Migratory movements of the Osprey across North America.
This animation depicts weekly relative abundance estimates, with brighter color representing higher abundance. The data for this animation were generated using Spatio-Temporal-Exploratory Models to predict population relative abundance at specific times and locations by relating observations of Ospreys from eBird to local environmental features derived from NASA remote sensing data.eBird (eBird.org) is a citizen science program run by the Cornell Lab of Ornithology.
Distribution in the Americas
Canada and United States . Figure 1. The breeding range is widespread and expanding. Indeed, the distribution in 2015 is considerably different from what it was only 15 years earlier. There are now breeding records or nesting attempts for all of the lower 48 states of the U.S. Broadly speaking, as at the turn of the 21 st century, the core of the population continues to be found along the Gulf and Atlantic coasts from eastern Texas through Florida (with the exception of the Everglades ( ) to Newfoundland, with highest densities along the Atlantic coast, particularly around Chesapeake Bay, where the population may exceed 9,000 pairs (B. Watts, personal communication). Florida Fish and Wildlife Conservation Commission 2003
From the east, the range extends across forested Canada and the Great Lakes region from the Avalon Peninsula in Newfoundland to Alaska, where breeding occurs to treeline south of the Brooks Range, and as far northwest as the Selawik River (C. MacIntyre, personal communication). Ospreys are largely absent in extreme western Alaska and the mountainous region from the eastern Kenai Peninsula to Glacier Bay (
). Walton et al. 2013
Range extends south through British Colombia (
) and western Alberta through appropriate habitat in Washington ( Campbell et al. 1990a ) and Oregon ( Smith et al. 1997 ) to the Sierras of California and into Nevada east of Lake Tahoe (Douglas, Carson City, and Lyon counties) ( Adamus et al. 2001b ); and through the Rocky Mountains in western Montana ( Floyd et al. 2007 Montana Natural Heritage Program), Wyoming, and Colorado and through most of Idaho (C. Molleen, personal communication), with high concentrations along Lake Coeur d’Alene ( ) into central Utah east of Salt Lake at least to the Kolob Reservoir (S. Lindsey, personal communication) and Lake Powell (C. Monson, personal communication), with a large population in the Flaming Gorge area on both sides of the Utah/Wyoming border (S. Lindsey, personal communication). Stephens and Sturts 1997
Across much of the continent, Ospreys have been expanding their range. In Canada, the range has expanded northward. Nests have been reported near the Labrador coast at 58.9°N (T. Chubbs, personal communication); in Churchill, Manitoba (
Manitoba Breeding Bird Atlas); in Yellowknife, Northwest Territories at 62.4°N ( ); and in Inuvik, Northwest Territories at 68.3°N ( Poole 1985b www.osprey-watch.org/nests/1477).
Along the coast of California, the range has expanded south to include a growing population around San Francisco Bay (
), and two small disjunct populations from San Diego to Irvine ( Brake et al. 2014 www.osprey-watch.org) and east and north of San Bernadino (L. Fields, personal communication). In the Sierra Nevadas breeding occurs along the western slope as far south as Bass Lake (37.3°N; www.osprey-watch.org), and a small population is established nesting on tufa towers at the fishless, hypersaline Mono Lake ( ). Fields and Pagel 2016
Along the eastern Rockies, the Montana population has expanded east as far as Fort Peck along the Missouri River and along the Yellowstone to Miles City (
Montana Natural Heritage Program) and from there down through the Bighorn valley into north-central Wyoming (M. Restani, personal communication). In Colorado the core population occurs in the northern Rockies out onto the east slope, especially from Fort Collins to Boulder in the north and along the Colorado River as far east as Pueblo ( ). Isolated pairs occur in northeastern Colorado and eastern Wyoming. A separate population is found in southwestern Colorado, perhaps contiguous with the species’ limited range in northern New Mexico around El Vado, Heron, and Navajo reservoirs (F. Stahlecker, personal communication). Dwyer 2016
In the Southwest, scattered pairs occur in central New Mexico (e.g., Cochita Reservoir, Fenton Lake, and Ramah Lake; F. Stahlecker, personal communication) and in central Arizona, where 25 breeding bird atlas quadrats have confirmed breeding records (
). Corman and Wise-Gervais 2005
In Saskatchewan, the range now extends south into the prairie region to the North Saskatchewan River and the northern arms of Diefenbaker Lake (M. Stoffel, personal communication). In Manitoba, breeding range has moved south to the southern suburbs of Winnipeg and west of Dauphin Lake and Winnipegosus Lake (
In eastern Canada, the range has extended south in Quebec (
www.atlas-oiseaux.qc.ca/donneesqc/cartes.jsp?lang=en) and Ontario (D. Lamble, personal communication) to Lake Ontario and Lake Erie and is contiguous with the population in Michigan, New York, and New England.
In the northeastern U.S., the formerly disjunct population in Long Island and southern New England has expanded west to merge with the population in New Jersey and up the coast as well as inland through much of Connecticut, Massachusetts as far west as the Connecticut River, merging with northern populations (
), where the species is now breeding in all of New Hampshire (C. Martin, personal communication) and northern Vermont, with an estimated 150 to 175 pairs nest in the Champlain Valley (J. Gobeille, personal communication). Bierregaard et al. 2014a
Maine has never conducted a statewide Osprey survey; available data are mostly incidental to other surveys or from local monitoring efforts. Based on current surveys and nests last documented in the 1970s, distribution is virtually statewide, but scarce in mountainous regions, and locally high density in some coastal regions. Population may exceed 1,500 pairs (C. Todd, personal communication). Casco Bay had 133 pairs in a 2012 survey (
). Distribution may be shifting inland owing to local increases in Bald Eagle population (C. Todd, personal communication). DeSorbo et al. 2013
In upstate New York, nesting Ospreys were formerly primarily concentrated in the Adirondacks, but now occur south along the Hudson River as far as West Park (M. Grant, personal communication) and west through the Finger Lakes to the shores of Lake Erie (
; L. Harper, personal communication) and south into northern Pennsylvania in Elk, Potter, and Tioga counties ( McGowan and Corwin 2008 ; Wilson et al. 2012 www.pabirdatlas.psu.edu/map/species_maps).
Through the Mid Atlantic states, breeding pairs are now found along rivers and reservoirs from southeastern Pennsylvania east of the Appalachians to Georgia. Similarly, most large lakes and reservoirs across the Gulf coast states host breeding Ospreys, especially along the Tennessee River in northern Alabama (R. Goss, personal communication). Scattered nesting records for larger lakes and reservoirs also are reported for Arkansas (
In the Midwest, nesting occurs across much of central Tennessee and Kentucky, southern and eastern Ohio, southwestern Pennsylvania, and northern West Virginia. Small, isolated populations occur in southwestern Missouri (
www.osprey-watch.org/nests, G. Swick, personal communication) where a reintroduction project was carried out, and Iowa ( www.osprey-watch.org/nests).
Around the Great Lakes, breeding occurs in all but southwestern and northwestern Minnesota (
mnbba.org/blockmap/cresults.php?species=Osprey), all but extreme southwestern Wisconsin and the Bayfield Peninsula ( www.uwgb.edu/birds/wbba/species/maps/OSPR.htm), and all but east-central Michigan and the Keweenaw Peninsula ( mibirdatlas.org/Portals/12/MBA2010/OSPRcompare.pdf). The noteworthy absence on the Keweenaw Peninsula may be related to high densities of nesting Bald Eagles (S. Postupalski, personal communication).
In the central prairie states, from North Dakota to Oklahoma, formerly unoccupied, scattered pairs have bred in North Dakota with recent unsuccessful attempts ongoing (L Igl, personal communication), a small population exists in western South Dakota (resulting from relocation efforts) mostly south and west of Rapid City (
www.osprey-watch.org/nests), and at least 10 nesting attempts with no confirmed successes have been made in Kansas (R. Mangile). Recent nesting attempts in the reservoirs of eastern Oklahoma (S. Sherrod, personal communication) suggest the range has expanded from neighboring Arkansas.
Mexico and Caribbean. A disjunct population of more than 1,000 pairs occurs on both coasts of the Baja California and the eastern shores of the Gulf of California in the Mexican states of Sonora and Sinaloa from roughly 31°N to the tip of Baja California ( ) and at least as far south as Mazatlán, ca. 23.2°N ( Henny et al. 2008b ). Another population ( Howell and Webb 1995 P. h. ridgwayi) is found on the east coast of Yucatán Peninsula (Quintana Roo) and south into Belize; ridgwayi also resident in the southern Bahamas (north to Exuma and Cat Island) and Cuba (northern cays, Isla de Pinas, Zapata Swamp, and coastal Oriente) ( ). Wiley et al. 2014 Overwintering Range
Bulk of North American population overwinters south of the U.S. in Mexico and Central and South America; overwintering birds distributed by geographic origin of nesting grounds and sex (
, Poole and Agler 1987 , Poole 1989a ); see also Migration: segregation on overwintering grounds, below). Most Ospreys nesting in western North America overwinter from southern Texas through Mexico south through Central America; rarely found in South America ( Washburn et al. 2014 ) and Cuba ( Johnson and Melquist 1991 ). Midwestern U.S. breeders found from Mexico into South America, with a small number in the Caribbean. Eastern breeders overwinter in small numbers in the Caribbean and predominantly in South America. Banding data show 22% in Caribbean and 78% in South America ( Bedrosian et al. 2015 ), but satellite tagged birds show higher proportion in South America ( Poole and Agler 1987 ). Washburn et al. 2014
United States. Scattered numbers overwinter along the Pacific coast from (rarely) northwestern Washington and Oregon (increasingly since the early 1990s; ) south to southern California and locally inland east to western foothills of Cascades and western edge of southeastern California deserts ( Gilligan et al. 1994 , Christmas Bird Count [CBC] data); east to Phoenix and Tucson, Arizona (CBC data); more abundantly along the entire Gulf coast and locally inland throughout much of eastern Texas as far north as Dallas, most of Louisiana, central Mississippi and Alabama; throughout Florida, and along the Atlantic coast (and very locally at inland sites) to the mouth of Chesapeake Bay (CBC data). Small 1994
Ospreys increasingly observed in recent years (at least through early winter) at scattered inland and coastal locations throughout the U.S. to extreme southern British Columbia, Alberta, Ontario, Quebec, Nova Scotia, and Newfoundland (
, Campbell et al. 1990a , Semenchuk 1992 , CBC data). Birds found north of the indicated overwintering range (Figure 1) may be in poor health or otherwise compromised. DesGranges 1996b
Mexico and Central America. Overwinters locally along the entire Pacific, Gulf of Mexico, and Caribbean coasts of Mexico and Central America, extending inland to eastern Coahuila and eastern San Luis Potosí in northeastern Mexico, throughout Isthmus of Tehuantepec, and Yucatán Peninsula ( ), but most numerous from Sinaloa and Veracruz south ( Howell and Webb 1995 , Martell et al. 2001a ). Martell et al. 2014
Caribbean. Overwinters throughout the Caribbean ( , Wiley 1984 ), but is rare in the Lesser Antilles east of Virgin Islands ( Poole 1989a ). Raffaele et al. 1998
South America. Overwinters locally throughout South America south to central Chile and Buenos Aires, Argentina ( , Poole and Agler 1987 , Martell et al. 2001a , Saggese et al. 2014 ). Band recoveries and satellite-telemetry data suggest most overwinter north of the equator, although significant numbers migrate farther south. Satellite data show a widespread dispersal across South America, suggesting a bias in band recovery data towards areas of higher human population density, especially the Magdalena River valley in Colombia ( Washburn et al. 2014 ). Mestre and Bierregaard 2009 Other Records
Casual in Aleutian Islands, Pribilof Islands, St. Lawrence Island (Bering Sea), Seward Peninsula, northern Quebec, Guadalupe Island (off Baja California), and Clipperton Island (1,000 km southwest of Colima, Mexico) (
) and encountered in all months in Bermuda (see American Ornithologists' Union 1998a eBird).
Distribution Outside the Americas
Breeding and Overwintering Ranges
Worldwide distribution; Nominate subspecies (
P. h. haliaetus) nesting across the Palaearctic from Subarctic (although north of the Arctic Circle in Finland; P. Saurola, personal communication) south to about 40°N. Found from the United Kingdom (predominantly Scotland), Scandinavia, the Baltics, France, Germany ( ), and Poland ( Hauff 1996 ), east across northern Russia and Asia to Kamchatcka and Japan. Reintroduction programs underway in Portugal, Spain, and Switzerland. Mizera and Szymkiewicz 1996
haliaetus Ospreys overwinter south of 20°N in Africa, India, and Southeast Asia.
Nonmigratory populations are found in the western Mediterranean (Corsica, Balearic Islands, Gibraltar, and coastal Morocco and Algeria), Red Sea, Persian Gulf, and Cape Verde (80 pairs) and Canary Islands (7 pairs) (
, Cramp and Simmons 1980a ). Coastal, resident populations ( Schmidt-Rothmund et al. 2014 P. h. cristatus) found in Australia and islands stretching from New Caledonia to Java ( , Poole 1989a , Marchant and Higgins 1993 ). Dennis and Clancy 2014 Other Records
Very rarely encountered in Greenland, Iceland, Faroe Islands, and eastern Atlantic islands (
). American Ornithologists' Union 1998a
Migration from North America displays broad spatial and temporal distribution, evidenced by data from band returns (
, Henny and Van Velzen 1972 , Poole and Agler 1987 ) and satellite telemetry ( Mestre and Bierregaard 2009 , Martell et al. 2001a ). Most of North American breeding range is migratory, with non-migratory individuals in southern Florida, the Caribbean, southern California, and the Baja Peninsula. Martell et al. 2014
Individuals typically migrate singly, except at concentration points.
reported that 80% of Ospreys sighted during fall over New York State were alone, 11% seen with other raptors (primarily Broad-winged Hawks [ Kerlinger and Moore 1989 Buteo platypterus]), with maximum flock size 2 individuals. A flock of 11 was reported at Cape May, New Jersey ( ). Larger flocks (up to 92 individuals) have been reported in Cuba ( Dunne et al. 1988 ) and Haiti ( Rodríguez-Santana et al. 2014 ) where the broad-front migration from eastern and (partly) central North America is concentrated as it passes through the Caribbean. Crouse and Keith 1999
Satellite telemetry data show: (1) male-female differences in timing of migration (females leave up to a month before males), distance traveled, and overwintering locations; (2) individuals followed for over 1 yr by satellite telemetry showed unwavering fidelity to overwintering sites and a general fidelity to migration flyways, but not to specific routes; (3) breeding pairs did not migrate or overwinter together (
, Martell et al. 2001a ); and (4) adults do not migrate with their offspring (MSM). Martell et al. 2014
Band recoveries (
) and satellite-telemetry data ( Poole and Agler 1987 , Martell et al. 2001a ) show separate, parallel migration routes for eastern, midwestern, and western populations (see Routes of Migration, below). Midwestern migration routes overlap with both eastern and western birds; with rare exceptions, there is no overlap between eastern and western routes. Also with rare exceptions, eastern and western breeding populations show no overlap on overwintering areas, whereas midwestern birds overlap with both eastern and western populations. For Ospreys nesting along the east coast of North America, most band returns for the overwintering period have come from South America and, to a lesser degree, the West Indies ( Martell et al. 2014 , Henny and Van Velzen 1972 ); over 50% were from locations east of 75°W (e.g., central Colombia). Some 19% of overwinter band recoveries of eastern individuals were from U.S., although the health of Osprey overwintering at temperate latitudes has been questioned ( Poole and Agler 1987 ); nonetheless, recent data from satellite-tagged individuals show that some viable individuals are known to overwinter in the southern U.S., especially Florida (ROB). Poole and Agler 1987
The fall migratory period extends from mid-July (Florida) to early November, spring migration from late February to May for northern birds. Florida birds migrate south in July and August, return in January and February (
). Band returns ( Martell et al. 2004 , Henny and Van Velzen 1972 , Poole and Agler 1987 ) and sightings of individuals on overwintering areas during the northern summer indicate first-year birds remain on overwintering grounds ca. 18 mo, only returning in their third calendar year. Osprey very rarely return to breeding grounds in the spring following their first migration south (AFP; M. McMillian, personal communication). Ewins and Houston 1992
Migrants tend to avoid overwintering in areas where non-migrant populations breed, but there is some overlap of breeding and overwintering ranges in the southern U.S. (see range map). Band returns (
) and satellite-telemetry data ( Poole and Agler 1987 ) do not indicate “leap-frogging” of southern nesters by northern nesters, at least in the East (excluding non-migratory populations). Martell et al. 2001a
Of 79 Ospreys tracked by satellite, 30.4% overwintered on coasts, 50.6% overwintered on rivers, and 19% overwintered on lakes or reservoirs, with differences based on both sex and region of origin (
). Washburn et al. 2014
Timing and Routes of Migration
Fall.Typically begins with adult females leaving in August ( , Poole and Agler 1987 ), although individuals (generally failed breeders) may leave earlier and wander widely before beginning true migration. Some individuals from failed nests commute > 100 km sporadically between their nesting area and alternative fishing sites prior to migration (ROB). Median departure date for 45 adult females satellite tagged from South Carolina to Maine was 19 August (12 July–21 September). From the same region, 54 tagged males had a median departure of 15 September (16 August–12 October) (MSM, ROB, B. Washburn, unpublished data). Later departures of males is likely due to males continuing to feed young postfledgings. Martell et al. 2001a
September reported as the migratory peak from New York to Maryland (
, Henny and Van Velzen 1972 ). Four nesting females satellite-tracked from southern Florida began migration in late July ( Poole and Agler 1987 ). Twenty-six Minnesota adults tracked by satellite telemetry had a median departure date of 7 September (range 1 August–1 October; Martell et al. 2004 ). Seventeen adults from Oregon and Washington tracked by satellite telemetry had a median departure date of 9 September (range 28 August–24 September; Martell et al. 2001a ). Departure from northern Idaho through September and early October ( Martell et al. 2001a ). Melquist et al. 1978
Median date of arrival in overwintering areas for satellite-tracked breeding adults from New York and New Jersey (
n = 21) was 14 September (range 30 August–18 November); from Minnesota ( n = 19) 9 October (7 September–5 November); from Oregon and Washington ( n = 15) 20 September (7 September–14 October). Individuals arrived on overwintering sites an average of 10 d apart in subsequent years (range 0–20, n = 7); mean difference in arrival dates between years was 12.6 d ( n = 3) for males, and 6 d ( n = 4) for females ( ). Martell et al. 2001a
Observations and satellite-telemetry data on entire family groups indicate that females left breeding territories before their young were completely independent (MSM). Median dates of arrival in overwintering areas were earlier for adult females than for males satellite-tracked from the East and Midwest. From the East Coast, 60% of females and only 25% of males arrived by the median date for the population. In the Midwest, 75% of females and 45% of males arrived by the median date. In 6 pairs, females arrived on average 36.8 d (range 20–70) ahead of their mates; in 3 pairs, males arrived an average of 9 d (range 2–20 d) before females.
Spring.Individuals return to breeding areas as early as January in southern Florida ( ) and as late as June in Labrador (J. Pfeiffer, personal communication). Most residents are on territories in San Francisco Bay, California, by February (D. Rooney, personal communication). Most territories are settled by late March in lower Chesapeake Bay (B. Watts, personal communication). In Kentucky, individuals arrive from early March to late April ( Martell et al. 2004 ); in northern Idaho, from late March to early April ( Palmer-Ball 1996 ). Individuals return to nest sites along southeastern Lake Huron around the second week of April, before total ice-out ( Melquist et al. 1978 ). Returns in Wisconsin peak from mid-April to mid-May ( Ewins and Cousineau 1994 ), mid-April peak in central Alaska ( Robbins 1991 ). In southeastern Massachusetts, first arrivals second week of March; on the Westport River, Massachusetts, 25% of the population is back by 1 April, 70% by 15 April, 95% by 30 April (AFP). Timm et al. 2012
Of 67 median spring departure dates for 49 satellite tagged individuals: East Coast (
n = 39) males 18 Mar (25 Feb–7 Apr), females 8 March (9 February–6 April); Midwestern ( n = 10) males 24 March (16 March–1 April), females 15 March (10–22 March); Western ( n = 18) males 10 March (13 February–19 April), females 1 April (10 March–11 April) ( ). Martell et al. 2014
) as well as satellite data show that juvenile Ospreys do not leave their overwintering areas until their second spring -- after 18 months at their overwintering site. Henny and Van Velzen 1972 ) reported that some European birds remain in Africa until their third spring, and South American band recoveries of 3- and 4-yr old North American Ospreys in May-July indicate an even further delayed return migration occurs occasionally here as well. That said, no satellite-tagged Osprey in North America or Europe has stayed in the overwintering range more than 18 mo. Prevost 1982 Routes
Fall, through the U.S. Breeders along the Eastern Seaboard–from Nova Scotia, Prince Edward Island ( ) and Labrador ( Ewins and Houston 1992 ) south–show a broad-front migration along the U.S. Atlantic Coast, from the Atlantic shore to Appalachian ridges ( Laing 2005 , Henny and Van Velzen 1972 , Poole and Agler 1987 ). Satellite data show some individuals from New England and Long Island, New York, follow the coast from Cape May, New Jersey, across Delaware and Chesapeake bays, continuing through eastern Virginia to the Outer Banks of North Carolina. Others fly inland of Delaware and Chesapeake bays arriving at the eastern seaboard around Charleston, South Carolina, or Savannah, Georgia. ( Martell et al. 2001a , ROB). Most adults that reach the Outer Banks cross the Georgia Bite from the North Carolina coast to Florida, a trip of 650–800 km, which typically takes 14–18 h (see Martell et al. 2001a www.ospreytrax.com/OspreyMainPage.html for details), but do not repeat that route the following spring. Neither band returns ( ) nor satellite-telemetry data indicate Ospreys breeding along the Eastern Seaboard move through northern Mexico or across the Gulf of Mexico, as is the case for some raptors, but migrant Ospreys do cross the Caribbean, see below. Poole and Agler 1987
Juveniles migrate alone and follow the general routes taken by adults, with notable exceptions. Over 50% of 21 satellite-tagged juveniles from Martha’s Vineyard, Massachusetts, flew due south over the Atlantic non-stop to the Bahama Islands, crossing as many as 2,400 km in flights as long as 58 h (
). No adults tagged on Martha’s Vineyard and no juvenile making a second migration south took this overwater route, suggesting that juveniles learn the coastal route on their first trip north (ROB). Horton et al. 2014
Presumably, juveniles from Nova Scotia and Newfoundland make similar long distance trans-Atlantic crossings, which would explain the frequency with which they are reported in Bermuda (see
eBird). Some juveniles, but no adults, that leave the east coast at the Outer Banks of North Carolina arrive in the Bahamas, rather than the Florida Coast, further indication that the “adult route” is learned over the course of several migration cycles.
Satellite-tracked breeders from Minnesota (
n = 27) moved along 3 pathways: (1) south through central Iowa and Missouri, then southwest through Oklahoma and Texas to the gulf coast of Mexico; (2) south along the Mississippi River and across the Gulf of Mexico; and (3) southeast through Illinois, across Kentucky, Tennessee, and Alabama and into Florida ( ). Minnesota juveniles ( Martell et al. 2001a n = 7) followed this same pattern (MSM).
Nestlings banded in Saskatchewan and British Columbia were recovered from the central U.S. to the northwestern shores of the Gulf of Mexico (
). Satellite-tracked adults nesting along the Columbia River and Willamette Valley in Oregon ( Ewins and Houston 1992 n = 15) moved south along a broad front through the western U.S. Eleven traveled directly south through central California to the west coast of Mexico; 4 headed southeast through Nevada, Utah, and Arizona, with 1 of these 4 moving as far east as eastern Utah and into New Mexico ( ). Adults tagged in British Colombia ( Martell et al. 2001a n = 31) followed the same broad front across the western U.S. ( ). Elliott et al. 2007b
South of the U.S.Satellite-telemetry data show virtually all eastern breeders, and some midwestern, crossing the Florida Keys to the north coast of Cuba, 100–250 km east of Havana, then flying southeast across Cuba before crossing to Hispaniola ( ). Only 2 out of 45 satellite-tagged adults crossing the Caribbean flew from Cuba to Central America, rather than to Hispaniola (ROB). Movement of 60 individuals over Gonaives Bay, Haiti, was noted on 28 Aug 1999 ( Martell et al. 2001a ). Crouse and Keith 1999
From the Dominican Republic, roughly 90% of satellite-tagged adults depart from Cabo Beata (17°36’N; 71°25’W), crossing approximately 600-km of the Caribbean Sea to South America (ROB). About 80% of these make landfall between Santa Marta, Colombia, and the Guajira Peninsula at the western mouth of the Gulf of Venezuela (
www.ospreytrax.com/OspreyMainPage.html). Returns of birds banded in North America have been reported from the islands of Cuba, Hispaniola, Trinidad, Puerto Rico, and the Bahamas ( , Santana and Temple 1987 ). Rodriguez et al. 2001
After landfall in South America, the main front of migration is east along the northern coast and southeast through Venezuela into Brazil, although final destinations vary. Based on band returns a significant number overwinter in the Magdalena Valley in northern Colombia (
), although the relative importance of this area may be overestimated because of the high density of human population there and related shooting ( Poole and Agler 1987 ). Mestre and Bierregaard 2009
Whereas 90% of satellite-tagged adults crossing the Caribbean left Hispaniola at Cabo Beata (the shortest route to South America), most naïve juveniles on the East Coast flew to the east end of Hispaniola and only headed south at land’s end in the Dominican Republic.
Some Midwestern breeding adults and juveniles tracked by satellite telemetry crossed the Gulf of Mexico to the Yucatán Peninsula, then south through Central America into Colombia and Peru. Others moved through Mexico along the Gulf Coast over Veracruz and Chiapas, then along the west coast of Honduras, El Salvador, and Nicaragua, crossing to the east coast of Costa Rica and Panama into Colombia or Peru. A third group moved southeast, joining the Atlantic flyway to overwinter in Cuba, or moved through Hispaniola to overwintering sites in Venezuela, Colombia, Brazil, and Bolivia (
, Martell et al. 2001a ). Less than 1% of band returns for midwestern (primarily Michigan and Wisconsin) Ospreys from west of 85°W (e.g. central Nicaragua; Washburn et al. 2014 ). Poole and Agler 1987 reported band recoveries of Wisconsin birds from Haiti, as well as Central and South America. Robbins 1991
Adults from British Columbia, Saskatchewan and Oregon continued their broad front movement from the U.S. into Mexico, moving to both the Pacific and Gulf coasts of Mexico. Some overwintered there (between 25°95’N and 15°15’N), while others continued into Central America as far south as El Salvador, with one bird as far as Venezuela—the only bird from western North America known to have overwintered in South America (
, Martell et al. 2001a , MSM and S. Houston, unpublished data). No band returns ( Elliott et al. 2007b ) indicate that western-breeding Ospreys use the West Indies for migration or overwintering, but one satellite-tagged bird from Wyoming overwintered in Cuba ( Poole and Agler 1987 ). Bedrosian et al. 2015
Satellite telemetry found no difference in fall or spring migratory routes by sex, although females tend to overwinter south of males from same breeding population (
). Females spent longer on their overwintering sites than males (167.0 ± 3.2 d v. 153.6 ± 3.7 d, Martell et al. 2001a n = 79) ( ). Washburn et al. 2014
Spring.Across all three populations, satellite-tracked adults generally followed the same broad fronts (but not specific routes) going north in spring as they took south in fall. The spring migration route for eastern birds was shorter than the fall route and the front was as much as 50% narrower than in fall, suggesting the birds expended more energy fighting wind drift en route to their breeding sites than travelling to overwintering areas ( ). Washburn et al. 2014
Three examples of “loop migration” were documented in satellite-tagged adults. One adult female migrated south from British Colombia through Idaho and Utah to western Mexico in the fall and returned via a route through California the following spring that was 700 km west of her track in fall (
). One adult male from the Midwest and one from Long Island, New York, migrated through Cuba and Hispaniola to South America in the fall and north through Central America in the spring ( Washburn et al. 2014 ) ( Washburn et al. 2014 www.ospreytrax.com/2015%20Spring%20All%20Birds.html).
The tracks of satellite-tagged juveniles on their first trips north were often vastly different from their tracks in their first fall, 18 mo before. No bird that crossed the Atlantic from Martha’s Vineyard to the Bahamas repeated that route. Rather, individuals either island-hopped through the Bahamas or traveled northwest through Cuba. Just as many missed the shortest route to South America in the fall leaving Hispaniola, most did not turn north in time to cross the Florida Straits (150 km) but rather continued on to western Cuba where they then had to make a 700–900 km crossing of the Gulf of Mexico (ROB).
Segregation during migration and overwintering periods
Geographic origin. Band recoveries ( ) and satellite-telemetry data ( Poole and Agler 1987 ) show separate, parallel migration routes for eastern, midwestern, and western populations (see Routes of migration, above). Midwestern migration routes overlap with both eastern and western birds; no overlap between eastern and western routes. Eastern and western breeding populations show no overlap on overwintering areas, while midwestern birds overlap with both eastern and western populations. Most winter-season band returns of Ospreys nesting along east coast of North America have come from South America and, to a lesser degree, the West Indies ( Martell et al. 2001a , Henny and Van Velzen 1972 ); > 50% were east of 75°W (e.g., central Colombia). Poole and Agler 1987 reported that 19% of winter band recoveries from eastern individuals were from U.S., although they questioned whether any Osprey overwintering at temperate latitudes is healthy. Nonetheless, 1 satellite-tracked male from Maine and 1 from New York overwintered near Vero Beach, Florida, in 1999 and 2000, while a breeding female from Maine overwintered near Naples, Florida (MSM). Midwestern winter band recoveries mostly from South and Central America ( Poole and Agler 1987 ). Winter band recoveries of western-breeding Ospreys from Mexico and Central America through Panama, with most occurring west of 85°W ( Niemuth 1991 , Melquist et al. 1978 ). Single band recoveries of western breeders from southern Texas and Ecuador ( Poole and Agler 1987 ). Satellite tracking of Ospreys breeding in British Columbia showed overwintering in southeastern Texas and Gulf coast of Mexico (J. Elliott, personal communication). Sixteen Oregon breeders tracked by satellite overwintered from Mexico ( Johnson and Melquist 1991 n = 14) to El Salvador and Honduras ( ). Band recoveries of western Canadian birds (3–17 mo old) primarily along Pacific coast between El Salvador and northern Ecuador, with 1 recovery from Florida ( Martell et al. 2001a ). Widely separated overwintering locations of individuals that nested in close proximity to each other found in North American satellite-telemetry study ( Ewins and Houston 1992 ); also noted for Swedish birds ( Martell et al. 2001a ). Kjellen et al. 1997
Sex. Median fall departure dates for satellite-tracked females preceded males in the East and Midwest; no comparisons could be made in the West. Among mated pairs, tagged females departed 7–39 d (mean 22.3 d) before their mates in all cases ( n = 11; ). Similar results from Sweden ( Martell et al. 2001a ). Most likely due to males continuing to feed young postfledging. Observations and satellite-telemetry data (MSM) on entire family groups indicate that females left breeding territories before their young were completely independent. Median dates of arrival on overwintering areas earlier for adult females compared to males satellite-tracked from the East and Midwest. From East Coast, 60% of females and only 25% of males arrived by median date for population. In Midwest, 75% of females and 45% of males arrived by median date for population. In 6 pairs, female arrived on average 36.8 d (range 20–70) ahead of their mates; in 3 pairs, male arrived an average of 9 d (range 2–20 d) before female. Satellite telemetry found no difference in fall or spring migratory routes by sex, although females tend to overwinter south of males from same breeding population ( Kjellen et al. 1997 ). No difference in time spent on overwintering grounds for adult males versus females. Oregon adults (mean 178 d, Martell et al. 2001a n = 7) spent more time on overwintering grounds than Minnesota (mean 159 d, n = 7) or New York (mean 155 d, n = 6) birds (MSM).
Age. Band recoveries of fall-migrating northeastern U.S. adults were south of juveniles in all months except October ( ). Adults and young from mid-Atlantic states migrated south at same rate. Minnesota family groups ( Poole and Agler 1987 n = 3) tracked by satellite telemetry showed juveniles leaving breeding ground after adult female, but before adult male (MSM). Most (96.5%) fall migrants over Bake Oven Knob, Pennsylvania, reported as adults ( ). Spring band returns for 2-yr-old eastern birds lagged behind adults ( Heintzelman 1986 ). Adults and juveniles from all breeding areas have similar overwintering distributions ( Poole and Agler 1987 ). Juvenile Ospreys do not leave their overwintering areas until their second spring ( Poole and Agler 1987 ), and many stay on until their third spring ( Henny and Van Velzen 1972 ). Prevost 1982
Juvenile on migration. Although closely tied to water and fish, Ospreys are sometimes seen far from suitable foraging areas during migration. This bird is migrating down the Goshute Mountain range in Nevada. Photo by Brian Sullivan, Elko County, Nevada, 24 September 2010.
© Brian Sullivan, Nevada, United States, 24 September 2010
Night and Over-Water Flights
North American satellite-tagged Ospreys rarely migrate at night over land (
), but inevitably migrate at night when undertaking longer (> 12 h) water crossings. Wing morphology and wing-loading characteristics ( DeCandido et al. 2006 , Kerlinger 1985b ) facilitate such long non-stop flights. Kerlinger 1989a
Routine overwater flights are well known in North American populations of this species, especially among eastern breeders.
found 7 fall band recoveries at sea, including some as far as 200 km off the Atlantic coast, and, historically, a few individuals were reported from ships in the Atlantic Ocean at least 150 km offshore from Cape Hatteras, North Carolina, and New England (Bent 1937). Recent analysis of satellite-tracking data shows that significant numbers of juvenile Ospreys make flights from fledging sites in southern New England over the Atlantic Ocean to the Bahama Islands (up to 1,200 km; Poole and Agler 1987 ). And satellite data show that most adults breeding east of the Mississippi River fly south from Florida to Cuba over the Florida Straights (ca. 90 km), and then over the Caribbean Sea from the southern coast of Hispaniola to the northern cost of South America (about 600 km), generally repeating this route on their northern migration ( Horton et al. 2014 , Martell et al. 2001a www.ospreytrax.com/OspreyMainPage.html).
Ospreys reported on Bermuda (
eBird) are presumably traveling from far eastern Canada and making a trans-Atlantic crossing well in excess of 2,000 km.
Fall migrations of juvenile Ospreys were tracked by satellite from their fledging sites in southern New England to the Bahama Islands (
). These naïve birds, most of which had never flown more than 10 km from their nest sites before starting migration, showed remarkable navigational precision over open ocean, despite the perturbing effects of winds (up to 70 km/h) and lack of landmarks. Of 11 birds tracked, every one demonstrated a navigational capacity that was no more than 10 km off course for every 1,000 km traveled. Such precision seems best explained by use of vertical plane and horizontal plane magnetic cues, defined in a transformed magnetic coordinate space relative to each bird’s hatching site. In addition, these birds were shown to compensate for the effects of headwinds, demonstrating that their movements are also paced through some means of keeping time. Horton et al. 2014 Timing
Reports from North American migration watch sites show no preference for time of day in Osprey movements (
). Heintzelman 1986
The number of days satellite-tracked individuals traveled on fall migration varied greatly, ranging from 7 to 68 d. No significant difference in days traveled was found between males and females, either within populations or pooled across areas. No difference between satellite-tracked birds from the Midwest (25.7 d) and East Coast (32.6 d); western Ospreys (13.7 d) traveled significantly fewer days on fall migration than Ospreys from either the Midwest or East Coast, due to the shorter distance they had to travel. Number of days traveled by satellite-tracked spring migrants (both adult males and females) from New York, Minnesota, and Oregon varied between 9 and 68 d, but sample sizes were too small for comparison between groups (MSM).
Mean distance traveled per day by satellite-tracked individuals during fall migration varied from 95 to 380 km/d, with no difference in distance traveled per day between males and females. Combining sexes, no difference in mean daily distance was found between East Coast (203 km/d) and Midwest (230 km/d) Ospreys. Western individuals traveled farther per day (296 km/d) than either Ospreys from the Midwest or East Coast. Average speed of travel as measured by distance covered per day varied from 95 to 380 km/d (
). Martell et al. 2001a
For spring migration, the number of days satellite-tagged Ospreys spent on migration ranged from 5 to 68 d (mean = 22 d), and the daily distance covered varied from 74 to 568 km (mean = 237 km), with no difference between sexes and regional (East Coast, Midwest, and western individuals;
). Martell et al. 2014
Satellite data showed that spring migrants from the East Coast covered shorter distances overall and spent fewer days migrating than during the fall, while no differences were seen for western birds (
). Martell et al. 2014
Across western and eastern populations, male Ospreys spent 4 times as many in stopovers on migration in the fall compared to spring; females showed no difference. Regionally, birds from the East Coast spent 4 times more days in fall stopovers than in spring, while western birds showed no difference (
). Martell et al. 2014
Air speed during ridge gliding over Pennsylvania and New Jersey in fall was 16.2 m/s (58.3 km/h; range 9.4–33.4 m/s; [33.8–120.2 km/h]); during inter-thermal gliding over New York, it was measured at 24.9 m/s ± 0.9 SE (
). Kerlinger 1985b
Mean elevation of Osprey migration was measured by glider altimeter at 818 m ± 343 SD over Massachusetts, 470 m ± 256 SD as measured by radar over Cape May, New Jersey, and 831 m ± 340 SD in fall and 880 m ± 243 SD in spring as measured by radar over New York (
). Twenty-five percent of Ospreys tracked over New York in spring were at altitudes > 1,000 m. Reports of sightings up to 1,500 m ( Kerlinger 1989a ). Flights within 5 m of waves were recorded during water crossing from Cape May, New Jersey, and Whitefish Point, Michigan ( Heintzelman 1986 ). Six adult males crossing the Caribbean in fall migration averaged 68.7 m above sea level (range 0–498 m; ROB). Kerlinger 1985b
While using thermals, Ospreys were measured climbing at 3.1 m/s ± 1.1 SD (
). Flapping flight (3.2% of total), flap and glide (9.7%), glide and flap (22.6%), and gliding (64.5%) were all used during migration over Pennsylvania and New Jersey ( Kerlinger 1989a ). Kerlinger 1989a
Highest sighting rates at Hawk Mountain, Pennsylvania, came on day of a cold-front passage, decreasing up to 4 d after that (
). Allen et al. 1996 Feeding
Information on foraging during migration is sporadic.
and P. Spitzer (personal communication) have reported large numbers of migrating Ospreys feeding on concentrations of menhaden ( Decandido 1991 Brevoortia tyrannus) around Long Island Sound in the fall. Ospreys have been seen carrying food on migration in Cuba (F. Rodríguez-Santana, personal communication). In Great Britain, an Osprey was observed to carry the uneaten portion of a trout ( Salmo spp.) as it soared up and drifted out of sight, apparently continuing its migration ( ). Fall data from 30 hawk watch sites across North America shows roughly 1.5% of all Ospreys seen were carrying fish (L. Goodrich, personal communication) Analyses of both high spatial and temporal resolution GPS telemetry data will shed light on the extent to which foraging is restricted to stopover periods or may take place during periods of movement. Blincow 1994
For further details on Ospreys tracked by satellite during migration, see
Habitat in Breeding Range
Details in Food Habits: Feeding and Breeding: Nest site. See also
and Bent 1937b . Few quantitative data on habitat structure or aquatic ecology; needs study, better documentation. Palmer 1988e
Habitat varies greatly (boreal forests and mountain valleys to temperate coasts, lakes and rivers to subtropical coasts to desert salt-flat lagoons), but common denominators are: (1) adequate supply of accessible fish within energetically adequate commuting distance (10–20 km) of nest; shallow waters (0.5–2 m deep) generally provide most accessible fish; (2) open nest sites free from predators (especially mammalian)–such sites generally elevated (e.g., trees, large rocks [especially over water], or bluffs), or predator-free islands, or, increasingly, artificial structures such as nest platforms, towers supporting electrical lines or cellphone relays, and channel markers; (3) ice-free season of sufficient duration to allow fledging of young.
In boreal forest of northern Quebec (
), nests along rivers in area of roughly half rock outcrop, half marine clays and sand. Vegetation about 10% riparian (mainly willow [ Bider and Bird 1983 Salix spp.]) in broad valleys; 5% peat hummocks, fens and bogs; remainder black-spruce ( Picea mariana) forests of varying but usually sparse density. In New Brunswick, 1974–1980 ( ), most nests (74% of 152) coastal (vs. inland rivers); largest concentrations in coastal lagoon and island in large bay, reflecting nest site/feeding demands. Historically, and probably increasingly in recent decades, beaver ( Stocek and Pearce 1983 Castor canadensis) a key factor in creation of Osprey habitat in boreal and other northern forested regions. Beaver floodings create dead snags for nesting and shallow ponds for fish.
In southeastern British Columbia, breeds along 2 water management areas: a shallow, productive warm-water marsh and a cold and deeper lake; at both locations, impoundment and control of water levels for hydroelectric or marsh management created extensive shallows, enlarging areas suitable for foraging (
). In south-coastal New England, eastern Long Island, New York, and New Jersey, often depends on saltmarsh ecosystems, which form in tidal rivers and embayments: individuals nest on saltmarsh islands or along marsh shorelines and feed in bay waters or tidal creeks ( Steeger et al. 1992 ; Stone 1937 , Poole 1982b [Fig. 8.2]; see Poole 1989b for details of saltmarsh ecology). Valiela 1984
Mono Lake in California supports a small breeding population in a hypersaline, fishless lake where Ospreys nest on exposed tufa towers and leave the lake to forage in the surrounding habitat (
). Fields and Pagel 2016
In South Florida, small mangrove (
Rhizophora sp.) islands provide nesting sites in Florida Bay, typical habitat for other coastal, subtropical populations (e.g., Belize; , Ogden 1977b ; see Poole 1982b for details on the biology of mangrove ecosystems). Both mangroves and salt marshes are detritus-based ecosystems, contributing to their productivity. In addition, Florida Bay Ospreys forage over shallow-water mudbanks, used as well by bottlenose dolphins ( Hogarth 1999 and Brown Pelicans ( Tursiops truncates) Pelecanus occidentalis) ( ). Torres 2009
In Bahia de los Angeles (northern Baja California, Mexico), the Osprey is essentially a coastal desert species: Nests on shoreline cliffs and pinnacles surrounded by sparse Sonoran Desert vegetation, dominated by cacti, which are also used for nesting (
). Farther south and west in Baja and across the Gulf of California in western Sinaloa and Sonora, pairs nest on islands in large lagoons 6–12 m deep, surrounded by dunes, natural and artificial brine pans and salt flats, saline marshes, and desert flats ( Judge 1983 ). Castellanos and Ortega-Rubio 1995
A wealth of data from individuals tracked by GPS telemetry (providing fine-scaled habitat details) awaits analysis at
https://www.movebank.org. For example, numerous males in southeastern New England, foraging for mates and nestlings, have commuted regularly between coastal estuarine sites and inland freshwater lakes – demonstrating this species’ flexibility, even when attached to nest sites.
Few published data; broad tolerance of different habitats, see Migration, above. Interestingly, analysis of observations across Cuba (
, Rodriguez et al. 2001 ) has shown migrants from the U.S. moving in a broad front (coastal and inland) south and east across Cuba, then migrating along slopes and through mountain passes of Cuban Sierra Maestra and Nipe-Sagua-Baroca (average annual counts of nearly 7,500 birds), while roughly 90% of satellite-tagged birds pass far north of the southern mountains where Rodríguez-Santana and his team maintain a hawk watch (see Rodríguez-Santana et al. 2014 www.ospreytrax.com/OspreyMainPage.html for details). Long-distance water crossings are regular for migrating Ospreys; e.g., from eastern North Carolina over the Georgia Bight to Florida ( www.ospreytrax.com/OspreyMainPage.html), and across the Caribbean Sea and Gulf of Mexico ( ). Martell et al. 2001a
Satellite-tagged birds on both northbound, and especially southbound, migrations show no indication of straying from a relatively direct path to their destination to take advantage of any particular habitat (
, Martell et al. 2001a ). Almost all Ospreys returning to the northeastern U.S. travel inland from Florida in a direct line to the Northeast, roosting and presumably feeding in freshwater habitats rather than following the coast ( Martell et al. 2014 ). Martell et al. 2014
Most satellite-tagged adults arrive in South America in the arid Guajira Peninsula, west of the Gulf of Venezeula (
www.ospreytrax.com/OspreyMainPage.html). In Venezuela, southbound migrants are concentrated in large numbers at Lake Valencia as they pass over the eastern cordillera of the Andes en route to overwintering grounds farther south ( ). North-bound migrants are funneled to the Guajira and Paraguana peninsulas on either side of the Gulf of Venezuela as they leave South America (see Gessner 2008 www.ospreytrax.com/OspreyMainPage.html for details).
Habitat in the Overwintering Range
Increasing amount of data for the Americas. Large, shallow, freshwater reservoir in southern Mexico (Oaxaca) hosted roughly 80 Ospreys each winter, 1981–1983 (
). Band recoveries from 1914 to 1984 show North American breeders almost equally divided between coastal (marine) and inland (freshwater lakes and rivers) overwintering locations ( Barradas 1984 ). Band recoveries from Brazil show extensive use of Amazonian rivers ( Poole and Agler 1987 ), including mangrove estuaries of southeastern regions ( Mestre and Bierregaard 2009 ). Silva E Silva and Olmos 2002
In a study of Ospreys tracked by satellite telemetry (
): 1) out of 79 individuals tracked, 50.6% found on rivers, 19.0% on lakes, and 30.4% along saltwater coasts; 2) forest was dominant habitat type within overwintering areas, followed by agricultural lands – generally away from human settlements (but see Washburn et al. 2014 for data on Ospreys at aquaculture facilities in Colombia). Of 25 breeders from the Pacific Northwest, 18 overwintered at freshwater, inland sites, and seven in overwintered in estuarine environments (12 along the western coast of Mexico, 7 from Texas through eastern Mexico, 4 on the western coast of Central America, and 1 in Venezuela near Lake Maracaibo ( Bechard and Márquez-Reyes 2003 , Elliott et al. 2007b ). Washburn et al. 2014
There is a wealth of unanalyzed data from individuals tracked by telemetry that could reveal fine-scale information on overwintering ecology (e.g., hour-to-hour changes in habitat locations); see
www.ospreytrax.com/OspreyMainPage.html for details.
Historical Changes to the Distribution
Persecution eliminated Ospreys from southern and central California and the northwestern Baja Peninsula of Mexico before 1920 (
). In Minnesota, shooting eliminated birds from southeastern portion of state ( Henny 1983 ). In Wisconsin, historical numbers declined and range shrank by the 1940s ( Roberts 1932c ). Around Lake Huron, shoreline development and tree clearance reduced numbers and distribution; population suspected to be at historic low 1940–1970 owing to contamination; Georgian Bay population eliminated 1940–1970, but rebounded strongly in 1980s and 1990s, aided by artificial nesting platforms ( Robbins 1991 ). Habitat changes eliminated nesting birds from the Niagara River in New York ( Ewins et al. 1995a ). Ewins 1997
DDT caused dramatic population declines worldwide (
); in North America these declines sometimes, but not always, were associated with changes in distribution. From New York City to Boston, Massachusetts, there was a decline from around 1,000 to 100 pairs from 1950s to mid 1970s ( Poole 1989a ). Historically more abundant along the lower Mississippi River than in recent decades ( Spitzer 1980 ). Houghton and Rymon 1997
Ban on DDT (1972) allowed population to rebound while relocation (hacking), ability to nest on artificial structures (poles, towers), and newly created reservoir habitats have allowed expansion into formerly unoccupied habitat. Relocation projects allowed or accelerated the species’ range expansion into Pennsylvania, northwestern South Dakota, southern Michigan, southeastern Minnesota, Iowa, Missouri, Tennessee, and Ohio (
) (see Conservation and Management, below). Reservoirs responsible for range expansion in the northern Rocky Mountains ( Simnor 2015 ), southeastern Montana ( Henny 1983 ), and other parts of West, as well as Tennessee ( Swenson 1981 ). Artificial nesting structures have promoted population increases in many places, including Oregon ( Nicholson 1997m ) and around the Great Lakes ( Henny and Kaiser 1996 ). In Saskatchewan, 1977–1993, Ospreys expanded south about 100 km from boreal lakes to aspen parkland, aided by nesting sites located on power poles ( Ewins 1997 ). Number of breeders in Virginia, North Carolina, South Carolina, and Georgia more than doubled between 1981 and 1994. Houston and Scott 2001
The population in Europe, northern Africa, and the Middle East is reported to have doubled from the early 1980s to 2010 (
). Schmidt-Rothmund et al. 2014
Wing bones from mid-Miocene (approximately 13 million years ago) in California were used to describe the paleospecies
Pandion homalopteron ( ). Other mid- to late-Miocene fossils include hind-limb bones from Florida, the remains of paleospecies Warter 1976 P. lovensis ( ), complete enough to show that Miocene Ospreys were quite similar to the modern species, although not so robust. Other fossil remains found at about 12 Pleistocene sites (2.0–0.1 million years ago) in western Europe, North America, and the Bahamas, a range essentially unchanged from that of the earlier Miocene ( Becker 1985e ). Bones collected from Native American middens around Lake Huron, 1500–1620 AD ( Brodkorb 1964a ). Ewins et al. 1995a