Evaluation of Canadian Bee Mortalities Coinciding with Corn Planting in Spring 2012

Transcription

1 HEALTH CANADA SANTÉ CANADA Evaluation of Canadian Bee Mortalities Coinciding with Corn Planting in Spring 2012 Category X.3; Submission Number /13/2013

2 Table of Contents Table of Contents... 1 List of Tables... 3 List of Figures... 4 Evaluation Summary... 6 Section 1 Introduction... 9 Section 2 Description of 2012 Canadian Mortality Reports Overview Information Provided in the Reports Ontario Quebec Section 3 Exposure Pesticide residue Analysis Ontario Quebec Summary of Analytical Results for Ontario and Quebec 2012 Honey bee Morality Use pattern Location of Honey Bee Mortalities Quebec Ontario Section 4 Toxicity Honey Bee Toxicity of the detected pesticides Mode of Action for Honey Bee Toxicity Section 5 Consideration of residue levels as related to toxicity Section 6 Additional Factors Considered Corn Production in Ontario Seed Corn Supply and Distribution In Ontario Machinery and planting processes for corn Vacuum seeders (pneumatic seeder) Positive air pressure seeders Air drills Finger planter (Mechanical Meter Planter) Gravity seeder (Drill Box) Use of Talc or Graphite in planting of seeds Dust Generation from treated seeds Weather conditions at time of the honey bee mortalities Temperature Precipitation Soil Moisture Wind speed Bee Biology and General Health Snapshot of Honey Bee Health in Ontario and in Canada Page 1 of 112

3 6.5.2 Overwintering Losses Environmental Cues on Honey Bee Populations Honey bee colony health inspections conducted by OMAF and MRA Pesticide Poisoning Effects and Symptoms in Honey Bees Discussion and Conclusions Ontario Bee Incident Follow up Inspection Program Section 7 Other Bee Mortality Reports Previously Evaluated Canadian Honey Bee Mortality Incidents Previous Incidents Occurring Outside of Canada Journal Articles Section 8 Conclusions Section 9 Mitigation Measures Section 10 References Appendix 1 Analytical Methods A 1.1 PMRA Analytical Method A Active Ingredients Analyzed, Reporting Limits and Detections A Results of individual Honey bee Samples from Ontario Analyzed by the PMRA laboratory A 1.2 MAPAQ LOD and LOQ Appendix Ontario Bee Mortality Follow up Inspection Program A 2.1 Objectives A 2.2 Background A 2.3 Program Delivery A 2.4 Results Page 2 of 112

4 List of Tables Table 1: Summary of Honey Bee Mortality that Occurred in Canada in Table 2: Number of yards and hives showing mortality as reported by beekeeper Table 3: Summary of Active Ingredients detected in Honey bee samples collected in Ontario in spring Table 4: Summary of number of samples with detections by beekeeper Table 5: Concentrations of Pesticides Detected in Vegetation Samples Collected at the site of the Bee Mortality Table 6: Results of Neonicotinoid analysis of seed planted near bee mortalities Table 7: Residues of Pesticides Detected in Honey bee Samples collected in Quebec Table 8: Summary of Honey Bee Mortality that Occurred in Canada in Table 9: The acute 48 hr honey bee contact and oral toxicity endpoints (LD 50 and NOEL) in ug ai/bee and ppm for selected active ingredients detected in the Ontario and Quebec honey bee samples Table 10: The Mode of Action of the Pesticide Active Ingredients that are Toxic to Honey bee and Detected in the Ontario and Quebec Samples Table 11: Sources of Variability in honey bee residue levels Table 12: The 2012 mean temperature and the Normal mean temperature for 4 cities across Ontario Table 13: Comparison between historical April average wind speeds and April 2012 average wind speeds for three cities in Ontario Table 14: Summary of OMAF and MRA health inspection reports for beekeepers who reported mortality events over the period of April to August Table 15: The level of Pesticide residues detected in Québec honey bee samples Table 16: Clothianidin Honey bee mortality reports from the US EPA. Information was downloaded from US EPA EIIS database (February 2013) Table 17: Summary of Honey Bee Incidents that have occurred outside of Canada Table 18: Active ingredient analyzed in the Ontario and Alberta honey bee mortality samples Table 19: Active Ingredients detected in the Ontario Honey Bee Samples Table 20: The Limit of Detection and Limit of Quantitation for the laboratory method from MAPAQ Page 3 of 112

5 List of Figures Figure 1: The Ontario Sample collection locations Figure 2: The Montérégie Region of Québec ( (The red star indicates the approximate location of St Dominique.) Figure 3: The Major Corn Growing Regions of Ontario and Québec ( 26 Figure 4: The location of the honey bee mortalities that were reported in spring Figure 5: The location of the honey bee mortalities relative to all the registered honey bee yard locations in Ontario (This map was produced in October 2012 by OMAF and MRA and does not include all of the locations where honey bee mortality was observed as some reports were received after this map was generated) Figure 6: The honey bee mortalities and the Cash Crop Area (Soybeans, corn and wheat rotation Figure 7: Corn and Fruit grower location in Ontario for 2012 (provided by Agricorp) Figure 8: The Major Fruit Producing Areas of Ontario (OMAF and MRA Fact Sheet What you should know about fruit production in Ontario htm) Figure 9: Map of Hectares of Corn grown in Ontario per county 2011 Census of Agriculture Statistics Canada ( 31 Figure 10: The relative proportion of bee yards affected per county Figure 11: Detections of Clothianidin in Dead Honey Bee Samples Figure 12: Detections of Fluvalinate in Dead Honey Bee Samples Figure 13: Detection of Coumaphos in Dead Honey bee samples Figure 14: Detection of Iprodione in Dead Honey bee samples Figure 15: Detection of Permethrin in Dead Honey bee samples Figure 16: Detection of phosmet in Dead Honey bee samples Figure 17: Location of detection of Acetamiprid in Dead Honey Bee Samples Figure 18: Acute Contact and Oral Toxicity Values (LD 50 ) of Selected Pesticide Active Ingredients Detected in Honey Bee Samples from Ontario and Quebec (Atkins et al criteria for bee toxicity grey area is highly toxic, blue area is moderately toxic, white area is relatively non toxic) Figure 19: An example of a Vacuum seeder Figure 20: An example of a Vacuum seeder Figure 21: National March 2012 Monthly Mean Temperature Difference from Normal 50 Figure 22: National April 2012 Monthly Mean Temperature Difference from Normal Figure 23: National May 2012 Monthly Mean Temperature Difference from Normal Figure 24: National June 2012 Monthly Mean Temperature Difference from Normal Figure 25: Ontario region percent of average precipitation ( ) for March Figure 26: Ontario region percent of average precipitation ( ) for April Figure 27: Ontario region percent of average precipitation ( ) for May Figure 28: Ontario region percent of average precipitation ( ) for June Page 4 of 112

6 Figure 29: Quebec region percent of average precipitation ( ) for March Figure 30: Quebec region percent of average precipitation ( ) for April Figure 31: Quebec region percent of average precipitation ( ) for May Figure 32: Quebec region percent of average precipitation ( ) for June Figure 33: Soil moisture for the month of April 2012; red indicates <5% soil moisture while white indicates >25% soil moisture Figure 34: Soil moisture for the month of May 2012; red indicates <5% soil moisture while white indicates >25% soil moisture Figure 35: Soil moisture for the month of June 2012; red indicates <5% soil moisture while white indicates >25% soil moisture Figure 36: Soil moisture for the month of July 2012; red indicates <5% soil moisture while white indicates >25% soil moisture Figure 37: Wind speed for spring 2012 in Sarnia, Ontario. Average daily daytime wind speed and 80 th percentile morning (6am 12pm) and afternoon (12pm 6pm) wind speeds represented by different coloured lines Figure 38: Wind speed for spring 2012 in London, Ontario. Average daily daytime wind speed and 80 th percentile morning (6am 12pm) and afternoon (12pm 6pm) wind speeds represented by different coloured lines Figure 39: Wind speed for spring 2012 in Toronto, Ontario. Average daily daytime wind speed and 80 th percentile morning (6am 12pm) and afternoon (12pm 6pm) wind speeds represented by different coloured lines Figure 40: Percentage above historical weekly daytime average wind speeds for the windiest days in London, Ontario in spring Average daily daytime wind speed data for 2012 is compared to historical weekly daytime average wind speeds from Data source is Environment Canada Figure 41: The Montérégie Region of Québec ( 81 Page 5 of 112

7 Evaluation Summary In the spring and summer of 2012, Health Canada s Pest Management Regulatory Agency (PMRA) received a significant number of honey bee mortality reports from the provinces of Alberta, Manitoba, Saskatchewan, Nova Scotia Quebec and Ontario. A portion of these mortalities were determined to be associated with spray drift, however, an unusually high number of reports of honey bee mortalities were received from beekeepers in corn growing regions of Ontario and Quebec. The majority of reports were from southern Ontario, involving over 40 beekeepers and 240 different bee yard locations. Additionally, one report was received from Quebec involving eight bee yards. Timing and location of these honey bee mortalities appeared to coincide with planting corn seed treated with insecticides. An evaluation was undertaken to assess whether pesticides may have contributed to the mortalities and whether regulatory action was required. This evaluation focussed only on pollinator mortalities that coincided with planting treated corn. To evaluate the role that pesticides may have played in the Ontario bee losses, Health Canada, supported by the Ontario Ministry of Environment (MOE) and the Ontario Ministry of Agriculture, Food and Rural Affairs (OMAF and MRA), collected samples for pesticide residue analysis, as well as information on the effects observed, bee health, and agricultural activities in the vicinity of affected bee yards. Affected Ontario beekeepers reported varying levels of mortalities and other symptoms consistent with pesticide exposure (twitching, unable to fly, extended proboscis). It was commonly reported that the bees were foraging at the time of the incidents and that the strongest colonies were the most affected, having the largest number of dead and dying bees, which were often observed to have pollen on their legs. Many of the beekeepers monitored their affected hives through the season and reported ongoing effects, including lack of recovery and colony build up and lack of honey production. Effects on queens were also reported, including queen mortality, high supersedure (replacement of the queen), and poor egg laying resulting in spotty brood. Additionally, some beekeepers noted drone mortality and brood removal from the colonies. Some beekeepers reported colony recovery after varying time periods (days to weeks). Prior to the mortality there were indications that the bees were healthy. Most of the beekeepers reported overwinter losses below 15%, the level generally considered to be acceptable and sustainable by most apiculturists. The 2012 province wide overwinter loss reported by OMAF and MRA was 12%. This is the lowest Ontario overwintering loss in the last 6 years (range 20% to 43%, as reported by OMAF and MRA). Canada wide, the overwintering loss in 2012 was 15%, also the lowest in the past 6 years. Page 6 of 112

8 Weather conditions in the areas where beekeepers were affected were unusual in spring 2012, particularly in Ontario. It was warmer and drier than normal as well being windy in April. OMAF and MRA reports indicate corn planting began two to three weeks ahead of schedule in Ontario, which coincided with the first honey bee mortalities. The unusual weather conditions may have been a contributing factor to the high number of mortality incidents. Corn planting was early, the bees overwintered well and began to increase hive populations early, and in many cases the bees were out foraging. As well, dry windy conditions could have facilitated exposure to bees if dust travelled further afield than would normally be the case. In almost all cases, there was evidence of corn planting near affected bee yards. Some affected beekeepers observed corn planting near their affected hives. Information collected from growers confirmed large areas of corn planted near these yards, and that negative pressure (vacuum) planters and talc seed flow lubricants were used. Information from OMAF and MRA and Agricorp confirmed a correlation between the bee mortalities and location of corn growers in Ontario. The reported honey bee mortality in Quebec was also located in a corn growing region. Residue analysis was conducted to determine whether bees were exposed to the insecticides used on treated corn seeds. Samples of affected bees, from many incident locations, were analyzed for pesticide residues by the PMRA Laboratory Services or Ministère de l Agriculture, Pêcheries et Alimentation Québec. Clothianidin was detected in approximately 70% of the samples analyzed in Ontario and clothianidin and thiamethoxam were detected in the samples analyzed from Quebec. On a bee yard basis, these residues were detected in approximately 80% of the bee yards where dead bee samples were collected and analysed. Samples of unaffected bees were also analysed and clothianidin was only detected in one sample at very low levels. Corn seed in Ontario and Quebec is treated in approximately equal quantities with either clothianidin or thiamethoxam. Since thiamethoxam is converted to clothianidin, the detection of clothiainidin in dead bees could indicate exposure to either clothianidin or thiamethoxam. Additional pesticides were detected in some affected honey bee samples, including acetamiprid, coumaphos, fluvalinate, permethrin, phosmet and thiabendazole. However, these pesticides were detected only in a small number of samples or in localized areas, whereas clothianidin was detected across all areas of reported honey bee mortalities. Acetamiprid, fluvalinate and permethrin were also detected in unaffected honey bee samples. With the exception of phosmet, which is toxic to honey bees and was detected at high levels in samples collected close to apple orchards where phosmet is commonly used, it was considered unlikely that these pesticides contributed significantly to the honey bee mortalities. Page 7 of 112

9 The information evaluated suggests that planting of corn seeds treated with the nitro guanidine insecticides clothianidin and/or thiamethoxam contributed to the majority of the bee mortalities that occurred in corn growing regions of Ontario and Quebec in Spring The likely route of exposure was insecticide contaminated dust generated during the planting of treated corn seed. The unusual weather conditions in the spring of 2012 were likely also a contributing factor. Measures have been implemented to reduce honey bee exposure to dust generated during planting of treated corn seed, including communication of best practices to reduce the exposure of honey bees, labelling of treated seed, a treated seed dust standard, and development of technical solutions to reduce dust, including developments in the areas of seed coating quality, seed flow lubricants, planting equipment, and disposal of treated seed bags. Please refer to Pollinator Protection: Reducing Risk from Treated Seed ( sc.gc.ca/cps spc/pubs/pest/_factfiche/pollinator protection pollinisateurs/index eng.php) for details. Additionally, the nitro guanidine neonicotinoids have been placed under re evaluation (REV ) and further regulatory action will be taken if required. Page 8 of 112

10 Section 1 Introduction Pest control products are only registered by Health Canada's Pest Management Regulatory Agency (PMRA) for use if there is reasonable certainty that no harm to human health or the environment will result from exposure to, or use of the product as directed on the label. As part of post registration oversight, Health Canada collects incident reporting data under the authority of the Pest Control Products Act. If a pesticide manufacturer receives information about an incident involving one of their products, they are required by law to submit that information to Health Canada. Members of the public may also submit information about an incident directly to Health Canada. It is important to note that the information presented in incident reports reflects the observations and opinion of the person reporting it, and does not include any assessment by Health Canada, nor does it confirm an association between the pesticide and the effects reported. The details reported to the PMRA are posted in a public registry after all personal information is removed. Health Canada evaluates the information reported to assess the likelihood of an association between the suspected pesticide and effects reported. Through this evaluation it is determined if there are potential health or environmental risks associated with a pesticide that necessitate mitigation. The conclusions of the evaluation of serious incident reports are communicated to the public. Between 2009 and 2011 the PMRA has received four reports of bee mortalities that were suspected to be associated with neonicotinoid pesticides and has monitored this issue. In the spring and summer of 2012 the PMRA received an unusually high number of honey bee mortality reports from Ontario, as well as additional reports from across Canada. Because of the large number of honey bee mortalities being reported in Ontario, Health Canada determined it was necessary to go outside the normal mandate of the Incident Reporting Program and examine the reported mortalities with the assistance of the Ontario Ministry of the Environment and the Ontario Ministry of Agriculture & Food and the Ministry of Rural Affairs to obtain additional information. Inspections continued throughout the season in order to gather further information about the mortalities, to determine whether pesticide exposure contributed to the mortalities, and whether mitigation measures were required. Page 9 of 112

11 These inspections included the following: Interviews or site inspections with all beekeepers Collection of samples from a number of affected bee yards (including dead honey bees, unaffected honey bees, vegetation, bee collected pollen). Inspectors from Health Canada and/or Ontario Ministry of the Environment collected samples. In some cases beekeepers collected samples. Sample collection information was considered in the evaluation. Pictures of many affected bee yards Detailed agriculture survey of selected affected bee yards. This included discussions with growers located within a 2 km radius of six affected bee yards to determine the agricultural practices in the area, and a survey of all fields located within a 5 km radius surrounding seven affected bee yards. Many follow up discussions and/or site visits occurred with beekeepers throughout the bee keeping season. The purpose of this document is to evaluate the honey bee mortality incidents in order to determine if specific pesticides contributed to honey bee mortality and what the potential pesticide exposure routes may have been. Although this document identifies all honey bee mortalities reported to the PMRA during the 2012 season, the focus of the analysis is on those incidents that occurred in Ontario and Quebec and were considered to have a potential link to the planting of corn seed treated with insecticides. The evaluation of other bee mortality incidents, potentially related to pesticide spraying, will be documented separately. Page 10 of 112

12 Section 2 Description of 2012 Canadian Mortality Reports 2.1 Overview Between April and August 2012 over 50 beekeepers, with more than 300 bee yard locations across Canada, reported honey bee mortalities to the PMRA (summarised in Table 1). Reports may have been filed per bee yard or per beekeeper. To facilitate the comparison the information in Table 1 is reported by beekeeper and by bee yard. Hives at which dead bees were observed are considered affected as such, each bee yard with at least one affected hive is classified as affected. More than 240 (41 beekeepers) of these bee mortality locations were situated in Ontario and Quebec and most coincided with the planting of treated corn seeds and potential exposure to dust emitted during planting. Other mortalities (14 beekeepers at 31 locations across the remainder of the country) were potentially related to spray application of pesticides and unlikely to be related to the planting of corn; full evaluations for these mortalities will be completed separately. The remainder of this evaluation focuses on the mortalities that were potentially related to planting of treated corn seeds. Page 11 of 112

13 Table 1: Summary of Honey Bee Mortality that Occurred in Canada in 2012 Location Sub. No. Date # of beekeepers # of bee yards affected # of hives affected Potential link (corn or other) Ontario* Quebec ** Spring >4550 *** Corn July Other April 19, Corn May 18, Other Alberta Saskatchewan May Other June Other unknown Other Other Other Other Other Other Other Other Other Other Other unknown Other Other Unknown Other unknown Other Other Manitoba June 12, Other Nova Scotia July Other July Other Number of bee yards with mortality potentially related to corn planting (Number of beekeepers) [Number of hives] Number of bee yards with mortality unlikely to be related to corn planting (Number of beekeepers) [Number of hives] 248 (41) [>5330] 31 (14) [>2081] * In Ontario there are 3100 registered Beekeepers with colonies managed at 6400 locations/bee yards (2012 provincial apiarist report) ** This is the submission number for the evaluation of the mortalities potentially linked with planting of corn and is linked to all the individual mortality reports submitted for Ontario *** The number of affected hives is unknown for 34 bee yards from 5 beekeepers. If assume that all hives in these yards were affected the total number of hives affected is ~ For many of the honey bee mortalities reported to the PMRA, the reporter did not specify a suspected pest control product; however, given the timing, location and the large cluster of the mortalities, the PMRA is considering all of these mortalities as potentially related to the planting of treated corn seed. Further analysis is conducted in the remainder of the document. Page 12 of 112

14 2.2 Information Provided in the Reports Information reported below is a brief summary of the information that was collated from the reports submitted to the PMRA. It is recognized that the information is variable from one report to the next and is often limited. The interpretation of information provided in the reports is discussed in other sections of this document Ontario During the spring and summer of 2012, the PMRA received honey bee mortality reports for > 4550 colonies from 40 beekeepers involving 240 bee yard locations throughout southern Ontario (Table 2). The PMRA and Health Canada regional inspectors, supported by the Ontario Ministry of Environment (MOE) and the Ontario Ministry of Agriculture, Food and Rural Affairs (OMAF and MRA), collected detailed information to evaluate the role that pesticides may have played in these bee losses. The focus was on mortalities that occurred between April and May of 2012 that coincided with planting of treated corn seed. In addition to honey bee mortality, beekeepers reported other symptoms in the honey bees that were considered indicative of chemical poisoning by the investigators, including the following (wording taken directly from reports): Honey bees lying on their backs, legs up, wiggling, vibrating Wings folded back Staggering around (ataxia) Twitching, can t stand on their feet (tremors) Rolling on their back, losing control of their bodily functions Difficulty flying, dozy, difficulty walking Moving slowly (lethargy) Moving erratically Extended proboscis Rolling out of colony still alive but curled up In some cases, beekeepers reported observing corn being planted around the same time as the honey bee mortalities. Additionally, in many cases, the weather conditions were reported to be dry and windy. In many cases, the affected bees were foraging bees that were returning to the hive with pollen on their legs. Some beekeepers later reported that colonies did not recover from the spring mortalities and that they continued to see various levels of mortality throughout the summer. Some of the reported information provided by the beekeepers indicated that some queen bees were negatively affected. Symptoms observed that indicated queens may have been negatively impacted included queen mortality, spotty brood pattern (queen not performing/laying eggs well), and supersedure (replacement of the queen by a new queen). The high number of mortality reports occurred despite the fact that a Page 13 of 112

15 number of the beekeepers indicated that the honey bee hives had overwintered very well and seemed very healthy at the start of Samples of dead and unaffected honey bees, treated seeds and vegetation samples were collected at some of the locations where honey bee mortality was observed. The results of the analysis are discussed under Section of this document. Table 2: Number of yards and hives showing mortality as reported by beekeeper Beekeeper # of yards showing # of hives in yards # of hives showing mortality mortality >358 (unknown for 9 yards total of 658 hives originally) >25 (only reported for one yard, the remaining yards account for a total of 459 hives) 9 51 >1328 (unknown for 4 yards; assume that these yards contained 28 hives each) >492 (unknown for 4 yards; assume that these yards contained 28 hives each)) >865 (unknown for 1 yard) >753 (unknown for 4 yards total of 94 hives) Unknown (there appears to be 10 hives in yard based on pictures) Unknown (assume 10 hives in yard) & (no dead bees observed, grass was tall around hives, the colonies stopped developing in population and seemed to go backwards) Totals 240 >8057 >4563 (5886 if assume all hives affected in those yards for which this number was not reported) Page 14 of 112

16 2.2.2 Quebec In the spring of 2012, honey bee mortality occurred in St. Dominique, Quebec that coincided with the planting of treated seed was reported to the PMRA. This mortality (sub # ) occurred between April 19, 2012 and April 22, At the time of the mortality, corn seed was being planted in the general area with about 10% of the corn planted by April 19. The beekeeper noted that corn planting took place on April 18 th close to his bee yards. The beekeeper indicated that there were 788 affected colonies in eight different bee yards. The beekeeper observed large numbers of dead bees in front of the affected hives as well as trembling bees. The beekeeper collected samples of dead honey bees for analysis from one of the yards which held a total of 20 colonies and was closest to his home. Unaffected honey bees were also collected. Results of the residue analysis are further discussed under Section At the time of the mortality the weather was very dry, which was unusual for that time of the year. No pit or pond water was available for the bees as a source of drinking water. The beekeeper indicated that there were no dandelion flowers available for the bees and the only flowers available for the honey bees were cherry blossoms. Page 15 of 112

17 Section 3 Exposure Summary of Overall Conclusions for Section 2 The most commonly detected pesticide in the honey bees was clothianidin, which was detected throughout all areas where mortalities were reported. Clothianidin is a known transformation product of thiamethoxam. Both of these pesticides are used as corn seed treatments. The location of the honey bee mortalities overlaps with the corn growing regions of both Ontario and Quebec. Other pesticides detected in the honey bees were found infrequently in localized areas. 3.1 Pesticide residue Analysis Pesticide residue analysis of environmental samples provides an instantaneous, point in time, indication of the level of the chemical in the sample and contributes as part of a line of evidence in determining potential exposure. Various factors can affect the amount of pesticide detected in an environmental sample, which are discussed in Section 5. The analysis of the Ontario samples was conducted by the PMRA Laboratory Services. Information on the residues detected in samples from Quebec was provided by the Ministère de l Agriculture, Pêcheries et Alimentation de Québec (MAPAQ). All sample analysis from Quebec was conducted by MAPAQ Ontario Samples of various matrices (dead honey bees, unaffected honey bees, pollen from bees, vegetation (primarily dandelion), and treated seed) were collected from the Ontario mortality locations and submitted to the PMRA laboratory 1 for analysis using a multi residue method (see Table 18 in Appendix 1 for the active ingredients analyzed and the reporting limits). HC PMRA compliance officers along with the Ontario Ministry of the Environment and the Ontario Ministry of Agriculture and Rural Affairs were instrumental in the collection and submission of all analytical samples. In instances when inspectors could not be present at a site, the beekeepers were instructed on collection techniques and an inspector picked up the frozen samples for submission to the PMRA laboratory. In some cases beekeepers initiated sample collection without instruction from inspectors and 1 The PMRA laboratory usually analyzes samples to determine compliance issues. No compliance issue was identified with these incidents; however, due to the nature of these incidents the PMRA laboratory provided analytical support for the honey bee incident samples collected this spring and summer. Page 16 of 112

18 these samples were also collected from the beekeepers by the inspectors. Figure 1 shows where samples were collected relative to all the bee mortality reported in the province of Ontario. Figure 1: The Ontario Sample collection locations Honey bee samples A total of 151 honey bee samples were collected from 74 affected honey bee yards (representing 27 bee keepers) with mortality in Southern Ontario 2. Of these samples, four samples were unfit for analysis (2 of these samples were from beekeepers with only one sample, therefore resulting in no samples available for analysis from their bee yard) 3. The remaining 147 samples consisted of 20 unaffected honey bee samples, two lethargic dozy bee 4 samples, and 125 dead honey bee samples. These 127 dead or dozy bees represent samples from 72 yards and 25 beekeepers. Of these 25 beekeepers, five beekeepers also had unaffected bees collected from 13 affected bee yards, totalling 20 unaffected bee samples. Bee samples # of samples # of yards # of beekeepers Initial collected Removal of unfit samples With the knowledge that these honey bee mortalities commenced around the same time as corn planting in Southern Ontario, the PMRA laboratory initially focused their attention on a screening of 38 samples for residues of clothianidin, thiamethoxam and 2 Some bee yards had multiple samples collected 3 These samples were delayed in transit and the samples were decomposed 4 These bees will be considered Dead in the analysis that follows Page 17 of 112

19 imidacloprid, which are used to treat corn seeds. These initial 38 samples, as well as the remainder of the 147 samples, were subjected to full multi residues analysis to ensure that other pesticide residues were not overlooked. Each sample contained approximately 2 g of honey bees which is roughly equivalent to 20 individual honey bees. Samples were analyzed as whole honey bees, and therefore included residues on the bee surface, any pollen loads the honey bees were carrying, and ingested pollen or nectar that may have contained residues. Table 3 summarizes the residue levels of the active ingredients detected in the honey bee samples. Details on the individual samples are provided in Table 19, Appendix 1. The majority of the dead bee samples (70%) contained detectable levels of clothianidin. On a bee yard basis 79% (57 of 72 bee yards) of the yards from which samples were analyzed were positive for clothianidin and similarly on a beekeeper basis 80% (20 of 25 beekeepers) of the beekeepers from which samples were analyzed had positive detection of clothianidin. Of the unaffected bee samples, only one sample (5%) had a positive detection of clothianidin. Table 3: Summary of Active Ingredients detected in Honey bee samples collected in Ontario in spring Dead honey bees Unaffected honey bees (Total 127 samples analysed) (Total 20 samples analysed) Active ingredient (LOQ ppm) # samples with positive detection (% of samples) Concentration range (ppm) # of samples with positive detection (% of samples) Concentration range (ppm) Clothianidin (0.001) 93 (73%) Trace (5%) Acetamiprid (0.003) 36 (28%) Trace (20%) Trace 0.02 Phosmet (0.050) 4 (3%) ND Fluvalinate (0.025) 28 (22%) Trace (10%) Coumaphos (0.025) 11 (9%) ND Permethrin (0.025) 1 (0.8%) (5%) Trace Iprodione (0.020) 1 (0.8%) ND Trace detection below LOQ and higher than LOD (see Appendix 1) Although the percent detection of acetamiprid (28%) is relatively high, these samples are limited to 3 beekeepers (12% of the beekeepers) with 90% of the positive detections of acetampirid coming from the samples collected from one beekeeper. Furthermore, acetamiprid was detected in both dead bee samples (28%) and in unaffected bee samples (20%). Similarly, fluvalinate detections (22% of samples) are limited to samples from 4 beekeepers (16% of all beekeepers). In contrast, 20 beekeepers (80% of the beekeepers) that had samples analyzed had positive detections of clothianidin (Table 4). Page 18 of 112

20 Table 4: Summary of number of samples with detections by beekeeper Beekeeper Reference # # of dead bee samples Dead bees Unaffected bees # of samples with detections # of bee yards sampled # of unaffected bee samples # of bee yards sampled # detects clothianidin acetamiprid phosmet fluvalinate Coumaphos permethrin iprodione Dead Unaffected dead unaffected dead unaffected dead unaffected Dead unaffected dead unaffected dead unaffected # of yards % of with samples detects % % % % % % % % ( % 26 1 (6%) 32 (30 T) T) % % % % % % % (T) % % (1 T) % 1 1 (T) % % # of beekeepers with detections % of beekeepers with detections Total Samples % detect T = Trace; There are 25 beekeepers with dead bee samples analyzed; There are 5 beekeepers with unaffected bee samples analysed Page 19 of 112

21 Plants/Vegetation/Pollen In addition to the honey bee samples, nine dandelion/vegetation samples and two honey bee pollen samples were analyzed using multi residues analysis (Table 5). Limited detections of pesticides were observed in the vegetation samples analyzed; however, clothianidin and myclobutanil were detected in one of the pollen samples collected from bees, using pollen traps. The majority of vegetation samples collected and analyzed were taken near the affected hives and may not reflect samples of vegetation in or near the fields being planted. The information suggests, in general, that there were no residues on vegetation in the bee yard or near the hives for those limited locations where vegetation was sampled. Table 5: Concentrations of Pesticides Detected in Vegetation Samples Collected at the site of the Bee Mortality Concentration (ppm) Lab Sample # Beekeeper # Matrix Sample notes Clothianidin Myclobutanil Atrazine Desethyl Atrazine Metolachlor 2012 ON Bee Pollen Collected from Pollen traps ND ND ND ND ND 2012 ON Bee Pollen Collected from Pollen traps ND ND ND 2012 ON Dandelion ND ND Same bee keeper different yards 2012 ON Dandelion ND ND ND ND ND 2012 ON Dandelion Collected in bee yard ND ND ND ND ND 2012 ON Dandelion Collected in bee yard ND ND ND ND ND 2012 ON Winter Wheat 250 m to bee yard ND ND ND ND ND 2012 ON Vegetation Near hives ND ND ND ND ND 2012 ON Dandelion All associated Perimeter of yard ND ND ND ND ND 2012 ON Dandelion with the Roadside ND ND ND ND ND 2012 ON Dandelion same bee Adjacent to planted yard soybean field ND ND ND ND ND Seed coating Six seed samples were analyzed for neonicotinoids to determine if the levels of pesticide in corn seed were within compliance standards (Table 6). Seed samples were collected from corn planting operations near 4 of the affected bee yards. Some seed was collected from the planter and others were collected from the field being planted. The levels of chlothianidin and thiamethoxam in the seed samples were within the range of levels registered for use. Page 20 of 112

22 Table 6: Results of Neonicotinoid analysis of seed planted near bee mortalities Concentration (ppm) Lab Sample # Beekeeper # Notes about seed Clothianidin Thiamethoxam 2012 ON ON ON 0020 A ON 0020 B ON ND ON ND 633 PPST 250 contains Maxim Quattro + Cruiser biological Field corn seed Pioneer P9910XR Treated with PPST 250 (thiamethoxam), Sample collected south of bee yard Collected from Seed Hopper (Case IH Air Seeder) Field corn seed Pioneer 35F44 Treated with PPST 250 (thiamethoxam), Collected from Seed Hopper, John Deere Air Seeder Field corn seed Orange Seed Pioneer treated with PPST 250 (thiamethoxam), Sample collected from seeded field immediately North of bee yard; planted with large Case IH Air Seeder Field corn seed Green seed Dekalb seed treated with clothiandin, Sample collected from seeded field immediately North of bee yard; planted with large Case IH Air Seeder Field corn seed Orange Seed Pioneer treated with PPST 250 (thiamethoxam), Sample collected from seeded field approx 150m North of Bee yard; planted by Air Seeder Pioneer 35F 38 Treated with Cruiser Extreme 250, (Thiamethoxam), Sampled from hopper of finger metre planter, Corn planted near bee yard Overall Ontario Observations Over 200 samples related to the 2012 Ontario spring bee mortalities were collected and analyzed by the PMRA laboratory. These samples included honey bees, vegetation, pollen and treated seeds. Seven different pesticides were detected in the honey bee samples, including the insecticides clothianidin, acetamiprid, phosmet, fluvalinate, coumaphos, and permethrin and the fungicide iprodione. With the exception of clothianidin, all of the detections were in a very small number of samples or in a relatively isolated location. Clothinanidin was detected in 73% of the samples, representing 80% of the beekeepers that had samples analyzed, and spanned the entire area of southern Ontario where samples were collected. On a bee yard basis, clothianidin was detected in 79% of bee yards that had samples analysed (57 of 72 bee yards). A limited number of vegetation samples were collected, the majority of which were collected in or near the affected bee yards and one sample collected near a recently planted soybean field. No insecticide residues were detected on any of these vegetation samples. Seeds collected from seed hoppers or fields planted with corn contained levels of insecticides within the range expected, and are within compliance standards. Page 21 of 112

23 3.1.2 Quebec The analytical information in Table 7 was provided in the incident reports submitted to the PMRA regarding the honey bee mortality that occurred in St Dominique. Dead honey bees were collected from the site where the incident occurred and were submitted for residue analysis. Unaffected bees were also collected and submitted for analysis to serve as controls. The analysis of the honey bee samples was conducted by (Ministère de l Agriculture, des Pêcheries et de l Alimentation Québec (MAPAQ). The analysis was reported to have been carried out using a multi residue method, which is capable of detecting over 200 different pesticides). Residues of clothianidin, thiamethoxam, thiabendazole and coumaphos were detected. Clothianidin residues were found in greater concentrations than the other pesticides. An overview of the use of these pesticides is presented in Section 3.2 of this document. Except for coumaphos, which is used to control honey bee parasites inside the hive, the pesticides detected in the St Dominique samples are associated with corn planting. This is consistent with observations made by the beekeeper indicating that corn was being planted in the general area at the time the bee mortality occurred. Table 7: Residues of Pesticides Detected in Honey bee Samples collected in Quebec IR # Incident Date Honey Bee Sampling date # of bee yards 8 # of hives showing mortality 788 MAPAQ Lab Sample Ref # Clothianidin (ppm) Thiamethoxam (ppm) Coumaphos (ppm) ND Thiabendazole (ppm) The results in this table were determined using a targeted method for which an LOD was not defined. The presence of the pesticide can be confirmed, however, the levels presented may not be precise Summary of Analytical Results for Ontario and Quebec 2012 Honey bee Morality Table 8 summarizes the pesticide residues detected in various matrices collected from Ontario and Quebec honey bee mortality locations. Common active ingredients detected in both provinces were clothianidin and/or thiamethoxam. As clothianidin is a breakdown product of thiamethoxam, it is not possible to know if residues of clothianidin detected resulted from exposure to thiamethoxam or clothianidin. Clothianidin was detected in approximately 70% of the Ontario dead honey bee samples and in the Quebec samples. There was one detection of clothainidin at LOQ in one unaffected bee sample from Ontario. Clothianidin was also detected in one of two pollen samples from Ontario. Thiamethoxam was detected in the Quebec samples at levels lower than the PMRA laboratory LOQ of ppm. Other pesticide residues were also detected infrequently in dead honey bee samples and include acetamiprid, Page 22 of 112

24 ipriodione, phosmet, permethrin, fluvalinate, coumaphos and thiabendazole. Acetamiprid, permethrin, fluvalinate and coumaphos were also identified in unaffected honey bee samples. Myclobutanil was identified in one of two pollen samples and atrazine, desethyl atrazine and metolaclor were identified in one of nine plant samples from Ontario analyzed. Further examination of the relevance of these detections is elaborated in this section. Table 8: Summary of Honey Bee Mortality that Occurred in Canada in 2012 Location Date # beekeepers # of bee yards # of hives affected Residues detected Active Matrix # of samples # of detects Sample Details LOQ/ LOD (ppm) * Min (ppm) Max (ppm) Ontario Quebec Spring 2012 April 19, > Dead bee / trace Clothianidin Unaffected bee / Pollen / Dead bee / trace Acetamiprid Unaffected bee / trace 0.02 Dead bee / ND 3.0 Phosmet Unaffected bee / 0.35 ND ND Dead bee / trace Fluvalinate Unaffected bee / Coumaphos Dead bee / Dead bee / Permethrin Unaffected bee / Trace Iprodione Dead bee / Myclobutanil pollen / Atrazine Plant / Atrazine Desethyl Plant / Metolachlor Plant / Clothianidin Dead bees / 0.03* Thiamethoxam Dead bees / 0.005* Coumaphos Dead bees / 0.001* Thiabendazole Dead bees 2 2 NA/ 0.003* No pesticides Unaffected detected bees * LOD and LOQ values provided by MAPAQ are the screening method targets. The results in this table were determined using a targeted method for which a new target LOD has not been defined. The presence of the pesticide can be confirmed, however, the levels presented may not be precise. Page 23 of 112

25 3.2 Use pattern There were nine pesticide active ingredients detected in the honey bee samples analyzed for Ontario and Quebec, including acetamiprid, iprodione, phosmet, permethrin, fluvalinate, coumaphos, thiabendazole and clothianidin. With the exception of the fungicides iprodione and thiabendazole, all of the active ingredients detected are insecticides used to control various insect pest species. In order to determine the potential source of exposure for the honey bees the use pattern for each of these active ingredients was considered. Acetamiprid is registered as a seed treatment on canola, however; this pesticide is not currently used as a canola seed treatment in Canada. The active ingredient is also registered as a foliar spray on a variety of crops including fruit trees, vegetable crops and cereal crops. The timing of acetamiprid foliar applications on pome fruit, grape and strawberry occur pre bloom which would have been during the timing of the mortalities. Due to the early spring weather, foliar applications of acetamiprid to certain pests on blueberry and stone fruit may have also occurred during the period of the bee incidences (Crop Specialists, OMAF and MRA). Phosmet is a foliar spray that is used on a variety of fruit and vegetable crops as well as on livestock. It is applied as a pre bloom foliar application to apple, pear and grapes. Bloom timing of orchard crops this past spring began in late April in Ontario; the timing of these applications coincided with the timing of a majority of the bee mortalities (Crop Specialist, OMAF and MRA). Due to the early spring weather, foliar applications of phosmet to certain pests on blueberries may have also occurred during the period of the bee incidences. Permethrin is applied as a foliar spray to a variety of crops including fruiting trees and some field vegetables; it is also used as a soil drench to control the presence of small hive beetle (Aethina tumida) near honey bee colonies. Permethrin foliar applications could have been applied to a wide variety of crops such as blueberry, wheat, barley and brassica vegetables at the time of the bee mortality if growers were targeting the larval insect life stages. Although registered for use, application of permethrin to tree fruits is very unlikely due to the negative effects this active ingredient has on beneficial arthropods such as mites (OMAF and MRA, 2012). Coumaphos and fluvalinate are used to control varroa mites (Varroa destructor) in honey bee hives. It would not be uncommon for residues of these actives to be detected in honey bees. The most common pesticide detected in the honey bee samples was clothianidin. Given that clothiandin is a transformation product of thiamethoxam the use pattern for Page 24 of 112

26 thiamethoxam was also considered. Both thiamethoxam and clothianidin are used to treat corn seed which is the main local crop that was being planted at the time of the honey bee mortalities. Clothianidin is also applied to potato seed pieces and oilseeds as a seed treatment. Thiamethoxam is also applied as a seed treatment to sugar beets, wheat, barley, sunflowers, canola, soybeans, and as a seed piece treatment to potatoes. Both active ingredients are registered for use as a foliar spray on a number of crops. Clothianidin is applied to grape pre bloom and thiamethoxam is applied to pome fruits pre bloom; the pre bloom period for these crops occurred at the same time as when a majority of the incidences occurred (Crop Specialist, OMAF and MRA). Additionally due to the early spring weather, foliar applications of clothianidin to certain pests on tree fruits may have also occurred at the time of the bee incidences. Given what is known about the use pattern and the commonly used crop production for the nine pesticides detected it is likely that the following explain the exposure to these pesticides. Seed treatment: clothianidin and thiamethoxam, Foliar applications o acetamiprid to pome fruit, grape and strawberry, o phosmet to apple, pear and grape, o clothianidin to grape, and tree fruits o thiamethoxam to pome fruit o Permethrin may have been applied to a wide variety of crops Bee hive management: o Permethrin for control of small hive beetle o Coumaphos and fluvalinate for varroa mite control 3.3 Location of Honey Bee Mortalities Quebec Honey bee mortality was observed in St Dominique and was linked to pesticides used on corn seeds. St Dominique is located in the Montérégie region which is a major corn growing region in Quebec (Figure 2 and Figure 3). Agriculture surveys in a 5 km radius around the affected bee yard showed that the predominant crop was corn. This is consistent with the pesticides reported in the residue analysis, which are all used to treat corn seed (clothianidin, thiamethoxam, thiabendazole) or used in beehives for varroa mite control (coumaphos). Page 25 of 112

27 Figure 2: The Montérégie Region of Québec ( (The red star indicates the approximate location of St Dominique.) Figure 3: The Major Corn Growing Regions of Ontario and Québec ( Page 26 of 112

28 3.3.2 Ontario Further analysis of the locations where honey bee mortalities were reported in Ontario indicates that there is a strong relationship between the location and intensity of corn growing and the location and intensity of the reported honey bee mortalities. The Ontario honey bee mortalities occurred mostly in southwestern Ontario with a few located in central Ontario, and two in eastern Ontario (Figure 4). There are over 6000 bee yards registered with the Ontario Ministry of Agriculture and Rural Affairs. Figure 5 shows the location of all the registered bee yards and highlights bee yards that experienced mortality in spring of This map illustrates that while registered bee yards are located throughout Ontario, the yards that experienced bee mortality are relatively localized or clustered in the major corn growing regions of Ontario (Figure 3). The location of the honey bee mortalities corresponds with the location of the cash crops in Ontario (Figure 6); cash crops are the rotation between corn, soybean and wheat. The cash crop map was provided by OMAF and MRA and includes information on cropping patterns for several years in the 1980s (OMAFRA 1983). The location of the growers who purchased corn seed in 2012 planting season (Figure 7) was provided by Agricorp. This map also shows the location of fruit growers in Ontario. Figure 7 suggests the amount of corn in Ontario is greater in 2012 compared to the cash crop map (1980 s), and indicates a strong relationship between the bee mortality locations and where corn was being grown in Ontario in There is not a strong relationship between bee mortality locations and the areas where fruit is grown (Figure 8). As well, the intensity of corn growing in Ontario by county (Figure 9; based on Census of Agriculture Statistics for 2011) is compared with the intensity of bee mortality reports by county (Figure 10). There also appears to be a strong relationship with intensity of reporting and intensity of corn growing. The significant overlap of bee mortalities with corn growing intensity and the widespread detection of insecticides used to treat corn seed supports a relationship between the honey bee mortalities and corn seed planting. The location and frequency of detections of other residues in the bees do not follow a pattern that supports a hypothesis that these pesticides were implicated in the majority of the honey bee mortalities. The location of all the residues detected in dead honey bee samples is shown in Figure 11 Figure 17. Clothianidin was detected in a large number of dead bee samples and in all the areas where Ontario honey bee mortalities were reported (Figure 11). This active ingredient was also detected in an unaffected 5 honey bee sample from one beekeeper. Pesticide residues other than clothianidin were detected infrequently and were localized to smaller areas (Figure 12 Figure 17). 5 This unaffected sample was a sample of live bees collected from a hive that was considered unaffected in a yard with affected hives. A follow up report on June 22 nd indicated that all hives in this yard were affected. Page 27 of 112

29 Fluvalinate (Figure 12) was detected in dead honey bee samples with the majority from one beekeeper. This beekeeper also had positive detections of fluvalinate in unaffected honey bee samples. Fluvalinate is used to control varroa mites within hives. The presence of fluvalinate in honey bee samples indicates that the honey bees were exposed to this chemical and that the beekeeper in these instances may have treated for varroa mites in the past. Occasional detection of coumaphos, iprodione, permethrin and phosmet in dead bees were observed (Figure 13 Figure 16). Coumaphos, like fluvalinate, is used in honey bee hives to treat for varroa mites. Coumaphos was detected in bee yards from two beekeepers and the detections in honey bees indicate positive exposure to coumaphos, thus the beekeepers likely used these products in their hives at some point in the past. Permethrin was detected in two samples from two different beekeepers; it was detected once in a dead bee sample, and once in an unaffected bee sample. Permethrin has a very broad use pattern including the registered use around honey bee hives for the control of small hive beetle. It is not known how permethrin exposure occurred in these instances. Iprodione was only detected once in dead honey bee samples. This active ingredient is used as a fungicide on a variety of crops and can be applied in combination with another fungicide thiram as a seed treatment on canola. Phosmet and acetamiprid detections (Figure 16 and Figure 17) generally occurred near fruit growing regions (Figure 8), which is consistent with their use pattern. Phosmet, which is commonly used on tree fruit and selected berry crops, was detected four times in dead honey bee samples. The four samples positive for phosmet were collected from bee yards of two beekeepers with their yards located in or near the Norfolk Brant and Huron Erie fruit growing regions (Figure 16). Phosmet was not detected in unaffected honey bee samples. Acetamiprid is registered for foliar use on a variety of crops, including apples, potato, alfalfa and on canola as a seed treatment. It was detected in 36 honey bee samples of which 32 were from one beekeeper (Figure 17) where acetamiprid was detected in both dead bee and unaffected bee samples. In dead bee samples where other insecticides were detected, clothianidin was also detected in most cases. Corn is also grown in the areas of Ontario that had positive detection of pesticides other than clothianidin, as shown in the Agricorp map. The unaffected bees that were analyzed contained residues of other pesticides with only one sample containing residues of clothiainidin. Overall, the residue information is consistent in supporting a relationship between corn growing areas and bee exposure to clothianidin and/or thiamethoxam. Page 28 of 112

30 Figure 4: The location of the honey bee mortalities that were reported in spring 2012 Figure 5: The location of the honey bee mortalities relative to all the registered honey bee yard locations in Ontario (This map was produced in October 2012 by OMAF and MRA and does not include all of the locations where honey bee mortality was observed as some reports were received after this map was generated) Page 29 of 112

31 Figure 6: The honey bee mortalities and the Cash Crop Area (Soybeans, corn and wheat rotation Figure 7: Corn and Fruit grower location in Ontario for 2012 (provided by Agricorp) Page 30 of 112

32 Figure 8: The Major Fruit Producing Areas of Ontario (OMAF and MRA Fact Sheet What you should know about fruit production in Ontario htm) Figure 9: Map of Hectares of Corn grown in Ontario per county 2011 Census of Agriculture Statistics Canada ( Page 31 of 112

33 Figure 10: The relative proportion of bee yards affected per county Figure 11: Detections of Clothianidin in Dead Honey Bee Samples Page 32 of 112

34 Figure 12: Detections of Fluvalinate in Dead Honey Bee Samples Figure 13: Detection of Coumaphos in Dead Honey bee samples Page 33 of 112

35 Figure 14: Detection of Iprodione in Dead Honey bee samples Figure 15: Detection of Permethrin in Dead Honey bee samples Page 34 of 112

36 Figure 16: Detection of phosmet in Dead Honey bee samples Figure 17: Location of detection of Acetamiprid in Dead Honey Bee Samples Page 35 of 112

37 Section 4 Toxicity Overall Conclusions for Section 3 Toxicity and mode of action were compared among the pesticides detected in dead and unaffected bee. Clothianidin and thiamethoxam have the highest toxicity to honey bees of all the pesticides detected, and in most cases the symptoms observed during the mortality events are consistent with neonicotinoid poisoning. Phosmet and permethrin are also toxic to honey bees and are likely to exhibit similar effects due to their neurotoxic mode of action. Other pesticides detected in the honey bees have limited toxicity to honey bees. 4.1 Honey Bee Toxicity of the detected pesticides The LD 50 value is the median lethal dose of a test substance, or the amount required to kill 50% of a given test population. The NOEL, or no observed effects level, is the dose level at which no mortality is observed. The honey bee NOELs and acute oral and contact LD 50s are determined in laboratory dose response studies with honey bees based on exposure to either a single oral dose or a single contact dose (Table 9). Exposure of organisms to doses below the LD 50 may still result in mortality. For example, at a dose level equivalent to an LD 10, one could expect mortality to result in 10% of the test population. Exposure at or below a NOEL would not be expected to result in mortality. Table 9 summarizes the honey bee toxicity information available for the active ingredients detected in the Ontario and Quebec honey bee samples. Based on the classification scheme of Atkins et al. (1981), acetamiprid and coumaphos are classified as moderately toxic 2<LD µg ai/bee and thiamethoxam, clothianidin, phosmet, fluvalinate and permethrin are classified as highly toxic given that the LD 50 values are between and 1.99 µg ai/bee. The LD 50 values for phosmet and fluvalinate are one to two orders of magnitude higher (less toxic) than clothianidin and thiamethoxam indicating a significant difference in toxicity between these chemicals even though their hazard classification is the same (Figure 18). Permethrin shows lower toxicity (i.e., higher LD 50 values) than clothianidin and thiamethoxam on an acute oral exposure basis, but shows a similar toxicity level on an acute contact exposure basis. The use pattern of fluvalinate and coumaphos is different from the other insecticides in that they are used in honey bee hives to control varroa mites. Even though these chemicals are toxic to honey bees, the registered rate and use pattern are intended to expose honey bees to levels that are not likely to pose a risk of acute mortality, but are efficacious to control the mites. Iprodione and thiabendazole are fungicides that are practically non toxic to adult honey bees on an acute contact exposure basis. Page 36 of 112

38 Table 9: The acute 48 hr honey bee contact and oral toxicity endpoints (LD 50 and NOEL) in ug ai/bee and ppm for selected active ingredients detected in the Ontario and Quebec honey bee samples. Active Ingredient Clothianidin Thiamethoxam Original study PMRA# LD h oral 48 h contact 95% 95% Confidence NOEL LD limits 50 Confidence limits Insecticides µg ai/bee ppm* µg ai/bee NC NC ppm* NC NC Acetamiprid µg ai/bee Phosmet µg ai/bee PACR NA NA NA NA NA NA NA Fluvalinate µg ai/bee Contact: NOEL NA NA NA NA 0.2 NC NC PACR NA NA NA NA NA NA NA NA Coumaphos µg ai/bee EPA ECOTOX 24 h 24 h NA NA NA NA Database 3 to Permethrin µg ai/bee NA NA NA NA NA NA Fungicides Thiabendazole ** µg ai/bee PRVD NA NA NA NA NA NA NA NA Iprodione µg ai/bee > 25 NA NA NA > 200 NA NA NA NA = Not Available NC = Not Calculated *based on weight of bees collected during sampling (0.12 g/bee based on 100 bees) ** Thiabendazole is a fungicide and is practically non toxic to honey bees ( Page 37 of 112

39 Thiabendazole and iprodione are practically non-toxic to honey bees Concentration (ug ai/bee) Clothianidin Thiamethoxam Acetamiprid Phosmet Fluvalinate Coumaphos Permethrin Iprodione Thiabendazole 0.03 oral LD50 Contact LD Concentration (ppm) Figure 18: Acute Contact and Oral Toxicity Values (LD 50 ) of Selected Pesticide Active Ingredients Detected in Honey Bee Samples from Ontario and Quebec (Atkins et al criteria for bee toxicity grey area is highly toxic, blue area is moderately toxic, white area is relatively non toxic) 4.2 Mode of Action for Honey Bee Toxicity Honey bees exposed to the active ingredients detected in Ontario and Quebec (Table 10) exhibited symptoms that are consistent with effects on the central nervous system of insects such as tremors, paralysis and eventually death. Many of the beekeepers in Ontario reported symptoms other than death that included trembling; inability to walk or fly, and proboscis extension among others. These symptoms are indicative of pesticide poisoning and are similar to those expected from exposure to neurotoxic insecticides, including neonicotinoids and pyrethroids. Page 38 of 112

40 Table 10: The Mode of Action of the Pesticide Active Ingredients that are Toxic to Honey bee and Detected in the Ontario and Quebec Samples Active Class Information Symptoms Reference Ingredient Clothianidin Thiamethoxam acetamiprid Fluvalinate Permethrin Coumaphos Phosmet Systemic Nitroguanidine Neonicotinoid insecticide Systemic Cyanoamidine Neonicotinoid insecticide Systemic synthetic pyrethroid insecticide non systemic synthetic pyrethroid insectide Non systemic organophosate insecticides Bind to the nicotinic acetylcholine receptors of the insect s nervous system and mimics acetylcholine (ACh). Normally ACh binds to these receptors and is broken down by acetylcholinesterase which cannot readily break down neonicotinoids Neonicotinoids are closer mimics for the insect ACh than for human ACh, giving this class of insecticides more specificity for insects and less for mammals Inhibit sodium channel modulators (keep sodium channels open in neuronal membranes) Contact pesticide that depends on disturbance of anoxic nerve impulse conduction Acetocholinesterase inhibitors Mortality, affect the mobility by inducing symptoms such as knockdown (moribund), trembling, tumbling, uncoordinated movements (staggering), hyperactivity, tremors, rotating and cleaning of abdomen while rubbing hind legs together, decreased walking, proboscis extension Initially stimulate nerve cells to produce repetitive discharges (tetany) and eventually cause paralysis and death Erratic movements, unable to fly, paralysis, moribundity and death. Bees will die in forage area, between forage area and hive or at the hive. Regurgitation, bees are disoriented, paralysis, death, high percentage of bees will die at colony Blacquière et al. (2012); Decourtye and Devillers (2010) US EPA RED, 2005 Atkins (1992) Atkins (1992) Page 39 of 112

41 Section 5 Consideration of residue levels as related to toxicity Residue analysis of environmental samples may be used to confirm exposure; however, residue levels in environmental samples are affected by many factors. Even though exposure is confirmed from the detection of the pesticide, it is noted that the residue concentration is typically below the LD 50. The residue levels detected in environmental samples are affected by many factors, resulting in differences between exposure levels (the actual concentrations of the chemical to which the organisms were exposed) and detected residues. New laboratory studies submitted by Bayer Crop Science (PMRA , PMRA ) and Syngenta (PMRA , PMRA ) provided some insight into the relationship between exposure and the observed residue levels and toxicity in honey bees. These studies demonstrated that thiamethoxam transforms to clothianidin in honey bees. As such, even though thiamethoxam was not detected in honey bee samples collected in Ontario, the exposure of the honey bees to thiamethoxam cannot be ruled out. Concentrations of clothianidin or thiamethoxam detected in the laboratory samples of honey bees were generally lower than the initial dose and tended to be variable. In addition, there is evidence to suggest that residues of clothianidin and thiamethoxam detected in honey bees following oral exposure decline over time. These studies suggest that even in laboratory studies the residue levels detected in honey bees may not reflect the initial exposure to the bees. The length of time between death and the sample collection can influence the potential for detection of pesticide residues. The longer this time period is the greater likelihood that the residues will decrease. In addition to the degradation of residues in the honey bees there are other weather and environmental conditions that can affect the levels of the pesticide detected. This includes physical removal of the chemical (via wash off, wind and contact with surfaces) and the degradation for the chemical on the surface of the honey bees via various routes including phototransformation and biotransformation from surface bacteria on the dead bees. Another factor that can affect the levels of pesticide detected in the honey bee samples is dilution. The samples analyzed consisted of approximately 20 honey bees which may all have differing residue levels, and may include bees that were not exposed, thus diluting the concentration detected. When interpreting the residues levels determined it is important to consider the length of time between death and sample collection and analysis. There may also be differences in toxicity depending on formulation or exposure type (e.g. dust versus liquid). Two of the laboratory studies submitted by registrants (PMRA , PMRA ) investigated thiamethoxam exposure to liquid end use Page 40 of 112

42 product and dust generated from treated seeds. These studies suggested that the observed mortality following oral exposure to dust tended to occur more quickly than with liquid end use product. However, it could not be determined whether there were differences in overall toxicity levels. Historical information on insecticides indicates that dust formulations aerially applied over crops are generally more toxic to honey bees than liquid formulations aerially applied over crops. A California study demonstrated that a change in formulation from dust to spray with an organophosphate insecticide resulted in 50% lower honey bee colony deaths from 1962 to 1963 when the change was implemented (Atkins, 1992). The type of exposure (dust, liquid) can affect toxicity of the pesticide indicating toxicity values determined with one exposure type (e.g. liquid) may not always reflect toxicity of another exposure type (e.g. dust). Differences in exposure routes may affect the toxicity noted in honey bees. Traditional laboratory toxicity studies are conducted based on a single oral (ingested) or contact dose (usually applied directly to the thorax). However, with foraging honey bees it is unknown exactly how the exposure occurred as there could have been one or multiple doses through a combination of ingestion and/or contact exposure. Additionally, in contrast to laboratory studies the duration of exposure in the field is variable. The amount of pesticide required to result in a toxic effect may be different than determined in the laboratory studies due to differences in the exposure duration, the number of exposures and the potential combination of exposure routes. The exposure level at which mortality can occur also depends on sensitivity of the individual honey bees. Laboratory studies demonstrate that there is a variation in the individual sensitivity of honey bees as there is always some variation in the calculated LD 50. The individual sensitivity of the honey bees can be affected by various factors including the health of the bees and the different ages and stages of development. Therefore, it is possible to observe mortality following exposure to pesticides at levels that are lower or higher than the laboratory LD 50. Based on available information, affected bees are generally expected to have been exposed to higher pesticide levels than what is detected in residue analysis. Residue analysis provides useful information for the interpretation of an incident report (e.g., confirms exposure); however, drawing conclusions regarding the causality of an incident through direct comparison of the residue detected with the laboratory derived LD 50 alone is not expected to lead to clear conclusions. The factors that must be considered when comparing detected residue levels to laboratory determined LD 50 s or NOELs are outlined in Table 11. Page 41 of 112

43 Table 11: Sources of Variability in honey bee residue levels Factors affecting level of pesticide detected in samples Effect on interpretation Considerations Degradation in Bees Metabolism of Lower detection than Laboratory studies submitted by the registrant the pesticide by honey bees exposure concentration. demonstrate: a difference between the levels detected in honey bees relative to the levels to which they are exposed. residues detected in the honey bee samples were often below the oral and contact doses administered levels of thiamethoxam and clothianidin decrease with time in the bee samples.(also supported by published literature) Detection of transformation product (i.e., clothianidin) may be detected even if exposure was to the parent (i.e., thiamethoxam) Degradation in Environment Time of death vs. time sample collected Individual exposure level variations Exposure routes (oral or contact) Lower detection than exposure concentration Dilution Cannot distinguish whether residues detected result from: Oral or contact exposure or both one or many exposures variation in time between death and sample collection may reduce or remove residues through the following processes: biodegradation on the surface of honey bees by bacteria weather/environmental conditions that can affect the levels of a pesticide detection a. Physical removal of chemical via wash off, wind, contact with surfaces, etc. b. Degradation by other routes (photolysis, etc.) some honey bees may carry higher or lower residue levels than others the sample residue levels stated in this report are based on an average per bee from a sample size of approximately 20 honey bees (2 g sample) honey bees dying from natural causes (unlikely to have detectable levels of residues) may be included in the approximately 20 honey bees used to make up a sample, thus diluting the residues that are detected Differences in Sensitivity the LD50 and NOEL are based on laboratory exposures of individual bees to either a single oral dose (ingestion) or contact dose (usually a liquid painted on thorax) When honey bees are foraging they could be exposed through both oral and contact routes simultaneously and the LD50 resulting from this Page 42 of 112

44 Factors affecting level of pesticide detected in samples Exposure type (technical active ingredient (TGAI), liquid end use product (EP), dust, residues on plants,) Differences in sensitivity of individual honey bees Effect on interpretation Lower residue levels may result in toxicity if exposure was through multiple routes (oral and contact and/or many exposure times) Exposure type may result in different toxicity. Exposure levels at which mortality occurs differs depending on sensitivity Considerations combined exposure is unknown It is unknown from the residue analysis how much is due to oral versus contact exposure. It is unknown if the honey bees were exposed multiple times or received a single dose. The duration of exposure is unknown. Laboratory studies submitted by the registrants demonstrate that the dust generated from treated seeds may have different toxicity than the liquid end use product Historical information indicates that dust formulations tend to be more toxic to honey bees than liquid spray formulations for certain insecticides There may be potential differences in toxicity resulting from exposure to technical grade active ingredients (TGAI), liquid end use products (EP), dust, or residues on plants. It may be misleading to compare toxicity values from one exposure type to residue levels from a different exposure type. other stressors, e.g., honey bee health individual variation in sensitivity as demonstrated by the confidence intervals around the LD50 s (oral LD50 = ug ai/bee (95% confidence limit = ug ai/bee), contact LD50 = ug ai/bee (95% CI = ug ai/bee) Exposure during various time periods of development and energy requirements of the colony (e.g., bees coming out of overwintering may be more susceptible compared to midseason, etc.) Page 43 of 112

45 Section 6 Additional Factors Considered Summary of overall conclusions for Section 5 In addition to the presented information on exposure and toxicity other factors such as corn production in Ontario, seed corn supply and distribution, machinery and planting processes, weather conditions and honey bee health and biology, were taken into account in making conclusions regarding the potential contributors to the honey bee mortalities. Due to unseasonably dry and warm conditions the fields were ready to plant corn approximately 2 to 3 weeks early in spring 2012 in Ontario. It was determined that corn planting began in early April and continued through to the first half of May which coincides with the honey bee mortalities. Most of the corn used and sold in Ontario is treated with clothianidin or thiamethoxam and it appears that the market split for these two pesticides is approximately 50/50. There are planting machinery options available for corn planting, however; most corn planting in Ontario uses either vacuum seeders or positive air pressure seeders. Both of these types of seeders tend to require a seed lubricant to ensure that the seeds move easily through the seeder. These lubricants are either talc or graphite. During planting these lubricants are exhausted from the planters either into the air or towards the ground. These lubricants can be contaminated with the pesticides present in the seed coating as there is a certain amount of seed coat abrasion that occurs during the handling and planting processes. Bees can come into contact with contaminated dust while flying across the field during planting or from the dust settling on water sources or nearby flowers that they are foraging on. At the time of the honey bee mortalities, the weather was often very dry and unusually windy. The unseasonably dry and warm early spring also meant that there was low soil moisture in fields. These unusual conditions could have resulted in increased movement of the dust generated during planting of treated seed. Spring of 2012 was also good for honey bees. The mild winter resulted in low overwintering losses with an overall average winter loss of 12% for Ontario; well within the acceptable range. This average was significantly lower than the 43% overwintering loss reported in 2011 and lower than the winter losses for which ranged between 14 and 36%. The beekeepers indicated that their hives were very healthy and populous and honey bee inspection reports indicated that the level of pests and diseases were at the same level or lower than previous years. Page 44 of 112

46 6.1 Corn Production in Ontario The OMAF and MRA Agronomy Guide for Field Crops indicates that the best yields are usually obtained from corn planted in late April and the first half of May, because the crop is able to use the full growing season. Early planting also results in early maturity in the fall, reducing the risk of damage from an early frost or adverse weather at harvest. The optimum planting date is on or before May 7 in South western Ontario and May 10 in Central and Eastern Ontario. Depending on the total number of days required to plant the farm s entire corn acreage, it is generally necessary to start planting corn well before the optimum date. Producers wanting to plant corn significantly earlier than optimum dates (i.e., April 15 25) must consider that soil temperature needs to reach a minimum of 10 o C before germination and emergence will occur. If the average soil temperatures are at or above 10 o C, the soil conditions are favourable and the weather forecast is predicting average to above average temperature, then early planting of at least a portion of the corn crop can be initiated. The spring of 2012 was unusually warm and planting was known to have begun earlier than usual by 2 to 3 weeks. According to OMAF and MRA, corn planting would have begun in early April and continued through the first half of May in 2012 (OMAF and MRA; pers comm 2012). 6.2 Seed Corn Supply and Distribution In Ontario Most of the corn seed sold and planted in Ontario are treated with either clothianidin (Poncho) or thiamethoxam (Cruiser) in approximately equal proportions. There are sixteen different suppliers of corn seed in Ontario with Dekalb, Pioneer Seed, Northrup King, Maizex and Hyland being the dominant suppliers. Approximatly 80% of the corn seed planted in Ontario is treated in the United States and imported. 6.3 Machinery and planting processes for corn There is a variety of seeding equipment that can be used for planting of corn. An important consideration for many corn growers is the ability of the planter to provide accurate spacing between plants and prevent skipping or multiple seeds delivered to a particular spot. Additionally, depending on the area to be planted, the seeder may have to be of sufficient size to simultaneously deliver large numbers of seed across a wide boom length. Therefore, a variety of seeding equipment options have evolved and are discussed below. Page 45 of 112

47 6.3.1 Vacuum seeders (pneumatic seeder) This type of seeder is used in Canada for planting corn, as it is efficient and precise in seed placement (CLC, 2012). Figure 20: 19: An example of a Vacuum seeder Vacuum planters have individual row unit seed boxes. Each row unit has an associated vacuum seed meter. The seed is drawn into the meter using a vacuum to hold the seed to the disk. This planter uses gravity to drop the seed into the seed trench. Air is drawn from individual row units with exhaust at a central location (AEM, 2012). Both central (bulk fill) systems and individual row unit systems are available. (AEM, 2012). Photo credits: AEM Positive air pressure seeders These planters are similar to vacuum planters, except that instead of creating negative air pressure (to hold the seeds onto the disc), they create positive air pressure (to push the seeds onto the disc). Air is drawn from a central location with exhaust downward at each row unit, where the seed is deposited in the seed trench. Both central (bulk fill) systems and individual row unit systems are available. (AEM, 2012). These seeders can be used for corn planting. Photo credits: AEM Page 46 of 112

48 6.3.3 Air drills Air seeders have a central seed tank with associated centrally mounted volumetric meters (one per row or multi row with downstream splitter devices). Pressurized air is used to convey the seed from the tank to the meter to the seed trench. Seed is blown into the seed trench. Air exhausts downward so that the seed is deposited in the seed trench (Association of the Equipment Manufacturers, 2012). Photo credits: AEM This type of seeder is typically used for cereal, canola and soybean. It releases the seed and air in furrow with no exhaust of dust. These types of seeders are very common in Western Canada because of the large amount of cereal, canola, lentils and pulses that are planted in this part of Canada. In Ontario, this seeder may be used to plant soybean and other crops but would not be used to plant corn. (Bayer Crop Sciences; personal communication) Finger planter (Mechanical Meter Planter) This type of seeder is an older style system which does not use an air delivery system and therefore, does not have air exhaust. It is a system where mechanical fingers pick up the seeds and turn them against plates. Triggers cause the fingers to open and close resulting in the movement of the seeds at the appropriate time. This planter uses gravity to drop the seed into the seed trench (AEM, 2012). This type of seeder is used for planting corn. Photo credits: ( graindrills~planter Drills.html) Gravity seeder (Drill Box) Box drills have a central seed box with associated volumetric mechanical meters (one per row). This type of seeder uses gravity to drop the seed into the seed trench and is typically used for canola, soybeans, cereals, etc. where the placement of the seed is not very important. This seeder is not used for corn planting (AEM, 2012). Photo credits: AEM Page 47 of 112

49 6.3.6 Use of Talc or Graphite in planting of seeds Seeding equipment for corn must be able to deliver a precise and consistent number of seeds. This planting equipment has evolved to allow growers to plant large acreages in a minimal amount of time. The type of equipment used to plant these large acreages requires a relatively fluid movement of seeds down long booms. To accomplish this, growers rely on dry seed lubricants. Seed lubricants such as talc or graphite are recommended to be used when planting treated corn seeds. These lubricants are recommended by equipment manufacturers as they decrease static electricity in vacuum (negative pressure) planters and air (positive pressure) seeders; they act as a lubricant for seed meters; they will reduce bridging (seed sticking together) in seed boxes and they enhance seed flow ability to help provide even planting in the field (Talcusa.com and stineseed.com). Both talc and graphite are absorptive; insofar as they scavenge loosened or abraded seed treatment material while interacting in a seed box. Because of their lightness, the lubricant used in planting seeds is typically blown into the air through the exhaust vent of the seeding equipment when seed is planted using a vacuum seeder. The talc or graphite used as a lubricant can become contaminated with the chemicals in the treated seed coat as some of the coating may be abraded off the seed to produce a certain level of dust (CLC, 2012). Bees can come into contact with contaminated dust while flying across the field during planting or from the dust settling on water sources or nearby flowers that they are foraging on Dust Generation from treated seeds The following information was provided by the registrant to give some information regarding the amount of dust produced by the seeds treated with their products. This information only considers the dust that can be removed from the treated seed using a Heubach test method. It does not consider the dust generated at the time of planting, or from the talc or other material used during the process Syngenta Cruiser treated seeds Syngenta monitored the performance of commercial seed treatment operations in Ontario in 2011 and 2012 (PMRA ). Samples of corn hybrids were obtained from Ontario growers and seed retailers. The level of dust off (i.e., dust from abrasion of treated seed coatings) for these seeds was determined based on the Heubach method as described by the European Seed Association (ESA, 2011). These levels of dust off were compared to the European Heubach industry standard dust reference value for corn of 0.75 g dust/ seeds (ESA, 2012). This guideline was established to evaluate seed treatment quality of seed collected immediately following application of the pesticide to the seeds. All of the corn samples in this study were collected several Page 48 of 112

50 months following application of Cruiser 5FS, however, the method still provides a valid estimate of treatment quality for comparison with ESA s maximum dust reference value for corn. The results of the dust off study conducted in support of this investigation indicated that all the hybrids were below the ESA reference for corn with an average of ± g dust/ seeds in 2011 and ± g dust/ seeds in Bayer Poncho treated seeds Bayer indicated they will be submitting additional information regarding dust formation. Until that information is available the following is a statement from Bayer Crop Sciences. The acceptable European standard mentioned below is the EU Heubach standard of 0.75 g/100,000 seeds.... we have worked with our customers to help ensure proper use of coatings since the introduction of Poncho 600 FS as a seed treatment. With the increasing concern over dust abrasion, we incorporated into our QA [Quality Assurance] program the monitoring of commercially treated corn seed for dust abrasion. We conduct these tests using an accepted standard protocol and equipment identified by the European seed treatment industry [Heubach test]. Our purpose is to verify that the results obtained by our customers are meeting the expected low levels of dust abrasion that were identified in our development process. Since 2009 we have results from over 1000 samples of commercially treated corn. These results show an average dust abrasion level that is well below the lowest standard currently accepted in Europe [Hueback standard]. We continue to monitor each year and work with our customers to make sure that these standards are being maintained. 6.4 Weather conditions at time of the honey bee mortalities The bee mortalities were reported to have occurred in April and May of 2012 across Ontario and in the region of Quebec south of Montreal. The spring of 2012 was warmer and dryer than the normal for this region which would have allowed for early planting of corn as confirmed by OMAF and MRA. The higher temperatures during March of 2012 warmed the soil to temperatures conducive to corn planting earlier in the planting season. Generally, in Ontario and Quebec, there was limited snow cover during the winter of and below average precipitation during the spring months, particularly in Ontario. These two factors contributed to lower soil moisture and increased potential for dust generation during manipulation of agriculture fields during planting and tilling. The lower amount of precipitation also resulted in less abundant water sources for organisms including honey bees. Wind records for southern Ontario demonstrate that the spring of 2012 was also windier than normal which would result in Page 49 of 112

51 conditions that would be conductive to the blowing of dust potentially contaminated with pesticides from the treated seeds off field Temperature Southern Ontario and Quebec experienced an early spring in The average temperature for the month of March was >5 C higher than the normal average temperature for that month (Figure 21). During the month of April the average mean temperature across southwestern Ontario was close to the average normal temperature, with some overnight temperature dipping to below freezing (Figure 22). The months of May (Figure 23) and June (Figure 24) warmed up again to be above normal for this area of Canada. Table 12 shows the 2012 average temperature and the normal average temperatures for the months of March to June for four cities across Ontario. The information in Table 12 was obtained from the Environment Canada weather site 6 Figure 21: National March 2012 Monthly Mean Temperature Difference from Normal 6 Page 50 of 112

52 Figure 22: National April 2012 Monthly Mean Temperature Difference from Normal Figure 23: National May 2012 Monthly Mean Temperature Difference from Normal Page 51 of 112

53 Figure 24: National June 2012 Monthly Mean Temperature Difference from Normal Table 12: The 2012 mean temperature and the Normal mean temperature for 4 cities across Ontario Windsor London Peterborough Ottawa March 2012 mean temp Normal mean temp April 2012 mean temp Normal mean temp May 2012 mean temp Normal mean temp June 2012 mean temp Normal mean temp Precipitation On average, the provinces of Ontario and Quebec were dry compared to normal during the spring of Figure 25 to Figure 28 present the percent difference between the 2012 total monthly precipitation and the average monthly total precipitation for Southern Ontario between 1971 and With some localized areas receiving more precipitation it can be concluded that the majority of the areas where the bee mortalities occurred received 60% or less of the normal average precipitation in Ontario for the months of March through to May. Similarly the region south of Montreal in Quebec where the bee mortalities were observed was very dry compared to normal in 7 Agriculture and Agri Food Canada (2012) GS/historicalhistoriques.jspx?lang=eng&jsEnabled=true Page 52 of 112

54 March and April, but received more precipitation in May compared to the normal average compared to Ontario(Figure 29 Figure 32) Figure 25: Ontario region percent of average precipitation ( ) for March Figure 26: Ontario region percent of average precipitation ( ) for April Page 53 of 112

55 Figure 27: Ontario region percent of average precipitation ( ) for May Figure 28: Ontario region percent of average precipitation ( ) for June Page 54 of 112

56 Figure 29: Quebec region percent of average precipitation ( ) for March Figure 30: Quebec region percent of average precipitation ( ) for April Page 55 of 112

57 Figure 31: Quebec region percent of average precipitation ( ) for May Figure 32: Quebec region percent of average precipitation ( ) for June Page 56 of 112

58 6.4.3 Soil Moisture Soil moisture information for 2012 was provided by Agriculture and Agri food Canada 8. There is a strong relationship between the soil moisture levels observed this year and the precipitation received in the specific areas. The information was collected from the Soil Moisture and Ocean Salinity (SMOS) Earth Observation (EO) mission 9. Figure 33 to Figure 36 show the variation in soil moisture through the months of April to July of The soil moisture variation is shown by the red squares with the darkest red square representing <5% surface soil moisture and the white area representing >25% surface soil moisture. Unfortunately at this time the PMRA does not have available historical surface soil moisture that will allow for a direct comparison with the 2012 data. Figure 33: Soil moisture for the month of April 2012; red indicates <5% soil moisture while white indicates >25% soil moisture. 8 Agriculture and Agri Food Canada (2012) ftp://ftp.agr.gc.ca/pub/outgoing/aesb eosgg/climate_data/soilmoisture/soilmoisture/2012/ 9 Canadian Space Agency (2012) ( csa.gc.ca/eng/programs/grip/archive_ asp). Page 57 of 112

59 Figure 34: Soil moisture for the month of May 2012; red indicates <5% soil moisture while white indicates >25% soil moisture. Figure 35: Soil moisture for the month of June 2012; red indicates <5% soil moisture while white indicates >25% soil moisture. Page 58 of 112

60 Figure 36: Soil moisture for the month of July 2012; red indicates <5% soil moisture while white indicates >25% soil moisture Wind speed The wind speed data cited in this section was sourced form Environment Canada s National Climate Data and Information Archive Data indicate that April was windier than normal in Ontario. Compared to historical average April wind speeds, the three cities investigated (Sarnia, London and Toronto) all had average wind speed that were higher than normal (Table 13). Table 13: Comparison between historical April average wind speeds and April 2012 average wind speeds for three cities in Ontario. Sarnia London Toronto Wind Speed km/h Historical April average ( ) Spring 2012 April average Difference Further analysis was conducted using available data for daytime wind speeds (between 6am and 6pm). Daytime wind speeds were selected for further analysis as this is the time when bees would be out of their hives and when agricultural activities would be taking place. Page 59 of 112