This report was prepared by Peter Frost and is supplementary to POP2021-03 Seabird population research: Chatham Islands. It is an aerial census of the northern royal albatross, involving photographs taken from a fixed wing aircraft and counted, and comparison of plots on Motuhara with ground counts and drone imagery.
An aerial photographic survey of Northern Royal Albatross | Toroa Diomedea sanfordi nesting on two groups of privately-owned islands, Rangitautahi and Te Awanui (both islands in the Rangitatahi/Sisters group) and Motuhara, was carried out on 1 February 2022. Together, these colonies hold >99 % of the global populations of this species.
The survey was carried out at the end of incubation–start of the chick brooding and guarding phase because poor weather prevented a survey being carried out at the start of incubation.
During the flight, 1295 photographs were taken from which 71, together covering all of each island, were selected for broad analysis. For each island, the selected images were partitioned into contiguous sections with no overlaps or gaps. In each section, all Northern Royal Albatrosses present on the ground were classified by their apparent behavioural state and tallied. A total of 3546 individuals (95% CL: 3272–3848) were counted across all three islands, of which 3257 were judged initially to be either still incubating eggs, brooding recently hatched chicks or guarding chicks more than a few days old. No lone chicks were seen.
A further 22 non-overlapping close-up images from all three islands were then selected and the birds clearly visible in them classified by behavioural state. Omitting those birds whose status was uncertain, the proportions of individuals clearly occupying a nest, standing or sitting around as a partner, or loafing, were then calculated for each island. These proportions were then used to adjust the initial numbers categorised and counted.
When these proportions were used to correct the initial counts, 3269 birds (95% CL: 3018–3547) were judged to be incubating eggs or brooding/guarding chicks, distributed as follows: Motuhara 1601; Rangitautahi 993; Te Awanui 675. This total is less than the average number of birds apparently occupying nests (AON) in 2016–2021 (4229 nests), but higher than the average number of chicks counted between 2016 and 2019 as being close to fledging (2103 chicks).
In addition to the images of Motuhara taken from the survey aircraft, a further 221 images, taken on 31 January 2022 from a DJI Mavic Air 2 drone, were received. Ten of these were wide-angle views of the western half of the island, whereas the rest were taken on 10 more-or-less parallel transects aligned approximately along the long axis of the island. Among this set, those images containing five recently established 20 x 20 m Northern Royal Albatross study plots and five of ten 10 x 10 m Northern Buller’s Mollymawk monitoring plots (some of which dated back to 2007/2008), were selected for detailed analysis. The number of active nests in these quadrats, which had been surveyed on the ground a couple of days earlier, were then identified and counted. Images taken of these quadrats from the aircraft were also assessed and the number of apparently nesting birds compared with the drone and ground counts.
In general, the average of these image counts for both species compared favourably with those made on the ground. Some differences stood out, however, with both over-counting and under-counting evident (range +7% to -8%), especially among the drone images. Apparent problems were in accurately deciding which birds were sitting on nests as opposed to standing or sitting around when viewed directly overhead in the drone images, and in determining if a bird sitting on a nest close to the superimposed boundary line was inside or outside a quadrat when viewed obliquely on the aircraft-derived images.
Despite the various uncertainties involved in surveying seabird populations from aerial photographs or drone-based imagery, there is little alternative at this stage for monitoring species such as Northern Royal Albatross and Northern Buller’s Mollymawk nesting on remote, difficult-to-access offshore islands. Continued monitoring of the survey quadrats set out on Motuhara for these species, both on the ground and from the air, can complement the wider coverage achieved through aerial survey, but further development of this approach is needed, particularly in ensuring that these quadrats are representative of these larger populations.
To ensure comparability of ground counts and those obtained from aerial images, taken either from an aircraft or a drone, a protocol is needed as to which individuals nesting on the quadrat boundaries should be included in the count.
While recognising the difficulties involved in doing this, conducting regular, twice-yearly, coordinated aerial and ground surveys of breeding Northern Royal Albatross—ideally in December, soon after egg laying has been completed, and in August, just before the chicks fledge—would enable more robust assessment of this species’ population dynamics.