4.5. Sandeel

4.5.1. Desktop Study

  1. There are five species of sandeel present in the North Sea which display differences in dwelling areas and abundances (e.g. intertidal zone versus deeper waters). While Raitt’s sandeel is the most abundant species supporting a large fishery in the North Sea, this section refers to sandeel species collectively, unless specified otherwise. Two sandeel species, Raitt’s sandeel and lesser sandeel are Scottish PMFs.
  2. The wider Forth and Tay SMR has been known historically to support important sandeel populations. The highest density of this population is focussed on the Wee Bankie (approximately 57 km west of the site boundary), however sandeel do range across much of the wider North Sea. In the early 1990s, there was a substantial industrial sandeel fishery on the Wee Bankie, Marr Bank and Berwick Bank sandbanks. By 1993, landings from this area had peaked at over 100,000 tonnes (Greenstreet et al., 2010b).
  3. In 2000, this industrial sandeel fishery was closed in response to concerns that the fishery was having a deleterious effect on sandeel stocks within the Forth and Tay SMR. The sandeel closure within this region (precautionary closure — Article 29a from Council Regulation No 850/88) had the effect of limiting sandeel fishing on most of the Forth and Tay SMR sandeel grounds. The fishery remains closed and sandeel abundance is monitored by Marine Scotland and ICES.
  4. In 2000, the first year of the closure of the Forth and Tay SMR sandeel fishery, high levels of recruitment, combined with a lack of any significant fishing activity resulted in an immediate and substantial increase in the biomass of sandeel on the Wee Bankie sandbank (Greenstreet et al., 2010b). However, between 2001 and 2010, sandeel biomass steadily declined to levels that were similar to those observed when the sandeel fishery was active (Greenstreet et al., 2010b). This was thought to be due to the absence of sustained recruitment, meaning that predation and other causes of natural mortality still exceeded population growth (Greenstreet et al., 2010b). More recently sandeel stocks have recovered leading to an increase in sandeel fishing adjacent to the closed area, however, ICES (2022c) recently stated “The escapement strategy [by which sandeel stocks are managed] is not sustainable for short-lived species unless the strategy is combined with a ceiling on fishing mortality”.
  5. Sandeel act as an umbrella species linking primary productivity to higher trophic levels and impacts on sandeel can cascade through the food chain. Sandeel feed exclusively on phytoplankton and zooplankton in the water column during the daytime and are an important prey items for many fish, seabirds, and marine mammals (Freeman et al., 2004; Engelhard et al., 2008). At night and during winter months, sandeel bury in the sediments, and high level site fidelity has been evidenced which puts them under potential threat of direct habitat loss (Jensen et al., 2011; Latto et al., 2013). This behaviour limits the habitat that sandeel can occupy to areas of very specific sediment particle sizes, where penetration into the sediment is possible. During the days of spring and summer months, sandeel tend to actively feed in schools within 10 km of their burying grounds (Wright et al., 2019). This is an adaptation to conserve energy and to avoid predation.
  6. Sandeel remain buried in the seabed between September and February with occasional emergence between December and February to spawn a single batch of demersal eggs that are deposited on the seabed (van der Kooij et al., 2008). Larvae hatch between February and April and drift with currents within the plankton for ten weeks (Wright and Bailey, 1996; Régnier et al., 2017; Proctor et al., 1998; Wright et al., 2019). After metamorphosis, juveniles return to the demersal environment and look for suitable areas of sand to inhabit. There are indications that the survival of sandeel larvae is linked to the availability of copepod prey in the early spring, especially Calanus finmarchicus, and that climate generated shifts in the Calanus species composition can lead to a mismatch in timing between food availability and the early life history of lesser sandeel (Wright and Bailey, 1996; van Deurs et al., 2009).
  7. Studies in the laboratory (Wright et al., 2000) and in the natural environment (Holland et al., 2005) have focussed on identifying the sediment characteristics that define the seabed habitat preferred by sandeel. Both approaches produced similar results, indicating that sandeel preferred sediments with a high percentage of medium to coarse grained sand (particle size 0.25 mm to 2 mm), and avoided sediment containing >4% silt (particle size <0.063 mm) and >20% fine sand (particle size 0.063 mm to 0.25 mm). As the percentage of fine sand, coarse silt, medium silt and fine silt (particles <0.25 mm in diameter) increased, sandeel increasingly avoided the habitat; this finding was also supported by Wright et al. (2000) as reported by Mazik et al. (2015). Conversely, as the percentage of coarse sand and medium sand (particles ranging from 0.25 mm to 2.0 mm) increased, sandeel showed an increased preference for this substrate.
  8. Work by Greenstreet et al. (2010a) draws on the research by Holland et al. (2005), to define four sandeel sediment preference categories, using hydro-acoustic seabed surveys and nocturnal grab surveys. This work merged fine sand, three silt grades and two coarser sand grades, to define two particle size classes, silt and fine sand and coarse sand, and then examined the combined effect of these two size grades of sediment particles on the percentage of grab samples with sandeel present. Latto et al. (2013) used this research, along with that described above by Wright et al. (2000) to produce four sandeel sediment preference categories, which were defined as: Prime, Sub Prime, Suitable and Unsuitable (refer to Table 4.5   Open ▸ ). To align with the Folk (1954) categories presented within the EUSeaMap seabed substrates layer, shown in Figure 4.14   Open ▸ to Figure 4.16   Open ▸ , these categories have been presented as preferred (prime and sub-prime), marginal (suitable) and unsuitable substrate classifications, taken from Latto et al. (2013).
  9. Further work has been completed by Langton et al. (2021) where a predicted distribution model for sandeel was developed, producing predicted density and probability of occurrence for sandeel around the British coastline. This modelling was undertaken based on the dependence of sandeel on particular habitat types, with the four main explanatory variables within the model being silt, depth, sand and slope, and was supported by sandeel fisheries data (e.g. data from Jensen et al., 2011). The results from Langton et al., (2021) within the fish and shellfish ecology study area were mapped, highlighting areas of importance for sandeel populations in the North Sea. Figure 4.13   Open ▸ presents the outputs of the modelling within the site boundary and shows that the whole site boundary has extremely low probability of sandeel presence, with areas where predicted density is high closer to the coasts or towards the Firth of Forth. These areas also correlate to previous studies which identified where marine mammals and birds are known to congregate and feed on sandeels (Langton et al., 2021).

Figure 4.13:
Model Derived Predictions of Density and Probability of Presence of Sandeel within the Site Boundary (Derived from Langton et al. (2021))

Figure 4.13: Model Derived Predictions of Density and Probability of Presence of Sandeel within the Site Boundary (Derived from Langton et al. (2021))


4.5.2. Site-Specific Surveys

  1. As outlined in section 3.3, site-specific survey data were collected and reviewed alongside desktop studies to assess the extent of suitable sandeel habitat within the site boundary. Grab sampling was undertaken (refer to section 3.3) and PSA completed on the sediment samples collected in 2022 within the site boundary which allowed classification of the sediment types according to Latto et al. (2013), as described in section 3.3. These classifications were originally developed for the marine aggregates industry, drawing on work from Greenstreet et al. (2010) and Holland et al. (2005), investigating spatial interactions between the aggregate application areas and sandeel habitat.
  2. Figure 4.14   Open ▸ illustrates the results of this site-specific analysis with sandeel habitat sediment preference classifications of preferred, marginal and unsuitable habitat denoted in Table 4.5   Open ▸ . The distribution of the habitat suitability shows that the site boundary is characterised by preferred (sub-prime; seven stations) and marginal habitat (37 stations), accounting for 56% of the site boundary, and the remaining 44% (34 stations) of sediments sampled were considered unsuitable for sandeel habitation. Excesses of mud prevent sandeel maintaining their burrows as the burrows are more likely to collapse and can reduce the ability of sandeel to respire due to fine particulate clogging gill tissues. The north-west section is mostly characterised by marginal and preferred habitats, while the south-east is covered by patches of unsuitable and marginal habitat.

 

Table 4.5:
Sandeel Habitat Sediment Classifications Derived from Latto et al. (2013)

Table 4.5: Sandeel Habitat Sediment Classifications Derived from Latto et al. (2013)

 

  1. Figure 4.14   Open ▸ illustrates the site-specific survey data alongside EMODnet seabed substrate data which can also be used to assign broadscale habitat suitability for sandeel. For the purposes of considering sandeel habitat suitability across the site boundary and surrounding areas, gravelly sand (between 30% and 5% gravel), slightly gravelly sand (between 5% and 1% gravel) and sand (under 1% gravel) in the EMODnet substrate data were classified as preferred habitat and sandy gravel (between 30% and 80% gravel) as marginal habitat. The substrates classified as preferred and marginal habitats all have a sand to mud ratio of nine to one or higher. Where no shading is present, the habitat in that area is considered unsuitable for sandeel.
  2. Overall, results from the site-specific surveys do not align completely with the EMODnet seabed substrate data; this is generally to be expected when comparing broadscale data with site-specific or point source information. The EMODnet seabed substrate data is of lower resolution and accuracy than the results of the site-specific survey and is based upon a high degree of interpolation, and so should be interpreted with caution due to not accounting well for local scale variance. Regardless, this data is a useful tool to support a broadscale regional characterisation of the general surrounding substrate. While the site-specific survey data shows the north-west portion as preferred and marginal habitat and south-east as a mosaic of unsuitable and marginal habitat, EMODnet data suggests that the whole site boundary is covered by slightly gravelly sand which is a preferred habitat for sandeel. These data highlight a degree of fine-scale variation that is not possible to resolve when working with broadscale data alone, and highlights the patchy nature of sandeel habitat within the site boundary.
  3. Further site-specific survey results from grab samples and epibenthic trawls have provided incidental data on the presence of sandeel within the site boundary; one grab sample and two beam trawl samples captured low numbers of sandeel (stations S058, BT001 and B002, see Figure 3.1   Open ▸ ) (see volume 3, appendix 8.1, annex A). These are shown in Figure 4.15   Open ▸ with records in grab samples shown as presence/absence and trawl data shown as abundances per 200 m trawled. The abundance data collected indicates higher abundances of sandeel in the north-west section of the site boundary which aligns with the composition of the sediments, which is less muddy and sandier compared to the south-east of the site boundary. However, it should be noted that neither grab sampling nor epifaunal trawling target sandeel specifically, therefore these results should be regarded as opportunistic. Conversely, whilst these opportunistic data may indicate higher abundances in specific areas (with regards to higher catchability due to higher density of burrows), it cannot be interpreted as low abundance or absence of sandeel where individuals were not recorded, due to the lack of specificity of sampling methods for sandeels. The site-specific survey data and desktop data indicate that sandeels are likely to be present across the site boundary, specifically in the north-west section, although almost half of the habitats recorded within the site boundary were assessed to be unsuitable or marginal grounds.
  1. As in section 4.4.2 for herring, Figure 4.16   Open ▸ expands on the site-specific data presented in Figure 4.14   Open ▸ , as it includes results of PSA conducted on sediment samples obtained from the OneBenthic Portal in the fish and shellfish ecology study area outside of the site boundary (Cefas, 2019a). Of the 669 samples obtained from the OneBenthic Portal, 335 were assessed as unsuitable, 162 as marginal, and 172 as preferred spawning habitat.

Figure 4.14:
Sandeel Habitat Preference Classification from EMODnet and Site-specific Survey Data

Figure 4.14: Sandeel Habitat Preference Classification from EMODnet and Site-specific Survey Data


Figure 4.15:
Sandeel Habitat Preference Classification with Site-specific Abundance Data

Figure 4.15: Sandeel Habitat Preference Classification with Site-specific Abundance Data


Figure 4.16:
Sandeel Habitat Preference Classification from EMODnet and Spawning Habitat Classifications from the Site-Specific Survey Data and the OneBenthic Portal

Figure 4.16: Sandeel Habitat Preference Classification from EMODnet and Spawning Habitat Classifications from the Site-Specific Survey Data and the OneBenthic Portal