Quantifying the frequency and effects of secondary exposure to rodenticides in barn owls

Study design and procedures are designed to collect appropriate type and amount of data for analyses surrounding barn owl survival, reproductive success, chick growth rate, nest attendance, diet, foraging ecology, and rodenticide exposure in an agricultural setting in Yolo and Solano Counties.

All work was conducted under appropriate federal (# 23406) and state permits (# 1259). Field sampling protocols were approved by UC Davis Institutional Animal Care and Use Committee (IACUC # 20516).

Study Sites:

Our study was conducted on agricultural properties in Yolo, Solano, Sonoma, Napa, and Monterey counties in California. The crops grown on these farms included perennial crops such as wine grapes and fruit and nut trees; forage crops such as alfalfa and winter wheat; and annual crops such as tomatoes, sunflowers, and corn. Properties varied in the amount of natural land cover, such as grassland, oak woodland, and oak savannah, in the vicinity. This work took place annually between the months of January and August for the years 2018-2021.

Nest monitoring and banding:

We monitored occupancy and nest status using a live-streaming camera (SONY Action Cam) attached to an extension pole (Kross et al. 2016; Wendt and Johnson 2017). This is an effective way to minimize nest disturbance and the chance of flushing adults from nest boxes. We recorded nests as unoccupied, roosting individuals or pairs, incubating females, eggs and/or nestlings. Nest boxes that were at appropriate stages were included in rodenticide exposure, nestling growth, telemetry, and/or video monitoring protocols.

We banded all roosting adults and nestlings (minimum of 4 weeks old; Varland et al. 2007) using US Geological Survey federal lock on leg bands. We did not disturb owls in the incubation period due to risk of nest abandonment (Taylor 1994). We used a stockinette head covering to minimize stress once owls were removed from the nest box (Colvin and Hegdal 1986). We took morphometric measurements, including mass, hallux length, wing chord, tail length, culmen length, and facial disc width. At this time, we took a blood sample for rodenticide screening and collected pellets from the nest box (see below). Individuals were returned to their nest boxes with the entrance holes temporarily plugged (i.e., wooden plug or towel) to prevent adults from flushing, which is a standard technique when capturing birds from nest boxes (Colvin and Hegdal 1986). The entrance hole was unplugged immediately before vacating the site.

Rodenticide exposure screening:

Because we wanted to understand rodenticide exposure in a living population, we used blood samples to screen for circulating rodenticide exposure. Compounds can be detected in the blood after consuming exposed prey for a window of time that may be as short as several days or as long as 2 weeks, depending on dose and compound. We sampled adults and nestlings over 4 weeks of age found in nest boxes. If nestlings were being studied for growth rates, they were sampled up to 2 times between the 4-9 week old nestling period. Individuals that were sampled multiple times were not sampled within the same 14 day period. For nests that had many nestlings and were not part of the growth or telemetry study, we selected two random nestlings for blood draws.

Prior to blood draws, we weighed each individual to ensure the maximum of 1% of body weight allowed by IACUC guidelines was not exceeded, although due the required amount of blood for rodenticide screening we collected much less than 1%. Blood was drawn via medial metatarsal venipuncture using a 25 gauge needle and a 3cc syringe. Past experience working with owls and other raptors, as well as discussion with raptor veterinarians, indicates that this is the best location for drawing blood and minimizes the likelihood of hematoma. If we were unable to obtain blood from this site we did not attempt another location. Prior to drawing blood, the metatarsus was wetted with disinfectant and pressure was applied to visualize the vein. One researcher restrained the bird (the handler) while the second performed the blood draw. Following the blood draw, the handler applied pressure to the draw site for a minimum of 1 minute and visually inspected the draw site to ensure bleeding had stopped. Prior to release of the owl back into the nest box, we inspected the draw site again to ensure bleeding had not resumed.

Blood was stored in a cooler and frozen in cryovials in the lab. Blood draws were conducted by researchers that were trained by UC Davis avian veterinarians. Rodenticide screenings took place at Texas A&M Veterinary Medical Diagnostics Laboratory via measurement by liquid chromatography/mass spectrometry (LC/MS), which detects occurrence and measures quantity of first- and second-generation anticoagulant rodenticides (Warfarin, Bromadiolone, Brodifacoum, Diphacinone, Chlorophacinone, Coumatetralyl, Difenacoum, and Difethialone).

We collaborated with Martinico et al. (GW19-200) to compare exposure rates from breeding owls to exposure rates of owls and hawks in fall and winter on the same properties. We also collaborated with Phillips et al. (in prep) to compare rodenticide exposure detection from blood samples to detection through regurgitated pellets found in the same nest boxes.

Nestling growth rates:

We visited a subset of nests with nestlings on a weekly basis to gather data to compile growth curves, test for rodenticide exposure (see above), and gather pellets for diet analysis (see above). Due to extreme heat of the Central Valley in spring and summer, all handling and measurements took place in the early morning and was completed before temperatures exceeded 80 degrees F to prevent hyperthermia. When possible, permanent and temporary shade structures were used to minimize nestlings' exposure to direct sunlight. Before entering a box, we plugged the nest box entrance to prevent any adults that might be present from flushing. We measured mass, hallux length, and culmen length for all nestlings. Once feathers emerged from the skin we took wing chord, tail length, and facial disk width measurements. Nestlings were marked with non-toxic fabric dye on downy feathers on the head for future identification prior to being old enough for banding. We banded nestlings once they reached 4 weeks old, when tarsi have fully developed (Taylor 1994, Varland et al 2007).

Diet monitoring:

We collected all fresh whole pellets from the inside of active nest boxes at each visit and subsequently froze pellets until further analysis. We dissected pellets to classify the diet of the barn owls living in different agroecosystems. We identified vertebrate prey items commonly found in the barn owl diet which included: deer mice (Peromyscus maniculatus), house mice (Mus musculus), California voles (Microtus californicus), pocket gophers (Thomomys spp.), Norway rats (Rattus norvegicus), and black rats (Rattus rattus). We did not measure or try to distinguish between house and deer mice and therefore used the estimated biomass of mice surveyed in a tangent study on the same vineyard (Phillips et. al., 2020). We estimated the biomass of voles and gophers using previous study methods (see Kross et. al. 2016). We chose to be conservative when dissecting the pellets and therefore jaw bones were the only identifying marker used. Any pellets that did not have jaw bones were not identified to prey species and therefore were not included in the final pellet results. We collaborated with Martinico et al. (GW19-200) to compare breeding diet to fall and winter diet on the same landscape. This collaboration assisted in if the proportions of prey shift during different times of the year, which may impact susceptibility to secondary exposure to ARs in barn owls.

Nest boxes with more than 11 intact pellets collected were used to identify diet proportion and biomass removed from the surrounding landscape (Charter et al. 2012). Jaw bone measurements from pellet dissections were used to estimate average biomass of prey items by nest box and the average biomass of prey items over the entire network of nest boxes. If a nest box had less than three gophers and/or voles found in pellets dissected, the average biomass for those nest boxes were based on the average gopher/vole mass from the entire nest box network rather than the individual nest box. Using video monitoring (see below) we were able to identify the average biomass per prey delivery (D). The average number of prey deliveries per nestling per night was calculated using prey delivery videos (4.19 deliveries per night per chick), and was multiplied by the average number of nestlings during the approximate eight week period they were in the nest box. The number was then multiplied by D to give us the average biomass delivered per night. Finally, this number was multiplied by the number of days in the nest box to give us the average biomass consumed by each nest box during the nestling period. If a nest box failed (i.e., all nestlings perished), the eight week period was decreased to the age of the nest box at failure.

Video monitoring:

Prey deliveries were recorded using a low-cost motion trigger video camera between 2019 and 2021 breeding seasons. The low-power Raspberry Pi Zero W was used to create an affordable alternative to higher-end wildlife cameras currently on the market. This micro-computer has Bluetooth and LAN technology and has an open-source operating system allowing for high customization opportunities. The computer language for coding the camera was written in Python 3 (van Rossum 1995). A tupperware container (Rubbermaid Brilliance Food Storage 3.2 cup) was used as a waterproof case. A passive infrared motion sensor was used to activate the motion capture video. Videos were saved on a microUSB (SanDisk Ultra Dual Drive m3.0 32GB) connected to the Raspberry Pi. The cameras were set up for motion capture and were set to save video that occurred three seconds prior to motion, and then ten seconds after the camera was triggered (13 seconds per recorded motion trigger).

Cameras were placed at the same time as nestling growth measurements were being taken to decrease disturbance of the nest box. The cameras were placed on the opposite side of where the nest box entrance was to capture prey deliveries brought back by adults. Batteries to operate the camera (Hiluckey 25000mAh Solar Charger, RAVPOWER 25000mAh Solar Power Charger RP-PB092; discontinued) were placed on top of the next box and were exchanged weekly during nest box visits.

Prey were identified in videos where items were clearly visible and not obstructed by the adult or nestling owls, and classified into general size categories when difficult to identify (small, medium, or large rodent, reptile, bird, invertebrate). During initial nest box monitoring, when we found an adult, we strategically placed a USGS metal band on different legs to identify male versus female later on the recorded videos. If we could not identify the adult during the video, the adult was marked as unknown.

Telemetry:

To document spatial hunting patterns and foraging in proximity to rodenticide bait stations, we used GPS transmitters (Alle-300 models, Ecotone Telemetry, Poland). Transmitters were programmed to record 1 GPS point every minute each night from between 20:00—06:00 for the duration of battery life of the unit (expected to last 5-7 days). Females were selected for transmitter deployment if they had greater than 2 week old nestlings and were heavy enough so that the total mass of the unit and harness did not exceed 3% of body mass (IACUC guidelines; USGS Bird Banding Laboratory; Kenward 2001; Boal 2014). Transmitters were placed on the bird's back and attached with teflon ribbons that crossed at the breast at the top of the keel (this ensures the crop is not obstructed) and under each wing to connect to the two loops at the bottom of the transmitter. The teflon ribbons were carefully adjusted for fit by working them underneath the feathers and leaving one finger width of space (this ensures the transmitter is not too loose or too tight). Once the ribbons were adjusted, they were clamped in place and glued with a small amount of super glue to prevent fraying. Transmitter fit was rechecked before release by ensuring normal movement of the owl and that there is one finger width between the transmitter and the bird. Our processing time for attaching and detaching GPS transmitters of each adult did not exceed 20 minutes and was only conducted by trained and permitted individuals. We downloaded data from the transmitter’s base station, installed beneath the nest box, to infer the duration of the battery and when we should return to retrieve the transmitter. We made every attempt to recapture owls for transmitter removal before nestlings fledged.