Matthew D - s3.amazonaws.com

advertisement
1
Brood Surveys and Hunter Observations Used to Predict Gobbling Activity of Wild Turkeys in
2
Mississippi
3
Matthew D. Palumbo,* Francisco J. Vilella, Bronson K. Strickland, Guiming Wang, Dave
4
Godwin
5
M.D. Palumbo, B.K. Strickland, G. Wang
6
Department of Wildlife, Fisheries, and Aquaculture, Mississippi State University, Mississippi
7
State, MS 39762
8
Present address of M.D. Palumbo: Westchester County Airport, 14 Loop Road Building 10,
9
White Plains, New York 106704
10
Francisco J. Vilella
11
USGS Cooperative Research Unit, Department of Wildlife, Fisheries, and Aquaculture,
12
Mississippi State University, Mississippi State, MS 39762
13
Dave Godwin
14
Mississippi Department of Wildlife, Fisheries, and Parks, 1505 Eastover Drive, Jackson, MS
15
39211
16
*
Corresponding author: Plums526@aol.com
17
18
Disclaimer: “This draft manuscript is distributed solely for the purposes of scientific peer
19
review. Its content is deliberative and predecisional, so it must not be disclosed or released by
20
reviewers. Because the manuscript has not yet been approved for publication by the U.S.
21
Geological Survey (USGS), it does not represent any official USGS finding or policy”.
1
22
Abstract
23
The Mississippi Department of Wildlife, Fisheries, and Parks utilize data from turkey hunter
24
observations and brood surveys from across the state to manage wild turkey (Meleagris
25
gallopavo) populations. Since 1996, the Mississippi Department of Wildlife, Fisheries, and
26
Parks’ (MDWFP) personnel and cooperating wildlife managers have collected gobbling, jake
27
observation, and brood survey data to better forecast poult recruitment and hunting quality. Our
28
objective of this study was to evaluate if MDWFP’s survey data can be used as a viable predictor
29
of gobbling activity and hunt quality. We used three mixed models to investigate the relationship
30
between number of gobblers heard per hour of hunting (expected hunt quality) and the number of
31
jakes observed per hour of hunting one year prior (model 1), the number of gobblers heard per
32
hour of hunting and the total number poults per total hens observed 2 years prior (model 2), and
33
the number of jakes seen per hour of hunting one year prior and the total number of poults per
34
hens 2 years prior (model 3) from 1996 to 2008 among 5 wild turkey management regions
35
encompassing the state. We incorporated year and region as random effects to account for spatial
36
and temporal variation. We found that model 1 (P = 0.001) and model 3 (P <0.001) displayed
37
significant relationships. We recommend managers incorporate hunter observations of jakes seen
38
per hour of hunting the previous year as an indicator of statewide gobbling activity, rather than
39
the brood survey data. Gobbling activity is influenced by a multitude of factors and further
40
research is warranted to determine statewide relationships.
2
41
Introduction
42
Many states index wildlife population abundance by collecting harvest data or use data
43
gathered during hunting seasons to estimate harvest, number of hunters, and days of hunting
44
(Strickland et al. 1994). However, many of these estimates vary greatly, making yearly and
45
regional comparisons difficult (Connelly et. al. 2005). Thus, wildlife agencies also rely upon
46
observations from field personnel to index wildlife populations. For some states, recruitment
47
estimates for wild turkey (Meleagris gallopavo) are indexed through brood surveys conducted by
48
volunteers on “routine duties” during summer (Shultz and McDowell 1957; Wunz and Shope
49
1980).
50
The Mississippi Department of Wildlife, Fisheries and Parks (MDWFP) uses a
51
combination of hunter harvest data and field observations to index wild turkey populations
52
throughout the state. Harvest data collected through a Spring Gobbler Hunting Survey (SGHS)
53
are used to index harvest rates, population size, age structure and gobbling activity on a yearly
54
basis across the state. Additionally, data collected during brood surveys conducted by the
55
MDWFP, Mississippi Forestry Commission, U.S. Forest Service, U.S. Army Corps of Engineers,
56
and Weyerhaeuser Company are used to index reproduction and estimate productivity of wild
57
turkey populations statewide. The MDWFP uses a combination of data from the SGHS and
58
brood survey to predict the age structure and population levels that spring hunters can expect.
59
The MDWFP also attempts to predict gobbling activity based on the brood survey of two years
60
prior and jakes seen one year prior from the SGHS, which is used to categorize expected hunt
61
quality. On average, adult gobblers gobble more than sub-adults (Bevill 1973; Hoffman 1990)
62
and male two and three year age classes are dominant over older sibling groups (Watts 1968).
63
This dominance can be displayed through increased gobbling (Bevill 1973). Thus, the MDWFP
3
64
assumes that the two year old age class of gobblers is the most vocal age group of male turkeys
65
(D. Godwin, A. Butler, and J. Koloski, MDWFP, unpublished report). However, gobbling
66
activity has been reported to be extremely variable and influenced by a variety of factors
67
(Hoffman 1990; Lint et al. 1995; Kienzler 1996; Miller et al. 1997a; Miller et al. 1997b)
68
Since 1996, MDWFP and cooperating personnel have recorded number of gobblers heard
69
per hour of hunting, number of jakes seen per hour of hunting, and number of poults per hen
70
from five turkey management regions in Mississippi. Our objective was to determine if a
71
relationship exists between counts of jakes observed and poults per hen and subsequent gobbling
72
activity. A relationship among these counts would make predictions of gobbling activity more
73
reliable and strengthen current strategies for turkey harvest recommendations.
74
Methods
75
In the SGHS, individual hunters voluntarily report throughout the spring season (typically
76
mid-March to 1 May) location of hunt (i.e., whether the hunt was on private or public land); date;
77
start and end time of hunting; county of hunt; number of gobblers heard; number of gobbles
78
heard; number of gobblers, jakes, hens and unknowns seen; time a gobbler was harvested; and
79
beard and spur lengths (D. Godwin, A. Butler, and J. Koloski, Mississippi Department of
80
Wildlife, Fisheries and Parks, unpublished report). From June to August while conducting daily
81
activities, brood survey participants record number of poults, hens with and without poults,
82
gobblers, and “unknown” (age and gender not discernible) observed, as well as location (county
83
and/or wildlife management area).
84
85
We calculated mean number of gobblers heard per hour of hunting, mean number of
jakes observed per hour of hunting, and total poults per total hens observed for each turkey
4
86
management region 1996-2008. The five turkey management regions were delineated by
87
differences in physiography and location of logistical support (Figure 1).
88
We estimated the mean number of gobblers heard per hour of hunting as a conservative
89
index of gobbling activity (Gobbling Activity Index, GAI) since the propensity to gobble may
90
differ among birds. We assumed that number of jakes seen per hour of hunting would be an
91
adequate index of recruitment of one-year-old gobblers to the two-year age class, when gobblers
92
are assumed to be most vocal (Jake Recruitment Index, JRI). Total number of poults per total
93
hens (Nest Success Index, NSI) was used as an index of fecundity and recruitment (Glidden
94
1977; Seiss 1989). Previous research at the Tallahalla Wildlife Management Area (TWMA) in
95
Mississippi found a high correlation (r = 0.82) between number of successful nests two years
96
prior and number of gobblers heard the current year (Miller et al. 1997b). Therefore, we assumed
97
that NSI from 2 years prior would index abundance of two-year-old gobblers.
98
We tested 3 separate mixed models to determine the relationships among NSI, JRI and
99
GAI from 1996 – 2008 using the SAS 9.2 software package (SAS Institute, Cary North Carolina,
100
USA). We used the mean count of GAI, JRI, and NSI for each region and year for analysis and
101
did not have NSI data for 1996. In model one we compared JRI to NSI to determine if poult
102
recruitment was a reliable predictor of jakes observed while hunting the following year. In
103
model two we compared GAI to JRI to determine if the mean number of jakes observed while
104
hunting was a reliable predictor of gobbling activity the subsequent year. Finally, with model
105
three we compared GAI to NSI to determine if poult recruitment was a reliable predictor of
106
gobbling activity two years later. We used year and region as random effects to account for
107
annual and regional variation in turkey counts and used the Kenward-Rodgers degrees of
108
freedom adjustment for models with multiple random effects (Littell et al. 2006). From each
5
109
model we output parameter estimates and associated errors for use predictive models.
110
Additionally, to gauge model fit we compared the predicted values from our mixed models to the
111
observed values for JRI in model one and GAI in models two and three using Pearson’s
112
correlation coefficient. We acknowledge this is not a true correlation between the response and
113
explanatory variables, but we provide this correlation statistic as a simple measure of model fit to
114
augment the F- and P-values for each model.
115
116
Results
All three models demonstrated relationships between current and prior turkey counts. In
117
model one, JRI was related to NSI one year prior (n = 60, F = 16.39, P < 0.001) with a
118
correlation of r = 0.65 between observed and predicted JRI counts. In model two, GAI was
119
related to JRI one year prior (n = 65, F = 11.46, P = 0.001) with a correlation of r = 0.91 between
120
observed and predicted GAI counts. Finally, in model three, GAI was related to NSI two years
121
prior (n = 60, F = 3.52, P = 0.067) with a correlation of r = 0.91 between observed and predicted
122
GAI counts (Table 1.)
123
Discussion
124
Previous research has focused on monitoring broods to index population parameters, such
125
as productivity (Speak et al. 1969; Gardner et al. 1972; Glidden 1977; Speake1980; Bartush et al.
126
1985). Miller et al. (1997b) attempted to correlate an estimate of productivity (i.e., nest success)
127
to breeding behavior (i.e., gobbling activity) for wild turkeys. Miller et al. (1997b) detected a
128
significant correlation between nest success two years prior to number of gobblers heard, but not
129
to number of calls heard. They hypothesized the proportion of two-year-old birds in the
130
population could be indexed by nest success two years previous and can contribute to the
131
potential of hearing an individual turkey gobble. We attempted to test this hypothesis at the
6
132
regional scale and did not detect the same relationship. But, we did determine the JRI used to
133
track the proportion of juvenile male turkeys in a population that will potentially be recruited to
134
the two year old age class was as suitable index of gobblers heard during a hunting season.
135
The Miller et al. (1997b) research was performed on TWMA in central Mississippi and nest
136
success was determined from radio-monitored hens from 1984 -1995 from Palmer et al. (1993)
137
and Miller et al. (1995). Palmer et al (1993) and Miller et al. (1997b) indexed nest success from
138
daily monitoring hens until they were found in the same location for 2 consecutive days,
139
whereupon the hens were considered to have initiated nesting. During incubation hens were
140
located twice daily to determine hatching date, nest destruction or desertion, or hen mortality. As
141
soon as incubation stopped researchers approached the nest to determine fate (Palmer et al. 1993
142
and Miller et al. 1995). These hens and their nests where intensely monitored (Palmer et al. 1993
143
and Miller et al. 1995) compared to the brood survey that the MDWFP can efficiently conduct.
144
The NSI of total poults per total hens seen 2 years prior, which is gathered from staff while they
145
are conducting routine duties throughout the state, is not as accurate. And therefore we
146
recommend not using the NSI as an indicator of expected hunt quality (i.e., gobbling activity) at
147
the regional level.
148
But the NSI was significantly correlated to the JSI which indicates that it can track population
149
size of wild turkeys at the statewide level. Furthermore, the JSI was also significantly correlated
150
to the GAI which indicates that the JSI, a surrogate for the gobbler population one year prior, is
151
an indicator of gobbling activity. But previous research by Lint et al. (1995) at TWMA reported
152
that gobblers heard per day was not related (P= 0.14) to estimates of gobbler population size but
153
was strongly correlated to harvest and harvest per effort.
7
154
Our statewide analysis resulted in different correlations than were documented at a single
155
management area, TWMA. The TWMA studies were at one research site where turkeys
156
observed, monitored, and hunted could be considered of one population. Therefore estimates of
157
these metrics are more consistent over years. Through the SGHS hunters may be visiting several
158
different areas and essentially recording data from several different populations. We attempted to
159
address the lack of spatial congruence by incorporating region and year as random effects, which
160
resulted in detecting significant statewide correlations.
161
The SGHS is useful to track long-term trends in wild turkey populations across the state.
162
But the amount of spatial variation between sources of data such as gobbling activity, brood
163
surveys, and jake encounters, likely can result in poor models to explain gobbling activity.
164
Standardization of brood, jake, and gobbling call count surveys should be warranted to decrease
165
this variation. The ability of observers to detect and identify birds has been recognized as a
166
source of variation in many studies trying to detect birds (Diefenbach et al. 2003), therefore,
167
unbiased observers (i.e. not hunters) that are consistently assigned to specific surveys may
168
reduce variation. Additionally, attempting to predict a behavioral response that is influenced by
169
such a variety of factors may consistently result in a high level of inaccuracy. The indices we
170
investigated can be used to predict regional gobbling activity with consideration that there is a
171
large amount of variation associated with this mating behavior. Therefore, we recommend
172
further research into potential sources of variation (i.e., observers, habitat, hunter effort and
173
success, brood survey effort, weather, population dynamics) at regional levels, so that more
174
accurate relationships can be determined.
175
Acknowledgements
8
176
We greatly appreciate the efforts of Mississippi’s sportsmen and the various natural
177
resource agencies that have gathered wild turkey observations. We thank S. L. Edwards, R. S.
178
Seiss, A. Butler, J. Koloski for their support. Funding for this project was provided by the
179
Mississippi Department of Wildlife, Fisheries, and Parks through project W-48-56, Study 58 of
180
the Federal Aid in Wildlife Restoration Program. Any use of trade, firm, or product names is for
181
descriptive purposes only and does not imply endorsement by the U.S. Government.
182
183
184
185
References
Bartush WS, Sasser MS, and Francis DL. 1985. A standardized turkey brood survey method for
Northwest Florida. Proceedings of the National Wild Turkey Symposium. 5:173-181.
Bevill WV Jr. 1973. Some factors influencing gobbling activity among wild turkeys.
186
Proceedings of the Annual Conference of Southeastern Association of Game and Fish
187
Commissioners. 27:62-73.
188
189
190
Bevill WV Jr. 1975. Setting spring gobbler seasons by timing peak gobbling. Proceedings of the
National Wild Turkey Symposium 3:198-204.
Connelly J W, Gammonley JH, and Peek JM. 2005. Harvest management. 6th edition. Pages
191
658-683 in C. E. Braun, editor. Techniques for wildlife management investigations and
192
management. Bethesda, Maryland: The Wildlife Society.
193
194
195
Diefenbach DR, Brauning DW, and Mattice JA. 2003. Variability in grassland bird counts
related to observer differences and species detection rates. The Auk. 120(4):1168-1179.
Gardner DT, Speake DW, and Fleming WJ. 1972. The effects of a spring “gobblers only”
196
hunting season on wild turkey reproduction population size. Proceedings of the Annual
197
Conference of Southeastern Association of Game and Fish Commissioners. 26:244-252
9
198
199
200
201
202
Glidden JW 1977. Net productivity of the wild turkey population in Southwest New York.
Transactions of the Northeast Section of the Fish and Wildlife Conference 34:13-21.
Hoffman RW 1990. Chronology of gobbling and nesting activities of Merriam’s wild turkeys.
Proceedings of the National Wild Turkey Symposium 6:25-31.
Kienzler JM, Little TW, and Fuller WA. 1996. Effects of weather, incubation, and hunting on
203
gobbling activity in wild turkeys. Proceedings of the National Wild Turkey Symposium.
204
7:61-68.
205
Lewis JC. 1967. Physical characteristics and physiology. Pages 45 – 72 in O. H. Hewitt, editor.
206
The wild turkey and its management. Washington D. C.: The Wildlife Society.
207
Lint RJ, Leopold BD, and Hurst GA. 1995. Comparison of abundance indexes and population
208
209
210
211
estimates for wild turkey gobblers. Wildlife Society Bulletin 23:164-168.
Littell, R. C., G. A. Milliken, W. W. Stroup, R. D. Wolfinger, and O. Schabenberger. 2006.
SAS for mixed models, second edition. SAS Institute Inc., Cary, North Carolina, USA.
Miller DA, Weinstein M, Pries SR, Leopold BD, and Hurst GA. 1995. Wild turkey reproductive
212
parameters from two different forest ecosystems in central Mississippi. Proceedings of
213
the Annual Conference Southeastern Association of Game and Fish Commissioners.
214
49:466-475.
215
216
217
Miller DA, Hurst GA, and Leopold BD. 1997a. Chronology of wild turkey nesting, gobbling,
and hunting in Mississippi. Journal of Wildlife Management. 61:840-845.
Miller DA, Hurst GA, Leopold BD. 1997b. Factors affecting gobbling activity on wild turkeys in
218
central Mississippi. Proceedings of the Annual Conference of Southeastern Association
219
of Fish and Wildlife Agencies 51:352-361.
10
220
Palmer WE, Hurst GA, and Lint JR. 1990. Effort, success, and characteristics of spring turkey
221
hunters on Tallahala Wildlife Management Area, Mississippi. Proceedings of the
222
National Wild Turkey Symposium 6:208-213.
223
Palmer WE, Priest SR, Seiss RS, Phalen PS, and Hurst GA, 1993. Reproductive effort and
224
success in a declining wild turkey population. Proceeding of the Annual Conference of
225
the Southeastern Association of Fish and Wildlife Agencies 47:138-147.
226
SAS. 2003. Version 9.1.3 SAS Institute, Cary North Carolina, USA.
227
Seiss RS. 1989. Reproductive parameters and survival rates of wild turkey hens in east-central
228
229
230
Mississippi. Master’s Thesis. Mississippi State: Mississippi State University.
Shultz V. and McDowell RD. 1957. Some comments on a wild turkey brood study. Journal of
Wildlife Management 21:85-89.
231
Speake DW, Barwick LH, Hillestead HO, and Stickney W. 1969. Proceedings of the Annual
232
Conference of Southeastern Association of Game and Fish Commissioners 23:46-58.
233
234
235
Speak DW. 1980. Predation on wild turkeys in Alabama. Proceedings of the National Wild
Turkey Symposium 4:86-101.
Strickland MD, Harju HJ, McCaffery KR, Miller HW, Smith LM, and Stoll R J. 1994. Harvest
236
Management. 5th Edition. Pages 445-473 in T.A. Bookhout, editor. Research and
237
management techniques for wildlife and habitats. Bethesda, Maryland: The Wildlife
238
Society.
239
240
241
242
Watts CR. 1968. Rio Grande turkeys in the mating season. Transactions of the North American
Wildlife and Natural Resource Conference. 33:205-210.
Wunz GA and Shope WK. 1980. Turkey brood study in Pennsylvania as it relates to harvest.
Proceedings of the National Wild Turkey Symposium 4:69-75.
11
Tables
Table 1.
Three mixed models with year and region as random effects to determine the relationship among gobblers heard per
hour of hunting, mean number of jakes seen per hour of hunting one year prior and the total poults per total hens seen 2
years prior from 1996 – 2008 across all 5 of Mississippi Department of Wildlife, Fisheries, and Parks wild turkey
management regions.
Model
Response
variable
Explanatory
variable
Intercept
estimate
SE
Slope
estimate
SE
Num
df
Den df
Fvalue
Pvalue
2
GAIa
JRIb
0.563
0.063
0.483
0.143
1
53.4
11.46
0.001
3
GAI
NSIc
0.609
0.066
1
JRI
NSI
0.098
0.028
a. Mean number of gobblers hear per hour of hunting.
0.029
0.051
0.016
0.013
1
1
47.8
57.9
3.52
16.39
0.067
<0.001
b. Mean number of jakes seen per hour of hunting one year prior.
12
c. Total poults per total hens observed 2 years prior.
Figure Captions
Figure 1.
The Mississippi Department of Wildlife, Fisheries, and Parks wild turkey
management regions from 1996 - 2008.
13
Download