Bird Study (1999) 46, 48–61

Estimating the population size and rate of decline of Kittiwakes Rissa tridactyla breeding in , 1981–97

MARTIN HEUBECK1*, R. MICHAEL MELLOR2, PAUL V. HARVEY3, ANTHONY R. MAINWOOD4 and ROGER RIDDINGTON5 1Aberdeen University Research and Industrial Services Ltd, c/o Sumburgh Lighthouse, Virkie, Shetland ZE3 9JN, UK, 2Fairview, Quendale, Shetland ZE2 9JB, UK, 3Scottish Natural Heritage, Stewart Building, The Esplanade, Lerwick, Shetland ZE1 OLL, UK, 413 Ben Bhraggie Drive, Golspie, Sutherland KW10 6SX, UK, 5Fair Isle Bird Observatory, Fair Isle, Shetland ZE2 9JU, UK

In the early 1980s, Shetland’s Kittiwake population totalled 54 000 pairs in 55 breeding stations of greatly varying size. Since 1985, periodic counts of nests have been made at all stations, mostly from an inflatable boat and in the same manner as a comprehensive survey in 1981. Different stations were surveyed in different years and at varying intervals. Changes in numbers at individual stations varied in size and direction. To estimate the total breeding population, and variation in the overall rate of change between 1981 and 1997, ‘missing’ nest counts at each station were calculated from mean annual rates of change between actual counts; the overall estimated population was expressed as the annual sum of actual and calculated nest counts. Shetland’s Kittiwake population declined by 50% during 1981–97. Although the decline was progressive, the rate increased markedly during 1989–93, then decreased between 1993–97. The main causes of the decline are considered to have been successive years of low breeding success, due to low availability of sandeels, increased rates of predation of nests by Great Skuas and increased predation of adults by skuas.

n 1981, the first comprehensive survey of repeated counts at fixed study plots, it was Inesting Kittiwakes Rissa tridactyla in the more appropriate to survey all known colonies Shetland Islands found 54 300 pairs. These (and suitable habitat where new ones might were dispersed among colonies of greatly form) at least once every three to five years.2 varying size, with 90% of colonies consisting of By the early 1990s, overall breeding numbers fewer than 250 apparently occupied nests had declined further, although only slightly on (AONs).1 Fair Isle and Foula, which together held 60% of When colonies were resurveyed for the Sea- the population in 1991–94.3 All colonies have bird Colony Register in 1985, it was apparent been resurveyed at least once since 1994: those that, although there had been an overall on Fair Isle by Fair Isle Bird Observatory, on decrease in numbers of AONs, rates of change Noss by Scottish Natural Heritage (SNH), had varied greatly, both between and within on Foula by A.R.M, and elsewhere by M.H. and different colonies. It was decided that, instead R.M.M, with assistance from S.N.H. This paper of attempting to infer population change from presents these latest counts and it estimates the average rates at which the overall Shetland *Correspondence author. population, and the largest colonies, have Email: [email protected] declined since 1981.

© 1999 British Trust for Ornithology Shetland Kittiwake decline 49

METHODS

Aggregations of breeding Kittiwakes are most frequently referred to as ‘colonies’, and Boyd’s4 definition of groups of nests separated by at least 50 m of unoccupied cliff (or sea) was used in analyses of the 1981 survey.1 However, the term ‘colony’ has social as well as geographic connotations and here we have used the term ‘breeding station’,5 defined as colonies or colony groups less than a mile apart. Colonies on Foula were regarded as a single breeding station, because of their remoteness from other colonies and because earlier surveys were recorded as whole-island counts. The locations of all known breeding stations during 1981–97 are shown in Fig. 1. Surveys were carried out as detailed previously,3 i.e. from an inflatable boat in suitably calm weather during June, although sections of Eshaness, Noss and Fair Isle were counted from land, while the 1996 and 1997 surveys of Foula were made from a larger, hard-hulled boat. Although the basic count unit was an AON (defined as a well-built nest capable of holding eggs), trace nests (partially built or disinte- 25 km grated) were also counted where it was possible to distinguish them and the following analyses are of the total number of nests at each breeding station. Some nest counts here there- fore differ slightly from previously published figures, which referred to AONs only. To estimate the total breeding population in any year, the number of nests at each breeding Figure 1. The location of all known Kittiwake breeding stations in Shetland during 1981–97. The station in years between surveys was calcu- numbers refer to stations named in the Appendix. lated from average annual rates of change Stations referred to in the text are, in alphabetical between actual sets of counts (Appendix 1). At order: 48 Clett Head, Whalsay; 22 Eshaness; 57 Fair Noss, Foula and Fair Isle (the only stations not Isle; 56 Foula; 34 Hermaness; 8 Kettlaness; 53 Noness; surveyed in 1981), values for the early 1980s 49 Noss; 17 Papa Stour; 27 Ramna Stacks; 21 Skerry of were calculated from the mean annual percent- Eshaness; 41 Strandburgh Ness; 1 Sumburgh Head; age change between 1980–85, 1976–87, and 54 Troswick Ness; 9 West Burra; 12 Westerwick. 1975–86, respectively. To calculate the rate of decline at a breeding station found to have 1994 and 1995 which would have projected to been abandoned, it was presumed to have an improbable 426 nests by 1997. A second consisted of a single nest the previous year. The population estimate was therefore calculated, total Shetland breeding population was then based on the annual total of the last actual assumed to have been the annual sum of actu- counts at breeding stations. al, and calculated, numbers of nests. At In analyses, breeding stations are categorized breeding stations not surveyed since 1995/96, according to their size in 1981: small (fewer the number of nests in 1996/97 was calculated than 250 nests), medium (250–749 nests), large from the last known average annual rate of (750–1749 nests) and major (more than 1750 change. However, one recently established nests). station increased from 40 to 88 nests between Kittiwake breeding success was monitored in

© 1999 British Trust for Ornithology, Bird Study, 46, 48–61 50 M. Heubeck et al. study plots at 11 breeding stations between 50.5%, from 54 600 in 1981 (slightly higher than 1986 and 1997, using methods developed by Richardson’s figure1) to 27 000 in 1997 (Fig. 2, Harris.6 The number of study plots and the Appendix), an overall average annual decline proportion of the total of nests that each of 4.2%. However, the rate of decline was far breeding station contained varied, and in most from uniform, either over time or between cases they were not selected randomly from a breeding stations. pool of potential plots. Plots were visited every During 1981–85 an increase (+1.0% p.a.) in four to six days throughout the breeding numbers at Fair Isle, by far the largest breeding season, until the last remaining chicks were station, offset a steeper decrease (–3.1% p.a.) within four to six days of fledging. Breeding elsewhere in Shetland (Fig. 3). Then, between success was defined as the number of young 1985 and 1988 there was an increase of 7.3% fledged per nest at which an adult was seen to (+1.8% p.a.) in the total number of nests at be, or appeared to be, incubating. small, medium and large breeding stations (Fig. 2). Other than Fair Isle, which continued to increase during this period, none of the RESULTS major stations was counted twice during Had all extant breeding stations (i.e. disregard- 1985–88, so we do not know whether any ing the years before stations were known to concurrent increases occurred at Noss, Foula, have formed, and after they were known Hermaness or Sumburgh Head. The annual to have disappeared) been surveyed annually percentage change in the estimated total popu- during 1981–97 there would have been a total lation (Fig. 4) suggests that a steep decline in of 908 counts; the total achieved was 340 overall numbers began in 1988 (after which (37.4%). However, sampling of the population numbers also began to decrease on Fair Isle) varied considerably between years, with no that averaged –4.8% p.a. between 1987 and counts made during 1982–84 and in some other 1991 (–7.7% p.a. for Shetland excluding Fair years predominantly the smaller breeding Isle), but which halted abruptly in 1992. After a stations being surveyed. Nevertheless, the pro- further large decrease in 1993 the rate of decline portion of the estimated population actually began to slow, to –7.4% p.a. between 1996 and sampled exceeded 20% in all but one year 1997 (Fig. 4, Appendix). (1993) during 1985–97 (Table 1). Apart from Fair Isle, numbers at the four The estimated number of nests at Shetland’s other major breeding stations all declined Kittiwake breeding stations decreased by between 1981–97 (Fig. 5). The rates of decline at

Table 1. The number of extant Kittiwake breeding stations in Shetland and the number surveyed annually, the number of nests counted, and the percentage of the estimated population counted each year.

No. of breeding No. of No. of nests Proportion of estimated Year stations stations surveyed counted population sampled (%)

1981 55 52 20 783 38.1 1982–84 55 0 0 0 1985 55 23 20 494 38.1 1986 56 30 2 889 43.4 1987 54 15 10 923 21.7 1988 54 18 3 681 46.4 1989 55 30 10 139 21.2 1990 54 12 9 153 20.3 1991 53 35 11 010 26.7 1992 51 15 8 506 66.0 1993 52 29 5 057 13.7 1994 51 20 10 107 29.9 1995 51 24 8 076 26.1 1996 51 17 5 891 20.5 1997 51 21 5 111 64.2

© 1999 British Trust for Ornithology, Bird Study, 46, 48–61 Shetland Kittiwake decline 51

Fair Isle Noss Foula Hermaness



Sumburgh 750–1749 250–749 <250 55 000





50 000

45 000

40 000

35 000

30 000

Number of nests 25 000 

 

20 000   



 



15 000 





10 000





5000 

 0 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 Year

Figure 2. The estimated size of the Shetland Kittiwake breeding population during 1981–97, showing the number of nests in the five major breeding stations, and the totals for those that in 1981 consisted of 750–1749, 250–749, and fewer than 250 nests. The figures for 1996/97 are projected trends.

Noss, Hermaness and Sumburgh Head were breeding station) declined by 20.3% p.a. broadly similar until at least 1993/94. After an between 1995–97. The main period of decline at increase (5.9%) in 1994, breeding numbers at Foula was later than at the other breeding Sumburgh Head stabilized. The 1994 and 1995 stations, averaging –1.3% p.a. between 1981–92 counts at Noss suggested a sharp decrease in and –11.3% p.a. between 1992–97. the rate of decline. This trend was reflected in Trends in numbers at the five large breeding the four breeding success plots, which in 1995 stations varied considerably between 1981–89 contained 12.1% of nests at the breeding (Fig. 6) and during this period the total number station, and where numbers of nests declined of nests in these stations only declined by 0.9% by –4.5% p.a. between 1995–97 (SNH, unpubl. p.a. After 1989, substantial decreases occurred data). at all five stations with only Noness showing The recent situation at Hermaness is un- an increase in 1994, similar to that at Sumburgh known, although the numbers of nests in two Head (Fig. 5). At the other four stations, all on breeding success plots (6.5% of the total at the the west coast, the rate of decline increased

© 1999 British Trust for Ornithology, Bird Study, 46, 48–61 52 M. Heubeck et al.

60 000 10 000 50 000

7500 40 000

5000 30 000

20 000 2500 Number of nests Number of nests

10 000 1000 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 Year Year Figure 3. Estimated size of Shetland Kittiwake Figure 5. Estimated size of Kittiwake populations at breeding population: (■) Shetland total; (●) Shetland Noss (■, ■), Foula (◆, ◆), Hermaness (●, ●) and excluding Fair Isle; (▲, ▲) Fair Isle, (▲) actual counts. Sumburgh Head (▲, ▲). ■, ◆, ●, ▲, actual counts. progressively after 1989, and between 1993/94 1500 and 1997 reached –7.8% p.a. at Papa Stour, –20.2% p.a. at Ramna Stacks, –37.7% p.a. at Eshaness and –45.2% p.a. on the Skerry of 1000 Eshaness. Breeding stations with fewer than 750 nests 750 comprised less than 20% of the Shetland Kittiwake breeding population in 1981, 1989 500

0

–2.5 Number of nests 250 –5 –7.5 –10 –12.5 Annual change (%) –15 100 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 Year Year Figure 6. Estimated size of Kittiwake populations at Figure 4. Estimated annual change in Shetland Noness (■, ■), Ramna Stacks (◆, ◆), Skerry of Kittiwake breeding population: (■) Shetland total; Eshaness (●, ●), Papa Stour (▲, ▲) and Eshaness (●+). (●) Shetland excluding Fair Isle. ■, ◆, ●, ▲, actual counts. © 1999 British Trust for Ornithology, Bird Study, 46, 48–61 Shetland Kittiwake decline 53

Table 2. The total number of nests and the percentage of the Shetland breeding population calculated to be in breeding stations of different sizes in 1981, 1989 and 1997, and the percentage changes for 1981–89 and 1989–97. Two values are given for the 1989–97 percentage change and the 1997 population, based on the last actual count at each breeding station (upper), and 1997 figures calculated from the average annual rate of change between the last two sets of actual counts (lower). Chi-squared tests were conducted on the 1981, 1989 and 1997 totals of nests in each size category in 2 × 6 contingency tables.

1981 1989 1997

No. of No. of Total Populn Change a No. of Total Populn Change b No. of Total Populn nests stations nests (%) (%) stations nests (%) (%) stations nests (%)

<50 15 367 0.7 –40.9 18 226 0.5 +72.6 17 238 0.9 +241.2 16 183 0.7 50–249 16 1789 3.3 –5.1 16 1836 3.8 –27.0 22 2819 10.4 –27.9 21 2723 10.1 250–749 14 5975 11.0 –24.1 9 4398 9.2 –68.4 4 2046 7.6 –70.4 5 2484 9.2 750–2499 5 6139 11.3 –6.9 6 7525 15.8 –54.7 3 5443 20.1 –54.5 3 5198 19.2 2500–4999 3 11 491 21.1 –21.6 2 7198 15.1 –44.2 1 4266 15.8 –47.6 1 4197 15.5 >5000 2 28 803 52.8 –7.8 2 26 560 55.6 –37.9 1 12 224 45.2 –38.2 1 12 224 45.3

Total 55 54 564 –12.5 53 47 743 –43.4 48 27 036 –43.4 47 27 009 χ2 1041.3 1858.1 1596.3 df 5 5 5 P <0.001 <0.001 <0.001 a1981–89. b1989–97. and 1997 (Table 2). Between 1981 and 1989, the DISCUSSION total number of nests at these stations Methodology decreased by 20.6% compared to 11.1% at larger stations, with the smallest (<50 nests in Many Kittiwake nests in Shetland are not read- 1981) experiencing the greatest percentage ily visible from land. In 1981, a comparison of decrease. Breeding stations were re-ranked by land and sea counts at 21 stations found the size in 1989 and, in contrast to the first period, total number of nests counted from land was the total number of nests in the smallest 37% lower than for counts from the sea.7 The stations (<50 nests in 1989) increased while the actual figure for the proportion of nests not largest proportional decreases occurred at visible from land probably exceeds 60%, since stations holding 250–2499 nests in 1989 (Table 2). breeding stations completely inaccessible from Breeding success varied greatly during land were excluded from the 1981 comparison 1986–97, both between years and between and because at some breeding stations there breeding stations in any one year (Table 3). has been an increase in the proportion of pairs There was a general decline in success between nesting in highly sheltered situations, such as 1986 and 1988–90, when there was widespread cave entrances, that are unlikely to be visible breeding failure except at stations in northwest from the clifftop.3 Thus it was decided to Shetland. Success improved markedly in 1991 monitor changes in breeding numbers from and 1992, but then began to decline at some the sea. stations and was uniformly low in 1997. With so many breeding stations dispersed

© 1999 British Trust for Ornithology, Bird Study, 46, 48–61 54 M. Heubeck et al.

Table 3. Kittiwake breeding success at monitored breeding stations, 1986–97, defined as the number of young fledged per nest at which the adult was seen, or appeared to be incubating. At stations with multiple plots, the mean of individual plot values is given. Shetland breeding success is the mean of individual station values.

Name No. (No. of plots) 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997

1. Sumburgh Head (4–9) 0.59 0.63 0.01 0.06 0.00 0.68 1.55 1.32 1.04 0.59 0.90 0.14 8. Kettlaness (1) 0.58 0.00 0.00 0.00 0.06 0.00 0.00 9. West Burra (1) 0.81 1.30 0.74 0.84 0.33 12. Westerwick (1) 0.03 0.00 0.00 0.00 0.72 0.55 1.13 1.18 1.04 0.28 0.28 22. Eshaness (1–2) 0.70 0.67 0.46 0.65 0.51 0.79 0.83 0.05 0.00 0.00 0.00 0.00 34. Hermaness (2) 0.47 0.48 1.19 1.20 0.64 0.52 0.22 0.13 0.09 41. Strandburgh Ness (1) 0.11 0.00 0.00 0.00 0.53 0.67 49. Noss (4–5) 0.35 0.27 0.00 0.00 0.00 0.17 0.31 0.36 0.51 0.50 0.74 0.07 53. Noness (1) 0.96 0.25 54. Troswick Ness (1) 1.11 0.46 0.00 0.02 0.00 0.44 1.14 0.96 0.10 0.00 57. Fair Isle (5–10) 1.10 1.15 0.09 0.39 0.00 0.94 1.32 1.08 1.20 0.88 1.24 0.67

Overall mean success 0.77 0.49 0.07 0.18 0.11 0.61 0.86 0.65 0.73 0.50 0.64 0.23 se 0.15 0.13 0.06 0.08 0.07 0.12 0.19 0.17 0.18 0.14 0.16 0.07 along 1450 km of coastline, it was impractical to larly when there are long intervals between survey the entire Shetland Kittiwake breeding surveys of major breeding stations. In this population in a single year. Furthermore, the context, it is probably advisable to survey the short survey period (ideally the first three largest stations such as Noss every three or four weeks of June) and the unpredictable nature of years, rather than to plan five-yearly surveys the weather meant that surveys had to be into work programmes. Thirdly, the latest carried out opportunistically rather than in a (1996–97) population estimates were hypo- planned, systematic manner. This resulted in thetical, being based on projected trends at a the problem of describing an overall popula- large proportion of breeding stations rather tion trend based on counts at different breeding than observed changes in numbers. stations in different years and at varying An alternative, more cautious, measure of intervals. population trend was the annual sum of the We believe that the method described here is latest actual count at each breeding station. appropriate for Shetland, although there are There was little difference (±5%) between this three weaknesses. First, it is more difficult to and the estimated total during 1981–88, when count large numbers of nests accurately from an increase at Fair Isle partly offset decreases the sea than from land, and extrapolating elsewhere in Shetland (Appendix). However, average annual rates of change from inaccurate when more widespread declines at breeding counts is undesirable, particularly when there stations began after 1988, annual differences are long intervals between counts or large between the estimated and last-known totals between-year differences in counts. However, increased to as much as 24%, particularly as the most individual colonies (the maximum length of time between counts at the major recording unit in the field) were of 100 nests or breeding stations increased. fewer and so rather easy to count. Moreover, a greater error than count accuracy may arise Timing, scale, and causes of the decline from surveys late in June when a high propor- tion of breeding attempts may have failed and Although the Shetland Kittiwake breeding nests been deserted, or early in June during a population declined during 1981–87, a far late breeding season. greater decrease occurred during the late 1980s Secondly, using average annual rates of and 1990s. This was probably due largely to a change to calculate population estimates will combination of poor breeding success in the have a considerable ‘smoothing’ effect, particu- late 1980s7 that created a shortage of recruits to

© 1999 British Trust for Ornithology, Bird Study, 46, 48–61 Shetland Kittiwake decline 55 the breeding population from about 1989 diet of Kittiwakes in Shetland during the onwards, together with an increase in the breeding season,18–21 although little is known of intensity of predation of both adult Kittiwakes the diet prior to breeding. During the late and nestlings by Great Skuas Stercorarius skua, 1980s, successive years of very low recruitment particularly from 1988 onwards.9,10 of young sandeels in Shetland waters led to a With the loss of 27 600 breeding pairs since shortage of sandeels in the area generally.22 1981 (equivalent to 5% of the most recent, This was followed by a very large recruitment 1985–87, estimate of the breeding population of in 1991,23 which resulted in an increase in the the British Isles),11 it is perhaps surprising that biomass of 1-group sandeels (hatched in the only 11 of the 55 breeding stations known in previous calendar year) around Shetland 1981 were abandoned. Eight were small, never during the first half of the year from 3200 having been known to contain more than 50 tonnes in 1991 to 97 000 tonnes in 1992; low nests. In contrast, Clett Head, Whalsay (No. 48, recruitment in 1992 reduced this again to 11 500 Appendix) is an example of the resilience of tonnes in 1993.24 Many experienced Kittiwakes some of these small stations. Adults were shot may have failed to breed in 1990 and 1991. This illegally during the 1970s and early 1980s and may have contributed to the relatively large numbers declined to three nests in 1989, but declines in estimated numbers of nests in those increased again by an average of 66% per years, while greatly improved feeding condi- annum over the following six years to 63 nests tions in the spring of 1992 may have led to a in 1995. very high proportion of birds of suitable age In Orkney, Kittiwake breeding numbers building nests that year. declined by 40% between 1969–70 and Breeding success and survival rates of adults 1985–87,11 with most of the decline occurring and immatures are key determinants of popu- between 1980 and 1985.12 Numbers at five lation dynamics, as are rates of emigration and breeding stations then increased by about 6% immigration, since the Shetland Kittiwake p.a. between 1985/86 and 1991, but declined population is unlikely to be completely closed. again by 3–4% p.a. during the 1990s (JNCC, Only data on adult survival and breeding suc- unpubl. data). Whether the increase in Orkney cess exist for Shetland. On Fair Isle, adult between 1985/86 and 1991 (a period of steep survival averaged 84.2% during 1992–97,25 decline in Shetland) involved breeders moving lower than on the Isle of May (89.0%),26 in from Shetland is unknown, but the coincidental northern Norway (92.2%)27 and Alaska timing of these opposing trends in neigh- (92.6%),28 but higher than in Brittany (80.8%)29 bouring island groups is intriguing. No consis- and over a 34-year period at North Shields tent regional declines have been recorded (78.3% for males, 81.8% for females).30 elsewhere in Scotland since the mid-1980s.13 Estimates of the level of productivity required Changes in numbers of nests may reflect to maintain a stable, closed Kittiwake popula- changes in the extent of non-breeding, as well tion vary from 0.88 fledged per nest (based on as changes in the numbers of potential immature survival to first breeding of 43% and breeders. At North Shields, the incidence of an adult survival rate of 81%)31 to 0.53 (imma- intermittent breeding in Kittiwakes was high- ture survival of 31% and an adult survival rate est among birds in the year after they first bred of 92%).27 Assuming a linear relationship (22%) and among birds that first bred at an between these estimates, a figure of about 0.8 early age (30%), but low among experienced may be appropriate for Shetland. There are no breeders (4%).14 However, non-breeding is data on immature survival to age of first more prevalent in years of food shortage breeding for Shetland Kittiwakes, but figures of during spring. On the Isle of May, 32% of site- 30%, 40% and an improbable 50% would holding birds did not build nests in a year of require breeding success of 1.1, 0.8 and 0.6, apparent food shortage compared to a mean of respectively, to compensate for the 15.8% 5% in six other years,15 while in similar circum- annual adult mortality observed at Fair Isle. stances up to 40% of pairs did not lay eggs in The poor breeding success during the late Newfoundland,16 and over 70% of birds did not 1980s was clearly sufficient to cause a marked build nests in Alaska.17 reduction in the adult population, although Sandeels Ammodytes marinus are the main movement of experienced breeders away from

© 1999 British Trust for Ornithology, Bird Study, 46, 48–61 56 M. Heubeck et al.

Shetland may also have occurred. Furthermore, 1981–1994. Scot. Birds, 17, 192–204. both immature and adult survival rates may 4. Boyd, J.M. (1960) The distribution and numbers of have been lower during this period of food Kittiwakes and Guillemots at St Kilda. Br. Birds 53, shortage than during the mid-1990s. 252–264. 5. Coulson, J.C. (1963) The status of the Kittiwake in Most Kittiwakes first breed when four or five 14 the British Isles. Bird Study, 10, 147–179. years old, and hence the improvement in 6. Harris, M.P. (1987) A low-input method of moni- breeding success in Shetland during 1991–93 toring Kittiwake Rissa tridactyla breeding success. probably contributed to the reductions in the Biol. Conserv., 41, 1–10. rate of population decline in 1996 and 1997. 7. Pritchard, D. (1981) Monitoring for conservation: During 1994–97, breeding success varied the Kittiwake in Shetland. MSc thesis, University greatly, ranging from nil to 1.30 at different of London. stations in any one year, but mean success was 8. Heubeck, M. & Ellis, P. 1986. Shetland Seabirds less than 0.8 in each of the previous four 1985. BTO News, 143, 10. years. In some years (notably 1997), Kittiwakes 9. Hamer, K.C., Furness, R.W. & Caldow, R.W.G. (1991) The effects of changes in food availability apparently experienced difficulty finding suffi- on the breeding ecology of great skuas Catharacta cient food for their young, while predation of skua in Shetland. J. Zool. Lond. 223, 175–188. nestlings and adults by Great Skuas has 10. Heubeck, M., Mellor, R.M. & Harvey, P.V. (1997) persisted at a high level and by 1997 occurred Changes in the breeding distribution and num- at all monitored breeding stations (pers. obs.). bers of Kittiwakes Rissa tridactyla around Unst, There was again a large recruitment of 0-group Shetland, and the presumed role of predation by sandeels (hatched in the current calendar year) Great Skuas Stercorarius skua. Seabird, 19, 12–21. into Shetland waters in August 1997 (SOAFD, 11. Lloyd, C., Tasker, M.L. & Partridge, K. (1991) The unpubl.). Previous experience suggests that Status of Seabirds in Britain and Ireland. T. & A.D. this may lead to increased sandeel abundance Poyser, London. 12 Benn, S., Tasker, M.L. & Reid, A. (1987) Changes and improved breeding success in 1998/99. in numbers of cliff-nesting seabirds in Orkney, However, the current level of predation by 1976–1985. Seabird, 10, 51–57. Great Skuas, particularly of adult Kittiwakes, is 13. Thompson, K.R., Brindley, E. & Heubeck, M. 32 considered unsustainable and seems likely to (1997) Seabird Numbers and Breeding Success in result in a continued downward trend in the Britain and Ireland, 1996. UK Nature Conservation Shetland Kittiwake population. No. 21. Joint Nature Conservation Committee, Peterborough. 14. Wooler, R.D. & Coulson, J.C. (1977) Factors affect- ACKNOWLEDGEMENTS ing the age of first breeding of the Kittiwake Rissa Surveys were carried out by the University of tridactyla. Ibis, 119, 339–349. 15. Harris, MP. & Wanless, S. (1997) Breeding success, Aberdeen, under contract to the Shetland Oil diet, and brood neglect in the kittiwake (Rissa tri- Terminal Environmental Advisory Group, by dactyla) over an 11-year period. ICES J. Mar. Sci., the Nature Conservancy Council and its 54, 615–623. successor Scottish Natural Heritage, by Fair Isle 16. Regehr, H.M. & Montevecchi, W.A. (1997) Interac- Bird Observatory and by Tony Mainwood. We tive effects of food shortage and predation on are grateful to all who assisted, and to Mike breeding failure of black-legged kittiwakes: indi- Harris, Chris Wernham, Kate Thompson and rect effects of fisheries activities and implications Jim Reid who commented on earlier drafts of for indicator species. Mar. Ecol. Prog. Ser., 155, the manuscript. 249–260. 17. Murphy, E.C., Springer, A.M. & Roseneau, D.G. (1991) High annual variability in reproductive REFERENCES success of Kittiwakes (Rissa tridactyla L.) at a colony in western Alaska. J. Anim. Ecol., 60, 1. Richardson, M.G. (1985) Status and distribution of 515–534. the Kittiwake in Shetland in 1981. Bird Study, 32, 18. Furness, R.W. (1982) Population, breeding biology 11–18. and diets of seabirds on Foula in 1980. Seabird 2. Heubeck, M., Richardson, M.G. & Dore, C.P. Rep., 6, 5–11. (1986) Monitoring numbers of Kittiwakes Rissa 19. Harris, M.P. & Riddiford, N.J. (1989) The food of tridactyla in Shetland. Seabird, 9, 32–42. some young seabirds on Fair Isle in 1986–88. Scot. 3. Heubeck, M. & Mellor, R.M. (1994) Changes in Birds, 15, 119–125. breeding numbers of Kittiwakes in Shetland, © 1999 British Trust for Ornithology, Bird Study, 46, 48–61 Shetland Kittiwake decline 57

20. Harris, M.P. & Wanless, S. (1990) Breeding success 27. Erikstad, K.E., Tveraa, T. & Barrett, R.T. (1995) of British Kittiwakes Rissa tridactyla in 1986–88: Adult survival and chick production in long-lived evidence for changing conditions in the northern seabirds: a 5-year study of the kittiwake Rissa tri- North Sea. J. Appl. Ecol., 27, 172–187. dactyla. In Ecology of Fjords and Coastal Waters (eds 21. Hamer, K.C, Monaghan, P., Uttley, J.D., Walton, P. H.R. Skjoldal, C. Hopkins, K.E. Erikstad & H.P. & Burns, M.D. (1993) The influence of food supply Leinaas), pp. 471–477. Elsevier Science, Amster- on the breeding ecology of Kittiwakes Rissa tri- dam. dactyla in Shetland. Ibis, 135, 255–263. 28. Hatch, S.A., Roberts, B.D. & Fadely, B.S. (1993) 22. Bailey, R.S., Furness, R.W., Gauld, J.A. & Kunzlik, Adult survival of Black-legged Kittiwakes Rissa P.A. (1991) Recent changes in the population of tridactyla in a Pacific colony. Ibis, 135, 247–254. the sandeel (Ammodytes marinus Raitt) at Shetland 29. Danchin, E. & Monnat, J.-Y. (1992) Population in relation to estimates of seabird predation. ICES dynamics modelling of two neighbouring Kitti- Mar. Sci. Symp., 193, 209–216. wake Rissa tridactyla colonies. Ardea, 80, 171–180. 23. Wright, P.J. & Bailey, M.C. (1993) Biology of 30. Aebischer, N.J. & Coulson, J.C. (1990) Survival of Sandeels in the Vicinity of Seabird Colonies at Shet- the Kittiwake in relation to sex, year, breeding land. Fisheries Research Report No. 15/93. experience and position in the colony. J. Anim. SOAFD Marine Laboratory, Aberdeen, Scotland. Ecol., 59, 1063–1071. 24. Suddaby, D. & Ratcliffe, N. (1997) The effects of 31. Chapdelaine, G. & Brousseau, P. (1989) Size and fluctuating food availability on breeding Arctic trends of Black-legged Kittiwake (Rissa tridactyla) Terns (Sterna paradisaea). Auk, 114, 524–530. populations in the Gulf of St Lawrence (Quebec) 25. Riddington, R., Bull, J., Newell, M.A & Reid, J.M. 1974–1985. Am. Birds, 43, 21–24. (1998) Fair Isle seabird studies 1997. JNCC Report, 32. Furness, R.W. (1997) The Impact of Predation by No. 74. Great Skuas on Other Seabird Species, with Special 26. Harris, M.P. & Calladine, J. (1993) A check on the Reference to Special Protection Areas in Shetland. efficiency of finding colour-ringed Kittiwakes Final Report on contract 5092AA, Scottish Natur- Rissa tridactyla. Ringing Migr., 14, 113–116. al Heritage, Lerwick.

(MS received 9 January 1998; revised MS accepted 22 June 1998)

© 1999 British Trust for Ornithology, Bird Study, 46, 48–61 58 M. Heubeck et al.

APPENDIX

Counts of Kittiwake nests (AONs plus traces) at Shetland breeding stations. Those in bold are actual counts, those not bold were calculated by average annual rate of change between actual counts. Values for Noss, Foula and Fair Isle in the early 1980s were calculated from counts in 1980, 1976 and 1975, respectively. Two figures for the Shetland population are given, derived from the annual totals of (a) the estimated numbers in breeding sta- tions calculated from mean annual rates of change, and (b) the last actual count at each breeding station.

No. Name 1981 1982 1983 1984 1985 1986 1987

1. Sumburgh Head 2641 2592 2544 2497 2451 2273 2107 2. Horse Island 594 573 554 534 516 327 371 3. Siggar Ness 0 00001 2 4. Fitful Head 500 438 384 336 295 258 268 5. St Ninian’s Isle 47 50 54 57 61 65 84 6. Ness of Ireland 203 193 184 176 167 160 152 7. South Havra 29 23 18 14 11 919 8. Kettla Ness 281 288 295 302 310 317 325 9. West Burra 615 657 701 749 800 854 912 10. Reawick 272 259 246 234 222 212 201 11. Skelda Ness 188 152 122 98 79 64 84 12. Westerwick 454 413 376 343 312 284 296 13. Burga Stacks 58 32 18 10 5 3 5 14. Vaila 255 264 272 282 291 327 370 15. Braga Ness 12 14 17 20 23 27 24 16. Wats Ness 11 10 10 9 8 8 7 17. Papa Stour 1059 1066 1073 1080 1087 1073 1060 18. South Muckle Roe 96 97 97 98 99 69 81 19. Swabie Stack 56 53 51 48 46 83 77 20. 375 363 351 339 328 340 353 21. Skerry of Eshaness 1266 1267 1268 1270 1271 1314 1359 22. Eshaness 837 809 782 756 731 831 944 23. 35 32 29 26 23 21 13 24. Gruna Stack 50 54 58 62 67 72 77 25. Uyea 731 614 517 434 365 424 492 26. 14 74210 0 27. Ramna Stacks 1350 1251 1158 1073 994 1047 1010 28. Trumba 2 211111 12 29. Varnadil 56 27 13 6 31010 30. Holm 78 48 30 18 11 7 11 31. Blue Mull 158 150 142 134 128 121 115 32. Lang Holm 32 20 13 8 5 3 2 33. South Holms 121 142 166 195 229 269 315 34. Hermaness 3872 3823 3775 3728 3681 3635 3589 35. Saxavord 141 151 162 174 186 293 462 36. Virdik 84 83 82 81 81 80 79 37. Burgar 217 218 218 219 220 220 221 38. Mooa Stack 53 59 65 72 79 88 97 39. Ramaberg 0 0 0 0 0 0 0 40. Clett Stack 30 29 28 27 26 50 26 41. Strandburgh Ness 273 226 188 156 129 34 54 42. Southeast Fetlar 32 10 3 1 02027 43. Lambhoga 37 43 49 57 66 67 47 44. Birrier 82 50 30 18 11 43 57 45. 392 318 257 209 169 168 133 46. Grunay 261 233 208 186 166 148 118 47. North Benelip 23 23 23 24 24 24 20

continued © 1999 British Trust for Ornithology, Bird Study, 46, 48–61 Shetland Kittiwake decline 59

No. 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997

1. 1954 1812 1638 1481 1397 1342 1421 1423 1425 1427 2. 421 478 387 314 145 71 66 60 55 50 3. 3 5 17 60 74 102 86 72 61 51 4. 279 290 230 182 182 182 185 187 190 193 5. 91 98 132 178 162 147 131 117 104 93 6. 139 127 119 112 124 137 154 173 194 218 7. 10 0 0 0 0 0 0000 8. 294 266 216 175 92 48 13 3 1 0 9. 779 665 514 397 349 306 290 275 261 247 10. 155 120 93 78 57 41 43 46 63 87 11. 110 92 77 77 67 58 55 53 59 66 12. 309 231 173 166 149 134 165 203 191 179 13. 8 10 12 17 17 18 22 26 25 24 14. 412 316 242 174 140 112 116 121 130 139 15. 21 18 15 10 6 4 2 2 1 0 16. 7 3 1 1111111 17. 1046 1033 998 965 932 897 827 762 703 648 18. 95 112 72 47 30 33 52 82 100 122 19. 71 66 73 81 90 60 26 11 6 3 20. 366 379 339 303 266 233 204 55 15 4 21. 1405 1453 1221 1026 933 850 773 417 230 133 22. 1072 1218 1178 1140 993 865 753 469 292 176 23. 8 5 210 1 3 210 24. 82 89 91 93 95 88 82 87 92 97 25. 570 662 658 655 651 555 474 438 404 373 26. 0 0 000 0 0 000 27. 974 940 922 904 886 773 675 539 430 343 28. 13 14 15 15 16 17 18 17 15 14 29. 10 10 10 10 11 12 14 15 17 18 30. 19 30 50 81 78 78 77 76 75 74 31. 107 100 93 87 64 48 35 26 19 14 32. 2 1 1 0 0000 00 33. 217 150 103 71 37 19 10 5 31 34. 3207 2866 2561 2288 2110 1947 1795 1656 1527 1409 35. 477 492 508 524 375 268 192 137 98 70 36. 61 48 37 29 23 19 15 12 10 8 37. 216 210 205 200 200 200 199 199 199 199 38. 90 83 77 71 56 45 35 28 22 18 39. 0 3 581424 40 88 194 426 40. 14 5 1 0 0 000 0 0 41. 87 111 92 76 73 70 50 39 31 24 42. 36 27 13 6 6 676 5 4 43. 33 18 23 30 30 30 11 14 18 23 44. 76 77 54 38 47 58 62 70 78 87 45. 106 95 68 49 71 102 111 112 114 116 46. 94 75 50 34 23 12 5 2 10 47. 16 13 14 16 17 22 16 12 97

continued © 1999 British Trust for Ornithology, Bird Study, 46, 48–61 60 M. Heubeck et al.

APPENDIX continued

No. Name 1981 1982 1983 1984 1985 1986 1987

48. Clett Head 25 20 16 13 10 8 6 49. Noss 10707 10374 10052 9740 9438 8895 8383 50. Millburn Geo 19 20 22 23 25 27 26 51. Hole of Bugars 19 94210 0 52. Mousa 148 92 58 36 22 14 9 53. Noness 1627 1591 1556 1522 1489 1465 1441 54. Troswick Ness 716 661 610 563 520 540 561 55. Boddam 256 261 266 272 277 321 371 56. Foula 4978 4867 4759 4654 4550 4449 4350

Total minus Fair Isle 36468 35121 33979 32988 32110 31473 31235

57. Fair Isle 18096 18286 18477 18671 18866 19064 19202

Calculated total (a) 54564 53407 52456 51659 50976 50537 50437 Last known total (b) 54403 54403 54403 54403 50943 51926 51111

Difference (%) –0.3 +1.9 +3.7 +5.3 –0.1 +2.7 +1.3

© 1999 British Trust for Ornithology, Bird Study, 46, 48–61 Shetland Kittiwake decline 61

No. 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997

48. 4 3 51018 33 46 63 87 120 49. 7900 7446 7017 6209 5494 4861 4301 4266 4321 4197 50. 26 12 5210 0000 51. 0 0 0000 0000 52. 6 1 0 00100 1 1 53. 1417 1069 806 862 819 544 623 670 682 694 54. 583 453 352 338 273 204 189 61 34 19 55. 430 397 366 370 368 277 244 218 212 206 56. 4341 4332 4324 4315 4306 3858 3457 3097 2775 2362

Total minus Fair Isle 30269 28629 26275 22681 22368 19813 18171 16513 15491 14785

57. 19340 19114 18890 18670 18451 16993 15649 14412 13273 12224

Calculated total (a) 49609 47743 45165 41351 40819 36806 33820 30925 28764 27009 Last known total (b) 51679 51694 47906 44774 43475 42316 38583 37378 35762 27036

Difference (%) +4.2 +8.3 +6.1 +8.3 +6.5 +15.0 +14.1 +20.9 +24.3 +0.1

© 1999 British Trust for Ornithology, Bird Study, 46, 48–61