Mike Klangman Mike The influx of into Western Europe, 1995/96 R. Riddington, S. C.Votier and J. Steele

ABSTRACT Winter 1995/96 saw a huge influx of Common Redpolls Carduelis flammea and Arctic Redpolls C. hornemanni into Britain and Western Europe. Two forms of were involved, the nominate race of C. f. flammea (‘Mealy Redpoll’) and the Eurasian race of Arctic Redpoll C. h. exilipes, and there is no evidence of any arrivals from other than northern Eurasian populations. Three main influxes were recorded in the north and east of Britain. The first two, in early and mid November, were immediately preceded by or coincided with substantial movements through Utsira in southwest Norway, Heligoland off the northwest coast of Germany, and Falsterbo in southern Sweden; the last of the three arrivals occurred in early December and was associated with the onset of very cold weather, but not with large movements farther north and east. On the British east coast, numbers were the largest since 1975.

Heading drawing. Two Arctic Redpolls and one Common Redpoll.

© British 93: 59-67, February 2000 59 Riddington et al.: Influx of redpolls in 1995/96

The irruption was absorbed in Britain, and no significant movements were recorded through observatories on southern and western coasts. Return spring migration was heavy through southern North Sea sites (which experienced little of the autumn arrival), but was much less apparent farther north.

wo species of redpoll occur in in Western Europe. A combination of the Western Europe (fig. 1). Common factors described above forced birds to move TRedpoll Carduelis flammea is a west and south from their breeding grounds familiar breeding bird and/or winter visitor in Norway, north and central Sweden, throughout much of the region, while Arctic Finland and northern Russia. Redpolls were Redpoll C. hornemanni is a rare visitor, recorded in large numbers throughout many mostly in autumn and winter, from breeding parts of Western Europe, with unusually high grounds on the northern tundras of North numbers of both Mealy and Arctic Redpolls America, Europe and Asia. Within both in many areas. Similar invasions have taken species, there are a number of recognisable place in previous winters (notably 1984/85 subspecies (Knox 1988; Svensson 1992), and 1990/91), but, while these pronounced some of which are migratory while others influxes have received a great deal of atten- are predominantly sedentary (Clement et al. tion from birdwatchers, they have been 1993). For example, both the nominate race rather poorly documented. This short paper of Common Redpoll C. f. flammea (widely aims to describe the scale, timing and sub- referred to as ‘Mealy Redpoll’), which breeds species composition of the 1995/96 influx, from Scandinavia (except the south) and with particular reference to Britain and Finland eastwards across north Asia, and Ireland. across much of northern North America, and the Eurasian form exilipes of Arctic Redpoll Methods migrate each year in varying numbers to There are several problems associated with wintering areas south of their breeding documenting invasions of common species. grounds. Conversely, Common Redpolls of Large numbers of observations, from many the races cabaret and islandica (known different sites and observers, are difficult to respectively as ‘Lesser Redpoll’and ‘Icelandic collate and put into context. Indeed, such a Redpoll’) are typically much more sedentary: blanket approach often tells us more about the former breeds from Britain and Ireland, patterns of observer coverage than about southern Norway and southwest Sweden ecological processes. Added to this, the south to the Alps,while islandica is confined highly variable appearance and somewhat to Iceland. confusing of the ‘redpoll complex’ In northern breeding areas, food make the task even harder. Acknowledging resources often vary markedly from year to these difficulties, we attempted to devise a year.As a consequence, summers when pop- methodology to minimise such biases. For ulation levels are high (a result of good Common Redpolls, therefore, we used data breeding success following a winter of abun- only from those sites where observer cov- dant seed availability) tend to be succeeded erage was systematic during the 1995/96 by an autumn of widespread seed failure. winter, and also during the preceding and When this happens, unusually large south- subsequent winters. Daily log counts from ward migrations occur which may be termed bird observatories in Britain and Western ‘eruptions’. Later in the autumn, severe Europe were chosen for this purpose. weather may compound these pressures and Although not free from many of the biases act as an additional stimulus for southward that affect observational fieldwork, bird movement. In such years, populations which observatories are typically sited in important are normally sedentary may become migra- coastal migration areas, and are covered on tory,too. most days, in a similar way, by the same, In the autumn and winter of 1995/96, an experienced observers. These factors mean unprecedented arrival of redpolls occurred that, for comparative analysis both within

60 British Birds 93: 59-67, February 2000 Riddington et al.: Influx of redpolls in 1995/96

Fig. 1. Breeding ranges in Europe of Common Carduelis flammea and Arctic Redpolls C. hornemanni. Nos. 1-16 indicate location of bird observatories which provided data on Common Redpolls: 1 Fair Isle; 2 North Ronaldsay; 3 Copeland; 4 Cape Clear; 5 Calf of Man; 6 South Walney; 7 Bardsey; 8 Portland; 9 Spurn; 10 Gibraltar Point; 11 Sheringham; 12 Sandwich Bay; 13 Dungeness; 14 Utsira; 15 Heligoland; 16 Falsterbo. and between seasons, our data are as accu- tively). Thirteen British and Irish observato- rate and reliable as possible without a pre- ries responded to our request for data, as did determined methodology. In essence, we three foreign stations, making 16 study sites have sacrificed greater geographical cov- in total (fig. 1). erage in favour of more systematic and rig- In addition, all records of Arctic Redpolls orous data from a smaller number of sites. accepted by the British Birds Rarities Com- All bird observatories in Britain and mittee (BBRC) for the winter 1995/96 (see Ireland, together with four others in Western Rogers et al. 1996-1999) were analysed. Europe, were sent a series of simple forms Although these data are not derived in the on which to record Arctic and Common same systematic manner as those for Redpoll sightings on a daily basis from Sep- Common Redpolls, they nevertheless tember 1995 to May 1996. Recorders were provide an instructive insight into the distri- asked to differentiate recognisable sub- bution of this tricky species (see Votier et al. species of redpoll where possible, and also 2000). to distinguish between those days when Results there was observer coverage of the site but To investigate the magnitude and timing of no redpolls were seen and those when no the arrival in 1995/96, and to compare it observers were present. Finally, they were with the winters immediately before and asked to provide a monthly summary of after, we analysed data from six sites in redpoll counts in the previous and subse- detail.These encompassed two from the east quent winters (1994/95 and 1996/97, respec- coast of Scotland (Fair Isle, Shetland, and

British Birds 93: 59-67, February 2000 61 Riddington et al.: Influx of redpolls in 1995/96 Fig. 2i. Monthly mean daily counts of Common Redpolls,Fig. 2ii.Utsira. Monthly mean daily counts of Common Redpolls, Fair Isle.

200 2i. Utsira, south 20 2ii. Fair Isle, Norway Shetland

150 15

100 10

50 5

0 0 S O N D J F MAM - S O N D J F MAM - S O N D J F MAM S O N D J F MAM - S O N D J F MAM - S O N D J F MAM Fig. 2iii. Monthly mean1994/95 daily counts 1995/96 of Common 1996/97 Redpolls, Fig.North 2iv. Ronaldsay. Monthly1994/95 mean daily counts1995/96 of Common 1996/97 Redpolls, Spurn.

20 2iii. North 40 2iv. Spurn, East Ronaldsay, Yorkshire Orkney 15 30

10 20

5 10

0 0 S O N D J F MAM - S O N D J F MAM - S O N D J F MAM S O N D J F MAM - S O N D J F MAM - S O N D J F MAM Fig. 2v. Monthly mean1994/95 daily counts 1995/96 of Common 1996/97 Redpolls,Fig. Sheringham. 2vi. Monthly mean1994/95 daily counts 1995/96 of Common 1996/97 Redpolls, Heligoland.

50 2v. Sheringham, 80 2vi. Heligoland, Norfolk off northwest 40 Germany 60

30 40 20

20 10

0 0 S O N D J F MAM - S O N D J F MAM - S O N D J F MAM S O N D J F M A M 1994/95 1995/96 1996/97 1995/96 Fig. 2. Mean daily counts of Common Redpolls Carduelis flammea at six sites in Western Europe for each month September-May in the three winters 1994/95-1996/97. For Heligoland (2vi), data are available only for 1995/96.

North Ronaldsay, Orkney), two from the east To analyse magnitude and timing of the coast of England (Spurn, East Yorkshire, and influx, counts of all Common Redpoll sub- Sheringham, Norfolk), one from the west species were pooled. At most sites, the coast of Norway (Utsira) and one from the majority of individuals were not assigned to northwest coast of Germany (Heligoland). subspecies with complete confidence.This is All of these sites are in areas where the quite understandable given the difficulties of redpoll invasion was pronounced. Summary identification if views are brief, as, for data are provided for remaining sites, which example, at Falsterbo in south Sweden and at either are in areas that were largely unaf- Sheringham, where a large proportion of the fected by the influx or did not have suffi- records are of birds passing over on active cient observer coverage to permit more migration. Consequently, we include all detailed analysis. ‘unknown’ redpolls (i.e. those not assigned

62 British Birds 93: 59-67, February 2000 Fig. 3. Autumn counts of Common Redpolls at Falsterbo,Riddington 1973-1996. et al.: Influx of redpolls in 1995/96

10000 30 days and 30 redpolls present for one day.Indeed, 8000 the whole concept of ‘bird- days’ has been widely criti- 6000 cised as a form of analysis, but for the data available 4000 to us this is the most mean- ingful way to present and

Count (total bird-days) 2000 compare differences bet- ween sites and years. 0 The data in fig. 2 illus- 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 trate dramatically the mag- Year nitude of the winter Fig. 3. Autumn counts of Common Redpolls Carduelis flammea at 1995/96 invasion. For all of Falsterbo, south Sweden, 1973-96. the British sites, plus Utsira, mean daily counts during to race) in these figures, but exclude all the peak month of the invasion were at least known Arctic Redpolls.While this may intro- one and often two orders of magnitude duce a small bias into our results, in that greater than those during the same month in some fly-over ‘unknowns’ may relate to the two non-invasion winters. The quantity Arctics, this is likely to be negligible and to of Common Redpolls arriving in Western have no effect on the general patterns Europe meant that record numbers were reg- shown here. istered at many sites. Fig. 3 presents 24 years of redpoll counts Scale and timing of the influx in autumn (11th August to 20th November) To assess the magnitude of the influx, we cal- at Falsterbo, where the daily visible-migration culated, for the six sites listed above, mean count has been carried out in exactly the daily counts of Common Redpolls for each same way and, for the most part, by the same month from September to May in each of the observer during that period. This clearly three winters 1994/95-1996/97 (for identifies 1995/96 as a major invasion year. Heligoland, data were available only for Numbers at Utsira were described as ‘extra- 1995/96). These equate to ‘mean bird-days’ ordinary’, and the same is true for some per month, but are corrected for the number British sites. For example, at Spurn, only of days on which a site was manned (fig. 2). three autumns since 1975 have produced This method of handling the data does not daily totals of 20 or more Mealy Redpolls (up differentiate between new arrivals and long- to 40 in October 1975, up to 46 in stayers, i.e. between one redpoll present for November 1984, and up to 94 in November

Table 1. Summary of Common Redpoll Carduelis flammea records at nine bird observatories in southern and western Britain and Ireland in winter 1995/96.

Observatory Peak Mean Peak % Mealy month daily count daily count C. f. (date) flammea

Gibraltar Point, Lincs. January 3.85 27 (10th) 29.0% Sandwich Bay, Kent December 1.45 8 (5th, 17th) 24.4% Dungeness, Kent November 3.03 21 (4th) 0 Portland, Dorset October 1.00 15 (18th) 0 Bardsey, Gwynedd May 2.03 10 (6th,17th) 0 South Walney, Cumbria December 0.39 5 (7th) 25.0% Calf of Man, I. of Man May 0.48 6 (7th) 0 Copeland, Co. Down October 0.89 3 (29th) 0 Cape Clear, Co. Cork October 2.26 40 (21st) unknown

British Birds 93: 59-67, February 2000 63 Riddington et al.: Influx of redpolls in 1995/96 Fig. 4i. Daily counts of Common Redpolls, Utsira, autumn 1995. Fig. 4ii. Daily counts of Common Redpolls, Helgoland, autumn 1995. 3000 4i. Utsira, south 600 4ii. Heligoland, Norway 500 off northwest Germany 2000 400

daily daily 300 count count

1000 200

100

0 0 0 10 20 30 40 50 0 10 20 30 40 50 Date (1 = 26 October) Fig. 4iv. Daily counts of CommonDate Redpolls, (1 = 26 October)North Ronaldsay, autumn 1995. Fig. 4iii. Daily counts of Common Redpolls, Fair Isle, autumn 1995.

50 4iii. Fair Isle, Shetland 100 4iv. North Ronaldsay, Orkney

40 80

30 daily 60 daily count count 20 40

10 20

0 0 0 10 20 30 40 50 0 10 20 30 40 50 Date (1 = 26 October) Fig. 4v. Daily counts ofDate Common (1 = 26 Redpolls,October) Spurn, autumn 1995.

300 4v. Spurn, East Yorkshire

200 daily count

100

Fig. 4. Daily counts of Common Redpolls 0 0 10 20 30 40 50 Carduelis flammea at five sites in autumn 1995. Date (1 = 26 October)

1985); compare this with the peak count in 7th, when a further 200 were recorded. 1995/96 of 250 Mealies on 12th November. The synchrony of these arrivals in the Numbers of Common Redpolls at observato- Northern Isles and Yorkshire clearly indicates ries in southern and western Britain, that redpolls were reaching Britain on however, were very low, and the 1995/96 a broad front. The small numbers in invasion largely petered out before reaching November at coastal sites south of Yorkshire these areas (table 1). Indeed, both Portland, (including both Gibraltar Point, Lincolnshire, Dorset, and Dungeness, Kent, reported that and Sheringham) suggest, however, that red- the winter of 1995/96 was a very poor one polls in the southern half of Britain had for migrant redpolls. already made landfall farther north, and that, The histograms in fig. 2 present a crude as the birds spread south and west, their view of the timing of the influx at the sites numbers were comparatively less concen- concerned, but, in order to provide a more trated. detailed picture of arrivals in autumn, we On the Continent, the major arrivals of plotted daily counts for late October to mid Common Redpolls were in late October/ December 1995 for five sites (fig. 4). In the early November. At Utsira, 600 on 29th Northern Isles, there were two main arrivals October were followed by 2,500 on 4th in November, on 4th and on 12th-14th; November; smaller movements of 200-300 counts on both Fair Isle and North Ronaldsay on 11th-12th November and 250 on 16th were then modest until December, when perhaps correspond with mid-November there were further arrivals on 4th and 8th.At arrivals in Britain.At Heligoland, too, the first Spurn, there was only a very small arrival of large counts were on 29th-30th October redpolls on 4th November, but a pronounced (138 on 30th), while the main arrival, of 503, influx on 11th-12th (with 250 on 12th), after occurred on 1st November; smaller move- which numbers decreased through the ments followed in mid November, with 103 month; in December, the main arrival was on on 10th and 105 on 14th.

64 British Birds 93: 59-67, February 2000 Riddington et al.: Influx of redpolls in 1995/96

Table 2. Percentages of Common Redpoll Carduelis flammea of nominate race (‘Mealy Redpoll’) recorded during autumn 1995. Note that all figures for Falsterbo refer to ringing data only.

Observatory No. of ‘bird-days’ % redpolls % Mealy 25th Oct.-15th Dec. identified C. f. flammea

Fair Isle, Shetland 486 100% 100% Spurn, East Yorkshire 1,881 100% 99.7% Utsira, south Norway 6,247 100% 99.8% Falsterbo, south Sweden 1,091 99.5% 63.5% (and 36.4% Lesser C. f. cabaret)

It is interesting that, at Sheringham, example, recorded its very first Mealy Red- counts were modest in November and polls (three singles) during December 1995. December, and the large numbers were of About a quarter of redpoll records at both birds returning to northern breeding Gibraltar Point and Sandwich Bay, Kent, grounds in spring (fig. 2v). That such move- during their peak months (January and ments were less marked farther north, at December, respectively) referred to Mealy least in Britain, perhaps indicates that north- Redpolls, this being a higher proportion than bound spring migrants opted for a shorter normal. crossing of the North Sea, from southeast England. Arctic Redpolls A total of 431 records of Arctic Redpolls has Subspecies composition been accepted by the BBRC for the period At three of the six sites studied in detail, 2nd November 1995 to 29th May 1996 together with Falsterbo in Sweden, virtually (Rogers et al. 1996-1999), completely all Common Redpolls were identified to sub- eclipsing the previous highest total, of 90 in species during the peak period of late winter 1990/91 (Rogers et al. 1991-1995). October to mid December (table 2).At these During the 1995/96 winter, there were no sites, the vast majority were Mealy Redpolls. accepted records of the nominate race of In addition, ‘almost all’ Common Redpolls Arctic, and it seems likely that most, if not all, seen at North Ronaldsay were also Mealy individuals were of the Eurasian race Redpolls (A. Duncan in litt.). Very small exilipes. numbers of Lesser Redpolls were reported Monthly totals of Arctic Redpolls are listed from northern sites with the exception of in table 3. There was an extremely broad Falsterbo, where over one-third of redpolls scatter of records, involving a total of 38 ringed were Lessers. counties across Britain, although with a It is interesting also that there were no marked bias towards northern and eastern confirmed sightings of Icelandic or Green- parts, and evidence of a general southward land individuals C. f. islandica/rostrata, and westward spread as the winter pro- though a claim of an Icelandic Redpoll gressed. Large arrivals in Shetland and trapped on Utsira was not accepted by the Orkney during November and December Norwegian rarities committee (G. Mobakken contrast with an absence of sightings there in litt.). This is in marked contrast to the during the rest of the winter, perhaps sug- autumn of 1997, at least in northern Britain, gesting a lack of suitable wintering habitat in where good numbers of these subspecies the Northern Isles. Early observations of indi- were recorded (Reid & Riddington 1999). viduals on the east coast of Britain were This suggests that the factors stimulating the widespread, as the birds arrived on a broad invasion originated in northern Eurasia and front. By December, they had penetrated the did not affect populations of redpolls in central areas of Britain, with records from Greenland and Iceland. western counties and, remarkably, even Small numbers of Mealy Redpolls were Scilly.The broad scatter of records through recorded at sites where they are normally January-March is probably due to the dis- extremely scarce. South Walney, Cumbria, for covery of large wintering flocks of redpolls;

British Birds 93: 59-67, February 2000 65 Riddington et al.: Influx of redpolls in 1995/96

Table 3. Monthly totals of Arctic Redpolls Carduelis hornemanni in Britain, November 1995 to May 1996.

County Nov Dec Jan Feb Mar Apr May Total

Angus 00240006 Caernarfon 00000101 Cambridgeshire 00300003 Cheshire 01100103 Cleveland 120073013 Cumbria 01100002 Denbigh 00002002 Derbyshire 094152021 Dorset 00020002 Essex 05001006 Fife 00020002 Gloucestershire 00010001 Greater London 00010001 Greater Manchester 00025108 Hertfordshire 00010001 Lancashire 02301006 Leicestershire 05000005 Lincolnshire 1 11 1 110015 Lothian 00100001 Moray & Nairn 06000006 Norfolk 4 12 16 36 7 0 0 75 Northeast Scotland 3 2 55 020062 Northamptonshire 01300004 Northumberland 170110010 Nottinghamshire 01110205 Orkney 13270000040 Outer Hebrides 00100012 Pembrokeshire 00010001 Perthshire 01210004 Scilly 01000012 Shetland 28 11 0 000241 Staffordshire 040460014 Suffolk 052030010 Worcestershire 00400004 Yorkshire, East 35000019 Yorkshire, North 12004108 Yorkshire, South 081920020 Yorkshire,West 153240015

TOTALS 56 134 104 70 51 11 5 431 many apparently new arrivals may simply prompted an irruption of the latter species reflect an increase in observer awareness also stimulated Arctics to move. and more interest in searching for the rarer species, rather than genuine new arrivals. Discussion The five Arctic Redpolls recorded during The autumn and winter of 1995/96 wit- May represent a tiny fraction of the many nessed one of the largest invasions of red- birds moving north, often too rapidly to be polls in the last 25 years, and one that was identified with certainty. probably more widespread than the irrup- More specific arrival times match closely tions of 1984/85 and 1990/91. Mealy Red- those of Common Redpolls, since virtually all polls formed the bulk of the invading birds, Arctic Redpolls were recorded within flocks although there were also large numbers of of that species. Clearly, those conditions that Arctic Redpolls, and the pattern of the move-

66 British Birds 93: 59-67, February 2000 Riddington et al.: Influx of redpolls in 1995/96 ments suggests that they originated entirely November-December. from the northeast. Among the Mealy Redpolls was a small Arrivals on the east coast of Britain but significant proportion of Arctic Redpolls, appeared to occur in three waves, in early with record numbers being noted by the end and mid November and in early December, of the winter. The percentage of Arctics is but the scale of the arrival varied at different impossible to determine, but the numbers times according to the site involved.The first reported are almost certainly a considerable wave was closely correlated with very large under-estimate of the true figures involved. movements through Scandinavia and Intense scrutiny of three large redpoll flocks Heligoland, and, although numbers moving which were relatively easy to observe (at on the Continent remained high through the Garlogie, Northeast Scotland, and at Langham period of the second arrival, the large totals and Cromer in Norfolk) revealed the true in eastern Britain in mid November may have extent of plumage variation within and been partly related to easterly weather con- between the two species (see Votier et al. ditions during that period. By the time of the 2000); it also resulted in much collective third wave, which was apparent across a head-scratching, and a considerable number very broad front in eastern Britain, passage of submitted and acceptable records. There through Utsira and Heligoland had all but were many other large flocks which did not dried up, suggesting that the severe weather receive anything like the same attention, but and easterly winds in early December may which must also have contained Arctic Red- have contributed in large part to the major polls. Numerous individual and some collec- arrival in Britain, or at least to the coastal tive lessons were learned by those who grounding of migrant birds. The invasion observed such flocks. Among the most itself appears to have been absorbed across important of these were the need for excel- Britain where there were sufficient food sup- lent views to confirm the key identification plies to support the immigrant birds, with features, and also that, in good views, some records from the south and west coasts relatively dull birds were identifiable as being very sparse. Return passage was noted Arctics, while others which may initially through Norfolk, particularly in March, in have looked ‘promising’ were clearly Mealy stark contrast to the lack of any significant Redpolls (see Votier et al. 2000). arrival there in the peak invasion period of

Acknowledgments References We owe a great debt of thanks to all the birders who Clement, P., Harris, A., & Davis, J. 1993. and carefully observed and counted redpolls at the study Sparrows: an identification guide. London. sites during the three winters, and particularly those Knox,A. G. 1988.The taxonomy of redpolls. Ardea 76: who took time to extract data and respond to our 1-26. request: David Anning (Bardsey Bird Observatory),Tim Reid, J., & Riddington, R. 1999. Identification of Bagworth (Calf of Man Bird Observatory), Martin Cade Greenland and Iceland Redpolls. Dutch Birding 20: (Portland Bird Observatory), John Cudworth (Spurn 261-271. Bird Observatory), Alison Duncan (North Ronaldsay Rogers, M. J., & the Rarities Committee. 1991-1999. Bird Observatory), Jochen Dierschke (Heligoland Bird Report on rare birds in Great Britain in 1990 Observatory), Fair Isle Bird Observatory, Bill Makin …1998. Brit. Birds 84: 449-505; 85: 507-554; 86: 447- (South Walney Bird Observatory), Neville McKee 540; 87: 503-571; 88: 493-558; 89: 481-531; 90: 453- (Copeland Bird Observatory), Geir Mobakken (Utsira 522; 91: 455-517; 92: 554-609. Bird Observatory), Ken Preston (Cape Clear Bird Svensson, L. 1992. Identification Guide to European Observatory), Sandwich Bay Bird Observatory, Kevin . 4th edn. Stockholm. Shepherd, Mike Young-Powell & Tim Wright Votier, S. C., Steele, J., Shaw, K. D., & Stoddart, A. M. (Sheringham Bird Observatory), Göran Walinder 2000. Arctic Redpoll Carduelis hornemanni (Falsterbo Bird Observatory), David Walker exilipes: an identification review based on the (Dungeness Bird Observatory) and Kev Wilson 1995/96 influx. Brit. Birds 93: 68-84. (Gibraltar Point Bird Observatory).

Dr Roger Riddington, Chapel Cottage, Dunrossness, Shetland ZE2 9JH S. C.Votier, Division of Environmental and Evolutionary Biology, IBLS, Graham Kerr Building, University of Glasgow, Glasgow G12 8QQ Dr J. Steele, 16 Oaklands, Gosforth, Newcastle-upon-Tyne NE3 4YQ

British Birds 93: 59-67, February 2000 67