male equivalent from the same sources. Expressing intake per adult male equivalent and comparing it to requirements is based on the assumption that food and nutrients are distributed to all household members proportionately to their requirements. The daily recommended intake of iron and zinc was based on low bioavailability and 5% absorption of the ingested minerals according to recommendations for developing countries (FAO/WHO 2004).

Since data for snacks and beverages were qualitative, no quantification of nutrient intake based on these data was possible, and these were not included in the calculation of the dietary intake. The daily household food consumption therefore represents a lower limit for the estimate of the total dietary intake.

Statistical analysis

Since the nutritional composition varies from day to day because different meals are consumed, data accumulated over 24 h cannot be used to estimate the long-term nutritional intake for individual families. Instead, we calculated the average across a village as an estimate of the average consumption for that village in a given month. Considering the three villages as replicates, all data are hence expressed as overall means ± SD.

The variables analyzed included the number of mentions of tree and palm foods in meals, the number of mentions of snacks and beverages, and the various nutrient intakes (total and intakes from tree and palm foods) expressed per adult male equivalent. Using the average values per village and month, repeated measures analyses were used to model the effects of month on the variables. We applied univariate tests for the within-village effects, as the tests for sphericity indicated that Huynh-Feldt conditions were satisfied in most cases. Where conditions were not met, the Huynh-Feldt-Lecoutre epsilon was applied to correct the degrees of freedom (SAS Institute Inc. 2013).

Results

Food consumption patterns and demographic characteristics of households

Three daily main meals (breakfast, lunch, and dinner) were reported. Most households reported the same dishes for breakfast and dinner. Between lunch and dinner, a meal (‘Ndiogonal’) was served for children. A total of 82 food items were reported as used to prepare the main meals (Appendix 1, supplementary material). Respondents took no meals outside the household, and children attending school were having their lunch at home. However, snacks and beverages were consumed between meals and mainly on an individual basis. We registered, but did not quantify the consumption of 50 types of snacks and beverages eaten between meals or outside the home (Appendix 2, supplementary material).

The three surveys covered 93, 99, and 100 households and 1267, 1535 and 1378 people in February, June, and October, respectively. The size of the households varied greatly; the smallest household had three individuals and the largest household included 47 people. The respondents for household food consumption were women with an average age of 36, and the average household sizes were 14, 11, and 17 persons for Boly Serere, Poukham Tock, and Samba Dia, respectively (Table 1). The population was mainly farmers (99%), with a few people having a small business, such as trading or tailoring.

Table 1. Numbers of households and their average numbers of individuals in different age classes in the three villages across the three periods.

Categories	Boly Serere			Poukham Tock			Samba Dia
	February	June	October	February	June	October	February	June	October
Number of households in recall	21	21	22	36	39	41	36	39	37
Children <10 years	4.0	5.1	3.6	3.9	4.7	4.0	5.0	5.6	4.9
Adolescents 10–18 years old	3.4	4.1	3.5	2.6	2.9	2.9	3.4	3.8	3.0
Adults >18 years	5.1	5.6	5.2	4.3	5.1	5.5	8.6	9.4	8.3
Total	12.5	14.8	12.3	10.8	12.7	12.4	17.1	18.7	16.2

Note:

The bold values are used to separate the overall figures (numbers of households and the total number of individuals) from their age composition.

Contributions of food groups to nutrient intake in the households

Cereals are the staple foods in the villages where millet, rice, and maize were consumed pure or mixed with 79 other food items, condiments and ingredients belonging to the 12 food groups (Table 2, Appendix 1). As expected, cereals were the main contributors to energy (54% of intake) and carbohydrates (82% of intake). They were also major contributors to vitamin B6, iron, and zinc nutrition, and vegetables were the main providers of vitamin C (Table 2). The meat and fish group was the major contributor to protein (39% of intake), and vitamins B12 and D. Meat was rarely eaten in the area but, due to the proximity of the Sine Saloum River, fish was an important part of the diet. Oil and fat was the food group with the highest contribution to vitamin E, lipids, and fats.

Table 2. Contribution of each food group to energy and nutrients from meals.

Food group	Vitamins						Minerals
	A	B6	B12	C	D	E	Iron	Zinc
Cereals	3.8(2.7)	45.4(2.5)	0(0)	0(0)	0(0)	6.1(4.5)	57(0.9)	58.8(1.0)
Roots and tubers	0(0)	0.3(0.2)	0(0)	4.7(2.8)	0(0)	0.1(0)	0.2(0.2)	0.2(0.1)
Legumes	0(0.1)	0.6(0.6)	0(0)	0.1(0.1)	0(0)	0.1(0.1)	1.2(1.1)	1.9(1.9)
Vegetables	17.5(8.3)	4.9(1.3)	0(0)	39.8(1.4)	0(0)	2.5(0.5)	4(1.8)	2.6(0.7)
Tree and palm foods	55.9(10.7)	16.9(2.4)	0(0)	45.3(9.9)	0(0)	10.8(4.2)	21.3(4)	9(1.7)
Meat and fish	3.2(1.0)	14.9(3)	88.6(7.6)	0(0)	99.6(0.3)	4.1(0.9)	5.6(1.8)	16.8(3.5)
Oil and fat	2(3)	0(0)	0(0)	0(0)	0(0.1)	47.8(13)	0(0)	0(0)
Cubes, spices and salt	12.3(4.8)	3.9(0.5)	1.8(0.2)	5.7(5.5)	0(0)	6(1.7)	3.5(0.9)	1.1(0.2)
Milk	5.1(3.3)	1.4(1.1)	9.6(7.7)	4.5(3.9)	0.4(0.3)	0.3(0.2)	0.3(0.3)	0.2(0.0)
Nuts and seed	0(0)	11.9(1.8)	0(0)	0(0)	0(0)	22.3(7.2)	6.9(0.9)	9.4(1.5)
Sugar	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0.1)	0(0)
Beverages	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)
Total	100	100	100	100	100	100	100	100

Notes: Values are expressed as a percentage of total intake across the three surveys and represent mean values (SD) based on village means (n = 3). Note that snacks and beverages were not included.

Tree and palm foods contributed, with small parts of the energy intake (5%), proteins (6%), and lipids, and carbohydrates (5%). For vitamins and minerals, however, the contribution from tree and palm foods was larger and, especially for vitamins A and C, contributions were very important with 56 and 45% of the daily intake, respectively (Table 2). The contribution of iron was also important with 21% of the intake.

A comparison of the mean nutrient intake (based on the meals consumed in the households) with the daily recommended intake indicated that the reported meals taken at home were not sufficient to cover requirements with respect to vitamins A and C as well as the minerals iron and zinc (Figure 1). However, requirements with respect to the vitamins B12 and D were met with mean consumptions per adult male equivalent of 2.7 and 17.9 μg, respectively. Requirements were also met for vitamin B6 (Figure 1). To this should be added contributions from snacks and beverages consumed outside the home.

Figure 1. Mean daily nutrient intakes in meals consumed in households across the three surveys (snacks and beverages excluded). Values are expressed per adult male equivalent and compared to mean requirements. Filled symbols indicate total nutrient intakes and open symbols indicate contributions from tree and palm foods. P-values indicate differences between survey periods according to a repeated measures analysis of within-village effects. (a) protein, (b) lipids and fats, (c) vitamin A, (d) vitamin B6, (e) vitamin C, (f) iron, and (g) zinc. Error bars denote standard deviation (n = 3).

The intake of nutrients was significantly different between the three survey periods for all analyzed nutrients except vitamins B12 and D, with the highest values observed in February and the lowest in October (Figure 1).

Tree and palm foods and their contribution to the household diet

Among the food items, snacks and beverages cited, 26 items belong to the tree and palm foods group (Table 3) and include 21 species. Eleven tree and palm food products were used in main meals, and 18 were eaten or drunk raw as snacks or beverages. Some tree and palm food products such as Mangifera indica, Adansonia digitata (fruit), and Tamarindus indica were used both as ingredients in main meals and as snacks or beverages.

Table 3. List of tree and palm foods reported by respondents.

Scientific name	Local name	French name	English name	Uses	Type of culinary uses	Consumption state
Exotic species
Anacardium occidentale	Darkase	Pomme d’acajou	Cashew apple	Fruit	Snacks	Raw
–	Damrass	Noix de cajou	Cashew nut	Nut	Snacks	Raw
Carica papaya	Papaye	Papaye	Papaya	Fruit	Snacks	Raw
Citrus limon	Lemon	Citron	Lemon	Fruit	Meals	Cooked
Citrus sinensis	Orange	Orange	Orange	Fruit	Snacks	Raw
Mangifera indica	Mangue	Mangue	Mango	Fruit	Meals/snacks	Raw/cooked
Moringa oleifera	M’bum	Ben oleifere	Drumstick	Leaves	Meals	Cooked
Psidium guajava	Goyave	Goyave	Guava	Fruit	Snacks	Raw
Native species
Adansonia digitata	Buy	Baobab	Baobab fruit	Fruit	Meals/snacks	Raw/cooked
–	Jus buy	Jus de baobab	Baobab juice	Fruit	Snacks	Raw
–	Laloo buy	Baobab	Baobab leaves	Leaves	Meals	Cooked
Balanites aegyptiaca	Sump	Datier du desert	Desert date	Fruit	Snacks	Raw
Borassus aethiopum	Rone	Ronier	African fan palm	Fruit	Snacks	Raw
Cordyla pinnata	Dimb	Poirier du cayor		Fruit	Meals	Cooked
Detarium senegalense	Ditakh	Ditax	Dattock	Fruit	Snacks	Raw
Diospyros mespiliformis	Alom	Kaki de brousse	Ebony	Fruit	Snacks	Raw
Elaeis guineensis	Tire	Huile de palme	Palm oil, red	Palm oil	Meals	Cooked
Ficus gnaphalocarpa	Gang	Figuier	Mulberry fig	Fruit	Meals	Cooked
Parinari macrophylla	New	Pommier du cayor	Cayor apple	Fruit	Snacks	Raw
Parkia biglobosa	Oule/nete	Nere	African locust bean	Fruit	Snacks	Raw
	Netetu	Nere	African locust bean	Seeds	Meals	Cooked
Saba senegalensis	Mad		Senegal saba	Fruit	Snacks	Raw
Sterculia setigera	Laloo mbeb	Platane du Senegal		Leaves	Meals	Cooked
Tamarindus indica	Daxar	Tamarin	Tamarind	Fruit	Meals/snacks	Raw/cooked
–	Jus Daxar	Jus de tamarin	Tamarind juice	Fruit	Snacks	Raw
Ziziphus mauritiana	Sidem	Jujube	Jujube	Fruit	Snacks	Raw

Tree and palm foods were used frequently in the meals taken in the households, and between 93 and 99% reported consumption of tree and palm products during the 24 h reporting period, depending on the season. Seasons had no significant effect on the average number of tree and palm foods reported per household (p = 0.24). By far, the most frequent tree food consumed at home was leaves. On average, across the three surveys, the use of dry leaves of A. digitata was mentioned by 85% of the families, while 14% reported having used the fresh leaves of Moringa oleifera and 2% had used the leaves of Sterculia setigera (Table 4).

Table 4. Tree and palm foods consumed as part of meals in the three villages (snacks and beverages not included).

Scientific name	No. reports per survey			Mean consumed – g individual^−1(SD)
	February	June	October	February	June	October
No. households in recall	93	99	100
Exotic species
Citrus limon (fresh fruit)	5	4	5	0.7(1.1)	0.8(1.4)	0.6(1.0)
Mangifera indica (fresh fruit)	0	1	0	0	0.1(0.2)	0
Moringa oleifera (fresh leaves)	25	12	3	11.6(7.2)	4.5(2.6)	1.8(2.4)
Native species
Adansonia digitata (fresh fruit)	1	0	0	0.4(0.6)	0	0
Adansonia digitata (dried leaves)	72	86	88	16.9(7.4)	10.9(4.1)	13.3(2.3)
Cordyla pinnata (fresh fruit)	0	1	0	0	0.5(0.8)	0
Elaeis guineensis (palm oil)	3	5	5	2.1(2.4)	1.6(0.8)	0.5(0.6)
Ficus gnaphalocarpa (fresh fruit)	12	6	0	5.9(10.2)	3.9(4.7)	0
Parkia biglobosa (seeds)	17	13	21	1.3(0.7)	0.5(0.2)	0.6(0.2)
Sterculia setigera (dried leaves)	1	2	2	0.1(0.2)	0.2(0.4)	0.3(0.5)
Tamarindus indica (fresh fruit)	27	15	15	2.0(1.9)	0.4(0.3)	0.2(0.2)

Notes: Number of reports per survey indicates the number of families reporting consumption of the product during the 24 h survey. Mean values indicate the intake per individual, based on averages from the three villages, followed by the standard deviation (n = 3).

Fruits were also consumed at home, either fresh or in meals. The most frequently used fruit was T. indica, used by 20% of the households at each survey on average. The second and third most consumed fruits were Ficus gnaphalocarpa and Citrus limon (6 and 5%, respectively), while M. indica, A. digitata (fruit) and Cordyla pinnata were each mentioned by a single family during the three surveys. Fermented seeds of Parkia biglobosa were used by 17% of the families across the three surveys, while palm oil from Elaeis guineensis were used by 4% of the families.

A variation was also found between the survey periods for the intake of nutrients from tree and palm foods (Table 4, Figure 1). The highest values of nutrient supply from tree and palm foods were registered in the post-harvest period (February), and differences between survey periods were significant for protein and iron, while there were borderline significant differences for vitamins A, B6, C, and also for zinc.

The single most important source of vitamin A was palm oil from E. guineensis which, although it was used by few families, contributed about a third of the total intake of vitamin A (Appendix 1, supplementary material). A. digitata and M. oleifera were also important sources of vitamin A. Furthermore, these two species were important contributors to the daily intake of vitamins B6 and C, and also iron. Since the fruits of A. digitata, F. gnaphalocarpa, and C. limon were consumed in only small amounts in the households, their contributions of vitamin C were only modest. The seeds of the legume P. biglobosa contributed less than 1% of the protein.

Tree and palm foods eaten as snacks and beverages

The number of respondents for snacks and beverages was 186, 198, and 200 in February, June, and October, respectively, consisting of 50% women, 30% children, and 20% men, on average.

Eighteen snacks and beverages reported by the respondents belonged to the category of tree and palm foods. Tree and palm snacks and beverages constituted 37% of the total number of mentions of snacks and beverages, and included 6 exotic species and 10 native species (Table 5). The number of reports varied significantly between months (p > 0.04), with respondents reporting on average 1.15 ± 0.23, 0.57 ± 0.38, and 0.10 ± 0.16 types of tree and palms snacks per day for February, June, and October, respectively.

Table 5. Number of reports of tree and palm foods consumed between meals as snacks and beverages across the three surveys.

Species	February	June	October
No. respondents	186	198	200
Snacks
Exotic species
Anacardium occidentale (fruit)	14	10	0
Anacardium occidentale (nut)	0	4	0
Carica papaya (fruit)	3	4	0
Citrus sinensis (fruit)	5	0	0
Mangifera indica (fruit)	0	61	4
Psidium guajava (fruit)	0	3	0
Native species
Adansonia digitata (fruit)	25	3	0
Balanites aegyptiaca (fruit)	3	0	0
Borassus aethiopum (fruit)	2	0	0
Detarium senegalense (fruit)	0	0	14
Diospyros mespiliformis (fruit)	51	0	0
Parinari macrophylla (fruit)	2	0	2
Parkia biglobosa (fruit)	0	4	0
Saba senegalensis (fruit)	0	3	0
Tamarindus indica (fruit)	18	0	0
Ziziphus mauritiana (fruit)	78	0	0
Beverages
Native Fruit Tree Species
Adansonia digitata (fruit)	9	2	0
Tamarindus indica(fruit)	1	0	0

All snacks and beverages from trees and palms were fruits, with the exception of four mentions of cashew nuts in June. In February, native species such as Ziziphus mauritiana, A. digitata (fruit), and Diospyrus mespiliformis were frequently consumed, whereas in June, the exotic M. indica dominated. In October, the number of mentions was low, with only three species mentioned, of which the most important was Detarium senegalense.

Origin of tree and palm foods

A. digitata leaves, the most frequently consumed tree and palm food (58 ± 40% of mentions across the three villages), were mainly collected (58 ± 40%) from nearby trees, while the rest were purchased at the market. The other tree and palm foods were mainly bought at the market (53 ± 29%), while smaller proportions were collected (41 ± 32%) and cultivated (6 ± 5%). Cultivated products included M. oleifera, C. limon and M. indica. Finally, almost half of the tree and palm snacks and beverages were collected (47 ± 38%) and 37 ± 28% were purchased from the market. The 15 ± 13% of the snacks and beverages that were cultivated were mainly M. indica fruits. The proportions were quite variable between villages, as evidenced from the large standard deviations. On the other hand, there were no significant differences between months in the proportion of collected, purchased, or cultivated tree and palm products.

Discussion

This study shows that almost all households in the surveys used tree and palm foods in their daily meals. Tree and palm foods made important contributions to a balanced diet as they provided more than half of the intake of vitamin A, c. 40% of vitamin C, and about 20% of vitamin B6 and iron intake from meals. Furthermore, tree and palm foods constituted a large proportion of snacks and beverages, which were not included in the total calculated nutrient intake. As they have high levels of vitamins A and C, consumption of fruits such as Z. mauritiana, A. digitata, Diospyros mespiliformis, and M. indica are likely to provide an important additional contribution to these nutrients.

Previous studies in the Sahelian countries have shown that edible plants are sources of essential fatty acids, vitamins, iron, zinc, and calcium (Becker 1983; Glew et al. 1997). The current study represents one of the first attempts to quantify the nutritional intake from trees and palms in the Sahel but is supported by findings from other parts of Africa. For example, in rural Tanzania and Congo, wild foods (principally from trees and palms) were used by virtually all informants and represented important sources of vitamins and minerals (Termote et al. 2012; Powell et al. 2013).

Our study recorded 82 food items, a level which corresponds to findings from Dakar, where 90 foods and beverages were identified (Anderson et al. 2010). In total, we recorded 21 tree and palm species used as foods and snacks in the area. Similar results were reported from the Democratic Republic of Congo with 11–15 wild edible species recorded and from south central parts of Burkina Faso where 17–21 tree species were used for food and snacks (Lamien et al. 2009; Termote et al. 2012).

The major limitation in our study is that snacks consumed in between meals could not be quantified, especially for smaller fruits, such as Z. mauritiana which were not counted during consumption. As stated above, this leads to an underestimation of the contribution of trees and palms to nutrition, and in future studies, it would be interesting to make a closer observational study about this. Snacks may be an important source of total daily nutrient intakes (e.g. Stockman et al. 2005; Anderson et al. 2010), but their contribution to nutrition remains to be verified for Sahelian countries. Another limitation is that the accuracy of recall of food may vary (Räsänen 1979). Previous studies have reported underestimations of approximately 20% by people participating in studies of diets (Martin et al. 1996; Sawaya et al. 1996). However, it is unlikely that underestimation and unrecorded consumption of snacks are the causes of the very low intake of nutrients such as iron and zinc.

The comparison between average daily nutrient intakes and recommended intakes indicate deficits for a number of minerals and vitamins (Figure 1). Leafy vegetables from trees play an important role for iron, vitamin A, and vitamin C intake, and A. digitata and M. oleifera stand out as the most important sources. Besides these, the oil of E. guineensis (red palm oil) has an extremely high content of vitamin A (Nagendran et al. 2000; Rao 2000; FAO 2012) and, on average, had a very large impact on vitamin A intake but was consumed by few families (4% across the three assessments). The pivotal role of A. digitata was underlined by it being the most reported tree and palm food. This apparently applies to most of the region as in Benin, Burkina Faso, Mali, and Senegal, 70–90% of respondents mentioned A. digitata leaves and fruits as a daily source of food (Buchmann et al. 2010; Thiombiano et al. 2013). A prominent feature of the baobab is that the leaves can be stored dry, making it an important contributor to balanced nutrition, irrespective of time of year. However, despite being used less frequently, the exotic M. oleifera has an even larger contribution to vitamins A and C nutrition (cf. Manguro & Lemmen 2007; FAO 2012) and an increased use of this species should be explored to reduce the frequent deficiency in vitamin A found in rural areas. It should be noted that because vitamin A is fat-soluble, the bioaccessibility of vitamin A from leaves may be limited by being served with low-fat dishes as was often the case in the households.

The question is whether increased use of tree and palm foods may alleviate lack of macro and micronutrients. Our study villages were deliberately selected because they had a range of different land uses within their borders, but tree densities were relatively low (Sambou et al. 2016). By increasing the numbers of trees and palms providing fruits and edible leaves in fields and fallows, households would get easier access to a range of nutritious foods and snacks. However, an important assumption for the increased use of trees in nutrition is that people appreciate and adopt an increased use of tree products in their meals and snacks, which remains to be explored. Food trees in the Sahelian agroforestry parklands in most cases tend to reduce the yields of crops grown directly underneath them (Bayala et al. 2014), but this may be offset by the increased availability of products with high vitamin and mineral content. In the studied villages, products were rarely from cultivated trees, showing that there is a large potential for an increased use of tree and palm foods if they are domesticated. Such domestication may even include the development of varieties with a higher content of vitamins (Ræbild et al. 2011; Korbo et al. 2013).

Besides this, most tree species appeared to be consumed in strongly seasonal patterns, related to their phenology, as only few species can be conserved and consumed around the year. With the exception of mango consumed in June, snacks from trees and palms were practically only consumed during the dry season (February). The cultivation of trees and palms selected on the basis of the seasonal calendar of fruit harvest times may help providing key nutrients around the year (Jamnadass et al. 2011).

In conclusion, tree and palm foods are consumed frequently in the study area and represent an important source of micronutrients that contribute toward a balanced nutrition of rural communities. In areas where malnutrition is highly prevalent, these sources of micronutrients might be essential as even small amounts of micronutrients may have significant effects on health and development. We believe that an improved use of trees and palms will contribute to alleviating malnutrition in this part of the Sahel, even though this remains to be confirmed by studies of food preferences. This is an important addition to the many other benefits provided by trees and palms, including ecosystem services, medicine, fodder, wood, and sources of income.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by Islamic Development Bank.