Looming-eye buoys temporarily reduce seabird depredation in pound nets (supporting dataset)

Materials and methodsThe experiment was carried out over a period of 46 days (2021-04-26 to 2021-06-11) at two pound net sites (’test’ and ’control’) spaced 1200 meters apart. The number of birds by species was counted at each site between 4 and 8 times a day on 8 selected dates in the time interval 07:30 to 11:40 with a total of 80 observations. Each count was made by one of three different observers, and a net was never observed by more than one observer on the same day, which precludes estimation of any observer effects. The looming eyes device was introduced at the test site between day 5 and 6 while the control site remained unchanged.While all species of birds were counted during the field trial, we focused our statistical analysis of the effects of LEB on the large piscivorous species common in the area at this time of the year and that are potentially attracted by the opportunity to feed on trapped fish. Specifically, we considered four species: the great cormorant, the greater black-backed gull (Larus marinus), herring gull, and lesser black-backed gull (Larus fuscus). The three gull species (family Laridae) were grouped together as “large gulls”. Following this, three independent statistical analyses were performed using different groupings of the bird counts as response variable:1. All (counts of great cormorants and all large gulls).2. Cormorants only3. Large gulls onlyFor each of these, a GAM model was used to test the effectiveness of the device:log(µi) = α(PoundNeti) + f1(Timei, PoundNeti) + f2(TimeOfDayi)Where μi is the mean number of unique birds counted in categories 1, 2, and 3 (see Figure 2 for detail) in the ith observation. Parameter α maps the ith observation to one of two intercepts, one for each pound net. For each pound net, a separate random process is estimated (represented by f1), describing the development in number of birds over time. Parameters f1 and f2 are Duchon splines with first derivative penalization, similar to a random walk random process (Duchon, 1977). First derivative penalization implies that splines go toward constant values beyond the data range as opposed to second derivative penalization, where trends are extrapolated. The effect of the LEBs is not explicitly modelled, but it can be derived from f1 (i.e. the change in relative number of birds between control and treatment site after the LEBs were introduced). Smoothness selection was carried out with the maximum likelihood (ML) method (Wood, 2011). The distribution of the response variable is assumed to be either Poisson or negative binomial. Model selection is based on the Akaike information criterion (AIC).References[1] Simon N. Wood. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 73(1):3–36, 2011.Versions1: Original submission.2: The description in "Material and Methods" is updated to reflect the changes in the data analysis. The results are similar to the first version, but are now presented in a clearer way.