SUSITNAHYDROELECTRIC PROJECTFEDERAL ENERGY REGULATORY COMMI8810NPROJECT No. 7114

STREAM FLOW ANDTEMPERATURE MODELING IN

THE SUSITNA BASIN, ALASKA

UNIVERSITY OF ALASKA

ARCTIC ENVIRONMENTALINFORMATION ANDDATA CENTER

UND." CONTRACT TO FINAL REPORT

SEPTEMBER 1983DOCUMENT No. 882

3~tA\=~~tA\®OO·NA JOINT VENTURE

TK1425.sa1"472

___n_O.8_62_- ALASKA POWER AUTHORITY~--,

r:!

i!I

I'"'"

I

--

SUSITNA HYDROELECTRIC PROJECT

STREAM FLOW AND TEMPERATURE MODELINGIN THE SUSITNA BASIN r ALASKA

Report by

Arctic Environmental Informationand Data Center

Under Contract toHarza-Ebasco Susitna Joint Venture

Prepared forAlaska Power Authority

Final ReportSeptember 1983

Document No. 862

T/<\\'1J..SI S ~

f'i?~_

1'1'0· ~ (O'~

....

....

r

....

....

....

NOTICE

ANY QUESTIONS OR COMMENTS CONCERNINGTHIS REPORT SHOULD BE DIRECTED TO

THE ALASKA POWER AUTHORITYSUSITNA PROJECT OFFICE

iI

rI

rII

If'Il''i''Ili

....

....

-

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

AC'KNOWLEDGEM.ENTS ••••••••••••••••••••••••••••••••••••••••••••••••••••viii

INTRODUCTION ••••••••••••••••••••••••• •-• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 1

WATER BALANCE 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 ••Method II. Areal Precipitation Weighting.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 ••••••••

335779

101112121313

STREAM NETWORK TEMPERATURE SIMULATION MODEL......................... 17

r!

-

....

N('t')1.1')r-~

((.)('I)

000\!)\!)~('t')....('t')

-------~-

~

INTR.ODUCTION•••••••••••••••••••••••••••••••••••••••••••••• _•••••DESCRIPTION OF THE STREAM TEMPERATURE MODEL•••••MODIFICATIONS '..••••.••••••••.•••••••.•.•••••••••STREAM. NETWORK•••••••••••••••••••••STREAM. STRUCTURE••••••••••••••••

Topographic Shading •••.••••Stream Widths ••••••••••••••Hydraulic Retardance •••Tributary Assumptions •••••••••

HYDROLO,GY •••• '•••••••••••••••••••••••Flows ' .Stream Temperatures ' .

Observed Temperatures ••••••••••••••••••••••Synthetic Temperatures •••••••••••••••••••••

USGS Cantwell gage on the Susitna River(RM 223.7) ' .

USGS gage data collected on the Chulitnaand Talkeetna rivers ••••••••••••••••••••

Temperatures of Distributed Flow.'METEOROLOGY •••••••••• '•••••••••••••••••••••••••••

Selection of Meteorologic Data •••••••••••Ground Reflectivity and Atmospheric Dust.Meteorologic Predictions •••.•••.••.•••••••••••••••

V~IDATION •••••••••••••••••••••••••••••••••••••••••••CALIBRATION •••••••••••••••••••••RESULTS .ANl) DISCUSSION •••••••••••••••••••••.•••••••••••••••.•.•

iv

1718212124242729292929303030

30

333339394040465154

-

....

r!

r

....

TABLE OF CONTENTS (Continued)

PAGE NO.

FUTURE APPLICATIONS AND ENHANCEMENTS •••••••••••••••••••••••••••••••• 55

BIBLIOGRAPHY. • . . • • • • • • • • • • • • . • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 57

APPENDICES

A. TOPOGRAPHIC SHADING••.••.••••••••••••••••'••••••••••.•••..•• A-IB. WIDTH/FLOW FUNCTIONS...................................... B-1C. LONGITUDINAL TEMPERATURE PROFILES, JUNE TO

SEPT'EM:BER 1981-1982....................................... C-l

v

LIST OF FIGURES

PAGE NO.

gage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 4

C coefficients, Cantwell to Gold Creek Basin.............. 11

Flow balance sub-basins, Cantwell gage to Sunshine

Period of record for gage stations used in H20BAL......... 6

Calculated mean annual precipitation and water-yield values, Cantwell to Sunshine Basin.................. 9

Mean pre- and postproject June flow profiles, Watanato Sunshine Station, using precipitation-weightingwater balance method....................................... 14

r FIGURENUMBER

'I

I"'" 1.

1

l2.

3.

!"'"

4.

5.

6. Mean preproject June flows, Watana to SunshineStation, using three weighting methods.................... 15

7. Vertical air temperature distribution. Anchorage andFairbanks 1968, 1969, 1970, 1980, 1981, 1982•••••••••••••• 22

- 8. Vertical relative humidity distribution. Anchorageand 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 toSept. 1980~ 1981, 1982••...•............•..•......•....... 31

r13. Monthly stream temperatures, usable data June to Sept.

1980, 1981, 198.2.......................................... 32

14. Temperature regression for Susitna River at Cantwellgage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 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

­,rI

FIGURENUMBER

LIST OF FIGURES (Continued)

PAGE NO.

r

19. Susitna observed humidities vs. humidities predictedfrom 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.

23.

Tributary temperatures; 3 C groundwater inflow assumed •••• 49

Tributary temperatures; postcalibration, includingdistributed flow temperature model•••••••••••••••••••••••• 52

-r

24. Temperature model calibration statistics for tributarypredictiona. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

vii

­I

ACKNOWLEDGEMENTS

This report is the result of a cumulative effort by a number ofpeople. Special appreciation is extended to Gail Heineman (ADF&G) forfacilitating necessary data transfers; Tim Quane and Jay Sauntner (bothof ADF&G) and Carl Schoch (R&M) for sharing their on-site knowledge;Fred Theurer (IFG) and Woody Trihey for their technical guidance; andJeff Coffin (R&M) for his continued assistance throughout the project.At AEIDC~ we would like to thank the Resource and Sciences staff fortheir ongoing support in all phases of this report; Information Servicesstaff for library and bibliographic services; Judy Brogan for hereditorial assistance; and Nancy Walters and Sherrie Cornett for theproduction of this report.

viii

rr

-

....I

....r

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 postproject 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 a1. 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

....

.-

P­

I

-

model and how it was modified to more accurately reflect Alaska environ­

mental conditions. It also includes an analysis of the temperature

model's performance to date. The last section is a discussion of the

future applications and enhancements of both the water balance and

stream temperature models. including how they will be applied for esti­

mating project effects. This report does not include actual estimates

of project impacts.

2

....I

-

.....

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 within 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

J ' 1 1 J 1 I ) ) -- ---1 -1 1 1 I

~

+

@,l USGS Gaqa/Nodaloco tlon

Gl USGS - GOQl Stat Ion

• Node locol jon

o Sub-Basin Idenllflcatlon

'- Sub-Bosln Bliundliry

I Dam SiI,

LFORK

--..-9/"

e-9

FLOW BALANCE SUB-BASINS,Canlwei I Gage 10 Sunshine Goge

Scale t, 8miles

,,'"","

,,~<)

_____ 1

I'%j1-"

~IiCD

.....

rtl'%jo I-'

ocn:E:§ 0'tn. Pl

~'lu'::l ::lCD n

CDOQPl tn

OQ CCD 0'

I0'Pltn1-"::ltn

C"lPl::l

~CDI-'I-'

OQPl

OQCD

...

...

-

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, Watana, 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 f1ow-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

~ -1 -j -J ) 1 1 1 J J 1 1 1 ]

(j\

Gage Station

Figure 2. Period of record for gage stations used in H20BAL.

Susitna near Cantwell

Susitna at Watana

Susitna at Gold Creek

Chulitna River

Talkeetna River

Susitna at Sunshine

Yentna River

Susitna at SusitnaStation

I"""I

where:

Q is the mean flow for the given period (L3/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,

and

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 diffe~ence in flow between these two sites. The CI"""

coefficients are defined by: .

....

....C =s

where:

As

(2)

....

....

2A is the planimetered area (L ), and subscripts refer to

sub-basin, s, and the total basin, b.

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

--

c =s

P As s (3)

-

....

8

where:

P is the mean annual precipitation (L). and the subscripts

and other variables remain as previously defined.

The methods employed to determine the mean annual

precipitat10n 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 map 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,OOo-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.

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 As s (4)

9

-

where:

Y refers to the mean annual water yield (L), 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)

-i where:

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 Method II Method III

0.5104 0.6759 0.4636

10

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 As s s s s s s....(mi 2)

Pb

Yb ~ Pb~ Yb~

Clarence 76.8 1.3660 .4741 .0383 .0524 .0181Kosina 485.1 1. 4049 .8954 .2421 .3401 .2153t·J!atana 242.4 1.4439 .8837 .1210 .1747 .1090Deadman 218.4 .9979 1. 0971 .1090 .1087 .1212Tsusena 191.5 .7248 1.2529 .0996 .0693 .1175Fog 175.0 .9394 1.0428 .0873 .0820 .0891Devil 174.5 .5603 1.0310 .0871 .0488 .0896Chin-Chee 94.2 .4739 .8425 .0470 .0223 .0399.... Portage 186.4 .5018 1.2548 .0930 .0467 .1150I

Indian 159.4 .6913 1.0358 .0796 .0550 .0852

....

The mean observed value of C for the 13 months of record when data werew

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

~

I

....

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,wfor 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

r

....

,.,.,

flows at Watana. Acres (1982), using a drainage area-based C value ofw

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 samew .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 wat.er-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 accept~ 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 •

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.

12

....

....

,....

r-,

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

14

3.

r Figure 6. Mean preproject June flows, Watana to Sunshine Station,using three weighting methods.

I--------Preproject Flows (cfs)--------

r Location River Area Precip. Water-YieldI

Name Mile Weighting Weighting Weighting

Watana 184.4 23034 23034 23034Tsusena 181.3 23999 24056 24081

I"'" Fog 176.0 24844 25266 24876r Devil Canyon 161. 3 25688 25986 25675

Chinchee 154.6 26143 26314 26031Portage 148.8 27044 27003 27056Indian 138.6 27815 27815 27815Mckenzie 116.8 28543 28655 28687Whiskers 101.4 28787 28917 28952Chulitna 98.6 52359 52526 52589Talkeetna 97.2 63916 64064 64103Trapper 91.2 64117 64280 64291

I"'" 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.

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 maps 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

rr

Fi

-i

..... 16

2.F-

F"'" 3.

­I

....I

....

.....I

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 a1.

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.

SNTEMP contains a calibration technique which provides the

ability to adjust low-confidence input parameters to obtain

minimum prediction error.

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, tI synthetic data are

used to approximate temperatures during the construction and postproject

phases of the proposed project.

17

3.

rj

DESCRIPTION OF THE STREAM TEMPERATURE MODEL

SNTEMP 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;

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 with 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) (aT/at) + aT/ax = (qd/Q) (Td - T) + (BEH)/(QPCp )

I<-dynamic term-->!<------steady state equation---------->I

!<----dynamic temperature - steady flow equation-------->\

where:

A = flow area, L2

Q = flow, L'3/ t

T = temperature, T

t = time~ t

x = distance, L

qd = distributed inflow, L2 /t

Td = distributed inflow temperature, T

18

(6)

r

r,

...,

B = stream top width, L

EH = net heat flux, {E/L2 )/t

p = water density, M/L 3

c = specific heat of water, {E/M)/Tp

and dimensions are:

M - mass

T temperature

L - length

t - time

E - energy

The assumption of steady state (aT/at = 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)

....I

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

(9)

F

'I

....

....II

-i

....

i'I

where:

T = equilibrium temperature, Te

K1

= first order thermal exchange coefficient, [«E/L2 )/t)/T]

K2

= second order thermal exchange coefficient, [«E/L2 )/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 (LH = D).

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

3D-day meteorologic data averaging period has been selected and that it

takes 3D days for water to travel from point A to point B. Water

passing point B on the first day of this 3D-day period left point A 3D

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 3D-day average

meteorologic conditions.

This data averaging versus travel time dilemma can be overcome

either by 1) selecting averaging periods greater than the network 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.

2D

-

-

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 ana 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

] ) - J 1 1 ] 1 1~~,

":EjJUNE f-'.

N I!ll1IIAIlJlE LYS£ I'll AUGUST ~N I£IQlI 1M o(llJS TEIlPEIlAlURE LAPSE RATE t'im • i HE:1GHI IN METalS lD

~aa [

I-..l

I ._l .. ., • 2Sll11 ••• •• ":Ej<lIII lD

I 20aB l:- I f-'. t'i1lIIlIt- • 'f...o , ............. t'i rt...tt f-'•III nI. l-

IP o~ °Illo I 15011 t- o • ","".,.0••• I [~

III10011 t- ~l:I .D. I -f-"

18 l- N • ,p.... ° - , I • \Ot'i0'\OOn

SllIIr I

'(""~.. lD

* t- ° ° ""- • • I .... ~OD n' D. \OlD

I 8 I ! I I I ! I ! I , I , I I , I , ,DR P !, 0'It'i...\Dill

-18 -8 -6 -4 -2 8 2 4 6 8 18 12 14 .. n-II .. .. -I ·z • . l 1 • • II Il II TEMPERA TURE WEGREES Cl a

.... t'iIDlWlUi£~D \OlD

-..l00... f-'•

[jJ

.... rt\Ot'i00 f-'.ottJULY SEPTEMBER .. a

rtT£H?UlATURE LAPSE RATE IEHPEIlATURE LAPSE RATE_f-'oHEIGHT IN METEAs

i /lEIGHT IN METERS \003000 I 3008 00:;:1.....I ..

2508 t- I I .., rI' I 258a• 1/1 •• .... :;p-• • \0:;:1

oon~I- I IIl'.~ I I 2000 I=- N::r'

• ..~ . at'i

I 1588 r- III1588 l- """..'- .' I OQ1/1 III 'Ir6' .'" lD

1000 l- • • 11--... • • • I ,. r §• • •• • 0.

~t.• :>J 588 • •• • aD • a

0," .1 • ·0I , I I I , I , I • I I I , I . I ,8

-18 -8 -6 -4 -2 0 2 4 & 8 III 12 14 -18 -8 -6 -4 -2 8 2 4 6 B 18 12 14TEllPERATlJRE WEGREES 0 TEHPERATURE WEGREES Cl

Figure 8. Vertical relative humidity distribution. Anchorage and Fairbanks1968, 1969, 1970, 1980, 1981, 1982.

..----------- -

~ ~.. '".... ..~ .. ""..

..oil a ..,,"" .. 'C ..:- ...

,p ...... ,p .... a g.. .. "1 .. ...a • §l§ e::::UJ

".... l- e !Xl .. ::<f) ... :::l;:

'" ;;;:=l .. ;; UJ '" ~~ ~ l-e....l!:-e:: !l! l.1J

~U'l

~:; '"

~ ....- :; ...... ~ .. !I ""II!

~

is 8,!ii ~ !i! I I , I I I I , I I , I I I I ! , ~

'" '"~ ~ II ~ g ; ~ ~ ~ '-! !'1.

.. ..-'~

~

:;..a

~~ a .... .... .. ..,jI .. ..a .... .. ..

.P-8 a:- :-

'C'l.

<P B .. ....~ ..

~l!I §w ... ~ >-~

Z.:::J ..

~......l ..:=l

~--, --,

1;(~;; i :; ~-

~.. ..N

N.. ..~ ffi

~ ; ..i ..

~:'!

5 :s!ii .. i.. ....~ ::! Iiil

23

rr

....

.....

-~

I

""'",

....

­II

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 thesreach 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.

FOG195.9

181. 3

176. l2l

WATANA DAM SITE184.4

0-~'----ebTSUSENA2l2l8.6

rI

DEVIL174.3 161. 3

H CHINCHEE163.3

flIII!!II"" PORTAGE175.5 148.8

-INDIAN

138.6159.6

-

-

llfflIIil!!<,

WHISKERS113.2

CHULITNA116.5

TRAPPER115. 4

HH.4

98.6

91. 2

SUNSHINERM 83.8

TALKEETNA11212. 121

H refers to tributary head~

waters as defined in thestream network

J refers to tributaryjunction with the mainstem

Numbers refer to River Mileas interpolated from R&MRiver Mile Index (1981).

25

Figure 10. Susitna mainstem reach definitions.

"'"

-.

::"-to en!<~ Z

'1-05'0 0Ezi:i:w0

z{> "'

.!! c:( .!?

~W "e

'" 0:: '"., "'" 0

a: .,::E --a:W

.,:-~e t- o

"ttl "'Zu «1;'J;~ ....,nQ ::E

c:(zt:ttl:::>en

26

rI

-.....

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

!F"Figure 11. Tabular values of width function parameters.

!

Stream Reach IF Start End a b(mile) (mile)

Susitna 1 184.5 179.5 98.26 0.1577

Susitna 2 ·179.5 175.5 105.40 0.1708rI

Susitna 3 175.5 166.0 98.13 0.1820u

r Susitna 4 166.0 163.0 189.96 0.077 4"I,

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

r Susitna 8 124.0 115.0 33.95 0.3117

- Susitna 9 115.0 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 50

Devil 1 174.3 161.3 35

Chinchee* 1 163.3 154.6 25

Portage 1 175.5 148.8 60

F"" Indian 1 159.6 138.6 50

Whiskers 1 113.2 101.4 20r-I 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) flow b If "b"= a . (ds) •is undefined. "a" represents a constant width (feet).

*A synthetic stream representing the combined Chinook and ~hechako tributaries.

28

r'J

rI

:1

r'"1L

riI

-

r!

-r

Hydraulic Retardance

SNTEMP does not require stream velocity estimates to predict

average daily downstream water temperatures (see uDescription 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 "Water 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

M""II'i

r~,

rI',

r,...I

....

rI

....

r

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). St'I'eam 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

r- Figure 12. Monthly stream temperatures, available data June toSept. 1980, 1981, 1982.

i MainstemlTributary Number of DaysI 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

r 28.012.0 Yentna R. 26 31 31 14

I 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,-Deshka R. 21 31 31 3040.6/1.2

49.8/4.9 "Deception Cr. near Willow 5 8 8

49.8111.6 ··Willow Cr. near Willow 5 18 22

50.5/1.0 Little Willow Cr. 7 31 31 30

50.5 Susitna R. above Uttle Willow Cr. 7 31 31 24

61.2 Susitna R. above Kashwitna R. 2 27

..., 77.210.0 Montana Creek 19 24 1

n.5 Susitna R. above Montana Cr. 19 3 2 30

83.8 Susitna R., east shore-Parks Hwy. 20 14 30

83.9 Susitna R., west shore-Parks Hwy. 23 9 10 3097.0 Susitna R.-LRX 1 17

97.215.0 ··Talkeetna R. near Talkeetna 1 - -97.0/1.0 Talkeetna R. 10 31 31 3097.211.5 Talkeetna R. 17 1 31 3098.5118.0 "Chulitna R. near Talkeetna 1 - 27 30 3 2098.6/0.5 Chulitna A. 11 17 2098.6/0.6 Chulitna R. 17 10 25

~

103.0 Susitna A.-TKA fish wheel 11 10 19 22 7 28 31 25113.0 Susitna R.-LRX 18 25 31 30120.7 Susitna R.-Curry 25 31 30126.0 Susitna R.-Slough 8A 4 31 30126.1 Susitna A.-LRX 29 22 31 30129.2 Susitna A.-Slough 9 4 31 24130.8 Susitna A.-LRX 35 23 4 17

1"'"131.3 Susitna A. above 4th of July Cr. 15 31 30 26136.5 "SJJsitna R. near Gold Cr. 30 31 31 30 8 25 29 12 30136.810.0 Gold Creek 7 3- 138.6/1.0 Indian R. 23 31 4 28138.610.1 Indian R. 10 25 14138.7 Susitna R. above Indian R. 11 29 16140.0 Susitna A.-Slough 19 5 13140.1 Susilna R.-LRX 53 23142.0 Susitna R.-Slough 21 4 29 4 31 30148.8 Susitna R. above Portage Cr. 13 31 29148.8/0.1 Portage Cr. 13 26 28181.3/0.0 Tsusena Cr. 12 7 31 30184.4 ·Susitna R. at Walana dam site 30 - 31 30194.1/0.0 Watana Cr. 11 31 15 16206.8/0.0 Kosina Cr 4 31 17 12223.7 ··Susitna R. near Cantwell - - - 27 31 31 22231.3/0.0 Goose Creek 31 31 30233.410.0 Oshetna Creek 31 31 30

"R&M gages• ·USGS gagt:s

All others are ADF&G gages.31

Figure 13. Monthly stream temperatures, usable data June to

r Sept. 1980, 1981, 1982.I;

Iil

MainstemITributary Number of DaysRiver Mile River name 1description 1980 1981 1982

J J A 'S J J A S J J A S

10.110.5 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

i 28.0/2.0 Ventna R. 20 31 31

i 28.014.0 Ventna A. 14 31 31 24

29.5 Susitna A. above Ventna A. 10 31 31 30

32.3 Susitna A. above Yentna R. 18 31 29 6....40.611.2 Deshka R. 10 31 31 30

49.814.9 "Deception Cr. near Willow 2

49.8111.6 "Willow Cr. near Willow 13 4.... 50.5f1.0 Little Willow Cr. 31 31 28

50.5 Susitna R. above Little Willow Cr. 31 31 10

61.2 Susitna A. above Kashwitna R. 22

77.210.0 Montana Creek 6 17

77.5 Susitna R.above Montana Cr. 8 30

83.8 Susitna A., 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

!"'" 97.211.5 Talkeetna R. 14 31 30

I 98.5/18.0 "Chulitna A. near Talkeetna - - - - 24 30 10[

98.6/0.5 Chulitna R. 3 12

98.6/0.6 Chulitna R. 14 18

103.0 Susitna R.-TKA fishwheel 17 13 21 31 16

113.0 Susitna A.-LRX 18 17 31 30120.7 Susitna R.-Curry 17 31 30

!"'" 126.0 Susitna A.-Slough SA 29 30126.1 Susitna A.-LAX 29 13 31 30129.2 Susitna A.-Slough 9 31 20130.8 Susitna A.-LAX 35 6r 131.3 Susitna A. above 4th of JUly Cr. 31 26 22136.5 • 'Susitna A. above Gold Cr. 30 31 31 30 24 24 30136.810.0 Gold Creek138.6/1.0 Indian A. 16 31138.6/0.1 Indian A. 17 8138.7 Susitna A. above Indian A. 21 10

r 140.0 Susitna R.-Slough 19140.1 Susitna A.-LRX 53 23142.0 Susitna A.-Slough 21 28 31 30148.8 Susitna A. above Portage Cr. 31 28- 148.8/0.1 Portage Cr. 15 25181.3/0.0 Tsusena Cr. 31 31 30184.4 ·Susitna A. at Watana dam site 30 - 31 30194.1/0.0 Watana Cr. 31 6206.8/0.0 Kosina Cr 31 3223.7 "Susltna A. near Cantwell - - - - 24 31 31 15231.3/0.0 Goose Creek 31 31 30233.4/0.0 Oshetna Creek 15 31 24

"R&M gages.... ··USGS gages

All others are AOF&G gages.

32

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

assumption where the

downstream water temperatures.

stream temperature

These models

model used

employ an

calculated

to predict

equilibrium

equilibrium

rtemperature (T) represents the value the stream is asymptotically

eapproaching. The standard regression model of SNTEMP uses the

calculated T and the rate of approach to T as independent variables.e eFor the Susitna River application, this model was simplified to use only

the equilibri.um 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 observled 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 ± 1.74i

EQUILIBRIUM TDfPERATURE

-15

F-

I r.::l..::~Eo-<

10f"'" ~r.::l

s:r.::lE-+- ~r.::l 5:>..::r.::len~0

o!~

­I!

,...

1"'" 34

5 10 15 20 25

....

....

Figure 15. Temperature regression for Chulitna River at USGS gage.

Regression95 Percent

Regression Confidence ValueObserved (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 datacollected by ADF&G.

EQUILIBRIUM TEMPERATURE

r

rI

iI

r,

15

oo 5 10 15 20 25

35

*Temperature at gage was adjusted so downstream simulation matched datacollected by ADF&G.

r

252015105o

o

15

~~~ 1121wp.,

~E-o

A~ 5 I-----f-----f--~--_I_---___t---~

wl;I1i=Qo

r---I

....

....

EQUILIBRIUM TEMPERATURE

.....

,..,.II

36

r,...,I.....

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):

T (x,t) = T +(~T /2) cos [(2~t/t ) - x/~/at ] exp(-x/~/at )g g goo 0

where:

(10)

T =gAT =

g

rl

t

to

x

average annual ground surface temperature (C)

annual range of ground surface temperature

variation (C)

= time from occurrence of peak temperature (days)

= time for one cycle of temperature variation (365 days)

= depth (m)

= thermal diffusivity (m2 /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) areg

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:

= 21r/t , Bo

= /1r/a.t ,o

Tg(x,t) = Tair + Toff + 14.1 cos (At-Bx) exp(-Bx) (11)

r-Ii

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 discussion 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:

(12)

where:

TZ = air temperature at elevation Z (C)

TO = observed air temperature at elevation Zo (C)

~ Zo = elevation of site where air temperature is known

(Z = 105 m for Talkeetna)0

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 cm2 Jsec. 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 cm2 Jsec for a and 0.8 m for depth, the above formula reduces to:

T (t,Z) = 4.9 - y(Z-lOS) + 10.3 cos(At-0.316)g

(14)

I"""I

I

-

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 temperature 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 of assumed values. AEIDC believes this model

currently provides the best available approximation of 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

.-.Ii

rl

i'"I

-

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&M.

As was previously discussed, SNTEMP 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&M (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

Figure 17. Ground reflectivity and atmospheric dust coefficients,Matanuska Agricultural Experiment Station, Palmer.

l"""

ii

r

I"""W!~-1<:>-,,- ..-Zl.LJ......U......u..u..w0u

7 '"l "-') -.~. /'

~. /'~.~\

-

.~.-

/ \( \ -

\\\/ -

I I\ ./ -

)/ -

./. \./ \ -

/ \( -

/l. / -"'.

I I I , '", I, I 1

.-

uw0

>-0z

..-u0

0-WU)

Co.:):=>-<.

-1:=> (J')....,

I:J-Z0

Z ::Ii::=>....,

>-<:::::E:

a::::a..-<.

a::::<::::Eo

CDl.LJ1...1-

Z<:....,

>­f-......

0>­z­:=>..­aua:::: wCo.:) -1

u­wa::::

41

Figure 18. Susitna Basin observed air temperatures vs. temperaturespredicted from Talkeetna data.

JOOE 198B. 1981. 1982

PREDICTED TEMP. OEG C

SHERIlAN 14

(l

12

IJ£VIL CJo.NYDIl

f-Iii

VArANA 8

•5

I<OSIHA-2

B,....

B 2 4 5 8 III 12 14

IlISERVED TElIP. OEG C

JULy 198B. 19B1. 1982

PREDICTED TEMP. !lEGe

SIlERI~"'jl 14

D0 A- 12

OEVIt c..a.NYCN

- A111

~8

!YAWl/. r....

t•6 I-

KCSIW.

~4•

2

Z

a 2 6 B 18 12 14

!""" OBSERVE:: m? (lEG C

4?

,....II

Figure 19. (Continued)

....

foUGUST 19aa. 1981. 19a2

~ PREDICTED RH. OEC1lVJ.I.II

9i~N

a.9Q

Il.8 D

DE~'IL CANYON • A A·3.7

A

11.5t"'"

VAUNA9.5

+

9.4

KOSI/i},&.3

•0.2

11.1

Iil.0

ll.lI 9.1 9.2 3.3 0. ~ 15 11.6 117 z.a IJ.9 l.ll

OBSERVED Ril. DE.W!f,L

SEPTEMar..R 19SD, 1981. 1982

PREDICTED RII, DE.C1HfL1.0

- SIi£Rllf.N1l.9

00

3.8,""

DEY IL CUl'/ON11

+A.....

11.5

Vm.NAa.S

+

3.4

..... j(OSlllAI 3.3

+

;1.2

Il.l

~.3

~. a 11.1 <1.2 3.3 :1. 1, 11.5 U 0.1 Il.B 0.9 !.II

~::aSF.R'I:D PH. ~EC;~f.L

45

-

r

i"i

,-

....

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

basin 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 a1. 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' (MiS), 1980, 1981, 1982.

1'!'"\ AVE. WIND SPEED5,...--------------------WATANA

4

DEVIL CANYON

3TALKEETNA

aco0'\.....

2

-KOSINA

------1

.....co0'\.....

-- - - --

,01-- --'-_--'-_-'-_-'-_-'--_"'--__'------'"_--'-_----'

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

AVE. WIND SPEED5 ~:..:...;;;:..:.......;;~=-..;;.;;;..=:::....--------------·l

I,

~~ /:J_, / I ;'", / !',--~ i

.,)

i,

WATANA

3TALKEETNA

-'-" 2

. KOSINA

._--1

4DEVIL CANYON

r(

i

-

OL-----'"_--'-_-'-_-'-_-'--_.J...-_L....-____"_--'-_--'-_--'

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

5 ,=A;..;...VE=.::-• ...;.WI;..::=ND=-'S::.;:P;..::E=ED=- _

-WATANA

4DEVIL CANYON

TALKEETNA

2SHERMAN

'-- -

Nco0'\.....

47

"'-'-,... ...... / ................

.....,..- ...----~ ..._- ..._~---

1 -

Io ,-I-~--'-_.i..----'-_-"-----'-_-"-----'-_--"-----'-_I

JAN FEB MAR APR HAY JUN JUL AUG SEP OCT NOV DEC

KOSINA

Figure 21. Predicted vs. observed solar predictions, 1980, 1981, 1982.

[J

[J

<l

lSIlSI--.......

r (..)UJ

I en"'~:;)

~

c=:<:.-J0en

~(..).....63c=:c- IS

LO

IS IS IS IS ISKl IS LO IS LON - -

r

r

....

.... 48

Figure 22. Tributary temperatures; 3 C groundwater inflow assumed.

N-c

I:S-c

c

!"'" co

-.(....).....

r 0-::Eu.JI-

0UJ

I"- :>0:::UJen§

r'" coI

r

9

N

N- 51- co N

49

flow gaging,

substantially

2.

3.

4.

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. This

determination was based on the following logic.

1. . Without the benefit of continuous tributary

present stream flow estimates cannot be

enhanced.

With the subsequent necessary assumption of zero flow at the

tributary headwaters, initial tributary temperatures have no

influence on the predictions.

Tributary lengths were measured from maps.

Stream widths are based on field estimates and initial tests

with SNTEMP demonstrated that this variable was not sensitive

I""'"I.

il

r'"!

.....

-

r

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 mainstem 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 •

50

iI

rI,

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 cm~/sec for the entire basin,

Z = 0.4 m for Kosina Creek, Z = O. 7 m for Watana Creek, and Z = 2.0 m

for the mainstem and remaining tributaries. Further analysis is

necessary to validate these values.

The goodness of fit was determined by using the following

statistics:

....where:

IJ. = I: o. /n1.

s = It(o. - 1J.)2/(n-1)1.

0i = difference between i th predicted and observedtemperatures, C

Ti = i th published temperature, C

thTi = i temperature predicted by 8NTEMP, C

IJ. = mean difference, C

n = number of observed temperatures

8 = 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 1J.±(1.645)8 of the observed values. Postcalibration

statistics for the tributaries indicate that predicted values are on the

average 0.28 C (IJ.) 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 (8 = 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

Figure 23. Tributary temperatures; postcalibration, includingdistributed flow temperature model.

s...

-.....u.......

~.- cl:!:!u....~

N

N S CQ N... ...

- CQ

-U.......~

ffi~

..... ~ffienCQc

(Q

.....

.....

.... N...

.....

.....

,f,i1!IIll'l.,

52

....!

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

~ = -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 tributarypredictions.

n ~ (C) S (C)

June 1981 2 -2.08 1. 39

July 1981 1 0.41

r-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

I"""!

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: ~ = 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

r

r

-

....

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 proj ect

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

r3.

rI,

4.~

2.

r

-

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 postproject

32-year flow records.

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.

Combinations of hydrology and meteorology which produce normal

and extreme stream temperature changes will be determined from

simulations using recorded meteorologic and hydrologic data.

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.

I"'"

Il

-

6.

7.

8.

9.

Weekly or daily prediction capabilities will be pursued if the

need is indicated by analysis of the monthly simulations.

Results of the 1983 field season will be incorporated into the

model and new model performance statistics calculated.

Techniques will be developed for improving the distributed

flow temperature model.

Fall and winter conditions will be used for water temperature

simulations to provide estimates of tbe most upstream limit of

ice cover. If the stream temperature model reliably predicts

the recorded limits J the model will be applied to proposed

project conditions. Ice observations by R&M will be used for

validation of these simulations.

55

-

-

-

r56

I""",

BIBLIOGRAPHY

Acres American, Inc. 1983a. Application for license for major project,Susitna Hydroelectric Project, before the Federal Energy RegulatoryCommission. Vol. SA. Exhibit E, Chaps. 1 and 2 (figures). AlaskaPower Authority. Susitna Hydroelectric Project. 1 vol.

1983b.Authority.

Slough hydrogeology report. Draft Report.Susitna Hydroelectric Project. 27 pp.

Alaska Power

1982. Susitna Hydroelectric Project. Feasibility report. Vol.4. Appendix A. Hydrologic studies. Final draft. Prepared forthe Alaska Power Authority. 1 vol.

Aitken, G.W.Alaska.Hanover,

1964a. Ground temperature observations, Big Delta,Cold Regions Research and Engineering lab., U. S. Army,

NH. CRREL Technical Report 104. 15 pp.

Alaska Dept. of Fish & Game. 1983. Susitna hydro aquatic studies,phase 2 basic data report. Vol. 4. Aquatic habitat and instreamflow studies, 1982. Preliminary draft report. Anchorage, AK.Alaska Power Authority. Susitna Hydro Aquatic Studies. Report forAcres American, Inc. 7 vols.

.-iII

1964b. Ground temperature observations, Gulkana, Alaska.Regions Research and Engineering Lab., U. S. Army, Hanover,CRREL Technical Report 106. 13 pp •

ColdNH.

1981.report.Studies.

Aquatic habitat and instream flow project. Final draftAnchorage, AX. Alaska Power Authority. SuHydro AquaticReport for Acres American, Inc. 2 vols. in 3.

Alaska, University, Arctic Environmental Information and Data Center.1983. Methodological approach to quantitative impact assessmentfor the proposed Susitna hydroelectric project. Alaska PowerAuthority. Susitna Hydro Aquatic Studies. Anchorage, AX. Reportfor Harza/Ebasco Susitna Joint Venture. 71 pp.

Bigelow, B.B. 1983. Letter, March 8.Anchorage, AK. 7 pp.

U.S. Geological Survey,

....

Bredthauer, S. 1983. Personal communication. Telephone conversation.R&M Consultants, Inc., Anchorage, AX.

Bredthauer, S., and B. Drage. 1982. River morphology. R & MConsultants Inc., Anchorage, AK. Alaska Power Authority. SusitnaHydroelectric Project. Report for Acres American, Inc. 1 vol.

Brett, J.R. 1971. Energetic responses of salmon to temperature. Astudy of some thermal relations in the freshwater ecology ofsockeye 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. Journalof the Fisheries Research Board of Canada. 32:485-491 •

Coutant, C. C. 1970. Thermal resistance of adult coho (OncorhynchusKisutch) and jack chinook (0. tshawytscha) salmon, and adultsteelhead 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 precipitation in Alaska.

National Weather Service, U.S. National Oceanic and AtmosphericAdministration, Anchorage, AK. NOAA technical memorandum •NWSAR-10. 1 vol.

McKendrick, J. 1982. Plant ecology studies. Final Report.Agri.cultural Experiment Station, University of Alaska, Palmer, AK.Alaska Power Authority. Susitna Hydroelectric Project. 124 pp.

....i

Nelson, G. 1983. Personal communication •U.S. Geological Survey, Anchorage, AK.

Telephone conversation.

Quane, T. 1983. Personal communication. Interview. Alaska Dept. ofFish & Game, SuHydro, Anchorage, AK.

R&M Consultants, Inc. 1982a. Field data collection and processing,Supplement 1. Task 3-Hydrology. Alaska Power Authority. SusitnaHydroelectric Project. Prepared for Acres American, Inc. 215 pp.

1982b. Field data collection and processing.3-Hydrology. Alaska Power Authority. SusitnaProject. Prepared for Acres American, Inc. 1 vol.

Vol. 3. TaskHydroelectric

1982c. Field data index. Anchorage,Authority. Susitna Hydroelectric Project.American, Inc. 1 vol.

AK. AlaskaReport for

PowerAcres

1982d. Hydraulic and ice studies. Anchorage, AK. Alaska PowerAuthority. Susitna Hydroelectric Project. Report for AcresAmerican, Inc. 1 vol.

Devil Canyon Station.Susitna Hydroelectric

1 vol.

1982e. Processed climatic data. Vol. 6.Anchorage, AK. Alaska Power Authority.Project. Report for Acres American, Inc.

5. Watana Station.Susitna Hydroelectric

1 vol.

1982f. Processed climatic data. Vol.Anchorage, AK. Alaska Power Authority.Project. Report for Acres American, Inc.

1981.AttachmentAuthority.can, Inc.

Susitna River mile index: Mouth to Susitna Glacier.D to Hydrographic Surveys. Anchorage, AK. Alaska PowerSusitna Hydroelectric Project. Report for Acres Ameri­

1 vol.

58

R&M Consultants, Inc. 1980. Field data index. Anchorage, AK. AlaskaPower Authority. Susitna Hydroelect~ic Project. Report for AcresAmerican, Inc. 1 vol.

Reiser, D.W., and T.C. Bjornn. 1979. Habitat requirements ofsalmonids. No.1. in U.S. Forest Service. Influence of forest andrangeland management on anadromous fish habitat in western NorthAmerica. Anadromous Fish Habitat Program. General TechnicalReport 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&MConsultants, Inc., Anchorage, AK.

Theurer, F., and K. Voos. 1982. An instream water temperature model ofthe upper Colorado River basin. Unpublished. Paper for presen­tation at the International Symposium on Hydrometeorology, Denver,CO. 6 pp.

....I:

Siefert, R.D. 1981. A solar design manual for Alaska.Water Resources, University of Alaska, Fairbanks, AK.6. 163 pp •

Institute ofBulletin No.

i!

Theurer, F., K. Voos, and W. Miller. 1983.model. Draft report. Instream FlowU. S. Fish and Wildlife Service, FortInformation Paper No. 16. 263 pp.

Instream water temperatureand Aquatic Systems Group,Collins, CO. Instream Flow

U.S. Geological Survey. 1981. Water resources data for Alaska.Anchorage, AK. Water-Data Report AK-81-1. 395 pp.

Interview.

.....

....I

Trihey, E.W. 1983. Personal communication.consultant, Anchorage, AK•

1980. Water resources data for Alaska •Water-Data Report AK-80-1. 373 pp.

Private

Anchorage, AK.

i

U.s. National Weather Service. 1980. Climatological data nationalsummary, 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

i'II

Williams~ G.P., and L.W. Gold. 1976. Ground temperatures. CBD180 inNational Research Council Canada. Canadian Building Digests151-200. Ottawa, Canada~ 1979.

Wise, J.L. 1983. Personal communication. Interview. ArcticEnvironmental Information and Data Center, University of Alaska,Anchorage ~ AK.

1979. Alaska solar radiation analysis. Arctic EnvironmentalInformation and Data Center, University of Alaska~ Anchorage, AK.27 pp.

1977. Mean annual precipitation in inches. Arctic Environ­mental Information and Data Center, University of Alaska~

Anchorage~ AK. 1 map (scale 1:2~500~OOO).

World Meteorological Organization. 1982. Monthly climatic data for theworld~ Vol. 35, No.1. National Climatic Center~ Asheville~ NC.

1981. Monthly climatic data for the world~ Vol. 34, No. 1.National Climatic Center, Asheville~ NC.

60

-"""I

-

....

....

~I

....

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-I

I I C>C] c~c- -1 1 -~1 -'-~-l --,-----) ~-1 ----] ] ) -----~-l J ------1

Reach 1RM 179.5-184.5

B4NL

-I I I I I I I I I I I I I I I I I I I I leo'

-I I I , I I I I I 'I I I I I I I I I I I i acr

-

-

-

:rN

I \ , I I I I I I I I I I I I I I I I I 170'UlW

I I I I I I I I I I I I I I I I I I I I IsO' G2

r---+-+-+--i--r--r--t----,jJ~~w.;.:.r~bant-___r-_r__i-_r-_r-I_t-_l50' et

~~~?-+_4--\-_+_-+-_+_+--+__+"'""""l''"c_~..:.......;_7l--t_-+_-_t_-_+___i 40' WoJ

~~~-+_¥-+--_+_+-_++--+---_H_-_H-~"k_+_-+.,..~_.l:::'=__+_-_+_-_t_-__t 30' ...-~

-J.£+.rt-+-.I;-....,..f--l:-+--\-1,..L-+l----\6""l"""'C~~::_1!_--H..---:::'+T-_I_+__f".t__+_~~"'t___-+_-+____f~· <!

BEARING ANGLESEAST SOUTH WEST

Sun palh diagram lor 64 oN lalitude.

-

-

-

-

-_1l nc pn

~Inn-

10~

0.10.21 'U, 21K::~pm -

. 9p In/W \ t ~ pm

~V I I An, 21 .a.. , 71 I I "f...N\ v-r rr-.

~j:lpms. \ ~

I -I

/, I I I

~'/ I I I r--.../, ..

V \ [/'1\ I I I I I~ I

~pm

I~ lK L--t Mar 2 S.'" , ,I If( ~

I I r,---... , -

A IB I \

~ I pm\ \ II !I\ Icy[/(

~\

V\I \

, ,~

I

~ "f.-~n15 \ I I I II

\Ir,h'l Or, "

,'v \

~I -

\ ): V- -r--.. N N t-.. !!E".\

, I41Jm/ \ ./ r-....... I

~ ~~~~~........... I "-

I

~~~~0

~ ~ ~Join 21 Nl)"21 I I

I

~ ~~~ ~II

~~~ ~J

~ ~ ~" ~'\..'\..'\ ~ N ~~-- - ,,"-. " ,,'

~I

lJ,l

--I

,

]

Reach 2RM 175.5-179.5

EAST

Sun path diagram for 64°N latitude.

'1 J

SOUTH

] 1 1

WEST

J

64NL

1

so"

00-

70"

UlW

sO" ..J(j

250" ct

40" WoJ

30" J--~

~. ct

"KJ'

0"

1

<---") J J .I I ~ 1 'I -- --=J --c-c---l ) - --1 1 1 J j ----J

Reach 3RM 166.0*175.5

64NL

-I I I I I I I I I I I I I I I I I I I I I so'

-I I I I I I I I I I I I I I I I I I I I I DO"

I I I I I I I I I I I -J-- I I I I I I I I I 70'

-

-

WESTSOUTHEAST

UlW

I I I I I I I I I I I I I I I I I I I I I 60' ffi2

~+-t--t---t--t--+--+---rlnu~~T~b*-t----r-t--t-T--t-f----j 50' q

~~-.,.~~-1-"-+-+----+--+-f--+---+-~-.::--~.:......;.-+---t:---1I---t--+------I 40' WoJ

------=eb--."L..J4---+--¥+-i:-t----l----t---+--I-+--t-+---I...:::..,.,..-l---r:-k--=~~=_____t_-+-_t_-__1 30' t--~

\----~-~-~,,,£-~__~_':_4--~~---++-~I6.-"~~~{.~~dJ-~~-+_.f___f'.t__+_+___"r_I_''i_-+_-+____l20·~

. --i--.",.('1 10'

BEARING ANGLES:»I

Po

Sun path diagram lor 64 ON latilude.

1 ) ] Co ~) »»

-D J "000>-1 ) J .» J J J J --~-l -J

Reach 4RM 163.0-166.0

64NL

I I I I I I I I I I I I I I I I I I I I lBO'

I I I I I I I I I I I I I I I I I , I I I DO'

I I I I I I I I I I I I I I I , I I I I '70'

WESTANGLES

60UTH

BEARINGEAST

1>00~~~~~ \ -I I :;;(~,),)tb~~~'\t>~~';:-'~0l1O'

I I I I I I I I IliL·~ll"T" 4::Jp"q I I I I I I I 150'

I I I -,- I '" --ti" 7'f i I I' I I! I i 1"-oc ""':'("' I I I I 140'W

oJ

I-- I I J-<'k'i.~:A0 I -:AL" I j I ! I I I ! I! I .....k I / "" " L I ~ I 130' I--~

I J!-----l~; />..I'J..'c*:~a'0,p~~·J~' ... ' ...'d':"":""..7Sl.. I~' c:t

U1W..J-l-t--1-II-~160' rJI I I 2I I

I I I q:--l-+-jr--,'- I II I IL_~-+'-rl- ~ II I

~\JI

Sun palh diagram lor 64 ON lalilude.

----] '--1 ----1 -] -1 , ---I ') --, "I ---] J ---1 J 1 J

Reach 5RM 146.5-163.0

S4NL

I I I I I I I I I I I I I I I I I I I I leo"

~ I I I I I I I I I I I I I I I I I I 100-

WESTANGLES

SOUTHBEARING

EAST

~~~f-+-~+~~~h~~~~~~~~~~-J--l---L~L-L-Jqo"illoJ

~0\."..'\Io\.'\,'1.~~ I >,_L 7/ J4 I, 1/1, R"'....'\,'\...)..J\?'x"J..:"""....l'..>0-l'~?\o.'.}X'~~ 30" I-

~l'I. ',~ ':I. ,-", ,1.:'1. ,"~ I I f I I 'I~"\i.~'lJ~~~1· . Ar""'X'T' 'C'I\."~ I 7 I ,::cv'~~K''''~I~~~''N''''''~1 ',:"'k I I I&?G" '"

R' '\ ,,r:......Y'{ X "~~X'.['<\; I/'\ l'''\~'''l''~"""*" ?R ""~,\("-J~ 'R " "~ ~Xf "'~'\.I~ I C\.~ ~~~" '01 '10','- ,)4\jX.'-, V '. ~Ah. " x~, ,,'," 1-." \: "i ,,~/ ~ J ,,1'-.. .,,'

I I I I 1 1 1 1 I I I I I \ I I I I I I 170"

I I I I I I I I I I I

;~~t=tj=tjtjt:b~;t~~'~tl=~I=tjtj~ U1

I I I WI Iom n 16" .J

~

250" ~

~0"\

Sun path diagram lor 64 oN latilude.

) --1 ~- ---~ -1 ----1 J J --1 ----] --1 '--, ---J

Reach 6RM 142.5·146.5

64NL

BO'

10'

70'

40' WoJ

30' t--~

~,~

oc:r

UJW

60" .J(!J

2so' <t

C'

WESTSOUTHEAST

-

-

-

-l---l~-1---+---+-----+---f-.--+---+--t----lr--+--+--+---+-+--+--+--t--i--i1

-.,.nc~n I

I~IC~~ "'. 21 JUI21~~pm _

. BO~/~ \ ; Ii(~ pm

I /[)(" \ I ~".21 ~"·'I I I "f.,i'--..'B.~/ \ I ~v-r n--~! K ",ppm -/IY \ .....V\ \ I I 'f'.../ I "{f'..

1,0~" r{7l)(\V\ - \ I 1",.,2

1S.n" I t l'"'If('"'-)!~pmIY \ \ \~I n--.. ' I I I 'f.,. -

6 • m ' / \ \ \ ~ \ ,/ ~ I I ! f"- c 6~m

e~V\ lX \V\ \ Lh" ~Ol"l ! ~! 1'- }~7~ -~h~/ 'J; \/ \ ); \ ~L.-- ~~r----. I If\. I N ~""~ D~'i--~

~_~ ;.(\ ~v \/~~~~~~~~~~!'~'~~~'" ~ i / \ -S(~-1~ ~~'I'~~~~~~~ \c"'-" '\I- ~'\~"\

/ ~ /,:r--J

Sun palh diagram lor 64 ON latitude.

~, I .. ) ] ] J ..,] -·~I 1

Reach 7RM 124.0-142.5

B4NL

-

-

-

~00

I I I I I I I I I I I I I I I I I I I I I so'

-I I J I I I I I I I I I I I I I I I I I Iocr

-

I I I I I I I I I I I , I I 1 I I I I I \70' mW

\ I I I I I I I I I I I I I I I I I I I 16 0" a2

I I I I 50' <lI I I I I -

l--+--+--+-..-+--+-=-~~~+-+-\-t---+----I--+-~:--If--if>-.c~C:-----,+--+--+--+----+---J. 40' WoJ

~ ~~-~

---l--+---P-.:,......f--+-"""~--t---t----f 20' <l

SEARING ANGLESSOUTH WEST

Sun palh diagram lor 64 oN latitude.

Reach 8RM 115.0-124.0

~··~i ] --~J ~J -J

64NL

1

-

-

-

-,

-"non InIT _

~ M.. 21 1"'21 r--f"""'o.......12pm

IO~n->- r--r---f:~ p~. /k I I ~~ _. • '''/., I ••,,, -''''' I I ,.."","".p~/'~ \ ---t~ f-r-r--.. J I '-("'-.•• :r'/' \ ~ I I I r--.. J I '); p~

L' Y I -- I I I i, I I~", _,.y./' VI \ \I""'- <'.",(____.. ! : ~ I ~",p~~x)(\ \ I I f ~ f I i

a. ~ . / \ \ \ Lt \ I i I'f-,. I ~ 0~'Frn _L • I I 'II " I I ,z; v.: I Y \ -i \ :"",,- -"'J',_ I h IN, ~'!n'r---

", / \ V \ ;" / \ ,______ .c ~, ''',....' I " J 'N."-l. p,",

y V \ /" I /' _'"" "" " "0 I I:"l.K~"'"' IV"'~~\V \ /v p." __~~~~~I'",-<"',,'\::~~' &.:: \ k::::::\~~~~~l'\."':,R§~~~ /~l\."''>'\,'" "~'

BO'

00'

70'

rnw

sO' ..J~

250' e:t

40' Wo:J

30' ....

~00' e:t

-.0'

0"

:J>I

\0 EAST

Sun path diagram lor 64 oN latitude,

SOUTH WEST

~ -~1

i ! 1 ~'l '1 1 1 -- -J I 1 I '--1 -1 --J ----~-l J

Chulitna River64NL

-I I r I I I I I I I I I I I I I I I I I leo'

-I I I I I I I I I I I I I I I I I I I I I DO'

-I I I I I I I I I I I I I I I I I I I I '70'

-

-

:;I-'o

U)

WI I I I I I I I I I I I I I I I I I I I I 60' cl

21--+--+--t-+--t--t--t--'\r.~~9~~JPiTt--r---r---t--r---jr---T-r~ 50' ct

1--~-I--l--l--l-:--~~-7'-t=---t-1--+-1--1--+-+-+-~~~~~+---+---+--+---+---j 40' WoJ

~ ~~-l-..J

l--H----jI-!-~:+_-J.__+-f"c----iL-.t~+>+-__+-_+_-__j 20· c::t

BEARING ANGLESEAST SOUTH WEST

Sun palh diagram lor 64 ON latitude.

i 1 -J --J 1 -) _.. - ~I

I,_-0.-

i _OJ '''--,--

1

Devil Creek. Upper lS$SSl

Lower lZZZZl84NL

:r 1I-'I-'

-

-

-

-

....,.

,'U nconIll~ In ....

JO"''P............ ..--.L ",.. 21 '''121 r--~.2~~1 ~L ~pm -

ao~/)r' \ ~~pm

~v \ I ••,21 "1 , I~,,·R.V \ j,..--t- rt--....J J ~~pm -

/\ \ ............. \ I I I.............. I , :--..

\/ I I I r--.../ ~

''''''/./ \ 1./1 \' I I f'......r/J;~ >bc:n ~~...+-r-,..,...-br-'7":VlK£'~./ yV \ I 1~'21 S,'B'I I J I ,(//>///./;~ 'lh~_ \ \ \ ~I ~J, VAX/S(k'l'/: -

/7 'l~V~~~WX/I//i A \ I ~~-'7'/ I VA/:><:)X'7;// ">-...

%~~~~~0~~~~\ J..... "",,: /~~a0~~~~~~:rv~'/- ~Vh~f0~~~ >< ~///:~/~~ f< ~~~"I~~~~Vflv~~ ~ ~ I :"ll ~... J~V/2 ,/'Q X

~ ..><; -x.llN. to-. '/) . A. X Y

-_. ---- --_. ---- '- --- -- --- -- --- /'- --- -- -- -- -_. ./ - -- --- --- --

00'

ocr

70·(f]

W60" ..J

~

2~. ct

40' illoJ

30· ....-......J

~. c:t

10·

0"

EAST

Sun path diagram lor 64 oN latitude.

SOUTH WEST

~l J ] » ~J ---I c----1

Indian River Upper bSSSSJLower V/7/A

64NL~-

90'

(,

DO"

-

-

-

I I I I I I I I I I I I I I I I I I I I 170'UJill..J

I I I I I I 16 0" l'lI I I I I I I I I 2I I I I I 5<> <l

'lJ'-1J-~-+---:;k4~t~~=:;;;~~If1r~~1-rTl~ 40' g30' J--_Jl-l-JL--,-,1;~~~~~rf~~~i1:",d=:;~H=jf~tt~~tt~rF~!~-r-1-1 -

~~ ~<l-~

~

:P­I

I-'N

~x

100' 1315'

:x1lO!O' 10.~ ~~;;,:;,;

EAST

15' ~""> 15' 3D' 45' 60'

~ ANGLESGOUTH

76' .::(> 105'

~WEST

120' 135' 150'

Suo patll diagram lor 64 ON latitude.

- -) 1 J 1 J J ) --1

Portage Creek84NL

eo'

70'

aa

({JW

so- ..JrJ2

so' ct

40' WoJ

30' I--~

~, ct

10'

0'

WESTSOUTHEAST

-

-

-

-

-., m: Ion

Ivr -~10~ l-+- 0.1.,.21 JU'2.

~i-pm -

~/Y I I ~. ell I pm

fV~ \ I An,21 ~". 2/ I i ~~B.\ ~

Vi n---~

I N "'pm -/

,I I

~\/ I I i'..;,..V \ n, . \ , , , ,""'" I N

pm

I~ l)( vt lA., 21 Son >/ I ,I

~ ~orrnI ~ r,---.. I -I A ~

Is. \\ \ II I /\\ I

ry l/\ lX,

I I

~I

~ ~~n~\

~~~lI Ie

kJ ~~ I

~\ ,

I I

V \~~@ ~ I ~ I N N~.q noy \ r--.. I

l1~~~~~~~~)0~~)

~..........

I "

~~~'~~~"

~ ~ ~~I ~N~.~'"~ ~~ t\: ,; "-.l'J'o;:'" ~'\. ['1".1'\:~"" l"\.l\.

.,,", """-_. --_. ---- ---- / ....... -- -- -- / ....... ./'1;J;­I.....

W

Sun path diagram lor 64 oN latitude.

) --J J J 1 j ,-- -, 1 1 --1

Tsusena Creek64NL

-I I I I I I I I I I I I I I I I I I I I 190°

-~ I I I I I I I I I I I I I I I I I I 100"

I I I I I I I I I I I I I II I I I I I 1'70°

-

-

-

WESTSOUTHEAST

UJW

I I I I I I I I I I I I I I I I I I I I IsO" G2

~ ~.~

"::"::~:"-"'+--t-----+---l-+---I---"-+--4- .~~~:.F--+---f----+----4--+--j 40° WoJ

~ ~~-~

----JL:.."L--I-.L.-,.+----+--+----+..y.::~+____1I~--+_-H---I~..::....j,_+____t__.it~~~~~·___::~.,..__+_-___1 ~. C!

BEARING ANGLES>I....~

Sun palh diagram for 64 oN latitude.

-

.....

APPENDIX B

WIDTH/FLOW FUNCTIONS

I"""

!

I"""I

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 (l982d) 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

i

~

en

r co

e-

~ m t"- <c.-t ~ ~.-t .-t .-t

Ln

.....

r-N ("IJ

~--....~ e- l:§ ~-- ;::

1""" -- ttl::r: IL:i

u..L.) ('I') L.)

-< ~ - -en §l.Ji.J --0=:~

CO ::z-- ~ ~-~ ...Iu.......

~<c

CleJ w

\0t-:J..... Q.,~ -au

....

-~z

"'" -~~-:-~-en f3

("IJ -- f3 ~..... uJ- -....

B-2

--1 ---I - ) I 1 1 -- -1

REACH 1CROSS SECTIONS 111 - 114 J GRAPH 2 (If 2

STREAM VIDTH IN FEETlE4 I • I

5

2

114IS -I I·------::;;>"=-er~ _

-'113

112

111- ---COMPUTED WIDTH

5

2

1£2 I I I I I I , I I I 'I I I , , I , I

tpI

Vol 1£32 3 4 5 6 7 8 9

lE4

FLOY IN llJBIC FEET/SECOND

2 3 4 5 6 7 9 9IF.5

.....

F"; ~0)

CD....\ -

tolfl

CD \~,... - --- -- in

....

N

- N! -

l:!s

~Sr'"~

("'\J V)

;:::- tl:i::c -- ;::'jI.L..

U I.~- -ce::: in

\~ en ~-lJ.J -0:::: ~

CD :z:

\; -- - 9.-t3 I.L.U')

~C

to

r- lJJt- in:Ja..::E0 ...u

~.-V)/---I-....I......L......I..--l.-..L.-.J--_..t- J-I....J......I.....J..---Ji..-....L.._....L..__...l- 4-

L1i- f:3- ~-f:3-

r B-4

J 1 . 1 ] ) 11

REACH 3

) 1 .···1 ~J 1

STREAM WIDTH IN FEETlE4

1£3

lE2

5

2

5

2

CROSS SECTIlIlS 119 - 121, GRAPH 1 Of 1

l-I-

I-

.1-

l-

I-

-

I-

~ 118l-

I-12121

I-

~I- 119- ----I- COMPUTED WIDTH /

l-

I-

I I I I I I I , , , I I , I I I

tdI

VI

21£3

3 4 5 6 7 9 91£4

FLOV IN ClIJIC FEET/SECOOD

2 3 4 5 6 7 9 9IF.5

) J J 1 . 'I J ~1 J '1 ]

2

5

STREAM WIDTH IN FEET1£4

1£3

1£2

5

2

REACH 4CROSS SECTIONS 120 - 121 I GRAPH 1 (f 1

t-l-

f-

f-

f-

f-

l-

I-

-l-

I-

l-

I-

I-

- 12(21

COMPUTED WIDTH -- ~--.--.-------'-----

I-

- 121

l-

I I I I I I I I I I I I I I I I

t:dI0\

21£3

3 4 5 6 7 8 91£4

FLOW IN UlJIC FEET/SEcm:J

2 3 4 5 6 7 8 91£5

--

~C\J""-

ISl (Y) -(0~(O(O

(0

-

\ -

\ -

\

\-

\\

\ --

\::I: -l-e -1-1

~ -elJ.lI- -::Ja...~0u -

-

f I I 1 • I I I • • I I I I I I

-

....,

-

'"""'

N

~-5:tn ~

::J:: •(Q

u a-<I..J.J fS0::: -..-

~

§

8.....z-

~- ['3- ~-

~-

-

N

~t.n-..-ttl

~.....u- -en §

CD :z:-t- ~

-J.....(Q

U")

•

['3-

B-7

--

\t--.-

tn co CD -(0 CO ro CO

-\ -

N -l\

i\ -l

\\

\ --

\:c -I-0 -toot

3 -0WI- -:::lQ.:£0U -

-

, I , , • , , I I , 1 • • I f f

~

N

\5N....

I

~t..n

::c CZ)(Q

L..)tS-<

Lt...J~,.. 0::: -t-t3en

.... ~

-

...,.u.J-

N N

~-

~-

-

N

i-.....t:tl~

.....C-)- -a)

~CD =-e- 9.....cc

U")

-

N

C3-

B-3

,_..- --J '~·--l - 1

REACH 6·CROSS SEenONS 57 - 59 I GRAPH 1 Of 1

2

5

5

STREAM VIDTH IN fEETlE4 .

1E3

-

59

COMPUTED WlDTH---

59

,.---------

57

2

1£2

b:lI

'-0

2lE3

3 4 5 6 7 8 9lE4

fLOV IN CUBIC fEET/SECOND

2 3 4 5 6 7 8 9lE5

1 --~1 ---1 -] --1 ~·]----1 -----1 J

REACH 7CROSS SECTIONS 28 - 32 r GRAPH 1 (J 6

STREAM VIDTH IN FEET1£4 . i • i

5

___ 3121

c:- ~ 311 ~ ·29

------COMPUTED WIDTH

2

5

28........-:::: :::: 32

1£3 I -- ----1""':::::""'-- ---.-::;:::00 I

2

JE2 J '" I I , I I_ _ I " I J I I I , ,

tpI

t-'o 1£3

2 3 4 5 6 7 8 91£4

flOW IN (1JBJC FEET/SEC{}I()

2 3 4 5 6 7 8 91E5

,....

t-

CCI

~N

~,....

~-<~

t'- ~

~-::::c F;; ~

i1iLa..

L.) U- --< ~ \ C)

~LJ.....la:::: ~ CD :z--

9-~

It'--~ La..

~CCI....

~U")

ClLIJ,.,.. t- -:Jtl.~0U

I!""~--

C)

CD~

It'--

('f) (0 r-. ~\ 10 CCI(Y) (Y) (Y) ('f) (Y)

In

r -l"'"

.... - Nl±ILa..

:z-- ::c.......l:::)-=-:II:,.... i:ia::.......

GJc.n -In N- ~ C:I...... L&.I- -

B-ll

- -) -I :---1 j 1 Cl § ) 1 1 1 1 1

REACH 7CROSS SECTIOllS 38 - 42 I GRAPH 3 If 6

lE~

5

STREAM VIDTtI IN fEET

1;;sJiJAi~ 40

--­COMPUTED WIDTH

2

5

41

J :/'- 4238

18 -I' . I z? c=e ;a=- 5?C 39 1

2

Jf2 I " I , I I I I I ' , I I I I I I I

b:lIt-'N

1£32 3 4 5 6. 7 8 9

lE4

FLOW IN UJBIC fEET/SECOND

2 3 4 5 6 7 8 9lE5

~-0)

CD

"¢'\ to en

r-

r-.. III <.c"¢' "¢' it "¢' "¢'

r ~r Ini

--

.....<.c

~

--.- if e~~

~ bS-I-:::J:: r- , ttl-- ?\ -- ~

t....) I uJ t~- --< ('"f)

\0) §--u...J saa::: cs CD %--.-

~ 9~r-.

u-

~<.c....

~

InCJIJJI-::J --a.:40U

~:z:-

;:::;-In

C3-In

~-al-

B-13

~~ J -1 'C] J - 1 1 . -) -1

REACH 7CROSS SECTIOlJS 48 - 52 I GRAPH 5 (f 6

STREAM VIOTJI IN FFIT1E4 I I ,

5

___ --- o:r ....COMPUTED WIDTH

2

5

/ 52

48

1£3 -I i /' 7C=;;;=- ---:- I_ 4t:1

51I _."..:;/~:'-;'~/L-_-------:;;-+- /pa<"'" 5# /'" 5122

2

1E2 , " I I I I I I , ' J I ( I I I , t

tJ:lI.....

.p.

1£32 3 4 5 6 7 B 9

1E4

FLOW IN CUBIC FEETlS£IDID

2 3 4 5 6 7 8 91ES

1 C-l 1 J , 1 --1 --1 I -] 1

STRH.14 VIDrtt IN FEET

REACH 7CROSS SECTlOOS 53 - 56. GRAPH 6 OF 6

ilE4 i i

5 H

2

1£3 llr------------------~, ----------:=-~-.....:::::

5 ~~

2

--- J:-. ~ - 53--- ::>COMPUTED WIDTH

1E2Iii iii I I i I ! i I I I I I I ,

t::dIt-'VI

1£32 3 4 5 6 7 8 9

lE4

FlOV IN ClIBIC FEET/SEmI)

2 3 4 5 6 7 8 9lE5

1 J 1 ] -c-~ --J 1

REACH 8CROSS SECTIONS 19 - 23 I GRAPH 1 (f 2

STREAM WIDTtI IN FEET1£4 I • I

5

2

COMPUTED "ttI"OTH"- I c= / -5

1E3 I I ;;»' .........-- 7" ...--- .-,.,;';-- I

2

1£2 I I I I I I I ! ! I I I I I I I I ! I

b:lJ.....0\

1£32 3 4 5 6 1 8 9

1£4

FlOY IN ClIJIC FEETISECOOD

2 3 4 5 6 7 8 9lES

-1 -J -] J I -1 I ~·1 . ·--1 - 1

REACH 8CROSS SECTIOOS 24 - l/ I GRAPH 2 (f 2

STREAM WIDTH IN fEETlE4 I • i

5

25

27J =- - 24----COMPUTED WIDTH

2

5

2

lE3 . I I 7" ......... -- I

1£2 I " I I , , I , I I I I I I , I I I

t:xlI--...J lE3

2 3 4 5 6 7 8 9lE4

FLOW IN ClIHe FEETISECCW

2 3 4 5 6 7 8 9lE5

~,

I

rT)

~

""" ~~0") ~-r :::r:: l:'--

~ffi

L) I U- ---e: rT)

\ c::n mLW ~ t:Sa:::. CO :z.- --u

\~ l:'-- ~~l:i

~ e CO

3:

Cl II')

Wt-:Ja.~0U

tn-c::n

co

\l:'--

toen ~ 11') rot--

\ II')

\ .......I

I

.....

..-~:z.-F i=~-=-

.... ~IV)

II') N II')

..". f:3w.II""" - -

N

f:3-

B-18

­i

F'

("I')

~N

en ~.- ::c N-<-..) I

<e::. CICI

l..L.J~Ct::: --ul:H

.-~

--

-l±iL&..

Z-~­:.

--

\ -~ CD::: N -.... en

\ -

\ -

\ -\

-1\

~~

"

\J:--

~--

C1UJ -I-:Ja.. -~0U

-

-

I , r , , r I , , , , , r r I

N

C:f-

-

N

e~-l±i- L&..

uJ U- §en

CICI z-t-- ~

~(Q

II")

-

[3-

B-19

J ---] - 1 --l -J

REACH 9

,- ---1 - "J --~ 1

STREAM WIDTH IN FEET1£4

1£3

lE2

5

2

5

2

CROSS SECTIONS 13 - 18 I GRAPH 3 OF 3

-l-I-l-

I-

-l-

I..

14

16

..-:::::::::.. 17

I-

~~15,-

I-

~

~ ;::::::::;-- ~/13

- l--- --

- -- --COMPUTED WIDTH~

18

l-

.1--

l-

I-

I I I I I I I I I I I I I I I I

tl::IItvo

2lE3

3 4 5 6 7 B 91£4

FLOW IN OJBIC FEET/SECOND

2 3 4 5 6 7 8 91£5

­,

....,

­[ 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-l

) J ~...J 1 1 -"1 "J ) J 1 1

Temperoture (D20

15

5

Predicted Longitudinal Temp. ProfilesJune 81 (957. confidonce intervals)

o190 180 170 160 150 140 130 120 110 100 g~ . 80

('")I

N River Mila

"~J] ~1] 1 '] '~"1 --" ~ J J 1 ") J

20

15

Temperature (D

Predicted Longitudinal Temp. ProfilesJuly 81 (95% confidence intervals)

10

5

o

--L--1 -0 L- -- T l~._

190 180 170 160 150 140 130 120 110 100 90 80(')I .

WRiver Mile

] c_] J j 1 ] J } - ] ---') _C_-~C'--l CJ C 1

Predicted Longitudinal Temp. ProfilesAugust 81 (95% confidence intervals>

20

15

10

5

o

Temperature (0

r 1 r t T I l_~

190 180 170 160 150 140 130 120 110 100 90 80(')I

.j:-.

Ri vel" Mi Ie

ISlc:c

ISl0')

ISc:c-

lSIen-

ISl-Lrl-=(".y

CDt.::l....ot.oa..e~

""" gs-00Q)~

• r-fIS

4- --0L

0-.

-. ISII CO N

~0-. 0E >

I. t.ID 0....

I-- c...Q)

lSI.- (J~ l: CT')

0 ~ CD-C.... .........

:IE:C• ...-f 0

t.U

(JCD

~ >::) ....Lrl

IS e:::C)

+J ....... "'I"

• r-f -c:cI'-- OJ t.

C 0...D

0 eCD

--.J ....ISla.. LrlQ -"..,. c.n

-0ill

+J.... 0

ISl• ...-f (,0

-0illL

0-.ISle----

C-5

J -J - J '1- 1 ] -. -] ) ) 1 -1 ) 1 1

20

15

10

5

Temperature (C)

Predicted Longitudinal Temp_ ProfilesJune 82 (95% confidence intervals>

·11

o

River Mi Ie

nI0\

190 180 170 160 150 140 130 120 11~ am 90 80

J ··1 1 1 ~. 1 -J I

213

15

Tempef'otuf'e (C)

Predicted Longitudinal Temp. ProfilesJuly 82 (95t confidence intervals)

113

5

o

___ ~ I I I_ --- .. lI

190 180 170 160 1513 140 1313 120 110 1130 90 80nI

-...J River Mile

CSllSI

(J)

ill...--i

r- .......<..j.....

CSl

0--

LCL

.o.....~

lSIN

E 0

ill.:>

L

r-- ~....I:

...--i ell(,)

lSI

G cCT'l

0

C -CCII... .-

....... ... ...C

x

U 0I""" (,)

L

::>CII

~

:>

4-> L.r'l

...0)

lSI a::::

-- ~

....... -~

OI~C ....0

(0:J

-.-JC"l:J~

lSIL.r'l

U-

Q)

4->()

~ • r-f

UlSl(C

Q)

LCL

lSlr-.-

~

......W--

CII

lSl

~ 5cc

.....0LCD0-eCD..... lSl

0)

Gl L.r'lN

lSl- lSl

C-8

'~-1 ""'] ) ~~'1 '-~"

Predicted Longitudinal Temp. ProfilesSeptombe~ 82 (95% confidence intepvcls)

'IC'~J 1 )

20

15

10

5

o

Tompe~ctu~e (C)

~-----------h I I -r

190 180 170 160 150 140 130 120 110 100 90 80

(')I

1.0 Rive~ Mile