Northern Prairie Wildlife Research Center
The combined breeding population of scaup has fluctuated markedly since the Waterfowl Breeding Ground Population and Habitat Survey (hereafter referred to as Breeding Ground Survey, BGS) was initiated in 1955 (Figure 1). The population declined during the 1960s, increased during the 1970s and early 1980s, then in 1984 began a nearly steady decline to record lows in 1998. During 1978-97, the population declined by 150,491 scaup per year (Afton and Anderson in review). In separate analyses of BGS data, Allen et al. (1999) found no linear trend in numbers from 1955 through 1987, but noted a significant decline since 1987.
Discerning whether both species are declining is problematic because of the lack of species-specific information. Allen et al. (1999) assumed strata 8, 9, 10, and 11 in Alaska and stratum 13 in the Northwest Territories contained primarily greater scaup, and that all other strata contained primarily lesser scaup. Those strata thought to contain primarily greater scaup had stable populations during 1955-98, with an increase during 1988-98, but populations in other strata (containing primarily lesser scaup) in the boreal forest have declined since 1988 (Allen et al. 1999). Independent analyses by M. MacCluskie, A. Afton, and M. Anderson (unpubl. data; hereafter referred to as MacCluskie et al.) indicate similar trends. However, knowledge about the proportion of greater versus lesser scaup in each surveyed strata, or at smaller substrata scales, is poor and often based on old information, and there are concerns that proportions have shifted in some areas.
The combined breeding scaup population has not declined uniformly across the breeding range. The Boreal Forest population (BGS strata 1-7, 12-25, 50, and 77) has declined markedly, whereas the Prairie-Parkland population (strata 26-49 and 75-76) has increased 50% during 1993-97. The Tundra population (strata 8-11) remained relatively stable during this period (Figure 2; Afton and Anderson in review). Within the surveyed areas, populations have significantly declined during the last 20 years (1978-97) in central and northern Alberta, northeastern British Columbia, and in Northwest Territories (combined, an area which hosts an average of 52% of the continental breeding population), and in southern Alberta, Montana, and western Dakotas.
MacCluskie et al. further refined these analyses by examining breeding population data at the transect level for 14 strata. Of those, 10 strata (strata 3, 4, 12-18 and 20) contained at least one transect with a significant positive or negative slope. Generally, scaup counted at the transect-level for strata east of the continental divide show a decline beginning about 1980. Stratum 12 (Old Crow Flats) shows significant increases in scaup numbers since initiation of surveys. Strata from interior Alaska (Strata 3 and 4) have mixed trajectories with some transects showing increasing and others decreasing populations. For transects with significant negative slopes, the decrease in scaup primarily is due to fewer groups and pairs and not to fewer lone males counted. The increase in Stratum 12 is due to increased number of groups since the early 1970s.
MacCluskie et al.'s analyses indicate that the decline of scaup is widespread in the western Canadian boreal forest and is not restricted to a few strata. The cause(s) of the decline in this region does not appear to be affecting populations in interior Alaska, or in boreal forest habitat west of the continental divide. Analyses by MacCluskie et al. are ongoing and will include comparisons with population trends of other waterfowl (e.g., American wigeon [Anas americana], green-winged teal [A. crecca], and scoters [Melanitta spp.]) that breed in boreal forests.
Biases in the BGS for scaup population estimates have not been adequately addressed (see Austin et al. 1998, Afton and Anderson in review). Scaup are among the latest migrants to move north in spring, and their migration may be delayed during cold springs. The BGS, designed primarily for mallards (Anas platyrhynchos), is conducted in mid-May (U.S. and southern prairie-parkland areas) to early June (northern strata), when some scaup are still migrating to breeding areas. Some scaup may be counted multiple times, i.e., in southern regions and again in northern areas. Without extensive mark-resighting or telemetry studies, evaluating this bias is difficult. Biologists suspect that in some years, particularly those with delayed migration due to cold or inclement weather, this bias results in large standard errors or, more seriously, in high population estimates for some strata (see also Austin et al. 1998, Afton and Anderson in review). Weekly waterfowl surveys in south-central North Dakota (1957-97) indicated a trend toward earlier chronology of scaup migrations and a weak relationship between peak scaup numbers and temperature (J. Austin, D. Granfors, M. Johnson, and S. Kohn, unpubl. data). These data and other observations (Afton and Anderson in review; D. Kay, Ducks Unlimited Canada, pers. obs.) also indicate that BGS counts are conducted in some years when migrant birds are still in the area.
Waterfowl production and habitat surveys, conducted during 1-21 July in prairie-parkland areas and 8-22 July in boreal forest areas (U.S. Fish and Wildlife Service and Canadian Wildlife Service 1987), are poorly timed to estimate scaup production. Most scaup clutches hatch from mid-July through August (Austin et al. 1998), so most scaup production is missed in surveys conducted before late July, particularly in wet years.
On average, 40% of scaup counted during midwinter surveys were in the Mississippi Flyway, 39% in the Atlantic Flyway, 14% in the Pacific Flyway, and 6% in the Central Flyway during 1955-97 (Afton and Anderson, in review). Midwinter counts of scaup have declined over this period (Figure 1), with declines in Atlantic and Mississippi Flyways accounting for most of the change. Long-term trends by state are mixed. Afton and Anderson (in review) noted that midwinter survey totals are strongly affected by number of scaup recorded in Louisiana, and these counts are influenced by weather, which occasionally prevents completion of surveys. Midwinter counts have substantial error and biases (Crissey 1975), and comparisons among years or states are not recommended (Eggeman and Johnson 1989). However, such counts provide an indicator of long-term trends and changing waterfowl distributions not otherwise available.
Analyses of the Christmas Bird Count (CBC) data indicate that lesser scaup numbers during 1955-95 were stable (Allen et al.1999), but greater scaup numbers declined 3.2% per year. However, numbers of greater scaup in the CBC have increased in the Atlantic Flyway during 1988-95. The only changes in CBC data for all scaup occurred around the Great Lakes, where they showed a 4.8% per year decline from 1955 through 1987 and a 3.5% decline per year for 1955-95. From 1988 through 1995, scaup numbers on the Great Lakes increased by 16% per year. This reflects a change in distribution associated with their exploitation of zebra mussels (Dreissena polymorpha) (Wormington and Leach 1992; C. Custer and T. Custer, pers. comm.).
Distribution of migrating and wintering scaup has shifted in the past 20 years probably because of changes in food resources. Such shifts could change exposure of scaup to harvest and also to contaminants. Rocque and Barclay (1999) and P. Castelli (New Jersey Div. Fish, Game, and Wildlife, pers. comm.) noted that numbers of wintering greater scaup have declined along the Long Island Sound, Connecticut, and New Jersey coast. These declines could be due to changing food resources, habitat, or disturbances. Food resources for scaup have increased in parts of the Great Lakes because of expansion of zebra mussel populations, but may have declined in some riverine or coastal areas because of contaminants, sedimentation, or other factors affecting invertebrate communities. Declines in migrant scaup numbers using the Illinois and Mississippi Rivers were attributed to degraded habitat quality - reduced food resources, water levels, and water quality (Mills et al. 1966, Bellrose et al. 1979, Korschgen 1989; S. Havera, Illinois Natural History Survey, pers. comm.). Wetland degradation also has affected coastal marshes of Lake Erie, the Detroit River, and southern Gulf coast. Changing nutrient and sediment loads in the Mississippi River have caused a large (up to 9,000 km2 ) hypoxia zone in the inner continental shelf near the Mississippi Delta, an area once used by wintering scaup; such areas now have low densities of fish and invertebrates (Turner and Rabalais 1994). This region may have once provided significant food for wintering scaup. Biologists know that scaup occur far off shore (e.g., off the Louisiana coasts), but these areas are not surveyed because of logistical difficulties and expense.
Harvest and Age and Sex Ratios
Harvest.--Harvests of greater scaup in the U.S (1961-97) and Canada (1974-97) have declined. Harvest of lesser scaup also declined in Canada since 1974 (Allen et al.1999). The Canadian declines may be linked to declining number of hunters there since 1974. Harvest of lesser scaup in the U.S. was variable but has increased in the most recent 4-5 years, likely due to longer seasons, more liberal bag limits, and increased hunter participation. Nonetheless, during 1988-97, harvest of scaup relative to population size was the lowest of all common duck species for which there are spring survey and harvest data (Allen et al. 1999). Scaup are lightly harvested compared to more r-selected species of ducks, suggesting that hunting has not played a significant role in their decline (see also Afton and Anderson in review).
Age ratios.--Age ratio (immatures to adults) of greater scaup in the U.S. harvest did not change during 1961-97; the mean ratio was 1.4 (Allen et al. 1999). For lesser scaup, the age ratio in the U.S. harvest averaged 1.46 and declined 1% per year during that period. Afton and Anderson (in review) found similar declines, with the greatest declines in the Mississippi Flyway. In Canada, age ratio for greater scaup was highly variable but showed no trend. Age ratio of lesser scaup there also was highly variable but declined about 1.5% per year during 1969-96 (Allen et al. 1999).
Sex ratios.--Afton and Anderson (in review) reported increased sex ratios of lesser scaup (more males relative to females) for both adults and immatures in the U.S. harvest. They interpreted this as evidence that reproductive success has declined and/or female survival rates have declined relative to male survival rates. Furthermore, if age ratio of males has declined and sex ratio has increased, female survival seemingly must have decreased.