Pesticide exposure can be harmful to insect pollinators and the ecosystem services they provide. As pesticide guidelines warn against applying such products when pollinators are active, it is important to determine how pollinator activity changes with the time of day and when the most appropriate time is to spray these chemicals. We walked transects from sunrise to sunset in oilseed rape (Brassica napus L.) fields in Ireland to capture the abundance of honeybees, bumblebees, solitary bees, and hoverflies across daylight hours. We also recorded the activity of representative species from the three bee groups at their nests across similar time periods to compare with field observations. Peak pollinator abundance was in the mid-afternoon with fewer individuals in the early morning and late evening for all groups. At the nest, we observed patterns of activity that broadly reflected field abundance but indicated that bees are active earlier and later than those observed on the crop. However, there were differences between pollinator groups. Overall, honeybee and solitary bee abundance and activity were found to peak in the middle of the day, while bumblebee abundance and activity were more consistent throughout daylight hours. Hoverflies were relatively abundant in the morning and increased in number towards the late afternoon and early evening. Our results confirm current recommendations that pesticide application should be avoided in the middle of the day when pollinators are most active. However, the diversity of responses within and between pollinator groups to time of day should be accounted for when shaping guidelines, and clearly defining optimal pesticide application timings for end users is difficult and needs further consideration as it will vary between regions and crops. Further research should also explore how time impacts pesticide efficacy and exposure of pollinators to residues post-application to allow full evaluation of how practical and beneficial timing of application may be when aiming to protect pollinators from pesticide exposure.

Pollinator abundance on oilseed rape

We observed three fields of mass-flowering spring-sown oilseed rape (canola; Brassica napus L.) in Co. Carlow, Ireland.

We visited each field six times, twice from 0600-1100, twice from 1100-1600, and twice from 1600-2200. Visits began when oilseed rape reached Growth Stage 6 (flowering) and was conducted in the best weather possible (i.e., dry, clear skies, still, warm) to replicate the recommended conditions for pesticide application which coincide with optimal conditions for pollinator abundance. All observations took place from 15 June to 07 July 2021.

On each visit, we walked three 15-minute transects (with 20 minutes in between) in two-hour blocks. Transects were walked at a slow, steady pace along field tramlines. Tramlines nearest the sides of the fields were designated ‘edge’ and the tramlines past 20 meters into the fields were designated ‘centre;’ composition of a pollinator assemblage can differ between the ‘centre’ and ‘edge’ of fields; an even number of transects in both locations were walked to get a representative measure of the area. Every insect pollinator visiting an oilseed rape flower within a one-meter radius of the observer was recorded. To establish the effect of time on pollen gathering, bees with visible pollen loads were distinguished from those without. For modelling purposes, we grouped insects into four main groups - honeybees, bumblebees, solitary bees, and hoverflies - as this represents the main functional groups of pollinators in the study region and is best suited to the resolution of our data.

Data collection and processing for nest data

To compare the effect of time on pollinators in oilseed rape crops with activity at the nest, we also recorded honeybee (Apis mellifera) and bumblebee (Bombus terrestris audax) activity at their nest entrances. Nest arrivals were used as the measure of activity since the number of arrivals is representative of the number of foraging trips. Activity of these two bee species was recorded at the UCD Rosemount Environmental Research Station from 01 to 31 July, for 24 hours a day in all weather conditions.

A. mellifera and B. terrestris were tagged with Radio Frequency Identification (RFID) tags (mic3–TAG 64-bit RO, BeeID system; Microsensys GmbH, Erfurt, Germany); we used queen marking cages to stick tags to their thorax with Uhu Endfest Plus Adhesive (UHU GmbH & Co. KG, Bühl/Baden, Germany). Bee activity was recorded 24 hours a day except when RFID console data were downloaded.

Three A. mellifera hives were provided by Rosemount beekeepers. A. mellifera returning to the hive entrance were captured and tagged. We only captured those bees observed returning with a visible pollen load on their corbiculae to ensure they were foragers. Tagging of A. mellifera continued throughout the study duration with 199 A. mellifera tags being read in total.

Three research  colonies of B. terrestris were obtained from Biobest Group NV (Westerlo, BE) via Agralan Ltd. (Wiltshire, UK) with each colony containing a queen and approximately 50 workers upon delivery. Before outdoor placement, every worker from each B. terrestris colony was tagged. Afterwards, colonies were placed outdoors and allowed to forage freely. On subsequent days, workers observed returning to the colonies without a tag were captured and tagged, and 182 B. terrestris tags were read in total. 

Data from every day of the experiment, regardless of weather conditions, were included to capture an overall pattern of activity for both species. Legitimate foraging trips were specified by setting a threshold of at least 5 minutes between arrivals of the same tag. A 5-minute threshold has been used as the minimum for a ‘foraging bout’ length in honeybees, as orientation flight length rarely exceeds 5 minutes; a 5-minute flight has been used for B. terrestris as it is estimated to accurately represent a foraging bout. The number of arrivals was then aggregated into 15-minute intervals. Any intervals during which bees were tagged were removed, as the colonies were disturbed. Although 600 A. mellifera were tagged across all three colonies (approximately 200 per colony), due to misreads by the RFID system some tags were not detected, or detection frequency was inconsistent. Only 199 honeybee tags were read across all colonies which resulted in low readings of activity overall. However, data reflected patterns observed in the field and are thus assumed to be a representative subset. 

We supplemented these data by also recording solitary bee (Osmia bicornis) nest activity from 29 to 31 July 2021. As an RFID system was not available to record solitary bee nest entrances, we used video cameras to record their activity from 0600 to 2200 on warm and sunny days which reflected the same timespan observed in the oilseed rape study.

O. bicornis activity was recorded in two flight cages in the UCD Rosemount Experimental Research Station from 29 to 31 July 2021 (average sunrise at 0546 and sunset at 2128). O. bicornis females emerged from commercially available cocoons were marked and placed in each cage following mating and had access to flowering potted Phacelia tanacetifolia for forage as well as mud and water for nest cavity construction. Bees were placed in the cages five weeks before observations as they were used as the control group in a prior experiment. Video was recorded of nest activity in both cages from 0600 to 2200h for three days, apart from times when the cameras were switched out. Seven bees, five in one cage and two in another, were recorded. Since the bees could enter multiple nest cavities before locating their own, a threshold of 30 seconds between nest arrivals was used to ensure the time spent away from the nest was more likely for foraging than return orientation.

Weather recording

For all experiments, we recorded temperature, relative humidity, solar radiation, precipitation, and wind speed at 30-second intervals throughout each observation day using a HOBO Micro Station Logger from the Onset Computer Corporation. In the field, we placed the logger at the edge of the oilseed rape field, while when recording A. mellifera and B. terrestris nest entrance activity, we situated the logger between the two sets of colonies. When observing O. bicornis we placed the logger outside the flight cages. Note that we only had one weather station; weather data recorded was only able to be recorded at one site at a time.