Provisioning activity of Common Guillemots (Uria aalge) was studied on Lundy between 2012 and 2015. A video camera was used to record 2 hour sessions in the morning at the same breeding ledge. Analysis of provisioning rates and parental attendance showed significant differences in activity between the four years. Provisioning rates were significantly lower in 2012 compared to 2013. Parental attendance was significantly lower in 2012 and 2015 compared to 2014. More detailed analysis of the 2012 data showed that, in the later part of the breeding season, provisioning rates were significantly lower than in the early part of the 2012 and in all subsequent seasons. Parental attendance in the later part of 2012 was significantly lower than in 2014 or in 2013. Chicks were intentionally left unattended on two occassions while their parents went out to sea to feed. Both occured in the later half of 2012. In one of these cases, brooding duties had been shared with a neighbouring, chickless pair. Comparing these observation to those in other colonies suggests that the later half of 2012 was a time of low food availability.
The Common Guillemot (Uria aalge Pont.) or Common Murre is a large (c.950g), long-lived seabird of the auk family (Gaston & Jones 1998). They occur in the north of the Pacific and Atlantic Oceans and in the Arctic Ocean. Guillemots breed on on sea-cliffs in tightly-packed ledges or in rocky crevices, often in mixed colonies with other sea-birds. Breeding adults make no nest but instead their egg or chick is sheltered under the body or wing of one parent. In the southern areas of their range, they return to these breeding colonies from mid-October onwards.
Breeding phenology varies with sea-temperature, the start of egg-laying in Arctic colonies can be a month later than those in Britain or California. A single egg is laid and parents take turns in incubating, with one parent present in all but exceptional times. Egg-laying is more synchronised within a ledge than between different ledges of the same colony. Eggs are lost to predators or accidently knocked from ledges due to high breeding density. A replacement egg (or occasionally second replacement egg) can be laid if the first egg is lost. Only one chick is raised per season, although Harris (et al. 2000) reported a case where a pair successfully raised an orphan as well as their own chick to fledging age. The pair inceased their foraging time at the expense of time spent at the colony
Parental attendance Once the chick has hatched, parents take turns in brooding, again with one parent present in all but exceptional times. The off-duty parent will either spend time at the ledge or away from the ledge (flying, washing, and foraging for themselves and for their chick). Guillemots can vary their foraging effort to compensate for variations in food availability, either by switching to alternative prey-fish, or by increasing the amount of time spend foraging (Burger & Piatt 1990). Decreased attendance at times of low food availability has been seen in Norway (Tschanz 1979), eastern Canada (Birkhead & Hatchwell 1984, Burger & Piatt 1990), Scotland (Uttley et al. 1994, Monaghan et al. 1994, Wanless et al. 2005, Ashbrook et al. 2008), and Alaska (Zador & Piatt 1999, Harding et al. 2007).
Pairs that have lost their egg or chick will sometimes brood a neighbouring chick (Tschanz 1979). This allo-brooding can be beneficial to the chick, especially in times of low food availability when chicks may left unattended while both parents forage (Birkhead & Hatchwell 1984).
Provisioning rates Foraging adults return to their chick with one fish at a time (single-prey loading), usually held lengthways in their bill. The returning adult half-opens its wings to 'tent' its chick whilst the fish is transferred. Older chicks will take the fish directly from their parent. Younger chicks will wait until the fish is dropped and then take it from the ground. Provision rates count the number of fish delivered per chick in a given time. Different colonies have different feeding areas, so normal provisioning rates vary from colony to colony. When combined with the size and species of fish provisioning rates can give an estimate of the total energy intake of a chick. Fish fed to chicks at UK colonies are mainly sandeels, clupeids (Sprats and Mackerel) or gadiods (Whiting, Cod) (Anderson 2014).
At times on severe food shortage, adults will abandon their chick to feed themselves. This behaviour allows long-lived seabirds to maximise their breeding success over their life-time at the expense of a single chick. Unlike large Gulls, adult Guillemots will only kill neighbouring chicks in times of extreme food stress (Ashbrook et al. 2008). In these cases, it is only birds feeding their own chick that will kill neighbouring chicks, failed breeders still show allo-parental behaviour towards unattended chicks.
'Junk Food' Hypothesis In some years, provisioning rates are in the normal range but parental attendance is low. Analysis of energy content has shown that prey-fish were malnourished (Wanless et al. 2005)
Fledging Guillemot chicks fledge after two to three weeks of being fed on the breeding ledge. At this age they can regulate their own temperature but are unable to fly. They go to sea with the male parent, who continues to feed his chick until it is able to feed itself. Captive chicks were able to pick up fish from the bottom of a tank a week after 'fledging'. Fledging can fail through predation, or through communication failure between the chick and adult (Greenwood 1963 & 1964)
Female Guillemots stay on the breeding ledge after their chick and partner have departed. They may stay for another two weeks in normal conditions, but can leave after a few days in times of low food availability. Their partners may return to the ledge if fledging fails.
Guillemots on Lundy Lundy is a three mile long granite island located where the Bristol Channel meets the Atlantic Ocean. It is the largest seabird colony in South-West England, with 4,114 Guillemots on breeding ledges in 2013 (Price and Booker, 2014) The first major study of Guillemots on Lundy was by Richard Perry in 1939 (Perry 1940). Perry made detailed descriptions of the breeding behaviour of Guillemots as well as Razorbills, Puffins and Kittiwakes. Since then Guillemot research has concentrated on ringing (LFS 1940s-1960s), first-flighting (Greenwood 1962), breeding success (Taylor, 1970s), population (David Price, RSPB, 1980s - present), comparative feeding ecology (Birkhead et al. 1986, Hatchwell et al. 1992) and productivity (Natural England Warden, 2007 – present).
The feeding ecology of Guillemot chicks on Lundy (specifically chick diet, provisioning rates, and feeding trip duration) was compared with nearby islands in the mid 1980s. In 1985, Lundy was compared with Skomer and Great Saltee (Birkhead et al. 1986). In a continuation of that study in 1986, Lundy was compared with Skomer (Hatchwell et al. 1992). These studies showed Guillemots from these colonies had different feeding areas: trip durations were too short, and relative proportions of Sprats and Sandeels in the diet of chicks were different.
Although feeding areas are different, there is movement of non-breeding birds between these three islands. An immature Guillemot ringed on Skomer was later seen on Great Saltee (Birkhead & Hudson 1977) At least four birds ringed as chicks on Skomer have been recorded on Lundy (Taylor), and one of these was re-sighted on Skomer five days after being seen on Lundy (Tim Birkhead personal communication). One of the adult birds for this study was ringed as a nestling on Great Saltee in 2000 (Taylor? 2015 LFS ringing report?)
Vedöy, Norway, 1956-1975 (Tschanz 1979) Chicks were only left unattended at times of food shortage.
Brunnich's Guillemot, Prince Leopold Island 1975-1978 & Eastern Digges Island 1980, Canada (Gaston & Nettleship 1982) Colony attendance correlated chick fledging weight. Hypothesis: colony attendance depends on food availability.
Gannet Clusters, Labrador, 1982-1983 (Birkhead & Hatchwell 1984) Nunber of days that chick were unattended was higher in 1982, when food was less abundant.
Witless Bay, Newfoundland, 1983-1985 (Burger & Piatt 1990) Although prey abundance of Capelin varied 10-fold, provisioning rates remained constant (mean 0.28 feeds per chick per hour). Birds compensated by switching to other prey species, and by increasing the amount of time spent away from the colony (mean attendance 22.3%). Parental attendance was lower in the morning and higher in the evening.
Isle of May , Scotland, 1983-1984 (Wanless & Harris 1986) Off duty birds were recorded at the site more often in 1984 than in 1983.
Witless Bay, Newfoundland, 1985 (Cairns et al. 1987, Cairns 1988) Time away from colony correlated with time spent diving. Birds spent most time on the water when not foraging or at the nest site.
Sumburgh Head, Shetland, 1990 & 1991 (Uttley et al. 1994, Monaghan et al. 1994) "Compared with the high food availability year, adult Guillemots in the year of low food availability spent much less time resting at the breeding colony" "rate of chick feeding, chick weight and fledging success were greater in the year of high food availability" "Time spent foraging in the poor food year was at the expense of time spent sitting at the colony"
Cook Inlet, Alaska, 1995 (Zador & Piatt 1999) "attendance time-budgets provide a more sensitive index of food availability than other breeding parameters"
Cook Inlet, Alaska, 1995-1999 (Harding et al. 2007) "Strong non-linear relationship between food density and colony attendance during chick-rearing"
Isle of May, Scotland, 2004 (Wanless et al. 2005) Provisioning rates were normal (4.5 feeds per chick per day) but energy values of fish were significantly lower than expected. Parental attendance (0.6%) was lowest on record and significantly lower than 1981-2003 values (10.9-23%). Percentage of chicks unattended at midday was 10.5% compared to < 0.01% mean for 1981-2003.
Isle of May, Scotland, 2007 (Ashbrook et al. 2008) Provisioning rates (2.32 feeds per chick per day) were significantly lower than 1981-2003 values. Parental attendance (0.8%) was significantly lower than 1981-2003 values (10.9-23%). Percentage of chicks unattended at midday was 13.1% compared to < 0.01% mean for 1981-2003.
Provisioning data was collected on Lundy for four years from 2012 to 2015, using a video recorder. Data collected for this study was from a site that has been recorded since 23rd June 2008 near St Philip's Stone on the west coast of Lundy (SS 130 462). This ledge is part of colony F4 on the Lundy Seabird Colony Register and was viewed from location Fe (SS 13180 46354, Price et al. 2008) . Up to 19 pairs of adults have been recorded on this ledge in any one year with a maximum of 15 chicks seen in any year. The ledge was mostly observed in the morning, but afternoon and evening sessions were also recorded. Only morning sessions were used for this study as provisioning rates can vary with the time of day and there were insufficient afternoon or evening session to give significant comparisons.
The camera was a Canon XL2 video camera which allows Canon SLR lenses to be used with an effective increase in focal length of around 7x. The site was first recorded with Sigma 70-300mm APO lens, and later with a Canon 100mm-400mm L lens. In 2015, a 1300mm focal length telescope was attached to the video camera which allowed one of the birds to be identified from its metal leg ring. The camera records Standard Definition 25p PAL video on to 60 minute miniDV tapes, in practice these tapes produce just over 62 minutes of usable video. The tapes were captured onto computer using Sony Video Capture 6.0e and edited using Sony Vegas Pro 8.0c. Corrupt video at the beginning or end of a recording was deleted. Only the first 60 minutes of video was used to calculate provisioning rates and parental attendance.
Focal sites were identified by the presence of an incubating or brooding adult, or by the presence of a pair of birds. Pairs were identified by mutual interactions; particularly by mutual preening, although adults also preen neighbouring birds (Lewis et al. 2007).Data collection:
Feeding ecology data used in this study was collected on the following days: 2012, 11 days between 11th June and 2nd July. 2013, 16 days between 19th June and 11th July. 2014, 11 days between 18th June and 8th July. 2015, 9 days between 15th June and 2nd July.Recording session
Data for each session was collected in two, one-hour tapes in the morning. Three types of data were transcribed from the tapes: arrival time, departure time, and whether an arriving adult was carrying a fish. Data was only recorded for adults known to have chicks.Provisioning Rate
A provisioning event was recorded as time an adult returned to the ledge with a fish. Adults with fish at the start of a recording session were not counted, nor were adults returning after 60 minutes 00 seconds of the tape had elapsed. The numbers of provisioning events, per chick, per recording session were tabulated:
Prov. 2012 2013 2014 2015 Events 0 59 51 58 51 1 33 51 45 47 2 5 15 7 10 3 1 1 2 2 4 1 1 1 0 N = 99 119 113 110 PR= 0.25 0.37 0.31 0.33Provisioning rates were significantly different between 2012 and 2013 (p < 0.1, Kolmogorov-Smirnov test)
Prov. 2012E 2012L 2013 2014 2015 Events 0 21 38 51 58 51 1 20 13 51 45 47 2 4 1 15 7 10 3 1 0 1 2 2 4 1 0 1 1 0 N = 47 52 119 113 110 PR= 0.37 0.14 0.37 0.31 0.33Provisioning rates were significantly lower in the later part of the 2012 breeding season, than in the early part of the 2012 (p < 0.05, Kolmogorov-Smirnov test) and in all subsequent seasons (2013 p < 0.01, 2014 p < 0.1, 2015 p < 0.05, Kolmogorov-Smirnov test) . Parental Attendance
Parental attendance was significantly lower in 2012 and 2015 compared to 2014.
Time 2012 2013 2014 2015 (mins.) -55 1 0 0 0 -25 1 0 0 0 5 62 64 51 65 15 15 20 13 17 25 7 12 14 13 35 2 5 8 3 45 4 4 9 3 55 2 2 3 2 65 1 2 5 2 75 1 3 2 2 85 2 2 3 1 95 0 1 2 1 105 1 2 3 0 115 0 2 0 1 N= 99 119 113 110 PA= 9.75% 15.19% 19.94% 12.29%
Parental attendance in the later part of 2012 was significantly lower than in 2014 (p < 0.01, Kolmogorov-Smirnov test) or in 2013 (p < 0.1, Kolmogorov-Smirnov test) .
Chicks were intentionally left unattended on two occassions while their parents went out to sea to feed. Both occured in the later half of 2012. In one of these cases, brooding duties had been shared with a neighbouring, chickless pair.
Time 2012E 2012L 2013 2014 2015 (mins.) -55 0 1 0 0 0 -25 0 1 0 0 0 5 25 37 64 51 65 15 8 7 20 13 17 25 4 3 12 14 13 35 2 0 5 8 3 45 4 0 4 9 3 55 1 1 2 3 2 65 0 1 2 5 2 75 1 0 3 2 2 85 2 0 2 3 1 95 0 0 1 2 1 105 0 1 2 3 0 115 0 0 2 0 1 N= 47 52 119 113 110 PA= 14.37% 5.55% 15.19% 19.94% 12.29%
Comparisons of data between Guillemots at different colonies should be taken with care. Birds from relatively close colonies can feed in different areas (Birkhead et al. 1986, Hatchwell et al. 1992). The physical characteristics of a particular ledge limit the number of birds present, and departing and returning birds can knock neighbours from the ledge. Comparisons of a particular sub-colony of a number of breeding seasons allow us to put unusual events into context. The 2012 event on Lundy showed lower than average parental attendance and provisioning rates as well as the only cases of intentional chick abandonment during the survey period.
During 'junk food' events, prey fish are availiable and provisioning rates are normal. Due to low energy values of the prey-fish, adult Guillemots spend more time away from the colony to obtain food for themselves and their chick. In extreme cases, adults will neglect their chick if their own survival is compromised. During 'low food availability' events, both provisioning rates and parental attendance are low. Adults will increase their foraging effort to provide for their chick. In extreme cases, adults will neglect their chick if their own survival is compromised.Possible causes of low food availability
The foraging areas of Lundy's Guillemots are not known. Comparisons between chick diet and feeding rates between Lundy and Skomer in 1985-1986 indicated little overlap between foraging areas (Hatchwell et al. 1992). The shortest feeding trip duration during this survey was 6m52s (unpublished data). This suggests that as least some foraging occurs close to Lundy.
Turbidity A multi-trophic survey of the Cook Inlet, Alaska found lowest fish densities in areas of high turbidity (Speckman et al. 2005). These were also the lowest density areas of Kittiwakes and Guillemots (Speckman 2004) Turbidity has been suggested as a cause of the auk 'wreck' during the winter 2013-2014 and also for a 'wreck' in the winter 1871-1872 (Lock 2014) High rainfall and flooding in the UK in 2012 may have increased sediment discharge from the Severn and other rivers that flow into the Bristol Channel. Comparisons between this study and archive satellite turbidity data for the Bristol Channel could be used to test this hypothesis.
Oceanic Fronts Oceanic fronts occur where water bodies of different of temperature and salinity meet. Lundy is situated at the edge of the geographical Bristol Channel. Variations in river flow can change the location of the front between these estuarine and oceanic waters. Locations of oceanic fronts in UK waters have been mapped at different temporal scales using composite satellite data (Miller 2009). These fronts have been suggested as foraging areas for seabirds (Miller 2011, Miller & Christodoulou 2014). Gannets foraging in the Celtic Sea were more like to search in predictable, seasonally persistant frontal zones (Scales et al. 2014). Comparisons between this study and the distance from Lundy to oceanic fronts could be used to test this hypothesis.
Over-fishing Over-fishing - (c.f. Sand eel fisheries North Sea)
Provisioning rates and parental attendance were significantly lower in the later half of the 2012 breeding season. This is consistent with low food availability, rather than a 'junk food' event. Further work could look for correlations with fish stocks, water quality, or the positions of tidal fronts in the Bristol Channel.