Demonstrating a novel technique for luring raptors using songbird distress calls. Jason Lacson and Kathryn E. Sieving College of Liberal Arts and Sciences University of Florida I present a novel method for attracting forest dwelling raptors The technique involves a combination of playback of songbird distress calls and an artificial songbird model. I used this technique to conduct camera trap surveys in forest habitat in north central Florida from early December 2014 to mid February 2015. I compared the effectiveness of distress calls from two different songbird species. I also conducted active sampling in early March 2015 to directly observe behavioral responses among raptors INTRODUCTION ffective surveying and capture of raptors is often necessary for studies of population ecology, anatomy, and behavior ( Marti, 2007; Bloom et al. 2007). Due to their diversity of ecological and behavioral characteristics, there is no single surveying or trapping method that is effective across all species. As a result, a wide variety of techniques have been developed Furthermore, co mpared to other birds, raptors often present a challenge for wildlife researchers due to their relatively low densities and cryptic nature (Fuller and Mosher 1981). In order to increase detectability or capture rates, a common component in many study desi gns is the use of an audio lure to elicit conspicuous behavioral responses. The most used form of audio lure i s playback of conspecific calls, which elicit territorial resp onses in many raptor species ( Bloom et al., 2007 ) Recent research has suggest ed that playback of the songbird distress calls may also serve as a potential lure (Burnett and Sieving, unpublished data ). Another less documented technique is the use of a stuffed prey item typically in place of live bait Stuffed prey has been used for trapping Northern Goshawks in bownets (Kenward and Marcstrom, 1983) as well as Great Gray Owls in cast lures (Nero, 1980). I explored two different novel methods to lure r a ptors to a point location (referred to as camera trap sampling and human observer sampling ). I used a combination of distress call playback and an artificial pre y model. I examined the potential of this technique for camera trapping forest dwelling raptors. Camer a trapping has previously been used for surveying large, ground Kinnaird 2008; Dinata et al. 2008), but has seen little documented success for surveying other avian taxa. With an effective lure, camera trapping may serve as a n effective method for determining occupancy Second, I also condu cted active sampling, in order to directly observe the behaviors of responsive raptors I sought to assess the potenti al of this techniq ue for raptor capture. METHODS Study Area I conducted passive s ampling mainly in hardwood habitat situated within the University of Florida campus in north central Florida (29.4 N, 82.0 W) from early December 2014 to mid February 201 5. I selected locations where at least one raptor species had been iden tified within the last year in eBird Raptors found in the sites sampled vary in their food preferences and foraging behaviors. I conducted additional samples in suburban residential yards to explore potential differences in predator composition Table 1 show s a list of potential raptor species within the sampling area. Table 1. Responses to passive sampling by raptor species. Species Main prey Hunting mode Barred O wl ( Strix varia ) Mam mals Dive Red shouldered H awk ( Buteo lineatus ) Mammals Dive Cooper 's H awk ( Accipiter cooperii ) Birds Aerial pursuit Sharp shinned H awk ( Accipiter striatus ) Birds Aerial pursuit Red tailed H awk ( Buteo jamaicensis ) Mammals Dive Great Horned O wl ( Bubo virginianus ) Mammals Dive Eastern Screech O wl ( Megascops asio ) Mammals; Invertebrates Dive; Aerial pursuit I conducted active sampling in the u pland mixed woodland habitat of north central Florida situated within the Ordway Swisher Biological Station, a 3700 ha managed research preserve, in early March. Woodland vegetation comprised a mixture of longleaf pine ( Pinus palustris ), shortleaf pine E
( P inus echinata ), and mixed hardwood spe cies (FNAI, 2010 ). Sampling Design Camera trap sampling: Each sampling set up consisted of a motion activated digital camera (Wingscapes Birdcam Pro) focused on an artificial prey model. I chose a Northern cardinal ( Cardinalis cardinalis ) model to serve as a visual lure due to its high visibility amongst the forest vegetation. The model was mounted on a 20 cm x 20 cm wooden platform raised 1.5 m from the ground by a metal pole to prevent potential disturbance from ter restrial mammals. Hous ed inside a waterproof plastic bag attached directly underneath each platform was a digital audio player (Transcend MP300) wired to a loudspeaker (Rad ioshack Mini Audio Amplifier). Figure 1 Camera trap platform with speaker and MP3 player underne ath and Cardinal model mounted on top Each player c ontained a WAV file recording o f the distress calls of either Tufted T itmice ( Baeolophus bicolor ; hereafter Titmouse) or Blue J ays ( Cyanocitta cristata ; hereafter Jay). These two species are common and widespread within the study area. Previous studies have suggested that the mean prey species weight for a raptor is proportion al to mean raptor species weight (Storer, 1966). Therefore, u nder the assumption that different raptor species may be more attracted to the distress calls of avian prey they typically hunt I chose these two species of disparate body size as my treatments to attract the widest range of potential predators. For each treatment, I combined recordings of different individuals into a single sound file, in o rder to eliminate variation in responsiveness that may result from variations in each recording. I used r ec ordings from Branch and Freeb e rg (2012) for the Titmouse treatment Exemplars for the Jay treatment came from record ings sourced from the Macaulay Library T o balance the likelihood of detection of the playback by foraging raptors and habituation to the treatmen t once a raptor has detected it, p layback consisted of 10 minutes of distress call broadcast repeated at 30 minute intervals. Each trial lasted 15 total hours, starting in the after noon (1500 1600) and ending in early morning (0600 0700). I set this sampling period so that both nocturnal raptors (Strigidae) and crepuscular ones would be exposed to the playback. Roth and Lima (2007) ha ve shown that two bird eating raptors the sharp shinned hawk ( Accipiter striatus ) and Cooper s hawk ( A. cooperii ) are most active around dawn and dusk. For each sampling site, I conducted up to six trials, each one spaced one day apart or more, to increase the potential of capturing raptors that pass through each sampling site but are not res ident. I selected the initial treatment type for each location randomly, then alternated treatments for consecutive trials. To decrease potential weather effects, I did not conduct sampling when rain occurred in the afternoon. I used a motion activated di gital camera (Wingscapes Birdcam Pro) to record presence and behaviors of raptors on each platform. Each camera captured four photos and four 10 second videos when triggered by motion detection. To reduce redundant captures, I set five minute delay interva ls between each trigger event I considered instances where a raptor triggered the camera as an attack, and treated only instances occurring in different intervals of the playback as separate attacks when the s ame species occurred consecutively. I fit the data with a g eneralized linear model in Statistica (StatSoft, 201 4 ) using a binomial distribution and logit link function, and included temperature and rainfall data from the Florida Automated Weather Network as predictor variables Human observer s ampling: For faster deployment, I substituted the raised platform with a mesh cage and placed the audio equipment inside. I mounted the cardinal model on top of the cage rather than inside so that responding raptors are less likely to attempt to carry and fly off with the cage. I substituted the audio player and speaker with a bluetooth capable audio speaker (brand), remotely connected to my phone (LG G3). I used only Jay distress ca lls as my single treatment, b ecause sampling was limited to daytime surveys during the spring when, s maller raptors, namely Eastern Screech O wls ( Megascops asio ) and Sharp shinned H awks ( Accipiter striatus ) would most likely be absent.
After setting up the pla tform, I retreated in a random direction to a location 20 m away from the platform. Once situated inside a portable blind I started the distress call playback. I broadcast the call continuously for up to 30 minutes or until I detected a raptor within a 50 m radius of the speaker. I then alternated between playback and silence on an interactive basis, based on the approach behavior of the responding raptor. This continued for the duration of the ck, up to 3 hours, in order to not be overly disruptive. Once all raptors had appeared to lose interest in the playback, I repeated the process, broadcas ting the call until no raptors we re detected within 30 minutes. During each survey, I noted the latency for first raptor detection, the time that raptors stayed within the survey area and the behaviors of each raptor RESULTS Passive Sampling Raptors attacked in 15 out of 47 trials (31.9%) across 10 sampling sites. C a meras detected 30 total attacks ( Table 2 ) B arred O wls ( Strix varia ) were detected in all 15 trials and were responsible for 29 a ttacks. In one trial, a single Red shouldered H awk ( Buteo lineatus ) was also detected, attacking once. No attacks were recorded during periods that playbacks were not broadcast. Among the 10 sampling sites, 8 received at least one attack across all trials, while 2 sites did not receive any responses across the six trials in ea ch site. Table 2 Responses by raptors to camera trap sampling Species Trials with Attacks Total Attacks Barred O wl ( Strix varia ) 15 29 Red shouldered H awk ( Buteo lineatus ) 1 1 Figure 1 Barred Owl (left) and Red shouldered Hawk (right) attacking prey model. Barred O wls attacked in 9 out of 23 trials (39.1%) with the Titmouse treatment. and 6 out of 18 trials (33.3%) with the Jay treatment. The average number of attacks per trial for the Titmouse treatment was greater ( p = 0.0 4 ) than th at of the Jay treatment Trials in locations where the Titmouse treatment was conducted first ( n = 30) had greater (p = 0.01 ) average number of attacks per trial than trials in locations where the Jay treatment was conducted first ( n = 18). Temperature and rainfall did not appear to have a significant effect on attack rate s Table 3 Effects of initial playback type on attacks by Barred Owls Treatment Trials with Attacks Total Attacks Mean Attacks per Trial Titmouse ( n = 23) 9 (39.1%) 20 0.867 Jay ( n = 24) 6 (25%) 9 0.375 * Significant difference ( p < 0.05) Active Sampling Raptors ( n = 8) approached within 30 m of the broadcast in 5 out of 11 trials (45.5%) Responding species include the Red tailed H awk ( n = 1), R ed shoulder ed H awk ( n = 3), s H awk ( n = 2), M erlin ( Falco columbarius ; n = 1), and Barred O wl ( n = 1). Time to first approach ranged from 5 to 30 minutes and did not appear to vary by species All responding raptors engaged in inspection behaviors in varying degrees. All moved from perch to perch around the speaker multiple times during the playback. In separate trials, two individuals s H awk and one Red shouldered H awk landed on the ground and approached within 5m of the br oadcast, but did not perch o n the platform. Most raptors ( n = 5) did not approach closer than 20 m from the speaker. Individuals remain ed within 30 m of the broad cast for 30 minutes to 2 hours The two raptors that r emained longest were a pair of Red shoul dered H awks one of which had a rodent in its mouth when it first approached. DISCUSSION Impact of distress call broadcasts The study by Branch and Freeberg (2012) concluded that Titmouse distres s calls do not attract raptors However, locations sampled were not tested for raptor presence prior to the study and playbacks were not conducted at night Burnett and Sieving ( unpublished data ) previously showed that playbacks of the same Titmouse distress calls significantly improved detectability of raptors in point
counts. Likewise, data from passi ve sampling in this study also indicates that distress calls attract raptors. Data from passive sampling indicates that treatment type may have had an effect on attack rates. Sites where the Titmouse treatment was conducted for the first trial receive d more responses overall compared to sites where the Jay treatment was conducted first In addition trials with the Titmouse treatment received more responses than trials with the Jay treatment. This may indic ate a stronger attraction among Barred Owls to the distress calls of the Tufted Titmouse. However, because multiple trials were conducted in each location, and because barred owls were the only attackers in nearly all trials, it was not possible to parse the potential effect of treatment type against variations in each location. Impact of artificial prey lure By employing an artificial prey model in addition to distress call playbacks I was able to lure some raptors to a specific perch (the platform). The model alone was not as strong in attracting raptors, as raptors only responded when distress calls were broadcast. Barred O wls appeared sufficiently "fooled" by the cardinal mode l when paired with playback Cameras recorded attempts to eat the model in all instances containing barred owls, and models often needed to be replaced after attacks due to disfigurement. Other raptors appeared to be less a ttracted to the model. The one Red shouldered H awk detected by camera was not recorded a ttempting to eat the model. One video showed the individual perched on the platform dropping directly underneath, then quickly perching back on the platform. I believe this was an attempt to locate the source of the broadcast. The following day I found the plastic bag containing the audio equipment had been torn in multiple spots, though the camera did not capture this event. Raptors that approached within 5 m of the broadcast during active surveys also did not appear to be attracted to the p rey model. Comp ared to live bait, artificial lures are likel y less attractive to raptors due to lack of movement Bloom et al. (2007) noted that wild rodent species are poor attractants for diurnal raptors because they tend to remain motionless The cardinal model used i n this study had artificially painted feathers, which may be less attractive to raptors than more natural, stuffed bait. Responding Species Barred owls were the most responsive species during passive sampling, whereas red shouldered hawks were most responsive during active sampling. These species are the most abundant nocturnal and diurnal raptors, respectively. I believe the absence of other species during passive sampling were due to actual absence within the sampling area The locations I selected for passive sampling were smaller patches of forest compared to the rural locations where I conducted active sampling. B oth Barred Owls and Red shouldered H awks primarily feed on small mammals, though birds constitute a small portion of their diet ( Dykstra et al. 2003; Livezey, 2007 ). Because distress call s presumably signal to raptors that a bird is incapacitated, raptor species that primarily hunt mammals may seize the opportunity to take an easier meal compared to one that must be captured Furthermore, as mammal hunters, they are adapted to capturing pr ey on the ground, which may make them more likely to land on a baited platform, compared to r aptors that are primarily bird eaters, which are adapted to aerial pursuit of prey This may be another factor contributing to the absence of these species during passive s ampling It could not be determined if any raptors responded but did not land on the platform. Future Work Given that diurnal raptors appeared less attracted to the prey model than barred owls, modifications should be made to make a more convinc ing visual stimulus Live bait would be most effective. However, in situations where live bait is unavailable or undesirable, a convincing prey model may serve as a viable alternative. One that is motion capable would likely be most effective. Changes to t he playback protocol may also increase the rate of attacks. Implementation and comparison of distress calls and prey lures with current techniques for surveying and capture would be most insightful for developing more effective designs. REFERENCES Bloom, P. H., W. S. Clark, and J. W. Kidd. 2007. Capture techniques. Raptor research and management techniques. Hancock Ho use Publishers, Blaine, WA USA. 193 219. Branch, Carrie L., and Todd M. Freeberg. 2012. Distress calls in tufted titmice (Baeolophus bicolor): are conspec ifics or predators the target?. Behavioral Ecology. ars041. Dinata, Y., Nugroho, A., A chmad Haidir, I., & Linkie, M. 2008 Camera trapping rare and threatened avifauna in west central Sumatra. Bird Conservation International, 18(01), 30 37. Dykstra, C. R., Hays, J. L., Simon, M. M., & Daniel, F. B. 2003 Behavior and prey of nesting Red shouldered Hawks in southwestern Ohio. Journal of Raptor Research, 37(3), 177 187. Florida Natural Areas Inventory (FNAI). 2010 Guide to the natur al communities of Florida: 2010 edn. Tallahassee, FL: Florida Natural Areas Inventory.
Kenward, R. E. Karlbom, M., & Marcstrom, V. 1983 The price of success in goshawk trapping. Journal of Raptor Research 17:84 91 Fuller, Mark R., and James A. Mosher. 1981. Methods of detecting an d counting raptors: a review. Estimating Numbers of Terrestrial Birds Cooper Ornithological Society Lawrence, KS USA. 235 246. Livezey, K. B. 2007. Barred owl habitat and prey: a review and synthesis of the literature. Journal of Raptor Research, 41(3), 177 201. Nero, R. W. 1980 Great gray owl. Smithsonian Institution Press. G., & Kinnaird, M. F. 2008 A picture is worth a thousand words: the application of camera trapping to the study of birds. Bird Conservation International 18(Supplement S1), S144 S162. Perrone, Michael. 1980. Factors affecting the incidence of distress calls in pa sserines. The Wilson Bulletin. 404 408. Roth II, T. C., & Lima, S. L. 2007. The predatory behavior of wintering Accip iter hawks: temporal patterns in activity of predators and prey. Oecologia, 152(1), 169 178. Storer, Robert W. 1966. "Sexual dimorphism and food habits in three Nort h American accipiters." The Auk. 423 436. StatSoft, Inc. (2014). STATISTICA Version 12.0A http://statsoft.com/.