Draft technical Report 2: Assessing the distribution and abundance of nonbreeding pelagic and near-shore waterbirds

Stephanie R. Schmidt

Daniel B. Breen

Katharine C. Parsons

generous support from

Sport Fish and Wildlife Restoration Programs of the U.S. Fish and Wildlife Service

and the American Bird Conservancy

29 Dec 2008

Technical Report 2: Assessing the distribution and abundance of nonbreeding pelagic and near-shore waterbirds
Introduction

Pelagic and near-shore waterbirds can be defined as any species of bird that is adapted to spend the majority of its life out at sea, returning to shore only during the breeding season. Common at-sea birds include loons, shearwaters, kittiwakes, petrels, murres, grebes, and phalaropes, among other species. Little is known about at-sea-birds during the non-breeding season, and universally standardized pelagic surveys of seabird distribution and abundance are still largely in the development process. Data on the near- and off-shore distribution and abundance of off- and near-shore seabirds are critical for understanding their basic ecology and role in marine ecosystems, monitoring population trends, and assessing threats to their population.

A standardized sampling scheme for pelagic seabird surveys that stratifies for ocean currents and major populations of zooplankton has been developed for use in the southern hemisphere (BIOMASS 1992). More research is still needed to effectively stratify for these variables in the northern hemisphere (BIOMASS 1992). Wintering seabird populations can be located by looking at previous surveys, consulting with experienced biologists or fishermen, searching through the published literature and any maps of migration patterns, and tracking birds that have already been fitted with electronic tags (BIOMASS 1992; Bart 2006). Wind currents, gyres, known areas with high concentrations of prey, and marine productivity, which can be measured by detecting amounts of chlorophyll-a, can also be used to identify the most ideal wintering habitats for seabirds (Gonzalez-Solis et al. 2007). Once a seabird population has been located, it is recommended to sample at least 80% of the birds and to use a simple or stratified random sampling technique to reduce environmental biases (Nur et al. 1999). A minimum of 30 observation points per habitat is recommended (Nur et al. 1999).

Pelagic seabirds are most commonly surveyed from vessels at sea during the non-breeding winter months (Kendall and Agler 1998; Holm and Burger 2002; Langen et al. 2005). Surveys should be conducted at similar tides and should begin 45 minutes prior to the peak of each tide phase to ensure that they are completed in time (Holm and Burger 2002). The vessel should travel along transects that are 300 m apart, starting at about 200 m offshore, with observers using binoculars and/or a spotting scope to record seabirds within 150 m of each side (Tasker et al. 1984; Gould and Forsell 1989; BIOMASS 1992; Holm and Burger 2002). Three consecutive, 10-min transects are recommended at a speed of about 10 kn, as higher speeds can make counting difficult while lower speeds can inflate the count (Gould and Forsell 1989). Wind speed should be below 28.5 km/hr, and waves should be no higher than 0.6 m so that they do not affect visibility or seabird movement patterns. Speaking into a tape recorder is recommended so that the observer does not take his/her eyes off the ocean to write down observations, while multiple observers with experience at identifying seabirds should be present to improve accuracy (Briggs et al. 1985). Observers should record all stationary and feeding birds at a 90^o angle from the front to the mid-ship, as some birds may fly away or dive underwater as the ship approaches (Gould and Forsell 1989; Kendall and Agler 1998). Results can be presented in terms of relative abundance per unit time or per distance traveled (Van Franeker 1994).

When surveying, observers should record all stationary and feeding birds but use a different counting technique for flying birds (Gaston et al. 1987; Gould and Forsell 1989; Van Franeker 1994). Repeat counts of flying birds and the attraction of birds to the boat can lead to an overestimation of the population if they are not factored into the protocol. Large flocks and ship-follower birds should only be counted once. An instantaneous “snapshot” count is recommended when surveying flying birds to reduce the biases caused by flight patterns, ship-following and avoidance, and patchy distribution patterns, particularly among smaller seabirds such as storm-petrels (Gaston et al. 1987; Gould and Forsell 1989). Ten instantaneous counts should be made during each 10-min observation period. Ship-follower birds should be recorded separately as “recounts” and not tallied multiple times (Powers 1982; Haney 1985). In addition to instantaneous counts, other supplementary methods can include skiff counts, in which all birds in one direction to a specified distance are recorded, or ship-follower counts (Gould and Forsell 1989).

Poor visibility is another variable that can generate biases and decrease the accuracy of seabird counts. Visibility can be compromised by wind direction, wind speed, height and direction of swell and surface waves, cloud cover, position and strength of the sun, position and height of the observer(s), number of birds per group, proximity of birds to each other, and size and coloration of birds (Dixon 1977). Observations may be divided into categories based on wind speed- calm, moderate, and rough- and analyzed separately (Dixon 1977). Surveys should also be conducted at similar tidal stages and weather conditions with clear visibility to eliminate other variables. Photographs can provide a lasting record and eliminate inconsistencies between observers.

In addition to visibility and bird movement, a few additional factors should also be taken into consideration. Because several types of seabirds, including fulmars, kittiwakes, and shearwaters, are attracted to fishing vessels, care should be taken to avoid known fishing areas so that the population estimate will not be inflated (Powers 1982). Surveys should not be conducted from a fishing vessel if possible, and any surveys from fishing vessels should be analyzed separately. Overall abundance is often considered more important than species identification, and visibility problems normally cause many individuals to be identified only to genus (Gould and Forsell 1989; Holm and Burger 2002). Finally, the unpredictable activity patterns of seabirds and frequent change in distribution patterns can make scheduling surveys difficult and cause surveys to be biased towards certain species (Gould and Forsell 1989). It is therefore recommended to repeat surveys over multiple years to identify long-term population trends (Kendall and Agler 1998).

For improved accuracy, vessel surveys may also be supplemented by land-based scan surveys performed from a clear observation point on an island (Holm and Burger 2002). Each land survey should involve instantaneous scans at 15-minute intervals looking through a 10-45 power spotting scope (Holm and Burger 2002). Land surveys should also be conducted during the non-breeding season and are particularly recommended when surveying in narrow or small channels where vessel surveys would be difficult (Holm and Burger 2002). Wind speed should again be below 28.5 km/hr, and waves should be below 0.6 m in height so that they do not impair visibility. Bird species, bird activities, water depth, and water types should be recorded at each of the three tidal stages: maximum flood, maximum ebb, and slack water (Holm and Burger 2002).

Another seabird survey technique that has gained prominence in recent years is tagging technology (Gonzalez-Solis et al. 2007; Burger and Shaffer 2008). This method involves attaching an electronic tracker to a bird during the breeding season, recording data for one year, and then typically recapturing the bird and removing the device (Gonzalez-Solis et al. 2007). These electronic trackers can provide unprecedented information for monitoring long-term demographic patterns of seabirds and locating populations to study. They can also be more accurate than vessel counts, which rely heavily on estimates and can be influenced by the presence of the vessel. The diverse types of tagging devices include geolocators, satellite telemetry, GPS loggers, and depth recorders (Burger and Shaffer 2008). Geolocators detect changes in ambient light to estimate the time of day and can also record temperature to improve their location accuracy. They consume minimum power because data are stored rather than transmitted. Satellite telemetry can more accurately document movement patterns away from the colony, but it produces more varied results. GPS loggers record a precise location every minute that is accurate within meters and reveals travel speed. Lastly, depth recorders record water pressure, temperature, and heartbeat to determine diving depths and give a comprehensive foraging profile, but the bird must be recaptured to receive these data (Burger and Shaffer 2008).

The advantages of electronic tags include an ability to efficiently track a bird’s migration patterns on a global scale, allow observation of underwater activities and from greater distances, provide precise latitudinal and longitudinal coordinates, and eliminate observer biases. Tagged seabirds can also serve as ocean sensors that provide insight into other important biological and physical processes (Gonzalez-Solis et al. 2007; Burger and Shaffer 2008). However, carrying electronic devices, especially the larger appliances involved in satellite telemetry and depth detection, can: impair bird movement, diving ability, foraging ability, and consequently health. It may also increase stress, reduce colony attendance, and decrease the probability of future reproduction (Burger and Shaffer 2008). Tagging technology can also be quite expensive, exceeding $20 per day per tag (Burger and Shaffer 2008), and may be beyond the budget of many seabird population studies. As newer and more miniaturized technologies are being developed, their smaller size and greater efficiency is expected to reduce many of these drawbacks.

Species Protocol Table:

The attached Species Protocol Table is a comprehensive reference to surveying pelagic and near-shore waterbirds. The first column in the table lists each species of pelagic or near-shore waterbird that is in the North Atlantic. The remaining columns cover each habitat type. Relevant field protocols for each species in each habitat type are then listed in the table. The numbers in parentheses after each survey method correspond to a different primary source. The full citations of each primary source are then listed in the following worksheet labeled “Sources.”

Data interpretation

A thorough analysis of the data is important to accurately assess seabird population trends. Density can be calculated by dividing the estimated number of birds by the area of the transect, and confidence levels can then be estimated from the sum of the variances (Spear et al. 1992; Kendall and Agler 1998). Density calculations should be performed separately by transect, depth, and season to control for temporal and spatial variables. The variance of mean ratio, which is equal to the product of variance in abundance across transects divided by the mean abundance, can quantify the degree of aggregation among transects (Langen et al. 2005). It is not possible to compare seabird density and abundance with the available habitat because pelagic habitats do not have discernable boundaries and constantly change along with the tides (Holm and Burger 2002). Instead, analyses should compare density and abundance with habitat use and depth (Holm and Burger 2002). Another way of estimating population size from survey data is to create general additive models, which can be more accurate than linear models because they use a Poisson distribution and control for data points that are not normally distributed (Clarke et al. 2003).

In addition, it is usually difficult to conduct separate analyses for each species, since limited visibility causes many birds to be identified only to genus. Seabirds should rather be divided into guilds based on their feeding habits- piscivores, planktivores, and benthic foragers- which have different habitat needs (Holm and Burger 2002). Distribution and abundance of each species can then be compared with the sum of all other species groups within the same guild using a chi-squared goodness of fit test or a paired t-test (Powers 1982; Holm and Burger 2002).

Non-parametric Mann-Whitney tests can compare densities and behaviors of birds with tidal currents (Holm and Burger 2002). Other statistical tests that may be useful include a logistic regression for presence/absence, a Poisson regression for total detection, ANOVA tests for relative abundance, Shannon’s index for species diversity, Fisher’s exact test for biases caused by fishing activity, and linear regression for the analysis of population trends (Powers 1982; Nur et al. 1999). These analyses may be performed in SPSS 10.0 (Holm and Burger 2002) or the free software WinBUGS (Kery 2008; Sturtz et al. 2005). The statistical programs MONITOR (Gibbs and Ramirez 2007) and TRENDS (Gerrodette and Brandon 2006), which employ power analysis and are also available for free on the Internet, are useful tools to assess population trends (Nur et al. 1999). Seabird distribution patterns can then be mapped using Arc GIS or with the aid of tagging technologies if they were used in the study. These maps can serve as a valuable resource for locating wintering seabird populations in future pelagic studies.

Bart, J. 2006. A sampling plan for the North American marsh bird monitoring program. Unpublished Manuscript available at http://greatbasin.nbii.gov/CBM_documents.htm.

BIOMASS Working Party on Bird Ecology. 1992. Recording distribution and abundance of seabirds at sea in the Southern Ocean: methods used in the BIOMASS Programme. Marine Ornithology 20: 51-59.

Briggs, K.T., Tyler, W.B., Lewis, D.B. 1985. Comparison of ship and aerial surveys of birds at sea. J. Wildl. Manage. 49: 405-411.

Burger, A.E. and S.A. Shaffer. 2008. Application of tracking and data-logging technology in research and conservation of seabirds. The Auk 125 (2): 253-264.

Burnham, K.P., Anderson, D.R., Laake, J.L. 1980. Estimation of density from line transect sampling of biological populations. Wildlife Monographs 72: 1-202.

Clarke, E.D., L.B. Spear, M.L. McCracken, F.F.C. Marques , D.L. Borchers, S.T. Buckland and D.G. Ainley. 2003. Validating the use of generalized additive models and at-sea surveys to estimate size and temporal trends of seabird populations. J. Appl. Ecol. 40: 278-292.

Collazo, J.A., T. Agardy, E.E. Klaas, J.E. Saliva, and J. Pierce. 1998. An interdecadal comparison of population parameters of brown pelicans in Puerto Rico and the US Virgin Islands. Waterbirds 21 (1): 61-65.

Dixon, T.J. 1977. The distance at which sitting birds can be seen at sea. Ibis 119: 372-375.

Gaston, A.J., Collins, B.L., Diamond, A.W. 1987. The “snapshot” count for estimating densities of flying seabirds during boat transects: a cautionary comment. The Auk 104(2): 336-338.

Gerrodette, T. and G. Brandon. 2006. Program TRENDS: Version 3.0. URL: http://swfsc.noaa.gov/textblock.aspx?Division=PRD&ParentMenuId=228&id=4740.

Gibbs, J. P., and P. Ramirez de Arellano. 2007. Program MONITOR: Estimating the Statistical Power of Ecological Monitoring Programs. Version 10.0. URL: http://www.esf.edu/efb/gibbs/monitor/monitor.htm.

Gonzalez-Solis, J., J.P. Croxall, D. Oro, and X. Ruiz. 2007. Trans-equatorial migration and mixing in the wintering areas of a pelagic seabird. Frontiers in Ecology and the Environment 5 (6): 297-301.

Gould, P.J., Forsell, D.J. 1989. Techniques for shipboard surveys of marine birds. Fish and Wildlife Technical Report 25: 20 pp. Washington D.C.

Haney, J.C. 1985. Counting seabirds at sea from ships: comments on interstudy comparisons and methodological standardization. Auk 102: 897-898.

Holm, K.J. and A.E. Burger. 2002. Foraging behavior and resource partitioning by diving birds during winter in areas of strong tidal currents. Waterbirds 25 (3): 312-325.

Kendall, S.J. and B.A. Agler. 1998. Distribution and abundance of Kittlitz's murrelets in southcentral and southeastern Alaska. Colonial Waterbirds 21 (1): 53-60.

Kery, M. 2008. Estimating abundance from bird counts: binomial mixture models uncover complex covariate relationships. The Auk 125 (2): 336-345.

Langen, T.A., M.R. Twiss, G.S. Bullerjahn, and S.W. Wilhelm. 2005. Pelagic bird survey on Lake Ontario following Hurricane Isabel, September 2003: observations and remarks on methodology. J. Great Lakes Res. 31: 219-226.

Nur, N., S.L. Jones, and G.R. Geupel. 1999. A statistical guide to data analysis of avian monitoring programs. U.S. Department of the Interior, Fish and Wildlife Service, BTP-R6001-1999, Washington, D.C. Available at http://www.nebirdmonitor.org/tools-resources/methods.

Powers, K.D. 1982. A comparison of two methods of counting birds at sea. J. Field Ornithol. 53(3): 209-222.

Spear, L., Nur, N., Ainley, D.G. 1992. Estimating absolute densities of flying seabirds using analysis of relative movement. Auk 109: 385-389.

Sturtz, S., U. Ligges, and A. Gelman. 2005. R2 WinBUGS, Version 1.4: A Package for Running WinBUGS from R. Journal of Statistical Software 12 (3): 1-16.

Tasker, M.L., Hope Jones, P., Dixon, T., Blake, B.F. 1984. Counting seabirds at sea from ships: a review of methods employed and a suggestion for a standardized approach. The Auk 101(3): 567-577.