A meta-analysis of parrot tracking studies reveals the difficulties of transmitter attachment to Little Corella (Cacatua sanguinea)

ABSTRACT

Little Corellas (Cacatua sanguinea) are notorious for damaging crops and ornamental trees, and are an overabundant nuisance native in some urban areas. Habitat and landscape use studies ideally require the attachment of tracking devices. Here, we describe an attempt to attach GPS transmitters to Little Corellas using the backpack and keel-strap harness method. All five tracking devices initially attached to Little Corellas were removed by the birds within 48-hours. Additionally, a conspecific was observed chewing on the Teflon ribbon of a GPS harness shortly after the release of a Little Corella; a novel record that suggests cooperative abilities for the species. To complement our study, we conducted a literature review of parrot (psittacine) tracking studies. We investigated attachment styles and potential reasons for past device removals. Of the 46-relevant studies, neck collars were the most common attachment method (38.8%), followed by backpack harnesses (24.5%). Transmitter removal was documented in 28.3% of studies, with removals being most common for medium-sized (500–1000 g) species, and for combination (simultaneous back- and tail-mounts; n = 4) and backpack harness attachments (n = 6). Although there are relatively few parrot telemetry studies to date, they are increasing, and future studies will require considerable innovation and collaboration to overcome attachment style hurdles for difficult species, like the Little Corella.

KEYWORDS:

• Attachment techniques
• Australia
• harness
• parrot movement
• telemetry
• tracking device

Introduction

The field of movement ecology has increased our understanding of species distribution, habitat selection, foraging behaviour and migration (Kays et al. 2015; Clements et al. 2021). Attaching tracking devices to birds has provided new insights into how species interact with the spatial environment, which would be near impossible to achieve using observation data from free-moving individuals (Groom et al. 2015; Bodey et al. 2018; but see; Davis et al. 2017; Aplin et al. 2021). Researchers have used a variety of methods to attach tracking devices to birds, with alterations being made that are species-specific and objective dependent (Barron et al. 2010). Psittaciformes have proven to be a challenging group for attaching tracking devices, given their intelligence, dexterity and strong beaks (Meyers 1996). Despite the initial difficulties, parrot species including Macaws (Davenport et al. 2021; Brightsmith et al. 2021), Cockatoos (Yeap et al. 2017) and Parakeets (Theuerkauf et al. 2009) have been successfully studied with tracking devices, generating a new understanding of the species themselves and how they use the landscape.

Little Corella (Cacatua sanguinea) is a water-dependant granivorous parrot species widely distributed across Australia, which has expanded its range in southern Australia during European colonisation. In South Australia, Little Corellas were previously restricted to the north-eastern corner of the state, however, since the 1920s, Little Corellas have expanded their range southward, as well as dramatically increasing in abundance (St John 1994). Little Corellas’ ability to exploit the wealth of permanent water (e.g. dams, stock troughs, lakes) and food resources (e.g. grain crops, stock feeders, ornamental plantings) available in human-modified landscapes has enabled it to become an efficient urban adapter as well as an over-abundant nuisance native (Temby 2010).

Little Corellas are reported to over-winter in the city of Adelaide, as well as breed in the area, indicating that they are beginning to live in close proximity with humans year round (Scanlon et al. 2017). Their obvious presence through their flocking behaviour, excessive noise levels, chewing and damage of infrastructure and defoliation of trees contributes to most of the reported human–wildlife conflict with Corellas (Scanlon et al. 2017). It is therefore important to understand what is driving their presence in urban areas and how they utilise the landscape. Fitting tracking devices to wild bird populations is arguably the most effective way to understand movement patterns, including foraging ranges and potential ‘hotspots’ of human–wildlife conflict (Kennedy et al. 2015). This is the first published study to attempt the attachment of tracking devices to the subgenus Licmetis, collectively known as Corellas.

Here, we document the observations made after tracking devices were attached to Little Corellas, as we believe that future parrot telemetry studies can benefit from our experience. Similar to other bird researchers who have previously described harness failures (Steenhof et al. 2006; Lameris and Kleyheeg 2017), we suspect that unsuccessful attempts to attach tracking devices to parrots are rarely reported. As part of our study, we present a literature review of parrot transmitter attachment studies. We aim to demonstrate the reasons for the harness failure we observed and to highlight that psittaciformes are a relatively understudied group in the practice of wildlife telemetry, requiring innovative thinkers to diversify attachment styles to enable effective tracking of parrots.

Methods

Capture method

We captured Little Corellas (hereafter Corellas) in Keith Stephenson Park (35° 04’ 14” S, 138° 51’ 21” E) in Adelaide, South Australia, over a 3-day catching period in March 2022, during the Corellas’ non-breeding season. The catching site was selected based on ease of access, >500 Corellas present, and food being scarce elsewhere in the local landscape.

Corellas were attracted to the catching site by free-feeding with commercial corn and millet (bait) over 1-week. Birds were then caught using a baited cage walk-in trap, that was set up for 3-hour catching sessions each day. The walk-in trap consisted of an enclosed mesh holding pen that is accessible by a funnel shaped entrance. A Corella was able to easily move from the wide exterior section of the funnel through the narrow constriction into the trap, but the reverse process was difficult, ultimately capturing the bird (Figure 1a). Once three birds had been captured or a single bird had been in the trap for 5 minutes, an observer would open the small doors in the top of the cage and manually remove the birds (one at a time) wearing protective wildlife handling gloves. All bird handling followed protocols approved by University of Adelaide Animal Ethics Committee (S-2020-095) and was conducted under a Department for Environment and Water Scientific Research Permit (Y27074–1) and Mount Barker Council approval to conduct activities on council land.

Figure 1. (a) The walk-in cage trap which successfully captured Little Corellas after 1 week of pre-feeding at the capture site. (b) GPS transmitter harness removal aided by a neighbouring Corella. The harness is difficult to see as the Corella with the transmitter (left) has preened its feathers over the harness; the adult Corella without a transmitter (right) is chewing the Teflon strap looped around the left wing (photos: Rowan Mott).

Transmitter attachment

We used Ornitela UAB (Vilnius, Lithuania) OrniTrack-10 GPS loggers, weighing 14 g (<4% of the body weight; captured Corellas weighed between 428 and 486 g). These tracking devices were chosen because they transmit data through the telecommunications network allowing for remote data download without the need to recapture or be in close proximity to the birds.

A harness incorporating a single weak link was attached to the Corellas using the ‘Backpack and Keel Strap’ method described by Kenward et al. (2001). Teflon tape (6.5 mm wide) was used to form the harness (Bally Ribbon Mills, USA) and mount the GPS tracker on the back of the bird. The tape was looped over the shoulders and under the wings and the two harness loops were secured with the keel strap. The keel strap was sewn using cotton stitching (single weak link) to ensure the transmitter fell off the bird (after c. 1 year). The loose tape from each harness loop was clamped using a compression clamp at the rear of the transmitter once the desired fit was achieved and the excess material was trimmed. Five GPS transmitters were fitted in total. Attachment took place on site after capture, with the total process taking c. 15 minutes per bird.

Literature search of parrot telemetry tracking

To evaluate the transmitter attachment styles used on psittacine species, a literature search in Google Scholar and Web of Science was conducted during April 2023 using the following search terms: (parrot OR psittacine OR cockatoo) AND (tracking OR GPS OR telemetry OR transmitter). The literature chosen contained all studies that used any electronic tracking device on parrots (including VHF-transmitters, PTT satellite transmitters and GPS transmitters) using any attachment method (including harness attachments, neck collars, back-mounts and tail-mounts). Our search resulted in 46 relevant studies, including three studies that were suggested during the peer review process (see Additional file 1 for the complete list). This list of parrot tracking studies provided a representative sample to draw conclusions for the common tracking methods previously used on parrots. From this set of studies, the following details were extracted (Additional file 2): study species, bird body mass, attachment performed in captivity or the wild, tracking device type, tracker attachment style, tracking device company, tracking device mass, materials used, retention time and whether removals were reported. Additionally, 15 wildlife transmitter manufacturers identified from the literature or recommended by experts were contacted (via email or inquiry forms on their webpage) to further understand the technology available, the experience they have had with parrot species and their willingness to trial custom solutions. This allowed us to obtain expert advice, relating to tracking devices attached to parrots, which may not have been published.

All data analyses and visualisations were conducted in the R software environment for statistical and graphical computing version 4.0.5 (R Core Team 2021). The influence of tracking device mass (as a percentage of parrot body mass) on the probability of attachment removal was tested using a generalised linear model with a logit link function and binomial errors.

Results

Harness removal

Of the five Corellas that had transmitters attached, we directly observed all five actively trying to remove their trackers immediately after deployment. The location at which the harness attachment failed varied among the five harnesses. In one instance the corella removed its harness 2-hours after attachment, and it was found under a tree, severed at the weak link. The transmitter attached to the harness was intact with no bite damage, so it was then refitted to another individual Corella on the next day of catching. The other four harnesses were removed within a 24–48-hour period with limited GPS tracking data being collected and visualised (Additional file 3). Harnesses were removed by Corella biting through the weak link (n = 3), or by chewing on both the back of the harness at the base of the transmitter and the keel strap (n = 1). The remaining transmitter was recovered after being run over by a vehicle in a neighbouring carpark, so the location of detachment could not be deciphered; but it can be assumed it was removed by the bird (n = 1).

Notably, less than 30-minutes after transmitter attachment, an individual Corella fitted with a harness was observed in the canopy trying to remove the transmitter. The bird was then approached by a neighbouring adult Corella (without a transmitter) which began to aid in removal of the transmitter by chewing on the Teflon tape looped under the wing of the harnessed Corella (Figure 1b). The behaviour continued until the Corellas left the area.

History of telemetry tracking of parrots

Our meta-analysis of 46 research articles highlighted that tracking studies of psittacine species have increased in frequency since 1987, when the first study was published (Snyder et al. 1987). The device type has changed as new technologies have been developed (Additional file 4) and the variety of attachment methods has remained relatively consistent over the years (Figure 2a). Neck collars, the first method used to attach tracking devices, are still the most common method (38.8%), followed by backpack harnesses (24.5%). Tail-mount attachments (16.3%) have been more frequently used in the last 10 years. The combination attachment of using two tracking devices and attachment methods, namely a GPS back-mount and a PTT tail-mount, was first developed by Yeap et al. (2017), and has subsequently been utilised by other researchers (12.2%). Back-mount attachments have been used the least in parrot telemetry studies (8.2%).

Figure 2. (a) The attachment style used in telemetry studies of parrot species since 1987. (b) Removal rate as a function of attachment style used in parrot telemetry publications. (c) The number of studies that reported the removal of a tracking device as a function of parrot body mass, compared to the total number of parrot tracking studies in each mass range. (d) The influence of tracking device mass as a percentage of parrot body mass on the probability of attachment removal.

Thirteen of the parrot tracking studies (28.3%) documented removal of tracking devices by birds chewing on the attachment and detaching the device or chewing on the transmitter, making it inoperable. Conspecific removal of trackers was not previously reported. Studies that used the combination attachment method reported the highest rate (66.7%) of removal, followed by studies using backpack harness attachments which reported removal of devices in 50% of the studies. In published literature using tail-mount attachments, 25% documented attachment removal. Similarly, 21.1% of studies using collar attachments reported tracking device removal. There were no reported removals of back-mount attachments in the reviewed parrot telemetry studies (Figure 2b).

Body mass of parrots also influenced the likelihood of removal of tracking devices by the species (Figure 2c). Medium to large sized parrots, in the weight range of 500–1000 g, removed their tracking devices the most (53.3%). All the study species in this weight range, apart from one (Blue-throated macaw Ara glaucogularis), were in the family Cacatuidae (commonly known as Cockatoos).

Transmitter mass as a percentage of body mass ranged from 1.1% to 7% for parrot telemetry studies, with five studies being outside the recommended 5% ‘rule’ when fitting devices (Bodey et al. 2018). Of the studies that documented transmitter removal, device percentage of body mass ranged from 1.1–4.3%. There was no significant influence of the tracking device mass (as a percentage of parrot body mass) on the probability of attachment removal (Figure 2d; slope estimate = −0.16, effect size (z value) = −0.57, p = 0.57).

Of the 15 wildlife transmitter manufacturers contacted, eight mentioned past success with attaching tracking devices to parrot species. However, many were sceptical about trialling custom solutions due to concerns that the Corellas’ beak would damage the transmitter body or antenna. Additionally, the companies acknowledged that the size of Little Corellas presents further difficulties, as they are too small to carry heavy reinforced metal attachments commonly used on larger parrot species such as Macaws.

Discussion

Prior to this study, there was no published record of Corella species being fitted with a tracking device. We were able to develop an effective method to capture Corellas, although our harness technique to attach a tracking device was unsuccessful and our harnesses were quickly removed by the birds. Here, we discuss the immediate and long-term implications of these observations for Corellas and parrots.

Harness removal

In urban areas, Corellas are frequently observed chewing infrastructure like wiring, antennas and solar panels (Temby 2010). Similarly, Kea Nestor notabilis are well known for chewing and causing bite damage to human-made objects, creating human–wildlife conflict (Diamond and Bond 1999). Yet Kennedy et al. (2015) demonstrated that Kea could be fitted with a backpack style harness and GPS transmitter. Unlike Kea, the Corellas in this pilot study are an over-abundant native species, living in an urban environment, which vastly increases their survival (Scanlon et al. 2017). The relative safety that Corellas receive in urban areas may have allowed them to devote more time to non-critical tasks such as harness destruction, allowing them to rapidly remove the harnesses within a 48-hour period.

A significant observation of our study is the chewing and pulling on the Teflon harness by a conspecific. Australian Magpies Gymnorhina tibicen, are the only other bird species where removal of a tracking device by an unharnessed individual has been documented (Crampton et al. 2022). Little Corellas are a highly social species in which a high degree of plasticity to novel objects occurs (Auersperg et al. 2015). As such, the observation may have resulted from adaptive cognition, whereby social groups that are naturally faced with cooperative challenges use social problem solving in new situations (Klump et al. 2021; Crampton et al. 2022). Alternatively, social play may have been the catalyst for the observation, where the individual pulling and prying on an object attracted a neighbouring bird to join in, a type of behaviour seen in highly social species, including Corellas (Kaplan 2020) and Kea (Diamond and Bond 1999).

Our review found no evidence for systematic differences in the likelihood of a transmitter being shed between cohorts of birds (e.g. age classes, sexes, breeding status). However, the sample size of reviewed studies was relatively small. Our observation of one Little Corella seemingly assisting another to remove a harness could mean that individuals in breeding pairs are more likely to be able to shed a telemetry device due to assistance from the partner bird, for example, through an extension of allopreening behaviour. If biases such as this do occur this would have implications for study design (e.g. avoiding tracking during breeding periods) and study interpretations (e.g. recognising that movements of non-breeding individuals may contribute disproportionately to inferences made from a tracking dataset). Future studies should report suspected biases in devices lost between cohorts to improve understanding of the factors that lead to successful parrot telemetry research.

History of telemetry tracking of parrots

For a bird group that is known for their intelligence, dexterity and strong beaks (Meyers 1996), there are seemingly low rates of tracking device removal. We suspect that this might be a result of unsuccessful attachment attempts not being reported and suppression of failed studies (Mlinarić et al. 2017). This emphasises the importance of not only reporting attachment failure, but also how the removals occurred (Steenhof et al. 2006).

The review highlighted that the combination attachment method of a back-mount and tail-mount had the highest removal rate by parrots among the different attachment styles. This style has only been attached to Black Cockatoo species Calyptorhynchus spp. Importantly, the back-mount had a lower retention time and was removed more often than the tail-mount which was left to moult out, the removals were attributed to the birds’ inquisitive nature, the tags prominence on the back and preening behaviour (Yeap et al. 2017; Rycken et al. 2022). This group also weigh between 500–1000 grams, which is the weight range that reported the highest number of transmitter removals. Furthermore, tracking device mass relative to the bird’s body mass did not influence the likelihood of transmitter removal. This highlights that the powerful bills of medium–large-sized birds are likely the reason for most attachment removals and researchers do not improve their chances of success by attaching relatively small devices, which are often more expensive. This has implications for project planning and budgeting.

Despite Corellas detaching our backpack harnesses and harnesses having a high removal rate in comparison to other attachments, previous research has shown that harnesses can be a successful attachment style for large parrot species (Le Souef et al. 2013; Kennedy et al. 2015; Aplin et al. 2021). Researchers may see harnesses as a more robust style and hence these might be more commonly trialled on larger species resulting in them being removed more often. Also, another reason for the prevalence of harnesses may be the requirements for specialist expertise (i.e. anaesthetist) for tail mounts. Blue-Throated Macaws Ara glaucogularis destroyed their harness attachment in minutes (Davenport et al. 2021). An alternative metal collar attachment, weighing 31 grams, was highly successful (Davenport et al. 2021). The species-specific transmitter design further highlights the challenges researchers still face when fitting tracking devices to parrot species, being limited to expensive options and small sample sizes (Davenport et al. 2021; Brightsmith et al. 2021). When samples sizes for parrot tracking studies are small, device removals can impact the precision of the results (Murdoch 2012).

Conclusion

Although transmitters have been successfully attached to other large, strong-billed parrots using backpack harnesses, this technology and style of attachment may not be suitable for Little Corellas, especially when their cooperative abilities are considered. Our literature review indicated that combination attachments and backpack style harnesses have high apparent removal rates. Also, species weighing between 500–1000 grams have higher rates of removal, and this is evident in Cockatoo species. Nonetheless, we still advocate for more parrot tracking studies and encourage continued collaboration and innovation of technology and attachment methods. As parrot telemetry becomes more widespread, we strongly encourage researchers to reliably report the failures and any deleterious effects of attaching tracking devices to Psittaciformes (see Bodey et al. 2018).

Acknowledgements

We acknowledge the Kaurna and Peramangk people as the Traditional Owners of the Land on which this research was conducted. We thank D. Whittaker, and K. Marker for their assistance in field work. This research was approved by the University of Adelaide Animal Ethics Committee (Approval Number S-2020-095) and The Department for Environment and Water (Scientific Research Permit Number Y27074-1).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/01584197.2024.2352367

Additional information

Funding

This research was financially supported by Landscape Boards South Australia: Hills and Fleurieu, Green Adelaide, Murraylands and Riverland; and South Australia Department for Environment and Water.