Abstract
This report’s goal is to summarize seabirds at sea data available for the Netherlands and evaluate if the data is sufficient for a Before and After Control Impact (BACI) analysis for the effects of offshore wind farms (OWF) on Common guillemots (Peschko et al. 2024).
In general, there seems to be a good spatio-temporal survey coverage of the Dutch North Sea. However, each survey method has its limitations.
Ship-based surveys in the Netherlands have been conducted since the early 1980s. However, following the publication of two atlases (Baptist and Wolf 1993, Camphuysen and Leopold 1994), survey efforts on the Dutch Continental Shelf (DCS) became increasingly ad-hoc and project-based (van Bemmelen et al. 2011). Ship-based surveys therefore no longer covered the entire Dutch EEZ, and only dedicated surveys primarily those dedicated to OWF were being conducted (ICES 2020). For example, ship-based surveys were conducted in and around OWF prior to construction, as pre-construction (T-0) monitoring was required (Leopold et al. 2010, Leopold 2012, van Bemmelen et al. 2015). Some ship-based surveys have also been carried out post-construction (T-1) to monitor impacts after the OWFs became operational, or to evaluate whether birds will recognise a corridor as a safe passage route through the wind farm (Leopold 2012, Leemans et al. 2024). In recent literature, ship-based survey data have been considered unbalanced and inconsistent, particularly for the last 10 to 15 years. Nevertheless, some studies still incorporated incidental and systematic ship-based survey data in their analyses (Leopold et al. 2015, van Kooten et al. 2019), while other opted not to use it (van Donk et al. 2024). It should be mentioned that modern regression methods (such as applied for the planned BACI analysis) can cope with a certain degree of unbalance.
Observer-based aerial surveys provide good spatial and temporal coverage and are generally considered to have more consistent coverage than ship-based surveys (van Bemmelen et al. 2011, van Donk et al. 2024). Post-construction plane surveys, as part of the MWTL surveys, are also conducted crossing the OWF throughout the shipping lane (Collier et al. 2022), the regular monitoring passes Gemini wind farm in a certain distance, at the standard altitude of 75 m. Note, however, that the methodology changed after 2014. Prior 2014, planes flew at higher altitudes (150 m) (Baptist and Wolf 1993, Arts 2010, Poot et al. 2016), and thus species such as Common guillemots and Razorbills could not be reliably distinguished (van Donk et al. 2024) but rather assessed on the level of auks only. Furthermore, the MWTL monitoring program started using bubble windows as a standard after 2014, which could suggest, that in surveys previous 2014, the transect line below the plane could not be properly observed, and consequently many birds on the transect line were probably missed (Poot et al. 2016). To address the issue of individuals not determined to species level, some studies using observer-based aerial surveys after 2014, have estimated the proportion of each species, and applied these proportions to assign species to undetermined Common guillemots / Razorbills (van Bemmelen et al. 2022). This might be appropriate if total species-specific numbers are estimated. However, if species-specific OWF-effects are evaluated, this approach is not appropriate, as it does not solve the problem that species-specific reactions cannot be extracted from individuals that are not determined at species level. Other studies, after exploratory analyses, have decided that aerial surveys conducted before 2014 yielded no usable species-level data (van Donk et al. 2024). Because assessing the effects of offshore wind farms (OWFs) via a BACI analysis following Peschko et al. 2024 typically requires a dataset spanning six years, this lack of usable pre-construction species specific data might pose a limitation for the analysis of some OWF, or analysis methods need to be adapted accordingly. It is however probable, that data collected in the same time span via ship-based surveys and observer-based aerial surveys can be combined. Although the analysis is highly likely, a final decision on how to deal with the data before 2014 can only be made once the data including survey effort, weather conditions, and the sightings of Common guillemots, Razorbills and undetermined Auks is available and can be evaluated in detail.
Digital aerial surveys are only project based, and no regular monitoring in the DCS is conducted. For digital surveys only two studies were found, and both cover the post construction period of OWFs. One study included eight surveys in Gemini: the first as pilot in March 2022, and seven following between October 2022 and March 2023 (Grundlehner et al. 2025). The other study was conducted in Borssele (Collier et al. 2022) which included surveys carried out between 2021 to 2022. There are more digital surveys coming in the next years (Leopold et al. 2021, Bos et al. 2023, Fijn pers. comm., Zomer pers. comm.), but at the present time, it was not clear when and where.
Based on the current overview, it appears highly likely that sufficient data have been collected to support a BACI analysis. Compared to the studies on OWF effects conducted in the German EEZ, effort and guillemot sightings collected via ship-based surveys (currently available through the ESAS data base) are available in a similar order of magnitude, which makes a successful analysis highly likely.
To date it is however impossible to assess the outline and precision of a potential BACI analysis, since several aspects concerning the amount of data available per OWF cluster and per data collection method (ship-based, aerial observer-based or digital surveys) are still partly unclear since the data is not yet entirely available to us. Relevant information per OWF cluster such as the surveys methods used, the area covered (in km2), the weather conditions during the survey, and detailed sightings data for Common guillemots, Razorbills and individuals recorded as undetermined Auks would be required. Some of this information is openly available in the ESAS database (mainly for ship surveys in the focused area). However, key parts of this information could be located in some, but not all, of the available reports in such detail that conclusions on the data availability and quality would be possible on the spatial level of OWF clusters. This probably was due to the reports targeting different questions as the one targeted here. Consequently, our understanding on data availability, particularly for observer based aerial surveys, remains incomplete. Once this data is accessible, the proportion of birds identified to species level in observer-based aerial surveys could be determined. As mentioned before, it would be necessary to evaluate the proportion of animals detected at species level on the specific area of interest. Without a better understanding on the proportion of species-level identification prior to 2014, it currently remains unclear whether these data can be included in the analysis. Thus, although we have identified necessary actions for a future analysis, such as compiling all data available across survey methods, and assessing the proportion of records identified to species level, some uncertainty remains regarding general aspects. These include the amount of survey effort available for each OWF and survey method, as well as the number of individuals determined to species level, during both the pre- and post- construction phases. Consequently, whether all operational OWF are to be included in the analyses, as well as the level of analytical precision that can be expected, is due to future analyses based on the entire dataset available.
Based on the current understanding of data availability, we assume that a BACI analysis will be feasible, either for single or clustered OWF. Given the temporal and spatial occurrence of Common guillemots in the Dutch North Sea, the analyses will likely focus on the winter and autumn seasons. Where possible, we aim to incorporate all available data collected through different survey methods into a combined analysis (Mercker et al. 2021, Garthe et al. 2023), while explicitly accounting for methodological limitations, such as those associated with aerial survey data collected before 2014.
Once all data sources have been compiled (a task that would be undertaken in a potential part B of the project), it will be possible to evaluate which OWF areas are adequately represented. Based on the available data, decisions will then be made regarding which OWF, and which survey methods can be included in the analysis. It will therefore be essential to obtain full access to all relevant datasets, in order to assess their quantity and quality and to further develop the analytical approaches. It could be valuable to include data collected by neighbouring countries that surveyed the areas of interest (see Figure S25 to S27). For example, data collected in Belgium, especially in and around the Belgian OWF located along the border to the Dutch EEZ, could be potentially included (see Figure S27 and S28). In the OWF as well as in control areas close by, data was collected monthly since the year 2009 via ship-based surveys (https://odnature.naturalsciences.be/mumm/en/windfarms/), thus a comparably good coverage would be available. This data could provide valuable additional information for the analysis because they covered the OWF regularly with ship surveys since 2009. Likewise, also data that was collected in the German EEZ could be included in an analysis.
Access to all available data is therefore important. The datasets should be provided as raw data collected with the different survey methods. Datasets should include the surveyed transects (latitude and longitude), the effort covered (in km2), together with the weather conditions during the survey such as seastate, glare and turbidity. Moreover, it should include the sightings of Common guillemots, Razorbills and individuals recorded as undetermined Common guillemots / Razorbills/Auks collected, accompanied to latitude and longitude of the sightings and a position key to link to the basic information on the survey effort.