Farmer knowledge and perceptions of agrochemicals use on honeybee community and honey production in Benin

Apiculture faces many constraints including the use of agrochemicals. Chronic exposure to chemical pesticide use results in the loss of honeybee colonies and alters detoxication mechanisms and immune responses, driving them more susceptible to parasites. This study aims at investigating the impact of agrochemical use on honeybee community and honey production in northern Benin. For this purpose, 100 beekeepers were surveyed using a survey questionnaire. Information on treated crops, the distance of beehives from these crops, the effect of insecticides and herbicides, adaptation patterns and honey production were collected through individual interviews. Socio-demographic data and farmer’s knowledge on agrochemical uses and their impacts on honeybee communities were determined. Information were also collected on alternative methods to agrochemicals, beehive types and honey taste.


Abstract
Background Apiculture faces many constraints including the use of agrochemicals. Chronic exposure to chemical pesticide use results in the loss of honeybee colonies and alters detoxi cation mechanisms and immune responses, driving them more susceptible to parasites. This study aims at investigating the impact of agrochemical use on honeybee community and honey production in northern Benin.

Methods
For this purpose, 100 beekeepers were surveyed using a survey questionnaire. Information on treated crops, the distance of beehives from these crops, the effect of insecticides and herbicides, adaptation patterns and honey production were collected through individual interviews. Socio-demographic data and farmer's knowledge on agrochemical uses and their impacts on honeybee communities were determined. Information were also collected on alternative methods to agrochemicals, beehive types and honey taste.

Results
Apicultural investigations have shown that chemical pesticides were applied in elds of crops and that beehives were mostly located near treated elds. The most used chemical pesticides were Kalach (glyphosate), Callifor G (glyphosate, uometuron, prometyn), Atrazila 80 wp (atrazin), Herbextra (glyphosate) and Cotonix (deltamethrin and chlorpyriphos). The majority of beekeepers (79%) recognize the effect of pesticides and use alternatives to pesticides to reduce the risk. In general, 49% of beekeepers found the decrease in the big breed of bees and 40% that of the small breed. The honey production was higher when the beehives were installed far from the treated elds.

Conclusion and implications:
This study allowed to knowing the direct and indirect effects of the agrochemical use on beekeeping.
Alternative pest management methods to the pesticide use practiced by farmers would lead to sustainable improvements of honey production in Benin.

Background
Achieving high yields in agriculture requires the use of pesticides for the weed control and plant health management. The sharp increase in the use of chemical pesticides in recent years endangers plant and animal biodiversity [1; 2]. The increased use of chemical pesticides coupled with the agricultural boom is raising the vulnerability of natural resources, plant and animal biodiversity [3]. Many pesticides are toxic to bene cial insects, birds, mammals, amphibians or shes [4,5]. The poisoning of non-target species depends on the toxicity of a pesticide and its other properties, the amount applied, frequency, timing and method of spraying, climate, vegetation structure and type of soil. Insecticides, rodenticides, fungicides and herbicides threaten the animal species that is exposed to them [6]. During these last years, the use of very toxic carbamates and organophosphates has increased signi cantly. In Africa, organochlorines such as endosulfan, which are highly environmentally persistent, moderately persistent in soil systems and not in aqueous systems are still widely used [7,8,9]. Like habitat change, pesticide poisoning can cause major decline of a population that threaten non-target species.
Pesticides are a major factor affecting biodiversity [10] and especially bene cial insects. Several studies have reported unusual bee colony decay and mortality in several countries around the world [11]. Broad spectrum insecticides including carbamates, organophosphates and pyrethroids can cause population decline of bene cial arthropods such as bees, spiders and beetles. Many of these species play an important role in the food web or as natural pests' enemies. Bees provide essential pollination.
Honeybees are also under pressure from parasitic mites, viral diseases, habitat loss and pesticides [12,13]. Intensive farming practices, habitat loss and agrochemicals are considered to be among the main environmental threats to honeybees and wild bees. The relative low number of detoxi cation genes in bees suggests greater sensitivity to pesticides compared to other insects [14]. It has been frequently reported that exposure to pesticides can have a direct effect on certain compounds of the immune system, even on physical defenses or behaviors preventing contamination in insects, particularly bees [15]. Thus, pesticides can have an effect on the humoral immune responses of insects [16]. They can also impact cellular responses. Thus, beekeeping professionals estimate that there was a loss of global honey production of 20 to 30% between the years 1997 and 2009 [17].
In Benin, pesticides are used in agriculture with the installation of beehives not far from elds treated with agrochemicals. These farming practices could also have adverse effects on the dynamics of the bee population and the honey production [18]. Chemicals are used to treat seeds of agricultural products.
When a treated seed grows, chemicals can integrate in the environment to poison the bees [19].
Chemicals residues in pollen from crop plants from treated seed may pose a high risk to bees [20]. Sublethal doses of chemicals may lead to reduced learning ability in bees [21]. These toxic molecules (neurotoxicants) were found in plant pollens [22,23] and their effects associated to the declining bee health remains to be determined. Indeed, the Northern Benin is an area of strong agricultural and apicultural production characterized by two breeds of bees, the rst race with small, yellow bees, more aggressive and producing more honey and the second race with larger, black, bees, less aggressive and producing less honey. Since these two income generating activities are carried out in the same agroecological zones and with the same farmers, they must be done sustainably. Agricultural activities must not pollute honey produced by beekeepers for consumption. These areas of agricultural production are polluted by agrochemicals. Crops are treated with chemical pesticides the most treated being cotton. The increase in the number of organic cotton producers would contribute to reduce the pollution by chemical pesticides. Farmers continue to use agricultural practices based on biopesticides based on extracts of plants as Azadirachta indica for example [24,25]. They also use compost and animal droppings to fertilize the soils. These agroecological practices used by some farmers would reduce the risk of contamination of beekeeping systems closer to organic beekeeping.
This work aims to determine the in uence of the use of chemical pesticides in agriculture on the honey bee abundance and their potential for honey production in Benin. Speci cally, this involves: (1) listing all the chemical products (insecticides, nematicides, herbicides, fungicides) used in agriculture in the study area, (2) collecting the folk perception on the effects of chemical use in agriculture on honeybee communities and honey production in the study area and nally (3) determining the in uence of the distance of beehives from cultivated elds on the honey production.

Study site
Four districts investigated in the study were Banikoara, Gogounou, Kandi and Segbana. These districts were located in Alibori region and characterized by agricultural and beekeeping production with a wide range of chemical pesticides used. Alibori is located between 11°19' north latitude and 2°55' east longitude. It is bordered to the north by the Republic of Niger, to the north-west by the Republic of Burkina Faso, to the south by the Borgou Department, to the east by the Federal Republic of Nigeria, and to the west by the Department of Atakora. Its area is 26242 km² (23% of the national territory). The far-north zone and the cotton zone of north Benin are the two agro-ecological zones. It has only one rainy season which lasts between 5 to 6 months with annual precipitation ranging between 700 mm and 1200 mm.
The vegetation is a mosaic of dense dry forests, clear forests, wooded savannas and shrubby savannas [26]. The most dominant crops are sorghum, rice, maize and cotton.

Selection of respondents
An exploratory survey was made to select districts for investigation. Choice of those districts was based on following criteria: i) agricultural production and beekeeping and (ii) pesticide use in agriculture. In each district, 25 beekeeper farmers were chosen randomly, making a total of one hundred (100) farmers in the study area. Then, data were collected about the socio-demographic characteristics of the farmers. Structured questionnaire was administered to 100 beekeepers selected using a multistage sampling procedure in villages of the four districts. Individual and group surveys were conducted among beekeepers.
Data collection of agrochemicals use on bee communities and honey production First, we have inventoried all crops produced that require the use of agrochemicals in the study area and also made an inventory of the agrochemicals used with their active ingredients. Secondly, we sought to understand the period during which farmers were heavily using chemicals and that of an intense beekeeping activity. Data were collected on whether these agrochemicals have contaminated bees and honey during the same year that both these two activities were conducted using farmer's perception.
Farmers were asked to list all the agroecological practices they use as an alternative to the use of chemicals. Third, we determined farmers' perceptions of the possible effects of using agrochemicals on bees and honey produced. In this part, the farmers gave their perception of the increase or decrease of each of the two breeds of bees in the study area. Then, the farmers provided information on the average abundance of the two breeds of bees per hive for the study area. Finally, the relationship between of the distance between the beehives and the elds treated on the honey production were determined. For that, all the beehives placed between 100 and 500 m from the treated elds and the quantity of honey produced per beehive were identi ed.
Data collection on beehive systems types and the honey taste Information has been collected on the types of beehive systems. Three categories of beehives (modern type, wood type, sheet metal type) have been identi ed in the study area [27]. The amount of honey produced by these three types of beehives has been determined. The taste of the honey produced by these beehives types was determined using scores (0.20: very poor taste; 0.40: poor taste; 0.60: good taste; 0.80: very good taste; 1: better taste).

Data analysis
Data collected was summarized using to descriptive statistics (mean, standard deviation, frequency and percentage). Correspondence Analyzes (CA) were carried out using the 'FactoMinor' package to determine the relationship between the honey production districts and i) types of herbicides used by farmers, ii) types of insecticides used by farmers. Correspondence Analyses (CA) were also carried out to show the relationships between the crops and the herbicides and insecticides used by farmers. GLMs were used to determine the relationship between pesticide use and abundances of small and large breeds of bees. A linear model (lm) was used to understand the relationship between the distance from the treated elds and the amount of honey collected per hive. This model was simulated using the PermTest function of the 'pgirmess' package. A probalistic function was used to estimate the quantity of honey due to uctuations in the distances from the treated elds [28]. Consequently, a stochastic Monte Carlo model was used to generate the honey quantity per beehive based on a random outputs model with a normal distribution function [29] with a 10,000 iteration model. GLMs have also been used to show the relationship between types of hives and the quantity and taste of honeys. The TukeyHSD test was used to show the signi cant difference between the honey tastes produced from modern beehives, beehives made from woods and beehives made from sheet metal. All statistical analyzes were performed with R version 3.2.2 [30] at the signi cance level of 5%.

Socio-demographic characteristics of surveyed farmers
Beekeeping is an activity mainly performed by men (97%). Women were among the surveyed farmers and represented 3%. Of surveyed farmers, 76% of farmers had no formal education and others farmers had primary (12%), secondary (9%) and university (3%) levels. The surveyed beekeepers had a mean age of 37.58 ± 1.48 (SD) years. The age group of 20-35 years had the largest number, with 45% of the farmers.
Inventory of the main treated crops and their requirements for agrochemical treatments in Northern Benin Whatever the crop, farmers necessarily applied some agrochemical to control weeds or pests of crops. The most popular crop combinations were cotton and maize, cultivated by all the beekepers, followed by sorghum, soya, cowpeas, yam and millet. Among these crops, cotton (100% of farmers), maize (100% of farmers), sorghum (60% of farmers), soy (45% of farmers), cowpeas (20% of farmers), yam (3% of farmers) and millet (2% of farmers) needed agrochemical treatment (Fig. 1).
Correspondence Analysis (CA) realized to show the relationship between honey production areas and herbicide types revealed that the rst factor plane formed by axes 1 and 2 explained 87.78% of the total variability. The rst axis contributed for 61.18% and the second 26.60%. The projection of the honey production areas in the factorial plan de ned by 1 and 2 revealed that the beekeepers from Segbana and Banikoara used the herbicides Atraforce, Callifor and Atrazila in their elds. The beekeepers from Gogounou used the herbicides Cottonex, Herbextra, Adwuma, Kabasate, Butaforce, Malik, Grasskiller and Glyphader. The beekeepers from Kandi used Kalach and Butaplus (Fig. 2). Correspondence analysis (CA) performed to show the association between honey production areas and insecticide types revealed that the rst factor plane formed by axes 1 and 2 explained 99.34% of the total variability. The rst axis contributed for 94.76% and the second 4.58%. The projection of the honey production areas in the factorial plan de ned by 1 and 2 revealed that the beekeepers from Segbana and Banikoara used the insecticides Thalis and LambdaSuper. The beekeepers from Gogounou used the insecticides Emacot and Cypercal. The beekeepers from Kandi used Cotonix (Fig. 3). Correspondence analysis (CA) performed to show the relationship between the treated crops and used herbicides types revealed that the rst factor plane formed by axes 1 and 2 explained 68.45% of the total variability. The rst axis contributed for 45.54% and the second 22.91%. The herbicides Kabasate, Adwuma and Grasskiller were used in the elds of soy and millet. The herbicides Callifor, Butaforce and Kalach were used in the elds of cowpea, yam and sorghum. Malik, Herbextra, Atraforce, Butaplus were used in the elds of maize. Cottonex and Glyphader were used in the eld of cotton. Paraeforce was used in the all elds (Fig. 4). Correspondence Analysis (CA) performed to show the relationship between the treated crops and used insecticides types revealed that the rst factor plane formed by axes 1 and 2 explained 89.41% of the total variability. The rst axis contributed for 66.17% and the second 23.24%. The insecticides LambdaSuper and Cypercal were used in the elds of soy cowpea and millet. The insecticides Cotonix and Thalis were used in the elds of cotton. Emacot was used in the elds of maize and sorghum. No product was used in the eld of yam (Fig. 5).

Decolonize beehives 5
Farmer's perception on negative effects of pesticide use on bee communities A gradual decrease in bees has been observed by beekeepers regardless of breed. By taking the study area in general, 49% of the farmers found the decrease in the big breed and 40% that of the small breed.
In the same area, some beekeepers mentioned an increase in the population dynamics of bees; 6% for the small breed and 5% for the big breed (Fig. 6). Statistical analyses showed negative effects of pesticide use (insecticide and herbicide) on big honeybee abundances (Estimate = − 2.45; P = 0.0042) and small honeybee abundances (Estimate = − 0.29; P = 0.001). Farmers collected honey (May -November) during a period of the year following the use of agrochemicals (Fig. 7). Given the gradual pollution of this honey production environment, some farmers adopted agroecological practices such as the installation of hives away from elds treated with pesticides, association of crops, biopesticide use and ecological beekeeping for sustainable production of honey (Fig. 8).
In uence of the distance of beehives from cultivated elds on the honey production and Monte Carlo modeling The beekeepers were grouped into two groups according to the proximity of the beehives to treated crops.
The average honey production per beehive in the rst group was 8.08 liters, the second group was 9.75 liters per harvest. The Monte Carlo method showed that the variation in the quantity of honey production with the beehive distance from elds was signi cant (F = 116.7; Df = 98; P = 0.00001; Fig. 9).
In uence of beehive systems on the quantity and taste of honey There was no signi cant effect of the beehive systems types on the quantity of honey produced (P = 0.32; Df = 98). However, there was a signi cant effect of beehive systems types on the honey taste (P < 0.00001; Df = 98). Tukey's test showed a highly signi cant difference between the tastes of honey produced from beehives made with sheet metal and beehives made with therapeutic plant woods. The beehives made with therapeutic wood contained honey with a very appreciable taste. The difference was moderately signi cant between honeys produced from beehives made with therapeutic woods and modern beehives. In addition, there was a low difference between the honey produced from modern beehives and those made from sheet metal ( Fig. 10; Table 3).

Discussion
The present study reveals that beekeeping is an activity mainly practiced by men (97%) and by illiterates.
These observations were similar to those of [18] who observed that honey production in the study area was predominantly male and had a low level of schooling. The average age of beekeepers surveyed was 37.58 ± 1.48 years. This result showed that this activity was more practiced by young people. The interest of these young people in beekeeping was an important asset for training in modern beekeeping, production and intensi cation of production. Beekeepers with one year of experience ranging from 1 to 5 are the most numerous, which shows beekeepers were taking over their business by involving more young people in this income-generating activity.
The farmers use chemical pesticides for the most cultivated crops in Northern Benin. Agrochemicals were heavily applied from June to October while the bees were in the process of collecting nectar for the honey production. This could lead to the high mortality in the bee community as shown in others studies [31,13,32] in which the negative effects of neonicotinoids on bee health has been shown. The honey harvesting activity occurred after the pesticide application. This could affect the quality of honey that could contain some traces of pesticides. In Uganda, a signi cant contamination of honey with insecticides including neonicotinoids, organophosphates, carbamates, organophosphorus, triazines and diacylhydrazines has been observed in contaminated samples [33].
The perception of beekeepers on the effect of pesticides on bees reveals that the beekepers comprehend that chemical pesticides affect bees according to their categories. Herbicides had an indirect effect, while insecticides act directly on them. Indirect effects, especially trophic effects, are related to the feeding of wildlife. By destroying weeds, herbicides affect the animal species that feed on them, even if they are not toxic to these animals. The use of pesticides in market gardening and probably other crops had a negative impact on the honey bee Apis mellifera adansonii [34]. Farmers thought that insecticides increased the bee mortality. These intoxications contribute to the massive depopulation of beehives and a loss of honey harvest [35]. Other studies have shown the presence of fungicides in bees, pollen and honey [36]. Results showed that bees could be contaminated by insecticides containing active ingredients such as cypermethrin, deltamethrin and emamectin. Additionally, these active ingredients had lethal and sublethal effects on forager worker honeybees in experimental conditions [37].
Some beekeepers applied alternatives to try adapt to the risk of bee extinction caused by agricultural pesticides. The forms of adaptation recorded were: keeping the beehives away from the treated elds and reserving a space especially for the bees breeding. These two methods could reduce the disappearance of bees [38]. The major problem is the land availability because of their occupation for the bene t of agriculture. Another method was the biopesticide use. Several plant extracts are used to control crop pests. These plant extracts are also used in combination with biological control methods to control crop pests in Integrated Pest Management (IPM) programs which could enhance the pollinators [39]. The development of agroforestry systems based on melliferous plants species in association with cultivated plants could be a better solution for the ecological and sustainable production of honey. A diversity of agroforestry honey plants exists in the study area [27] and could be an asset for the design of agroforestry systems with crop planting.
The study has also showed the farmer's perception on the decrease in the populations of the two breeds of bees existing in Benin also reported by [40]. About 49% of farmers found the decrease of the big breed (yellow breed) and 40% that of the small breed (black breed). This possible decrease is not only due to the use of pesticides but also the agricultural pressure with the bee biotope destruction [41]. Others factors including pests like Varroa which is known most important pest of bees, decline both the bee and the honey production [42,43]. The in uence of the distance of the apiaries from the elds which was evaluated showed that the effect was signi cant. As a result, beehives set up far from treated elds produce on average the most quantity of honey. Far from the treated elds, bees are less exposed to agrochemicals and can survive for a long time and increase their populations. They have the possibility of looking for nectar on several honey plants without being contaminated. Several factors threaten the honey production such as the lack in maintenance for beehives. Some beekeepers abandon their beehives installed far from the elds for the bene t of agriculture and do not return to beehives until harvest. The quality of the hives could contribute to the taste of honey [44]. In our study area, we have identi ed 3 categories of beehives that could affect the honey taste according to beekeeping producers. Beehives made of wood provide a good taste to honey compared to hives made of sheet metal and modern hives. The woods have therapeutic properties which could be yielded to honey.

Page 12/25
Beekeeping is an activity mainly carried out by cotton farmers in the north of Benin. These farmers use chemical pesticides to treat almost all of the cultivated crops. Herbicides and insecticides are the most widely used pesticides. Chemicals coming mainly from the informal supply networks are more applied during the months when bees collect nectar and pollen. The present study has shown that the pesticides used in the study area can cause a decrease in the two bee breeds encountered in northern Benin. A minority of beekeepers are developing forms of adaptation to the risk of the bees disappearing. It is therefore necessary to reduce the damage of pesticides on bees using biopesticides which are not necessarily less harmful for bees.

Declarations
Ethics approval and consent to participate No ethical approval was needed for this study. Prior to data collection, participants gave oral consent to participate in the study.

Competing interests
The authors declare that they have no competing interests.

Availability of data and materials
Data generated during this study are available from the corresponding author.  Correspondence analysis (CA) performed to show the relationship between honey production areas and herbicide types