Gemini Wind Park

Visual observations demonstrated that the FaunaGuard-PM successfully deterred harbour porpoises up to distances of at least 1000 meters, with porpoises returning to the vicinity shortly after device deactivation; acoustic monitoring data partially corroborated these findings, though equipment limitations reduced data value.

Nathusius' pipistrelle was the dominant species offshore, exhibiting highly seasonal migration patterns (peak late August-end September in autumn; late March-end June in spring), with presence significantly predicted by low-to-moderate wind speeds, favorable wind directions indicating tailwind-seeking behavior, higher temperatures, and increased moon illumination.

Nathusius' pipistrelle was the dominant species offshore, exhibiting highly seasonal migration patterns (peak late August-end September in autumn; late March-end June in spring), with presence significantly predicted by low-to-moderate wind speeds, favorable wind directions indicating tailwind-seeking behavior, higher temperatures, and increased moon illumination.

Nathusius' pipistrelle was the dominant species offshore, exhibiting highly seasonal migration patterns (peak late August-end September in autumn; late March-end June in spring), with presence significantly predicted by low-to-moderate wind speeds, favorable wind directions indicating tailwind-seeking behavior, higher temperatures, and increased moon illumination.

Most seabird species occurred at low densities in the Gemini area, making significant wind farm impacts unlikely; however, Common Guillemots were recorded at high densities during their post-breeding exodus and through autumn and winter, identifying them as the species most likely to be affected by displacement and the focus of recommended future operational-phase (T-1) monitoring.

Except for common guillemots, most seabird species occurred at low densities within the study area, with breeding-colony and migrant species notably sparse. Common guillemots showed peak densities during autumn and winter post-breeding movements, identified as warranting continued monitoring for future wind farm impact assessment.

Except for common guillemots, most seabird species occurred at low densities within the study area, with breeding-colony and migrant species notably sparse. Common guillemots showed peak densities during autumn and winter post-breeding movements, identified as warranting continued monitoring for future wind farm impact assessment.

Anthropogenic noise sources — shipping, geotechnical surveys, and pile driving from other nearby activity — dominated the acoustic environment, increasing sound levels by up to 20 dB in some frequency bands; GEMINI 1 recorded more shipping noise while GEMINI 8 recorded more pile-driving noise, and several underwater explosions with the potential to adversely affect marine animals were also detected. Biological sounds were rarely evident in the recordings.

Anthropogenic noise sources – shipping, geotechnical surveys, and pile driving from other nearby activity – dominated the acoustic environment, increasing sound levels by up to 20 dB in some frequency bands; GEMINI 1 recorded more shipping noise while GEMINI 8 recorded more pile-driving noise, and several underwater explosions with the potential to adversely affect marine animals were also detected. Biological sounds were rarely evident in the recordings.

Visual observations demonstrated that the FaunaGuard-PM successfully deterred harbour porpoises up to distances of at least 1000 meters, with porpoises returning to the vicinity shortly after device deactivation; acoustic monitoring data partially corroborated these findings, though equipment limitations reduced data value.

Visual observations demonstrated that the FaunaGuard-PM successfully deterred harbour porpoises up to distances of at least 1000 meters, with porpoises returning to the vicinity shortly after device deactivation; acoustic monitoring data partially corroborated these findings, though equipment limitations reduced data value.

Anthropogenic sources—primarily shipping traffic, geotechnical surveys, and pile driving from nearby wind park construction in German waters—dominated the underwater noise environment, with man-made noise elevating sound levels up to 20 dB above ambient in certain frequency bands; biological sounds were rarely detected, and several underwater explosions were also recorded at both stations.

Aerial surveys recorded 691 harbour porpoises with densities ranging from 0.3-1.5 individuals/km² (mean 1.0/km²), while passive acoustic monitoring detected harbour porpoises on 93.2% of sampled days, revealing strong seasonal patterns with peaks in spring and lowest activity in summer and winter.

Part 1 – No significant behavioral response to the FG-FM sound or the white noise control was detected: swimming depth, distance to the speaker, swimming speed, and group coherence did not differ significantly across sound levels (ANOVA, all P > 0.4), though brief individual startle responses were observed qualitatively at the highest amplitude tested. The authors concluded that the FG-FM module may need higher sound levels or a redesigned acoustic signal to achieve a reliable deterrent effect.

Part 2 – Kurtosis proved an unsatisfying predictor of behavioral response: crest factor showed only a non-significant trend on swimming depth (p = 0.054), and pulse rate had no significant effect on any of the four behavioral parameters measured. The authors concluded that an alternative index weighting multiple acoustic parameters would likely outperform kurtosis for future impact assessments.

Part 1 – No significant behavioral response to the FG-FM sound or the white noise control was detected: swimming depth, distance to the speaker, swimming speed, and group coherence did not differ significantly across sound levels (ANOVA, all P > 0.4), though brief individual startle responses were observed qualitatively at the highest amplitude tested. The authors concluded that the FG-FM module may need higher sound levels or a redesigned acoustic signal to achieve a reliable deterrent effect.

Part 2 – Kurtosis proved an unsatisfying predictor of behavioral response: crest factor showed only a non-significant trend on swimming depth (p = 0.054), and pulse rate had no significant effect on any of the four behavioral parameters measured. The authors concluded that an alternative index weighting multiple acoustic parameters would likely outperform kurtosis for future impact assessments.

Multiple wind farm sites in the Dutch North Sea were identified as suitable for flat oyster restoration, and the integration of oyster restoration with aquaculture for protein production emerged as a viable approach warranting pilot investigation.

Multiple wind farm sites in the Dutch North Sea were identified as suitable for flat oyster restoration, and the integration of oyster restoration with aquaculture for protein production emerged as a viable approach warranting pilot investigation.

Multiple wind farm sites in the Dutch North Sea were identified as suitable for flat oyster restoration, and the integration of oyster restoration with aquaculture for protein production emerged as a viable approach warranting pilot investigation.

Five of ten grey seals tracked in autumn/winter used or approached the Gemini windpark area, while only one and two harbour seals respectively showed windpark overlap, indicating greater spatial overlap with the proposed development for grey seals despite harbour seals being more abundant in the region.

Baseline – Aerial surveys recorded 691 harbour porpoises with densities ranging from 0.3-1.5 individuals/km² (mean 1.0/km²), while passive acoustic monitoring detected harbour porpoises on 93.2% of sampled days, revealing strong seasonal patterns with peaks in spring and lowest activity in summer and winter.

Construction – Aerial surveys recorded 469 harbour porpoises (densities 0.05-3.3 individuals/km²), showing a seasonal decline from June to October 2015, while PAM detected porpoises on 96.2% of sampled days and revealed avoidance behavior at 10-20 km from pile driving sources, with rapid return to avoided areas after pile driving ceased. The construction-phase report notes that the same underlying factors explained most of the variation in acoustic activity during both the baseline and construction periods.

Baseline – Aerial surveys recorded 691 harbour porpoises with densities ranging from 0.3-1.5 individuals/km² (mean 1.0/km²), while passive acoustic monitoring detected harbour porpoises on 93.2% of sampled days, revealing strong seasonal patterns with peaks in spring and lowest activity in summer and winter.

Construction – Aerial surveys recorded 469 harbour porpoises (densities 0.05-3.3 individuals/km²), showing a seasonal decline from June to October 2015, while PAM detected porpoises on 96.2% of sampled days and revealed avoidance behavior at 10-20 km from pile driving sources, with rapid return to avoided areas after pile driving ceased. The construction-phase report notes that the same underlying factors explained most of the variation in acoustic activity during both the baseline and construction periods.

Seals were found to potentially deplete local prey resources by up to 50%, and accelerometer-derived algorithms successfully classified behavioral states; the model identified a tipping point where reduced foraging efficiency rapidly leads to energy loss and mortality, and revealed that seals showed avoidance of offshore wind farms even before pile-driving, suggesting sensitivity to preparatory activities or prior disturbance events.

Seals were found to potentially deplete local prey resources by up to 50%, and accelerometer-derived algorithms successfully classified behavioral states; the model identified a tipping point where reduced foraging efficiency rapidly leads to energy loss and mortality, and revealed that seals showed avoidance of offshore wind farms even before pile-driving, suggesting sensitivity to preparatory activities or prior disturbance events.

Seals were found to potentially deplete local prey resources by up to 50%, and accelerometer-derived algorithms successfully classified behavioral states; the model identified a tipping point where reduced foraging efficiency rapidly leads to energy loss and mortality, and revealed that seals showed avoidance of offshore wind farms even before pile-driving, suggesting sensitivity to preparatory activities or prior disturbance events.

Baseline – Five of ten grey seals tracked in autumn/winter used or approached the Gemini windpark area, while only one and two harbour seals respectively showed windpark overlap, indicating greater spatial overlap with the proposed development for grey seals despite harbour seals being more abundant in the region.

Construction – Grey seals exhibited significant avoidance behavior during pile-driving, swimming away more frequently than towards active pile-driving sites within 30-40 km, while both species demonstrated shallower dive depths during construction; harbour seals showed reduced offshore habitat usage during the construction year compared to previous years.

Baseline – Five of ten grey seals tracked in autumn/winter used or approached the Gemini windpark area, while only one and two harbour seals respectively showed windpark overlap, indicating greater spatial overlap with the proposed development for grey seals despite harbour seals being more abundant in the region.

Construction – Grey seals exhibited significant avoidance behavior during pile-driving, swimming away more frequently than towards active pile-driving sites within 30-40 km, while both species demonstrated shallower dive depths during construction; harbour seals showed reduced offshore habitat usage during the construction year compared to previous years.

Porpoises exposed to pile-driving noise swam at approximately 7 km/h, potentially travelling 14 km from a piling site during a typical two-hour event, with associated increases in energy expenditure. The updated expert elicitation concluded that disturbance effects on vital rates were less severe than estimated in 2013, with the largest revisions seen for harbour porpoises. These updated transfer functions were integrated into a revised iPCoD model and used to update the Dutch cumulative impact assessment framework for offshore wind farm construction.

Previous research and monitoring programmes (Shortlist, Master Plan VUM, and mandatory Monitoring and Evaluation Programmes) established foundational ecological data on offshore wind farm impacts, with findings informing the updated 2014-2015 Master Plan and guiding the design of the 2017-2021 monitoring priorities focused on reducing uncertainties around long-term impacts, upscaling effects, and mitigation measure effectiveness.

Previous research and monitoring programmes (Shortlist, Master Plan VUM, and mandatory Monitoring and Evaluation Programmes) established foundational ecological data on offshore wind farm impacts, with findings informing the updated 2014-2015 Master Plan and guiding the design of the 2017-2021 monitoring priorities focused on reducing uncertainties around long-term impacts, upscaling effects, and mitigation measure effectiveness.

Previous research and monitoring programmes (Shortlist, Master Plan VUM, and mandatory Monitoring and Evaluation Programmes) established foundational ecological data on offshore wind farm impacts, with findings informing the updated 2014-2015 Master Plan and guiding the design of the 2017-2021 monitoring priorities focused on reducing uncertainties around long-term impacts, upscaling effects, and mitigation measure effectiveness.

The study demonstrated proof-of-concept for active oyster reintroduction with viable populations showing high survival, growth, and larval production, and identified that storm-driven seafloor turbulence disperses loosely deployed oysters in shallow waters (40 m depth). It also validated multiple novel monitoring methods including environmental DNA sampling for fish biodiversity detection and larval presence confirmation, while discovering evidence of a previously unidentified relic oyster population capable of reproduction.

The study demonstrated proof-of-concept for active oyster reintroduction with viable populations showing high survival, growth, and larval production, identified that storm-driven seafloor turbulence disperses loosely deployed oysters in shallow waters (40 m depth), and validated multiple novel monitoring methods including environmental DNA sampling for fish biodiversity detection and larval presence confirmation, while also discovering evidence of a previously unidentified relic oyster population capable of reproduction.

The study demonstrated proof-of-concept for active oyster reintroduction with viable populations showing high survival, growth, and larval production, identified that storm-driven seafloor turbulence disperses loosely deployed oysters in shallow waters (40 m depth), and validated multiple novel monitoring methods including environmental DNA sampling for fish biodiversity detection and larval presence confirmation, while also discovering evidence of a previously unidentified relic oyster population capable of reproduction.

Median sound exposure levels (SEL) at 750 m were approximately 180 dB re 1 µPa²s for the two monopiles, about 7 dB higher than the smaller pin-pile foundation, reflecting its smaller sound-radiating surface. Sound attenuation over long distances exceeded standard model predictions; only the Thiele & Schellstede rough-sea propagation model accurately predicted measured levels beyond 5 km. Ambient noise (111–155 dB re 1 µPa²) was dominated by vessel traffic, and piling noise exceeded ambient levels by at least 6 dB at the closest measurement positions.

Median sound exposure levels (SEL) at 750 m were approximately 180 dB re 1 µPa²s for the two monopiles, about 7 dB higher than the smaller pin-pile foundation, reflecting its smaller sound-radiating surface. Sound attenuation over long distances exceeded standard model predictions; only the Thiele & Schellstede rough-sea propagation model accurately predicted measured levels beyond 5 km. Ambient noise (111–155 dB re 1 µPa²) was dominated by vessel traffic, and piling noise exceeded ambient levels by at least 6 dB at the closest measurement positions.

Harbour porpoises showed detectable behavioural avoidance at considerably smaller distances (7-15 km) than predicted by current methodology (~30 km), and analysis suggested that unweighted sound pressure level (SPL) may better predict behavioural response than frequency-weighted metrics, though practical limitations in measuring high-frequency components and distinguishing pile-driving noise from ambient noise complicate direct application of these findings.

The acoustic dataset was successfully compiled to evaluate frequency weighting requirements for harbour porpoise avoidance thresholds and to assess how water depth and mitigation measure effectiveness influence the underwater noise frequency spectrum at distances from pile driving locations.

Aerial surveys recorded 469 harbour porpoises (densities 0.05-3.3 individuals/km²), showing a seasonal decline from June to October 2015, while PAM detected porpoises on 96.2% of sampled days and revealed avoidance behavior at 10-20 km from pile driving sources, with rapid return to avoided areas after pile driving ceased.

Harbour porpoises showed detectable behavioural avoidance at considerably smaller distances (7-15 km) than predicted by current methodology (~30 km), and analysis suggested that unweighted sound pressure level (SPL) may better predict behavioural response than frequency-weighted metrics, though practical limitations in measuring high-frequency components and distinguishing pile-driving noise from ambient noise complicate direct application of these findings.

Harbour porpoises showed detectable behavioural avoidance at considerably smaller distances (7-15 km) than predicted by current methodology (~30 km), and analysis suggested that unweighted sound pressure level (SPL) may better predict behavioural response than frequency-weighted metrics, though practical limitations in measuring high-frequency components and distinguishing pile-driving noise from ambient noise complicate direct application of these findings.

Initial response thresholds were lower for small fish (131 dB re 1 μPa² s SELss) compared to large fish (141 dB re 1 μPa² s SELss), but no consistent sustained behavioral responses were observed at any sound level tested.

Initial response thresholds were lower for small fish (131 dB re 1 μPa² s SELss) compared to large fish (141 dB re 1 μPa² s SELss), but no consistent sustained behavioral responses were observed at any sound level tested.

Initial response thresholds were lower for small fish (131 dB re 1 μPa² s SELss) compared to large fish (141 dB re 1 μPa² s SELss), but no consistent sustained behavioral responses were observed at any sound level tested.

The study documented underwater noise levels as a function of distance from the pile driving source and compared noise characteristics across different pile designs, with data intended to inform validation of TNO underwater noise propagation models (Aquarius and Zampoli).

The FG-FM sound series failed to elicit consistent behavioral responses in seabass, while intermittent brown noise pulse trains produced significant changes in swimming depth; kurtosis was evaluated as a potential acoustic impact metric but was ultimately deemed an inadequate standalone measure of behavioral responsiveness.

Grey seals exhibited significant avoidance behavior during pile-driving, swimming away more frequently than towards active pile-driving sites within 30-40 km, while both species demonstrated shallower dive depths during construction; harbour seals showed reduced offshore habitat usage during the construction year compared to previous years.

ROV grid surveys were conducted across designated deployment areas to document the spatial distribution and current condition of oyster shell material, BESE-reef paste substrates, and living O. edulis populations established at the wind farm site. The assessment aimed to determine restoration success metrics including oyster presence, health status, and substrate stability at multiple months post-deployment. Ongoing analysis of these survey data is currently being conducted as part of a broader, long-term monitoring initiative to establish standardized nature-enhancement practices within offshore wind farms.

ROV grid surveys were conducted across designated deployment areas to document the spatial distribution and current condition of oyster shell material, BESE-reef paste substrates, and living O. edulis populations established at the wind farm site. The assessment aimed to determine restoration success metrics including oyster presence, health status, and substrate stability at multiple months post-deployment. Ongoing analysis of these survey data is currently being conducted as part of a broader, long-term monitoring initiative to establish standardized nature-enhancement practices within offshore wind farms.

ROV grid surveys were conducted across designated deployment areas to document the spatial distribution and current condition of oyster shell material, BESE-reef paste substrates, and living O. edulis populations established at the wind farm site. The assessment aimed to determine restoration success metrics including oyster presence, health status, and substrate stability at multiple months post-deployment. Ongoing analysis of these survey data is currently being conducted as part of a broader, long-term monitoring initiative to establish standardized nature-enhancement practices within offshore wind farms.

Both razorbills and guillemots demonstrated consistent avoidance of the wind farm areas and surrounding zones, with birds consistently absent or at very low densities within turbine parks and the intervening zone, while forming dynamic hotspots only outside the wind farm footprint. Significant habitat loss was documented for both species, with species-specific variations in magnitude and substantially lower bird densities in the immediate vicinity of the wind farm compared to greater distances.

Both razorbills and guillemots demonstrated consistent avoidance of the wind farm areas and surrounding zones, with birds consistently absent or at very low densities within turbine parks and the intervening zone, while forming dynamic hotspots only outside the wind farm footprint. Significant habitat loss was documented for both species, with species-specific variations in magnitude and substantially lower bird densities in the immediate vicinity of the wind farm compared to greater distances.

Lesser black-backed gulls spend the vast majority of foraging time on land rather than at sea, with 17% of offshore flight occurring within the rotor-swept zone; detailed behavioral mapping through the wind farm identified variable movement patterns and micro-avoidance behaviors that provide insight into collision risk factors including flight altitude, speed, and turbine proximity.

Both razorbills and guillemots demonstrated consistent avoidance of the wind farm areas and surrounding zones, with birds consistently absent or at very low densities within turbine parks and the intervening zone, while forming dynamic hotspots only outside the wind farm footprint. Significant habitat loss was documented for both species, with species-specific variations in magnitude and substantially lower bird densities in the immediate vicinity of the wind farm compared to greater distances.

Lesser black-backed gulls spend the vast majority of foraging time on land rather than at sea, with 17% of offshore flight occurring within the rotor-swept zone; detailed behavioral mapping through the wind farm identified variable movement patterns and micro-avoidance behaviors that provide insight into collision risk factors including flight altitude, speed, and turbine proximity.

Lesser black-backed gulls spend the vast majority of foraging time on land rather than at sea, with 17% of offshore flight occurring within the rotor-swept zone; detailed behavioral mapping through the wind farm identified variable movement patterns and micro-avoidance behaviors that provide insight into collision risk factors including flight altitude, speed, and turbine proximity.

Bird flight patterns exhibited significant temporal variation across daily, seasonal, and annual scales, with notable differences in abundance, altitude distribution, and traffic rates during migration periods.

Bird flight patterns exhibited significant temporal variation across daily, seasonal, and annual scales, with notable differences in abundance, altitude distribution, and traffic rates during migration periods.

Bird flight patterns exhibited significant temporal variation across daily, seasonal, and annual scales, with notable differences in abundance, altitude distribution, and traffic rates during migration periods.

The automated analysis successfully identified foraging events in sufficient numbers to detect temporal patterns (including diel cycles) and spatial differences between monitoring stations, demonstrating the viability of using CPOD-derived metrics to assess harbour porpoise foraging behavior across wind farm areas.

The automated analysis successfully identified foraging events in sufficient numbers to detect temporal patterns (including diel cycles) and spatial differences between monitoring stations, demonstrating the viability of using CPOD-derived metrics to assess harbour porpoise foraging behavior across wind farm areas.

The automated analysis successfully identified foraging events in sufficient numbers to detect temporal patterns (including diel cycles) and spatial differences between monitoring stations, demonstrating the viability of using CPOD-derived metrics to assess harbour porpoise foraging behavior across wind farm areas.

The study evaluated the establishment and development of native oyster reefs within an offshore wind farm environment, generating knowledge applicable to future reef restoration efforts in similar settings.

The study evaluated the establishment and development of native oyster reefs within an offshore wind farm environment, generating knowledge applicable to future reef restoration efforts in similar settings.

The study evaluated the establishment and development of native oyster reefs within an offshore wind farm environment, generating knowledge applicable to future reef restoration efforts in similar settings.

Elasmobranch eDNA was detected in 8.5% of samples, with five species identified: two sharks (Mustelus asterias and Cetorhinus maximus) and three skates (Raja clavata, R. montagui, and R. clavata). R. clavata was the most widespread species, detected across three OWFs with highly variable detection rates (1.6–57.1%), while M. asterias showed strong seasonal patterns corresponding to known migratory periods. Species composition and detection rates varied substantially among OWFs, reflecting differences in sampling effort and temporal coverage.

No consistent behavioral response patterns were observed across the four measured parameters, with no significant changes in swimming depth, distance to speaker, swimming speed, or group coherence during sound exposure trials. While some individual sudden behavioral changes occurred at sound onset, particularly at higher sound levels, no significant deterrent effect was detected and no dose-response relationship could be established.

Elasmobranch eDNA was detected in 8.5% of samples, with five species identified: two sharks (Mustelus asterias and Cetorhinus maximus) and three skates (Raja clavata, R. montagui, and R. clavata). R. clavata was the most widespread species, detected across three OWFs with highly variable detection rates (1.6–57.1%), while M. asterias showed strong seasonal patterns corresponding to known migratory periods. Species composition and detection rates varied substantially among OWFs, reflecting differences in sampling effort and temporal coverage.

Elasmobranch eDNA was detected in 8.5% of samples, with five species identified: two sharks (Mustelus asterias and Cetorhinus maximus) and three skates (Raja clavata, R. montagui, and R. clavata). R. clavata was the most widespread species, detected across three OWFs with highly variable detection rates (1.6–57.1%), while M. asterias showed strong seasonal patterns corresponding to known migratory periods. Species composition and detection rates varied substantially among OWFs, reflecting differences in sampling effort and temporal coverage.

Only 12 oyster spat were recovered and identified as Pacific oysters rather than the introduced flat oysters, while notable associated fauna included crabs, limpets, clams, and other benthic species, with no clear evidence that BESE treatment enhanced settlement density.

Due to challenging weather conditions, vessel availability constraints, and wind farm access restrictions, only three of four planned deployment trips were completed, with frame retrieval ultimately unsuccessful after 13 retrieval attempts; consequently, no oyster larvae were detected in samples and the effects of deployment timing and Reefpaste addition on spat abundance could not be determined.

Only 12 oyster spat were recovered and identified as Pacific oysters rather than the introduced flat oysters, while notable associated fauna included crabs, limpets, clams, and other benthic species, with no clear evidence that BESE treatment enhanced settlement density.

Due to challenging weather conditions, vessel availability constraints, and wind farm access restrictions, only three of four planned deployment trips were completed, with frame retrieval ultimately unsuccessful after 13 retrieval attempts; consequently, no oyster larvae were detected in samples and the effects of deployment timing and Reefpaste addition on spat abundance could not be determined.

Only 12 oyster spat were recovered and identified as Pacific oysters rather than the introduced flat oysters, while notable associated fauna included crabs, limpets, clams, and other benthic species, with no clear evidence that BESE treatment enhanced settlement density.

Due to challenging weather conditions, vessel availability constraints, and wind farm access restrictions, only three of four planned deployment trips were completed, with frame retrieval ultimately unsuccessful after 13 retrieval attempts; consequently, no oyster larvae were detected in samples and the effects of deployment timing and Reefpaste addition on spat abundance could not be determined.