Guillemot

155. For guillemot, there was a similar relationship to that of kittiwake between adult return rate at the Isle of May and sandeel stock biomass. Adult guillemot return rates correlated with sandeel stock biomass that was best described by the function y = 0.0532ln(x) + 0.224 ( Figure 1.9   Open ▸ ). This non-linear relationship indicated that adult survival could increase rapidly when sandeel total stock biomass was increased from below 300,000 tonnes. Further increasing sandeel abundance would improve guillemot survival less when the stock biomass was above 300,000 tonnes.
156. From the equation in Figure 1.9   Open ▸ , the predicted change in adult survival was calculated for each scenario ( Table 1.6   Open ▸ ).

Table 1.6 Predicted change in adult return rate of guillemots on the Isle of May with potential changes in sandeel stock.

157. From the difference in adult survival predicted due to sandeel stock biomass it is possible to predict the number of additional adult birds that could survive as a result. For each SPA in SA4 between the East Caithness Cliffs SPA in the north of SA4 and the Farne Islands SPA in the south of SA4 the additional adult guillemots predicted to survive per annum is shown in Table 1.7   Open ▸ .

Table 1.7 Additional adult guillemots predicted to survive per annum for SPAs in SA4.

158. Adult guillemot productivity also correlated with sandeel stock biomass ( Figure 1.10   Open ▸ ). This non-linear relationship indicated that productivity could increase rapidly when sandeel total stock biomass was increased from below 300,000 tonnes. Further increasing sandeel abundance would improve guillemot productivity much less when the stock biomass was above 300,000 tonnes.
159. From the equation in Figure 1.10   Open ▸ , the predicted change in adult guillemot productivity was calculated for each scenario ( Table 1.8   Open ▸ ).

Table 1.8 Predicted change in productivity of guillemots (chicks per pair) on the Isle of May with potential changes in sandeel total stock biomass.

160. From the difference in productivity predicted due to sandeel stock biomass it is possible to predict the number of additional adult guillemots that could be produced as a result. For each SPA in SA4 between the East Caithness Cliffs SPA in the north of SA4 and the Farne Islands SPA in the south of SA4 the predicted additional adult birds predicted to be produced per annum is shown in Table 1.9   Open ▸ . This was calculated from the predicted increase in fledged birds per annum combined with the predicted guillemot survival rate from fledging to age at first breeding (0.4243).

Table 1.9 Additional adult guillemots predicted to be produced per annum for SPAs in SA4.

Razorbill

161. For razorbill, Furness et al. (2013) concluded that the closure of sandeel and sprat fisheries was likely to be effective but with insufficient evidence to have a high confidence in that assessment.
162. For razorbill, there was also a relationship similar to that of kittiwake and guillemot between adult return rate at the Isle of May and sandeel stock biomass. Adult razorbill return rates correlated with sandeel stock biomass that was best described by the function y = 0.0438ln(x) + 0.3372 ( Figure 1.12   Open ▸ ). This non-linear relationship indicated that adult survival could increase rapidly when sandeel total stock biomass was increased from below 300,000 tonnes. Further increasing sandeel abundance would improve razorbill survival much less when the stock biomass was above 300,000 tonnes.
163. From the equation in Figure 1.12   Open ▸ , the predicted change in adult razorbill survival was calculated for each scenario ( Table 1.10   Open ▸ ).

Table 1.10 Predicted change in adult return rate of razorbills on the Isle of May with potential changes in sandeel total stock biomass.

164. From the difference in adult survival predicted due to sandeel stock biomass it is possible to predict the number of additional adult razorbills that could survival as a result. For each SPA in SA4 between the East Caithness Cliffs SPA in the north of SA4 and the Farne Islands SPA in the south of SA4 the additional birds predicted to survive per annum is shown in Table 1.11   Open ▸ .

Table 1.11 Additional adult razorbills predicted to survival per annum for SPAs in SA4.

165. Unlike kittiwake and guillemot, adult razorbill productivity did not correlate with sandeel stock biomass ( Figure 1.13   Open ▸ ). There was therefore no evidence to support the hypothesis that sandeel fisheries management would result in changes in razorbill productivity.

Puffin

166. For puffin, Furness et al. (2013) concluded that the closure of the sandeel and sprat fisheries was likely to be effective but with insufficient evidence to have a high confidence in that assessment, at the time of publication.
167. For puffin, there was a similar relationship to that of kittiwake, guillemot and razorbill between adult return rate at the Isle of May and sandeel stock biomass. Adult puffin return rates correlated with sandeel stock biomass that was best described by the function y = 0.0963ln–x) - 0.3563 ( Figure 1.15   Open ▸ ). This non-linear relationship indicated that adult survival could increase rapidly when sandeel total stock biomass was increased from below 300,000 tonnes. Further increasing sandeel abundance would improve puffin survival much less when the stock biomass was above 300,000 tonnes.
168. From the equation in Figure 1.15   Open ▸ , the predicted change in adult puffin survival was calculated for each scenario ( Table 1.12   Open ▸ ).

Table 1.12 Predicted change in adult return rate of puffins on the Isle of May with potential changes in sandeel total stock biomass.

169. From the difference in adult survival predicted due to sandeel stock biomass it is possible to predict the number of additional adult puffins that could survival as a result. For each SPA in SA4 between the East Caithness Cliffs SPA in the north of SA4 and the Farne Islands SPA in the south of SA4 the additional adult puffins predicted to survive per annum is shown in Table 1.13   Open ▸ .

Table 1.13 Additional adult puffins predicted to survive per annum for SPAs in SA4.

170. Adult puffin productivity also correlated with sandeel stock biomass ( Figure 1.16   Open ▸ ). This non-linear relationship indicated that productivity could increase rapidly when sandeel total stock biomass was increased from below 300,000 tonnes. Further increasing sandeel abundance would improve puffin productivity much less when the stock biomass was above 300,000 tonnes.
171. From the equation in Figure 1.16   Open ▸ , the predicted change in adult survival was calculated for each scenario ( Table 1.14   Open ▸ ).

Table 1.14 Predicted change in productivity of puffins (chicks per pair) on the Isle of May with potential changes in sandeel stock.

172. From the difference in productivity predicted due to sandeel stock biomass it is possible to predict the number of additional adult puffins that could be produced as a result. For each SPA in SA4 between the East Caithness Cliffs SPA in the north of SA4 and the Farne Islands SPA in the south of SA4 the additional adult puffins predicted to be produced per annum is shown in Table 1.15   Open ▸ . This was calculated from the predicted increase in fledged birds per annum combined with the predicted puffin survival rate from fledging to age at first breeding (0.4342).

Table 1.15 Additional adult puffins predicted to be produced per annum for SPAs in SA4.

Discussion and Conclusions

173. A scenario testing approach was used as future predictions of the absolute sandeel TSB in SA4 cannot be made with sufficient certainty. The tested scenarios were an effective approach to determining whether a likely worst-case benefit from reduction or removal of fishing pressure in SA4 would result in sufficient change in seabird demography to compensate for the likely worst case impact prediction. Thus, the approach here was not to attempt to model a likely future sandeel TSB in SA4, and therefore likely future seabird population sizes, but to demonstrate that the minimum gain from compensation would sufficient.
174. Likely gains to the SPA populations of kittiwake, guillemot, razorbill, and puffin varied across the five compensation scenarios. The scenario that produced the smallest benefit to SPA populations was consistently the change in sandeel TSB from 300,000 to 400,000 tonnes. The largest predicted change to SPA population was from the change in sandeel TSB from 100,000 to 200,000 tonnes. This was due to the shape of the relationships between species survival and productivity on the Isle of May and sandeel TSB in SA4.
175. The predicted additional seabirds from predicted changes in demography as a result of the worst case compensation scenario was shown to be more than sufficient to compensate for the worst case predicted impacts from the proposed development.

1.9. Assessment of population level effects on SPA qualifying features

176. The assessment of population level effects on SPA qualifying features was completed using three approaches to provide different information on the predicted effects of compensation measures.

• Additional adult survival and productivity - based on the predicted benefits to adult survival and productivity from reducing or removing fishing pressure in SA4 the change in adult population size, as a result of the combined negative effects from the predicted impacts and the predicted benefits from the compensation measures, was used to calculate the number of additional adult birds per annum for each SPA qualifying feature. The predicted additional number of birds surviving per annum was compared with the predicted impact to estimate the compensation ratio from the proposed compensation measure.
• Comparison of PVA outputs from predicted impacts – PVA models were used to compare the effects of impacts from the Proposed Development alone and the effects of these impacts combined with the positive effects on population demography from the proposed compensation measures. The PVA output metrics used were the counterfactual of population size (CPS) and the counterfactual of population growth rate (CGR).
• Comparison of PVA outputs from a range of impact levels – by varying the impacts on the modelled population it was possible to determine the highest impact level that did not result in a positive population outcome from the predicted benefits from the proposed compensation measure.

SPAs & features

177. For each SPA within SA4, the qualifying features were identified and the population size, population change and site condition monitoring (SCM) results of the features predicted to benefit from sandeel fisheries compensation measures were summarised.

Forth Islands SPA

178. The Forth Islands SPA is a group of islands in the Firth of Forth on the east coast of Scotland designated for its breeding seabird populations. The site qualified for its Annex I breeding populations of Arctic tern Sterna paradisaea, roseate tern Sterna dougallii, common tern Sterna hirundo and Sandwich tern Sterna sandvicensis. It further qualified for its breeding migratory populations of northern gannet Morus bassanus, lesser black-backed gull Larus fuscus and in excess of 20,000 individual seabirds, including razorbill Alca torda, common guillemot Uria aalge, black-legged kittiwake Rissa tridactyla, herring gull Larus argentatus, great cormorant Phalacrocorax carbo, European shag, and Atlantic puffin.
179. Four qualifying features in the SPA could be positively influenced by sandeel fisheries changes: kittiwake, guillemot, razorbill and puffin. Among these four features only kittiwake has declined since the site was designated ( Table 1.16   Open ▸ ) and this is the only feature in unfavourable site condition.

Table 1.16   Open ▸ Current population size, population change and site condition of key qualifying features of the Forth Islands SPA.

Fowlsheugh SPA

180. The Fowlsheugh SPA is a stretch of 30 – 60 m high sheer cliffs on the east coast of Scotland designated for its breeding seabird populations. The site qualified for its breeding migratory populations of more than 20,000 individual seabirds, including common guillemot, black-legged kittiwake, razorbill, northern fulmar Fulmarus glacialis, and herring gull Larus argentatus.
181. Three qualifying features in the SPA that could be positively influenced by sandeel fisheries changes: kittiwake, guillemot, and razorbill. Among these three features only kittiwake has declined since the site was designated ( Table 1.17   Open ▸ ). All qualifying features are in favourable site condition according to NatureScot Sitelink V3, although the decline in kittiwake breeding numbers is large.

Table 1.17 Current population size, population change and site condition of key qualifying features of the Fowlsheugh SPA.

St Abb's Head to Fast Castle SPA

182. The St Abb's Head to Fast Castle SPA is a stretch of cliffs on the east coast of Scotland designated for its breeding seabird populations. The site qualified for its breeding migratory populations of more than 20,000 individual seabirds, including common guillemot, black-legged kittiwake, razorbill, herring gull and European shag.
183. Three qualifying features in the SPA that require compensation due to predicted impacts from the Proposed Development could be positively influenced by sandeel fisheries changes: kittiwake, guillemot, and razorbill. Among these three features only kittiwake has declined since the site was designated ( Table 1.18   Open ▸ ) and this is the only feature in unfavourable site condition.

Table 1.18 Current population size, population change and site condition of key qualifying features of the St Abb's Head to Fast Castle SPA.

Buchan Ness to Collieston Coast SPA

184. The Buchan Ness to Collieston Coast SPA is a stretch of cliffs on the east coast of Scotland designated for its breeding seabird populations. The site qualified for its breeding migratory populations of more than 20,000 individual seabirds, including common guillemot, black-legged kittiwake, herring gull, European shag and northern fulmar.
185. Two qualifying features in the SPA could be positively influenced by sandeel fisheries changes: kittiwake and guillemot. Between these two features only kittiwake has declined since the site was designated ( Table 1.19   Open ▸ ) and this is the only feature in unfavourable site condition.

Table 1.19 Current population size, population change and site condition of key qualifying features of the Buchan Ness to Collieston Coast SPA.

Farne Islands SPA

186. The Farne Islands SPA comprises a group of low-lying islands off the coast of Northumberland in north-east England. The site qualified for its Annex I breeding populations of Arctic tern, roseate tern, common tern and Sandwich tern. It further qualified for its breeding migratory population of common guillemot and further for holding in excess of 20,000 individual seabirds, including black-legged kittiwake, great cormorant, European shag, and Atlantic puffin.
187. Three qualifying features in the SPA that could be positively influenced by sandeel fisheries changes: kittiwake, guillemot and puffin. Among these features only kittiwake has clearly declined since the site was designated ( Table 1.20   Open ▸ ). The change in guillemot abundance shown is likely within the error of the count, so the population has changed little since designation.

Table 1.20 Current population size, population change and site condition of key qualifying features of the Farne Islands SPA

Troup, Pennan and Lion's Heads SPA

188. The Troup, Pennan and Lion’s Heads SPA is a stretch of cliffs along the north coast of Aberdeenshire. The site qualified for its breeding migratory populations of kittiwake, guillemot and assemblage of breeding seabirds including razorbill, fulmar and herring gull.
189. Three qualifying features that could be positively influenced by sandeel fisheries changes: kittiwake, guillemot and razorbill ( Table 1.21   Open ▸ ). Both kittiwake and guillemot have declined at this site and are in Unfavourable condition. Despite a population increase in recent years, the condition of razorbill is still listed by NatureScot as Unfavourable Declining.

Table 1.21 Current population size, population change and site condition of key qualifying features of the Troup, Pennan and Lion’s Heads SPA

East Caithness Cliffs SPA

190. The East Caithness Cliffs SPA includes most of the sea cliffs from Wick or Helmsdale on the east coast of Caithness in north of the Scottish mainland. The site qualified for its Annex I breeding population of peregrine, its breeding migratory populations of guillemot, razorbill, kittiwake, herring gull and shag, and its breeding seabird assemblage of 300,000 individual seabirds including guillemot, razorbill, kittiwake, herring gull, great black-backed gull, fulmar, cormorant and shag ( Table 1.22   Open ▸ ). The kittiwake population has declined, though it is listed as Favourable and in Maintained condition by NatureScot. Both guillemot and razorbill populations have increase substantially and both are in Favourable condition.

Table 1.22 Current population size, population change and site condition of key qualifying features of the East Caithness Cliffs SPA

Flamborough and Filey SPA

191. The RIAA found that the in-combination impact on kittiwake and razorbills populations at the Flamborough and Filey Coast (FFC) SPA from the Proposed Development, and other reasonably foreseeable plans and projects, could be sufficient that it may not be possible for the competent authority to conclude that there was no adverse effect on site integrity. The FFC SPA is not adjacent to SA4, where the proposed compensatory measures will be applied. It therefore not possible to apply the same evidence of changes in sandeel TSB and seabird demography to the FFC SPA, as the majority of individuals will be foraging within SA1r, not SA4. The results of the RIAA of impacts from the Proposed Development on the FFC SPA is, in part, due to the apportioning of impacts based on, among other parameters, the mean of the maximum foraging range of these species overlapping with the footprint of the Proposed Development. This apportioning uses foraging range information based on tracking of seabirds during the breeding season and applies this distance equally in all directions.
192. Tracking of kittiwakes from the FFC SPA indicates that birds are foraging entirely in SA1r (Wischnewski et al. 2017) and not in SA4, during the chick rearing phase of the breeding season at least. The apportioning would also suggest that only a relatively small proportion of the FFC SPA populations for both of these species would forage in SA4. Any direct benefit to the FFC SPA from changes in sandeel fisheries management in SA4 would therefore only directly benefit a relatively small proportion of the FFC SPA populations, but it would be expected that any birds from FFC SPA that did forage in SA4 would benefit from the proposed sandeel fisheries compensation measures. However, this proportion is difficult to estimate and would be a major source of uncertainty in any quantitative assessment of the impacts and benefits of the proposed sandeel fisheries compensation measures to that part of the FFC SPA populations. However, it is likely that the whole of the FFC SPA colony would benefit from two key consequences of the proposed sandeel fisheries compensation measures:

• Increase in the number of recruits from colonies in SA4 available to immigration in to the FFC SPA; and
• Spill over effects from the reduction in sandeel fishing mortality in to SA1r.

193. The predicted increases in kittiwake and razorbill populations in SPAs where birds forage in SA4 would very likely result in large increases in populations size, as a result of increases in both adult survival and productivity for kittiwake and adult survival for razorbill. This would likely result in general population increases across all colonies that forage in SA4 (i.e. including colonies not in SPAs) resulting in a large number of recruits available for emigration into other colonies on the North Sea coast. Ringing information shows that kittiwake (Coulson & Coulson 2008) tends to recruit to colonies away from their natal colony. However, natal recruitment appears to be relatively high in razorbill (Lavers et al. 2007), in the western Atlantic at least. Though these authors also stated that dispersal and recruitment into new colonies partly explained the population dynamics in these populations of razorbill.
194. Spill over effects in sandeels are somewhat limited, as adult fish are highly sedentary. However, larval spill over is likely to occur from sandbanks in SA4, over distances of about 100 – 150 km (Sørensen et al. 2009). The net water flow in the North Sea is southwards along the North Sea coast of Scotland and England, and so from SA4 to SA1r, which would tend to move some of any increased abundance of sandeel larvae in the southern part of SA4 into SA1r. This has the potential to increase sandeels in SA1r, where most of the kittiwakes and razorbills from the FFC SPA forage. It is difficult to predict the level of increase in sandeels available to seabirds in SA1r as a result of spill over from SA4 as a result of changes to fisheries management. However, Sørensen et al (2009) modelled the closure of a single ICES rectangle on the Dogger Bank (ICES 37F2). They stated that this would result in, “the total southern North Sea [sandeel] yield increasing by 16% based on a crude assumption of effort response”. Thus, it is very likely that changes to the management of sandeel fisheries in SA4 would result in more substantial spill over effects in to SA4.
195. Given the predicted impacts on the FFC SPA for both are extremely small from the Project alone, the proposed sandeel fisheries compensation measures are very likely to be sufficient to also provide adequate compensation to the FFC SPA.

Additional adult survival and productivity

196. The number of additional adult birds in each SPA population as a result of the increase in adult survival and productivity predicted to occur from the increase in sandeel TSB from 300,000 tonnes to 400,000 tonnes was calculated. The gain from the increase in adult survival was calculated from the change in survival applied to each SPA population size. The gain from the increase in productivity was calculated by multiplying the change in productivity by the number of pairs of birds in each SPA and then adjusting the number of birds by the overall survival from fledging to age at first breeding. This provides an estimate of the additional birds available for recruitment into the population. Not all birds that fledge and survive to recruit will be added to either each SPA population or the SPA network. However, the proposed compensation measure is very likely to result in increased adult survival and productivity to all colonies that forage within SA4, thus the total number of birds available to recruit into the whole meta-population should increase. As a result, the predicted increase in population size due to productivity increases from the compensation measure is likely to be a reasonable estimate.
197. These predicted increases were compared with the predicted impacts across the three impact scenarios. These are summarised in Table 1.23   Open ▸ .

Table 1.23 Predicted increase in the number of adult birds surviving per annum as a result of the combined compensation measures minus the predicted impacts for each SPA qualifying feature impacts and in total.

198. Since the aim of the compensation measures is to ensure the coherence of the UK SPA network, the overall benefit of the compensation measure compared with the overall impact to the SPA network was used to estimate the compensation ratios for each species for each impact scenario ( Table 1.24   Open ▸ ). This shows that the potential lowest increase due to compensation (based on the increase in sandeel TSB from 300,0000 to 400,0000 tonnes) would result in at least a 1:7 compensation ratio (for guillemot) and up to a 1:40 ratio for puffin, if the higher scoping opinion impact level was assumed. Ratio are even larger for the other impact scenarios.

Table 1.24 Estimated compensation ratios for each SPA qualifying feature with potential to be impacted across all three impact scenarios

Using PVA to assess compensation

Scenario testing approach

199. The assessment of additional adult survival and productivity provided good evidence that the increase in adult survival alone or the increase in adult productivity alone would be more than sufficient to compensate for the predicted impacts from the Proposed Development. However, it was considered important to show that the effects of the compensatory measure on demographic rates result in longer term increases in population size of each impacted SPA. The approach to assessing the effects of compensation measures was to use a scenario testing approach using PVA models.
200. There is very strong empirical evidence to demonstrate that reducing or removing fishing pressure in SA4 would have a highly beneficial effect on the impacted seabirds. However, it is necessary to quantify this benefit to demonstrate that the benefits generated offset the potential impact on seabirds. In order to provide as much certainty as possible that reducing or removing fishing pressure in SA4 will be effective, a number of scenarios were developed based on plausible changes in the sandeel stock biomass. Consequently, scenarios were based on a range of plausible changes in sandeel stock biomass. The historic stock biomass of sandeels in SA4 was about 900,000 tones. Therefore, an estimated Cury et al. (2011) one-third-for-the-birds threshold was approximately 300,000 tonnes. Five potential scenarios were considered useful in assessing the benefits of the proposed compensation measure on seabird SPA populations. Increases in sandeel stock biomass from below 300,000 tonnes were considered useful, with changes from 100,000 to 200,000 tonnes and 200,000 tonnes to 300,000 tonnes estimated (see Figure 1.26   Open ▸ ). In addition, increases in stock biomass from 300,000 tonnes to 400,000, 600,000 and 800,000 tonnes provided information on plausible future sandeel stock biomass levels being maintained above the Cury threshold but below the historic maximum stock biomass.
201. The PVAs were run using five scenarios for how the compensation could affect demographic rates (where evidence allowed). These scenarios allowed for positive changes to adult survival only, positive changes to productivity only, and positive changes to both adult survival and productivity simultaneously. The positive effects on these demographic parameters are not independent as both adult survival and productivity are predicted to be affected by sandeel stock biomass. As such the scenarios where only adult survival was increased, or only productivity was increased, were precautionary. However, the scenarios where both survival and productivity were increased were likely to be the most realistic. The effects of change in sandeel stock biomass were based on correlations between adult return rate (a proxy for survival) or productivity and the sandeel stock biomass estimated by ICES. The relationships used in the primary testing of compensation described here were based on publicly available data from the Isle of May ( Table 1.25   Open ▸ ). These relationships were estimated from the data from 2004 to 2019, during which time the seabird populations were recovering from previously very low sandeel stock biomass levels in SA4. Note that there was no apparent relationship between sandeel stock biomass and productivity of razorbills, so this parameter was kept static in all scenarios.

Table 1.25 Summary of the correlations between demographic parameters and sandeel stock biomass in SA4 from 2004 – 2019, with some exceptions (see text). Data from the Isle of May studies by UKCEH.

202. Precaution was incorporated into the scenario testing approach through the lack of change to immature survival and age at first breeding. Both of these demographic elements would be expected to benefit from sandeel fisheries compensation measures. However, there was no empirical information to parameterise the increase in immature survival or the decrease in age at first breeding. In addition, the colonies of seabirds within SPAs do not occur as discrete populations but are part of a larger meta-population. The predicted changes to adult survival and productivity should be expected to positively affect all seabird colonies where birds forage within SA4, including those colonies not designated as SPAs. This overall change in population size across SPA and non-SPA colonies would likely increase population stability and resilience to future pressures.

Baseline

203. The PVA assessments of the effects of reducing or removing fishing pressure from SA4 were based on the values used in the PVA for the RIAA, with the exception of the survival and/or productivity values obtained from the relationships between adult return rate (as a proxy for survival, referred to as survival hereafter) and productivity. The baseline against which impacts, and compensation for those impacts, was compared, was dependent on the change in survival and/or productivity from the relationships summarised in Table 1.25   Open ▸ . The baseline survival and productivity values were calculated from the relationship between sandeel stock biomass and productivity for the TSB values of 100,000, 200,000 and 300,000 tonnes. These values were used for each scenario as appropriate. The starting population sizes for each SPA qualifying feature, and the year of the count, are shown in Table 1.26   Open ▸ .. All PVA input demographic values are provided in Table 1.27   Open ▸ .

Table 1.26 Initial population size and year of count from each SPA colony assessed.

Table 1.27 Demographic values used in PVA assessments of the efficacy of proposed compensation measures.

Impacts on adult survival

204. For each SPA qualifying feature tested, as the level of impact required to result in a counterfactual of population growth rate value less than one (i.e. the population growth rate would be less than the baseline). This change in sandeel stock biomass was found to result in the smallest predicted change in demographic rates of the scenarios tested above, so was the most precautionary estimate to use. Based on the most recently available estimate of population size the impact on adult survival was calculated for a series of three impact levels up to the level where the CGR value was less than one. These are summarised in Table 1.28   Open ▸ . For each SPA qualifying feature, the overall change in adult survival and productivity predicted to occur due to both the negative effects of the predicted impacts from the Proposed Development offshore wind farm and the positive effects predicted from changes in sandeel biomass from 300,000 to 400,000 tonnes in SA4 were calculated Table 1.28   Open ▸ .

Table 1.28 Summary of the predicted mortality of adult birds from SPAs and the calculated effect on adult survival

PVA projections

205. The results are presented below for projected CGR and CPS values based on the assumption that impacts begin in 2027 and end in 2077 (i.e. 50 years). A longer population model projection was used for the compensation measures assessment than for the RIAA as compensation measures should be provided in the long term. Seabirds are particularly long-lived birds, so longer population projections were used to assess the long-term effects of the measure. The impacts were assumed to continue for the entire duration of the run, so that the results were both precautionary and in relation to the predicted impact levels. In reality, impacts would cease after the decommissioning of the wind turbines. The absolute population size increases from the PVA’s described here are all likely to be unrealistic, as they are assumed to be density independent and closed populations (i.e. no immigration or emigration). However, PVA model results are best interpreted as relative differences rather than as absolute predictions of a likely future condition. In particular, the difference in population growth rates is likely to be the most informative.

Comparison of PVA outputs from predicted impacts

206. For each of the SPAs predicted to be impacted by more than a trivial amount, PVAs were run for the scenarios described above.

Kittiwake

207. The median counterfactual metrics (CGR and CPS) for the impact scenarios were all below one for all three impact scenarios ( Table 1.29   Open ▸ ). All the confidence interval metrics were also below one. The CGR and CPS median values were greater than one for all compensation scenarios where predicted compensation effects from a change in sandeel TSB from 300,000 tonnes to 400,000 tonnes were combined with predicted impacts. This is strongly indicative that the proposed compensatory measures will overcome the predicted impacts for kittiwakes at all SPAs assessed. In all cases the combination of the most precautionary (i.e. largest) impact and the most precautionary estimates of compensation (i.e. smallest increase in stock biomass) generated overall increases in population growth and population size.

Table 1.29 PVA metrics (CGR & CPS) from kittiwake model projections of impacts from the Proposed Development alone and impacts from the Proposed Development  alone minus the beneficial effects of proposed compensation metrics (based on a change in sandeel TSB from 300,000 tonnes to 400,000 tonnes). Shaded cells are larger than one.

Guillemot

208. The median CGR and CPS metrics for the impact scenarios were all below one for all three impact scenarios ( Table 1.30   Open ▸ ). Some of the upper confidence intervals for the CGR and CPS values were greater than one for some impact scenarios at some SPAs. The CGR and CPS median values were greater than one for all compensation scenarios where predicted compensation effects from a change in sandeel TSB from 300,000 tonnes to 400,000 tonnes were combined with predicted impacts. This is strongly indicative that the proposed compensatory measures will overcome the predicted impacts for guillemots at all SPAs assessed. In all cases the combination of the most precautionary (i.e. largest) impact and the most precautionary estimates of compensation (i.e. smallest increase in stock biomass) generated overall increases in population growth and population size.

Table 1.30 PVA metrics (CGR & CPS) from guillemot model projections of impacts from the Proposed Development alone and impacts from the Proposed Development alone minus the beneficial effects of proposed compensation metrics (based on a change in sandeel TSB from 300,000 tonnes to 400,000 tonnes). Shaded cells are larger than one

Razorbill

209. The median CGR and CPS metrics for the impact scenarios were all below one for all three impact scenarios ( Table 1.31   Open ▸ ). Some of the upper confidence intervals for the CGR and CPS values were greater than one for some impact scenarios at some SPAs. The CGR and CPS median values were greater than one for all compensation scenarios where predicted compensation effects from a change in sandeel TSB from 300,000 tonnes to 400,000 tonnes were combined with predicted impacts. This is strongly indicative that the proposed compensatory measures will overcome the predicted impacts for razorbills at all SPAs assessed. In all cases the combination of the most precautionary (i.e. largest) impact and the most precautionary estimates of compensation (i.e. smallest increase in stock biomass) generated overall increases in population growth and population size.

Table 1.31 PVA metrics (CGR & CPS) from razorbill model projections of impacts from the Proposed Development alone and impacts from the Proposed Development alone minus the beneficial effects of proposed compensation metrics (based on a change in sandeel TSB from 300,000 tonnes to 400,000 tonnes). Shaded cells are larger than one

Puffin

210. The median CGR and CPS metrics for the impact scenarios were all below one for all three impact scenarios ( Table 1.32   Open ▸ ). Most of the upper confidence intervals for the CGR and CPS values were greater than one for all but the largest impact scenario at both SPAs. The CGR and CPS median values were greater than one for all compensation scenarios where predicted compensation effects from a change in sandeel TSB from 300,000 tonnes to 400,000 tonnes were combined with predicted impacts. This is strongly indicative that the proposed compensatory measures will overcome the predicted impacts for puffins at both SPAs assessed. In both cases the combination of the most precautionary (i.e. largest) impact and the most precautionary estimates of compensation (i.e. smallest increase in stock biomass) generated overall increases in population growth and population size.

Table 1.32 PVA metrics (CGR & CPS) from puffin model projections of impacts from the Proposed Development alone and impacts from the Proposed Development alone minus the beneficial effects of proposed compensation metrics (based on a change in sandeel TSB from 300,000 tonnes to 400,000 tonnes). Shaded cells are larger than one

Conclusions for all Potentially impacted SPAs

211. The ability of the proposed compensatory measure, reducing or removing fishing pressure in SA4, was tested using relationships between sandeel TSB in SA4 and adult return rate (as a proxy for adult survival) or productivity. Across the range of likely changes in sandeel TSB in SA4 as a result of reducing or removing fishing pressure, positive effects on adult survival and productivity were shown. The assessment here was based on the smallest demographic changes predicted, which corresponded to the change in sandeel TSB from 300,000 to 400,000 tonnes. The positive effects of these predicted changes in demographic parameters were compared with the negative effects of three predicted impact scenarios from the Proposed Development alone. Three different approaches to this assessment were made:

• Predicted increase in number of adult birds in each SPA population and in the SPA network based on increased adult survival;
• Predicted change in populations growth rate and size due to the effects of compensation and impacts combined using PVA; and
• Predicted relationship between CRG and impact level compared with three impact prediction scenarios, using PVA.

212. For all three approaches, for all species and all SPAs, it was clear that the predicted minimum benefit from reducing or removing fishing pressure in SA4 was sufficient to compensate for all predicted impact scenarios.
213. For the FFC SPA, the predicted impacts were very small from the Project alone. A qualitative assessment concluded that the combination of spill-over effects of sandeel larval drift from SA4 to SA1r and the increased population size of kittiwake and razorbill colonies in SA4, resulting in increased emigration from those colonies to FFC SPA, would very likely more than compensate for the small predicted impacts.
     

1.10. Dealing with uncertainty, precaution and confidence

214. Key to reaching robust conclusions on the potential value of reduction or removal of fishing pressure on sandeels as a compensation measure is understanding the uncertainties in the assessment. This section addresses the uncertainties around both the predicted impacts and predicted benefits of the proposed sandeel management measures. In reaching conclusions with uncertain information, it is important to apply suitable levels of realistic precaution to the assessment as this helps to deliver confidence in the conclusions reached on the efficacy of the proposed compensation measures.

Uncertainty

215. Two key sources of uncertainty were identified in assessment of the effects of the proposed sandeel fisheries compensatory measures and their ability to address the predicted impacts: the uncertainty around the seabird data and uncertainty around the sandeel data.
216. In addition, the application of correlations between these data sources presents its own uncertainties. The effectiveness in reaching conclusions based on correlative data is therefore addressed separately.

Seabird data

217. Among the key seabird data used in assessing the proposed sandeel compensation measures, the three key seabird parameters where the uncertainty needed to be assessed were:

• Adult survival data;
• Productivity data; and
• Population size data.

Adult survival

218. In the relationships between sandeel TSB and seabird demographic parameters, adult survival has been based on the return rate of adult birds to the Isle of May breeding colonies in each year. This proxy will underestimate the apparent adult survival of birds as it does not take into account the resighting probability. From year to year there will be birds present in the colony that were present but not resighted but are then seen on the colony in subsequent years. The return rate fails to account for these birds that were present but not seen in a particular breeding season. Modelling of adult survival does account for the resighting probability to provide a better estimate of the “apparent” survival. This would likely increase the estimated value of apparent adult survival compared to the return rate values used in the correlations between sandeel TSB and survival. However, it is likely that all of the values would have been increased only slightly, as the resighting rate on the Isle of May is very high because that colony is studied in great detail by expert ornithologists, not only intensively but also following very systematic protocols. Birds that were colour ringed at the Isle of May for survival study were selected to be in locations within the colony where resighting is relatively easy. It is also likely that any increase would be similar each year because the effort put in is consistent across years. Therefore, the relationships between sandeel TSB and return rate would likely be very similar to the relationship between sandeel TSB and apparent adult survival. The modelling of apparent adult survival was not possible for this report
219. While the use of return rates as a proxy for adult survival did introduce some uncertainty in the assessment it was thought to be small, mainly due to the high resighting rate achieved by UKCEH researchers on the Isle of May. However, this uncertainty is accounted for by considering worst case scenarios rather than mean estimates of impact and gain (see below).
220. Another element of uncertainty is the assumption that survival rates of seabirds at other colonies in the region are similar to those monitored at the Isle of May. There are no data on survival rates at other nearby colonies to compare with the Isle of May time series, but it is likely that patterns correlate among colonies as colonies are exposed to the same major drivers of variation. Correlations in time series of breeding success across neighbouring colonies have been reported in several studies (e.g. Olin et al. 2020), and this is therefore likely also to apply to adult survival except where colony-specific impacts may alter that general pattern at specific individual colonies. Geographically widespread correlations between annual survival time series from different colonies have been reported for several seabird species, strengthening the case that such time series tend to respond strongly to wider ecosystem/environment drivers.

Productivity

221. There are two sources of productivity used in this report. Firstly, productivity was obtained from publicly available data provided by UKCEH from their long term study of seabirds on the Isle of May. Secondly, productivity from the JNCC SMP database was obtained for kittiwake colonies in SA4 both adjacent to and away from the sandeel box.
222. The productivity data from UKCEH is of a very high standard using consistent methods and regular fieldwork throughout the breeding season. There will still be some uncertainty in the data, which will vary between years and species. Productivity of seabirds will be influenced by multiple factors, which will include food supply, weather and predation. The data used here did not account for the effects of weather or predation on productivity. These effects will add noise to the correlation between productivity and sandeel TSB. For some species the effects of predation or weather can cause productivity to be well below the average in specific years. Despite this noise in the data there was good evidence of a non-linear relationship between productivity on the Isle of May and sandeel TSB in SA4.
223. The productivity data from the SMP database were collected from multiple colonies by a variety of people and organisations. There was variability in the number of years of data between colonies. As with the data from the Isle of May, weather and predation will add noise to these data.

Population size

224. Population size data were obtained from publicly available sources. This was a combination of data from UKCEH and from NatureScot. As with return rate and productivity data, the population size data were of a very high standard, collected by very experienced staff to standard methods each season. There will still be some uncertainty in these data, as for example counts of large colonies, or colonies that are difficult to view from vantage points will be more likely to miss some nesting birds than smaller colonies. These issues are well known and have been reviewed many times (e.g., Mitchell et al. 2004).

Sandeel data

225. Among the sandeel data used in assessing the proposed sandeel compensatory measures, the key sandeel parameter where the uncertainty needed to be assessed was the estimate of TSB.
226. Sandeel TSB data from SA4 were obtained from the most recent stock assessment report (ICES 2022). This is a modelled output based on a variety of data, including the previous years’ catch and effort data. Modelled outputs will have different uncertainties than empirical data but are often preferred as either empirical data cannot be collected (the case with sandeel TSB) or cannot be sampled sufficiently to provide data the is robust enough for the intended purpose.
227. The ICES approach is to use a Stochastic Multi-Species (SMS) assessment model. This is the best available assessment model and is considered to be a “state-of-the-art” model; it is regularly reviewed by ICES in Benchmark Working Groups and is improved when possible (ICES 2017). The model is run in single species mode using seasonal time-steps, necessary to distinguish the fishing season. The model integrates catch data, effort time series data, biological sampling from catches (e.g., age-class abundance, fish maturity, weight), survey data such as from modified dredge sampling within the sandeel box area, and estimates of natural mortality based on predatory fish diet sampling and estimates of predatory fish stock biomass. ICES coordinate a regular working group to peer review the multi-species assessment methods, which aims to enable research to improve the ecosystem approach to stock management. The Working Group on Multispecies Assessment Methods (WGSAM) continues to recommend the use of the SMS for stock size prediction. Details of the SMS methodology and uncertainties are provided in ICES 2017 and 2021b.
228. The sandeel TSB predictions from the SMS is clearly the best scientific knowledge in the field. While it does contain uncertainties that will affect the relationships between seabird demographics and sandeel TSB these are minimized through a rigorous and transparent peer review process undertaken by ICES. Two processes can particularly affect confidence in the assessment outputs. One is where a change is introduced to the assessment that alters all previous outputs in a systematic way. For example, revision of the diet composition data may alter estimated natural mortality, leading to an increase or decrease in TSB that is proportional for all years (or more disconcertingly creates a step change between two time periods). The other is the fact that all previous estimates alter when a new set of annual sampling is added to the model. This latter effect can especially alter the assessment model estimates for the most recent year or two but tends to have minimal influence on TSB estimates from earlier years. In practice, TSB estimates are revised every year when the stock assessment incorporates one further year’s new data, but the correlation between the TSB time series published in year x is very highly correlated with the TSB time series published in year x-1. Changes to the data each year can be disconcerting but are very small relative to the very large variability in TSB over decades.

Correlations

229. The key evidence used to demonstrate that sandeel fisheries management changes can be used to compensate for predicted impacts from the Proposed Development was correlative. It is hypothesized that these correlations are based on causative relationships. The nature of the relationship between seabird demography and the accessibility of prey is clear, seabirds need to feed in order to maintain themselves, female birds need to gain enough energy to produce an egg, and both the male and female birds need enough energy to be able to additionally forage to provide for chicks while they are nutritionally dependent. Where the prey availability is low this would be expected to have an effect on foraging birds, initially affecting productivity as birds would be expected forgo current productivity to maintain survival, but if forage conditions remain poor then adult survival can become affected as birds starve.
230. Walton (2008) proposed seven types of critical questions that can help to strengthen the argument from correlation to causation:

1. Is there a positive correlation between A and B?
2. Are there a significant number of instances of the positive correlation between A and B?
3. Is there good evidence that causal relationship goes from A to B, and not just from B to A?
4. Can it be ruled out that correlation between A and B is accounted for by some third factor (a common cause) that causes both A and B?
5. If there are intervening variables, then can it be shown that the causal relationship between A and B is indirect (mediated through other causes)?
6. If the correlation fails to hold outside a certain range of causes, then can the limits of this range be clearly indicated?
7. Can it be shown that the increase or change in B is not solely due to the way that B is defined, the way that entities are classified as belonging to a class of Bs, or changing standards, over time, of the way Bs are defined or classified?

231. Walton (2008) points out that it can, “always (be) suggested that there might be some other factor at work that might throw doubt on the causal relationship between A and B. As each of the above seven critical questions is adequately answered …the causal claim is strengthened.”
232. Each of these questions was therefore considered and responses provided to determine whether, by answering these questions, the correlations that underlie the evidence that reducing fishing mortality of sandeels in SA4 were sufficiently strong to be able to conclude that there will be sufficient population change to compensate for the predicted impacts from the Proposed Development (section 1.9).

Table 1.33 Responses to Walton’s (2008) critical questions.

233. Walton (2008) points out that we may have a strong suspicion that there is a causal link between two correlated parameters, as is the case with the relationships between seabird demography and sandeel TSB. By adequately answering each of the seven critical questions in Table 1.33   Open ▸ , Walton (2008) suggests that suspicions of a causal link become “more and more highly strengthened as an argument”. While it is “not easy to establish conclusively that there is a causal link between two” variables, the aim of the monitoring and adaptive management (as outlined in the Implementation and Monitoring Plan) will be to provide the demonstration of that causal link.
234. In conclusion, while this report is based on correlations between variables, each of Walton’s (2008) critical questions can be answered robustly, and therefore there are strong grounds for concluding that these correlations are highly likely to be causative.