Northern Prairie Wildlife Research Center
Assessing Breeding Populations of Ducks by Ground Counts
Appraisal of Factors Influencing Inventories
On the basis of field programs on the two study areas, a number of corrections
to potential sampling errors and biases and solutions to problems of duck population
estimation present themselves.
Unmated Males and the Census
Although the proportion of unmated males that remain through the breeding season
on the pothole breeding habitat is unknown, there are observations that May sex
ratios may be more distorted in large marshes where unpaired mallard and pintail
males congregate (Hochbaum, 1944; Ellig, 1955). For parkland pond habitat, Diem
and Lu (1960) discussed the errors involved in censuses because of distorted sex
ratios and enumeration of unmated males as indicated pairs. For the present study,
I assumed that the greater proportion of unmated males remained on the study areas
until immediately prior to the brood season. Firstly, the assumption was based
on some incomplete data on marked, unpaired drakes. Of 23 mallard, pintail, gadwall,
widgeon, and blue-winged teal unpaired males marked in 8 years, 11 were subsequently
seen on or around the study areas where they were marked, for periods up to 42
days. Furthermore, observation of marked pairs showed that as many as one in four
may disassociate themselves from the home range because of the disturbance due
to marking and capture. Utilizing a similar ratio for unmated drakes, I concluded
that from 48 to 65 per cent of the unpaired drakes may also remain on or near
a study area, at least during the laying and early incubation periods and be available
for censusing. Secondly, in 1964 and 1965, at Waubay, South Dakota, Drewien (in
litt.) established that unmated blue-winged teal drakes composed 15 to 20
per cent of the resident population of drakes of a study block and remained there
throughout the breeding season. Unmated males showed varying degrees of site tenacity
and localization of activity through the month of May to early June, as did the
mated males and breeding pairs. Thirdly, Bossenmaier (1951: 61), who intensively
censused a major moulting marsh Whitewater Lake in southern Manitoba, did not
record the first flocks of 10 or more mallard drakes until May 28, 1950, and May
21, 1951. While these may have been either previously paired, postbreeding or
unpaired drakes, he gives no records of earlier congregations of drakes. I suggest
that the greater proportion of both mated and unmated drakes of most dabbler species
do not leave the breeding ponds for the moulting lakes until the nesting season
is well under way. They are thus located on the breeding grounds and can be erroneously
assessed as potential breeding pairs.
On both study areas, marked, unmated males were seen to aggregate with mated
males whose hens were incubating and to take part in GFAC and ARF. They may
also associate with pairs as novice drakes (Hochbaum, 1944) or form groups of
two, three or more. In early April, associations of two or more unmated pintail
drakes were common but fewer small groupings of unmated drake mallards were
observed. During the prenesting period, groupings of two unmated males of other
dabbler species were also uncommon except for blue-winged teal.
As all males are counted as potential breeding pairs, the use of a correction
factor to disregard unpaired drakes should be valid, in order to assess only
true pairs.
With all species the number of unmated drakes observed in the prebreeding period
varies slightly from year to year (Table 3). If sufficient
counts are made to show a significant difference, then the yearly sex ratio
correction factor should be applied to the indicated population based on all
drake and pair counts. If no counts are available the average prebreeding male
to hen ratios found in Bellrose et al. (1961) and in Table
3 might be used.
Sex Ratio Correction Factors
The problem of overestimation by enumerating unmated lone males as indicated mated
pairs becomes acute in those dabbler and diving duck species which are regularly
known to have distorted sex ratios on the breeding grounds (Bellrose et al.,
1961). Since it is almost impossible to separate lone mated from lone unmated
drakes on the basis of plumage or obvious behavioural characteristics, I propose
that all drakes should be enumerated as breeding pairs 1 to 2 weeks before the
first appearance of broods, and that a prelaying sex ratio correction factor should
be applied. For yearly trend information, in those species whose sex ratio may
not fluctuate widely, enumeration of all drakes should not distort measures of
relative abundance. For more sophisticated and accurate measures of populations
and the factors limiting recruitment on special study blocks some correction for
the unmated male component must be made (cf. Table 9).
The errors involved in assuming that all unmated drakes are indicated pairs
have been recognized by many workers (Hochbaum, 1944; Murdy, 1953; Sowls, 1955;
Diem and Lu, 1960; Bellrose et al., 1961; and Hammond, in litt.,
1966). Hawkins, Gollop, and Wellein (1951) concluded that obtaining prebreeding
sex ratio data offered a logical method of correcting aerial counts for hens
not seen associated with drakes. However, Kiel (1951) cautioned against the
use of sex ratio correction factors because egress of premoulting males might
equal disappearance of nesting females from the transects, the result "being
a constant sex ratio but a decreasing countable population". On the Yellowknife
Study Area, Northwest Territories, Murdy (1962) enumerated all drake lesser
scaup and ring-necked ducks and applied a sex ratio correction factor for the
imbalance of males. The method assumes that sufficient prebreeding sex ratio
data are available. Murdy (1964) later utilized counts of pairs and drakes on
waiting sites to estimate breeding populations.
Late Influxes of Pairs
In May, the separation of late migrant pairs or drought-displaced birds from residents
is nearly impossible to make. The criterion I used revolved around whether pairs
were grouped or spaced and general behaviour, such as whether hens gave the persistent
quacking call or drakes showed hostility toward hens of pairs. However, during
the transitory, postmigration period pairs may be spaced from other pairs but
do not yet show activity localization. Such pairs could be included as residents.
Late May and June influxes of breeding pairs nesting for the first time or
influxes of renesting pairs are apparently common on some marsh and pond study
areas (Jessen, Lindmeier, and Farmes, 1964; Hammond, 1959; Kirsch, in litt.).
Such influxes give rise to a multitude of sampling problems in the accurate
assessment of the seasonal or total number of pairs which attempt to nest on
a specific area. Hammond (1959; letter June 11, 1967) noted influxes of birds
into the marsh habitat of Lower Souris National Wildlife Refuge, when: (1) sheet
water disappeared from fields surrounding the marsh, (2) mass destruction of
nests occurred through farming practices, (3) drakes and hens made premoult
movements away from their breeding grounds, possibly after only one nesting
attempt. Yearly, I recorded early June influxes of mallard pairs on the Roseneath
Study Area, where ponds were intensively surveyed, but could not differentiate
late migrants from resident pairs on the Kindersley Study Area because of much
larger populations of all species. At present, the sampling problems involved
with adequate census of seasonal populations and the differentiation of residents
from new arrivals appear insurmountable. Where influxes occur indirect methods,
concerned with nest counts and brood numbers, may have to be used to estimate
breeding populations.
GFAC, ARF, TBF and the Census
Both mated and unmated drakes join spring and summer aerial flights, i.e., group
flights associated with courtship and attempted rape flights, and all drakes in
such flights temporarily on ponds have been enumerated on the Kindersley Study
Area. As such flights may extend for 1 to 4 miles and gather drakes from a number
of ponds they would tend to distort population estimates from narrow one-quarter-mile
roadside strip counts and therefore might be deleted. Observations of behaviour
of drakes in aerial flights temporarily on ponds should be noted so that late
season, postbreeding groups of males and pairs are not classed in error as breeding
birds in flights. The number of flights observed is small in relation to over-all
population numbers, except for ARF of pintails and GFAC of lesser scaup, where
groupings of males with one hen are common through May. Attempted rape flights
may be more frequent in all species during late afternoons, at a period when incubating
hens are taking their recesses and are more prone to attract drakes. If censuses
are conducted in the 0800- to 1200-hour interval the number of three bird flights
seen is much reduced over the early morning (0530-hour) and late evening (1800-
hour) periods. Therefore, sampling errors due to mobility should be minor. Most
mallard ARF were observed in the early mornings and late afternoons, after hens
started to incubate. Few were enumerated in the 0800- to 1200-hour census interval.
Mallard and pintail GFAC are usually seen in early and mid-April and are commonly
associated with migrating flocks. The number observed per day on pond breeding
habitat, during any optimum census period in May, will be negligible, especially
if transects or study blocks are located at some distance from habitat where migrants
congregate. For other dabbler species few GFAC were recorded after May 15 or during
the optimum census period.
Pair Bond Duration and the Census
Much variation of pair bond duration exists between and within species of Anatidae.
The duration of the pair bond, size of home ranges, and period of localized activity
on a waiting site all affect accuracy of spring pair censuses. Hochbaum (1944)
recorded the early abandonment of hens by drake mallards and pintails, usually
after laying of the clutch. Sowls (1955: 101) noted that blue-winged teal and
shoveler drakes did not abandon their hens until late incubation, while mallard,
pintail, and gadwall did not associate with the hen after clutches were complete.
Similar observations were made by Smith (in Bellrose et al., 1960: 427).
Leitch (1952) noted that mallard and pintail drakes remained on territories the
shortest time, while blue-winged teal, widgeon, and shoveler drakes remained on
their territories much longer and population figures for these species, taken
from ground counts, should be quite accurate. Bue (1952) pointed out wide variations
in the times that species and individual drakes abandon home ranges. Similar variations
in stage of incubation at which drakes desert hens were also reported by Oring
(1964). McKinney (1965) summarized the available literature on pair bond duration
in North American Anatidae. He showed that many diver and dabbler drakes did not
associate with their hens after the first week of incubation. However, blue-winged
teal, shoveler, gadwall, and lesser scaup generally abandoned their hens only
after the second or third week of incubation. Gates (1962) reported that drakes
of renesting pairs of gadwall abandoned their hens sooner than drakes of initial
nestings, a conclusion I have substantiated in following renesting mallards, pintails,
and blue-winged teal. In Maine, Stotts and Davis (1960) noted that seven drake
black ducks attended hens from 7 to 22 days, averaging a minimum of 14.3 days,
during early, first nesting attempts. For eight examples of late renesting, drake
attendance varied from 4 to 16 days and averaged only 9.1 days. I noted that the
pair bond may also be retained through a renesting attempt. Nine hens, of dabbler
pairs which renested, were trapped and marked over an 8-year period. Five of the
nine renesting hens were subsequently seen with their original drake while four
had reformed pair bonds with another unmarked drake. Kirsch (in litt. )
suggested that some mallard and pintail pairs retain the bond for periods of up
to 2 months and renest several times.
I also noted wide yearly variations in time of breaking of the pair bond. In
1958, an exceptionally early breeding season with hatching starting on May 15,
22 mallard hens and broods were observed with associated drakes, whereas in
years when hatching peaked after May 25, fewer than five were recorded annually.
In 1955, 1956, and 1958, I have recorded newly hatched broods associated with
pairs of blue-winged teal and shoveler in late June, 2 weeks after the first
broods of this species were recorded. Kirsch (letter, June 27, 1967) noted "many"
blue-winged teal broods with associated drakes on the Woodward Study Area, North
Dakota. Most drakes were seen with early season broods but seldom remained with
the brood beyond the first few days. At Roseneath, I regularly saw drake ruddy
ducks associated with hens and broods up to 10 days old. In all cases of drakes
with broods the hen did not take the "repulsion posture" and it was therefore
assumed to be a mated pair. Thus, some drakes have stronger hen attachments
and are associated with their waiting areas for longer periods than other drakes.
Therefore, they are more readily seen and available to be censused as indicated
pairs.
Pair bond duration is related to strength of site tenacity by the drake and
the daily re-use of the activity centre by the hen and drake. Pair bonds are
periodically re-enforced through the early and mid-incubation period as long
as the drakes return to the waiting sites while hens are on recess. Tight pair
bonds are maintained by frequent and joint activity through the migration, postmigration,
and laying period but progressively weaken as the pair is associated less and
less during incubation. Strong pair bond attachment in males is associated with
strong attachment for the waiting site of the home range. Site tenacity to the
breeding home range leads to faster pair bond re-establishment whenever the
hen returns to the home range to feed, bathe and preen during her infrequent
recesses. Pintail drakes appear least attached to the waiting site; other dabbler
drakes return regularly to it or to favoured feeding areas in early morning
and late afternoon to rejoin the hen on her recesses during these periods. Other
dabbler drakes, especially mallards, whose hens are in the same period of incubation
do not disassociate themselves from drake aggregations and are therefore found
less frequently on the waiting sites. Such differences appear due to individual
behaviour.
In short, the period of drake desertion of the home range varies with the species,
the individual pair, the nesting phonology of a season and perhaps density of
pairs. There is no sharp break but only a general waning of the pair bond and
attachment to home range. Therefore, no accurate and predictable period of drake
desertion can be given for a species or for any breeding season. In general,
mallard and pintail drakes should be censused before their hens are in their
second week of incubation, whereas widgeon, shoveler, gadwall, and blue-winged
teal can be accurately censused by counting drakes through the second or third
week of incubation. The correlation of census periods with breeding phenology
is imperative.
Grouped Drakes and the Census
In spring the period of strongest pair bond attachments is associated with periods
of maximum pair spacing and maximum drake intolerance. These periods extend from
the time of dissolution of the migrating flock to the beginning of incubation.
Paired mallard drakes rarely associated with each other in the prenesting and
early laying periods. Unmated drakes of all species remain gregarious while even
some mated pintail drakes may associate with each other during the laying period
(Smith, 1963). At Kindersley, there was a strong tendency for groups of four to
seven drake pintails to associate, although this may have been a reflection of
the higher percentage of unmated males found in this species (Table
3). At Kindersley, more groups of two to eight drakes with a single hen, in
GFAC or ARF, were seen of pintail than of mallards (Tables
8a, 8b). In mallards the period of drake intolerance
is followed by a period of drake sociability. Mated drakes form small aggregations,
usually two or three but up to ten. Such morning and mid-day drake associations
are characteristic of mallards and pintails toward the end of the egg laying period
and through incubation (Tables 8a, 8b).
However, mated drake associations were rarely observed in widgeon, shoveler, gadwall,
and blue-winged teal until mid- and late incubation periods (Tables
8c, 8d, 8e, 8f).
Bue (1952) and Evans and Black (1956) noted that grouped mallard and pintail drakes
were conspicuous before broods appeared. These authors noted that with blue-winged
teal, whose drakes stayed with hens until nearly hatching time, grouping of drakes
coincided with the appearance of broods. Thus, as the drake-hen bond wanes, drake
intolerance for other drakes also wanes and males form aggregations for varying
periods of the day. A number of inter-related phenomena occur with drakes after
the hen begins incubation: (1) the hostility of drakes toward hens and other drakes
decreases, (2) the tendency to behave sexually toward all hens except his own
mate increases, (3) drakes disperse widely from their waiting area sites, (4)
they begin to associate with other drakes in larger and larger groups. Most drakes
make a final break with the breeding ground from 2 to 5 weeks after the hen has
started to incubate when they form large premoulting flocks and migrate to moulting
lakes. Few drakes of any species, except small groups of blue-winged teal, were
recorded moulting in the parkland or grassland study area ponds.
Most mallard drakes with incubating hens tend to form associations with other
drakes during the mid-day hours, 0800 to 1600. Before and after this interval
drakes either avoid association with other drakes or remain isolated on their
waiting areas. Morning counts, i.e., 0800 to 1200 hours, showing a breeding
population of predominantly pairs and lone drakes would indicate that the greater
portion of the pairs are in the prelaying and laying interval. Censuses which
show a preponderance of lone males and grouped males would indicate that most
pairs are laying or incubating.
I made observations in parkland, grassland, and large marsh habitats which
suggested that males associated more readily under dense than under sparse population
situations. The opportunity for drake association is increased or drakes may
be forced to group by lack of surface waters or common loafing spots. Also,
if large marsh areas serve as congregating areas for unmated males, they may
be seen associated throughout the breeding season (Hochbaum, 1944; Ellig, 1955).
In summary, on block-type study areas, enumeration of groups of males of five
or less, especially mallards and pintails, before the appearance of first broods
in mid-May, is a valid measure of indicated breeding pairs. They should be enumerated
on large sample plots (4 square miles or over) but further investigation of
their distribution and activity patterns should be made before all such groups
are tallied on narrow transects. For other dabbler species in which groupings
of drakes are generally less than five, i.e., two's and three's, all such drakes
should also be enumerated as indicated pairs on block areas prior to recommended
cut-off dates.
Optimum Census Periods
I have recommended that optimum census periods be established yearly for each
species. These should be based on migration chronology and nesting phenology.
Counts should be conducted when the greatest proportion of any species is in the
prenesting (including renesting), laying, or early incubation stages. For 1956
through 1959 at Kindersley, Saskatchewan, the interval from May 8 to 20 was considered
optimum for censusing mallards and pintails, from May 20 to June 5 for widgeon
and shovelers, and from May 25 to June 10 for gadwall and blue-winged teal. In
years with no April cold snaps, counts for mallards and pintails could be initiated
2 to 3 weeks after the first few hens start to lay or are seen dropping into nesting
cover. Egg laying may start from 10 days to 3 weeks after the first pairs migrate
into an area. For other dabbler species, counts should start 1 to 2 weeks after
the first eggs are noted. All censuses should terminate before the first few broods
appear or better still when the first nesting hens are in their third week of
incubation. Censuses taken in the above intervals will assess populations of pairs
in their first nesting attempt at a relatively stable level. Exceptionally late
migrants or drought-displaced birds moving into an area in mid-June would still
not be adequately enumerated without later periodic counts.
Murdy (1953) concluded that optimum census periods occurred after the migrants
had left and before emergent vegetation and pair behaviour changed. For 1951,
he pointed out that the optimum time for the annual duck survey in South Dakota
was during the week of May 13 to 23 when mallard, pintail, shoveler, blue-winged
teal, and gadwall pairs were all in residence and populations were relatively
constant. The optimum survey period for mallard and pintail pairs extended from
April 28 to May 28 while for the remaining three species it extended from May
15 to 28. Hammond (1966) noted that optimum census periods vary annually by
7 to 10 days. He recommended a May 7 to 17 census period in North Dakota for
mallards, pintails, canvasbacks, and wood ducks and a May 25 to June 7 census
period for gadwalls, blue-winged teal, redheads, lesser scaup, and other species.
For a study block in the forested habitat of the Northwest Territories, Murdy
(1964) recommended two censuses, one immediately after ice break-up between
May 20 and 25 for all puddle ducks except shovelers (but including canvasback)
and the second between June 6 and 12 for late nesting divers. Where possible,
single censuses should be avoided, although Salyer (1962) felt a single valid
census could be conducted in early June at the period of nest initiation of
blue-winged teal.
Number of Seasonal Counts Required
Intensive studies of waterfowl breeding chronology and optimum census periods
show that all duck species do not migrate into a region or initiate nesting at
the same time (Kiel, 1949; Lynch, 1951; Murdy, 1953; Bue, 1952; Stoudt, 1952;
Smith, 1956; Evans and Black, 1956; and others). All dabbler species show peak
indicated populations at different times, usually associated with the period of
strongest site attachment, i.e., prenesting, laying, and early incubation periods.
Mallards, pintails, and canvasback are generally early breeders; widgeon, shoveler,
redhead, and ring-necked duck intermediate; while gadwall, blue-winged teal, lesser
scaup, and ruddy ducks are late breeders. Regional differences in start of nesting
may occur as Kirsch (in litt.) noted that blue-winged teal on the Woodward
Study Area, North Dakota, could be considered intermediate breeders. Generally,
one census cannot accurately assess peak populations of all species. I concluded
that a minimum of two, and possibly three, counts may be necessary to assess pair
numbers of a multi-species breeding population with asynchronous nesting periods.
Very early, Mendall (1948) had noted that more than one count might be necessary
to enumerate early hatching black ducks and golden-eyes and late hatching ring-necked
ducks and teal (A. carolinensis; A. discors). Kiel (1949) conducted as
many as four breeding pair censuses on 12 Manitoba transects. The highest mallard
and pintail populations were recorded in the April 21 to 25 count while the
highest indicated populations of widgeon, shoveler, gadwall, and blue-winged
teal were noted from May 14 to June 6. Lynch (1951) recognized that late breeding
blue-winged teal and lesser scaup could not be adequately surveyed in mid-May
at the time of the mallard and pintail aerial census. For over 10 years in the
parklands of Saskatchewan, Stoudt (1964) used two censuses to assess breeding
populations, one in May for early breeders and the second in June for all other
species. Similar double censuses were conducted on various Alberta study areas
by Smith (1957). Evans and Black (1956) made periodic censuses, April through
August, on the Waubay Study Area, South Dakota. All authors utilized the peak
number of indicated pairs for any one period as the estimated breeding population
for the respective study or transect area. However, as discussed later, I suggest
four or five replicate counts and a mean population estimate would better describe
species pair numbers than a maximum count taken from one census. To establish
trends, and where time and manpower is limited, a single count can be successfully
used to estimate total populations, as has been proposed by Hammond (1959; in
litt.).
Time of Day for Counts
I have recommended that in grassland areas where emergent vegetation does not
affect visibility of ducks, ground counts be conducted between 0800 and 1200 hours.
Between these times, pairs and drakes localize their activity and are least mobile;
most laying and incubating hens are on their nests, winds are low, and light is
favourable. Diem and Lu (1960) recommended that for censusing mallards on transects
in parkland habitat, counts be conducted in morning hours after 0530 but before
0930 hours as both pairs and lone drakes are more visible in early morning and
leave transect ponds in the forenoon. They showed no statistically significant
differences in indicated numbers of blue-winged teal and lesser scaup for counts
conducted at 0530, 0930, and 1330 hours. They further concluded that aerial counts
made at mid-day may be from 30 to 50 per cent lower than early morning counts
and recommended all censuses be made from 0600 to 1200 hours. Smith (1956) concluded
that for parkland areas of Alberta mid-day aerial counts, when pairs were most
inactive, did not give accurate coverage. Numbers of paired ducks observed in
mid-day decreased from early morning counts while numbers of single males increased.
In Manitoba, Rogers (1964) began lesser scaup counts between 0800 and 0900 hours,
completing them in 2 to 3½ hours. For marshes, pond-habitat blocks, and transects,
Hammond (1966) recommended all-day counts starting after 0900 hours, as some duck
pairs were in nesting cover prior to this time. Late afternoon counts in large
marshes were to be avoided. Kirsch (in litt.) censused ducks from 0800
to 1500 hours on the Woodward Study Area. For aerial surveys of ducks of prairie
small pond habitat, J.D. Smith (1964) points out that transect counts should be
completed by noon of each day. Optimum daily and seasonal times for aerial pair
surveys are also discussed by Stewart et al. (1958) who point out that
winds generally rise toward mid-day, while light is poor in the early morning
and evening.
Daily patterns of use of marsh areas or ponds may affect optimum census times.
In North Dakota, Lacy (1959) noted that peak use of ditches by nesting pairs
occurred 2 hours after sunrise with decreasing evidence of pair use thereafter,
as birds retired to an adjacent large marsh. By 1300 hours fewer than half the
pairs enumerated in the morning were assessed. In 1967 counts, on part of Lacy's
study area, Hammond (letter, June 11, 1967) found substantially more pairs present
in the morning than afternoon. Many pairs retired to a nearby marsh after egg
laying and drakes of incubating hens also tended to fly to the marsh in afternoons.
Hammond pointed out that late afternoons were a good time to find redhead drakes
on open bays, especially during the laying period. Sowls (1955: 54) recorded
maximum populations of pairs in a roadside ditch from 0400 to 0800 hours, decreasing
pair numbers through mid-day, and minimum numbers from 1600 to 2000 hours. He
also stressed that sharing of a single loafing spot by several pairs of blue-winged
teal and gadwall may occur at different times of the day. A pair located in
one spot during one census period need not be the same pair in the spot at a
later period. For any one species, a turn-over of pairs occurred with the early,
intermediate, and late nesting hens and their drakes using the same spot for
varying portions of the morning. Heavy vehicular traffic on transect routes
may also tend to flush birds, forcing them away from well-travelled roads after
the mid-morning hours.
Generally, high post-noon temperatures and winds tend to affect mobility, visibility,
and therefore, countability of ducks. On the Kindersley Study Area many pairs,
lone, and grouped drakes rested on shore lines whenever temperatures exceeded
60°F and winds were low, a situation also reported by Diem and Lu (1960).
This behaviour and their general inactivity made them difficult to locate visually
during the afternoon, from 1200 to 1600 hours. Activity increased after 1700
hours. For parkland and marsh habitats estimates made from early morning counts,
0400 to 0800 hours, when all ducks are most active and visible, may more closely
approximate absolute breeding populations. Although visibility is increased,
estimate biases will occur because of mobility of pairs and absence of some
pairs in nesting cover or distant feeding grounds. Standardization of census
times between habitats need not be important if statistical testing of counts
shows little hourly variation in countableness or if correction factors can
be used. Most studies suggest that more consistent and accurate counts are obtainable
in the early and mid-morning than in the afternoon and evening.
Duck and Pond Distribution
Ducks and ponds are not regularly or randomly distributed over the parkland and
grassland pond habitat. Potholes do not occur "in neat patterns or regular numbers"
(Smith, Stoudt, Gollop, 1964). Pond distribution and numbers change seasonally
as some small, temporary potholes dry through April and May. Ducks themselves
tend to be found in aggregations on favourable potholes or portions of any habitat
block, i.e., a contagious distribution (Grieg-Smith, 1964; Southwood, 1966). Pairs
of some species, e.g., divers and blue-winged teal, appear more social and are
found more closely associated on "primary waiting areas" (Dzubin, 1955), than
other species. In grasslands, pair and drake pintails are more closely aggregated
than mallards. Other dabbler species tend to be more dispersed because of inherent
behavioural spacing mechanisms. Since the degree of spacing and intensity of coactions
varies with the breeding phase of each pair and the density of pairs (Dzubin,
pers. obs.), the spatial distribution of pairs, lone drakes, grouped drakes, and
grouped prebreeding and postbreeding birds, throughout the breeding season, is
a constantly changing phenomenon. In spring, newly arrived pairs are aggregated.
With the advent of nesting, pairs space themselves from other pairs. Spacing mechanisms
promote regularity of distribution while sociability leads to contagious arrangement
of indicated pairs over the habitat. As early as 1951, Lynch (1951) had recognized
seasonal spatial distribution as a major sampling problem for aerial transect
counts. He noted that when drakes and pairs start to group in late May the optimum
period for census has passed, especially since enumeration of "ganged" drakes
on transects biases the resultant density figures.
Any inferences or predictions of population densities based on single, seasonal
counts during an extremely complex period of spatial distribution are subject
to wide error, unless sampling conditions are intimately known. The accurate
measurement of density of a multi-species population in any habitat is extremely
difficult (Odum, Cantlon, and Kornicker, 1960; Preston, 1948, 1962; Williams,
1953) and requires further investigation in pond-type waterfowl habitat.
The statistical concepts underlying strip intersect methods as measures of
bird density have been discussed by Moore (1955) and Davis (1963). With waterfowl,
transect censuses over one-quarter- or one-eighth-mile-wide strips dissect a
large number of home ranges. Ponds on which ducks are enumerated may be resting,
feeding, loafing, nesting, waiting, or social congregation areas. The hourly,
daily, and seasonal use of ponds varies but replication of counts can be used
to pool data and determine average densities of breeding pairs. Density of pairs
affects the distribution pattern of pairs or lone drakes, as do favoured feeding
or loafing spots. All these factors should be weighed to better plan transect
surveys and predict the accuracy and precision obtainable from any strip census.
Previous Section -- Duck Census Techniques
Return to Contents
Next Section -- Recommendation for Block Study Area Census
U.S. Department of the Interior |
U.S. Geological Survey