Skip booklet index and go to page content

COSEWIC Assessment and Status Report on the Chimney Swift in Canada

Population Sizes and Trends

Abundance

Canada

Estimates of the Canadian population of Chimney Swifts can be made from Breeding Bird Survey (BBS) data. These surveys are designed for relative abundance measures, so total abundance estimates made from them are somewhat imprecise. Peter Blancher (in litt.) has provided an estimate for the Canadian population based on BBS data from 2000 to 2005. This calculation assumes that the survey sampled all the swifts within a 400-m detection distance. It also used a time-of-day correction of 1.12 to compensate for points sampled before swifts were active in the morning. Finally, a pair correction of 1.5 was used to compensate for adults that were on nests during the surveys, assuming that about a third of the adult population is non-breeding. The result is a Canadian estimate of 17 250 breeding individuals for the years 2000-05; an estimate of 20 250 for the 1990s can be generated using the same methods.

More precise estimates can likely be made from provincial and regional population surveys.

Quebec

The annual Chimney Swift Survey Program (Figure 8) in Quebec has never counted more than 5000 swifts: 1572 in 1998; 3508 in 1999; 3687 in 2000; 2095 in 2001; 3496 in 2002; 3850 in 2003; 3131 in 2004 and 4700 in 2005. In 2006, effort was increased but only 2415 Chimney Swifts were counted, the decline possibly a result of high mortality in Hurricane Wilma the previous fall. Effort increased greatly from 2004 on, but few new sites, none of them with large numbers of swifts, have been found. With such a high level of coverage and effort, the probability of discovering new large roosts or many small ones is low. 

Based on the monitoring of 26 roosts (where data are sufficient), it was determined that immatures constitute about 55% of the birds (Gauthier et al. in press). Although the coverage and survey effort in Quebec were high, not all sites were surveyed every year--about 80% of the sites were monitored in 2005. To come up with a good population estimate, we added together the maximum number of swifts observed at each site in 2005 and the maximum number seen between 1998 and 2004 for all known sites not monitored in 2005 and obtained an total of 5700 individuals. If we remove the number of immatures (55%), the total breeding population for the province of Quebec then becomes 2520 adults or 1260 pairs for 2005 (Gauthier et al. in press).

Ontario

Population estimates for Ontario can be made in a variety of ways, but the strongest dataset for this is likely that from the second Ontario Breeding Bird Atlas. Abundance estimates collected for the first Ontario Breeding Bird Atlas (Cadman et al. 1987) allow an estimation of the Ontario population of between 29 010 and 265 384 during the period 1981-1985. Using the same method of calculation as used in the Maritimes Atlas (which took order-of-magnitude estimates for each square and calculated an overall population estimate for the region; Erskine 1992), the 1985 Ontario population size would be about 35 000 individuals. The second Ontario Breeding Bird Atlas used point counts to more accurately estimate abundance of bird species across the province. This is a much stronger source of Chimney Swift data than the BBS, because point counts were more randomly dispersed across the landscape, included off-road counts, and were much more numerous than BBS stops.  There were 47 901 point counts in 1 635 10x10 km atlas squares in southern Ontario, where almost all squares were covered by atlassers, and where by far the majority of Chimney Swifts were detected.

Peter Blancher (in litt.) calculated a single point count average for each square that had at least 10 point counts to avoid biasing results to squares most heavily surveyed by atlassers.  He then stratified the square averages by provincial ecoregion, giving estimates for each Ontario ecoregion (the five most southern ecoregions had Chimney Swifts on point counts, with about half found in the Carolinian Forest ecoregion).  To calculate area covered by an atlas point count, he used the same distance, time and pair corrections as above, including the assumption that all swifts within 400 m of the point count were detected in 5 minutes.  Atlassers recorded 77% of Chimney Swifts within 100 m, 23% outside of 100 m, suggestive of a smaller radius of detection, but given the mobility of swifts during a 5-minute count, a 400 m distance seems reasonable.  If the actual detection distance is smaller than 400 m, then the population estimate is conservative. The result is an Ontario estimate of 7500 breeding individuals for the years 2001-05.

A second method of estimating numbers of Chimney Swifts in Ontario would be to use the chimney method based on an extrapolation from figures obtained in Quebec, which are based on systematic surveys of roosting and nesting sites since 1998. Because buildings in Ontario are similar to those in Quebec (structure and chimneys), the number of chimneys and, therefore, the number of swifts will be proportionate to the number of people. Based on this method, there should be about 2988 breeding individuals in Ontario. Figures from the Atlas point-count method (7500 individuals) are those that we retain for this report

Table 3: Total Number of Chimney Swifts Observed at Sites Identified During the Quebec Chimney Swift Survey Between 1998 and 2003
Administrative RegionNumber of Chimney Swifts
Lower St Lawrence406
Saguenay-Lac St Jean35
Quebec City553
Mauricie-Bois Francs147
Eastern Townships309
Montréal353
Outaouais574
Abitibi-Témiscamingue0
North Shore0
Northern Quebec0
Gaspé-Magdalen Islands28
Chaudière-Appalaches1134
Laval2
Lanaudière208
Laurentians2296
Montérégie698
Total6743


Atlantic Canada and Prairies

In the Atlas of Breeding Birds in the Maritimes, Erskine (1992) converted orders-of-magnitude population estimates from each atlas square into an overall estimate for each of the Maritime provinces. The outcome of this exercise was an estimate of approximately 20 000 ± 3 000 pairs in the Maritimes as a whole (New Brunswick: 12 000 ± 2 300, Nova Scotia: 8 500 ± 1 900, Prince Edward Island: 15?). These numbers seem overestimated and are perhaps characteristic of the method used. After applying the BBS trend (Table 4) to these figures we get a swift density about ten times higher than the one for Quebec. The Chimney Swift population estimate from Quebec is based on a systematic inventory and is therefore more reliable. Also, the habitat trend (degradation of traditional chimneys) is similar in both areas. Using an extrapolation of the number of swifts per potential building a total of about 345 pairs is obtained; it is likely that this is low so we estimate a population of about 450 breeding pairs (Gauthier et al. in prep).

There are no population estimates for Chimney Swifts in Newfoundland (Montevecchi and Tuck 1987) but sporadic breeding may still occur (J. Brazil, pers. comm.). No quantitative data exist for the province of Saskatchewan and Chimney Swifts are not common (Smith 1996). In Manitoba, Chimney Swifts seem fairly common in urban areas but only sporadic counts exist (Taylor et al. 2003); some 200 swifts were observed in Winnipeg in 1980. In light of this information, a total of about 450 breeding pairs is estimated for Newfoundland and the Prairies.

Gauthier et al. (in press) present evidence that the Canadian population could be as low as 8000 breeding individuals, but it is likely that the total is higher than that. Using the regional calculations above, a total of 11 820 individuals is estimated for all of Canada (Quebec 2 520; Ontario 7 500; Maritimes 900; all other provinces 900), although the estimates from the Prairie and Atlantic provinces should be considered educated guesses at best.


Trends

Breeding Bird Survey (BBS) routes cover the Chimney Swift’s entire breeding range and provide more than three decades of data. According to the BBS, the Canadian Chimney Swift population declined 7.8% annually between 1968 and 2005 (Downes et al. 2005) (Table 4), resulting in a cumulative decline of 95% over that 37-year period. The decline accelerated until 1998; the trends for the last 25 years, 10 years and 5 years of the 1968-1998 analysis are -7.4%, -11.4% and -15.2% respectively (Dunn et al. 2000). The decline seems to have moderated in recent years with a non-significant annual decline of 3.4% in the 1995-2005 period (Downes et al. 2005). In the last 3 generations (13.5 years) the decline has been 2.37% per year (Sauer et al. 2005), resulting in a total decline of about 28%. All these BBS trends are calculated by comparing numbers on the same routes with the same observers. 

Table 4: Chimney Swift Population Trends (% change/year) in Canada Based on Breeding Bird Survey Data
PeriodRegionTrendPN
1968-20051Canada-7.8*207
1968-20051Ontario-8.4*82
1968-20051Quebec-4.9*60
1968-20051New Brunswick-4.9 30
1968-20051Nova Scotia-8.5*29
1968-19851Canada-8.0*138
1968-19851Ontario-6.4*49
1968-19851Quebec-9.3*37
1968-19851New Brunswick-2.7 27
1968-19851Nova Scotia-10.8*23
1985-20051Canada-6.8*159
1985-20051Ontario-10.2*65
1985-20051Quebec-3.1 46
1985-20051New Brunswick-8.6 21
1985-20051Nova Scotia-5.6 21
1991-20052Canada-2.37 -
1991-20052Ontario-11.3*-
1991-20052Quebec+0.88 -
1995-20051Canada-3.4 117
1995-20051Ontario-9.0*47
1995-20051Quebec+6.8 38
1995-20051Nova Scotia-17.3 16

1 Downes et al. 2005
2 Sauer et al. 2005
N: number of routes used in analysis;
*: P < 0.05.

The annual population index, calculated using BBS data, has declined steadily since 1970 (Downes et al. 2005) (Figure 5). This decline can be observed in all provinces where BBS data is available (Figure 6).


Figure 5: Annual Population Indices for the Chimney Swift in Canada Based on Breeding Bird Survey Data

Figure 5: Annual population indices for the Chimney Swift inCanadabased on Breeding Bird Survey data.

Downes et al. 2005).


Figure 6: Chimney Swift Annual Population Indices in Four Canadian Provinces Based on Breeding Bird Survey Data

Figure 6: Chimney Swift annual population indices in four Canadian provinces based on Breeding Bird Survey data.

Downes et al. 2005.

This decline is occurring throughout the breeding range of the Chimney Swift (DeGraaf and Rappole 1995). The BBS data show a significant downward trend in the entire Chimney Swift population of 1.6% annually between 1966 and 2005 (Table 5). The decline has accelerated, reaching -2.5% annually for the period 1980-2005. In the United States, although Chimney Swifts are considered common in almost all the states where they breed, the population has also declined by 1.5% per year since 1966 (Table 5). Of the 38 US states for which data are available, 22 (58%) show a significant downward trend for the 1966-2005 period (Sauer et al. 2005). Within states showing a significant decrease for the longer period (1966-2005), 16 out of 22 (73%) saw the decline accelerate in the last 25 years (1980-2005; Table 5). Sauer et al. (2005) caution that the trend data from Illinois, Indiana, Kentucky, Michigan, Rhode Island, Tennessee and West Virginia are deficient (e.g. small number of birds per route and fewer than five routes sampled).

Table 5: Chimney Swift Population Trends (% change/year) in the United States Based on Breeding Bird Survey Data
Regions1966-20051980-2005
North America-1.6 ***-2.5 ***
United States-1.6 ***-2.5 ***
Alabama-1.5 *-2.8 ***
Delaware-2.0 *-1.2
Georgia-1.1-2.0 ***
Illinois-2.5 ***-3.6 ***
Indiana-2.8 ***-3.8 ***
Iowa-1.9-3.1 ***
Kentucky-2.4***-4.9 ***
Maine-2.5 **-2.6
Maryland-1.6 **-0.7
Missouri-1.5-2.2
New Hampshire-1.9 *-2.9 ***
New Jersey-3.0 **-2.1 **
New York-1.7 ***-1.0
North Carolina-0.5-1.0
Ohio-0.7-1.6 ***
Oklahoma-3.1 ***-3.3 **
Rhode Island-11.2 **-12.6 **
Tennessee-2.0 ***-3.2 ***
Texas-2.4 ***-3.4 ***
Vermont-2.4 *-0.3
Virginia-1.3 ***-1.4 ***
West Virginia-1.6 **-2.5 ***

Sauer et al. 2005).
*: P < 0.10;
**: P <0.05;
***: P < 0.01

Rodriguez (2002) re-analysed the 1966-1993 BBS data to study changes in the range size of significantly declining birds. During this period, Chimney Swift populations declined by 21% while its range contracted by 32.2%. Rodriguez noted that Chimney Swifts decreased more rapidly at the edge of its distribution than in the centre, where numbers are higher.

Breeding Bird Atlas projects have also reported declines: Palmer-Ball (1996) in Kentucky, Hess (2000) in Delaware and Mulvihill (1992) in Pennsylvania. In Connecticut, Zeranski and Baptist (1990) note that the species began to decline in the 1960s and 1970s. The Driftwood Wildlife Association (2000), a Texas organization conducting a research on the Chimney Swift, reports a population decline since the mid-1980s. Some authors cite the decrease in available chimneys used for nesting sites as the reason for the decline (Zeranski and Baptist 1990; Hess 2000). Sibley (1988) also reports a significant decline in the number of swifts in New York State, particularly in New York City and area. In Ohio, Peterjohn and Rice (1991) reported that the Chimney Swift was widespread, but that the population decline had become obvious in many parts of the state in the 1980s. In Colorado, Kingery (1998) observes that the species has been less and less evident in recent years. The population drop in New Hampshire and Maryland has led the authors to declare that the species should be monitored in the next few years and that a follow-up should be carried out (Sutcliffe 1994; Zucker 1996).

Data collected for the second Ontario Breeding Bird Atlas (2001-2005) indicate that the species has been reported in only 48% as many 10-km squares as it was during the first atlas (1981-1985), even though at least 85% as much effort has been expended on field surveys. When corrected for effort, the data indicate a decline of 46% in the Area of Occupancy over the last 20 years; that decline has occurred throughout the species range in Ontario.

On the basis of the data gathered for the Quebec Breeding Bird Atlas in the 1980s (1984 to 1989), Lemieux and Robert (1995) describe the Chimney Swift as a fairly common species in Quebec. However, they also note that the population is showing a downward trend. According to 1966-1998 BBS data (Table 4), the Quebec Chimney Swift population is falling 6.3% annually. Since 1993 the downward trend seems to have reversed, but this recent trend is not statistically significant (Table 4). This positive trend could be the result of a few good years, giving a short term increase. It could also be a consequence of a limited number of routes in the province or the biased estimation due to short time intervals discussed earlier and it is therefore best to consider trends that cover the longest time periods.

Breeding Bird Survey data are not alone in showing a negative trend in Quebec. For the 1969-1989 period, ÉPOQ (Étude des populations d’oiseaux du Quebec) file data show a significant annual decline of 1.44 percent (p < 0.01) (Cyr and Larivée 1995). After learning about the sharp decline in the range of distribution in Ontario, a subsample (n=200) of the 1995 census squares from the Atlas of breeding birds in Quebec (n=790; Gauthier and Aubry 1995) was made in 2004 to see if a reduction in the Chimney Swift’s range of distribution also occurred in this province. Results showed that the range of distribution significantly declined by 33% between 1989 and 2004 (p<0.001; CWS-QC Unpublished data).


Figure 7: Chimney Swift Breeding Distribution in Southern Ontario in 1981-1985 (top) and 2001-2005 (bottom)

Figure 7: Chimney Swift breeding distribution in southern Ontarioi n 1981-1985 and 2001-2005.

Ontario Breeding Bird Atlas.

Since 1998, the Canadian Wildlife Service has been surveying Chimney Swift roosting and nesting sites in Quebec. The purpose of the survey is to estimate swift population sizes, monitor population changes, locate and characterize the sites used, evaluate how many of these sites have been abandoned or closed over the years and develop conservation measures for the species. The results of the survey are presented in Table 3; 258 confirmed sites were identified between 1998 and 2005. These sites are spread out over most of Quebec’s regions, with the exception of Northern Quebec and the North Shore, Laval and Abitibi-Témiscamingue. All of the sites are shown on Figure 8. The largest roosts in the province are located in St-Georges de Beauce and Mont-Laurier, where at least 1000 birds have been counted. The region where most of the sites have been found to date is Chaudière-Appalaches. There are few sites south of Montréal and in the Eastern Townships, but the network is still being formed. More than a third of the sites identified are roosts and the swifts use church or rectory chimneys in 40% of cases. This survey was made possible thanks to a network of volunteers and 93 observers have participated since 1998.


Figure 8: Distribution of Parishes Surveyed During the Quebec Chimney Swift Survey Program Between 1998 and 2005 and During the Chimney Survey of Religious Buildings in 2000

Figure 8.  Distribution of parishes surveyed during the Quebec Chimney Swift Survey Program between 1998 and 2005 and during the chimney survey of religious buildings in 2000.

Empty circles represent parishes where no swifts were detected, while full stars indicate parishes with a known Chimney Swift site (nest or roost), active, abandoned or closed.

In the case of three roosts, located at least 200 km apart, in St-Raymond de Portneuf, Jonquière and La Pocatière, we have historical data on the presence of Chimney Swifts. At all three sites the downward population trend observed confirms other existing Canadian data (Figure 6). In St-Raymond de Portneuf, located 50 km west of Quebec City, swifts are occupying the chimney of a former convent that has been converted into a senior citizens’ residence. The chimney is still being used by the birds. In 1981, approximately 1200 swifts were counted, while in 2002 the maximum number was 234 (Figure 3). Their numbers have thus fallen 81% in 21 years, or 3.8% per year.

Figure 3 also presents the changes in the number of Chimney Swifts in Arvida (town merged with Jonquière in 1975) from 1958 to 1999. Observations between 1958 and 1986 are related to a roost in a supermarket chimney in Carré Davis, the downtown shopping district. In the case of the 1988-1997 observations, the exact location is not specified, but there is every reason to believe that it is the same chimney since Chimney Swifts exhibit strong site fidelity. The number of swifts frequenting this area of town has fallen significantly in the last 40 years. There were approximately 1000 birds in the late 1950s (Browne 1967), a few hundred in the early 1980s, a few dozen at the end of the 1980s and only about 15 since 1991. The population has dropped 99% in 42 years (2.4%/year). The supermarket chimney, the location of the large roost in 1958, was capped in fall 1998. The chimney was still available as a roost before that date, but that did not prevent the number of swifts from declining. A follow-up visit in summer 1999 confirmed this decline; no more than 11 birds were observed entering the chimney of a school in the same area. Since a roost usually draws birds from a large area, this number is indicative of the population in this area of the town since the observers searched it systematically over one entire summer.

The third roost is located at the François Pilote Museum in La Pocatière in the Lower St. Lawrence region. The chimney is as old as the structure, which was built in 1925, but is no longer used for heating. The roost has been used by Chimney Swifts since 1940 and has been protected since then (Tanguay 1964-65). Although the chimney is still available, the number of birds using it has declined significantly since the late 1950s (Figure 6). Only two were seen entering the chimney in 1999, down from the 1200 birds observed in 1958, a drop of over 99.8% in 42 years (2.4%/year). It is highly possible that the number of swifts present in 1957 was over 500 because the observer was only able to witness the start of flocking (R McNeil, pers. comm.). Twenty years ago in the Rimouski region, 100 or so swifts were regularly seen in flight; today just over 20 can be observed after much effort (J. Larivée, pers. comm.).

Including historical data, the species was reported in at least 40 different locations in the Saguenay-Lac St-Jean region between 1971 and 1997 (Savard 1999) and in six or seven municipalities annually in the 1980s (Savard 2000). However, in the summer of 1999, swifts were spotted in only three municipalities (Jonquière, La Baie and Roberval) (Savard 2000). The species seems to have completely deserted Chicoutimi, where it had been present in the past (Savard 2000). In short, all the available sources of information in Quebec show a dramatic and continuing decline in the Chimney Swift population.

According to Erskine (1992), the Chimney Swift population in the Maritime Provinces has declined markedly in the last 30 years. They are less numerous in Prince Edward Island and New Brunswick than in the past. A. Erskine (pers. comm.) also mentioned the species has been less frequently observed and even disappeared from certain areas since the publication of the Atlas of Breeding birds of the Maritimes in 1992. Figure 4 presents maximum numbers counted at three historical roosts in the Maritimes. Chimney Swifts have almost completely disappeared at the University Hall roost (Acadia University), while numbers of birds in the two other sites showed considerable variation over the years with no clear trend. Chimney Swift numbers at the RTNC chimney in Wolfville peaked in 1989 but then declined following a severe cold spell in 1990 (Wolford 1996).


Figure 3: Maximum Number of Chimney Swift in Time at Three Historical Roosting Sites in Quebec

Figure 3: Maximum number of Chimney Swift in time at three historical roosting sites in Quebec.


Figure 4: Maximum Number of Chimney Swifts at Three Historical Roosting Sites in Nova Scotia

Figure 4: Maximum number of Chimney Swifts at three historical roosting sites in Nova Scotia.


In the Atlas of Saskatchewan Birds, Smith (1996) classifies the Chimney Swift as uncommon in Saskatchewan and notes that historical data suggest that the species was once more widespread in the province than it is today. In Manitoba, swifts were also more abundant historically (Taylor et al. 2003). There are no data on population trends in Newfoundland.

Population trends in the Chimney Swift’s winter range are unknown (Cink and Collins 2002). Logging in the Amazon forest could cause problems for the species in the future. This aspect will be addressed in greater detail in Limiting factors and Threats.


Extinction Probability

Gauthier et al. (in press) have undertaken population viability analyses for the Canadian Chimney Swift population. While these analyses use some data that are several decades old and from the southern United States, the results suggest that if only swifts of at least two years of age reproduce, between 40 and 60% of these birds must breed in order to maintain a viable population over the next century. These results not only suggest how a decrease in the availability of nesting sites (or successful reproduction) would influence the viability of the population but also the possible effects of habitat enhancement measures on this species (e.g., building artificial chimneys).

The analyses indicate that if only 30% of birds two years of age and older breed, the chance of extinction over the next 100 years would be higher than 10% regardless of how many first year birds breed. If we use the maximum mean number of fledged young per year, the threshold extinction of 10% in the next 100 years is exceeded only when no first year birds breed. In Quebec, the number of breeding sites now seems insufficient to the point where there are probably very few first year birds that breed.