36 Supporting data: At-sea survey data
Dedicated surveys to sample the abundance and distribution of seabirds are a basic requirement to understand bird densities within coastal and offshore marine environments.
Distribution data can be obtained from observations collected during surveys within a predetermined spatio-temporal frame of reference or by sampling locations of individual animals using animal‐borne tracking devices. The former provide an area-based sampling perspective, also referred to as a Eulerian perspective in the oceanographic literature, whereas the latter provides an individual based sampling perspective, also referred to as a Langragian perspective (Phillips et al. (2019)).
Area-based or Eulerian survey designs sample across a predetermined survey area using predetermined sampling locations or transects, which may or may not be replicated through time. The primary objective of this sampling approach is to obtain animal distribution and abundance data in a predefined area and time period. For seabirds, at-sea such surveys are generally conducted using ships or aircraft (Buckland et al. (2012); Camphuysen et al. (2004); Tasker et al. (1984)), although land-based surveys may be used to cover near-shore environments (Arranz et al. (2014); Smith et al. (2015)).
36.1 At-sea survey data and tracking data
Due to the differing strengths and weaknesses of the two methods (at survey data vs. direct animal tracking), distributions derived from both may be combined to provide a more complete picture of seabird distribution (Louzao et al. (2009)), but this is not straightforward and formal analytical frameworks for combining area-based and individual-based data remain an active research topic in ecological statistics (Chandler et al. (2022); Glennie et al. (2021)).
Synoptic comparisons of both approaches are rare, but the results of studies (Carroll et al. (2019); Phillips et al. (2019)) suggest that tracking data can yield comparable results to area-based survey data when the tracked individuals are a representative sample of the birds using the area of interest. In particular, single-colony tracking appears to yield comparable results to area-based approaches for relatively isolated colonies, whereas multi-colony tracking or tracking of individuals captured at-sea in the area of interest is recommended to get an unbiased picture of seabird distributions in areas used by individuals from dispersed colonies. Assessments of the comparability of sampling approaches therefore remain critical to the interpretation of differences in seabird distribution, when data have been collected using different methods.
36.2 At-sea survey methodology
This section provides a brief review of existing census techniques for area-based or Eulerian surveys of seabirds at sea.
Two primary observation tools are discussed: aerial and ship-based line transect surveys, which both have the potential to provide similar data outputs, i.e. spatially referenced counts of birds.
The European Seabirds At Sea protocol outlines a standardised and broadly agreed upon approach for collecting offshore monitoring data on seabirds and marine mammals through both aerial or ship-based surveys at sea.
Additional observation methods exist, such as land-based human observers (Arranz et al. (2014); Smith et al. (2015)) or terrestrial or marine radars (Lilliendahl et al. (2003); Orben et al. (2019)) but are not discussed in detail due to their inability to monitor large offshore areas, and or limitations with regard to species identification.
The at-sea survey census techniques generally rely on a common principle: line transect surveys, but different survey methodologies have strength and weaknesses both concerning the nature and quality of the collected data, and in terms of their logistics.
Data quality is primarily affected by:
the level of detectability of individual birds a method affords, and
the level of taxonomic identification of detected birds.
Logistical characteristics include differing costs, survey speeds, and allowable environmental conditions under which surveys can be undertaken.
Aerial surveys are quick, enabling coverage of larger areas per unit time and therefore providing a more synoptic view of seabird distributions than slower shipboard surveys. Bird data obtained during aerial surveys may be combined with remotely sensed environmental parameters in a correlative approach, but in-situ collection of environmental parameters is usually very limited, whereas shipboard surveys can allow the simultaneous collection of in-situ oceanographic data. Both aerial and shipboard surveys generally rely on visual survey methods (human observers, camera systems; Buckland et al. 2012), which restricts them to daylight hours, and are generally limited to relatively benign wind and wave conditions (generally sea states below 3-5). This may fundamentally limit the understanding of realized seabird distributions under the full set of prevailing environmental conditions. Survey platforms further differ in their effect on seabird behaviour. Surface vessels may elicit avoidance or attraction effects on birds (Spear et al. (2004)), whereas the degree of avoidance to aircraft is largely governed by their flight height, so that surveys may have little to no impact on bird behaviour when executed at a sufficient aircraft altitude (Buckland et al. (2012)).
36.3 At-sea survey data for KBAs or IBAs
While at-sea survey data only provides a snapshot of distribution and numbers at any given time (unless collected over many years), it can play a vital role in identifying the locations of potentially important marine sites, especially if threshold numbers of animals are exceeded.
Raw data or those converted into density estimates can be used to consider if important criteria thresholds are met. Applying a scale which includes the relevant important criteria threshold will clearly show the locations of observations that have recorded adequate numbers of animals.
This approach may be most useful for threatened species with low population thresholds, given these species are more likely to trigger relevant important criteria because of lower thresholds needing to be met.