Habitat modelling

Baseline studies in 2007-2008 in relation to Horns Rev 2 OWF modelled the distribu-tion of prey species to Common Scooter (Melanitta nigra) (Skov et al., 2008) Later, as part of environmental monitoring programmes for large scale offshore wind farms in Denmark, The Environmental Group commissioned a special report on wind farm im-pacts on sea birds and their food resources (Leonhard & Skov, 2012). In these re-ports, a number of dependent models were developed for measuring the impacts of wind farms. The offshore wind farms covered in the 2012 report are Horns Rev 1 and 2. The original modelling framework in this report consisted of:

 A regional and local hydrodynamic model, which delivers input to →

 An ecological model, which delivers input to →

 A deterministic filter-feeder model and

 A habitat suitability model

In the present work at Horns Rev 3, the habitat suitability models are expanded to cover a geographical area, which now includes the planned Horns Rev 3 project area.

HR3-TR-024 v3 30 / 121 Habitat Suitability model

Habitat suitability models were developed on top of the filter-feeder models in order to estimate more precisely the distribution of cut trough shell Spisula subtruncata and American razor clams Ensis directus. This was done within the frame of habitat suita-bility modelling, using empirical samples of biomass (trough shells, ash-free dry weight) and abundance (American razor clams, number of individuals) as response variables; and modelled filter-feeder indices, sediment data and data on the depth and relief of the sea floor as predictor variables.

Suitability functions were computed using Ecological Niche Factor Analysis (ENFA) (Hirzel et al., 2002). In suitability functions, the distributions of American razor clams and trough shells in the multivariate oceanographic space encompassed by recorded abundance data are compared with the multivariate space of the whole set of cells in the modelled area (Hirzel, 2001). On the basis of differences in means and variances of the bivalve ‘spaces’ and the global ‘space’, marginality of bivalve records was iden-tified by differences to the global mean and specialisation by a lower species variance than global variance. Thus, for large geographical areas like the Horns Rev area of the North Sea studied here, ENFA approaches the concept of ecological niche, defined as a hyper-volume in the multi-dimensional space of ecological variables, within which a species can maintain a viable population.

In the “Food Resources for Common Scoter. Horns Reef Monitoring 2009-2010” report (Leonhard & Skov, 2012), the following nine eco-geographical variables were found to be of significance for the model:

1. The modelled filter-feeder index for each of the two species (averages for the entire growth season from March to November);

2. Modelled sediment structure: median grain size (mm);

3. Modelled sediment structure: proportion (pct.) silt fraction;

4. Modelled sediment structure: proportion (pct.) fine sand fraction;

5. Modelled sediment structure: proportion (pct.) medium sand fraction;

6. Modelled sediment structure: proportion (pct.) coarse sand fraction;

7. Water depth;

8. Slope of the sea floor slope (in %);

9. Complexity of the sea floor calculated for 5x5 kernel: F = (n-1)/(c-1) where n=number of different classes present in the kernel, c= number of cells.

The main focus in relation to the Horns Rev 3 OWF is to expand the model to cover the new area, rather than document year-to-year changes. It was therefore decided not to run filter-feeding models isolated for the year 2013. Instead, index values from the original report were supplemented with values from 2011 and 2012 to calculate mean values for 2001-2012.

Marginality (M) was calculated as the absolute difference between the global mean (Mg) and the mean of the bivalve presence data (Ms) divided by 1.96 standard devia-tions of the global distribution (g):

M =

HR3-TR-024 v3 31 / 121 while specialisation (S) was defined as the ratio of the standard deviation of the global distribution to that of the species distribution:

S =

s g





To take multi-colinearity and interactions among eco-geographical factors into ac-count, indices of marginality and specialisation were estimated by factor analysis. The first component, being the marginality factor, was passed through the centroids of 1) all bivalve presence records and 2) all background cells in the study area. The index of marginality being a measure of the orthogonal distance between the two centroids.

Several specialisation factors were successively extracted from the n-1 residual di-mensions, ensuring their orthogonality to the marginality factor, while maximising the ratio between the residual variance of the background data and the variances of the bivalve occurrences.

A high specialisation would indicate restricted habitat usage compared to the range of conditions measured in the studied part of Horns Reef. This is obviously highly sensi-tive to the location and size of study area.

A habitat suitability index was computed on the basis of the marginality factors and the first four specialisation factors. A high proportion of the total variance was explained by the first few factors, by comparison to a broken-stick distribution. The habitat suita-bility algorithm allocated values to all grid cells in the study area. These values were proportional to the distance between the cells position and the position of the species optimum in factorial space. Habitat suitability computation was done using the medi-ans algorithm.

Sediment models

Besides the 56 sediment samples and 26 infauna samples from the present study, data from a total of 262 samples from the sampling campaigns from Horns Reef 2001-2010 (Skov et al., 2008; Leonhard & Skov, 2012) and data from the Danish national environmental monitoring scheme was used in the models, see Figure 4.3.

Data layers showing the proportion of each seabed type (silt/clay/sand, etc.) were developed from the sediment samples using variogram-based kriging models.

HR3-TR-024 v3 32 / 121 Figure 4.3 Positions of the sediment samples used in the habitat suitability modelling. The samples were

taken in previous sample programmes 2001-2010 and in the present study 2012-2013.

The definitions for the seabed types characterised by grain size are shown in Table 4.1.

Table 4.1 Seabed type characterised by grain size.

Seabed type Grain size (mm)

Silt and clay < 0.063 mm

Sand, fine 0.063 mm – 0.200 mm

Sand, medium 0.2 mm – 0.6 mm

Sand, coarse 0.6 mm – 2 mm

Gravel > 2 mm