The multivariate statistical models developed for baseline conditions were used for impact assessment by imposing the predicted changes in those variables, such as current speed that were significant for the benthic habitat groups. Because of low predictability of PLS models for carnivores/omnivores, deep deposit-feeders and for abundances within all feeding-groups impacts were only estimated for biomass with-in filter-feeders and surface deposit-feeders. Basically, only permanent effects are considered and inclusion of sediment grain size changes resulting from dredging ac-tivities as a long-lasting effect can be questioned, as we do not know the persistence of accumulation at a site.
6.2 Impacts during operation
6.2.1 Impacts on benthic habitats and functional groups during operational phase Habitat modeling based on hydrodynamic and water quality modeling of baseline conditions in combination with in situ data collected during survey in spring 2009 examined a large range of variables that potentially could affect important functional groups in the benthic communities. For the two benthos variables, biomass of ben-thic filter-feeders and surface suspension-feeders four governing variables were identified:
• Changes in sediment grain size resulting from sediment accumulation
• Degree of sediment sorting resulting from sediment accumulation
• Changes in mean current speed above seabed
• Changes in combined filter-feeder index (FF)
In addition to potential impacts identified during benthic Habitat modeling
• Occupation of seabed by foundations and scour protection will destroy habitats for infauna which include surface deposit-feeders and a major part of the filter-feeder community
• Introduction of hard substrate, i.e. wind mill foundation potentially will favor epibenthic filter-feeders
• Small-scale changes in current patterns around foundations and scour protec-tions surely will change habitats for all functional groups
The small-scale changes are not dealt with because changes in governing factors cannot be resolved at appropriate scale using a model grid size of 600m.
The potential impacts on the two important functional groups during the operation phase are summarized in Table 6-1.
Table 6-1: Sources of impact and potential impacts on benthic fauna in the operation phase.
Notice that sediment grain size changes related to construction phase potentially
Project Activity Sources of Potential impact Potential environmental
6.2.2 Changes in governing variables and functions
6.2.2.1 Change in benthic habitats caused by current speed changes
Permanent changes in current speeds above seabed will affect benthic filter-feeders and surface deposit-feeders according to the statistical models developed on basis of baseline data (see Table 5-3). The magnitude of change in currents is dependent on the size and shape of the wind mill foundations and the distance between the wind mills compared to the radius of the wind mills. The change in mean current speeds was modelled for the two different wind park layouts in /16/ and is briefly summa-rised.
The wind park acts as an extra roughness or a partial blockage of the overall current field. The blocked water volume is forced around the park which leads to a decrease in the flow inside the park and an increase in flow velocities outside of the park.
The largest impact is expected at the surface where the flow velocities are signifi-cantly higher than at the bottom. The flow resistance at the bottom is smaller even though the geometry of the concrete foundations is larger at the bed.
The main flow direction is north-south. The results in Figure 6-1 show very small effects with average velocity changes less than 1 – 2 mm/s and in most areas even lower. The mean surface velocity changes are very small and insignificant compared to the mean flow velocity, which for the surface flow is in the order of 0.2 m/s. At
bottom mean currents are lower at 0.071 m/s and mean reduction within the wind park is 0.0003 m/s.
Figure 6-1: Annual mean surface velocity changes in the north-south velocity component in 2005 for the two wind farm layouts. Model results from the local 3D model (grid spacing ap-proximately 600 m). Green-blue colours indicate a velocity reduction and red colours indicate an increase in current velocity.
6.2.2.2 Change in benthic habitats caused by sediment accumulation and changes in grain size
Permanent changes in grain size of surface sediments will affect biomass of filter-feeders and surface deposit-filter-feeders and, change in degree of sorting (i.e. indexed by the slope at median grain size) will affect biomass of surface deposit-feeders (Table 5-3). Changes in sediment grain size may result from permanent change in current speeds and from accumulation of spilled sediment on the sediment surface. Reduc-tion in current speeds above sea bed is very modest averaging 0.0003 m/s within the wind farm areas, and a comparable increase outside the wind farm area. Such small changes are not likely to affect surface sediment significantly, especially con-sidering that mean current speed (modeled) and median sediment grain size (meas-ured) are only weakly correlated within the wind farm area (r = 0.17; p = 0.12, n=74).
Accumulation of fine sediment (20µm) on the sea bed resulting from dredging is very modest at either wind farm layouts (Figure 6-2) with only a limited area where ac-cumulation exceeds 100 g/m2 (eqv. to 0.10mm thickness) within the wind farm area, while accumulation rarely exceeds 25 g/m2 outside the farm area. Theoretical-ly, an accumulation of 0.10mm consisting of 20µm particles on the seabed will redu-ce the calculated median grain size from a typical 0.320mm to 0.3195mm (using data from ‘Sample D-4’ from benthic and sediment survey). Correspondingly, a worst case scenario with 0.45mm accumulation will reduce the median grain size of a
typical sediment from 0.320 to 0.317mm. The reduction in median diameter can be entered into prediction models to for estimating impacts.
The persistence of changes in grain size is unknown and given the relative minor changes it is doubtful if such changes can be measured. Monitoring after the con-struction of Nysted Offshore Wind farm generally showed no changes in sediment composition after the dredging operations. However, at few stations close to the construction area the content of silt/clay was increased after the earthworks /18/.
‘Unsorted’ sediments collected in the wind park area typically contain a large proportion of gravel (>10%) and on average a median grain size 0.42mm compared the an average median size of 0.34mm for the entire pool of samples. In addition to a low median grain size biomass of surface deposit-feeders also correlated
significantly to proportion of gravel in a sample. The degree of ‘sorting’ was described by the slope (%/mm) in the cumulative grain size curve.
Accumulation of fine sediments had very little influence on the slope even in the worst case (0.45mm accumulation) and, in combination with a low regression coefficient the predicted increase in deposit-feeder biomass due to reductions in slope of the grain size curve was less than 0.02%. Therefore, this effect was not considered.
Figure 6-2: Deposition pattern after dredging operation for wind farm layout 1 (left) and layout 2 (right). Model results from the sediment spreading calculations. Information from /3/.
6.2.2.3 Change in Filter-Feeder index
The combined filter-feeder indices (‘Mytilus-FF’, ‘Spisula-FF’) were significant for biomass of benthic filter-feeders and for surface deposit-feeders (Table 5-3). The two indices were closely inter-correlated (r = 0.86, p<10-5), but as the Mytilus-FF index had the largest contribution to model R2 and also was the most robust in PLS models this index was chosen for impact predictions. Spatial distribution and
changes in the Spisula-FF index for the two wind farm layouts is shown in Annex A.
The Mytilus-FF index encompasses a current speed function in addition elements of food concentration and oxygen impacts. The fact that the PLS models for filter-feeder
biomass and surface deposit-feeder biomass also include current speed as an inde-pendent variable (Table 5-3) underline that near bed currents are very important and that a power function rather that linear function probably would be more appro-priate.
The spatial distribution of change in Mytilus-FF is shown in Figures 6-3 and 6-4 for the two wind farm layouts. Changes are small but consistent with reduction in values within the wind farms and downstream (i.e. north of farms) while FF-values are slightly increased outside the farms, overall reflecting the small changes in average current speed.
Figure 6-3: Difference in Mytilus-FF index between baseline condition and wind mill park layout 1 (squared formation). Right: absolute values, left: % change.
Figure 6-4: Difference in Mytilus-FF index between baseline condition and wind mill park layout 2 (arched formation). Right: absolute values, left: % change.
6.2.2.4 Impact on filter-feeders and deposit-feeders within wind farm area
Based on the predicted changes in governing variables the impact on filter-feeders and deposit-feeders can be estimated using the PLS regression models developed for the baseline conditions. Calculations are carried out for the average changes within the wind park and for the worst case referring to model grid cells with the maximum accumulation of fine sediment, i.e. 0.45 mm.
6.2.2.5 Filter-feeders
Prediction of future biomass of filter-feeders within wind farm area is calculated as:
Filt-Feedwf = Filt-FeedBas - %cur-red * regcur – grain-red * reggrain - %FF-red*regFF
where
Filt-FeedBas are the average biomass of filter-feeders under baseline conditions (from survey),
%cur-red is the reduction in near-current speed in % of baseline,
regcur is the PLS regression coefficient for current influence on filter-feeders, grain-red is the reduction in median grain size (mm)
reggrain is the PLS regression coefficient for grain size influence on filter-feeders,
%FF-red is the reduction in FF-index in % of baseline, and
regFF is the PLS regression coefficient for FF-index influence on filter-feeders
Inserting values in equation gives for average condition within the wind park:
Filt-Feedwf = 441–0.0003/0.071*3089 – 0.000073*1163 – 0.003*2844 = 419 gdw/m2
and for worst conditions (highest sediment accumulation):
Filt-Feedwf = 441–0.0003/0.071*3089 – 0.001*1163 – 0.003*2844 = 418 gdw/m2 The reductions in biomass amounts to 5 and 6% of the baseline value for average condition and worst case, respectively. However, these predictions must be taken with some precaution because of the low predictability of the PLS-model (Q2 = 0.15).
The physical loss of seabed habitats by occupation of wind mill foundations including scour protection was estimated to a max of 0.5% of the entire project area /1/ and compared to this figure the indirect effects mediated through current reductions, sediment accumulation and reduction in FF-index were notable higher.
Blue mussels are expected to colonize and populate the concrete foundations from the pycnocline (≈13 m) to 2 m below water surface with an average biomass of 23.5 kg wet weight/m2 /16/. The available surface area of concrete for mussel settlement is 384 m2 for a single wind mill and 66.814 m2 for the entire wind park and assuming a steady-state biomass of 23.5 kg /m2 the total mussel biomass amounts to 1.570 tons. This figure can be compared to the loss of benthic filter-feeder biomass due to physical loss and indirect loss predicted using statistical models (Table 6-2).
Table 6-2: Estimated change in filter-feeder biomass due to direct effects (physical habitat loss and gains) and indirect effects. Dry weight is converted to wet weight (both including shells) by multiplying by 2.
Impact Causes Calculation
(kg wet weight)
Tons Wet Weight Physical loss Wind mill footprints 0.005*88*106*0.441*2 -388 Indirect loss Current reduction
Sediment change
(0.441-0.414)*2*88*106 -4752 Additional hard
Despite a predicted much higher biomass per area of filter-feeders on concrete foun-dation (i.e. 23.5 kg/m2) than the baseline biomass on seabed (i.e. 0.44 kg/m2) the calculated loss of filter-feeder biomass in the wind park area as a whole exceeds the gain in biomass due to introduction of additional hard substrate with a factor of about 3, . because of a much higher area of seabed compared to concrete founda-tions. However, as mussels on foundations are exposed to higher current speeds and higher algal concentrations than filter-feeders on and in seabed the ecological activ-ity in terms of biomass of phytoplankton filtered and yearly net production probable are more comparable than differences in biomass suggest.
6.2.2.6 Surface deposit-feeders
Prediction of future biomass of surface deposit-feeders within wind farm area is cal-culated as:
Dep-Feedwf = Dep-FeedBas + %cur-red * regcur + grain-red * reggrain + %FF-red*regFF
where
Dep-FeedBas are the average biomass of deposit-feeders under baseline conditions (from survey),
%cur-red is the reduction in near-current speed in % of baseline,
regcur is the PLS regression coefficient for current influence on deposit-feeders, grain-red is the reduction in median grain size (mm)
reggrain is the PLS regression coefficient for grain size influence on deposit-feeders,
%FF-red is the reduction in FF-index in % of baseline, and
regFF is the PLS regression coefficient for FF-index influence on deposit-feeders
Inserting values in equation gives for average condition within the wind park:
Dep-Feedwf = 1.14+ 0.0003/0.071*19.6+ 0.000073*11.5 + 0.003*6.5 = 1.24 gdw/m2
and for worst conditions (highest sediment accumulation):
Dep-Feedwf = 1.14+0.0003/0.071*19.6 + 0.0010*11.5 + 0.003*6.5 = 1.25 gdw/m2 The increases in biomass amounts to 9 and 10% of the baseline value for average condition and worst case, respectively. Overall, the influence of sediment accumula-tion was insignificant, due to low accumulaaccumula-tion rates and relative coarse sediment.
Hence, to change median grain size significantly accumulation rates should exceed say 5 mm. Compared to the model for filter-feeders the predictability of the deposit-feeder model is higher at Q2 = 0.23, but in absolute terms the predictability must be considered to be low and the predictions should be taken with some precaution.
6.2.2.7 Impact on filter-feeders and deposit-feeders outside the wind farm area
Average current speed and Mytilus FF-index will increase west and east of the wind farm with numerical values comparable to decreases within the farm (Figures 1, 6-3, 6-4), while sediment accumulation resulting from spill will not extend outside the wind farm (Figure 6-2). Assuming that the PLS-model developed for the wind farm area also applies to the area outside the wind farm, it follows that biomass of benthic filter-feeders will increase and biomass of surface deposit-feeders will decrease, probably in same proportions as the changes predicted within the wind farm.
In popular terms we could say that biomass loss and gains within the farm area are mirrored by gains and losses outside the farm area.
6.2.3 Summary of impacts during the operation phase
The expected impacts on benthic filter-feeders and surface deposit-feeders due to sediment changes, introduction of hard substrate, and changes in current pattern and food fluxed are summarized in Table 6-3. The duration of sediment changes in terms of grain size change cannot be predicted with certainty as sediment spread modeling was not carried out beyond the dredging period.
Table 6-3: Summary of impacts on the benthic filter-feeder and deposit-feeder habitats and biomasses
Sediment change Minor Local Long-term? No impact
Solid substrate Minor Local Long-term Minor impact
Change in currents and food-fluxes
Minor Regional Long-term Minor impact