Red knot (Calidris canutus) COSEWIC assessment and status report: chapter 5

• Previous page
• Table of contents
• Next page

Species Information

• Name and Classification
• Morphological Description
• Genetic Description
• Designatable Units

Name and Classification

Subspecies

Red Knots are currently classified into six subspecies, each with distinctive morphological traits, breeding areas, migration routes, wintering areas, and annual cycles (Piersma and Davidson 1992; Tomkovich 1992, 2001; Piersma and Baker 2000; Piersma and Spaans 2004; Figure 1). Three subspecies occur in Canada. They include C. c. rufa, C. c. roselaari, and C. c. islandica.

C. c. rufa breeds in the central Canadian Arctic and winters in southern Patagonia and Tierra del Fuego.

In this report, C. c. roselaari will include three biogeographic populations that will be assessed as one designatable unit: i) the Pacific coast population, which breeds in northwest Alaska and on Wrangel Island and migrates down the Pacific coast through Canada and the northwestern US and winters from California to the Pacific northwest region of Mexico, and possibly the Gulf of Mexico; ii) the Florida/SE US population, which likely breeds in Alaska or the western Canadian Arctic and winters in Florida, Georgia and South Carolina and iii) the Maranhão, Brazil population, which likely breeds in Alaska or the western Canadian Arctic and winters in Maranhão, on the north-central coast of Brazil. The three groups clearly form separate biogeographic populations; there is currently uncertainty over the taxonomic status of knots in the Florida and Maranhão populations.

C. c. islandicabreeds in the northeastern Canadian High Arctic probably as far west as Prince Patrick Island and south to Prince of Wales Island (Godfrey 1992; Morrison and Harrington 1992) and in the high Arctic of Greenland from the northwest around the north coast to about Scoresby Sound on the east coast: it winters in the UK and the Netherlands and migrates to the breeding grounds through Iceland and northern Norway.

C. c. rogersi breeds on the Chukotski Peninsula in eastern Russia and winters in south-east Australia and New Zealand.

C. c. piersmai breeds on the New Siberian Islands in north-central Russia and winters in northwest Australia.

Figure 1: Worldwide Distribution of the Six Recognized Subspecies of the Red Knot

Arrows connect non-breeding (“wintering”) areas (dots, scaled to population size) with breeding areas (dark gray shading). Migration system of rufa knots between wintering areas in Tierra del Fuegoand breeding areas in the central Canadian Arctic is outlined. Note that former wintering populations of rufa on the coast of Patagonia in Argentina have almost disappeared, with the bulk of the population currently confined to Tierra del Fuego. (Adapted from a map drawn by Dick Visser, provided by Jan van Gils, see Niles et al. 2005).

The nominate subspecies C. c. canutus breeds on the Taymyr Peninsula in western Siberia and winters in west and southwest Africa.

Morphological Description

The Red Knot is a medium-sized shorebird with a typical calidridine sandpiper profile: proportionately small head, bill straightish, tapering from thicker base to thinner tip, and not much longer than head; short neck, short tibia, stout tarsus, long tapered wings giving an elongated streamlined profile to the body. It is the largest of the calidridine sandpipers in North America (length 23-25 cm, mass about 135 g though highly variable).

In breeding, or alternate plumage, knots are highly distinctive, with face, neck, breast and much of the underparts coloured a distinctive rufous chestnut-red (Figure 2). The lower belly and vent behind the legs tends to be light, especially in rufa compared to the other subspecies, and some whitish or brownish feathers may be scattered through the breast (thought to be more common on females). Feathers on the upperparts have dark brown-black centres, edged with rufous and grey, giving the bird a spangled appearance that provides a remarkably effective camouflage on the sparsely vegetated High Arctic breeding grounds. Flight feathers range inwards from dark brown/black in the primaries to grey in the secondaries and tail feathers, and there is a narrow whitish wingbar. Males tend to be more brightly coloured than females, with more extensive rufous on the underparts.

Figure 2: Adult Red Knot in Breeding (alternate) Plumage

In winter, or basic plumage, knots are much plainer, with white underparts and a pale grey back. The upper breast has greyish or brownish streaking, extending laterally along the flanks, and the head has dull greyish patterning with a whitish supercilium.

Juveniles have similar plumage, but can be distinguished by dark subterminal bands on the feathers of the mantle, scapulars and coverts, giving the bird a characteristic scaly appearance. Juveniles may also have a pale dull buffish colour suffusing the breast.

Knots may be distinguished from the superficially similar dowitchers (Limnodromus species) by their shorter bill, paler crown, whitish rump barred with grey (vs a white lower back forming a distinctive “V” in flight in dowitchers), and voice; and from the smaller Dunlin (Calidris alpina) by their straighter bill (Dunlin’s appear proportionately longer with a droop at the tip).

Genetic Description

Genetic differences among knot subspecies have been investigated by sequencing the fast-evolving control region of the mtDNA molecule (Buehler and Baker 2005). Most haplotypes differed by a single base change, producing a minimum spanning network diagram with a star-like pattern characteristic of a species that has undergone a recent bottleneck with subsequent expansion (Slatkin and Hudson 1991; Figure 3). Despite the apparent lack of sorting of haplotypes into discrete genetic lineages in each subspecies, knots showed low but significant population differentiation using both conventional F-statistics and exact tests. Four genetically distinct groups were found, corresponding to C. c. canutus, C. c. piersmai, C. c. rogersi and a North American group containing C. c. roselaari, C. c. rufa and C. c. islandica (Table 1, Pooled Exact test, P < 0.001, Buehler and Baker 2005).

Above the diagonal “+” indicates significance at the 0.01 level, and – indicates not significant (P> 0.01). From Buehler and Baker (2005).

Genetic differences between subspecies are also apparent in nuclear DNA. A genomic scan of 836 loci using amplified fragment length polymorphisms (AFLPs) detected different frequencies of the dominant markers at 129 loci, and showed significant genetic differentiation among subspecies (F_ST = 0.089). The genetic distance between C. c. roselaari and C. c. rufa is small (0.1), but similar to the genetic distance between C. c. rogersi (southeast Australia and New Zealand) and C. c. canutus (Eurasia).

The demographic history of knot populations can be deduced from the genetic signature in the control region sequences, providing they are selectively neutral, which appears to be the case in knots. This can be done by computing the number of mutational differences between each pair of sequences in individual birds. These pair-wise differences in knot subspecies have a single peak pattern that would be expected when a population expands after a recent bottleneck (Avise 2000, Figure 4).

Coalescent modelling of the sequence variation using a rate of molecular evolution calibrated for shorebirds estimated that divergence times of populations representing all six subspecies of knots occurred within the last 20 000 years (95% CI: 5 600-58 000 years ago), thus corresponding to the Last Glacial Maximum 18 000-22 000 years ago. This basal split separated C. c. canutus breeding in central Siberia and migrating to western Africa from a lineage that expanded into eastern Siberia and began to migrate to Australia (the ancestor of C. c. rogersi and C. c. piersmai).

Figure 3: Minimum Spanning Network Diagram Showing the Relationships Between Haplotypes from the Mitochondrial Control Region of Knots

Ovals represent haplotypes and connecting lines represent a single base pair change between haplotypes. Small open circles on lines represent multiple base pair changes between haplotypes. From Buehler and Baker (2005).

Figure 4: Observed and Expected Mismatch Distributions of Mitochondrial Control Region Sequences in Knots

Knots closely match the pattern expected under population growth in the recent past. From Buehler and Baker (2005).

As the ice retreated, the latter lineage eventually expanded across Beringia into Alaska and established the North American lineage about 12 000 (95% CI: 3 300-40 000) years ago. At this time, an ice-free corridor that had opened between the ice sheets covering the Rockies to the west and the Great Plains to the east served as a dispersal route for an assortment of organisms, including humans. This corridor was oriented NW-SE, and may thus have guided the evolution of a new migratory pathway between Alaska or the western Canadian Arctic and the southeast United States. As the ice sheets retreated farther eastwards across the High Arctic of Canada, the ancestral population was fragmented sequentially within the last 5500 years into three breeding populations, corresponding today to C. c. roselaari, C. c. rufa and C. c. islandica. If this is correct, then the present wintering flocks in the southeast US are properly attributed to C. c. roselaari and would be predicted to return annually to their ancestral breeding grounds in Arctic northwest North America. Furthermore, the migration pathways of C. c. rufa and C. c. islandica are newly evolved responses to the eastward expansion of their breeding ranges. The divergence of C. c. piersmai and C. c. rogersi was estimated to have occurred about 6 500 (95% CI: 1 000-23 000) years ago, probably as a consequence of their isolated breeding ranges in the New Siberian Islands and the Chukotski Peninsula in Russia.

Given the recent nature of these divergence times among knot subspecies, it is not surprising that the level of genetic differentiation in the neutral mtDNA sequences and nuclear AFLP is small. There has simply not been enough time for mutations to accumulate in these DNA regions to track evolutionary changes operating in the more immediate scale of ecological time. In such cases, the apparently small genetic differences in neutral DNA sequences should not be misinterpreted in defining subspecies (Avise 1989). Instead, morphological and ecological differences are more likely to reflect adaptive changes that represent responses to positive natural selection. The situation is summarized by Buehler and Baker (2005) thus: “While the six currently recognized subspecies of Red Knots cannot be distinguished by their control region sequences, they are beginning to sort into different lineages and it would be inadvisable to lump them into a single evolutionary unit on genetic grounds alone. Given the passage of time, evolution of neutral genetic markers in Red Knots should catch up with morphological and plumage differences, different migration routes, separate breeding grounds, and different moult schedules to more clearly distinguish subspecies …“

Designatable Units

This assessment will be based on three designatable units that correspond to the three subspecies of Red Knot that occur in Canada; C. c. rufa, C. c. roselaari and C. c. islandica.  Although genetic differences between the three subspecies may be small owing to the relatively short time they have been separated (see above), they occupy widely separated geographic areas (breeding and wintering; Salomonsen 1950; Godfrey 1953, 1986; Morrison 1975), have different migration and life history schedules, differences in morphology and plumage (Conover 1943; Morrison and Harrington 1992; Harrington 2001) and lack interchange between populations (Baker et al. 2005a,b).

The main uncertainty with the assigned designatable units concerns the inclusion of the Florida/SE US and the Maranhão, Brazil populations within C. c. roselaari.  The taxonomic status of the Florida/SE US population is currently under revision. The genetic evidence suggests, however, that it is closer to C. c. roselaari than to C. c. rufa (Niles et al. 2005). This population also differs from rufa in morphology (Niles et al. 2005) and banding of thousands of birds in the flyway has clearly shown no detectable interchange between the two groups on the wintering grounds (Baker et al. 2005a). For these reasons, we will include the Florida/SE US population of knots with C. c. roselaari.

The taxonomic status of the Maranhão, Brazil population is also uncertain. Genetic (Baker et al. 2005a; Niles et al. 2005) and stable isotope evidence (Atkinson et al. 2005) from feathers suggests a closer affinity to the Florida/SE US population than C. c. rufa.  Also, this population differs from rufa in morphology (Baker et al. 2005a; Niles et al. 2005) and migration scheduling. Like the Florida/SE population, there appears to be no interchange between the Maranhão populations and rufa wintering in Tierra del Fuego. Given the wide geographic separation from rufa wintering in southern South America, the lack of interchange, the relative proximity to more northerly wintering groups in Florida/SE US, and the proposed evolution of migratory patterns based on genetic evidence (see above), we also assign the Maranhão birds to the roselaari group. We will refer to this group as the roselaari type for purposes of assessment.