Two Goldcrest Regulus regulus recoveries in conjunction

This note advocates joint treatment of Goldcrest recoveries from different countries: the recovery picture obtained with recoveries from one single country is always fragmentary and inadequate. Birds from Norway and Sweden may fly against easterly winds when leaving the breeding areas; birds from Lithuania and Poland may fly before the same winds when correcting for earlier dislocation. It is not until e.g. Norwegian, Swedish, Lithuanian and Polish recoveries are joined, that the full picture of any particular season can emerge. Meteorological data should always be added and considered: pooling and mapping of rather anonymous recoveries from different years with different weather conditions is not really informative. The argument is supported by a couple of recoveries from Ljunghusen on the Falsterbo peninsula, S. Sweden. Goldcrests occurring here in the autumn of 2001 must have flown against easterly winds, prevailing in the preceding weeks, but it seems as if they turned in the south Baltic area and flew before the wind to Jutland. One bird covered 288 kms in a single night, flying before southeasterly winds of force 3, this is the longest distance so far covered in a single night by a Goldcrest in the south Baltic area. Åkesson (1993) suggests a "use of the coastline to compensate for wind drift"; the Swedish west coast may be the backbone of a corridor, where Norwegian migrants regularly "wind drive", - by easterly or northwesterly winds - to S. Scania. Or maybe better: the coastline acts as a guidance in any kind of correction, be it with diagonal headwinds, sidewinds or diagonal tailwinds. On the other hand the same idea is completely rejected in Zehnder et al. (2001), where Åkesson is one of the co-authors; conclusions obviously are liable to drift by the wind as well.

Introduction

In the pinewood area of Ljunghusen (55º 24' N, 12º 55' E) on the Falsterbo peninsula, S. Sweden, Goldcrests Regulus regulus are very numerous on calm mornings in October, and they are easy to catch with the help of a tape lure. I do not catch the species on a regular basis, but in some years I make a limited effort, ringing 1 - 2000 individuals. The birds are caught in a single mist-net of 10 m length, a total of 18.000 has been ringed in that way. In years when I am active or stimulated by Goldcrest abundance, I may take a little more than 5 % of the annual Swedish Goldcrest catch with one single net in my backyard. If the tape plays song, the catch will be biased towards males, at least where Goldcrests are involved - but the sex ratio on the Falsterbo peninsula is biased from the beginning, almost two males in every female when birds are caught without lure (Ljunghusen; neighbouring Falsterbo had a male : female ratio 55 : 45 between 1970 and 1978, Karlsson 1980). The major advantage with tape lure catch is that the time expenditure may be reduced at will (Goldcrest ringing is extremely boring, hard labour); when a good sample has been taken the tape recorder is switched off and no more birds will be caught.
Goldcrest recoveries in the Baltic area are rather predictable, when you have seen twenty you have seen them all: the birds seem to fly by the wind - or be driven by the wind if it is too strong, they stick to a certain N - S corridor in calm weather, and correct westwards in anticyclonic weather with winds from the east. Up till and including 1999 Hammarö Bird Station (new adress) at 59º 15' N, 13º 30'E in central Sweden had 14 direct autumn recoveries of Goldcrests from Russia (Rybachi), Latvia, Lithuania and Poland, all these birds flying SE, and this is not a "correct" direction of movement for the species in autumn. Other birds reach the south Baltic sites from Swedish catching stations at their own longitude; Stora Fjäderägg at 63º 49' N, 21º 00' E and Haparanda Sandskär at 65º 34' N, 23º 46' E are two examples.
The recovery rate of ringed Goldcrests is not impressive; even with a Swedish annual catch of 20.000 birds there will be no more than 20 recoveries per annum, all but a few of them retraps between catching stations in the Baltic area, distances as a rule 200 - 1000 kms, time elapsed 1 - 3 weeks. This is recoveries en route, even on early route, and what they show is the migration pattern in an area where two or three lows may pass each week, bringing strong northwesterly winds in their wake - or where a high is parked over Norway or Sweden and easterly winds prevail through most of October. In recent years there has been a third option: warm weather, caused by lows following very northerly trajectories. Here strong southwesterly or westerly winds and warm nights may cause extreme delay in Goldcrest migration; in the autumn of 2000 the species culminated in SW Scania with an unprecedented outburst on 3 November, and there were still migrating Goldcrests by mid-November; in 2001 a small dawn arrival fell as late as November 13th.
Given the - not unpredictable - given the arbitrary weather, Goldcrests have every reason to try to make the most of the predominant wind forces and wind directions in any autumn. They should be able to migrate in anticyclonic weather with night fog or easterly winds force 4 - 5 lasting for twenty days. And they should be able to evacuate e.g. Norway in an autumn with twenty low passages and periods of NW winds force 4 - 5 lasting at most 3 - 4 hours. These two situations must involve completely different options: in one case migration before the wind seems to be the best way to part with the breeding area, in the other some sort of zigzaging against the wind must be necessary. In the first case, there is still a choice to be made: flight before strong westerly winds will increase the distance to the goal area of e.g. Norwegian birds. But two steps forward and one aside (200 kms to the south and 100 kms to the east) seem to be quite acceptable. How about one step forward and two steps aside? Judging from recoveries this is not an option favoured by Goldcrests, but they may be overtaken by the wind and confronted with it. In general I think that it is quite acceptable for Goldcrests to fly with winds from NNW - and they will make good progress with this direction. And, judging from arrivals on the Falsterbo peninsula, many Goldcrests may even depart with winds from WNW, maybe fly on a southwesterly course and make good progress in a southeasterly direction that way. Later on, if they find themselves too dislocated, they correct with easterly winds. This seems to be the way many Norwegian Goldcrests are brought to the Falsterbo peninsula in autumn. But can this picture be sustained when confronted with recoveries?

A situation in October 2001

In the morning of 23.10.01 I stood in my garden by about 6 o'clock, listening to wind and birds. There was a strong whistling from the crowns of the trees, an easterly wind, force 4 (5.5 - 8 m/s) had blown all night, and no birds could be heard. I pulled up a net and switched on a tape player with Goldcrest song, expecting nothing but the local territory bird, which is sedentary, but always mingles with the migrants. What happened took me by complete surprise: some thirty seconds later I had the net completely filled with Goldcrests and had to prop up the net poles. The density of birds surrounding the net was immense, they clung to every branch like locusts of Egypt; there must have been tens of thousands of Goldcrests in the immediate vicinity. It took me an hour and twenty minutes to free and ring 170 birds, standing before the net, with rings between my teeth. The tape player had been switched off after thirty seconds and was never switched on again that day, the total catch of the day was 198 Goldcrests, 139 of them (= 70%) males. By 10h I took a stray bird with Stavanger ring, indicating the ultimate origin of the night's migration, and later on I had a remarkable recovery of one of the dawn birds, the longest distance ever covered by a Goldcrest in a single night (both recoveries shown in Fig. 1):

1. Stavanger AH 2090 V 1y female, Ljunghusen 23.10.01 10h - ^. Revtangen (58.45 N, 05.30 E), Rogaland, NORWAY 14.10.01, 12h. 584 km SE, 9 days. (584 km "as the bird flies"! But this bird and other birds from Norway must have flown 700, 800 kms before reaching the Falsterbo peninsula.)
2. Stockholm RH 4774 ^. 1y male, Ljunghusen 23.10.01 07h - x Saedden (55.30 N, 8.23 E), Esbjerg, Jylland, DENMARK 24.10.01 11h. 288 km W, 28 hours. (found by a ringer, Erik Hansen).

Fig. 1. Two Goldcrest recoveries in the autumn 2001: Revtangen - Ljunghusen and Ljunghusen - Esbjerg. The connection Hammarön - Rybachi has been drawn; it is parallel to the line connecting Revtangen with the Falsterbo peninsula. This angle has something to do with the the standard direction of Goldcrests and the maximum tailwind/sidewind in which they migrate.

North of Oslo the wind had blown around the clock in the preceding night, force 0, at Gothenburg direction E, force 2 and at Malmö direction E, from force 3 increasing to 4. A typical anticyclonic situation, rather common when highs establish themselves in October. Of course the Goldcrests had not arrived from Norway this particular night - but from what proximate area did they arrive, and what had brought them to the Falsterbo peninsula? A closer look at the weather (Fig. 2) reveals, that there had been no pronounced "drift" situation with strong northwesterly winds between 14 and 23 October, easterly winds force 2 to 4 had prevailed during the whole period, the center of the high in four nights located to the north of Oslo. There is one calm night, 17 October, but as the Stavanger bird moved southwards it must have flown against the wind for several nights, making rather slow progress. Given this meteorological prehistory I do not think, that wind drift was involved in the night 22-23 either, rather the birds had followed the Scanian west-coast, flying on a southeasterly course and effectively migrating south. Then, reaching the latitude of the Falsterbo peninsula, one bird (and, if it was representative: the whole cohort) turned and flew before the wind in one single night, making extremely good progress and returning to the right "corridor" of Norwegian birds. There is even some "logic" in the resultant direction if we assume, that the bird flew SW before a SE-wind. With 15 km/h wind velocity and 25 km/h air speed of the bird, the westward resultant will be appr. 29 km/h, allowing the bird to cover 290 km in 10 hours, from 19h in the evening to 05h in the morning. In my opinion, these two recoveries amply illustrate the complexity of the options facing Goldcrests migrating in the Baltic area. "Drift" need not be involved with easterly winds, and drift may be involved with easterly winds.



Fig. 2. Winds at 03h between 15.10.01 and 24.10.01. The centre of a high was situated near Oslo in at least four nights, easterly winds prevailed in Scania. Beaufort scale 3 is 3.4 - 5.4 m/s, 4 5.5 - 7.9 m/s. From the German Wetter online.

There are similar recoveries from earlier years, showing that Norwegian birds occur in S. Scania, and that birds from e.g. Poland may "correct" northwestwards:

1. Stavanger A26546 V 1y female, Ljunghusen 24.10.88 11h - ^. Jomfruland (58.52 N, 09.36 E), Kragerö, Telemark, NORWAY 16.10.88. 435 km SSE, 8 days.
2. Oslo A10804 V 1y male, Ljunghusen 31.10.90 14h - ^. Mölen (58.58 N, 09.49 E), Brunlanes, Vestfold, NORWAY 18.10.90. 435 km SSE, 13 days.
3. Stavanger L26923 V 1y female, Ljunghusen 15.10.93 11h - ^. Hodne (58.47 N, 05.33 E), Klepp, Rogaland, NORWAY 30.9.93. 587 km SE, 15 days.
4. Gdansk LB94673 V 1y female, Ljunghusen 16.10.90 11h - ^. Stacja "Bukowo-Kopan" (54.28 N, 16.25 E), Koszalin, POLAND 9.10.90. 250 km NW, 7 days.

I have no weather data for the situations preceding these recoveries, so I do not know what factors may have been involved. They are put here as complementary indications of the apparent "hypotenuse-cathete"-migration of Goldcrests in the Baltic area - both legs by all likelihood flown against the wind as often as before the wind. With easterly winds force 4 or 5 the arrival of Goldcrests in Ljunghusen is very hard to predict: there may be huge amounts of birds or none at all. The prehistory of any particular contingent is crucial here; in some years there is no need for correction and what we see is ordinary arrival from areas to the north of Falsterbo, often arrival by delayed birds under strong coercion, in other years the birds have been severely dislocated and correct even from areas to the south of Falsterbo, like the above bird from Poland.

Discussion

Alerstam (1982) gives an introduction to the concept "pseudodrift". It is caused by the fact that different populations are to a greater or less extent favoured by a particular wind direction; the most favoured groups will migrate, and compensate in relation to their specific goal-area, other groups will be less active. In passing he remarks: "Is it possible, that birds in some regions and weather situations do compensate for wind drift but allow themselves to be wind-driven in others?"
This remark is inspired, anticipatory in a fruitful sense, and I believe this is what can be expected in the case of the Goldcrest. Norwegian birds are quite capable of crossing Skagerak; Cramp et al. (1992) outline a crossing from southernmost Norway to Jutland, the birds flying on a SSE course, and Grenmyr (1997) in turn assumes that birds reaching the British Isles are to some extent such migrants, dislocated by adverse weather in the Skagerak/North Sea area. But it is obvious from recoveries, that many Norwegian Goldcrests - like Norwegian Tree Pipits Anthus trivialis (To: Viktstudier på trädpiplärkor Anthus trivialis under höststräcket i sydvästra Skåne) and Willow Warblers Phylloscopus trochilus - shun the dangerous Skagerak passage, or perhaps rather: the perspective of veering winds forcing them to cross the 400 km wide water area separating Norway from Scotland. I believe that the overall weather pattern of a particular autumn decides the ratio between S- and SE-directed (initially E-directed) migration from Norway, but in any autumn substantial amounts from even the southwesternmost counties of Norway (Hordaland, Rogaland, Vest-Agder) will migrate by way of the Swedish west coast and the Sound area. Still, the populations involved do not decide prematurely; I have a feeling that there is always a delay of 5 - 10 days relative to the normal median when Norwegian Yellow Wagtails Motacilla flava or Goldcrests flood the Scanian migration scene. They have waited for favourable or low-risk winds, waited in vain, and finally they have flown before or against the wind to Sweden, working themselves down the Swedish west coast. Such departure is catastrophic, all birds sharing the same preparedness for migration are evacuated simultaneously and arrive as a migration "wave". And here are no considerations creating pseudodrift; the birds involved are running out of time and work themselves southwards in all kinds of weather. Then, suddenly, the longed-for high or favourable wind direction appears and all migrational worries are done with. Migration before the wind may be an option again, e.g. with easterly winds from Scania.

Did Alerstam have material (e.g. from radar studies) indicating that small passerines will fly against easterly winds under certain circumstances, or was his remark just a stroke of luck?¹ The theoretical framework he has spun around migration tended to be one-sided and mechanistic in the early versions, still many researchers allowed themselves to be guided by it like blind men. I have read Gezelius & Hedenström (1988) ten times, their guiding-star obviously is Alerstam (1979): "Wind as a selective agent in bird migration" - and the statement of the paper is practically zero. (I do not hold this against the instigators; it is always better to try to think than just bluntly presenting a set of data. And the remark three years later - that I never noticed before - makes Alerstam seem less rigid, almost wise). Ottenby doesn't receive many migrants from some "far east" region, what easterly winds generally achieve in the Baltic is to bring havoc in the normal southbound migration over the Baltic from Sweden and Finland. In the same way as they are a nuisance to Robins or Goldcrests migrating from Norway to Denmark. Fifty years of ringing recoveries from the station demonstrate this, leaving no room for doubt. There are eastern rarities, Great Tits flying before the wind, but the non-Scandinavian element among Robins or Goldcrests at Ottenby is quite insignificant. The myth of "eastern origin" is cherished among bird-watchers, they need it in order to make their world exotic, but in the case of Ottenby it means little more than birds of Finnish origin. And I find it quite likely, that small passerines moving southwards along the coastline of Öland in easterly winds will appear to be wind-driven when caught at Ottenby in the early morning hours. Åkesson (1993) makes an interesting comment to this particular situation: nocturnal passerine migrants... ...may use the coastline to compensate for wind drift. In my interpretation she hints, that it should be easier to compensate for drift - or maybe better: compensate for any kind of adverse wind - when there is a landmark of some kind, e.g a coastline, and that real, unfavourable drift ocurs over vast, rather anonymous areas - or at high altitudes. Unfortunately she does not elaborate this idea with more material. The same thing is indicated in at least one paper in Optimal Migration (J. Av. Biol. 29: 337 - 636), I lost the reference but will find it again. The Swedish west coast could be the backbone of a corridor, where Norwegian passerine migrants can compensate for wind-drift or adverse headwind (from NW or E) on their course to S Scania.
In Zehnder et al. (2001) the whole idea is rejected, however; at Falsterbo night migrants flew on a constant course, independant of wind direction, and therefore where drifted by wind. The mean course angle deviated from the angle of the coast-line, this was interpreted as indication that topography is not used to compensate for wind drift. As far as I can see, the authors have been rash here, making sweeping statements of a specific Falsterbo pattern; the whole matter cannot and should not be decided at Falsterbo, but at extremely "neutral", in-between sites.
1. Goldcrests are caught in (or collide with) wader-nets over water only when flying against the wind; at the spits of Skanör their flight altitude is little more than 1 - 1 1/2 m above the sea level with force 3 headwinds. Such migration must be very difficult to record by radar. Wheatears are even worse, flying like Storm Petrels just above the water surface when moving against force 4 - 5 winds (beacon at the Falsterbo Canal, spits of Skanör), and Blackbirds arriving from north in sudden headwinds of force 5 in January were caught in the lowest shelf of wader nets set for Iceland Redshanks. Even Goldcrests arriving in spring may fly very low over water when encountering head-winds and collide with wader-nets set along the water-front. In calm weather or when migrating before the wind Goldcrests have never been caught at night. I guess this is just another example of birds choosing the altitude, where the headwind is weakest or least harmful to flight.

Of course my own material is extremely poor and only indicative. But I am convinced, that more material and better material is already at hand, and that we only need to compile it from different sources (and maybe enter into some sort of coalition with Operation Baltic workers in order to merit use of their material). The material should be constrained to short-time recoveries with clear-cut meteorological prehistory. And, important, the theoretical prejudice should be locked up; in each case the unbiassed question should run: What is the most likely route of this self-propelling creature, weighing six grams and making its progress between two points under the given external conditions - and how did it fly relative to the wind? No strategies (there may be "strategies"), no selection (there may be "selection", but what selects most: the joint departure for Sweden instead of Denmark - or the winds met with?) The answer will not be correct in some cases: if the Revtangen bird had been recovered at Esbjerg, the interpolated route would be erroneous. Still, interpolation will give a good approximation if the project starts with recoveries within 2 - 3 days of ringing and gradually expands the material to recoveries within a week, ten days. And the whole thing is not only about correct description or sound knowledge, in the end even the theory of bird migration might profit from the suggested approach.

Litteratur/Literature:

ALERSTAM, T. (1979): Wind as a selective agent in bird migration. Orn. Scand. 10: 76-93. º ALERSTAM, T. (1982): Fågelflyttning. Signum. Lund. º CRAMP, S. (1992): The birds of the Western Palearctic. Vol. VI. Oxford Un. Press. Oxford. º GEZELIUS, L. & A. HEDENSTRÖM (1988): Vindens inverkan på fångsten av rödhake Erithacus rubecula och kungsfågel Regulus regulus vid Ottenby. Vår Fågelvärld 47: 9-14 º GRENMYR, U. (1997): Sex differences in recovery pattern and migratory direction of Goldcrests Regulus regulus ringed in northern Europe during autumn migration º KARLSSON, L. (1980): Kungsfågelns höstflyttning över Falsterbo: Tidtabell, könskvot och årliga fluktuationer. Anser 19: 139 - 146. º ZEHNDER, S., S. ÅKESSON, F. LIECHTI & B. BRUDERER (2001): Nocturnal autumn bird migration at Falsterbo, south Sweden. J. Avian Biol. 37: 239 - 248. º ÅKESSON, S. (1993) Coastal migration and wind drift compensation in nocturnal passerine migrants. Orn. Scand. 24: 87 - 94.

The first version of this note was written by Christer Persson and published on the web on 31.1.02, my thanks to Kenneth Bengtsson for providing the Grenmyr paper.

Till artiklar om flyttning/Articles on migration

Till startsidan/Back to start page