Author: Melanie Steinkamp, Wetlands International, 4401 North Fairfax Drive, Room 730, Arlington, Virginia 22203, melanie_steinkamp@fws.gov, 703-358-1953

Pelagic Surveys
Boat Surveys
       Purpose
       Parameters
       Survey Design
       Data Recording
Beached Bird Surveys
Systematic Reconnaissance Flights

For some pelagic species, the population cannot be counted at the breeding colonies, or information on species needs outside of the breeding colonies is desired. For these birds, counts at sea can be conducted by air or boat. While open ocean surveys are beyond the scope of most cooperating land and water managers, in some cases, they are the only method available to allow managers to gain insight into seabird population status and trends. Additionally, managers are beginning to recognize the importance of changes in ocean habitats, whether caused by shifting ocean regimes, or commercial and/or recreational fisheries activities on marine and coastal ecosystems. In the past decade, it has become increasingly important to use pelagic surveys to track ocean habitat use by species whose populations appear to be declining, such as the Black-capped and Bermuda Petrels, but also for tracking concentrations of species, such as shearwaters.

There are a number of problems associated with detecting seabirds in the open water. Tasker et al. (1984) divided detection problems into five interrelated categories: size, color, behavior, weather, and observer ability. Each of these factors, by themselves or in some combination, can result in significant bias and variability in counts. For example, larger birds are easier to see than smaller birds at some set distance; a storm petrel, whose wing span is one-fifth that of sulid is much harder to detect. A dark colored bird such as a murre, in a dark sea or a kittiwake against a light sky, have less chance of detection than a light colored bird against a dark sea. The behavior of particular species can make them easier or more difficult to detect; procellariformes, who spend more time flying or moving through an observer's plane of view will be easier to detect than an auk, who spends portions of the time diving or stationary on the surface of the water. Additionally, flocking birds are easier to detect than species who are solitary most of the time.

It has been documented that the presence of the ocean vessel can influence species' behavior and bias its detectability (Bailey and Bourne 1972). Some species, such as Northern Fulmars, albatrosses, and Tufted Puffins, are attracted to ocean vessels while storm petrels and other species avoid ships. Weather may influence an observer's ability to detect birds; winds generating waves may obscure birds sitting on the water from the view of an observer. Sun, glare, and fog may limit may limit visibility. Weather may also affect bird behaviors; during calm conditions, birds may sit on the water instead of flying, decreasing their chances of detection. All of these factors, either singly, or in combination, lead to variability in counts and makes it difficult to interpret whether changes in numbers are the result of bias or true population change.

Perhaps one of the most important biases influencing counts of birds from ships is caused by birds flying through the total area surveyed. A count of all birds passing through a count zone within a count period will overestimate the total number of birds within that zone at any one moment by incorporating flux. Counting flux overestimates actual bird density. To account for this, coefficients of detection should be established for each observer per species. Correction factors derived from these coefficients of detection can be applied to the data.

Transect methods were developed to minimize many of the biases and variables discussed above. Transects reduce the area of sea examined at any one time so that a substantial proportion of the birds within the transect are detectable. However, most band-transect methods don't address the bias associated with flux, the movement of birds through a transect during a count period. Ideally, the observer should make an instantaneous count of all birds within the transect band. Tasker et al. (1984) has suggested that an instantaneous count is impossible to accomplish, especially at higher ship speeds. However, van Franeker (1996) reported on a snapshot method used for counting seabirds in the southern ocean. We recommend the snapshot method, below.

As with colony surveys, the pelagic survey design will depend on the objectives of the monitoring program; if quantitative results are desired, the program must attempt to control for unwanted systematic Standardization of methodologies must be adopted if data from different surveys are to be comparable.

The purpose is population indices for detecting trends, with an ability to detect a greater than or equal to 20 % change in the numbers of birds represented by open water transects.

Continuous counts of all stationary birds (swimming, sitting on ice, or actively feeding) within the transect limits and (2) snap-shot counts of all flying birds within the transect limits.

Birds should be observed from an observation post aboard the vessel which give the greatest angle of clear view and permits the additional use of sound to detect less conspicuous birds. Ten minute blocks have been the standard count period over the past decade for many pelagic surveys. Count periods longer than this will make it difficult to record changing conditions with bird numbers. Birds within a 150 meter transect on one side of the ship should be counted in 10 minute blocks of time. Although many surveys have used a 300 meter transect, previous work has reported difficulties in detecting small, inconspicuous birds at distances of greater than 150 meters (Briggs et al. 1985; van Franeker 1994). A range finder can be used to determine this distance. Detection of birds should be done with an unaided eye; binoculars (10 x 40 suggested) should be used for species confirmation and for aging or looking at plumage. Counts should be made: (1) continuously of all stationary birds (swimming, sitting on ice, or actively feeding) within the transect limits and (2) in a snap-shot fashion for all flying birds within the transect limits. The speed of the ship determines the forward limit of the snapshot area within a range of 150 meters. Longer or shorter forward distances are avoided by adapting the frequency of the snapshot counts. Birds following and circling the ship should be omitted from both snapshot and continuous counts. If birds arrive and then follow the ship, they should be included in the count only if their first sighting falls within a normal snapshot or continuous count of the transect area.

Ship Information: Information on the ship's position, course and speed, and the starting time of observation should be recorded. Also recorded should be the height of the "eye" above the water and viewing arc. Environmental factors must be recorded such as wind speed and direction, cloud cover, barometric pressure and tendency, precipitation type and intensity, visibility, sea state, swell height and direction, air temperature, and an assessment of the sun's effect on the observation areas, based on the strength of the sun and its direction relative to the direction of viewing. Observation notes should be repeated at least every 100 minutes, with major changes noted as they occur.

Species Information: For each bird observation record species, number of individuals present, activity, plumage and age of bird where possible, approximation of bird's flight direction, and notes on other distinguishing characteristics, such as whether a bird was oiled.

Target Population: number of birds on the sea in a continuous count and a snapshot of the birds flying within a defined band transect.

Analysis: Densities may be calculated as the sum of stationary plus snapshot birds per surface area.

(To be added)

(To be added)

Other Methodologies to be added….