Insecticides may kill not only the target species but also other invertebrates on which birds rely on for their food. Additionally, herbicides are designed to manage weed species which also kill many other plant species in fields, including the essentially beneficial species, which give both shelter and food for the members of wildlife. Amphibians are now considered as the foremost threatened and rapidly decreasing species on earth (Brühl et al., 2013).

In Ethiopia, due to the intensification of agricultural activities, inputs like pesticides and fertilizers use are increasing at an alarming rate (Mengistie 2016; Amera and Abate, 2008). The influence of pesticides on the environment consists of the harmful and toxic effects of pesticides to non-target plants and animals. Residues of pesticides may contaminate all the environmental compartments including soil, air and surface or ground water (Gupta et al., 2003; Konstantinou et al., 2006). Thus, the global use of pesticides may have contributed to environmental degradation and depletion of biodiversity negatively impacting the wellbeing of the global flora and fauna (Al-Shaalan et al., 2019; Naik and Wanganeo, 2014; Zhang et al., 2011; Ali and Jain, 1998).

Among the widely impacted non-target animals by pesticide application, bees are the main concern. There are 20,000 species of bees on earth, pollinating 90% of the 107 main crops in the world. Bee numbers have declined dramatically in recent years (Sheridan, 2017). It is estimated that 75% of the world's honeybees have been found to have traces of bee-damaging insecticides, particularly to neonicotinoids, such as acetamiprid, clothianidin, imidacloprid, thiacloprid, and thiamethoxam (Sheridan, 2017). 

Ethiopia is known for its immense variety of agro-climate and biodiversity conditions that favored the life of diversified honeybee flora and large numbers of colonies of honeybees (Nuru, 2007). Thus, beekeeping may be a long-standing tradition in Ethiopia's rural communities (Yirga and Teferi, 2010). Being  an  export  commodity, it has significant contribution for household wealth and poverty reduction as well as the economy of the country. Central Statistical Agency (2011) report shows that Ethiopia is among the four largest beeswax producing countries. For example, the honey export in 2010/2011 production year was estimated to be 620,101 kg from which the country on average generated 420 million Ethiopian Birr on annual basis from the sale of honey. It is estimated that the overall honeybee colonies population in the country is estimated to be10 million, of which 7.5 million are tamed, while the remaining are from wild colonies found in forests (Kenesa, 2018).

However, with the introduction of pesticides in Ethiopia, the poisoning effect of the agro-chemical on honeybees has been increasing over time, where some beekeepers have even lost all their colonies (Kerealem et al., 2009). In connection to this, Melaku et al. (2008) attributed colonial death and absconding with insecticides and herbicides. Chauzat et al. (2006) also showed that improper use of insecticides results in the demise of honeybee. Study conducted by Fikadu (2020) attributed the declining of honeybees’ pollinators with unwise use and practices of pesticides to lack of knowledge of pest, and predators’ management causes the misuse of pesticide.

Pesticides are harmful compounds with a common mode of action, which means that they are primarily engineered to regulate a target group of species by interacting with certain biochemical pathways. Pesticide effect on species can be categorized based on the lethal dose (LD50) values which determine the dose that kills 50% of the exposed animal after a given time of exposure. The sub-lethal effect induced by the chemical to the exposed organisms can also cause other irregularities in their behavioural and physiological activities including stress paralysis or irregular habits without killing them like exposure to neurotoxic insecticides. This effect works for bees as it does for any other organism (Chakrabarti et al., 2015; Zaluski et al., 2015; De Grandi-Hoffman et al., 2013).

Various tools are used to determine risks of agrochemicals to non-target organisms. Among them is an already developed pesticide risk assessment tool for an Ethiopian situation known as Pesticide Risks In the Tropics for Man, Environment and Trade (PRIMET). This tool was developed in collaboration with Wageningen University of the Netherlands, considering specific scenarios in Ethiopia and can be taken as a pioneer in Africa. It can estimate risks of pesticides currently registered or to be registered in Ethiopia for non-target organisms including bees. The software calculates the Exposure Toxicity Ratio (ETR) for each chemical pesticide for both off-crop and in-crop scenarios (Wipfler et al., 2014).

The impact of harmful pesticides on bees and assessment of risks need to be studied in Ethiopia. Even though there are some studies that tried to show the impact of pesticides on bees in Ethiopia, they cannot objectively quantify the level of risks to bees via application of pesticides for the control of other pest incidences. For example, Fikadu (2020) studies pesticides use, practice and its effect on Ethiopian honeybee; nevertheless, the study was based on secondary data and thus the finding was more general. The objective of this study is to assess pesticide application, use and its implications on honey bees.

Target population, sampling design and sample size

The sampling frame for this particular study was rural farmers found in Gudeya Bila wereda. Multi- stage sampling technique was used to select the representative samples. The study area was selected purposefully and carefully so as to represent the wereda in terms of economic, socio-cultural, and physical factors like agro-ecology, and familiarity of the researcher to the study area. Therefore, at the first stage, the rural kebeles of the wereda were stratified by agro-ecology as dega, weyna dega and kolla and then the sample rural kebeles were selected randomly to represent the agro-ecological zones (Figure 1) by lottery method. Households or respondents were selected randomly from the sample kebeles agriculture office lists of farmers engaged in beekeeping. In addition to this, pesticide retailers were included in the sampling for FGD and KII to get information on pesticides selection and use status. Hence, it was appropriate to have a deep understanding of the pesticides use practice, application practice and beekeeping status of the study area.

The sample size of the study was determined or calculated using Taro Yamane sample size determination formulas with household number of sample kebeles, as given in Equation 1.

n = 312 sample sizes used for this study

where n-sample size, N-is number of households of sample kebeles, and e-the precision or sampling error which is 0.005.

As the proportion of respondents/households that is food secure is not known, 0.5 was used as p-value to obtain the sample size (312).

Out of the total respondents involved in the questionnaire survey: respondents were pesticide traders from whom data on pesticides were obtained and 304 respondents were farmer households who were interviewed in relation to pesticide use.

Data analysis

Qualitative and quantitative techniques were used to analyse the study data. Information obtained from key informant interview, focus group discussion and personal observation was analysed qualitatively. The SPSS software version 20 was used to analyse the quantitative data obtained from household survey.

For the risk assessment of in crop and off crop, an oral LD50 and a contact LD50 were available and taken from Pesticide Properties Data Base (PPBD) from which the lowest value was determined. Moreover, pesticide application rate was obtained from Ministry of Agriculture (Ethiopian) Plant Health Regulatory Directorate database (Equations 2 and 3) (Wipfler et al., 2014). The European Union (EU) trigger value of 50 was used, which was established based on empirical research. An assessment of observed bee killings (colony sound effects) for various pesticides and different application rates showed that for sprays a factor of the Exposure Toxicity Ratio (ETR) below 50 is always safe as no field incidents at ETR < 50.  While a trigger value of above 400 for the ETR is considered highly risky for bees. This value is taken as the upper limit of the risk classification as shown for in-crop and off-crop situations of beehives from the pesticide application spots (Wipfler et al., 2014).

ETRin-crop = dose rate/ LD50                                           (2)

ETRoff-crop = (dose rate × drift factor)/ LD50                            (3)

where Low risk: if ETR in-/off-crop situation is < 50, possible risk: if ETR in-/off-crop situation is between 50 and 400, gigh risk: if ETR in-/off-crop situation > 400.

Thus, risks of some frequently used pesticides were assessed using PRIMET software version 1.1.1.1 (Wipfler et al., 2014) and the household survey data as an input.

Depending on the ETR values, decision was made regarding the pesticide  application  using  PRIMET software to determine the risk level of a given chemical (pesticide) to honeybee. Hence, quantitative household survey data were analysed in Statistical Package for Social Sciences (SPSS) and the outputs of SPSS such as frequency of application, time interval for application, and methods of applications were subjected to PRIMET test. Among most frequently used pesticides, insecticides and herbicides were widely used. Insecticides are the more damaging types of pesticides to honeybees (Leska et al., 2021). Therefore, six types of insecticides which are most frequently used in the study area and anywhere in Ethiopia were selected and their risks were analyzed using PRIMET Version 1.1.1. Following Wipfler et al. (2014) method, LD₅₀ name of pesticide chemical, name of the target crop, number of application, application methods, time interval of the application, scenarios of exposure (in-crop or off-crop), category of risk (as chronic or acute risk), pesticide drift factors, rate (that is, concentration of active ingredient per hectare) and target organism were entered into PRIMET software for risk assessment.

The current study showed that almost all farmers (97.4%) did not move their hives to safe place during pesticide application (Table 3). This implies that honeybees are exposed to the damage of agro-chemicals related with application methods and safety measures. This result is supported by Fikadu (2020)’s finding who indicated that majority of Ethiopian beekeepers do not use any control measures for poisoning honey bees with chemical. Likewise, Sánchez-Bayo and Goka (2016) indicated that most of the time, bees are exposed to pollutants by the ingestion of pollen and nectar residues from infected seeds, whether from crops or from weeds across the fields.  It is important to note that bees eat chemicals wherever they go and search for the most fitting flowers that provide ample pollen and nectar. Such crops are also more attractive than others; for example, canola (rape seed oil) yellow flowers, sunflowers, and many of the weeds that grow in and around the crops are more attractive to bees than potato plant flowers (Dötterl and Vereecken, 2010). In converse, farmers apply chemicals to these crops to control pest, insect, disease and weed. The forager bees take pesticide residues in pollen and nectar to their colonies and live inside the beebread and honey for quite some time (Orantes-Bermejo et al., 2010). These residues are then fed to the larvae and the queen as well, that are influenced by the forager bees in comparable ways. Bees also consume water in addition to sugar, to control their temperature (Schmaranzer, 2000). Pesticide compounds in the soil gradually get into the water and emerge in and above the lakes, creeks and rivers in rural areas; they are polluted with a combination of agrochemicals that can eventually be consumed by bees (Belden et al., 2007; Sánchez-Bayo and Goka, 2016; Zhu et al., 2014). Additionally, water pollution from spray applications, particularly from insecticides can impact honey bees. This is very critical for bumblebees and wild bees that prefer to drink from puddles, drainage ditches, rivers and lakes, and they are often eaten by forager bees if these waters are polluted with pesticide residues (Woodrow et al., 1989). Thus, these exposures of honey bees to pesticides cause the collapse of bee colonies.

Risk assessment results of selected pesticides using PRIMET software

According to FGDs and KII, carbaryl, chlorpyrifos, diazinon, fipronil, malathion and profenofos are among the most frequently used insecticide in the study area. Other similar works also confirmed the frequent use of these pesticides in Ethiopia (Teklu et al., 2021). Of these investigated pesticides that are risky to bees, carbaryl and chlorpyrifos are mostly used for the treatment of maize stalk borer. The risk assessment of carbaryl and chlorpyrifos using PRIMET software in this study area revealed 9107 and 3254 Exposure Toxicity Ratio (ETR)  value respectively for carbaryl and chlorpyrifos in-crop scenario (Table 4). This showed that both insecticides are highly risky to honey bees. Similarly, assessment in United Kingdom by Mineau et al. (2008) revealed 50% probability of bee mortality at a trigger value of 400 for the ETR for numerous pesticides at different application rates.

Likewise, for off-crop scenario, the ETR value was 255 for carbaryl and 91.12 for chlorpyrifos sprayed on maize. This implies that there is possible risk if expected from both insecticides (Table 4). Pesticide exposure can have a sizable impact on the nutritional composition of royal jelly produced by honey bees and as a result can influence queen development. Oral exposure to pesticides in adult workers has been shown to influence nurse bee glandular physiology and could therefore impact royal jelly production (Böhme et al., 2018). As shown in Tables 2 and 3, the study area farmers do not displace (move) beehives and select appropriate time during pesticide applicant. These practices would expose honey bees to pesticide poisoning. Both pesticides are used foliar in controlling crop pest and have a relatively high toxicity to bees compared to other pesticides (Johnson et al. 2010). Worker honey bees can forage in range up to 12 km around hive and, therefore, are frequently exposed to a dispersal of pesticide residues present in water, nectar and pollen (Mullin et al., 2010).

Diazinon is an insecticide registered to treat maize and sorghum stalk borers and armyworm in Ethiopia. In the study area, it was mainly used to treat maize and sorghum stalk borer. The ETR value of diazinon for in-crop scenario was 6667 and this value shows that diazinon is highly risky to honey bees (Table 4). For off-crop scenario the value of ETR was 186.7 and this value indicated diazinon can be classified in possible risk category. As a matter of fact, diazinon was known to be highly toxic to terrestrial invertebrates, bees and other beneficial insects following acute contact exposure, where acute LD50 for bees was 0.22 μg/one bee as per University of Hertfordshire (2013). In general, the toxicity of insecticides to honey bees increased with increase in the exposure time.

Fipronil is among the insecticides used mainly to treat termites in rice and aphids on cabbage. Fipronil ETR is 12×E4 or 120,000  and this value indicates fipronil was highly risky to honey bees for in-crop scenario (Table 4). For off-crop scenario, the ETR value of fipronil is 335.5 which can be classified under possible risk for honey bee. According to Narahashi et al. (2010), fipronil has an antagonistic action on gamma amino butanoic acid (GABA) neurotransmitters and glutamate-activated chloride channels (GluCls). Therefore, this pesticide/ insecticide can cause interactive changes in bees that embrace agitation, spasms, tremors, and paralysis (Zaluski et al., 2015). Fipronil is more noxious in sublethal doses, spoiling the motor activity of bees. Experimental exposure to dietary fipronil caused dose-dependent reductions in the longevity (days of exposure survived)  of adult honey bees and fipronil can be lethal to honey bees in dietary exposures to the trace residues that typify those in nectar and pollen from treated crops (Mullin et al., 2010). Including fipronil, all insecticides assessed as indicator by PRIMET software were risky to honey bee. Therefore, protection measures must be taken to keep honey bees from pesticides poisoning during or after application. 

Malathion is one of the most frequently used pesticides and is formulated locally by Adamitulu Pesticide processing company of Ethiopia in addition to imported ones. According to FGDs and KIIs, Malathion, in form of 50% EC formulation type, was mostly used in the study area to treat maize and sorghum stalk borer. For the in-crop scenario, the ETR value of malathion was 4688 which can be categorized as highly risky to honey bee (Table 4). Concerning off-crop scenario, the ETR of malathion was 131.3 that can be classified under possible risk to honey bee. As indicated earlier, bees eat chemicals including malathion wherever they go and search for the most fitting flowers that provide ample pollen and nectar.

Profenofos is another pesticide categorized under organophosphate chemical group. As shown in Table 4, the ETR values of profenofos for in-crop and off-crop scenario were 7579 and 212.2 respectively and thus categorized under high and possible risk to honey bees in the respective order.  From the above PRIMET output as an ETR value showed that all (six) pesticides types namely: carbaryl, chlorpyrifos, diazinon, malathion, fipronil and profonefos are highly and possibly risky to honey bees for in-crop and off-crop scenarios considering each chemical (Table 4).

Other exposure routes of bees to pesticides come from water pollution due to drift from spray applications, particularly from insecticides (Woods et al., 2003). Honey bees, bumblebees and wild bees prefer to drink from puddles, drainage ditches, rivers and lakes, and they are often eaten by forager bees if the waters are polluted with pesticide residues. These facts indicate that when bees are exposed to pesticides their colonies are under risk of damage by pesticides poisoning. Wild bees (Osmia bicornis) exposed to thiamethoxam and clothianidin sub lethal levels had their reproductive performance reduced by 50% (Sandrock et al., 2014a), while honey bee queens experienced exceptionally high supersede rates (60%) (Sandrock et al., 2014b); colonies of bumble bees (Bombus terrestris) exposed to thiamethoxam sub lethal levels did not perform and produced 85% (Whitehorn et al., 2012). In forager bees, sub lethal doses of neonicotinoid insecticides often induce disorientation in the state of mind (Decourtye and Devillers, 2009). Recent global decline on pollinators including honey bees has been reported owing to several factors and unwise use and practices of pesticides honey bee productivity is affected by the indiscriminate use of agrochemicals, lack of knowledge, pest and predators (Fikadu, 2020).

In this study, six pesticides were identified to be used frequently by farmers that are considered to be used frequently and at the same time risky to bees namely, carbaryl, chlorpyrifos, diazinon, malathion, fipronil and profonefos. Farmers’ awareness in considering the presence of bees in pesticide applications and caring for bee hives that are present in/off-crop situation is found to be less. Moreover, concern to protect human and environmental health in general is minimal. The dominant type of bee hive present in the study area is traditional followed by transitional one. Thus it is common to see bee hives hanged in trees in the middle of farming areas. Thus, in-crop risks form application of pesticides is found to be high. This is evidenced by farmers reporting to observe serious physical, behavioural and physiological changes of the honey bees after application of pesticides in the area.

Risk assessment for the selected six pesticides revealed that all the six pesticides pose high risk to bees if applied in an in-crop situation. Possible risks are also indicated for bee hives available in an off-crop situation. Therefore, farmers and retailers need to be constantly informed about hazards of theses pesticides so that the current haphazard handling of pesticides is improved. Concerning application practice, there must be legal bindings or directives that obligate pesticide users and farmers to notify beekeepers of the neighbouring areas before application of these pesticides. Protection measures like removing or covering beehives during application, avoiding application of pesticide on time of flowering of either to main or fodder crops, application only after sunsets need to be adopted for honey bees risky pesticides. Pesticide registration and control department of Ministry of Agriculture should notify and regulate the registrants or agents of pesticides traders to indicate on the labels clearly the toxicity status of pesticides. Moreover, it is important to translate labels to languages appropriate to local situation.

The authors declare no financial, commercial and legal conflict of interest directly or indirectly related to this published article.

The authors thank the Addis Ababa University and the Ministry of Agriculture of Ethiopia for the successful completion of this study.