USGS Home
Contact USGS
Search USGS

A. Adults

Examinations of predator food habits have provided useful information for assessing mortality of adult waterfowl. However, because most predators scavenge as well as capture prey and tend to select disadvantaged prey (e.g., crippled birds, incubating hens), the circumstances leading to death of waterfowl taken by predators are seldom known with certainty (Keith 1961, Sargeant et al. 1984). Studies of raptors (McInvaille and Keith 1974, Schmutz et al. 1980), mink (Eberhardt and Sargeant 1977, Arnold and Fritzell 1987), and especially red foxes (Sargeant et al. 1984) have provided information on the magnitude of mortality of adult ducks taken by these predators. The relative vulnerability of waterfowl species and sexes to predation by mink and red foxes has also been estimated.

Searches for waterfowl dying at certain high-risk sites have been used to investigate impacts of mortality caused by collisions with power lines (Krapu 1974), hay-cutting machinery (Labisky 1957), and vehicles on roads (Sargeant 1981). In such studies, relatively intact, fresh carcasses are often available for determining body condition, breeding status, and cause of death.

Radiotelemetry techniques have provided a means for estimating daily mortality rates of adult waterfowl during the breeding season and for determining times and causes of death (Ringelman and Longcore 1983, Cowardin et al. 1985, Kirby and Cowardin 1986). Comparisons of recovery rates of birds banded at both the beginning and end of the breeding season and resighting rates of marked birds from comparable time periods can also be used to estimate mortality rates of adults of some species. This procedure generally involves marking or observing birds during periods when they are difficult to trap or see. These methods have recently been employed to estimate the spring-summer mortality of Cackling Canada Geese in Alaska (Raveling and Zezulak 1988) and female Mallards in the prairie-parkland region of Canada (Blohm et al. 1987).

B. Eggs

Most determinations of nest success have been subject to considerable bias. Until recently, the traditional method for estimating nest success was to find a sample of nests and return to them later to determine how many of those nests hatched one or more eggs. The proportion of nests found that hatched young is commonly referred to as "apparent nest success." Mayfield (1961, 1975) recognized that apparent nest success rates were often inflated, because many nests were destroyed before they were found. The solution Mayfield proposed was to calculate daily survival rates of nests. Although the need to apply Mayfield's suggestions to studies of waterfowl was recognized almost immediately (Hilden 1964, Townsend 1966), his procedures were not widely adopted by waterfowl investigators until Miller and Johnson (1978) further discussed the subject and showed that the bias could inflate nest success estimates by 100% or more.

Calculation of daily survival rates of nests typically requires determination of age of eggs in each nest when found. This can be accomplished by flotation of eggs in water (Westerskov 1950), candling of eggs (Weller 1956), or examination of embryos (Caldwell and Snart 1974). Johnson (1979), Klett and Johnson (1982), and Johnson and Klett (1985) discussed biases of, and refinements in, methods for estimating nest success. These methods and techniques for finding and examining waterfowl nests were summarized by Klett et al. (1986).

Because of the potential severity of biases, apparent nest success should not be used except in special situations, such as when every nest sampled was found at the time the first egg was laid in it or the fate of all nests present in a given area (including those destroyed before they were found) was determined. Ely and Raveling (1984) discussed an exception for White-fronted Geese in which relatively high visibility of nests, synchronous nesting, no renesting, and unpredictable flooding resulted in calculation of apparent nest success being superior to that obtained using the Mayfield method. In many studies where nests are conspicuous (e.g., geese, eiders, and waterfowl nesting in man-made structures), the probability of finding nearly all destroyed nests is high and, therefore, calculation of apparent nest success likely produces reasonably accurate results. Unfortunately, few authors have reported their search methods in sufficient detail to judge the suitability of their estimates of nest success.

Creation of artificial nests has been used to assess egg mortality and to identify causes of mortality (e.g., Hammond and Forward 1956, Schranck 1972, Crabtree and Wolfe 1988). The method is attractive because of the ease with which desired samples can be obtained, but probably results seldom mimic actual mortality. Differences between artificial and natural nests include investigator rather than waterfowl choices of nest sites and nest construction, investigator choice of egg type (usually chicken), and absence of attending adult waterfowl that may defend nests from or reveal nests to predators.

Nest success should be coupled with other information, such as renesting rates and clutch size, to best assess impact of egg mortality on reproductive success. For situations in which waterfowl do not renest, nest success equals the proportion of nesting hens that hatch one or more eggs. When renesting occurs, hen success may be much higher than nest success (e.g., Cowardin et al. 1985).

Investigators often report egg success (proportion of eggs found that hatch) as well as, or instead of, nest success, especially for waterfowl with high nest success, such as swans and geese and ducks nesting in sites protected from predators. Egg success is seldom reported when nests are heavily affected by predators (e.g., most ground-nesting ducks in temperate regions), because partial predation of clutches is common, thereby making it difficult to determine how many eggs were laid and subsequently hatched.

C. Prefledged Young

The most accurate estimates of survivorship of prefledged waterfowl are from studies in which individually marked young or sizes of broods of individually marked parents can be observed repeatedly throughout the brood-rearing period. Radiotelemetry of adults (Ball et al. 1975, Ringelman and Longcore 1982, Cowardin et al. 1985, Duncan 1986) and ducklings (Talent et al. 1983) has assisted observation of broods of secretive species in habitats where young are difficult to locate.

Studies of food habits and behavior of predators have confirmed that predation is an important mortality agent on prefledged waterfowl (e.g., Vermeer 1970b, Sargeant et al. 1973, Eberhardt and Sargeant 1977, Arnold and Fritzell 1987). Use of these methods for determining mortality of prefledged young has not been as revealing as for adult waterfowl, because the young are often totally consumed (Sargeant 1978) and almost completely digested.

Analyses of recoveries of marked birds have been used in a few studies to assess survivorship of prefledged waterfowl (e.g., Grice and Rogers 1965, Haramis and Thompson 1984). This method requires marking one sample of young at time of hatching and another sample near the time of fledging, in order to compare recovery rates for the two samples. Young can be marked with small web-tags while still in eggs at time of hatching (Alliston 1975) or just after hatching, whereas older young are generally marked with standard leg-bands. Because of differences in observability of the two markers, recovery rate data (hunter reports) are likely biased. Grice and Rogers (1965) avoided this problem by capturing adult birds during the year after marking and examining each bird for presence of both types of markers. Brown (1972) and Rockwell et al. (1987), in addition to using this method, used a double-marking, successive capture-recapture scheme that enabled estimation of survival during year of marking.