Influence of lithology and climate on spring chemistry in

the Upper Deschutes River watershed, Oregon

Julia E. Schwarz

Senior Integrative Exercise

March 10, 2010

Submitted in partial fulfillment of the requirements for a

Bachelor of Arts degree from Carleton College, Northfield, Minnesota

Table of Contents Abstract Introduction . . . . . . 1 Quantifying weathering controls . . 1 Constraining groundwater age . . 3 Geologic Setting . . . . . 3 Methods . . . . . . 9 Results . . . . . . 11 Geologic Map . . . . 11 Chlorofluorocarbons(CFCs) . . . 11 Field Data . . . . . 14 Trace Element Data and Lab Analyses . 16 Discussion . . . . . . 21 pH. . . . . . . 21

Conductivity and Alkalinity. . . . 23 Spring Temperature. . . . . 24 Dissolved Oxygen Content (DO) . . 25 Dissolved Organic Carbon (DOC) . . 26 Trace Elements . . . . 26 Rare Earth Elements (REEs) . . 26 Other Trace Elements . . . 28 Fe, Al, Mn, Sm, and Nd . . 28 Lithium and Arsenic . . 29 Strontium . . . 30 Vanadium . . . 30 Groundwater Residence Time . . . 30 Weathering Rates . . . . 32 Trace elements and climatic influences . 33 Lithologic influences on trace elements . 33 Conclusions . . . . . . 35 Future Work . . . . . 36 Acknowledgements . . . . . 37 References Cited . . . . . 37

Influence of lithology and climate on spring chemistry in the Upper Deschutes River watershed, Oregon

Julia E. Schwarz

Senior Integrative Exercise

March 10, 2010

Advisors:

Bereket Haileab and Cameron Davidson, Carleton College Department of Geology ABSTRACT Fourteen springs in the Upper Deschutes River watershed, Oregon, were sampled to determine factors contributing to trace element distribution within the watershed. In particular, two areas were compared for differing climate and lithology; the Ochoco Mountains in the east had higher concentrations of trace elements compared to springs within 30 km of the Cascades Mountains crest. Specific trace elements (Fe, Al, Mn, Sm, Nd, Li, As, V) within these spring waters were examined for links with precipitation, air temperature, and catchment lithology. Rock composition plays a large role in element concentrations. For example, vanadium concentrations act as a tracer for certain rock units, particularly the youngest Quaternary basalts and Tertiary andesites. Bedrock age may also play a role in element concentrations, with older rocks in the Ochocos contributing to higher trace element concentrations. Throughout all springs, the weathering relationships between Fe, Al, Mn, Sm, and Nd are constant, though the weathering rate is higher in the Ochocos. High spring temperature and arsenic concentrations indicate that some springs are part of a deeper regional groundwater flow.

Keywords: Deschutes River, ground water, springs, geochemistry, chemical weathering, trace elements, rare earth elements

INTRODUCTION

Knowledge of silicate mineral chemical weathering rates is important to

understanding of individual watersheds and the global carbon cycle, because silicate

weathering creates bicarbonate from the uptake of atmospheric CO2 (Walling and Webb,

1992a; Bluth and Kump, 1994; Suchet et al., 2003; Velbel and Price, 2007). Within

watersheds, silicate rock weathering is essential to soil development and water quality

(White and Brantley, 1995), with silicate weathering contributing an estimated 60% of

major dissolved constituents (e.g. Ca2+

, Na+, Si) in rivers (Walling and Webb, 1992b).

Clearer understanding of silicate mineral weathering rates will help quantify its role in

the global CO2 budget and silicate mineral contributions to solute fluxes within

watersheds.

This paper examines silicate weathering fluxes from springs as part of a larger

project that aims to better understand the interplay between weathering reactions and

products in rocks, soil, and water in the Upper Deschutes River watershed in Oregon.

This study investigates weathering reactions on mineral and outcrop scale, soil formation

from rock and tephra deposits, and the role of regional rock chemistry on water chemistry.

The purpose of this paper is to explore factors affecting weathering rates of volcanic

rocks in aquifers by looking at the influence of climate, lithology, and residence time on

trace element concentrations in springs.

Quantifying weathering controls

The rate of mineral dissolution reactions is governed by thermodynamics,

kinetics and the exposure time of the water to the soil or rock body (Walling and Webb,

1

1992b). The reaction rate depends on the solute concentrations already in the water, and

the residence time determines if the reaction runs to equilibrium.

The rate of weathering is also determined by lithology (Walling and Webb, 1992b;

Bluth and Kump, 1994; Bowser and Jones, 2002), because varying mineral compositions

will lead to variable weathering rates. Spatial variation in weathering rates are found due

to a variety of climactic factors, including elevation, temperature (ambient and mean

annual), precipitation, and topographic features (Walling and Webb, 1992b; Bluth and

Kump, 1994).

One of the most reliable and straightforward ways of quantifying weathering

reactions in the field is a geochemical mass-balance model, sometimes referred to as

input-output or solute budgeting (Garrels and Mackenzie, 1967; Bricker and Jones, 2005;

Velbel and Price, 2007). Mass-balance model objectives include quantifying fluxes in

and out of the system, interpreting mineral weathering reactions, and determining

weathering rates of minerals (Bricker and Jones, 2005). Mass-balance models generally

use concentrations (this study) or mass flux of solutes (Bricker and Jones, 2005). Studies

based on concentration of elements in the system are useful in determining the

weathering reactions that are occurring, but they do not provide sufficient information to

determine quantitative weathering rates. In mass-balance models, a closed system is

assumed and all system inputs and outputs must be defined (White and Brantley, 1995;

Bricker and Jones, 2005; Velbel and Price, 2007).

System inputs include all processes that introduce minerals or elements into the

system. Precipitation is one of the major system inputs. Spring chemistry is affected by

2

precipitation chemistry, since groundwater is recharged by surface water. Further system

inputs include dissolution of minerals through weathering processes (Velbel and Price,

2007). The number of weathering reactions depends on the mineral assemblage and thus

the lithology (Bluth and Kump, 1994). Using local rock chemistry instead of idealized

chemistry is necessary to obtain more accurate results (Bowser and Jones, 2002; Bricker

and Jones, 2005). Rare earth elements (REEs) have been increasingly used in hydrologic

studies, and the source of dissolved REEs in ground water is generally assumed to be

weathering of the substrate rock (Garcia et al., 2007).

Constraining groundwater age

The residence time of water in the subsurface also plays an important role in

mineral weathering. Since the 1930’s, chlorofluorocarbons (CFCs) have been used in

refrigeration and other industries. CFCs are released into the atmosphere, where they last

from 50 to 100 years. Waters exposed to the atmosphere have CFC concentrations

corresponding to the year the water was last exposed (Phillips and Castro, 2005). CFCs

are useful for dating young groundwaters, though the method is insensitive to dispersion

and mixing of waters (Phillips and Castro, 2005). Since the groundwater in the Deschutes

area is known to be fairly young (Gannett et al., 2001), CFC dating is a useful tool to

determine the underground residence time of these groundwaters.

GEOLOGIC SETTING

The Deschutes River and tributaries drain an area of over 28,000 km2 (Gannett et

al., 2001) from central Oregon to the Oregon-Washington border. This area is drained

from the west by the Deschutes and Metolius Rivers, and from the east by the Crooked

3

River (Fig. 1). The watershed is bounded on the west by the Oregon Cascade Mountains

crest, on the south between the drainage divide with the Klamath Basin, and on the east

by the Ochoco Mountains (Fig. 1). The northern boundary of the Deschutes watershed is

its confluence with the Columbia River near The Dalles, Oregon. This study concentrates

on the area south of the Metolius-Crooked-Deschutes River confluence area.

The majority of water in the western part of the watershed is from precipitation

falling on the Cascade Range, the principal recharge area for the Deschutes River

(Gannett et al., 2001). Highly permeable igneous rocks of the high Cascades allow

precipitation to enter the subsurface and flow eastward through the basin (Gannett et al.,

2001). Much of this groundwater is released via springs into surface streams at the edge

of the La Pine basin, a structural graben filled with low permeability sediments (James et

al., 2000; Gannett et al., 2001). The La Pine basin is the main source of water into the

Deschutes River. The majority of the remaining groundwater is released into surface

streams near the confluence of the Metolius, Crooked, and Deschutes Rivers, where the

Green Ridge fault juxtaposes high permeability Quaternary and Tertiary volcanics to the

south with the low permeability devitrified John Day Formation to the north (Gannett et

al., 2001). In the eastern part of the study area, groundwater is not from the high

Cascades, but instead falls on the Ochoco Mountains.

In a recharge area, water is absorbed into the subsurface, and can flow

underground for periods of time on the scales of hours to years (Bethke and Johnson,

2008). Water flows down gradient until it flows into surface water, generally through a

spring. The flow of water underground is known as the flow path. Some common reasons

4

La P

ine

Basi

n

Quinn

TyeeBlacktail

Des

chut

esRiver

CASC

ADES

RAN

GE

CRES

T

OCHOCO MOUNTAINSMetolius

Rock

Whiskey Bandit

44º00'

43º30'

44º30'

122º 00' 121º00'

DeschutesBasinStudy

area

0 5 10 MILES0 5 10 KILOMETERS

Jack Creek

Tumalo Creek

Ranger

John

Fall

SnowCultus

Figure 1. Schematic map of the study area with spring sample sites marked as red circles. The Cascades Range crest borders the study area on the west, and is the recharge area for springs on the west side of the study area. For the purposes of this study, springs are separated into two main groups, the Ochoco springs (Rock, Whiskey, and Bandit Springs) and springs within 30 km of the Cascades crest. The La Pine Basin is a subsection of the Cascades springs area, and is shown on this map by a grey back-ground. Map modi�ed from Gannett et al. (2001).

N

5

for water to resurface are when the aquifer level is above the ground level, or when there

is a permeability change of the rocks in the subsurface (Gannett et al., 2001).

The drainage basin of the Deschutes River is primarily composed of basalt with

some andesites, dacites, and rhyolites (Fig. 2). The area has been a volcanic center for the

last 35 million years (Gannett et al., 2001), resulting in a complex assemblage of volcanic

deposits such as tephra, ignimbrites, and flows.

This study is primarily concerned with the chemistry of 14 springs sampled in the

headwaters of the Deschutes River. Eleven of these springs are within 30 km east of the

Oregon Cascades crest, and three (Bandit, Rock, and Whiskey Springs) are located in the

Ochoco Mountains, over 100 km from the Cascade crest.

Central Oregon is part of the Oregon high desert, which averages less than 5 cm

of rainfall per month, with June through August the driest months (Table 1) (NOAA,

2006). Precipitation is highest in the western half of the study area and lower in the

Ochocos. Average snowfall follows a similar trend, with Wickiup also receiving the

highest average annual snowfall (NOAA, 2006).The mean annual temperature for the

study area is 6.6 °C for the western half of the study area (Wickiup) and 6.3 °C for the

eastern half (Ochocos), averaged over a period of 1971-2000 (Table 2) (NOAA, 2006).

During the sampling period of July-August, the average temperature was about 18 °C,

with highs of 24 °C and lows of 7 °C (NOAA, 2006), and there was no precipitation.

Precipitation chemistry is sampled every month by the National Atmospheric Deposition

Program (NADP). Chemistry is averaged from 1983 to 2009 (Table 3). The average

precipitation pH is 5.31 across the study area.

6

50

1.0

0.5

2.0

4.0

Ts 4Ts 4

Ts 4

Qs

Qs

Tct

Tct

Tct

Tca

Tca

Qal

Tct

Tct Ks

Figure 2. Surface geology map of the study area, with spring sites marked as black dots. Geologic formations are grouped by age. Blues are Quaternary basalts, and are the major units for spring catchments within 30 km of the Cascades crest. Brown is Tertiary andesite and clastic rocks, and is the major unit for spring catchments in the Ochocos. Yellows and oranges are Quaternary sediments, pink is rhyolite. Map modi�ed from Walker and MacLeod (1991), Sherrod and Smith (2000), Gannett et al. (2001), and Lite and Gannett (2002).

121º00'

44º00'

43º30'

44º30'

122º00'

DeschutesBasinStudy

area

0 5 10 MILES

0 5 10 KILOMETERS

Tct Tertiary tu�

Ts4 Tertiary sediments

Ks Cretaceous sediments

Qb1

Qb2

Qb4

Qb3

Qb5

Quaternary basalts

Qs Quaternary sediments

Qg Quaternary granite

Qr Quaternary rhyolite

Qd Quaternary dacite

Qs1 Quaternary sediments

Other Quaternary rocks

Tb1 Tertiary basalt

Tca Tertiary andesite and clastic rocks

Tr4 Tertiary rhyolite

Tertiary and Cretaceous rocks

7

Jul

Aug

Sep

Oct

Nov

Dec

Jan

Feb

Mar

Apr

May

Jun

Yea

rly

(cm

)(c

m)

(cm

)(c

m)

(cm

)(c

m)

(cm

)(c

m)

(cm

)(c

m)

(cm

)(c

m)

(cm

)

Wic

kiup

2.03

2.11

2.03

3.38

7.92

8.59

8.86

6.58

5.08

3.33

2.97

2.57

55.4

7

Och

oco

2.08

2.08

2.16

2.92

5.56

5.21

5.28

4.11

3.45

2.82

3.28

2.46

41.4

0

Jul

Aug

Sep

Oct

Nov

Dec

Jan

Feb

Mar

Apr

May

Jun

Yea

rly

(°C

)(°

C)

(°C

)(°

C)

(°C

)(°

C)

(°C

)(°

C)

(°C

)(°

C)

(°C

)(°

C)

(°C

)

Wic

kiup

16.7

16.5

12.5

7.6

1.6

-2.0

-2.2

-0.4

2.0

5.0

9.0

13.1

6.6

Och

oco

16.4

16.7

12.2

7.3

0.9

-2.9

-3.3

-0.7

2.4

5.4

8.8

12.7

6.3

Ca

Mg

KN

aN

H4

NO

3C

lS

O4

pHC

ondu

ctiv

ityP

reci

pita

tion

(mg/

L)(m

g/L)

(mg/

L)(m

g/L)

(mg/

L)(m

g/L)

(mg/

L)(m

g/L)

(uS

/L)

(cm

)

Ann

ual A

vera

ge0.

0563

0.01

010.

0243

0.05

570.

0726

0.35

930.

0664

0.21

855.

3163

4.12

4624

.558

8

Sta

ndar

d D

evia

tion

0.01

910.

0052

0.02

260.

0431

0.02

730.

1160

0.03

720.

0762

0.10

130.

7384

6.65

40

Tabl

e 1.

Mon

thly

and

Yea

rly P

reci

pita

tion

Ave

rage

s fro

m N

OA

A, 1

971-

2000

Tabl

e 2.

Mon

thly

and

Yea

rly A

vera

ge T

empe

ratu

re, 1

971-

2000

from

NO

AA

Tabl

e 3.

Yea

rly P

reci

pita

tion

Che

mis

try a

t Silv

er L

ake,

OR

(198

3-20

09) f

rom

NA

DP

8

The higher elevations of the study area (mostly the western and southern regions)

have coniferous forests of predominantly ponderosa pine with some lodgepole pine.

Much of the area has been managed for timber production (Gannett et al., 2001). Lower

areas are dominated by juniper, with some sagebrush and grassland, though much of the

area is cultivated (Oregon Department of Forestry, 1997; Gannett et al., 2001). The

Ochocos to the east have predominantly lodgepole pine and juniper (Oregon Department

of Forestry, 1997).

The Deschutes headwaters are primarily wilderness areas. The human impacts are

relatively few, though many of the sample sites were located in close proximity to roads.

In the Ochocos, the main land use is farming and open range grazing.

METHODS

In June-August 2009, 51 streams and 14 springs in the headwaters of the

Deschutes, Crooked, and Metolius Rivers were sampled. Sampling locations were chosen

based on accessibility and location within the watershed.

Samples were tested for dissolved oxygen, conductivity, pH, and temperature.

Dissolved oxygen data was collected by a YSI 58 Lab/Field Dissolved Oxygen meter,

which was also used for water temperature. Conductivity was collected with a PCSTestr

35 conductivity probe. pH was recorded with a Thermo Orion 3 Star pH meter with a

glass Orion Ross combination pH electrode.

Samples were run on a Dionex ion chromatograph (ICS_2500) at the University

of Arizona, Tucson, AZ and ICP-MS analyses were run on a PerkinElmer/Sciex Elan

6100 DRC at Union College in Schenectady, NY. Water samples were analyzed for 60

trace elements. Samples were also tested for dissolved organic carbon (DOC), alkalinity,

9

and chlorofluorocarbons (CFCs). Samples for CFC dating were prepared and tested using

the method laid out by the USGS Reston Chlorofluorocarbon laboratory (Plummer and

Busenberg, 2009). Alkalinity samples were collected and tested within 36 hours of

collection using the Gran titration method (Gran et al., 1981). All samples except for

DOC and CFCs were filtered with a 0.45 micron nylon filter. IC, ICP-MS, and alkalinity

samples were collected in 30mL HDPE bottles. ICP-MS samples were prepared with 5

drops of nitric acid. DOC were collected in glass boston round bottles. HCl was added to

DOC bottles to inhibit growth and drive off the dissolved inorganic carbon component.

A USGS surface geologic map of Oregon (Fig. 2) (Walker and MacLeod, 1991;

Sherrod and Smith, 2000; Lite and Gannett, 2002) and a groundwater flow map (Lite and

Gannett, 2002) were used to approximate the different rock types within each spring

drainage area. There is no data on subsurface composition in the Deschutes headwaters

area, so this study uses surface lithology as a proxy for subsurface composition. Two

maps (Walker and MacLeod, 1991; Sherrod and Smith, 2000) were correlated to give

surface geology throughout the area. The two maps have different classifications and map

areas for geologic units, which have been simplified into one classification scheme from

Sherrod and Smith (2000). Walker and MacLeod (1991) separate Quaternary basalt into

four general categories: QTba, Quaternary basalt and basaltic andesite; Qyb, youngest

basalt and basaltic andesite; Qba, Quaternary basaltic andesite and andesite; and QTmv,

Quaternary mafic vent complexes. These roughly correlate to five age groups from

Sherrod and Smith (2000): Qb1, 0-12 k.y. (Qyb); Qb2, 12-25 k.y. (Qyb); Qb3, 25-120 k.y.

(QTba); Qb4, 120-780 k.y. (Qba, QTba); Qb5, 780 k.y.-2.m.y. (Qba, QTba). QTmv from

Walker and MacLeod (1991) is denoted in Sherrod and Smith (2000) as a vent symbol on

10

top of the surficial geology. The age range and area of QTba and Qba overlaps in the

Sherrod and Smith (2000) classification, and is grouped together for estimation purposes.

Precipitation data is compiled from the National Atmospheric Deposition

Program (NADP) from their collection site in Silver Lake, OR, and from the National

Oceanic and Atmospheric Administration (NOAA).

RESULTS

Geologic Map

Percentage estimates of surface geology for each spring catchment is given in

Table 4 from north to south, with the springs in the Ochocos last. Of the 11 springs

sampled within 30 km of the Cascades crest, the primary surface formation is Quaternary

basalt (Fig. 2). Tumalo Creek, Tyee, and Blacktail Springs are exceptions. For Tumalo

Creek Spring, the surface rock in the catchment is predominantly Quaternary rhyolite,

with some basaltic andesite. Tyee and Blacktail Springs are near outcrops of dacite,

andesite, and rhyolite (Fig. 2). Springs more than 100 km from the Cascades crest are

predominantly Tertiary andesite flows and clastic rocks from the Clarno Formation.

Chlorofluorocarbons (CFCs)

CFC results are given in years (Table 5), with the majority of samples between 21

and 27 years. Three samples were taken at each sample site. For each sample bottle, three

results are given, one for each of the main CFC compounds (CFC-12, CFC-11, CFC-113).

For all CFC sample sites, there is some variability in age within each type of CFC

compound, though all are within the error range of two years. With the exception of

Bandit Spring, the variability at each sample site across all age estimates is 4 years or less.

11

QT

mv†

Qb1

,Qb2

§Q

b3/Q

b4/Q

b5#

Qr*

*Tob††

Tca§

§Tr

##Tc

t***

Ks†††

(%)

(%)

(%)

(%)

(%)

(%)

(%)

(%)

(%)

Cas

cade

s cr

est

Jack

Cre

ek S

prin

g10

0

Met

oliu

s S

prin

g10

8010

Tum

alo

Cre

ek S

prin

g30

70

Tyee

Spr

ing

8020

Bla

ckta

il S

prin

g80

20

Qui

nn S

prin

g60

40

Sno

w S

prin

g20

80

Cul

tus

Riv

er S

prin

g20

80

Fall

Riv

er S

prin

g20

5030

John

Spr

ing

2080

Ran

ger C

reek

Spr

ing

100

Och

oco

Mou

ntai

ns

Roc

k S

prin

g90

10

Whi

skey

Spr

ing

9010

Ban

dit S

prin

g90

55

*Est

imat

ed p

erce

nt ig

nore

s al

luvi

al a

nd g

laci

al s

edim

ents

†Q

uate

rnary

and T

ert

iary

mafic v

ent

deposits

§ Q

uate

rnar

y ba

salt,

0-2

5 k.

y.

# Q

uate

rnar

y ba

salt,

25

k.y.

- 2

m.y

.

** Q

uate

rnar

y rh

yolit

e

†† T

ert

iary

oliv

ine b

asalt

§§ T

ertia

ry c

last

ic ro

cks

and

ande

site

## T

ertia

ry rh

yolit

e

***

Terti

ary

tuff,

pre

dom

inan

tly o

f the

Cla

rno

Form

atio

n

††† C

reta

ceous s

edim

enta

ry r

ocksTa

ble

4. E

stim

ated

per

cent

age

of s

urfa

ce ro

ck ty

pe b

y sp

ring

catc

hmen

t, no

rth to

sou

th*

12

Spring Name Sampling Date CFC12 error CFC11 error CFC113 error

(years) (years) (years)

Cultus River Sp. 7/1/2009 21 2 21 2 22 2

Cultus River Sp. 7/1/2009 17 2 23 2 22 2

Cultus River Sp. 7/1/2009 20 2 22 2 23 2

John Sp. 7/1/2009 27 2 27 2 25 2

John Sp. 7/1/2009 27 2 29 2 25 2

John Sp. 7/1/2009 25 2 24 2

Blacktail Sp. 7/2/2009 20 2 23 2 22 2

Blacktail Sp. 7/2/2009 19 2 23 2 21 2

Blacktail Sp. 7/2/2009 20 2 24 2 22 2

Tyee Creek 7/2/2009 21 2 24 2 22 2

Tyee Creek 7/2/2009 21 2 22 2

Tyee Creek 7/2/2009 20 2 23 2 22 2

Quinn Creek Sp. 7/2/2009 22 2 25 2 23 2

Quinn Creek Sp. 7/2/2009 21 2 22 2 22 2

Quinn Creek Sp. 7/2/2009 22 2 26 2 23 2

Snow Sp. 7/2/2009 25 2 28 2 24 2

Snow Sp. 7/2/2009 24 2 28 2 24 2

Snow Sp. 7/2/2009 23 2 24 2

Bandit Sp. 7/3/2009 15 2 30 2 24 2

Bandit Sp. 7/3/2009 15 2 29 2 24 2

Bandit Sp. 7/3/2009 16 2 31 2 24 2

Tumalo Sp. 7/5/2009 23 2 24 2

Tumalo Sp. 7/5/2009 23 2 25 2 24 2

Tumalo Sp. 7/5/2009 23 2 23 2

Metolius Sp. 7/5/2009 37 2 36 2 32 2

Metolius Sp. 7/5/2009 37 2 32 2 32 2

Metolius Sp. 7/5/2009 37 2 30 2 32 2

Jack Sp. 7/5/2009 25 2 17 2 24 2

Jack Sp. 7/5/2009 25 2 24 2

Jack Sp. 7/5/2009 25 2 27 2 25 2

Fall River Sp. 7/6/2009 28 2 28 2 25 2

Fall River Sp. 7/6/2009 28 2 25 2 26 2

Fall River Sp. 7/6/2009 28 2 24 2 26 2

Fall River Sp. August 8/11/2009 28 2 32 2 26 2

Fall River Sp. August 8/11/2009 28 2 32 2 25 2

Fall River Sp. August 8/11/2009 28 2 31 2 26 2

Supersaturated

Table 5. CFC-Derived Recharge Age (years before sampling date)

Supersaturated

Supersaturated

Supersaturated

Supersaturated

Supersaturated

13

Across the study area, CFC-measured ages are within a range of 6 years with the

exception of the Metolius Spring (Fig. 3). Though the Ochoco springs are only

represented by one CFC date, the age suggests a similar residence time to springs within

the La Pine basin. Fall River and John Springs show an age estimate above the average.

Figure 3. Spring water age determined by CFC dating, plotted from north (Bandit Spring) to south (John Spring). Metolius Spring water is the oldest at an average of 33 years; most other CFC ages range from 21 to 29 years. Fall River Spring shows some variability from June to August.

Field Data

Data collected in the field, such as temperature, conductivity, pH, and dissolved

oxygen content are given in Table 6. The average sampled water temperature in the

Deschutes headwaters is 6.0 °C for springs. Water temperature in the Ochocos and the

Metolius River was higher with an average of 12.0 °C.

Conductivity was an average of 190.2µS in the Ochoco springs, higher than the

average conductivity of 58.9µS in spring water near the Cascades crest. pH was neutral to

slightly basic, with a range from 6.4 to 7.8 with an average of 7.29 (Table 6).

10

15

20

25

30

35

40

Ye

ars

14

Sam

ple Nam

eS

ampling D

ateTim

eE

astingW

estingE

levationTem

p (DO

) C

onductivity D

OpH

Alkalinity

Dissolved O

rganic Carbon

UTM

(10T)U

TM (10T)

(m)

(°C)

(µS)

(%)

(mg/L)

Cultus R

iver Spring

7/1/2009m

orning596648

48543701382

3.865

95.87.41

0.625

John Lake Spring

7/1/20094:30pm

5950994841360

13914.6

56.598.9

7.240.592

0.875

Blacktail S

prings7/2/2009

10:20am596763

48755271699

2.837.5

95.96.64

0.3432.104

Tyee Creek S

prings7/2/2009

12:15pm597919

48768201734

2.628.2

96.96.43

0.3470.940

Quinn S

prings7/2/2009

afternoon597831

48723001553

3.544.4

947.16

0.3550.904

Snow

Springs

7/2/2009afternoon

5990484859246

15756.2

38.393.4

7.260.511

0.873

Rock S

pring*7/3/2009

lunch690694

49283631130

10.489

99.57.68

0.9524.665

Whiskey S

pring*7/3/2009

afternoon692163

49302621725

1936.6

907.32

0.4383.780

Bandit S

pring7/3/2009

5:45pm706819

49293411747

8.4445

73.77.37

5.0382.713

Tumalo C

reek Spring

7/5/200911:15am

6159024876645

14905.3

53.693.6

7.410.534

1.782

Head M

etolius Spring

7/5/20093:45pm

6084754920837

148610.2

123.196.9

7.791.235

1.483

Jack Creek S

pring7/5/2009

6:30pm601437

49256761001

4.761.9

109.97.49

0.6112.299

Ranger C

reek Spring

7/6/200911:10am

5901724827747

13485.5

54.2110.4

7.10.571

1.008

Fall River S

pring7/6/2009

2:30pm609996

48470151304

7.371.8

95.37.76

0.7781.761

*Sam

pled downstream

from spring

Table 6. Deschutes w

atershed spring field data, July 2009

15

pH was slightly basic in 12 out of 14 spring samples. Tyee and Blacktail Springs showed

a pH of less than 6.65. Dissolved oxygen content (DO) ranged from 73% to 110% (Table

6). Out of 14 spring samples, 11 have dissolved oxygen content between 90% and 100%.

The springs outside of this range are Bandit Spring with a low DO content, and Jack and

Ranger Creek, with higher DO content.

Trace Element Data and Lab Analyses

Dissolved Organic Carbon (DOC) ranged from .8 to 4.6 mg/L (Table 6). DOC

values were higher in the Ochocos than near the Cascades crest.

Spring samples were run for 60 trace elements (Table 7). Rare earth elements

(REE) were normalized to upper continental crust (UCC) values (Fig. 4) (Taylor and

McLennan, 1985; Garcia et al., 2007). Springs in the Ochocos have higher concentrations

of REEs than springs close to the Cascades crest (Fig. 4). All spring samples have higher

concentrations of heavy rare earth elements (HREE) compared to light rare earth

elements (LREE). In general, all elements had higher concentrations in the Ochocos.

Specific trace elements are addressed below.

With the exception of Bandit Spring, Fe concentrations are below 50 ppm (Table

7). When compared to values close to the Cascades crest, all values in the Ochocos are

high (Fig. 5). Al concentrations are below 10 ppm with the exception of John and Tyee

Springs near the Cascades crest and Rock and Whiskey Springs in the Ochocos (Fig. 6).

Bandit Spring has a concentration of less than 10 ppm (Fig. 6). Mn concentrations are

below .9 ppm with the exception of Whiskey Spring (9.987 ppm) and Tyee Spring (1.694

ppm) (Fig.6 , Table 7).

16

LiB

Al

Mn

FeN

iC

oC

uZn

As

Se

Rb

Sr

Mo

Cd

Ba

La

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

(ppm)

Cultus R

iver Spring

1.4180.752

0.9310.024

3.9340.056

0.0160.133

0.7900.085

bdl3.709

21.9910.197

bdl*1.160

bdl

John Spring

0.8521.061

19.4700.808

9.1320.037

0.0150.905

0.7940.061

0.0493.241

26.6130.172

0.0020.913

46.119

Blacktail S

pring1.662

1.5692.509

0.2044.142

0.049bdl*

1.3622.681

0.0730.021

3.60315.972

0.2450.023

0.8187.087

Tyee Creek S

prings15.780

4.57414.973

1.6954.347

0.4170.007

2.37516.896

0.481bdl

7.8266.516

0.8090.021

1.69434.362

Quinn S

prings15.258

10.0273.317

0.1450.346

0.0580.021

1.0461.991

3.0130.018

3.1877.376

0.6980.004

0.82019.448

Snow

Springs

2.4382.525

8.3420.472

5.5370.129

0.0010.423

4.3640.251

bdl2.715

11.6430.341

0.0040.859

16.146

Rock S

pring0.884

3.07519.537

1.07624.123

0.1410.013

0.7735.372

0.138bdl

1.32446.130

0.0450.002

16.514335.294

Whiskey S

pring0.993

1.96063.120

9.98844.863

0.2590.039

4.3444.685

0.1450.079

1.39825.043

0.0200.003

14.350983.148

Bandit S

pring0.863

12.3910.426

0.251124.591

0.2630.049

1.9516.456

0.0940.063

0.973476.977

0.2620.003

63.95832.396

Tumalo C

reek Spring

2.0941.759

1.038bdl*

2.6400.024

bdl*bdl

0.7400.239

0.0132.950

16.6560.145

bdl0.753

bdl

Metolius S

pring4.650

50.2894.325

0.1067.519

0.0580.015

8.1666.081

1.315bdl

4.75033.495

0.5470.003

2.1380.285

Jack Creek S

pring0.866

0.7081.611

0.0766.554

0.042bdl*

0.1180.734

0.061bdl

2.61428.028

0.2100.000

0.565bdl

Ranger C

reek Spring

2.3561.594

2.2020.087

4.7770.034

bdl*0.665

0.6620.071

bdl2.684

25.5110.154

0.0011.337

7.768

Fall River S

pring1.819

3.3073.986

0.2117.153

0.0670.001

7.2881.903

0.202bdl

2.22124.728

0.3220.002

1.4274.549

Fall River S

pring Aug

1.7763.017

1.9670.069

5.9220.020

0.0020.065

0.1880.193

0.0472.072

24.1460.326

-0.0011.351

bdl

*below detection lim

it

Table 7. Trace Elem

ents in Springs, D

eschutes Watershed

17

Pr

Nd

Sm

Eu

Gd

TbD

yH

oE

rTm

YbLu

WP

bU

Cr

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

Cul

tus

Riv

er S

prin

g0.

527

5.17

20.

038

0.28

88.

504

0.00

05.

584

19.8

7611

.660

bdl*

61.0

7532

.258

0.01

4bd

lbd

l3.

5977

John

Spr

ing

67.5

3477

.790

2.22

20.

998

87.4

6121

.461

111.

026

58.0

5911

6.78

3bd

l69

.474

13.8

710.

015

bdl

bdl

3.70

90

Bla

ckta

il S

prin

g14

.984

18.6

850.

543

0.28

128

.962

34.0

7369

.777

25.7

4934

.550

bdl

28.8

4321

.578

0.01

0bd

lbd

l0.

1342

Tyee

Cre

ek S

prin

gs75

.817

366.

108

2.42

61.

194

163.

961

31.2

8656

1.99

219

3.15

066

.589

bdl

638.

065

6.00

70.

075

bdl

bdl

0.57

72

Qui

nn S

prin

gs38

.383

56.2

311.

589

0.69

918

7.57

30.

000

39.4

41bd

l26

.415

bdl

60.7

3525

.398

0.04

2bd

lbd

l0.

8877

Sno

w S

prin

gs41

.278

232.

529

1.33

20.

923

64.1

339.

994

226.

236

56.5

9823

.023

bdl

135.

731

23.3

680.

035

bdl

bdl

1.70

78

Roc

k S

prin

g52

5.46

086

2.56

027

.571

9.12

611

91.8

1234

9.73

316

95.5

4691

7.23

917

33.6

54bd

l17

15.8

74bd

l0.

007

bdl

bdl

1.28

26

Whi

skey

Spr

ing

1446

.329

2482

.306

63.1

4417

.309

8352

.390

0.00

019

18.4

39bd

l14

56.3

14bd

l34

41.6

47bd

l0.

001

bdl

bdl

1.03

29

Ban

dit S

prin

g36

.270

62.9

8015

.499

13.4

3234

4.26

00.

000

83.6

40bd

l11

5.96

4bd

l27

5.75

2bd

l0.

001

bdl

0.06

82.

2366

Tum

alo

Cre

ek S

prin

g1.

518

3.17

40.

143

0.34

222

.446

0.00

051

.034

29.0

9320

.245

7.70

9bd

lbd

l0.

021

bdl

bdl

2.42

52

Met

oliu

s S

prin

g7.

062

22.4

490.

679

0.94

310

.893

9.85

954

.878

37.6

0012

.707

bdl

73.3

5527

.053

0.05

00.

122

0.05

79.

3744

Jack

Cre

ek S

prin

g5.

781

6.66

30.

262

0.15

742

.704

0.00

010

.179

bdl

13.9

2724

.542

21.5

3037

.033

0.01

5bd

lbd

l1.

5463

Ran

ger C

reek

Spr

ing

18.8

6274

.165

0.65

10.

559

34.4

8610

.871

127.

494

41.8

0211

.777

4.34

118

7.20

624

.322

0.01

6bd

lbd

l5.

2175

Fall

Riv

er S

prin

g10

.387

14.5

880.

580

0.33

713

.813

0.00

04.

108

18.9

8119

.825

bdl

23.7

1618

.246

0.02

2bd

lbd

l7.

3657

Fall

Riv

er S

prin

g A

ug1.

569

3.61

30.

039

0.30

147

.650

10.4

976.

901

33.1

888.

643

5.06

614

.831

1688

.264

0.02

0bd

lbd

l12

.537

7

*bel

ow d

etec

tion

limit

Tabl

e 7

(Con

tinue

d). T

race

Ele

men

ts in

Spr

ings

, Des

chut

es W

ater

shed

18

0.0

01

0.0

1

0.1 1

10

10

0

10

00

10

00

0

LaC

eP

rN

dSm

EuG

dTb

Dy

Ho

ErTm

YbLu

log(ppm/UCC)

Cu

ltus R

iver Sprin

g

Joh

n Sp

ring

Blacktail Sp

ring

Tyee Creek Sp

rings

Qu

inn

Sprin

gs

Sno

w Sp

rings

Tum

alo C

reek Sprin

g

Meto

lius Sp

ring

Jack Creek Sp

ring

Ran

ger Creek Sp

ring

Fall River Sp

ring

Fall River Sp

ring A

ugu

st

Ban

dit Sp

ring

Ro

ck Sprin

g

Wh

iskey Sprin

g

Figure 5. Plot of rare earth elements in spring sam

ples, normalized to upper continental crust values (Taylor and

McLennan, 1985). O

chocosprings are show

n with dotted lines. R

EE concentrations in the Ochocos are higher than in the

Cascades.All springs show

higher concentrations of heavy rare earth elements (H

REE), over light rare earth elem

ents (LR

EE), which reflects the high H

REE solubility in w

ater. Com

paring dysprosium (D

y) concentration to holmium

(Ho)

concentrations separate out two different types of w

eathering trends.

19

Figure 5. Iron concentration in springs, organized by sampling date. Rock, Whiskey, and Bandit Springs are in the Ochoco Mountains and show higher Fe concentrations than other springs. In particular, Bandit Spring shows a Fe concentration above 120 ppm.

0

20

40

60

80

100

120

140

Fe p

pm

Figure 6. Spring water concentrations of Al, Mn, and Sm, by sampling date. Springs in the Ochoco Mountains are Rock, Whiskey, and Bandit Springs. Bandit Spring has one of the highest alkalinities of any sampled spring, but a low Al and Sm concentration. Tyee and John Springs have moderate concentrations of Al. Most springs have a low concentration of Mn, but Whiskey Spring has a moderate concentration.

0

10

20

30

40

50

60

70

pp

m

Sm

Mn

Al

20

Metolius and Quinn springs both have above average Li and As concentrations

(Fig. 7, Table 7). Li concentrations are also above the average in Tyee Spring (Table 7).

The concentration of vanadium is above 2000 ppm in Tumalo Spring waters, and below

200 ppm in the Ochocos and Blacktail and Tyee Springs (Fig. 8). Sr concentrations are

below 50 ppm except at Bandit Spring, with a concentration of 477 ppm (Table 7).

DISCUSSION

pH

The average precipitation pH from the NADP Silver Lake station from 1983-2009

is 5.3 (Table 3), and the average sampled spring pH is 7.29. None of the sampled pH

values are less than the average precipitation pH, so there must be some basic addition

into the system, probably from silicate weathering.

The two springs with the lowest pH are Blacktail and Tyee Springs, which

correspond to a specific spring drainage lithology, with predominantly the youngest

Quaternary basalts (Qb1, Qb2) (Table 4, Table 6). One possible reason these springs may

have a particularly low pH is that the very young basalts are weathering at a slower rate

and thus contribute fewer elements to the subsurface water. Other spring catchment

lithology groups, such as John Lake Spring and Ranger Creek Spring, or Cultus River

Spring and Snow Spring, show similar pH to each other; this suggests that regardless of

the age of the rocks, the type of rock may control pH and dissolved content of waters.

21

0

0.5

1

1.5

2

2.5

3

3.5

As

pp

m

Figure 7. Arsenic concentrations in spring water, organized by sampling date.

Quinn and Metolius Springs have concentrations above 1 ppm, all other springs

have concentrations under 0.5 ppm. The Metolius Spring also have spring water

temperatures that are higher than the average annual air temperature. High arsenic

concentrations may indicate a deeper flow path.

0

500

1000

1500

2000

2500

V p

pm

Figure 8. Vanadium concentration in springs, organized by sampling date. Rock,

Whiskey, and Bandit Springs are in the Ochoco Mountains. These springs and

Tyee and Blacktail Springs also have low concentrations of Vanadium. Tumalo

Creek Spring has a concentration above 2000 ppm, the remaining springs have

concentrations between Tyee and Quinn Springs have concentrations between 800

ppm and 1700 ppm.

22

Conductivity and Alkalinity

In the 14 springs sampled, conductivity and bicarbonate (HCO3-) alkalinity are

strongly correlated (r2=.997). Conductivity is a measure of the solutes in water; the higher

the conductivity, the more dissolved elements. Alkalinity is the water’s ability to

neutralize acids. If the HCO3- alkalinity is high, there are more cations in the water to

balance the negative charge, and it follows that the conductivity will be higher.

The average precipitation chemistry given by the NADP estimates in Table 3

shows the average conductivity to be 4.12 µS, whereas the average for all sampled

springs is 58.5 µS. Thus, there must be some input into waters from the time they fall as

precipitation to the time they emerge from the springs. Bandit Spring is not included in

this average because it skews the average to the right; if Bandit Spring were included, the

input should be even greater.

Spring water temperatures in the western half of the study region are strongly

correlated to alkalinity (r2=.835) and conductivity (r

2= .733). This suggests that near the

Cascades crest, temperature can control the concentrations of dissolved elements in the

water (e.g. Walling and Webb, 1992a) because temperature controls the reaction rates of

the mineral dissolution. There is no correlation between temperature and alkalinity or

conductivity in the Ochocos. There are two statistical reasons for this lack of correlation.

First, it is likely that the small number of data points in the Ochocos contributes to its

lack of statistical significance. Second, while Bandit Spring shows the highest

conductivity and alkalinity by a significant margin, the water is not significantly warmer

than Rock Spring, also in the Ochocos, which has a much lower alkalinity. The highest

temperature is that of Whiskey Spring, with a low alkalinity. This temperature value is

23

misleading, since Whiskey Spring was sampled downstream from the spring and in full

sun, and the velocity and discharge were both small, allowing for a large temperature

increase from the spring to the sampling site. This spread of values means that the

Ochoco springs do not fit the trend from the western springs. Bandit Spring has a much

higher conductivity and alkalinity than any other sampled spring, and since the

temperature is not significantly higher than any other stream in the Ochocos, there is

probably some other local factor controlling the concentrations of elements in this spring.

Spring Temperature

Spring temperatures should approximate the local mean annual temperature

(Kathryn Szramek, pers. comm., October 29, 2009). This holds true for the La Pine basin

with a mean spring water temperature of 6 °C and a mean annual temperature of 6.6 °C

for Wickiup. In contrast, the mean temperature of springs in the Ochocos is 12.6 °C, and

the mean annual air temperature 6.3 °C. This may be influenced by the sampling site,

because Rock and Whiskey spring samples were taken downstream from the actual

springs and had relatively small discharge, so the water temperature could have increased

downstream of the spring. Bandit Spring’s temperature of 8.4 °C is still anomalously high,

but within a more reasonable range. The Metolius Spring temperature of 10.2 °C is also

high relative to the mean annual temperature.

James et al. (2000) suggest that water temperatures warmer than the mean annual

temperature in a region can indicate deeply circulating groundwater, which may absorb

more geothermal heat than shallower groundwater. They suggest the Metolius Spring as

an example of a geothermally warmed spring with an estimated heat slightly more than

the measured background heat flux in the central Oregon Cascades (Blackwell et al.,

24

1982; Ingebritsen et al., 1992; James et al., 2000). This study’s measured discharge is less

and the temperature difference greater, which indicates a smaller geothermal heat flux

(e.g. James et al., 2000).

Bandit Spring also shows a higher than expected temperature, though it is

unknown whether it is another example of regional groundwater flow. Given a shorter

residence time and smaller drainage area than the Metolius Spring, it is likely that the

high temperature of Bandit and other Ochoco springs are due to a different mechanism,

such as a shallow flow path. Water at shallow depths can be influenced by current surface

temperatures. Shallow ground water can be warmer than mean annual temperature

during the summer months if the flow is close to the surface.

Dissolved Oxygen Content (DO)

Dissolved oxygen solubility generally depends on temperature, pressure, and

elevation, since the partial pressure of O2 in the atmosphere changes with elevation

(Walling and Webb, 1992a). Sampled dissolved oxygen content is not correlated to

spring temperature and is moderately negatively correlated with elevation (r2=.468).

Bandit Spring in the Ochocos has a lower than average dissolved oxygen content.

Low dissolved oxygen can indicate microbes in the water, or contamination of water by

livestock waste (Judi Schwarz, pers. comm., March 4, 2010). Bandit spring has a high

water temperature compared to the average air temperature, indicating a shallow flow

path, which are more susceptible to contamination by surface sources. The Ochoco

Mountains are open rangeland for cattle, and some water with cattle manure may be

infiltrating the flow path for Bandit Spring.

25

Dissolved Organic Carbon (DOC)

Dissolved organic carbon content in water often controls the concentrations of

some elements in waters (Gaillardet et al., 2005). However, though springs that had high

alkalinity also had high DOC, there were no elements that directly correlated to DOC

concentrations. The source of dissolved inorganic carbon can be from weathering of

silicate rock, organic breakdown in soils, or pollution (Meybeck, 2003), so it is possible

that the lack of ties between DOC and concentrations in spring water are due less to rock

weathering and more to other inputs. Dissolved organic carbon contents are higher in the

Ochocos, which could be due to contamination, particularly because the flow path is

shallow.

Trace Elements

Rare Earth Elements (REEs)

All springs are heavy REE (HREE) enriched over light REEs (LREE) (Luucc/Ybucc)

compared to the initial rock ratio, which was also weakly HREE enriched. This

corresponds to trends of REE in river water—HREE are more soluble in water than

LREE (Gaillardet et al., 2005). Within the REE graph (Fig. 4), the Cascades and Ochocos

show two different trends for HREEs. Springs near the Cascades crest conform more

closely with the distribution of HREE in the upper continental crust. Spring waters in the

Ochocos are similar to each other but not as similar to the upper continental crust values.

Springs in the Ochocos have higher concentrations of REEs than springs close to

the Cascade crest (Fig. 4). In particular, Rock and Whiskey springs consistently have the

highest concentrations of REEs, with the exception of Lu, where they have anomalously

low values and Fall River Spring (August) has the highest. However, the concentration

26

for the Fall River Spring in August may be suspect since the July value is significantly

less. Alternatively, it could be due to seasonal variation. Tyee Spring is mostly closely

matched with the Cascades trend, though it also shows some characteristic shapes of the

Ochoco trend, particularly for the heavy REEs (HREEs). One element that seems to

differentiate two types of springs is Dysprosium (Dy). Many of the springs show about

equal concentrations of Ho and Dy, but a few (the springs in the Ochocos and Tyee

Spring) show a higher Dy and some a higher Ho concentration (Jack Creek Spring, Fall

River Spring (both), Cultus Spring). This may be due to a difference in the weathering

minerals.

Lithology may play a role in the higher concentrations and trends of HREEs in the

Ochocos. The Ochoco surface lithology is older and probably compositionally and

chemically different from the surface lithology in the Cascades crest area.

Tyee Spring water is not significantly different in composition from other springs

in the Cascades crest region, and is similar to Blacktail Spring both in location and

probable composition of bedrock. However, Blacktail Spring does not show the same

REE trend as Tyee Spring. Some maps suggest that the regional groundwater flow in this

area may come from two different directions, and depending on the exact flow path of

water emerging from these two springs, it is possible that Blacktail Springs is fed from

the high Cascades crest and Tyee Springs is fed largely off of Mt. Sheridan (Lite and

Gannett, 2002).

27

Other Trace Elements

The section below groups elements according to behavior in spring water, and

concentrates on several elements that are of particular interest. These elements are either

abundant in common rock-forming minerals or elements that are often tracers of rock

weathering. Li, Al, Fe, and Mn were chosen because of their relative abundance in

common rock-forming minerals. Sm and Nd concentrations are often indicative of

weathering, particularly when found in suspended sediments (Gaillardet et al., 2005). Sr

was chosen because it is also a classical tracer of rock dissolution, and because it is

highly mobile (Gaillardet et al., 2005). Arsenic is common in some magmatic systems

and is also highly mobile in river waters (Gaillardet et al., 2005). When silica

concentrations are available, the V/Si ratio is also indicative of rock weathering

(Gaillardet et al., 2005), so vanadium concentrations are evaluated even though there is

no available Si water data for this area.

Fe, Al, Mn, Sm, and Nd. Elements in this group are well correlated to each other, and

show higher concentrations in the Ochocos than closer to the Cascade crest.

When considered without Bandit Spring, Fe has a strong positive linear

association with water temperature (r2=.840), as well as with Mn (r

2=.814), Al (r

2=.871),

Sm (r2=.963), and Nd (r

2=.923). Fe has a moderate positive linear association with DOC

(r2=.576) and a moderate negative linear association with DO (r

2=.574). Fe is not well

correlated with Li, V, As, Sr, alkalinity, or pH.

Al has a strong positive linear correlation with Mn (r2=.925), Sm (r

2=.810) and

Nd (r2=.922), and a moderate positive linear correlation with water temperature (r

2=.603);

Al concentration is not correlated with DOC, pH, alkalinity, Li, V, As, or Sr.

28

When considered without Whiskey Spring, Mn concentrations have a moderate

positive correlation with Nd (r2=.571). Mn is strongly correlated to Al and Fe. No

correlation is shown between Mn and water temperature, DOC, pH, alkalinity, Li, V, As,

Sr, or Sm.

Sm and Nd are strongly correlated (r2=.929). Both elements have high

concentrations at Bandit and Rock Springs. They are strongly correlated to Fe and Al. Nd

is moderately correlated with Mn.

Elements abundantly found in common rock-forming or accessory minerals (e.g.

Fe, Al, Mn) are well correlated to each other. The strong correlation suggests that these

elements weather at constant rates relative to each other. Sm and Nd are correlated to Fe

and Al, suggesting that they are also tied up in similar rock weathering reactions. In

springs where the concentrations of these elements are high, there may be higher

concentrations or less water, but the weathering rate relationship between these three

elements stays constant. Five springs have anomalous concentrations for one or more of

these elements. These springs are Bandit Spring (high in Fe, low in Al), Whiskey Spring

(high Al, high Mn), and Rock, Tyee, and John Springs (high in Al). It is logical to

conclude that there is another factor controlling these elements’ concentrations, but Fe,

Al, and Mn concentrations correlate well in all springs but Bandit Spring. Since it is one

single point, it could be that the sample was contaminated in some way, or that local

factors contribute strongly to a locally high iron concentration.

Lithium and Arsenic. Li has a moderate positive linear correlation with As (r2=.526). Li

and As are not correlated with DOC, water temperature, alkalinity, Fe, Al, Mn, Sm, Nd,

V, or Sr. Arsenic is common in fluids within magmatic systems, and is found in high

29

concentrations in Quinn and Metolius Springs, suggesting that the flow path for these two

springs may be closer to some deeper magmatic source. This is consistent with the

previous theory postulated by James et al. (2000) that the Metolius Spring waters may be

part of a deeper basin flow (James et al., 2000; Gannett et al., 2001; Lite and Gannett,

2002).

Strontium. Sr is not well correlated with the minerals mentioned above. It is moderately

correlated with pH when considered without Bandit Spring, and strongly correlated with

alkalinity. It is not correlated with pH, water temperature, or DOC. The high strontium

concentration at Bandit Spring further suggests that factors other than weathering may

influence element concentrations at Bandit Spring.

Vanadium. Vanadium concentrations are not correlated to alkalinity, DOC, water

temperature, pH, or any of the minerals addressed above.

Groundwater Residence Time

With the exception of the Metolius Spring, all of the ages of water emerging from

sampled springs are between 20 and 28 years (Fig. 3). Waters from the Metolius Spring

are older than other spring waters, with an average residence time of 33 years, an outlier

when compared with other spring ages. However, site-specific ranges of estimates show

no outliers. Metolius Spring CFC age estimates are left skewed, suggesting that the actual

range of age estimates trends younger than the average suggests. The youngest spring

waters are in the La Pine Basin.

Previous studies suggest that the Metolius Spring water surfaces because of

underlying geologic structures in the area, particularly the Green River fault system,

30

which juxtaposes more permeable rocks next to less permeable rocks to the north (Fig. 2)

(Gannett et al., 2001; Lite and Gannett, 2002). Additionally, they suggest that the

recharge area for the Metolius is in the high Cascades, similar to springs draining into the

La Pine basin (Fig. 2). However, the Metolius River is farther to the north and the water

must travel farther and longer underground until it is discharged into the Metolius River.

Studies of oxygen isotopes support the high Cascades as the recharge area for the

Metolius Spring based on elevation (James et al., 2000). This hypothesis works in

conjunction with the older age estimate for the Metolius Spring (Fig. 3, Table 5).

When considered without the Ochocos, CFC age has a strong positive linear

correlation with water temperature (r2=.855). This corroborates the idea of regional

groundwater flow leading to warmer spring water temperatures for springs with longer

residence times.

A strong positive linear correlation between CFC age and alkalinity (r2=.817) for

spring waters with Cascade recharge suggests that waters flowing in the subsurface

accrue elements at a relatively constant rate. The Metolius Spring fits this linear trend

even though Metolius Spring water has a longer residence time, which suggests that

spring water in the region does not reach saturation.

Mean residence time for spring water in the Ochocos is similar to spring water

residence time in the La Pine basin. Based on topography, the recharge area for springs in

the Ochocos is not the same as for springs of the La Pine basin, and thus either the flow

path for Bandit Spring is similar in permeability and length to that of the La Pine basin,

or both the permeability and the length are different. The bedrock in the Ochocos is

mostly Tertiary and Cretaceous age, which differentiates it from the predominantly

31

Quaternary basalt and basaltic andesite closer to the Cascade crest. Additional

compounding factors are climatic differences between the two sides of the study area,

including temperature differences, less rainfall, and less discharge from springs in the

Ochocos, which may also influence the velocity of flow within the aquifer.

CFC age results for Bandit Spring have the highest variability between the three

CFC compounds, with an age difference of 15 years (Table 5), more than the 2 year CFC

dating error margin. No other spring shows this wide spread of ages between CFC

compounds. Many factors can influence the accuracy of CFC age estimates, such as the

recharge temperature and the proximity to urban areas (Phillips and Castro, 2005). One

factor that may influence the dates for Bandit Spring is microbial degradation. The

dissolved oxygen content in Bandit Spring is much lower than for any other spring, at

73%, which, while not low enough for anaerobic microbes, may suggest some microbial

growth that affects the water much more than at other springs. In springs with high

concentrations of microbes, CFC-12 is relatively stable, though CFC-11 and CFC-113

results appear older than the actual CFC age (Plummer and Busenberg, 2009). This

suggests that the recharge date for Bandit Spring is not 1986 as the CFC average suggests,

but more likely 1993 or 1994.

Weathering Rates

Assuming a shorter residence time combined with much higher concentrations of

trace elements, conductivity, and alkalinity suggest that the weathering rate in the

Ochocos is much higher than for springs to the west.

32

Trace elements and climatic influences

Data on the concentration of trace elements in precipitation is not available for

this area. Without precipitation concentrations of trace elements, we cannot know what

percentage of trace elements found in springs waters are from sources other than

weathering. Assuming that the precipitation chemistry throughout the basin is the same,

the sites that are most easily discussed are those with much higher or much lower than

average element concentrations. The three springs sampled in the Ochocos nearly always

have concentrations much higher than those springs nearer the Cascades crest.

Climatologic reasons for this could be different temperature, precipitation, or evaporation

conditions, or differences in biomass, or land use.

The average annual temperature difference between the Ochocos and Wickiup is

0.3°C, which does not account for the difference in concentrations of trace elements or in

water temperature. Spring temperature controls trace element concentrations for springs

within 30 km of the Cascades crest, but not for the Ochocos. The annual average

precipitation difference is 14 cm; the Ochocos are somewhat drier than Wickiup. The

annual average snowpack difference also follows this trend, with Wickiup receiving

much more snow. Less water in the system may mean less water overall enters the

groundwater system, allowing the amount of water that is there to become more saturated.

Lithologic influences on trace elements

The majority of rocks within the Deschutes Basin are Quaternary and Tertiary

basalts and andesites (Fig. 2). The actual ages of each formation are only somewhat well

constrained, and since the actual area of each spring catchment and the subsurface

33

geology are not known, the surface lithology is a rough approximation of the rocks within

the flowpath. Nevertheless, it is detailed enough to consider large scale trends in the data.

The vanadium concentrations do not correspond to other elements mentioned

above, but outlying data does correspond to different lithology types. Rock, Whiskey, and

Bandit Springs are all very low in V, and they have approximately the same percentage of

Tertiary age lithologies, with about 70% Tertiary clastic rocks and andesite flows (Table

4). Tumalo Creek Spring has the highest concentration of V, and it is the only spring

catchment that is made up predominantly of rhyolite (80% Qr). Tyee and Blacktail

Springs also have very low V concentrations, and these two springs have about the same

percentage of the youngest Quaternary basalts (80% Qb1, Qb2). pH also corresponds to

spring catchments with high percentages of Qb1and Qb2; For example, Tyee and

Blacktail Springs are the only two springs that have a pH of less than 6.65 (Table 4,

Table 6).

Most notably, the Ochocos, which consistently have higher concentrations of

trace elements compared to springs near the Cascades crest, have a different surface

geology from spring catchments near the Cascades crest. The Ochocos are made of

predominantly Tertiary rocks, whereas the western part of the study area is

predominantly Quaternary. In particular, the Ochocos are composed of Oligocene,

Eocene, and Paleocene clastic rocks and andesite flows, with some Cretaceous

sedimentary rocks and Middle Tertiary tuffs; the Cascades crest is nearly all Quaternary

basalt, basaltic andesite, Quaternary alluvium, and glacial deposits. This difference in age

and composition may be a major factor in the high concentrations of trace elements in the

Ochocos compared to the Cascades.

34

CONCLUSIONS

An increase in pH from 5.3 to 7.3 and increase in conductivity from 4.1µS to

86.0µS indicate that there is an input to the system between precipitation and discharge.

The main source of solutes is related to the release of elements during mineral weathering

within the watershed.

The concentrations of many trace elements are higher in the Ochoco springs than

in the headwaters of the Deschutes River. The difference in concentrations between the

two different hydrogeologic units may be due to a difference in lithology, weathering

rates, or to differences in water budgets, including less precipitation. The age or

composition of the lithology may contribute to the difference in weathering rates because

rocks in the Ochocos are older. Additionally, some areas are clastic rocks, which may be

more susceptible to weathering due to easier water infiltration.

The concentrations of Fe, Al, Mn, Sm, and Nd are related even in waters with

high concentrations of these elements, suggesting that the processes that control the input

of these elements are similar, and not limited by concentration.

Low vanadium concentration may act as a tracer for certain rock units,

particularly Qb1, Qb2, and Tca. This may indicate a lack of vanadium in these rocks, or

that elements containing vanadium weather at a slower rate in these areas. High pH may

indicate a spring catchment with predominantly Qb1and Qb2. Both the rocks and spring

waters are HREE enriched over LREEs, which follows solubility trends of REE in water.

CFC based residence times indicate that all springs except the Metolius River

Spring are recharged within 21-27 years, though the age for Bandit Spring may be

younger than the results suggest due to microbes in the water. The Metolius Spring is

35

recharged from the Cascades crest and travels further; the warmer temperature and high

arsenic concentration may be evidence for a deeper groundwater flow. Correlations

between CFC age dates and alkalinity for springs near the Cascades crest suggest that

waters pick up trace elements at a steady rate, never approaching equilibrium. Trace

element concentrations and CFC age dates suggest that the weathering rate is higher in

the Ochocos than near the Cascades crest.

Near the Cascades crest, spring water temperature is strongly correlated to

concentrations of dissolved elements in spring water. Air temperature does not explain

high spring water temperatures in the Ochocos. Instead, Ochoco spring temperatures are

indicative of a shallow flow path, and water temperature does not control the

concentration of dissolved elements.

Future Work

In the future, more detailed spring sampling coupled with a detailed petrologic

sampling campaign would help constrain the link between spring chemistry and lithology.

In particular, further sampling of rocks within the Ochocos is needed, as well as

examination of all outcrops for jointing and permeability to water. In addition to rock

sampling, springs should be sampled several times, over several days and throughout the

year. This would help to identify seasonal variation within springs as well as sampling

errors. Sampling a larger number of springs within the Ochocos would help to determine

if contamination of spring waters is local or widespread. Further studies should also

consider spring discharge, flow rate, and trace element concentrations in precipitation.

Additional information on major elements in spring water would help to determine the

weathering reactions that are occurring.

36

ACKNOWLEDGEMENTS

I would like to thank the Keck Geology Consortium for providing funding and the

opportunity to conduct this research, as well as Kati Szramek, Bereket Haileab, and Cam

Davidson for their invaluable input. Thanks also to the entire Carleton Geology

Department and to my family for their support throughout the process.

REFERENCES CITED

NADP/NTN Monthly Data Data: http://nadp.sws.uiuc.edu/nadpdata/monthlyRequest.asp?site=OR09 (last accessed December 20, 2009)

1997, Oregon Vegetation 1988: Oregon Department of Forestry. 2006, NOWData - NOAA Online Weather Data:

http://www.weather.gov/climate/xmacis.php?wfo=pdt (last accessed January 3, 2010) Bethke, C. M., and Johnson, T. M., 2008, Ground water age and ground water age dating: Annual

Review of Earth and Planetary Sciences, v. 36, p. 121-152. Blackwell, D. D., Bowen, R. G., Hull, D. A., Riccio, J., and Steele, J. L., 1982, Heat-Flow, Arc

Volcanism, and Subduction in Northern Oregon: Journal of Geophysical Research, v. 87, p. 8735-8754.

Bluth, G. J. S., and Kump, L. R., 1994, Lithologic and Climatologic Controls of River Chemistry: Geochimica Et Cosmochimica Acta, v. 58, p. 2341-2359.

Bowser, C. J., and Jones, B. F., 2002, Mineralogic controls on the composition of natural waters dominated by silicate hydrolysis: American Journal of Science, v. 302, p. 582-662.

Bricker, O. P., and Jones, B. F., 2005, Mass-balance Approach to Interpreting Weathering Reactions in Watershed Systems, in Drever, J. I., ed., Surface and Ground Water, Weathering, and Soils: Treatise on Geochemistry: Oxford, Elsevier, p. 119-132.

Gaillardet, J., Viers, J., and Dupre, B., 2005, Trace Elements in River Waters, in Drever, J. I., ed., Surface and Ground Water, Weathering, and Soils: Treatise on Geochemistry: Oxford, Elsevier, p. 225-273.

Gannett, M. W., Lite, K. E., Morgan, D. S., and Collins, C. A., 2001, Ground-Water Hydrology of the Upper Deschutes Basin, Oregon, U.S. Geological Survey: Water-Resources Investigations Report 00-4162.

Garcia, M. G., Lecomte, K. L., Pasquini, A. I., Formica, S. M., and Depetris, P. J., 2007, Sources of dissolved REE in mountainous streams draining granitic rocks, Sierras Pampeanas (Cordoba, Argentina): Geochimica Et Cosmochimica Acta, v. 71, p. 5355-5368.

Garrels, R. M., and Mackenzie, F. T., 1967, Origin of Chemical Compositions of Some Springs and Lakes: Advances in Chemistry Series, v. 67, p. 222-242.

Gran, G., Johansson, A., and Johansson, S., 1981, Authomatic titration by stepwise addition of equal volumes of titrant, part VII: potentiometric precipitation titrations: Analyst, v. 106, p. 1109-1118.

Ingebritsen, S. E., Sherrod, D. R., and Mariner, R. H., 1992, Rates and Patterns of Groundwater-Flow in the Cascade Range Volcanic Arc, and the Effect on Subsurface Temperatures: Journal of Geophysical Research-Solid Earth, v. 97, p. 4599-4627.

James, E. R., Manga, M., Rose, T. P., and Hudson, G. B., 2000, The use of temperature and the isotopes of O, H, C, and noble gases to determine the pattern and spatial extent of groundwater flow: Journal of Hydrology, v. 237, p. 100-112.

37

Lite, K. E., and Gannett, M. W., 2002, Geologic Framework of the Regional Ground-Water Flow System in the Upper Deschutes Basin, Oregon Report 02-4015: Portland, U.S. Geological Survey.

Meybeck, M., 2003, Global Occurence of Major Elements in Rivers, in Drever, J. I., ed., Surface and Ground Water, Weathering, and Soils: Treatise on Geochemistry: London, Elsevier.

Phillips, F. M., and Castro, M. C., 2005, Groundwater Dating and Residence-time Measurements, in Drever, J. I., ed., Surface and Ground Water, Weathering, and Soils: Treatise on Geochemistry: Oxford, Elsevier, p. 451-498.

Plummer, L. N., and Busenberg, E., 2009, USGS CFC Lab: http://water.usgs.gov/lab/chlorofluorocarbons/background/ (last accessed January 20, 2010)

Sherrod, D. R., and Smith, J. G., 2000, Geologic Map of Upper Eocene to Holocene Volcanic and Related Rocks of the Cascade Range, Oregon: U.S. Geological Survey.

Suchet, P. A., Probst, J. L., and Ludwig, W., 2003, Worldwide distribution of continental rock lithology: Implications for the atmospheric/soil CO2 uptake by continental weathering and alkalinity river transport to the oceans: Global Biogeochemical Cycles, v. 17, p. 1038-1051.

Taylor, S. R., and McLennan, S. M., 1985, The Continental Crust: Its Composition and Evolution: Oxford, Blackwell.

Velbel, M. A., and Price, J. R., 2007, Solute geochemical mass-balances and mineral weathering rates in small watersheds: Methodology, recent advances, and future directions: Applied Geochemistry, v. 22, p. 1682-1700.

Walker, G. W., and MacLeod, N. S., 1991, Geologic Map of Oregon: U.S. Geological Survey, scale 1:500,000.

Walling, D. E., and Webb, B. W., 1992a, Water Quality I: Physical Characteristics, in Petts, G. E., and Calow, P., eds., The Rivers Handbook: Hydrological and Ecological Principles: Oxford, Blackwell.

-, 1992b, Water Quality II: Chemical Characteristics, in Petts, G. E., and Calow, P., eds., The Rivers Handbook: Hydrological and Ecological Principles: Oxford, Blackwell.

White, A. F., and Brantley, S. L., 1995, Chemical weathering rates of silicate minerals: an overview, in White, A. F., and Brantley, S. L., eds., Reviews in Mineralogy and Geochemistry; Chemical Weathering Rates of Silicate Minerals.

38