HomeMy WebLinkAboutAPA230EIDC 707 A STREET . ANCHORAGE. ALASKA.99501; (9071 279-4523
STREAM FLmv AND TEMPERATURE MODELING
IN THE SUSITNA BASIN, ALASKA
FINAL REPORT
ARCTIC ENVIRONMENTAL INFORMATION AND DATA CENTER
STREAM FLOW AND TEMPERATURE MODELING
IN THE SUSITNA BASIN, ALASKA
FINAL REPORT
STREAM FLO'i\T AND TEMPERATURE MODELING
IN THE SUSITNA BASIN, ALASKA
FINAL REPORT
Prepared by:
Arctic Environmental Information
and Data Center
University of Alaska
707 A Street
Anchorage, Alaska 99501
Submitted to:
Harza-Ebasco Joint Venture
8740 Hartzell Road
Anchorage, Alaska 99507
For:
The Alaska Power Authority
334 W. 5th Avenue, Second Floor
Anchorage, Alaska 99501
June 30, 1983
This report was prepared by the following
AEIDC staff
William J. Wilson, Principal Investigator
Ken A. Voos, Ph.D., Environmental Engineer
Paul R. Meyer, Hydrologist
Lynn D. Leslie, Climatologist
Beverly J. Valdez, Information Coordinator
TABLE OF CONTENTS
PAGE NO.
LIST OF FIGURES.. . . . . . . . . . • • . . . . . . . • . . . . . . • . . . . . . . • . . . . . . . • • . . . . . . . • . vi
ACKNO"'VlLEDGmf.ENTS •••••••••••••••••••••••••••••••••••••••••••••••••••• viii
INTRODUCTION. • • • • • • • • • • • • • • • • • • • • • • • • • • . • • • • • • • • . .. • • • • • • • • • • • • • • • • • • 1
WATER BALM~CE ACCOUNTING FOR THE SUSITNA BASIN...................... 3
INTRODUCTION ••••••••••••••..•.••••••••••
DESCRIPTION OF THE WATER BALANCE MODEL ••
METHODS TO APPORTION SUB-BASIN WATER .•••
Method I. Linear Watershed Area Contribution.
Areal Precipitation Weighting •• Method II.
Method III. Water-Yield Weighting ••••
TESTING THE C COEFFICIENTS ••••••••
C COEFFICIENTS TEST RESULTS .•
USE OF RELATIVE PRECIPITATION WEIGHTING •••
May through September •.
October through April ••
RESULTS AND DISCUSSION ••••••
3
3
5
7
7
9
10
11
12
12
13
13
STREAM NETWORK TEMPERATURE SIMULATION MODEL ••••••••••••••••••••••••• 17
INTRODUCTION ••..•••••.••••••••••.•.•••••••••
DESCRIPTION OF THE STREAM TEMPERATURE MODEL.
HODIFICATIONS •••
STREAM NETWORK •••
STREAM STRUCTURE ••
Topographic Shading.
Stream Widths ••••••••
Hydraulic Retardance.
Tributary Assumptions.
HYDROLOGY ••••••
Flows •••.•
Stream Temperatures •••
Observed Temperatures.
Synthetic Temperatures.
METEOROLOGY •••
USGS Cantwell gage on the Susitna River
(ID1223.7) ............................ .
USGS gage data collected on the Chulitna
and Talkeetna rivers •••••••.••••
Temperatures of Distributed Flow.
Selection of Meteorologic Data .•••
Ground Reflectivity and Atmospheric Dust •.•
Meteorologic Predictions... • ••••..
VALIDATION ..•.
CALIBRATION.
RESULTS AND DISCUSSION ••.••••••••
iv
17
18
21
21
24
24
27
29
29
29
29
30
30
30
30
33
33
39
39
40
40
46
51
54
TABLE OF CONTENTS (Continued)
PAGE NO.
FUTURE APPLICATIONS AND ENHANCEMENTS .•••••••••••••••••••••••.••••••• 55
BIBLIOGRAPHY. . • . • • • • • • • • • • • • • • • • • • • • . • • • • • • • • • • • • . • • • • • • • • • • • • • • • • • • 57
APPENDICES
A. TOPOGRAPHIC SHADING. . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . A-1
B. WIDTH/FLOW FUNCTIONS. • • • • • • • • • • • • . • • • • • • • • • • • • • • • • • • • • • • • • B-1
C. LONGITUDINAL TEMPERATURE PROFILES, JUNE TO
SEPTEJ1BER 1981-1982 ..........•..•......................... C-1
v
FIGURE
NUMBER
1.
LIST OF FIGURES
PAGE NO.
Flow balance sub-basins, Cantwell gage to Sunshine
gage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2. Period of record for gage stations used in H20BAL......... 6
3. Calculated mean annual precipitation and water-
yield values, Cantwell to Sunshine Basin.................. 9
4. C coefficients, Cantwell to Gold Creek Basin •.•.•••••••••• 11
5. Mean pre-and postproject June flow profiles, Watana
to Sunshine Station, using precipitation-weighting
water balance method...................................... 14
6. Mean preproject June flows, Watana to Sunshine
Station, using three weighting methods •••••••.••.••••••••• 15
7. Vertical air temperature distribution. Anchorage and
Fairbanks 1968, 1969, 1970, 1980, 1981, 1982 ••••••.••••.•• 22
8. Vertical relative humidity distribution. Anchorage
and Fairbanks 1968, 1969, 1970, 1980, 1981, 1982 •••.•••..• 23
9. Stream network from Watana to Sunshine ••••••.•••••••..•••• 25
10. Susitna mainstem reach definitions •.•••••.••••.•••••••.•.. 26
11. Tabular values of width function parameters ••••••••••••••• 28
12. Monthly stream temperatures, available data June to
Sept. 1980, 1981, 1982 .................................... 31
13. Monthly stream temperatures, usable data June to Sept.
1980, 1981, 1982 .......................................... 32
14. Temperature regression for Susitna River at Cantwell
gage...................................................... 34
15. Temperature regression for Chulitna River at USGS gage •.•• 35
16. Temperature regression for Talkeetna River at USGS gage ••• 36
17. Ground reflectivity and atmospheric dust coefficients,
Matanuska Agricultural Experiment Station, Palmer •••.••••• 41
18. Susitna Basin observed air temperatures vs. tempera-
tures predicted from Talkeetna data •..•••••••••••..•.•••.• 42
vi
FIGURE
NUMBER
19.
LIST OF FIGURES (Continued)
PAGE NO.
Susitna observed humidities vs. humidities predicted
from Talkeetna data....................................... 44
20. Average monthly wind speeds (M/S), 1980, 1981, 1982 ••••••• 47
21. Predicted vs. observed solar predictions, 1980, 1981,
1982...................................................... 48
22. Tributary temperatures; 3 C groundwater inflow assumed •••• 49
23. Tributary temperatures; postcalibration, including
distributed flow temperature model •••••••••••••••••••••••• 52
24. Temperature model calibration statistics for tributary
predictions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
vii
ACKNm\ILEDGEMENTS
This report is the result of a cumulative effort by a number ·of
people. Special appreciation is extended to Gail Heineman (ADF&G) for
facilitating necessary data transfers; Tim Quane and Jay Sauntner (both
of ADF&G) and Carl Schoch (R&M) for sharing their on-site knowledge;
Fred Theurer (IFG) and Woody Trihey for their technical guidance; and
Jeff Coffin (R&M) for his continued assistance throughout the project.
At AEIDC, we would like to thank the Resource and Sciences staff for
their ongoing support in all phases of this report; Information Services
staff for library and bibliographic services; Judy Brogan for her
editorial assistance; and Nancy Walters and Sherrie Cornett for the
production of this report.
viii
INTRODUCTION
The Alaska Power Authority, through Harza-Ebasco Joint Venture,
contracted with the University of Alaska's Arctic Environmental Informa-
tion and Data Center (AEIDC) to simulate postproj ect physical habitat
conditions in the Susitna River drainage with a computerized model
system. Water balance and stream temperature models permit the simu-
lation of unmeasured water discharges and temperatures at various
locations downstream from the proposed Watana or Devil Canyon dams
(AEIDC 1983). These predictions are necessary for the analysis of
project impacts on downstream fishery populations and habitats and will
allow identification of appropriate streamflow regimes to minimize
negative effects and aid mitigation efforts.
Determination of stream temperatures requires flow data at various
mainstem and tributary locations. This is the main purpose of the
Susitna water balance model. Water temperature is important because it
has various effects on fish behavior, including habitat selection,
migration, movement patterns, food selection, and the physiological
functions associated with growth and metabolism. It has a direct effect
on the time required for salmonid egg development. Many studies have
illustrated the relationship between small temperature change over long
periods of time and salmonid egg incubation (Reiser and Bjornn 1979).
Temperature has also been implicated as a factor affecting the timing of
outmigration of smolts and inmigration by adult spawners (Brett 1971;
Coutant 1970; Cherry, et al. 1975; Reiser and Bjornn 1979). These
physiological and behavioral functions may be altered by temperature
changes of as little as 0.5 to 1.0 C.
For these reasons, it is important to predict downstream tempera-
tures accurately and at the specific locations where fishery habitat may
be affected. Tributary flows and temperatures also should be simulated
so that the dilution or buffering effect of tributaries on the mainstem
can be understood. Water balance and temperature predictions will also
be critical to the river ice modeling efforts of Harza-Ebasco.
This report is organized into three major sections. The first
section describes the water balance model and hydrologic data synthesis.
The second section provides a description of the stream temperature
1
WATER BALANCE ACCOUNTING FOR THE SUSITNA BASIN
INTRODUCTION
The task of water balance accounting in the Susitna Basin is one of
defining the methodology to assign inflows between known flows at
mainstem gage stations. The lack of hydrometeorologic data in this
region makes this a difficult task, subject to a number of gross
assumptions. Three basin water apportionment methods have been explored
and are discussed in this section. AEIDC developed a computer program
to employ these apportionment methods, generating time series of flows
at a number of mainstem and tributary locations within the Susitna
Basin. Output files containing these flows are directly usable as input
to the stream network temperature model (SNTEMP).
DESCRIPTION OF THE WATER BALANCE MODEL
The water balance accounting program, H20BAL, was designed to
operate on the Susitna Basin between the USGS gages at Cantwell (Vee
Canyon) and Susitna Station. AEIDC's initial modeling efforts focus on
the reach from the Watana dam site to the USGS gaging station near the
Parks Highway bridge at Sunshine. The Chulitna and Talkeetna river
flows are incorporated into the system at the gage station on each river
near Talkeetna.
The basin between Cantwell and Sunshine Station was divided into 16
sub-basins (excluding the Chulitna and Talkeetna basins above their
respective USGS gages) for the purpose of water apportionment. These
basins center around the larger tributaries and are defined by drainage
divides (Figure 1). They do not necessarily follow the watershed
boundaries of any single stream, often including drainages of three or
more streams. In most of the sub-basins, a node location on the
mainstem river was chosen, representing the point source for all inflow
to the mainstem. For the few sub-basins without a dominant tributary,
inflow is linearly distributed along the adjacent mainstem reach.
The accuracy associated with assigning flow ~•ithin a basin between
gage stations increases as the distance between gage stations decreases.
Thus, it is advantageous to use as many data stations within the basin
as are available. Gaps in historical data records exist at some
3
USGS G<JQe/Node
Locatio~
USGS GaQa SIOtion
Sub-Basin ldentlflccrtlon
Sub-Ba:r.ln Boundary
Oam Sihll
)
FLOW BALANCE SUB-BASINS,
Cantwell Gage lo Sunshine Gage
Scale i"' 8 miles
(Ill
p.l
rJQ
I'D
(ll
~
cr'
f
cr'
Ill
Ul
1-'• p
Cfl
(")
Ill p
r1"
~
(1)
t-'
t-'
00
p.l
rJQ
(1)
stations within the basin (see Figure 2 for historical flow data
periods). Rather than discarding all data at a gage with occasional
gaps, we used linear regression to fill them.
The H20BAL program requires input data for the following USGS gage
stations:
Susitna River near Cantwell (Vee Canyon)
Susitna River at Gold Creek
Susitna River at Susitna Station
Chulitna River near Talkeetna
Talkeetna River near Talkeetna
We used flows at the Yentna River gage for the period that they are
available. For the present extent of simulation, flow data at Watana
are preferable to those at Cantwell, and flows at Sunshine are used
instead of those at Susitna Station. These additional stations provide
for greater accuracy by effectively reducing the size of the basin under
consideration. Usable statistically-filled 32-year data sets are
available for the Cantwell, \Vatana, Devil Canyon, Chulitna, Talkeetna
and Susitna Station sites (Acres 1983a).
A filled data record is also available for Sunshine gage but was
not used in H20BAL because of resulting flow deficits in the Gold Creek
to Sunshine reach. These deficits occur when the sum of flows at the
Gold Creek, Chulitna, and Talkeetna gages exceed the synthesized flow at
Sunshine Station. The alternate method used to assign flows at Sunshine
was to assume that the flow-per-unit-drainage-area contribution to the
mainstem was the same for the Gold Creek to Sunshine basin as it was for
the summed Gold Creek, Chulitna, and Talkeetna drainages. The limited
accuracy of this method is acceptable considering this sub-basin
comprises only 3.3 percent of the total drainage area defined at
Sunshine.
METHODS TO APPORTION SUB-BASIN WATER
Once data records are collected or filled for the skeletal gage
station network, interstation flows incrementally increase downstream by
the following relationship:
(1)
5
Figure 2. Period of record for gage stations used in H20BAL.
Gage Station Water Years
Susitna near Cantwell
Susitna at Watana
Susitna at Gold Creek
Chulitna River
Talkeetna River
Susitna at Sunshine
Yentna River
Susitna at Susitna
Station
Beginning '70 '7l '72 '73 '74 '75 '76 '77 '78 '79 '80 '81 '82
5/61-------------6/80-----
7/80-----
10/49------------------==-----------~~~~~---------------------------
2/58---------------------5/80-·---
6/64----------------------~---------------------·-----------------
5/81---
10/80-----
10/74----·----------~-------------------
where:
and
Q is the mean flow for the given period (L 3 /t),
C is a fractional constant determined from some combination of
watershed area, areal precipitation, and water yield esti-
mates (decimal),
s, s - 1 are subscripts referring to mainstem locations,
numerically increasing for each sub-basin downstream,
1, 2 are subscripts referring to mainstem gage locations,
numerically increasing downstream,
L, t refer to dimensions of length and time respectively
The node structure defining the network of sub-basins is fixed
(nonvariable) within the water balance model. The different values of
the C coefficients are selected as input options. We developed three
different methods for determining values of the C coefficients.
Method I. Linear Watershed Area Contribution--Acres (1982) used
this method to determine flow series at proposed dam sites
between the USGS gages at Cantwell (Vee Canyon) and Gold
Creek. A sub-basin that drains 10 percent of the basin area
between gage stations is consistently assigned 10 percent of
the difference in flow between these two sites.
coefficients are defined by:
The C
c = s
where:
A s
A is the planimetered area (L 2), and subscripts refer to
sub-basin, s, and the total basin, b.
(2)
Method II. Areal Precipitation Weighting--The purpose of this
method is to incorporate the weight of relative sub-basin
precipitation into the C coefficients. These coefficients are
now defined by:
7
8
c s
where:
p A s s
Pis the mean annual precipitation (L), and the subscripts
and other variables remain as previously defined.
(3)
The methods employed to determine the mean annual
precipitation for each sub-basin are important to note, since
a great amount of subjectiveness is involved. The primary
data source for the precipitation distribution was a
statewide, annual precipitation isohyetal nap prepared by
James Wise (1977), Alaska state climatologist. This map is
contoured in 10-in intervals for the 10-to 40-in annual
precipitation range, and 20-in intervals above 40-in annual
precipitation. These isohyetals were redrawn on a
1:250,000-scale map of the Susitna Basin. Additional
isohyetals were interpolated between each of the existing
ones, resulting in 5-in contour intervals in the 10-to 40-in
range, and 10-in intervals in areas with over 40-in of annual
precipitation.
To find average precipitation for each basin, we assumed
that the total precipitation between two isohyetals could be
estimated as the product of the area between the isohyetals
(found by polar planimetry) and the average of the two
isohyetal values. These products were summed for all of the
intercontour areas within a sub-basin and then divided by the
sub-basin area to determine the average annual precipitation.
The same process was used to find the mean annual precipi-
tation for the entire basin (Figure 3).
Figure 3. Calculated mean annual precipitation and water-yield values,
Cantwell to Sunshine Basin.
Mean Annual Mean Annual
Sub-Basin Name Precipitation Water Yield
(in.) (in.)
Clarence 49.0 10.1
Kosina 50.4 19.1
Watana 51.8 18.9
Deadman 35.8 23.4
Tsusena 26.0 26.8
Fog 33.7 22.3
Devil 20.1 22.0
Chin-Chee 17 .o 18.0
Portage 18.0 26.8
Indian 24.8 22.1
Curry 31.8 25.0
Whiskers 30.3 22.8
Chulitna 30.7 24.0
Talkeetna 24.7 14.4
Trapper 30.5 20.3
Sunshine 17.5 12.7
Method III. Water-Yield Weighting--A report by Evan Merrell of
the U.S. Soil Conservation Service (1982) suggests a third
method for determining the C coefficients. In this report
Merrell uses precipitation and evapotranspiration estimates to
develop a mean annual water-yield map of the Susitna Basin.
To incorporate the relative weights of sub-basin water yield
estimates, the C coefficients are defined as:
c s
y A
s s (4)
9
where:
Y refers to the mean annual water yield (1), and the
remaining variables are as defined previously.
The mean annual water-yield values for each sub-basin
were determined in the same manner as the mean annual
precipitation values. The water yield isopleths were redrawn
on a base map of the basin, along with the sub-basin outline.
The exception to note is that no isopleths were interpolated
between those given by Merrell. Once again, polar planimetry
was used to determine the areally-weighted basin water-yield
values (refer to Figure 3).
TESTING THE C COEFFICIENTS
The C coefficients determined for any of the methods will sum to
the value 1.0 over the basin defined by two gage stations. A variety of
basins can be defined within the area of concern by using different
pairs of gaging stations. As previously discussed, increased accuracy
results from using data at all available gage stations.
The applicability of each method was tested by determining the
three sets of C coefficients for the Cantwell to Gold Creek basin and
applying these methods to the period for which historical records are
available at the Watana dam site. The predicted values were then
compared to the historical record at Watana. Figure 4 gives the C
coefficients for the Cantwell to Gold Creek basin. Predicted flow at
Watana is given by:
(5)
where:
10
subscripts w, c and g refer to Watana, Cantwell and Gold Creek
respectively, and Q and C are as previously defined.
The calculated C for each of the three methods is:
w
Method I
0.5104
Method II
0.6759
Method III
0.4636
Figure 4. C coefficients, Cantwell to Gold Creek Basin.
c Coefficients
Meth I Meth II Meth III
Sub-Basin Name Area p y A p A y A s s s s s s s
(mi 2)
pb yb \ Pb\ Yb\
Clarence 76.8 1.3660 .4741 .0383 .0524 .0181
Kosina 485.1 1. 4049 .8954 .2421 .3401 .2153
Watana 242.4 1.4439 .8837 .1210 .1747 .1090
Deadman 218.4 .9979 1.0971 .1090 .1087 .1212
Tsusena 191.5 • 7248 1. 2529 .0996 .0693 .117 5
Fog 175.0 .9394 1.0428 .0873 .0820 .0891
Devil 174.5 .5603 1.0310 .0871 .0488 .0896
Chin-Chee 94.2 .4739 .8425 .0470 .0223 .0399
Portage 186.4 .5018 1.2548 .0930 .0467 .1150
Indian 159.4 .6913 1.0358 .0796 .0550 .0852
The mean observed value of C for the 13 months of record when data were w
collected at all three stations (Cantwell, Watana, and Gold Creek) was
0. 6034, with a standard deviation of 0.1119. It is important to note
that these data were collected during the June through November period,
and may not be representative for the entire year. However, since
approximately 82 percent of the annual flow occurs during this period
(based on the 1950 through 1979 flow record at Gold Creek), this period
of record appears adequate.
C COEFFICIENTS TEST RESULTS
One conclusion that can be drawn from this simple test is that none
of the three methods show clear superiority. Based on the estimates of
C , we preferred Method II, the relative precipitation weighting scheme, w
for determining C coefficients; however, a couple of points concerning
these three methods should be mentioned. One concerns the differences
resulting from use of the linear drainage area method and the observed
11
flows at Watana. Acres (1982), using a drainage area-based C value of w
0.515, calculated a synthesized mean annual flow at Watana of 8023 cfs
(Acres 1983a). The observed C value of 0. 6034 applied to the same w
32-year period results in an annual flow of 8338 cfs. This constitutes
a 3. 9 percent increase in available water for the Watana reservoir.
Though the magnitude of this increase seems insignificant, it indicates
that any error would probably be on the side of underestimating water
supply at Watana. Second, the water-yield map used for Method III was
developed to consider the smaller topographic features of the Susitna
Basin, while the precipitation map used for Method II has considerably
less topographic resolution. Consequently, greater utility would be
expected from the increased sophistication of Method III. The
water-yield map, however, apparently underestimates the contributions of
the upper basin (Cantwell to Watana) substantially. In calibrating the
map, Merrell was restricted to the available gage data at Cantwell and
Gold Creek.
If used on a small scale sub-basin such as Cantwell to Watana,
Method III might prove to be much more accurate than Method II.
However, the lack of flow data for the smaller tributaries presently
makes this assumption untenable.
USE OF RELATIVE PRECIPITATION WEIGHTING
Method II accepts the premise that the sub-basin watersheds
contribute to mainstem flow in amounts relative to the distribution of
mean annual precipitation. However, actual watershed conditions exhibit
strong seasonal influences which must be considered. Consequently, the
year was divided into two periods for application of this method.
12
May through September. Flow in the early part of this period (May
through June) is dominated by the melt of winter precipitation.
During July through September, when storm events contribute a large
amount to tributary flow, the accuracy of this method depends on
the matching of storm precipitation with average annual
precipitation patterns.
October through April. Most tributary flow during this period is
generated by groundwater baseflow; very little is a direct result
of precipitation or of snowpack melting. Consequently, annual
precipitation patterns are not used to weight relative basin
contributions. For this period we have continued to use linear
drainage area weighting (Method I).
RESULTS AND DISCUSSION
The water balance accounting model is largely a support program,
providing input flows to other component models. As such, it operates
on a specific scenario, generating an output flow time series for each
nodal location in the system.
To generate the postproject flow time series, H20BAL runs through
two cycles. A time series at each node is first determined based on the
natural input flows. Tributary contributions are determined in this
step. The next cycle reassigns postproject output flows to the dam node
and flows at the remaining mainstem nodes are re-adjusted.
Figure 5, longitudinal profiles of the pre-and postproj ect mean
June flow regimes, provides graphic representation of H20BAL output.
Figure 6 gives tabular comparison of the three apportionment methods for
the same preproject mean June flows.
Since filled flow records for the 32-year period of simulation
exist for the Talkeetna and Chulitna rivers, flow from these systems can
be treated as point source inputs to the mainstem basin. The Yentna
River, however, cannot be treated in this way, except when simulating
the period covered by the two-year gage record. When extending the
water accounting system downstream from Sunshine, Yentna River flow must
be apportioned as a fraction of the difference between Susitna Station
and Sunshine Station gage flows.
The decision to use the area weighting procedure in the lower
basin, regardless of the method used in the upper basin (i.e., upstream
from Sunshine Station), was based on the following considerations:
1. The large size of the Yentna Basin (6180 mi2 ) makes the task
of developing C coefficients for water-yield or precipitation
weighting formidable.
13
Figure 5. Mean pre-and postproject June flow profiles, Watana to Sunshine Station, using precipitation-weighting water balance method.
,.....__
C/J
1.1--'
C)
'-"
~ ,.....,
~
14
0
0
0
0
)'-.
.IJ
C)
Q) Cll
·n ~ 0 0
1-lr-1
p.,~
Q)
1-l p,
0
0
0
0
.0
0
0
0
0
l!")
.IJ u
Q)
'---
·n C/J
~ ~
P.,.-l
.IJ~
C/J
0 p,
0
0
0
0
-.::t
--
0
0
0
0
C'l")
\
\
0
0
0
0
N
-1
\
\
0
0
0
0
.--{
\
\
0
0
0'\
0
.--{
.--{
0
C'l")
.--{ Q)
0
l!")
.--{
0
)'-.
.--{
0
0'\
.--{
r-1 .,.;
;::;::
l-1
Q)
:>
•ri
~
Figure 6. Mean preproject June flows, Watana to Sunshine Station,
using three weighting methods.
--------Preproject Flows (cfs)--------
Location River Area Precip. Water-Yield
Name Mile Weighting Weighting Weighting
\._Tatana 184.4 23034 23034 23034
Tsusena 181.3 23999 24056 24081
Fog 176.0 24844 25266 24876
Devil Canyon 161.3 25688 25986 25675
Chinchee 154.6 26143 26314 26031
Portage 148.8 27044 27003 27056
Indian 138.6 27815 27815 27815
Mckenzie 116.8 28543 28655 28687
Whiskers 101.4 28787 28917 28952
Chulitna 98.6 52359 52526 52589
Talkeetna 97.2 63916 64064 64103
Trapper 91.2 64117 64280 64291
Sunshine 83.8 64555 64555 64555
2. The lack of gage data for the Yentna River with which to
calibrate makes any selection of a weighting scheme somewhat
arbitrary.
3. The confluence of the Yentna River is at the downstream end of
the Susitna Basin, far from the dam sites. Consequently, this
is the region least sensitive to differences in flow
apportionment methods.
Enhancement of the apportionment methodologies might be undertaken
in a number of ways. The relative precipitation weighting method could
be improved by using monthly or seasonal precipitation distribution
maps. Presently, however, these ~aps are not available. Kilday (1974)
developed mean monthly precipitation maps for the State, but they do not
have the resolution necessary to be used on the Susitna Basin.
Sub-basin water yield would be determined most directly using
Method III, relative water-yield weighting. Improvement of the present
water-yield map is possible as additional precipitation and streamflow
data become available. Continued enhancement could lead to monthly or
seasonal water-yield estimates.
15
STREAM NETWORK TEMPERATURE SIMULATION MODEL
INTRODUCTION
AEIDC selected the Stream Network Temperature Simulation model
(SNTEMP), developed by the U.S. Fish & Wildlife Service (Theurer et al.
1983), for use in the Susitna simulation model. SNTEMP predicts average
daily and daily minimum and maximum water temperatures at selected
points within a river network. The model requires meteorologic,
hydrologic, and stream geometry data to compute heat flux relationships
and to transport heat through the system.
Several features of SNTEMP make it particularly applicable for use
in the Susitna system.
1. SNTEMP contains a temperature regression technique which
allows use of incomplete or noncontinuous input temperature
data. Much of the Susitna water temperature data are point
measurements or incomplete records.
2. SNTEMP contains a calibration technique which provides the
ability to adjust low-confidence input parameters to obtain
minimum prediction error.
3. Daily average, maximum, and minimum water temperatures can be
predicted for periods ranging from as short as one day to as
long as one year (continuously variable in one-day incre-
ments). Thus, short yet critical river reaches could be
modeled in daily detail, but the full length of the system is
simulated with longer averaging periods.
For the Susitna system, SNTEMP has been configured to simulate mean
monthly temperatures at any location between the Watana dam site and the
Parks Highway bridge at Sunshine Station. The model utilizes either
historical or synthetic hydrologic and meteorologic data. In this
latter mode of simulation, referred to as "gaming," synthetic data are
used to approximate temperatures during the construction and postproject
phases of the proposed project.
17
DESCRIPTION OF THE STREAM TEMPERATURE MODEL
SNTEMF is a collection of several submodels:
1. a solar model which predicts solar radiation based on the
latitude of the stream basin, time of year, basin topographic
characteristics, and prevailing meteorologic conditions;
2. a meteorologic correction model accounting for changes in air
temperature, relative humidity, and atmospheric pressure with
elevation;
3. a heat flux model accounting for all significant heat sources
and sinks;
4. a heat transport model to move the water and its associated
heat content downstream;
5. a regression model for smoothing or completing observed water
temperature data; and
6. a flow mixing model for merging tributary flows and heat
content 1;.dth those of the mainstem.
A complete description for each of these components is provided in
the model description/ documentation available from the U.S. Fish and
Wildlife Service (USFWS) (Theurer et al. 1983). A brief description of
the heat transport model will be provided since it is this component,
more than any other, which determines the model's limitations. The heat
transport model used in SNTEMP is based on the following dynamic
temperature-steady flow equation (Theurer et al. 1983):
(A/Q) (dT/Clt:J + ClT/Clx = (qd/Q) (Td -T) + (Bl:H)/(Qpcp) (6)
[<--dynamic term-->[<------steady state equation---------->[
[<------dynamic temperature-steady flow equation-------->[
where:
A flow area, L2
Q = flow, L3 /t
T temperature, T
t time, t
X distance, L
qd distributed inflow, L2 /t
Td distributed inflow temperature, T
18
B = stream top width, L
LH =net heat flux, (E/L 2 )/t
P water density, M/L 3
c = specific heat of water, (E/M)/T p
and dimensions are:
M -mass
T -temperature
L -length
t -time
E -energy
The assumption of steady state (8T/8t = 0) can be used to reduce
the order of Equation (6) when 24-hour average temperature predictions
are sufficient, resulting in:
(7)
It is significant that this equation does not contain a stream
velocity term. SNTEMP does not require stream velocities for prediction
of average daily temperatures downstream from a known temperature.
Dynamic temperature predictions are possible if steady state is not
assumed. Equation (6) can also be solved by the method of
characteristics (Theurer et al. 1983) which results in a solution
identical in form to Equation (7). Dynamic temperature predictions
require Equation (7) to be solved along the characteristic line equation
as follows:
dx (Q/A) dt (8)
The factor Q/ A is stream velocity. Dynamic temperature predictions
require an estimate of stream velocity which SNTEMP computes using
Manning's equation. Closed form solutions of Equation (7) are obtained
by assuming that 1) the flow is uniform within a reach and 2) a second
order approximation of the heat flux is valid. This heat flux
approximation can be expressed mathematically:
19
~H (9)
where:
T equilibrium temperature, T e
K1 first order thermal exchange coefficient, [((E/1 2 )/t)/T]
K2 second order thermal exchange coefficient, [((E/L 2 )/t)/T2 ]
The equilibrium temperature is the theoretical temperature the
stream would approach if all heat transfer processes were held constant
with time. If the water reached equilibrium temperature, the rate of
heat input to the water would equal the rate of heat loss (~H = 0).
Equilibrium temperature and steady flow assumptions constrain the
methods used to average input data. The input hydrologic and
meteorologic conditions must be representative throughout the travel
time from the initial to final points of the model network. If the
travel time from the most upstream point to the downstream end of the
network becomes significant compared to the data averaging time, then
model prediction becomes less reliable. For example, assume that a
30-day meteorologic data averaging period has been selected and that it
takes 30 days for water to travel from point A to point B. Water
passing point B on the first day of this 30-day period left point A 30
days earlier. Therefore, the meteorologic conditions which determine
the daily average water temperature at point B on the first day are not
included in the time period averages. Only the last day's water column
can be considered to have been influenced by the 30-day average
meteorologic conditions.
This data averaging versus travel time dilemma can be overcome
either by 1) selecting averaging periods greater than the netw·ork travel
time or 2) dividing the network into serially connected subnetworks, or
reaches, and using moving average input conditions. The first technique
is the standard way of operating SNTEMP. If short-term average water
temperature predictions are necessary, the second technique can be
accomplished with SNTEMP by simulating an upstream reach with
appropriate average input data, and using this simulation's output as
input to the next downstream reach.
20
MODIFICATIONS
AEIDC modified SNTEMP to more accurately simulate conditions
specific to Alaska and the Susitna Basin, including techniques to
approximate the seasonal variation in canyon wall shading and winter air
temperature inversions which normally occur in the Susitna River basin.
The original design of SNTEMP assumed topographic shading to be
constant. Since solar altitude angles are so acute in Alaska, resulting
in extreme shading during the winter months, SNTEMP was modified to
accept a monthly topographic shading parameter.
SNTEMP originally featured a constant lapse rate to simulate air
temperature and humidity change at elevations other than those where
data were recorded. Radiosonde data from Fairbanks and Anchorage
indicated this approximation to be a. poor predictor of actual
conditions, especially in the colder months (U.S. National Weather
Service 1968, 1969, 1970, 1980; World Meteorological Organization 1981,
1982). AEIDC modified SNTEMP to accept monthly, nonconstant lapse
rates. Local monthly temperature lapse rates were determined by
regressing temperature on elevation using data recorded above Anchorage
and Fairbanks (1968 through 1970; 1980 through 1982) by U.S. National
Weather Service balloons. The temperature lapse rate curves for June,
July, August, and September are shown in Figure 7. Piece-wise linear
humidity lapse rate curves were also determined from the balloon data
and are presented in Figure 8.
In addition, we also adjusted the normal SNTEMP operating method to
accommodate the limited water temperature data available throughout the
study area. Typically, a built-in regression model provides missing
water temperature data and smooths the data but we had to bypass this
feature since it required more data than were available at any of the
water temperature collection sites. This will be discussed further in
the section entitled "Synthetic Temperatures."
STREAM NETWORK
The stream network as defined for SNTEMP is designed to allow easy
manipulation of flows and water temperatures at specific locations.
This network can be used for simulations with either or both Watana and
Devil Canyon reservoirs. Using expected water temperatures and outflows
21
r:rj JUNE f-'•
N 1ERW!\K LJi'lt !.11E t.UGUST ~ N t£l~T lN M[J[.qs TElWERATURE LAPS( RATE H -HEIGHT IN HETERS (1)
3330
-....) ,.. • 6' 2500 10
D oO .. r:r:i<:
Pl (D
f-'• H laJil ill 2030 ,., .. •• H rt "' cr' f-'·
Pl n
::J Pl I~ •" ~. 1500 •"' 01 D 0 ?;'I-' riJ
CJJ
Pl . . . ...... f-'· 1000 D 1.0 H li11111 •• ill
(J\
00 rt
500 D D D 0 (D
~ •• s 1--''I:;J
DOn£ {J I.O(D
ooP 0'\H
I.OPl ·10 -B -6 -4 -2 10 12 14 rt
·II -e .. ·4 -1 II ll 14 TEMPERATURE COEGREES Cl c ...... H
1Elf'ERAT\K o:= 0 1.0 (D
-....)
0 0..
f-'•
CJJ ...... rt
1.0 H
00 f-'•
0 cr' JULY SEPTEMBER c T£i4PERATURE LAPSE RATE T£i4PERATURE LAPSE RATE rt
HEIGHT IN METERS HEIGHT IN METERS ...... f-'•
1.0 0 3300 3000
00 ::J .......
250Z .. "" .. . 2530 ...... gr • .... a
1.0
00 n 23ZZ • "' .. 2ZZZ N ::r ... .. 0
H
Pl I500 D Clt!Do .. • 1530 ao .. q, '1>6' '"' (D
1020 1000 Pl •• ::J
0..
500 530 .. .. • •
"'" •o
-10 -B -6 -4 -2 10 12 14 -10 -8 -6 -4 -2 10 12 14
TEI<PERA TURE (IJEGREES Cl TEI<PERA TURE COEGREES Cl
r..j
JUNE AUGUST f-l·
QQ
~
l'i
ll'l;HT lN M(f[RS l{!QlT IN ME!fRS ro -313~0
00
Do~ 25ili D o "o "a 1-' <l
\.0 (I)
0'\ l'i
00 rt
D D 11ll!ll 1-'•
D D D D ('l
1-' Pl
\.0 t-'
0'\
]Sill! Do l5e0 \.0 l'i
(I)
t-'
1-' Pl
\.0 rt
llllli D I{J " 1-'•
0 <:
(I)
5e0 CD D C 1-' ::r'
1.0 ~ co
0 1-'•
D llJ 0.. _a P f-l•
1-' rt
u 1.1 1.2 ll 1.4 1!.5 1!.6 1.7 i.B l9 u u B.l 2.2 8.3 B.• 1!.5 0. 6 1!.7 1!.8 l9 I. 0 \.0~ co
R£LATII'f IU<ICJTY, ll:CIIW. R£LITIVE tu<IOITY. OCCINAL 1-' 0..
> f-l•
Cll
1-' rt
\.0 l'i
JULY SEPTEMBER 00 1-'•
N o'
~
f£1Ghl IN >i:T!lll I£11ID IN I£1£RS rt
1-'•
3llllol 0
::J ..
D 'h D Dill 'IJo D D 'a ~
(')'
1:1' a
.l'i
Pl
QQ
(I)
1500 q, II Pl
::J
0..
liOO lllOO DO DO Ctj
Pl
1-'•
l'i
dl tP D D 5111! D 0 oo D D D o'
Pl
N ::J
w D D 0 ~
oa OJ (I)
0 D -----'--1__,__ rnj_q_p,_ll_i_o
B.B II. I 1!.2 1!.3 a.• 0. 5 l6 B.7 1!.8 11.9 I.e u B. I B. 2 B. 3 8. 4 B. 5 B.6 ll.7 B.B 8.9 l.B
lifLATIV[ fUIIOITY, [][CIHAL RELITI'If: IIJIIOJTY. O[CIMI.L
from Watana reservoir (RM 184.4), the model predicts the water
temperatures at any specific location downstream to Sunshine Station (RM
83.8). This network (Figure 9) is easily modified for simulations with
flows and water temperatures at Devil Canyon. To obtain starting
temperatures at Watana for validation and calibration simulations, we
defined a 40-mi reach from the USGS Cantwell gage (RM 223.5) downstream
to Watana. River mile distances are based on interpolations of maps in
the Susitna River Mile Index (R&M 1981).
Tributaries between Watana and Sunshine Station were included in
the Susitna stream network. The flow and thermal contributions of
smaller tributaries not explicitly included were estimated, and either
incorporated into a nearby tributary or were linearly distributed to the
neighboring mainstem reach. A more detailed description of these
hydrologic approximations appears in the section "Water Balance
Accounting for the Susitna Basin."
STREAM STRUCTURE
Segmenting the system network into reaches with similar physical
characteristics (Figure 10) provided the physical model of the system.
Reach selection was based predominately on orientation and local topo-
graphy with consideration of significant slope change, width change, and
elevation drop.
Topographic Shading
Topographic shading may significantly affect Susitna River tempera-
tures especially in the winter months. The orientation of the reach and
the elevation of surrounding canyon walls limits the amount of sunshine
the stream surface receives. As previously mentioned, SNTEMP was
modified to accept changes in stream shading for each month. The
variable which defines the amount of shading is the sunrise/ sunset
altitude angle (a). We chose a representative midchannel point in the s
reach to compute this angle. A compass rose was centered on this point
and terrain elevation versus distance transects collected from 30
degrees east of North to due South to 30 degrees west of North at 15
degree increments. Maximum terrain altitude angles were determined from
24
Figure 9. Stream network from Watana to Sunshine.
TSUSENA
208.6
DEVIL
174.3
PORTAGE
175. 5
INDIAN
159. 6
WHISKERS
113. 2
CHULITNA
116. 5
TRAPPER
115. 4
WATANA DAM SITE
184. 4
181. 3
1 76. l2l
161. 3
148. 8
138.6
101. 4
98.6
91. 2
SUNSHINE
RM 83.8
H
FOG
195. 9
CHINCHEE
163.3
TALKEETNA
112l2.12l
H refers to tributary head-
waters as defined in the
stream network
J refers to tributary
junction with the mainstem
Numbers refer to River Mile
as interpolated from R&M
River Mile Index (1981).
25
RM
I Rtvermtle
111.0
! :
1
------G) Reach
SUS ITNA -:-::--:---0-E_F IN IT I 0 N s I
Scale In; 8 miles -
each of these transects and then transferred to solar altitude versus
bearing angle plots (Siefert 1981). We computed the average
sunrise/sunset altitude angles for each month from these plots
(Appendix A).
Stream Widths
The quantity of radiative energy entering or leaving the stream is
a function of the stream surface area. An estimate of the stream width
is necessary for surface area determination. Mainstem wetted widths
used in SNTEMP from the Talkeetna River confluence to Watana were
determined from the R&M cross sections and HEC-2 simulations (R&M
1982d). The stage-discharge relationships developed by ADF&G (1983)
were not available when our width analysis was being performed.
However, since the stage discrepancies noted between the R&M simulations
and ADF&G observations would not result in significant width
differences, we do not propose to modify the width functions at this
time.
Water surface widths simulated by R&M were measured from the cross
section diagrams (R&M 1982d) and plotted as a function of flow (Appendix
B). We calculated width/flow functions from these plots.
Other methods were used to estimate top width for other mainstem
reaches and tributaries. USGS (1980, 1981) observations at Cantwell,
Chulitna, and Talkeetna provided some stream width and flow data. Width
data at the Chulitna and Talkeetna gages were available for several
flows. Several width measurements within a narrow range of flows
provided a constant width estimate for the Susitna River between
Cantwell gage and the Watana dam site. The width of the reach below the
Chulitna junction to Sunshine Station was determined from transects
collected by R&M (Coffin 1983). This width was also assumed constant
with flow. Field personnel estimated widths of the tributaries
(Sauntner 1983; Schoch 1983; Quane 1983) which were assumed constant
with flow.
Figure 11 presents width/flow functions in tabular form with
graphic presentations in Appendix B. The plots present data points
connected by line segments and the computed function.
27
Figure 11. Tabular values of width function parameters.
Stream Reach It Start End a b
(mile) (mile)
Susitna 1 184.5 179.5 98.26 0.1577
Susitna 2 179.5 175.5 105.40 0.1708
Susitna 3 175.5 166.0 98.13 0.1820
Susitna 4 166.0 163.0 189.96 0. 077 4
Susitna 5 163.0 146.5 144.88 0.1005
Susitna 6 146.5 142.5 98.15 0.1845
Susitna 7 142.5 124.0 13.16 0.4078
Susitna 8 124.0 115 .o 33.95 0.3117
Susitna 9 115 .o 99.5 29.77 0.3390
Susitna 10 99.5 83.8 1256
Tsusena 1 208.6 181.3 80
Fog 1 195.9 176.0 so
Devil 1 174.3 161.3 35
Chinchee* 1 163.3 154.6 25
Portage 1 175.5 148.8 60
Indian 1 159.6 138.6 50
Whiskers 1 113.2 101.4 20
Chulitna 1 116.5 98.6 60.70 0.2086
Talkeetna 1 102.0 97.2 97.92 0.1761
Trapper 1 115.4 91.2 18
Values for "a" and "b" in the function width (feet) b If "b" = a . flow (cfs) •
is undefined, "a" represents a constant width (feet).
*A synthetic stream representing the combined Chinook and Chechako tributaries.
28
Hydraulic Retardance
SNTEMP does not require stream velocity estimates to predict
average daily downstream water temperatures (see "Description of the
Stream Temperature Model"). On the other hand, daily minimum and
maximum temperature predictions do require estimates of stream
velocities. If daily maximum and minimum temperature estimates are
desired later, it will become necessary to obtain the Manning's n values
to compute stream velocities.
Tributary Assumptions
Except for the Chulitna and Talkeetna rivers, all Susitna
tributaries simulated by SNTEMP are essentially self-starting.
Simulation of these tributaries starts from their estimated headwaters
where a constant headwater temperature of 0 C is assumed. Since the
headwater flow is assumed to be zero, this seasonally constant initial
water temperature is not critical (the heat content of zero mass would
be zero, exclusive of the temperature assigned). Flow is added to these
tributaries based on the flow balance schemes discussed in the section
"Water Balance Accounting for the Susitna Basin." Predicted tributary
temperatures are highly sensitive to the temperature assumed for distri-
buted flow. Techniques for estimating these temperatures will be
discussed in the section "Temperatures of Distributed Flow." Tributary
widths were based on field estimates and lengths were measured from
topographic maps with an opisometer. Each tributary in the model is
assumed to be a single stream. For branched tributaries we estimated a
sub-basin area-weighted average length. Tributary reaches were defined
based on 300 m elevation drops.
HYDROLOGY
Flows
As described in the section "l-Jater Balance Accounting for the
Susitna Basin," we investigated three types of flow balancing techniques
for supplying flow estimates to the temperature model. These techniques
are used both with historical flows and for gaming with reservoir
releases.
29
Stream Temperatures
Observed Temperatures. SNTEMP uses observed water temperature data of
two types--initial water temperatures necessary for starting the model
and validation/calibration water temperatures. Only three initial
temperatures are required for the Watana (or Cantwell) to Sunshine
Station simulations. These are Susitna River at Watana (or Cantwell),
Chulitna River at the USGS Gage, and Talkeetna River at the USGS gage.
The remaining observed water temperatures are essential in determining
how well the mainstream and tributary temperatures are being simulated
and in serving as a calibration target.
Most of the validation/calibration temperature data for this study
are being collected by ADF&G (1981, 1983); USGS (USGS 1980, 1981;
Bigelow 1983) collected the three initial water temperatures.
Unfortunately, most of these initial temperatures are unusable as a
result of incomplete records or discrete sampling. Usable data are
defined as those data which are complete for the month or, if not
complete, symmetric around the middle of the month. Data which cluster
evenly around the middle of the month should result in an unbiased
measure of the monthly mean. Figure 12 presents the available data, and
Figure 13 presents usable data collected for the June to September
periods of 1980 through 1982. Data collected by USGS at Gold Creek were
not used in this study since it had been observed that the temperature
recorder was in the plume of the Gold Creek tributary (Trihey 1983) and
thus not representative of mainstem flow. USGS recently relocated this
temperature recorder, and future data provided by USGS and ADF&G should
allow adjustment of the historical USGS data to be representative of the
mainstream temperatures.
Synthetic Temperatures. USGS Cantwell gage on the Susitna River (RM
223. 7). Stream temperature data were recorded at the Cantwell gage
during the 1980 and 1982 June through September periods. To verify
downstream temperature predictions with stream temperatures observed by
ADF&G (1981) and R&M (1982b), we estimated water temperatures at
Cantwell for 1981. SNTEMP incorporates a regression technique for data
filling, but, as discussed previously, more data than are available are
necessary for this technique to produce physically reasonable results.
30
Figure 12. Monthly stream temperatures, available data June to
Sept. 1980, 1981, 1982.
Mainstem/Tributary Number of Days
River Mile River name I description 1980 1981 1982
J J A s J J A s J J A s
10.1/0.5 Alexander Cr. 25 31 31 30
10.1 Susitna above Alexander Cr. 25 31 31 1
25.8 Susitna R., Su Station 30 31 31 30 10
28.0/2.0 Yentna R. 26 31 31 14
28.0/4.0 Yentna R. 23 31 31 27
29.5 Susitna R. above Yentna R. 20 31 31 30
32.3 Susitna R. above Yentna R. 25 31 31 12
40.6/1.2 Deshka R. 21 31 31 30
49.8/4.9 **Deception Cr. near Willow 5 8 8
49.8/11.6 **Willow Cr. near Willow 5 18 22
50.5/1.0 Little Willow Cr. 7 31 31 30
50.5 Susitna R. above Little Willow Cr. 7 31 31 24
61.2 Susitna R. above Kashwitna R. 2 27
77.2/0.0 Montana Creek 19 24 1
77.5 Susitna R. above Montana Cr. 19 3 2 30
83.8 Susitna R., east shore-Parks Hwy. 20 14 30
83.9 Sus it na R., west shore-Parks H wy. 23 9 10 30
97.0 Susitna R.-LRX 1 17
97.2/5.0 **Talkeetna R. near Talkeetna
97.0/1.0 Talkeetna R. 10 31 31 30
97.2/1.5 Talkeetna R. 17 1 31 30
98.5/18.0 **Chulitna R. near Talkeetna 27 30 3 20
98.6/0.5 Chulitna R. 11 17 20
98.6/0.6 Chulitna R. 17 10 25
103.0 Susitna R.-TKA fishwheel 11 10 19 22 7 28 31 25
113.0 Susitna R.-LRX 18 25 31 30
120.7 Susitna R.-Curry 25 31 30
126.0 Susitna R.-Siough SA 4 31 30
126.1 Susitna R.-LRX 29 22 31 30
129.2 Susitna R.-Slough 9 4 31 24
130.8 Susitna R.-LRX 35 23 4 17
131.3 Susitna R. above 4th of July Cr. 15 31 30 26
136.5 ••susitna R. near Gold Cr. 30 31 31 30 8 25 29 12 30
136.8/0.0 Gold Creek 7 3
138.6/1.0 Indian R. 23 31 4 28
138.6/0.1 Indian R. 10 25 14
138.7 Susitna R. above Indian R. 11 29 16
140.0 Susitna R.-Siough 19 5 13
140.1 Susitna R.-LRX 53 23
142.0 Susitna R.-Siough 21 4 29 4 31 30
148.8 Susitna R. above Portage Cr. 13 31 29
148.8/0.1 Portage Cr. 13 26 28
181.3/0.0 Tsusena Cr. 12 7 31 30
184.4 *Susitna R. at Watana dam site 30 -31 30
194.1/0.0 Watana Cr. 11 31 15 16
206.8/0.0 Kosina Cr 4 31 17 12
223.7 ··susitna R. near Cantwell -- -
27 31 31 22
231.3/0.0 Goose Creek 31 31 30
233.4/0.0 Oshetna Creek 31 31 30
*R&M gages
··usGs gages
All others are ADF&G gages.
31
Figure 13. Monthly stream temperatures, usable data June to
Sept. 1980, 1981, 1982.
Mainstem/Tributary Number of· Days
River Mile River name I description 1980 1981 1982
J J A s J J A s J J A s
10.1/0.5 Alexander Cr. 18 31 31 26
10.1 Susitna above Alexander Cr. 18 31 27
25.8 Susitna R., Su Station 30 31 31 30
28.0/2.0 Yentna A. 20 31 31
28.0/4.0 Yentna A. 14 31 31 24
29.5 Susitna A. above Yentna A. 10 31 31 30
32.3 Susitna A. above Yentna A. 18 31 29 6
40.6/1.2 Deshka A. 10 31 31 30
49.8/4.9 ··Deception Cr. near Willow 2
49.8/11.6 • *Willow Cr. near Willow 13 4
50.5/1.0 Little Willow Cr. 31 31 28
50.5 Susitna A. above Little Willow Cr. 31 31 10
61.2 Susitna A. above Kashwitna A. 22
77.210.0 Montana Creek 6 17
77.5 Susitna R. above Montana Cr. 8 30
83.8 Susitna R., east shore-Parks Hwy. 8 30
83.9 Susitna A., west shore-Parks Hwy. 14 30
97.0 Susitna A.-LAX 1 14
97.215.0 **Talkeetna A. near Talkeetna
97.0/1.0 Talkeetna A. -31 31 30
97.2/1.5 Talkeetna A. 14 31 30
98.5/18.0 **Chulitna A. near Talkeetna ---24 30 10
98.6/0.5 Chulitna R. 3 12
98.6/0.6 Chulitna A. 14 18
103.0 Susitna A.-TKA fish wheel 17 13 21 31 16
113.0 Susitna R.-LAX 18 17 31 30
120.7 Susitna R.-Curry 17 31 30
126.0 Susitna R.-Siough SA 29 30
126.1 Susitna A.-LAX 29 13 31 30
129.2 Susitna A.-Slough 9 31 20
130.8 Susitna A.-LAX 35 6
131.3 Susitna A. above 4th of July Cr. 31 26 22
136.5 ••susitna A. above Gold Cr. 30 31 31 30 24 24 30
136.8/0.0 Gold Creek
138.6/1.0 Indian A. 16 31
138.6/0.1 Indian A. 17 8
138.7 Susitna A. above Indian A. 21 10
140.0 Susitna A.-Slough 19
140.1 Susitna R.-LRX 53 23
142.0 Susitna A.-Slough 21 28 31 30
148.8 Susitna A. above Portage Cr. 31 28
148.810.1 Portage Cr. 15 25
181.310.0 Tsusena Cr. 31 31 30
184.4 *Susitna A. at Watana dam site 30 -31 30
194.110.0 Watana Cr. 31 6
206.8/0.0 Kosina Cr 31 3
223.7 ··susitna R. near Cantwell 24 31 31 15
231.310.0 Goose Creek 31 31 30
233.410.0 Oshetna Creek 15 31 24
*R&M gages
·*USGS gages
All others are ADF&G gages.
1?
To fill this missing year, we simplified, but retained, the logic of the
SNTEMP regression technique.
SNTEMP uses what may be termed a "physical process" regression
model for data filling and smoothing. The regression model is based on
a simplified version of the heat transport model used to predict
downstream water temperatures. These models employ an equilibrium
stream temperature assumption where the calculated equilibrium
temperature
approaching.
(T ) e represents the value the stream is asymptotically
The standard regression model of SNTEMP uses the
calculated T and the rate of approach to T as independent variables. e e
For the Susitna River application, this model was simplified to use only
the equilibrium temperature (Figure 14).
USGS gage data collected on the Chulitna and Talkeetna rivers.
Only three usable water temperatures were available for the Talkeetna
and Chulitna rivers during the June to September periods. These
temperatures were recorded on the Chulitna River during June, July, and
September of 1982 (Bigelow 1983). Because of the limited data at these
stations, regressions similar to those used for the Cantwell gage were
of little value (Figures 15 and 16). However, the values predicted by
these regressions were used to fill in the missing data and to smooth
those observed data points with only one observation per month. Where
available, ADF&G temperature data were used to adjust the temperatures
at the gages so that simulated temperatures matched the observed data at
the ADF&G sites. Figures 15 and 16 list the values assumed by the
model, but the reader should note the low confidence associated with
these values.
Temperatures of Distributed Flow. Flow accretions from groundwater
or surface inflow are included in the network as continuous additions to
the stream flow, referred to as distributed flows. This is the primary
mechanism for simulating Susitna tributary flows. Water temperature
predictions for smaller tributaries depend on the water temperatures
assigned to tributary distributed flows. Thus, the accuracy of
temperatures assigned to distributed flow is critical to the simulation.
Contribution from surface or groundwater flows have not been quantified
33
Figure 14. Temperature regression for Susitna River at Cantwell gage.
Month Regression 95 Percent
Regression Prediction Confidence Intervals
(C) (C)
June 81 8.82 ± 1.53
July 81 8.96 ± 1.54
August 81 8.20 ± 1.51
September 81 5.30 ± 1. 74
15
~ p:::
;:J
H 10 22
>4
~ w
H
A
j;:il 5 > p:::
~
'JJ
p:)
0
0
lZl 5 10 15 20 25
EQUILIBRIUM TENPERATURE
34
Figure 15. Temperature regression for Chulitna River at USGS gage.
Regression
95 Percent
Regression Confidence Value
Observed (C)/ Prediction Intervals Used
Month Sample Size (C) (C) (C)
June 81 7.4/1 6.68 ± 2.85 6.68
July 81 -/0 6.90 ± 2.89 7.10*
August 81 7.2/1 6.64 ± 2.84 6.64
September 81 -/0 4.95 ± 3.83 5.25*
June 82 7.3/24 6.53 ± 2.84 5.45*
July 82 5.7/31 7.01 ± 2.92 5.7
August 82 -/0 7.01 ± 2.92 7.01
September 82 4.6/10 5.22 ± 3.56 4.6
*Temperature at gage was adjusted so downstream simulation matched data
collected by ADF&G.
15
0
!Zl 5 10 15 20 25
EQUILIBRIUH TEMPERATURE
35
Figure 16. Temperature regression for Talkeetna River at USGS gage.
Regression
95 Percent
Regression Confidence
Observed (C)/ :Prediction Intervals
Month Sample Size (C) (C)
June 81 10.2/1 8.37 ± 5.85
July 81 9.0/1 9.28 ± 5.96
August 81 9.7/1 8.14 ± 5.84
September 81 1. 5/1 2.86 ± 7.38
June 82 -/0 8.54 ± 5.86
July 82 -/0 9.67 ± 6.04
August 82 -/0 9.52 ± 6.01
September 82 -/0 2.74 ± 7.45
*Temperature at gage was adjusted so downstream simulation matched data
collected by ADF&G.
36
15
0
0 5
@
10 15 20 25
EQUILIBR.IUN TEMPERATURE
Value
Used
(C)
8.37
8.60*
8.40*
5.70*
7.00*
9.67
9.20*
5.50*
in Susitna tributaries; therefore, they must be estimated. Presently,
two techniques can be used to estimate these temperatures. The first is
to assume groundwater inflow at a constant temperature for all time
periods and all locations. G. Nelson (1983) of the USGS suggested a
value of 3 C as representative of a wide range of conditions encountered
by that organization in adjacent ·drainages. This assumption does not
allow for 1) seasonal ground temperature variation, 2) ground tempera-
ture variation with site elevation, or 3) the possibility of surface
runoff.
Rather than assuming a constant temperature for distributed flows,
an alternative technique is to vary temperature by location and depth.
AEIDC modified the ground temperature function presented by Williams and
Gold (1976):
where:
T (x,t) = T +(~T /2) cos [(2~t/t ) -xl~tat ] exp(-xlrr/at ) (10) g g g 0 0 0
T average annual ground surface temperature (C) g
~T = annual range of ground surface temperature g
variation (C)
t time from occurrence of peak temperature (days)
t time for one cycle of temperature variation (365 days)
0
x depth (m)
a thermal diffusivity (m 2 /day) = thermal
conductivity/volumetric heat capacity
This formula can be used to predict ground temperatures at variable
depths and times if the average annual ground surface temperature (T ) g
and annual range of ground surface temperature variation (~T ) are g
known. The annual range of ground temperature can be assumed to be the
same as the annual range of air temperature variation (Williams and Gold
1976) which is 28.2 C at Talkeetna. Data presented in Williams and Gold
(1976) indicates that the average annual ground temperature is
approximately 1 to 7 C warmer than the average annual air temperature in
regions with persistent snow cover. If, for notational purposes, we
37
designate this 1 to 7 C offset by Toff and define A
the formula becomes:
2Tr/t , B
0
ITI/at ,
0
T (x,t) = T . + Toff + 14.1 cos (At-Bx) exp(-Bx) g al.r (11)
Air and ground temperature data collected at Gulkana, Alaska
(Aitken 1964b) and Big Delta, Alaska (Aitken 1964a) suggest that this
offset temperature is in the range of 4. 3 to 4. 9 C. For purposes of
further discus'sion in this paper a value of 4.6 C will be assumed,
although in the SNTEMP implementation of this ground temperature model
Toff will be used as a calibration variable.
The mean annual air temperature of an arbitrary location at
elevation Z can be computed from the mean annual air temperature at
Talkeetna (0.3 C) using the lapse rate equations discussed in the
modifications section:
where:
T -y(Z-Z ) 0 0
T2 air temperature at elevation Z (C)
T0 observed air temperature at elevation z0 (C)
z0 elevation of site where air temperature ,is known
(Z0 = 105 m for Talkeetna)
(12)
Z elevation of site where air temperature is desired (m)
y air temperature lapse rate (C/m)
By substituting the air temperature lapse rate expression for air
temperature at elevation Z, the ground temperature formula can be
rewritten as:
T (x,t,Z) = 4.9 -y(Z-105) + 14.1 cos(At-Bx) exp(-Bx) (13) g
If a value is assumed for the thermal diffusivity, the only
undefined variable for any location and time period is the depth of the
ground temperature. There are two depths of interest which correspond
38
to two separate forms of heat flux--conduction to and from the streambed
and mass transfer of heat (distributed flow). Streambed conduction is a
function of the depth at which the ground temperature variation is
essentially zero for the simulation time period. Given an estimate of
a., a depth can be computed where daily temperature fluctuations are
essentially zero. Williams and Gold (1976) give an a.-value for wet sand
of 0.01 cm 2 /sec. This value is also used to represent the thermal
diffusivity of sand, gravel, cobbles, and boulders in the Susitna slough
hydrogeology study (Acres 1983b). Using this value, daily temperature
fluctuations penetrate to a depth of approximately 0.8 m. Substituting
0.01 cm 2 /sec for a. and 0.8 m for depth, the above formula reduces to:
T (t,Z) = 4.9 -y(Z-105) + 10.3 cos(At-0.316) g (14)
The distributed flow heat flux is a function of the average depth
from which the water flows. Rather than assume a value, this depth has
been retained as a variable for calibration purposes.
This ground temperatur~ model must be considered provisional as the
assumptions made cannot be tested or validated without further data
collection. Temperature at depth data at several locations within the
Susitna Basin would be required for validation of this model and
improving estimates
currently provides
of assumed values. AEIDC believes
the best available approximation of
this model
the physical
conditions existing in the Susitna Basin and will be applied without
validation until better estimates of existing conditions are obtained.
METEOROLOGY
Selection of Meteorologic Data
The SNTEMP model is designed for climatic data input from only one
representative meteorologic data station per stream network. The only
long-term meteorologic data station within the Susitna Basin is the U.S.
National Weather Service station located in Talkeetna. This station has
summarized monthly data (air temperature, wind speed, relative humidity,
and percent cloud cover) --the data required by SNTEMP--for the period
1968 to 1982. In addition, unreduced data are available from 1950 to
1968 on computer tape from the National Climatic Data Center. This
39
period of record allows stream temperature simulations under extreme and
normal meteorologic conditions once these data are adjusted to better
represent Susitna Basin conditions. We used meteorologic data collected
specifically for the Susitna study (R&M 1980, 1982a, 1982b, 1982c,
1982e, 1982f) to validate this meteorologic data adjustment and SNTEMP
solar model predictions.
Ground Reflectivity and Atmospheric Dust
The stream temperature model predicts solar radiation based on site
latitude, period of the year, cloud cover, ground reflectivity, and
atmospheric dust. AEIDC determined monthly ground reflectivity values
for the Susitna Basin using the percent area groundcover vegetation
types presented in McKendrick (1982) and Bredthauer and Drage (1982).
The remaining component necessary to predict solar radiation is an
estimate of atmospheric dust. Dust was estimated by calibrating monthly
average predicted solar radiation to observed values using the published
solar radiation and percent possible sunshine data collected at the
Palmer Agricultural Experiment Station (Matanuska Station as recorded in
Wise 1979). Figure 17 presents these coefficient values.
Meteorologic Predictions
Conditions observed at Talkeetna are not necessarily representative
of the entire basin. SNTEMP adjusts most of the recorded variables to
better represent the local conditions within the basin. For example,
the predicted solar radiation considers local topographic shading. The
following discussion compares Susitna Basin meteorologic predictions
with data collected by R&H.
As was previously discussed, SNTEHP has been modified to accept
monthly air temperature/elevation and humidity/elevation functions. The
air temperatures and humidities predicted by these equations using
observed data are compared to the data collected by R&H (Figures 18 and
19). From these plots it appears that the humidity lapse model is a
poor predictor of basinwide conditions; however, we retained it in
SNTEMP for three reasons: 1) Talkeetna humidity data are based on
wet/dry bulb measurements which are inherently more accurate than
ceramic plate recorders (Wise 1983); 2) balloon-carried radiosondes are
40
OUST
COEFFICIENT
GROUND
REFLECTIVITY
1.0
0.5
COEFFICIENT VALUE
-
-
r-··-
-.r--·~ I / ·~ I 7' "\ .
\ I ~\_ --~ ----1 "\ ·~·-----·--·--: \ --.--· \ /
./
I I I I I t I I I I
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
MONTHS
Figure 18. Susitna Basin observed air temperatures vs. temperatures
predicted from Talkeetna data.
JUNE 1980. 1981, 1982
Figure 18. (Continued)
AUGUST 1980, 1981, 1982
PREDl CTED TEJ.!?, OEG C
SHEPMr.N 14
0 0
12
DEVIL Ci.NYON /J.
/J.
~~ +
/J. •
•
WW.NA 8
6
KDS!NA
4 •
2
2
~ 2 6 a 12 12 14
OBSERVED TEMP, OEG C
SEPTEMBER 198~. !981. 1982
PREDICTED TEIW. DEG C
SHERMAN 14
0
12
CEVIL CANYON
IJ. I~
~~iAN.~
+
6
KCSINA
4 •
2
a
• IJ 2 6 8 13 12 14
OBSERVED TEl!?. CEG C
43
Figure 19. Susitna observed humidities vs. humidities predicted
from Talkeetna data.
JUI;E 199~. 1991. 1992
Figure 19. (Continued)
WGUST 1983. 198i. !932
PREDIC1Ell RH. DECIW.L
1.11
SHERJI.AN
0.9
0
!!.8 0
DEV 1 L C.WYON A A t
D. 7
t..
ll.6
WWiiA
~.5
+
ll.4
KDSI/l/,
1!.3
ll.2
~.I
0.0
0.0 ll. I 3.2 a3 ~-( ll.S O.fi ll.7 ll.B ~. 9 i.C
OBSERVE.lJ RH. DEW.,'.L
SEPTEHBER 198~. !991. !9S2
PREDICTED RH. OECI~JL
u
SHERMI.N
ll.9
0 0
2.8 • A
DEVIL CliftON
2. 7 •
+
A
ll. 6
~ATWA
3. 5
a. 4
KGSit:A
2. 3
•
~-2
~-1
H
J. z ~-1 z. 2 3.3 ~-4 a.s 0. 6 e. 1 a.e ~-9 u
~8SER'I:D PP.. :::c:~~'-45
calibrated at the time of release and resultant data are the means of
twice-daily observations; and 3) erratic behavior (e.g., daily 0 to 100
percent oscillations) was noted in several of the R&M humidity
recordings.
The wind speeds at Talkeetna are not currently adjusted in any way
to better represent winds within other parts of the Susitna River basin.
Wind speeds recorded at Talkeetna were compared to wind speeds recorded
by R&M at various locations within the basin (Figure 20). It would be
relatively simple to incorporate a linear adjustment equation to
translocate observed Talkeetna wind speed data to locations which would
be better represented by the observed R&M data. However, the wind speed
data collected by R&M does not necessarily represent the wind speeds
which occurred directly above the water surface and are responsible for
the rates of convective and evaporative heat flux. Since it appears to
be impractical to collect wind speed data within the canyons below the
existing meteorological data sites (Bredthauer 1983), the wind speed
data collected at Talkeetna will be used as representative of average
bas in winds.
Figure 21 compares observed solar radiations to predicted solar
radiation. The simulated data are a reasonable approximation of the
field measurements.
VALIDATION
The purpose of model validation is to locate systematic prediction
errors. Systematic errors result when observed or assumed data for a
particular study do not represent actual conditions. Since the stream
temperature model has been verified with previous applications (Theurer
and Voos 1982; Theurer et al. 1983) and, since some adjustments have
been made to SNTEMP to account for conditions particular to the Susitna
application, it is assumed that any remaining systematic errors are the
result of nonrepresentative input data.
An initial validation run of the Susitna-modified SNTEMP
demonstrated a tendency to underpredict the upper tributary temperatures
(Figure 22). Since most of the data defining these tributaries are
assumed or estimated values, much uncertainty exists in the definition
of each tributary. Several poorly defined variables which might be
46
Figure 20. Average monthly wind speeds (M/S), 1980, 1981, 1982.
WATANA
DEVIL CANYON
TALKEETNA
KOSIN A
5
WATANA
4
DEVIL CANYON
3
TALKEETNA
-·-·· 2
KOSINA
·--1
WATANA
DEVIL CANYON
TALKEETNA
SHERMAN
KOSINA
AVE. WIND SPEED
5
ok---L-~-~--~-~-~-~--~--~--~~
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
AVE. WIND SPEED ~:...:...=~:.:...=.:=--=-=-=-=:.:::...._--------------------"j
"~I /~~I>"! "=·-·" . . i
."" I ! ·'-..---.............. ! ..)
JAN FEB }~ APR MAY JUN JUL AUG SEP OCT NOV DEC
AVE. WIND SPEED 5 r-=-=:..:..=.:_:..:.=c:..::__.==..._ _____ ------
/ 1 .
' _/··----... "' /
'-/-. ~-------------
oL-~--~--~--~-~--~--~--~~--~~
JAN FEB MAR APR 1-IAY JUN JUL AUG SEP OCT NOV DEC
0
00
0"1
.-l
47
~
00
~
1-'•
()q
R
PREDICTED SOLAR (J/M2/SEC) ro
253 N
f-' .
WATANA t-c:1
~ ro
0 0.
0 1-'·
(')
rt ro p.
200 <
DEVIL CANYON (f.) .
0
0"'
1::.
Cll ro
~ < ro
0.
(f.)
SHERMAN 0
1--' 153 P>
~
+ '0
1-1 ro
0.
1-'•
(')
rt
1-'•
0
::1
(f.)
~
133 f-'
\.0
00
0
~
f-'
\.0
00
f-' ..
f-'
\.0
50 00
N .
50 100 150 200 250
OBSERVED SOLAR (J/M2/SEC)
PREDICTED TEMP (C)
12
1~
8
6
4 D D
D
2
2 4
D
D
D
6 8
OBSERVED TEMP (C)
D
D
10 12
N
N
adjusted to improve model predictions are 1) stream flow, 2) initial
stream temperature, 3) stream length, 4) stream width, and
5) distributed flow temperatures. An effort has been made to adjust
other variables to better represent prevailing conditions (e.g., air
temperature, relative humidity, and topographic shading).
Of the five poorly defined variables, most improvement could be
gained from focusing on temperatures of distributed flows.
determination was based on the following logic.
This
1. Without the benefit of continuous tributary flow gaging,
present
enhanced.
stream flow estimates cannot be substantially
2. With the subsequent necessary assumption of zero flow at the
tributary headwaters, initi;:l tributary temperatures have no
influence on the predictions.
3. Tributary lengths were measured from maps.
4. Stream widths are based on field estimates and initial tests
with SNTEMP demonstrated that this variable was not sensitive
enough to remove the existing predictive bias.
Rather than arbitrarily modifying the constant 3 C estimate of
groundwater temperature, the ground temperature model previously
described was employed to generate physical process-based temperature
estimates. This model introduced three variables which must be
estimated--the average annual air/ground temperature offset (Toff),
thermal diffusivity (a.), and depth of inflow (x). AEIDC is currently
seeking techniques and data for estimating values of these variables.
Until solid estimates can be obtained, these variables will be adjusted
to calibrate to observed water temperature data.
While the tributary temperatures have a relatively small influence
on the natural rnainstern temperatures, this influence could increase in
importance during construction or operation of the dams. Future
temperature simulations will provide an indication of this effect and
perhaps suggest alternate modeling techniques or point out the need for
more tributary temperature data.
so
CALIBRATION
Tributary temperature predictions were improved by adjusting the
three groundwater temperature parameters (Figure 23). The resulting
values were: Toff = 1.0 C and a= 0.01 cm2 /sec for the entire basin,
Z = 0. 4 m for Kosina Creek, Z = 0. 7 m for Watana Creek, and Z = 2. 0 m
for the mainstem and remaining tributaries.
necessary to validate these values.
Further analysis is
The goodness of fit was determined by using the following
statistics:
where:
0.
1
T.
1
fj, = l: o./n
1
s
0.
1
T.
1
n
s
difference between ith predicted and observed
temperatures, C
ith published temperature, C
ith temperature predicted by SNTEMP, C
mean difference, C
number of observed temperatures
standard error estimate, C
(15)
(16)
(17)
These statistics can be combined with Z values to define prediction
confidence intervals. For example, 90 percent of the predicted values
fall within /j,±(1.645)S of the observed values. Postcalibration
statistics for the tributaries indicate that predicted values are on the
average 0.28 C (fj,) higher than the published values, and 90 percent of
the predicted values can be expected to fall between 2.10 above and
-1.54 C below the published water temperature (S = 1.11 C, n = 12). The
model fit could be improved with additional adjustment of Toff' Z, and
a. However, it was decided that additional calibration be postponed
51
52
Figure 23. Tributary temperatures; postcalibration, including distributed flow temperature model.
g
~ w
1--
c::::l w
1--w --til
8:
('.J lSI co (C "'¢
('.J
IS!
co
0 ........
0...
:::E
LL..I
I-
c::l
LL..I > 0:::
LL..I
(J) cc
0
(C
"'¢
('.J
('.J
until research is completed to define reasonable physical limits of
these parameters.
Once the tributary predictions had been improved, the entire
mainstem/tributary system essentially was calibrated, and no additional
parameter adjustments were attempted. Statistics for the mainstem are
8 = -0.05 C, S = 0.90 C, n = 28. Figure 24 presents these statistics as
computed for each month.
Figure 24. Temperature model calibration statistics for tributary
predictions.
n 8 (C) S (C)
June 1981 2 -2.08 1.39
July 1981 1 0.41
August 1981 6 0.59 0.42
September 1981 6 -0.002 0.74
June 1982 2 0.08 0.44
July 1982 3 -0.89 0.28
August 1982 4 0.19 0.80
September 1982 4 0.14 0.20
Average 28 -0.05 0.90
The statistics for June 1981 indicate a poor fit. This is
understandable since the three required initial water temperatures
(Cantwell, Chulitna, and Talkeetna) were synthesized with linear
regression models. This is the only month which had all three initial
temperatures synthesized. A more reasonable estimate of the simulation
performance for the mainstem is obtained by eliminating this month from
the computations: 8 = 0.10 C, S = 0.66 C, n = 26.
percent confidence interval is 0.10 ± 1.09 C.
The corresponding 90
Appendix C provides
53
longitudinal temperature predictions for the 1981 and 1982, June through
September periods.
RESULTS AND DISCUSSION
The Susitna River temperature model has been validated and
calibrated for the months of June through September 1981 and 1982. We
estimate that mainstem temperature predictions will be within 1. 09 to
-0.99 C of actual values, and upper tributary temperature predictions
will be within 2.10 to -1.54 C of actual values (90 percent confidence
intervals). This estimate assumes that the statistics computed from
simulations using two years of historical data will apply to project
conditions and there is no way of knowing if this assumption is valid.
Nevertheless, these statistics are a measure of the model's performance
given the best possible conditions and the available input data.
Tributary and mainstem temperature data from the 1983 field season are
expected to improve estimates of the model's accuracy and precision.
Additional analysis of distributed flow and temperature regimes and
tributary flow regimes will be required if the model's predictive
capabilities are to be improved, especially with respect to the upper
basin tributaries. We used a ground temperature model to estimate the
temperature of distributed flow. This model has not been validated with
data from within the Susitna Basin. If the parameter values defining
the model can be measured, or at least assigned physically relevant
constraints, the model can be applied with confidence to simulations of
the proposed project.
54
FUTURE APPLICATIONS AND ENHANCEMENTS
AEIDC will continue the Susitna flow and stream temperature
analysis by the following steps.
1. Normal and extreme flow regimes within the basin will be
defined by statistical analysis of the pre-and postproj ect
32-year flow records.
2. Using statistical analysis, AEIDC will determine the location
where postproject flows are significantly different from
natural flows. This will identify the area facing possible
hydrologic/hydraulic impacts.
3. Combinations of hydrology and meteorology which produce normal
and extreme stream temperature changes will be determined from
simulations using recorded meteorologic and hydrologic data.
4. Ranges of expected flows and temperatures resulting from the
filling and operational phases of the project will be used as
input to the temperature model for simulating downstream
effects. These simulations will use normal and extreme basin
hydrology and meteorology.
5. Results of these simulations will be analyzed and a zone of
predictable impacts identified. This zone will be partially
defined by estimates of the model's performance statistics.
6. Weekly or daily prediction capabilities will be pursued if the
need is indicated by analysis of the monthly simulations.
7. Results of the 1983 field season will be incorporated into the
model and new model performance statistics calculated.
8. Techniques will be developed for improving the distributed
flow temperature model.
9. Fall and winter conditions will be used for water temperature
simulations to provide estimates of the most upstream limit of
ice cover. If the stream temperature model reliably predicts
the recorded limits, the model will be applied to proposed
project conditions. Ice observations by R&M will be used for
validation of these simulations.
55
56
BIBLIOGRAPHY
Acres American, Inc. 1983a. Application for license for major project,
Susitna Hydroelectric Project, before the Federal Energy Regulatory
Commission. Vol. SA. Exhibit E, Chaps. 1 and 2 (figures). Alaska
Power Authority. Susitna Hydroelectric Project. 1 vol.
1983b. Slough hydrogeology report. Draft Report. Alaska Power
Authority. Susitna Hydroele~tric Project. 27 pp.
1982. Susitna Hydroelectric Project. Feasibility report. Vol.
4. Appendix A. Hydrologic studies. Final draft. Prepared for
the Alaska Power Authority. 1 vol.
Aitken, G.W. 1964a. Ground temperature observations, Big Delta,
Alaska. Cold Regions Research and Engineering lab., U.S. Army,
Hanover, NH. CRREL Technica~ Report 104. 15 pp.
1964b. Ground temperature observations, Gulkana, Alaska •
Regions Research and Engineering Lab. , U.S. Army, Hanover,
CRREL Technical Report 106. 13 pp.
. Cold
NH.
Alaska Dept. of Fish & Game. 1983. Susitna hydro aquatic studies,
phase 2 basic data report. Vol. 4. Aquatic habitat and instream
flow studies, 1982. Preliminary draft report. Anchorage, AK.
Alaska Power Authority. Susitna Hydro Aquatic Studies. Report for
Acres American, Inc. 7 vols.
1981.
report.
Studies.
Aquatic habitat and instream flow project. Final draft
Anchorage, AK. Alaska Power Authority. SuHydro Aquatic
Report for Acres American, Inc. 2 vols. in 3.
Alaska, University, Arctic Environmental Information and Data Center.
1983. Methodological approach to quantitative impact assessment
for the proposed Susitna hydroelectric project. Alaska Power
Authority. Susitna Hydro Aquatic Studies. Anchorage, AK. Report
for Harza/Ebasco Susitna Joint Venture. 71 pp.
Bigelow, B.B. 1983. Letter, March 8. U.S. Geological Survey,
Anchorage, AK. 7 pp.
Bredthauer, S. 1983. Personal communication. Telephone conversation.
R&M Consultants, Inc., Anchorage, AK.
Bredthauer, S., and B. Drage. 1982. River morphology. R & M
Consultants Inc., Anchorage, AK. Alaska Power Authority. Susitna
Hydroelectric Project. Report for Acres American, Inc. 1 vol.
Brett, J.R. 1971. Energetic responses of salmon to temperature. A
study of some thermal relations in the freshwater ecology of
sockeye salmon (Oncorhynchus nerka). American Zoologist.
11:99-113.
57
Cherry, D.S., K.L. Dickson, and J. Cairns. 1975. Temperatures selected
and avoided by fish at various acclimation temperatures. Journal
of the Fisheries Research Board of Canada. 32:485-491.
Coutant, C.C. 1970. Thermal resistance of adult coho (Oncorhynchus
Kisutch) and jack chinook (0. tshawytscha) salmon, and adult
steelhead trout (Salmo gairdneri) from the Columbia River.
Battelle, Pacific Northwest Labs., Richland, WA. USEAC R&D Dept.,
BNWL-1508.
Kilday, G.D. 1974. Mean monthly and annual
National Weather Service, U.S. National
Administration, Anchorage, AK. NOAA
NWSAR-10. 1 val.
precipitation in Alaska.
Oceanic and Atmospheric
technical memorandum.
McKendrick, J. 1982. Plant ecology studies. Final Report.
Agricultural Experiment
Alaska Power Authority.
Station, University of Alaska, Palmer, AK.
Susitna Hydroelectric Project. 124 pp.
Nelson, G. 1983. Personal communication. Telephone conversation.
U.S. Geological Survey, Anchorage, AK.
Quane, T. 1983. Personal communication. Interview. Alaska Dept. of
Fish & Game, SuHydro, Anchorage, AK.
R&M Consultants, Inc. 1982a. Field data collection and processing,
Supplement 1, Task 3-Hydrology. Alaska Power Authority. Susitna
Hydroelectric Project. Prepared for Acres American, Inc. 215 pp.
58
1982b. Field data collection and processing. Vol. 3. Task
3-Hydrology. Alaska Power Authority. Susitna Hydroelectric
Project. Prepared for Acres American, Inc. 1 val.
1982c. Field data index. Anchorage,
Authority. Susitna Hydroelectric Project.
American, Inc. 1 vol.
AK. Alaska
Report for
Power
Acres
1982d. Hydraulic and ice studies. Anchorage, AK. Alaska Power
Authority. Susitna Hydroelectric Project. Report for Acres
American, Inc. 1 vol.
1982e. Processed climatic data. Vol. 6. Devil Canyon Station.
Anchorage, AK. Alaska Power Authority.
Project. Report for Acres American, Inc.
Susitna Hydroelectric
1 vol.
1982f. Processed climatic data. Vol. 5. Watana Station.
Anchorage, AK. Alaska Power Authority. Susitna Hydroelectric
Project. Report for Acres American, Inc. 1 val.
1981.
Attachment
Authority.
can, Inc.
Susitna River mile index: Mouth to Susitna Glacier.
D to Hydrographic Surveys. Anchorage, AK. Alaska Power
Susitna Hydroelectric Project. Report for Acres Ameri-
1 vol.
R&M Consultants, Inc. 1980. Field data index. Anchorage, AK. Alaska
Power Authority. Susitna Hydroelectric Project. Report for Acres
American, Inc. 1 vol.
Reiser, D.W., and T.C. Bjornn. 1979. Habitat requirements of
salmonids. No. 1. in U.S. Forest Service. Influence of forest and
rangeland management on anadromous fish habitat in western North
America. Anadromous Fish Habitat Program. General Technical
Report PNW-96. 54 pp.
Sauntner, J. 1983. Personal communication. Interview. Alaska Dept.
of Fish & Game, SuHydro, Anchorage, AK.
Schoch, C. 1983. Personal communication. Telephone conversation. R&M
Consultants, Inc., Anchorage, AK.
Siefert, R.D. 1981. A solar design manual for Alaska.
Water Resources, University of Alaska, Fairbanks, AK.
6. 163 pp.
Institute of
Bulletin No.
Theurer, F., and K. Voos. 1982. An instream water temperature model of
the upper Colorado River basin. Unpublished. Paper for presen-
tation at the International Symposium on Hydrometeorology, Denver,
co. 6 pp.
Theurer, F., K. Voos, and W. Miller. 1983. Instream water temperature
model. Draft report. Instream Flow and Aquatic Systems Group,
U.S. Fish and Wildlife Service, Fort Collins, CO. Instream Flow
Information Paper No. 16. 263 pp.
Trihey, E.W. 1983. Personal communication.
consultant, Anchorage, AK.
Interview. Private
U.S. Geological Survey. 1981. Water resources data for Alaska.
Anchorage, AK. Water-Data Report AK-81-1. 395 pp.
1980. Water resources data for Alaska. Anchorage, AK.
Water-Data Report AK-80-1. 373 pp.
U.S. National Weather Service. 1980. Climatological data national
summary, Vol. 30, No. 9. Washington, DC.
1970. Climatological data national summary, Vol. 20, No. 8.
Washington, DC.
1969. Climatological data national summary, Vol. 19, No. 7.
Washington, DC.
1968. Climatological data national summary, Vol. 18, No. 6.
Washington, DC.
U.S. Soil Conservation Service. 1982. Precipitation and water yield.
Alaska Rivers Cooperative Study, Susitna River Basin,
Matanuska-Susitna Borough, Alaska. 4 pp.
59
Williams, G.P., and L.W. Gold. 1976. Ground temperatures. CBD180 in
National Research Council Canada. Canadian Building Digests
151-200. Ottawa, Canada, 1979.
Wise, J.L. 1983. Personal communication. Interview. Arctic
Environmental Information and Data Center,
Anchorage, M...
University of Alaska,
1979. Alaska solar radiation analysis. Arctic Environmental
Information and Data Center, University of Alaska, Anchorage, M...
27 pp.
1977. Mean annual precipitation in
mental Information and Data Center,
Anchorage, M... 1 map (scale 1:2,500,000).
inches. Arctic Environ-
University of Alaska,
World Meteorological Organization. 1982. Monthly climatic data for the
world, Vol. 35, No. 1. National Climatic Center, Asheville, NC.
1981. Monthly climatic data for the world, Vol. 34, No. l.
National Climatic Center, Asheville, NC.
60
APPENDIX A
TOPOGRAPHIC SHADING
TOPOGRAPHIC SHADING
These plots present the solar shading characteristics of the
Susitna reaches (refer to Figure 10). Mainstem reaches 9 and 10 and the
Talkeetna and Trapper tributaries were estimated to be unshaded for all
months. Fog Creek was assigned the same shading characteristics as
reach 1. The synthetic tributary (Cheechin) was assigned the same
characteristics as reach 4. The continuous curves represent the path of
the sun for each month. The hatched area represents the potential
shading of the surrounding terrain.
A-1
:>
I
N
1
Reach 1
RM 179.5-184.5
64NL
-
-
-
-
-
.?! nc: !on
10 ~ ~ M•y21 Jut 21 rr: ~ pm -
~/ ~ I ~-Sa I I pm
~ v I I Ao< ll -" t 21 ~ I '-,l ~ I Vi ~ ~pm Ba \ v. I -I
L \ I I I
"""
1/ I t ' ~/ ,,. v I V', I I i j I" I I~ pm
lX vt M•' 21 s •• 2! ri---I I ~ ~ ~ I I ..-I -I ~ rh I )pm 6• I I I I I I / I -~ I
tY. v lX I A I ~ ~ ~ I t' I ~ ('~~ t5 I \ ~ I I I I I ' ~ ~ ~ I -'y \ ~ ~ ~ ~ I N ['.. Op_m ~ .. m./ \ \ I
B~ ' /\ ~ ~ ~ ~ ~ ~~ ~ ~ ~ ~ I " ~ t" I,~ I y \ ~ ~~ I~~ ~ ~ ~ ~"i '\ ~
1:10' 130' 1'l0' 105' H 715' B ;;~ ;; N3~ 15' ~ 15' 30· <ls· so· 75' g 105' 120' 135' 15
o· ANGLES DO
EAST SOUTH WEST
Sun path diagram lor 64 °N latitude.
eo·
'70'
en w
..J
60' (j
2
so· c::t
40' w
0
J
30' ....
~
~·c::t
'lO'
iP
I w
Reach 2
RM 175.5-179.5
64NL
-
~--+---4---~---+--~--~~--+---+---~---+---4--~~--~--4---~---+--~----r---T-~BO·
-
-
713" ~~ 105" 120• 135" 150"
~
WEST
Sun palh diagram for 64 "N lalilude.
Reach 3
RM 166.0-175.5
64NL
so·
Ul w
_J
1---+---~---+--~~--+---~---+--~~--~--4----+--~~--+---4----+--~,---+---~---T--~Ga ~
2
c:l:
~--~--~--~--~--~~~-r~~~~~--~--+4--+--~~~~~+---~--~--~--~~40·w
0
:J
~---+---+---~~~~~~-4~~~~+-~4---~---+~~,~--~k-+-~~~~---+--~----r-~30· r
~
~~~4-~~-+~~--~----~--20· ~
BEARING ANGLES
EAST SOUTH WEST
Sun path diagram for 64 •N latitude.
Reach 4
RM 163.0-166.0
BEARING
EAST SOUTH
Sun path diagram for 64 •N latitude.
64NL
ANGLES
WEST
(})
w
..J
~
2
c:l
w
0
J
~
~
.J
~
Reach 5
RM 146.5-163.0
BEARING
EAST SOUTH
_Sun path diagram for 64 "N latitude.
S4NL
ANGLES
WEST
Ul
w
.J
(!)
2
c:t
w
0
J ....
.....
.J
<!
Reach 6
RM 142.5~146.5
B4NL
-
-
-
-
-
,21 no ion
~ -~ ID ~ !J.a 21 Ju121 ~ N---pm -
tn/ w I ~ Be I I ~prn
~ ~ I v\-Anr 21 :21 t+--I > ~ \ ~ ~prn a. I Vt I -
/ I I I I ~ 1/ I \ I f"-.,/
'" tv L)( l/\ I I I I ,"'-..., I ~ pm
l--t ~r2 s.o 21 t+--I It( ~~m ir,; I I I I -I ~ ~ I
6• I I I I \ I / I I
tY. [):' lX I V\ I \ I I ~ I (::; ~ I I I I I I I I I frh ll ~Ort_]l I ~ I t,---' I -v ;; ~ ~ f--._ ~ I I \ I I I N !"-.. B_E 4.,m/ \ I / "-" 1:"'0. t
~ k(y \ V\ ) v I ~ ~ ~'\. "\ ~) ~ ~ " 7 '" {'~ l ~ ~ ~ ~ )<: \
...KT'S{ ~~ ~~ ~ I ~ I Spn:;
\ I J_ t
f'..A'\.. "-KI' 'II. 1\..
1 z=so· 13z=s• 'UO!O" 1015·g 7!::i"B ~~~s~ Na~ 15· B 15" ~o~ ~5~ E s; 75· H '105· 120· '13!:5· '15
00
EAST SOUTH \NEST
Sun palh diagram lor 64 •N lalilude.
BCI'
70"
Ul w
.J
60' (.!)
2
so· cl:
40' w
0
J
aa· r -r
.J
20" <t
')Q"
~
I
00
1
Reach 7
RM 124.0-142.5
64NL
-
-
-
-
-
,21 n~ bn
10 ~ V\ Mov21 JuJ 21 ~ N: ~ rl--pm -
~/ w I ~·
Sa I I ~pm ;
~ v I vr Mr 21 Au, 21 t+--I "!--~ I ~ ~pm a. I ~ -\ I N / I I I I
""' 1/ I \ I r---...1
7,. v I [/"'\ I I I I
''""""
I I~ pm
~~ L)( vt ..... 21 S•p 21 I I I ~ ~ I I +-r,---... I -I ~ ~ I pm a. \ I I I \ II / I I
8( l);' lX I V\ I \ I I ~ I ~ ~ t'5 I I I I I I I I I frh 21 On 1r J I I I -'v I /: I I ~ k0 ~ ~"'-.'\ ~ ~ I Nl 4t~m/ I I / ~ I i'..~ ~ I Y\ ) v ·~ / ~ ~~ ~~) ~ ~ ~ ~ ~ /~ / ~ I
3fttn ){ ~ ~ ~ ~ I ~ ~~ I
I lL j I ~I "''' '\.h. ~' ~K :'\j
~o· 130' '1l20' 10!5' g 715' B ~o~ ;;s~ N3~ 15' 16 15' ~~ ~5~ EG; 75' g "105' "120' "135' "15
00
EAST SOUTH WEST
Sun path diagram lor 64 •N lalilude.
8Cl'
70'
en
w
..J
so-(.!)
2
so· q:
40' w
0
J
3CI' 1--s
2o· c:t
'l()•
Reach 8
RM 115.0-124.0
S4NL
-
-
-
-
'
-
.11 no pn
I ~ -H:: 10 ~ ~ Ma 21 Ju121 N: -pm -~/ w I I ~; Sa I I pn"l
~ v I v, ~o,21 ~u~ 2J I ( ~ N I N--~ ftpn"l Ba \ vr-I -
/ I I I I " 1/ I I "-./
"" fY I ~\ I I I I '""" I N Pn"l
lK l--t .~ .. 21 S•o 11 I I ~ ~fm ~~ I I +-~ I I -I ~ ~ I
6• I I I I \ d I I I
I~ [);' lX I V\ I \ i I ~ I ·~ ~~ ~ I I I I I I I \ I
f•h ll Oct 11 I \ v t:L -\/ ;; -~ I I I I I I N N ~ f3pn\ ·'~ ~"X~ ~ I I / A .~ ..... I
~ ~ ~ ~ ~ / I v ~ ~ ~ ~ ~ ~ 7 " IK: ~ "'" 21 ~ ~ rK ~ I ~ l:4$ ~ ~ I l ~ N. "--" ""'\.'X'\ 1"\.J'\J ~ N
1 oo· 1::Ja· '1020' 10l:i'H 7!:S'B ~o~,;s~ N3~ 15' {aS ~5· ~~ ~5~ Es~ 75' -g 105' 120' 135' 15
00
EAST SOUTH WEST
Sun path diagram for 64 •N latitude.
so·
70'
tn w
-' 60" t:l
2
50' c:t
40' w
0
J
30' ... -1-
.J
~· ct
'lCI'
Chulitna River
64NL
-
~--~---+----~--4---~--~ii----+---~--~----+----r--~----t----r---t----T---~---t----J---jDO•
-
-
WEST
Sun path diagram for 64 •N latitude.
Devil Creek. Upper ·rssssr
Lower 17ZZ2I
BEARING
EAST
Sun path diagram lor 54 •N latitude.
SOUTH
S4NL
ANGLES
WEST
en w
..J
(!)
2
~
Indian River Upper ~51
Lower V/7%3
S4NL
-
~--4---~--~~--+----b--~----+----+"--~----+---~--~----+---~--~----+---~--~----+---~ao·
-
~--+----~--~~--+----b--~----+----b--~----+---~--~~--+---~--~----+---~--~----+---~70'
SOUTH WEST
Sun path diagram lor 64•N latitude.
Portage Creek
64NL
~--~--~--~----~--~--~----+---~---4----+---~--~----~--~---f----t----r---t----t---~BC·
--+----+----~---4----~---4----4----4----4---~----4-----r----r----r----r--~~
~--~--~~--~----~--~----~--~----r---4-----t----+----+----+----+----+----+----+----+----t--~70·
Ul
w _,
~--~---~--~,---+---~--~----+---~--~----+---~---4----+----r--~----+----r---+----T---~6~ ~
2
q:
~--~--~--+---+---~~~~~~-r--~---+--~~~,-r-~~c---T---~--~---r---r---!40' w
0
:J
~-~--~--~~~~~Lr~~~r-~r-~~--~--r.-~rT--~K-r-~k-~~~T---T---T-~30· ~
1-_,
r-~---t--/1~111/~~t--t~~~~~---~--~~--~~{-T---r+-4~~~~---r---r--~20·~
BEARING ANGLES
EAST SOUTH WEST
Sun palh diagram for 64 •N lalilude.
Tsusena Creek
64NL
[J)
w
..J
~
2
<t
w
0
J
I-
I-
.J
~--~--+---~~4-~~~~--rvL-4+--~~--~--~~~~~~--~~~~~~~~~---r--~zo· ~
BEARING ANGLES
EAST SOUTH WEST
Sun path diagram for 64 •N latitude.
APPENDIX B
WIDTH/FLOW FUNCTIONS
WIDTH/FLOW FUNCTIONS
These graphs represent the relationship of wetted river width to
flow on a log/log scale. The solid lines connect HEC-2 predicted widths
for the six different flows used in the R&M (1982d) simulations. The
numbers associated with these solid lines are R&M cross-section identi-
fiers. Several R&M cross-sections were used for each reach as defined
for the SNTEMP network (refer to Figure 10). For more readable plots,
several plots are presented for a single reach when necessary. The
dashed line presents the flow/width function used in the SNTEMP simu-
lations.
B-1
tJj
I
N
1£4
1£3
STREA!f VJDTH IN F££T
5
2
5
2
2 3
1E3
REACH 1
CROSS SECTIONS HJ6. 3 -11.0: GRAPH 1 OF 2
11121
11219 ---__;;::;:::.::::-:::::==--------1 06. 3
COMPUTED WIDT 108
11217
4 5 6 7 8 9 2 3 5 6 7 8 9
1£4 lES
FLOV IN CUBIC F££T/SE!llil
b:!
I w
REACH 1
CROSS SECTIONS 111 -114 : GRAPH 2 OF 2
STREAM WIDTH IN FEET
1E4 ~------------------------------------------~------------------------------------~
5
2
114
1~ 4------------------------------
111 ---===-===--------=-113 5 -COMPUTED WIDTH 112
2
2 3 4 5 6 7 8 9 2 3 5 6 7 8 9
1~ 1E4 1E5
FLOV IN !lJBIC FEET/SECOND
REACH 2
CROSS SECTIONS 115 -117 I GRAPH 1 OF 1
STREAM VIDTH IN FEET
1E4
5
2
115
1~ 4--------------------------------r--------------------+----------l
5
2
2 3 4 5 6 7 8 9 2 3 5 6 7 8 9
1~ 1E4 IES
FLOV IN CUBIC FEET/SE~ID
td
I
Ln
1£4
1E3
1£2
5
2
5
2
STREAM WIDTH IN FEET
r-
r-
'-
-
-
-
-
r-
'-
--
--
r-
r-
r-
l
2
1E3
REACH 3
CROSS SECTIONS 118 -131 : GRAPH 1 Of 1
118
120 ~ --119 ---
COMPUTED WIDTH /"
I I I I l I I l I I l I I I I
3 4 5 6 7 8 9 2 3 4 5 6 7 8 9
1E4 lES
FLOW IN UJBIC FEET/SECOOD
REACH 4
CROSS SECTIONS 1il1 -121 : r&PH 1 Of 1
STREAM WIDTH IN FEET
1E4
f-
r'
'-
-
5 -
-
-
2 -
1E3 ----120
5 r------r-f.-------121 COMPUTED WIDTH
r--
2 r-
1E2 I I I I I I I I I I I I I I I I
2 3 4 5 6 7 8 9 2 3 4 5 6 7 8 9
1E3 1E4 lES
FLOV IN UJBIC FEETISECml
REACH 5
CROSS SECTIONS 60-64 a GRAPH 1 Of 2
1E4 STREAM VIDTH IN FEET -
1-
1-
r
1-
5
~
~
2 1-
1E3
1-
1-
1-
61Z1
5 1-
1---1----- --- --::::---64 62
I-COMPUTED WIDTH 61
63
2 1-
1E2 I I I I I I I I I I I I I I I I
2 3 4 5 6 7 8 9 2 3 4 5 6 7 8 g
1E3 1E4 1E5
FLOV IN OJBIC FEET/SECOND
b:l
I
CXl
1E4
1E3
1E2
5
2
5
2
STREAM WIDTH IN FEET
---
t-
t-
.....
-
t-
r ---
-
-
-
I
2
1E3
REACH 5
CROSS SECTIONS 65-68 r GRAPH 2 OF 2
65
____-/ L -___::---L.----~
COMPUTED WIDTH ~ 68
I I I I I I I I I I I I I I I
3 4 5 6 7 B 9 2 3 4 5 6 7 B 9
1E4 1£5
FLOV IN fl.IBIC FEET/SECOND
REACH 6.
CROSS SECilONS 57 -59 : GRAPH 1 Of 1
STREAM WIDTH IN FEET
1E4
1-
1-
1-
r-
5 f-
-
-
2 1-
1E3 59 --..-=:::: 59
r-~ ------f----·---·· -57
f--COMPUTED WIDTH 5
-
2 r-
1E2 I I I I I I I I I I I I I I I I
2 3 4 5 6 7 8 9 2 3 4 5 6 7 8 9
1E3 1E4 1£5
FLOV IN CUBIC FEET/SECOND
t:d
I ......
0
REACH 7
CROSS SECTIONS 28 -32 : GRAPH 1 (f 6
STREAM VIDTH IN FEET
1E4 ~------------------------------~-------------------------------,
5
30
2
28
32 ---1£3
31
29
5 ------COMPUTED WIDTH
2
2 3 4 5 6 7 8 9 2 3 4 5 6 7 8 9
1E3 1E4 lES
FLOV IN llJBIC FEET/SEClliD
td
I
1-'
1-'
IE4
IE2
5
2
STREAM VIDTH IN FEET
2
IE3
REACH 7
CROSS SECTIONS 33 -rt : GIUPH 2 OF 6
33
36
37
3_!-
COMPUTED WIIJ'TH ~-------------------35
3 4 5 6 7 8 g 2 3 4 5 6 7 8 9
1£4
FLOV IN CUBIC FEET/SECLHJ
IES
0::1
I ......
N
REACH 7
CROSS SECTIONS 38 -42 : GRAPH 3 OF 6
STREAM WiDTH JN FEET
1E4 ~------------------------------~----------------------------~
5
2
5
2
41
--------7-42
38
--~ ~------------------------------4-------------~-=~~~~~-----
2 3
1E3
---COMPUTED WIDTH
4 5 6 7 8 g
1E4
FLOV 1N tllHC FEET/SECOND
, _____ 4f21
2 3 4 5 6 7 8 9
1E5
REACH 7
CROSS SECTIONS 43 -47 : GRAPH 4 OF 6
STREAM WIDTH IN FEET
1E4 .--------------------------------.-------------------------------,
2
47
5 ---COMPUTED WirffH --------45
2
lE2
2 3 4 5 6 1 a g 2 3 4 5 6 7 8 9
lE3 1E4 1E5
FLO'i IN CUBIC FEET/SECOND
STREAM WIDHI JN FFIT
1E4
5
2
2
1E2
2 -3
1E3
REACH 7
CROSS SECTI~ 48 -52 z GRAPH 5 (f 6
---__..-
~-=----''--~
51
~-~~~~~----------~50
------L------------COMPUTED WIDTH
4 5 6 7 8 g 2 3
1E4
FLOV IN ClJBIC FEET/SECOND
4 5 6 7 8 9
1£5
t;tl
I ......
Ln
REACH 7
CROSS SECTIONS 53 -56 : GRAPH 6 OF 6
STREA~ VIDTtl IN FEET
1E4 ,-------------------------------~-------------------------------,
5 -
2
55
18 4--------------------------------~--------------~~~~~4------
5
2
2 3
18
-----56
___. L,..,.....,~L-==-------53 ---COMPUTED WIDTH
4 s 6 1 a 9 2 3 4 5 6 7 8 9
1E4
FLOW IN CUBIC FEET/SEWID
1£5
REACH 8
CROSS SECTIONS 19 -23 : GRAPH 1 Cf 2
STREAM VIDTtl IN FEET
1E4
5
2 3 4 5 6 7 8 9 2 3 4 5 6 7 8 9
1E3 1E4 lES
FLOV IN CUBIC FEET/SECOND
REACH 8
CROSS SECTIONS 24 -ll : GRAPH 2 (f 2
STREAM VJDTH IN FEET
JE4 ~-------------------------------.------------------------------~
5
2 -25
5 -----COMPUTED WIDTH
27
L_____,~---==--======= 24
2
2 3 4 5 6 7 9 9 2 3 4 5 6 7 9 9
1E3 1E4 lES
FLOV IN CliHC FEET/SEctNJ
STREAM WIDTH IN FEET
IE4
5
2
5 ·-
2
2 3
1E3
REACH 9
CROSS SECTIONS 3 -1 : GRAPU l (f 3
----COMPUTED wl15fH
4 5 6 1 8 9
lE4
FLOV IN !lJSIC FEETIS£cml
3
--
2 3 5 6 7 8 9
lES
REACH 9
CROSS SECTIONS 8 -12 1 GRAPH 2 ('f 3
STREAM VIDnl IN FEET
IE4 -
1-
f-
1-
5 ,.
.
r-
2 f-'
.-----1E3 ---10 _.;::;::-= -
f-/----8
t-/ 11
1---WIDTH 12
1-COMPUTED 5
/
-
9
r
2 -
1£2 I I I I I I I I I I I I -' I I I
2 3 4 5 6 7 8 9 2 3 4 5 6 7 8 9
J£3 JE4 JES
FLOV IN CUBIC FEET /SECOO
td
I
N
0
1£4
l£3
1E2
STREJ\1~ VIDTH IN FEET
r-
1--
1-
5 -
-
-
2 -
'-
1-
'-
-
5 -
-
-
2 -
I I
2 3
1E3
REACH 9
CROSS SECTIONS 13 -18 , GRAPH 3 OF 3
14
16
~17
_;;=-~5 -
""'' =-L13 ~------. 18 ---.......--COMPUTED WIDTH ,____.
I I I I I I I I I I I I I I
4 5 6 7 8 9 2 3 4 5 6 7 8 9
1E4 lES
FLOW IN CUBIC FEET/SECOND
APPENDIX C
LONGITUDINAL TEMPERATURE PROFILES
JUNE TO SEPTEMBER 1981-1982
LONGITUDINAL TEMPERATURE PROFILES
JUNE TO SEPTEMBER 1981-1982
These graphs represent both the predicted and observed temperatures
for the June, July, August, and September period of 1981 and 1982.
The observed data points are shown with 95 percent confidence
intervals. These confidence intervals are measures of the monthly
variations in the usable historical data for the Susitna Basin
(Figure 13).
Predicted temperatures are from the postcalibration simulations
with SNTEMP.
C-1
(")
I
N
20
15
10
5
Temperature (C)
190 180 170
Predicted Longitudinal Temp. Profiles
June 81 (957. confidence intervals)
I
160 150 140 130 120 110 100 90 80
River Mile
('")
I
UJ
20
15
10
5
Temperature (()
190 180 170
Predicted Long i tud i no l Temp. Profiles
July 81 (957. confidence intervals)
I
160 150 140 130 120 110 100 90 80
River Mile
Predicted Longitudinal Temp. Profiles
August 81 (957. confidence intervals)
T emporature (C)
20
15
10
r~--~~--~f--r+r-rr------rr'l~--
5
190 180 170 160 150 140 130 120 110 90 80
River Mile
n
I
L11
15
10
5
Predicted Longitudinal Tempo Profiles
Septambor 81 (95% confidence intorvals)
Tom perature (()
I 1~----~I~ri~I----~L ~
190 180 170 16fl 150 140 130 120 110 100 90 sa
River Mile
Predicted Long i tud i no l Temp. Profiles
June 82 (957. confidence intervals)
T emperoture (C)
20
15
10
5
190 180 170 160 150 140 130 120 110 100 90 80
River Mile
n
I
"
20
15
10
5
T omperature (C)
190 180 170
Predicted Long i tud ina l Temp. Profiles
July 82 (957. confidence intervals)
I I I
160 150 140 130 120 110 100 90 80
River Mile
CJ
I
CXI
20
15
10
5
Predicted Longitudinal Temp. Profiles
August 82 (957. confidence intervals)
T omperature (C)
_J:L-____ I ~ ---------
190 180 170 160 150 140 130 120 110 100 90 80
River Mile
Predicted Longitudinal Temp. Profiles
September 82 (95% confidence intervals)
T omperature (C)
20
15
10
5
190 180 170 160 150 140 130 120 110 100 90 80
River Mile