AN ABSTRACT OF THE THESIS OF 
Alan D. Weinbrenner for the Master of Science Degree 
in Biology presented on 19 November 1991 
Title: Nest site Selection on Flint Hills Impoundments by 
Canada Geese 
Abstract approved: 
- V 
Twenty-four habitat variables considered important to 
the selection of a nest site by Canada Geese were evaluated 
at 59 impoundments (n = 55 stockponds, n = 4 watershed 
lakes) in northern Greenwood County, Kansas, during the 1989 
and 1990 nesting seasons. Impoundments with nesting pairs 
present, termed "occupied," were compared to impoundments 
without nesting pairs present, termed "vacant." 
Eleven pairs of Canada Geese during 1989 and 27 pairs 
during 1990 produced broods. Twenty-four nests during 1989 
and 26 nests during 1990 were located on artificial 
structures. 
Occupied impoundment size was significantly larger than 
vacant impoundment size during both 1989 and 1990 when 
watershed lakes were included in the analysis (E < 0.05). 
When watershed lakes were removed from analysis, occupied 
impoundment size was not significantly larger than vacant 
impoundment size during 1989, however, occupied impoundment 
size was significantly larger than vacant impoundment size 
during 1990 . 
.l. 
~ 
r 
Analysis of 18 habitat variables by stepwise 
discriminant function analysis indicated that brood use of 
an impoundment, presence of an artificial structure, 
impoundment fenced to exclude livestock, and the average 
distance to the nearest three vacant impoundments explained 
49% of the variation between occupied and vacant 
impoundments during 1989. During 1990, the presence of an 
artificial structure, brood use of an impoundment, 
impoundment size, and the percent of shoreline having shrubs 
explained 41% of the variation between impoundment classes. 
Thirty-eight vacant impoundments (97.4%) and eight 
occupied impoundments (13.8%) were correctly classified by 
discriminant function analysis during 1989; 12 occupied 
impoundments were misclassified as vacant impoundments. 
During 1990, 38 vacant impoundments (100%) and 21 occupied 
impoundments (100%) were correctly classified. 
Brood use of occupied and vacant impoundments was 
significantly different during both years. Climatic 
ponditions, previous nesting success, presence of an 
ificial structure, and brood use of an impoundment 
,~peared to be important factors influencing nest site 
selection by Canada Geese. 
L
 
NEST SITE SELECTION ON
 
FLINT HILLS IMPOUNDMENTS BY
 
CANADA GEESE, Branta canadensis
 
A Thesis
 
Presented to
 
the Division of Biological Sciences
 
EMPORIA STATE UNIVERSITY
 
In Partial Fulfillment
 
of the Requirements for the Degree
 
Master of Science
 
by 
Alan D. Weinbrenner 
19 December 1991 
iv 
//J;U·~/!d 
r Major Department 
j~ n. Uoweii 
Appro ed for Graduate Council 
f 
I 
J.
 
-~
 ACKNOWLEDGMENTS
 
The completion of this study would not have been 
possible without the valuable assistance of numerous 
individuals. I am greatly indebted to Dr. Carl Prophet, my 
major professor, for advice, help, understanding, and 
encouragement during this study. I thank Dr. Dwight Moore 
for help with editing and advice on the statistical analyses 
of my data. Thanks and appreciation go to Gerald Horak, 
Kansas Department of wildlife and Parks (KDWP), for 
encouragement and information on the Canada Goose 
restoration program that he coordinated in Kansas, and to 
Lloyd Fox, (KDWP), for help with statistical analyses 
throughout this study. I also thank Marvin Kraft (KDWP) for 
his suggestions and assistance in completing this study. 
The field investigation during the study was greatly 
enhanced by the assistance of retired rancher Mr. Charles 
"Bud" Wiggins and his family of Eureka, Kansas. Bud's help 
in locating nest structures, nesting geese, and in acquiring 
permission for access to impoundments from the numerous 
landowners was invaluable during this study. I thank the 
Greenwood County Department of Agriculture Soil Conservation 
Service for the use of aerial photographs of the study area 
and help in locating area landowners. I also thank Mr. 
Richard Tush (KDWP) for help during the study. special 
thanks go to the ranchers of Greenwood County who gave me 
permission to study impoundments on their land. 
I am grateful to fellow Emporia state University 
students Rolland Fraser and Jo Anne Pritchard for their help 
in the field investigation part of this study. Financial 
support and equipment for this study were provided by the 
KDWP and an ESU Graduate Research Assistantship during the 
1989-1990 school year, and also by an ESU Educational 
opportunity Scholarship provided by the Delta Kappa Chapter 
of Beta Beta Beta. Finally, special thanks go to my wife 
Tamara, for her help and patience in completing this study 
and to my family and friends for their support during my 
college career. 
vii 
TABLE OF CONTENTS
 
PAGE
 
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . .viii
 
LIST OF FIGURES
 ix
 
GENERAL INTRODUCTION 1
 
Historical Range of the Giant Canada Goose. . . . . . 1
 
Introduction Program in Kansas 4
 
Objectives of Study. . . . . . . . . . . . . . . . . 5
 
DESCRIPTION OF STUDY AREA . . . . . . . . . . . . . . . . 6
 
METHODS . . . . . . . . . . . . . . . . . . . . . . . . . 10
 
Habitat Characteristics 10
 
statistical Analyses. . . . . . . . . . . . . . . . 15
 
RESULTS . . . . . . . . . . . . . . . . . . . . . .. . . 18
 
1989 and 1990 Nesting Seasons 18
 
Nesting Success. . . . . . . . . 20
 
Artificial Nest Structure Use 22
 
Habitat Characteristics. . . . . . . . . . . . . . . 23
 
Analyses of Habitat 27
 
DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . 37
 
Nest Site Selection. . . . . . . . . . . . . . . . . 37
 
Nesting Success. . . . . . . . . . . . . 39
 
Artificial Nest Structure Use. . . . . . . . . . . . 41
 
Brood Use of Impoundment. . . . . . . . . . . . . . 42
 
Habitat Characteristics
 44
 
Analyses of Habitat. . . . . . . . . . . . . . . . . 46
 
viii 
TABLE OF CONTENTS CONTINUED 
MANAGEMENT. • • • • • • • • • • • • • • • • • • • • • • • 49 
LITERATURE CITED • • • . • • • • • • • • • • • • • • • • 50 
ix 
LIST OF TABLES 
TABLE PAGE 
1
 Identification number, legal description, and
 
location of the study impoundments ..
 11 
2	 The 24 habitat variables selected for study during 
1989 and 1990 12 
3	 Comparison of 20-year precipitation and temperature 
averages to conditions during the fall of 1988, all 
of 1989, and spring 1990. . . . . . . . . . . . . . 19 
4	 Use of the study impoundments by Canada Geese during 
the 1989 and 1990 nesting seasons 21 
5	 Impoundment surface areas during 1989 and 1990. . . 24 
6	 Results of habitat variables evaluated during the 
study (see Table 2 for definition of variables) . . 26
 · 
7	 Results of habitat variables during 1989 using 
tc-tests (see Table 2 for definition of variables) . 28
 · 
8	 Results of habitat variables during 1990 using 
tc-tests (see Table 2 for definition of variables) . 28
 · 
9	 The contribution of 18 habitat variables used to 
distinguish between impoundment classes during the 
1989 nesting season using, stepwise discriminant 
function (see Table 2 for definition of variables). 30 
10	 The contribution of 18 habitat variables used to 
distinguish between impoundment classes during the 
1990 nesting season using, stepwise discriminant 
function (see Table 2 for definition of variables). 32 
11	 Presence or absence of Canada Geese on impoundments 
as predicted by discriminant function analysis and 
based on 18 habitat variables during the 1989 and 1990 
nesting seasons . . . . . . . . . . . . . . . . . . 33 
12	 Results of 1989 Pearson product correlation 
coefficients for 18 habitat variables (see Table 2 for 
definitions of variables). . . . . . . . . . . . . . 34 
13	 Results of 1990 Pearson product correlation 
coefficients for 18 habitat variables (see Table 2 for 
definitions of variables). . . . . . . 35 
LIST OF FIGURES 
FIGURE PAGE 
1 Historical range of the Giant Canada Goose
 
(shaded area, adapted from: Hanson 1965) ..
 3
 
2	 Location of the study area (outlined in black) in
 
northern Greenwood County, Kansas, during 1989
 
and 1990. . . . . . . . . . . . . . . . . . . . . . 8
 
INTRODUCTION 
The Canada Goose (Branta canadensis) has been 
subdivided into numerous subspecies based on differences in 
body size, plumage, weight, and distribution on wintering 
and nesting grounds. Bellrose (1976) recognized 11 
sUbspecies, while Johnsgard (1978) described 12 subspecies. 
Regardless of the actual number of sUbspecies that have 
existed, the Giant Canada Goose (~. £. maxima) was perhaps 
the only subspecies that nested in Kansas (Hanson 1965). 
Historical Range of the Giant Canada Goose 
The historical range of the Giant Canada Goose probably 
included the tall grass and mixed grass prairies of the 
United states and Canada, as well as the parkland region of 
Canada (Hanson 1965, see Figure 1). Johnsgard (1978) 
concluded this subspecies originally nested from Manitoba 
south into Kansas, Missouri, Arkansas, and Tennessee. Also, 
Hanson (1965) believed that the Giant Canada Goose nested 
along rivers in Kansas before it became extinct, in Kansas, 
during the early 1900s. He stated that Cheyenne Bottoms in 
Barton County and McPherson Bottoms in western McPherson 
County may have been important nesting grounds for this 
subspecies prior to the arrival of settlers in Kansas. As 
agricultural activity in the area increased, McPherson 
Bottoms disappeared; but Cheyenne Bottoms continued to be an 
important nesting, breeding, and stopover point. 
Records of nesting by Giant Canada Geese in other parts 
of North America during the late 1800s and early 1900s were 
Figure 1. Historical range of the Giant Canada Goose 
(shaded area, adapted from: Hanson 1965). 
2 
gOl 008 006 0001 
011 X. L/ 
008 Jl6 
.... , 
\ 
" 
\ -... , J ~ 
... ,
 ,
 ,
 _---::a...J 
4 
also incomplete. According to Cooper (1978), the 
disappearance of nesting Giant Canada Geese in southern 
Manitoba probably occurred during the period of rapid 
settlement from 1870 to 1900. 
By the mid-1930s the Giant Canada Goose was believed 
extinct throughout its entire range in North America (Hanson 
1965). However, surviving members of this subspecies were 
discovered in a city park in Rochester, Minnesota, during 
January 1962. Biologists of the Minnesota Department of 
Conservation and the u.s. Fish and Wildlife Service noticed 
differences in size, weight, and plumages among Canada Geese 
that they captured during banding operations; and they 
concluded that some of the captured birds were Giant Canada 
Geese (Hanson 1965). 
Introduction program in Kansas 
In 1980, the KDWP initiated a program to restore 
nesting populations of Canada Geese in eastern Kansas. 
Adult Canada Geese that were used in the program were 
obtained mainly from Colorado and Wisconsin. However, 
Gerald Horak (KDWP, pers. comm.) stated that these geese may 
not have been purebred Giant Canada Geese, but rather a 
mixture of several different sUbspecies of Canada Geese. 
Therefore, in my study no attempt was made to distinguish 
between the separate subspecies. These birds were wing­
 clipped and held at the Mined Land wildlife Area in Cherokee 
County and at facilities near Cheney, EI Dorado, Fall River, 
L
 
5 
Melvern, and Glen Elder reservoirs. After one year, these 
birds were released. Other adult geese were pinioned and 
held in goose pens at various sites across the state, 
including Cedar Bluff Reservoir, Marais Des Cygnes Wildlife 
Area, and at facilities near Pratt (Gerald Horak, pers. 
comm.). Goslings produced by these captive adult geese were 
removed from the pens before they could fly, at ages 8-9 
weeks, and then released at selected sites within the Flint 
Hills. It was hoped that after sexual maturity the goslings 
would return to the same area where they learned to fly and 
attempt to nest (Brakhage 1965). 
Objectives of study 
Because little was known about the nesting behavior of 
Canada Geese in Kansas (Hanson 1965), a study of the nesting 
of this species in the Flint Hills was conducted. The 
primary objectives of the study were to determine whether 
impoundments used as nesting sites shared certain common 
features and to determine if it were possible to distinguish 
between impoundments which were suitable and which were 
unsuitable for nesting. Knowledge of the preferred nesting 
habitat of the Canada Goose may be used during future 
efforts to restore and manage this species in the Flint 
Hills of Kansas. 
DESCRIPTION OF STUDY AREA 
The study was conducted in an area of approximately 374 
km2 in northern Greenwood County, Kansas, during the Springs 
of 1989 and 1990 (see Figure 2). This site was selected for 
study because KDWP personnel had initially released the 
majority of the Canada Geese adult and goslings in this 
area. 
Fortner et ale (1982) described the soils of the area 
as deep to moderately deep, gently sloping to moderately 
sloping with a clay subsoil. The Clime-Sogn-Martin soils 
complex, with a slope range of 0 to 30% and the 
Eram-Labette-Kenoma soil complex, with a slope range of 0 to 
8%, made up 75% of the soil types found in Greenwood County. 
Impoundments in Greenwood County had a clay substrate 
overlying a limestone bedrock that was usually less than 100 
cm below ground level. These shallow soils were easily 
eroded causing siltation of impoundments. 
The major land use in this area was livestock grazing, 
with some crop production on the more suitable soil types. 
Dominant native grasses of this region included big bluestem 
(Andropogon gerardi), little bluestem (Andropogon 
scoparius), indiangrass (Sorghastrum nutans), switchgrass 
(Panicum virgatum), and sideoats grama (Bouteloua 
curtipendula). Smooth brome (Bromus inermis) and tall 
fescue (Festuca arundinacea) were domestic grasses that were 
introduced for early and late season grazing and haying 
7 
Figure 2.
 Location of the study area (outlined in black) in 
northern Greenwood County, Kansas, during 1989 
and 1990. 
8 
09& 
00MN331:19 
811-' Je6unW"£ 
811-' All:> -1I8Jn3"Z 
'181:1'J:> Inol 
I 
09& 
9 
purposes; however, they accounted for only seven percent of 
the pasture acreage in Greenwood County (Fortner et ale 
1982) • 
Vegetation types that were found along the shoreline of 
the study impoundments included aquatic macrophytes, shrubs, 
trees, or domestic or native grasses. Aquatic vegetation 
included cattails (Typha sp.), barnyardgrass (Echinochloa 
crusgalli), sedge (Carex sp.), curly dock (Rumex crispus), 
and smartweed (Polygonum sp.). Shrub species included 
aromatic sumac (Rhus aromatica), smooth sumac (Rhus glabra), 
wild plum (Prunus americana), and rose (Rosa sp.). Common 
tree species included cottonwood (Populus deltoides), red 
cedar (Juniperus virginiana), mulberry (Morus sp.), and 
willow (Salix sp.). 
METHODS 
Selection of the study impoundments began 24 March 1989 
and continued through 31 May 1989. After permission was 
received from area landowners and operators, impoundments 
were visited to determine if they were used for nesting 
sites by Canada Geese. The impoundments that were visited 
were divided into two classes; occupied and vacant. 
Impoundments where a ground nest was found or where an 
artificial nest structure was used were classed as occupied, 
even if the nest was later abandoned or destroyed. If there 
was no evidence of nesting activity at an impoundment, it 
was designated vacant. Each impoundment that was visited 
was assigned an identification number for future reference 
(see Table 1). An aerial survey of the study area was 
conducted on 13 April 1989 to identify those impoundments 
that were occupied by Canada Geese. 
Habitat Characteristics 
Twenty-four habitat variables that were assumed to 
influence the selection of a nesting site by Canada Geese 
were selected for evaluation at each study impoundment. 
Variables that were chosen were based on previous nesting 
studies of Giant Canada Geese (Hamilton 1978, Bultsma et al. 
1979, Poly 1979, and Stiefel 1980), on other waterfowl 
species (Lokemoen 1973, Mack and Flake 1980, Rumble and 
Flake 1983, Lokemoen, Duebbert, and Sharp 1984, and Belanger 
and Couture 1988) and on numerous discussions with Gerald 
Horak (KDWP). These variables were summarized in Table 2. 
11 
Table 1. Identification number, legal description, and 
location of the study impoundments. 
Impoundment: Township: LegaL Description: 
BAN-12-1 BACHELOR NORTH W1/2E1/2SW1/4 Sec.12,T25S, R11E. 
BAN-12-4 BACHELOR NORTH W1/2E1/2SW1/4 Sec.12,T25S, R11E. 
BAN-12-5 BACHELOR NORTH NW1/4NW1/4SW1/4 Sec.12,T25S, R11E. 
BAN-14-1 BACHELOR NORTH NE1/4NE1/4 Sec.14,T25S, R11E. 
BAN-17-1 BACHELOR NORTH S1/2N1/2N1/2S1/2SW1/4 Sec.17,T25S,R11E. 
BAN-17-2 BACHELOR NORTH SE1/4N1/2SW1/4 Sec.1?,T25S, R11E. 
BAN-17-3 BACHELOR NORTH NE1/4S1/2SW1/4 Sec.17,T25S, R11E. 
BAN-17-4 BACHELOR NORTH SW1/4SW1/4NE1/4 Sec.17,T25S, R11E. 
BAN-17-5 BACHELOR NORTH S1/2S1/2SW1/4 Sec.17,T25S, R11E. 
BAS-5-1 BACHELOR SOUTH SW1/4SW1/4NW1/4 Sec.5,T26S, R11E. 
BAS-5-2 BACHELOR SOUTH N1/2S1/2NW1/4 Sec.5,T26S, R11E. 
BAS-5-3 BACHELOR SOUTH NW1/4NW1/4SW1/4 Sec.5,T26S, R11E. 
BAS-5-4 BACHELOR SOUTH NW1/4NW1/4SW1/4 Sec.5,T26S, R11E. 
BAS-5-5 BACHELOR SOUTH E1/2SW1/4 Sec.5,T26S, R11E. 
BAS-6-1 BACHELOR SOUTH S1/2S1/2NW1/4S1/2S1/2NE1/4 Sec.6,T26S, R11E. 
BAS-6-2 BACHELOR SOUTH E1/2NW1/4W1/2NE1/4 Sec.6,T26S, R11E. 
BAS-6-3 BACHELOR SOUTH NE1/4SW1/4 Sec.6,T26S, R11E. 
EUN-2-2 EUREKA NORTH S1/2NE1/4 Sec.2,T25S, R10E. 
EUN-3-1 EUREKA NORTH NE1/4NW1/4 Sec.3,T25S, R10E. 
EUN-4-1 EUREKA NORTH N1/2S1/2NW1/4 Sec.4,T25S, R10E. 
EUN-4-2 EUREKA NORTH W1/2E1/2SE1/4 Sec.4,T25S, R10E. 
EUN-5-1 EUREKA NORTH NE1/4SW1/4 Sec.5, T25S, R10E. 
EUN-5-2 EUREKA NORTH SW1/4SW1/4 Sec.5, T25S, Rl0E. 
EUN-5-3 EUREKA NORTH W1/2SE1/4 Sec.5, T25S, R10E. 
EUN-5-4 EUREKA NORTH SW1/4NE1/4 Sec.5, T25S, R10E. 
EUN-7-1 EUREKA NORTH W1/2NE1/4 Sec.7, T25S, R10E. 
EUN-7-2 EUREKA NORTH W1/2NE1/4 Sec.?, T25S, R10E. 
EUN-7-3 EUREKA NORTH E1/2NW1/4 Sec.7, T25S, R10E. 
EUN-7-4 EUREKA NORTH E1/2NW1/4 Sec.?, T25S, R10E. 
JAW-34-1 JANESVILLE WEST S1/2S1/2SE1/4 Sec.34,T24S, Rl0E. 
JAW-34-2 JANESVILLE WEST Sl/2SW1/4 Sec.34,T24S, Rl0E. 
JAW-34-3 JANESVILLE WEST Wl/2Sl/2NE1/4 Sec.34,T24S, Rl0E. 
JAW-34-4 JANESVILLE WEST N1/2S1/2SE1/4 Sec.34,T24S, R10E. 
JAW-34-5 JANESVILLE WEST SE1/4NE1/4 Sec.34,T24S, R10E. 
JAW-34-6 JANESVILLE WEST NE1/4NE1/4NE1/4 Sec.34,T24S, R10E. 
JAW-34-7 JANESVILLE WEST Nl/2Sl/2NE1/4 Sec.34,T24S, Rl0E. 
JAW-34-8 JANESVILLE WEST NW1/4SW1/4 Sec.34,T24S, R10E. 
SSE-3-1 SOUTH SALEM EAST E1/2Wl/2W1/2E1/2 Sec.3, T25S, R9E. 
SSE-3-2 SOUTH SALEM EAST NW1/4NW1/4SE1/4 Sec.3, T25S, R9E. 
SSE-3-3 SOUTH SALEM EAST E1/2Nl/2SE1/4 Sec.3, T25S, R9E. 
SSE-3-4 SOUTH SALEM EAST SW1/4N1/2SE1/4 Sec.3, T25S, R9E_ 
SSE-3-6 SOUTH SALEM EAST SW1/4SW1/4 Sec.3, T25S, R9E. 
SSE-10-l SOUTH SALEM EAST NW1/4NW1/4 Sec.l0, T25S, R9E. 
SSE-10-2 SOUTH SALEM EAST SE1/4SE1/4NW1/4 Sec.l0, T25S, R9E. 
SSE-10-3 SOUTH SALEM EAST SE1/4NE1/4 Sec.l0, T25S, R9E. 
SSE-10-4 SOUTH SALEM EAST SE1/4SE1/4 Sec.10, T25S, R9E. 
SSE-ll-1 SOUTH SALEM EAST SW1/4SW1/4 Sec.l1, T25S, R9E. 
SSE-14-1 SOUTH SALEM EAST NW1/4 Sec.14, T25S, R9E. 
SSE-22-1 SOUTH SALEM EAST NW1/4NW1/4SE1/4 Sec.22, T25S, R9E. 
SSE-23-1 SOUTH SALEM EAST El/2NW1/4 Sec.23, T25S, R9E. 
SSE-27-1 SOUTH SALEM EAST E1/2W1/2S1/2NE1/4 Sec.27, T25S, R9E. 
SSE-27-2 SOUTH SALEM EAST Wl/2E1/2S1/2NE1/4 Sec.27, T25S, R9E. 
SSE-33-l SOUTH SALEM EAST NW1/4SE1/4 Sec.33, T24S, Rl0E. 
SSE-33-2 SOUTH SALEM EAST Nl/2S1/2SE1/4 Sec.33, T24S, R10E. 
SSE-33-3 SOUTH SALEM EAST SW1/4SE1/4 Sec.33, T24S, Rl0E. 
SSW-28-1 SOUTH SALEM WEST Sl/2NE1/4Nl/2SE1/4 Sec.28, T24S, R9E. 
SSW-28-2 SOUTH SALEM WEST SW1/4SW1/4NE1/4 Sec.28, T24S, R9E. 
SSW-28-3 SOUTH SALEM WEST Wl/2SE1/4 Sec.28, T24S, R9E. 
SSW-29-1 SOUTH SALEM WEST N1/2Sl/2SE1/4 Sec.29, T24S, R9E. 
12 
Table 2. The 24 habitat variables selected for study during 1989 and 1990. 
label: Variable Description: 
IMSIZE 
IMTYPE 
SHII 
SSl 
PSHlEX 
ISPR 
AGI 
PSHlEV 
PSHlSV 
PSHlIN 
lAUA 
CUCT 
NIW1.6 km 
IDSRS 
DISN31 
DISNF 
DISNCC 
NNPI 
PRW 
IMBROOO 
GNP 
PRAN 
ART 
IMFENCE 
Iq>oundment Size 
Iq>oundment Type 
Shoreline Irregularity Index 
Slope of Surrounding land 
Percent of Shoreline Exposed or Dry 
Presence of NaturaL or Man Made Islands 
Age of Impoundment 
Percent of ShoreLine Emergent Vegetation 
Percent of ShoreLine Shrub Vegetation 
Percent of Shoreline Trees 
landuse of Area 
Cultivated Crop Type 
Number of Impoundments Within 1.6 km Radius of each Study Impoundment 
Impoundment Distance to Release Site 
Distance to Nearest 3 Vacant Impoundments 
Distance to Nearest Farmstead 
Distance to Nearest Cultivated Crop 
Number of Nesting Pairs on Impoundment 
Presence of Other Waterfowl 
Impoundment Use by Broods 
Ground Nest Present 
Presence of Artificial Nest Structure 
Artificial Nest Structure Type 
Impoundment Fenced to Exclude Livestock 
13 
Physical data recorded at each impoundment included 
size (IMSIZE), type (IMTYPE), shoreline development (SHII), 
slope of the surrounding land (SSL) , percent of the 
shoreline exposed (PSHLEX), presence of islands (ISPR), and 
age of the impoundment (AGI). Shoreline development was 
determined according to Wetzel (1975). The percent of 
shoreline exposed was determined using a standard measuring 
wheel. This value was then subtracted from the impoundment 
size value estimated from Agricultural Stabilization and 
Conservation services aerial photographs, using a 
planimeter. The resultant value expressed the percentage of 
an impoundment as exposed or dry. The age of impoundments 
was estimated from discussions with area landowners and 
operators and the Greenwood County Soil Conservation 
Service. 
Vegetational data included the percent of the shoreline 
having vegetation classified as emergent macrophytes 
(PSHLEV), shrubs (PSHLSV), trees (PSHLWV), or as bareground, 
or domestic or native grasses. The values for these 
variables were estimated either visually or from aerial 
photographs. 
Land use categories (LAUA) surrounding, or near, an 
impoundment were rangeland, pasture, and cultivated cropland 
(Mack and Flake 1980). The cultivated crop type (CCUT), 
such as wheat, milo, alfalfa or soybeans, were determined 
visually. 
14 
Variables that characterized the degree of isolation or 
the availability of impoundments, included the number of 
impoundments within a 1.6-km radius (Rumble and Flake 1983) 
of each study impoundment (NIW1.6 km), the distance to the 
nearest release site (IDSRS), and the average distance to 
the nearest three vacant impoundments (DISN3I). The values 
for these variables were estimated from aerial photographs. 
The distance to the nearest farmstead (DISNF) and the 
distance to the nearest cultivated field (DISNCC) were also 
estimated from aerial photographs. 
Variables that characterized the use of impoundments by 
nesting pairs of Canada Geese and other waterfowl included 
the number of nesting pairs present on an impoundment 
(NNPI), the presence of waterfowl other than the nesting 
pair (PRW), brood use of an impoundment (IMBROOD), and the 
presence of at least one ground nesting pair (GNP). The 
presence of an artificial nest structure (PRAN) and the 
artificial nest structure type (ART) were observed and 
recorded. The variable (IMFENCE) impoundment fenced to 
exclude livestock was included to see if the presence of 
livestock negatively influenced the presence of a nesting 
pair (stiefel 1980), or if fenced impoundments where the 
vegetation surrounding the impoundment was not grazed by 
livestock, or burned, or mowed negatively influenced the 
presence of a nesting pair (Gerald Horak, pers. comm.). 
15 
Nest sites at occupied impoundments were examined for 
the presence of either eggs or goose down. Usually, clutch 
size and distance from the nest structure to water were 
recorded during the initial visits to reduce the number of 
times a nest site was visited. This procedure reduced the 
probability of the adults deserting the nest because of 
frequent disturbance (Poly 1979). Impoundments were visited 
weekly (weather permitting) during the nesting season until 
all habitat variables were measured or evaluated. When the 
opportunity arose, initial brood size was recorded; however, 
individual broods were not followed closely throughout the 
summer. 
Weather data for Eureka, Kansas (closest reporting 
weather site) were obtained from the National Oceanic and 
Atmospheric Administration (NOAA) Climatological Data for 
Kansas (1988, 1989, 1990). Weather data pertaining to the 
20-yr precipitation and temperature averages were obtained 
from Fortner et al. (1982). 
Statistical Analyses 
All data were recorded in a dBASE III PLUS file 
(ASHTON-TATE 1985) and later converted to a SAS-program file 
(SAS Institute 1979) for the statistical analyses. Basic 
statistics (mean, standard deviation, standard error, 
minimum, maximum) and Pearson product correlation 
coefficients were calculated for each variable using SAS 
(SAS Institute 1979). Pearson product correlation 
16 
coefficients represented associations between variables. 
The higher the absolute value of the correlation coefficient 
value between two variables the greater the association 
between those two variables. Negative coefficient values 
represented negative associations between two variables. 
An ~-max. test was used to see if the variances of 
occupied and vacant impoundments were equal. If the 
variances were equal a standard t-test was used to determine 
if the means of occupied and vacant impoundments were 
significantly different. If the variances were unequal then 
an approximation to the t-test was used (Behrens-Fisher 
problem, in SAS; SAS Institute 1979) to determine if the 
means of occupied and vacant impoundments were significantly 
different. After initial analysis of the habitat variables 
by ~-tests, all the data pertaining to watershed lakes were 
removed from the analysis to determine if inclusion of the 
watershed lakes influenced the results of the t-tests. 
Stepwise discriminant function analysis was used to 
measure the discriminating power of each variable. In this 
test, dependent variables were not continuous but were 
separated into classes (Mack and Flake 1980). At each step 
in the procedure the variable with the highest ~-statistic 
was entered into the analysis, and this stepwise selection 
continued until no more variables were entered (~ < 0.05). 
The resulting list of variables "best explained" the 
differences between occupied and vacant impoundments. 
17 
Discriminant function was used to analyze the data set. 
The first part of the discriminant function analysis was a 
classification technique used to measure how well the 
variables separated the two classes of impoundments: 
occupied and vacant impoundments. The program gave each 
impoundment a classification score based on the value of 
each recorded variable. Each impoundment received a 
classification score for each of the two classes, and then 
it was assigned to the class for which it obtained the 
highest score. The percent of impoundments that were 
correctly classified indicated the usefulness of the 
selected habitat variables for the prediction of 
impoundments used or not used by nesting Canada Geese. 
After initial testing of the 24 habitat variables by 
discriminant function analysis and stepwise, 18 habitat 
variables that produced similar results to the original 24 
variables were used to compare occupied and vacant 
impoundments. 
RESULTS 
1989 and 1990 Nesting Seasons 
Nesting by Canada Geese within the area began on or 
near 1 April in 1989 and on or near 19 March in 1990. First 
reported hatching of goslings was 30 April in 1989 and 16 
April in 1990, 14 days earlier than the first hatching in 
1989. Latest known hatching dates of goslings were 17 May 
in 1989 and 5 May in 1990. 
Nesting conditions during 1989 were poor because of 
previous and existing weather conditions. Temperatures 
fluctuated from above to below normal during the latter 
parts of March and early April (see Table 3). Precipitation 
falling during 1988 and early 1989 was below normal, and 
caused water levels to fall in most impoundments, thereby 
reducing their surface areas. The distance from a nest 
structure to the nearest point of water averaged seven m 
(min. = 0, max. = 100 m). 
Nesting conditions during 1990 were considerably better 
than 1989. Early warm temperatures during February and 
March, moderate temperature fluctuations, and slightly above 
normal low temperatures may have contributed to improved 
nesting success. No known desertion of nests occurred. In 
addition, normal to above normal rainfall during the summer 
and fall of 1989, and early spring 1990 kept impoundments 
nearly full, to running over. Unlike the spring of 1989, 
20 
when most nest structures were surrounded by dry ground, in 
1990 nearly all nest structures were surrounded by water. 
Nesting Success 
During 1989, 20 impoundments were occupied by at least 
one pair of Canada Geese, compared to 21 impoundments being 
occupied during 1990. Thirty-nine impoundments were vacant 
during 1989 compared to 38 during 1990. Fourteen 
impoundments occupied during 1989 were nested on again 
during 1990 (see Table 4). Twenty occupied and nine vacant 
impoundments were used by the geese for brood rearing during 
1989, compared to 21 occupied impoundments and 17 vacant 
impoundments being used during 1990. 
Usually, only one nesting pair was present at each 
impoundment. However, during 1989 two of the stockponds 
were occupied by two pairs, and one watershed lake was 
inhabited by five pairs of nesting geese. During 1990, 
three stockponds were inhabited by two pairs, and one 
watershed lake was inhabited by four pairs of Canada Geese. 
The four watershed lakes during 1989 and three watershed 
lakes during 1990 were occupied by one or more nesting pairs 
of Canada Geese; one watershed lake was not nested on during 
1990. Waterfowl other than the nesting pair were observed 
on 13 occupied and seven vacant impoundments during 1989, 
and on 15 occupied and 22 vacant impoundments during 1990. 
--
 21 
Table 4. Use of the study impoundments by Canada Geese during the 
1989 and 1990 nesting seasons. 
Impouncinent : 1989 1990
 
Occupied Vacant Occupied Vacant 
BAN-12-1 X X
 
BAN-12-4 X X
 
BAN-12-5 X X
 
BAN-14-1. X X
 
BAN-17-' X X
 
BAN-17-2 X X
 
BAN-17-3 X X
 
BAN-17-4 X X
 
BAN-17-5 X X
 
BAS-5-' X X
 
BAS-5-2 X X
 
BAS-5-3 X
 
BAS-5-4 X X
 
BAS-5-5. X X
 
BAS-6-' X X
 
BAS-6-2 X X
 
BAS-6-3 X X
 
EUN-2-2 X X
 
EUN-3-1 X X
 
EUN-4-1 X X
 
EUN-4-2 X X
 
EUN-5-1 X X
 
EUN-5-2 X X
 
EUN-5-3 X X
 
EUN-5-4 X X
 
EUN-7-1 X X
 
EUN-7-2 X X
 
EUN-7-3 X X
 
EUN-7-4 X X
 
JAIJ-34-,1 X X
 
JAIJ- 34- 2. X X
 
JAIJ-34-3 X X
 
JAIJ-34-4 X X
 
JAIJ-34-5 X X
 
JAIJ-34-6 X X
 
JAIJ-34-7 X X
 
JAIJ-34-8 X X
 
SSE-3-' X X
 
SSE-3-2 X X
 
SSE-3-3 X X
 
SSE-3-4 X X
 
SSE-3-6 X X
 
SSE-l0-l X X
 
SSE-'0-2 X X
 
SSE-'0-3 X X
 
SSE-'0-4 X X
 
SSE-"-' X X
 
SSE-'4-' X X
 
SSE-22-'. X X
 
SSE-23-1 X X
 
SSE-27-1 X X
 
SSE-27-2 X X
 
SSE-33-1 X X
 
SSE-33-2 X X
 
SSE-33-3 X X
 
SSIJ- 28-' X X
 
SSIJ-28-2 X X
 
SSIJ-28-3 X X
 
SSIJ- 29-' X X
 
• 
- lJatershed lakes. 
- Nest structure rusted out and fell apart after the 1989 season. 
22 
During 1989, 11 (42%) of 26 pairs of Canada Geese, 
produced broods; 10 broods were fledged from artificial 
structures, and one brood was produced from a ground nest. 
A successful nest was identified as a nest from which at 
least one or more goslings was produced. Of 15 unsuccessful 
nests (58%), 11 were deserted for unknown reasons; two nests 
were deserted after human disturbance, and two nests were 
destroyed by predators. Initial brood size for the 11 
broods averaged 4.5 goslings (min. = 1, max. = 6). 
During 1990, 27 pairs of Canada Geese produced broods; 
26 broods were incubated on artificial nest structures, and 
one brood was produced from a ground nest. Initial brood 
size for 11 of the 27 broods averaged 5.1 goslings 
(min. = 1, max. = 7). 
Artificial Nest structure Use 
During 1989, 24 pairs of Canada Geese nested on 
artificial structures, compared to 26 pairs nesting on 
artificial structures during 1990. Nest structures were 
located on 35 impoundments during 1989 and on 34 
impoundments during 1990. Two watershed lakes and two 
stockponds each had at least two nest structures on them; 
one watershed lake had seven nest structures on it, and one 
watershed lake did not have any nest structures on it. The 
three types of nest structures commonly used in the 
area were floating nest structure, barrel-type tub nest 
structure, and tire-type tub nest structure. 
23 
Numerous artificial structures within the area were not 
properly maintained during both years of the study. For 
example, the bottoms on several of the barrel type nest 
structures were completely rusted through. The supporting 
poles of some nest structures were leaning and the tops of 
the nest structures were almost touching the surface of the 
water. One floating structure that was used in 1989 was 
either damaged or destroyed by weather and unusable in 1990. 
The bottom of one barrel-type nest structure used during 
1989 had completely rusted out, making it unusable in 1990. 
Also, the nesting material in some nest structures was not 
replaced during both years of the study. 
Habitat Characteristics 
Fifty-nine impoundments (n = 55 stockponds, n = 4 
watershed lakes) were identified and studied during 1989 and 
1990. During 1989, the average surface area of 16 occupied 
stockponds was 0.439 ha, compared to an average of 0.323 ha 
for 39 vacant stockponds. The average surface area of four 
occupied watershed lakes was 4.063 ha (min. = 0.615, 
max. = 5.83 ha): there were no vacant watershed lakes. 
However, the surface area of one of the watershed lakes was 
reduced because of drought conditions (see Table 5). The 
percent of the shoreline exposed on 59 impoundments averaged 
51% (min. = I, max. = 87%). 
24 
Table 5. Impoundment surface areas during 1989 and 1990. 
Impouncinent: 1989 (ha) 1990 (ha) 1990 DL' 
BAH-12-1 0.32 1.29 1.20 
BAH-12-4 0.04 0.19 1.03 
BAH-12-5 0.12 0.19 1.03 
BAH-14-1. 0.40 0.91 1. 19 
BAH-17-1 4.92 7.44 1.58 
BAH-17-2 0.26 0.26 1. 11 
BAH-17-3 0.19 0.77 1.29 
BAH-17-4 0.06 0.39 1.45 
BAH-17-5 0.08 0.19 1.00 
BAS-5-1 0.32 0.52 1.26 
BAS-5-2 0.89 1.10 2.38 
BAS-5-3 0.39 0.39 2.55 
BAS-5-4 1.23 1.23 1.64 
BAS-5-5. 2.01 2.01 1.28 
BAS-6-1 5.83 5.83 2.63 
BAS-6-2 0.52 0.52 1.89 
BAS-6-3 0.48 0.78 1.54 
EUH-2-2 0.52 1.15 1.39 
EUH-3-1 0.72 0.72 1.00 
EUH-4-1 0.75 1.15 1.42 
EUH-4-2 0.14 0.16 1.34 
EUH-5-1 0.10 0.58 1.34 
EUH-5-2 0.24 0.72 1.40 
EUH-5-3 0.94 0.94 1.05 
EUH-5-4 1.09 1.30 1.04 
EUH-7-1 0.45 1.15 1.42 
EUH-7-2 0.12 0.61 1.29 
EUH-7-3 0.87 1.22 1.07 
EUH-7-4 0.03 0.08 1.99 
JAW-34-1 0.07 0.20 1.26 
JAW-34-2. 0.16 0.43 1.03 
JAW-34-3 0.61 3.17 1.61 
JAW-34-4 0.06 0_29 1.00 
JAW-34-5 0.16 0.58 1.11 
JAW-34-6 0.06 0.07 1.00 
JAW-34-7 0.08 0.29 1.00 
JAW-34-8 0.04 0.15 1.00 
SSE-3-1 0.06 0.29 1.26 
SSE-3-2 0.12 0.43 1.03 
SSE-3-3 0.91 1.87 1.00 
SSE-3-4 0.16 0.43 1.03 
SSE-3-6 0.01 0.22 1.78 
SSE-10-1 0.07 0.10 1.89 
SSE-10-2 0.07 0.10 1.00 
SSE-10-3 0.58 0.58 2.01 
SSE-10-4 0.81 1.15 1.26 
SSE-11-1 0.07 0.07 2.52 
SSE-14-1 0.24 0.58 1.34 
SSE-22-1. 0.04 0.07 1.26 
SSE-23-1 4.89 5.76 1.69 
SSE-27-1 0.48 0.72 1.00 
SSE-27-2 0.12 0.43 1.29 
SSE-33-1 0.26 0.30 1.60 
SSE-33-2 0.30 0.72 1.00 
SSE-33-3 0.06 0.07 1.26 
SSW-28-1 0.14 0.43 1.03 
SSW-28-2 1.44 1.44 1.13 
SSW-28-3 0.79 1.87 1.23 
SSW-29-1 0.36 1.30 1.19 
• 
- Watershed lakes. 
- Values for the 1989 were not determined because of 
drought conditions. 
25 
During 1990, the average surface area of 18 occupied 
stockponds was 0.963 ha, compared to an average of 0.554 ha 
for 37 vacant stockponds. The average surface area of three 
watershed lakes was 5.46 ha (min. = 3.17, max. = 7.44 ha), 
compared to 5.83 ha for one vacant watershed lake 
(see Table 5). The percent of the shoreline exposed on 59 
impoundments averaged 2.5% (min. = 0, max. = 74%). 
Forty-nine impoundments (83%) were located in gently 
rolling rangeland during 1989 and 1990. Seven impoundments 
were located in steep-sloped rangeland; three impoundments 
were located in flat to slightly rolling rangeland. 
Eighteen occupied impoundments (90%) and 37 vacant 
impoundments (95%) were located in rangeland. Of the 59 
impoundments, 29 were located near fields planted to grain 
sorghum; 14 of those impoundments were occupied and 15 were 
vacant. 
All impoundments were estimated to be approximately 10 
years of age, or older. In addition, some of the 
impoundments were thought to be 40 - 60 years old (Charles 
Wiggins, pers. comm.). This variable was removed from 
analysis by SAS (SAS Institute 1979) because it was a 
"constant" variable. No natural or man made islands were 
found on the study impoundments. This variable was also 
removed from analysis by SAS (SAS Institute 1979). The 
results of other habitat variables that were not summarized 
in the results section were listed in Table 6. 
26 
Table 6. Results of habitat variables evaluated during the study (see Table 2 for definition of 
the variabLes). 
1989 1990 
VariabLe: Occupied Vacant min. max. Occupied Vacant min. max. 
PSHLEV 42%1 61%1 1% 100% 63% 52% 1% 100% 
PSHLSV 3% 3% 0% 40% 2% 4% 0% 40% 
PSHLWV 28% 14% 0% 88% 22% 17% 0% 88% 
NIW1.6 km 4.3 4.1 0 8 4.3 4.0 0 8 
DISN31 0.6 km 0.5 0.2 1.9 0.6 0.5 0.2 1.9 
* *IMSIZE (stockponds) 0.44 ha 0.32 0.03 1.23 0.96 0.55 0.07 2.22 
IMSIZE (wished Lakes) 4.1 ha --- 0.61 5.83 5.5 5.6 3.2 7.44 
IDSRS 2.0 km 2.4 0 6.4 2.2 2.3 0 > 6.4 
DISNF 0.5 km 0.7 0.1 2.5 0.6 0.7 0.1 2.5 
DISNCC 0.9 km 1.2 0.1 3.3 1.2 1.1 0.1 2.1 
IMFENCE 3 0 --- --- 2 
SHII --- --- --- --- 1.33 1.38 1.00 2.63 
IMBROOD 20* 9 - -- - - - 21 * 17 
ISPR 0 0 --- --- 0 0 
PRW 13 7 -- - --- 15 22 
PRAN 19 16 1 7 20 14 1 7 
- Aquatic vegetation was measured at 13 occupied and 16 vacant impoundments during 1989. 
* 
. Indicated significant differences (~ < 0.05) between occupied and vacant impoundments. 
27 
Analysis of Habitat 
During 1989, the average occupied impoundment size 
(0.439 ha, n = 16) was not significantly different from the 
average vacant impoundment size (0.323 ha, n = 39) when the 
four watershed lakes were removed from the analysis 
(~ = -1.198, df = 53.0, £ > 0.05, see Table 7). However, 
the average occupied impoundment size (1.16 ha, n = 20) was 
significantly different from the average vacant impoundment 
size (0.323 ha, n = 39) when the four watershed lakes were 
included in the analysis (t = -2.098, df = 19.7, £ < 0.05). 
The percent of the shoreline having aquatic macrophytes 
along vacant impoundments (61%, n = 16) was significantly 
different from occupied impoundments (42%, n = 13) when the 
watershed lakes were removed from the analysis (t = 2.221, 
df = 23.0, £ > 0.05). In addition, brood use of occupied 
impoundments was significantly different from brood use of 
vacant impoundments when watershed lakes were included 
(~ = -4.681, df = 53.0, £ > 0.05), and removed (t = -8.447, 
df = 57.0, £ > 0.05) from the analysis. 
During 1990, the average occupied impoundment size 
(0.964 ha, n = 18) was significantly different from the 
average vacant impoundment size (0.554 ha, n = 37) when the 
four watershed lakes were removed from the analysis 
(~ = -2.600, df = 27.0, £ > 0.05, see Table 8). In 
addition, the average occupied impoundment size (1.60 ha, 
n = 21) was significantly different from the average vacant 
28 
Table 7. Result~ of 1989 habitat variables 
using i-tests (see Table 2 for definition 
of variables). 
Variable: n OF T Prob > ITI 
Watershed lakes included 
IMSIZE 59 19.7 -2.0972 0.0491 
PSHLEV 29 27.0 1.9453 0.0622 
IMBROOO 59 57.0 -8.447 0.0000 
Watershed lakes removed 
IMSIZE 55 53.0 -1.198 0.2361 
PSHLEV 25 23.0 2.226 0.0365 
IMBROOO 55 53.0 -4.681 0.0000 
* 
- Variables reported based on significant 
differences of unequal (E < 0.05) or 
equal (E > 0.05) variances. 
Table 8. Result~ of 1990 habitat variables 
using i-tests (see Table 2 for definition 
of variables). 
Variable: n OF T Prob> IT! 
Watershed lakes included 
IMSIZE 59 26.3 -2.125 0.0431 
IMBROOD 59 54.6 -5.044 0.0001 
Watershed lakes removed 
differences of unequal (E 
IMSIZE 55 53.0 -2.852 0.0062 
IMBROOO 55 53.0 -4.853 0.0001 
* 
- Variables reported based on significant 
< 0.05) or 
equal (E > 0.05) variances. 
29 
impoundment size (0.693 ha, n = 38) when watershed lakes 
were included in the analysis (t = -2.125, df = 26.3, 
~ < 0.05). Brood use of occupied impoundments was 
significantly different from brood use of vacant 
impoundments when watershed lakes were included (t = -4.853, 
df = 53.0, E < 0.05) and removed (t = -5.044, df = 54.6, 
~ < 0.05) from the analysis. 
Analysis of 18 of the habitat variables by stepwise 
discriminant function indicated that brood use of an 
impoundment, presence of an artificial nest structure, 
impoundment fenced to exclude livestock, and distance to the 
nearest three vacant impoundments, were correlated with 
impoundment use by nesting pairs of Canada Geese 
(B2 = 0.486, Wilks' Lambda = 0.514, E > 0.0001, see Table 
9). This analysis showed that 49% of the variation between 
occupied and vacant impoundments was best explained by these 
four variables. However, a significant portion of the 
variation (51%) between impoundment classes was not 
explained. Brood use of an impoundment and presence of an 
artificial nest structure explained 39% of the variation 
between impoundment classes. The remaining two variables 
explained an additional 10% of the variation. 
During 1990, the presence of an artificial nest 
structure, brood use of an impoundment, impoundment size, 
and percent of shoreline with shrub vegetation along it, 
were correlated with impoundment use by nesting pairs of 
30 
Table 9. The contribution of 18 variables used to distinguish between impoundment classes during 
the 1989 nesting season using stepwise discriminant function analysis (see Table 2 for 
definition of variables). 
Variable: Partial R2 F - Statistic Prob > F IJilks' lambda Prob < lambda ASCC 1 Prob > ASCC 
IMBROOD 0.34 28.444 0.0001 0.663 0.0001 0.337 0.0001 
PRAN 0.08 4.904 0.0309 0.609 0.0001 0.391 0.0001 
IMFENCE 0.12 6.545 0.0134 0.543 0.0001 0.457 0.0001 
DISN31 0.05 2.970 0.0907 0.514 0.0001 0.486 0.0001 
CUCT 0.04 2.104 0.1529 a2 a a a 
SHII 0.02 1.173 0.2839 a a a a 
DISNF 0.02 1.083 0.3029 a a a a 
IMTYPE 0.02 0.986 0.3253 a a a a 
SSl 0.01 0.793 0.3773 a a a a 
PSHlW 0.01 0.785 0.3796 a a a a 
PSHlEX 0.01 0.684 0.4119 a a a a 
PSHlSV 0.01 0.571 0.4534 a a a a 
IMSIZE 0.01 0.522 0.4732 a a a a 
PRIJ 0.01 0.390 0.5351 a a a a 
NIIJ1.6KM 0.00 0.264 0.6094 a a a a 
DISNCC 0.00 0.225 0.6372 a a a a 
lAUA 0.00 0.192 0.6633 a a a a 
IDSRS 0.00 0.048 0.8277 a a a a 
1 
- Average Squared Canonical Correlation. 
a2 - Values were not calculated by stepwise because they were not entered into 
the equation. 
31 
Canada Geese (B2 = 0.407, wilks' Lambda = 0.592, E > 0.0001, 
see Table 10). This analysis showed that approximately 41% 
of the variation between occupied and vacant impoundments 
was best explained by these four variables. However, a 
significant portion of the variation (59%) between 
impoundment classes was not explained. Presence of an 
artificial structure, and brood use of an impoundment 
explained approximately 34% of the variation between 
impoundment classes. The remaining two variables explained 
an additional seven percent of the variation. 
During 1989, analysis of the 18 habitat variables by 
discriminant function correctly classified eight (13.8%) 
occupied impoundments (see Table 11). Twelve occupied 
impoundments were misclassified as vacant impoundments. 
Thirty-eight vacant impoundments (97.4%) were correctly 
classified. One vacant impoundment was removed from 
analysis by SAS (SAS Institute 1979) because of missing 
data. Discriminant function analysis was able to correctly 
classify occupied and vacant impoundments an average of 80% 
(n = 46) of the time. During 1990, discriminant function 
analysis correctly classified 21 occupied impoundments 
(100%) and 38 vacant impoundments (100%, see Table 11). 
Several of the habitat variables were significantly 
correlated with each other (~ < 0.05, see Tables 12 and 13). 
Impoundment size was correlated with the percent of the 
shoreline supporting trees during 1989 (K = 0.502) and 
32 
Table 10. The contribution of 18 variables used to distinguish between impoundment classes during 
the 1990 nesting season using stepwise discriminant function analysis (see Table 2 for 
definition of variables>, 
Variable: Partial R2 F - Statistic Prob > F Wi lks' Larrbda Prob < Larrbda ASCC 1 Prob > ASCC 
PRAN 0.24 18.418 0.0001 0.756 0.0001 0.244 0.0001 
IMBROOO 0.13 8.792 0.0044 0.653 0.0001 0.347 0.0001 
IMSIZE 0.05 3.240 0.0773 0.617 0.0001 0.383 0.0001 
PSHLSV 0.04 2.235 0.1407 0.592 0.0001 0.408 0.0001 
PSHLEX 0.02 1.312 0.2571 a2 a a a 
PRW 0.02 1.196 0.2790 a a a a 
SSL 0.02 0.936 0.3377 a a a a 
DISN31 0.01 0.459 0.5010 a a a a 
DISNCC 0.01 0.379 0.5409 a a a a 
NIW1.6KM 0.00 0.268 0.6071 a a a a 
CUCT 0.00 0.141 0.7091 a a a a 
IMFENCE 0.00 0.073 0.7875 a a a a 
DISNF 0.00 0.055 0.8148 a a a a 
SHII 0.00 0.050 0.8231 a a a a 
IDSRS 0.00 0.033 0.8563 a a a a 
IMTYPE 0.00 0.015 0.9044 a a a a 
PSHLWV 0.00 0.004 0.9524 a a a a 
LAUA 0.00 0.000 0.9986 a a a a 
1 
- Average Squared Canonical Correlation. 
a2 - Values were not calculated by stepwise because they were not entered into 
the equation. 
33 
Table 11. Presence or absence of Canada Geese on impoundments 
as predicted by discriminant function analysis, and based 
on 18 habitat variables during the 1989 and 1990 
nesting seasons. 
Vacant Occupied 
1989 1990 1989 1990 
Vacant: 38* 38 o o
 
Occupied: 12 o 8 21
 
Total: 50 (86.2%) 38 (100%) 8 (13.8%) 21 (100%) 
Prior Prob.: 0.5000 0.5000 
* 
- One vacant impoundment ommited from analysis by 
SAS (SAS Institute 1979). 
34 
Table 12. Results of 1989 Pearson product correlation coefficients for 18 habitat variables (see Table 2 for definition of variables). 
Variable: DISNF DISNCC IMSIZE PSHLEX PSHLIN NIW1.6Icm SHII IMTYPE LAUA PRAN SSL CCUT IMFENCE PRW 0ISN31 IDSRS PSHLSV IMBROOO 
DISNF 1.000 0.426 -0.059 0.097 -0.282 -0.215 -0.278 -0.093 -0.277 0.283 0.503 -0.028 -0.157 -0.258 0.068 0.007 -0.276 -0.227 
DISNCC 0.426 1.000 0.105 -0.087 0.179 -0.006 0.170 -0.061 -0.264 0.180 0.402 -0.009 0.129 -0.247 0.182 -0.060 0.170 -0.104 
IMSIZE -0.059 0.105 1.000 0.017 0.502* -0.093 *0.282 *0.737 0.139 -0.010 0.042 -0.003 0.617 0.082 0.084 -0.121 0.243 0.084 
PSHLEX 0.097 -0.087 0.017 1.000 -0.107 -0.031 - O. 148 O. 000 -0.059 0.049 -0.049 -0.102 -0.223 0.008-0.028 -0.175 -0.105 0.009 
PSHLWV -0.282 0.179 0.502 -0.107 1.000 0.229 *0.385 *0.450 0.041 0.017 0.002 0.176 0.456 0.107-0.091 -0.127 0.720 0.166 
NIW1.6Icm -0.215 -0.006 -0.093 -0.031 0.229 1.000 0.082 -0.240 -0.099 -0.073 -0.026 0.119 0.104 0.140-0.313 -0.401 0.302 0.326 
SHII -0.278 0.170 0.282 -0.148 0.385 0.082 1.000 0.140 -0.025 -0.209 0.094 0.062 0.478 0.045 0.001 -0.010 0.507 0.234 
IMTYPE -0.093 -0.061 0.737 0.000 0.450 -0.240 0.140 1.000 0.349 0.043 -0.240 -0.069 0.361 0.042 0.100 -0.070 0.195 0.022 
LAUA -0.277 -0.264 0.139 -0.059 0.041 -0.099 -0.025 0.349 1.000 -0.171 -0.388 -0.135 -0.054 0.104 0.511 0.081 0.000 -0.060 
PRAN 0.283 0.180 -0.010 0.049 0.017 -0.073 -0.209 0.043 -0.171 1.000 0.041 0.000 -0.037 0.063-0.141 0.058 -0.049 0.169 
SSL 0.503 0.402 0.042 -0.049 0.002 -0.026 0.094 -0.240 -0.388 0.041 1.000 0.025 0.230 0.155-0.372 0.983 -0.125 -0.102 
CCUT -0.028 -0.009 -0.003 -0.102 0.176 0.119 0.062 -0.069 -0.035 0.000 0.025 1.000 0.093 -0.093-0.118 -0.081 -0.156 0.273 
IMFENCE -0.157 0.129 0.617* -0.023 0.456* 0.104 *0.478 *0.361 -0.054 -0.037 0.230 0.093 1.000 0.105-0.102 -0.199 0.560 0.072 
PRW '0.258 -0.247 0.082 0.008 0.107 0.140 0.045 0.042 0.104 0.063 0.155 -0.093 0.105 1.000 0.029 0.278 0.170 -0.072 
DISN31 0.068 0.182 0.084 -0.028 -0.091 *-0.313 0.000 0.100 0.511 * -0.141 -0.107 -0.118 -0.102 0.029 1.000 0.274 0.130 -0.107 
IDSRS 0.007 -0.061 -0.121 -0.175 '0.127 *-0.401 -0.010 -0.070 0.081 0.058 0.983 -0.081 -0.199 0.278 0.274 1.000 0.105 -0.493 
PSHLSV -0.276 0.170 0.243 -0.105 0.720 0.30t 0.507* 0.196 0.000 -0.049 -0.125 -0.156 0.560 0.170 0.130 0.105 1.000 0.067 
IMBROOO -0.227 -0.104 0.084 0.009 0.166 0.326* 0.234 0.022 -0.060 0.169 0.000 0.273 0.072 -0.072-0.107 -0.493 0.067 1.000 
* 
- Indicates significant correlation between habitat variables (~ < 0.05). 
35 
Table 13. Results of 1990 Pearson product correlation coefficients for 18 habitat variables (see Table 2 for definition of variables). 
Variable: DISNF DISNCC IMSIZE PSHLEX PSHLWV NIW1.6km SHII IMTYPE LAUA PRAN SSL CCUT IMFENCE PRW DISN31 IDSRS PSHLSV IMBROOD 
DISNF 1.000 0.409 0.089 0.211 -0.259 -0.225 -0.256 -0.106 -0.272 0.361 0.484 -0.031 -0.108 -0.286 0.088 -0.013 -0.286 -0.217 
DISNCC 0.409 1.000 0.167 -0.087 0.217 0.008 0.219 0.009 -0.248 0.129 0.397 0.008 0.177 -0.130 0.205 -0.076 0.181 -0.462 
IMSIZE -0.089 0.167 1.000 *-0.264 *0.527 -0.128 *0.390 *0.810 0.149 -0.162 0.055 -0.063 0.653 0.210 0.050 -0.030 0.278 0.027 
PSHLEX 0.211 -0.087 -0.264 1.000 *-0.334 -0.084 0.506* -0.037 0.119 0.273 -0.012 -0.101 -0.181 -0.065 0.092 -0.105 -0.390 0.029 
PSHLWV -0.259 0.217 0.527 -0.334 1.000 0.208 *0.416 *0.462 0.013 -0.165 0.018 0.148 0.419 0.196 -0.135 -0.110 0.720 0.012 
NIW1.6km -0.225 0.008 -0.128 -0.084 0.208 1.000 0.145 -0.051 -0.104 -0.106 -0.015 0.106 0.079 -0.052 -0.349 -0.40 0.303 0.119 
SHII -0.256 0.219 0.390 -0.506 0.416 0.145 1.000 *0.325 -0.028 -0.378 0.145 0.118 0.479 0.165 0.017 0.025 0.583 -0.015 
IMTYPE -0.106 0.009 0.810 -0.037 *0.462 -0.051 0.325 1.000 0.174 -0.100 0.000 -0.169 0.490 0.268 0.001 -0.122 0.242 0.072 
LAUA -0.272 -0.248 0.149 -0.119 0.013 -0.104 -0.028 0.174 1.000 -0.108 -0.371 -0.141 -0.098 0.215 0.499 0.092 0.005 -0.063 
PRAN 0.361 0.129 -0.162 0.273 -0.165 -0.106 -0.378* -0.100 -0.108 1.000 0.020 -0.145 -0.318 -0.034 -0.005 -0.002 -0.222 0.181 
SSL *0.484 *0.397 0.055 -0.012 0.018 -0.015 0.145 0.000 *-0.371 0.020 1.000 0.021 0.202 0.015 0.033 -0.100 -0.013 -0.121 
CCUT -0.031 0.008 -0.063 -0.101 0.148 0.106 0.118 -0.169 -0.141 -0.145 0.021 1.000 0.000 -0.098 -0.137 0.014 0.199 -0.139 
IMFENCE -0.108 0.177 *0.653 -0.181 *0.419 0.079 *0.479 *0.490 -0.098 -0.318 0_202 0.000 1.000 0.225 -0.085 -0.054 0.256 0.060 
PRW -0.286* -0.130 0.210 -0.065 0.196 -0.052 0.165 *0.268 0.215 -0.034 0.015 -0.098 0.225 1.000 0.103 -0.104 0.105 0.279 
DISN31 0.088 0.205 0.050 0.092 -0.135 *-0.349 0.017 0.001 *0.499 -0.005 0.033 -0.137 -0.085 0.103 1.000 0.087 -0.165 -0.218 
lDSRS -0.013 -0.076 -0.030 -0.105 -0.110 *-0.405 -0.025 -0.122 0.092 -0.002 -0.100 0.014 -0.054 -0.104 0.087 1.000 -0.107 -0.105 
PSHLSV -0.286 0.182 *0.278 *-0.390 *0.720 0.303 0.583* 0.242 0.005 -0.222 -0.013 0.199 0.256 0.105 -0.165 -0.107 1.000 -0.132 
IMBROOD -0.217 -0.462 0.027 0.029 0.012 0.119 -0.015 0.072 -0.063 0.181 -0.121 -0.139 0.060 0.279 -0.218 -0.105 -0.132 1.000 
* 
- Indicates significant correlation between habitat variables (f < 0.05). 
36 
1990 (~= 0.527). Impoundment type was correlated with the 
percent of the shoreline supporting trees during 1989 (~ = 
0.450) and 1990 (~= 0.462). Brood use of an impoundment 
was correlated with the number of impoundments within 1.6­
 krn radius of each study impoundment during 1989 (~= 0.326), 
however, it was not significantly correlated with any other 
variable during 1989 or 1990. 
DISCUSSION 
Nest site Selection 
In my study, the majority of Canada Geese nested on 
artificial structures. I believed that the selection of the 
nest site included the selection of; 1) a particular 
impoundment or area as a nest site, and 2) the nest site 
type based on previous nesting success or imprinting as 
goslings on the nest structure or nest site. 
Canada Geese usually returned to the same area to nest 
year after year (Brakhage 1965), however, there is more of a 
tendency for females, than males, to return to their natal 
home when they begin nesting (Johnsgard 1978). MacInnes 
(1962) suggested that pairs returned to nest in the same 
area but that they may not use the same nest. If a pair 
nested on the impoundment the previous year and were 
successful at producing a brood then they would probably 
return and nest at this same impoundment, but they may not 
use the same nest site. Several of the impoundments within 
the study area had more than one nest structure on them; 
however, usually the same nest structure was used each year. 
Studies by MacInnes (1962) and Cooper (1978) showed that 
Giant Canada Geese pairs usually did not nest in the same 
site year after year. Hanson (1965) believed that it was 
the rule rather than the exception for ~. ~. interior to 
return to the same nesting site. 
38 
Canada Geese successful at producing a brood on an 
artificial structure one year may be more likely to nest on 
a structure the following year. Cooper (1978) reported that 
73% of the Giant Canada Geese at Marshy Point nested on the 
same type of structure each year; also older, used, 
structures were used more often than newer structures. 
observed that most of the artificial structures used were of 
the tub type. Also, Canada Geese were observed more often 
on artificial structures used the previous year compared to 
structures not used the previous year. I also believed that 
older nest structures, or those used the previous year(s) , 
were used by early nesting Canada Geese. Nest structures not 
selected by the early nesting Canada Geese were used more by 
younger pairs. 
Studies by Hanson (1965) and Stoudt (1971) showed that 
Canada Geese and other waterfowl species had a tendency to 
nest on or near the same body of water where they learned to 
fly. I believed that this was part of the imprinting 
process that goslings experience. If the nest site that the 
goslings had imprinted on was an artificial structure then 
they would more likely return and nest on an artificial 
structure. Conversely, had the goslings imprinted on a 
ground nest site then they would be more likely to return 
and nest on the ground. For example, during 1989 I observed 
a pair of Canada Geese nesting on the ground on an 
impoundment outside the study area (located along K-99 
I 
39 
Highway). Again, in 1990 I observed a pair nesting on the 
ground at this same impoundment, but the nest site was 
different. It was possible that I observed the same pair 
nesting both years and that previous nesting success and 
imprinting influenced the selection of the nest site. 
Cooper (1978) concluded that nesting success apparently 
influenced the distance from the old nest site to the new 
nest site. He also found that the average distance between 
nests, nest concealment, distance to and height above open 
water, and the visibility of the surrounding habitat 
strongly influenced the selection of the nest site by the 
female Giant Canada Goose. Bellrose (1976) concluded that 
prerequisites to the selection of the nest site included 
cover for the nest and an exposed view for the incubating 
bird. 
Nesting Success 
The lower nesting success rate (42%) that I found in 
1989, in comparison to 1990, was caused by desertion and 
predation. Two nests were deserted after disturbance by 
humans, while 11 nests were deserted for unknown reasons, 
and two nests were destroyed by predators. However, I 
assumed that desertion was influenced by several factors: 1) 
periods of above normal high and below normal low 
temperatures combined with unusually dry weather prior to 
and during the egg-laying and incubation periods, 2) 
disturbance by humans or livestOCk, or 3) an observed lack 
40 
of suitable nesting material in the nest structure. It was 
possible that the lack of nest material in some of the 
artificial structures increased the chances of freezing or 
overheating of the eggs during 1989. Inspection of a nest 
deserted during 1989 revealed four unhatched eggs surrounded 
only by the down from the female Canada Goose. Conversely, 
during 1990 absent nesting material did not appear to 
influence nesting success. It was possible that improved 
weather conditions contributed to the higher nesting 
success. The aerial survey flown 13 April 1989 revealed 
numerous artificial nest structures located in Butler and 
Greenwood Counties that did not have nest material in them. 
The results of my study were comparable to that of other 
researchers like Szymczak (1975) who reported nesting 
success rates ranging from 0 to 100% for Giant Canada Geese 
in the Fort Collins-Loveland, Colorado area. Cooper (1978) 
reported nesting success rates varying from 65 to 82% for 
Giant Canada Geese at Marshy Point, Manitoba, from 1969 
through 1971. stiefel (1980) studied nesting Giant Canada 
Geese in western South Dakota and he indicated a success 
rate of 78%. 
Despite the fact that only two ground nests were 
observed in 1989 and one in 1990 management of the 
vegetative cover along impoundments was important to the 
success of ground nesting Canada Geese. In both 1989 and 
1990 ground nests located in dense vegetation were 
41 
successful, while the ground nest located in short 
vegetation in 1989 was unsuccessful. Poly (1979) found that 
dense vegetation near the nest resulted in a low desertion 
rate. Cooper (1978) concluded that the female Giant Canada 
Goose selected her nest site where vegetation would screen 
the nest. 
Artificial Nest structure Use 
The majority of Canada Geese during 1989 and 1990 
nested on artificial structures. Mackey et ale (1988) 
reported a nesting success rate of 85%, from 1983 through 
1987, for Canada Geese nesting in Montana using artificial 
structures located in trees. He also found that the nesting 
population increased when artificial structures were 
provided for them. 
During 1989, 13 pairs of Canada Geese that nested on 
artificial structures deserted their nests. Cooper (1978) 
reported that nests on artificial structures, at Marshy 
Point, were deserted more often than nests located on 
natural sites. Perhaps artificial structures placed in open 
areas were subjected to more disturbance than nests or 
structures that were well concealed. However, Cooper (1978) 
also believed that the high nesting success rate at Marshy 
Point was due to the nest sites being located such that 
disturbance, predation, and flooding were reduced. 
Most nest sites were located on artificial structures, 
either in open water or on exposed mudflats during both 
42 
years of study, therefore, vegetative cover near the nest 
seemed of little importance. Cooper (1978) concluded that 
the female Giant Canada Goose selected a nest site with high 
visibility and with little or no concealment only when 
surrounded by water or surrounded by a haystack. Hamilton 
(1978) reported nesting success rates of 65% for nests 
located in exposed sites compared to 35% for nests located 
in concealed sites. In my study, success rates for exposed 
sites (artificial structures), were 42% in 1989 and 100% in 
1990 compared to 50% in 1989 and 100% in 1990 for concealed 
sites (ground nests). 
Many of the artificial structures not used by Canada 
Geese in the area were unsuitable as nest sites. Likewise, 
some of those used in 1989 were no longer suitable nest 
sites in 1990. Cooper (1978) suggested that the physical 
condition of a nest structure influenced its use by nesting 
Giant Canada Geese; structures lacking stability 
were often rejected. This may be one of the reasons why 
numerous nest structures were unused. 
Brood Use of Impoundments 
Brood use of occupied and vacant impoundments was 
significantly different during both 1989 and 1990. Broods 
were observed more often on previously occupied impoundments 
than vacant impoundments, and some impoundments were 
frequented more than others. Watershed lakes, for various 
reasons (i.e., larger surface area, more food or cover), 
43 
usually attracted the most broods. In addition, brood use 
of an impoundment and the number of impoundments within 1.6 
km were correlated with each other (~= 0.326). Broods were 
more often associated with impoundments in close proximity 
to other impoundments. This observation supported results 
reported by Rumble and Flake (1983) that indicated that the 
number of natural pond basins within a 1.6 km radius of a 
study pond was positively associated with use by duck 
broods. Stiefel (1980) found that several Giant Canada 
Goose pairs and their broods usually used a common 
stockpond, and that overland travel of 1.6 to 4.8 km 
occurred within three to seven days after hatching. 
Hamilton (1978) reported that brood movement was minimal 
under optimal habitat conditions, however, brood movement 
became more extensive when conditions were poor. In my 
study brood travel was not negatively influenced by the 
drought conditions (1989) because brood travel occurred 
during improved habitat conditions (1990) as well. I 
concluded that this was the case for adults and broods 
located in areas where watershed lakes were absent. 
However, broods that were hatched on watershed lakes had a 
tendency to stay there all summer. 
Brood use of impoundments was dependent on numerous 
influential variables of which food and cover were probably 
the most important. Also, I believed that brood use was 
limited for impoundments whose surface areas were reduced. 
44 
Numerous impoundments that had no shrub or tree cover along 
the shoreline were unused during both years of the study. 
Habitat Characteristics 
The smaller impoundment surface areas of 1989 were 
caused by the drought conditions that were prevalent during 
1988 and early 1989. Impoundment surface areas were 
approximately one-half of their normal size. In addition, 
occupied and vacant impoundment sizes were smaller than the 
minimum impoundment size of 0.61 ha recommended by Lokemoen 
(1973) to maximize waterfowl use of an impoundment in the 
Northern High Plains. However, during 1990 conditions were 
different because above normal precipitation during the 
summer and fall of 1989 filled impoundments to within two 
percent of their normal surface areas. The average 
impoundment surface area reported in my study was also 
smaller than the reported mean size of 1.6 ha for stockponds 
used by Giant Canada Geese, and 0.6 ha for stockponds not 
used by Giant Canada Geese in South Dakota (Stiefel 1980). 
Although, occupied impoundment size was significantly 
larger than vacant impoundment size during 1990, I do not 
think that impoundment size alone was an indicator of 
impoundment use by Canada Geese, because Canada Geese nested 
on impoundments ranging in size from less than 0.5 to 7.44 
ha during both years of the study. 
45 
Several of the habitat variables were significantly 
correlated with each other. Impoundment size was correlated 
with percent of shoreline supporting trees during 1989 and 
1990 (see Tables 12 and 13). As impoundment size increased 
the presence of trees along the shoreline increased. The 
lack of difference between the 1989 and 1990 coefficient 
values was probably due to the lack of change in the percent 
of shoreline having trees. Impoundment type was also 
correlated with percent of shoreline supporting trees during 
1989 and 1990. An example of this was watershed lakes that 
were fenced to exclude livestock. Livestock had a tendency 
to graze or trample vegetation along the shoreline of 
impoundments, thus limiting tree or shrub growth. 
Because numerous broods were observed on watershed 
lakes I concluded that trees and shrubs along the shoreline 
of watershed lakes helped hide the young goslings from 
predators. Conversely, Rumble and Flake (1983) found that 
the occurrence of trees along the shoreline or an 
impoundment was associated with reduced use of the 
impoundment by duck broods. Also, in my study I found that 
adult Canada Geese, at impoundments where trees or shrubs 
were limited or absent, often took their broods to the 
middle of an impoundment when disturbed. 
Impoundment size was correlated with shoreline 
development during 1989 and 1990 (see Tables 12 and 13), 
yet, because of the drought, an index of approximately 1.00 
46 
would have more accurately reflected the shoreline 
development of most impoundments during 1989. In comparison 
to most stockponds, watershed lakes had a more irregular 
shoreline, providing more food and cover for adult geese and 
broods. Mack and Flake (1980) reported that shoreline 
length and surface water area were both indicative of 
impoundment size, however, shoreline length predicted brood 
occurrence better than actual impoundment size. This may 
have been the reason some adult geese took their broods to 
the watershed lakes soon after hatching. 
Analysis of Habitat 
Analysis of the 18 habitat variables by stepwise 
discriminant function produced somewhat different results 
between years. The two variables that explained the most 
variation between use of occupied and vacant impoundments by 
Canada Geese during both nesting seasons were brood use of 
an impoundment and presence of an artificial structure. 
other important variables included impoundment fenced to 
exclude livestock and the average distance to the nearest 3 
vacant impoundments during 1989 or impoundment size and 
percent of shoreline supporting shrubs during 1990. 
believed that the features influencing Canada Geese to 
select an impoundment as a nest site included: l)the 
impoundment area provided food and cover for the nesting 
pair and brood, and 2) an artificial structure (in good 
condition) was present. 
...
 1

 I 
47 
The drought-reduced impoundment surface areas forced 
the adult geese and broods to move to impoundments that 
provided more food and cover. During 1990 when impoundments 
were full, the variables impoundment size and percent of 
shoreline with shrubs were more important than IMFENCE and 
DISN3I (see Table 2). Also, during 1990, when water levels 
were higher the vegetation along the shoreline possibly 
provided escape cover for the broods. Shoreline vegetation 
was missing along many ponds during 1989; this would explain 
why some impoundments were unused during both years of the 
study. Numerous vacant impoundments that were not used by 
Canada Geese had no trees or shrub cover along their 
shorelines. until the breeding population within the area 
became large enough to force nesting on these vacant 
impoundments or shrubs or trees become established along the 
shorelines, nesting or use of these impoundments by adult 
geese and broods probably will not occur. 
Analysis of the 18 habitat variables by discriminant 
function resulted in an overall average of 80% and 100% of 
the impoundments being correctly classified during 1989 and 
1990, respectively. However, the 80% average reported for 
1989 was misleading because only eight occupied impoundments 
were correctly classified as being occupied. It was 
apparent that discriminant function had difficulty in 
distinguishing between occupied and vacant impoundments, 
because 12 occupied impoundments were misclassified as 
48 
vacant impoundments. Eleven of the 12 misclassified 
impoundments were stockponds. I assumed that the 
misclassification was caused by the occupied impoundments 
having a smaller surface area comparable to that of the 
vacant impoundments, although, the surface area of the 
watershed lake that was misclassified was 85% of normal. If 
the impoundments were misclassified because of their size, 
then the occupied watershed lake whose surface area was 
reduced because of the drought should have been 
misclassified. During 1990, using the same habitat 
variables, discriminant function accurately distinguished 
between occupied and vacant impoundments. It was important 
to note that impoundment surface areas were significantly 
larger, enabling discriminant function to correctly classify 
impoundments. It was possible that drought years produced 
results that made it hard to discern between impoundment 
classes, while years when good habitat conditions were 
present, enabled discriminant function to distinguish 
between the two classes. The results of my study were 
similar to those reported by Stiefel (1980) who, was able 
predict the presence and absence of Giant Canada Geese on 
stockponds in South Dakota 83.3% and 97.6% of the time, 
respectively. Mack and Flake (1980) studying duck brood use 
of stockponds in South Dakota were able to classify 90% of 
their study ponds correctly as either brood or nonbrood 
stockponds. 
MANAGEMENT
 
The results from both years of the study indicated that 
nest structures were important to the selection of an 
impoundment as a nest site. If the structures were not 
important the geese would have never used them. I believed 
that the program would have been more successful had the 
structures been maintained and kept filled with nesting 
material. If the KDWP did not have the manpower to maintain 
the structures then they should have looked to other 
organizations such as Ducks Unlimited for help. 
There were many impoundments in the Flint Hills that 
could provide suitable nesting habitat for Canada geese if 
nest structures were provided and maintained. The absence 
of maintenance was apparent during both years of my study. 
Nest structures were either leaning over in the water, were 
submerged under water, were rusting out, or had little or no 
nest material in them. The nest structures I saw in good 
condition were maintained by the area ranchers, primarily by 
Mr. Charles Wiggins. 
j
 :1 
1
 I
 
j
 ,
 i 
J
 
I
 ! 
..
 
50 
LITERATURE CITED 
ASHTON-TATE. 1985. dBASE III PLUS, Torrance, CA. 
Be11rose, F.C. 1976. Ducks, geese and swans of North 
America. Stackpole Co., Harrisburg, PA., and wildl. 
Manage. Inst., Washington, D. C. pp.141-165. 
Brakhage, G.K. 1965. Biology and behavior of tub-nesting 
Canada Geese. J. wildl. Manage. 29:751-771. 
Belanger, L., and R. Couture. 1988. Use of man-made ponds 
by dabbling duck broods. J. wildl. Manage. 52:718­
 723. 
Bultsma, P.M., R.L. Linder and T.L. Kuck. 1979. 
Reproductive success of Giant Canada Geese in western 
South Dakota. Proc. S.D. Acad. Sci. 58:35-38. 
Cooper, J.A. 1978. The history and breeding biology of 
the Canada Geese of Marshy Point, Manitoba. Wildl. 
Monogr. 61. 87pp. 
Fortner, J.R., J.T. Neill, and S.C. Ekhart. 1982. Soil 
survey of Greenwood County, Kansas. United States 
Department of Agriculture Soil Conservation Service. 
116pp. 
Hamilton, A.L. 1978. Spring, summer and fall use of 
stockponds in southeastern Montana by Canada Geese. 
Montana Dept. Fish and Game, Helena. P-R Rep. proj. W­
 120-R-8, 9. 71pp. 
Hanson, H.C. 1965. The Giant Canada Goose. Southern Ill. 
Univ. Press, Carbondale. xix + 226pp. 
Johnsgard, P.A. 1978. Waterfowl of North America. 
Indiana Univ. Press, Bloomington, pp.76-80. 
Lokemoen, J.T. 1973. Waterfowl production on 
stock-watering ponds in the northern plains. J. Range 
Manage. 26:17-184. 
1 , H.F. Duebbert, and D.E. Sharp. 1984. 
:f Nestspacing, habitat selection, and behavior of ~~ 
waterfowl on Miller Lake Island, North Dakota. J . .~ 
I
 .~ 
.~ Wildl. Manage. 48:309-321 . 
MacInnes, C.D. 1962. Nesting of small Canada Geese near
 '! Eskimo Point, Northwest Territories. J. Wildl.i Manage. 26:247-256. 
....................
 
1 
51 
Mack, G.D., and L.D. Flake. 1980. Habitat relationships 
of waterfowl broods on South Dakota stock ponds. J. 
wildl. Manage. 44:695-700. 
Mackey, D.L., W.C. Matthews, Jr., and I.J. Ball. 1988. 
Elevated nest structures for Canada Geese. Wildl. Soc. 
Bull. 16:362-367. 
Poly, D.M. 1979. Nest site selection in relation to water 
level and some aspects of hatching success in Giant 
Canada Geese (Branta canadensis maxima). M.S. Thesis, 
Middle Tennessee State Univ., Tenn. Wildl. Resour. 
Agency Tech. Rep. 79-13 39pp. 
Rumble, M.A., and L.D. Flake. 1983. Management 
considerations to enhance use of stock ponds by 
waterfowl broods. J. Range Manage. 36:691-693. 
SAS Institute. 1979. SAS users guide. SAS Institute, 
Raleigh, N.C. 494pp. 
Stoudt, J.H. 1971. Ecological factors affecting waterfowl 
production in the Saskatchewan Parklands. u.S. Fish 
and Wildl. Servo Resour. Publ. 99 56pp. 
Stiefel, D.M. 1980. Nesting Giant Canada Geese in western 
south Dakota. M.S. Thesis, South Dakota State Univ., 
Brookings. 44pp. 
Szymczak, M.R. 1975. Canada Goose restoration along the 
foothills of Colorado. co. Dept. of Natural Res. Fed. 
Aid Proj. W-88-R. 63pp. 
u.S.
 Dept. of Commerce. National Oceanic and Atmospheric 
Administration. Climatological Data: Kansas. Vol. 102. 
Washington: Govt. Printing office, 1988. 
National Oceanic and Atmospheric 
Administration. Climatological Data: Kansas. Vol. 103. 
Washington: Govt. Printing Office, 1989. 
National Oceanic and Atmospheric 
Administration. Climatological Data: Kansas. Vol. 104. 
Washington: Govt. Printing Office, 1990. 
Wetzel, R.G. 1975. Limnology. W.B. Saunders Company, 
Philadelphia, 743pp . 
..	 
52 
~Q.~ 
Signature of Graduate Student 
&rc£ 
I, Alan Dale Weinbrenner hereby submit this thesis 
to Emporia State University as partial fulfillment of the 
requirements of an advanced degree. I agree that the 
Library of the University may make it available for use in 
accordance with its regulations governing materials of this 
type. I further agree that quoting, photocopying, or other 
reproduction of this document is allowed for private study, 
scholarship (including teaching), and research purposes of a 
nonprofit nature. No copying which involves potential 
financial gain will be allowed without written permission of 
the author. 
~ Q. ~~~ 
Signature of Author 
19 November L 1991 
Date 
Nest site Selection on Flint Hills ImRoundments 
by Canada Geese (Branta canadensis) 
Title of Thesis 
)( /I~~L~.t~Lk.// 
Signat'tr 
~'WrrJL/U)9, )CJq) 
Date'Received I 
_................