Can Wild Forest Foods Contribute to Food Security and Dietary Diversity of Rural Populations Adjoining Forest Concessions? Insights from Gabon, DR Congo and Cameroon

HIGHLIGHTS A high diversity of forests foods for consumption exists in DR Congo, Cameroon and Gabon. Despite a diverse pool of wild foods, more than 80% of households engaged in logging suffer from food insecurity. A significant inverse correlation exists between food insecurity and forest foods consumption meaning that forest foods play a role in ensuring food security and nutrition among populations near forest concessions. Age and access to income, illegal logging and agriculture on logged land are significant determinants of food security among forest communities. Investing in strategies to provide income outside illegal logging and practicing agriculture on logged land, can contribute to addressing health, nutrition and food security. SUMMARY The potential of forest foods to address malnutrition, food insecurity and poor dietary intake is increasingly being recognized. However, most existing data presents average results of proximate analysis, overlooking the opportunities to document how forest foods contribute to nutrition, food security and dietary intake. In this study, food security was estimated using the HFIAS score while dietary intake was estimated using FFQ recalls among 720 households in and around six logging concessions in Cameroon, DR Congo and Gabon. There was a high diversity of forests foods consumed, with DR Congo (FFCS=16.2) registering the highest number. Cameroon had the highest number of food groups (HDDS=12.5) while Gabon had the highest number of food varieties (FVS=24.5). Despite a highly diverse pool of wild foods in the region, most households (82%) practicing farming on logged land suffered from food insecurity. A significant inverse correlation was observed between the HFIAS score and the FFCS (r2=–0.26, P=0.0002), revealing that forest foods can play a role in ensuring food security and nutrition among populations near forest concessions. The chi-square test revealed that age >46 years, and access to income from business or salaries were associated with food security. Further work is required to document how wild forest foods can contribute to improved diets among these populations, including documenting the nutrient composition of more forest foods, integrating them into regional dietary guidelines and encouraging their domestication. Pour faire face à la malnutrition, à l'insécurité alimentaire et à un apport alimentaire appauvri, le potentiel des aliments forestiers est de plus en plus reconnu. Toutefois, la plupart des données existantes présentent des résultats moyens d'analyse approximative, faisant fi des opportunités de documenter la manière dont les produits forestiers peuvent contribuer à la nutrition, à la sécurité alimentaire et à l'apport alimentaire. Dans cette étude, la sécurité alimentaire a été estimée en utilisant le score HFAIS, alors que l'apport alimentaire était estimé en utilisant les rappels FFQ auprès de 720 foyers autour et au sein de six concessions de coupe au Cameroun, en République Démocratique du Congo et au Gabon. On trouva une haute diversité d'aliments forestiers consommés, la République Démocratique du Congo en enregistrant le plus grand nombre (FFCS=16.2). Le Cameroun comptait le plus grand nombre de groupes d'aliments (HDDS=12.5), alors que le Gabon pouvait se vanter de la plus grande variété d'aliments (FVS=24.5). Malgré la forte diversité d'aliments sauvages dans la région, la plupart des foyers (82%) pratiquant du fermage sur terres défrichées souffrait d'insécurité alimentaire. Une corrélation inverse importante était observée entre le score HFIAS et le FFCS (r2=–0.26, P=0.0002), révélant que les aliments forestiers peuvent jouer un rôle en assurant la sécurité alimentaire et la nutrition dans les populations vivant proches des concessions forestières. Le test chi-square révéla qu'un âge >46 ans et l'accès à des revenus provenant de commerce ou de salaires étaient associés à la sécurité alimentaire. Un travail supplémentaire est nécessaire pour documenter la manière dont les aliments sauvages forestiers peuvent contribuer à améliore les régimes dans ces populations, incluant la composition nutritive de davantage d'aliments forestiers, en les intégrant dans les lignes de conduite alimentaires régionales et en encourageant leur domestication. Cada vez se reconoce más el potencial de los alimentos del bosque para hacer frente a la malnutrición, la inseguridad alimentaria y una dieta deficiente. Sin embargo, la mayoría de los datos existentes presentan resultados promedio de análisis de constituyentes que pasan por alto las oportunidades de documentar la contribución de los alimentos del bosque a la nutrición, la seguridad alimentaria y la dieta. En este estudio, la seguridad alimentaria se estimó utilizando la puntuación de la HFIAS, mientras que la ingesta dietética se estimó utilizando cuestionarios de Frecuencia de Consumo de Alimentos (FCA) en 720 hogares dentro o alrededor de seis concesiones madereras en el Camerún, la República Democrática del Congo y el Gabón. Se observó una gran diversidad de Alimentos Forestales Consumidos (AFC=16,2), siendo la República Democrática del Congo la que registró el mayor número. Camerún mostró la mayor Diversidad Elevada de Dieta (HDDS=12,5) mientras que Gabón tuvo el mayor número de Variedades de Alimentos (VA=24,5). A pesar de la gran diversidad de alimentos silvestres de la región, la mayoría de los hogares (82%) que practican la agricultura en terrenos donde se ha talado el bosque sufren inseguridad alimentaria. Se observó una correlación inversa significativa entre la puntuación de la HFIAS y los AFC (r2=–0,26, P=0,0002), lo que revela que los alimentos forestales pueden desempeñar un papel para garantizar la seguridad alimentaria y la nutrición de las poblaciones cercanas a las concesiones forestales. La prueba de chi-cuadrado reveló que la edad >46 años, y el acceso a ingresos procedentes de negocios o salarios, estaban asociados con la seguridad alimentaria. Es necesario seguir trabajando en documentar la forma en que los alimentos silvestres del bosque pueden contribuir a mejorar la dieta de estas poblaciones, como la composición de nutrientes de un mayor número de alimentos del bosque, su integración en las directrices dietéticas regionales y el fomento de su domesticación.

snook@cgiar.org HIGHLIGHTS • A high diversity of forests foods for consumption exists in DR Congo, Cameroon and Gabon.
• Despite a diverse pool of wild foods, more than 80% of households engaged in logging suffer from food insecurity.
• A significant inverse correlation exists between food insecurity and forest foods consumption meaning that forest foods play a role in ensuring food security and nutrition among populations near forest concessions. • Age and access to income, illegal logging and agriculture on logged land are significant determinants of food security among forest communities. • Investing in strategies to provide income outside illegal logging and practicing agriculture on logged land, can contribute to addressing health, nutrition and food security.

SUMMARY
The potential of forest foods to address malnutrition, food insecurity and poor dietary intake is increasingly being recognized. However, most existing data presents average results of proximate analysis, overlooking the opportunities to document how forest foods contribute to nutrition, food security and dietary intake. In this study, food security was estimated using the HFIAS score while dietary intake was estimated using FFQ recalls among 720 households in and around six logging concessions in Cameroon, DR Congo and Gabon. There was a high diversity of forests foods consumed, with DR Congo (FFCS=16.2) registering the highest number. Cameroon had the highest number of food groups (HDDS=12.5) while Gabon had the highest number of food varieties (FVS=24.5). Despite a highly diverse pool of wild foods in the region, most households (82%) practicing farming on logged land suffered from food insecurity. A significant inverse correlation was observed between the HFIAS score and the FFCS (r 2 =−0.26, P=0.0002), revealing that forest foods can play a role in ensuring food security and nutrition among populations near forest concessions. The chi-square test revealed that age >46 years, and access to income from business or salaries were associated with food security. Further work is required to document how wild forest foods can contribute to improved diets among these populations, including documenting the nutrient composition of more forest foods, integrating them into regional dietary guidelines and encouraging their domestication.

BACKGROUND
The most recent global food security report revealed that severe or moderate food insecurity is on rising among Central and West African countries, raising from 45.3% in 2014 to 53.6% in 2019 (FAO, IFAD, UNICEF, WFP and WHO, 2019). Food insecurity, undernutrition and micronutrient deficiencies undermine health, psychological wellbeing, work capacity and economic development (Mbhenyane 2017, Vollmer, et al. 2016. Then again, reports elsewhere indicate that undernutrition and micronutrient deficiencies are a persisting public health challenges in Central Africa (Kamgaing et al. 2018), with Cameroon documented to have the highest stunting (a measure of height-for-age (HAZ) for individuals) (33%) and iron deficiency (45%) rates. While DR Congo has the highest vitamin A deficiency (42%) among women in the region. For the purposes of this paper, food security refers to the ability of communities, households, individuals and households to have physical and economic access to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active and healthy life (World Food Summit 1996).
Millions of people in Central Africa, including the Congo Basin, rely on forests for food and income (Pimentel et al. 1997, Laird 1999, Ndoye and Tieguhong 2004, Tieguhong and Ndoye 2007, Fungo et al. 2016a. The economic growth attributable to the forest sourced foods and their products, ranges between 6 to 10% per annum economic growth (Sonwa et al. 2012). The World Bank additionally estimates that, more than 50% of the Congo Basin forest population depends, to varying degrees, on forests for livelihoods, not just for food but also for fuel, livestock grazing and medicine (World Bank 2006). Then again, Angelsen et al. (2014) and Tata-Ngome et al. (2017), reported that wild foods obtained from tropical forests, are as vital to the livelihoods of rural people in developing countries as agriculture. Forests can contribute about 30.3% household food (about $304) including; fish, bush meat, fruits, vegetables and mushrooms important sources of proteins, vitamins and essential minerals for the rural poor households (Angelsen et al. 2014). The superiority of forest foods in contributing to essential nutrients and bioactive contents than the processed and imported foods has been documented in several African countries (Onimawo et al. 2003, Blaney et al. 2009, Fungo et al. 2015. For example, in Gabon and Cameroon, forest foods have been reported to address both macro and micronutrient deficiencies and non-communicable disease disorders related with inadequate intake of bioactive compounds , Blaney et al. 2009, Fungo et al. 2005. In Gabon, forest fruits of Poga oleosa, Panda oleosa, Gambeya lacourtiana and Afrostyrax lepidophyllus if consumed in adequate daily amounts by a non-lactating and non-pregnant woman, they can provide their 100% daily nutrient requirements (DRI), for magnesium (1000 mg/day), iron (58.8 mg/day) and zinc (12 mg/day). Also, 100% DRI for vitamin E requirements of 0.4 mg/day for children and 19 mg/day for women can be met by consuming adequate amounts of fruits of P. oleosa, P. oleosa and G. lacourtiana. In neighbouring Cameroon, children aged between 1-3 years and women of reproductive age can get considerable amounts of vitamins C, A and E, selenium, calcium, iron and zinc from forest foods (Fungo et al. 2016). Tata et al. (2019), further revealed that forests can provide foods that are protective against anaemia in women and children. Bush meat and fish are the major animal source foods providing significant amounts of digestible iron that is responsible to address anaemia among communities residing near forest concessions (Nasi et al. 2011, Fa et al. 2015. Furthermore, in the last decade, studies have documented positive correlations between household dietary diversity with presence of forests (Powell et al. 2012, Johnson et al. 2013, Ickowitz et al. 2014, Fungo et al. 2016a. For example Ickowitz et al. (2014), reported positive correlations between forest cover and dietary diversity, among forest dependent communities in 21 African countries.
It has been reported that forests provide food and other subsistence products to approximately 60 million people who live within and near them (De Wasseige et al. 2014, Marquant et al. 2015 as well as contributing to the nutrition of another 40 million people who live in urban areas near the forest estates (Nasi et al. 2011). However, to date few dietary intake and food security assessment studies have documented how wild forest foods can be used to address food insecurity and improve dietary intake among populations residing near concessions. In particular, studies that combine assessment of forest food consumption scores with food insecurity assessment techniques and dietary assessment methods are limited. It consequently remains unclear how forest biodiversity contributes to food security and dietary intake for rural populations in Central Africa. Therefore, the overall objective of this study was to investigate the role of wild forest foods in diet and food security of villagers living in forest environments in Gabon, Cameroon and DR Congo. It describes the dietary indices of women, evaluates their food insecurity security categories and their coping mechanisms and documents the relationship between food consumption indices and food security indicators. Furthermore, the determinants of food insecurity among forest foods consumers were analysed.

Study concessions
The study was carried out among communities residing in or around six purposefully selected forest concessions (Table 1), adjoining forests in Cameroon, Gabon and DR Congo. These three countries account for more than 80% of the forest cover of Central Africa (Tchatchou et al. 2015). Study concessions were selected based on: existence of trees of multiple use (timber/food value); ease of access to sites for the research team; the willingness of the forest concessionaires to allow the study to be implemented in or around concession villages; and the presence of village populations and various ethnic groups willing to participate in the study. Concessions were selected to determine whether their logging activities were reducing access to forest foods for villagers, concessions that were actively logging were selected.

Concessions, districts, villages and households' selection design
A four-stage cluster sampling technique involving two stages of purposeful selection and two stages of randomization were deployed during the selection of the study villages and households. The first stage involved selecting concessions, following the criteria outlined above. The second stage involved purposefully selecting districts within each concession on the basis of their accessibility, ethnicity and willingness of the community to participate in the study and ongoing logging activities in the surrounding forests. The third stage involved selecting five villages randomly within the two chosen districts. With the assistance of the village authorities including village chiefs, study villages were mapped and listed using village transect walks. Households and their occupants were listed in each village for random selection. The fourth stage involved selecting 24 households randomly within each village. This resulted into a total of 720 households (6 concessionaires*5 villages per district*24 households per village=720 households). The survey was simultaneously conducted in and around the six forest concessions.

The interviews
Structured questionnaire-based interviews were administered to collect data on socio-demographic characteristics, household food insecurity access scale (HFIAS) and food consumption from 720 non-pregnant and non-lactating women older than 18 years who were married or cohabiting. For polygamous households, one respondent among the wives of the household head was selected to respond on behalf of the rest of household members in case the entire household (including her co-wives) were preparing one meal for the entire family. In case house wives were preparing food separately for their children and husband the respondent responded to questions with reference to only those she took care of in her household. Women were interviewed because in rural African settings they are responsible for food collection and preparation and are custodians of knowledge on wild and forest foods, that is passed on from generation to generation (Maundu 1992, Becquey et al. 2009). Interviews took place in the respondents' homes, during the months of May-October 2012. The long rainy season in the region commences in May and ends in November (United Nations Environment Programme 1999). The questionnaires were pre-tested and translated into local languages for those that could not speak French by educated and locally recruited Field Assistants.

Household characteristics
Household characteristics documented included information on the age, gender, economic occupation, education level of the head of household, source of water used in household and period of stay in the study area.

Dietary diversity scores
Dietary diversity scores were obtained using a seven day recall food frequency questionnaire that was designed preceding the survey to capture information. A total of thirteen food groups were assessed including the twelve food groups specified in the FAO guidelines for measuring household and individual dietary intake, in addition to the food group for healthfully important forest foods (Arimond et al. 2010, FAO 2011. The food groups were: (i) cereals and wheat products; (ii) roots and tubers; (iii) legumes and lentils; (iv) nuts; (v) dairy and fats; (vi) meat; (vii) poultry; (viii) fruits; (ix) fish and seafood; (x) vegetables; (xi) alcoholic drinks; (xii) nonalcoholic beverages ; and (xiii) forest foods (including bush meat, honey, caterpillars, wild fruits, roots and vegetables). In the present study, forest foods were defined as foods of plant and animal origin, which were growing wildly and collected from forests (Agbogidi 2010). The food items consumed were distributed among the thirteen food groups and the household dietary diversity score (HDDS) was calculated by summing the number of food groups consumed. The food variety score (FVS) was computed as the number of different food varieties or items consumed over the recall period of 7 days. The forest food consumption score (FFCS) was calculated by summing the occurrences and incidents of consuming of forest food items and varieties (FAO 2011 did not have access to protected water sources. Despite living in forests, more than 90% of the respondents in each country were farmers who also practiced illegal logging. More than three quarters of the respondents in all the countries were natives of the study sites and 72% or more had lived there for more than 10 years.

Household dietary diversity, food variety and forest food consumption scores
Respondents' FVS for the seven days preceding the interview ranged from 11 to 23 in Cameroon, 16 to 28 in DR Congo and 17 to 30 in Gabon (Table 3). During the seven day recall period, each individual reported consuming from 11 to 23 food items in Cameroon; from 16 to 28 food items in DR Congo; and from 17 to 30 food items in Gabon. In Cameroon, the majority of the respondents (67%) consumed 15 to 19 different food items over the seven day period, while in DR Congo about 50% of the respondents consumed 17 food items, and in Gabon 67% of the respondents consumed 23 to 25 food items. In terms of food groups, about 50% and 60% of the respondents in DR Congo and Cameroon, respectively, ate items from all the thirteen food groups during the seven day recall period. In Gabon, respondents reported consuming from only 11 food groups. While 94% of respondents in Cameroon had DDS of 12-13 (consumed food items from thirteen food groups), with a mean of DDS 12.5±0.7 food groups, in Gabon 79% respondents had DDS of 9-10 (mean DDS 9.6±0.8 food groups) and in DR Congo 100% respondents consumed DDS of 11-13 (mean DDS 11.9±1.0 food groups). As regards forest food consumption, the highest was registered in DR Congo with FFCS ranging from 15 to 18 forest foods consumed during the seven day recall period. Ninetyfive percent of respondents in DR Congo had high forest food consumption (FFCS of 16-17), with a mean FFCS of 16.5±0.59 forest food varieties. In DR Congo animals sourced foods were the most consumed including; caterpillars (Gonimbrasia belina), snails (Gastropoda) and cat fish (Siluriformes). Wild mushrooms (Fungi) commonly consumed forest foods. In Cameroon the FFCS ranged from one to eight, with the largest group (31%) reporting having eaten three forest food items (mean FFCS 3.97±1.6) during the seven day recall period. Most consumed forest foods in Cameron included; Irvinga gabonensis (bush mango), a wildly growing fruit used for cooking oil production and as a food thickening agent in soups and stews, and wildly growing oranges (Citrus spp). Other forest foods included wildly growing yams (Dioscorea spp.) and forest harvested nuts, Bambara groundnuts (Vigna subterranea). Among the forest vegetables, Solanum nigrum was the widely consumed greens. Among the meats, (porcupine and rats), were the most commonly consumed in Cameroon. Gabon registered the fewest forest food varieties and items consumed, with FFCS ranging from one to three. In Gabon, taro (Colocasia esculenta), porcupine and wild birds were the most consumed. Seventy-three percent of the respondents interviewed did not consume any type of forest food.

Household food insecurity assessment
The Household Food Insecurity Access Scale (HFIAS) was assessed following the method described by Coates et al. (2007). A set of nine standard questions was posed to women, who responded on behalf of other household members. The HFIAS score is a continuous indicator ranging from 0 (food secure) to 27 (maximum food insecurity) (Coates et al. 2007), with the score categorized into four levels of household food insecurity: food secure (score=0), mildly food insecure (score=1-13), moderately food insecure (score=14-16) or severely food insecure (score=17-27). The household food insecurity access scale score (HFIAS Score) and the household food insecurity access prevalence categories (HFIAP categories) were calculated by country. The household food insecurity occurrence and coping mechanisms and food insecurity categories per concession were also calculated. The HFIAS score was calculated for each household by summing the codes for each frequency of occurrence question. The higher the HFIA score, the more food insecure (access) the household experienced. The HFIAP categories per country and between consumers and non-consumers of forest foods consumers were also computed by adding the occurrence of different categories of food insecurity.

Data analysis
All data were analysed using the statistical software package IBM SPSS Statistics Version 21.0 and statistical significance was set at P < 0.05 for all tests. The mean values were computed for continuous variables while proportions were computed for the categorical variables. Differences between means or proportions were considered statistically significant if P < 0.05. To assess how forest foods are related to food security, logistic regression analysis and Spearman's correlation analyses were performed. Chi-square tests were further performed to assess the determinants of food insecurity among forest food consumers. Respondents were dichotomized into food secure and food insecure, the latter including those who were suffering from mild, moderate and severe food insecurity.

Characteristics of study populations
A total of 720 households were surveyed. The mean age of the sampled female respondents was 46 years (SD = 10.4) with a range from 17 years in both DR Congo and Cameroon to 89 years in Gabon ( Table 2). The majority of respondents interviewed were monogamously married with about 85% of these marriages registered in Cameroon, 58% in Gabon and 100% in DR Congo. In terms of household size, Gabon had the fewest household members (5) while DR Congo had the most (8). About a quarter of respondents in each country had no formal education while three quarters of the total respondents in both Cameroon and Gabon and 100% in DR Congo,

Household food insecurity access and prevalences scales
According to the HFIAP categorization, nearly half (48%) of the respondents in Cameroon, and 100% and 88% of households in DR Congo and Gabon, were respectively, severely food insecure (Table 4). Cameroon had the highest number of food secure households (20%), followed by Gabon with only 4%. None of the households surveyed in DR Congo were registered as food secure. Categorizing all forms of food insecurity of HFIAS by forest concessions revealed that the average HFAIS scores were significantly different between the two forest concessions in both Gabon and Cameroon (Table 5). For instance, in Cameroon, the proportion of households that were food-insecure around SCTB concession without a forest management plan, was significantly higher (at P  0.0001) than the proportion that were food-insecure in FIPCAM (74 vs. 5%, respectively). In Gabon, the average proportion of households that were food-insecure was significantly higher around CEB without a forest management plan, compared to those registered in Bayonne (96% vs. 89%, respectively) at P  0.03. However, in DR Congo, the populations in and around both concessions were all (100%) food-insecure.

Occurrence and coping strategies for Household Food Insecurity Access
With regards to HFAIS occurrences, several coping strategies to address food insecurity were used in varying degrees by the respondents from the three countries. The respondents in Cameroon were the most food secure, thus fewer respondents employed the coping strategies listed under the HFIAS scale compared to the respondents in DR Congo and Gabon. For example, in Cameroon, during food-insecurity episodes, just over half of the households deployed three coping strategies including 53% who reported eating just a few kinds of foods, 62% who ate smaller meals and 54% who ate fewer meals in a day. On the other hand, in DR Congo, more than 80% of the households used seven coping strategies including: expressing anxiety and uncertainties about accessing food; eating a few kinds of foods; eating foods they do not want to eat; eating a smaller meal; eating fewer meals per day; and not eating any kind of food. In Gabon all the nine strategies listed HFIAS were used by more than 70% of the households interviewed.

Relationship between forest food consumption and food insecurity and dietary intake
Logistic regression analysis revealed that forest food consumers were 90% (OR=0.9; 95% CI 0.71, 4.01; p=0.001) more likely to be food secure compared with non-forest food consumers (Not in table). There was no significant difference in the proportion of consumers of forest foods who were food secure as compared to non-consumers (12% vs 11%, respectively; p=0.74) (Table 6). However, fewer consumers of forest food reported severe food insecurity (60% as compared    Household food insecurity access scale score 1 *, **: Correlation is significant at 5% and 1% level respectively. a : The maximum score includes 13 food groups b : Only forest food species were included in the count group to 69%, p0.0001; Table 6). When the results were further subjected to Spearman's correlation analysis, it was observed that among forest food consumers, the mean HFIAS score was significantly and inversely correlated with the FFCS (r 2 =−0.26, p=0.0001) ( Table 7). This suggests that households that consumed more forest foods were less food insecure. Statistically significant positive correlations between the FFCS and the FVS (r 2 =0.29, p=0.05) and between the FFCS and the HDDS (r 2 =0.25, p=0.0001) further support the association between higher forest food consumption and reduced risk of food insecurity.

Socio-demographic factors affecting food security among consumers of forest foods
The chi square test analyses in the present study revealed that among forest foods consumers, only a minority of food secure households (31%) had a younger household head (<46 yrs) while the majority of food insecure households (81%) have a younger household head. (Table 8). A high proportion of respondents (~82 %) who were food insecure relied on illegal logging and subsistence farming on previously logged and abandoned land. Food insecure households the share is significantly lower (18%). The 18% of those who were foodsecure described themselves as business owners and 15% as salaried employees.

Contribution of forest foods to dietary indices
To the best of our knowledge, this is the first regional study in which the relationships between food insecurity and HDDS, FVS and FFCS have been assessed among forest dwelling populations in Central African countries of Cameroon, Gabon and DR Congo consuming all types' of foods that were both plant and animal sourced. Although, Tata-Ngome (2016) quantified food insecurity in Cameroon, the quantification in Cameroon was a derivative of consuming only fruits. Furthermore, Tata is considered a useful indicator of household food security (Mbhenyane 2017, Hoddinott andYohannes 2002) and a key proxy of nutrient adequacy (Rathnayake et al. 2012), which is apparent in most dietary patterns across Africa. In the present study, a high consumption score for forest foods (FFCS) was registered in DR Congo and Cameroon, while in Gabon, a high food variety score (FVS) was registered. Previous studies in DR Congo (Termote et al. 2012) and Cameroon (Tata-Ngome 2015, Fungo et al. 2016b, documented how forest foods can mitigate food insecurity and health disorders related to inadequate intake of nutrients, if consumed in adequate amounts. In Benin, a higher diversity score and higher food security were documented among the population residing around the precincts of the government owned forest reserve in the North West of the country, than among the populations in urban centres of Parakou city. This was attributed to increased access to forest tree foods ( Van Liere et al. 1995). In Benin, the government grants periodic permits to residents to collect forest foods from the forest reserve.

Household food insecurity and coping strategies
The most recent United Nations food security report revealed that more than half (54%) of the total population in the Congo Basin Region were food insecure (FAO, IFAD, UNICEF, WFP and WHO 2019). The present study found a much higher prevalence of food insecurity, with 78 % of the populations participating in this study being categorized as severely food insecure, 17% as moderately and 8% as mildly food-insecure. The high prevalence of food insecurity among the respondents in the present study may be due to the majority having low levels of education, to the high proportion of respondents (~82 %) relying on illegal logging and subsistence farming on previously logged and abandoned land for income in the three countries and having a large number of household dependents. Previous studies in Cameroon (Fungo et al. 2016b, Tata-Ngome et al. 2017) and in DR Congo (Termote et al. 2012) have attributed high food insecurity among forest food consumers to low volumes of forest foods consumed, seasonal availability of forest foods and practicing agriculture on previously logged degraded land. This was corroborated in studies carried out by Levang et al. (2015), Tata-Ngome (2015) and Leakey (2019), who reported that the forest food and fruits gatherers and also practicing agriculture on degraded logged land, suffer worse food insecurity than those practicing agriculture alone. This may reflect the fact that those who are short of food from abandoned logged agriculture land rely on forest foods as a safety net Shackleton 2004, Kuhnlein et al. 2007). Similar studies conducted among the forest communities of Benin (Boedecker et al. 2014) and Burkina Faso (Maisonneuve et al. 2014) revealed that despite the communities' having access to highly diverse forest diets, low dietary intake and high food insecurity were rampant among these communities. Food insecurity in Burkina Faso and Benin were attributed to the seasonality of forest foods, the annual and regional variations in food availability, and the small portion sizes of forest foods consumed.
Coping strategies are the methods used by households or individuals to meet their food and nutrient requirements, or survive when faced with food scarcity (Ellis 2000). In the present study, different coping strategies listed under the HFAIS scale (Coates et al. 2007) were adopted by foodinsecure households. In Cameroon, among respondents who were the most food-secure, fewer employed the coping strategies, as compared to the respondents in Gabon and DR Congo, who were faced with more severe food scarcity. For instance, worrying about not having enough food in the household was experienced by fewer than 50 percent (43%) of the food insecure respondents in Cameroon, compared to 98% and 94% of DR Congo and Gabon respondents, respectively. On the other hand, communities in DR Congo and Gabon who suffered from severe food insecurity, deployed progressively more coping strategies as food insecurity levels increased in these two countries. For example, almost all (94% to 100%) of the food-insecure respondents in DR Congo and Gabon employed four coping strategies: consumption of only a few kinds of foods, not eating preferred foods, eating a smaller meal and consuming foods that the respondents did not really want. In Cameroon, DR Congo and Gabon, about 30%, 40% and more than 70%, respectively, of the respondents that were food-insecure would go a whole day and night without eating and went to sleep hungry; some did not have any kind of food. Tata-Ngome (2016) reported a much higher population of 50% going to sleep hungry. Food insecurity results in decreased nutrient intake (Berman et al. 2014). This was confirmed in this study, with the majority of the food-insecure respondents from Cameroon (62%), DR Congo (~100%) and Gabon (96%) cutting portion sizes, and 54%, 99% and 95%, respectively, consuming fewer meals.

Relationship between food consumption indices and food security
Spearman's correlation analysis and regression modelling in the present study revealed that greater forest food consumption was significantly and positively associated with increased dietary diversity and food security among consumers of forest foods. The significant inverse correlation between FFCS and HFIAS, and the positive correlation between FFCS and HDDS and FVS, indicate that forest foods may be playing a significant role in household and community food security and nutrition. These findings are in agreement with findings among forest dependent women of reproductive age from Embolowa region and Bertoua region in the South and East, in Cameroon (Tata-Ngome 2017, Fungo et al. 2016b) and among children and women residing in Awajún forests of Peru (Roche et al. 2008) and among rural household farmers of northern Ethiopia (Maxwell et al. 2014). In Cameroon, Fungo et al. (2016a) and Tata et al. (2017) observed a positive relationship between the forest food consumption score and household dietary diversity score, food variety score and food security. Among the Awajún community of the Amazonian forests of Peru, Roche et al. (2008) reported a positive association between the traditional forest food diversity score and food security, with dietary intakes higher among the forestdependent communities residing in the lower Cenepa River region of the Awajún community. The majority of previous studies relating dietary diversity to household food insecurity focused on assessing household dietary diversity, individual or women's dietary diversity by measuring conventional food groups or individual food items over a fixed period of time (Ruel 2003). However, in the present study and the previous study reported by Fungo et al. (2016b) in Cameroon, the score of forest foods consumption (FFCS) was added to the list of commonly used dietary diversity indicators that assessed household food security. As a result of high biodiversity in the study areas in Cameroon, Gabon and DR Congo, and the high nutrient content of several forest foods (Kana-Sop et al. 2008, Djoulde et al. 2012, Fungo et al. 2015, inclusion of the FFCS indicator is a useful complementary measure for assessing dietary diversity and food security among the forest dependent communities.

Determinants of food insecurity among forest foods consumers
The proportion of respondents in the present study that prepared household meals using forest foods (79%) was considerably higher than what has been reported elsewhere in the Congo Basin countries of Cameroon (Fungo et al. 2016) and DR Congo (Termote et al. 2012). In the villages of Turumbu and Kisangani (DR Congo), 22% of the forestdwelling population was reported to prepare meals with forest-sourced foods (Termote et al. 2012) while in the South and East regions of Cameroon, about 40% of the population prepared meals with forest foods (Fungo et al. 2016b). The higher consumption of forest foods documented in the present study may be attributed to greater number of respondents (720) sampled from villages close to forest concessions, as compared to the 278 in Cameroon and 241 in DR Congo sampled from towns and cities by Termote et al. (2012). The chi-square test analysis in the present study revealed that food security of households consuming forest foods was higher when the head of household was aged 46 years and above. Previous studies have documented improved food security status among older household heads who consumed forest and wild foods (Pelto et al. 2004, Tata-Ngome et al. 2017, Fungo et al. 2016b. A study from eMantlaneni village in the East of South Africa, revealed that elderly household heads were not only more knowledgeable about wild foods but also these households were more food-secure than households with younger household heads (Dweba and Mearns 2011). Other studies revealed that older household heads experience more stable household food security (Egger and Dixon 2014) and associated this with their better access to land and capital, a major factor to producing their own foods, than younger household heads (Egger andDixon 2014, Tata-Ngome 2015). This study found, also, that households with income from salaried employment or business ownership, were more likely to be food-secure.
The high food insecurity recorded among the younger respondents in the present study may be attributed to rapid westernization of diets in Congo Basin forest countries, easier access to imported and processed foods than forest foods and associating consumption of forest or wild foods with poverty (Rensburg et al. 2007, Fungo et al. 2016a. The aggressive promotion of imported and processed food crops, by industrialists, agricultural research centres and government extension officers have been documented as some of the bottlenecks hindering use of forest foods and traditional Indigenous foods to address the high food insecurity in rural Africa (Dweba and Mearns 2011, Keller et al. 2006, Rensburg et al. 2007). This is associated with an inability of the older generation have poor form and others are inaccessible; and thirdly, local people tend to gather forest foods within a limited radius of their villages (Noutcheu et al. 2016, Muvatsi et al. 2017, Taedoumg et al. 2018.
The potential of forest foods to address food insecurity has been neglected by governments, donors, non-governmental organizations and community-based initiatives for decades, yet it represents an opportunity worth exploring further (Fungo and Tieguhong 2019). Integration of forest foods into agro-forestry interventions and policies for better food and nutrition security coupled with sustainable use of forest biodiversity, could contribute to addressing food insecurity among rural populations in Central Africa (Leakey , 2014(Leakey , 2017. A limitation of this study is that it focuses on forest dependent communities residing around logging concessions. The results can thus not be generalized to all forest dependent communities in general. Furthermore, this was a cross-sectional study and no causal pathways between food insecurity and dietary diversity could be drawn. Although this study has revealed that some socio-demographic factors are associated with food and nutrition security among consumers of forest foods, the study does not provide the reasons for the rampant food insecurity in Central Africa. Longitudinal studies are thus recommended to further explore relationships. Finally, this study was carried out at during the peak of the rainy season within the region, which relates with the most food secure period of the year in Central Africa. To increase the significance of the results, the same research could be repeated during the dry season because forest foods availability is dependent on seasonality.

CONCLUSION
This study points to a serious problem of household food insecurity among communities residing adjacent to the forest concessions in Congo Basin, affecting about 90% of respondents. The results provide some of the first insights into the food insecurity status and dietary diversity measures among communities adjoining forest concessions. Across the region, a high food insecurity score was found among both those who consume forest foods, who were most frequent in DR Congo, and among those who did not, in Gabon. A high forest foods consumption score (FFCS) in DR Congo did not result in food security. However, fewer consumers of forest foods were severely food insecure than was the case for those who did not consume forest foods. Furthermore, an inverse relationship between FFCS and HFIAS and the positive correlation between FFCS and HDDS and FVS, implies that forest foods have the potential to contribute to addressing food insecurity in the Congo Basin. Given that access to income from sources other than agriculture on logged land and illegal logging was associated with higher food security, investing in ways to provide income outside of these areas could contribute to addressing health, nutrition and food security. To sustainably address food insecurity in the three countries, urgent action is required to promote, domesticate and conserve nutrient-rich forest foods that are locally acceptable.
to successfully pass on their Indigenous knowledge about African forest foods to the younger generation, exacerbating the substitution of forest foods with imported foods in Africa (Fungo et al. 2016a).
Studies elsewhere in Africa, including among rural communities in South Africa (Pelto et al. 2004), Cameroon (Fungo et al. 2016a, Tata-Ngome 2015 and Uganda (Tabuti et al. 2004) have reported better food security among forestdependent communities with little or no formal education. The findings in the present study corroborate previous findings in Cameroon, which revealed that the uneducated consume more of the wild forest foods than the educated (Fungo et al. 2016a). In contrast, the food-secure coastal Mediterranean populations of Europe with a higher nutritional knowledge, consume more nutrient-rich wild and forest foods than the less educated in Europe (Serra-Majem et al. 2007). Fungo et al. (2016a) further attributed the difference in the results between the Mediterranean study and the present African studies to socio economic status, health consciousness among the populations of the two continents. African elites with higher income have tended to move away from eating traditional forest foods with strong cultural identity, to consuming less nutritious western foods (Pingali 2007). In many such communities in central Africa, forest foods are increasingly being replaced by imported and refined foods that have high saturated fats and sugars (Frison et al. 2006). For instance, in Gabon, where forest foods are widely being replaced with refined imported diets, diet-related non-communicable diseases including obesity, type 2 diabetes and cardio-vascular diseases are on the rise, especially among the urban elite (Siawaya et al. 2015). Studies among forest populations of Central Africa have suggested that populations residing in or around forests that are designated as concessions for timber logging may be restricted from obtaining nutrient-rich forest sourced foods (Laird 1999, Blaney et al. 2009, Guariguata et al. 2010, Tata-Ngome et al. 2017). This restrictions can have an impact on the food security, health and welfare of these populations. In Cameroon and Gabon, about 50% of the forests are allocated to logging concessions, which are legislatively protected areas. It has been hypothesized that this may, or has, limited the surrounding populations' access to nutrient rich forest foods (Blaney et al. 2009, Rist et al. 2012, Tata-Ngome et al. 2017. For example, in the North of DR Congo, Hardin and Auzel (2001), reported how one logging company that employed about 650 workers, consumed 390 tons of bush meat (live animal weight), or close to 35 000 animals per year, depriving the surrounding communities of an important source of proteins and micro nutrients. Consumption of wild animals by logging company employees deprives communities of important sources of dietary fat and other energy sources, leading to severe food insecurity (Bailey and Peacock 1988). However, recent studies have found that the situation for other forest food resources is not so clear: for one thing, trees below the minimum cutting diameter provide both fruits and edible caterpillars, but are not cut by industrial loggers; for another, not all commercialsized trees are removed in harvesting operations, as some