Habitat suitability index models: Gadwall (breeding)

              Biological Report 82(10.100)
September 1985
HABITAT SUITABILITY INDEX MODELS: GADWALL (BREEDING)
Patrick J. Sousa
Habitat Evaluation Procedures Group
Western Energy and Land Use Team
U.S. Fish and Wildlife Service
Drake Creekside Building One
2627 Redwing Road
Fort Collins, CO 80526-2899
Western Energy and Land Use Team
Division of Biological Services
Research and Development
Fish and Wildlife Service
U.S. Department of the Interior
Washington, DC 20240
Thfs repart should be cited as:
Sausa, P. J. 1985.
U.S. Fish Wildl.
Habitat suitability index models:
Serv. Biol. Rep. 82(10.100). 35 pp.
Gadwall (breeding).
PREFACE
This document is part of the Habitat Suitability Index (HSI) Model Series
[Biological Report 82(10)] which provides habitat information useful for impact
assessment and habitat management. Several types of habitat information are
provided. The Habitat Use Information Section is largely constrained to those
data that can be used to derive quantitative relationships between key environ-mental
variables and habitat suitability. This information provides the
foundation for the HSI model and may be useful in the development of other
models more appropriate to specific assessment or evaluation needs.
The HSI Model Section documents the habitat model and includes information
pertinent to its application. The model synthesizes the habitat use informa-tion
into a framework appropriate for field application and is scaled to
produce an index value between 0.0 (unsuitable habitat) and 1.0 (optimum
habitat). The HSI Model Section includes information about the geographic
range and seasonal application of the model, its current verification status,
and a list of the model variables with recommended measurement techniques for
each variable.
The model is a formalized synthesis of biological and habitat information
published in the scientific literature and may include unpublished information
reflecting the opinions of identified experts. Habitat information about
wildlife species frequently is represented by scattered data sets collected
during different seasons and years and from different sites throughout the
range of a species. The model presents this broad data base in a formal,
logical, and simplified manner. The assumptions necessary for organizing and
synthesizing the species-habitat information into the model are discussed.
The model should be regarded as a hypothesis of species-habitat relationships
and not as a statement of proven cause and effect relationships. The model
may have merit in planning wildlife habitat research studies about a species,
as well as in providing an estimate of the relative suitability of habitat for
that species. User feedback concerning model improvements and other sugges-tions
that may increase the utility and effectiveness of this habitat-based
approach to fish and wildlife planning are encouraged. Please send suggestions
to:
Habitat Evaluation Procedures Group
Western Energy and Land Use Team
U.S. Fish and Wildlife Service
2627 Redwing Road
Ft. Collins, CO 80526-2899
iii
iv
CONTENTS
PREFACE ................................................................ iii
FIGURES ................................................................ vi
ACKNOWLEDGMENTS ........................................................ vii
HABITAT USE INFORMATION ................................................
General ...........................................................
Food ..............................................................
Water .............................................................
Cover .............................................................
Reproduction ............. . ........................................
Interspersion .....................................................
Special Considerations ............................................
HABITAT SUITABILITY INDEX (HSI) MODEL ..................................
Model Applicability ...............................................
Model Description .................................................
Application of the Model ..........................................
SOURCES OF OTHER MODELS ................................................
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Page
11
122
27
:
8
11
24
31
V
FIGURES
Number Page
1 Geographic applicability of the gadwall HSI model within
the United States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 The relationships between values of variables used to
evaluate gadwall pair habitat and suitability indices
for the variables . . . . . . ..*..............*....................... 16
3 The relationships between values of variables used to
evaluate gadwall nesting habitat in a given field and
suitability indices for the variables . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4 The relationship between the equivalent optimum area of
gadwwall nesting habitat and an overall nesting habitat
suitability index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5 The relationships between values of variables used to
evaluate gadwall brood habitat and suitability indices
for the variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
6 Summary of equations used in the gadwall HSI model . . . . . . . . . . . . . . 25
7 The relationships between habitat variables, derived
variables, life requisites, and an HSI for the gadwall . . . . . . . . . . 26
8 Definitions of variables and suggested measurement
techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
vi
ACKNOWLEDGMENTS
The basis for this HSI model was developed in a workshop that included
the following waterfowl biologists from the Northern Prairie Wildlife Research
Center, Jamestown, ND: Leo Kirsch (retired), John Lokemoen, and George
Swanson. These individuals contributed freely of their experience so that we
could develop the most reasonable HSI model possible with the current knowledge
of waterfowl habitat requirements. L. Kirsch, J. Lokemoen, and G. Swanson
also reviewed the model that resulted from the workshop. L. Kirsch provided
unpublished data that were used in the development of the nesting component
portion of this model.
In addition to the waterfowl authorities, the following potential users
of the models participated in the workshop and/or the model review: Michael
McEnroe and Steven Young (U.S. Fish and Wildlife Service, Bismarck, ND);
Richard McCabe and Robert Schultz (U.S. Bureau of Reclamation, Bismarck, ND);
and Fred Ryckman and Terry Steinwand (North Dakota Department of Game and
Fish, Bismarck, ND). The inputs of all these individuals contributed to the
content of this model. Michael Armbruster and Arthur Allen (U.S. Fish and
Wildlife Service, Ft. Collins, CO)
workshop.
served as facilitators for the modeling
The cover of this document was illustrated by Jennifer Shoemaker. Word
processing was provided by Carolyn Gulzow, Dora Ibarra, and Elizabeth Graf.
Kay Lindgren assisted with literature searches.
Funding for the development of this model was provided to the U.S. Fish
and Wildlife Service's Habitat Resources Program (Region 6) by the U.S. Bureau
of Reclamation under the provisions of the Fish and Wildlife Coordination Act.
Additional funds were provided by the U.S. Bureau of Reclamation's Program
Related Engineering and Scientific Studies (Environmental, Evaluation, and
Planning Project).
vii
GADWALL (Anas strepera)
HABITAT USE INFORMATION
General
Extensive breeding populations of the gadwall (Anas strepera) in the
United States are limited to the northern prairies and to the marshes of the
intermountain valleys of the western United States (Bellrose 1979). Isolated
breeding populations exist along the Atlantic and Alaskan coasts and in other
inland locations (Bellrose 1976). The largest numbers of breeding gadwalls
occur in the mixed-grass prairies of the Dakotas and the Prairie Provinces of
Canada. Range expansion to the eastern United States has apparently resulted
from creation of suitable habitat in the form of impoundments on Federal
refuges and state management areas (Henny and Holgersen 1974). Recent range
expansion also has been noted west of the Cascades in the northwestern United
States (Canning and Herman 1983).
Food
The diet of gadwalls during fall and winter is predominantly vegetation
(Gates 1957; Landers et al. 1976; Paulus 1982). Vegetative material accounted
for over 95% of the diet of gadwalls on a Louisiana wintering area and included
algae, dwarf spikerush (Eleocharis parvula), common widgeongrass (Ruppia
maritima), spiked watermilfoil (Myriophyllum spicatum), and baby pondweed
(Potamogeton pusillus) (Paulus 1982). The two most prominent plants in the
diet of gadwalls in South Carolina were fragrant flatsedge (Cyperus odoratus)
and Carolina redroot (Lachnanthes caroliniana) (Landers et al. 1976). Impor-tant
food plants in Utah during the fall were sago pondweed (IJ. pectinatus),
widgeongrass, and inland saltgrass (Distichlis stricta) (Gates 1957).
Animal food makes up a larger proportion of the spring and summer diet of
adult gadwalls (Serie and Swanson 1976) and ducklings (Sugden 1973). From
early spring through late August, animal material accounted for 46% of the
diet of adult gadwalls on saline wetlands in North Dakota (Serie and Swanson
1976). Crustaceans, especially those belonging to the order Anostraca, and
insects, especially adult and larval chironomids (Chironomidae), were prominent
in the animal portion of the diet. Hens ate more crustaceans and dipterans,
and less vegetation, than their mates during the egg-laying period. The
proportion of animal food in the diet reached a peak of 72% in females during
the egg-laying period. Important plants during the spring and summer period
were filamentous algae, widgeongrass, muskgrass (Chara spp.), and sago
pondweed.
Recently hatched gadwall ducklings in Alberta predominantly fed on
invertebrates, but were essentially herbivorous by 3 weeks of age (Sugden
1973). Major animal foods included adult and larval chironomids, water boatmen
(Corixidae), beetles (Coleoptera), and cladocerans (Cladocera). Important
plants in the ducklings' diet were baby pondweed, green algae (Cladophoracea),
duckweed (L-emna mino-r), and seeds of American sloughgrass (Beckmannia
syzigachne).
Gadwalls typically feed by dabbling, tipping, surface picking, and
filtering (Serie and Swanson 1976). Gadwalls in North Dakota concentrated
their feeding activities in the littoral zones of deeper, permanent wetlands
and throughout the entire basin of shallow wetlands (Serie and Swanson 1976).
Ephemeral ponds may be important sources of accessible and abundant planktonic
crustaceans early in the breeding season. Wintering gadwalls in Louisiana fed
96.7% of the time in water 15 to 67 cm deep (Paulus 1982). Broods in Alberta
also avoided feeding in areas < 15 cm deep and concentrated their feeding
activities in water 17 to 46 cm deep (Sugden 1973). Broods fed 86% of the
time over areas of submerged vegetation,
Water
The distribution and density of waterfowl is influenced to a large degree
by water permanence in available wetlands (Kantrud and Stewart 1977); wetlands
are considered to be the primary factor in waterfowl production (Higgins
1977).
The use of stock ponds by gadwall broods in South Dakota primarily was
influenced by the amount of open water (Mack and Flake 1980). Open water
sites are preferred by gadwal Is for loafing (Duebbert 1966) and as escape
cover by broods (Evans and Black 1956, cited by Mack and Flake 1980) and
adults (Flake et al. 1977, cited by Mack and Flake 1980). Gadwall pairs in
Manitoba were found in greatest abundance when the ratio of open water to
emergent vegetation was approximately 50:50, compared to wetlands with 30:70
and 70:30 ratios (Kaminski and Prince 1981). However, preference of gadwall
pairs for habitats with a 50:50 ratio of water to cover was significant for
only 1 year of the Z-year study. Trauger (1967) recommended that open water
compose at least 40% of a wetland for brood use by dabbling ducks. Murkin
et al. (1982) concluded that a water to vegetation ratio of 50:50 resulted in
the maximum density of waterfowl pairs.
Cover
Gadwalls wintering in Texas utilized fresh water habitats and use was
concentrated in the deeper waters (maximum water depth was 2.5 m; average
water depth was 1.5 m) with abundant aquatic vegetation and sparse emergent
vegetation (White and James 1978). In contrast, late summer molting cover in
Manitoba was characterized by dense stands of cattail (Typha spp.) or bulrush
(Scirpus spp.) (Oring 1969).
Reproduction
Pair habitat. Seasonal and semipermanent wetlands accounted for 54.6 and
35.4%, respectively, of use by gadwall pairs in North Dakota (Kantrud and
Stewart 1977) and 33 and 18%, respectively, of the wetland area (Stewart and
Kantrud 1973). A Z-year average of 61% of gadwall pairs in South Dakota used
natural basin wetlands, which accounted for 75% of the area of all wetland
basins (Ruwaldt et al. 1979). Semipermanent and seasonal wetlands accounted
for a 2-year average of 43.7 and 13,0%, respectively, of gadwall pair use and
32.1 and 13.4% of the wetland area. Constructed wetlands (dugouts and stock
ponds) in the same study accounted for a Z-year average of 15% of the wetland
area and 32.4% of use by gadwall pairs. Gadwall pairs in South Dakota used
ponds with scattered dense patches of emergents and avoided ponds with no
emergent vegetation (Flake and Vohs 1979). Use of wetlands by gadwall pairs
also was positively correlated with the presence of round-stem bulrushes
(ScirPus spp.) and shoreline irregularity.
Numbers of gadwall pairs over a 15-year period in Saskatchewan were
Positively correlated with the number of pairs present in the previous year
(third-order, Spearman-rank, partial correlation coefficient = 0.31; P < 0.10)
(Leitcb and Kaminski 1985). Pair numbers were not significantly correlated
with the number of wetlands available in May or in August of the previous
year.
Nesting habitat. Gadwalls nest on islands, on dikes in marshes, and in
fields and meadows, but rarely nest over water (Bellrose 1976). Fields of
seeded native grasses supported the highest number of initiated nests of
gadwalls in North Dakota, followed by seeded introduced grasses and unplowed
prairie (Klett et al. 1984). Seeded native and seeded introduced grasses
supported about the same number of initiated nests in South Dakota, followed
by unplowed prairie. Ninety-five percent of gadwall nests in an intensively
farmed area of North Dakota were in untilled uplands, with the remaining 5% in
growing grains (Higgins 1977); summer fallow areas, mulched stubble, and
standing stubble were not used. Nests may be placed in herbaceous vegetation
(Duebbert and Lokemoen 1976; Kirsch et al. 1978) or on the ground under shrub
clumps (Duebbert et al. 1983; J. T. Lokemoen, U.S. Fish and Wildlife Service,
Northern Prairie Wildlife Research Center, Jamestown, ND; pers. comm.). Most
nests are located on the driest sites available (Miller and Collins 1954;
Gates 1962; Oring 1969).
Presence of residual herbaceous vegetation may be an important habitat
factor in nest site selection (Duebbert and Lokemoen 1976; Kirsch et al. 1978;
Voorhees and Cassel 1980), although new growth may partially compensate for
the lack of residual herbaceous vegetation (Martz 1967) because gadwalls begin
nesting after new plant growth has begun (Kirsch et al. 1978; Giroux 198113;
Hines and Mitchell 1983). Data (Table 1) provided by L. M. Kirsch (U.S. Fish
and Wildlife Service, retired, Woodworth, ND; unpubl.) revealed an increase in
gadwall nesting density with an increase in the average height and density of
residual herbaceous vegetation as evaluated by a visual obstruction measure-ment
(the height at which a round pole 3 x 150 cm is totally obscured by
vegetation when viewed from a distance of 4.0 m) (Robe1 et al. 1970). Linear
regression analysis of the data in Table 1 resulted in a regression equation
of Y = 1.68 + 2.42x, a correlation coefficient (r) of 0.77 (P < O.OS), and a
coefficient of determination (r') of 0.59. Nesting densities within a given
class of visual obstruction measurements of residual vegetation varied widely,
however, as indicated by the ranges in nesting density in Table 1, suggesting
that other factors also had a major influence on nest density. One reason for
the high variability in nest densities within given visual obstruction classes
3
Table 1. Gadwall nesting densities by classes of residual
vegetation for fields on the Woodworth Study Area, North
Dakota, 1974-1978 (data provided by L. Kirsch).
x visual obstruction Number of
measurement (range) observations,
in dm in class
Mean number of
gadwall nests/40.5 ha
(range)
0.17 (0.12-0.24)
0.30 (0.25-0.34)
0.42 (0.35-0.49)
0.55 (0.50-0.60)
0.66 (0.62-0.72)
0.78 (0.74-0.83)
0.91 (0.86-0.98)
1.06 (1.01-1.14)
1.32 (1.18-1.44)
1.52 (1.45-1.60)
1.73 (1.62-1.91)
2.31 (2.01-2.86)
3.70 (3.18-4.22)
7 3.28 (0.00-14.28)
7 2.66 (0.00-8.33)
7 1.36 (0.00-3.33)
7 1.52 (0.00-5.00)
7 3.99 (0.00-9.09)
7 3.12 (0.00-8.33)
7 3.38 (0.00-8.33)
7 1.36 (0.00-4.21)
7 10.07 (0.00-25.00)
7 6.85 (0.00-16.67)
7 4.85 (0.00-13.04)
7 6.51 (2.17-13.04)
2 10.36 (6.06-14.67)
4
3s that some nests may have been initiated after new growth had begun (Kirsch,
pers. comm.), resulting in situations where residual vegetation was not a cue
used in nest site selection. Kirsch (pers. comm.) and Lokemoen (pers. comm.)
indicated that a field with an average visual obstruction measurement of
residual vegetation 1 2.5 dm would be ideal nesting habitat for gadwalls.
A study on 15 areas in North Dakota, Saskatchewan, and Manitoba suggested
a direct but weak (p = 0.06) relationship between gadwall nest density and
average visual obstruction measurement of vegetation in late spring (late
May-early June) (Shaffer et a?. 1985).
area effects,
Visual obstruction measurements, study
and number of pairs explained 26% of the total variation in
gadwall nest density. Shaffer et al. (1985) suggested the following model to
determine the number of gadwall nests (N) to be found in a given field:
N = (0.0052 + 0.0045 x X) x P x A
where X = the average visual obstruction measurement in late spring
P = the estimated number of gadwall pairs within 0.6 km of the center
of a field
A = the area of the field
The above model was considered to be most useful when comparing fields that
shared similar study area effects, i.e., fields that were located in close
proximity to each other (Shaffer et al. 1985).
High densities of gadwalls nested on a North Dakota island where
herbaceous vegetation averaged 15 to 25 cm tall at the initiation of nesting
but 1.5 to 1.8 m tall during the late incubation stages (Duebbert 1966).
Fifty-one percent of nests in North Dakota nesting fields were in herbaceous
cover from 30 to 60 cm tall, while 47% were in cover r 60 cm tall (Duebbert
and Lokemoen 1980). No nests were found in herbaceous cover < 15 cm tall.
Most gadwall nests in a California study were in vegetation 33 to 91 cm tall
that provided concealment on all sides, as well as from above (Miller and
Collins 1954). Canopy cover at gadwall nests in Saskatchewan exceeded 25% in
vegetation > 30 cm tall that provided lateral concealment on three or four
sides (Hines and Mitchell 1983). Vegetation : 20.3 cm tall is considered too
short for nest concealment (Martz 1967). Kirsch et al. (1978:492) stated that
they "... have not found grassland vegetation that was too tall and dense for
use by nesting ducks nor have .-. [they] found evidence that such conditions
exist in the prairies." Duebbert (1982:236) concluded that gadwall nesting
cover on islands '... can consist of brush, forbs, or grasses if the
vegetative structure is tal? and dense."
The highest nesting densities of gadwalls have been reported from island
habitats (Hammond and Mann 1956; Duebbert 1966, 1982; Giroux 1981a; Duebbert
et al_ 1983; Hines and Mitchell 1983), in response to the lack of mammalian
predation on the islands (Duebbert et al. 1983). Nest success on islands is
higher than reported for nests in mainland habitats (Duebbert et al. 1983;
Mines and Mitche?? 1983). For example, nest success on islands and isolated
5
ditch banks in Saskatchewan was 65%, while no nests were successful on uplands
(Hines and Mitchell 1983). High densities and success of gadwall nests in
mainland habitat in North Dakota resulted from intensive control of mammalian
predators (Duebbert and Lokemoen 1980).
Gadwall nests on islands in Alberta were in forbs and grass-forbs cover
(Giroux 1981a), and gadwall nests on Miller Lake (North Dakota) islands were
concentrated in patches of western snowberry (Symphoricarpos occidentalis) -
woods rose (Rosa woodsii) (Duebbert et al. 1983). Gadwall nests on a 2.2 ha
island in Saskatchewan were concentrated in patches of western snowberry and
slim nettle (Urtica gracilis) (Hines and Mitchell 1983). Shrub clumps in
nonisland habitats are readily used for nesting cover by gadwalls (Lokemoen,
pers. cornm.).
Preferred nesting cover is eliminated by activities such as grazing
(Kirsch 1969) or mowing (Martz 1967; Kirsch 1969; Voorhees and Cassel 1980).
Although a strong relationship has been demonstrated between duck nesting
densities and undisturbed cover (Kirsch et al. 1978), mowing may be useful for
maintaining vegetative cover in earlier, more productive successional stages
(Voorhees and Cassel 1980). Duebbert et al. (1981) recommended periodic
disturbance to native and introduced grassland nesting habitat to maintain
optimum conditions, although annual mowing or grazing was not recommended.
Brood habitat. Preferred escape cover for gadwall broods is large areas
of open water, rather than water with emergents (Evans and Black 1956, cited
by Mack and Flake 1980). Gadwall broods in Utah used deep-water marshes and
the edges of large impoundments (Gates 1962); broods in Washington used large
alkaline lakes with steep walls, as well as other wetlands (Yocom and Hansen
1960). Sixty-one percent of 1,073 gadwall broods observed over a 20-year
period in North and South Dakota were in semipermanent wetlands (Class IV of
Stewart and Kantrud 1971), 18% were in seasonal wetlands (Class III), and 9%
were in permanent wetlands (Class V) (Duebbert and Frank 1984). The proportion
of total wetland area accounted for by these wetland types in North Dakota in
1967 was 18% semipermanent, 36% seasonal (including 3% in tilled condition),
and 3% permanent (Stewart and Kantrud 1973). However, wetland availability
figures were for 1 year only, apparently reflected the availability of wetlands
to pairs, and may not be a valid estimate of the wetland distribution available
to broods. Amount of open water and number of wetland basins/O.65 km2 plot
were the primary factors that determined use of stock ponds by gadwalls in
South Dakota (Mack and Flake 1980). The mean open water area on stock ponds
used by gadwall broods was 1.4 ha, compared to a mean of 0.6 ha on ponds not
used by gadwall broods. Wetland basins averaged 5-O/0.65 km* on study plots
used by gadwall broods, but only 2.8/0.65 km* on plots not used by broods.
Gadwall broods over a 15-year period in Saskatchewan were positively correlated
with the number of pairs in the preceding spring (second-order, Spearman-rank,
partial correlation coefficient = 0.56; P < 0.01) and the number of wetlands
containing water in August (second-order, Spearman-rank, partial correlation
coefficient = 0.43; P < 0.05) (Leitch and Kaminski 1985).
The gadwall is the primary waterfowl species in North Dakota that uses
saline lakes for brood-rearing (Swanson et al. 1984). Brood use is closely
tied to the presence of freshwater seeps or areas of lower salt content.
6
These areas provide fresh water for drinking and support dense emergent
vegetation which provides cover for broods.
Interspersion
The average distance from nest sites to water was c 45.8 m in several
studies of gadwalls (Miller and Collins 1954; Gates 1962; Vermeer 1970).
Gadwall nests in North Dakota averaged 351 m from water (Duebbert and Lokemoen
1976), including SOme fW?StS in fields up to 2.4 km from water (Duebbert and
Lokemoen 1980). G. A. Swanson (U.S. Fish and Wildlife Service Northern
Prairie Wildlife Research Center, Jamestown, ND; pers. comm.) sugg,&ted that
selection of nesting habitat by gadwalls is based on proximity to pair feeding
habitat rather than on proximity to brood-rearing habitat.
Gadwall hens in Utah moved their broods an average of 0.9 km and a maximum
of 1.85 km from the nest to brood habitat (Gates 1962). Hens and broods in
South Dakota dispersed into wetlands 1.6 to 3.2 km from the nests (Duebbert
and Lokemoen 1980). Breeding home ranges for five hens in Utah averaged 27.1
ha (range 13.8 to 35.2 ha) and included at least one feeding pond and a ditch
or channel used for loafing (Gates 1962).
Island-nesting gadwalls may reach high densities, for example: 494 nests/
ha on a 0.32 ha island in North Dakota (Hammond and Mann 1956); 139 to 237
nests/ha in preferred island habitat in North Dakota (Duebbert et al. 1983);
and 74 nests/ha on a 2.2 ha island in Saskatchewan (Hines and Mitchell 1983).
Nest density in the latter study was 284 nests/ha in two patches of snowberry,
totalling 0.5 ha (Hines and Mitchell 1984). Nest parasitism by other gadwalls
(31 of 355 nests) at this high nest density reduced nesting success from 76 to
54%, egg success from 74 to 45%, and hatchability of eggs from 97 to 91%
(Hines and Mitchell 1984). Parasitism by lesser scaup (Aythya affinis) reduced
egg success from 74 to 67%. Nest densities elsewhere on the island ranged
from 13 to 52 nests/ha; only one parasitized nest was found at the lower nest
densities.
Nest densities in nonisland habitats are generally much Tower than those
observed on islands. For example, Kaiser et al. (1979) recorded only 0.67
nests/km' in native grasslands in South Dakota and 3.88 nests/km' in tame
grasslands. The estimated number of initiated gadwall nests in North Dakota
was 28/km2 in seeded native grasses, 14/km2 in seeded introduced grasses, and
4/km2 in unplowed prairie (Klett et al. 1984). Gadwalls in South Dakota
initiated an estimated 10 nests/km2 in seeded native grasses, II/km2 in seeded
introduced grasses, and O/km2 in unplowed prairie. Nest density in untilled
uplands and growing grains in North Dakota was 4.33 nests/km' and o-23 n@sts/
km2, respectively (Higgins 1977). A density of 7.5 nests/km' (61 nests on a
8.13 km2 study area) was observed in an area of intensive Predator control
(Duebbert and-Lokemoen 1980).
Areas with diversified land uses are better for duck production
large expanses of tilled grain monocultures (Duebbert and Lokemoen 1976).
than
Special Considerations .
Island-nesting waterfowl, such as gadwalls, require suitable wetlands for
pair and brood-rearing habitat (Duebbert et al. 1983). Creation of nesting
habitat in the form of small islands should consider the carrying capacity of
surrounding wetlands for pairs (Hines and Mitchell 1983). Habitat management
for waterfowl production must involve both wetland and upland habitat (Leitch
and Kaminski 1985).
HABITAT SUITABILITY INDEX (HSI) MODEL
Model Applicability
Geographic area. This HSI model was originally developed for use in
central and eastern North Dakota.
Prairie Pothole Region,
It is considered applicable throughout the
occur (Fig. 1).
where the greatest breeding densities of gadwalls
Within the United States this region includes the mixed-grass
prairie of North and South Dakota; the tallgrass prairie in western Minnesota,
eastern North and South Dakota, and the sandhills of Nebraska: and the short-grass
1979).
Canada
gadwal 1
prairie west of the Missouri River through Montana '(Bellrose 1976
The model also should be applicable within the Prairie Provinces oi
and may be applicable in other portions of the breeding range of the
Figure 1. Geographic applicability of the gadwall
the United States (corresponds to areas of highest
densities, as shown in Bellrose 1976).
8
HSI model within
gadwall breeding