Skip booklet index and go to page content

COSEWIC Assessment and Status Report on the Red Knot in Canada

Population Sizes and Trends

Search Effort

Wintering Grounds

Counts on the wintering grounds are considered the best method of assessing population size, since the birds are resident and little movement is likely to occur between sites; it is the method of choice recommended for the species in technical documents accompanying the US Shorebird Conservation Plan (Brown et al. 2001). Coverage of the entire wintering range in effect enables a census to be carried out, in which all of the birds are counted. Uncertainties in the total are related to counting and detection errors, since there is no sampling or extrapolation involved. Following surveys in the 1980s, counts on the major wintering areas were resumed in 2000 and have continued through 2006. The entire wintering range in Patagonia and Tierra del Fuego has been covered in three of those years in order not only to count all birds, but also to determine whether any changes at the principal sites may have been caused by redistribution within the range.

A major feature of the South American wintering surveys is that they have all been carried out by the same observers, Morrison and Ross, thus eliminating concerns that changes in numbers may be caused by an observer effect.

Counts from Tierra del Fuego and estimates of the population made through re-sighting studies conducted during migration have been in close agreement (Gonzalez et al. 2004). Ground and aerial counts conducted at Bahia Lomas, Chile and at Rio Grande, Argentina have also agreed closely (RIGM, RKR, and LJN unpublished data).


Although counts at migration areas may be affected by turnover (birds moving through the site at different times, or at different rates in different years) and uncertainties regarding the proportion of the population using the site (Bart et al. 2005), they can nevertheless provide valuable additional data on trends in a migratory population (Morrison et al. 1994; Morrison 2001). Maximum counts or mean counts (sometimes during a specific time window, e.g., to estimate adult numbers, where their passage time differs from juveniles) have been used as indices of annual abundance. Trends in rufa populations have been estimated from counts at various locations in both South and North America; trend estimates from migration areas in southern South America have the advantage that they involve only rufa knots.

South America

San Antonio Oeste, Rio Negro Province, on the coast of Patagonia in Argentina, is a major stopover area for rufa knots during northward migration. A complete “local census” is conducted by experienced observers approximately weekly during March and April, covering all the beaches used by knots and thus providing a comprehensive count of birds in the area. Data are available for the 16-year period 1990-2005. Counts have been conducted at Fracasso Beach, on the Valdes Peninsula, Chubut Province, Patagonia, Argentina, between 1994 and 2005 at regular intervals during the passage period in March and April. Finally, a series of counts spanning the years 1995 to 2003 are available from the National Park at Lagoa do Peixe, Rio Grande do Sul state, in southern Brazil, another major stopover area for rufa knots.

Delaware Bay

The most comprehensive data on knots during northward migration in the US comes from Delaware Bay, the critically important final stopover area before the birds take off for their breeding areas. Aerial surveys have been conducted since 1982, and a standard survey procedure was adopted in 1986 involving weekly flights (6) during the period early May to mid-June over areas used by knots in New Jersey and Delaware (Clarke et al. 1993, 2001). The same observers have taken part in all flights. Rufa knots may mix with the Florida/SE US and Maranhão, Brazil populations of C. c. roselaari during migration through the bay. However, knots from the latter two groups generally pass through this site somewhat earlier than rufa (Atkinson et al. 2005). 

Eastern North America

In Canada, information on knot trends has come from the Maritimes Shorebird Survey (MSS) scheme, a survey program in which volunteers collect data on shorebirds in a consistent manner at their study site at two-weekly intervals throughout the period of southward migration (Morrison et al. 1994; Morrison 2001; Morrison and Hicklin 2001). The scheme has operated for over 30 years (1974 – ongoing), and a number of sites have had coverage by the same observers throughout that period. The MSS, along with the International Shorebird Survey (ISS) program collecting similar data in the US (Howe et al. 1989), is the longest running shorebird survey program in North America. In Quebec, knot trends have been obtained from analysis of checklist data submitted to the Étude des populations d’oiseaux du Québec (ÉPOQ) for areas along the St. Lawrence River system between 1976 and 1998 (Aubry and Cotter 2001a,b).

Breeding Grounds

The only information on recent trends in knot numbers on the breeding grounds in the central Canadian Arctic has come from studies conducted by Larry Niles on Southampton Island as part of the international Red Knot project. Surveys were conducted on a 9.2 km2 study area annually between 2000 and 2005 (Niles et al. 2005).


C. c. rufa

Initial estimates of the size of the rufa population in the 1980s, derived from band re-sighting data and counts, suggested the population numbered between 100 000 and 150 000 (B.A. Harrington unpubl. results in Morrison and Harrington 1992). This estimate included rufa wintering in southern South America, as well as the Florida/SE US and Maranhão, Brazil populations (Harrington et al. 1988; Morrison and Harrington 1992; Morrison et al. 2001a). During this period, Morrison and Ross (1989) reported aerial survey totals for South America of 76 400 overall, with 67 500 of this total being rufa and the remaining birds part of the Maranhão, Brazil population.

By the latter part of the 1990s, the population counts from the southern wintering grounds of rufa were closer to 60 000 (Baker et al. 2001; Morrison et al. 2001b). The most recent population estimate for rufa, derived from the wintering ground aerial census in 2005, suggests that this group may now number about 18 000-20 000 individuals, of which 75% or 13 500-15 000 are adult birds (Table 3). Estimates of the population size from band re-sighting data during this period were consistent with the aerial counts (Gonzalez et al. 2004).

Losses on the wintering grounds were particularly evident at the “peripheral” sites along the coast of Patagonia compared to the major sites in Tierra del Fuego. In 1982/85 the Patagonian sites held 14 314 (21.2%) of the overall total. By 2002 this had fallen to just over 2000 (6.9%), and was reduced further to less than 1000 in 2003 and 2004 (1.8% and 2.8%, respectively; Morrison et al. 2004).

C. c. roselaari

An aerial survey of the Maranhão coast of north-central Brazil in 2005 produced a total of 7600 knots of which 5700 are mature adults (Table 3; Baker et al. 2005a), similar to the 1982/85 total of 8150 (Morrison and Ross 1989). Estimates for the Florida/SE US population over the past 20 years have ranged between 10 000 (Morrison and Harrington 1992), 7 500 (B.A. Harrington pers. comm. 2005), and 4 500 (Sprandel et al. 1997; Niles et al. 2005). The most recent estimate based on aerial and ground surveys during the winter of 2005-2006 is 4500, of which 3375 are mature individuals (Table 3; LJN unpublished data). Finally, the Pacific coast population numbered 6000-9000 in 1981. The current total for the entire Pacific coast is 2000-4000, of which 1500-3000 are mature (Table 3; R.I.G. Morrison unpubl. data; G. Page and D. Paulson pers. comm. 2005), plus an additional 1000 birds from other areas (Table 3). The current overall population estimate for C. c. roselaari involving the populations wintering on the Pacific coast, in Florida, and in Maranhão is 17 100 individuals of which 12 825 are mature individuals (Table 3).

C. c. islandica

The current estimate for C. c. islandica is approximately 270 000, of which 202 500 are mature, based on midwinter counts in the United Kingdom and the Netherlands of 250 000, plus a further 20 000 in France (Table 3; T. Piersma pers. comm.); this estimate is somewhat lower than recent estimates by Wetlands International (2005a,b), though similar to the estimate produced in 2003 using the TRIM (Trend analysis and Indices for Monitoring data) method (Wetlands International 2005b). Canada is thought to hold about 81 000 or 40% of the breeding population of islandica (Morrison et al. 2007). 

Table 3: Current Population Estimates of the Three Subspecies of the Red Knot that Occur in Canada
SubspeciesEstimated population sizeSource
islandica270 000
(202 500 mature;
81 000 breed in Canada)
T. Piersma, pers. comm.
Morrison et al. 2007
roselaari17 1001
(12 825 mature)
present estimates, involves separate biogeographical populations wintering in Florida/SE US (4 500 – counts from surveys in winter 2005-2006, LJN pers. comm.), Maranhão, north-central Brazil (7 600 – Baker et al. 2005b), Pacific coast2 (2 000 - 4 000 – G. Page and D. Paulson pers. comm. 2005), and other areas (1 000 - see note1)
Tierra del Fuego & Patagonia
18 000 - 20 0003
(13 500 -15 000 mature)
Morrison et al. (2004, and unpublished data)

Bracketed numbers indicate uncertainties and possible overestimates – see footnotes.

1 Roselaari appears to have a much smaller population than that previously suggested (35 000-50 000) by Wetlands International (2005). The Wetlands International estimate included birds wintering from Florida through the Caribbean to N. South America, which were considered to belong to this subspecies, as well as the Pacific coast. This total was based on previous estimates, some of which now appear to be too high: previous totals for Florida (7 500 B.A. Harrington pers. comm. 2005), Texas (3 000 B.A. Harrington pers. comm. 2005), Pacific 6 000-9 000 (Page et al. 1999), North-central Brazil (8 200 Morrison and Ross 1989) and birds wintering in northern Venezuela (520, Morrison and Ross 1989) and Panama (200, Buehler 2002), come to approximately 28 400. Current estimates are closer to 17 100 and would include distinct biogeographic populations in Florida (4 500), Maranhão (7 600), Pacific coast (2 000-4 000) and miscellaneous other areas (Texas 300 (Skagen et al. 1999), northern Venezuela 520 (Morrison and Ross 1989), and Panama 200 (Buehler 2002)). Knots are also known to occur on some Caribbean islands (e.g., Cuba) during the winter; there is little information available, but numbers are not thought to be large.

2 Previous estimate of 9000 is likely an overestimate as it was based on counts summed across seasons (Page et al. 1999, G. Page pers. comm.): previous counts have included 6000 migrating northwards through the Pacific Northwest USA, but there have been no recent reports of such large numbers. Winter counts suggest the total may be closer to 2000-4000 for the entire Pacific coast (R.I.G. Morrison unpubl. data; G. Page and D. Paulson pers. comm. 2005), an estimate compatible with observations of 2000 - 3000 knots on southward migration in the Pacific Northwest USA in 2006 (J. Buchanan pers. comm.).

3 The estimate for rufa in Wetlands International (2002) would include the approximately 18 000 in Tierra del Fuego/Patagonia and the approximately 7 600 on the north-central coast of Brazil. Note that these are considered separate biogeographic populations and that the birds in Brazil are presently included in the roselaari group.

Fluctuations and Trends

C. c. rufa

Wintering Grounds

The best trend information for rufa knots comes from surveys conducted on the wintering range, where there is no mixing with other populations. Overall, the wintering population of C. c. rufa has decreased from approximately 67 500 birds in 1982 (Morrison and Ross 1989) to approximately 18 000 in 2006, a decrease of 73.4% in the last 24 years. The rate at which this decline has occurred differs between the core and peripheral wintering areas and over the above time period (Figure 8). That is, the rate of decline was relatively slow between 1982 and 2000, particularly at the core sites, and then increased dramatically between 2000 and 2006 (Figures Figure8 and Figure9).

Specifically, the annual rate of decline over all wintering sites between 1982 and 2000 was -1.2%. The rate, however, increased to -17.0%/year between 2000 and 2006 at the core sites, where most of the population now resides (Figure 9). Assuming a linear decline in population size between 1982 and 2000 (see Figure 8), the estimated rufa population in 1991 (i.e. 15 years or approximately 3 generations ago) would have been approximately 60 900 individuals. This means the population would have decreased by 70.5% in the last 15 years or 3 generations, given the current population size of 18 000 birds.

Figure 8: Population Estimates of Red Knots rufa Subspecies in Southern South America from Aerial Surveys Conducted in 1982/1985 and 2000 – 2006

Figure 8: Population estimates of Red Knots rufa subspecies in southern South Americafrom aerial surveys conducted in 1982/1985 and 2000 – 2006.

Solid symbols: counts from aerial surveys; open symbols, estimated totals (value for Patagonian sites in 2000 from regression equation, values for Patagonian sites in 2005 and 2006 taken as average of counts for Patagonian sites in 2002-2004). Upper line (squares) – overall wintering population; middle line (circles) – major wintering sites in Tierra del Fuego; lower line (diamonds) – “peripheral” sites in Patagonia.

Figure 9: Counts of C. c. rufa at Major Wintering Sites in Tierra del Fuego

Figure 9: Counts of C. c. rufa at major wintering sites in Tierra del Fuego.

Data suggest little change from 1982 to 2000, followed by rapid declines.

Migration Trends

Although migration counts are not independent of the counts on the wintering grounds (i.e. they include the same birds), they do show a highly consistent pattern to those described above. In turn, this suggests that declines have not been caused by redistribution of birds on the wintering grounds, but rather reflect a true population decrease. Indeed, no redistribution of birds was found on aerial surveys across the southern wintering areas (Morrison et al. 2004).

South America

Survey data from stopover areas used as the birds move to their wintering grounds in South America (i.e. two on the coast of Patagonia, Argentina and one in southern Brazil) show similar trends to those described above for the wintering grounds in Tierra del Fuego. For instance, rufa knots passing through the Valdes Pennisula, Argentina showed a significant annual rate of decline of -22.4% between 1994 and 2005 (r = 0.725, R2 = 0.526, p < 0.017). If the rate of decline for this 11-year period is extrapolated to cover a 15-year period, the population at this site would have decreased by 97.8% in the last 3 generations. Similarly, birds at the second stopover site in Argentina, San Antonio Oesto, showed a significant annual rate of decline of -8.0% between 1990 and 2005 (R2 = 0.717, p = 0.003). At this rate of decline, the population at this site would have decreased by 71.5% over the last 15 years. Finally, rufa passing through the third site in southern Brazil showed a significant annual rate of decline of -30.6% between 1995 and 2003 (r = 0.928, R2 = 0.862, p < 0.0009). Again, if this rate of decline is applied over a 15-year period, the population at this site would have decreased by 99.6% in the last three generations.

In summary, migration counts at three sites in South America also showed strong declines in the last 8 to 15 years, with annual rates of decline ranging from -8.0% to -30.6%. At these rates, the populations at the above sites would have decreased between 71.5% and 99.6% over the last three generations.

Delaware Bay

Delaware Bay on the east coast of the US is the key final refuelling stop for C. c. rufa wintering in South America before the final leg of their journey to the Arctic breeding grounds. Knots from the Florida/SE and Maranhão, Brazil populations of C. c. roselaari (see below for wintering ground trends) also pass through this site, though probably slightly earlier than the rufa knots (Atkinson et al. 2005). Survey data from this site show a significant annual rate of decline of -4.4% between 1982 and 2006 (R2 = 0.283, p = 0.009), with a peak of 95 000 in 1982 to less than 15 000 in 2006. 

Surveys between 1992 and 2006 show a significant annual rate of decline of -5.85% (Figure 10). At this rate of decline, the population would have decreased by 59.5% over the last 15 years (three generations).

The prolonged decline in counts in Delaware Bay appears to reflect the declines in the rufa population wintering in Tierra del Fuego, possibly accelerating declines in the Florida population, and the suspected declines in the Maranhão population.

Figure 10: Maximum Counts Observed During Aerial Surveys of Red Knots During Northward Migration in Delaware Bay, 1992-2006

Figure 10: Maximum counts observed during aerial surveys of Red Knots during northward migration in Delaware Bay, 1992-2006.

Adapted from Niles et al. (2005) and R.I.G. Morrison and R.K. Ross (unpublished data). Knots from separate biogeographic populations wintering in Tierra del Fuego, Maranhão, and Floridapass through Delaware Bay.

Eastern North America

Combined data from the International (east coast US) and Maritimes (Atlantic Canada) shorebird survey schemes collected during fall migration have shown a significant long-term annual rate of decline of -13.3% between 1972 and 1998 (more recent data are not available; Figure 11). Like Delaware Bay, these trends reflect changes in populations of rufa and the Florida/SE US and Maranhão, Brazil populations of roselaari (see below).

Figure 11: Mean Counts of Red Knots Observed During Southward Migration on the East Coasts of the US and Canada (from International Shorebird Survey and Maritimes (Atlantic Canada) Shorebird Survey datasets), 1972-1998

Figure 11: Mean counts of Red Knots observed during southward migration on the east coasts of the US and Canada (from International Shorebird Survey and Maritimes (Atlantic Canada) Shorebird Survey datasets), 1972-1998.

(B.A. Harrington unpublished data; R.I.G. Morrison, unpublished data). Counts likely include knots from separate biogeographic populations wintering in Tierra del Fuego, Maranhão, and Florida.

Breeding Grounds

The only information available on trends from rufa’s central Canadian Arctic breeding grounds was obtained by L.J. Niles from Southampton Island, where a 9.2 km2 study area was censused between 2000 and 2004. Breeding densities of knots fell from 1.16 nests/km2 in 2000 to 0.33-0.55 nests/km2 in 2003-2004. The difference in mean density over the first three years compared to the last two years was significant (ANOVA, F1,3 = 10.09, p = 0.0502).

Population Modelling and Possibility of Extinction

Demographic studies by Baker et al. (2004) showed that survival of the Tierra del Fuego adults declined from an average of 85% in 1994-1998 to only 56% during 1999-2001. Population projections were developed using a matrix model under two scenarios (Figure 12): (a) The “best case scenario”, in which adult survival returned to 85%, juvenile survival was assumed to be half that of adults, and fecundity averaged 0.29; and (b) a “worst case scenario” in which adult survival remained at 56%, with juvenile survival half that of adults and a mean fecundity of 0.29. Under the best case scenario, the Tierra del Fuego population (including immature birds wintering farther north) would remain stable at about 70 000 birds, based on the 2000 count data from South America (Morrison et al. 2004). However, if the worst case scenario prevailed, with seriously depressed levels of adult survival, the population would fall rapidly to very low levels by 2010, putting it at immediate risk of extinction. Between 2000 and 2002, the model predictions using the adult survival of 56% closely followed the field data. Between 2002 and 2004, the population apparently stabilized somewhat in the 27 000-31 000 range, but the dramatic drop to only 17 650 in 2005 and 17 211 in 2006 brought it back close to the track of the worst case scenario once again, suggesting a trajectory heading once more towards extinction. Atkinson (2005) analysed band returns in Delaware Bay and suggested that survival of the southern wintering population may have recovered somewhat in 2002-2004, when the population stabilized on the wintering grounds. It is clear, however, that increased mortality and/or lack of recruitment seriously impacted the population again in 2005, and that problems for the rufa knot population are ongoing.

Figure 12: Predicted Population Trends of Adult rufa Red Knots (middle grey line), Juveniles (lower grey line), and the Sum of these Two Age Groups (top grey line) for 10 Years from 2000; Associated 95% Confidence Limits for Adults are Shown as Dashed Lines

Figure 12: Predicted population trends of adult rufa Red Knots (middle grey line), juveniles (lower grey line), and the sum of these two age groups (top grey line) for 10 years from 2000; associated 95% confidence limits for adults are shown as dashed lines.

Two scenarios are presented: (a) constant adult survival of 85%, with juvenile survival being half that of adults (λ = 1); and (b) constant adult survival of 56%, with juvenile survival being half that of adults (λ = 0.66). The dots represent aerial census counts of over-wintering knots inTierra del Fuegoduring 2000-2006 (Morrison et al. 2004, and unpubl. data). The 95% upper and lower confidence limits are based on 1000 bootstrap iterations. Modified from Baker et al. (2004, and see Baker et al. 2005a).

C. c. roselaari

Wintering Grounds

The best trend information for C. c. roselaari also comes from counts on the wintering grounds.

Florida/SE US population: The portion of the population wintering in Florida has dropped from 10 000 up to the early 1990s (Harrington et al. 1988; Morrison and Harrington 1992) to 3 020 (2005-2006, LJN pers. comm.). This represents a decrease of 70% over the last 15 years. The number of birds wintering outside Florida in the early 1990s is not known.

Little information is available from individual sites from which to derive trend estimates. The only site at which data are available from the early 1990s to the present (i.e., 15 years or 3 generations) is from Honeymoon Island, Florida (Niles et al. 2005). Here the population showed a significant annual rate of decline of -33.1% (R2 = 0.604, p = 0.002). At this rate of decline, the population would have decreased by 98.1% over the last 15 years.

Maranhão, Brazil population: In the early 1980s, 8150 knots were counted on the Maranhão coast on the north-central coast of Brazil (Morrison and Ross 1989). Aerial surveys of the same area were carried out by different observers in 2005, and resulted in a count of 7600 birds (Baker et al. 2005a), a decrease of about 7% from the earlier survey. There is concern, however, that the most recent count may have involved some degree of overestimation, and it is likely that this population has also undergone a larger decline than the counts suggest (A.J. Baker unpubl. data.).

Pacific coast population: The most recent assessment of the overall population size of the Pacific coast population of C. c. roselaari is 2000-4000 birds based on counts of knots on wintering and migration areas (J. Buchanan pers. comm. 2006, Morrison et al. 2007). The population is thought to have declined from previous estimates of 6000 – 9000 birds in the 1990s (Paulson 1993; Page et al. 1999). This represents a decrease of approximately 60%, using the mid-point in the range of the counts. At Grays Harbor Washington, counts have declined from a high of some 6000 knots recorded in 1981 (Paulson 1993) to only about 250 in 2006 (Buchanan 2006). Here the population showed a significant annual rate of decline of -12.0% (R2 = 0.92, p = 0.0002). If this rate of decline were consistent over the most recent 15 year period, the population at this site would have decreased by 72.0% over the last 3 generations. 

In summary, the Pacific coast and Florida/SE US populations of C. c. roselaari have shown a 60 to 70% decrease in population size, respectively, in the last 15 years. The Maranhão, Brazil population decline has been about 7% since the early 1980s, but a suspected overestimation in recent surveys indicates declines may be larger than indicated. Overall, the combined decline for the three populations is approximately 47% over the last 15-20 years based on a decline from a combined population of about 25 650 (10 000 + 8 150 + 7 500) in 1990 to about 13 620 currently (3 020 + 7 600 + 3 000). This number may, however, be conservative, if the Maranhão population was overestimated in recent surveys.

C. c. islandica

Wintering Grounds

Estimated numbers of C. c. islandica wintering on the European seaboard have ranged between about 300 000 and 400 000 between the early 1970s and late 1990s. A series of poor breeding years in the 1970s led to declines, but the population recovered to a peak in the 1980s and has again fallen somewhat up to the late 1990s (Boyd and Piersma 2001; Collier et al. 2005; Wetlands International 2005b). Between 1998 and 2003, numbers wintering in the UK and Netherlands have fallen about 17% from about 300 000 to about 250 000 (Figure 13). The pattern over the last 15 years is presumably similar, given that the population was likely at the low end of the above range (i.e. 300 000) throughout the 1990s.

Figure 13: Declines of Red Knots Subspecies islandica on Wintering Grounds in the United Kingdom and the Netherlands Between 1998 and 2003

Figure 13: Declines of Red Knots subspecies islandica on wintering grounds in the United Kingdomand the Netherlands between 1998 and 2003

(T. Piersma pers. comm.).

Rescue Effect

The present evidence, discussed elsewhere in this report, supports the view that rufa, roselaari, and islandica all form distinct biogeographic units with no detectable interchange. A rescue effect would not be anticipated for rufa and roselaari, whose range is restricted to North America. There may be some rescue potential, however, for islandica from breeding populations in Greenland.