Abstract
Normandeau was tasked to design and test a system combining thermal imagery and acoustic and ultrasound sensors to survey bird and bat species potentially affected by offshore developments. Monitoring birds offshore has been limited worldwide due to difficulty of access and high cost. Boat-transect surveys and “ships of opportunity” are subject to potentially large sampling error and are too limited in scope to provide sufficient information. Traditional visual aerial surveys are expensive and also subject to substantial sampling error. An effective and economical way to monitor bird presence offshore would be to use specially designed, strategically positioned and remotely operated acoustic microphones and thermographic cameras attached to offshore structures such as meteorological towers, oil and gas platforms, or wind turbines. Acoustic microphones and thermographic cameras could monitor vocalizations of birds both day and night at all seasons of the year and in any weather conditions including periods of low visibility that would prevent effective visual monitoring. This report describes the initial development of the system and the results from test deployments.
The Acoustic Thermographic Offshore Monitoring (ATOM) system is designed to gather data through all weather conditions both day and night. Deployment at the remote Frying Pan Shoals Light Tower (FPSLT), 29 mi offshore (Figure ES1), provided a challenging work arena with costly and limited access to the system for installation, maintenance, repairs, and retrieval. The restricted ability to access the system increased system down-time, delaying repairs and increasing the cost of deploying and maintaining the system. These factors forced subsequent improvements and modifications to both hardware and software to create a more robust unit that was able to withstand harsh offshore conditions.
A number of seabird species, including gulls, terns and frigate birds, were expected to occur during the offshore deployment and were identified as expected by ATOM. The dataset of land bird species identified by ATOM is a significant contribution to filling gaps in knowledge about these migrants, and includes herons, bitterns, and many passerines. The data show a clear pattern of migrant occurrence in the offshore environment, with April and October showing peak density, using combined acoustic and thermographic data. Peak in fall density of migrating birds occurred during periods of north to northwest winds (i.e., with a tail wind). Flight bearing in passerines showed seasonal differences but similar trends were not evident with non-passerines. Passerines showed strong tendencies to fly to the south and southeast during the fall and to the northwest during the spring. Mean flight direction during Apr was 286º (NW) and in Oct was 151º (SSE).
Most birds appear to fly higher in the evenings with an estimated 1.8 times increase in flight height between 8 PM and 12 AM than at all other times. Flight altitude seems unaffected by wind speed. Instead, from both acoustic and thermographic data, there is more bird activity during wind speeds of less than 10 km/hr with no discernable alteration in altitude. Flight direction is affected by wind speed and direction with data showing birds inclining to fly into head wind. Flight speed data are consistent throughout the year as well as throughout the day with an average speed of 23 km/hr.
Bats were not recorded at FPSLT. Although there were large data gaps in the ultrasonic data, no bats were seen in the thermographic data either. Bats have been encountered this far offshore and away from any terrestrial habitat; however, it is unlikely that they occur at remote stations like FPSLT with any regularity.
The system is designed to survey birds and bats within the rotor swept area of a turbine, and consequently most flight altitude data are within this detection area. Acoustic data also fill information gaps on small birds flying higher than 150 m that might otherwise be missed by thermographic methods due to the decay in detection over distance for small birds. Information from these two detection methods provides new data on peak migration times for both vocal and silent species.
Although an original goal was that ATOM would give species-specific information on flight altitude, velocity, and bearing, sufficient data were not collected that would match many species level identifications with all detectors. Increased system reliability should augment the amount of data that could be matched, and longer deployment would gather more data from all sensors. However, species-specific data collected show Yellow-rumped Warbler with flight altitudes of 103.9 m and 46.3 m (n=2), and Laridae with flight altitudes ranging from 49.1 m to 193.9 m, mean 87.43 m (n=35).
The results presented in this report are evidence of progress in the use of acoustic and thermographic monitoring to understand the ecology of large-scale migrations and apply that knowledge to conservation planning. Particularly novel is the dataset itself, the first of its kind from the offshore environment in the western Atlantic Ocean.