Abstract
The Offshore Wind farm Egmond aan Zee (OWEZ) located 10618 km off the coast comprises of 36 wind turbines with steel monopile foundations. The OWEZ was taken into operation in 2007. As part of a Monitoring and Evaluation Program (NSW6MEP), this study focuses on the behaviour of fish in response to the operation of the wind farm. Wind farms can have either negative or positive effects on fish, for instance by disturbance due to noise or by acting as a refuge because fisheries are banned within the wind farm. An important feature to determine if positive effects might occur is residence time. The longer individual fish spend in the wind farm, the stronger potential benefit of wind farms can be expected. To our knowledge this is the first study on individual behaviour of fish within wind farms.
To study the potential effects of wind farms on fish behaviour, we used two approaches: tagging experiments (mark6recapture) and telemetry experiments by following individual fish with small transmitters in time. We selected two target fish species that are important for fisheries: sole Solea vulgaris as a target species potentially representing fish that use sand habitats and Atlantic cod Gadus morhua as a target species potentially representing fish that use artificial reefs such as the monopile and scour bed habitats in the wind farm.
With tagging experiments with sole, we compared return rates of fish caught, tagged and released within the wind farm to return rates of fish caught, tagged and released in a reference area outside the wind farm. If individual residence time in the wind farm is larger, then a significant lower return rate is expected for the wind farm batch over the reference area batch. And in addition, a stronger difference is expected between return rates in the period directly following the release than at longer time intervals after release. Two paired tagging experiments were performed: 300 tagged sole (150 caught tagged and released in OWEZ and 150 in a reference area) in October 2007, and 800 (400 in OWEZ and reference area) in June 2008.
With telemetry experiments with sole and cod, we assessed individual residence time and individual behavioural patterns. During early 2008, different telemetric deployment methods were tested in the wind farm and found to be robust against severe winter storm conditions. In July 2008, professional divers installed an array of detection stations on the sea floor covering 16 out of 36 monopiles. In August 2008, transmitters were implanted in 40 sole (length range sole 25634 cm), 40 cod in September 2008, and 7 cod in January 2009 (length range cod 226 46 cm, i.e. predominantly juveniles). In June6July 2009, all 16 detection stations could be retrieved and the telemetric data was extracted. We determined the duration between first and last detection in the wind farm for each fish as a proxy for individual residence time. We also determined the fraction of the time between first and last detection (detection rate). We compared detection rates as observed from fish with transmitters with as hypothesized for different behavioural scenarios: a) if random movement within the wind farm (associated expected detection rate would then be 7 %, i.e. detection area/wind farm area); b) if random movement occurred within the area where monopiles were present (expected detection rate: 14 %, i.e. detection area/wind turbine area); c) if strong attraction to the monopile habitats occurred (expected detection rate: 44 %, i.e. 16 out of 36 wind turbines covered with detection stations); or d) if extreme site fidelity occurred (expected detection rate: 100 %, i.e. stationary at catch and release site).
For tagged sole, there was no overall significant difference in return rate between OWEZ and reference area batches. Our combined results of the tagging and telemetry for sole indicate that the majority of movements take place at spatial scales larger than the wind farm area of OWEZ. Some individuals use the wind farm area for periods up to several weeks during the growing season, which indicates that there is no large scale avoidance of the wind turbines, at least in part of the sole population. On the other hand there were no indications found for attraction to the monopile habitats either. All of the individual soles showed detection rates well below the 44 % as expected when attraction to monopile habitats had indeed occurred.
For cod, as measured by telemetry, large variation in individual behaviour was observed. About 30 % of the cod were detected for only a few days after release and appeared to use spatial scales larger than the wind farm. About 55 % of the cod with transmitters were detected for several weeks to just over two months. About 15 % of cod with transmitters was detected in the wind farm for 869 months (the duration of the experiment). Individual detection rate averaged 46 %. Typically, cod stayed within a detection area for prolonged periods and sometimes switched to a different detection area within a short time interval. Cod staying within the wind farm showed clear cyclic daily patterns that changed throughout the seasons. Our results show that at least part of the juvenile cod population spends long periods within the OWEZ. No larger adult cod were caught in the wind farm. Whether this is due to a difference in behaviour between juveniles and adults or due to the ‘young’ age of the wind farm (just over a year at the start of the telemetry experiments) and subsequent later development or colonization of a ‘resident’ adult cod population within the wind farm, can not be determined at this stage. We also compared presence of cod near monopiles prior, during and after events when wind turbines were temporarily out of order and found no evidence for disturbance by the operation of wind turbines. Moreover, there appears to be strong attraction to the newly created monopile habitats. Cod behaviour as observed in OWEZ in combination with the fact that all fisheries are banned within the wind farm, make it that the wind farm acts as a refuge against fisheries for at least part of the cod population.