In 2001, the Swedish Government authorised the construction of an offshore wind farm at Lillgrund in the Öresund Strait between Denmark and Sweden. In 2002, the Environmental Court defined the final terms and conditions for the wind farm development and the extent of the monitoring programme required. Lillgrund wind farm came into full operation in 2008, and is currently the largest offshore wind farm in operation in Sweden. The Swedish National Board of Fisheries conducted a monitoring programme, in the area, in the years before (2002–2005) and after (2008– 2010) the construction of the wind farm; a base line study and a study when the wind farm was operational, respectively. No investigation was conducted during the construction phase. The aim was to investigate the impact of the wind farm during the operational phase on the benthic and pelagic fish as well as on fish migration. These studies have partly been integrated into work conducted as a part of the Vindval Research Programme, funded by the Swedish Energy Agency.
Acoustics (sound)
• The overall sound energy from the wind farm under water is mainly generated by vibration from the gearbox.
• An analysis of the sound pressure level for the wind farm area, showed a correlation between noise level and the number of turbines in the wind farm (the so called park effect), where each individual turbine helps to increase the overall noise level in the area.
• Sound measurements from Lillgrund wind farm showed that noise levels within a distance of 100 metres from a turbine at high wind speeds are high enough to be a risk for some species of fish to be negatively affected, e.g. in the form of direct escape behaviour, or masking of vocal communication between individuals.
• Stress reactions can also occur at distances of more than 100 metres from a turbine. This is due to the fact that the noise from the turbines is continuous and louder than the ambient noise levels within some frequencies.
Measurements of the underwater noise levels were carried out at varying distances from individual turbines, from longer distances away from the entire wind farm as well as within a reference site (Sjollen) 10 km north of the wind farm. The results show that the wind farm produces a broadband noise below 1 kHz as well as one or two tones where the 127 Hz tone is the most powerful (vibrations from the first stage in the gear box). The majority of the overall underwater sound energy from the wind farm lies around the tone of 127 Hz. The maximum noise levels, generated by the wind turbine, working at full production (12 m/s), at 1 m were 136 dB re 1µPa(RMS) for the dominant tone of the turbine which was 127 Hz (integrated across 123–132 Hz) and 138 dB re 1µPa(RMS) at the full spectrum (integrated across 52–343 Hz). At a distance of 100 m from the turbine, the noise levels are reduced to 104–106 dB re 1µPa(RMS) across the full spectrum, which is close to the locally measured ambient noise in the Öresund Strait, but the noise level was still around 23 dB above the background level for the 127 Hz tone.
An analysis of the sound pressure level for the wind farm area showed a correlation between noise level and the number of turbines in the wind farm (called the park effect). Close to the wind farm (<80 m), the noise environment was dominated by the individual wind turbine with a calculated sound propagation loss of 17•log (distance). At greater distances (80 m to 7000 m) the sound propagation loss was non-linear and less than 17•log (distance). This is explained by the fact that the other turbines in the wind farm contributed to the total noise level. At even greater distances (>7 km) the entire wind farm functioned as a point source and the sound propagation loss was once again measured as 17•log (distance). The noise levels equivalent to those recorded and calculated from Lillgrund wind farm have not been shown to cause any physical injury to fish according to the current published scientific literature. It was only within some 100 metres from a turbine at high wind speeds that the noise levels were high enough to result in the risk of negative effects on some species of fish in the form of direct escape behaviour or possible masking of communication. The response depends upon the individual species’ sensitivity to sound. Fish have been shown to become stressed when they find themselves in a consistently noisy environment, which in turn can result in for example, lower growth rates or can have an impact on reproduction. Stress in general can also, in combination with other negative factors, make them more susceptible to disease etc., due to an impaired immune system. Animals can choose however, to remain in an area despite the disturbance, if the area is sufficiently important for their survival or reproduction. Based on the calculated sound propagation around the wind farm, salmon and eel could theoretically detect the 127 Hz tone at 250 m and 1 km distances respectively at a productivity rate of 60 and 100 %, which is equivalent to a wind speed of approximately 6 and 12 m/s. The calculated distances would be limited by the hearing ability of both fish species and not the background noise levels in the Öresund Strait. For herring and cod, the theoretical detection distance was calculated to be between 13 and 16 km respectively for a production rate of 60 and 100 %. This distance should have been greater, but is limited for these species due to the ambient noise levels in the area. These calculations indicate that fish can potentially detect sound from the wind farm at relatively long distances. Local variations with regard to depth and physical barriers such as peninsulas, e.g. Falsterbonäset in the southern end of the Öresund Strait, can however, have a large impact on the actual sound propagation.
Benthic Fish
• The temporal development of the fish community in Lillgrund was similar to that observed in the reference areas during the study period. For the wind farm as a whole, no effect was observed on species richness, species composition or on the abundance of fish.
• Several species of fish however, showed an increase in abundance close to the wind turbines compared with further away, especially eel (yellow eel) (Anguilla anguilla), cod (Gadus morhua), goldsinny wrasse (Ctenolabrus rupestris) and shorthorn sculpin (Myoxocephalus scorpius). The results reflect a redistribution of fish within the wind farm, rather than a change in productivity or migration from surrounding areas. The increase in abundance is probably due to the wind turbine foundations providing an opportunity for protection and improved foraging. The distance within which an increased abundance could be observed was estimated, for different species, to be between 50– 160 metres from a wind turbine.
• Fish distribution may to some extent have been influenced by the local acoustic environment, as a lower degree of aggregation close to the wind turbines at higher noise levels. The effect was most obvious for eelpout and eel (yellow eel). No response was seen for cod in relation to sound levels.
Changes in the species composition of the fish communities over time were studied in comparison with two reference areas. Of these, the northerly reference area (Sjollen) had a larger marine component than the southern reference area (Bredgrund). The species composition at Lillgrund had similarities with both of the reference areas. The results from fish sampling with fyke nets and gill net series indicate that there have been no significant changes in the number of species, the species composition or the fish abundance after the wind farm was built, looking at the wind farm as a whole. Some changes have however been noted in relation to individual species. An increased catch of shore crab and eel (yellow eel) was observed during the first two years of production, but not in the third year. The catch of eelpout increased in all areas during the period studied, but to a slightly lesser extent at Lillgrund when compared to the reference areas. For the other species, the changes observed at Lillgrund were similar to at least one of the reference areas. These results suggest that the fish communities within the wind farm were primarily affected by the same general environmental conditions as the fish communities within the reference areas, rather than by the effects of the wind farm. An analysis of the distribution patterns of fish close to the turbines showed an increased abundance in the immediate vicinity of the wind turbines in four of the eight species of fish studied: specifically shorthorn sculpin, goldsinny wrasse, cod and eel (yellow eel). The effects were seen already after the first year and were similar over all three years studied. An effect was also identified for eelpout, but only in 2010. The aggregation effect was seen within a distance of 50–160 metres from the wind turbines, different for the different species. A comparison of the relative effect of different factors, based on the data from an extended survey in 2010, showed that the observed distribution pattern could be explained to a larger extent by the presence of the turbines rather than the underwater topography of the area. The analysis also indicated weak effects of the local acoustic environment on fish distribution patterns, with a reduced presence of fish at higher noise levels. The response was strongest for eelpout and eel. No response in relation to noise level was seen for cod. For shorthorn scuplin and common shore crab a response was seen only 11 Swedish Agency for Marine and Water Management Report 2013:19 during the autumn. The magnitude of the effect of noise was, however, lower than the aggregation effect. Hence, fish aggregated close to the wind turbines in all conditions, but the effect was weaker when the noise levels were higher. It is recommended that the the wind farm area is reinvestigated after a number of years to follow the long-term development of the fish populations, and to see if the aggregation effect observed continues and potentially also increases over time. A prerequisite for a long term positive development of fish abundance is that the removal of fish, such as from fishing or predation by marine mammals and fish-eating birds, does not increase in the area.
Pelagic Fish
• There was a dramatic increase in commercial fishing for herring north of the Öresund Link (close to the north of the wind farm) in the first years of operation of the wind farm, in contrast to south of the bridge that forms a part of the Öresund Link, where it virtually completely stopped. This change may imply that the Rügen herring migration was affected by the Lillgrund Wind Farm. Due to the fact that there were other factors in addition to the wind farm contributing to the herring movements, it proved difficult to identify any correlation.
The evaluation was based on catch statistics from the commercial fisheries in the Öresund Strait (ICEs subdivision SD 23) and fisheries independent statistics from ICES for adult herring (Rügen herring) (ICES subdivision SD 21–23, western Baltic Sea and southern Kattegatt) and density of juvenile fish (ICES subdivision SD 24). There was a dramatic increase in commercial fishing for herring north of the Öresund Link in the first years of operation of the wind farm, in contrast to south of the bridge where it virtually completely stopped. The reason may be largely explained by the regulations banning drift-net fishing and a favourable market for herring, but potentially also because of the Öresund Link which was completed in 2000.The potential impacts of the wind farm are therefore difficult to distinguish from the impacts of these other factors because detailed resolution in the catch statistics are missing from the years before 1995 prior to the start of the building work on the Öresund Link. The statistics independent of commercial fishing from ICES showed no significant correlation between the density of herring juveniles in the western Baltic Sea and the number of adult herring (3 years old or more) in the following years in the Öresund Strait (ICES SD 21–24). There was however a weak tendency towards a negative development of the fish population over the period 1993 – 2010. The presence of Rügen herring and their migration through the Öresund Strait is likely strongly influenced by the fact that the population shows large fluctuations between the years. In addition, there is a possible overlapping effect on the soundscape from the wind farm and the Öresund Link, which has been in use since 2000. Overall, the variety of factors together mean that it is difficult to identify any clear results with regard to if the migration of Rűgen herring is influenced by Lillgrund wind farm.
Fish Migration
• According to the results from this work, the wind farm at Lillgrund is not a barrier for the migration of the eels that come into contact with it. An equally large proportion of the tagged and released silver eels (approximately one third) passed the transect line with receivers, at Lillgrund both before the wind farm was constructed (baseline study) and after it was in operation.
• There was no statistically significant difference indicating any alteration in the migration speed of eels, but there were occasional longer migration times when the wind farm was working at higher levels of production (>20 % of maximum) which may indicate that some eels are affected by the wind farm. The fact that the eels also showed a tendency towards being noted on fewer occasions than expected within the wind farm at low productivity (<20 %) and on slightly more occasions than expected at higher productivity (>20 %), could indicate that they have greater difficulty in navigating past the wind farm at higher levels of productivity than lower.
The impact of the wind farm on migration was studied via tagging of migrating silver eels. In total, 300 acoustically individually tagged eels were included in the study and of these, 100 contributed with useable information. The baseline study period started on a small scale in 2001 and ended in 2005. The majority of the eels were tagged and monitored during the production period (2008– 2010). All tagged silver eels were released south of the wind farm.
The results showed that an equally large proportion of the tagged and released silver eels; approximately one third, passed a transect with receivers at Lillgrund wind farm, both during the baseline period 2001–2005, and when it was in production 2008–2009. The greatest proportion of eels passed through the deeper part of the transect by the navigation channel Flintrännan close to the Danish border at Drogden during the production phase (31 %) and baseline period (43 %). A somewhat larger proportion of the eels were registered passing the most easterly part of the transect, close to Klagshamn, during the production phase (14 %) compared with the baseline period (5 %). A behaviour which occurred during the production phase, was that some individuals moved back to the release site, after being in the vicinity of wind farm. The most commonly observed behaviour during the study in 2010 was that an eel was registered moving south of the wind farm in a more or less northerly direction, but without being registered to the north of the wind farm. The range in the time taken for the movement of the eels from the release site to the transect running through the wind farm was very great, from four to more than 1000 hours. There was no statistically significant difference in the time taken to travel, between periods with low production (<20 % of maximum) and periods with high production (>20 %) or for individuals which passed through or outside of the wind farm. Even if the eels did not show any statistically significant behaviour, changes in movement patterns may occur for some individuals. The fact that there was a tendency towards longer periods of time taken for movement at higher production levels (not statistically significant) (>20 %) could indicate that some individual eels are influenced by the wind farm. The proportion of eels that took more than a week (168 hours) to make the journey was 48 % during the period with higher production (>20 %) compared with 28 % at lower production. No significant difference in the proportion of passes within or outside of the wind farm respectively could be shown. The eels showed however, – a tendency of being recorded on fewer occasions than expected inside the wind farm at low production levels (<20 %) and on more occasions than expected at higher production levels (>20 %). The irregularities in the proportions, compared with the expected result, could indicate that individual eels stayed longer in the wind farm when it was functioning at higher productivity. If the eels discover the wind turbine only when they are very close and do not change course, then other factors such as the speed of the current across the shallow marine areas become significant and can mean that the time spent in the area is shorter and records fewer. At high productivity, the eels may hesitate and/or divert their course and be recorded from close to or within the area, to then be recorded on the transect outside of the wind farm. The mechanisms that lie behind the possible impact from the electromagnetic field or the noise pattern are difficult to distinguish, as both can have an impact on the same areas. Travelling speed showed no linear relationship with the level of production in the wind farm.
Conclusions
The study at Lillgrund has resulted in an increase in the understanding of how offshore wind farms can affect fish, which is very valuable. Even within an international context, there are currently very few experience-based studies of offshore wind farms in operation. The results from three years of monitoring during the operational phase show that the effects of the wind farm on fish populations and fishing were limited. One of the clearest results showed that some benthic fish species were attracted to the foundations of the wind turbines with their associated scour protection (reef effect). In addition, the effect on the local noise environment in the form of increased noise in the Öresund Strait was documented. The results of the eel tracking study may indicate that some eels are influenced by the wind farm on their migration. Some care should be taken however, when applying the results of these studies in other offshore environments and on a larger scale. The monitoring has only been carried out for three years and thus reflects only a short-term perspective. Lillgrund wind farm is also one of the first large-scale wind farms and is situated in an area with regular and noisy shipping traffic and both frequent and large variations in environmental factors such as salinity and currents. A key knowledge gap that remains after the completion of this work is the lack of studies over a longer period of time, to help identify the long term ecological effects of, for example, the reef effect. Ideally, the wind farm should be re-visited after a number of years to see how the fish populations have developed over the longer term, and see if the observed aggregation of certain fish species close to the wind turbines continues, and to possibly see if any quantitative effects have taken place. Studies are also required in relation to how stress may affect fish species/individuals which choose the reef-like foundations and their noisier environment. Additional studies, primarily for the Baltic Sea, are also required to establish if there are any cumulative effects on migratory fish such as silver eels.