Our photographic survey design was developed to provide maximum coverage of Izembek Lagoon on a single flight without having to refuel the aircraft, thus maximizing efficiency and minimizing the time window during which bird movement might confound the counts. We used a systematic sampling design with regularly spaced points taken along 49 transects over the open water of the lagoon. In 2017, we used 5,468 planned photo locations over the open water of the lagoon; in 2018 and 2019, we expanded some transects to cover the full width of the lagoon, resulting in 5,649 photo locations. We used a pair of cameras, described below, that were triggered automatically and simultaneously to obtain 2 non-overlapping images at each photo point, one on each side of the transect line. Our total potential sample was thus 10,944 (2017) to 11,298 (2018 and 2019) photos per survey replicate that would cover a combined footprint of 14.6–15.1% of the area of the lagoon. We aimed to center each replicate (spanning about 4 hours) on low tide, when we expected geese to be mostly stationary while feeding on exposed eelgrass beds. However, greater flexibility in timing could allow more replicates to be completed during suitable weather windows (which can be scarce in this study area), so we also completed one replicate at high tide to evaluate the consequences for the resulting population estimate.
Photo survey replicates were typically conducted on different days than the ocular replicates, but still within the fall staging period (late Sept to Oct). We conducted the photo flights in an amphibious Cessna 206, modified with dual belly camera ports, on days when the cloud ceiling was above 460 m with no more than occasional light rain to allow good views of the lagoon. Ground speed averaged 165 km/hr, and we chose a target altitude of 457 m ASL to avoid disturbing geese.
For each photo survey replicate, a camera operator was on board in addition to the pilot. We implemented our flight plan with the Aviatrix Flight Management System (AeroScientific, Blackwood, SA, Australia), which provided visual flight-line guidance to the pilot and automatically triggered cameras (through Aeroscientific Trigger Box 2018-011) when the aircraft passed within 150 m of preset locations. The system included a ublox-7P GNSS module to record latitude and longitude (expected accuracy ±5-10 m) and altitude (±15 m) in a log file when each photo was triggered. The camera operator monitored one screen showing the flight line and status of each camera trigger point, and another on which photos could be previewed to ensure camera settings were appropriate for the current light conditions.
Our camera setup consisted of two Canon EOS 5D R bodies equipped with 24 × 36 mm 50.6-megapixel sensors and either Canon EF 70-200mm f/2.8L IS III USM lenses (2017) or Canon EF 200mm f/2.8L II USM prime lenses (2018-2019). As part of the testing phase for this monitoring program, subsets of photos from 2017 were captured at nonstandard altitudes (<350 m or >550 m) or focal lengths (<200 mm), thus affecting the size and resolution of geese pictured in photos. We mounted the cameras inside the aircraft such that the distal ends of the lenses were just inside portholes in the aircraft body. We set the cameras to manual mode with focus set near infinity, aperture 2.8 to 4.0 depending on current conditions, shutter speed 1/5000 to 1/8000 sec, automatic ISO, and exposure compensation -2.0 to -2.3 to avoid overexposing birds against the dark water. Each camera saved .JPG images to a CompactFlash and/or SD card (about 200 GB per replicate). To avoid overlap among photos, we tilted each camera away from the vertical axis (plumb line) by approximately 5.6° and set camera trigger points at spatial intervals of 62 m, which corresponded to about one photo every 1.35 sec at our average ground speed of 165 km/hr. With the specifications of our camera sensors, 200 mm focal length, and our target aircraft altitude of 457 m, the expected footprint of each photo was 4555 m2 (83.1 m wide by 54.8 m parallel to the path of the aircraft).