Kittiwake breeding - Marine online assessment tool

The UK target was met at kittiwake colonies on the UK mainland coast of the North Sea, but no colonies passed the assessment in Shetland and Orkney, where the population is in steep decline. The reasons for the poor breeding success in the Northern Isles are unclear.

Background

UK target on kittiwake breeding success

This indicator on kittiwake breeding success (Figure 1) is used to assess progress against the following target, which is set in the UK Marine Strategy Part One (HM Government, 2012): “At the scale of the Marine Strategy Framework Directive, sub-regions marine bird productivity is not significantly affected by human activities: Annual breeding success of black-legged kittiwakes should not be significantly different, statistically, from levels expected under prevailing climatic conditions such as sea surface temperature.”

Black-legged kittiwake (photo from Matt Parsons)

Figure 1. Black-legged kittiwake (photo from Matt Parsons)

Key pressures and impacts

Seabirds breeding in the UK and elsewhere in the North-East Atlantic have experienced declines in abundance and frequent, widespread breeding failures over the last two decades. These declines are thought to be caused by a shortage of the small shoaling fish that seabirds feed on, which has been attributed to both warming seas and commercial sandeel fishing.

This indicator aims to distinguish between the effects of prevailing climatic conditions and those that may have resulted from human activities such as fishing. Kittiwakes are particularly susceptible to food shortages because they rely on fish being close to the surface to feed on them. Kittiwake breeding success data are therefore used in this indicator.

Measures taken to address the impacts

Since 2000, an area off North-East England and East Scotland has been closed to sandeel fishing to help with the recovery of breeding success of local kittiwake colonies. Voluntary bans on sandeel fishing in Shetland have also been in place, and no sandeels were landed there during 2004 to 2015. A seasonal ban on sandeel fishing across the entire North Sea was imposed for the first time from August to March 2016. This indicator could help to monitor the effectiveness of such measures.

stuff

Further information

Introduction

Kittiwake, like other marine bird species that feed at the surface of the sea were found to have experienced frequent, widespread breeding failure, and therefore failed the UK and OSPAR assessments of breeding abundance (HM Government, 2012; OSPAR Commission, 2017). This is in contrast to water-column feeding species who also feed on fish but can dive below the surface to catch prey at a variety of depths. These results suggest that it is not a sheer abundance of prey that is affecting their productivity and breeding numbers, but the availability of the prey to predators confined to the surface. Indeed, there is only weak evidence linking prey abundance to kittiwake breeding success (ICES, 2015). The poor breeding success of kittiwakes has been related to increases in sea surface temperature (Frederiksen, 2004; 2007; Cook and others 2014a) and stratification within the water column (Carroll and others, 2017), suggesting these environmental factors are important in determining the abundance or availability of sandeels to kittiwakes and other surface-feeders. There is also evidence to suggest an additional negative impact of industrial sandeel fishing on the breeding success of kittiwakes at nearby colonies (Frederiksen, 2004; 2008; Cook, and others 2014a; Carroll and others, 2017). This indicator aims to distinguish between the effects of prevailing climatic conditions from those that may have resulted from human activities, such as fishing.

Justification for the indicator

This indicator was developed by Cook and others (2014b) and refines and improves on a previously proposed draft OSPAR Ecological Quality Objective on local sandeel availability to Black-legged kittiwakes. In the North Sea, the lesser sandeel (Ammodytes marinus) is a key prey species for kittiwakes and other fish-eating seabirds, such as shag, guillemot, razorbill, puffin and terns. The Ecological Quality Objective was not adopted because it failed to distinguish between changes in sandeel availability that were an impact of fishing from those that resulted from variation in prevailing climatic conditions. The indicator developed by Cook and others (2014b) considers the changes in kittiwake breeding success that are related to changes in sea-surface temperature.

The indicator is based on work by Frederiksen and others (2004; 2005; 2007), which found that kittiwake breeding success at seven colonies along the North Sea coast of the UK was significantly negatively correlated with local mean sea-surface temperatures during February and March of the year preceding each breeding season. The relationship was thought to be related to larval sandeel survival and their subsequent availability, impacting kittiwakes in the rearing of their chicks. Frederiksen and others (2004) proposed that warmer winters resulted in lower recruitment of sandeels to the current year cohort (referred to as group 0), subsequently leading to reduced availability of sandeels in the following year (referred to as group 1). Kittiwake breeding success is not affected immediately following a warm winter because neither adults nor chicks feed on Group 0 sandeels until June. However, birds are affected in the following year when they feed on the same cohort of sandeels (now classed as Group 1) at the start of the breeding season, which leads to poorer breeding conditions and lower breeding success. A recent study at the Bempton Cliffs colony in North East England found sea surface temperature to be negatively related to sandeel biomass, as well as to kittiwake breeding success (Carroll and others, 2017).

In an area of the North Sea, off the eastern UK, where sandeel fishing occurred between 1991 and 1998 but has been banned since 2000, there was a significant negative effect of the presence of fishing (Frederiksen and others, 2004; 2008). An analysis of this kittiwake indicator in UK North Sea waters (Cook and others 2014a) provides further evidence of links to fishing pressure. They found a significant effect on kittiwake breeding success of a fishing pressure factor denoted by the interaction between the annual North Sea stock size of sandeels and the proportion of the stock that was harvested had a significant effect on kittiwake breeding success. Carroll and others (2017) further found an additional negative effect of increased sandeel fishing mortality on kittiwake breeding success, even when sea surface temperature had been taken into account. These studies suggest that kittiwake breeding success could be used as an indicator of fishing impacts on the availability of prey to kittiwakes and other sandeel reliant seabirds.

Population context

The last census of kittiwakes breeding in the UK, conducted during 1998 to 2002 recorded a total of 380,000 pairs, or 8% of the global population (Mitchell and others 2004; Figure 2). This total was 25 % less than the number previously counted in the mid-1980s. Since 2000, numbers have continued to fall. Table 1 shows changes in abundance at the colonies included in this kittiwake breeding success indicator. Declines in Shetland have been as much as 90% even at previously substantial colonies as Noss and Fair Isle. There have also been similarly steep declines on Orkney Island, including a decline at the large colony in the Northern Isles at West Westray from 33,000 pairs in 1999 to 12,000 pairs in 2007. Declines on the mainland North Sea coast have been lower than in the Northern Isles but still substantial in places (see Table 1).

Figure 2. Abundance and distribution of kittiwake in the UK and Ireland in 1998 to 2002 (Source: Mitchell and others, 2004. Provided under copyright by P. I. Mitchell).

Table 1: Changes in the number of breeding kittiwake pairs at the colonies included in this indicator of kittiwake breeding success. The year during which populations were counted are given as superscripts to the count number. Source: JNCC, 2016, Seabird Monitoring Programme Database.

	Colony Name	Seabird 2000 Census count ^(year)	Most recent Count ^(Year)	Change (%)	Used to assess period 2010-2015
Shetland	Hermaness NNR	643 ¹⁹⁹⁹	304 ²⁰⁰⁹	-53%	Yes
Shetland	Noss	2,395 ²⁰⁰⁰	179 ²⁰¹⁵	-93%	Yes
Shetland	Foula	1,934 ²⁰⁰⁰	277 ²⁰¹⁵	-86%	Yes
Shetland	Noness (aka. Ramna Geo to Skerry of Rest)	633¹⁹⁹⁹	139²⁰¹⁵	-78%	Yes
Shetland	Whale Wick to Sandwick	139¹⁹⁹⁹	30²⁰¹⁵	-78%	Yes
Shetland	Westerwick	76¹⁹⁹⁹	12²⁰⁰³	-84%	No
Shetland	Cross-Voe-Sand to Caves (aka Eshaness)	11¹⁹⁹⁹	N/A	-100%	No
Shetland	Sumburgh Head	877 ²⁰⁰¹	362 ²⁰¹⁴	-59%	No
Shetland	Fair Isle	8,204 ²⁰⁰¹	859 ²⁰¹⁵	-90%	Yes
Orkney	Marwick Head	5,573 ¹⁹⁹⁹	526 ²⁰¹³	-91%	Yes
Orkney	Gultak	249 ²⁰⁰⁰	34 ²⁰¹⁵	-86%	Yes
Orkney	Mull Head	559 ²⁰⁰⁰	0 ²⁰¹⁵	-100%	No
Orkney	Row Head (aka HY2218)	1565 ¹⁹⁹⁹	15 ²⁰¹⁵	-99%	Yes
Orkney	Costa Head	1256 ²⁰⁰⁰	298 ²⁰¹⁵	-76%	Yes
Orkney	Papa Westray - North Hill RSPB	149²⁰⁰⁰	18²⁰¹²	-88%	Yes
Mainland	North Sutor of Cromarty	447²⁰⁰⁰	275²⁰¹⁵	-38%	Yes
Mainland	Buchan Ness to Collieston Coast SPA	14,091 ²⁰⁰¹	12,542 ²⁰⁰⁷	-11%	No
Mainland	Sands of Forvie	601²⁰⁰⁰	496²⁰¹⁵	-17%	No
Mainland	Fowlsheugh	18,800 ¹⁹⁹⁹	9,655 ²⁰¹⁵	-49%	Yes
Mainland	Isle of May	4618²⁰⁰⁰	3433²⁰¹⁵	-26%	Yes
Mainland	St Abb's Head NNR	11,077 ²⁰⁰⁰	4,209 ²⁰¹⁵	-62%	Yes
Mainland	Dunbar Coast and Harbour	1191²⁰⁰⁰	1155²⁰⁰⁷	-3%	Yes
Mainland	Farne Islands	5,096 ¹⁹⁹⁸	3,956 ²⁰¹⁵	-22%	Yes
Mainland	Coquet Island	80²⁰⁰⁰	326²⁰¹⁵	308%	Yes
Mainland	River Tyne (all colonies)	351²⁰⁰⁰	622²⁰¹⁵	77%	Yes
Mainland	Saltburn Cliffs (huntcliff)	3375¹⁹⁹⁹	1330²⁰¹⁵	-61%	No
Mainland	Flamborough Head and Bempton Cliffs	42,692 ²⁰⁰⁰	37,617 ²⁰⁰⁸	-12%	Yes
Mainland	Lowestoft	150²⁰⁰⁰	196²⁰¹⁵	31%	Yes
Mainland	St Aldhelms Head - Durlston Head	49²⁰⁰⁰	23²⁰¹⁵	-53%	No

Assessment method

Data used in the assessment

This indicator uses a known relationship between estimates of annual breeding success of kittiwakes and the local mean sea-surface temperature in February and March of the previous year (see Frederiksen, 2004; 2007; Cook and others, 2014a) to predict what annual breeding success should be if it is in line with prevailing climatic conditions, as stated in the UK target (HM Government, 2102). A mathematical model describing this relationship was established for each kittiwake colony included in the assessment. These models were used to predict annual baseline values of breeding success, for example those expected under prevailing climatic conditions. If the observed values for kittiwake breeding success were significantly lower than the baselines, this might indicate other impacts, possibly anthropogenic.

Kittiwake data

This indicator was constructed from a time-series of annual estimates of breeding success at a sample of 29 colonies along the UK North Sea coast during 1986 to 2015. The Tyne colony is an aggregation of data from five colonies nesting on human-made structures in Newcastle and along the River Tyne, including the Tyne Bridge and the Baltic art gallery.

Breeding success was measured by the mean number of chicks fledged per pair or nest at a colony in any given year. Data were collected as part of the Seabird Monitoring Programme of the UK and Ireland and extracted from the Seabird Monitoring Programme database Not all colonies in the samples were observed every year in the times-series. Missing annual observations were predicted by statistical models. Initial runs of these models indicated that for a power of 0.8 and a significance level of 0.05 (see below for details), a minimum of 14 years data are required to detect a significant relationship between kittiwake breeding success and sea surface temperature. Twenty-nine colonies in the North Sea had 14 or more years of data.

Data from 13 colonies in the Celtic Seas were also analysed, including one colony in the Isle of Man (Glen Maye to Peel) and four colonies in Ireland (Ram Head, Rockabill, Dunmore east to Red Head, and Portally). However, a significant relationship between sea surface temperature and breeding success was found in only four of these. There was also no overall significant correlation between kittiwake breeding success and sea surface temperature in the Celtic Seas, which is consistent with previous studies (Lauria and others, 2012; Cook and others, 2014a).

Sea surface temperature data

Data were obtained from the Hadley Centre Sea Ice and Sea Surface Temperature dataset (Rayner and others, 2003) and processed using the ncdf package within R (Pierce, 2011). Annual mean winter temperatures (during February and March) were calculated for the adjacent sea areas. These areas were chosen to coincide with sandeel spawning grounds (Ellis and others, 2012). Each colony was paired with temperature data from the adjacent sea area which showed the strongest relationship between breeding success and sea surface temperature.

Baseline and Regression Model

The baseline for this indicator is different for each colony and varies between years. The baseline is the annual mean breeding success at a colony in any given year as predicted by the annual mean winter sea surface temperature (measured during February and March) of the preceding year (Figure 3). There is a strong relationship between breeding success and sea surface temperature at the North Sea colonies. This relationship was used to construct a baseline, which predicts what annual breeding success should be if it is in line with prevailing climatic conditions. If breeding success is significantly lower than the baseline, it is considered not to be in line with ‘prevailing climatic conditions’. The UK target was met if, at a significant proportion of kittiwake colonies, the threshold was exceeded in five years out of six between 2010 and 2015.

Figure 3. A stylised version of the relationship between kittiwake breeding success and sea-surface temperature during February and March of the year preceding each breeding season (SST^-1). The light grey area indicates values of breeding success that is in line with prevailing climatic conditions and the dark grey area indicates values that are not in line.

A General Linear Model was run on the breeding success data from all colonies in both sub-regions and the annual mean winter sea surface temperature data from the corresponding adjacent sea areas. The following R code was used, in which GNS refers to Greater North Sea:

KI.GNS.mm = glmer (cbind(Chicks_fledged,Plot_size*2) ~ SST

+(1|ColonyName)+(0+SST|ColonyName), weights=Plot_size, family = binomial, data=KI.GNS)

Before modelling, the values of chicks fledged per nest were divided by two (the normal clutch size of kittiwakes) to estimate the number of young fledged per egg. This was necessary because the number of chicks fledged per nest (an integer number, either 0, 1 or 2) did not follow a standard distribution, for example, a Poisson distribution would include a substantial frequency > 2, which is impossible for a bird that lays only two eggs. The estimates of chicks fledged per egg had a binomial distribution, for example each egg will either produce a fledged chick or not, which could easily be incorporated into the model. As colonies differed greatly in size, weighting was introduced in both models to account for a different number of nests found in each colony.

There was a strong significant negative relationship between breeding success and mean winter sea surface temperature across the 29 North Sea colonies (slope = -0.56, P < 0.0001; Table 2). But there was no significant relationship for the 13 Celtic Seas colonies (Table 2).

Table 2. Parameters of General Linear Models incorporating kittiwake breeding success and mean winter sea surface temperature (SST^-1) for 29 North Sea colonies and 13 Celtic Seas colonies.

	Effect	P
Greater North Sea
Intercept	3.73 (±0.49)	<0.0001
SST-1	-0.56 (± 0.13)	<0.0001
Celtic Seas
Intercept	-0.54 (± 1.33)	0.6820
SST-1	-0.16 (± 0.23)	0.4640

The mean effect size for the Greater North Sea (Table 2) was used within a power analysis framework in the R statistical package pwr (Champely, 2012) to calculate the minimum number of years of observations that were required in the time-series to detect a relationship between kittiwake breeding success and mean winter sea surface temperature that was significant at the level of 0.05 and had a power of 0.8. Based on this initial analysis, colonies with a minimum of 14-years data were selected for further analysis and inclusion in the kittiwake breeding success indicator. A total of 29 colonies along the North Sea coast were found to have sufficient data to detect a significant relationship.

The annual baseline values for each colony North Sea colony were predicted using the Greater North Sea model, but with colony-specific slopes and intercepts. The relationship between breeding success and mean winter sea surface temperature for each colony are shown in the plots in Figure 4. The slope or strength of the relationship varies between colonies, with stronger negative relationships occurring further north (Figure 5).

Figure 4. Farne Island example of the relationship between colony-specific kittiwake annual mean breeding success and winter mean sea surface temperature. The red line denotes the baseline and the black line is the lower 95% confidence limit of the relationship and is used to calculate the threshold values in this assessment.

Figure 5. Locations of kittiwake colonies for which sufficient data were available. Plots indicate the strength of the effect of mean winter sea surface temperature on breeding success, with Celtic Seas colonies in the left-hand plot and Greater North Sea colonies in the right-hand plot. Colonies are plotted according to latitude, aligned to the map. Grey boxes show sea areas in which annual mean winter sea surface temperature was calculated from the Hadley Centre Sea Ice and Sea Surface Temperature HadISST dataset: from top clockwise - Shetland/Orkney, East Scotland & Wee Bankie, Dogger Bank, English Channel, Bristol Channel, Liverpool Bay/Irish sea and West Scotland.

Assessment of UK target on kittiwake breeding success

The schematic in Figure 6 shows how the UK target for kittiwake breeding success is assessed.

Figure 6. Schematic showing how the UK target for kittiwake breeding success was assessed. MSFD = Marine Strategy Framework Directive.

Determining the indicator threshold for each colony

To assess the UK target for kittiwake breeding success, the breeding success at each colony was first assessed against the following supporting target: “Annual breeding success is not significantly lower statistically, from the level expected in the prevailing climatic conditions in five years out of six.” (HM Government, 2012)

If breeding success is significantly lower than the baseline, it is considered not to be in line with ‘prevailing climatic conditions’ (see Figure 3). To determine if breeding success was significantly lower, a threshold was set for each year at each colony. This threshold was the minimum detectable decrease in breeding success from what was predicted from the baseline relationship with sea-surface temperature (see Figure 3). The minimum decrease in breeding success is what would be detected with a statistical significance of p=0.05 and with a power of 0.8 such that if the assessment were repeated, you would expect to detect a significant difference in at least 80% of the repeat assessments. This was determined using the pwr.t.test function in the R package pwr (Champely 2017). If breeding success at a colony each year was below the threshold, the colony was considered to be not achieving the level of breeding success expected given the prevailing climatic conditions. A colony was considered to have failed to achieve the supporting target if breeding success was below the threshold in more than one year between 2010 to 2015 (inclusive). This target of five years out of six was designed to allow for failures resulting from stochastic events, such as very high rainfall which can lead to hypothermia and death of chicks.

Determining the number of colonies required to be ‘above threshold’

Out of the total sample of 29 colonies, 22 had data for every year during 2010 to 2015. To achieve the UK target all colonies would be expected to achieve the supporting target or be ‘above threshold,’ during 2010 to 2015 except for a small number that statistically, would be expected to fail by chance. The number of colonies expected to fail by chance was calculated using binomial probability theory (see example in Greenstreet and others, 2012 for a similar application). Given that the threshold for each colony in each year was set at a value that would be attained with a probability of 0.05, the probability of annual breeding success being below the threshold by chance, would also be 0.05 (or 5 %). With p=0.05, the function dbinom in R 3.4.0 (R Core Team, 2017) was used to estimate the cumulative probability that breeding success at a colony would be below the threshold in two, or more, years out of six by chance. This probability was estimated at 0.03 (or 3%).

With p=0.03, the dbinom function in R (R Core Team, 2017) was used to estimate the number of colonies that would be expected to fail to meet the supporting target by chance, with a probability of greater than 0.05. Figure 7 shows that up to three colonies out of 22 would be expected to fail by chance. Therefore, to achieve the UK target, breeding success of kittiwakes at 19 or more colonies of the 22 assessed would need to be in line with prevailing climatic conditions in five years out of six during 2010 to 2015.

Figure 7. The cumulative probability of the number of Kittiwake breeding colonies (out of 22) failing, by chance, to achieve the target that breeding success should be in line with the prevailing environmental conditions in at least 5 out of the preceding six years.

Retrospective assessment of the UK target

There were sufficient data to assess the UK target during each six-year period from 1986 to 1991 up to the most recent period of 2010 to 2015. To be included in the assessment for each period, a colony needed six consecutive annual estimates of breeding success. Therefore, the number of colonies included in the assessment varied across the assessment periods. A minimum of 17 colonies were assessed during 1987 to 1992 and a maximum of 28 colonies in 1998 to 2003 and 1999 to 2004. Statistically, up to three colonies would be expected to fail to meet their colony targets by chance with a sample size of 17 to 28 colonies. Therefore, the number of colonies during each six-year period that would be required to meet their colony targets varied depending on the number of colonies assessed and ranged from 14 to 24. The number of colonies in each regional assessment varied between nine to 13 and 6 to 12 for colonies in the Northern Isles (Orkney and Shetland) and the UK mainland coast bordering the North Sea, respectively. The number of colonies that would be expected to fail to meet their colony targets by chance was reduced to two when the number of colonies assessed was between seven and 16 and reduced to one when only six colonies were assessed.

Results

Findings from the 2012 UK Initial Assessment

This indicator was not considered as part of the 2012 Initial Assessment (HM Government, 2012).

Latest findings

Status assessment

During 2010 to 2015, 22 colonies were assessed (Figures 8 and 9). To meet the UK target 19 kittiwake colonies needed to breed successfully in at least five years out of six. Unfortunately, only nine of the 22 colonies followed this pattern, so overall the UK target has not been met. Nine out of ten colonies assessed on the British mainland did meet the UK target in 2010 to 2015 with only the St Abbs colony failing. All 12 colonies assessed in Orkney and Shetland failed to achieve the UK target.

Figure 8. Proportion of years in which the target breeding success was met in each colony during the most recent 6-year assessment period c) 2010 to 2015, compared to examples of other 6-year periods: a) 1986 to 1991 during the Shetland sandeel collapse and b) 1994 to 1999 when the Wee Bankie sandeel fishery was in operation. Pie charts: Grey denotes no breeding success values, green denotes breeding success was as expected by prevailing climatic conditions and red denotes that it was not. Red line denotes limits of sandeel fishing ban in place since the year 2000.

Figure 9. The proportion of assessed colonies where breeding success was in line with conditions (“pass”) and not in line with “fail”) prevailing climatic conditions in at least 5 of the previous six years. Each annual value represents an assessment for the period of the preceding six years, for example 1991 = 1986 to 1991 inclusive. The red line is the threshold for the number of colonies ‘passing’, at which the UK target is met.

Trend assessment

The assessment for the period 2010 to 2015 above was repeated for each six-year period from 1986 onwards (Figure 9). Colonies on the British mainland are ‘improving,’ and colonies in Orkney and Shetland have been ‘declining.’ During the late 1980s, failure to meet the UK target was largely confined to Shetland colonies (see Figure 9a, assessment period 1986 to 2001). This was mostly likely due to a crash in the Shetland sandeel stock at that time which was caused by a change in currents and independent of sea-surface temperature. Breeding success at colonies in Shetland subsequently recovered temporarily during the early 1990s. During the 1990s, failure to meet the UK target was mainly restricted to colonies in eastern mainland Scotland that were adjacent to an area of high sandeel fishing pressure, which was subsequently closed to fishing from the year 2000 onwards (Figure 9b, assessment period 1994 to 1999). During the mid-2000s, failure to reach the UK target was widespread (Figure 9). Breeding success at UK colonies subsequently improved to be in line with prevailing conditions (Figure 9b). In contrast, all colonies in Orkney and Shetland, have not met the UK target in any six-year period since 2001 to 2006, except at one Shetland colony between 2004 and 2009 (Figure 9c).

Further information

An example of colony-specific results is provided in Figure 10.

Farne Islands Colony example of colony specific annual mean breeding success. The dashed line denotes the lower 95 % confidence limit of breeding success if it was in line with prevailing climatic conditions.

Figure 10. Farne Islands Colony example of colony specific annual mean breeding success. The dashed line denotes the lower 95 % confidence limit of breeding success if it was in line with prevailing climatic conditions.

Conclusions

Knowledge gaps

This indicator could be used with more confidence to inform management through improving understanding of:

the processes within the marine ecosystem which lead to the significant relationship between kittiwake breeding success and sea-surface temperature
the causes of the lower than expected breeding success of kittiwakes in Orkney and Shetland.

Further information

Filling the two knowledge gaps detailed below will enable this indicator to be used with more confidence in a management framework.

Understanding the ecosystem processes driving the relationship between kittiwake breeding success and sea surface temperature

This indicator uses a statistical relationship between mean winter sea surface temperature and kittiwake breeding success to account for prevailing climatic conditions. Carroll and others (2015) found that higher breeding success at some of the colonies included in this indicator was also associated with weaker stratification before breeding and lower mean water sea surface temperature during the breeding season. Frederiksen and others (2004) first suggested that the link between temperature and kittiwake breeding success was a result of processes further down the food chain that affected the abundance and/or availability of sandeels, on which kittiwakes feed. Recently, Carroll and others (2017) found sea surface temperature to be negatively related to sandeel biomass, as well as to kittiwake breeding success. However, the processes which lead to this correlation are not understood. Arnott & Ruxton (1992) suggested warm winters lead to low sandeel recruitment, perhaps by reducing the food supplies available to larval sandeels. But recently, Eerkes-Medrano and others (2017) found that winter sea temperature in an area of the North Sea, east of the Firth of Forth was not a reliable indicator of the abundance of sandeels or of their prey (Calanus copepods in the zooplankton). Further work is needed to determine the nature of a causal link between sea temperature and the abundance of sandeels. This will help to determine whether kittiwake breeding success can be used to indicate changes at lower trophic levels.

Understanding the causes of low breeding in Orkney and Shetland

It is unclear why kittiwake breeding success has been poorer than expected across all colonies in Orkney and Shetland since the early 2000s. It is unlikely to have been due to fishing pressure since this was low or even absent due to voluntary bans during this period (ICES, 2017). In some years, extreme weather events such as heavy rain that washes nests from cliffs may also have lowered breeding success, but such events are unlikely to have caused such widespread and sustained reductions in breeding success.

At some colonies, particularly on Shetland, predation from great skuas is likely to be a major cause of poor breeding success and declines in colony size (Furness, 1997; Heubeck and others, 1999). During the 1990s depredation of seabirds by skuas increased as other food sources declined, including discards and natural prey such as sandeels (Votier and others, 2004). Levels of high predation by great skuas are likely to continue as discards are eventually eliminated in UK waters by EC Landings Obligations.

In the late 1980s, changes in ocean currents led to low levels of sandeels in waters around Shetland and therefore, the poor breeding success of kittiwakes and other seabirds at Shetland colonies (Hamer and others 1993; Monaghan and others, 1989; Wright and Bailey, 1993). The reductions in kittiwake breeding success during the late 1980s were evident in this indicator. At such times when sandeel availability is low, kittiwake may struggle to find alternative food source, such as sprat, which are relatively scarce around the Northern Isles (Macdonald & Napier, 2014 ).

Further examination of available evidence on the above impacts is required to determine their relative importance at each of the colonies included in this assessment in Orkney and Shetland. This will help to identify any possible management that may help to enhance breeding success of kittiwakes in the Northern Isles.

This indicator could not be developed for the Celtic Seas because no relationship between sea surface temperature and kittiwake breeding success could be found by this and other studies (Cook and others, 2014b, and Lauria and others, 2012). This is likely because sea surface temperature is not the primary driver of food availability on the west coast of the UK as a result of the convoluted tidal currents in coastal areas. Kittiwakes in the Celtic Seas are also more reliant on other species of small fish, such as sprat and herring, that are differently affected by sea surface temperature compared to sandeels. It is currently unknown what the primary drivers of change in breeding success of kittiwakes are on the west coast of the UK, or whether these operate via changes in food availability. A multi-trophic level, region-wide approach would most likely aid our understanding of the ecological processes regulating marine food webs in response to climate change.

Understanding the ecosystem processes driving kittiwake breeding success in the Celtic Seas.

References

Arnott SA, Ruxton GD (2002) ‘Sandeel recruitment in the North Sea: demographic, climatic and trophic effects’ Marine Ecology Progress Series, 238, 199–210 (viewed on 24 October 2018)

Carroll MJ, Butler A, Owen E, Ewing SR, Cole TJ, Green A, Soanes LM, Arnould JPY, Newton SF, Baer J, Daunt F, Wanless S, Newell MA, Robertson GS, Mavor RA, Bolton M (2015) ‘Effects of sea temperature and stratification changes on seabird breeding success’ Climate Research 66: 75–89 (viewed on 24 October 2018)

Carroll MJ, Bolton M, Owen E, Anderson GQA, Mackley EK, Dunn EK, Furness RW (2017) ‘Kittiwake breeding success in the southern North Sea correlates with prior sandeel fishing mortality’ Aquatic Conservation: Marine Freshwater Ecosystems 27: 1164-1175 (viewed on 24 October 2018)

Champely S (2017) ‘pwr: Basic Functions for Power Analysis’ R package version 1.2-1 (viewed on 24 October 2018)

Cook ASCP, Dadam D, Mitchell I,Ross-Smith VH, Robinson RA (2014a) ‘Indicators of seabird reproductive performance demonstrate the impact of commercial fisheries on seabird populations in the North Sea’ Ecological Indicators 38: 1–11 (viewed on 23 October 2018)

Cook ASCP, Robinson RA, Ross-Smith VH (2014b) ‘Development of MSFD Indicators, Baselines and Target for Seabird Breeding Failure Occurrence in the UK (2012)’ JNCC Report 539, ISSN 0963 8901 (viewed on 23 October 2018)

Ellis JR, Milligan SP, Readdy L, Taylor N, Brown MJ (2012) ‘Spawning and nursery grounds of selected fish species in UK waters’ Science Series Technical Report Number 147, Cefas Lowestoft: 56 pages (viewed on 24 October 2018)

Erekes-Medrano D, Fryer RJ, Cook KB, Wright PJ (2017) ‘Are simple environmental indicators of food web dynamics reliable: Exploring the kittiwake-temperature relationship’ Ecological Indicators 75: 36-47 (viewed on 24 October 2018)

Frederiksen M, Wanless S, Harris MP, Rothery P, Wilson LJ (2004) ‘The role of industrial fisheries and oceanographic change in the decline of North Sea black-legged kittiwakes’ Journal of Applied Ecology 41: 1129–1139 (viewed on 24 October 2018)

Frederiksen M, Wright PJ, Harris MP, Mavor RA, Heubeck M, Wanless S (2005) Regional patterns of kittiwake Rissa tridactyla breeding success are related to variability in sandeel recruitment’ Marine Ecology Progress Series 300: 201–211 (viewed on 24 October 2018)

Frederiksen M, Mavor RA, Wanless S (2007) ‘Seabirds as environmental indicators: the advantages of combining data sets’ Marine Ecology Progress Series 352: 205–211 (viewed on 24 October 2018)

Frederiksen M, Jensen H, Daunt F, Mavor RA, Wanless S (2008) ‘Differential effects of a local industrial sand lance fishery on seabird breeding performance’. Ecological Applications 18: 701–710. (viewed on 24 October 2018)

Furness, R. W. 1997. The impact of predation by great skuas on other species with particular reference to Special Protection Areas in Shetland. Report to Scottish Natural Heritage.

Greenstreet SPR, Rossberg AG, FoxCJ, Le Quesne WJF, Blasdale T, Boulcott P, Mitchell I, Millar C, Moffat CF (2012) ‘Demersal fish biodiversity: species-level indicators and trends-based targets for the Marine Strategy Framework Directive’ ICES Journal of Marine Science 69(10): 1789–1801 (viewed on 24 October 2018)

Hamer KC, Monaghan P, Uttley JD, Walton P, Burns MD (1993) ‘The influence of food supply on the breeding ecology of Kittiwakes Rissa tridactyla in Shetland’ International Journal of Avian Science 135: 255-263 (viewed on 24 October)

Heubeck M, Mellor RM, Harvey PV, Mainwood AR, Riddington R (1999) ‘Estimating the population size and rate of decline of Kittiwakes Rissa tridactlya breeding in Shetland, 1981-97’ Bird Study 46: 48-61 (viewed on October 24, 2018)

ICES (2015) ‘Report of the Joint ICES/OSPAR Working Group on Seabirds (JWGBIRD)’, 17–21 November 2014, Copenhagen, Denmark. ICES CM 2014/ACOM:30. 115 pages (viewed on 22 October 2018)

ICES (2017) ‘Sandeel (Ammodytes spp.) in Division 4.a, Sandeel Area 7r (northern North Sea, Shetland)’ ICES Advice on fishing opportunities, catch, and effort, Greater North Sea Ecoregion. Code: san.sa.7r (viewed on 24 October 2018)

JNCC (2016) ‘Seabird Population Trends and Causes of Change: 1986-2015 Report’ Updated September 2016 (viewed on 24 October 2018)

Lauria V, Attrill MJ, Pinnegar JK, Brown A, Edwards M, Votier SC (2014) ‘Influence of Climate Change and Trophic Coupling across Four Trophic Levels in the Celtic Sea’ PLoS ONE 7(10): e47408. https://doi.org/10.1371/journal.pone.0047408 (viewed on 24 October 2018)

Macdonald P, Napier I (2014) ‘Survey trends of fish species in Shetland coastal waters’ NAFC Marine Centre, Shetland.

Mitchell P I, Newton S F, Ratcliffe N, Dunn T E (Editors) (2004) ‘Seabird Populations of Britain and Ireland: results of the Seabird 2000 census (1998-2002)’ Published by T & AD Poyser, London. (viewed on 24 October 2018)

Monhaghan P, Uttley JD, Okil, JD (1989) ‘Terns and sandeels: seabirds as indicators of changes in marine fish populations’ Journal of Fish Biology 35:339-40 (viewed on 24 October 2018)

Pierce D (2011) ‘ncdf: Interface to Unidata netCDF data files’ R package version 1.6.6 (viewed on 24 October 2018)

R Core Team (2017) ‘R: A Language and Environment for Statistical Computing’ R Foundation for Statistical Computing. Vienna, Austria (viewed on 24 October, 2018)

Rayner NA, Parker DE, Horton EB, Folland CK, Alexander LV, Rowell, DP, Kent EC, Kaplan A (2003) ‘Global analyses of SST, sea ice and night marine air temperature since the late nineteenth century’ Journal of Geophysical Research: Atmospheres, 108:4407 (viewed on 24 October 2018)

Votier SC, Furness RW, Bearhop S, Crane JE, Caldow RWG, Catry P, Ensor K, Hamer KC, Hudson AV, Kalmbach E, Klomp NI, Pfeiffer S, Phillips RA, Prieto I, Thompson DR (2004) ‘Changes in fisheries discard rates and seabird communities’ Nature, 427: 727-730 (viewed on October 24 2018)

Wright PJ, Bailey MC (1993) ‘Biology of sandeels in the vicinity of seabird colonies at Shetland’ Fisheries Research Report No. 15/93. SOAFD Marine Laboratory, Aberdeen

Acknowledgements

Assessment metadata

Assessment Type	UK Marine Strategy Framework Directive Indicator Assessment
	D1 - Biological Diversity D6 – Seafloor Integrity
	Condition of soft sediment invertebrate communities in coastal waters determined using Water Framework Directive methods
	Please see links provided in ‘References’ section above.
Point of contact email	marinestrategy@defra.gov.uk
Title	Kittiwake breeding
Resource abstract	Kittiwake breeding success data was extracted from the UK Seabird Monitoring Programme Database. Data consists of annual counts of black-legged kittiwake (Rissa tridactyla) chicks fledged at UK North Sea colonies during 1986-2015.
Linkage	Please see links provided in ‘References’ section above.
Conditions applying to access and use	© Crown copyright, licenced under the Open Government Licence (OGL).
Assessment Lineage	This indicator is constructed from data on the annual breeding success of kittiwakes (number chicks fledged per pair) collected at 29 colonies along the North Sea coast in at least 14 years during the period 1986-2015. For each colony, annual breeding success during 1986-2015 was plotted against sea-surface temperature. There is a strong relationship between breeding success and sea surface temperature at the North Sea colonies. This relationship was used to construct a baseline, which predicts what annual breeding success should be if it is ‘in line with prevailing climatic conditions’.
Indicator assessment results
Dataset metadata
Dataset DOI	Contact marinestrategy@defra.gov.uk

The Metadata are “data about the content, quality, condition, and other characteristics of data” (FGDC Content Standard for Digital Geospatial Metadata Workbook, Ver 2.0, May 1, 2000).

Metadata definitions

Assessment Lineage - description of data sets and method used to obtain the results of the assessment

Dataset – The datasets included in the assessment should be accessible, and reflect the exact copies or versions of the data used in the assessment. This means that if extracts from existing data were modified, filtered, or otherwise altered, then the modified data should be separately accessible, and described by metadata (acknowledging the originators of the raw data).

Dataset metadata – information on the data sources and characteristics of data sets used in the assessment (MEDIN and INSPIRE compliance).

Digital Object Identifier (DOI) – a persistent identifier to provide a link to a dataset (or other resource) on digital networks. Please note that persistent identifiers can be created/minted, even if a dataset is not directly available online.

Indicator assessment metadata – data and information about the content, quality, condition, and other characteristics of an indicator assessment.

MEDIN discovery metadata - a list of standardized information that accompanies a marine dataset and allows other people to find out what the dataset contains, where it was collected and how they can get hold of it.

Recommended reference for this indicator assessment

Ian Mitchell¹, Aonghais Cook⁶, Andrew Douse,² Simon Foster,² Melanie Kershaw,³ Neil McCulloch⁴, Matty Murphy⁵, & Jane Hawkridge¹ 2018. Kittiwake breeding success. UK Marine Online Assessment Tool, available at: /biodiversity-food-webs-and-marine-protected-areas/birds/kittiwake-breeding-success/

¹Joint Nature Conservation Committee

²Scottish Natural Heritage

³Natural England

⁴Department of Environment, Agriculture & Rural Affairs, Northern Ireland

⁵Natural Resources Wales

⁶British Trust for Ornithology