HomeMy WebLinkAboutAPA3382for
LGL Alaska Ltd.
Anchorage and Fairbanks, Al'aska
SUSITNA HYDROELECTRIC PROJECT
DRAFT REPORT
TERRESTRIAL ENVIRONMENTAL
MITIGATION PLANNING SIMULATION MODEL
by
Robert R. Everitt
Nicholas C. Sonntag
ESSA Environmental and Social Syst.ems Analysts Ltd.
Vancouver, B.C., Canada
Gregory T. Auble
James E. Roelle
U.S. Fish and Wildlife Service
Fort Collins, Colorado
William Gazey
LGL Ecological Research Associates
Bryan, Texas
April 27, 1983
e material in this report is preliminary in nature
.rl should not be cited in technical publications.
[
r
[
[
[
[
[
6
[
c
8
~
c
~
c
u
[
u
l
..-
~
~ c.o
0
0
0
~
t-
("')
("')
I '-1 ZS
'58
The model described herein was developed at a series of
workshops at which representatives of LGL Alaska, the Alaska
Department of Fish and Game, ESSA Ltd. and others contributed many
ideas and suggestions. The material presented in this report is
preliminary in nature and should not be cited in any technical
publications without the written approval of both LGL Alaska and
the Alaska Department of Fish and Game.
ARLIS
Alaska Resources
Library & Information Serv1ces
Am::horage, Alask;:;
[
[
[
r L
[
[
[
0
[
[
rJ
~
c
u
[
E
6
[
ACKNOWLEDGEMENTS
We would like to thank the over fifty different participants
at the workshops who devoted much time and considerable energy to
the process of building the model. In particular, we thank Warren
Ballard and SuzAnne Miller of the Alaska Department of Fish and Game
for allowing us to use their modelling work on moose as a basis for
the moose submodel described in this report. They also contributed
the technical Appendix I on moose population modelling in the Upper
Susitna Basin.
Once again, Jean Zdenek showed patience and wizardry in
typing and correcting this and earlier drafts of this report.
[
[
[
[
[
[
E
[
[
u
1.0
2.0
3.0
TABLE OF CONTENTS
INTRODUCTION. . . . . . . . . . . . . . . .
1.1 Objectives ............ .
1.2 Relationship to Mitigation Planning ........ .
1.3 Simulation Modelling Workshops .
1.3.1 Workshop Activities .
1.3.2 Beyond the Workshop .......... .
BOUNDING. . . . . . . . . . . . . . . . . . .
2.1 Actions. . . . . . . . . . . . . . . . . . .
2.2 Indicators . . . . . . . . .
2.3 Spatial Considerations . . . . . . . . . . .
2.4 Temporal Considerations. . . . . . .
2.5 Submodel Definition. . . . .
2.6 Looking Outward. . . . . . . . . . . . .
SUBMODEL DESCRIPTIONS . . . . . . . . . . .
3.1 Physical Processes/Development/Recreation.
3.2
3.3
3.1.1 Physical Processes ......... .
3.1.2
3 .1. 3
3 .1. 4
3.1.1.1 Reservoir Elevations ....... .
3 .1 . 1 . 2 Stage. . . . . . . . • . . . . .
· 3.1.1.3 Water Surface Area in the Downstream
Floodplain (Devil Canyon to Susitna-
3.1.1.4
3.1.1.5
3.1.1.6
Chulitna Confluence) ...... .
Ice Dynamics . . . . . . . . . . .
3.1.1.4.1 Formation of Ice Cover.
3.1.1.4.2 Ic~ Staging .. .
3.1.1.4.3 Break-up ...... .
Flood Events • . . . . . . . . . .
Downstream Effects . . . . . . . .
3.1.1.6.1 Beaver Overwintering
Habitat ....... .
3.1.1.6.2 Vegetation Succession .
3 . 1 . 1 . 7 Snow . . . . . . . . . . . .
Hydroelectric Development Activities.
Other Land Use Activities
Disturbance to Wildlife . • .
3.1.4.1 Recreational Use
Vegetation . . . . . . . . . . . . . . . . . . .
3. 2 .1 Structure . . . . . . . . . . . . . . . .
3. 2. 2 Classification System . . . . . . . . . .
3.2.3 Development Activities ........... .
3.2.4 Riparian Succession .......... .
3.2.5 Wildlife Habitat. . .. .
Furbearers and Birds . . . . . . . . . . . . . .
3.3.1 Beaver .... e o ••••••••••
3.3.1.1 Beaver Carrying Capacity ... .
3.3.1.2 Intrinsic Growth Rate (r) ..... .
1
2
3
3
4
6
7
7
7
10
13
13
15
18
18
18
19
19
22
22
22
24
24
27
27
27
32
32
32
32
36
36
39
41
41
44
45
48
51
51
52
56
[
[
[
[
[
[
[
c
[
4.0
3.4
3.5
3.6
3.3.2
3.3.3
Moose
3.4.1
3.4.2
3.4.3
3.4.4
3.4.5
3.4.6
TABLE OF CONTENTS (cont'd.)
3.3.1.3
3.3.1.4
3.3.1.5
Marten .
3.3o2.1
Birds. .
3.3.3ol
3.3.3.2
3.3.3.3
Mortality . . . . . . . . . . . . .
Beaver Migration. . . . . . . . . .
Beaver's Impact on Vegetation .
Population Structure.
Passerine Birds .
Trumpeter Swan.
Golden Eagle.
Structure. . . . . . ·. . . . . . . . . . .
Wolf Population. . . . . . . . . . . . .
Moose Reproduction
Mortality ....... • ..
3.4.4.1 Neo-Natal Mortality .
3.4.4.2 Spring Wolf Predation .
3.4.4.3 Summer Wolf Predation ...... .
3.4.4.4 Bear Predation .... .
3.4.4.5 Harvest ........... .
3.4.4.6 Post-Harvest Population Statistics.
3.4.4.7 Winter Wolf Predation ... .
Winter Carrying Capacity . . . . . . . .. .
Winter Mortality . . . • . . . . . . . . . .
3.4.6.1 Winter Mortality as a Function of
Carrying Capacity ........ .
3.4.6.2 Winter Mortality as a Function of
Snow Accumulation. . . . . . . ..
57
62
62
63
63
66
66
72
72
72
73
74
77
77
77
79
80
80
82
82
82
83
84
84
Bear Submodel . . . . . . . . . . . . . . . . .
87
87
87
97
99
99
3.5.1 Population Structure . . . . . .... .
3.5.2 Initial Population Equilibrium o ..... .
3. 5. 3 Indices. . . . . . . . . . . . . . . . .
3.5.3.1 Summer and Fall Food Index.
3.5.3.2 Spring Food Index ...... . 103
3.5.3.3 Disturbance and Hunting Effort
Indices. . . ......... 105
3.5.4 Reproduction . . . ........... 105
3.5.5 Mortality. . . . . . . . 107
3o5.6
Model
3.6.1
3.6.2
3.6.3
3.6.4
3o6.5
3.5.5.1 Hunting Mortality . . ..... 107
3.5.5.2 Natural Mortality . . . . . 110
3.5.5.3 Nuisance Kill . . . . 0 •• 110
Dispersal. . . . . . . . . . . . . . 110
Results . . . . . . . • . . . . . . . . . 113
Physical Processes/Development/Recreation. . 115
Vegetation . . . . . . . 12 0
Furbearers and Birds . o . . . . . . 0 • 127
Moose. . . . . . ........ 136
Bears. . o . . . . . . . . . 136
CONCEPTUAL MODEL . . . . . 149
[
r
[
[
[
[
[
E _j
[
c
[J
~
E
c
[
[
[
b
r
5.0
TABLE OF CONTENTS (cont'd.)
Page
MITIGATION PLANNING. . . . . . . . • . . . . . . . . . 151
5.1 Physical Processes/Development/Recreation . . 151
5.1.1 Model Refinements. . . .......... 151
5.1.1.1 Recreation .............. 151
5.1.1.2 Development and Land Use ....... 151
5.1.1.3 Physical Processes .......... 152
5.1.2 Information Needs. . . . . . . . . . . 154
5. 1 . 3 ~-1i tiga tion . . . . . . . . . . . . . . . . . . 15 5
5 . 2 Vegetation. . . . . . . . . . . . . . . . . . . . 15 6
5.2.1 Model Refinements/Information Needs. . .. 156
5.2.1.1 Spatial Resolution. . . . . . 157
5.2.1.2 Ice Processes and Riparian Succession 157
5.2.1.3 Resolution of Development Activities. 158
5.2.1.4 Wildlife Food . . . . . . . . . . 158
5.2.1.5 Dynamics of Upland Vegetation .... 158
5.2.2 Planned Studies. . . . . . . . . . 159
5.2.2.1 Phenology . . . . . . . . .... 159
5.2.2.2 Food Habits ......•...... 160
5. 2. 2. 3 Browse Sampling . . . . . . . . . 160
5. 2. 2. 4 Browse Mapping. . . . . . 16 0
5. 2. 2. 5 Energetics Modelling. . . . . . . . . 161
5.2.2.6 Carrying Capacity . . . . . . .. 161
5.2.2.7 Monitor BLM Burn Site . . .. 161
5.2.3 Needed Studies . . . . . . . . . . . . . . 161
5. 2 .. 3 .1 ?-1oni tor Other Vegetation
Manipulations. . . . . . . . . . . . 16 2
5.2.3.2 Ice Processes and Riparian Vegetation 162
5.2.4 Mitigation and Monitoring. . . . . .... 163
5. 3 Furbearers and Birds. . . . . . . . . . . . . . . 16 3
5 . 3. 1 Beaver . . . . . . . . . . . . . . . . . 16 4
5. 3 .1.1 Model Refinements . . . . . . . . 164
5.3.1.2 Information Needs/Research ...... 165
5.3.lo3 Mitigation. . . . . o .... 167
5 . 3 . 2 Marten . . . . . . . . . . . . . . . . 16 8
5.3.2.1 Model Refinements . . . . .... 168
5.3o2.2 Information Needs . . 168
5.3.2.3 Mitigation. . . . . . .... 169
5 . 3 . 3 Birds . . . . . . . . . . . . . . . . . 16 9
5. 3. 3.1 Information Needs . . . . . . . . 169
5.3.3.2 Mitigation/Monitoring . . .... 170
5 . 4 Moo s e ~~~ . . . ., . . . . . . . . . . . . . . . . . . . 1 71
5.4.1 Model Refinements. ~ ........ o . . 171
5.4.1.1 Spatial Definition. . . . . . . . 171
5.4.1.2 Bear Predation. . . . 171
5 o 4 . 1 . 3 Wolf Predation. . . . . . . . . . 17 2
5.4.1.4 Winter Mortality ........... 173
5.4.1.5 Model Testing and Evaluation. 175
5.4.2 Planned Studies. . . ........... 175
5 . 4 . 2 .1 Moose . . . . . . . . . . . . 17 5
[
[
[
[
[
[
[
E
[
c
E
[
6.0
7.0
8.0
TABLE OF CONTENTS (cont'd.)
5 o 4 . 2 . 2 Wolves . o • • • ·• • • • • • • • 17 6
5.4.3 Needed Studies. . . • . . . . . . 176
5.4.4 Mitigation and Monitoring . . 177
5 . 5 Bears. . . . . . o • • • • • • • • • • • 17 8
5.5.1 Model Refinements . . . . . . . . . . 178
5.5.1.1 Bioenergetics and Foraging ... 178
5.5.1.2 Initial Equilibrium. . . . . 178
5.5.1.3 Berry Production .......... 179
5.5.1.4 Spatial Resolution . . . . . . . 179
5.5.1.5 Prairie Creek Salmon Resource. . 180
5. 5 .1. 6 Dispersal and Harvesting . . . . 18 0
5.5.1.7 Composite Food Index ........ 180
5.5.2 Mitigation and Monitoring . . . . . . 181
FUTURE WORK . . . 18 2
REFERENCES. . . 184
LIST OF PARTICIPANTS .. . . 18 6
APPENDIX I -UPPER SUSITNA RIVER BASIN MOOSE POPULATION
MODELLING . . . . . . . . ,. . . . . . . • . . . . . • . . . . 19 4
[
F
E
[
[
[
[
[i
[
[
B
~
c
u
[
E
[
L
r
2.1
2.2
2.3
2.4
2.5
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
3.14
3.15
3.16
3.17
3.18
3.19
3.20
3.21
3.22
3.23
3.24
LIST OF TABLES
Actions identified at workshop . . • . . . • . . . . . .
Indicators identified at workshop ........... .
Fourteen vegetation types associated with the spatial
areas. . . . . . . . . . . .
Submodel components decided on by workshop participants.
Looking Outward Matrix . . . . . . . . . . . . . . . . .
Hydroelectric development project actions. . ..
Estimates of current land use and recreational use in
geographic area considered in the model ..... .
Disturbance associated with construction workers and
vehicle traffic. . . . . . . . . . • . . . . . . .
Estimated recreation demand •.......•..
Initial conditions for vegetation types .....
Estimates of average values for potentially available
browse standing crop and annual berry production in
each land class ....•...............
Various parameters for marten population model . . . . .
Passerine bird density and number of species associated
with different vegetation types ........... .
Number of bird territories/10 ha for 12 bird species for
each of-the vegetation types represented in the model.
Avian territories/ha used in model .......... .
Moose mortality rates at various depths of snow
8
9
14
16
17
33
37
38
40
43
49
64
67
70
71
accumulation . • . • • . . . . . • . . . . . . . . . . 8 8
Assumed proportion of bears reaching maturity by age 95
Brown bear base natural mortality estimates. . . . . 98
Black bear base natural mortality estimates. . . . . 98
Assumed brown bear initial population size . • • . . . . 100
Assumed black bear initial population size . . ... 100
Brown bear dispersal weight by class and sex . . . . . . 101
Black bear dispersal weight by class and sex . . 101
Brown bear harvest weight by class and sex . . . 102
Black bear harvest weight by class and sex . . . . . . . 102
Assumed relative preference of vegetation types. . . 104
Brown bear nuisance kill weights by class and sex .... 112
Black bear nuisance kill weights by class and sex .... 112
Scenarios used in the simulations •..••.•..... 114
[
[
[
[ 2.la
[ 2.la
[ 3.1
3.2
3.3
[ 3.4
3.5
6 3.6
3.7
[ 3.8
3.9
3.10
6 3.11
IJ 3.12
3.13 u 3.14
c 3.15
3.16
[ 3.17
j 3.18
3.19 c 3.20
3.21
3.22
E 3.23
3.24
[ 3.25
3.26
b
r
LIST OF FIGURES
Upper Susitna Basin showing the Devil Canyon and
Watana impoundments .•....•.•.
-Lower Susitna Basin showing Devil Canyon to Talkeetna
riparian zone designated for the model . . . .
Gold Creek flows for preproject, case A, case C, and
Case D, assuming both dams operating . . . . .
Watana Reservoir elevations throughout the year.
Stage-discharge rating curves for Gold Creek Station
Water surface area in the downstream floodplain as a
function of discharge measured at Gold Creek Station .
Hypothetical relationship of area of maximum ice cover
as a function of discharge . • . . . . . . . . . . . .
Simulated timing of events affecting break-up ..... .
Potential overwintering habitat as a function of stage .
Calculation sequence for the vegetation submodel . . . .
Successional sequence in the Talkeetna to Devil Canyon
riparian zone. . . . . . . . . . . . . ....
Time dynamics of a population based on the logistic
growth model . . . . ~ . . . . . . . . . . . . . . .
Percent survival of beaver colonies on main channel as
a function of maximum change in water level from
summer to winter . . . . • . ~ • . . • . . . . . .
Mortality as a function of ice scouring area for slough
and main channel beaver populations ..•..•..
Maximum beaver trapping mortality as a function of a
user specified price factor. • • . . . . . . .
Trapper access factor as a function of the number of
people using the area ..•...••.......
Density dependent mortality rate for marten population .
The relative value of species in any given vegetation
type . . . . . . . . . . . . . . . . . . . . . . . . . .
Relative value of bird density in any given vegetation
type . . . . . . . . . . . c;l • • • • • • • • • • • • •
General structure of the moose submodel ........ .
Wolf fecundity rate as a function of population size ..
Wolf mortality rate as a function of snow accumulation .
Moose fecundity rates as a function of population size .
Bear predation rates as a function of moost. population
s~ze . . . . . . . . • . . ~~~ . . • . . • . . • . . . .
Forage availability as a function of snow accumulation .
Male winter mortality rates as a function of forage
availability . . . . . . . . . . . . . . . . . . . . .
Female winter mortality rates as a function of forage
availability . . . . . • . . . . . . . . . . . . . . .
Diagrammatic representation of the division of the bear
population into vulnerable and non-vulnerable numbers.
11
12
20
21
21
23
25
26
31
42
46
53
58
59
61
61
65
68
68
75
76
76
78
81
85
85
86
89
r
[
[
3.27
3.28
[ 3.29
3.30
3.31
[ 3.32
3.33
[ 3.34
[ 3.35
3.36
3.37
E 3.38
3.39
3.40
[ 3.41
3.42
6 3.43
3.44
u 3.45
3.46
~ 3.47
E 3.48
3.49
6 3.50
[ 3.51
3.52
E 3.53
3.54
[ 3.55
3.56
3.57
6 3.58
3.59
C
LIST OF FIGURES (cont'd.)
Life structure of female brown bear ....... .
Life structure of male brown bear •.
Life structure of female black bear ..
Life structure of male black bear ..... .
Sequence of bear submodel operations .
Reproduction relationships as a function of the index
91
92
93
94
96
of food. . . . . . . . . . . . . . . . . . . 106
Hunting mortality rate as a function of the hunter
index with the effect of a lower sensitivity
illustrated. . . . . . . . . . . . . • . . . . . 109
Number of nuisance kills as a function of construction
activity ........... " ........ .
Maximum annual change in stage at Gold Creek station
Amount of reservoir clearing (ha) per year . . .
Construction personnel on site at any one time ...
Recreational use days in the Upper Susi tna Basin. . .
Potential overwintering habitat for beaver in sloughs
. 111•
. 116
. 117
. 118
. 119
and side channels. . ..........•..... 121
Area subject to ice scouring in the downstream reach . 122
Minimum surface area covered with water in the
downstream reach . . . . . . . . . . . . .
The areal extent of deciduous and mixed forest
The area extent of low mixed shrub in the Upper
. . . 123
124
Susitna Basin. . . . . . . . . . . . . . . . . . 125
Winter forage availability for moose in the Upper
Susitna Basin. . . . . . . . • . . . . • • . . 126
The areal extent of deciduous and mixed forest in the
downstream floodplain. . • . . . . . . . . • . . 128
The areal extent of tall shrub in the downstream
floodplain . . . . . . • • • • . . . • . • . . . 129
The areal extent of low mixed shrub in the downstream
floodplain . . . . . . . . . . . . . . . • . 130
The areal extent of pioneer vegetation in the
downstream floodplain. . . . . . . . . . . . . . . . 131
Beaver colonies utilizing the sloughs and side
channels and the corresponding carrying capacity
in the downstream riparian zone •.......... 132
Beaver colonies utilizing the main channel and their
carrying capacity in the downstream riparian zone. . 134
Total marten population in the modelled project area 135
Number of bird territories associated with brown
creeper in the total modelled area . . . . . . . . . 137
The number of bird territories associated with
northern water thrush in the total modelled area . . 138
The total number of bird territories associated with
the area represented by the model. . . . . • . . 139
Post harvest fall moose population . . . . . 0 140
Wolf population. . . . . . . . . . . o • • 141
Bear kills and wolf kills of moose . . . . . . 142
Age ratio and sex ratio. . . o • • • • • • • • 143
Moose harvest .................... 0 144
[
E
[
c
[
[
L
L
b
r
LIST OF FIGURES (cont'd.)
3.60 Black bear populations •.........
3. 61 Brown bear populations • . . • • . . ·. . . •
3.62 Index of summer and winter black bear food .
. 145
0 • • • 14 6
0 • • • • • 14 8
4.1 Conceptual model of major components and linkages
included in the model of the terrestrial environment
in the Susitna Basin .•.....•... o ••••• 150
•
[
f
[
L
[
[
[
D
[
[
E
~
c
c
[
G
L
l
- 1 -
1.0 INTRODUCTION
The technical feasibility, economic viability, and
environmental impacts of a hydroelectric development project in
the Susitna River Basin are being studied on behalf of the
Alaska Power Authority. As part of these studies, LGL Alaska.
Research Associates Inc. has been contracted to coodinate the
terrestrial environmental studies being performed by the Alaska
Department of Fish and Game and, as subcontractors to LGL,
several University of Alaska research groups. LGL is responsible
for further quantifying the potential impacts of the project on
terrestrial wildlife and vegetation, and for developing a plan to
mitigate adverse impacts on the terrestrial environment. The
impact assessment and mitigation approach is included as part of
a license application to the Federal Energy Regulatory Commission
(FERC), submitted in February, 1983.
The quantification of impacts, mitigation planning, and
design of future research is being organized using a computer
simulation modelling approach. Through a series of workshops
attended by researchers, resource managers, and policy-makers,
a computer model has been developed and is being refined for use
in the quantification o~ impacts on terrestrial wildlife and
vegetation, and for evaluating different mitigation measures
such as habitat enhancement and the designation of replacement
lands to be managed as wildlife habitat. This report describes
the current status of the model.
A preliminary model was developed at the first workshop
held August 23-27, 1982 in Anchorage. Considerable refinements
for the model were proposed in a series of technical meetings
held from November, 19~2 to February, 1983. Many of these
refinements were incorporated into the computer simulation model
and this refined version was presented at the mitigation planning
workshop held February 28 -March 2, 1983 in Anchorage. This
- 2 -
report describes the current status of the model, needed
refinements, andmakes suggestions about studies for the
terrestrial program.
1.1 Objectives
The ultimate purpose of the workshops and simulation
modelling is to develop a framework that can be used as a basis
for assessing impacts of and evaluating mitigation options for
the effect of the Susitna Hydroel~ctric Project on the terrestrial
environment in the Susitna Basin.
to:
The specific objectives for achieving this purpose are
a) develop an understanding of the biophysical
processes of the Susitna Basin with respect
to wildlife and vegetation;
b) develop this understanding by integrating information
on big game, furbearers, small mammals, birds, and
plant ecology into a computer simulation model;
c) refine the model during a series of technical meetings;
d) update the model as new information becomes available
from field studies; and
e) use the model as a framework and guide to assess
terrestrial impacts of the Susitna Hydroelectric
Project and to evaluate ways of mitigating impacts.
The workshops play a major role in attainment of these
objectives. They provide a systematic approach to organizing
information and people. As such, they are a major tool for
consensus building and interdisciplinary coordination.
[
r L~
c
[
I
l
[
c
[
c
[
[
[
E
[
[
[
[
[
[
[
c
E
l
b
c
- 3 _-
1.2 Relationship to Mitigation Planning
Many aspects of mitigation planning will be accomplished
outside of the simulation modelling workshop process. Many
mitigation measures, such as controlling dust along ro.ads,
leaving clumps of trees along the reservoir margin for eagle
nesting, minimizing aircraft disturbance, locating recreation
facilities away from critical wildlife areas, and deciding
upon environmentally sound access road design criteria can
easily be developed without a quantitative model. Most of
these measures to be incorporated into engineering design and
construction plannirig have been developed or will be developed
prior to the submittal of the FERC application.
However, certain mitigation measures, such as hab.i tat
enhancement or compensation lands for habitat lost, may
require several years of analysis and discussion. The primary
purpose of the simulation modelling workshop process· is to
incorporate these more complex issues into the mitigation
planning. Recognizing that these issues will not be
resolved prior to the license application, the workshop
process allows for an adaptive approach to planning. It
provides a framework for increased communication, and a
mechanism for designing and utilizing the results of future
research and monitoring studies.
1.3 Simulation Modelling Workshops
There has been an enormous increase in public concern
over environmental impacts of development projects in the past
two decades. One consequence of this concern has been thJ
use of detailed environmental impact assessments as an integral
part of major resource development activities. These impact
assessments are always multidisciplinary, but, in most cases,
little effort is made to develop a coordinated, interdisciplinary
- 4 -
approach. Consequently, vital _information required to make
predictions of impacts encompassing more than one discipline
is often overlooked or not collected.
Over the past ten years a group of environmental
scientists and systems analysts at the University of British
Columbia and the International Institute for Applied Systems
Analysis (IIASA) in Austria have developed a methodology to
deal explicitly with interdisciplinary ecological problems
(Holling, 1978). The core of the methodology is a five day
workshop involving a team of four or five experienced simulation
modellers and a group of fifteen to twenty specialists. The
focus of the workshop is the construction of a quantitative
simulation model of the system under study. The development
of the simulation model forces specialists to view their area
of interest in the context of the whole system. This promotes
an interdisciplinary understanding of the system, and allows
ecological and environmental knowledge to be integrated with
economic and social concerns at the beginning, rather than
at the end, of an impact assessment.
Simulation models require unambiguous information.
In the workshop setting specialists are forced to be explicit
about their assumptions. This objectivity exposes critical
conceptual uncertainties about the behavior of the system,
and identifies research needs.
1.3.1 Workshop Activities
The first step in the workshop is to clearly define
a:·d bound the problem. Bounding makes the modelling problem
more explicit, thereby making it easier to decompose the
system into manageable components or subsystems. In bounding,
development actions (alternate controls available to management as
well as development strategies) and indicators (those measures used
by management in evaluating system performance in response to
[
r I L~
[
[
r-j -
L_;
'0
[
c
[
'Tl bJ
[
6
[
[
[
[
[
[
[
- 5 -
various combinations of actions) are generated. The model
embodies the biophysical rules required to transform the
actions into indicator time streams. Bounding also involves
defining the spatial extent and resolution.required to
adequately represent the system, and specification of the
temporal extent or time horizon and an appropriate time
step.
The final bounding exercise of the workshop is called
"looking outward". It focuses attention on the subsystems
_defined by the actions and indicators and those variables
required by each subsystem from the other subsystems. In
looking outward, the standard question of analysis is recast.
Instead of asking "what can you provide to the other subsystems
from subsystem X?", the question "what do you need to know
about all other subsystems in order to predict how subsystem X
will behave?" is asked. This question demands a more dynamic
view and forces one to describe a particular subsystem in the
context of the entire system. The looking out~ard exercise
generates, for each subsystem, a list of "inputs" it needs
from the other subsystems and a list of "outputs" it must
provide to the other subsystems.
The second step of the workshop is submodel construction.
The workshop and each subgroup develops submodels for one of
the subsystems. One workshop facilitator works within each
subgroup and acts as the submodel programmer. The submodel
must be able to generate the output variables required by
other submodels and the appropriate indicator variables
identified earlier.
The final step of the workshop is to put each of the
submodels into the computer and link them into the system
model. The system model is run under a variety of development
scenarios to explore the consequences of various actions and
hypotheses about system structure. The principal objective
-. 6 -
of this exercise in an initial workshop is to point out model
deficiencies and identify areas requiring better understanding
and information.
1.3.2 Beyond the Workshop
The first workshop can be followed by a period of
independent work on identified research needs by collaborating
individuals which will lead to a second workshop and possibly
subsequent ones in a phased sequence. Early in the sequence,
workshops concentrate on technical issues, but later, they
focus more and more on communication to policy advisors and
the affected constituencies. The emphasis on communication
enables an effective and logical move to implementation,
either in a pilot project or a full-scale program.
Throughout the workshop sequence, the simulation model
is an expression and synthesis of new information and the
changing mental models of scientists, managers and policy
makers. The involvement and interaction of these groups
means that learning becomes as much a product as does problem
solving.
[
[
'l
L
r··
'j
~ -
[
[
[
[
B
L
L
[
[
E
[
[
[
[
L
[
c
E
u
l
c
[
6
l
u
l
- 7 -
2.0 BOUNDING
All systems are hierarchial in nature; each is comprised
of smaller parts, and is, in turn, embedded in, or part of a
larger system. The most critical decisions that are made in
planning research and analysis are the choices of components to
be explicitly addressed. The same is true for modelling.
Within simulation modelling workshops, these choices
are made during an exercise called bounding. Bounding forces
the participants in the workshop to define lists of actions
and indicators and place them in an appropriate spatial and
temporal framework. Once accomplished, the "looking outward"
exercise defines the key interrelationships between components
of the system under scrutiny.
2.1 Actions
Actions, in the context of modelling, are normally thought
of as human intervention into the environment. With regard to
the proposed developments on the Susitna, four major categories
of actions (Table 2.1) were identified for inclusion into the
model. The first relates to the construction and operation of
reservoirs; the second; relates to recreational development, use,
and control; the third relates to development other than
hydroelectric; and the fourth corresponds to mitigation options.
2.2 Indicators
Indicators are th~se quantities which are used to
evaluate the performance or health of a system in response to
the defined actions. The set of indicators (Table 2.2)
identified by participants in the workshops are primarily
related to wildlife populations and wildlife habitat measures,
although instream flows and indicators of recreational use are
included.
- 8 -
Table 2.1: Actions identified at workshop.
I. Reservoirs
a. Construction
· roads
· borrow pits
• transmission lines
• camp sites
• village sites
river bed mining
· reservoir clearing
• air strip construction
• aircraft use
· staging areas
b. Operation
• operating rule curves
II. Recreation/Access
III. General
• reservoir recreational ~evelopment (access and
facilities)
· recreational use (back packing, hunting, fishing)
• increased traffic on existing roads/railroads
• changes in land use patterns (mining, oil and
gas development)
• increased population in surrounding communities
IV. Mitigation
• habitat enhancement
· controlled burn
• replacement lands
• vegetation crushing
• flow regulation for fish and wildlife
• fire protection
· control of access
• hunting/fishing regulation
• scheduling of construction activities
• siting of roads
• reclamation/revegetation
[
[
[
[
[
c
c
[
c
L
\ '
L
[
[
r
[
[
[
[
[
[
[
[
G
~
[
L
[
G
[
b
L
- 9 -
Table 2.2: Indicators identified at workshop.
Hydrology
· instream flows
Vegetation
· acres of selected vegetation types
Wildlife
· populations of: moose raptors
black bear beaver
brown bear marten
wolves birds
• carrying capacity for the above populations
• numbers of animals harvested by hunters
· habitat quality
Recreation
· number of user days
• non-consumptive uses of wildlife
-10 -
The predicted changes in indicators are used to help
determine the impacts of the actions over time, and. in turn,
evaluate the quantity, quality, and timing of mitigative actions.
2.3 Spatial Considerations
Defining the spatial extent and reoslution of any
research or analysis is a critical step. It determines the
level of detail arid places geographical limits on what is to
be considered. Simulation models require an unambiguous
definition of the spatial extent and resolution.
The spatial. extent of the model was guided by estimated
home ranges of brown bear and moose. An area corresponding to
all of a home range was included. With this criterion, the
Upper Susitna Basin, extended to include the Prairie Creek-
Stephan Lakes region, was chosen as the area for assessing
impacts upstream of the Devil Canyon Darn site. Within this
upstream area, the Watana and Devil Canyon impoundments are
considered separately and the remaining land is designated as
a third spatial unit (Figure 2.1). Downstream (Devil Canyon
Darn site to Cook Inlet), an area corresponding to moose horne
range was defined using estimates from Modafferi (1982).
Moose home range probably occurs in a band 60 km wide; 30 krn
on each side of the Susitna. The model simulates this band
as far downstream as Talkeetna. The Susi tna floodplain is
considered separately within the downstrearnarea. Areas down-
stream of Talkeetna were not included because the present and
future hydrologic regime there, and its influence on vegetation
dynamics, was considered too complex to construct an adequate
predictive model.
Therefore, there are five spatial areas in the model:
a) the Watana impoundment;
b) the Devil Canyon impoundment;
[
r~
L
u
[
c
L
L
r--"'!
l L ·'''I
UPPER SUSITN
RIVER BASIN
0 10 l 20
Kilom erG
·--,, r--:-1 I. I ~
20
30
Figure 2.la: Upper Susitna Basin showing the Devil Canyon and Watana impoundments (shaded
area) .
___,
' )
0
0
-12 -
10 20
Kilometers
I
20
N
I 30
COOK INLET
Figure 2.lb: Lower Susitna Basin showing Devil Canyon to
Talkeetna riparian zone (shaded area) designated
for the model.
r
[
[j
L
\ -
L.i
L
[
b
[
c
c
E
-13 -
c) the remainder of the Susitna Basin upstream of Gold
Creek;
d) the floodplain from Devil Canyon Dam to Talkeetna;
and
e) the remaining land in a 60 km strip from Devil Canyon
Dam to Talkeetna.
Within each of the spatial areas, fourteen vegetation
types (Table 2. 3 )· were defined.
2.4 Temporal Considerations
The choice of the temporal resolution or time step for
the model is always problematic because of widely different
time scales of important processes. Many biological processes
depend on water levels at critical times throughout the year
requiring monthly, and sometimes daily, water level estimates.
However, wildlife and waterfowl populations do not change
substantially from one day to the next making daily population
estimates.unnecessary. These considerations, combined with the
necessity of representing much slower successional processes,
led to a mixed temporal structure. Average and peak flows are
available monthly from hydrology. All other submodels have a
one year time step but may implicitly include seasonal dynamics
when needed. A time horizon of 50 -80 years was chosen (to
capture the successional effects).
2.5 Submodel Definition
The breakdown of the system into component subsystems
is reflected in the breakdown of the simulation model into
the submodels:
-14 -
Table 2. 3: Fourteen vegetation types associated with the
spatial areas.
Conifer forest
· woodland
· open
Deciduous and Mixed Forest
Tundra
Tall shrub -alder
Medium shrub
Low shrub
· birch
• willow
· mixed
Unvegetated
• water
· rock/snow/ice
Disturbed
· temporary
· permanent
Pioneer
[
[
[
·c
['
u
t
[
L
r u
D
[
[
L
L
[
f
f
F •/
[
[
[
[
G
[
~
~
l
E
L
t
[
t
[
-15 -
a) physical processes/development/recreation;
b) vegetation;
c) furbearers/birds;
d) moose; and
e) bears.
The major components of each submodel (Table 2.4) were decided
upon through discussion by workshop participants.
2.6 Looking Outward
The purpose of "looking outward" is to define the pieces
of information that a particular subsystem requires from all
other subsystems to predict its dynamic behavior. This is a
qualitatively different question than the tradition.al one which
generates lists of factors which affect a particular component
of a system. The product of "looking outward" is an interaction
matrix, with columns specifying what information a subsystem
requires from each of the other subsystems (Table 2.5). The
diagonals are blank because they represent the internal dynamics
of each subsystem.
Each piece of information listed in the matrix represents
a specific hypothesis about system behavior. For example, the
furbearers/birds submodel requires information on the length of
sloughs and side channels that maintain at least .5 m of ice-
free water throughout the winter from the physical processes/
development submodel. The underlying hypothesis is that ~his
represents potential overwintering habitat for beavers.
-16 -
Table 2.4: Submodel components decided on by workshop
participants. ·
1. Physical Processes/Development/Recreation:
• flows
• stages
• ice processes
• reservoir elevations
• aquatic furbearer habitat
hydroelectric development scenarios
• other development scenarios
• recreational use
• recreational development
2. Vegetation:
• areal extent of vegetation types . browse production
• berry production . ecological succession
3. Furbearers/Birds:
. beavers . marten . golden eagles . passerine birds
4. Moose:
. moose . moose habitat
5. Bears:
. bears . bear habitat
[
J'
I -
[
c
c
[
L
CT J •::---1 ,~l
' ' 1
Table 2.5: Looking Outward Matrix. Major information transfers between submodels.
PHYSICAL
PROCESSES/
DEVELOPMENT/
RECREATION VEGETATION FURBEARERS/BIRDS MOOSE BEAP.S
-location & areas -length (km) of -snow depth (ft) -recreational use
(ha) of develop-slough, side (days)
rrent activities channel, & mainstem
-minirrum water habitat with >
• 5 m ice-free PHYSICAL surface area (ha) water PROCESSES/ in floodplain
DEVELOPMENT/ during growing -reservoir
RECREATION season elevations (ft)
-area (ha) of ice -human disturbance
scouring in down-
stream floodplain
-areas of -areas of -production of
vegetation types vegetation types berries (kg/ha)
(ha) (ha) -area (ha) of
-pro{X)rtion of -standing crop berries suitable
slough, side (kg/ha) & areas for bear food
channel, & mainstem of: -areas of VEGETATION habitats that have Paper Birch vegetation types balsam {X)plar or
birch Lowbush Cran-(ha)
berry
Balsam Poplar
Willow Shrub
Aspen
FORBEARERS/ -number of beaver
BIRDS colonies
-consllllption
MOOSE (kg/ha) of browse
species by season
arrl type
-consumption -bear {X)pulation
BEARS (kg/ha) of forage (numbers)
species by season
and type
-18 -
3.0 SUBMODEL DESCRIPTIONS
The five submodels, described in this section, hydrology/
development/recreation, vegetation, furbearers/birds, moose, and
bear, are an interdisciplinary representation of the terrestrial
biophysical processes of the susitna Basin. In some cases, the
relationships described are based on good scientific evidence: in
other cases, they are simply crude hypotheses or educated guesses.
These models require critique and refinement before a reasonable
representation of important terrestrial processes is achieved.
3.1 Physical Processes/Development/Recreation
The Susitna hydroelectric development will impact the
terrestrial environment directly through disturbance and vegetation
loss on lands needed for project facilities, and indirectly through
alteration of the hydrologic and ice regimes ?f the Susitna River.
Another possible and perhaps major impact on the terrestrial
environment will occur through increased recreational opportunities
that may result from increased access and the development of
recreational facilities at or near the reservoir. Also, while
development associated directly with the hydroelectric project
may have a substantial impact and is the primary focus of this
proj ec.t, it is important to place this development in the context
of development activities that are indirectly related to the
project, such as mining, oil and gas exploration and production,
and new recreational facilities.
3.1.1 Physical Processes
Almost all the physical processes considered in the model
are related to the flow regime or climate or the interaction of
both factors. Currently, the model simulates the flow regime at
three stations (Gold Creek, Sunshine, and Susitna) for four
different cases:
[
[
L
[
[
[
[
[
[
[
g
l
-19 -
a) preproject flows;
b) Case A, which corresponds to optimum power generation;
c)· Case C, which corresponds to case used in the PERC
license application; and
d) Case D, which corresponds to the ·best development for
meeting instream flow targets.
The post project cases A, C, and D can be used assuming
Watana operating alone or with both Devil Canyon and Watana in
place. Thus, the model uses one of seven possible flow regimes
downstream of Devil Canyon. The flows are based on historical
preproject flow data and estimates provided by Acres American
Ltd. (Dave Crawford, pers. comm.) for post project flows under
different operating conditions. Thirty-two years of data for each
case are used and repeated. Figure 3.1 is a comparison among the
four cases using the data used for simulation year 12. Average
monthly flow is usually a poor indicator of the stress on an
ecosystem and, in many cases, extreme flows (minima and maxima)
are more important. The model makes daily and 3 day minimum and
maximum flow estimates using data supplied by R & M Consultants
(pers. comm.).
3.1.1.1 Reservoir Elevations
The operation of the dams causes the reservoirs to vary
throughout the year as seen for the simulation year 12 in Figure
3.2. The model provides the reservoir elevations for Watana
Reservoir based on monthly sstimates provided by Acres American.
3.1.1.2 Stage
The calculation of stage is based on stage~discharge
rating curves like the ones shown for Gold Creek (Figure 3.3).
36
30 .. 24 .... u
0 z IB <t
(I)
::>
0 12 :r:
I-
6
0
36
30
• ....
u 24
0 z
<( IB
(I)
::>
0 12 :r: ....
6
0
PRE PROJECT 36 CASE A
30
411 .... 24 u
0 z
<t. 18
(I)
::>
0 12 :r: ....
6
0
OCT. DEC. FEB. APR. JUNE AUG. OCT. DEC. FE'B. APR. JUNE
TIME TIME
CASE C 36 CASE 0
30
• .... 24 u
0 z 18 <t
(I)
::>
0 12 :r: ....
6
0
OCT. DEC. FEB. APR. JUNE AUG. OCT. DEC. FEB. APR. JUNE
TIME TIME
Figure 3.1: Gold Creek flows for preproject, case A, case C, and case D, assuming
both dams operating.
(~
'~ ' ' " 'j
AUG.
N
0
AUG.
-------• I
' '
[
[
[
[
[
u
[
[
c
t
[
b
r
-• • -
11.1
t!)
<t ....
(I)
2190
2170
.... 2150
11.1 w
~
2130
2110
-21 -
OCT. DEC. FEB. APR. JUNE AUG.
TIME
Figure 3.2: Watana Reservoir elevations
throughout the year.
1!5
12
9
6
3
0
0 10 20 30 40 ~0
DISCHARGE ( 000 etc )
Figure 3.3: Stage-discharge ratirig curves for
Gold Creek Station. Open water case
based on USGS data gathered since
October 1, 1967. Ice case estimated
from data in the FERC license application
(Exhibit E, Chapter 2). The dotted line
indicates uncertainty for the given
discharge ranges.
-22 -
Both the open water and ice covered curves shown are used by the
model. The open water case is based on USGS data gathered since
October, 1967; the ice cover case is estimated from the FERC
license application (Exhibit E, Chapter, Figure E.2.185).
3.1.1.3 Water Surface Area in the Downstream Floodplain
(Devil Canyon to Susitna-Chulitna Confluence)
Total area of water surface between Devil Canyon and the
Susitna-Chulitna confluence was estimated at various flow levels
using the U.S. Corps of Engineers HEC-2 runs (dated February 2,
1982) (R & M Consultants, pers. comm.). Figures were computed by
using the average width of adjacent cross sections and multiplying
by the length between them. The steep slope around a.flow of
20,000 cfs shown in Figure 3.4 exists due to the addition of
sloughs to the flow regime of that level.
Knowledge of the water surface area and an estimate of the
total area in the floodplain allows the vegetation submodel to
estimate the total surface area exposed in the floodplain.
3.1.1.4 Ice Dynamics
The ice dynamics in the downstream area are considered to
be the critical determinants of the suitability of fish and
furbearer habitat and vegetation succession. The introduction
of the project is expected to change the timing of freeze-up,
ice staging, ice scouring, the timing of break-up, and create
year round open water in part of the downstreamarea (Devil Canyon
to Talkeetna).
3.1.1.4.1 Formation of Ice Cover
Under preproject conditions, the model assumes that the
entire downstream reach (Devil Canyon to Talkeetna) is completely
covered with ice by mid-December. Under post project conditions,
)~
L~
[
r
['
-.,.,/
('
1 ... •
[
F'
[,
[3
b
c
[
c
t='
c;
Q
[
[
L
L
[
[
F
[
[
[
[
[
E
g
[
-23 -
3000
2500
c:r ..c
<( 2000 l.lJ
0:::
<t
LLI u 1500 ~
0:::
::J
(/)
0::: 1000
LLI ...
~
500
0 10 20 30 40 50
DISCHARGE ( 1000 cfs )
Figure 3.4: Water surface area in the downstream floodplain
(Devil Canyon to Susitna-Chulitna confluence)
as a function of discharge measured at Gold Creek
Station.
-24 -
an ice front is formed by mid-January delineating the ice
covered and open water stretches of the reach.· If Watana
alone is operating, this front is formed somewhere between
Portage Creek and Sherman; if both projects are operating, the
front is formed somewhere between Talkeetna and Sherman. The
exact position of the front is dependent on climatic conditions
simulated using a uniform random number.
3.1.1.4.2 Ice Staging
The formation of ice cover causes significant ice staging,
that is, a significant increase in stage over what would be
present under open water condition. This condition, illustrated
by Figure 3.3, has implications for maintenance of groundwater
upwelling in sloughs and for vegetation damage caused by the ice
as the river stages. As the river stages, it lifts the ice
already in place and tears or scours the vegetation along the
edges of the channel. To make a rough estimate of the area
affected, the model calculates the difference between the water
[
[
r
-, --
~'
D
[
[
surface area assuming open water ·(Figure 3.4) and the area _
covered at maximum ice cover (Figure 3.5). This area is considered [:
to be area subjected to potential vegetation damage due to ice
during freeze-up.
3.1.1.4.3 Break-up
Prior to the project, the model assumes that break-up
occurs in early May and more often than not is triggered by high
inflows from tributary streams. After the projects are operating,
break-up will occur in mid-April and more often than not the ice
cover will melt in place before the high inflows from tributary
streams occur. As a result, there will be significantly less
ice scouring after the project. To simulate the break-up
processes and the occurrence of ice scouring, the model
stochastically generates the timing (Figure 3.6) of melting and
high inflow from tributary streams. If the ice melts in place
before the high inflows occur, no ice scouring occurs; if high
''
[
[
[
L
4000
3000
(/)
LLI
a::
< 2000 ....
0
IU
:r:
1000
-25 -
--------------------------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
0 10 20 30 40
DISCHARGE ( 000 c f s )
Figure 3.5: Hypothetical relationship of area of
maximum ice cover as a function of
discharge. The dotted lines indicate
uncertainty for the given discharge
ranges.
-26 -
high tributary inflow
ice completely melted
Day90
March 31
Day90
March 31
120
April 30
a) PRE PROJECT
high tributary inflow
1----+-------tl ice completely melted
120
April30
b) POST PROJECT
Figure 3. 6: Simulated timing of events affe.cting break-up.
150
May 30
150
May30
c
[
[
[
[
[
[
[
[
[
[
c
[
6
~
~
c
u
[
h
-27 -
inflows occur before the ice has completely melted, then the area
subject to ice scouring is calculated using the water surface
area-discharge relatio"nship for the open water case (Figure 3.4).
3.1.1.5 Flood Events
The model calculates the area flooded based on the water
surface area curve (Figure 3.4) at various times throughout the
year. In particular, the maximum flooded area is calculated and
usually occurs in June, July, or August. The minimum flood area
during the growing season is calculated and provided to the
vegetation submodel.
3.1.1.6 Downstream Effects
The processes represented in the physical submodel are
important because of their effects downstream of Devil Canyon.
In the reach extending as far as Talkeetna, the model is currently
concerned with how changes in the hydrologic regime w·ill effect
beaver overwintering habitat and vegetation succession.
3.1.1.6.1 Beaver Overwintering Habitat
Side channels and sloughs that retain greater than .5 m
in depth of unfrozen water throughout the winter provide potential
overwintering habitat for beaver. In the major area of concern,
downstream of Devil Canyon Dam to Talkeetna, the amount of this
habitat is directly related to water level (stage) and ice
thickness. The stage depends on flow (Section 3.1.1.2), and the
ice thickness depends on flow and the severity of the winter. In
the model, the effect of the severity of winter was simulated as
a random process that increased or decreased the amount of
habitat for beaver.
Before discussing the relationships used to estimate the
amount of potential overwintering habitat for beaver, a careful
definition of mainstem, side channel, and side slough habitat is
necessary. The following definitions are adopted from Trihey
(November, 1982).
-28 -
Mainstem habitat consists_ of those portions of the Susitna
River which normally convey streamflow throughout the year. Both
single and multiple channel reaches are included in this habitat
category. In general, this habitat category is characterized by
high-velocity streamflows and well armored streambeds. Substrates
generally consist of boulder and cobble size materials with
interstitial spaces filled with a grout-like mixture of small
gravels and glacial sand. Suspended sediment concentrations and
turbidity are high from late May through early October due to the
influence of glacial melt water. Streamflows recede, and the
water appreciably clears in the early to mid fall before an ice
cover forms on the river in late November or December. Groundwater
and tributary inflow appear to be inconsequential contributors to
the overall characteristics of this habitat category. Seasonal
temperatures of the mainstem river respond primarily to air
temperature and solar radiation. Mainstem surface water appears
to establish mainstem intragravel water temperatures.
Side channel habitat consists of those portions of the
Susi tna River which normally convey streamflow during the open
water season but which become appreciably dewatered during periods
of low flow. The controlling streambed e~evations at the upstream
entrance to the side channels are less than the water surface
elevations of the mean monthly flows for June, July and August.
Side channel habitats are characterized by shallower depths,
lower velocities and smaller streambed materials than mainstem
habitats. In general, the streamflow, sediment, and thermal
regimes of the side channel habitats reflect attenuated mainstem
conditions. Tributary and groundwater inflow may prevent some
side channel habitats from becoming completely dewatered when
mains ·em flows recede. However, the presence of these limited
inflows could conceivably not be considered a critical component
of side channel habitat. A winter ice cover, similar to that
which forms on the mainstem, generally exists in the side channels.
[
f
..
...
L
D
[
E
c
c
[
L
L
L
[
[
[
[
[
I'
LJ
[
c
[
c
6
c
t
G
[
L
L
L
l
-29 -
Side slough habitats are found in spring-fed perched
overflow channels which only convey glacial meltwater from the
mainstem during median summer and high flow periods. At
intermediate and low flow periods, the side sloughs convey clear
water from small tributaries and/or upwelling groundwater. The
controlling streambed/streambank elevations at the upstream end
of the side sloughs are slightly less than the water surface
elevations of the mean monthly flows for June, July, and August.'
Side sloughs generally exist along the edge of the floodplain,
separated from the mainstem by well-vegetated bars. An exposed
alluvial berm often separates the head of the slough from mainstem
or side channel flows where as the water surface elevation of the
river generally causes a backwater to extend well up into the
slough from its lower end. It is important to note that, even
though a substantial backwater exists, _hydraulically the sloughs
function very much like small stream systems. Several hundred
feet of the slough channel often conveys water independent of
mainstem backwater effects.
Except when the discharge in the maintstem river is
sufficient to have overtopped the upper end of the slough, surface
water temperaturesinthe side sloughs appear to be independent of
those in the mainstem river. surface water temperatures in the
side sloughs during summer months are principally a function of
air temperature, solar radiation, and the temperature of the local
runoff. During winter months, surface water temperatures are
strongly influenced by upwelling groundwater. The large deposits
of alluvium through which the upwelling water flows appear to
act as a buffer or thermal reservoir, attenuating summer temperatures
and providing very stable winter temperatures.
The model assumes that all side slough habitat that retains
at least .5 m of ice free water throughout the winter can support
beavers. The side channels are only considered suitable if the
velocity is low enough (less than 4.4 ft/sec) in addition to
maintaining sufficient depth of ice free water.
-30 -
Apparently, the amount of ice free water in sloughs and
side channels is related to the amount of warm groundwater inflow.
The groundwater inflow is related directly to the hydraulic head
between the mainstem and the sloughs and side channels. The
hydraulic head is physically dependent on the mainstem stage ..
Under present conditions, the model assumes that the increased
stage associated with a winter ice cover makes it possible for
the same hydraulic head to exist between the mainstem and adjacent
side slough habitats during the winter as exists during late
summer.
In the model, the amount of suitable overwintering habitat
is functionally related to stage. In the case where the reach
is ice covered, the ice staging curve is used; in the case where
there is open water, the open water curve is used (Figure 3.4).
The relationship between the amount of habitat and the stage
(Figure 3.7) .saturates at high stages under the assumption that
increased groundwater inflow does not make a given area any more
desirable, although it may make areas that were formerly unsuitable,
desirable habitat.
Under current conditions, the entire reach becomes ice
covered during the winter; with the project, anice front will form
far downstream from Devil Canyon. The exact location depends on
the scale of the project and the severity of the winter. In any
case, only a portion of the reach will be ice covered. Because
of this, the model calculates the available habitat for the ice
covered portion of the reach and for the open water portion. In
addition, the model makes separate calculations for sloughs,
side channels, and mainstem habitat. The slough and side channel
habitat numbers are aggregated before being provided to the beaver
submodel.
[
[
[
[
r
l
c
[
fi
[
[
[
[
[
r' .3
['
··'
[
c
c
[
E
E
t
~
[
h
L
L
t:
Hmax.
... ~
ID
c:( z
~ z
~
11.1 ... z
:r: .5 Hmax.
a:
11.1 > 0
..J
c:( ... z
11.1 ...
0
Q.
-31 -
0 3 6 9 12
STAGE (feet) AT GOLD CREEK
Figure 3.7: Potential overwintering habitat as a
function of stage. Hmax represents
the maximum for a given habitat.
15
-32 -
3.1.1.6.2 Vegetation Succession
The regular flooding and ice scouring in the downstream
reach provides a regular stress to the vegetation types that
occur at lower elevations relative to the· water surface elevation.
Pr~vious sections (3.1.1.4.2, 3.1.1.4.3, 3.1.1.5) have discussed
how the extent of ice scouring and flooding is determined in the
model. The description of the vegetation model will discuss how
these processes affect succession.
3.1.1.7 Snow
Snowfall is simply generated stochastically because there
was insufficient conceptual understanding of snow dynamics. This
is a major model deficiency because snow levels can seriously
affect utilization of moose winter range.
3.1.2 Hydroelectric Development Activities
The timing, location, and areas affected by project
activities considered by the model are listed in Table 3.1.
The areal values in Table 3.1 are from the PERC license application,
Exhibit E, Chapter 3; Tables E.3.80, E.3.83, E.3.84, and E.3.85.
At the appropriate time and location, the model alters the
vegetation classification for the area associated with the site
for the activity to the "disturbed" category (c.f. Table 2.3).
The site may be permanently disturbed or may be reclaimed or
revegetated at a later date.
3.1.3 Other Land Use Activities
There are a number of current and potential uses for the
land with the geographic area being considered by the model.
These include agriculture, forestry, recreation, settlement, coal
development, mining development, oil and gas development, and
transportation. There appears to be little potential for
agriculture, coal development, and oil and gas development
[
[
r·
L,
r L
[
c
[
b
[
[
L
[
L
Table 3.1: Hydroelectric development project actions.
ACTION AREA AFFECTED TIME LOCATION
1. TRANSMISSION CORRIDORS (clearing)
• Watana to Devil canyon 380 hectares 1989-1990 watana to Devil canyon
• Devil Canyon to Intertie 132 hectares 1989-1990 Devil canyon to Chulitna Pass/Indian River
2. CAMPS
• Watana 63 hectares 1985-1994 Between Tsusena & Deadman Creeks
Reclamation starts 1994
(No pennanent
structures)
• Devil canyon 36 hectares 1994-2002 South of Susitna River on plateau opposite
Reclamation starts 2002 Portage Creek
(No pennanent
structures)
3. VILLAGES w
• Watana (pennanent) 70 hectares 1986-Between Watana camp site and Tsusena bJ
creek, surrounding Small lake
• Devil Canyon (no pennanent 39 hectares 1996-2002 South of Susitna River on plateau opposite
buildings) Portage Creek
4. RESERVOIR CLEARING
• Watana 3405 hectares 1989 watana impoundment
3642 hectares 1990 Watana impoundment
3642 hectares 1991 watana impoundment
4047 hectares 1992 Watana inpoundrrent
• Devil Canyon 1000 hectares 1999 Devil canyon impoundment
1196 hectares 2000 Devil canyon impoundment
1000 hectares 2001 Devil Canyon impoundment
ACTION AREA AFFECTED
5. STAGING AREAS
• Access Plan #13 (north) 61 hectares
• Access Plan #16 (south) 61 hectares
61 hectares
• Access Plan #17 (Denali) 61 hectares
• Access Road (FERC) 61 hectares
6. CDNTRACI'OR OORK AREAS
• Watana 77 hectares
146 hectares
77 hectares ·
• Devil Canyon (including 61 hectares
hatching plant) 61 hectares
61 hectares
12 hectares
7. CDNI'AINMENT STRUCIURES
• Watana 14 hectares
36 hectares
26 hectares
3 hectares
10 hectares
4 hectares
• Devil canyon 1 hectare
5 hectares
13 hectares
8. AIRSTRIPS
• Watana 17 hectares
TIME
1985-2002
1985-2002
1985-2002
1385-2002
1994-2002
1985-1994
1986-1994
1987-1994
1994-2002
1995-2002
1996-2002
1997-2002
1986-
1987-
1988-
1989-
1990-
1991-
1996-
1997-
1998-
1585-
LOCATION
Hurricane
Hurricane
Gold Creek
cantwell
Gold Creek
Between Watana Camp and Dam Site
Between Devil canyon Camp and dam site
w
ol»
Watana Dam site including floodplain
Devil Canyon Dam site including floodplain
Adjacent to Watana camp
-I '··
.~ -, -'1
J
,..._._,
' J
ACTION AREA AFFECTED
9. ACCESS ROAD (clearing) 192 hectares
189 hectares
29 hectares
10. BORROW AREAS WATANA
·A 333 hectares
. D 287 hectares
. E 180 hectares
• F 280 hectares
• H 489 hectares
• I 34 hectares
• Devil Canyon K 148 hectares
TIME
COnstruction: 1985
Intensive use: 1985-2002
Construction: 1991-1993
Intensive use: 1994-2002
COnstruction: 1991-1993
Intensive use: 1994-2002
1985-1993
1985-1993
1985-1993
1985-1993
1985-1993
1985-1993
1995-1999
.~....,
' ' '
LOCATION
Denali Hwy to Watana
Denali Hwy to Watana
Watana to Devil Canyon
watana to Devil Canyon
1:-J
Devil Canyon to Gold Creek
Devil Canyon to Gold Creek
w
U1
-36 -
although lease sales have been proposed. Forestry and settlement
may increase in the downstream portion of the Susitna. Perhaps
the greatest potential is for increased mineral development and
recreational opportunities.
Currently, the model only considers additional lands needed
for settlement, mining development, and recreational development.
Present use of the area is low, although substantial growth is
expected if the Susitna project goes ahead. Estimates of current
use (Table 3.2) are unsubstantiated, and must be revised when
better estimates appear.
3.1.4 Disturbance to Wildlife
Associated with project activities and other land use
activities is disturbance to wildlife as a result of the presence
of humans. The model keeps track of three major classes of
di"3turbance:
a) disturbance from recreational use;
b) disturbance due to the influx of construction workers;
and
c) disturbance from vehicle and aircraft movements.
The disturbance from construction workers and vehicle traffic
is provided in Table 3.3. Recreational disturbance is based on
the use information from the FERC license application, Exhibit
E, Chapter 7.
3.1.4.1 Recreational Use
In the model, recreational use is divided into eight
categories consisting of (FERC license application, Exhibit E,
Chapter 7): big game hunting, waterfowl hunting, freshwater
fishing, developed camping, canoeing/kayaking, hiking, picnicking,
[
[
r
L"
[
L"
['"
f
u
[
[
[
[
[
[
[
[
[
L
[
[
[
[
r
[
c
c
[
[
[
-37 -
Table 3.2: Estimates of current land use and recreational use
in geographic area considered in the model.
Mining (hectares)
Settlement (hectares)
Recreational Use (use days)
Big Game Hunting
Waterf~wl Hunting
Freshwater Fishing
Developed Camping
Canoeing/Kayaking
Hiking
Picnicking:
Cross-country Skiing
Upper Susitna
Basin
10,000
2,021
Downstream
(Devil Canyon-Talkeetna)
14,000
6,064
Project Area
800
100
1500
4000
. 200
100
-38 -
Table 3.3: Disturbance ~ssociated with construction workers and
vehicle traffic. [
DISTURBANCE
Construction workers
Vehicle traffic
Big Game Harvests
Diversion Structures
-Blasting -
LOCATION
Watana Camp &
Construction Area
Devil Canyon Camp
& Construction
Area
To Watana
TIME
1983
84
85
86
87
88
89
90
91
92
93
94
95
1994
95
96
97
9A
99
2000
01
02
1985-1995
To Devil Canyon 1994-2002
Gold Creek to 1994-2002
Devil Canyon
Game Management
Unit #13
Present
Natana Dam site 1985-1987
Devil Canyon Dam 1995-1996
site
MAGNITUDE [
180
192
690
780
workers on si tef,
at one time
__ .'
1,140
1,500
1,680
2,070
1,920
1,500
780
360
48
60
240
480
750
990
workers on ~itep
at one tlme G
1,020
900
540
48 c
53 trucks per week r
each direction ~
92 trucks per week
each direction U
4 trains per week
each direction (if
Denali Route is
chosen)
[
caribou -750/year b
Moose -750/year
Brown Bear -100/year
Black Bear -60/yea1J
Unknown
Unknown
[
L
L
[
[
[
E
[
[
[
G
c
c
E
g
E
b
[
G
L
li
~
-39 -
and cross country skiing. Estimates of current recreational use
(Table 3.2) are based on FERC license application, Exhibit E,
Chapter 7 (1983). The reliability of these estimates is
questionable.. In particular, the estimate of big game hunting
appears to be grossly understated.
The model assumes that recreation demand will approximately
double by the year 2000 without the Susitna hydroelectric project.
If the project goes ahead and the proposed recreation plan is
adopted, recreation demand will be approximately sevenfold by the
year 2000. These projections are based on the FERC license
application, Exhibit E, Chapter 7, and are summarized in Table 3.4.
The model allows for a choice of·access routes (Table 3.1).
The choice of the access route will affect the amount and level
of vegetation impacted and may impact critical wildlife areas.
Another aspect is whether public access to the project area via
the new access road is desirable. The model allows for open or
restricted access.
3.2 Vegetation
The vegetation submodel is a set of rules for simulating
vegetation and land use processes in response to direct Susitna
development activities and indirect changes of the hydrologic
regime in the downstream floodplain. The model is based on a
land classification system in which areas in each land class are
updated annually in response to human activities and processes
of natural vegetation change. The Looking Outward .Matrix (Table
2.5) identifies the processes simulated by the vegetation
submodel in terms of information required by other submodels.
The information consists of area of various land classes for
each spatial unit, berry production in each land class, the
standing stock of potential browse for moose in each land class,
and a measure of the proportion of both main channel and sloughs
or side channels with associated vegetation preferred by beaver.
Table 3.4: Estimated recreation demand (adapted from FERC license application,
Exhibit E, Chapter 7).
Assumed 1980 Use
of the Project
Recreation Area,
User Days
Estimated 2000
Use of the
Project
Recreation Area
Without SUsi tna
Hydroelectric
G GAME BI
HUNr Il\K}
800
Project, User Days 1, 300
WATERFOWL FRESHWATER
HUNrll'K} FISHING
100 1,500
170 2,500
'DEVEIDPED CANOEII'K}/ X -couNTRY
CAMPil'K} KAYAKil'K} HIKil'K} PICNICKING SKIIl'K}
4,000 200 ----100
8,000 370 ----220
'IOTAL
6,700
12,540
oj:::..
Estimated 2000 o
Use of the
Project
Recreation Area
With SUsitna
Hydroelectric
Project
Proposed
Recreation Plan,
User Days
__,..,...,
I I ' [,
2,
2,
200 -4,800 -12,000 -
400 170 5,200 14,000
12,000 -12,000 -
100 14,000 14,000 350 43,520
~
' j
I
[
t~
r
[
f~
[
[~
6
[
[
~
E
E
L
c
b
[~
b
l
-41 -
The only actions for which the vegetation submodel is directly
responsible are controlled burning and vegetation crushing.
3.2.1 Structure
The sequence of calculations for the vegetation submodel
is outlined in Figure 3.8. Given the 50 -80 year time horizon
for model runs, long-term successional dynamics in upland areas
were not simulated in the absence of development activities. An
attempt was.made to simulate shorter-term riparian vegetation
dynamics despite a limited understanding of riparian succession
and the effects of ice processes.
3.2.2 Classification System
The classification system was developed from work described
in the Plant Ecology Phase I F~nal Report (McKendrick et al.,
1982). The classification system in the model distinguishes 14
classes of land, primarily defined on the basis of vegetation
type, in each spatial unit (see Section 2.3). Initial conditions
(Table 3.5) were estimated for all spatial-units, except the one
representing moose range in the area downstream from Devil Canyon.
The impoundment areas estimated are slightly larger than the areas
that would be cleared if the development proceeds. In addition to
the spatial units described above, total areas in the upper
Susitna Basin were calculated as the sum of the two impoundment
areas and the rest of the upper Susitna unit.
The land classification was expanded slightly from
McKendrick et al. for this project. A medium shrub class was
defined in order to calculate bird indicator variables. Two
disturbed classes were defined to represent land disturbed by
construction of permanent facilities or by temporary activities
which would be followed by artificial or natural revegetation.
A pioneer class was added to represent the initial stages of
herbaceous vegetation in riparian areas and following temporary
human disturbance.
LAND DEMANDS
FOR VEGETATION~
MANIPULATION
ACTION
-42 -
MAKE DIRECT
TRANSFERS AMONG
LAND CLASSES TO
MEET DEMANDS
CALCULATE REVEGETATION
TRANSFERS ON
DEVELOPED LAND
CALCULATE RIPARIAN
SUCCESSION TRANSFERS
CALCULATE BROWSE AND
BERRY PRODUCTION IN
EACH LAND CLASS
CALCULATE PROPORTION
OF RIPARIAN CHANNELS
WITH ASSOCIATED BEAVER-
PREFERRED VEGETATION
CALCULATE TOTALS
FOR UPPER BASIN
LAND DEMANDS FOR
RESERVOIRS, FACILITIES,
~ BORROW PITS,
TRANSMISSION CORRIDORS ,
AND ROADS FROM
DEVELOPMENT SUBMODEL
Figure 3.8: Calculation sequence for the vegetation submodel.
[
[
[
[
c
c
Table 3.5: Initial conditions for vegetation types. All values are in hectares.
REST RIPARIAN ZONE
WATANA DEVIL CANYON OF UPPER TALKEETNA TO
LAND CLASS IMPOUNDMENT AREA IMPOUNDMENT AREA SUSITNA BASIN DEVIL CANYON
Coniferous Forest-
woodland and closed 4275 153 183963 0
Coniferous Forest-
open 3633 633 114607 0
Deciduous and Mixed Forest 2911 1516 36218 3500
Tundra 84 11 394590 0
Tall Shrub 537 3 128495 300
Medium Shrub 44 5 3306 0
Low Birch Shrub 400 44 29750 0
,j:>.
Low Willow Shrub 66 14 10565 0 w
Low Mixed Shrub 673 4 470784 400
Unvegetated-water 2060 813 36967 600
Unvegetated-rock, snow, ice 60 15 203478 0
Disturbed-temporary 0 0 0 0
Disturbed-permanent 1 1 1 0
Pioneer 1 1 1 200
-44 -
3.2.3 Development Activities
The vegetation submodel responds to demands for land
associated with reservoir development, road construction,
transmission corridor construction, borrow pits, and construction
of permanent facilities. These demands, calculated each year by
the development submodel, result in transfers of land among
various land classes within the respective spatial units.
[
[
[
[
. ["
Generally, the development land demands in a given spatial unit f'
are met from the various land classes in the spatial unit according ·
to their relative proportions in that unit. However, land demands
for roads are specified as proportions of various classes
associated with specific routes.
Clearing for reservoirs is simulated by subtracting the
appropriate proportions of the reservoir land demand from the
respective land classes and adding the total to the inundated
land class.
The development demand for facilities is met by
transferring land to the permanently disturbed class.
Access road construction is simulated by taking land from
various land classes according to development submodel demand and
route-specific land class proportions. Land for roads is added
to.the low mixed shrub class under the assumption that the
biggest areal change is in the associated right-of-way.
The demand for transmission corridors is met by initially
transferring land to the low mixed shrub class. This land is
then subject to succession to the medium shrub class at an
annual proportional rate of 20%.
Borrow pits are developed by transferring land to the
temporarily disturbed class. User specified fractions of the
borrow pit land are then subject to either inundation or
revegetation. Inundated borrow pits are transferred to the
[
c
fj
[
[
c
[
c
[
L
f"
L
,.-,
i
[
[
[
[
r,
[
c
[
c
G
E
E
r:
[
c
L
c
~
-45 -
water class, while revegetation of borrow pits consists of an
initial transfer to the pioneer land class followed by a
transition to low mixed shrub at a proportional rate of 10% per
year.
Finally, the action of vegetation manipulation (controlled
burning and crushing) transfers land from the deciduous and
mixed forest class to the low mixed shrub class. · This land is
then subject to succession to the medium mixed shrub class (at
a rate of 20% of the low mixed shrub class per year), followed
by transfer to the deciduous and mixed forest class (at a rate
of 7% of the medium shrub class per year) • The area of land
transferred by vegetation manipulation is provided as an action
to the model as,a whole, rather than as a value calculated by the
development submodel. This action is intended to roughly simulate
controlled burning and vegetation crushing which were discussed
as possible mitigation measures designed to increase wildlife
habitat value. The land is transferred only from the deciduous
and mixed forest land class. It was felt that this would be the
preferred . land for vegetation manipulation because of relative
increase in habitat value resulting from converting this land
class to earlier successional stages.
3.2.4 Riparian Succession
Dynamics of vegetation in the riparian zone from Devil
Canyon to Talkeetna are represented as the net effect of two
opposing processes; natural succession and disturbance due to
erosion and ice processes.
represented in Figure 3.9.
(Figure 3~9) wer.e estimated
The successional sequence is
Annual transfers among land classes
from observed ages of individual
trees and shrubs within the various vegetation types.
The effects of ice processes on riparian vegetation are
poorly understood. However, an attempt was made to include
these effects in the model, primarily as a mechanism to help
identify what information and studies might be required to
-46 -
LOW MIXED TALL
PIONEER 10% SHRUB 20% SHRUB
200 ha 400 ha 300 ha
7%
. DECIDUOUS
AND
MIXED FOREST
3500 ha
Figure 3.9: Successional sequence in the Talkeetna to
Devil Canyon riparian zone. Numbers within
each compartment are the estimated initial
conditions. Numbers on the solid arrows
represent the annual percentage transfers of
land.
[
[
r
[
u
c
c
L
[
r' l
L;
L
L
r
[
[
[
l
""
__ j
[
[
-47 -
understand these effects sufficiently for mitigation planning.
It was assumed that the vegetation communities are arrayed along
an elevational gradient with pion~er vegetation occupying the
lowest portion of the gradient and deciduous and mixed forest
the highest. Based on this assumption and the surface area
covered by ice (estimated by the physical processes submodel},
the amount of each vegetation type scoured by ice is calculated.
The total amount of vegetation scoured is the area covered by
ice minus the area of the river channel. Because it is lowest
on the elevational gradient, pioneer vegetation is assumed to
be scoured first. If the area scoured is greater than the amount
of pioneer vegetation, then some low shrub is also scoured. If
the area scoured is greater than the amount of pioneer and low
shrub, then some tall shrub is also scoured, and so on. The effect
of scouring (i.e. the amount of vegetation conve,rted to pioneer)
depends on the vegetation type. Early successional stages are
assumed to be less resistant to scouring than later successional
stages at the same flow. However, later successional stages are
assumed to be scoured only during high flow events when the energy
of scouring is very great. The vegetation subgroup did not have
sufficient information to determine the net effect of resistance
to scouring/energy of scouring. However, they felt for the pre-
project situation, it was reasonable to assume the riparian
sucoesoional stagea were in appro.xlmdle equilibrium (i.e. no net
long-term changes in the amount of land in each vegetation type}.
The model parameters controlling ice process effects were therefore
adjusted until an approximate equilibrium was obtained.
The amount of scouring and the water level during the
growing season determines how much new pioneer vegetation becomes
established each year. If water levels are the same as last year,
~hen the new pioneer vegetation is that created by scouring.
If water levels are lower, new pioneer vegetation is that
created by scouring plus those additional areas in the ,river
channel exposed because of lower water. If water levels are
higher than last year, new pioneer is only the portion
created by scouring which remains above the higher water.
-48 -
If water levels are much higher, then there may be no new pioneer
vegetation established (even if scouring occurred) and some areas
of existing pioneer vegetation may be flooded long enough to
eliminate the vegetation.
3.2.5 Wildlife Habitat
The wildlife submodels required a measure of browse, a
measure of berry production, and an index of the suitability of
vegetation along channels in the riparian zone (for beaver) as
measures of habitat.
An estimate of potential browse (kg dry weight/ha) is
obtained for each land cTass by multiplying the relative cover
of the primary browse species in each of the land classes by
·the quantity (kg/ha) of browse (measured to the average point
of browse) associated with each species (Table 3.5). Random
variation (standard deviation of 10%) is applied to these
estimates to yield annual values. Annual berry production (kg
dry weight/ha) is calculated in a similar fashion by applying
the same random annual variation to an average production
estimate (Table 3.6) based on production of berry species and
their relative cover in the various land classes.
The suitability of channel vegetation in the riparian
zone for beaver was difficult to calculate given the available
information and the spatial scale of the model. The furbearer/
bird submodel requires the proportion of both main channel and
sloughs/side channels, with certain substrate conditions, which
have willow or balsam poplar in close proximity to the channel.
While it was not possible to make distinctions between main
and sloughs/side channels or substrate conditions, an
examination of aerial photographs indicated approximately 25%
of the channels in the riparian spatial unit (Talkeetna to
Devil Canyon) currently have willow or balsam poplar vegetation
in close proximity to the banks. Cover values for willow and
balsam poplar in each of the land classes in the riparian zone,
[
\'
L.
F
c
[
[
[
r L
r
L
L
[
L
[
[
[
[
[
L
[
c
c
[
l
[
c
-49 -
Table 3.6: Estimates of average values for potentially available
browse (to average po.int of browse) standing crop and
annual berry production in each land class. Average
values are modified in the model by a random variation.
LAND CLASS
Coniferous Forest -
woodland and closed
Coniferous Forest -
open
Deciduous and Mixed Forest
Tundra
Tall Shrub
Medium Shrub
Low Birch Shrub
Low Willow Shrub
Low Mixed Shrub
Unvegetated -water
Unvegetated rock, snow,
Disturbed -temporary
Disturbed -permanent
Pioneer
ice
POTENTIALLY
AVAILABLE BROWSE
(kg dry weight/ha)
198
283
144
111
200
588
588
300
275
0
0
0
0
0
BERRY PRODUCTION
(kg dry weight/ha
66
66
25
99
0
50
70
30
45
0
0
0
0
0
-50 -
as estimated from data in McKendrick et al. (1982), are combined
to yield a total cover value for the vegetation preferred by
beaver for each land class. These cover values are then averaged
across the various land classes, weighting each value by the
relative area in that land class:
where,
TBC = total cover value (percent) of beaver
preferred species;
BCt = cover value (percent) of species preferred
by beaver in each land class;
HAt= area of each land class (hectares);
THA = total non-water area in riparian zone
(hectares) ; and
t =land class type (1 through 14).
( 6)
TBC increases if vegetation changes increase the
proportions of riparian area in land classes with high cover
values for willow and balsam poplar and decreases if vegetation
changes result in proportio~ally more areas with low cover values
for willow and balsam poplar. Encouragingly, the value of TBC
calculated from the initial areas in each land class is within
0.5% of the independentl; estimat.:::d 25% of channel currently
having willow or balsam poplar in close proximity. Since a
value of 0 for TBC would also imply that 0 percent of the channels
had willow or balsam poplar in close proximity, TBC was assumed to
be a reasonable, direct indicator of the percent of channels in
the riparian zone which had associated vegetation characteristics
suitable for beaver.
[
[
[
[
[
[
c
[
L
L
[
[
r
[
[
~~
[
[
c
[
c
E
u
c
[
[
-51 -
3.3 Furbearers and Birds
The Susitna hydroelectric development will impact
furbearers and birds primarily through habitat changes, although
increased access may cause increased trapping intensity on
furbearers. Habitat changes will result from habitat losses due
to impoundments and to alteration of the downstream hydrologic
and ice regimes.
At the first workshop, the participants decided to
concentrate on the population dynamics of one furbearer, the
beaver, and to utilize a habitat approach for birds. In the
intervening period between workshops, a simple population model
for marten was added and the beaver and bird aspects were refined.
3.3.1 Beaver
The major sources of impact on beaver were hypothesized
to be:
1) a change in the amount of appropriate habitat for
food and denning sites; and
2} an increase in beaver trapping intensity due to
improved access to the region.
A simple beaver population model was built to simulate
the effects of these two sources of impact. A simple but
rigorous approach, neglecting some detailed biology (i.e.
ingestion rates, growth rates, fat content, fecundity, etc.),
is appropriate given the current state of knowledge. A more
detailed representation of beaver may be needed when more data
and understanding are available.
The model chosen is commonly used in biology -the
logistic growth model with an additional mortality term:
where,
-52 -
dB B dt = rB(l -K) -M
B = number of beaver colonies;
r·= intrinsic growth rate (yr-1 );
K = carrying capacity (number of beaver colonies) ;
and
M = mortality term.
The group chose the number of beaver colonies (also
called dens or lodges) as the measure of population because the
number of beaver in a colony is extremely variable. The
population time trajectory is easily. predicted (Figure 3.-10) if
the carrying capacity, intrinsic growth rate, and mortality are
constant over time. However, the trajectory is more complex if
the parameters change with time. The remainder of this section
describes how the subgroup chose to represent the variation of
these parameters as a function of the information available
from the other subsystems.
3.3.1.1 Beaver Carrying Capacity
In the context of this model, carrying capacity is the
maximum number of beaver colonies that can be supported within
the floodplain from Devil Canyon to Talkeetna. To determine
this number, it is necessary to first define good beaver habitat
and second, to e~timate the maximum· number of colonies that can
successfully use that habitat.
Beaver habitat was defined as kilometers of shoreline
satisfying the following conditions:
[
r
L
r~
I
L~
I
L
0
c
[
c
c
[
[
[
L
[
L_
[
[
[
[
[
c
D
[
[
z
0
~
...J K
:::>
Cl.
0
Cl.
-53 -
------------
TIME t
Figure 3.10: Time dynamics of a population based on the
logistic growth model. A population that starts
above its carrying capacity (K) will decline to
its carrying capacity. A population that starts
below its carrying capacity will increase towards
its carrying capacity.
-54 -
a) willow and balsam poplar are the dominant vegetation
adjacent to a shoreline with a bank composed primarily
of silt (from the vegetation submodel);
b) the water adjacent to the bank is ·sufficiently deep
so there is at least .5 m of unfrozen water below the
maximum ice cover (from the physical processes/
development/recreation submodel); and
c) water velocity adjacent to the bank does not exceed
4.4 feet/second between mid August and freeze-up.
The willow and balsam poplar vegetation is required by
beaver both as a source of food as well as lodge construction
material. Only vegetation in the riparian zone on either side
of the river is of interest because beaver rarely travel more
than 100 m from th.eir lodge location. The silty bank is
hypothesized to be an indicator of suitable slope for den
construction and lack of ice scouring.
The severe annual ice scour under the present flow and
ice regimes prohibits development of suitable habitat along the
main channel, and beaver habitat is only associated with the
proper vegetation in sloughs and side channels. However,
severe ice scour will likely be a rare event after impoundment.
This will probably result in more willow and balsam poplar
stands along the main channel which, given the predicted
stabilitation of water levels between Devil Canyon and Talkeetna,
could result in beaver establishing colonies on or near the
main channel. Therefore, a proportion factor for willow and
balsam poplar along the main channel, provided by the vegetation
s:.1bmodel, is used to convert shoreline length to appropriate
habitat.
Ice-free water is a critical condition to the definition
of habitat. Because a beaver den entrance is below the water
line, ice-free water is the route by which the beaver leave
[
[
[
[
[
F
[J
c
[
[J
c
c
c
c
[
c
[
L
r
[
[
[
[
[
[
[
[
c
c
c
[
b
-55 -
their den in the winter to feed. The hypothesis is that the
beaver will not survive the winter if there is less than .5 m
of ice-free water.
~he velocity criteria is likely only critical along the
main channel where velocities often exceed 4.4 feet/second.
This condition represents a maximum velocity, above which beaver
would probably not build a den since they would not be able to
swim upstream to forage the vegetation (Phil Gi:pson, pers. commQ) "
To arrive at an actual carrying capacity for beaver
colonies, it was assumed that the maximum colony density is
2 colonies/km of habitat. Therefore, the total carrying capacity
for beaver in each spatial unit is:
where,
K = ( (S * V ) + (2 * S * V ) ) * 2 s s m m
K = carrying capacity;
S = km of sloughs and side channels that do not s
freeze to within .5 m of the bottom (supplied
by the hydrology submodel);
Vs = proportion of willow and balsam poplar with
silty banks associated with S (supplied by s
the vegetation submodel);
sm = km of suitable main channel that do not freeze
to within .5 m of th~ bottom nor have velocities
greater than 4.4 ft/sec (supplied by the hydrology
submodel); and
Vm = proportion of willow and balsam poplar associated
with S (supplied by the vegetation submodel). m
-56 -
3.3.1.2 Intrinsic Growth Rate (r)
The intrinsic growth rate is the maximum rate at which
the population can increase. It assumes ideal conditions (i.e.
plentiful resources, no competition for habitat, etc.). This
growth rate is only realized in the logistic model when the
population is very much smaller than the carrying_capacity
(i.e. when B is much less than K in the logistic equation,
page 52) . The intrinsic growth rate (r) can be estimated as
the exponential growth rate in the equation:
where,
= N 0
Nt = number of beaver colonies after t years;
N0 = number of initial beaver colonies; and
r = exponential growth rate.
Participants hypothesized one beaver colony would spawn a
second colony in a minimum of two years if there was a surplus
of appropriate habitat and no other beaver colonies competing
for space. Therefore, a doubling of colony size in 2 years
implies:
N2 = N * er*2 = 2N 0 0
and ln2 r = -2-
-. 3
The intrinsic growth rate was assumed constant for this
model.
[
[
"' II
'I l:
[
[
[
[
c
c
[
c
[
l
L
[
[
b
[
[
[
[
c
c
B
6
c
[
c
[
[
~
l:
-57 -
3.3.1.3 Mortality
Wat:er Levels
Beaver colonies are vulnerable to changes in water level
within the year. Increases in water level on the order of a
few meters can result in the flooding of a den (in summer) or
the freezing of a food cache. (in winter). Similarly,. a drop
in water level will expose the colony to increased predation or,
even more likely, severe winter temperatures if the water level
falls below the den entrance. This is likely not a problem in
the sloughs and side channels but is definitely a major factor
(along with ice scouring) currently preventing establishment of
beaver colonies along the main channel. Since decreased
fluctuations in water level are predicted after impoundment, the
simulated beaver colonies which may have established themselves
in available habitat along the main channel are subjected to a
mortality factor from water level changes (Figure 3.11). Total
mortality of main channel colonies is possible with sufficiently
extreme water level fluctuations.
Ice Scouring
The mortality on the beaver is assumed directly proportional
to the total land area scoured between Devil Canyon and Talkeetna
(Figure 3.12). This mortality is applied to the appropriate
population in the spring of each simulated year.
Predat: ion
After some rtiscussion, the subgroup felt that predation on
beaver probably is insignificant. Beaver is a minor food item
for both wolves and bear. Therefore, predation is not presently
included in the model.
-58 -
0 ~--------------------~------------------~ 0 2
MAXIMUM CHANGE IN WATER LEVEL ( m )
Figure 3.11: Percent survival of beaver colonies on main
channel as a function of maximum change in water
level from summer to winter.
[
[
r
[
[
E
c
c
0
c
f' w
c
r
f L~
[
[
[
[
r--
b'
[
c
[
c
h
[
IIJ
~--= <t • a: c c >-0 ,_.c
-g
..J c ~a a: E o ....
:E
-59 -
0 ~-----------------------------1800 5000
AREA SCOUR E 0 (hectares)
a) SLOUGH BEAVER POPULATIONS
0
0 1800
AREA SCOURED (hectares)
b} MAIN CHANNEL BEAVER POPULATIONS
Figure 3.12: Mortality as a function of
ice scouring area for slough
(a) and main channel (b)
beaver populations.
-60 -
Trapping
Trapping is certainly one of the major potential sources
of beaver mortality. Beaver are especially vulnerable to
trapping during the winter when traps can be set over the
beaver's access hole in the ice. The rapid decline of beaver
populations in the lower 48 states when beaver trapping was a
viable occupation is evidence of high vulnerability to trapping.
Three factors were hypothesized to influence trapping effort:
1) beaver pelt prices;
2) knowledge about the location of beaver colonies; and
3) the number of other trappers in the area.
Price is certainly a key factor. Participants suggested
that the beaver population in the Susitna Basin would probably be
decimated within one year if beaver pelts were suddenly worth 5
to 10 times their current price (given the trappers knew where to
go) .
A maximum trapping mortality is calculated (Figure 3.13)
using a price factor between 0 and 1. The price factor is model
input and can be changed to explore the effect of a sudden price
shift. This maximum mortality is modified by an access factor
(Figure 3.14) expressed as a function of the number of people
using the spatial area (i.e. construction workers plus public).
For any given population, the access factor will change as a
function of the user-specified price factor. The assumption is
that access becomes less important as the rqlative price for
beaver increases. Therefore, if the price factor reaches 1,
then the beaver will experience the maximum trapping mortality
(i.e. maxT). At present, maxT is equal to .9 and maxA is equal
to 1. To limit access, an identified mitigation possibility,
the user must specify a lower value for maxA.
[
[
r
[
L
8
[
c
G
c
[
L
[
MAX. T
>-!:::
...J
~ a::
0
:::=
a::
IJ.J a. a.
<(
a::
1-
::E
;:)
:::!!
X
<(
:::= 0.1
0
-61 -
0
PRICE FACTOR
Figure 3.13: Maximum beaver trapping mortality as a function
of a user specified price factor.
MAX. A
a::
0 t-o
<(
ll..
(/)
(/)
IJ.J
(.)
(.)
<(
Price Factor = I
0 L---------------------------------~----------0 10,000
NUMBER OF PEOPLE
Figure 3.14: Trapper access factor as a function of the
number of people using the area.
-62 -
3.3.1.4 Beaver Migration
Since the water level changes are large before impoundment,
the main channel population invariably suffers total mortality
each year. Similarly, the population associated with sloughs can
experience higher mortalities in years of extreme ice scouring.
During periods of high mortality, it is expected that the non-
utilized beaver habitat in the riparian zone will be colonized by
migrants from other populations in the Susitna watershed.
This is incorporated into the model by increasing the
number of colonies associated with both the main channel and.
sloughs by 25% of the difference between the carrying capacity
and the spring population times the trapping survival factor.
If the colony population exceeds the carrying capacity, the model
assumes no migrants.
3.3.1.5 Beaver's Impact on Vegetation
The quality, quantity, and kind of streamside vegetation
is critically important to beaver. The critical vegetation types
[
[
[
r l~~
c
G
[
c
are felt to be balsam poplar, willow, and cottonwood. Observations
indicate that the balsam poplar and willow are generally concentrated '[
in a band running more or less parallel to the channel and often
within 40 m of the water's edge. The representation of appropriate
vegetation along the water:s edge (i.e. proportion -see Section
3.3.1.1) needs to be revised to include the information included in
the river cross sections available from the hydrology field work.
These cross sections identify specific vegetation zones relative
tothewater and permit a more acceptable approximation of the
percent of good beaver habitat near the water's edge (see
vegetation submodel description) •
It was also hypothesized that high densities of beaver
could have a substantial impact on the vegetative successional
progression in the riparian zone. Evidently, an average sized
beaver colony will forage approximately .4 ha of tall to low
shrub in a year which then usually reverts to low shrub.
[ -·
f~
L
fj
L
[
[
[
[
[
[~
[
c
c
c
~
6
e
c
[
[
r \~
-63 -
3.3.2 Marten
3.3.2.1 Population Structure
Three age classes are represented: ·o-1 year, 1-2
years, and older than 2 years. At the end of each simulated
year, the population remaining in each class is advanced to the
next category and the new litters are added to the first age
class.
The population processes represented are reproduction,
trapping, and natural mortality.
Reproduction is a functionofthe proportion of the
females which conceive, the litter size (Table 3.7), and the
male to female ratio (assumed constant at 50:50). Reproduction
is calculated as follows:
where,
Total of
all litters =
n
E
i=l
Pregnancy
rate. *
l
i = age group i; and
n = number of age groups.
Litter
size.
l
*
M/F
ratio
# marten in
* age group i
Trapping mortality is assumed to be fixed at 20% of the
total marten population per year. The proportion removed from
each age class to make up that 20% is fixed (Table 3.7).
Observation has shown that marten are very territorial.
It was estimated from available data that a maximum marten
density in their preferred habitat (i.e. forest) would be of
the order of .009 per hectare. Therefore, a density dependent
mortality function was incorporated into the model to ensure
the densities did not exceed this number (Figure 3.15).
-64 -
Table 3.7: various parameters for marten population model.
AGE
CLASS
0 - 1
1 - 2
2 +
PREGNANCY
RATE
0
.69
.79
LITTER
SIZE
0
3.3
3.8
PROPORTION
TRAPPED
.67
.23
.1
[
[
[
G
L
c
r. k1
[
L
[
[
[
[
[
[
[
[
[
c
[
[
-65 -
.9
liJ
~ a:
> t: .5
~
ct
&:
~
0
-to -2 _,
LOG 10
Figure 3.15: Density dependent mortality
rate for marten population.
-66 -
Although structured arbitrarily, it succeeds in maintaining the
marten population levels at acceptable densities for an otherwise
unstable population model.
Although an extremely simple population model, it does f~
permit evaluation of how the potential marten population might ·
be impacted by impoundment. The model also facilitates accumulation r
of the total number of marten trapped over the simulated time ·'
horizon, therefore indicating the total amount of marten production
lost as a consequence of the project.
3.3.3 Birds
At the first workshop, the subgroup participants identified
the golden eagle, yellow-rumped warbler, tree sparrow, fox sparrow,
and the trumpeter swan as key bird species for discussion. However,
after considerable discussion, participants concluded that the
limited state of knowledge about these birds precluded a dynamic
population model description of how they might be impacted by the
project. Also, many critical survival processes for these species
are controlled by events and conditions external to the model
because they are migratory. Therefore, impacts were simulated as
changes in habitat.
3.3.3.1 Passerine Birds
At the first workshop, the approach used for this group
was the Habitat Evaluation Procedure (HEP). The number of species
and bird density were identified as important to establishing the
value of any particular habitat. Average magnitudes for these
two criteria were specified for each vegetation type (Table 3.8)
using data from field studies in 1980 and 1981 in the upper basin.
A per hectare suitability index is calculated for each
vegetation type by taking the sum of 1/3 of the species number
value from Figure 3.16 and 2/3 of the bird density value from
Figure 3.17.
[
E
E
c
[
0
[J
[:
c
L
L
I ?
L-.i
L
l
[
r
f
[
[
[
[
c
[
-67 -
Table 3.8: Passerine bird density and number of species
associated with different vegetation types.
DENSITY SPECIES
VEGETATION TYPE #/10 ha lt/10 ha
Coniferous Forest
Open 15.7 8
Woodland 34.3 17
Deciduous and Mixed Forest 43.9 22
Tundra 3.9 7
Tall Shrub 12.5 10
Medium Shrub 39. 6
Low Shrub
Birch 10.6 6
Willow (10.6)
Mixed (10.6)
-68 -
0 ~--------------------------------------~ 0 25
NUMBER OF SPECIES /10 ha
Figure 3.16: The relative value of species in any given
vegetation type.
lJ.J
:::::>
_J
~
>-1-
IJ)
z
lJ.J
0
0
0 75
DENSITY (NUMBER /10 ha)
Figure 3.17: Relative value of bird density in any given
vegetation type.
L
L
[
[
[
r
F
t
f
c
[
[
-69 -
The relative weights for each criterion selected by the
subgroup indicate that bird density is somewhat more important
than number of species.
A total number of habitat units is then calculated within
each spatial unit:
where,
Habitat E
units = i TUi * Areai
= suitability index for a given hectare of
habitat i (from Figures 3.16, 3.17); and
Area. = area of habitat i in spatial unit.
~
This representation assumes the birds, on average~ will
use land of any given vegetation type in exactly the same way
each year. Although this is probably not a reasonable assumption,
there is not enough informationtotake the model much further at
this time.
At the second workshop, it was requesLed U1dt the passerine
birds be incorporated from the perspective of the number of avian
territories per unit area. Then, by multiplying these numbers by
the area of each vegetation group (some of which will change after
impoundment), the change in the total number of bird territories
could be predicted. This was done for certain species (Table
3.9) and total territories for all passerine birds (Table 3.10).
SPRUCE
GROUSE
HAIRY
WOODPECKER
BROWN
CREEPER
SWAINSONS
THRUSH
YELLOW-RUMPED
WARBLER
BLACK POLL
WARBLER
NORTHERN
WATERTHRUSH
WILSONS
WARBLER
SAVANNAH
SPARROW
DARK EYED
JUNCO
TREE"
SPARROW
FOX
SPARROW
Table 3.9: Number of bird territories/10 ha for 12 bird species for each of
the vegetation types represented in the model (FERC license
application, Exhibit E, Chapter 3, Table E.3.136).
DECIDUOUS
AND CONIFEROUS CONIFEROUE
MIXED FOREST-FOREST-
FOREST CLOSED OPEN -
1
.
1
1.5
6.5 3
8.5 1.7 1
2.7 1.9 1
2.2
3 9.4
3.2 2 2.5
5
2.3 3.2
LOW
TALL WILLOW
SHRUB SHRUB
.
.1
1.2 9.2
12.3
2.8
1.5 15
1.6
MEDIUM
SHRUB
8.8
3
11.8
LOW
BIRCH
SHRUB
5.8
2.5
~
' ' -,/
TUNDRA
1
-....J
0
[
[
[
[
1
~,
__ _j
[
E
-71 -
Table 3.10: Avian territories/ha used in model (taken from
FERC license application, Exhibit E, Chapter 3,
Table E.3.139).
AVIAN CENSUS
PLOT
Balsam Poplar
Forest
White Spruce-
Paper Birch
Mixed Forest II
White Spruce-
Paper Birch
Mixed Forest I
Paper Birch
Forest
White Spruce
Woodland
Black Spruce
Woodland
Open White
Spruce Forest
Tall Shrub
Low-Medium
Willow Shrub
Medium Birch
Shrub
Dwarf-Low Birch
Shrub
Alpine Tundra
MODEL
VEGETATION CATEGORY
Deciduous & Mixed
Forest
Deciduous & Mixed
Forest
Deciduous & Mixed
Forest
Deciduous & Mixed
Forest
Coniferous Forest
Closed
Coniferous Forest
Closed
Coniferous Forest
Open
Tall Shrub
Low Willow & Low
Mixed Shrub
Medium Shrub
Low Birch Shrub
Tundra
NUMBER OF
TERRITORIES/HA
6.1
3.5
4.2
3.8
4.4
. 2. 5
1.6
1.3
4.5
3.3
1.1
• 4
-72 -
3.3.3.2 Trumpeter Swan
Trumpeter swans are very sensitive to human disturbance.
Although there are only a few breeding pairs in the area, it is
known that Stephan Lake is a favored staging area during the
spring and fall migration. Participants felt that the construction
and use of roads and the transmission line would cause the major r~.
L" impacts. It was concluded that because potential impacts are
known and predictable, the concern involved proper siting of roads
and transmission lines to ensure minimum interference with nesting/
staging areas. This was not included in the model.
3.3.3.3 Golden Eagle
The major impact of the Susitna project on the golden eagle
will probably by the destruction of their traditional cliff nesting
sites due to inundation.
Most of the good eagle nesting sites that may be affected
have been found in the Watana impoundment area. Representation
of this impact in the model is done by comparing the elevation of
each active site to the maximum elevation of the reservoir. If
the nest .elevation is less than the maximum reservoir level, then
the nest site is counted as flooded. No attempt was made to
determine just which sites had an active nest in any given year.
Instead, this indicator shows the potential reduction in existing
eagle nest carrying capacity as a consequence of impoundment.
3.4 Moose
Development of the moose submodel was 7uided by two
fundamental considerations:
1) the need to produce indicators for evaluating both the
impacts of Susitna hydrelectric development on moose and
the potential effectiveness of various mitigation measures;
and
l~
IJ
[
u
[
[
L
[
[
[
E
c
r
L
[
[
6
c
D
[l
8
u
[l
[
[
c
5
C
-73 -
2) the desire to represent population processes in a manner
consistent with the information and understanding generated
by Alaska Department of Fish and Game (ADF & G) studies in .
the Susitna area.
Fortunately, this moose submodel for mitigation planning
was developed in parallel with the ADF & G Upper Susitna Basin
moose population modelling (Ballard and Miller, 1983). A detailed
description of the ADF & G moose model is provided in a technical
appendix (Appendix I). Most of the data and many of the
relationships incorporated into the moose submodel are based on
the work described in Appendix I.
The bounding exercise (Table 2.2) identified three general
types of indicators that should be responsive to impacts of
development and mitigation alternatives:
~) measures of numbers of animals (e.g. population size
and harvest);
2) indices or measures of habitat carrying capacity; and
3) indices or measures of habitat quality.
The present formulation of the moose submodel deals with the
first two of these indicator categories.
3 . 4 ·.1 Structure
The basic structure of the moose submodel is a life table
(based on the structure of the ADF & G moose model described in
Appendix I) that represents the birth and death processes for three
age classes (calves, yearlings, and adults) of moose of each sex,
combined with a simple model of winter carrying capacity. The
spatial area considered by the population model was defined based
on home range data for moose utilizing the impoundment area (Ballard r
et al., 1983). This area includes approximately 1200 mi 2 surrounding
74 -
the river from the Devil Canyon darn site to the east end of the
Watana impoundment. Carrying capacity is computed for this area
as well as for the five spatial areas defined in Section 2.3.
Project impacts and mitigation measures can thus -be evaluated
either as they affect the carrying capacity and moose population
in the area immediately around the impoundments, or as they affect
carrying capacity of the more broadly defined spatial areas.
The computational sequence for the model (adapted from
Appendix I) is illustrated in Figure 3.18. The biological year [
begins with calving. Animals surviving from th_e previous year
are advanced to the next age class and calf production is calculated [~
based on the number of females or reproductive age in the herd. The
remainder of the model consists of removal of animals due to a
series of age and sex specific mortality agents.
3.4.2 Wolf Population
Because wolf populations are not considered elsewhere in the
model, it was necessary to incorporate a very simple representation
of their dynamics in order to simulate their impacts on moose. The
number of wolves is calculated from a reproductive function based
on density and a mortality function based on snow accumulation.
The wolf fecundity rate is computed from Figure 3.19 based
on the wolf population in the previous winter. The declining
portion of this curve is hypothetical in nature and was incorporated
only to keep _the simulated population within reasonable limits. The
calculated fecundity rate is then multiplied times the number of
wolves remaining at the end of the previous 'year to produce a new
population.
All of the mortality agents acting on wolves are encapsulated
E
D
[
u
[
u
in a single mortality function dependent on snow accumulation (generated[~
by the physical processes subrnodel) (Figure 3. 20) . While this __ J
representation is overly simplistic, it does capture the idea that
wolves are harvested at higher rates (due to better visibility and
landing conditions for ski planes) in years of greater snow accumulation.
[
[
[
[
[
[
[
8
C
[
[
[
g
C
NEXT YEAR
-75 -
INCREMENT
AGE CLASSES
NEO-NATAL CALF MORTALITY I
I
SPRING WOLF PREDATION
I
SUMMER WOLF PREDATION
l NUMBER OF GRIZZLY
BEAR PREDATION j+---BEARS FROM
~-----------1.-----------~ BEAR SUBMODEL
HARVEST
I
POST-HARVEST POPULATION,
AGE RATIO, AND SEX RATIO
LAND CLASS ACREAGES
WINTER AND BROWSE
CARRYING CAPACITY +---AVAILABILITY FROM ~--~~~~~~~~~--~ VEGETATION SUBMODEL
WINTER WOLF PREDATION
Figure 3.18: General structure of the moose submodel
(adapted from Appendix I).
-76 -
2.0 ~~------------------~
\Observed Rate = I. 93
0 ~--------_.----------~--------~~--------~ 0 10 20 30 40
NUMBER OF WOLVES IN PREVIOUS WINTER
Figure 3.19: Wolf fecundity rate as a function
of population size.
LIJ
~ a:
>-
!::: _,
~ a:
0
::iE
IL. _,
0 :r:
1.0
0.8
0.6
0.4
0.2
0~--------_.----------~----------~--------~ 0 15 30 45 60
SNOW ACCUMULATION (Inches)
Figure 3.20: Wolf mortality rate as a function
of snow accumulation.
[
[
[
t
r
G
[
D
c
[
[
E
[
[
[
[
[
l
[
B
[;
t:
~
[
[
b
g
C
-77 -
3.4.3 Moose Reproduction
Reproduction is calculated separately for yearlings (those
females 2 years old at the time calves are dropped) and adults
(those 3 years or older at the time calve~ are dropped. A fecundity
rate for each group was derived from Ballard, et al. (1983) and
Blood (1973). Based on the fecundity rate data, a relationship
based on.the numb~r of moose present in the previous winter was
developed (Figure 3.21). The declining portions of these curves
were incorporated only to prevent unlimited increase of the
simulated population. Moose populations in the 10,000 -15,000
range have never been observed in the field. As long as the
simulated population remains within reasonable bounds, the effect
of these curves is to produce constant fecundity rates. Fecundity
rates are multiplied times the number of females in each cohort to
arrive at the number of calves born. The sex ratio in the calf
crop is assumed to be 50:50.
3.4.4 Mortality
Each mortality factor is represented by a series of mortality
events described in detail in Appendix I. Specific mortality events
considered are: neo-natal mortality, spring wolf predation, summer
wolf predation, winter wolf predation, .bear predation, and hunting.
Organization of the model allows calculations of sex and age ratios
and population size following each mortality event. This allows
for comparison with composition counts done in the field, and
provides a useful check on the simulation results.
3.4.4.1 Neo-Natal Mortality (based on Appendix I)
Calves are assumed to suffer a natural (non-predation)
mortality of 6% in the period shortly after birth, reflecting
accidents, abandonment, and a variety of other processes.
Provisions are also made for mortality of other age classes at
this time, but these are presently not invoked.
1.5
1&.1
~ a: 1.0
> t:
Q z
::I
(.)
1&.1
LL.
~ 0.5
0
0
~
-78 -
Adult Females = 1.19\
YearlinQ Females= 0. 29\
0 __________ _. __________ ~--------~~---------
0 s,ooo 10,000 15,000
NUMBER OF MOOSE IN PREVIOUS WINTER
Figure 3.21: Moose fecundity rates as a
function of population size.
[
L
[
r l:
[
[
[
[
[
[
[
[
[
[
[~
[
[
c
G
[
[
L
[
-79 -
3.4"4.2 Spring Wolf Predation (based on Appendix I)
Spring wolf predation on moose is calculated before the wolf
population is incremented by reproduction. This is consistent with
the fact that pups do not kill moose. Numbers of calves and older
moose (yearlings plus adults) are computed in the following manner<
The total weight of prey items required by the wolf population is
calculated as:
where,
K = weight of prey items required by wolf population (kg);
N =number of wolves (excluding pups);
C = weight of prey items required each day by an
individual wolf (7.1 kg/wolf/day); and
D =number of days in the predation period (80).
The number of calves or older animals killed is then:
M = (K * Pc)/Wc c
M = (K * Po)/Wo 0
where,
M = number of calves killed; c
M = 0
number of older animals killed;
K = weight of prey items required by wolf population (kg);
Pc =proportion of wolf diet composed of calves (0.35);
-80 -
P0 = proportion of wolf diet composed of older animals
(0.47);
W = average weight of calves (39 kg); and c
W0 = older animals (kg) •
The number of calves killed is distributed evenly between the two
sexes. The number of older animals killed is distributed in
proportion to their number, by sex, in the population.
3.4.4.3 Summer Wolf Predation (based on Appendix I)
Summer wolf predation.is calculated in the same way with
the following parameter changes:
1) pups are included in the wolf population;
2) the number of days in the predation period is changed
to 108;
3) proportions of the wolf diet are changed to 0.12 (calves)
and 0.755 (older animals); and
4) the average weight of a moose calf is changed to 94 kg.
3.4.4.4 Bear Predation (based on Appendix I)
The number of moose killed by grizzly bears is a function
of both the number of bears and the number of moose present. The
number of bears (excluding cubs and yearlings) is obtained from
the bear submodel. Daily predation rates per bear on calves and
older animals (adults and yearlings) are then computed from Figure
3.22. The number of moose killed is the product of the number of
bears, the predation rate per day, and the number of days in the
predation period (60). Calf losses are distributed evenly between
the two sexes. Losses of older animals are distributed among
the cohorts in proportion of their number in the population.
f~
[
f.
[
r·
[
E
§
t
[
[
c
C
8
[~
lj
f
[
[
[
[
-81 -
0.2
/Yearlings and Adults
0 ~==~----~--------~--------~----------J 0 17000 27000 37000 4,000
NUMBER OF MOOSE PRESENT
Figure 3.22: Bear predation rates as a function
of moose population size.
-82 -
Provisions are also made to adjust the shape of the curves in
Figure 3.22 as a function of snow a~cumulation in the previous
winter (reflecting increased vulnerability of moose following
severe winters), but these are not presently used.
3.4.4.5 Harvest
Moose harvest is specified-as either a specific rate to be
applied to each cohort, or a specific number of animals to be
removed from each cohort. The model presently assumes an annual
harvest of 30% of the adult males. It was not considered important
to relate moose harvest to human population in the project area,
since harvest will likely be closely regulated to prevent detrimental
impacts on the moose population.
3.4.4.6 Post-Harvest Population Statistics
The age ratio, sex ratio, and size of the herd are calculated
following the narvest. The age ratio is obtained by dividing the
number of surviving calves by the number of adult females. The sex
ratio is obtained by dividing the number of adult bulls by the
number of adult cows. These ratios are expressed as calves/100 cows
and bulls/100 cows respectively. The simulated age ratio, sex
ratio,· and population size calculated after the harvest correspond
roughly in time to composition counts actually done in the field,
and provide a useful check on the reasonableness of simulations.
3.4.4.7 Winter Wolf Predation
Winter wolf predation is calculated in the manner described
in Section 3.4.4.2 with the following parameter changes:
1) the wolf population is estimated by the average of the
populations before and after the wolf mortality function
(Section 3.4.2) is appliedi
2) the number of days in the predation period is changed
to 196i
f
G
[
E
D
[
L
[
~
[
[
[
l'
-83 -
3) proportions of the wolf diet are changed to 0.18 (calves)
and 0.714 (adults); and
4) the average weight of a moose calf is changed to 148 kg.
3.4.5 Winter Carrying Capacity
The winter carrying capacity for each spatial subunit is
calculated as the number of moose-days of browse available:
where,
14
U = ~ A. * B. * (1-L)/F j=l J J
U = moose-days of browse available;
Aj =area in land class j (ha);
Bj =available browse in land class j (kg dry weight/ha);
L = proportion of available browse at end of summer lost
due to leaf fall; and
F = individual moose forage requirement (kg dry weight/
day).
The vegetation submodel provides the area (Aj) and amount
of browse available at the end of the summer (B.) for each land
J
class. Available browse is defined as the standing crop of plant
material of species, height, and size (measured to the average
diameter at which browsing stops) suitable for moose forage. The
amount of browse available in the winter is the amount available
at the end of the summer reduced by a proportion representing
leaf fall. If browse is measured without leaves, L can be set
to zero. Division of a daily forage requirement produces the
number of moose-days of winter forage available.
-84 -
3.4.6 Winter Mortality
Winter mortality rates for moose can.be calculated in two
ways:
1) as a function of the winter carrying capacity with an
additional availability component depending on snow
accumulation; or
2) directly as a function of snow accumulation.
3.4.6.1 Winter Mortality as a Function of Carrying Capacity
The amount of brows·e available in the 1200 mi 2 herd area
is calculated as discussed above. Proportions of each land class
in this area were estimated from the proportions measured in a
16 km band surrounding the impoundment areas. When development
is initiated in the model, the amounts of vegetation inundated
are subtracted from the available range and hence from the
available browse. Browse availability is further modified by snow
accumulation (Figure 3.23). The total amount of browse available
is then divided by the number of moose in the post-harvest
population and the number of days in the winter period (180) to
arrive at the forage available per moose per day. Winter mortality
rates are then determined from Figures 3.24 and 3.25 using forage
available per moose per day as the independent variable.
This approach to determining winter mortality has the virtue
of attempting to relate mortality to the most obvious project
impact (i.e. vegetation removal}. It must be used with caution,
however, since both the relationship between snow accumulation and
the proportion of forage available, as well as the relationships
between forage availability and mortality, are poorly understood
in a quantitative sense.
[
f
1' ( ....
L
[
[
[
[
[
[
[
[
[
c
[
[
-85 -
1.0
ILl
..J m
c(
_J
~ <
ILl
C)
c( a: 0.5 0 u..
u..
0
z
0
1-a:
~
0 c:: a..
0
0 15 30 45 60
SNOW ACCUMULATION (inches)
Figure 3.23: Forage availability as a function
of snow accumulation.
ILl
~ a:
>-!:::
_J
~ 0.5
0::
0 :::=
ILl
...J
c( :=
3 5
FORAGE AVAILABLE PER MOOSE PER DAY
(Kg dry weight)
7
Figure 3.24: Male winter mortality rates as a
function of forage availability.
IU
~ a:
>-....
:J
1.0
~ a: 0.5
0
~
IU
..J
~ :e
IU
IL
0
-86 -
3 5
FORAGE AVAILABLE PER MOOSE PER DAY
(Kg dry weight)
7
Figure 3.25: Female winter mortality rates as
a function of forage availability.
[
[
t
[
[
[
L
[
[
F
[
[
[
[
[
f]
C
[
§
~
E
u
[
t
-87 -
3. 4. 6. 2 Winter Mortality as a Function of Snow Accumulation
As noted above, winter mortality rates may also be calculated
directly as a function of snow accumulation. Three levels of
mortality are distinguished for three ranges of snow accumulation
(Table 3.11).
This approach to the winter mortality calculation has the
virtue of being more directly related to field observations.
Mortality rates for the first two levels of snow accumulation were
determined from radiotelemetry data (Ballard, et al., 1983).
However, the actual snow accumulations at the time the radiotelemetry
data were obtained are unknown. Second, snow accumulations and
mortality rates for the third level (> 39 in) are purely hypothetical
at this time. And finally, of course, this approach does not relate
mortality to any project impacts.
3.5 Bear Submodel
The bear submodel relates population responses of black
and brown bears to changes in habitat structure and to the more
direct human influences of hunting and dispersal from disturbance.
Due to the limited time available at ·.the first workshop, only
female bears were considered and hunting was not included in the 1
first modelling attempt. Subsequent technical meetings have
corrected these simplifications as well as adding substantial
complexity to the structure of the model. Field data upon which
some of the parameters of this submodel are based are presented
in Miller and McAllister (1982) and Miller (1983).
3.5.1 Population Structure
The brown bear population in the study area is stratified
into two groups: those using the area that will be directly
affected by the impoundment (vulnerable population) , and
those that will not (non-vulnerable population) (Figure 3.26).
-88 -
Table 3.11: Moose mortality rates at various depths of snow
accumulation (modified from Ballard, et al., 1983
and Appendix I) .
SNOW
ACCUl-!ULATION
> 32 in
32 -39 in
> 39 in
~MALES
CALVES YEARLINGS
6 6
57 10
95 80
MORTALITY RATE (%)
ADULTS CALVES
3.6 6
7.2 14
70 95
FEMALES
YFARLnx3S
2.4
2.4
80
ADULTS
3.6
3.6
50
[
[
[
G
D
[
u
[
L
[
I ~
Q
! .
I.
! ~
NON-VULNERABLE
POPULATION
-89 -
VULNERABLE
POPULATION
STUDY AREA
OUTSIDE STUDY AREA
Figure 3.26: Diagrammatic representation of the
division of the bear population
into vulnerable and non-vulnerable
numbers.
-90 -
Dispersal between the vulnerable and non-vulnerable population
and between a population outside the study area (i.e. a "buffer"
population) is allowed. A similar structure is utilized for
black bears, however, the entire population in the study area is
considered to be vulnerable. This structure is used to mimic the
idea that specific geographical regions may be net producers (i.e.
sources) or sinks for bears. The resulting population in any
given area depends, in part, on the rate at which the area "leaks"
bears to less productive areas or acquires bears from more
productive surrounding areas.
The submodel relates the underlying processes of
reproduction, hunting mortality, natural mortality and dispersal
to changes in conditions and food supplies which operate on
specific maturity, age and sex classes of the vulnerable and
non-vulnerable populations. These classes are linked in the
form of a simple life table and are portrayed in Figures 3.27
through Figure 3.30 for brown female, brown male, black female
and black male bears respectively. Mature females are partitioned
into groups based on the presence or absence of offspring (three
groups for brown bears (Figure 3.27), two groups for black bears
(Figure 3.29)). Immature black bears are partitioned into four
age classes and immature brown bears are partitioned into six
age classes.
The proportion of bears in a given age class that have
reached maturity (Table 3.12) is assumed constant. For example,
a three year old immature female brown bear that survives the
year must become either a mature animal with no offspring or a
four year old immature animal (Figure 3.27). Mature animals
without offspring either rema~n in that condition or produce cubs.
The sequence of calculations for the submodel is diagrammed
in the form of a flowchart in Figure 3.31. Each calculation is
described in further detail below.
[
[
[
[
['
L
F
l~
[
[
G
c
c
c
[
L
UJ
0:
:J
~
::E
::E
NO
OFFSPRING
6-
YEAR
5-
YEAR
I
WITH
CUB
CUB
4-
YEAR
I ...
3-
YEAR
WITH
YEARLING
YEARLING
2-
YEAR
Figure 3.27: Life structure of female brown bear. Each arrow represents
a time step of one year.
E
,.-----.,
' ' J
\.0
I-'
w a:
:J
!;(
:I!
~
MATURE
I cua I .. YEARLING
ti-5-4-3-2-
YE~R YEAR YEAR YEAR YEAR
Figure 3.28: Life structure of male brown bear. Each arrow represents
a time step of one year.
r----T"'.
I J
;-------\
I J
E
\D
N
...-..---,
\. .J
UJ a:
:::l
t(
:E
:E
NO
OFFSPRING
4-
YEAR
3~
.YEAR
2-
YEAR
WITH
CUB
CUB
1-
YEAR
Figure 3.29: Life structure of female black bear. Each arrow
represents a time step of one year.
E
1.0 w
MATURE
4-
YEAR
3-
YEAR
2-
YEAR
CUB
1-
YEAR
rue ure o rna e lack bear. Each arrow Figure 3.30.· L1"fe st t f 1 b
represents a time step of one year.
E
~ . .J ::-J
[
[
f
[
[
L
[
c
c
c
[
c
[
E
t
.-95-
Table 3.12: Assumed proportion of bears reaching maturity by age.
PROPORTION REACHING MATURITY
AGE BLACK BROWN
2 0.5
3 0.75 0.44
4 1.0 0.76
5 0.9
6 1.0
-96 -
~-------------------------START
OTHER YEARS
1
FIRST YEAR
l
SET AGE AND SEX SPECIFIC
HARVEST AND DISPERSAL RATES
SOLVE FOR STEADY POPULATION
LEVEL BY ADJUSTING IMMIGRATION
SET BASE HUNTING, DISTURBANCE
AND FOOD INDICES
OVERALL FOOD INDEX
I
REPRODUCTION
I
HUNTING ANO NATURAL
MORTALITY RATES
UPDATE LIFE STRUCTURE
DISPERSE BETWEEN VULNERABLE
AND NON-VULNERABLE POPULATIONS
SUMMER AND FOOD INDEX
END
Figure 3.31: Sequence of bear submodel operations.
[
[
l '
[
[
t
[
[
[;
[
L
L
[
[
[
[
-97 -
3.5.2 Initial Population Equilibrium
During the first modelled year (1980), the population is
assumed· to be in equilibrium with the surrounding populations
such that if all factors that affect bears were to remain the
same, the total population size after each cycle of the life table
(i.e. 10 years for brown bears and 7 years for black bears) would
reamin constant. In other words, immigration and recruitment are
in balance with losses due to natural mortality, hunting mortality
and immigration. This assumption may be unrealistic if populations
in the surrounding areas are in fact declining.
To obtain the above conditions, all factors, with the
exception of immigrat±on, were preset and a constant immigration
level was found by utilizing a non-linear algorithm (Simplex
varying step size). Recruitment was obtained by letting half the
females without offspring produce a litter size of two (one male,
one female) the following year. ~he natural mortality constants
are presented in Tables 3.13 and 3.14 for brown and black bears.
On the other hand, hunting mortality and dispersal rates were set
through a more cumbersome method. A generalized formulation was
used to determine both :the hunting mortality and dispersal rates:
L:L: G ..
R .. = R w .. ~:l
~J ~J IL: G .. w ..
~J ~J
where,
i = the age class;
j = the sex;
R ..
~J = the specific rate;
R = an overall mean rate;
-98 -
Table 3.13: Brown bear base natural mortality
CLASS FEMALE
Mature (no offspring) • 05 .
Mature (with cub) .05
Mature (with yearling) .05
Cub .15
Yearling .1
Immature (2-year) .08
Immature (3-year) .06
Immature ( 4-year) .05
Immature (5-year) .OS
Immature (6-year) .05
Table 3.14: Black bear base natural mortality
CLASS FEMALE
Mature (no offspring) .08
Mature (with cub) .08
Cub .15
Immature (1-year) .1
Immature (2-year) .08
Immature (3-year) .08
Immature (4-year) .08
estimates.
MALE
.04
.15
.1
.07
.05
.05
.05
.05
estimates.
MALE
.07
.15
.1
.08
.08
.08
I
(.
[
[
f
c
[
[
r-,
b
L
[
[
F
[
[
[
c
[
-99 -
w .. = a relative unitless weight; and
~J
G .. = a population level from which the overall mean ~J
was derived~
In other words, an overall mean rate is partitioned into the
various classes according to a set of weights consistent with
an initial population level.
Tables 3.15 and 3.16 depict the initial population levels,
Tables 3.17 and 3.18 the relative weights for dispersal, and Tables
3.19 and 3.20 the relative weight for hunting. The relative
weights can be viewed as the propensity for that event to occur.
For example, Table 3.17 declares that an immature three year old
male brown bear is 10 times more likely to disperse than a mature
animal.
3.5.3 Indices
The primary factors that affect the processes of reproductionr
mortality and dispersal of bears can be identified. However,
quantitatively little is known about the functional form and
parameter values for these relationships. Therefore, indices
relative to 1980 (assumed to be an "average year") are
utilized for each of the primary factors (summer and fall food,
spring food, disturbance, and hunting effort).
3.5.3.1 Summer and Fall Food Index
Since summer and fall foods are thought to be primarily
bluebe~ries, the index for any year t is defined as:
total berry production in year t
total berry production in 1980
-100 -
Table 3.15: Assumed brown bear initial population size.
CLASS FEMALE MALE
Mature (no offspring) 30 so
Mature (with cub) 13
Mature (with yearling) 12
Cub 12 12
Yearling 12 12
Immature (2-year). 10 11
Immature (3-year) 9 9
Immature ( 4-year) 4 6
Immature (5-year) 1 3
Immature (6-year) 1 1
Table 3.16: Assumed black bear initial population size.
CLASS FEMALE MALE
Mature (no off·spring) 39 54
Mature (with cub) 16
Cub 16 15
Immature (1-year) 17 18
Immature ( 2-year) 14 24
Immature (3-year) 8 14
Immature (4-year) 4 6
[
[
[
[
t~
['
I
I.
[
[
c
c
. [
[
l
c
[
[
[
[
[
L
c
n L
r
6
[
b
L
L
r:
-101 -
Table 3.17: Brown bear dispersal weight by class and sex.
CLASS FEMALE MALE
Mature (no offspring) 1 1
Mature (with cub) 1
Mature (with yearling) 1
Cub 1 1
Yearling 1 1
Immature (2-year) 2 5
Immature (3-year) 3 10
Immature (4-year) 3 9
Immature (5-year) 2 8
Immature ( 6-year) 1 1
Overall dispersal rate = .1
Table 3.18: Black bear dispersal weight by class and sex.
CLASS FE1-1ALE MALE
Mature (no offspring) 1 1
Mature (with cub) 1
Cub 1 1
Immature (1-year) 2 3
Immature (2-year) 3 10
Immature (3-year) 3 7
Immature (4-year) 2 3
Overall dispersal rate = .2
-102 -
Table 3.19: Brown bear harvest weight by class and sex.
CLASS F.EMALE MALE
Mature (no offspring) 4 5
Mature (with cub) 1
Mature (with yearling) 3
Cub 1 1
Yearling 3 3
Immature (2-year) 4 10
Immature (3-year) 8 10
Immature (4-year) 8 9
Immature (5-year) 8 9
Immature ( 6-year) 7 8
Overall hunting mortality = .1
Table 3.20: Black bear harvest weight by class and sex.
CLASS FEMALE MALE
Mature (no offspring) 4 5
Mature (with cub) 1
Cub 1 1
Immature (1-year) 6 8
Immature (2-year) 8 10
Immature (3-year) 7 9
Immature (4-year) 6 8
Overall hunting mortality = .1
[
[
l~
t
6
[J
c
f'
l
[
L
[I
\
L
[
[
[
[
[
c
[
c
0
[
[
[
[
E
[
-.103-
The total berry production for a given year is the sum of total
berry production in each vegetation type. The vegetation submodel
provides berry production per hectare for each vegetation type and
the area in each vegetation type which allows calculation of total
production. For brown bears, the total ~tudy area is utilized,
while for black bears, only the two impoundment areas and the 60
km strip from Devil Canyon Dam to Talkeetna are used.
The summer food index for brown bears is modified by use
of the salmon resource from Prairie Creek. Twenty-five percent
of brown bears in the study area are assumed to use this resource
during one-third of their summer feeding periods. It is assumed
that future recreational developments or material sites in the
area will preclude bear use of this resource. Because the level
of disturbance (number of recreational use days per year) necessary
to preclude use could not be determined, it was arbitrarily assumed
that this resource would be lost if recreational use doubles the
1980 level. If this recreational use level is reached, the summer
food index is reduced by 8%.
3.5.3.2 Spring Food Index
Spring food (which includes such items as Equisetum, moose
calves, small mammals, skunk cabbage, roots, and cottonwood buds}
is more vulnerable to inundation than summer food. The index
relates preference of vegetation types utilized per bear to the
base year 1980 and is calculated as:
total area of vegetation in year t weighted by preference
total area of vegetation in year 1980 weighted by preference
* # of bears in 1980
# of bears in year t
The assumed relative preference weights are depicted in Table 3.21
for brown and black bears. For brown bears, the total study area
is utilized, while for black bears, only the two impoundment areas
and the 60 km strip from Devil Canyon Dam to Talkeetna are used.
-104 -
Table 3.21: Assumed relative preference of vegetation types.
VEGETATION TYPE
Conifer Woodland
Conifer Open
Deciduous and Mixed
Tundra
Tall Shrub-Alder
Medium Shrub
Low Birch
Low Willow
Low Mixed
Water
Rock/Snow/Ice
Temporary (Disturbed)
Permanent (Disturbed)
Pioneer
BROWN BEAR
5
5
7
5
3
3
5
5
5
0
0
0
0
8
BLACK BEAR
8
8
10
0
3
2
2
3
3
0
0
0
0
10
L
[
B
0
[
c
[
c
c
L~
r
LJ
[
L
[
[
[
r
[
c
c
-lOS -
3.5.3.3 Disturbance and Hunting Effort Indices
Total disturbance and hunting effort in user days are
provided directly by the recreational submodel. The indices
are the simple ratio with the base 1980 year.
3.5.4 Reproduction
The proportion of females emerging with cubs is a function
of the previous summer's food index while cub survivorship is a
function of the current spring food index. In the model, the
combined effect of these processes is simulated as a function of
a composite index of the previous summer's food and the current
spring food. For vulnerable populations, the composite index
consists of 80% summer food and 20% spring food. For the non-
vulnerable brown bear populations, the index consists of 80%
summer food with a constant 20% added on to represent mean spring
food.
The proportion of females emerging with cubs as a function
of the composite index is shown in Figure 3.32a. Fifty percent of
the females emerge with cubs when the food index is 1.0,
representing an avArngP year. The a parameter governs the
sensitivity of pregnancy rate to food availability. When the food
index (Figure 3.32a) is near l -a, the proportion with cubs is
near 0; when it is near l + a, the proportion is close to 1.0.
In the current version of the model, a is 0.2 for black bears and
0.5 for brown bears; black bears are assumed more sensitive to
changes in berry production.
At present, the model employs a constant litter s~ze of
two. However, an option is available for mean litter size to be
determined as a function of the food index (Figure 3.32b). The
maximum mean litter size is 2.5 for brown bears and 2.7 for black
bears. The number of cubs is the product of the number of females
emerging with cubs and the mean litter size. It is assumed that
50% of the cubs are male and 50% are female.
-106 -
(a)
1 -----------------------·-------rn
Cl
:J
(..)
~ ,...
3:
~ z a a: w
~ .s w
rn w
...1
c(
~ w
"" "" 0
z
Q ,...
~ a:
"" 0
1-~ 1 1+9(.
INDEX OF FOOD
(b)
max ----------------------..,....------
w 2
~ rn
a: w ,... ,...
::!
1.1
1
INDEX OF FOOD
Figure 3.32: Reproduction relationships as
a function of the index of food:
(a) proportion of females emerging
with cubs;
(b) mean litter size.
[
[
f' L,
[
c
c
c
[
L
~
r~
r
[
[
[
[
E
c
c
G
6
c
u
E
6
[
u
r
-107 -
3.5.5 Mortality
3.5.5.1 Hunting Mortality
The method for devising the hunting mortality rate is
discussed in detail in this section since the same rationale is
utilized for natural mortality and dispersal rates.
Mortality rates can always be expressed in terms of the
complement survivorship; i.e.:
where,
HMt = hunting mortality; and
HSt = survival from hunting in any year t.
Suppose that the effective hunting effort doubled over the base
year (1980) with all bears in a population remaining equally
vulnerable. Then, the fraction of bears surviving is:
HSt = (1 -~) 2
where,
~ = the base hunting mortality.
In other words, the bears must be subjected to the base hunting
rate exactly twice since the effective h\:nting ef~ort doubled.
This scheme may be generalized to any increase or decrease in
hunting effort; i.e.:
-108 -
where,
EV = the effective hunting vulnerability.
However, a change in hunting effort may not translate into an
equal vulnerability of bears. An increase in hunters may produce
interference of an individual hunter's effectiveness, a portion
of the bear population may become wary because of disturbance, or
regulation may introduce inefficiency. This phenomenon can be
mimicked by multiplying any increase or decrease of the hunting
index from the base year (1980) by a sensitivity constant:
EV = (Hunting Index -1) Sensitivity Constant + 1
Thus, a sensitivity of 1 produces a direct relationship between
the number of hunters and the vulnerability of bears to hunting,
while a sensitivity of 0 results in no change from the base rate,
regardless of the number of hunters. Figure 3.33 depicts the
effect upon the mortality rate from a decrease in sensitivity.
The base rates partitioned by age, maturity and sex were
those obtained for the equilibrium conditions. All populations
(vulnerable and non-vulnerable) are assumed to be subjected to
hunting. However, at present, the sensitivity of brown bears
to hunting is set to 0.02 to reflect the workshops participants'
belief that hunting of brown bears can be largely controlled
through regulation. Similarly, the sensitivity of black bears
is also small (0.2, i.e. a five-fold increase in hunters only
doubles the effective vulnerability), but somewhat larger than
for brown bears since their range is restricted and kills are
often the result of chance encounters by hunters while targeting
upon other species.
r
L
r·
L
G
[
[
c
f1 L
c
c
L
L
>-
!::
..J
~
0::
0
:t
(!) z
1-z
::l
:I:
1
1:-r:J
BASE RATE
0
,..._,.....-------,
' J
\ SENSITIVITY
1
HUNTER INDEX
Figure 3.33: Hunting mortality rate as a function of the hunter
index with the effect of a lower sensitivity
illustrated.
-110 -
3.5.5.2 Natural Mortality
All animals of the non-vulnerable populations and animals
of the vulnerable populations two years of age or greater are
assumed to have a constant natural mortality r~te (see Tables
3.13 and 3.14). The mortality rates of the remaining cubs and
yearlings of the vulnerable population are calculated in the same
manner as hunting mortality with the reciprocal of the spring
food index replacing the hunting index in the mortality equation,
since spring food is more vulnerable to inundation than summer
food, and the base rates presented in Tables 3.13 and 3.14. The
cubs and yearlings are considered to be completely susceptible
to changes in spring food availability (i.e. sensitivity).
3.5.5.3 Nuisance Kill
Only nuisance kills associated with construction work are
considered explicitly. At maximum activity, it is assumed that
five brown bears and seven black bears will be killed each year.
For construction activity less than maximum, a simple proportionate
number of animals are killed (Figure 3.34). The total kill is
then partitioned into the appropriate sex, maturity and age classes
according to the relative weights given in Tables 3.22 and 3.23.
3.5.6 Dispersal
Brown bears disperse between vulnerable and non-vulnerable
populations at a constant relative rate of 15% each year. There
is no such dispersal of black bears since all are considered
vulnerable to inundation. In addition, all bears can disperse
to the "buffer" population outside the study area. Base dispersal
rates, as calculated for initial population equilibrium conditions,
are assumed to be constant for the non-vulnerable bear populations.
[
[
c
8
c
L
L
a:
ct
~
a: w
Q.
(/)
..J
d
~
UJ
0 z
~
5 z
u.
0
a:
UJ m
:::E
:::l z
..J
~
0
t-
(MAXIMUM (BROWN=5, BLACK=-7)) ·-------------------------~-----------
2500
NlJiMBER OF CONSTRUCTION WORKERS
Figure 3.34: Number of nuisance kills as a function of construction
activity.
-112 -
Table 3.22: Brown bear nuisance kill weights by class and sex.
CLASS
Mature (no offspring)
Mature (with cub)
Mature (with yearling)
Cub
Yearling
Immature (2-year)
Immature (3-year)
Immature (4-year)
Immature (5-year)
Immature (6-year)
FEMALE
7
7
7
7
7
4
4
4
4
4
MALE
2
7
7
4
4
4
4
4
Table 3.23: Black bear nuisance kill weights by class and sex.
CLASS FEMALE MALE
Mature (no offspring) 4 2
Mature (with cub) 4
Cub 4 4
Immature (1-year) 7 7
Immature (2-year) 7 7
Immature (3-year) 7 7
Immature (4-year) 7 7
[
I . l
[ u
~
c
c
L~
L
L
[ -113 -
For the vulnerable populations, the base rates and the disturbance
index are used in the same manner as hunting mortality to calculate
dispersal rates of sex, maturity and age classes each year. The
sensitivity of brown bears (0.4) to disturbance is assumed to be
much greater than for black bears (0.1).
While the dispersal of bears is modelled explicitly, other
mortality factors, such as the result of disturbance (e.g.
nuisance kills), are implicitly included since the bears that do
disperse are no· longer members of the study area population.
3.6 Model Results
The model, in its current state, consists of numerous
functional relationships of the biophysical processes operating
in the Susitna Basin. Lack of data and understanding forced an
overly simplistic representation of many of these processes. As
a result, great care must be taken in evaluating the results
presented in this section. We caution against considering the
results to b~ valid projections of what might happen in the Susitna
Basin.
Two scenarios (sets of actions) to be simulated were
developed:
a) a baseline or no project scenario; and
b) the full project, Case C, power generation scenario
with little mitigation.
The major differences between scenarios (Table 3.24) relate to
flow regime, number of dams constructed, choice of access route,
and control of access.
-114 -
Table 3.24: Scenarios used in the simulations.
Flow Regime
Access Route
Access Control
Dams Constructed
NO PROJECT
preproject
none
no increased
access
none
FULL PROJECT
case C (optimum
power generation)
plan used in FERC
license application
open access
Watana, Devil
Canyon
[
[
[
L
c
l
[
[
r~
[
[
~~
~~
u
L
Q
e
~
-115 -
The following figures compare indicators for the two
scenarios. It may ultimately be desirable to compare the
quantitative results but, at present, only the qualitative results
should be considered. It is more appropriate to examine the
general temporal differences in the indicators among the scenarios,
rather than to focus on their actual values.
3.6.1 Physical Processes/Development/Recreation
The maximum annual change in stage measured at Gold Creek
Station (Figure 3.35) is considerably less under the regulated
scenario (Figure 3.35b). The drop that occurs at simulation year
12 is associated with the commencement of the operation of the
dam~.
The amount of reservoir clearing in a year (Figure 3.36}
follows the schedules outlined in Table 3.1. The large jump in
reservoir clearing in the development scenario (Figure 3.36b) is
associated with the clearing for Watana; the smaller jump later
in years 21 -24 is associated with clearing for the Devil Canyon
impoundment.
Influx of construction personnel is associated with dam
construction (Figure 3.37). In the model, this influx is simulated
using the schedule outlined in Table 3.3. The large peaks are
associated with the construction of Watana (Figure 3.37b); the
lesser peak is associated with the construction of Devil Canyon
(Figure 3.37b).
Recreational use of the area is assumed to increase gradually
without the project (Figure 3.38a). Under the full project scenario
with no restriction on access (Figure 3.38), there is a steeper increase
in recreational us.e for ten years after construction of Watana is
completed.
5 1 ..
I . ,.
!
t
i
1-
!.l.
i r
.t.
I
l·
I •" /• ·\ 1-} '..,-"~i" r • ' .. ,
j-I
I •
5 1-
1
i-
1
t
I
.1-
i
j-
!.
Figure 3.3S·
-116 -
(a,) No Project
TI t·1E
(b) Full Project
31
TI 1'1E
Maximum annual change in stage at Gold
Creek 3tation. The maximum value on
y-axis is 10 feet.
[
[
l
[
r~
[
[
6
c
r:
lJ
[
[
u
[
-117 -
F:~ s '· 1 ) t·'!~ :~<= 5 a a a
~:E s ( = ! t··H~ ::-::= s 0 il !!
i ....
!
4-
i + i
! + i + l
5 + ! (a) No Project
t
j
+ I + ' + i
il
i 2. 1 = 1 21 sn
T I t-'E
F: E $ ( 1 ) r--!J.".i ::(:: 5 !! iJ !! •
RE$(2) t·'!J.".I::<::: 5£l!l!l.
iT
+ ! ·1•
I + i + I
~ + . :1 !
+
\
-+ I + I
::!. i ;
Figure 3.36:
(b) Full Project
21. su
T I 1'1:
Amount of reservoir clearing (ha) per
year. The maximum value on the y-axis
is 5000 ha.
:!. -! ..;.
+ i . •i•
i + . .
5 + I ·t
!
J.
!
i
i +
il
!
; --I
..j.
t
!
"t
I + l
. 5 i·
I
·+ ! + I + I + j
1
i ..
I
l
_,I
I
~c~ !.
! i
.i,
' ' ! iw
I ~
I ' t I ;
I I
t ,.
' •' ' ~ . ..
-118 -
ST ( ~ ! t·-:~~ >=~= 2 ~ i1 £!
(a) No Project
:1 l 3:!. su
T I r"£
(b) Full Project
..•
;. \ . ' I I ..
' p ~
I "' ;
; J " I r ,
,, ,. !
' ,o I \.
511
TIt·'£
Figure 3.37: Construction personnel on site at any
one time. The maximum on the y-axis
is 2500 workers.
[
r'
L
r=
' >
0
[
-119 -
~ ---
[ -~-~------~~----t...l-------------!
il
1 .. -I
i-
1
l-
!
' 1""
I
~ ..
1.!.
T! t·1E
i
5 t _ ... --·--... -------... ---"'-
1-I _,• ~
{"' ,;
I /
+-:r i I r-.... -----... -... "
---__ _, ..... --
n ~~~~-~~~~~~~~~~~++~~~~~~
i:!. ~ i 3 i
T! t·1E
(a) No Project
(b) Full Project
Figure 3.38: Recreational use days in the Upper Susitna
Basin. The maximum on the y-axis is
100,000 use days.
-120 -
Potential overwintering habitat for beaver (Figure 3.39)
appears to show a slight decrease after the project is introduced.
This small decline occurs in the model because of the lower
hydraulic head between the open water section of the downstream
reach and adjacent slough and side channel habitat. However, this
relationship is. a candidate for refinement (c.f. Section 5.1.1).
The area of the downstream reach subjected to ice scouring
(Figure 3. 40) shows considerable variation under the natural
hydrological regime (Figure 3.40a). With the project, the frequency
of ice scouring is reduced as a result of the ice melting in place
before the high tributary inflows have an opportunity to trigger
break-up.
The minimum surface area covered by water during the growing
season (Figure 3.41) is an important determinant of the process of
riparian succession. The introduction of the project reduces the
amount and variability of the flooded area (Figure 3.4lb).
3.6.2 Vegetation
In the Upper Susitna Basin, available winter range for moose
is assumed to be located at 4,000 feet in elevation. Changes in two
vegetation types that make up much of the food available on the
winter range are illustrated in Figures 3.42 (deciduous and mixed
forest) and 3.43 (low mixed shrub). The deciduous and mixed forest
shows a substantial decline (Figure 3.42b), while the low mixed
shrub (Figure 3.43b) shows only a slight decline.
While the deciduous and mixed forest declines, it has a low
browse value. As a result, the change in available forage for moose
(Figure 3.44) is difficult to discern from the natural variability.
However, much of the deciduous and mixed forest that will be
inundated occurs at lower elevations in the valley bottoms. It is
believed that during severe winters (high snow accumulation), moose
will utilize the valley bottoms during the early spring.
[
[
'.·
L
u
L
[
[
L
L
[
[
[
[
[
E
L
r
L
[
+ ! + i
-121 -
(a) No Project
o~·~--~~~--~~~~~~~----~--~
·t
I +
I
t
T I 1'1:
Sll
(b) Full Project
0~~~----------------~--------~--~~~ !
Figure 3.39:
31 sn
TIt·"£
Potential overwintering habitat for beaver
in sloughs and side channels. The maximum
on the y-axis is 30 km.
....
I + t
il
.;
' ... -i
l
i
·-t.
j ~\ .or. + .. '.' '. ,' ·,,...,,.
i' I + ~ ~~
I ~ I
5 t 1~
! 'I + ) . 'I
Figure 3.40:
-122
(a) No Project
21
(b) Full Project
2 1 31 S!!
T.I 1'-iE
Area subject to ice scouring in the
downstream reach. The maximum on the
y-axis is 2500 ha.
[
[
[
[
[
r
[
[
L
t
i + I + i
R~LOOD MAX: ~~DO.
-123 -
(a) No Project
n ~·~~~~~~++~~~~~~~~~~~-~~~~
i
t
! ·+
!
1
il
i
11 S!J
RFLOOD MAX= 2500.
:t.i
Figure 3.41:
(b) Full Project
S!l
Minimum surface area covered with water
in the downstream reach. The maximum on
the y-axis is 2500 ha.
1 "r -I
.j.
I
f
I + I + I
. s + I + r
-124 -
r.: ~1 ti 6 E r: 3 ;. t·11~ >~= s s o u a .
t--------------------------------------------· i .
+ i + i
1 T
t + I
·l·
I +
. 5 ~-
·l·
I
11
t--------.,
T! t·1:
Sll
I -~------~~ t --------------------------
t c .L-.. -............................. _.. ....... __ ...._ _____ ...._ __ _
1 11 31 Sll
TIt·~
{a) No Project
(b) Full Project
Figure 3.42: The areal extent of deciduous and mixed
forest (less than 4000 feet elevation in
the Upper Susitna Basin). The maximum
on the y-axis is 65,000 ha.
L
r~
L
L
[
[
[
\'
'L~
[
r·
['
[
[
L
L
[
L
u
[
; ....
- i I ..;.
-125 -
+ I
·1·
L --------------------------------------------· + i
. 5 t
+
+ I + !
t
< .,.
~ i
+
·"!"
I -----------+ ----------------------------------·
. 5 ~-
1
' ""!"
I .,.
I + i + !
0~~~----~------~~.---~~~----~~
:!. 1 i 31
T I t-'E
(a) No Project
(b) Full Project
Figure 3.43: The areal extent of low mixed shrub (less
than 4,000 feet elevation) in the Upper
Susitna Basin. The maximum on the y-axis
is 100,000 ha.
-126
F ')F: f·~:~<= 'of. E S
. -• I
1·
' ! +
t
a
i
; --i
! -+ !
' ·t
I
+
I +
i!
Figure 3.44:
• ·; h
• I
TIr-E
3i
h
~· .I . '
' I
,,_" I
(a) No Project
~-·.. 1"1
' ~ ~ fl
• \ t" I r ~ I I 1.,1 I
f I ' I
" l t. ;. i--..,: I I (b) Full Project
' ' I f i
I tl 1 t
I II
I I !
~' I
Winter forage availability for moose in
the Upper Susitna Basin. The maximum on
the y-axis is 4,000,000 ha.
[
~-·.
L
r: !. _ _;
L
[
[
[
[
[
r,
~ '
f
,
-"
[
F'
L
[
t
G
E
L
[
L
L
u
L
-127 -
In the downstream reach (Devil Canyon to Talkeetna), the
within year variability (Figure 3.34) and maximum stage will be
significantly reduced as a result of the project. The effect on
riparian succession is to move the vegetation types to a new,
much less variable, dynamic equilibrium. In .summary, the deciduous
and mixed forest shows a constant increase (Figure 3.45b); tall
shrub shows an initial increase and then gradually decreases as it
succeeds to deciduous and mixed forest (Figure 3.46b); low mixed
shrub shows a gradual decline as it succeeds to tall shrub (Figure
3.47b); and the pioneer species which are subjected to considerable
variability (Figure 3.48a) under natural conditions show a constant
decline to very low levels once the project is introduced (Figure
3.48b).
3.6.3 Furbearers and Birds
Under the current assumptions in the model, the number of
beaver colonies associated with sloughs and side channels in the
downstream riparian zone oscillate about the carrying capacity for
both scenarios (Figure 3.49). A major reason for the population
being nearly equal to the carrying capacity is the way in which
carrying capacity is defined. Since the hydrology group provides
the length of shoreline with greater than . 5 m of i r.A free wa.tl9r
under the maximum ice cover, a major source of overwinter mortality
(i.e. beaver colonies frozen out due to insufficient water depth)
is incorporated in the determination of carrying capacity. In
reality, the carrying capacity during the den construction period
(i.e. late summer) is likely much higher, although the effective
carrying capacity (which the model generates) is decreased
substantially by the ice free depth criteria. Therefore, the only
process which could result in a substantial drop in the population
from the carrying capacity is a severe scouring event; and, in fact,
the model predicted drops in population are a consequence of ice
scouring events (Figures 3.49a and 3.49b).
-128 -
f·1!=t ::{: :Jl!llil ...... ~-.
. --!
+ ~-------------------~----------~---~---------·-+ I + i + i
. 5 t
I ,.
f +
+
!
a~~._~--~--~~~~._~~~--~·-.~~
2 1 S!l
T! t-it
i i
! --... -----------· t-----------~--------~---------
"1"
!
"'t
+
!
~ +
• :"! i
+
i +
+
+
(a) No Project
(b) Full Project
Figure 3.45: The areal extent of deciduous and mixed
forest in the downstream floodplain. The
maximum value on the y-axis is 3,500 ha.
[
[
[
C
L
L
L
L
[
r
L
[
l '
[
[
. ~ -i
..j.
i + I
.. i·-'t
~ I
+
. ... -I
i
~
!
I
T
I
+~,
I •. + !
... ..l. i
!J i ~
' ~ + I
1i
I t
I
'=a-... ,.~.,.J .. 'a ___ ......
+ I +
!
Figure 3.46:
-129 -
(a} No Project
3i
T I I'~
(b) Full Project_
T I I''E
The areal extent of tall shrub in the
downstream floodplain. The maximum value
on the y-axis is 300 ha.
i-:"
I
-1.
I ..;.
i + I
·~ r.
5 ·f. 1', + "'r I I + ~
I 1
t t ...... . ' \ t tol
I
::!.:!.
1:!.
Figure 3.47:
-130 -
!ill
T I rt:
21 31
The areal extent of low
downstream floodplain.
y-axis is 200 ha.
(a) No Project
(b) Full Project
mixed shrub in the
The maximum on the
l'
L
[
f'
L
L
[
~--
:!.
r
:;
--.,
I
r
b
...
I
t
I
' "1•
I
' ...
l
1 +
~"' • ~
II
! ;
i"J
! ...
+
. .... -I
..!.
I
t
! +
.;.
I
s + . i
... ""rl ,, ." ~
l • " ,.
!I
P•
fl . ' I ' I ! I
' I
I
I
''-l
11
,.
p
i'
li '.
I ' I I
! !
' I r I
ii
Figure 3.48:
2i
' ~.
! ,.
•' I ~
li! ,.
'I if.
T T k«"' ......
T I t'·'E
31
-. 131 -
~
~
'J !, ,,
lr
'!
It
I I r ,
f i
I I
' ..
I
' I ....
.. ••
I '• .. ~ ,. ..
.,
' '
so
(a) No Project
(b) Full Project
The areal extent of pioneer vegetation
in the downstream floodplain. The
maximum on the y-axis is 700 ha.
.. J.
. :t \
+
+
I + j
+ i
-132 -
~C C:tl. .: i . S ) t·iA ~=<= "! !l .
E: C A F: ( 1 : .. S ) f'-1!~ ::-::: 1.! !l ..
(a) No Project
a~~-~~~_..-~-.~~._~~~~~~~~~~
i
. s +
i +
c
I
I +
+
i + ' '
i
11
T I 1''£
E: C 0 L ( ! ,. S ) t·1J,:J ::-::: l.f !l ..
B•:HR(i,S) l"·it=~::<: 't'!l.
(b) Full Project
11 21 31
T I 1''£
Figure 3.49: Beaver colonies utilizing the sloughs and
side channels (solid line) and the
corresponding carrying capacity (broken
line) in the downstream riparian zone.
The maximum on the y-axis is 40 colonies.
[
[
r·
[
~--
[
[
[
[
[
r ,
L
L
L
[
c
[
[
[
b
c
-133 -
The apparent stabilization in the beaver population after
project construction in year l2 (Figure 3.49b) is a direct
consequence of the reduction in ice scouring events. It is
inte~esting to note the "stable" population is lower, on average,
with the full project situation (i.e. 25 c.olonies versus 35 colonies),
although with a less dramatic shift from year to year. This is a
direct consequence of a reduction in the number of shoreline miles
meeting the ice free depth criteria, as determined by the hydrology
submodel (see Figure 3.39).
The main channel colonies, despite a viable carrying
capacity, are not in evidence for· the no project scenario. This is
a consequence of both ice scouring and wide fluctuations in stage,
which, in concert, result in a zero beaver colony population in
the spring. (What is not shown here is the fact that beaver colonies
are established along the main channel in the summer, but are
destroyed by the above mentioned hydrologic events.) For the full
project scenario, the reduction in tne magnitude of the scouring
event, as well as the reduction in stage fluctuation over the year,
result in a viable, although small, main channel population (about
two colonies -Figure 3.50). It is significantly lower than the
carrying capacity since the model prediction shown is for after
the impact of stage fluctuation on the cnlnniPs.
The model predictions for marten are essentially the same
for both scenarios (Figure 3.51). The population quickly reaches
its maximum density and, as such, is directly dependent on changes
in the amount of forest habitat. The loss of forest habitat, due
to the project impoundments, accounts for the slight drop in the
population after year 11 (Figure 3.5lb).
As described in the submodel description, the prediction of
bird territories is a direct functionofhabitat availability.
Therefore, any change in any one of the habitats identified as
important to a particular bird species (see Table 3.9) will result
in a proportionate change in the predicted number of bird territories o
; ~ -~
4.
I
' ~
i + l + ! .sf
I + i + !
-134 -
r-11:: >::= 2 s ~
t-1H::-:::: 2 S ..
(a) No Project
~--~,---~--~ -----¥--~-1 . .,"'_
+
1. T
..j.
I
4·
i + i + i
' . 51"
+
I + i
1! = 1 3i su
T It·iE
:;: C 0 L ( 2 .. 5 ) r·iA ::-::: 2 !% • s c ~~ R •: 2 .. s ;. r·N ::-:;: 2 s .
(b) Full Project
~---,------------------1 +
0 l,,,,,,,,,,,'ll"i"":""",,,,;7,:::::. =:-:. :
1:!.
Figure 3. 50:
2 1 31 so
T H·1E
Beaver colonies utilizing the main
channel (solid line) and their carrying
capacity (broken line) in the downstream
riparian zone. The maximum on the y-axis
is 25 colonies.
[
[
r·
L
c
c
[
[
c
L
L
[
[
[
[
[
[
l'
[
c
c
E
[
b
[
. ~ -' ' ' +
-+ fi8"'si~
! ..
1~
! + !
. s ·+
j
+
t
! + !
t a , .
-135 -
~~--~------------------------------------
1:!. 21 Sll
TIr-E
r o r t1 r1~ >::= 1 tl uu u .
!T
+ ! + ~ ~~~----~-----------------------------------' .....
I• ·1~
I
5 +
I . ·+
+ ! + I +
T If'~
(a) No Project
(b) Full Project
Figure 3.51: Total marten population in the modelled
project area. The maximum on the y-axis
is 10,000.
-136 -
This is evidenced for brown creeper (Figure 3.52), northern water
thrush (Figure 3.53), and the total number of bird territories
(Figure 3.54). These figures are nothing more than a cumulative
surrogate indicator appropriate to birds demonstrating cumulative
changes in the various land classifications as a consequence of the
[
r
[
[
project. It should be noted that although the drop in bird .. "
territories is small relative to the maximum of 4 x 10 6 (Figure 3.54),[
that drop does represent tens of thousands of birds and should not be
viewed as insignificant.
3.6.4 Moose
The post harvest fall moose population appears to increase
with the project (Figure 3.55b). This occurs because the simulated
grizzly bear population shows a decline (Figure 3.61). This
apparently results in the reduction of bear predation on moose
(Figure 3.57b). In the model, the simulated wolf population ~s
unaffected by the project (Figures 3.56a and 3.56b). Wolf predation
on moose is also unaffected (Figure 3.57).
The age ratio (calves/100 cows) shows a more rapid increase
under the full project scenario (Figure 3.58), indicating that while
the population numbers remain unchanged, there is a shift to younger
age distribution. The moose harvest shows a slightly different
pattern between scenarios, but the absolute numbers are similar
(Figure 3. 59) .
3.6.5 Bears
The total population of bears in the study area over the
first 50 years of the simulation with no project remair .3 stable
(Figure 3.60). However, under full development, there is a marked
drop of the black bear population and a lesser dropforbrown bears
(Figure 3. 61) .
r
b
[
[
[
[
[
[
[
[
[
r:
t
u
L
[
[
l
.
'
6
C
i ..,. -' l
.J.
-137 -
L--------------------------------------------· + i
-l·
i + I
j
t; ·T ... i
+
.,.
!
i
f
!
(a) No Project
a+·~-..-~--~._~~-~-.~~~~~~+-~-
1 i :1. 21 31
T! 1''£
i ~
-i .
~ .
f------~-~~~---------------------------------·
+ ! + i
j . 5 ·-t
!
i + l + i + I + i
(b) Full Project
~ i 3i Sll
T I ri:
Figure 3.52: Number of bird territories associated
with brown creeper in the total modelled
area-~ The maximuni. on the y-axis is
150,000.
i ...
-i l
i
-138 -
1------~-------------------------------------·
+ I + I
i
1
~ .l. . !I •
l + i
i"
I +
+
i ...
-1
1i 21 3:i.
T I r"E
E: £:I R 0 0: 7 :0 r·1A >~:: 5 !l !l !l !l •
t-----------------~--------------------------· + i +
!
·+
T I t-iE
(a) No Project
(b) Full Project
Figure 3.53: The number of bird territories associated
with northern water thrush in the total
modelled area. The maximum on the y-axis
is 50,000.
[
[
r-
L
[
B
t
[
[
c
[
L
L
l
I l. ~
c·
[
-139 -
' ~ -!
i . 1--------------------------------------------·
''!'
+
f -+ . :J i
t + i + l
. ~
-i
..j.
i
(a) No Project
t----------~---------------------------------· +
I + .
I
5 ·i·
I
-1·
! +
(b) Full Project
Figure 3.54: The total number of bird territories
associated with the area represented
by the model. The maximum on the y-axis
is 4 x 10 6 •
. .,. -' I
' ..;.
!
·+
+ i + !
~ +
:! i
' ·;-
PSIPOP MAX= 1DUDil.
T I l'i:
PSTPOP MAX= 1DUUO.
2 i 3:!.
T I i'i:
-140 -
I J . ,.
'
• i_/
... ~ t •
' r ~ t !
I ' t
I •
•
'
' •'
• ''•
I ' • i
• I ~ ' t I
' •
(a) No Project
(b) Full Project
Figure 3.55: Post harvest fall moose population. The
maximum on the y-axis is 10,000 animals.
[
[
L
[.
[
L
L
\ '
L
l
[ i
I
I' 5
ll
.---,
I
'-'
' :
[
E
[
-i
' ...
i
i ;
+ .
I + i I .,.
-·
-141 -
...
f I • i ~~ ..
. • .. ... (a) No Project
+ ' ' • i / !··
f.
i + !
i'
I
·~
:!. ii 21.
T I i'".t
31
' '
' . ._,
SB
. . ..
·~"' .; •
(b) Full Project
I r
". • .1
11 2:!. 3!
T I i"·".t
Figure 3.56: Wolf population •
. y-axis is 50.
50
The maximum on the
+
f + ~\ I .J._,-"" :.·
t::~::ILL J··1:=t::{:: 2U!JU.
T!.,.fK!LL. t·1H>::= ~UUU.
'"~; I, -\
;.d" 1\ ./ ..,,. ... \~J-"~t .. ,.; s ~*..1
I + !
1 ./ ---/
...
' . "
i _,.-"""' ~-----f.. -· _... ~
I
!1 21 31
T I t"lt
8 K I L L r·1!=t >~= : !l 1! ll .
1.-! + i .j.
! + I ~-,/.
I .,.-• '1. ! I ,
5 ..:.,..;
! + I
. .j.
I
! +
I,-"
f'
i
TI.,Jt<!LL t·1H::<= ~!IUD.
li 2 1
T I !'it
-142 -
(a) No Project
(b) Full Project
..... _.
SIJ
Figure 3.57: Bear kills (upper line) and wolf kills
(lower line) of moose. The maximum on
the y-axis is 2,000 animals.
[
[
c
[
L
[
=
[
i =
!;
I
.l.
! + I + I
t
r.: !~E. f: r-~::,::{: 1 au.
s t >=:F: r·i!:Z ::<= 1 a u .
~ ~~ j-.. ... ,.-, ............ .....
~ ~~ .~ J
I
-143 -
a ++~~++++~~~~~~~~~~~~~~~~~~~~
:l.i S!l
T I l''f:
A G E ~: t-1~ }::: 1 il !1 .
S E :~<R t·1::a_::<: 1 !lll . . .. • !
+
t
-+ ~ ! '-" ..... .,,. \ .i· \ i' ~/
I
! ! 1
Figure 3.58:
I •-.
31 S!l
TIt·~
Age·ratio (calves/100
(males/100 females)
y-axis is 100.
(a) No Project
(b) Full Project
cows) and sex ratio
The maximum on the
! T
4.
1
-144 -
+ I~
J I I
I I
f I • I I"~ t .. ., + ~1· .. t I I :
1r /
5 .f II f. I A f ~ I
I I / 1
1 11 ,-I f ~,! 1 ~ I I -.. J I 1 1 r
' \ 1 f 1 1 \;~1 : II I I I .l, _. • I .. r ... . I L' , ~ ~,..·,/ ~., ·~-, , . ' . ,_, ! ; :1 ,'1 ~ : I (~I • • j \ 1 ~ I I I f
-t ~-! • .. / I~· I .. ,
+ I
:!. 11 31 Sll
·rHAr:V t11D:;: i!!iD.
' '" .. -1 I If
1i
I
i
I
f ; I
r. : ~ 'I ~" :~ ,.. .. '
I IH / 1 I \ i l \ •.JI• ! I : I f ~
I
\
r • • I I \ I
I '.1 I I I ~
I ~ : : I I
' t
\ ' ...
T I t·t:
~ I I • I / I I
'• I ~' .
Sll
(a) No Project
(b) Full Project
Figure 3.59: Moose harvest. The maximum on the y-ax·is
is 250 animals.
[
[
[
[
[
[
L
L
" '
[
[
[
[
c
c
E
[
L
I +
t
j
..l.
!
;..;.·-. ··.··-. "'" , .. -2 5 u .
-145 -
(a} No Project
a ~-~ .. ~~~~~~~++~~~~~~~~++~~~~~~
:!. i:!.
' .. -i ~
-!i ~ ' I r •. , ~~, ,,
.'II l ,, J -+ ~~~!.~ i ~ ' ,,. I II + ~..r 'i' ~; ' ..
1 r -r r.
+ \
s .1. \
I + I ;
I + '
., ., -..
Figure 3.60:
...
511
(b} Full Pro-ject
...... ., ___ ~
.,--------------~-----
T I i'i:
Black bear :.opulations. The maximum on
the y-axis is 250 animals.
·+
-146 -
(a) No Project
T! l'i:
o;. T ri r·~ ::-:;: 2 S !l.
•
Figure 3.61:
(b) Full Project
T ! l'iE
Brown bear populations. The maximum on
the y-axis is 250 animals.
[
[
[
c
[
[
[
[
f
[
[~
[
[
[
[
c
n
6
E
c
E
[
c
6
r
-147 -
For black bears, the decrease in populationismostly
attributable to decreased reproduction (an increase in the
reproductive interval) and increased mortality of cubs and
yearlings. Both these processes are controlled by the food
indices; however, the reduction in spring food availability from
inundation shows a more dramatic response (Figure 3Q62).
For brown bears, the slight decrease in population is
attributable to the increased dispersal from disturbance. There
is a marked increase in recreational use (Figure 3.38b) resulting
in an increase in the dispersal rate, leading to a decline in
total population (Figure 3.61).
[l
-+
+ i
t
l
·+ i
:!.i
T I r'f:
t::SF t-1f:::(: C.
E: '.•.IF t·iil ;<;: 2 .
T I i''f:
-148 -
(a) No Project
31
(b) Full Project
Figure 3.62: Index of summer (solid line) and winter
(broken line) black bear food. The maximum
on the y-axis is 2.0.
[
I
!__
[
\ '
[;
0
c
L
[
[
r L_.
[
[
[
5
c
t
L
[
6
-149 -
4.0 CONCEPTUAL MODEL
The looking outward matrix (Table 2.5) provides the
framework for linking the component submodels. The completely
integrated model is a complex set of relationships within and
between submodels. To gain a broad understanding of the major
processes included in the model, the simulation model has been
translated through a process of simplification and compression
into a conceptual model of the terrestrial environment in the
Susitna Basin (Figure 4.1).
In the conceptual model, the major components (boxes)
and the major linkages (arrows) represent the processes and
information transfers considered to be important to understanding
the biophysical system in the Susitna Basin. In the diagram
(Figure 4.1), solid lines represent linkages that are included
in the numerical simulation model; broken lines represent
critical linkages that are not presently included into the
numerical simulation.model.
The model depicted in Figure 4.1 represents an inter-
disciplinary perspective of the potential impact of the Susitna
hydroelectric project on the terrestrial environment in the
Susitna Basin. As such, it provides an overall framework for
assessing deficiencies in our current understanding.
nesting, resting ,-----
1
I
I
I
I
I
I
I
habitat ... ..
I
Waterfowl
Passerlnes
___________________ J ..
Rec:reatlon
Susltno
Hydroeledrlc 1---"._._.. ~
Project
Figure 4.1:
Human
dlsturbonu dl spersal summer f·eedlng __ ..._
vehlc:les,alr-1----------!-------------,. c;roft, people . food ~vest
Land for
fa~:UIIIes,
roods,
reservoirs
Flows
vege lotion alteration . """
"" Vegetation ...
home range.., Bears
~ ..._-r--:r---'
predation ·food
''IF
food ~ .. .4~
I home range.... Moose t....._
snow depth ~ l..,.~orvest Erosion I _ _J
Sedimentation
Flooding
rF.;;...;;.;;:;~·~L __ _j
habitat ...
flooding of colonies . =::
lc:e
Regime
l<:e sc;ourlng
Ice thickness
flooding of nest sites ..
Climatic
Effects
.. Raptors
~arvest ,
Hunting .
Trapping
Conceptual model of major components and linkages included in
the model of the terrestrial environment in the Susitna Basin.
__ ,
~-.
'
I-'
Ul
0
[
~
[
[
l-
l~
[
c
[
[
[
6
[
c
[
[
L
b
r
-151 -
5.0 MITIGATION PLANNING
At the mitigation planning workshop held February 28 to
March 2, 1983, discussion centered around five major areas:
mitigation, monitoring, informationneeds., planned studies, and
model refinements. This section reports the discussion that
took place in each of the subgroups.
5.1 Physical Processes/Development/Recreation
5.1.1 Model Refinements
5.1.1.1 Recreation
Currently, the model contains little credible information
with respect to recreation. Information available (in FERC
License Application,Exhibit E, Chapter 7) on existing or future
recreational use in terms of numbers of use days or amounts of
land needed appears to be unreliable. Data on current use and
credible projections of future use and need are critical to
better understanding of the impact of recreation on wildlife in
the Susitna Basin.
5.1.1.2 Development and Land Use
To adequately reflect habitat disturbance and loss, the
model must use accurate up to date information about various
project features. This is particularly true of access road
locations, areas alienated by the activities described in Table
3.1, and air, road, .and train traffic estimates. Current
estimates are based on data from the FERC License Application,
Exhibit E, Chapter 3.
-152 -
At present, the model contains only scanty information
about current land use patterns in the study area. Because of
the dynamic nature of land ownership in the area brought about
primarily by the Alaska Native Claims Settlement Act, it is
extremely difficult to make projections about £uture land use
patterns. However, a credible development scenario requires
that the model make projections about changing land use patterns
with and without the project. This is inadequately represented
in the present model.
5.1.1.3 Physical Processes
Restructuring of Ice Processes:
The model contains a simplistic representation of the
positioning of the ice front, the formationof ice cover, spring
break up, and ice scouring with its subsequent impact on
vegetation. While this part of the model must be refined, there
is still considerable uncertainty surrounding the mechanisms
affecting the ice processes. As the uncertainty is resolved
through further hydrologic, hydraulic, and ice studies, the
model will be refined.
Spatial Resolution in the Downs~ream Reach (Devil Canyon to
Talkeetna)
At present, the downstream reach is represented in the
model by a single spatial unit. It is now clear that this is
inadequate. This reach needs to be divided into a number (not
less than five) of smaller reaches. In addition, it appears
desirable to represent t~e sloughs explicitly within each of the
smaller reaches.
I l~
r -.
l~
[
c
[
u
0
G
[
L
I,
L
L
l~
_j
I
I"
F
b
c
[
~
-153 -
Overwinter Habitat for Beaver
At present, the suitability of slough, side channel,
and mainstem habitats for beaver is indirectly related to flow.
In the model, the amount of suitable overwintering habitat is
functionally related to stage. However, this relationship is
a crude hypothesis and does not adequately represent the
underlying hydraulic processes. A more realistic representation
requires a more detailed spatial resolution of sloughs and the
dynamics of groundwater inflow as influenced by main channel
stage.
Climatic Effects
The importance of· climatic effects to understanding
processes that might be affected by the project can not be
overstated. The most important climatic influences are snow
and ice. The interrelationship between the ice regime, flow,
and vegetation was discussed earlier.
Snow, or rather the amount of snow on the ground, affects
the ability of moose and caribou to utilize winter range. In
the model, the amount of snow on the ground is stochastically
generated and does not provide a realistic representation of
what actually occurs. What is required is the amount of snow
on the ground by elevation class. Anal ternate approach is to
use a more robust snow model similar to one developed by McNamee
{1982) for simulating the effect of snow in elk dynamics. Such
a model consists of three components: snowfall, snowmelt, and
snow interception. In the simplest version of the model, snow
is assumed to be general in nature, such that snow depth (not
density, crusting, etc.) would be the only influence on ungulate
dynamics. The general model would be:
SN t = SN t l-MR * SR * f(CC ) +SOt * f{CC ) s, s, -s s s
-154 -
where,
SN = snow depth on site s in time step t; s,t
!1R = maximum snowmelt;
SRs = snowmelt factor specific to site characteristics
(e.g. elevation) ;
sot = snowfall; and
ccs = crown closure.
In simple terms, the model suggests that the snow depth in a
given time step is equal to what was there the time step before
less what has melted plus what has fallen through to the ground.
Work of Harestad and Bunnell (1981) relates the level of snow
interception to snowfall and canopy closure; the work 'of Haverly
et al. (1978) and Leaf and Brink (1973) can provide guides for
defining snowmelt. A similar model needs to be developed to
better understand how moose and caribou will adapt to the loss
of winter range as a result of the impoundments.
5.1.2 Information Needs
There are four major information needs related to the
model refinements:
a) better estimates of current and future recreational
use;
b) better est~mates of the maximum amount of suitable
overwintering habitat for beaver in each of the
slough, side channel, and mainstem habitats;
c) data on snow accumulation by elevation in the Upper
Susitna Basin; and
r
L
r~
L.
[
[
I
)
l_ -·
r .
I Lj
[
r
C
l
[
[j
c
L
c
[
[
L
l~
[
[
[
r
[
[
[
[
[
c
D
o
c
[
[
L
6
C
-155 -
d) data on river morphology in relation to water surface
area.
Of these needs, the last is of critical importance.
Currently, the model represents fourteen vegetation types, one
of which is designated water. In its simplest division, the
water is made up of three qualitatively different aquatic
habitats: slough, side channel, and mainstem. For a given
stage at any transect along the river, the model needs to
predict the proportion of the transect that is comprised of each
of the terrestrial vegetation and aquatic habitat types. Both
the data and the conceptual understanding to do this are currently
lacking.
5.1.3 Mitigation
For recreation, concern centers around the maintenance
and enhancement of recreational opportunities. Specific concern
is focused on canoeing and kayaking.
Existing and future land use pattern may conflict with
proposed mitigation measures. Two examples are: potential bear
mitigation at Prairie Creek may conflict with private development,
and the burning and clearing for moose may be prevented if there
are competing land uses.
It is also possible that the plans to set aside twelve
sloughs for aquatic mitigation may conflict with beaver utilization
of the same areas.
-156 -
5.2 Vegetation
Workshop discussions concerning model refinements and
information needed to represent vegetation changes associated with
the project, studies planned or required to provide that information,
and additional work with respect to mitigation and monitoring
activities are summarized ·bedow. While the studies described are
vegetation oriented, much of the work is being done to provide
information to assess project impacts on moose and to better plan
mitigation activities for those impacts.
5.2.1 Model Refinements/Information Needs
Information needs associated with vegetation can be
divided into two major categories: information required to better
define project related impacts, and information required to determine
appropriate mitigation activities. Impact related information
includes:
1) what vegetation do wildlife need and use;
2) what vegetation is currently available; and
3) what vegetation will be lost as a result of project
construction and subsequent operation.
Mitigation related information includes:
1) what areas in the Susitna Basin do wildlife use
that could be manipulated in some way; and
2) how will browse production and wildlife use increase
with d~fferent types of manipulation in various
vegetation types.
r-~ L
[
[
E
[
[
c
c
[
,-
L
r
[
[i
E
r=
h
L
[
-157 -
A number of refinements to the current vegetation submodel
have been discussed. The two major refinements involve better
spatial representation of the project area (especially the riparian
zone below Devil Canyon), and a better representation of ice
processes and their effects on riparian successi9n. Less important
model refinements include better representations of development
activities, wildlife food, and dynamics of upland vegetation.
5.2.1.1 Spatial Resolution
The spatial units and land classification system in the
model are compromises. Clear suggestions for improvement emerged
at the workshop with respect to birds (more detailed resolution
of vertical stratification in the land classification sy~tem) and
beaver (more detailed spatial resolution of vegetation in areas
close to channels and sloughs). The need for spatial units more
appropriate for moose (e.g. winter range) was also discussed at
the workshop. These issues must be resolved before proceeding to
a more precise estimate of variables within various spatial units
.and vegetation types.
5.2.1.2 Ice Processes and Riparian succession
The model currently represents riparian vegetation and
succession and the effects of ice processes very simplistically.
The assumptions incorporated in the model represent hypotheses
about ice process effects but they are largely untested. The
representation of these succession/disturbance processes could be
greatly improved if the riparian vegetation and channel morphology
were incorporated in more detail spatially and if work was
initiated to study ice processes. The aquatic assessment of the
Susitna project is utilizing hydraulic simulation mnnPls and
supporting channel cross section data for instream flow studies
and also has need to conduct ice process related studies. A
cooperative effort between the aquatic and terrestrial assessment
groups could be mutually beneficial and should be considered.
-158 -
5.2.1.3 Resolution of Development Activities
Land is removed for development activities from various
land classes based on the relative proportions in the respective
spatial units or, in the case of roads, based on proportions
specific to a given route. The model could be refined to provide
additional activities or to provide a finer resolution of the land
class changes associated with an activity given its specific
location within a spatial unit. An example is the transfer of
land in the impoundment spatial areas to the water class. This
[
[
transfer iS CUrrently based On the development SUbmodel IS CalCUlatiOn r,
of land cleared for vegetation, rather than on a calculation of the '
amount of area actually covered by water.
5.2.1.4 Wildlife Food
Currently, the model simulates the variation in browse
standing crop and berry production as a random process. This
-simple representation could be improved by adding mechanisms that
incorporate the effects of consumption of vegetation by wildlife.
This is particularly true in the case of moose consumption of
browse and to some extent, beaver alteration of habitat in the
riparian zone. Further improvements in the model would result
if the productivity of browse and berries can be functionally
related to climatic variables such as temperature, snowfall, or
total precipitation. However, current understanding of the
E
c
determinants of productivity in the area may not be sufficient to b
fully develop these relationships.
~.2.1.5 Dynamics of Upland Vegetation
The current hypothesis is that the areas in various upland
land classes are constant except for changes associated with
specific development activities or vegetation manipulation actions.
While this is a weak assumption, current understanding of upland
successional processes is not sufficient to suggest a more dynamic
approach.
[
L
L
l
[
[~
[
[
r~
[
~
[
c
[
c
6
E
-1S9 -
The most serious drawback of this approach may be an
underestimate of the importance of natural fire in the area along
with its consequent effects on the natural variability of wildlife
habitat. Van Cleve and Viereck (1981) have stated that:
"The taiga of interior Alaska is dominated by young
stands in various stages of succession -mature
stands of over 200 years in age are rare. Fire is
the main cause of the young ages of the stands -
in some areas fire that kills all of the above
ground vegetation can be expected every SO -100
yea~s."
If this is the situation in the study area, the natural
fire regime needs to be represented in a SO year simulation. The
long-term habitat value of inundated areas may not be fairly
represented by their current species composition if fire periodically
converts them to earlier successional stages in the absence of
inundation.
S.2.2 Planned Studies
Vegetation studies planned for the coming field SP-nsnn
address the information needed to better define impacts of the
project above the Devil Canyon site (i.e. not changes in r~parian
vegetation resulting from project operation) and associated
mitigation measures.
S.2.2.1 Phenology
It has been hypothesized that early green-up of vegetation
at lower elevations is a primary reason why a lot of moose are
found in the proposed impoundment area in early spring. It has
been further hypothesized that inundation of this area could result
in a shortage of moose browse during this period. The study would
consist of running transects down elevational gradients to the
river and noting phenological stage by species and utilization by
-160 -
moose if. evident. Results of the study should better define project
impacts on early spring food supply for moose.
5.2.2.2 Food Habits
This study will help define what vegetation moose are eating
at different times of the year. The study involves fecal samples .
for percent composition by vegetation species. Some fecal samples
collected during the winter.and early summer are already available
for analysis. Additional samples will be collected this spring
and in late summer. This information will be used to define project
impacts and as a basis for designing mitigation activities.
5.2.2.3 Browse Sampling
The purpose of this study is to determine the amount of
browse in different vegetation types and the energy content of that
browse. A pilot project will be conducted during the upcoming field
season to determine the best techniques to use with the full study
to be conducted the following summer. Prior to the pilot project,
the people doing this study will meet with several moose biologists
to determine the appropriate measure of browse (e.g. current year's
growth, to point of average browse, etc.). This information will be
used in conjunction with the carrying capacity work described below.
5.2.2.4 Browse Mapping
Browse mapping will be done to evaluate how much browse
(areal extent of vegetation types) is currently available and how
much will be lost as a result of project activities. A core area
around and including the impoundment area will be ~apped at a
scale of 1:24,000 and a larger area will be mapped at a scale of
1:63,360. The mapping contractor will work with vegetation
specialists, and moose and bear biologists to identify appropriate
vegetation mapping categories.
[~
[
[
L
[
[
[
E
[
E
[
c
[
[
[
[
[
r~
f~
[
[
r .
c
c
8
[
L
[
[
r:
c
-161 -
5.2.2.5 Energetics Modelling
An energetics model for moose will be developed from an
existing model and-validated with informationfromthe Kenai
Peninsula. The modelling will help define browse requirements
for a moose in this area. Results of the modelling will be used
in the carrying capacity work described below.
5.2.2.6 Carrying Capacity
The browse sampling, browse mapping, and energetics modelling
results will be integrated to determine current carrying capacity
of the Susitna area for moose and the reduction in carrying capacity
caused by project activities. These results will help define
mitigation needs.
5.2.2.7 Monitor BLM Burn Site
The BLM is planning to conduct a control burn in the Alphabet
Hills area. Vegetation sampling pre-burn will be done to initially
characterize the area with respect to canopy cover, tree and shrub
density, and browse production. Repeated sampling following the
burn will provide information on successions and browse production
following different severities of burns in different vegetation
types. This information should be very useful for evaluating the
potential of using burning as a mitigation measure for lost moose
habitat. If burning is shown to be an effective mitigation tool,
this study should also help determine what vegetation types should
be burned and how severe a burn should be planned to achieve a
maximum increase in browse production.
5.2.3 Needed Studies
In addition to the studies already planned, a number of
additional studies were discussed which would help to better define
project impacts and possible mitigation alternatives.
-162 -
5.2.3.1 Monitor Other Vegetation Manipulations
A number of areas downstream from Devil Canyon have been
disturbed in the past for 'different reasons. Some vegetation
sampling in these areas would provide information on succession
and browse production subsequent to these disturbances. In
addition, ADF & G is planning a chaining operation in the Palmer
area. If pre-and post-chaining sampling could be arranged at
this site, it would provide information to evaluate chaining as
a possible mitigation alternative.
5.2.3.2 Ice Processes and Riparian Vegetation
The effects of ice pr0cesses on riparian vegetation and
the potential impact of regulated flows (and associated changes
in ice processes) on riparian succession are not well understood.
Prediction of project impacts downstream from Devil Canyon and
L
design of sul:table mitigation measures requires a better understanding p
and representation of these processes. Currently, available l_;
geomorphological cross section information with associated
vegetation information could be used to better represent what
vegetation gets scoured at different flow and ice levels. Periodic
surveying of data at these cross sections could be used in
conjunction with ice surveys to define how ice processes affect
different vegetation types.
Aquatic and terrestrial environmental assessment studies
of the Susitna project, which are currently being conducted
independently, require much of the same information. This is
especially true of the hydraulic, hydrologic, and geomorphological
information produced by Acres American and ~ & M Consultants. The
two studies also have similar information needs, such as effects
of ice processes on fish and wildlife habitat and changes in these
processes post-project. Some coordination between these groups
to cooperatively develop and use this information could be mutually
beneficial and would result in analyses which are more logically
consistent and compatible with each other and therefore more useful
to APA.
[
c
c
f'
L
L
r
L
r.
Li
I
-163 -
5.2.4 Mitigation and Monitoring
Mitigation and monitoring activities for vegetation losses
which are addressed in the PERC license application are justified
primarily as they pertain to impacts on moose.. The studies
discussed above should better define these impacts and provide
valuable information for designing mitigation.
A concern was expressed, however, that the independent
aquatic and terrestrial assessment studies may result in
inconsistent mitigation recommendations (e.g. fish mitigation
release scenarios, which are detrimental to downstream vegetation
and wildlife). While these conflicts may ultimately be unavoidable,
a cross analysis of mitigation options by the other assessment
group would at least indicate potential areas of conflict early in
the mitigation process while a variety of mitigation options are
still available. If all major environmental impacts are to be
adequately considered in the design, licensing, and operation of
the project, an integration of aquatic and terrestrial analysis
and design of mitigation activities should be started.
5.3 Furbearers and Birds
The following section summarizes workshop discussions
concerning model refinements and information needed to represent
the biology of the furbearer and bird system, studies needed to
provide some of that information, suggested mitigation strategies
to minimize potential impacts, and monitoring procedures that
would help evaluate the impact of a mitigation or other action.
-164 -
5.3.1 Beaver
5.3.1.1 Model Refinements
Habitat Definition
Currently, beaver habitat is structured as a function of
two major criteria: proportion of sloughs and side channels with
greater than . 5 m of ice-free water below the maximum ice .cover;
and the proportion of shoreline length with balsam poplar and
birch vegetation adjacent to it.
The first criteria is the key determinant in the
appropriateness of an area for beaver habitation. Reduction in
the amount of ice-free water would almost certainly result in
a direct reduction in the number of colonies that could be
supported in any given area.
The vegetation criteria lacks any firm hypothesis about
what aspects of vegetation (i.e. type, quality, quantity, and
location) make one area more suitable than another. To more
clearly define this criteria it was suggested that the
vegetation and furbearer subgroups take a detailed look at the
available river cross sections and attempt to better establish
the proportion of the various vegetation types that are found
within 40 m of the shoreline. The result will be a more precise
representation of the appropriate vegetation (i.e. balsam poplar
and birch) as it is now defined for beaver habitat, thereby
improving the vegetation submodel 's prediction of how the adjacent
vegetation characteristics might change after impoundment. This
will, in turn, improve thE: model's capability to predict how
beaver colonies might be impacted by alterations in vegetative
structure. It should be noted that it may be necessary to
complement the analysis of the available cross sections with some
ground truthing.
[
I
(
I"
L
t
[
l
[
c
[
L
[
[
[
r
r:
[
[
b
8
r l~
-165 -
There is also a need for refinement of the hydrology aspect
of the beaver habitat criteria. Evidently, beaver will not build
adenadjacent to water with velocities greater than approximately
4.4 ft/sec between mid-August and freeze-up; this velocity being
the maximum a beaver can effectively swim against for any prolonged
period. This added criteria will require the hydrology and
furbearer groups to coordinate their field programs to ensure some
velocity information is obtained for critical reaches of the
river.
Price Index
In the model, a potential major source of beaver mortality
is trapping success which is a direct result of an externally
set price index. A high price index could conceivably result in
a complete decimation of the beaver population in one year.
Historically, the price for beaver pelts has oscillated regularly
with a period on the order of 15 years. Since a period of high
trapping intensity in conjunction with a shift in the hydrology
of the region could result in a severe impact on beaver, the
p~rticipants suggested an oscillating price index be introduced
into the model.
5.3.1.2 Information Needs/Research
Overwinter Survival
Currently, one of the major data needs is actually
determining how many beaver colonies there are along the Susitna
River Basin and their overwinter survival. Therefore, a concerted
effort should be made to count the number of caches in the fall,
and the following spring (before and after break-up) to establish
what proportion of the colonies survived the winter. This
survival would be related to three major factors:
-166 -
1) the degree of ice scouring during break-up;
2) maintenance of ice-free water under the ice cover;
and
3) the change in quality of the food cache over the
winter period.
The impact of all three of these factors could be assessed
through the above proposed site visits. However, a first step in
this direction could be made this year by coordinating a planned
April -May visit by the hydrologists with one or two of the
researchers in the furbearer study.
The third factor is of special interest since it directly
relates to the need to better understand the re~ationship between
the beaver and the nearby vegetation. Different vegetation types
have very different overwintering qualities and could be a
determining factor in a beaver colony's survival.
Charac~erize Habi~a~
The quality of the food cache is directly related to the
availability of appropriate vegetation. There is a definite
need to better characterize what it is that makes an area good
for beaver. Therefore, site visits designed to count beaver
colonies and/or caches should also measure:
1) the vegetation available to the colony and of that
available, how much was utilized (i.e. what is
actually found in the cache);
2) the characteristics of the adjacent water body (i.e.
bank structure, water depths, depth of ice cover,
water velocity, etc.); and
[
r
[
r~
-L
\ .
l .
r
[
d
Q
-r ~
[
[
[
r
L
[
L
[
[
[
u
c
8
~
E
u
£
E
f'
b
b
[
-167 -
3) is there any evidence of trapping?
5.3.1.3 Mitigation
Besides trapping control, mitigation specifically for
beaver was judged to be a minor issue for the region between
Devil Canyon and Talkeetna. Generally, it was felt that changes
due to impoundment in this reach of the river would have a
positive impact on beaver and would likely increase the number
of potential beaver colonies. However, in light of this
prediction, there was considerable concern expressed regarding
proposed destruction of beaver darns in the 12 sloughs which have
been selected as optimal salmon rearing habitat by the fisheries
studies. The furbearer group felt other control options should
be expl:-ored and requested some coordination between the fisheries
and furbearer studies.
Given the predicted increase in beaver, it was also
suggested that this might be viewed as compensation out of kind
for the probable loss .of marten due to impoundment.
Monitoring·
The monitoring recommendations were very much related to
enhancing the information needs and research described earlier.
Specifically, these are:
a) cache counts in the fall;
b) determination of the overwinter colony survival by
counting the viable colonies pre-and post break-up;
c) continual observation and evaluation of the nearby
vegetation and its utilization;
-168 -
d) interactions between beaver and salmon, specifically
in rearing areas -does the existence and persistence
of a beaver darn have an identifiable impact on salmon
rearing ·success?; and
e) the level of trapping in the region -this requires a
survey coordinated with the Alaska Department of Fish
and Game to obtain better information on the intensity
of beaver trapping and associated harrassrnent.
5.3.2 Marten
5.3.2.1 Model Refinements
As it now stands, the marten population model is a
simplistic representation based on very little information.
Therefore, refinement of the model is not practical until some
of t~e critical information gaps are addressed.
5.3.2.2 Information Needs
Mart:en Habit;at;
Marten generally depend on the availability of forest
habitat for both cover and food and it is suspected that the
forest lost due to impoundment is prime habitat. However,.this
suspicion is based on very qualitative information that requires
further investigation. The recommended first step is to
coordinate a marten specialist with the vegetation group to
better characterize marten habitat and then direct themselves
to improving the methodology for detecting that habitat.
Determination of how much habitat is available in the region is
important to taking a first step at predicting the impact of
impoundment on marten.
[
['
\ .
r
8
E
[
L
c
[
[
c
f:
l;;c
r
[
rc.
L
[
[
r'
[
[
[
c
c
5
p
y
L
[
L
~--
6;
-169 -
Population
Once the habitat types have been identified, the marten
densities associated with each type should be established,
possibly expressed as high/low estimates. This would then
permit a first cut estimation of the probable loss due to
impoundment. In the longer term, there is a need to improve our
understanding on how marten relate functionally to the available
habitat (i.e. fecundity, mortality, dispersal, density dependencef
etc.) •
Trapping
Marten are very vulnerable to trapping. As with beaver,
there is a need to get better information on trapping intensity
and projections of future levels of effort.
5.3.2.3 Mitigation
Given marten's dependence on forested lands, attempts
should be made to minimize the reduction in forest land due to
impoundment. Once more information on the expected losses in
numbers is available, it should be brouqht to the attention of
ADF & G and the Alaska Board of Game. High losses may require
exploration of enhancement strategies or trapping regulations.
There was also a concern expressed regarding the proposed
burning of forest to generate more moose browse in the area.
This would definitely have a negative impact on marten and, if
implemented, should be monitored before and after the event.
5.3.3 Birds
5.3.3.1 Information Needs
For the raptors (primarily golden eagle), there is a need
for more information of the location and elevation of potential
nesting cliffs and existing nesting sites, primarily around the
-170 -
Watana reservoir, on or near the water's edge. Also, there is a
need to confirm the location of the bald eagle nests downstream
of Indian River. Currently, it is not clear which of the
documented sites are actual nest locations and which are alternate
nest locations. Also, there are some discrepancies between the
documented information and more recent observations.
For the purposes of possible mitigation, there is a need
to document the location, distribution, and number of cliffs and
exposed bedrock above the maximum reservoir level available for
possible modification to make additional potential nesting sites.
These cliffs need to be typified as to suitability for modification
and level of effort to do so.
There is also a need to refine the available information
on location and extent of potential bald eagle nesting sites,
primarily in riparian poplar stands and hillside white spruce.
These should also be assessed for current suitability and
potential for modification.
5.3.3.2 Mitigation/Monitoring
The major mitigation strategies for the raptors have
already been identified, namely the creation of new nesting
sites to compensate for losses due to construction and/or
impoundment. The success of this approach is not predictable
since it depends greatly on how the birds react both to the
new site and the actual disturbance activity. Therefore, it is
important that the nests and nest sites be monitored each
spring to assess the effectiveness of the modification (i.e. are
the new sites utilized?) and determine what further action
might be necessary, if any.
For swans, mitigation involves at least minimization of
the disturbance to the nesting and staging areas, if not total
avoidance of those areas. Monitoring would involve annual
[
r ,_'
r~
L
L
t
[
F
c
[
-171 -
observation of the swan's utilization of those areas as well as
conformance of the public and.project staff to established
disturbance criteria.
5.4 Moose
The following section summarizes workshop discussions
concerning additional .model refinements and information needed
to adequately represent the biology of moose in the Susitna area,.
studies either planned or needed to provide that information, and
additional work that requi.t:es planning as mitigation and
monitoring proceeds. It is important to note that much of the
corresponding discussion concerning the vegetation submodel is
directly applicable to moose.
5.4.1 Model Refinements
5.4.1.1 Spatial Definition
The present moose model represents an ill-defined area of
1200 mi 2 with an assumed distribution of vegetation types. This
representation can be improved quite easily in the following way.
Existing radiotelemetry data can be used to define a herd area
by drawing a line connecting the outermost (farthest from the
impoundment) radiotelemetry locations for each moose whose home
range ovelaps the impoundment area. Amounts of each vegetation
type within this herd area can then be determined from the
vegetation mapping that is to be done this spring and summer.
5.4.1.2 Bear Predation
There are three fundamental deficiencies in the representation
of bear predation. First, the model assumes that only brown bears
prey on moose. While it is known that black bear can and do take
moose, the extent to which this actually occurs in the Susitna
area is uncertain.
-172 -[
Second, while a mechanism is incorporated in the model to [
alter vulnerability of moose calves to bear predation as a function
of severity of the previous winter, this mechanism is not presently r
used. Studies in the Susitna area indicate lower calf/cow ratios
in years following heavy snowfall. The relationship is fairly ['
consistent except in one year during which there was a bear
removal program. In that year, the fall calf/cow ratio was high
despite a hard previous winter. These observations thus seem to
indicate a relationship between winte.r severity and vulnerability
of moose calves to bear predation. Unfortunately, the
observations are not sufficiently well quantified at this time
to allow incorporation in the model.
Finally, the brown bear submodel considers a population
that occupies a spatial area somewhat larger than the moose herd
area described above. A method is needed to determine what
proportion of the brown bears in the model should be considered
effective predators on moose in the defined herd area. Radio-.
telemetry information from the bear studies may be useful in this
regard.
5.4.1.3 Wolf Predation
The current representation of wolf dynamics has similar
deficiencies.. The spatial area occupied by the wolf population
represented in the model may not completely coincide with that
for moose. More careful definition of the proportion of the
wolf population actually preying on moose in the herd area
described above is needed.
The model wolf population is presently not affected by
any model variables pertaining to development. Mechanisms ~f
impact on the wolf population need to be considered more
carefully.
[
L
L
r
I
I
-173 -
Finally, wolf predation rates on moose in the model are
unaffected by moose density, caribou density, or winter severity§
all of which are thought to be important in determining the
number of moose taken.
5.4.1.4 Winter Mortality
Win-ter severi-ty
Both methods of calculating winter mortality in the moose
submodel use snow accumulation·as an index of winter severity.
At the present time, the value for snow accumulation is estimated
by the physical processes submodel from the mean and standard
deviation of accumulations reported at 12 stations in 4 months
(January, February, March, and April) for varying (by station)
numbers of years. These records need to be examined carefully
in the context of known historic patterns of moose mortality to
see if other combinations of months and/or stations might provide
a better estimate of winter severity. For example, the sum of
snow accumulations for the 4 month period may be a better index
of severity than the average value for the 4 months. Methods
for incorporating other factors (e.g. hardness of snow,
temperature) that contribute to wi nt.Pr sPvF>ri.ty should a.lso be
examined.
Win-ter Mor-tali-ty as a Func-tion of Snow Accumula-tion
The above examination of historic snow accumulation
patterns with respect to observed moose mortality should provide
information useful in constructing a more realistic relationship
bei..-qeen winter severity and winter mortality rates (Table 3.11).
Winr. er Mor r. ali r. y as a Func'tio n of Carrying Capacity
The second method of calculating winter mortality rates
for moose uses snow accumulation to modify forage availability.
-174 -
The present relationship (Figure 3.23) is largely hypothetical
and needs to be refined to represent explicitly two aspects of
this phenomenon. First, snow accumulation influences the
availability of forage in'the vertical dimension; that is,
different snow depths cover different proportions of potentially
available forage. The currently planned browse studies (see
vegetation submodel) will provide information useful in this
regard through vertical stratification of browse samples.
Second, snow accumulation influences the availability of
forage in the horizontal dimension; that is, different snow
depths restrict moose to different altitudes and/or cover types.
}
\.-
[
\ :
More intensive monitoring of radio-collared moose in the impoundment n
area .should provide additional information useful in better L:
defining habitat use relationships under different snow conditions.
Given the availability of certain proportions of the total
browse present, the moose model then requires two additional types
of information: the utility of the available browse in supporting
moose, and relationships between consumption rates and mortality.
The utility of the available browse to the moose population is
currently estimated on a biomass (kg dry weight) basis. The
total available biomass of browse species is divided by the
number of moose use _days (moose population times number of days
on the winter range) to obtain daily consumption rates per moose
(assuming that all available forage can be found and consumed) .
Digestible energy and nitrogen are probably better estimates of
diet suitability than biomass. The browse sampling program to
be initiated this summer will provide plant materials that will
be analyzed for digestible energy and nitrogen, which will then
be used in the model in place of biomass as a measure of the
quality and quantity of forage available.
The second step, estimating mortality rates from consumption
rates, is more problematic. One possibility is to use a bio-
energetics model along with the above data on forage quantity and
quality to estimate weight loss at different consumption levels.
[
[
[~
L.
l -
~
[
[
[
L
c
[
f
.,
-~
L
c
[
c
E
B
t
6
[
b
L
E
[
-175 -
Mortality rates would then be estimated for various levels of
weight loss. Note that this approach may also be useful if the
desired output from the model is simply an estimate of carrying
capacity. The bioenergetics model can be used to estimate daily
forage consumption rates (assuming that rnoose·foraging is bulk
limited by rumen volume, rather than by forage availability).
Estimates of available nitrogen and digestible energy can then
be divided by the daily consumption rates to obtain the number
of moose use days available.
5.4.1.5 Model Testing and Evaluation
The above needs for information and model refinement were
identified in the absence of extensive experience with the
present formulation of the model. Additional model testing and
evaluation by ADF & G personnel will likely identify other
refinements. The current version of the workshop model has been
.made available to ADF & G for this purpose.
5.4.2 Planned Studies
5.4.2.1 Moose
Moose radiotelemetry studies to date have been aimed at
better definition of the subpopulations using the Susitna Basin.
In response to the need for better habitat use information and
better definition of the home ranges of animals using the .
impoundment areas, monitoring schedules are presently being
changed. Radio-collared moose whose horne ranges overlap the
impoundment areas will be monitored twice weekly. Other radio-
collared animals will be located less frequently. In addition,
monitoring of radio-collared animals in the proposed burn area
in the Alphabet Hills will be continued. · Studies of the
utilization of this area pre-and post-burning should provide
valuable insights into the effectiveness of burning as a mitigation
technique. A lat_e winter census of the number of moose in the
proposed burn area is also planned.
-176 -
In addition to these studies, most of the planned.work
dealing with vegetation mapping and browse sampling is directly
applicable to moose (see vegetation submodel) .
[
5. 4. 2. 2 Wolves l-,
As mentioned above, one of the principal information needs [
regarding wolves is more careful definition of the numbers preying
on moose in the herd area being modelled. Radiotelemetry studies r,
aimed at better population definition will be continued. Additional L
food habits information directed toward better estimates of
predation rates will also be collected. Finally, results of a
separate study examining relationships between presence of prey
items in the wolf diet and occurrence of those same items in
fecal samples will be useful in estimating predation rates.
5.4.3 Needed Studies
In addition to the work briefly outlined above, the
following studies would be very useful in more carefully estimating
the potential impacts of Susitna hydroelectric development on
moose and in evaluating the potential effectiveness of various
mitigation measures.
1) A fall census of moose in the composition count areas
in the Upper Susitna Basin would provide a useful
check on parameter values currently used in the moose
submodel. Many of the current parameters were
estimated from a single census conducted in the fall
of 1980.
2) More intensive study of calves of cows that are· already
radio-collared would be useful in refining estimates of
the sources and magnitude of calf mortality (e.g.
predation by both bear species), as well as the importance
of dispersal from the Susitna area. Preliminary results
from radiotelemetry work suggest that movement out of the
c
r ..__,..
L
c
L
[
L
I
[
F
-177 -
area is more common than movement into the area. If this
is so, the Susitna area may serve as a source of individuals
for a region much broader than that expected to be directly
impacted by hydroelectric development.
3) Additional information on moose utilization of so-called
"rehabilitation" areas downstream of the dam sites would
be useful in evaluating the potential effectiveness of
various possible mitigation measures.
4) Plans need to be formulated to allow more intensive
monitoring of moose behavior during a severe winter,
should one occur.
5.4.4 Mitigation and Monitoring
A variety of other factors will eventually be important in
the specification of an adequate plan for mitigation and monitoring
of the impacts of hydroelectric development on moose. First, it
is important that mitigation options· other than vegetation
manipulation continue to be given adequate consideration. Second,
successful use of vegetation manipulation techniques will require
additional information on the relative merits of options such as
burning and crushing. These techniques need to:be evaluated more
carefully with respect to site-specific criteria influencing their
probable success in producing additional browse at times and
places where it can be utilized by moose. Finally, it must be
remembered that impacts on forage availability may not be the
principal effect on downstream moose. Destruction or modification
of critical habitats, such as islands used for calving, may be
more important for these animals. Additional work is needed in
assessing both the probable impacts of development on these
areas, as well as their importance to moose.
-178 -
5.5 Bears
5.5.1 Model Refinements
5.5.1.1 Bioenergetics and Foraging
•
Reproduction and natural mortality of cubs and yearlings,
which are food related, are two important factors influencing
the population dynamics of bears. To completely represent these
processes, the bioenergetic requirements and foraging behavior
of bears must be understood better than is currently possible.
For instance, the prediction of fat reserves (i.e. condition)
for a bear would involve the knowledge of at least the search
efficiency, handling time, and digestibility of the major food
items in the bear's diet. Unfortunately, the expense of bear
research precludes this level of knowledge in the near future.
5.5.1.2 Initial Equilibrium
The tactic of assuming an initial population equilibrium
and relating indices to this equilibrium level effectively reduces
the number of processes to be quantified. The drawback, however,
is that the assumed equilibrium condition is, at best, tenuous.
A concrete suggestion made at the workshop, that partially
addresses the drawback, was to explore the sensitivity of the
bear population in the study area (with and without the project)
to changes over time of the immigration from the outside "buffer"
population. Then, sensitivity to absolute changes in immigration
was explored at the workshop with the conclusion that impacts
will be more severe (i.e. greatest relative change in population
level with aLi without the project) when immigration is minimal.
Nevertheless, the ability to increase or decrease immigration
over time should be incorporated into the model.
r
(
[
[
[
L
L
L
L
[
[
[
[
[.
[
[
[
[
c
B
E
t
c
[
[
[
E
[
-179 -
5.5.1.3 Berry Production
Another shortcoming of the model is the portrayal of
·fluctuations in berry production. At present, the production
in each vegetation type is subject to random variation each
year. A more realistic approach would be to simulate a berry
failure every few years.
However, the information used to predict berry production
was derived indirectly from data on stems. Stems do not have a
direct relationship to berries in any particular year. As suchi
the data currently available on vegetation types and berry
production cannot be used with any confidence.
5.5.1.4 Spatial Resolution
The current spatial resolution is gross in comparison to
the finer scale processes that may impact bears. In particular,
disturbance of brown bears will not occur evenly over the entire
study area. For instance, localized areas of disturbance would
likely disperse more brown bears out of the study area (or deplete
them through nuisance kills) than a more diffuse disturbance.
While it may be desirable to develop a finer spatial
resolution in the Upper Susitna Basin, it may be possible to
disregard the downstream reach in the analysis. The downstream
area is markedly different in terms of patterns of bear use and
vegetatiqn. For example, it is suspected that downstream black
bears use predominantly salmon and Devil's club berries in the
late summer, both of which are unavailable to bears in the
Upper Susitna Basin impoundment areas., assuming th~t this
habitat in the downstream may understate the project's impacts.
-180 -
5.5.1.5 Prairie Creek Salmon Resource
Another spatial problem is the portrayal of Prairie Creek
as a food source outside the study area. If development at
Prairie Creek can indeed be attributed to the project, then it
can be argued that all bears that utilize the resource should be
included in the model and not only those that chiefly reside in
the study area.
The assumption that a doubling of 1980 recreational use
would mean that the salmon resource would be completely
eliminated is much too conservative. However, the assumption
that the Prairie Creek salmon represent only one-third of the
summer energy intake for bears that use the resource could easily
be understated. Both assumptions are highly speculative.
The model currently distributes the loss over the entire
bear population by reducing the summer food index. A refined
approach would be to reduce the reproductive potential of a
significant proportion of the female bears that use Prairie Creek.
This would cause the number of females that use Prairie Creek to
decline and the population as a whole would also be reduced.
5.5.1.6 Dispersal and Harvesting
The we1ghts for dispersal and harvesting presented in
Tables 3.17 to 3.20 are, at best, educated guesses of the
relative propensities of the various classes to disperse or be
harvested. It would be valuable to test the sensitivity of the
model to the assumed weights.
5.5.1.7 Composite Food Index
The composite food index does not adequately portray the
importance of both spring and summer food to bear reproduction;
both foods must be adequate in a given year or bears will be
unable to reproduce the following spring. Another reason for
f
L
L
[
-L
L"
['
8
[
c
G
c
c
c
[
L
L
L
L
' I • I
J
E
L
[
[
L
L
r:
-181 -
treating spring and summer food separately is because the
importance of predation on moose calves (spring food) is
unknown. The availability of moose calves in the spring may
be impacted by the project as much as vegetation.
5.5.2 Mitigation and Monitoring
The model demonstrates that the major mechanisms of impacts
for blaqk bears is the loss of spring habitat from inundation
which results in a larger reproductive interval and increased
mortality of cubs and yearlings. Obviously, habitat manipulation
as a mitigation measure should target upon the production of
spring foods that can be enhanced (Equisetum, small mammals,
skunk cabbage, roots and cottonwood buds) through increased
acreage of forest and pioneer vegetation types.
Further, monitoring· during the construction and post-project
stages should focus upon these predictiOns. In addition,
inundation will displace black bears from traditional denning
sites which, in the model, either experience a longer reproductive
interval or disperse from the study area. Monitoring of these
displaced bears should present a research opportunity to document
their behavior.
For brown bears, the major impact mechanism is dispersal
or associated mortality from disturbance generated by increased
human usage (e.g. recreational, hunting) of the study area.
Therefore, the model would indicate that such mitigation measures
as controlled access and the minimization or limitation of
disturbance would be effective. Unfortunately, the model does
not have sufficient spatial resoluti~n to aid in the specific
design of these measures. However, the planned development of
the Prairie Creek area may serve as an opportunity to monitor
the effects of both dispersal from disturbance and the subsequent
effect upon reproduction of the lost salmon food resource. On
the other hand, Prairie Creek is viewed by many participants as
a potential site for out-of-kind (preservation) mitigation.
-182 -
6.0 FUTURE WORK
The model that existed at the completion of the second
workshop held February 28 -March 2, 1983 was greatly improved
over the preliminary model constructed during the first workshop.
In particular, the moose submodel has a much sounder empirical
basis and the bear submodel has a more realistic structure. The
hydrology submodel has been improved to incorporate linkages
between the vegetation and furbearer submodels. The vegetation
submodel itself has a more reasonable representation of riparian
succession in the downstream reach.
The discussions in the subgroups were fruitful, as evidenced
[
[
[
L
[
by the material presented in the previous section (5.0) on mitigation n
planning. The workshops allowed for examination of current and C
future study programs in the context o·f · the model and mitigation
[ planning.
Future modelling and mitigation planning is dependent upon [
a reevaluation of the spatial and temporal structure. The geographical
areas into which the model is currently divided are too large to c
address some of the critical questions regarding moose, bears, beaver, '
and riparian succession. A new spatial representation must be
developed before much more effort is put into model refinement.
Future modelling and mitigation planning now depends upon a
program of effective coordination between the aquatic and terrestrial
programs. At meetings held in late March, a program of coodination
was proposed by LGL, ESSA, AEIDC, and R & M Consultants. One of
the first priorities of this program is to develop a common spatial
and temporal structure for the aquatic and terrestrial models.
It is currently planned to hold a workshop in the fall of
1983 to integrate the results of the 1983 summer field season into
the mitigation planning and modelling. The workshops and the
[
f'
lJ
L
[
[~
r-,
[
[
l-
r'
[
[
c
E
B
B
G
[
[
l
E
l
-183 -
modelling will continue to be the focus for the terrestrial
mitigation planning by adapting to new information and enhancing
collective understanding.
-184 -
7.0 REFERENCES
Ballard, W., J.S. Jackson, N. Tankersley, L. Aumiller, P. Hessing,
1983. Big Game Studies, Vol. 3, Moose-Upstream, Alaska Dept.
of Fish & Game Special Report to the Alaska Power Authority.
Blood, D.A., 1973. Variation in reproduction and productivity of
an enclosed herd of moose (Alces alces). XI International
Congress of Game Biologists, Stockholm.
PERC License Application, 1983. Exhibit E, Chapter 2, 202 pp.
pl~s appendices, tables, and figures.
PERC License Application, 1983. Exhibit E, Chapter 3, 603 pp.
plus appendices, tables, and figures.
PERC License Application, 1983. Exhibit E, Chapter 7, 117 pp.
plus appendices, tables, figures, and references.
Harestad, A.S. and F.L. Bunnell, 1981. Snow: canopy cover
relationships in coniferous forest. Can. J. For. Res.
Haverly, B.A., R.A. Wolford, K.N. Brooks, 1978. A comparison of
three snowmelt prediction models. 46th Ann. Meeting, Western
Snow Conf., pp. 78-84.
Leaf, C.F. and G.E. Brink, 1973. Computer simulation of snowmelt
within a Colorado subalpine watershed. u.s. Dept. Agr. For.
Ser. Res. Pap. RM-99, 22 pp.
McKendrick, J. W. Collins, D. Helm, J. McMullen and J. Koranda,
1982. Alaska Power Authority, Susitna Hydroelectric Project,
Environmental Studies-Subtask 7.12, Plant Ecology Studies,
Phase I Final Report. University of Alaska Agricultural
Experiment Station. Palmer, Alaska.
[
[~
r'
r·
r
[
[
G
r~
[
[
c
r
u
[
[
L
L
[
c
[
[
-185 -
McNamee, Peter J., 1982. Description of habitat, deer, and elk
microcomputer models for the integrated wildlife-intensive
forestry research program. Prepared for the Technical
Working Group (IWIFR), Province of British Columbia.
Miller, Sterling D. and Dennis C. McAllister, 1982. Big Game
Studies, Volume VI. Black and Brown Bear, Susitna Hydroelectric
Project, Phase I Final Report, Alaska Department of Fish and
Game. 233 pp.
Miller, Sterling D., 1983. Big Game Studies, Volume VI. Black
and Brown Bear, Susitna Hydroelectric Project, Phase II
First Annual Progress Report, Alaska Department of Fish and
Game. 99 pp.
Trihey, E. Woody, 1982. Preliminary Assessment at access by
spawning salmon to side slough habitat above Talkeetna.
Draft report prepared for Acres American Inc., Buffalo,
New York, November, 1982.
Van Cleve, K. and L.A. Viereck, 1981. Forest succession in
relation to nutrient cycling in the boreal forest of
Alaska. Pages 185-211 in D.C. West, H.H. Shugart and
D.B. Botkin, editors. Forest Succession. Springer-Verlag
New York.
[
-186 -
8.0 LIST OF PARTICIPANTS
Attending the Susitna· Terrestrial Modelling Workshop
August 23-27, 1982
NAME
Tom Arminski
Greg Auble
Warren Ballard
Keith Bayha
Bruce R. Bedard
Steve Bredthauer
Leonard P. Corin
Ike Ellison
AFFILIATION
Alaska Power Authority
USFWS -Welut
Alaska Department of
Fish & Game
USFWS
Alaska Power Authority
R & M Consultants
USFWS
USFWS -Welut
ADDRESS & PHONE NO.
344 West 5th Avenue
Anchorage, Alaska
99501
(907)277-7641
2625 Redwing Road
[
Fort Collins, Colorado r
80526 t
(303)226-9431 ~
P.O. Box 47
Glennallen, Alaska
99588
(907)822-3461
1011 East Tudor Road
Anchorage, Alaska
99507
(907)276-3800
334 West 5th Avenue
Anchorage, Alaska
99501
(908)277-7641
P.O. Box 6087
5024 Cordova
Anchorage, Alaska
99503
(907)279-0483
605 West 4th, #G-81
Anchorage, Alaska
99501
(907)271-4575
[
[
[J
L
2625 Redwing Road r·~.·
Fort Collins, Colorado
80526 ~
(303)226-9431
["
L
NAME
John Ernst
r Bob Everitt
Steve Fancy
Richard Fleming
Bill Gazey
Philip S. Gipson
George Gleason
Michael Grubb
John Hayden
Dot Helm
-187 -
AFFILIATION
LGL
ESSA Ltd.
LGL
Alaska Power Authority
LGL
Alaska Cooperative
Wildlife Research
Unit
Alaska Power Authority
Acres American
Acres American
University of Alaska
Agriculture
Experiment Station
ADDRESS & PHONE NO"
#305 -1577 "C" Street
Anchorage, Alaska
99501
(907)274-5714
678 West Broadway
Vancouver, B.C.
V5Z 1G6
(604)872-0691
P.O. Box 80607
Fairbanks, Alaska
99708
(907)479-6519
334 West 5th Avenue
Anchorage, Alaska
99501
(907) 277-7641
1410 Cavitt Street
Bryan, Texas
77801
(713)775-2000
University of Alaska
Fairbanks, Alaska
99701
(907)474-7673
334 West 5th Avenue
Anchorage, Alaska
99501
(907)277-7641
900 Liberty Bank Buildins
Buffalo, New York
14202
(716)853-7525
1577 "C" Street
Anchorage, Alaska
99501
(907)276-4888
P.O. Box AE
Palmer, Alaska
99645
(907)745-3257
NAME
Brina Kessel
Sterling Miller
Suzanne Miller
Ron Modafferi
Robert Mohn
Carl Neufelder
Ann Rappoport
Wayne Regelin
Butch Roelle
David G. Roseneau
Karl Schneider
-188 -
AFFILIATION
University of Alaska
Museum
Alaska Department of
Fish &. Game
Alaska Department of
Fish & Game
Alaska Department of
Fish & Game
Alaska Power Authority
Bureau of Land
Management
USFWS
Alaska Department of
Fish & Game
USFWS -Welut
LGL
Alaska Department of
Fish & Game
ADDRESS & PHONE NO.
P.O. Box 80211
College, Alaska
99708
(907)474-7359
333 Raspberry Road
Anchorage, Alaska
99502
(907)344-0541
333 Raspberry Road
Anchorage, Alaska
99502
(907) 344-0541
333 Raspberry Road
Anchorage, Alaska
99502
(907)344-0541
334 West 5th Avenue
Anchorage, Alaska
99501
(907)277-7641
[
4700 East 72nd Avenue [
Anchorage, Alaska
99501
(907) 267-1200
605 West 4th, #G-81
Anchorage, Alaska
99501
(907) 271-4575
1300 College Road
Fairbanks, Alaska
99701
(907)452-1531
2625 Redwing Road
[
r' G
[
L
Fort Collins, Coloradolr'
80526 •
(303-226-9431
P.O. Box 80607
Fairbanks, Alaska
99708
(907)479-6519
333 Raspberry Road
Anchorage, Alaska
99502
(907)344-0541
L
L
t t=
L,
G L . ..i
[
I~
E
NAME
Robin Sener
Nicholas Sonntag
Robert N. Starling
Gary Stackhouse
Bill Steigers
Nancy Tankersley
Thomas W. Trent
Joe Truett
Larry Underwood
Jack Whitman
Marjorie Willits
-189 -
AFFILITATION
LGL
ESSA Ltd.
NORTEC
USFWS
University of Alaska
Agriculture
Experiment Station
Alaska Department of
Fish & Game
Alaska Department of
Fish & Game, SU Hydro
Aq11atiC"
LGL
AEIDC
Alaska Department of
Fish & Game
Alaska Department of
Natural Resources
ADDRESS & PHONE NO.
#305 -1577 "C" Street
Anchorage, Alaska
99501
(907} 274-5714
678 West Broadway
Vancouver, B.C.
V5Z 1G6
(604)872-0691
#100 -750 West 2nd Ave.
Anchorage, Alaska
99501
(907) 276-4302
1011 East Tudor Road
Anchorage, Alaska
99507
(907)276-3800
P.O. Box AE
Palmer, Alaska
99645
(907) 745-3257
333 Raspberry Road
Anchorage, Alaska
99502
(907) 344-0541
2207 Spenard Road
Anchorage, Alaska
99503
(907) 274-7583
Rural Route 1, Box l~A
Flagstaff, Arizona
86001
(602)526-5055
707 "A" Street
Anchorage, Alaska
99501
(907} 279-4523
P.O. Box 47
Glennallen, Alaska
99588
(907)822-3461
555 Cordova Street
Anchorage, Alaska
99510
(907) 276-2653
-190 -
LIST OF PARTICIPANTS
r ,
I L.
Attending the Susitna Terrestrial Mitigation Planning Workshop r
February 28, March 1-2,·1983
NAME
Warren Ballard
Bruce R. Bedard
Steve Bredthauer
Bob Burgess
Leonard P. Carin
Ivlalcolm Coulter
Rosanne Densmore
Bob Everitt
Randy Fairbanks
AFFILITATION
Alaska Department of
Fish & Game
Alaska Power Authority
R ' M Consultants
LGL
USFWS
LGL
Envirosphere
ESSA Ltd.
Envirosphere
ADDRESS & PHONE NO.
P.O. Box 47
Glennallen, Alaska
99588
(907)822-3461
334 West 5th Avenue
Anchorage, Alaska
99501
(907)277-7641
P.O. Box 6087
5024 Cordova
Anchorage, Alaska
99503
(907) 279-0483·
P.O. Box 80607
Fairbanks, Alaska
99708
(907)479-6519
605 West 4th, #G-81
Anchorage, Alaska
99501
(907)271-4575
[
L
r
L:
.#305-1577 "C" Street r
Anchorage, Alaska ~
99501
(907) 274-5714
1227 West 9th Avenue
Anchorage, Alaska
99501
(907) 227-1561
678 West Broadway
Vancouver, B.C.
VSZ lG6
(604)872-0691
1227 West 9th Avenue
Anchorage, Alaska
99501
(907)227-1561
[
~ .,
L
L
G u
[:
bi
[
[
NAME
Steve Fancy
Richard Fleming
Bonnie Friedman
Bill Gazey
Philip s. Gipson
David Hamilton
John Hayden
Dot Helm
Dick Hensel
Brina Kessel
Gary Lawley
-191 -
AFFILIATION
LGL
Alaska Power Authority
LGL
LGL
Alaska Cooperative
Wildlife Research
Unit
USFWS -Welut
Acres American
University of Alaska
Agriculture
Experiment Station
Arctic Environmental
Information & Data
Center (University
of Alaska
University of Alaska
Museum
Envirosphere
ADDRESS
P.O. Box 80607
Fairbanks, Alaska
99708
(907) 479-6519
334 West 5th Avenue
Anchorage, Alaska
99501
(907)277-7641
P.O. Box 80607
Fairbanks, Alaska
99708
(907)479-6519
1410 Cavitt Street
Bryan, Texas
77801
(713)775-2000
University of Alaska
Fairbanks, Alaska
99701
(907) 474-7673
26.25 Redwing Road
Fort Collins, Colorado
80526
(303)226-9431
1577 "C" s-:.reet
Anchorage, Alaska
99501
(907)276-4888
P.O. Box AE
Palmer, Alaska
99645
(907)745-3257
555 Cordova Street
Anchorage, Alaska
99501
(907)274-4676
P.O. Box 80211
College, Alaska
99708
(907)474-7359
1227 West 9th Avenue
Anchorage, Alaska
99501
(907)227-1561
NAME
Sterling Miller
Suzanne Miller
Ron Modafferi
Ann Rappoport
Martha Raynolds
Wayne Regelin
Butch Roelle
David G. Roseneau
Karl Schneider
Robin Sener
Nicholas Sonntag
-192 -
AFFILIATION
Alaska Department of
Fish & Game
Alaska Department of
Fish & Game
Alaska Department of
Fish & Game
USFWS
LGL
Alaska Department of
Fish & Game
USFWS -Welut
LGL
Alaska Department of
Fish & Game
LGL
ESSA Ltd.
ADDRESS & PHONE NO.
333 Raspberry Road
Anchorage, Alaska
99502
(907)344-0541
333 Raspberry Road
Anchorage, Alaska
99502
(907)344-0541
333 Raspberry Road
Anchorage, Alaska
99502
(907)344-0541
605 West 4th, #G-81
Anchorage, Alaska
99501
(907)271-4575
[
[
( .
I
L
#305-1577 "C" streetr
Anchorage, Alaska t
99501
(907)274-5714
1300 College Road
Fairbanks, Alaska
99701
(907)452-1531
r L~
2625 Redwing Road ~
Fort Collins, Colorado l
80526 .
(303)226-9431
P.O. Box 80607
Fairbanks, Alaska
99708
(907)479-6519
333 Raspberry Road
Anchorage, Alaska
99502
(907)344-0541
[
r~
I.
l.j
[
#305 -1577 "C" Street [
Anchorage, Alaska
99501
(907)274-5714
678 West Broadway
Vancouver, B.C.
V5Z 1G6
(604)872-0691
L
NAME
Bill Steigers
Nancy Tankersley
Jack Whitman
Larry Wright
l_;
-193 -
AFFILIATION
University of Alaska
Agriculture
Experiment Statipn
Alaska Department of
Fish & Game
Alaska Department of
Fish & Game
Alaska Department of
Natural Resource,
State Parks
ADDRESS
P.O. Box AE
Palmer, Alaska
99645
(907)745-3257
333 Raspberry Road
Anchorage, Alaska
99502
(907)344-0541
P.O. Box 47
Glennallen, Alaska
99588
(907)822-3461
555 Cordova Street
Anchorage, Alaska
99501
(907) 274-4676
-194 -
APPENDIX I
UPPER SUSITNA RIVER BASIN
MOOSE POPULATION MODELLING
by
Warren Ballard
Alaska Department of Fish and Game
P.O. Box 47, Glennallen, Alaska 99588
SuzAnne Miller
Alaska Department of Fish and Game
333 Raspberry Road, Anchorage, Alaska 99502
r~
L ,
c
" r : u
L
[
L
[
[
r
r
E
[
[
[
[
[
c
B
[
[
[
[
L
b
l
-195 -
Introduction
The Upper Susitna River Basin of Game Management Unit 13
(GMU-13) has received considerable attention in recent years by
the Alaska Department of Fish and Game (ADF & G). Information
on the distribution, abundance, and sex and age characteristics
of moose {Alces alces) populations have been routinely collected
since the early 1950s for harvest management. Since 1975, research
on the population status and food habits of· two important predators,
brown bears {Ursus arctos) and wolves, has been in progress. In
addition, several other intensive research projects have been
conducted in the area to identify predator-prey relationships and
other moose and predator population dynamics parameters. The
availability of such information presents a unique opportunity to
examine the structure and dynamics of the moose population occupying
the Upper Susitna River Basin and GMU-13.
ADF & G is currently developing a computer simulation model
to synthesize historical information rela.ted to the Upper Susi tna
River Basin and GMU-13 moose populations. Development of the model
has been motivated by several factors:
1) the model should recreate as closely as possible the
historical data base; and
2) analysis of model results should lead to the basis
for a predictive model which can be utilized in the
Susitna Hydroelectric Project Big Game Studies.
The moqel.has, therefore, concentrated on explaining existing
historical information, rather than futu:e predictions .. Increased
understanding of the historical conditions can then be used to
develop a satisfactory relational model for examining potential
development impacts.
-196 -
Because information and analyses presented in this report
are of a preliminary nature, they should not be used in scientific
technical publications without the approval of the authors.
Simulation Model -General Format
The preliminary version of the computer simulation· model
was designed to provide maximum flexibility with regard to both
structure and parameter estimation .. This was accomplished by
dividing the annual dynamics of the moose ~opulation into a series
of discrete events. These events describe the birth and death
processes of the population. The birth process is described by a
single component, whereas the death process consists of four
different components -death by:
1) natural causes;
2) hunting;
3) wolf predation; and
4) bear predation.
These events can be arranged in any sequence to describe the
annual cycle of the population_ In addition, detailed printouts
of the modelled population can be requested at any time to compare
with historical field data.
The simulation model divides the moose population into
six sex-age categories: calves, yearling~, and adults of each
sex. This reflects the level of classification attainable in
the field. Each time an event is invoked, the standing population
resulting from the previous event (or the initial population) is
subjected to the changes described by that event. The specific
changes for each event are as follows:
[
l.
[
[
[
l~
[
b
c
c
r L
c
[
L
[
l~
L
[
L
i ---" _j
L=
i: u
[
c
-197 -
A. Reproduction
The reproduction component involves two changes in the
following order:
1. New calves are created by the following equations:
TOTAL
CALVES = FECUNDITY *
(YEARLINGS)
FEMALE FECUNDITY
YEARLINGS + (ADULTS) *
FEMALE
ADULTS
MALE CALVES = (SEX RATIO AT BRITH) * TOTAL CALVES
FEMALE CALVES = TOTAL CALVES -MALE CALVES
2. The standing population is advanced one year in age:
ADULTS = ADULTS + YEARLINGS
YEARLINGS = CALVES
CALVES = TOTAL NEW CALVES (from step 1)
Parameters necessary for reproduction are:
1. Fecundity rate for yearling females.
2. Fecundity rate for adult females.
3. The sex ratio at birth.
B. Death by Natural Caus:s
A natural mortality rate for each sex-age category is
used to determine the number of deaths by natural causes:
DEATHS
(SEX, AGE)
= MORTALITY RATE *
(SEX, AGE)
NUMBER
(SEX, AGE)
-198 -
The number of survivors is simply:
NUMBER (SEX, AGE) = NUMBER (SEX, AGE) -DEATH (SEX, AGE)
Parameters necessary for natural mortality are:
1. Mortality rate for each sex-age category.
c. Death by Hunting
Since historical harvest information is available, the
number of deaths by hunting is an input parameter and is
simply subtracted from the standing population.
NUMBER (SEX, AGE) = NUMBER (SEX, AGE) -HARVEST (SEX, AGE)
D. Deat~ by Wolf Predation
Most of the information on moose mortality due to wolf
predation has been gathered through food habits studies of
wolf populations. This information, coupled with estimates
of the numbers of wolves occupying the same area as the
moose population, is used by the model to estimate the number
of deaths due to wolf predation.
The following equations constitute the wolf predation
component:
Total kgs prey
Consumed by wolves
= Daily consumption * Number of *
rate per wolf wolves
Number
of days
[
[
[
[
[
c
[
Number of
calves killed =
Proportion of
Total kg * diet consisting
prey consumed of moose calves
Average wei.Jht
of moose calf [
Number of
yearlings and =
adults killed
Total kgs
prey consumed
Proportion of
diet consisting
* of moose yearlings
and adults
Average
weight of
yearlings
and adults
I
•.. .1
b
r
l
E.
-199 -
The number of deaths due to wolf predation fs subtracted from
each sex category in proportion to their availability in the
population.
Parameters necessary for the wolf predation component are:
1) Number of wolves.
2)-Daily consumption rate per wolf.
3) Number of days of wolf predation.
4) Proportion of wolf diet consisting of moose calves.
5) Proportion of wolf diet consisting of moose yearlings
and adults.
6) Average weights of moose calves.
7) Average weight of yearlings and adults.
Death by Bear Predation
Bear predation rates have been estimated from studies on
both moose populations and bear populations. Preliminary
estimates of daily consumption rates were judged too high
to be realistic. In an effort to limit bear predation
within realistic bounds, a relationship between daily
consumption rates and moose abundance was hypothesized. The
bear predation component of the model adjusts the daily
consu~ption rates for both calves, and yearlings and adults
using the following relationship:
Adjusted _ ( Maximum
consumption rate -~consumption rate
Moose abundance)
at which maximum *
rate occurs
Moose
abundance
-200 -[
The adjusted consumption rates are then utilized in the
following equations: \ .
Number of
calves killed
= Adjusted daily calf *· Number of days * Number [
consumption rate of predation of bears .
Number of
yearlings and
adults k.illed
Adjusted daily Number of days
= yearling and adult * of predation *
consumption rate
Number
of bears
The number of deaths due to bear predation is subtracted from
each sex category in proportion to their availability in the
population.
Parameters necessary for the bear predation component are:
1) Number of bears.
2) Maximum daily consumption rate on calves.
3) Abundance of calves at which maximum daily consumption
rate occurs.
4) Maximum daily consumption rate on yearlings and adults.
5) Abundance of yearlings and adults at which maximum
daily consumption rate occurs.
6) Number of days of predation.
The number of events, and the specific sequences of events,
needed to define an annual moose population cycle can be changed
at any time during a simulation run. Similarly, the parameters
necessary for any event can be changed. This allows the modeller
to use historical information to recreate conditions as they
appear to have existed.
r
l
r·
L~
[ ..
.
[
c
[
f'
L
L
[
[
[
,,
',!-'
-201 -
In specifying the sequence of events and the event parameters,
it is important to remember that the events are independently
processed. This is not a problem for events that in nature occur
at distinct and separate time periods (spring reproduction and
fall hunting, for example). For those events that occur
simultaneously or that overlap in time (early summer wolf predation
and early summer bear predation, for example), care mus~ ·be taken
to ensure the proper order of events and the event parameters may
need to be altered.
Upper Susi tna R.i ver and GMU-13 Simulation Moose Model
Because longer, more intense moose population studies to
assess the impacts of predation on moose were previously conducted
in an adjacent portion of GMU-13 (Ballard, et al., 1981 a,b), that
area was used as the basis for the Upper Susitna River model.
Boundaries of the area were previously described by Ballard, et alo
(198la). Briefly, the boundaries are the Alaska Range on the
north, Brushkana and Deadman Creeks on the west, Susitna River on
the south and the Maclaren River on the east. Although this area
extends beyond the impact zones, we believe that the biological
characteristics of the area are representative of the project
area. Also, an attempt was made to model the entire GMU-13 moose
population as well, in an effort to provide a comparison to the
Susitna model and allow·assessment of the percentage of the GMU-13
moose population to be impacted by the project. Both models will
be published elsewhere (Ballard, et al., In prep.).
Both population models start with an estimate of population
size, and sex and age structure, and proceed through an annual
cycle of reproduction and mortality factors which, for these
models, are termed "events" (Figure 1). Population estimates
are calculated for each year at calving and subsequently the
population declines as mortality factors act on the population.
-202 -
.
Pre-calving moose
population estimate
~
Event 1 -Reproduction -
~
E:vent 2 -Early spring and. summer
mortality (excluding predation)
~
Event 3 -Spring wolf predation
(15 May -15 July)
-~
Event 4 -Summer wolf predation
(15 July -1 Nov.)
~
Event 5 -Brown bear pr-edation
~
Event 6 -Black bear pr.edation
~
Event 7 -Hunter harvest
t
Event 8 -Winter mortality
(excluding predation)
+
Event g· -Winter wolf predation
(1 .Nov. -15 May)
Figure 1: Timing and sequence of factors used in the models to
determine the annual population dynamics of moose
in the Susitna River Study Area and the entire
w~U 13 in southcentral Alaska.
r L
[
[
[
c
0
r
r
Q
c
L
[
[
[
[
r
. 203 -
Population Estimates
Population Size
The starting 1975 population size estimate (X) for each
model was derived from the following formula:
where,
X = (A) (B)
c
A = the number of moose observed/hour during the
1975 autumn composition counts;
B = the 1980 area population estimate for either
the study area or GMU-13; and
C = the number of moose observed/hour during the
1980 autumn composition counts which were
conducted immediately before the census.
We assumed that the numbers of moose observed/hour during fall
composition counts reflected annual changes in moos~ density.
Variable B was estimated from a census during November, 1980.
Approximately 8,142 km 2 of GMU-13, which included all of the
7,262 km 2 wolf removal area, were stratified and censused to
determine the number of moose, using quadrat sampling techniques
described by Gasaway (1978) and Gasaway, et al. (1979). Moose
density estimates derived during the census in 1980 were used as
the basis for grossly estimating numbers of moose within the
Susitna Study Area and within GMU-13 from 1975 -1981. The
actual moose population estimate in fall, 1980 was used as a
check for the population size generated by the project model. It
was assumed that for the model to be valid, the fall, 1980
population estimate derived from the model should closely coincide
with the census estimate.
-204 -
A different approach was ased for the GMU-13 model. Those
portions of GMU-13 not censused in 1980 were stratified into 4
density categories (none, low~ moderate, and high). The
stratification was based upon a combination of distribution and
numbers of moose observed during composition counts conducted
from 1975 -1981, and the knowledge of 5 biologists with experience
in this area (more than 24 man-years). Density estimates· for the
4 categories derived from sampling were then applied to the non-
sampled area to arrive at a GMU-13 population estimate of 23,000
moose for fall, 1980. The GMU-13 model was modified so that the
fall, 1980 population size generated by the model would conform
with the estimate derived from censusing and stratification.
Event 1 -Reproduction and Sex and Age Structure
The sex ratio of calves at birth was assumed to be 50:50
while the sex ratio of yearlings and adults was determined by the
previous year's estimate of reproduction and mortality. In the
case of year 1 (1975), the sex ratio was determined by the fall
moose composition count and back-calculated to correspond with
population size at calving (Figure 2). All age classifications
were directly extrapolated from sex and age composition count
data except for the percent of calves in the herd. This was
adjusted upward by 5% because calves are often located away from
large groups of moose and are usually underestimated in composition
counts (Ballard, et al., 1982 a,b; and Gasaway, pers. comm.).
Also, because preliminary runs revealed that in both models,
populations declined to extinction, initial estimates of numbers
of yearlings were doubled. Estimates of yearlings based upon
composition counts were drastically underestimated, probably
because they were incorrectly aged as adults.
Pregnancy rates of cow moose were determined from rectal
palpation of captured animals in 1976, 1977, and 1980
(VanBallenberghe, 1978; Ballard and Taylor, 1980; and Ballard,
et al., 1982a,b). Although some minor variations in rates were
noted, we assumed that 88% of the sexually mature cows ( ~ 2 yr ~ge:)
were pregnant each year.
[
L.
r .
L
L
[
[
[
[
[
[
C.
I-
[
b
L
r
Yearling
Females
-205 -
Male
Calves
Yearling Fecundity Rate · Proportion Males
Newborn
Calves
Adult Female
Females Calves
Input Variables:
(1) Fecundity Rate for Yearlings
(2) Fecundity Rate for Adults
(3) Sex Ratio at Birth
Figure 2: Schematic diagram of Event 1 (reproduction) for
the moose model.
-206 -
Estimates of moose productivity were determined during
calf collaring·programs from 1977-1979 (Ballard, et al., 1980;
198la) and were estimated at 135 calves/100 pregnant cows or
1.19 calves/adult cow. Productivity of 2 year olds was estimated
at 0.29 calves/cow (from Blood, 1973). For the models, we
assumed that productivity remained constant each year (which was
probably not the case). In fact, in that portion of the .susitna
River Study Area where brown bears were transplanted, there was
a significant (P < 0.01) negative relationship between the
preceding winter's snow depth and the following fall's calf:cow
ratio (Ballard, et al., 1980), suggesting that some fluctuations
inproductivityoccur due to winter severity. However, because
of large variations in snow depth between drainages, and because
calf survival has been significantly increased by predator
reduction programs following severe winters, we were unable to
modify productivity estimates based on available data.
Event 2 -Early Spring and Summer Mortality (Excluding Predation)
Following birth, both calf and adult mortality estimates
(Figure 3) were subtracted from the population. Immediately after
birth, 6% of the calves were assumed to die from natural factors
other than wolf and bear predation, such as stillbirth, drownings,
and other accidents (from Ballard, et al., 198la).
Events 3, 4, 9 -Wolf Predation
Estimates of annual moose mortality due to wolf predation
for each model were divided into 3 time periods to correspond
with pup production, human exploitation and natural mortality,
and changes in diet composition (Figure 4). The time periods
we~e·,as foliows:
*1) May 15-July 15 (Event 3);
*2) July 15 -November 1 (Event 4); and
#3) November 1 -May 15 (Event 9).
[
[
r~
t~
[
r l:
L
L
0
[~
[
t '
L
[
[
c
c
r~
6' -
~
-207 -
Number of
Moose by
sex and age X . Mortality Rate
by sex
and age
Input Variables:
Number of _,
Deaths by
sex and age
(1) Mortality Rate for each sex and age group
Figure 3: Schematic diagram of Events 2 and 8 (early spring
and winter mortality) for the moose model.
-208 -·[
Number
of
Wolves X
Consumption
rate per
wolf per day X
J
Number of C
Days of ,)
Wolf PredatioriJ
. [
Te.tal kgs wolf
consumption
(--
t
[
Prop6~r.....---------l~ Yearlings and Adul[s
Average
Weight of
Calf
Number of
Calves killed
Input Variables:
(1) Number. of Wolves
(2) Consumption Rate of Wolves
(3) Number of Days of Wolf Predation.
Average
Weight of
Yearlings and
Adults
Number of
Yearlings and
Adults killed
(4) Proportion of Wolf Kill Cons~sting of talves
(5) Proportion of Wolf Kill Consisting of Yearlings and Adults
(6) Average Weight of Calves
(7) Average Weight of Yearlings and Adults
Figure 4: Schematic diagram of Events 3, 4 and 9 (wolf
predation) for the moose model.
F
c
c
L
[
I , , __ _
l~
[
[
[
F ~-
f-~
--7
L
L
-209 -
Period #1 encompasses the wolf denning period and represents
the annual low in the wolf population. Because pups are quite
small during this time period, no food consumption was allocated
for them. Period #2 encompassed the post-denning period and
represents the highest level of the wolf population (adults plus
pups prior to hunting and trapping season) during the year. For
this latter time period, we assumed that pups had similar. food
requirements as adults. Period #3 encompassed both the population's
highest level during the year (prior to hunting and trapping season)
but also the lowest level (post-hunting and trapping season) .
Consequently, we used the mid-point between the ~wo population
estimates to provide an average number of wolves for the winter.
Wolf population levels were derived from Table 30 from Ballard,
et al. (In Prep.) for the Susitna River Study Area while the GMU-13
estimates were derived from Tables 22 and 30 (op. cit.).
~stimates of percent biomass of moose consumed by wolves
for Period· #1 were based entirely on scat analyses according to
methods described by Floyd, et al. (1978). The analyses indicated
that 91% of the biomass of prey consumed by wolves from May 15 -
July 15 was comprised of ungulates, with calf and adult moose
comprising 35% and 47%,-respectively, of the total biomass
consumed. Estimates of percent biomass of calf and adult moose
consumed by wolves during Periods #2 (July 15 -November 1) and
#3 (November 1 -May 15) were determined from kills observed
while monitoring radio-marked packs. The estimates for the
study were divided into 2 time periods to correspond with the
increased importance of caribou as wolf prey from 1979 -1981.
From 1975 -1978, we estimated that from July 15 -November 1
(Period #2), calf and adult moose comprised 12% and 78%,
respectively, of the prey biomass, while from November 1 -May
15 (Period #3), calf and adllt moose comprised 18% and 73%,
respectively, of the biomass. During Period #2 from 1979 -1981,
percent biomass of adult moose declined to 73%, while the percent
of calf moose remained constant. Percent biomass declined to
17% and 68% calf and adult moose, respectively, during Period #3
from 1979 -1981.
-210 -
The estimated biomass of calf and adult moose killed by
wolves during each time period per year was extrapolated from
wolf population estimates for each period multiplied by the
numbers of days in each period multiplied by the estimates of
wolf daily consumption rates. For all 3 time periods, it was
assumed that wolves consumed 7.1 kgs prey/wolf/day (Table 20
op. cit.). Estimates of percent biomass by prey species were
then multiplied to derive estimated biomass. For each time
period, the number of moose killed was estimated by dividing the
average weight of each age class for each period derived from
literature and field studies into the estimated biomass. The
wolf daily consumption rate used is relatively high in relation
to that reported in the literature and thus, we consider the
estimates of number of moose killed per year to be inflated.
Event 5 -Brown Bear Predation
Predation rates of brown bear on both adult and calf moose
were derived from observations of kills during daily relocation
flights of 23 adult radio-collared bears (Ballard, et al., 198la
and Table 35 from Ballard, et al., In Prep.). The relocation
flights were done between May 15 and July 15, the period of most
brown bear predation on moose {Ballard, et al., 198la). Kill
rates of adult moose were calculated by assuming that all adult
moose killed by the 23 radioed bears between May 15 and July 15
were observed (N = 28), and after this time, no adult moose
were killed. Observed rates of calf moose killed were 1 calf/
9.4 days/adult bear. These kill rates were extrapolated to the
adult bear population estimates for the Susitna Study Area and
GMU-13 (derived from Miller and Ballard, 1982). No information
was available on annual bear population fluctuations, so for these
models, we assumed a stable population from 1975 -1981 (Figure 5).
Preliminary runs of the model indicated that kill rates
of calf moose were too high. It seems more likely that estimates
of bear kill rates on calf moose would be underestimated even
[
r·
[
r ' ' ·~·
from daily relocation flights because many bears remained on calf L
L:
r
I.
I ~
l
b
Maximum Bear
Consumption
Rate per Bear
per Day on
calves
0 2000
calves
Adjusted
Consumption
Rate on Calves
Number of Calves
Killed
Input Var~ables:
..
-211 -
0
Number of Bears
Number of Days
Bear Predation
Maximum Bear
Consumption Rate
per Bear per Day
on Yearlings and
Adults
1 2ooo t Yearlings plus Adults
_ ....
Adjusted
Consumption
Rate on Yearlings
and Adults
Number of Year-
lings & Adults
Killed
(1) Maximum Consumption Rate on Calves
(2) Maximum Consumption Rate on Yearlings and Adults
(3) Num.'.>er of Bears
(4) Number of Days of Bear Consumption
Figure 5: Schematic diagram of Events 5 and 6 (brown bear
and black bear predation) for the moose model.
-212 -
kills less than 24 hours (Ballard, unpub. data) . Therefore, we
modified the estimates of calf kill rate by·assuming that the
magnitude of bear predation was partially dependent on the density
of moose calves. For the. study area model, it was assumed that
bears preyed upon 50% of the estimated number of calves produced
for 1977 and 1978. This was based upon estimates derived from
moose composition counts (0.14 calves/bear/day for 60 days and
0.02 adults/bear/day, for 60 days). At higher levels of calf
production than the 1977 and 1978 levels, we assumed that the
numbers preyed upon remained constant. At lower levels of calf
production, we assumed that a linear relationship existed between
percent calves taken by bears and calves produced. During 1979
only, we reduced brown bear predation on calves to 0.10 calves/
bear/day to correspond with removal of 47 transplanted bears from
the Susitna Study Area for a 2 month period in late spring and
early summer (Miller and Ballard, 1983).
Preliminary runs of the project model sugges±ed that our
estimates of bear predation on adults were also too high. The
original kill estimates meant than an excess of 20% annual adult
moose mortality occurred from brown bear predation alone. Such
estimates, compared with all of the other mortality factors, were
obviously greatly exaggerated. Because many bears remain with
adult moose kills for 5 - 6 days, periodic relocation of bears
could tend to overestimate kill rates, similar to overestimation
of wolf kill rates (Fuller and Keith, 1980). However, most of
our data were collected during contiguous daily flights and
because individual carcasses and bears could usually be identified,
the rates should not have been greatly exaggerated. Possibly the
23 adult radio-collared bears had kill rates greater than the
rest of the bear population, but we have no evidence to support
this idea. Predation estimates on adult moose were modified in
a similar way to those for calf moose except that we assumed that
at the 1977 and 1978 moose population estimates, brown bears were
responsible for 7% adult mortality.
[
r:
[
L
L
[
r
l J
[
r
l..__:;
n
I
-213 -
Preliminary runs of the GMU-13 model suggested that the
estimates of bear predation'derived for the Susitna area were
also too high for the entire unit. This was not unexpected.
since we originally applied bear density estimates obtained for
the Susitna area (Miller and Ballard, 1983} to the entire unit.
Undoubtedly, variations in both brown bear density and predation
on calves occur within the unit. Consequen~ly, both the number
of bears and predation rates were subjectively adjusted downwards
to 708 adult bears preying on calf and adult moose at a rate of
0.10 calves/bear/day and 0.01 adult moose/bear/day during May 15 =
July 15.
Event 6 -Black Bear Predation
Although black bears (Ursus americanus) occur in GMU-13
and they have been observed preying on moose (Ballard and Miller,
unpub. data), they were rare and were considered an insignificant
source of mortality within the Susitna River Study Area. However,
because black bears were quite numerous in other portions of
GMU-13, they were incorporated into the GMU-13 model (Figure 5).
Based on existing density estimates and observed rates of
predation from one portion of the unit, we originally estimated that
1,650 black bears occur in the unit and that they were preying on
calf and adult moose at a rate of 0.021 and 0.012/bear/day,
respectively. Similar to brown bear· predation rates, preliminary
runs suggested that perhaps both the population estimates and the
predation rates for black bear were too high. Consequently, they
were subjectively reduced to a population of 1,000 black bears
preying on moose at 0.003 calves/bear/day and 0.001 adults/bear/
day for 60 days following birth.
Event 7 -Hunter Harvest
Annual hunting mortality, which during this study affected
bulls only, was determined fo~ each year of study from "man~atory
harvest reports" (Figure 6). Harvest reports from successful and
Number of
Moose by
sex and age
Input Variables:
-214 -
minus
(1) Number of Moose Harvested by sex and age
Number of
Moose Harvested
by sex and age
Figure 6: Schematic diagram of Event 7 (hunting mortality)
for the moose model.
[
[
[
b
L
L
F
L
L
[
r--
1 J
r v
[
[
[
[
[
[
c
6
E
c
[
b
L
b
r
-215 -
unsuccessful moose hunters are required by law in GMU-13, however,
this is not enforced and compliance is less than 100%. To
encourage moose hunters to report results of their hunt, reminder
letters are sent to all those who took a harvest ticket but did
not report their hunt results. Because no·. reminder letters were
sent in 1980, the harvest for that year was determined by
extrapolating from return and non-return reports in previous years
to reports returned in 1980.
Antler measurements on harvest reports since 1978 provided
a basis for grossly estimating the number of yearlings killed,
although some measurements were undoubtedly false. Antler
measurements of ~ 30 inches were considered to be yearlings or
younger. Beginning in 1980, only bulls with antler spreads of
36 inches,or at least 3 brow·tines,were legal for harvest. For
the 1978 and 1979 hun'ting seasons, 55.4% of the measured moose
had·antlers of 30 inches or less; therefore, we assumed that
annually from 1975 -1979 half of the harvest was .comprised of
yearling bulls.
The annual hunting mortality rate for adult bulls was
estimated at 25% based on radio-collar data (N = 28) .
Event 8 -Winter Mortality (Excluding Predation)
Estimates of winter mortality in the model (Figure 3) were
subtracted from the estimated number of moose present each
November following hunter harvest. The magnitude of winter
mortality (usually by starvation) was initially estimated from
radio-collared moose by methods described by Hayne (1978) and
Gasaway, et al. (In press). Winter mortality was calculated as
follows (from Gasaway, et al., In press):
where,
a Percent mortality = b
-216 -
a = number of winter mortalities of radio-collared
moose; and
b = estimated number of collared animal months.
b estimated as follows: (c) (d)
e
where,
c = mean # months collars transmitting (exluding dead
moose) ;
d = total # radio-collared moose (including dead
moose); and
e = time interval for annual mortality.
Winter mortality data was available from 1977 -1981 for
calf moose and from 1979 -1982 for yearling moose (Table 1).
For modelling, it was assumed that during mild winter
(1975 -1976 through 1977 -1978 and 1979 -1980) calf mortality
was 6% .. Winter 1978 -1979 was considered relatively severe (Eide
and Ballard, 1982) with high rates of calf mortality during late
winter (Table 1). These higher rates for males and female calves
were used for 1978 -1989 in the models. For yearling females, we
utilized the calculated rate of 2.4%, and for yearling bulls, we
utilized the calculated mortality rate of 6% (Table 1). Even
though the yearling bull mortality rate was attributable to
hunting, which theoretically would have been illegal, it was used
because bulls usuall•r suffer proportionately larger natural
mortality than females and we suspected the calculated rate was
low.
L
[
[
[
[
B
8
r
E
·r\
_ _;
-.. ;
L
[
[
L
L
L
r-."-. J L --:-:. l l .. c-,,J
Table 1. Mortality rates due to winter starvation of radio-collared calf and yearling moose in the
Nelchina and Susitma River Basins, 1977-1982.
# mort ali ties
i mos. collars
transmitting (excluding
mortal! ties)
Total # radio-collared
moose (including
mortalities)
Time interval
(# mos.)
% mortality
1/
21 Mild winters
Sex
3/ Severe 1-linters
Both mortalities from hunting
1977-78 y
1979-80 y
1980-81
F 14
1 1
5.0 5.6
25 26
7 7
5.6 4.8
Calves Yearlings
1978-79 3./ 1979-80 y
1980-81
1981-82
f' M F M
3 8 1 2-y
2.6 2.7 9.9 10.5
41 26 50 37
5 5 12 12
14.1 57.1 2.4 6.2
l-J ( -J
N
I-'
-...J
-218 -
Annual winter mortality rates for adult cows varied from
0 to 5.6% during 1976 -1982 (Table 2). Overall, the winter
mortality rate was estimated at 3.6% and this was used for each
year of the study. Apparently the winter of 1978 -1979 was
severe enough to cause significant increases in calf morta~ity
but not for adults.
It was assumed that during mild winters, adult bulls
suffered rates of winter mortality identical to that of cows
(3.6%). Duringse.v.erewinter~, we assumed that adult bulls would
suffer higher rates of mortality than cows, so the 1978 -1979
winter mortality was subjectively estimated at 7.2%.
Project Population Model Analyses
Population Size Estimates
Between 1975 and 1981, estimates derived from fall
composition counts and the model suggest that the area's moose
population increased (Figure 7). The model indicates that the
fall moose population increased by 24%, while population estimates
based on the composition counts indicated a much larger increase
of 101%. Projected population estimates beyond May, 1981 (Figure
7) assume that all mortality factors remain identical to those of
1980 -1981.
Each year's independent moose population estimate based
upon composition counts were compared to those generated by the
model (Figure 8). From this comparison, it becomes quite evident
that the annua~ population estimates based on composition counts
were not accurate. Using both the 1975 and 1976 data with
docun1ented levels of productivity and mortality, the population
eventually becomes extinct. Based upon the 1980 census estimate
and the composition of the population at that time, no winter
mortality could have occurred for the moose population to have
increased up to the 1981 or 1982 estimates based on the composition
counts. Because this is highly unlikely, it suggests that the
G
[
L
L
[
L
L ·~· J J ~ • I' .} l j 'J
Table 2. Hortality rates of iidult (>2 yr.) radio-collared cow moose due to winter starvation and unidentified
mortality in the Nelchina and Susitna River Basins of southcentral Alaska from 1976-1982.
Year 1976-77 1977-78 1978-79 1979-80 1980-81 1981-82 Total
# Mortalities 0 1 1 1 2 4 9
x mos. collars
transmitting (eKcluding
mortalities) 5.~ u.s 10.6 6.0 10.0 10.4 24.1
'l'otal # radio-collared
moose (including
mortalities) 36 42 45 52 80 ' 82 126
Time Interval
(II mos.) 12 12 12 12 12 12 12
% Mortality 0 2.5 2.5 3.9 3.0 5.6 3.6
-220 -
3800
(ll ....
z 3800
::;:)
0
(J
z 3400
0
.... w
co 3200 co 0 0 0 a. ~ ~
o 7oT u.. 3000
(J . 0
La. I a:
0 sal w 2800 I
a: I ca
I ~ ::;:) I ::;:) I 0
J: so; z 2800 I .... c I c I w. w .... > -401 ~ 2-400 a: I ::t w
02 I ....
~ 30;
co
w 2200
w I -mooae model
~ 20J. --mooae obaerved/hour of
0 J 2000 compoaltlon counta
-eathwate baaed upon annual ~ I number of moo•• obaerved/hour
of compoaltlon counts In relation
Ll. 101 to 1980 compoaltton count and
0 1800 cenaua
J a:
w
ca
::t
::;:) y E A A z
Figure 7. November mooae population estlmatea aa derived from modeling versua
[
[
t
L
c
[
L
l'
compoaltlon counts for the Sualtna River Study Area of aouthcentral Alaak.a, L-'
1975-1988. -
[
-C/)
0
z
<(
C/)
::l
0
:r:
J-
w
C/)
0
0
~
u..
0
0:
w
m
:::E
::l
z
···' ······
1875 count
1878 count
1877 count
------1078 count
tD78 count
1880 count
model·
Individual competaJtlon counta
J. ~-, c~
.~ ,,~
/ "' ~· ~~~/
.,.,.,. 411',..
~-~· ;,.,.,.,..""'-
.. .;
,_]
.,.,. . .,., . ---.-;-:-
•••• I__.,• ............ --;;.••• ..,......---_.. __ .. ~ . ......--... "-'_._.._.~....... ----~----
--.--·--• .· ._. .... ._._u.,_,_._..._'-*a.. --..-------
--·--• • .... _,.._ua..a...l...}.at.uJ..J.l.'-'\.J -.-------,.,........ ···--. ----~.:o•.-.!".··· .:--!------------••••••....... ····· ~·
•• 0 • ..:...:..,:___-----..-;-:-: ---------------.--------------
S U A V E Y Y .E A A
N
N ......
figanrG ta. Fall mooeiill P«lll,tW!IiilUon trends dorlv«~d from modeiln@ l.!l&in(WJ tumual compoaiilon count data for lnllial population elze
for the Sueltna River Study Area, 1075-1891.
-222 -
number of moose observed/hour in composition counts is probably
not an accurate index of change in annual moose density. Also,
it suggests that the relationship between moose observed per hour
in composition counts versus population estimates obtained from
censusing may be quite variable from year to year. All other
population estimates suggested an increasing population trend
although the rates of increase were quite different.
Sex and Age Structure
Comparison of several sex-age parameters between the
model and composition counts suggest that at least three sex-age
classifications are underestimated during composition counts.
Calf:cow ratios, as estimated from the model, were higher than
those obtained from composition counts (Figure 9). Even though
composition count ratios were adjusted upward based upon
observed differences between composition surveys and census data,
the model suggests that th~ discrepancy between these two counts
may be larger than existing data suggest (Gasaway, et al., 1982;
Ballard, et al., 1982). The discrepancy occurs because cow:calf
pairs are often segregated from larger groups of moose and have
alowerprobability of being observed with either survey method.
Also, the model suggests that both survey estimates tend
to underestimate the proportions of yearling bulls (Figure 10)
and cows present in the population. This could occur for at
least 3 reasons:
1) counts are often made following hunting mortality, so
that usually an unknown proportion of yearling bulls
has been removed and remains unaccounted for;
2) an unknown proportion of the yearling bulls cannot
be identified from fixed-wing aircraft because antlers
are comprised of either buttons or short spikes; and
f
[
[
[
[
[:·
[~
b
E
c
[
c
[
G
[
L
[
L
[
r-:n:
60
60
en
;:
0 40
0
0
0 ,.....
Cl)
w
>
_J
-c(
0
30
20
10
1975 1976 1977
r ---~; r-------.,
\..._ ,, .) ... J " "J
- 1
'
~~
' ' '
--model
--oo_mpoaltlon oounta
1978 1979 1980 1981
Y E A A
figure 9. Eatlmated mooae calf:oow ratloa derived from modeling venue calf:cow ratloa obtained
from annual oompoelt!on count• In the Bueltna River Study Area, 197G-1982.
_----,
i
N
N w
--
en
...J
...J
::J
en
" z
...J
a:
<
w
>
-224 -
-model
--composition counta
'\ / \
I
I
I
---I --I --I ---
Y E A R
,, ·I
I ' I I '\. I
I .'"1
I
I
I
I
I
I
I
I
I
I
I
I
F1gure 10. Percent yearling bulls In moose populations each fall aa determined from
modeling versus composition counts for the Sualtna River Study Area, 1975-1982.
[
[
c
c
L
L
[
f
[
f~
r
[
[
[
[
[
E
b
[
[
[
[
l
L
-225 -
3) during the 1975 and 1976 composition surveys, the
criteria utilized for·estimating ages of yearling
bulls were not accurate according to antler
configuration data (Gasaway, pers. comm.).
Because the proportion of yearling females is based upon the
estimates of yearling males, this sex-age class would also be
underestimated.
Calf Mortality
Predation by brown bears was the single most important
calf mortality· factor during the study period. Because of the
manner in which brown bear mortality was calculated, the numbers
of calves killed by bears each year varied (Figure 11), but the
actual percentage of calves killed remained constant each year,
except in 1979 when bears were temporarily transplanted from the
area.
Calf mortality attributable to wolf predation declined
from 9.1% in 1975 to 4.1% in 1978 (Table 3). This suggests that
during th~ years that wolves were experimentally killed (1976 -
1978), calf survival increased slightly. Following termination
of wolf control and repopulationofthe area by wolves, calf
mortality attributable to wolf predation increased and slightly
exceeded precontrol levels by 1981. During the same period,
starvation accounted for 1.9 -3.2% of the total calf mortality
except during the winter of 1978 -1979. This was considered
a moderately severe winter, and at least 14.9% of the calves
died of starvation.
Yearly Mortality
Trends in yearling moose mortality were similar to those
of calves, except the magnitude of the mortality was substantially
less (Table 3). From 1975 -1979, hunting mortality (assuming
that half of the bull harvest was comprised of yearlings) was the
r~ ,, .
GO -w
(!)
<(
I-z 40
w
0
0: w a; ......
30
>-
I-
...J
<(
I-
0:
0 20
~
w
C/)
0
0 ·10 ~
u.
...J
<(
0 --------------/
0
1975 1976 1977
/
/'
/ ',
/ '
/ '
/ '
1978
YEAR
brown bear predation
woH predation
winter kin
'---. ------------.
1979 1980 1981
Figure 11. Annual rate• of calf moo•• mortality due to predation and winter kill a• determined from modeling
the 8u•ltna River Study Area moon population, 1876-UUU.
,....._.-, ' . ' I
,.-.--.-,
\ l
,,..__...,
r
N
N
0'1
t: I ,, .i (, \l.l \, '· ' J tJ > .J ., .. ) ) J .l J
Table 3. Estimates of spring moose population size, and causes and magnitude of mortality by sex and age class as determined from modeling the
Susitna River Study Area moose population from 1975-76 to 1981-82.
Year 1975-76 1976-77
Age Class Calves Yrlgs. Adults Total Calves Yrlgs. Adults Total
Sex M F 11 F H F Both M F M F H F Both
Spdng Population Est. 811 811 274 274 93 1365 3628. 699 699 272 272 197 1349 3488
l1or-ta lit y
Early Spring and Summer 48 48 0 0 0 0 96 41 41 0 0 0 0 82
Spring ~lolf Predation 36 36 2 2 1 8 85 21 21 1 1 1 4 49
Summer Wolf Predation 18 18 9 9 3 46 103 10 10 5 5 4 24 58
Brown Bear Predation 399 399 19 19 7 96 939 343 343 18 18 13 91 826
!hinting 0 0 51 0 52 0 103 0 0 41 0 42 0 83
muter Wolf Predation 20 20 10 10 4 52 116 13 13 6 6 4 31 73
Winter Kill 18 18 11 5 1 43 60 17 17 2 5 4 44 89
Subtotal 539 539 102 45 68 245 1502 445 445 67 35 68 194 1254
% of Population 66.5 66.5 37.2 16.4 73.1 17.9 41.4 63.7 63.7 24.6 12.9 34.5 14.4 36.0
Year 1977-78 1978-79
Age Class Calves Yrl2s. Adults Total Calves Yd9s. Adults 1'otal
Sex M F M F M F Both M F M F H F Both
N
•N
Spring Population Est. 721 721 254 254 318 1392 3660 753 753 272 272 396 1437 3883 -..,J
Horta lily
Ear-ly Spring and Summer 43 43 0 0 0 0 86 45 45 0 0 0 0 90
Spring Wolf Predation 17 17 1 1 1 4 41 15 15 1 1 1 3 36
Summer Wolf Predation 7 7 3 3 4 18 42 6 6 3 3 4 14 36
Brown Bear Predation 354 354 16 16 20 88 848 370 370 16 16 23 85 880
Hunting 0 0 52 0 52 0 104 0 0 74 0 74. 0 148
Hinter Wolf Predation lO 10 4 4 5 24 57 10 10 4 4 6 23 57
IHntcr Ki 11 18 18 10 5 8 46 105 181 44 17 16 21 48 317
Subtotal 449 449 86 29 90 180 1283 627 490 115 30 129 173 1564
% of Population 62.3 62.3 33.9 11.4 28.3 12.9 35.1 83.3 65.1 42.3 11.0 32.6 12.0 40.3
Table 3. (cont'd)
Year 1979-80 1980-81
Age Class Calves Yr'l:gs. Adults Total Calves Yrlgs. Adults Total
Sex M F f.! F H ~-· Both M F M F M F Both
Spring Population Est. 787 787 126 263 424 1506 3893 796 796 386 386 311 1512 4187
Nortality
Early Spring and Summer 47 47 0 0 0 0 94 47 47 0 0 0 0 94
Spring Wolf Predation 21 21 0 1 1 4 48 32 32 2 2 1 6 75
Summer Uolf Predation 14 14 3 6 9 33 79 18 18 9 9 a 37 99
Brown Bear Predation 276 276 a 16 26 91 693 391 391 21 21 17 82 923
Hunting 0 0 82 0 82 0 164 0 0 0 0 134 0 134
~linter Holf Predation 18 18 4 8 12 44 104 23 23 13 13 10 50 132
Winter Kill 25 25 1 5 11 49 116 18 18 21 8 5 49 119
Sul>total 401 401 98 36 141 221 1298 529 529 66 53 175 224 1576
'1. of Population 51.0 51.0 77.8 13.7 33.3 14.7 33.3 66.5 66.5 17.1 13.7 56.3 14.8 37.6
Year 19al-82
Age Class Calves Yrlgs. Adults Total
Sex !1 F f.! F' H F' BOfh
Spring Population Est. al4 814 267 267 456 . 1621 4239 1
Nortality N
Early Spring and Summer 48 48 0 0 0 0 96 N
Spring Wolf Pn!dation 40 40 1 1 2 8 92 00
Summer Wolf Predation 18 18 7 7 11 40 101
Brown Bear Predation 400 400 14 14 25 87 940
Hunting 0 0 0 0 153 0 153
Hinter Wolf Predation 20 20 8 8 13 46 115
~linter Kill 18 18 14 5 9 53 117
Sul>tota 1 544 544 44 35 213 234 1614
\ of Population 66.8 66.8 16.5 13.1 46.7 14.4 38.1
[
f
I.
L
[
b
L
L
r~
-229.-
largest source of overall mortality (Figure 12), even though only
affecting males. Beginning with the 1980 season, yearlings were
theoretically protected by antler regulations and, therefore,
hunting mortality declined to insignificant levels. Mortality
attributable to wolf predation declined from 7.6% in 1975 to a
low of 3% while wolf control was in effect. Following termination
of wolf control, yearling mortality attributable to wolf predation
increased. Yearling mortality attributable to brown bears
declined during the study period primarily because the model
assumed a stable bear population and the moose popualation was
increasing. Winter mortality (starvation) was quite variable
even during mild winters. The highest winter mortality occurred
during the severe winter of 1978 -1979.
Adult Mortality
Trends in adult mortality were quite similar to those of
yearlings because for both types of predation, it was assumed.
that the sex-age class of kills was dependent on availability
(Figure 13) .
GMU-13 Population Model Analyses
Population Size Estimates
The 1978 -1982 GMU-13 post-calving moose population
trend (15.8% increase) was similar in many respects to that of
the Susitna River Study Area (16.8%). However, the population
declined between 1975 -1976 and 1976 -1977 and again in
1978 -1979 (Table 4). The largest increases occurred between
1979 -1980 (7.5%) and 1980 -1981 (9.9%). The estimated fall
population size based on the model differed considerably from
the population estimate derived from composition counts,
particularly for 1975 and 1976 (Figure 14). This was believed
due to underestimation of both yearlings and calves during
composition counts.
22
20
18
-w
~ 16 < ..... z w
(.)
a: 14
w a. -
> ..... ·12
...J
< .... a:
0 10
:E
w r.n
0 o a
:E
~ z
...J 6 a:
< w
>
2
··. . . . . .. ..
-230 -
hunting
brown bear predation
wolf predation
winter kill
·. . . ... ·
. .
. . . .
. . .
. . . .
. . . •
. .
. .
. . . . . • . • . .
. .
. . .
. -: .. . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. .
• • . .
. . .
' . . . . .-----~ ,_,_-
.(
/ .
/ : . \ / .
\ / ../
' .
0~~------~----~------~------~------~----~------~ 1975 1978 1977 1978
Y E A R
1979 1980 1981
Figure 12. Annual percent yearling bull moose mortality due to several mortality
factors as deter-mined from modeling the Susltna River Study Area In southcentral
Aluka, 197e-1981. ·
[
[
[
[
[
c
[
[
L
l
L
"3
~
~-,
::5
-,
~ :
2:1
I'
--
=
"""
-w
0
< .... z w
0 a: w
0. -
>
1-
-l
<
1-a:
0
~
w en
0
0
~
1-
-l
:;:)
c <
. •
. . •
. -______ ...,. ....
· .....
-231 -
. ... ...
huntlnQ
brown beu predation
wolf predation
winter kill
o•••••••••••o•oo
'-----. ---------
0~~----~------~------+-----~------~------~----~ 1975 1976 1977 1978 1979 1980 1981
YEAR
FIQure 13. Annual adult moose mortality rates by cause as determined from modellnQ
the Su.Jtna River Study Area moose population In aouthcentral Alaska, 197e-198 1.
Table 4. Estimates of spring moose population size, and causes and magnitude of mortality by sex and age class as determined from modeling the moose
population in GMU 13 of southcentral Alaska from 1975-76 to 1981-82.
97 -77
Calves Adults Total Calves Yrlca:s. Adults Total
M F M F Both M F R F M F Both
Spring Population Est. 7230 7230 1098 1098 1269 11822 .29807 5598 5598 3356 3356 1129 10062 29099
Mortality
Early Spring and Summer 433 433 0 0 0 0 866 335 335 0 0 0 0 670
Spring Wolf Predation 486 486 11 11 13 123 1130 535 535 33 33 11 98 1245
Summer Wolf Predation 209 209 57 57 66 615 1213 156 156 111 111 37 333 904
Brown Bear Predation 2124 . 2124 61 61 70 658 5098 2124 2124 159 159 54 477 5097
Black Bear Predation 90 90 4 4 5 46 239 90 90 11 11 4 34 240
Hunting 0 0 358 0 358 0 716 0 0 366 0 366 0 732
Winter Wolf Predation 299 299 80 80 92 865 1715 250 250 176 176 59 526 1437
Winter Kill 233 233 36 23 27 375 927 141 141 160 73 23 328 866
Subtotal 3874 3874 607 236 631 2682 11904 3631 3631 1016 563 554 1796 11191
% of Population 53.6 53.6 55.3 21.5 49.7 22.6 39.9 64.9 64.9 30.3 16.8 49.1 17.9 38.5
1977-78 1978-79
Calves Yrlca:s. Adults Total Calves Yrlgs. Adults Total
M F M F M ·F M F M F M F M F Both
N
Spring Population Est. 5322 5322 1657 1967 2915 11059 28552 5751 5751 1972 1972 3231 10930 29607 w
Mortality N
Early Spring and Summer 319 319 0 0 0 0 638 345 345 0 0 0 0 69
Spring Wolf Predation 333 333 12 12 18 67 775 247 247 9 9 14 49 575
Summer Wolf Predation 157 157 65 65 97 368 909 128 128 53 53 87 294 743
Brown Bear Predation 2124 2124 93 93 138 525 5097 2124 2124 93 93 152 513 5099
Black Bear Predation 90 90 7 7 10 37 241 90 90 7 7 11 36 241
Hunting 0 0 428 0 428 0 856 0 0 432 0 432 0 864
Winter Wolf Predation 190 190 78 78 116 440 1092 173 173 70 • 70 115 390 991
Winter Kill 137 137 81 42 80 362 839 1608 397 137 43 182 361 2728
Subtotal 3350 3350 764 297 887 1799 10447 4652 4652 801 275 993 1643 11868
% of Population 62.9 62.9 38.8 15.1 30.4 16.3 36.6 80.9 60.9 40.6 13.9 30.7 15.0 40.5
,. '
•llhiJ
Table 4. (cont'd)
Spting Population Est.
Mortality
Early Spring and Summer
Spring Wolf Predation
Summer Wolf Predation
Brown Bear Predation
Black Bear Predation
Hunting
Winter Wolf Predation
Winter Kill
Subtotal
% of Population
Spring Population. Est.
Mortality
Early Spring and Summer
Spring Wolf Predation
Summer Wolf Predati.>n
Brown Bear Predation
Black Bear Predation
Hunting
Winter Wolf Predation
Winter Kill
Subtotal
% of Population
Calves
8 F
5571 5571
346 346
281 281
88 88
2124 2124
90 90
0 0
117 117
170 170
3216 3216
55.7 55.7
Calves
M F
6307 6307
378 378
218 218
97 97
2124 2124
90 90
0 0
123 123
204 204
3234 3234
51.3 51.3
IL "'" j ~;. :I. I J w ,,JJ
1979-80
Yrlgs. Adults
R F 8 F
1036 2247 3409 10984
0 0 0 0
5 12 18 57
18 40 61 195
50 108 164 528
4 8 12 37
500 0 500 0
25 55 83 267
27 49 95 366
629 272 933 1450
60.7 12.1 27.4 13.2
1981-82
Yrlgs. Adults
M F M F
2720 2720 4155 12312
0 0 0 0
9 9 13 40
43 43 66 195
105 105 161 477
7 7 11 34
0 0 794 0
56 56 86 255
153 61 111 416
373 281 1242' 1417
13.7 10.3 29.9 11.5
'· .J I j l i
1980-81
Total Calves Yrlgs. Adults Total
BCifil 8 F H F H F BCifil
29218 5958 5958 2555 2555 2833 ll509 31418
692 337 337 0 0 0 0 674
654 258 285 11 11 12 50 600
490 123 123 57 57 65 258 683
5098 2124 2124 111 111 126 501 5097
241 90 90 8 8 9 35 240
1000 0 0 0 0 557 0 557
664 106 106 51 51 58 231 603
877 180 180 142 56 76 383 1017
9716 3218 3218 380 294 903 1458 9471
33.3 54.0 54.0 14.9 11.5 31.3 12.7 30.1
Total N
Both w
w
34521
756
507
541
5096
239
794
699
1149
9781
28.3
I""""": \. ,I,,
C/)
1-z
:::>
0 o eo...,-
z
0
1-
~ 60_!_
o. I ~ I
o I
lL
-C/)
0 z
.10
< 60
C/)
:::>
0
J:
1-
0 401 -<tO
0:
:::>
0
J:
I
I a ao-1-
~ I
ffi I
~ 20-1
I ~ I g I
~ 101
~ I
0:
UJ m
~
:::> z
I
UJ
C/)
0
0
:::e 30
lL
0
IJC w
m
~
:::> z
tO
Y E A A
-mooae model
--population baaed
on compoaltlon counte
--mooae obaerved/hour
of compoaltlon counh
fiauro 14. fall moou population eatlmatea •• derived from modeling veraua annual composition counta for Game
Management Unit 13 of aouthcentral Aluka, 1Q78-1981S~
r----"l
I
.-. ;
-
-235 -
Calf Mortality
Brown bear predation was responsible for more calf mortality
than wolf predation or winter mortality (Figure 15). Except during
the severe winter of 1978 -1979, wolf predation was the second
most important cause of calf mortality (Figure 15). Mortality of
calf moose was higher in the GMU-13 than in the wolf control area,
particularly in 1976-1977 when wolves preyed upon 17.3% of the
estimated number of calves produced. As wolf densities declined
intheunit, primarily from hunting and trapping activities, the
estimated percentage of calves preyed upon by wolves declined
each year, reaching a low of 7.0% during 1981-1982. Calf
mortality studies conducted in 1977 and 1978 suggested that 3% of
the calf mortalities during the first 6 weeks following birth
were attributable to wolf predation (Ballard, et al., 1981).
Independent modelling estimates suggested that calf mortality
attributable to wolf predation ranged from 4.3 to 6.3% during
the same years. Therefore, both approaches suggested that wolf
predation on newborn moose calves was a secondary source of calf
mortality.
Adult Mortality
Wolf predation on adult moose in the GMU-13 also declined
during the study period (Figure 16), ranging from 13.5% in 1975
to 4.0% in 1981. The decline in wolf-related adult mortality
was due to a decrease in the wolf population and concurrent
increases in the moose population. Similarly, percent annual
adult mortality from brown bear predation also declined (5.5 to
4.8%) but this was primarily the result of increases in the
moose populetion since we assumed that bear populations were
stable during the study.
During the study, adult mortality attributable to hunting
increased primarily because of changes in hunting regulations in
1980 which placed all harvest pressure on adult bulls only.
-w
(!}
<( ..._
z w
0 a: w
0. -
>-._ --I
<( ..._
a:
0
:::E
w
(f)
0
0
:::E
lL
-I
<(
0
40
30
20
10
---
I
I -.---------J
I
I
I
I
1\
I \
I \
I \
\
\
\
\
\
\
--brown beer predation
-wolf predation
- -winter kll
"-------------
0_.~-------+------~--------~------._------~------~------~ 197~ 1978 1977 1978 1979 1980 1981
YEA A
Figure 15. Eatlmated .annual rate a of calf mortality from predation and winter kill determined from modeling the Game
Management Unit 13 mooae population of aouthcentral Alaaka, 1078-108L
~
L., , , J r---"1 \.,' ,,J ,:-=J ,.---,
' j
,,....,_.._..,
N w m·
......
w
(!)
<( .... z
UJ
0
0:
UJ a. -
L_. ..J l. .. J
14
12
10
IJ .! .. J
••••• hunting
brown bear predation
wolf predation
winter ktn
>-8 ....
...J
<( ....
0:
0
:::E
UJ
(/)
0
0
~
1-
...J
::J
0
<(
8
2
.--......------....___ .......... . ......
... . . .
------..... ~_!. ---__ :""T,_....a.t..!.!.!.:...L ............. ~ •• ;-;;-:.......... •• •• • • • • ----------------..... . . . . . .
'0~+-------~------~--------+-------~------~--------+-------~ 197t5 1978 1977 1978
Y E A A
1979 1980· 1981
Figure 16. Annual Game Management Unit 11 adult mooae mortamy ratea from four factora eatlmated from
modeling. 1911!i-UUU.
N
w
.......
-238 -
Wolf. Predation
Earlier analyses of the effects of decreased wolf
densities. (from wolf control) on moose calf survival suggested
that no significant increases had occurred because ratios of
various sex and age classifications had fluctuated similarly
between control and non-control areas (Ballard, et al., 1~81).
Although the reductions in wolf density were substantially
larger in the wolf control area, wolf densit~es in both the
wolf control area and GMU-13 decreased from 1975 levels, while
moose populations in both areas increased (Figur~ 17). · Reductions
in both calf mortality from 9 -17% annual mortality to 4 -7%,
and adult moose mortality from 8 -10% to 3 -4% annual mortality
probably contributed to the increases in the moose populations.
Because wolf d~nsities declined in both areas, it would be
expected that the sex-age ratios would fluctuate similarly.
Although wolf predation was not the primary sou~ce of moose
mortality, its reduction, in combination with several mild winters,
appears to have allowed both moose populations to.increase.
Substantially larger increases could probably be anticipated if
the level of bear predation was also reduced.
From November 1 through May 15 each year, mortality of
moose from wolf predation is relatively high on a superficial
basis, but on a population level, is relatively minor. For
[
[
[
[
[
[
b
[: _,
6
c
[ example, in both the experimental area and GMU-13, wolf predation
accounted for 6.5 and 7.7% mortality, respectively, of the calves
present on November 1, 1975. However, of the total calves produced, U
this source of mortality represented only 2.3 and 4.1% respectively.
From this comparison, it would be easy to conclude from flights [ __ ·
made during winter when wolf kills are most noticeable that wolf .
predation was a much more important source of moose mortality
than what it actually represents on a population basis. L
L
L
GAME MANAGEMENT UNIT 13 ·MOOSE (THOUSANDS)
>., ~ ~ • .. r; 0 0 0 0 0 ~ c -,-r--r-t---r-·t---r-1 II -.
0 0 -... SUSITNA STUDY AREA MOOSE (THOUSANDS) • "-l .
0 • .. tla Cot • 01 s. > -~-~-r--l--r--l--r-l---r-1 ::r::s
0 :I
0 c GAME MANAGEMENT UNIT 13 WOLVES :I • --... ,_ . --~ Cot • 01 -· >= 0 0 0 0 0
0 0 0 0 0 !i"a
.. 0
:ll" 0 !' •
0 SUSITNA STUDY AREA WOLVES ...
Cot • C) :I -.. ~~ 0 0 ·0 0
l ~
e2.
F
ca-~
:*'0
0 ...
'0 CD r-c .. ..., -Q
0 '-
::t -~ ... I 0 ... :I
~ CD I b • ....
co ct \ 0 -\ ~ • \ • ... :I \ C) CD
• ..... \ a ...,
• \ -< L
~ \ • :I • m ... \ c CD
0 ..... I 3 > CD 0 I :I -:0 I c: :. ... I --CD I II I 1
Cot ..... \ II CD I \ ::t \ ~ CDC) Cit C) \ \ ·~ -c~c~ '::1' \ 0 --!.c:~c \ CD --\ CD CD ::1~::1 .. \ c ·-·-!. 0 CDc.acac.a \ ' -:I - -' II c:~C:3 \ ~ 0 ~0 I ::! '<-'<o \ c --c •. I 0 co > o >o \ ... CD .. . ... ' 0 0 \ CD --II II ' -c: E a 0 \ ~ • '< 0 0
0 c • 0 0 • -6£(: -.
-240 -
Summary
Development and ref~nement of the models has identified
a number of areas where our understanding of moose population
dynamic processes is incomplete. Probably the most important
data gaps relate to the importance of various types of predation.
Although black bears are quite numerous in the western half of
the hydroelectric project study area, their importance as predators
of moose has not been investigated. If black bears are in fact
significant predators of moose, the addition of this factor to
the model could greatly alter our interpretations of the potential
impacts of the project on moose. Also, it became quite evident
that our 1978 estimates of brown bear predation on moose were
much too high, requiring additional study. Although a considerable
volume of information has been collected on wolf populations,
additional refinement of the relationships between snow conditions
and wolf population processes is needed.
Both models relied heavily on the moose population estimates
derived in 1980. To provide a validation of the model, the areas
should be recensused in 1983 or 1984. Moose studies should be
continued up to and through a severe winter. Currently, our
[
[
[
r
[
[~
[
b
[
[
[
estim~tes of starvation mortality during severe winter conditions h
u
are little more than guesses.
[
[
[
[
L
C
-241 -
REFERENCES
Ballard, W.B. and K.P. Taylor, 1980. Upper Susitna Valley moose
population study. Alaska Dept. Fish and Game. P-R Proj ..
Final Rep., W-17-9, W-17-10, and W-17-11. 102 pp.
Ballard, W.B., S.D. Miller, and T.H. Spraker, 1980. Moose calf
mortality study. Alaska Dept. Fish and Game. P-R Proj.
Final Rep., W-17-9, W-17-10, W-17-11, and W-21-1. 123 pp.
Ballard, W.B. T.H. Spraker, and K.P. Taylor, 198la. Causes of
neonatal moose calf mortality in southcentral Alaska.
J. Wildl. Manage. 45(2): 335-342.
Ballard, W.B., R.O. Stephenson, and T.H. Sprake~, 198lb. Nelchina
Basin Wolf Studies. Alaska Dept. Fish and Game. P-R Proj.
Final Rep., W-17-9 and W-17-10. 201 pp.
Ballard, W.B., C.L. Gardner, J.H. Westlund, and J.R. Dau, 1982a.
Big Game Studies, Vol. V, Wolf. Alaska Dept. Fish and
Game Spec. Rept. to Alaska Power Authority. 220 pp.
Ballard, W.B., C.L. Gardner, and S.D. Miller, 1982b. Nelchina
Yearling Moose Mortality Study. Alaska Dept. Fish and
Game. P-R Proj. Final Rep., W-21-1 and W-21-2. 37 pp.
Ballard, W.B., R.O. Stephenson, S.D. Miller, K.B. Schneider, and
S.H. Eide, In Prep. Ecological studies of timber wolves
and predator-prey relationships in southcentral Alaska.
Wildl. Monogr.
Blood, D.A., 1973. Variation in reproduction and productivity
of an enclosed herd of moose (Alces alces) . XI Intern.
Congress of Game Biologists, Stockholm.
-242 -
Eide, s~ and W.B. Ballard, 1982. Apparent case of surplus
killing of caribou by gray wolves. Can. Field-Nat.
96: 87-88.
Floyd, T.J., L.D. Mech, and P.A. Jordan, ~978. Relating wolf
scat content to prey consumed. J. Wildl. Manage. 42(3):
528-532.
Fuller, T.K. and L.B. Keith, 1980. Wolf population dynamics
and prey relationships in northeastern Alberta. J.
Wildl. Manage. 44:583-602.
Gasaway, W.C., S.J. Harbo, and S.D. Dubois, 1979. Moose survey
pro"cedures development. Alaska Dept. Fish and Game. P-R
Proj. Rept. 87 pp.
Gasaway, W.C., S.D. Dubois, and S.J. Harbo, 1982. Moose Survey
Procedures Development. Alaska Dept. Fish and Game. P-R
Proj. Final Rept. 66 pp.
[
[
[
[
c
[
[
Gasaway, W.C., R. Stephenson, J. David, P. Shepherd, and 0. Burris, [
1983. Inter-relationships of moose, man, wolves and
alternate prey in Interior Alaska. Wildl. Mongr. In Press. b
Hayne, D.W., 1978. Experimental designs and statistical analysis
in small mammal population studies. Pages -3-10 in A.P.
Snyder, ed. Populations of small mammals under natural
conditions. Vol. 5, Spec. Publ. Ser., Rymatuning Lab.
of Ecol., Univ. of Pittsburg, Pittsburg.
Miller, S.D. and W.B. Ballard, 1982. Homing of transplanted
Alaska b.cown bears. J. Wildl. Manage. 46: 869-876.
Miller, S.D. and W.B. Ballard, 1983. Density and biomass
estimates for an interior Alaskan brown bear population.
Can. Field Nat.: In Press.
[
L
L
d
l
l
b
-~4J -
Mohr, c.o., 1947. Table of equivalent populations of North
American small mammals. Am. Midl. Nat. 37(1): 223-249.
VanBallenberghe, v., 1978. Final report on the effects of the
Trans-Alaskan Pipeline on moose movements. Alaska Dept.
Fish and Game. 44 pp.
. '-j .• ~ ' I •