DRAFT THESIS PROPOSAL (08/21/02)

Contributions of Glacial Meltwater to Streamflow at Thunder Creek, Using a Distributed Hydrology Model, North Cascades National Park,

Washington

Thesis Proposal for the Master of Science Degree, Department of Geology, Western Washington University, Bellingham, Washington

Jay W. Chennault August, 2002

Approved by Advisory Committee Members:

________________________________________________________________________

Dr. Robert Mitchell, Thesis Committee Chair

________________________________________________________________________

Dr. Douglas Clark, Thesis Committee Advisor

________________________________________________________________________

Mr. Jon Riedel, Thesis Committee Advisor

________________________________________________________________________

Dr. Pascal Storck, Thesis Committee Advisor

TABLE OF CONTENTS

1.0 PROBLEM STATEMENT

2.0 INTRODUCTION

3.0 BACKGROUND

3.1 Thunder Creek Watershed

3.1.1 Topography

3.1.2 Geology

3.1.3 Glaciers

3.1.4 Soils

3.1.5 Vegetation

3.1.6 Climate

3.2 Previous Work

3.2.1 Glacier Inventories and Mass Balance Estimates

3.2.2 Glacier-Runoff Studies

3.2.3 DHSVM Studies

4.0 PROPOSED RESEARCH

5.0 METHODS AND TIMELINE

5.1 DHSVM Input Requirements

5.2 Modeling

5.3 Expected Outcomes

6.0 SIGNIFICANCE OF RESEARCH

7.0 REFERENCES

8.0 FIGURES

FIGURE 1: Location of the Thunder Creek Watershed FIGURE 2: Conceptual Diagram of DHSVM FIGURE 3: Shaded Topography of the Thunder Creek Watershed FIGURE 4: Digital Elevation Model of the Thunder Creek Watershed FIGURE 5: Boundary of the Thunder Creek Watershed FIGURE 6A: Soil Type from Miller and White (1998) in the Thunder Creek

Watershed FIGURE 6B: Soil Type from Surficial Geology Mapping in the Thunder

Creek Watershed FIGURE 7: Soil Thickness in the Thunder Creek Watershed FIGURE 8A: Vegetation in the Thunder Creek Watershed, LIA Glacier

Extent FIGURE 8B: Vegetation in the Thunder Creek Watershed, 1950’s Glacier

Extent FIGURE 8C: Vegetation in the Thunder Creek Watershed, 1998 Glacier

Extent

1.0 PROBLEM STATEMENT

I propose to examine the relationship between the streamflow and short-term

climate changes observed in the glacial activity in the Thunder Creek watershed, using

the newly developed Distributed Hydrology-Soil-Vegetation Model (DHSVM). The

model will be calibrated to historical data and used to estimate glacial contributions to

streamflow in the basin during the Little Ice Age maximum in the North Cascades. I will

also apply the model to examine the impact of glacial retreat on streamflow.

2.0 INTRODUCTION

Glacial runoff studies are important because glaciers represent valuable

resources that are vital components to hydrologic systems and aquatic ecosystems.

Glaciers also influence soil development, the distribution of vegetation, and are indicators

of climate change. Runoff from some glaciers in the North Cascades is currently being

utilized for hydroelectric potential (e.g., Diablo, Ross, and Gorge Reservoirs by Seattle

City Light; Baker Lake and Lake Shannon by Puget Power) and municipal capabilities

(e.g., Nooksack River by the cities of Lynden, Ferndale, and Bellingham). Glaciers can

be considered frozen reservoirs of water that have the ability to mediate yearly

fluctuations in runoff by providing water in warm and dry years and storing water in cool

and wet years. They also play an important role in seasonal drought buffering.

Therefore, the study of glacier runoff is of both theoretical and practical importance

(Fountain and Tangborn, 1985).

Thunder Creek is a glacially fed stream, near the mouth of which the U.S.

Geological Survey (USGS) has operated a continuous gauging station since October of

1930 (Zembrzuski et al., 2001). It is located in North Cascades National Park (NOCA)

and is a tributary to Diablo Reservoir on the Skagit River (Figure 1). The drainage area

upstream from the gauge is 271.9 square kilometers and has a mean elevation of 1,768

meters (Fountain and Tangborn, 1985). A recent glacier inventory of NOCA (Granshaw,

2002) found that the watershed contained 51 glaciers that occupy an area of 36.8 square

kilometers (approximately 13.4% of the basin). Thunder Creek’s glacial influence is

primarily from North and South Klawatti, McAllister, Inspiration, Forbidden and Boston

Glaciers, although contributions are made from smaller glaciers. NOCA has been

1

collecting detailed mass-balance data from North Klawatti glacier for the past eight years

(Riedel et al., 1999). Estimates of the glacial contribution to summer runoff (May-

September) in Thunder Creek based on this mass-balance data ranged from 23– 45% of

total summer runoff.

Hydrologic simulation (sometimes termed rainfall-runoff) modeling began in the

1950s and 1960s with the advent of the digital computer (Storck et al., 1998). The

purpose was to predict streamflow based on observed meteorological variables (notably

precipitation) at time scales short compared to catchment storm response times (usually

sub-daily). Applications of hydrologic simulation models include design and planning

(e.g., flood protection, hydroelectric power production), extension of streamflow records

as well as stream flow forecasting. The first models were spatially lumped, meaning that

they represented the effective response of an entire basin, without attempting to

characterize the spatial variability of the response explicitly (Storck et al., 1998). In other

words, lumped models average physical and vegetation characteristics over the entire

basin or portions of the basin. HSPF (Hydrologic Simulation Program-Fortran) is an

example of a lumped model and is widely used today by the USGS and other agencies.

The expansion of desktop computing power and the introduction of the Geographic

Information System (GIS) as a way to represent spatial data have paved the way for

distributed hydrology models like DHSVM to capture the spatial variability of a

watershed.

DHSVM is a physically based model that implements a digital elevation model

(DEM), and soil type, soil thickness, and vegetation GIS databases as basin parameters

(Figure 2). DHSVM models canopy interception, evaporation, transpiration, snow

accumulation and melt, subsurface flow and runoff generation based on basic

meteorological inputs of incoming solar radiation, air temperature, wind, humidity, and

precipitation. The model uses established hydrological relationships to perform water

mass balance simulations for each individual grid cell at a user defined time step. Digital

elevation data are used to model topographic controls on incoming solar radiation, air

temperature, precipitation and down slope water movement. Canopy evapotranspiration

is represented via a two-layer Penman-Monteith formulation that incorporates solar

radiation, and soil and vegetation characteristics (Wigmosta et al., 1994). Snow

2

accumulation and ablation are modeled using an energy balance approach that includes

the effects of local topography and vegetation cover. Both saturated and unsaturated

subsurface flow is also incorporated into the model using Darcy’s Law. Surface runoff is

generated in a cell when the input of throughfall and snowmelt exceeds the infiltration

capacity of the soil, or occurs on a saturated pixel, or when the water table rises above the

ground surface (Wigmosta and Perkins, 2001).

A detailed hydrological evaluation of the Thunder Creek basin can be achieved

using DHSVM, which could yield accurate assessments of the relationship between

glaciers and their meltwater streams. Such models can help to predict and quantify future

glacial contributions to streamflow if glaciers continue to retreat as they have during the

last century. This, in turn, will improve our ability to assess the responses of alpine

stream systems to changes in glacierization due to climate change.

3.0 BACKGROUND

3.1 Thunder Creek Watershed

To characterize stream discharge in Thunder Creek basin, it is important to

consider the influence of topography, geology, glaciers, soils, vegetation and climate in

the watershed. This is especially important when using DHSVM, since these are some of

the main variables that the model uses.

3.1.1 Topography

The upstream boundaries of the watershed are defined by topographic highs

(Eldorado, Forbidden and Boston Peaks, Mt. Logan, Mt. Arriva, and Buckner Mountain)

and the downstream boundary was chosen to be just upstream of where Thunder Creek

meets the high-water mark of Diablo Reservoir (Figure 3). Elevation ranges between

2770 meters (Mt. Logan) and 365 meters (Diablo Reservoir).

3.1.2 Geology

The bedrock in the Thunder Creek watershed is part of the crystalline core of the

North Cascades. Rocks consist of Paleozoic and Mesozoic sedimentary and volcanic

rocks which have been highly metamorphosed into schists and gneisses. These orogenic

3

rocks include mainly banded gneiss derived from the strata of the Chelan mountains

terrane, orthogneiss, Eldorado orthogneiss, and minor pegmatite (Haugerud, 1989). The

structure of the crystalline core is dominated by two north and northwest trending fault

systems, the Strait Creek and Ross Lake fault systems. Bedrock composition and

structure influence both soil development and topography of the watershed, two

important factors in watershed modeling.

3.1.3 Glaciers

During the last 800,000 years there have been nine major glaciations in western

North America (Imbrie et al, 1984). During the Pleistocene, alpine glaciers advanced and

joined the Cordilleran Ice Sheet, forming deep, U-shaped valleys. At higher elevations,

glacial activity eroded a complex system of arêtes, horns and cirques (Granshaw, 2002).

During the last 10,000 years, glacier advances occurred at 4.7 ka, 2.8-2.6 ka, and during

the last 700 years (Porter and Denton, 1967). The last period is known as the Little Ice

Age (LIA) (Grove, 1988). Glaciers have been retreating throughout NOCA since the

LIA ended in the late 19th century (Granshaw, 2002). Riedel (1987) identified eight major

periods of glacial recession in the Cascade Range during the last 700 years, the two most

recent of which occurred between 1820-1860 and 1880-1920.

3.1.4 Soils

There have not been any detailed soil surveys completed for the Thunder Creek

basin to date. The only data available for the basin are from general map that divides the

basin into three dominant soil texture classes, sandy-loam, loamy-sand and loam (Miller

and White, 1998).

However, NOCA has recently completed surficial geology mapping and

developed a hydrologic response model from that mapping for the Thunder Creek basin.

The hydrologic response model predicts the ability of landtypes to absorb precipitation

and snowpack melt.

4

3.1.5 Vegetation

Vegetation in the North Cascades is highly diverse, reflecting the spatial

variability in climate and topography in the region. Three broad vegetation zones are

represented in the Thunder Creek watershed including mixed conifer forest, subalpine

forest, and alpine forest. The first zone consists mainly of western hemlock, western red

cedar, Douglas fir and grand fir. The subalpine forest is dominated by mountain

hemlock, subalpine fir and Engelmann spruce (Franklin and Dyrness, 1973). The alpine

zone has been only been deglaciated since the LIA retreat and is being reoccupied by

stunted trees and alpine meadow vegetation including subalpine fir and heather.

3.1.6 Climate

The climate in the Thunder Creek watershed is considered a maritime climate

with mild winters of high precipitation and cool, dry summers. The location of the study

area is within the belt of prevailing westerly winds, which bring moist air in from the

Pacific Ocean. The Skagit River valley permits the moist maritime air to penetrate

further east than at other places along the Cascade crest (Porter, 1977). Precipitation

increases at high altitudes as the air cools and condenses as it rises over the North

Cascades. The Thunder Creek watershed has a major hydrologic divide (between

Thunder Creek and Fisher Creek) creating east-west variability in precipitation as well.

Snow accumulation of 8-10 meters near an altitude of 2,000 meters is common (Fountain

and Tangborn, 1985). Maximum precipitation occurs during the winter and the minimum

occurs in midsummer. Since most of the winter precipitation occurs as snow, maximum

runoff in the watershed is delayed until spring. The mean annual temperature at an

elevation of 1844 meters (South Cascade Glacier) in the North Cascades is ~ 2˚ C and the

mean July temperature is ~10˚ C (Porter, 1977).

3.2 Previous Work

3.2.1 Glacier Inventories and Mass Balance Estimates

The first comprehensive inventory of glaciers in the North Cascades was

completed by Post et al. (1971) and was based on aerial photographs taken in the late

1950’s. According to the inventory, the Thunder Creek watershed contained 51 glaciers

5

with a total area of 39.3 square kilometers (approximately 14.2% of the basin) making it

the most heavily glaciated basin in the North Cascades. Accumulation sources for the

glaciers in the basin include direct snowfall and minor amounts of snowdrift. North

Klawatti and McAllister glacier were classified as valley glaciers, and all others were

classified as cirque, slope/irregular topography, or small ice/snow patch (Post, 1971).

Post (1971) classified the termini of most glaciers to either be stationary or retreating

slightly. Tangborn (1980) concludes that, for the period of 1884-1974, the Thunder

Creek glaciers lost on average about 1.1 meters of water equivalent of ice per year.

However, most of that loss occurred between 1900 and 1940 (Tangborn, 1980).

Granshaw (2002) has recently completed another glacier inventory of NOCA

based on 1998 aerial photographs. According to this recent work, the watershed still had

51 glaciers that occupied an area of 36.8 square kilometers (approximately 13.4% of the

basin). Granshaw (2002) also digitized his work, as well as Post’s (1971) inventory, into

GIS datasets and digitized Post’s preliminary mapping of LIA glacier limits for a

majority of NOCA. These LIA glacier limits correspond to recent LIA moraine mapping

completed by NOCA. LIA extents for about 20% of the Thunder Creek basin are missing

from this database. I filled in the missing data by examining aerial photographs to map

LIA glacial features including moraines and trim lines.

The North Cascades National Park Service Glacier Monitoring Program began in

the spring of 1993 to develop an understanding of regional mass balance of glaciers.

Among the glaciers monitored are Noisy Creek, Silver, North Klawatti and Sandale

because they represent a large elevation range and are located at the headwaters of

watersheds with large hydroelectric operations. North Klawatti Glacier is in the Thunder

Creek watershed. It is a valley glacier located south of Primus Peak, with an area of 1.46

square kilometers, or about 4% of the total glacier cover in the basin. The glacier feeds

Klawatti Lake, which drains into Thunder Creek. Methods for the mass balance

investigations include at least three visits to each glacier per year to measure winter

accumulation and summer snowmelt. Measurements are taken at a series of points down

the centerline of each glacier then integrated across the entire glacier surface to determine

the annual mass balance for the glacier (Riedel, 1999). Results from the study of North

Klawatti Glacier have compared well with mass balance records from South Cascade

6

Glacier (~15 km south of Thunder Creek), which the USGS has monitored since the mid

1950s.

3.2.2 Glacier-Runoff Studies

The importance of studying glacier runoff has been recognized for a number of

decades. An enormous reserve of water is stored in the form of glacier ice: about three-

fourths of all the fresh water in the world, or equivalent to precipitation over the entire

globe for about 60 years (Meier, 1969). Glacial runoff studies in the North Cascades

(Tangborn et al., 1975; Tangborn, 1980) have attempted to refine techniques to

accurately predict glacial activity based on hydrological and meteorological records.

Tangborn et al. (1975) compared glacial mass balance calculations for South Cascade

Glacier by glaciological, mapping and hydrological methods. The hydrological method

proved to be unreliable, predicting a 38% greater mass loss than the other methods.

Tangborn (1980) developed a runoff-precipitation model to calculate glacier mass

balances for glaciers in the Thunder Creek basin. The model compared runoff between

Thunder Creek and South Fork Skykomish River (an unglacierized basin) and determined

the difference in runoff between the two basins was due to storage (or release) of water in

the glaciers of the Thunder Creek basin. This model shows general agreement with

observed trends in the glaciers; however it is highly simplified and does not consider

heterogeneities between the two basins (e.g., vegetation, topography and soils).

Meier (1969), Fountain and Tangborn (1985), Pelto (1991), and Krimmel (1992)

have also examined the effect of glaciers on runoff variations in the North Cascades by

comparing basins with and without glaciers. Their results indicated the presence of

glaciers has a profound impact on the seasonal timing, volume and quality of runoff in a

basin. Although these studies are important in showing the influence of glaciers on the

hydrology of a basin, they do not attempt to quantitatively predict runoff.

As a part of their glacier-monitoring program, NOCA has been using mass

balance measurements of North Klawatti glacier, glacier area and elevation to estimate

the contribution of glacial runoff in the Thunder Creek basin from 1993 to the present

(Riedel, 1999). NOCA’s estimates are based on assumptions that North Klawatti glacier

mass balance measurements are representative of all the glaciers in the basin and a melt

7

season of May through September each year. One limitation with the melt season

assumption is that glaciers actually begin to melt at different times from year to year and

at different elevations within a year.

Scientists have also attempted to predict seasonal runoff in glacierized basins by

specifically modeling snow and/or glacier melt (Rango et al., 1979; Martinec and Rango,

1986; Rango, 1991; Arnold, 1996). These models take advantage of daily cycles of

temperature and solar radiation and generally yield acceptable results of glacier melt.

However, they fall short of incorporating many important aspects of the hydrologic

influences of a basin (e.g., vegetation, soils, precipitation, etc.).

3.2.3 DHSVM Studies

I plan to use DHSVM to simulate hydrologic responses in the Thunder Creek

watershed because the model is one of the most advanced hydrologic models available

and has been validated in other Cascade mountain basins. Wigmosta et al. (1994)

originally developed DHSVM using the Middle Fork Flathead River in northwestern

Montana. The model was run at a 3-hour time step over 180-meter grid spacing.

Simulated results showed acceptable agreement (~ 2%) with recorded streamflow over a

four-year calibration and verification period. DHSVM has since been successfully

applied in some of the Pacific Northwest’s maritime and mountainous watersheds (e.g.,

Snoqualmie River, Little Naches River, Hard and Ware Creeks, Carnation Creek) (Storck

et al., 1995; 1998; Wigmosta and Perkins, 2001). These basins ranged in size between

5.2 – 2900 km2 and typically had high relief and were dominated by spring snowmelt and

rain-on-snow events.

To successfully simulate runoff in the Thunder Creek watershed it is important to

understand how DHSVM models snow/ice melt. DHSVM uses a two-layer energy and

mass balance approach to simulate snow accumulation and melt (Storck et al., 1995).

The energy balance components are used to model snowmelt, refreezing, and changes in

the snowpack heat content. Energy exchange between the atmosphere, forest canopy and

snowpack occurs to and from a surface layer. Energy exchange between the surface layer

and the pack layer occurs via melt water, which percolates from the surface layer into the

pack. The mass balance components represent snow accumulation/ ablation, changes in

8

snow water equivalent (depth of water that would result from the complete melting of the

snow in place (Dingman, 2002)), and water yield from the snowpack (Wigmiosta, et al.,

1994). Precipitation occurring below a threshold temperature is assumed to be snow.

During melting conditions the snow pack is assumed isothermal at 0° C (Storck et al.,

1995). DHSVM accounts for energy advected to the snowpack by rain as well as net

radiation, sensible heat, and latent heat. DHSVM is capable of simulating snow

accumulation and melt over hourly, daily, and seasonal time scales. It can be applied

where rain-on-snow is of concern and where spring snowmelt is dominant (Storck et al.,

1995).

Glaciers in DHSVM can be treated in a number of ways. In short-term

forecasting applications they are modeled as inexhaustible snowpacks (e.g., infinite

supply of water) that are static in aerial extent. In this application, DHSVM maintains a

snow-water equivalent of at least one meter in depth in all pixels that are specified as

glaciers. The main limitation of the inexhaustible approach is that in long term modeling

scenarios, glaciers will keep delivering water given the right melting conditions, but

never decrease in size. Glaciers can also be modeled by specifying an initial amount of

snow-water equivalent in each pixel of a glacier. The feature that maintains a snow-

water equivalent for a glacier (inexhaustible model) can be turned off in the DHSVM

source code, allowing the retreat or advance of glaciers to be simulated.

The major limitation of using DHSVM to model glaciers lies in the two-layer

representation of a snowpack. To get snowmelt in DHSVM the entire snowpack must

become isothermal and to get meltwater release, the liquid water content must increase

above saturation (Storck, pers. comm.). Both of these assumptions begin to deteriorate

when trying to model very deep snowpacks (e.g., glaciers). To address this limitation, I

will examine strategies for incorporating a better representation of glaciers into DHSVM

as modeling progresses.

4.0 PROPOSED RESEARCH

I propose to use DHSVM to examine the relationship between the streamflow and

short-term climate changes observed in the glacial activity in the Thunder Creek

watershed. Basin parameters will be represented in approximately 110,000 individual

9

50-meter grid cells for the watershed. I will calibrate the model to existing streamflow

data collected by the USGS and meteorological (solar radiation, air temperature wind,

humidity, and precipitation) data. I will use the model to estimate historical streamflow

in the basin during the Little Ice Age maximum in the North Cascades. I will also apply

the model to examine the impact of glacial retreat on streamflow. The main objectives of

the research are to develop a working hydrology model of the Thunder Creek watershed

that can be used for planning (hydroelectric power production), extending streamflow

records back to the Little-Ice Age, and forecasting streamflow.

5.0 METHODS AND TIMELINE

The methods used to complete this study will be completed with technical

assistance from the Puget Sound Regional Synthesis Model (PRISM) at the University of

Washington (Storck, pers. comm.). Thus far, assistance has been in gathering and

formatting the GIS data that DHSVM requires as described below. These steps were

completed using the GIS software ARC/INFO.

5.1 DHSVM Input Requirements

DHSVM requires five GIS grids (DEM, watershed boundary, soil type, soil

thickness and vegetation) a flow routing model and meteorological time-series data as

inputs. A grid resolution of 50 by 50 meters was chosen to optimize processing speed of

the model. DHSVM runs most efficiently when analyzing between 100,000 and 150,000

cells. A 50-meter grid resolution yields approximately 110,000 cells for the Thunder

Creek watershed.

1. DEM: The DEM is the fundamental foundation on which DHSVM and all of its

distributed parameters are based (Storck et al., 1995). The DEM for the study

area was compiled using eight USGS 7.5-minute, 10-meter DEM files. I merged,

filled, and converted them to 50 by 50 meter resolution using ARC/INFO

software. The DEM was then clipped to the watershed boundary (Figure 4).

2. Watershed Boundary: The watershed boundary is used in as the analysis domain

in DHSVM. It is also used as a mask to clip all other input grids to. This ensures

that all input grids contain an identical number of pixels. I delineated the

10

watershed boundary grid from 50-meter DEM using the WATERSHED command

in ARC/INFO (Figure 5).

3. Soil Type: I compiled the soil type grid from the Statsgo CONUS soils database

representing dominant soil texture class (Miller and White, 1998). This is a

countrywide soil grid with at 1000-meter resolution. I resampled the grid into 50-

meter pixels and clipped it to the watershed boundary (Figure 6 A). Because of

the coarse resolution of these data, I also used it in combination with a recent

surficial geology map (from NOCA) of the Thunder Creek basin to create a more

detailed soil type grid (Figure 6 B). DHSVM uses only the dominant soil type of

each grid cell. All grid cells with identical soil classifications are then assigned

one set of soil-dependant hydraulic parameters through a lookup table (Storck, et

al., 1995).

4. Soil Thickness: Soil thickness data do not exist for the watershed, so I created

this grid using an AML file (Arc-Macro-Language file that automates a number of

ARC/INFO commands) provided by PRISM that was written for use in the Skagit

River watershed. The AML uses a simple regression that made soil depths deeper

on shallow slopes and areas of high flow accumulation (Figure 7). This method

has provided acceptable results in basins similar to Thunder Creek. Minimum

and maximum soil depths were set at 0 and 3 meters respectively.

5. Vegetation: Vegetation data were compiled from a landcover database for NOCA.

The database is the result of a comprehensive inventory of the vegetation for the

park using Landsat Thematic Mapper (TM) satellite imagery and field data

collected in 1992. The data were in 25-meter resolution, so I resampled it into 50-

meter pixels and clipped it to the watershed boundary. I also reclassified the

vegetation classes in the dataset to the classification scheme used by DHSVM

with the help of Dr. David Wallin (Huxley College, WWU). Lakes and glaciers

are incorporated into the vegetation grid, so three vegetation grids were created

representing LIA, 1950s (based on Post, 1971) and 1998 (based on Granshaw,

2002) glacial conditions (Figures 8 A-C). Although the Little Ice Age refers the

time period (~1700-1850) where glaciers in the North Cascades were advancing, I

am modeling LIA runoff as if all the glaciers in the Thunder Creek basin were at

11

their LIA maximum at one time step. Actual vegetation characteristics between

each vegetation grid were assumed to be static and not changed. DHSVM uses

only the dominant vegetation type of each grid cell. All grid cells with identical

vegetation classifications are then assigned one set of vegetation-dependant

hydraulic parameters through a lookup table (Storck, et al., 1995).

6. Flow-routing Model: Flow in stream channels will be routed using be using a

flow routing model which represents a stream network by a series of cascading

linear reservoirs. This linear storage algorithm has provided satisfactory results to

a large range of basin sizes and topographic characteristics (Wigmosta and

Perkins, 2001). I created the flow network by running an AML file provided by

PRISM.

7. Meteorological Data: During the Summer of 2002, meteorological data will be

compiled from the Diablo Dam weather station, Thunder Basin Snowtel snow

survey site and Winter snow accumulation measurements made by NOCA on

North Klawatti Glacier. Using all three data sources will allow me to estimate

variations with topography due to the orographic effect in the basin. The west-

east variability in precipitation will be evaluated by also analyzing data from the

Ross Dam weather station, Swamp Creek and Rainy Pass Snowtel sites, and

Sandalee Glacier, east of the Thunder Creek watershed. Neither the Diablo Dam

weather station, nor the Thunder Basin Snowtel currently measures shortwave

radiation. I will need to develop a strategy to estimate shortwave radiation over

the basin based on radiation measurements from weather stations outside the

basin (Harts Pass, Park Creek Ridge and Marblemount) and albedo measurements

from South Cascade Glacier.

DHSVM can distribute point measurements of meteorological data over a basin in

a number of ways. Interpolation methods such as nearest neighbor, inverse

distance, and variable radius crestman are included as default choices in the

DHSVM source code (Storck, pers. comm.). DHSVM also has the capability to

use precipitation climatologies and temperature lapse rates to interpolate

12

precipitation and temperature vertically. Decisions on which methods to use will

be made as modeling progresses.

5.2 Modeling

Modeling will take place starting in the Fall of 2002. DHSVM is written in

ANSI-C and requires a Pentium 200 Mhz processor with 128 MB of RAM to run. I will

be running DHSVM on a SUN E450 in the Computer Science department via a

workstation in Dr. Robert Mitchell’s lab.

The model will be calibrated using historical streamflow and meteorological data

from a two-three year period 1997-2000. The calibration period was chosen because it

represents both negative and positive mass balance years for North Klawatti Glacier. I

will use DHSVM to asses the glacial contribution to stream flow in Thunder Creek by

comparing discharge from a glaciated and non glaciated representation of the basin. The

model will also be used to evaluate the hydrologic characteristics of the basin during the

Little Ice Age and to predict future runoff conditions based on potential glacier behavior

(e.g., continued glacier retreat).

5.3 Expected Outcomes

Final products from this project will include:

1. A working distributed hydrology soils-vegetation model of the Thunder Creek

watershed.

2. Comparison hydrographs for Thunder Creek with simulated and measured

streamflow.

3. Comparison of simulated and NOCA estimated glacial runoff contributions.

4. Simulated hydrographs from the period of LIA glacial maximum.

5. Simulated hydrographs for future streamflow based on current climate trends and

future climate predictions.

6. Analyses of the results and conclusions.

7. Assessment of a similar work on the Nooksack watersheds?

13

6.0 SIGNIFICANCE OF RESEARCH

As demand for water increases with increasing populations in the Northwest,

sophisticated hydrology models are needed to accurately examine water recourses. The

complexity of processes involved in the hydrologic cycle requires that hydrology models

incorporate details about precipitation, evapotraspiration and runoff in order to make

accurate predictions of future water resources needed for planning purposes. Humans in

the northwest use water for residential, agricultural, industrial, hydroelectrical, and

recreational purposes. Water is also needed to sustain delicate aquatic and terrestrial

ecosystems.

The influence of glaciers on all of the above make hydrologic modeling in the

North Cascades even more complex and important. A glacier is a frozen reservoir of

water, were potential runoff resides from a few hours to thousands of years. Since

glaciers can be indicators for climate change, they can be a link between climate change

and hydrology. If a sophisticated hydrology model can accurately characterize a glaciers’

impact on the hydrologic cycle, it can be used to characterize the impact of climate

change on they hydrology of a basin.

14

7.0 REFERENCES Arnold, N.S., Willis, I.C., Sharp, M.J., Richards, K.S., and Lawson, W.J., 1996, A distributed surface energy-balance model for a small valley glacier. I. Development and testing for Haut Glacier d’Arolla, Valais Switzerland, Journal of Glaciology, v. 42, n. 42, p. 77-88. Benn, D.I. and Evans, D.J.A., 1998, Glaciers and Glaciation, Arnold, London, 734 p. Dingman, S.L., 2002, Physical Hydrology, Prentice Hall, 646 p. Fountain, A.G. and Tangborn, W. V., 1985, The effect of glaciers on streamflow variations, Water Resources Research, v., 21, n. 4, p. 579-586. Franklin, J.F., and Dyrness, C.T., 1973, Natural Vegetation of Oregon and Washington, U.S. Forest Service, 452 p. Granshaw, F.D., 2002, Glacier Change in the North Cascades National Park Complex, Washington State, USA, 1958-1998, University of Portland Master’s Thesis, 104 p. Grove, J.M., 1988, The Little Ice Age, Methuen & Co. Ltd., London, 498 p. Haugerud, R.A., 1989, Geology of the Metamorphic Core of the North Cascades, in Geologic Guide book for Washington and other areas: Washington Division of Geology and Earth Resources Information Circular 86, Joseph, N.L., et al., editors, p. 119-127. Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L., and Shackleton, N.J., 1984, The orbital theory of Pleistocene climate: Support from revised chronology of the marine δ18 O record, in Berger, A.L., et al., eds., Milankovitch and climate, Part 1, D. Reidel Publishing, Dordrecht, pp. 269-305. Krimmel, R.M., Tangborn, W.V., Sikonia, W.G., Meier, M.F., Mayo, L.R., and Trabant, D.C., 1979, Glaciology, U.S. Geological Survey Professional Paper, Report: P 1150, p. 203-205. Krimmel, R.M., 1992, Runoff from two mountain basins in the North Cascades, Washington, U.S.A., AGU 1992 fall meeting, EOS, Transaction, American Geophysical Union, v. 73, n. 43, p. 180. Martinec, J. and Rango, A., 1986, Parameter values for snowmelt runoff modeling, Journal of Hydrology, v. 84, n. 3-4, p. 197-219. Meier, M.F., 1969, Glaciers and Water Supply, Journal American Water Works Association, v. 61, n. 1, p. 8-12.

15

Miller, D.A., and White, R.A., 1998, A Conterminous United States Multi-Layer Soil Characteristics Data Set for Regional Climate and Hydrology Modeling, Earth Interactions, vol. 2. (http://EarthInteractions.org) Nijssen, B., Lettenmaier, D.P., Wigmosta, M.S., and Perkins, W.A., 1996, Testing an imposed channel network algorithm for hydrograph prediction with a distributed hydrological model, AGU 1996 fall meeting, EOS Transactions, American Geophysical Union, vol. 77, no. 46, F232. (www.ce.washington.edu/~nijssen/docs/DHSVM/AGU.1996/agu1996.htm) Pelto, M.S., 1991, Glacier runoff into Baker Lake, Washington, AGU-MSA 1991 spring meeting, EOS, Transactions, American Geophysical Union, v. 72, n. 17, p. 130. Porter, S.C., and Denton, G.H., 1967, Chronology of the Neoglaciation in the North American Cordillera, American Journal of Science, v. 265, p. 177-210. Porter, S.C., 1977, Present and Past Glaciation Threshold in the Cascade Range, Washington, U.S.A.: Topographic and Climatic Controls, and Paleoclimatic Implications, Journal of Glaciology, v. 18, n. 78, p. 101-116. Post, A.P., Richardson, D., Tangborn, W.V., and Rosselot, F.L., 1971, Inventory of glaciers in the North Cascades, Washington, U.S. Geological Survey Professional Paper, 705 A, 26p. Rango, A., Hannaford, J.F., Hall, R.L., Rosenzweig, M., and Brown, A.J., 1979, Snow-covered area utilization in runoff forecasts, Journal of the Hydraulics Division, v. 105, n. HY 1, p. 53-66. Rango, A., 1988, Evaluating the effects of climate change on snowmelt hydrology of mountain basins, AGU 1988 fall meeting, EOS, Transactions, American Geophysical Union, v. 69, n. 44, p. 1203. Riedel, J.L., 1987, Chronology of Late Holocene Glacier Recessions in the Cascade Ranfe and Deposition of a Recent Esker in a Cirque Basin, North Cascade Range, Washington, University of Wisconsin-Madison Master’s Thesis, 91 p. Riedel, J.L., Fountain, A., and Krimmel, B., 1999, Glacier Monitoring Program: North Cascades, Progress Report. (www.nps.gov/noca/massbalance) Storck, P., Lettenmaier, D.P., Connelly, B.A., Cundy, T.W., 1995, Implications of forest practices on downstream flooding: Phase II Final Report, Washington Forest Protection Association, TFW-SH20-96-001, 100 p. Storck, P., Bowling, L., Wetherbee, P., and Lettenmaier, D., 1998, Application of a GIS-based hydrology model for prediction of forest harvest effects on peak streamflow in the Pacific Northwest, Hydrology Processes, v. 12, n. 6, p. 653-658.

16

Storck, P., 2002, personal communication. Tangborn, W.V., Krimmel, R.M., and Meier, M.F., 1975, A comparison of glacier mass balance by glaciological, hydrological and mapping methods, South Cascade Glacier, Washington, IAHS-AISH Publication, n. 104, p. 185-196. Tangborn, W.V., 1980, Two models for estimating climate-glacier relationships in the North Cascades, Washington, U.S.A., Journal of Glaciology, v. 25, n. 91, p. 3-21. Tangborn, W.V., Fountain, A.G. and Sikonia, W.G., 1990, Effect of area distribution with altitude on glacier mass balance; a comparison of North and South Klawatti glaciers, Washington State, U.S.A, Annals of Glaciology, v. 14, p. 278-282. Wigmosta, M.S., and Perkins, W.A., 2001, Simulating the effects of forest roads on watershed hydrology, American Geophysical Union, Water Science and Application Volume 2, p. 127-143. Wigmosta, M.S., Vail, L.W., and Lettenmaier, D.P., 1994, A distributed hydrology-vegetation model for complex terrain, Water Resources Research, v. 30, n. 6, p. 1665-1679. Zembrzuski Jr., T.J., Wiggins, W.D., Smith, R.R., Ruppert, G.P., Knowles, S.M., and Renslow, V.F., 2001, Water Resources Data, Washington, Water Year 1999, U. S. Geological Survey Water-Data Report WA-99-1, 508 p. Sources to Find Krimmel, R.M. and Meier, M.F., 1976, Measuring snow-covered area to predict reservoir inflow, U.S. Geological Survey Professional Paper, Report: P 929, ERTS-1, a new window on our planet, p. 173-175. Meier, M.F., 1986, Snow, ice and climate; their contribution to water supply, U.S. Geological Survey Water-Supply Paper, Report: W2300, p. 69-82. Tangborn, W.V., 1968, Mass Balances of some North Cascade glaciers as determined by hydrologic parameters, 1920-1965, Proceedings of the reports and discussions of the international commission o snow and ice, International Association of Scientific Hydrology, Geodesy ad Geophysics, n. 79, p. 267-274.

17

Other Sources * Cochra, O.D. and Harrison, W.D., 1994, The relationship between the hydrology and motion of Black Rapids Glacier during the 1993 melt-season, AGU 1994 fall meeting, EOS, Transactions, American Geophysical Union, v. 75, n. 44, p. 284. * Kunakhovitch, M.G. and Sokalskaya, A.M., 1997, The reaction of mountain glaciers to climatic change under continental conditions, Annals of Glaciology, v. 24, p. 415-420. March, R.S., 1996, Mass balance, meteorological, ice motion, surface altitude and runoff data at Gulkana Glacier, Alaska, 1992 balance year, U. S. Geological Survey Water Resources Investigations Report: WRI 95-4277, 32 p. March, R.S., 1997, Mass balance, meteorological, ice motion, surface altitude and runoff data at Gulkana Glacier, Alaska, 1993 balance year, U. S. Geological Survey Water Resources Investigations Report: WRI 96-4299, 30 p. March, R.S., 1998, Mass balance, meteorological, ice motion, surface altitude and runoff data at Gulkana Glacier, Alaska, 1994 balance year, U. S. Geological Survey Water Resources Investigations Report: WRI 97-4251, 31 p. Rango, A. and Martinec, J., 1979, Application of a snowmelt-runoff model using Ladsat data, Nordic Hydrology, v. 10, p. 225-238. Rango, A. and Martinec, J., 2000, Hydrological effects of a changed climate in humid and arid mountain regions, World Resource Review, v. 12, n. 3, p. 493-508. * Trabant, D.C., 1979, Alaska Glaciology Studies, The United States Geological Survey in Alaska, accomplishments during 1975, U. S. Geological Survey Circular, Report: C 0733, p. 45-47. * Young, G.J. and Ommanney, C.S.L., 1984, Canadian glacier hydrology and mass balances studies; a history of accomplishments and recommendations for future work, Geografiska Annaler, Series A: Physical Geography, v. 66, n. 3, p. 169-182. * indicated sources that I have

18

NOCA

Diablo Reservoir

Gauging Station

Highway 20

Watershed Boundary

N. Klawatti Glacier

Figure 1: Location map of the Thunder Creek Watershed located in North Cascades National Park

(northern and southern sections, dark green), in northwest Washington State.

19

Figure 2: Conceptualized diagram of the DHSVM structure. Linked one-dimensional moisture and energy

balance equations are solved independently for each model grid cell. Grid cells are allowed to exchange saturated subsurface flow with their eight adjacent neighbors. (Modified from Wigmosta et al., 1994;

Nijssen et al., 1996).

Figure 3: Shaded relief map showing the characteristically steep terrain of the Thunder Creek watershed.

20

Figure 4: The digital elevation model is used as the fundamental input grid in DHSVM.

Figure 5: The watershed boundary is delimited from the digital elevation model (Figure 4). It represents

110,807 50-meter grid cells that will be analyzed by DHSVM.

21

Figure 6A: Soil type from Miller and White (1998) of the Thunder Creek watershed. The resolution of

these data is considered very coarse (1000-meter) for modeling purposes of a basin this size.

Figure 6B: Soil type derived from NOCA surficial geology mapping of the Thunder Creek Watershed.

The data are similar to Miller and White (1998), however the smaller resolution allows for the incorporation of bedrock and sand classifications.

22

Figure 7: Soil thickness grid generated from an AML file provided by PRISM. Soil thickness was set to

vary from zero to three meters.

Figure 8A: Vegetation classifications of the Thunder Creek Watershed. LIA glacial extent of 22.5%

glacier coverage.

23

Figure 8B: Vegetation classifications of the Thunder Creek Watershed. Mid 1950’s glacial extent of

13.3% glacier coverage.

Figure 8C: Vegetation classifications of the Thunder Creek Watershed. 1998 glacial extent of 12.8%

glacier coverage.

24