Connecticut Warbler Oporornis agilis Scientific name definitions

Jay Pitocchelli, Julie L. Jones, and David C. Jones
Version: 2.0 — Published June 2, 2023

Demography and Populations


The global breeding population was estimated at 1,800,000 individuals in Canada and United States from 2006–2015, with 1,700,000 individuals in Canada. Information needed on reproductive success, survivorship, and population regulation.

Measures of Breeding Activity

Age at First Breeding

Information needed.

Intervals Between Breeding

Information needed.

Clutch Size and Number of Clutches per Season

Information needed. Clutch is 3–5 eggs (10); mean 4.0 eggs (n = 6; WFVZ, JB, DJ). Clutch size in eastern Quebec averaged 4.4 eggs (range 4–5, n = 7 nests) (27). There are no reports of breeding pairs laying a second clutch.

Annual and Lifetime Reproductive Success

Information needed; very little data. In Michigan, one nest with 5 eggs fledged 5 young (25). In Quebec, 2 of 7 nests were successful and produced 8 fledglings from 9 eggs. Of the remaining 5 nests, 3 nests were abandoned and 2 were victims of unidentified predators (27). Fledgling success rate was 44% for 4 nests in Quebec (27).

No information on lifetime reproductive success.

Number of Broods Normally Reared Per Season

A single brood per year is most likely across the breeding range but deserves further study. No mention by Saulnier (27) of birds attempting a second brood in a thorough study of several breeding pairs in Quebec, “probably” only a single brood in British Columbia (89).

Proportion of Total Females that Rear at Least One Brood to Nest-Leaving or Independence

Information needed.

Life Span and Survivorship

A bird banded in New Jersey (sex and age unknown) and subsequently recovered dead in Pennsylvania had an estimated minimum age of 4 yr 3 mo (228). This is much shorter than the 7 yr 1 mo estimated for Mourning Warbler (Geothlypis philadelphia), 8 yr 0 mo for Kentucky Warbler (Geothlypis formosa) (229), and 10 yr 11 mo estimated for MacGillivray's Warbler (Geothlypis tolmiei) (229).

Information needed on survivorship.

Disease and Body Parasites

Body Parasites

Hamer et al. (230) found a single larvae of the black-legged tick (Ixodes scapularis) attached to a juvenile Connecticut Warbler during late summer, near Vicksburg, Michigan. Black-legged tick larvae removed from Connecticut Warblers in Michigan were found to be hosts of the spirochete bacterium, Borrrreliia andersonii. Based on this finding, Hamer et al. (231) believed these larval ticks obtained the spirochete from their Connecticut Warbler hosts and that this species is a reservoir host for this bacterium. Connecticut Warbler was not parasitized by four common species of eastern North American ticks during spring and fall migration in Lyme Connecticut, Ixodes scapularis, Ixodes dentatus, Haemaphysalis leporispaustris, and Dermacentor variabilis (232). Two larvae of the rabbit tick (Haemaphysalis leporispalustris) were found on a Connecticut Warbler on Block Island, Rhode Island in 2018; Tufts and Diuk-Wasser (233) suspected that immature stages of this tick may have spread to Block Island via migratory, ground-feeding birds like the Connecticut Warbler.

Connecticut Warbler was not a host of four families of nasal mites (Rhinonyssidae, Ereynetidae, Cytoditidae, and Turbinoptidae) based on individuals sampled in Alberta and Manitoba (234).


Two juveniles in fall migration tested negative for West Nile virus at the Powdermill Nature Reserve in western Pennsylvania (235). The eastern equine encephalomyelitis virus was detected in the blood of a fall migrant in Maryland (236).

Causes of Mortality


Information needed.


Information needed. Eastern Chipmunk (Tamias striatus) may be a nest predator in Quebec (27), few details known; see Behavior: Predation.


Information needed. Few details known; see Disease and Body Parasites.

Direct Human Impacts

There have been numerous reports of birds striking lighthouses, buildings, communications towers, and other structures during migration (see Effects of Human Activity in Collisions with Stationary/Moving Structures or Objects). Juveniles were 5.8 times more likely than adults to be victims of window collisions in Chicago, differences that may be due to adult experience with buildings and windows (237). There were no mortality differences between sexes.

Population Spatial Metrics

Individual Distance

Information needed.

Territory Size

Information needed. Territory size of birds in the Great Lakes region was estimated to be 0.5 ha (USDA report cited in 80).

Home Range Size

Information needed. The average home range size for males (3.05 ha, n = 17) was significantly larger than that for females (1.29 ha, n = 5) (75).

Population Status


Estimates of Population Size

Partners in Flight (238) estimated the global population at 1,800,000 (95% CI: 970,000–2,700,000) individuals in Canada and United States from 2006–2015, with 1,700,000 individuals in Canada. Point-count abundance data collected during the Atlas of Breeding Birds in Ontario estimated the population at 250,000 individuals in Ontario from 2001–2005 (42). The population in northern Wisconsin was estimated at 9,430 individuals in 2010 (239).

Estimates of Local Abundance

The Connecticut Warbler has been part of several larger studies evaluating different methods of estimating abundance, density, and population size. A common theme of these studies is that the Connecticut Warbler is among the rarest species and is the most problematic for making population or density estimates. In Minnesota, densities in closed control (no disturbances) spruce forests ranged from 2.9–5.6 pairs/10 ha (mean 4.2, n = 3 years); higher than in closed treatment spruce forests with transmission lines, where 1.7–2.7 pairs/10 ha (mean 2.1, n = 3 years) (217). Mean numbers of individuals per 100 point counts in the western Great Lakes from 1995–2010 were: 4.64 for Chippewa National Forest, 1.67 Superior National Forest, and 0.73 Chequamegon National Forest. The mean numbers of individuals per 100 point counts from 1991–2010 was 0.26 for Nicolet National Forest (96). Lapin et al. (100) recorded a mean abundance of 4.3 birds/point count (n = 86) over 18 years of observations south of the Cloquet Valley in northern Minnesota. Some of the highest densities in Minnesota were from the Agassiz lowlands at 16.7 pairs/40 ha in semi-productive black spruce-tamarack bogs and 5.1 pairs/40 ha in stagnant black spruce-tamarack bogs. Densities were much lower at 0.6 pairs/40 ha and 0.2 pairs/40 ha in the Chippewa National Forest and Superior National Forest, respectively (240). Mean abundance ranged from 0.03–0.05 on the Taiga (n = 10) and Boreal plains (n = 106) respectively of northeastern British Columbia (241). Abundances in Alberta from 1997–2019 were highest (0.08 birds detected per point count) in undisturbed deciduous habitats 60–80 years old, undisturbed deciduous habitats 80–100 years old, and deciduous habitat 60–80 years post-harvest; abundances were lowest (0.01 birds per point count) in each of the following undisturbed habitats: grass, white spruce 0–10 years old, white spruce 100–120 years old, white spruce 120–140 years old, and shrubby bog (242). Highest abundances in Manitoba were from the Boreal Softwood Shield, east of Lake Winnipeg, ranging from 1.7–30 individuals/15 point counts (243). Lower densities in the 1990s were reported for Manitoba, 0.1–0.9 breeding pairs/10 ha (244). Breeding Bird Survey data suggest low densities, average 19.3 birds/route (n = 3 years) and 11.4 (n = 7 years) birds/route in Manitoba and Saskatchewan, respectively, and 9.7 birds/route (n = 6 years) in Wisconsin (245). Density across Ontario was 0.5 birds/25 point counts. The highest densities of greater than 1.5 birds/25 point counts came from the northern and western regions of the province (246). Blake et al. (247) compared abundance among populations from western Wisconsin and the Upper Peninsula of Michigan; populations varied annually from 1986–1992; breeding populations also varied within each state (local level) and between states (regional level) in the same years.

Information needed on density in overwintering areas, but generally reported to be thinly distributed (e.g., 54).

Methods to Estimate Populations and Field Survey Design

Sólymos et al. (248) were concerned about the limitations of population estimates from Partners in Flight that primarily used roadside data collected through the North American Breeding Bird Survey (BBS). Their primary concerns were about detection distances, limitations of habitat sampling along roads, the effects of roads on bird abundance, and whether roadside surveys were equivalent to off-road surveys. They used a new pixel-based approach (PIX) that predicted abundance at each pixel using bird-count data, land cover, and forest age. The bird-count data was a combination of roadside and off-road surveys from the BBS, Boreal Avian Modeling Project (BAM) and Alberta Biodiversity Monitoring Institute. They found that the PIF underestimated the number of Connecticut Warblers in northern Alberta. The PIF estimate of 198,000 birds was much lower than their PIX estimate of 642,000. They recommended their PIX method over the PIF because of its ability to monitor populations like the Connecticut Warbler that breed in remote areas where BBS are not available

Sólymos et al. (249) noted that population estimates of birds from different studies over large geographic regions could be biased because they failed to take into account the effects of local or regional differences in ecology, day-length, and time of day. Point-count durations also varied among studies (e.g., 3–10 minutes/count). These geographic and duration differences could affect availability or the probability that birds provide visual or auditory cues (e.g., songs and singing rate) of their presence. Availability estimates are important for adjusting survey counts and making them comparable. They studied availability estimates of the Connecticut Warbler and 151 other boreal species using different time-removal models. They compared availability estimates from a conventional-removal model versus a finite-mixture-removal model. The conventional model assumes constant singing rate of all birds while the finite-mixture model takes into account singing rate differences among birds and proportions of birds singing at different rates. They found that both models produced similar availability estimates for the Connecticut Warbler. They recommended using the finite-removal-mixture model because it better takes into account differences in availability caused by geography, singing rate, and count duration when using data across broad geographic scales.

Crosby et al. (250) pointed out a problem with estimating densities and spatial distribution of the Connecticut Warbler given that it shows regional differences in breeding habitat (differential habitat selection). It occurs in wet, coniferous forests in the eastern part of its breeding range, and in drier, deciduous forests in the western part of its range. The three regions in this study were: Western (northern Alberta), Central (western Ontario), and Eastern (southern Quebec). Each region had different temperatures, precipitation amounts, and dominant vegetation. Crosby et al. (250) noted that typical models assume that habitat selection is similar across all breeding populations. They showed that models that do not incorporate differential habitat selection result in overestimates of breeding densities for several species of boreal warblers, including the Connecticut Warbler.

Toms et al. (251) found a linear relationship between point-count abundance and the actual number of territorial males from spot mapping with grids in northern Alberta, but urged caution using point-count methods to estimate abundance of Connecticut Warblers and other species because of sampling errors at the level of point-count stations. On-road bird counts similar to those used by the North American Breeding Bird Survey tended to underestimate Connecticut Warbler densities compared to off-road counts made > 200 m from roads (252).

Efforts have been made to try to improve the usefulness of point-count methods to estimate abundance and population size. Adjustments to the number of birds counted have been proposed to take into account the fact that some birds could be missed on a point-count survey because of a declining probability of being detected as distance from the observer increases. Matsuoka et al. (253) compared different adjustment techniques used to estimate detection distances from point-count surveys and their effects on landbird population size estimates. They compared effective detection radius (EDR) using binomial distance sampling versus maximum detection distance (MDD) used by Partners in Flight (PIF) from across Canada. Average EDR’s for Connecticut Warblers were 75.0 m in open deciduous forest, 74.5 m in closed conifer forest and 73.7 m in non-forest habitats, compared to 200 m from MDD and the PIF classification system. Population size estimates of the Connecticut Warbler using EDR were 6.7 times greater than MDD estimates.

Hobson et al. (254) studied the use of omnidirectional microphones in acoustic surveys of birds. They compared the number of bird encounters between field observers versus recordings made simultaneously in the same area during the survey. They reported comparable numbers of encounters of Connecticut Warblers between field observers and recordings. Follow up study of variability among experts identifying birds from recordings made in northwest Ontario showed an increase in identification errors made for rare species like the Connecticut Warbler (255).

Venier et al. (256) compared the results of boreal forest bird surveys using 2 different automated recording units (ARUs), SM-1 bird song recorder, E3A Acoustic Monitor Kit, versus observers on the ground. Only three Connecticut Warblers were detected: 1 by the SM-1 bird song recorder versus 2 by observers. Based on these data and the other 65 species in this study, the SM-1 slightly underperformed the observers and the E3A Acoustic Monitor Kit. However, the differences were not considered important enough to rule out using ARUs in the future, particularly in remote areas or in cases where few observers are available. More support for using ARUs came from a study by Van Wilgenburg et al. (257). They compared detection data from ARUs versus human observers point-count detections for 35 boreal forest species in the Boreal Plains and Boreal Shield of northern Saskatchewan. Initial estimates using human observers underestimated Connecticut Warbler densities by 0.02 birds/ha compared to results from ARUs. However, after applying correction factors for detection radius estimated from Generalized Linear Mixed Models, they did not find differences between the two survey methods. They concluded that ARUs could be useful for quantifying a species’ abundance in remote areas, but point-count data from these ARUs for reclusive species like the Connecticut Warbler would need corrections for the detection radius to be effective.


Based on BBS data, the survey-wide population experienced a long-term (1966–2019) decrease of –0.8% per year, though the 95% confidence interval (–2.3, 0.6) included zero, indicating that it was not statistically significant (n = 262 survey routes) (258). Nonetheless, over a 45-year period (1970–2014), an analysis of BBS data indicated that the Connecticut Warbler population declined by an estimated 60% (259), illustrating that a statistically insignificant decline can amount to a large change in population size over a longer time period. Hallworth et al. (64) found declines in 8 natural populations of the Connecticut Warbler in Canada and the Great Lakes region of the United States based on BBS data. They estimated that different populations declined by 12.48% to 5.02% per year from 2000–2017. They estimated that the entire population has declined approximately 70% since 1966. Rosenberg et al. (259) also predicted that in less than 50 years the population size will reach 50% of its current numbers.

Walker and Taylor (260) reported declines using models of eBird and BBS data. They studied how well annual indices from eBird data could be used to track population trajectories in 22 boreal and Arctic species that included the Connecticut Warbler. The spring and fall trajectories of annual percent changes from the eBird data were in general agreement with each other (260). Models using checklists and eBird data from the time series 1928–2015 indicated a negative annual percent change of –0.79% in the spring while the fall data suggested a similar annual decline of –0.99%. Models from 1970–2015 also showed declines of –0.84% and –1.11% in the spring and fall respectively. There was a weak but significant positive correlation of 0.274 between the amount of forest defoliated by spruce budworm (Choristoneura fumiferana) and annual populations indices from eBird data during the fall, suggesting that Connecticut Warbler may have benefited from budworm defoliation. There was no similar relationship during the spring.

Niemi et al. (96) reported significant declines for Connecticut Warbler in northern Minnesota from 1995–2011. The average annual percent changes were –7.11% and –8.42% in the Chippewa National Forest and Superior National Forest, respectively. A subsequent analysis for 1995–2018 by Grinde et al. (261) reported similar results: declines of 8.01% and 7.04% for Chippewa and Superior National Forests, and a regional decline of 2.86%. Kovatch (262) studied BBS data in Saskatchewan and Manitoba from 1970–2011 in order to describe how bird populations responded to habitat loss in Moist-mixed Grassland and Aspen Parkland regions due to fire and grazing. She reported a decline of –3.54% annual trend in Connecticut Warblers over this 40-year period.

Modeling the impact of climate change on abundance forecasted a rapid decline for the Connecticut Warbler in Alberta by the year 2100: –25% from 2010–2040, –61% from 2041–2070, –84% from 2071–2100 (72). The mid-century declines from 2041–2070 by natural region would be highest in the grasslands and foothills at –66% and –94% respectively and lowest in the mountains at –21%.

Population Regulation

Information needed. Densities of breeding birds may be regulated by size and isolation of breeding habitat in Saskatchewan. Johns (86) found that breeding density of Connecticut Warbler in aspen groves decreased with increasing distance between groves, and that the species preferred larger groves 3 ha or larger.

Recommended Citation

Pitocchelli, J., J. L. Jones, and D. Jones (2023). Connecticut Warbler (Oporornis agilis), version 2.0. In Birds of the World (P. G. Rodewald, Editor). Cornell Lab of Ornithology, Ithaca, NY, USA. https://doi.org/10.2173/bow.conwar.02