Comparison of alternatives

This section will present the criteria relevant to the evaluation of seabird monitoring and censusing methods. Two additional methods for seabird monitoring and censusing will be presented, namely, the wildlife camera method and the human ground survey method. All three methods will be graded against the aforementioned relevant criteria and a preferred alternative will be selected.

Evaluation Criteria

One of the criteria used in this comparison is cost. When bringing novel techniques into general use, high costs can often be the prohibitive factor. When that is not the case, cost can still be the most important criterion when all alternatives are graded relatively equally by the other criteria. The attribute used for this criterion will be daily cost [€/day].

Temporal frequency is defined as the frequency at which measurements are taken. When censusing seabirds on tidal flats, there is a need to be able to perform population counts as often as possible. This need is due to the non-stationary nature of foraging seabirds and tidal differences (i.e. height and time of occurrence) between days which can result in large discrepancies between counts taken only a day apart. The attribute used for this criterion will be the number of measurements per month [# meas./month].

The amount of disturbance caused to the observed subjects is arguably the most important criteria in any bird censusing study. The stress induced by disturbance can lead to lower reproductive rates in birds (Borrelle, & Fletcher, 2017). Therefore, any population censusing method for conservation purposes must minimise disturbance, lest it be counterproductive. This criterion, unfortunately, does not have an obvious natural attribute. Therefore, a logarithmic- like scale of 0 to 2 will be used, where 0 means that virtually no birds are disturbed, 1 means that some birds are disturbed, and 2 means that most or all birds were disturbed.

Related to temporal frequency is the concept of the deployability of a censusing method. Certain censusing methods may not be able to be reliably deployed in certain weather conditions or in remote locations. Regular, high frequency measurements require a censusing method with good deployability. A natural attribute may be found for deployability through statistical analysis of historical weather data. However, such an analysis is outside the scope of this work. Therefore, a likelihood of deployability scale ranging from 0 to 1 will be used.

Spatial coverage, as its name indicates, pertains to the size of the area a censusing method can cover. Tidal plains can extend for many kilometres, sometimes tens of kilometres of coastline. With that in mind, it is important that a censusing method have good spatial coverage to completely and thoroughly inventory a tidal flat. The natural attribute for spatial coverage is area covered per measurement [m2/meas.]

Resolution is the final criterion. In a general sense, resolution may be defined as the amount of detail one can see through a lens, be it the lens of a human eye, a binocular or a camera. For specific methods, resolution can be more precisely defined as the ground sampling or, more abstractly, the number of pixels in an image. High resolution is necessary to count and identify the species of seabirds. Though in most applications, pixel resolution (ground sampling distance) is the natural attribute for resolution, that attribute cannot be applied here because not all of the alternatives considered use a camera. Instead, a scale based on the size of the birds identifiable by each method will be used. The scale ranges from 1 to 3 and is defined as follows: 1 for methods that can identify large birds, 2 for methods that can identify medium and large birds, 3 for methods that can identify small, medium and large birds.