University of Southern California Dissertations and Theses

Groundwater modeling and geochemical tracer (CFC-12 and tritium) distribution in the Abalone Cove landslide, Palos Verdes, California

(USC Thesis Other)

Groundwater modeling and geochemical tracer (CFC-12 and tritium) distribution in the Abalone Cove landslide, Palos Verdes, California

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content NOTE TO USERS
T his reproduction is the best copy available.
®
UMI
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
GROUNDWATER MODELING AND GEOCHEMICAL TRACER (CFC-12 AND
TRITIUM) DISTRIBUTION IN THE ABALONE COVE LANDSLIDE, PALOS
VERDES, CALIFORNIA
Copyright 2004
by
Erica Lynne DiFilippo
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(EARTH SCIENCES)
December 2004
Erica Lynne DiFilippo
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
U M I N um ber: 1 4 3 0 3 8 8
IN F O R M A T IO N T O U S E R S
T h e quality of this reproduction is d ep en d en t upon th e quality of th e copy
subm itted. B roken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard m argins, and im proper
alignm ent can ad versely affect reproduction.
In the unlikely ev e n t that the author did not send a co m p lete m anuscript
and th ere are m issing pages, th e s e will be noted. A lso, if unauthorized
copyright m aterial had to be rem oved, a note will indicate th e deletion.
®
UMI
U M I M icroform 1 4 3 0 3 8 8
C opyright 2 0 0 6 by P ro Q u est Inform ation and Learning C om pany.
All rights reserved. This m icroform edition is protected against
unauthorized copying und er Title 17, U nited S tates C ode.
P ro Q u est Inform ation and Learning C o m p an y
3 0 0 North Z e e b R oad
P .O . Box 134 6
A nn A rbor, M l 4 8 1 0 6 -1 3 4 6
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
ii
ACKNOWLEDGEMENTS
I would like to thank my thesis committee (Dr. Douglas Hammond, Dr.
Robert Douglas, Dr. Dennis Williams and Dr. Johnson Weh) for their guidance and
reviews of this thesis. I would like to give a very special thank you to Doug
Hammond for his endless support, knowledge and encouragement. Without Doug,
this experience would not have been as enriching or enjoyable.
This project could not have been accomplished without the hard work and
kindness of Dr. Jordan Clark and Dr. Dror Avisar at the University of California,
Santa Barbara and Roy Dunker and the entire staff of the Idaho State University
Environmental Monitoring Laboratory. I would like to thank them for allowing me
to run samples in their labs and their guidance on sampling and measurement
techniques. I would also like to thank Daphne Clark, Gerry Smith, and Rick and
Martha Schwartz for their help in collecting samples.
Financial assistance for this project was provided by the City of Rancho
Palos Verdes, the Department of Earth Sciences Graduate Student Research Fund,
the Gundaker Foundation Rotary District 7450, and the Sonosky Fellowship.
Lastly, but certainly not least, I would like to thank my family and friends for
their endless support and love. Whenever I was down or stressed, they helped me
focus on what was important and brought me back to sanity. Thank you for your
encouragement and love!
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
iii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS..................................................................................... ii
LIST OF FIGURES.................................................................................................. v
LIST OF TABLES.................................................................................................... viii
ABSTRACT.............................................................................................................. ix
CHAPTER I: INTRODUCTION............................................................................ 1
A. Purpose of Study......................................................................................... 2
B. Geologic Setting.......................................................................................... 2
C. Landslide History........................................................................................ 4
CHAPTER II: GROUNDWATER MODELING................................................. 9
A. Introduction.................................................................................................. 9
B. Steady State Model Parameters.................................................................. 14
C. Steady State Model and Model Calibration.............................................. 21
D. Transient Model........................................................................................... 33
E. MODPATH.................................................................................................. 44
F. Summary....................................................................................................... 49
CHAPTER III: GEOCHEMICAL TRACERS:
CHLOROFLUOROCARBONS................................................... 50
A. Introduction.................................................................................................. 50
B. CFCs as Hydrologic Tracers....................................................................... 51
C. CFC Sampling.............................................................................................. 54
D. CFC Analysis............................................................................................... 55
E. Application of CFC Dating to the Abalone Cove Landslide.................... 58
F. Summary of CFC Data................................................................................ 59
CHAPTER IV: GEOCHEMICAL TRACERS: TRITIUM.................................. 70
A. Introduction.................................................................................................. 70
B. Tritium as a Hydrologic Tracer.................................................................. 71
C. Tritium Analysis.......................................................................................... 73
D. Application of Tritium Dating to the Abalone Cove Landslide............... 75
E. Summary of Tritium Data............................................................................ 80
CHAPTER V: DISCUSSION AND SYNTHESIS............................................... 86
A. Introduction.................................................................................................. 86
B. Groundwater Mixing Model........................................................................ 87
C. Conclusions.................................................................................................. 99
D. Recommendations....................................................................................... 102
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
iv
REFERENCES........................................................................................................... 103
APPENDICES:
A. SAMPLNIG PROCEDURE FOR TRITIUM AND CFC WATER
SAMPLES.................................................................................................... 108
B. CFC SAMPLE ANALYSIS........................................................................ 122
C. CALCULATION OF CFC PARTIAL PRESSURE................................ 126
D. PROCEDURE FOR TRITIUM ENRICHMENT..................................... 130
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
V
LIST OF FIGURES
Figure 1 -1. Location map of the Palos Verdes peninsula and the Abalone Cove
Landslide............................................................................................... 3
Figure 1-2. Detail of the Abalone Cove landslide with the location of
groundwater wells................................................................................ 5
Figure 1-3. Cross section of the Abalone Cove landslide along line A-A' from
Figure 1-2.............................................................................................. 7
Figure 2-1. Planar view of the Abalone Cove landslide and ancient
landslide................................................................................................ 10
Figure 2-2. Digitized elevation map (DEM) of the land surface......................... 11
Figure 2-3. Reconstructed digitized elevation map (DEM) of the landslide
base....................................................................................................... 13
Figure 2-4. Conceptual model for groundwater flow in the Abalone Cove
landslide and ancient landslide........................................................... 15
Figure 2-5. Model grid constructed with Groundwater Vista............................. 16
Figure 2-6. Model calculated groundwater contours using a uniform hydraulic
conductivity (K) of 0.4 ft/day (0.12 m/day)....................................... 22
Figure 2-7. Schematic diagram of a slump landslide........................................... 24
Figure 2-8. Newly constructed hydraulic conductivity (K) zones...................... 25
Figure 2-9. Model calculated groundwater contours for the steady state
model..................................................................................................... 28
Figure 2-10. MODFLOW generated cross section along column 18 for steady
state model.......................................................................................... 29
Figure 2-11. MODFLOW calculated head vs. observed head calibration
curve........................................................................................................ 30
Figure 2-12. Cumulative rainfall from NOAA NCC station Palos Verdes
Est FC43D....................................................................................... 35
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
v i
Figure 2-13. Calibration curve used to calculate percolation as a function of
rainfall.............................................................................................. 37
Figure 2-14. MODFLOW generated cross-section along column 18 for the
2003 compared to the groundwater elevation generated for the
steady state model........................................................................... 43
Figure 2-15. MODPATH calculated particle flow paths and rates................... 46
Figure 2-16. MODPATH generated cross-section along column 18
displaying the vertical flow path of two particles........................ 47
Figure 3-1. Atmospheric CFC concentration through time............................... 52
Figure 3-2. Chromatograph for well WW-12 (9-03)......................................... 60
Figure 3-3. CFC partial pressures for each well per sampling date.................. 64
Figure 3-4. Mixing curves for February 2003 samples...................................... 66
Figure 3-5. Photograph of Kelvin Spring............................................................ 69
Figure 4-1. Reconstructed profile of tritium activity in precipitation plotted
with measured data from the Santa Maria, California GNIP
station........................................................................... 77
Figure 4-2. Reconstructed profile of tritium activity in precipitation at Santa
Maria, California............................................................................... 78
Figure 4-3. Tritium activity and precipitation vs. sampling date...................... 82
Figure 4-4. CFC-12 partial pressure vs. tritium activity of wells..................... 85
Figure 5-1. Schematic diagram of groundwater flow in the numerical model
illustrating boxes i-1 to i+1.............................................................. 89
Figure 5-2. Numerical model predicted CFC-12 partial pressures................... 94
Figure 5-3. Schematic diagram of hypothesized groundwater recharge in the
basin................................................................................................... 95
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
v ii
Figure A -l. Aluminum plates for CFC sampling............................................. 109
Figure A-2. Stainless steel clamps for CFC sampling...................................... I l l
Figure A-3. Photograph of the faucet assembly for CFC sampling................. 112
Figure A-4. Photograph of the complete assembly of aluminum plates,
stainless steel clamps and copper tubing for CFC sampling 114
Figure A-5. Sampler designed for non-pumping wells..................................... 115
Figure A-6. Schematic diagram of CFC sampling system for Kelvin
Spring................................................................................................ 119
Figure A-7. Schematic diagram of system used to transfer CFC samples
from ground glass syringes to copper tubing................................. 120
Figure B -l. Schematic diagram of the purge and trap system in Jordan
Clark’s laboratory at the University of California, Santa
Barbara............................................................................................... 123
Figure C-l. Calibration plot for standard CFC-12............................................ 127
Figure D-1. Diagram of tritium enrichment cell and electrode assembly 131
Figure D-2. Photograph of first distillation for tritium analysis...................... 132
Figure D-3. Diagram of vacuum distillation...................................................... 134
R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission.
v iii
LIST OF TABLES
Table 2-1. Average annual recharge for the Abalone Cove landslide 19
Table 2-2. Average annual recharge for the ancient landslide complex
excluding the Abalone Cove landslide (65% less than values
from Hill 2000)................................................................................ 20
Table 2-3. Hydraulic conductivity values obtained through calibration of
the steady state model..................................................................... 26
Table 2-4. Summary of model calculated inputs and outputs........................ 31
Table 2-5. MODFLOW calculated subsurface flow across the head scarp
of the Abalone Cove landslide........................................................ 33
Table 2-6. Calculated areal recharge for each stress period........................... 38-39
Table 2-7. Pumping rates for wells (ft3 /d)........................................................ 40
Table 2-8. Water budget for transient state model.......................................... 42
Table 3-1. Field measured temperature, conductivity, dissolved oxygen
and salinity....................................................................................... 57
Table 3-2. CFC-12 concentrations and apparent ages.................................... 61-62
Table 3-3. Percent change in allocthonous water in wells sampled in
February 2003.................................................................................. 65
Table 4-1. Tritium history reconstruction for Santa Maria, CA.................... 79
Table 4-2. Tritium activities for wells in the Abalone Cove and ancient
landslides....................................................................................... 81
Table 5-1. Transient state MODFLOW model calculated groundwater
elevations and flow rates................................................................ 91
Table 5-2. Estimated rate of input to uphill groundwater basin..................... 96
Table C -l. Sample loop volumes for CFC purge and trap system................ 126
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
ix
ABSTRACT
The Abalone Cove landslide occupies 80 acres of an ancient landslide
complex on the Palos Verdes peninsula. The influx of water into the slide mass is a
short-term catalyst for movement. The objective of this study is to constrain the
sources and flow rates of water entering and adjacent to the landslide complex.
MODPATH modeling indicates that water flows through the ancient
landslide complex in 22-28 years. MODFLOW models and measurements of
hydrologic tracers, CFC-12 and tritium, demonstrate that most of the water in the
complex is recent precipitation falling within the basin. The uphill groundwater
basin contributes minor flow to the ancient complex, with the majority of its water
flowing beneath the ancient slide plane. However, some communication between the
shallow and deep aquifers must occur. The deep circulating water may exert a strong
hydraulic pressure on the landslide slip surface, increasing the potential for
movement.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
1
CHAPTER I
INTRODUCTION
The Palos Verdes peninsula has been plagued with landslides for the past
500,000 years. A portion of this area, the Abalone Cove landslide, is currently
undergoing active movement. Water infiltration through the slide mass is a short
term catalyst for mass movement in the area (Merriam, 1960; Ehlig, 1992).
Therefore, it is important to determine the nature of ground water movement within
the slide. A few hydrologic studies have been performed in the area (Proffer, 1992;
Hill, 2000); however, these studies did not constrain the rate of recharge to the basin
or the source of water to the slide plane. Currently, it is unclear if slide movement is
triggered by shallow water infiltrating down and through the slide mass or deep
water rising from a deeper aquifer. Two potentially valuable approaches to
deciphering ground water movement in the slide plane are (1) construction of a
groundwater model and (2) the determination of ground water residence time (age)
using hydrologic tracers, such as tritium and chlorofluorocarbons (CFC-12). The
objective of this study is to constrain the source of recharge for the landslide. Water
that has percolated downward through the slide mass should be younger than water
that has percolated upward from below the slide mass. Therefore, the origin of the
water that triggers slide movement may be constrained by establishing the age of the
ground water throughout the slide mass. Currently, removing water from the slide
mass is the preferred remediation action. Constraining the source of water to the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2
slide plane will help determine the best location to place future dewatering wells in
the area.
A. Purpose of Study
This study attempts to construct a reasonable computer model of hydrologic
flow, as well as determine the residence time of water in the basin using geochemical
tracers. It begins with a geologic overview of the Palos Verdes peninsula and the
history of the active landslides. Next, the study presents computer groundwater
models for the Abalone Cove landslide, followed by measurements of CFCs and
tritium concentrations that were made to constrain groundwater residence times.
Finally, results of the three approaches are integrated into a description of the
hydrogeology of the basin.
B. Geologic Setting
Abalone Cove is located on the south side of the Palos Verdes peninsula,
southwest of the city of Los Angeles (Figure 1-1). The Palos Verdes peninsula is a
doubly plunging anticline with beds dipping seaward (Pipkin and Nash, 1967).
During the Pleistocene, the peninsula was an island, similar to the Channel Islands of
today (Pipkin and Nash, 1967). The peninsula emerged during the late Pleistocene
(Kiessling, 1963). Wave cut platforms (marine terraces), which range in elevation
from 100-1300 ft (30 - 400 m) above mean sea level (absl), formed around the island
(Kiessling, 1963). Current uplift of the peninsula is evident by deformation of the
youngest (lowest) of these marine terraces (Pipkin and Nash, 1967). Uplift occurs
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
3
Los Angeles
i m m
*------ ; M W H i l l . FiM ‘
3. ■ 8 5 V .'.* /? •; I % v < y * (•. ••,.....
• • • • « ■ - : - '■ * 7 * . V *
f W ? .v T ■ ' V
B M 2 f t ]
! a * > i
g.f\ - ; ^ z . i
***!' • 3 3 / ' - . 3 l;&'?S8fcv&«s^
': - - v i. 'Jfi '' . v.vJf .> > .# 3 'i
u i i»*rr «b»iu
LifhthO JM
Point Vicente
N A 4 I N E I A N O ' ■ i
Lone Point
Legend
- Approximate exent
of Abalone Cove
landslide
■ Approximate exent
of ancient landslide
Contour Interval 20 feet
2000 0 4000 Feet
1 Kilometer
U
Approximate exent
of Portuguese Bend
lanslide
Figure 1-1. Location map of the Palos Verdes pennisula and the Abalone Cove
Landslide. (Topographic map compiled from USGS Redondo Beach, California, San
Pedro, California, and Torrance, California 7.5 minute quadrangles (U.S. Geological
Survey, 1981).)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
4
along the Palos Verdes fault, a reverse fault that separates the peninsula from the Los
Angeles Basin (Keissling, 1963).
The bedrock of the peninsula consists of metamorphosed seafloor and marine
sediments. The Catalina schist is the oldest unit that outcrops on the peninsula.
Lying unconformably above the Catalina schist is the Altamira shale member of the
Monterey Formation (Pipkin and Nash, 1967). The Altamira shale is the oldest
exposed member of the Monterey. This unit is approximately 1000 ft (300 m) thick
on the peninsula and is interspersed with basalt sills (Pipkin and Nash, 1967). The
dominant lithofacies are cherty shale and silty sand; however, silty shale, tuff,
tuffaceous siltstone, dolostone and dolomitic sandstone are also common (Merriam,
1960; Proffer, 1992). Most of the pyroclastic tuffaceous units are extensively altered
to bentonite. Bentonite is an inherently weak clay and behaves plastically when
saturated with water. The most prominent tuffaceous layer in the Palos Verdes area
is the Portuguese Tuff. This tuff layer is fairly continuous with approximately 40 ft
(12 m) thick bentonite beds (Pipkin and Nash, 1967; Proffer, 1992).
C. Landslide History
Landslides have occurred sporadically on the Palos Verdes peninsula for the
past 120,000 years (Ehlig, 1992). The Abalone Cove Landslide occupies 80 acres of
an ancient landslide complex and did not undergo any historic movement prior to
present slide activation (Proffer, 1992) (Figure 1-2). In February 1974, the Abalone
Cove
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
5
118°23'
33°45‘ 30“
33° 44
118 23
X X F
\S 0 { I /1 — 1" —
U ^ B tf______
k es* / 'i • "*
> \ '••N . ^
L O S C Z K O
: /W ater
0 \ r^rtugues*
’ Riding
t x j k 0?wws
?f* , ^ M o n a h a n
■ 4J A R C
X
alone
V ‘ ‘ '''& k Ak'*
'P o r tu g u e s e 7 ■ ~
P o in t Hi)
In s p ira tio n
P o in t
118°21,30"
33? 45' 30“
33° 44'
118 21'30"
Contour Interval 20 Feet
1000
0.5
3000 Feet
1 Kilometer
| - Pumping well
A - Monitoring well
★ - Spring
Legend
- - - - Extent of Abalone Cove landslide
- — - Extent of ancient landslide
■ - - - - - Extent o f Portuguese Bend landslide
Figure 1-2. Detail of Abalone Cove Landslide with location of ground water wells.
A-A' indicates the location of the cross section represented in Figure 1-3. (Topographic
map compiled from USGS Redondo Beach, California, San Pedro, California, and
Torrance, California 7.5 minute quadrangles (U.S. Geological Survey, 1981).)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
6
Landslide was activated in response to heavy rainfall. In addition to increased
rainfall, previous studies suggest that the introduction of water into the slide mass by
domestic septic tanks and landscape irrigation are also responsible for mass
movement (Merriam, 1960; Proffer, 1992). The sliding in the Abalone Cove area
has occurred completely in the Altamira shale, at or near the base of the Portuguese
tuff (Proffer, 1992). Based on GPS surveys conducted by the City of Rancho Palos
Verdes and the Abalone Cove Landslide Abatement District, the slide exhibits small
creep (a few mm/yr) mostly in the areas near the beach (Robert Douglas, personal
communication).
Introduction of water into the old slide complex was the factor responsible
for slide reactivation (Merriam, 1960; Ehlig, 1992). Long-term agents responsible
for mass movement are wave erosion at the toe of the landslide, seaward dipping
beds and inherently weak bentonite layers (Merriam, 1960). In the Abalone Cove
area, dewatering of the slide mass has been used to reduce movement. In 1979,
homeowners in the area funded the installation of eight dewatering wells (Proffer,
1992). The Abalone Cove Landslide Hazard Abatement District (ACLAD)
maintains the dewatering wells. After installation of the dewatering system in March
1980, the slide mass slowed to a few millimeters a year (Robert Douglas, personal
communication) and it is presently moving in the form of creep (Proffer, 1992).
Currently, there are thirteen dewatering wells that pump an average of 300,000
gallons per day (gpd) from the active slide mass (Hill, 2000; Robert Douglas,
personal communication) (Figure 1-3).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Distance from shoreline (m)
-400 -200 1000 1200 1400 1600 1800 -600 200 400 600 800
1300
- - 400
1100 —
- - 350
1000
- - 300
9 0 0 - -
- - 250
7 0 0 - -
I
- - 200
( N U
C
o
500
- - 150 -o
m 4 0 0 - -
c n
100
300
o >
s Approximate ancient
landslide rupture plane
200
- - 5 0
Active
.rupture plane
100 - -
✓ -
-- 0
!T
-100 —
50
-200
6000 -2000 -1000 1000 2000 3000 4000 5000 0
Distance from shoreline (feet)
LEGEND
- Approximate
water table
elevation as o f
2003
I -W ell o f known
_ L depth
- Well o f unknown
cf depth
^ - Control points
(w ells where water
table elevation was
measured.)
Figure 1-3. Cross section of Abalone Cove landslide along line A-A' from Figure 1-2. (4x vertical exageration).
The water table elevation was approximated from the control points. The location of the landslide surface was estimated using
drillin logs taken from the Geologic/Geotechnical Data Technical Reviw peropared by Cotton, Shires and Associates (2001). -j
8
The ground water boundaries of the Abalone Cove landslide are
approximately the same as the ancient landslide complex (Proffer, 1992). In general,
ground water flow is southwest along the flank of the anticline toward the ocean
(Proffer, 1992). The graben and scarp area at the head of the slide consist of highly
fractured bedrock in a clay matrix (Proffer, 1992). This clay is largely material
washed in from uphill and deposited since the initiation of the slide movement
(Proffer, 1992). Therefore, permeability at the head of the slide has decreased over
time (Proffer, 1992). In addition, listric faults oriented east-west tend to provide a
trap for ground water and create a stepping down of the water table at the southern
end of the slide (Proffer, 1992). As the slide block has moved, fracturing within it
has increased permeability. Nearly vertical fractures that are oriented oblique to
slide movement have greatly increased horizontal permeability in the lower portion
of the Abalone Cove landslide (Proffer, 1992).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
9
CHAPTER II
GROUNDWATER MODELING
A. Introduction
Groundwater models are useful tools for evaluating the flow of groundwater.
Model simulations can be developed to illustrate observed data, such as groundwater
elevations and predict groundwater flow. The results of these simulations can
support the validity of budgets for water balances.
Previous studies by Proffer (1992) and Hill (2000) have constrained water
budgets for both the Abalone Cove Landslide (ACL) and the ancient slide mass.
However, both of these studies only calculated water budgets for the region and did
not constrain rate and direction of groundwater flow. The model constructed in this
study attempts to constrain flow rate and direction.
On a planar view, the boundaries of the groundwater basin are the landslide
scarps (Figure 2-1). The boundaries of the model were set to represent the edges of
the entire landslide region. Since the aquifer is an unconfined system, the upper
boundary of the model was set at the land surface elevation. A digitized elevation
map (DEM) was constructed for the top elevation using the program Didger (Figure
2-2). The slide plane is located in bentonite layers of the Altamira Shale. Bentonite
is a 2:1 smectite clay that is often used in the drilling of wells because it absorbs
water. Since bentonite can absorb large amounts of water and expand to fill pore
spaces, it is reasonable to assume that water will not flow vertically through the slide
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
10
,W aier_,V
T A n k
P o in t
Contour Interval 20 Feet
1000 0 3000 Feet
0.5__________________ 0______________________________________ 1 Kilometer
Legend
■ -Pum ping well
- - - . - Extent of Abalone Cove landslide
A - Monitoring well — — -E xtent of ancient landslide
★ -Spring
Figure 2-1. Planar view of the Abalone Cove landslide and ancient landslide.
The groundwater basin boundaries are defined by the boundary of the landslides. The
ancient landslide boundaries are based on Hill (2000). Current speculation places the
upper boundary of the the ancient landslide farther uphill in the community of Rolling
Hills (Robert Douglas, personal communication). (Topographic map compiled from
USGS Redondo Beach, California, San Pedro, California, and Torrance, California 7.5
minute quadrangles (U.S. Geological Survey, 1981).)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
8000 -
7000 -
680
6000 -
5000 -
4000 -
3000 -
2000 -
1000 -
7000 5000 6000 4000 2000 3000 1000 0
Figure 2-2. Digitized elevation map (DEM) of land surface.
(a) Topographic representation of land surface. The dashed lines define the boundaries
of the ancient and Abalone Cove landslides (long dashes) and the Portuguese bend
landslide (short dashes). The solid line defines the coastline.
(b) 3-D representation of landsurface. The boundaries are the same as (a).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
12
plane. Therefore, the slide plane was assumed to be completely within the bentonite
layer in order to simplify the model and provide an impermeable lower boundary.
The elevation of the slide plane was estimated using drilling logs taken from the
Geologic/Geotechnical Data Technical Review prepared by Cotton, Shires and
Associates (2001). Using the estimated depths of the landslide surface, the bottom
elevation for the entire basin was reconstructed. These estimated landslide
elevations were then placed into the program Didger in order to construct a digitized
elevation map (DEM) (Figure 2-3).
A groundwater model was constructed with MODFLOW 96 by the US
Geological Survey (USGS) using water budgets computed by Proffer (1992) and Hill
(2000). MODFLOW is a block-centered finite, difference model, meaning it uses
the center of the cell as the node location and it approximates the partial differential
for the governing equation (Equation 2-1):
d(^dh) df ^ dh) df^dh) v dh D !k _ .
— — I K x — — H — -— K y —— H — - — K z — — S s h R (Equation 2-1)
1 d
r
dh" d f
+ — K y
------
) dy
V
dz v
,Kx , -
dx{ d x) dy[ d y ) dz{ d z) dt
where h is the hydraulic head (length (L)), t is time (time (t)), K is the hydraulic
conductivity in the x, y, z direction (L/t), Ss is the specific storativity of the aquifer
(1/L), and R * is the volume of inflow per unit volume of the aquifer per unit time
(l/t) (Anderson and Woessner, 1992). For unconfined aquifers, MODFLOW follows
the Dupuit assumptions, presuming that: (1) hydraulic head does not change with
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
(a)
3^
7000-
380
6000-
5000-
$ s d
4000-
3000-
o
T O O
2000-
1000-
6000 7000 4000 5000 2000 3000 1000 0
13
(b)
1050
1000
950
900
850
600
750
700
650
600
550
500
450
400
350
300
■250
200
150
100
-50
0
- 5 0
100
Figure 2-3. Reconstructed digitized elevation map (DEM) of the landslide base.
(a) Topographic representation of landslide surface. The dashed lines define the ancient
and Abalone Cove landslides (long dashes) and the Portuguese Bend landslide
(short dashes). The solid line defines the coastline.
(b) 3-D representation of landslide base. The boundaries are the same as (a).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
depth (z)
14
dh \
— = 0 , and (2) flow lines are horizontal and equipotential lines are
dz
vertical (Anderson and Woessner, 1992).
MODFLOW is designed to read pre-written packages that contain
information regarding boundary conditions, aquifer parameters, inputs and outputs.
The program Groundwater Vistas (Student Version 3.40) created by Environmental
Simulations International (ESI) was used to generate the necessary packages for
MODFLOW (2000). Both the student version of Groundwater Vistas and
MODFLOW are available for free download from ESI’s website.
The first step for establishing a representative groundwater model for the
basin is developing a steady-state model. The water year used to establish this model
was 1976-77. Precipitation for this year (14.26 in (36.22 cm)) is close to the 100-
year mean (13.95 in (35.43 cm)). All input parameters for the model were taken
from the water budget established by Proffer (1992) for Abalone Cove and Hill
(2000) for the ancient landslide (Figure 2-4).
B. Steady State Model Parameters
A 50 x 50 cell grid was established for the groundwater basin, with cell
dimensions of x=138 ft (42 m) and y=165 ft (50 m) (Figure 2-5). Since both the
Abalone Cove and ancient landslides were modeled as unconfined aquifers, a one
layer model was used. The lateral and basal boundaries of the Abalone Cove and
ancient landslides were assumed to be impermeable and no flow was allowed to
cross these boundaries. Along the flanks of the basin, this assumption is reasonable
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
15
Surface
runoff
Abalone
Surface Cove
" “o ff head
western
fla nk
Surface f .
runoff_ _ '
eastern^.
flank * 4
Surface
runoff
Ancient
landslide
head
scarp
cipitation
I
Evapotranspiration
t
_Subsurface
inflow
Subsurface i
outflow ^
Subsurface
inflow
Figure 2-4. Conceptual model for groundwater flow in the Abalone Cove landslide
and ancient landslide. The black line defines the boundaries of the groundwater basin.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
16
^ f/f i7 > I
f \ \s.-.~ ( f t '(i ^ka & rr* re
,4 -^ \P^rtuauestf 8 fcS ?"' I /A\ '" -t^
y j J Ridine''CllSibinl I M ^ n • ,
r<— ~ = v * 8v I !i* '
\ \ " w t i r . m r *
t a r ~ 7 f " ■ a j r s ■ :x »
Legend
■ - Subsurface inflow and surface percolation
0 - Assigned groundwater elevation of 0 ft absl (Constant head boundary)
4" - Target
O - Abalone Cove landslide surface (Horizontal Flow Barrier)
- No flow boundary
Figure 2-5. Model grid constructed with Groundwater Vistas. (Topographic map
compiled from USGS Redondo Beach, California, San Pedro, California, and
Torrance, California 7.5 minute quadrangles (U.S. Geological Survey, 1981).)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
17
because the slide surface is steep, and it has been shown that the Portuguese Bend
landslide and the Abalone Cove landslide are not hydrologically connected (Proffer,
1992). Along the head scarp of the ancient slide complex, groundwater flow from
uphill must contribute some water to the groundwater basin. This contribution was
modeled using injection wells at the upper boundary. The inputs for these injection
wells were taken from subsurface inflow estimates by Hill (2000) for the ancient
landslide.
A constant head boundary was placed at the toe. This package denotes a
groundwater elevation for specified cells. The head along this boundary was set at 0
ft absl, presuming that groundwater elevations along the coast in an unconfined
aquifer should be at sea level. As groundwater flows seaward through the basin, it
encounters the slide plane for the Abalone Cove Landslide. Since this slide plane
rests in bentonite, it provides a barrier to groundwater flow. MODFLOW has a
horizontal flow barrier that was used to simulate this barrier. A low hydraulic
conductivity of 0.01 ft/d (0.003 m/d) was assigned to this wall function. This is
consistent with silt lithology. The barrier has a thickness of is 165 ft.
A hydraulic conductivity for Abalone Cove of 0.4 ft/d (0.12 m/d) was
calculated by Proffer (1992) using Darcy’s law (Equation 2-2) and her estimated
flow rate for the Abalone Cove landslide.
Q = - a k
\d x
(Equation 2-2)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
18
where Q is the flow rate in the aquifer (L /t), A is the area of the basin (L ), K is the
hydraulic conductivity (L/t), h is the hydraulic head (L) and x is the horizontal
distance (L). This hydraulic conductivity is consistent with sandy silt. Proffer
(1992) also estimated an effective porosity of 7%.
Surface inflow rapidly infiltrates after it enters the landslide area. Recharge
to the basin occurs through subsurface inflow, percolation of surface runoff, rainfall,
and delivered water (which includes water main losses, sewage and applied water for
landscape irrigation). Subsurface inflow into the basin was modeled using injection
wells (Table 2-1).
Surface percolation along the scarps and flanks of both the ancient and
Abalone Cove landslides was also modeled using injection wells. Proffer (1992)
demonstrated that most of the surface percolation in the Abalone Cove landslide
occurs along the head scarp and flanks: 97% along the head scarp, 2.5% along the
western flank and 0.5% along the eastern flank (Table 2-1). The total volume
injected was distributed evenly among the number of cells used to model the input.
Since rainfall and delivered water should generally provide widely distributed
c 'y
recharge into the basin (a total of 80 acres (3.2 x 10 m )), they were combined and
modeled as total areal recharge (Table 2-1). Rainfall percolation estimated by
Proffer (1992) is the net amount of percolation after consumption by low rooted
plants.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
Table 2-1. Average annual recharge for the Abalone Cove landslide
(Proffer, 1992).
Recharge Estimated Estimated No. of Model
Supply Supply
/ J\
(m )
Cells* Input per
(acre-ft) cell
Boundary Infiltration
from Surface Runoff
Head Scarp 2.33 2874 12 22 ft3 /d
West Flank 0.06 74 7 0.9 ft3 /d
East Flank 0.01 12 10 0.134 ft3 /d
Total 2.4 2960
Distributed Recharge
Rainfall 16.6 20476
Water Main Losses 1.7 2097
Sewage 9.2 11348
Landscape 6.2 7648
Total 33.7 41569 124 0.004ft/d
* The number of cells used to model the designated input. The total annual
distributed recharge was evenly distributed among all active cells in the
model.
In the ancient landslide, input parameters were estimated from Hill’s (2000)
budget, which was calculated during the 1997-98 water year. This was a year when
rainfall (26.83 in) was nearly twice the long term average. Rainfall percolation in
the Abalone Cove landslide estimated by Proffer (1992) is 35% of the percolation
estimated by Hill (2000) for the ancient landslide. Therefore, the recharge values
used in the model were estimated to be 65% less than the values estimated by Hill
(2000) (Table 2-2).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
Table 2-2. Average annual recharge for ancient landslide complex excluding
the Abalone Cove landslide (65% less than values from Hill, 2000).
Recharge Estimated Estimated No. of Model
Supply
(acre-ft)
Supply
(104 m3 )
Cells* Input per
cell
Boundary Infiltration
from Surface Runoff
Altamira Canyon 51.1 6.3 20 243.6 ft3 /d
Distributed Recharge
Rainfall 623.5 76.9 817 0.004 ft/d
Subsurface inflow 39.5 4.8 50 96.3 ft3 /d
* The number of cells used to model the designated input. The total annual
recharge was evenly distributed among the cells.
As previously mentioned, the steady state model for the basin was established
using the budget for the 1976-77 water year. The dewatering system in the area was
activated in 1980; therefore, the only outputs from the model system are subsurface
outflow and evapotranspiration. Proffer (1992) estimated an evapotranspiration rate
of deep rooted vegetation for the Abalone Cove landslide of 9.7 acre-ft/yr (3.6
cm/yr) using the Blaney-Criddle method. The Blaney-Criddle method is an
estimation of consumptive use related to the mean monthly temperature and
vegetation type. Hill (2000) used the Turc (1955) and Thomwaite (1948) methods
and estimated an evapotranspiration of 1403-1508 acre-ft/yr (5.48-5.09 cm/yr). The
Turc method relates evapotranspiration to global radiation, soil wetness and mean
daily temperature while the Thomwaite method calculates evapotranspiration using
only mean annual temperature. As previously mentioned, the values obtained from
Hill (2000) are from a year of higher than average rainfall. Since vegetation type
does not change dramatically from the Abalone Cove landslide to the lower portion
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
of the ancient landslide, evapotranspiration rates were assumed to be those calculated
by Proffer (1992) with an extinction depth of 15 feet (4.5 m). The upper portion of
the landslide is mostly grass covered; therefore, no evapotranspiration rate was
applied to this area.
C. Steady State Model and Model Calibration
Once the model parameters have been established, Groundwater Vista
generates the necessary packages and runs MODFLOW. Groundwater Vista
generates groundwater elevation contours based on the MODFLOW outputs. Using
the previously mentioned inputs, MODFLOW predicts extremely high groundwater
elevations (Figure 2-6). Groundwater elevations of 600 ft absl in an unconfined
aquifer when land surface elevation is only 150 ft absl are unreasonable.
In order to adjust the model to receive more reasonable results, target points
must be established. Targets are points at which the groundwater elevation is
known. For the steady state model five targets were used. Two target points are
located along the coast where groundwater elevation is 0 ft. Another target point is
located at Kelvin Spring. This spring is flowing all year round; therefore,
groundwater elevation is always at land surface elevation (668 ft absl). Two other
target points are located at two monitoring wells (LC-4 and MW-1) near the head
scarp of the ancient landslide. Groundwater elevations were measured in December
of 2002 and November 2003. Both of these wells are located up gradient of the
pumping wells and it is assumed that pumping does not have an effect on these
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
P o rtu g u e se
•Point
Inspiration
P o in t
_________________________Legend____________________________
■ - Cells with no groundwater
- Groundwater contours. Contour Interval = 20 ft
— — - Model boundary
| - Outline of area where groundwater elevation exceeds land
surface elevation
Figure 2-6. Model calculated groundwater contours using a uniform hydraulic
conductivity (K) of 0.4 ft/day (0.12 m/d). (Topographic map compiled from
USGS Redondo Beach, California, San Pedro, California, and Torrance, California
7.5 minute quadrangles (U.S. Geological Survey, 1981).)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 3
wells. Groundwater elevations were 673 ft absl and 738 ft absl for well LC-4 and
MW-1, respectively.
One poorly constrained model parameter is the hydraulic conductivity. The
groundwater basin is in a shale formation and 0.4 ft/d (0.12 m/d) appears reasonable
for shale. However, landslide mechanics have produced an extremely complicated
groundwater system. To assume that the hydraulic conductivity is the same
throughout the entire basin is unreasonable. There are areas that should be highly
fractured and areas that have been pulverized by movement of the slide mass.
Highly fractured areas should have much higher hydraulic conductivity values than
non-fractured areas. Pulverized areas should be very fine-grained and, hence, have a
very low hydraulic conductivity. Therefore, a possible source of error in the initial
run of the model is the hydraulic conductivity value.
Typical slump landslides usually consist of one major head scarp and several
minor scarps (Figure 2-7). Therefore, there can be several disassociated blocks
within the slide mass as a whole. Each of these blocks can have different
characteristics depending on the extent of their displacement. As previously
mentioned, dissimilar movement of these blocks can produce zones of different
hydraulic conductivity. Therefore, hydraulic conductivity zones were established for
the model to account for the movement of minor slumps throughout the basin (Figure
2-8). Each of these zones was assigned its own hydraulic conductivity (Table 2-3)
and Kx = Ky = Kz.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 4
Transverse
cracks
Radial
cracks
Figure 2-7. Schematic diagram of a slump landslide. From Kehew, 1998. p. 372.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 5
Portuguese 5 j\
^Point ^
Inspiration
Point
Legend
■
-K = 100 ft/day (30 m/d)
■
- K = 10 ft/d (3.0 m/d)
m
- K = 0.1 ft/d (0.3 m/d)
□
- K = 0.4 ft/d (0.12 m/d)
□
- K = 0.2 ft/d (0.06 m/d)
□
-K = 0.15 ft/d (0.05 m/d)
— -
- Model Boundary
Figure 2-8. Newly constructed hydraulic conductivity (K) zones. The hydraulic
conductivity values were estimated from model runs. (Topographic map compiled
from USGS Redondo Beach, California, San Pedro, California, and Torrance,
California 7.5 minute quadrangles (U.S. Geological Survey, 1981).)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 6
Table 2-3. Hydraulic conductivity values obtained through calibration of the
steady state model.
Hydraulic
Conductivity
Zone
Hydraulic
Conductivity (K)
(ft/d)
Hydraulic
Conductivity (K)
(m/d)
Comparable
Lithology
1 100 30 Coarse sand
2 10 3.0 Fine sand
3 1.0 0.3 Sandy silt
4 0.4 0.12 Sandy silt
5 0.2 0.06 Silt
6 0.15 0.05 Silt
The definition of zones in Table 2-3 was based on location. The Abalone
Cove landslide is one of the currently active portions of the ancient landslide
complex. This entire area has undergone prehistoric mass movement; therefore,
landslide mechanics changed the hydrology in this area prior to the reactivation of
the Abalone Cove landslide. Hydraulic conductivity values prior to reactivation
should reflect this movement. High hydraulic conductivity values were assumed
near the head scarp of both the ancient slide and Abalone Cove landslide, because
these areas contain fissures that act as a direct conduit for water to enter the
groundwater system. Hydraulic conductivity values and zone sizes were adjusted
until reasonable groundwater elevations were achieved. Directly below the head
scarp of the ancient slide, hydraulic conductivities had to be lowered to 0.15 ft/d
(0.04 m/d) and lowered to 0.2 ft/d (0.06 m/d) along the western flank in order to
obtain groundwater elevations below land surface. The hydraulic conductivity
estimated by Proffer (1992) is reasonable for the area directly above the Abalone
Cove landslide head scarp. This indicates that these areas have either (1) not
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 7
suffered much deformation from mass movement or (2) many of the fissures
produced by mass movement have been filled in with silty material. The Abalone
Cove landslide must have a high hydraulic conductivity, because the toe area of the
slide should have many transverse and radial fissures allowing rapid movement of
water in the subsurface. This is consistent with the highly fractured toe area
observed by Merriam (1960).
With the aforementioned hydraulic conductivity values, MODFLOW
calculates more reasonable groundwater elevations within the basin (Figure 2-9 and
2-10). Quantitatively, the model can be calibrated using the aforementioned targets.
MODFLOW calculates the hydraulic head for the targets and compares the
calculated head with the observed head (Figure 2-11). In an ideal case, a plot of
calculated head vs. observed head should show a 1:1 ratio. Groundwater Vista
calculates the residual standard deviation (RSD) between calculated and observed
heads along the line calculated head = observed head in the following manner:
where y(x) is the y value estimated from the straight line and n is the number of
calibration points. The RSD for the steady state model is 18.81. The percent of
model error can be calculated:
1/2
n — 1
(Equation 2-3)
E r r o r • 100% (Equation 2-4)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
_________________________ Legend__________________________
■ - Cells with no groundwater (inactive)
-----------Groundwater contours. Contour Interval = 20 ft
«■» • - Model Boundary
Figure 2-9. Model calculated groundwater contours for steady state model.
(Topographic map compiled from USGS Redondo Beach, California, San
Pedro, California, and Torrance, California 7.5 minute quadrangles (U.S. Geological
Survey, 1981).)
■ ;< ! ' * ' ]!
" v ,a ^ " V
P o r tu g u e s e ■ 3 ’ e . — u g u e s e £..*< M
* P o in t
Inspiration *
P o in t
% l
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 9
800 -
7 0 0 -
6 0 0 -
500 - E
d 4 0 0 -
b 300 -
100 -
-100 p ' l I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 1 I I I T I I I I I I I f T I I I
0 5 10 15 20 25 30 35 40 45 50
Row Cell Number
Legend
- Abalone Cove landslide head scarp
- Groundwater Table
- Outside of groundwater boundaries
Figure 2-10. MODFLOW generated cross-section along column 18 for Steady State
model.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Calculated H ead (ft)
3 0
Model Calculated Head vs. Observed Head
8 0 0
600
400
200
0
0 100 200 300 400 500 600 700 800
Observed Head (ft)
Figure 2-11. MODFLOW calculated head vs. observed head calibration curve.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
where ROH is range in observed head. The range in observed head for the model is
738 ft. Therefore, the percent error for the model is 2.5%. A model is generally
considered good if the error is 10% or less (Environmental Simulations, Inc., 1999;
Spitz and Moreno, 1996).
Another quantitative measure of the accuracy of the model is a comparison of
the model calculated input and output rates (Table 2-4). In a steady state system,
inputs equal outputs.
Table 2-4. Summary of model calculated inputs and outputs for the ancient
landslide complex including Abalone Cove.
Flux
(acre-ft/yr)
Flux
(104 m3 /yr)
Inputs
Subsurface inflow and surface runoff 147.9 18.2
Areal recharge 563.9 69.5
Total 711.8 87.7
Outputs
Subsurface outflow 700.9 86.5
Deep rooted evapotranspiration 9.6 1.2
Total 710.5 87.7
Inputs - Outputs 1.3 0
Percent Discrepancy 0.18 0.18
In general, a percent discrepancy of 1% is considered acceptable for
MODFLOW (Anderson and Woessner, 1992). MODFLOW calculates a total
outflow of 700 acre-ft/yr. This includes subsurface underflow and discharge to
surface at the toe where the groundwater table is above land elevation. Proffer
(1992) estimated subsurface groundwater outflow at the toe to be 1.7 acre-ft/yr for
the water year 1976-77 and discharge to surface to be approximately 40.3 acre-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 2
ft/year. One source of discrepancy between the model calculated outflows and
Proffer’s (1992) of hydraulic conductivity. Proffer (1992) estimated a hydraulic
conductivity of 0.4 fit/d for the landslide; however, this value is probably much
higher due to fracture flow created by landslide movement. An increase in Proffer’s
(1992) hydraulic conductivity would increase her estimated subsurface outflow.
Another parameter that can be calculated by the MODFLOW model is the
amount of subsurface underflow that occurs across the head scarp of the Abalone
Cove landslide. The rate of flow across the barrier can be calculated:
saturated thickness just upstream of the barrier (L), K' is the hydraulic conductivity
thickness of the barrier (L). The cumulative rate of subsurface flow across the head
scarp is calculated to be approximately 7.8 acre-ft/yr (Table 2-5); however, Proffer
(1992) estimated this subsurface flow to be 39.8 acre-ft/yr.
(Equation 2-5)
where Q is the flow rate across the barrier (L3 /t), dx is the cell width (L), z is the
of the barrier (L/t), Sh is the change in head across the barrier (L) and 6b' is the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
Table 2-5. MODFLOW calculated subsurface flow across the head scarp of the
Abalone Cove landslide.
Cell Upstream
head (ft)
Downstream
head (ft)
Flow
(ft3 /d)
Flow
(acre-ft/yr)
15 66 48 9.9 0.1
16 92 46 35.4 0.3
17 112 50 58.1 0.5
18 112 55 53.4 0.4
19 179 99 119.8 1.0
20 172 115 82.0 0.7
21 175 115 87.8 0.7
22 185 121 99.0 0.8
23 150 95 69.0 0.6
24 157 107 65.7 0.6
25 129 87 45.3 0.4
26 134 92 47.1 0.4
27 150 110 50.2 0.4
28 135 103 36.1 0.3
29 148 123 30.9 0.3
30 164 138 35.7 0.3
Total 925.4 7.8
*The thickness of the barrier is 165 ft and the hydraulic
conductivity is 0.01 ft/d.
D. Transient Model
Both qualitatively and quantitatively, the steady state model displays good
agreement with previous studies in the basin. However, the system is not in steady
state because dewatering of the slides has increased the outputs. Therefore, a
transient model was constructed to reproduce yearly fluctuations in groundwater
flow throughout the basin.
The transient model was constructed from the steady state model starting in
1976. The model was subjected to 28 stress periods, which are discrete intervals of
time in which aquifer parameters such as areal recharge and well pumping rates can
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 4
be changed. The length of each stress period was 365 days. Each stress period was
broken down into 12 time steps using a time step multiplier of 1.2. The time step
multiplier is a geometric progression (Equation 2-6).
SP = Y (* -2( l )^i (Equation 2-6)
where SP is the length of the stress period, i is the number of the time period and At/
is the length of the first time step. Therefore, the length of each time step increases
through the stress period. In this mode, the first time step is 9.22 days and the final
time step is 68.5 days.
One parameter that varies significantly each year is percolation from rainfall.
Rainfall data was obtained from the National Oceanographic and Atmospheric
Administration’s (NOAA) National Climatic Data Center (NCC) for station FC43D
(Palos Verdes Estates). Station FC43D is located at an elevation of 216 ft absl just
west of the Abalone Cove landslide. Since data from NOAA were only available
until 1996, data for the years 1996 - 2003 were obtained from the Los Angeles
County Department of Public Works for station 1216 (Rancho Palos Verdes). This
station is located at an elevation of 780 ft. A hydrograph plotting rainfall vs. time
reveals that precipitation varies dramatically from year to year (Figure 2-12). The
100-year mean is 13.95 inches (Proffer, 1992). Assuming a uniform rate of
evapotranspiration, the rainfall that percolates into the groundwater is dependent
upon the amount of precipitation. During years of high rainfall, a larger fraction of
the rainfall will percolate into the groundwater. This fraction should depend on
rainfall intensity and total precipitation. Proffer (1992) has taken these factors and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 5
Hydrograph for Palos Verdes Pennisula
30
Model
Siihulation
25
20
15
O O -lv fca rM ;an
10
5
0
1930 1940 1950 1960 1970 1980 1990 2000
Year
Figure 2-12. Annual rainfall from NOAANCC station Palos Verdes Est FC43D.
1996-2003 rainfall data from Los Angeles County Department of Public Works
Station 1216. The 100-year mean rainfall is 13.95 in (Proffer, 1992)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 6
estimated percolation for 1976-1982 (Figure 2-13). A quadratic formula was fit to
the data (Equation 2-7).
y = 13.495 -2.042x + 0.17275x2 R = 0.98 (Equation 2-7)
where x is rainfall (inches) and y is percolation (acre-fit). Using this relationship, the
percolation from rainfall was calculated for different years (Table 2-6).
Sewage, water main losses, and applied water are also components in total
areal recharge in the model. The amount of water discharged to the groundwater
from the septic tank system did not vary significantly during the six year study
period of Proffer (1992). Therefore, an average value for sewage discharge of 8.9
acre-ft/yr (10978 m3 /yr) was used. A uniform population density was assumed for
the entire basin. This assumption should be valid for 75% of the area. The area near
the head scarp of the ancient landslide does not have the same population density.
Therefore, areal recharge in this portion of the basin may be overestimated by the
model. The Abalone Cove area used septic tanks until the year 1998, at which time,
the community switched to an above ground sewer system. This change was made
in order to reduce the amount of water entering the subsurface. Since the septic tank
system went offline in 1998, sewage was not considered as an input after this point
in the model. Water main losses were estimated by documenting abnormally high
water usage for a particular month (Proffer, 1992). The community moved its water
mains above ground in 1980; therefore, any water main losses after 1980 would be
corrected before any significant amount of water entered the subsurface and not
considered an input after 1980. Applied water percolation was also taken as the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 7
Rainfall Percolation versus Percipitation
100
✓ — \
i
O
u
60
o
c
o
C S
o
o 40
lx
< u
a,
20
0 5 10 15 20 25 30
Rainfall (in)
Curve Fit values
Y = M0 + Ml*x + M2*x2
M0 13.495
Ml -2.042
M2 0.17275
R 0.98469
Figure 2-13. Calibration curve used to calculate percolation as a function of rainfall.
Data from Proffer (1992). The dashed lines represent the percent of total rainfall that
percolates into the groundwater basin.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 2-6. Calculated areal recharge for each stress period.
Year Rainfall
(in)
Rainfall
Percolation
(acre-ft)
Water
Main
Losses
(acre-ft)
Sewage
(acre-ft)
Applied
Water
(acre-ft)
Total Areal
Recharge
(acre-ft)
Model Input
(10-3 ft/d)
Model
Input
(in/yr)
1978 27.05 84.7 3.1 8.9 3.15 99.8 3.42 14.98
1979 15.27 22.6 2.4 8.9 3.15 37 1.27 5.56
1980 18.62 35.4 8.9 3.15 47.4 1.62 7.10
1981 10.36 10.9 8.9 3.15 22.9 0.79 3.46
1982 13.08 16.3 8.9 3.15 28.4 0.97 4.25
1983 28.78 97.8 8.9 3.15 109.9 3.76 16.47
1984 6.48 7.5 8.9 3.15 19.6 0.67 2.93
1985 9.23 9.4 8.9 3.15 21.4 0.73 3.20
1986 15.95 24.9 8.9 3.15 36.9 1.26 5.52
1987 18.6 8.9 3.15 30.7 1.05 4.60
1988 18.6 8.9 3.15 30.7 1.05 4.60
1989 4.63 7.7 8.9 3.15 19.8 0.68 2.98
1990 18.6 8.9 3.15 30.7 1.05 4.60
1991 15.588 23.6 8.9 3.15 35.7 1.22 5.34
1992 20.38 43.6 8.9 3.15 55.7 1.91 8.37
1993 22.24 53.5 8.9 3.15 65.6 2.25 9.86
1994 7.44 7.9 8.9 3.15 19.9 0.68 2.98
1995 24.96 70.2 8.9 3.15 82.2 2.82 12.35
Bold values indicate missing data. Where data was missing, the 100-year mean for rainfall (13.95 in)
was used.
U )
oo
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 2-5 (cont.). Calculated areal recharge for each stress period.
Year Rainfall
(in)
Rainfall
Percolation
(acre-ft)
Water
Main
Losses
(acre-ft)
Sewage
(acre-ft)
Applied
Water
(acre-ft)
Total Areal
Recharge
(acre-ft)
Model Input
(10-3 ft/d)
Model
Input
(in/yr)
1996 10.47 11.1 8.9 3.15 23.1 0.79 3.46
1997 13.58 17.6 8.9 3.15 29.7 1.02 4.47
1998 26.83 83.1 8.9 3.15 95.1 3.26 14.28
1999 9.54 9.7 3.15 12.9 0.44 1.93
2000 12.36 14.6 3.15 17.6 0.61 2.67
2002 4.19 8 3.15 11.1 0.38 1.66
2003 10.38 10.9 3.15 14.1 0.48 2.10
Bold values indicate missing data. Where data was missing, the 100-year mean for rainfall (13.95 in)
was used.
M 3
4 0
average value for the years of Proffer’s (1992) study (3.15 acre-ft/yr (3885 m3 /yr).
Together with rainfall percolation, the total areal recharge for the basin for each year
(stress period) was calculated (Table 2-6). This calculated areal recharge was
applied to both the Abalone Cove and ancient landslides.
As previously mentioned, in 1980 the Abalone Cove Landslide Abatement
District (ACLAD) established a dewatering system that removes water from the slide
mass. Pumping rates vary from well to well and from year to year. If data were not
available, pumping rates for each stress period were input into the model. Where
data were not available, the average value for pumping from that well for the years
that data exists was used. For wells where pumping data did not exist, an average
value for all of the pumping wells was used (Table 2-7).
Table 2-7. Pumping rate for wells (ft3 /d)
Well 1980* 1981* 1982* 1997T 1998t Average
Pumping
Rate
WW-1 610 591 307 136 241 337
WW-2 161 52 30 132 68 88
WW-3 563 847 250 114 91 373
WW-4 291 207 123 78 117 163
WW-6 107 139 71 44 77 88
WW-7 230 109 89 50 57 107
WW-8 63 40 37 35 35
WW-11 294 350 129
WW-12 2627 2965 1118
WW-13 1729 1574 661
Tolosckzo 925 912 367
Petak 597 833 286
Monahan 283 270 111
* - indicates data from Proffer (1992)
t - indicates data from Hill (2000)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
The hydraulic conductivity values obtained from the steady state model were
considered to be constant throughout the transient model. Most of the movement in
the groundwater basin has occurred in the Abalone Cove landslide, as opposed to the
ancient landslide. Therefore, hydraulic conductivity values in the ancient landslide
should not change during the duration of the transient state model. Since the
Abalone Cove landslide is creeping slowly (a few millimeters per year), the
hydraulic conductivity in the landslide should not change significantly through the
course of the transient state model.
By varying the recharge and pumping rates for each stress period, an image
of groundwater elevations through time can be obtained by running the transient
state model. The water budget for each year of the model is presented in Table 2-8.
Looking at the groundwater elevations for the model simulated year 2003, one can
see the impact that pumping has had on the basin (Figure 2-14). MODFLOW
predicts that groundwater elevations near the large producing wells above the head
scarp of the Abalone Cove landslide have declined 100-150 ft. Attempts to monitor
the groundwater elevation suggest that the groundwater table has dropped only about
50-60 ft; however, measurements are limited and no measurements were made
before the dewatering system was initiated (Robert Douglas, personal
communication). Therefore, the groundwater table may have dropped more than the
measured value. Most notable in this area is the decrease in the hydraulic gradient.
This decrease in hydraulic gradient will decrease the rate of groundwater flow in
area.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 2
Table 2-8. Water budget for transient state model
Year INPUTS
Subsurface Inflow
and Surface
Runoff (acre-ft)
Total
Areal
Recharge
(acre-ft)
Total
(acre-ft)
OUTPUTS
Subsurface
Outflow
(acre-ft)
Pumping
Wells
(acre-ft)
Deep
Rooted ET
(acre-ft)
Total
(acre-ft)
1975 60.1 99.8 159.9 1348.2 0.0 36.3 1384.5
1976 47.7 99.8 147.5 962.6 0.0 33.0 995.6
1977 47.7 99.8 147.5 849.4 0.0 21.4 870.8
1978 96.1 99.8 195.9 724.5 87.6 22.8 834.9
1979 114.6 37.0 151.6 676.5 87.6 17.6 781.7
1980 140.4 47.4 187.8 603.4 68.8 10.9 683.1
1981 150.5 22.9 173.4 511.0 254.0 4.8 769.8
1982 141.2 28.4 169.6 488.7 294.4 4.1 787.2
1983 167.5 109.9 277.4 366.8 294.4 1.5 662.7
1984 177.7 19.6 197.3 312.1 294.4 0.7 607.2
1985 178.2 21.4 199.6 298.3 294.4 0.6 593.3
1986 187.6 36.9 224.5 238.7 294.4 0.3 533.4
1987 193.4 30.7 224.1 201.9 294.4 0.1 496.4
1988 197.1 30.7 227.8 172.7 294.4 0.1 467.2
1989 200.1 19.8 219.9 145.9 294.4 0.1 440.4
1990 196.3 30.7 227.0 154.9 294.4 0.1 449.4
1991 191.5 35.7 227.2 168.9 294.4 0.1 463.4
1992 187.7 55.7 243.4 180.1 294.4 0.2 474.7
1993 198.2 65.6 263.8 127.7 294.4 0.1 422.2
1994 186.9 19.9 206.8 177.2 294.4 0.2 471.8
1995 196.5 82.2 278.7 131.2 294.4 0.2 425.8
1996 196.6 23.1 219.7 163.4 217.7 0.1 381.2
1997 185.5 29.7 215.2 225.3 244.6 0.2 470.1
1998 198.7 95.1 293.8 162.5 224.4 0.2 387.1
1999 199.9 12.9 212.8 155.8 224.4 0.1 380.3
2000 197.7 17.6 215.3 166.1 224.4 0.1 390.6
2002 203.6 11.1 214.7 140.2 208.7 0.1 349.0
2003 204.3 14.1 218.4 136.7 208.7 0.1 345.5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Row Cell Number
_________________ Legend__________________
• - Abalone Cove landslide head scarp
— - Groundwater Table for 2003
- - Groundwater Table for steady state model
■ - Outside of groundwater boundaries (inactive)
Figure 2-14. MODFLOW generated cross-section along column 18 for 2003 compared
to groundwater elevation generated for steady state model. In the Abaone Cove
landslidegroundwater elevations have declined approximately 40 feet. In the ancient
landslidegroundwater elevations have declined approxmately 120 feet.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 4
E. MODPATH
Another useful modeling program that can provide insight into the flow paths
and flow rate of water in the basin is MODPATH. MODPATH is a 3-D particle
tracking model which uses groundwater elevations calculated by MODFLOW.
MODPATH uses a linear interpolation, which allows it to predict the position of a
particle assuming that the rate of change is constant. The governing equation for the
linear interpolation approach used by MODPATH is:
where vx is the linear velocity in the x direction, xp is the coordinate of the particle,
and i and j represent node locations in the x and y direction, respectively (Anderson
and Woessner, 1992). In order to solve for the coordinate of the particle (xp),
MODPATH uses a semi-analytical integration method:
and (v*)i and (vx )2 are velocities at each end of the node (Anderson and Woessner,
1992). Similar equations for Equations 2-8 and 2-9 can be derived for the y and z
direction. In short, MODPATH uses the groundwater velocities of the surrounding
node to predict the position and rate of particle movement.
v
X
(Equation 2-8 a)
/
where f x = x n - x , / Ax, ,
p i-L i ™
V 2 )
(Equation 2-8 b)
(Equation 2-9 a)
where Ax = [(v j2 -(v J J/A x , , (Equation 2-9 b)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 5
The particles tracked in MODPATH are non-reactive. Therefore, processes
affecting the movement of particles, such as adsorption/desorption, are not
considered. This is a reasonable assumption when tracking the flow path of water.
Particles were placed along the head scarp of both landslides and along the unlined
portion of Altamira Canyon. MODPATH was run from 1976 - 2003. The output
from MODPATH is flow paths and arrows which represent a specified period of
time (Figures 2-15 and 2-16). The time interval of 364.5 was chosen to best
represent one year of flow (Johnson Yeh, personal communication). MODPATH
was set for forward tracking, meaning that it projects the path of the particle since
the time of release (start of the model). In order to make the output from
MODPATH easier to read, the output was reduced to show groundwater flow after
every 5 years.
MODPATH predicts it should take water approximately 15-20 years to flow
from the head scarp of the ancient landslide to the area on the east side above the
head scarp of the Abalone Cove landslide. Water traveling to the area directly above
the head scarp of the Abalone cove landslide on the west side should take
approximately 10-15 years, depending on the starting point of the particle. At the
toe, MODPATH predicts that water in the basin should take approximately 22-28
years to travel from the head scarp of the ancient landslide to this point. One
interesting point is that water above the head scarp of the Abalone cove landslide
appears to take longer to flow on the eastern flank than on the western flank. This
apparent discrepancy in age is a consequence of location. The area on the west side
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
________________ Legend________________
m m m m m - Model Boundary
- Particle Path
Figure 2-15. MODPATH calculated particle flow paths and rates. The flow rate of
the particle depends on its starting point. (Topographic map compiled from USGS
Redondo Beach, California, San Pedro, California, and Torrance, California 7.5
minute quadrangles (U.S. Geological Survey, 1981).)
* P o in t ' W t f
In s p ira tio n *
P o in t
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 7
8 0 0 -E
7 0 0 -E
6 0 0 -E
5 0 0 -E
4 0 0 —
45
s
0
1
5 300 -4
53
2 0 0 -
1 0 0 -E
-100
Particle
starting
location
Particle
starting
location
15 20 25 30 35
Row Cell Number
Legend
«■» - Abalone Cove landslide head scarp
- Groundwater Table for 2003
— - Flow path for one particle. Each arrow represents 5
years of flow
■ - Outside of groundwater boundaries
Figure 2-16. MODPATFI generated cross-section along column 18 displaying the flow
vertical flow path of two particles. The depth of particle location increases as it flows
downgradient due to the large input of areal recharge to the basin.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
of the basin is located closer to the sources of recharge. Altamira canyon, which is
located on the western side of the basin, adds a significant amount of water to the
system. Therefore, water in this portion of the basin should have a younger age than
areas that are located farther from this of recharge.
Both the horizontal and vertical movement of water in the basin are affected
by changing input of areal recharge, hydraulic conductivity values and changes in
basin area. As these particles move down gradient through the basin, their depth
increases (Figure 2-16), which is due to the recharge throughout the region.
MODPATH predicts rapid flow of water through the Abalone Cove landslide. This
should be expected because the hydraulic conductivity in the Abalone Cove landslide
is 2 orders of magnitude greater than the area directly above the Abalone Cove
landslide head scarp. In addition, the basin area transmitting flow decreases as water
is channeled into the Abalone Cove landslide. This decrease in basin area should
also increase the rate of groundwater flow.
MODPATH does not predict the amount of mixing of water bodies of
differing ages. Mixing of water bodies might become significant at the toe of the
Abalone Cove landslide. This area receives subsurface flow originating in the
ancient slide and water infiltrating through the head scarp of the Abalone Cove
landslide. MODPATH predicts that water in this area supplied by the ancient
landslide should be approximately 22-28 years old, whereas water supplied from
percolation of water through the head scarp of the Abalone Cove landslide should
only be 1 year old. Therefore, a maximum age constraint can be placed on this water
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 9
of 28 years. However, water may be younger in this region depending on the amount
of mixing of the two source bodies.
F. Summary
The MODFLOW calculated heads are reasonably representative of
groundwater elevations in the basin. The transient model illustrates a large decrease
in the groundwater elevations due to initiation of the dewatering system. In the area
above the head scarp of the Abalone Cove landslide, MODFLOW calculates a
decrease of 100-150 ft in groundwater elevation relative to pre-pumping elevation.
Finally, MODPATH can be used to predict the flow rate of water in the basin.
MODPATH predicts that flow from the head scarp of the ancient landslide to the
eastern flank of the ancient landslide complex directly above the head scarp of the
Abalone Cove landslide should take approximately 15-20 years. On the western
flank, water should flow faster (taking 10-15 years). This discrepancy in age is
produced by the close proximity of Altamira Canyon, which is a source of
groundwater. A maximum age of 28 years can be placed on water at the toe of the
Abalone Cove landslide; however, water in this area may be younger depending on
the mixture of subsurface inflow from the ancient landslide and water infiltrating in
the Abalone Cove landslide. The models do not mix water vertically through the
basin and MODPATH predicts that water should become older as depth increases.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5 0
CHAPTER III
GEOCHEMICAL TRACERS:
CHLOROFLUORCARBONS
A. Introduction
Chlorofluorocarbons (CFCs) are synthetic, halogenated alkanes developed by
Dupont in the 1930’s. Plummer and Busenberg (2000) have reviewed pertinent
characteristics and their utility for age-dating groundwater. The two most common
forms of CFCs are CFC-11 (CC13 F), and CFC-12 (CC12 F2 ). CFCs are inert and have
been used in a wide variety of industrial applications, including coolants, foam
blowing agents, packing material, propellants and solvents. Since they are stable,
CFCs have atmospheric lifetimes of approximately 45 years (CFC-11) to 87 years
(CFC-12). The main process for the breakdown of CFCs is a chain of
photodissociation reactions in the stratosphere (Reaction 3-la-c):
CC12F2 - uv ■ > CC1F2 + Cl (Reaction 3-1 a)
Cl + 0 3 ---------» 0 2 + CIO (Reaction 3-lb)
CIO + O ---------> Cl + 0 2 (Reaction 3-lc)
The production of the free chlorine (Cl) leads to the decomposition of ozone
(0 3 ) in the stratosphere. It was recognized in the mid 1970’s that CFCs were a major
contributor to the depletion of the ozone layer. In 1987, thirty-seven nations signed
the Montreal Protocol on Substances that Deplete the Ozone Layer, agreeing to limit
the use of CFCs and reduce emissions by 50% by the year 2000. The Clean Air Act
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
(1996) ceased the production of CFCs in the United States; however, they are still
used in other countries. Production of CFCs in the United States peaked in 1987.
The atmospheric concentration of CFC-11 peaked in 1993-94 and has begun to
slowly decrease, although the CFC-12 atmospheric concentration continues to rise
slowly (Figure 3-1).
Atmospheric concentrations of CFCs in the northern hemisphere have been
measured since 1976. Atmospheric concentrations have been estimated prior to
1976 using the amount of CFCs manufactured, estimates of release into the
atmosphere and hydrosphere, the rate of photolysis in the stratosphere and the rate of
removal by the soils, hydrosphere and biosphere. Atmospheric concentrations are
measured by the National Oceanographic and Atmospheric Administration (NOAA)
atNiwot Ridge, Colorado (elevation 9,885 ft (3,013m)).
B. CFCs as Hydrologic Tracers
CFCs have been used as tracers for oceanic circulation and mixing processes
since the late 1970’s. In the early 1990’s, this technique was adopted for
groundwater studies. CFCs are useful tracers because: (1) atmospheric
concentrations have been well documented through time, (2) their solubilities are
well known, and (3) concentrations are high enough to measure.
Examples of the use of CFCs as hydrologic tracers in unconfined aquifers
include Cook et al (1995), Szabo et al (1996) and Beyerle et al (1999). Szabo et al
(1996) demonstrated that CFC derived groundwater ages were strongly correlated to
3 H/3 He derived groundwater ages and flow model-based travel times. Both Cook et
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Concentration (pptv)
5 2
600
500
400
300
200
100
0
2000 1980 1960 1940
Year
Figure 3-1. Atmospheric CFC concentration through time
(from Plummer and Busenberg, 2000).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
al (1995) and Szabo et al (1996) demonstrated a strong correlation between CFC
concentrations and depth, with concentration decreasing with depth. This should be
expected because shallow water can readily exchange CFCs with the atmosphere or
soil gas. If water in a well is derived from multiple depths in the unconfined aquifer,
the CFC derived age will be a mixture of old deep water and younger shallow water.
Beyerle et al (1999) demonstrated that CFCs ages in an unconfined aquifer depend
greatly on the amount of recent infiltration, with younger groundwater ages obtained
during periods of high infiltration. At the same time that fluctuations were observed
in the unconfined aquifer, a deeper confined aquifer displayed uniform CFC derived
ages (Beyerle, 1999). This study indicates that CFC derived groundwater ages may
display temporal variation and young ages might be expected following periods of
rainfall.
The concentration of CFCs in groundwater is dictated by Henry’s Law. For
CFC-11 the dimensionless Henry’s Law constant (Hd) is 3 (at 20° C) and for 12 for
CFC-12 (Warner and Weiss, 1985). The solubilities of CFCs were determined by
Warner and Weiss (1985) for a wide range of temperature and salinities.
Water continues to equilibrate with atmospheric concentrations until it has been
removed from contact with the atmosphere. The low solubility of CFCs keeps air in
the vadose zone close to the composition in the atmosphere, so water in the vadose
zone should be in equilibrium with the atmosphere. Consequently, the CFC
concentration obtained from a water sample indicates the time at which the sample
became isolated from the unsaturated zone (Busenberg and Plummer, 1992). As
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5 4
previously mentioned, CFCs are degraded in the stratosphere. They are also
degraded in anaerobic conditions (Lovely and Woodward, 1992). CFC-11 degrades
more readily than CFC-12. As CFCs degrade, their concentrations decrease. Low
concentrations of CFCs are indicative of old water. Therefore, CFC samples from
anaerobic aquifers may give a false old age.
C. CFC Sampling
Such high Henry’s constants indicate that CFCs partition strongly into the gas
phase. Therefore, care must be taken when sampling groundwater in order to ensure
that the samples have not been exposed to a gas phase, particularly the atmosphere.
In addition, groundwater samples can be contaminated with organics, including
CFCs, by exposure to gas permeable material such as Teflon and Tygon (Johnson et
al, 1987). Water samples were collected from pumping wells (8" OD PVC pipe with
a stainless steel submersible pump) and sealed in copper tubing, using the technique
of Clark et al (1997) for noble gas samples. The copper tubing was connected
directly to the outflow faucet of the pumping wells and crimped with stainless steel
knife-edge clamps in order to seal the water sample inside the copper tubing. (For a
detailed sampling procedure see Appendix A). Samples can be stored for
approximately 3-4 months in the copper tubes.
Temperature, conductivity, salinity and dissolved oxygen measurements were
performed in the field with the YSI (Model 85-100 FT) Oxygen, conductivity,
salinity and temperature meter. Samples were taken in a 500 ml plastic beaker
immediately after the CFC samples were taken. The salinity calculated by the meter
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5 5
assumes solutes have ratios equivalent to seawater. This assumption has little effect
on the calculated solubilities.
D. CFC Analysis
CFCs were analyzed following the procedure developed by Bullister and
Weiss (1988) (Details in Appendix B). All samples were run in Dr. Jordan Clark’s
lab at the University of California, Santa Barbara (UCSB). Extraction and
concentration of CFCs from the water samples were performed using a purge and
trap system. Samples of approximately 10 ml, stored in copper tubing, were
introduced into a stripper chamber and flushed with bubbles of Ultra High Purity
(UHP) N 2 gas. The sample was stripped for 4 minutes (flow rate 50-60 ml/min).
The effluent passed through a dessicant column (packed with Mg(C1 0 4 )a) and then
into a trap (12" x 1/8" OD stainless steel tube packed with unibeads of 80/100 mesh
glass beads) cooled with a dewar containing a mixture of dry ice (~ -78° C) and
methanol. The sample was trapped for 4.1 minutes. The trap was isolated, heated
with a dewar of boiling water (100° C), and back flushed into a precolumn (6" x 1/8"
OD stainless steel tube packed with Porapak C). The backflush stream (flow rate of
50-60 ml/min) then entered the gas chromatograph column (6' 1/8" OD 0.028" ID
stainless steel column packed with 1% AT-1000 on Carbograph 1, 60/80 Mesh). The
pre-column and column were heated to 70° C. The gas chromatograph (GC
Shimadzu Model 14A) was equipped with a 6 3 Ni electron capture detector (ECD)
heated to 300° C. Finally, the eluted compounds were plotted and the peak areas
were integrated using an integrator (Shimadzu C-R8-A Chromatopac).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5 6
A standard of known concentration was run through loops of varying
volumes, through the trap, and into the GC. A series of standards were run at the
beginning and end of each sampling period. From these standards, a calibration
curve was constructed in order to relate sample peak area to moles of CFCs. (For
details see Appendix C). The curve was quite linear over the range of 0 to 2 xlO'1 3
moles of standard gas. Once the number of moles per sample was established, the
concentration was obtained by dividing the number of moles by the volume of the
water sample based on the weight change resulting from draining the copper tube.
Using the sampling procedure discussed in Appendix A, sample volumes were
approximately 10 ml.
Next, the solubility of CFCs was calculated for each sample using the
solubility determined by Warner and Weiss (1985) (Equation 3-1):
i ^ ( 100 " l ,
InA: = ax +a2\ — l + n3ln
100.
b t + bsy I --- I 4 " b*i
1 HiooJ 3 100.
(Equation 3-1)
where K' is the solubility of the CFC in mol/L-atm, T is the water temperature in
degrees Kelvin, S is salinity in parts per thousand (ppt) and the a’s and b’s are
constants. The temperature and salinity were measured at the time of sampling
(Table 3-1). Finally, using the measured concentrations and the solubility from
Equation 3-1, the partial pressure of the CFC was calculated for each sample
(Equation 3-2).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5 7
Table 3-1. Field measured temperature, conductivity,
dissolved oxygen and salinity
W ell
T em perature
(°C )
C onductivity @
25° C (m S)
O xygen
(p M )
Salinity*
(PPt)
WW-1 (4/28/03) 23.00 5.25 13.8 3.0
WW-1 (2/19/04) 21.90 5.14 43.1 3.0
WW-2 (4/28/03) 23.50 4.60 2.7
WW-2 (9/18/03) 24.00 4.90 68.8 2.6
WW-2 (11/20/03) 23.30 4.60 111.6 2.6
WW-2 (2/19/04) 23.70 4.60 109.4 2.5
WW-5 (11/20/03) 23.60 6.59 137.5 3.8
WW-5 (3/29/04) 24.10 6.81 64.1 3.7
WW-5 (5/27/04) 22.90 6.65 57.8 3.8
WW-8 (4/28/03) 23.60 5.09 134.4 2.9
WW-8 (11/20/03) 23.30 5.06 208.4 2.8
WW-8 (2/19/04) 23.70 5.03 95.0 2.8
WW-8 (5/27/04) 23.10 5.03 137.5 2.7
WW-12 (4/28/03) 23.50 5.40 115.0 3.0
WW-12 (9/18/03) 23.50 5.60 54.7 3.0
WW-12 (11/20/03) 23.40 5.43 141.6 3.0
WW-12 (2/19/04) 23.60 5.46 125.0 3.0
WW-13 (11/20/03) 21.70 4.40 125.0 2.5
WW-13 (2/19/04) 22.40 4.50 109.4 2.5
Tolocskzo (4/28/03) 23.40 4.70 71.9 2.7
Tolocskzo (9/18/03) 23.30 4.70 101.6 3.0
Tolocskzo (11/20/03) 22.00 4.90 114.7 2.6
Tolocskzo (2/19/04) 22.90 4.70 78.1 2.6
Tolocskzo (5/27/04) 22.50 4.50 63.8 2.4
Disgusting (4/28/03) 23.10 12.20 15.3 7.0
Digusting (11/20/03) 23.60 17.46 27.2 10.1
Disgusting (2/19/04) 23.50 13.31 74.4 7.8
Disgusting (5/27/04) 23.20 13.07 62.5 7.5
Monahan (3/15/04) 22.30 5.06 131.3 2.9
Kelvin Spring - C (12/17/03)* 16.70 0.17 213.1* 1.0
Kelvin Spring - C (3/12/04)* 19.70 3.77 214.1* 2.0
Kelvin Spring - C (6/3/04)* 22.40 3.65 56.6 1.9
Kelvin Spring - A & B (12/17/-03)* 21.10 0.17 207.2* 1.0
Kelvin Spring - A & B (3/12/04)* 21.00 3.60 122.8 1.9
Kelvin Spring - A & B (6/3/04)*
22.30 3.62 20.3 1.9
* Samples aerated during collection,
t - Salinity calculated assuming 1:1 ratio of ions in seawater
$ - For location of A, B and C see Figure 3-5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5 8
P , = (Cw j (Equation 3-2)
where P, is the partial pressure of CFC (atm) and Cw is the concentration in water
(moles/L).
Once the partial pressure of CFC for each sample was known, it was
converted to parts per trillion by volume (pptv) and compared to the established
atmospheric concentration curve. The apparent age of each water sample was
estimated by matching the partial pressure of the sample to year in which
atmospheric concentrations were the same. This assumes that once water becomes
groundwater it is a closed system and has not mixed with water of different ages.
E. Application of CFC Dating to Abalone Cove Landslide
When using CFCs as a groundwater tracer, it is assumed that there is no
anomalous local source of CFCs, the sample is not contaminated during collection,
and CFC concentrations have not been modified by geochemical, biological or
hydrologic processes (Busenberg and Plummer, 1992). In the Abalone Cove and
ancient landslides these assumptions should generally hold true. The landslide is
located on the Palos Verdes peninsula which protrudes from the southern California
coast, south of the city of Los Angeles. Most of the atmospheric influence in this
area is marine and not urban. This is an important assumption because it has been
shown that urban areas may have higher atmospheric concentrations of CFCs than
recorded at Niwot, Colorado (Hester et al, 1974 and Ho et al, 1998). Samples
contaminated during sampling and sample storage were recognized by either visual
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
inspection of the gaps in the clamps or anomalously high concentrations. Samples
that were not properly sealed at the time of sampling will equilibrate with the
atmospheric air in the laboratory during storage. Since the laboratory is located in
downtown Los Angeles, high concentrations of CFC would be expected in unsealed
samples. Finally, dissolved oxygen was measured for each sample at the time of
sampling (Table 3-1). Most wells have high concentrations of dissolved oxygen.
Since there is no evidence of aerobic degradation of CFCs (Plummer and Busenberg,
2000), these wells are assumed to have unmodified CFC concentrations. However,
two wells (Disgusting and WW-5) did have low dissolved oxygen concentrations and
a strong odor of H2 S at times (Table 3-1). At times, Kelvin Spring also had low
dissolved oxygen. CFC concentrations at these sites may be degraded, showing false
old ages.
F. Summary of CFC Data
Samples were collected from a suite of wells above and below the Abalone
Cove landslide head scarp, as well as the Kelvin spring. When possible, duplicate
samples were taken. During gas chromatographic analysis of each sample, several
unknown analytes eluted along with the CFCs. One analyte eluted at approximately
the same time as CFC-11, swamping the CFC-11 peak (Figure 3-2). Therefore, it
was not possible to determine the concentration of CFC-11 in the samples. CFC-12
was easily distinguished from other analytes in the samples (Figure 3-2).
Sample standard deviation (ssd) was calculated for each pair of
uncontaminated duplicate samples (Table 3-2). Several samples were found to be
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Tim e (min)
0
1
Air
2
- CFC-12
2.637
3
4
5
6
7
CFC-11
8
9
10
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Microvolts (|iV)
Figure 3-2. Chromatograph for well WW-12 (9-03). The CFC-11 peak is swamped
by an unknown analyte.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 3-2. CFC-12 concentrations and apparent ages
61
Sample ID
Cone, of
CFC-12
(PM)
Mean
Cone.
(PM)
Sample
Standard
Deviation
(PM)
Percent
SSD
Partial
Pressure
(pptv)
Age
(years)
WW-1 (4-03) #1 1.41 1.21 0.28 23 463 (± 108) 15 (± 5)
#2 1.01 23 332 (± 76) 21 (± 7)
WW-1 (9-03) #1 1.37 7f 448 (± 33) 15 (± 1)
WW-1 (2-04) #1 1.17 1.25 0.11 9 369 (± 33) 19 (± 1.5)
#2 1.33 9 419 (±33) 16 (± 1 )
WW-2 (4-03) #1 8.3*
#2 8.24*
WW-2 (9-03) #1 4.88*
#2 16.8*
WW-2 (11-03) #1 16.8*
#2 16.4*
WW-5 (11-03) #1 2.45*
WW5 (3-04) #1 1.09 1.01 0.11 11 375 (± 43) 20 (± 2)
#2 0.93 319 (±36) 23 (± 3)
WW5 (5-04) #1 0.84 0.98 0.19 20 276 (± 54) 26 (± 3)
#2 1.11 20 144 (± 28) 33 (± 2)
WW-8 (4-03) #1 1.10 1.14 0.05 4 368 (± 16) 19 (± 1)
#2 1.17 4 391 (± 16) 18 (± 1)
WW-8 (11-03) #1 2.62*
#2 2.82*
WW-8 (2-04) #1 1.56 7t 524 (± 37) 9 (± 4 )
#2 2.55*
WW8 (5-04) #1 2*
WW-12 (4-03) #1 1.18 1.18 0.00 0 393 (± 0) 18 (± 0 )
#2 1.18 0 393 (± 0) 18 (± 0 )
WW-12 (9-03) #1 1.53 7f 511 (±38) 11 (± 3 )
#2 1.99*
WW-12 (11-03) #1 1.51 1.56 0.06 4 503 (± 21) 11 (± 2 )
#2 1.60 4 531 (±21) 8 (± 3)
WW-12 (2-04) #1 2.09*
#2 3*
WW-13 (9-03) #1 1.9*
#2 1.93*
* - indicates samples where CFC-12 concentrations exceed possible atmospheric concentrations
t - indicates the average percent error calculated from duplicates
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 3-2 (cont.). CFC-12 concentrations and apparent ages
Sample ID
Cone, of
CFC-12
(PM)
Mean
Cone.
(PM)
Sample
Standard
Deviation
(PM)
Percent
SSD
Partial
Pressure
(pptv)
Age
(years)
WW-13 (11-03) #1 1.9*
#2 1.64 7* 508 (± 36) 11 (± 3)
WW-13 (2-04) #1 3.37*
#2 3.51*
Petok (6-03) #1 2.47*
Tolocskzo (4-03) #1 1.01 1.02 0.01 1 334 (± 5) 21 (±0.5)
#2 1.03 1 344 (± 5) 19 (±0.5)
Tolocskzo (9-03) #1 3.78*
Tolosckzo (11-03) #1 1.01 7* 319 (±23) 22 (± 2)
Tolocskzo (2-04) #1 1.54 7* 502 (± 35) 11 (± 3 )
#2 2.65*
Tolocskzo (5-04) #1 2.66*
Disgusting (4-03) #1 0.26 0.25 0.01 6 87 (± 5) 36 (± 0)
#2 0.24 6 81 (± 5 ) 36 (± 0)
Disgusting (9-03) #1 0.49 0.46 0.04 9 166 (± 15)
31 (± 1)
#2 0.43 9 147 (± 15) 32 (± 1)
Disgusting (11-03) #1 0.22 0.25 0.04 17 77 (±13) 37 (± 1)
#2 0.28 17 98 (± 13) 34 (± 1)
Disgusting (2-04) #1 1.11 7f 386 (± 27) 18 (± 2 )
#2 1.8*
Disgusting (5-04) #1 1.47 1.17 0.43 36 507 (± 184) 12 (± 11)
#2 0.87 36 300 (± 109) 24 (± 6)
Monahan (3-04) #1 2.69*
#2 2.74*
Kelvin Spring* A 3.47*
(12-03) B 0.46 0.66 0.28 17 137 (± 23) 32 (± 1)
C 0.86 17 254 (± 23) 26 (± 1)
Kelvin Spring* A 2.84*
(3-04) B 1.60 7* 482 (± 60) 12 (± 3)
C 0.17 7f 51 (± 4 ) 41 (± 1)
Kelvin Spring* A 1.49 7f 472 (± 33) 14 (± 2)
(6-04) B 0.11 7* 36 (± 3) 43 (± 2)
C 0.00
* - indicates samples where CFC-12 concentrations exceed possible atmospheric concentrations
f - indicates the average percent error calculated from duplicates
J - For location o f A, B and C see Figure 3-5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 3
contaminated after gas chromatographic analysis. Contaminated samples were
identified because they indicated unreasonably high CFC-12 partial pressures.
Contamination may have occurred because some clamps were not properly
tightened. When neither duplicate sample was contaminated, the ssd of duplicates
ranged from 23% to 0%, with an average ssd of 7%.
CFC partial pressures were calculated and compared to the atmospheric
mixing ratios for Niwot Ridge, Colorado to determine apparent age (Table 3-2).
Samples collected from the area directly above the Abalone Cove landslide (wells
WW-1, WW-8, WW-12, WW-13 and Tolosckzo) have CFC apparent ages of 8-23
years. This range of ages extends throughout the area, meaning that wells on both
the east and west side of the basin have waters with similar characteristics. The
youngest water was obtained from samples collected on February 19,2003, a few
days after a rain event (Figure 3-3). The higher CFC-12 partial pressures are thought
to be due to mixing of old, subsurface water (allocthonous water) with young
rainwater (autochthonous) that has percolated into the slide mass.
The sensitivity of different wells to recent rainfall can be estimated by
looking at the change in the fraction of autochthonous water of the well after a rain
event. As water flows through the landslide complex, it is continually recharged
with percolating precipitation. Therefore, water in each well contains a fraction of
autochthonous water. The average annual fraction of autochthonous water in each
well was calculated based on the transient state MODFLOW model results for the
year 2003:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Q .
Q.
C M
O
» + -
O
c
Q >
O
c
o
O
CFC-12 Partial Pressure and Precipitation vs. Sampling Date
5
600
' r
500
400
300
200
100
0
C O CO C O < o
22
M
64
7 3
0 3
3*
5 T
5 ’
£
3 J 3
E
Q.
©
C O
4 3
E
©
>
o
z
©
o
£ •
r a
3
£ P
( 0
E ^
4 3 ^
©
U.
Sampling Date
Legend
■ - WW 1
► - WW 5
♦ - WW 8
▲ - WW 12
X - WW 13
▼ - Tolosckzo
• - Disgusting
M - Kelvin Spring
- Precipitation
Figure 3-3. CFC-12 partial pressures for each well per sampling date. Results for
duplicates are shown. The letters next to the symbols represent samples from Kelvin
Spring A, B and C. For the location of these samples see Figure 3-5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 5
faU ochonous = (Equation 3-3)
c e l l
where V p erc is the volume of water added to the cell through rainfall percolation in
•5 o
one year (L ), and Vc e u is the calculated volume of water in the cell (L ) (Table 3-3).
Mixing curves were constructed in order to determine the change in the
fraction of allocchthonous water contributing to the samples collected in February
(Figure 3-4). The low concentration end-member was established by averaging all
measurements, excluding February 2003. The high concentration end-member was
assumed to be equivalent to present day atmospheric concentration (540 pptv).
Table 3-3. Percent change in allocchthonous water in wells sampled in
February 2003.
Well Saturated
Thickness
(ft)
Total Cell
Volume
(105 ft3 )
Average
CFC-12
Partial
Pressure
(pptv)
February
CFC-12
Partial
Pressure
(pptv)
Percent
Change
in of
Alloc.
Water
WW-1 164 37.3 414 419 + 4%
WW-8 158 25.9 379 524 + 90%
Tolosckzo 127 28.9 332 502 + 82%
Disgusting 5 1.1 109 386 + 64%
Percolation of rainfall has the greatest impact on wells WW-8 and Tolosckzo,
with percent changes in allocthonous water of 90% and 82%, respectively. Well
WW-8 is located just down gradient of well Tolosckzo (see Figure 1-2). As
previously mentioned, landslide induced fractures can play an important role in
subsurface flow. Rainfall enters the groundwater quickly through these fractures
and, because the hydraulic conductivity should be very high, these fractures act as a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
(a) Well WW-1
660
y = 414 + 126X R= 1
500
Z
a
C M
O
450
400
1 0.6 0.8 0 0.2 0.4
Fraction of Recent Percolation
(b) Well Disgusting
600
y® 109 + 431X R= 1
500
Z 400
200
100
1 0.6 0.8 0 0.2 0.4
Fraction of Recent Percolation
(c) Well WW-8 (d) Well Tolosczko
550
y= 332 + 208X R= 1
500
Z 450
a
C M
6
& 400
350
300
0.6 1 0 0.2 0.4 0.8
550
y = 379 + 161x R= 1
500
$ 2 450
400
350
1 0.6 0.8 0.2 0.4
Fraction of Recent Percolation Fraction of Recent Percolation
Figure 3-4. Mixing curves for February 2003 samples. The low concentration
end-member was established by averaging all measurements excluding February 2003.
The high concentration end-member was assumed to be equivalent to present day
atmospheric concentration (540 pptv). The star on each graph represents the actual
February 2003 measured value.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 7
channel for groundwater flow. Since both wells WW-8 and Tolosckzo are close in
proximity and are greatly affected by percolating rainfall, it is hypothesized that
these wells are hydrologically connected through fracture flow. Well Disgusting is
also influenced by percolating rainfall, with a 64% percent change in autochthonous
water in February 2003. This is expected because the toe of the landslide is highly
fractured and flow in this section of the basin is very rapid. Despite the rapid travel
of recent rainfall, it is interesting that conductivity varies little in February 2003
compared to previous months. Ionic concentration must be very well buffered.
Water from the toe of the landslide (well Disgusting) has CFC ages of 18-37
years. As previously mentioned, these may be false old ages because well
Disgusting has very low concentrations of dissolved oxygen, leading to possible
degradation of the CFC-12. The water sampled from Kelvin spring shows CFC ages
of 14-43 years. When compared to the area directly down gradient, Kelvin spring
has substantially older water.
It was expected that the oldest water would be located at the toe of the
landslide and Kelvin Spring would be young because it is located the farthest uphill.
The older age derived for Kelvin spring probably indicates that the spring is supplied
with a substantial component of older groundwater originating outside of the
landslide basin. One piece of evidence supporting this hypothesis is the constant
year round flow of water in the spring. The years 2001-2003 were periods of
reduced rainfall (30-75% less than the 100-year average). If the spring was supplied
solely by water infiltrating through the head scarp and percolating precipitation, the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
68
spring might be expected to dry up in years of low rainfall. Consequently, some of
the water supplying Kelvin spring probably flows througb the head scarp of the
ancient landslide, where little bentonite should be present. However, multiple
sources probably feed this spring. Subsurface samples were collected from about 15
cm depth at sites only 2 meters apart (Figure 3-5). Large differences were observed
for CFC concentration. One sample had rather low oxygen concentration, so some
CFC degradation cannot be ruled out.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
Samples A, B and C
Figure 3-4. Photograph of Kelvin Spring. The arrow is pointing to the location where
samples A, B and C were taken. The spring was always sampled in sequential order.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 0
CHAPTER IV
GEOCHEMICAL TRACERS:
TRITIUM
A. Introduction
Tritium (3 H, often referred to with the chemical symbol T), the radioactive
isotope of hydrogen, is a low energy |3- emitter. With a half-life of 12.23 years
(Lucas and Unterweger, 2000), tritium decays to the stable isotope 3 He
(L’Annuziata, 1998):
3H -» 3 He + j3 -+ v + 0.0\SMeV (Reaction 4-1)
Naturally occurring tritium was first detected in 1950 by Fairings and Harteck, and
later shown by Kaufman and Libby (1954) to be present in precipitation.
Atmospheric tritium occurs naturally as a result of cosmic radiation induced
reactions in the stratosphere (Libby, 1946):
l4N + n m > H + n C (Reaction 4-2)
Craig and Lai (1961) calculated the natural global production rate of tritium
in the northern hemisphere to be 0.5 ± 0.3 atoms T/cm -s. Stewart and Farnsworth
(1968) demonstrated that stations near the ocean have precipitation that is low in
tritium due to the dilution effects of the ocean. Large peaks in the tritium activity of
precipitation occur in the late-spring to early summer due to greater vertical mixing
between the stratosphere and troposphere (Stewart and Farnsworth, 1968).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
Atmospheric tritium concentrations increased sharply in the 1950’s and
1960’s due to nuclear weapons testing. Since the largest thermonuclear detonations
occurred in the Northern Hemisphere, stations in the north generally received higher
tritium concentrations than those in the south (Stewart and Farnsworth, 1968). The
bomb produced tritium reached its apex in 1963 due to large thermonuclear tests in
late 1962 and the 1963 nuclear test ban treaty (Stewart and Farnsworth, 1968). Since
that time, the tritium concentration in precipitation has decreased and is currently
near pre-bomb levels.
B. Tritium as a Hydrologic Tracer
Tritium naturally bonds with oxygen by exchange and oxidation reactions to
form molecules of HTO. These molecules enter the hydrologic cycle in the form of
precipitation (Erikson, 1965). The isotopic ratio of precipitation is affected by
altitude, latitude and seasonal variations. Heavier isotopic species (deuterium ( H or
D) and T) have lower vapor pressures than hydrogen (*H). This decrease in vapor
pressure results in the preferential concentration of the heavier isotopes in the liquid
phase. This process is amplified at lower temperatures. As an air mass moves across
a continent, the first precipitation will be preferentially enriched in the heavier
isotope. This process leads to Rayleigh fractionation. Typically, Rayleigh
fractionation is used in the study of the ratio of stable isotopes of hydrogen and
oxygen (1 8 0 and 1 6 0). Craig and Lai (1961) demonstrated that the ratio of HTO/H2 O
in precipitation is even more enriched relative to water vapor than HDO/H2 O due to
the larger vapor pressure difference between tritium and hydrogen. The pre-bomb
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 2
tritium concentration of precipitation was estimated to be 15 TU (1 TU = 1 atom
T/101 8 atoms 1 H) at Ottawa, Canada and 4 TU near the equator (Brown, 1961). The
present day tritium activity of precipitation at Ottawa, Canada is approximately 18
TU, indicating that tritium activities have approached pre-bomb levels (IAEA/WMO,
2001).
When molecules of HTO fall as rain they may enter surface waters,
groundwater or re-evaporate. Once in groundwater, tritium can leave the system by
radioactive decay, discharge to surface waters or evapotranspiration. The net loss of
tritium by evapotranspiration is considered negligible (Stewart and Farnsworth,
1968). Molecules of HTO are considered good hydrologic tracers because they not
selectively adsorbed in the subsurface (Cariston, 1964). However, Solomon and
Sudicky (1991) demonstrated a slight attenuation of tritium can occur in clay-rich
sediments due to ion exchange with hydroxyls on the clays.
The activity of tritium in precipitation has been well documented for a variety
of stations in both the northern and southern hemispheres as part of the International
Atomic Energy Agency/World Meteorological Organization’s (IAEA/WMO) Global
Network of Isotopes in Precipitation (GNIP) program. As previously mentioned,
precipitation in the late 1950’s and early 1960’s had increased tritium activity due to
nuclear weapons testing. The activity of tritium was as high as 5000 TU (non-decay
corrected) in Ottawa Canada (IAEA/WMO, 2001). If groundwater has tritium
activities higher than natural levels, then it is presumed that the water contains a
component of this bomb-produced tritium. This component of bomb-produced
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
tritium tags the water with a historically datable age. This method of hydrologic age
dating is considered to have a limit of 50 years (Haskell et al, 1966).
C. Tritium Analysis
The tritium concentration of groundwater samples was determined using
liquid scintillation counting. Because most natural waters contain low
concentrations of tritium, it was necessary to enrich the samples with electrolysis
before counting (Brown and Grummitt, 1956). (For detailed procedure see Appendix
D.) Samples were enriched using the protocol of Stencel et al (1995), which was
first developed by Brown and Grummitt (1956) and Ostlund and Dorsey (1977). The
electrolysis system works on the principle that bonds involving heavier isotopes are
stronger than those of lighter isotopes. Therefore, molecules of HTO are not as
readily decomposed into H2 and O2 as are molecules of H2 O. After electrolysis, the
residual liquid is typically enriched by a factor of 20 (Cook et al, 1998).
In order to oxidize the organic matter and remove any suspended solids
before electrolysis, potassium permanganate (KMnO^ was added, and the water was
quantitatively distilled. If the samples are not distilled prior to electrolysis, the
electrodes are subjected to increased corrosion. Next, 50 ml of sample was placed in
a glass cell with an iron cathode and nickel anode and made slightly alkaline with
sodium hydroxide (NaOH). The cells were cooled in a bath (~8° C) in order to
minimize evaporation losses and to improve the separation factor. The cell was
connected to a DC regulated power supply (Extech Instruments Model 382213) and
voltage was increased to create a current of 3.0 A. Because H2 and O2 gases are a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 4
product of the electrolysis process, samples must be enriched inside an active fume
hood. The enrichment process runs at a rate of approximately 1 ml/hr at a current of
3.0 A. More sample is added with a pipette to the cell each day until a total sample
volume of 250 ml has been processed and reduced to 25 ml (approximately 9 days).
Once the sample is reduced to 25 ml, the current is reduced to 0.3 A to prevent
overheating of the electrodes. After the sample is reduced to 10 ml (approximately 6
more days), it is vacuum distilled to remove any NaOH, which will cause
chemiluminescence and quenching during the counting process. After the vacuum
distillation, the final mass of the enriched sample is measured. Finally, in a plastic
counting vial with an aluminum foil lined cap, 9 ml of enriched sample is added to
11 ml of Ultima Gold LLT, an efficient tritium liquid scintillation cocktail (Cook et
al, 1998). The enriched samples were counted using a low background ( 3 particle
counter (Packard Tri-Carb model 3170TR/SL liquid scintillation analyzer). Sample
enrichment was performed by the author at the University of Southern California.
The samples were counted at Idaho State University by Roy Dunker.
For each batch of samples counted, blank samples (background) and tritium
standards were also counted. Trinity Springs™ water (www.trinitysprings.com) was
used as the standard blank. The tritium activity of a sample is calculated:
A s a = (D) (Equation 4-1)
V . S '7 ’ KZ, )
where Asa is the activity of the sample (Bq/kg), Nsa is the net number of counts per
minute for the sample (cpm), Asr is the activity of the standard (Bq/kg), Nst is the net
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 5
number of counts per minute for the standard (cpm), Z/ is the enrichment factor, and
D is the decay correction for the sample (Rozanski and Groning, 2004). The net
number of counts per minute is calculated:
Ns a = Ng s a - Nb g (Equation 4-2)
where Ng s a is the gross number of counts per minute (cpm) and NB g is the number of
counts per minute of the background (cpm). The decay correction for a sample
calculates the decay of the sample between the time of sample collection and
measurement:
D = eM (Equation 4-3)
where X is the decay constant for tritium (1.54 x 10-4 days’1 ) and t is the number of
days between the date of sample collection and the date of sample measurement
(days) (Lucas and Unterweger, 2000). The activity of the sample is reported in
tritium units (TU) using a conversion factor of 1 TU = 0.11919± 0.00021 Bq/kg
(Rozanski and Groning, 2004).
D. Application of Tritium Dating to the Abalone Cove Landslide
Before the tritium activity of a sample can be used as a hydrologic tracer, a
source function for tritium in the basin must be established. The tritium activity of
precipitation has been measured for a variety of stations as part of IAEA/WMO’s
GNIP program. The closest station to Palos Verdes is in Santa Maria, California,
located 190 miles north along the California coast. The tritium activity in
precipitation was measured at this station from 1962 to 1976. The samples were
analyzed by IAEA in Vienna, Austria (IAEA/WMO, 2001). Since this record is
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 6
incomplete for the time period of this study, a tritium history for the station was
reconstructed using the approach of Doney et al (1992). This method predicts the
tritium history for a particular station using the fractional contribution of two
separate reference curves, one for the northern and one for the southern hemisphere.
cp 01 ) = j\c p 0,1) + f 2cp 0,2) + ea 0) (Equation 4-4)
where cp(t) is the tritium activity of precipitation at the station (TU), cp(t, 1) is the
normalized reference curve for the northern hemisphere (year'1 ), cp(t,2) is the
normalized reference curve for the southern hemisphere (year'1 ),// and/ are the
loading factors for each reference curve (TU/year) and ea(t) is a random deviation of
the predicted value from the observed values (Doney et al, 1992). The reference
curves used were GNIP’s Ottawa, Canada station (samples analyzed by the
University of Waterloo, Waterloo, Canada) as the northern hemisphere reference and
Kaitok, New Zealand (samples analyzed by the Institute of Geological and Nuclear
Sciences, Lower Hutt, New Zealand) as the southern hemisphere reference
(IAEA/WMO, 2001). The values off l and f2 were calculated to be 7.750 and 0.245,
respectively. This is equivalent to 0.23 fractional similarity to Ottawa, Canada and
0.13 fractional similarity to Kaitik, New Zealand. The reconstructed curve was
calibrated for the time period of 1962-1976 using measured data from the Santa
Maria station (Figure 4-1) (IAEA/WMO, 2001). The final reconstructed tritium
profile was decay-corrected to mid-2003 (Figure 4-2, Table 4-1). Calculated natural
tritium activities for Santa Maria rainfall are between 2.8 and 4 TU (average of 3.5
TU) (Table 4-1).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 7
80
70
60
50
40
30
20
10
0
1962 1964 1966 1968 1970 1972 1974 1976 1978
Year
Legend
—■— - Measured Santa Maria Data
- Reconstructed tritium activity
Figure 4-1. Reconstructed profile of tritium activity of precipitation plotted with
measured data from the Santa Maria station. The data are decay corrected to 2003 .5.
Data from IAEA/WMO's GNIP database.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 8
8 0
70
60
50
40
30
20
10
Q i I I I i l > I I I I— l - i i I i i i. t I i i < I t - J- L i i I I I i i I I I I i I I I I I
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
Year
Figure 4-2. Reconstructed profile of tritium histroy in precipitation at Santa Maria,
California. The data are decay corrected to 2003.5. Data from IAEA/WMO's GNIP
database.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
Table 4-1. Reconstruction of rainfall tritium for Santa Maria, CA
Tritium Tritium
Activity for Normalized Activity for Normalized Predicted Observed
Ottawa, Ottawa Kaitok, New Kaitok Santa Maria Santa Maria
Year Canada (TU) Curve Zealand Curve Curve (TU) Curve (TU)
1960.5 12.4 0.369 1.1 0.563 3.0
1961.5 19.7 0.586 0.8 0.400 4.6
1962.5 96.3 2.872 1.3 0.659 22.4 27.4
1963.5 304.7 9.090 2.2 1.153 70.7 71.7
1964.5 171.1 5.106 3.7 1.917 40.0 44.4
1965.5 100.3 2.993 4.0 2.042 23.7 22.4
1966.5 71.6 2.137 4.3 2.201 17.1 11.5
1967.5 40.4 1.207 3.4 1.779 9.8 9.6
1968.5 29.1 0.867 3.7 1.915 7.2 7.0
1969.5 35.5 1.058 4.6 2.377 8.8 6.7
1970.5 31.4 0.938 4.1 2.092 7.8 5.7
1971.5 31.7 0.945 3.8 1.940 7.8 8.6
1972.5 15.2 0.454 2.7 1.410 3.9 5.3
1973.5 16.9 0.504 2.3 1.168 4.2 4.2
1974.5 19.7 0.588 1.8 0.941 4.8 4.0
1975.5 15.9 0.474 2.1 1.066 3.9 3.6
1976.5 12.6 0.376 1.5 0.773 3.1 3.3
1977.5 17.2 0.513 1.7 0.868 4.2
1978.5 18.7 0.557 1.5 0.798 4.5
1979.5 12.4 0.371 1.4 0.701 3.0
1980.5 15.3 0.457 1.3 0.692 3.7
1981.5 16.2 0.483 1.2 0.595 3.9
1982.5 14.2 0.423 1.2 0.595 3.4
1983.5 15.6 0.467 1.1 0.587 3.8
1984.5 15.0 0.446 1.2 0.614 3.6
1985.5 12.4 0.371 1.2 0.611 3.0
1986.5 15.4 0.460 1.1 0.557 3.7
1987.5 15.7 0.469 1.1 0.584 3.8
1988.5 16.3 0.487 1.1 0.577 3.9
1989.5 20.0 0.595 1.1 0.586 4.8
1990.5 18.3 0.547 1.1 0.589 4.4
1991.5 18.9 0.564 1.2 0.599 4.5
1992.5 11.4 0.341 1.2 0.616 2.8
1993.5 10.9 0.326 1.4 0.720 2.7
1994.5 11.9 0.354 1.3 0.687 2.9
1995.5 11.7 0.348 1.6 0.823 2.9
1996.5 11.2 0.334 1.6 0.807 2.8
1997.5 15.7 0.468 1.8 0.905 3.8
1998.5 16.5 0.491 1.8 0.904 4.0
1999.5 14.9 0.444 1.7 0.861 3.7
2000.5 19.1 0.571 1.7 0.856 4.6
2001.5 18.3 0.546 1.7 0.872 4.4
All activities are decay corrected to mid-2003. Data from IAEA/WMO, 2001.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8 0
E. Summary of Tritium Data
Samples were collected from a suite of wells above and below the Abalone
Cove landslide including the Kelvin spring. As previously mentioned, sample
temperature, conductivity and dissolved oxygen concentration of the samples were
measured at the time of sampling (Table 3-2). Samples were collected at the faucet
outlet for the pumping wells. Several monitoring wells were sampled using a
discrete down-hole sampler (Appendix A).
Tritium activities were calculated using Equation 4-1 and sample errors were
calculated using error propagation (Table 4-2). The enrichment factor for
electrolysis is generally between 15 and 30 (Cook et al, 1998); therefore, an
enrichment value of 22.5 was chosen for typical sample enrichment. The samples
counted were very low in activity, with the calculated uncertainties for most samples
nearly indistinguishable from background. A rainfall sample from Palos Verdes has
a measured tritium activity 1.2 ± 0.97 TU which is lower than the predicted tritium
activity for rainfall at Santa Maria (2.8 to 4 TU with an average of 3.5 TU).
The highest tritium activity in the basin came from wells WW-2 and
Disgusting, 5.8 ± 2.9 TU and 4.0 ±1.0 TU respectively (Figure 4-3). Both of these
wells are located near the toe of the Abalone Cove landslide and should have the
oldest water in the basin. The measured activities of these wells are similar to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
Table 4-2. Tritium activities for wells in the
Abalone Cove and ancient landslides
Sample ID
Net Counts per
Minute -
Sample (cpm)
Net Counts
per Minute -
Standard
(cpm)
Final
Mass
(g)
Decay
Correction
Tritium
Activity (TU)
WW-1 (4-03) 0.474 (±0.159) 60.221 12.38 1.04 2.1 (±0.7)
WW-1 (4-03) U 0.082 (±0.159) 60.221 7.25 1.04 -0.4
WW-1 (9-03) -0.081 55.949 10.00 1.04 0.4 (± 0.9)
WW-2 (4-03) 0.303 (±0.156) 60.221 14.88 1.04 1.3 (±0.7)
WW-2 (4-03) U 0.048 (±0.152) 60.221 7.25 1.04 0.2 (± 0.7)
WW-2 (9-03) 0.409 (± 0.189) 45.752 7.86 1.05 2.4 (± 1.1)
WW-2 (11-03) 0.351 (±0.177) 15.846 7.25 1.03 5.8 (± 2.9)
WW-5 (2-03) 0.138 (±0.184) 45.752 7.89 1.08 0.8 (± 1.1)
WW-5 (11-03) 0.003 (±0.181) 45.752 7.25 1.03 0.01 (± 1.0)
WW-12 (11-03) 0.585 (± 0.193) 45.752 7.32 1.03 3.3 (±1.1)
WW-13 (2-03) -0.242 52.163 8.16 1.08 -1.3
WW-13 (11-03) 0.27 (±0.189) 52.163 8.25 1.03 1.4 (± 1.0)
Disgusting (2-03) 0.327 (±0.196) 55.949 9.92 1.08 1.6 (± 1.0)
Disgusting (4-03) 0.672 (±0.195) 45.752 10.00 1.07 4.0 (± 1.2)
Tolocskzo (4-03) 0.093 (±0.191) 55.949 9.84 1.07 0.5 (±1.0)
Petak (2-03) 0.074 (± 0.85) 52.163 8.52 1.08 0.4 (± 1.0)
Petak (4-03) -0.072 45.752 7.12 1.08 -0.4
Kelvin Spring (2-03) 0.524 (±0.160) 60.221 10.03 1.05 2.3 (± 0.7)
Kelvin Spring (2-03) U -0.289 60.221 7.25 1.05 -1.3
Kelvin Spring (12-03) 0.126 (±0.192) 55.949 11.07 1.03 0.6 (± 0.9)
LC4 (2-03) 0.036 (±0.155) 60.221 7.25 1.05 1.0 (±0.7)
LC4 (2-03) U 0.225 (±0.152) 60.221 12.60 1.05 0.2 (± 0.7)
MW-1 (12-02) 0.069 (±0.191) 55.949 9.92 1.09 0.3 (± 1.0)
Rain (2-04) 0.249 (±0.194) 55.949 10.25 1.02 1.2 (± 1.0)
The activity of the standard was 684.65 Bq/kg and the blank was 2.3 cpm.
Negative activities indicate samples that were below background.
The symbol U represents unernriched aliquots of sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8 2
10
8
0
CO CO CO
o
CO CO
o
o
CM
o
CM
O
CM CM CM
0
C O
> »
< 0
5
U
3
&
O.
©
C O
E
8
&
Sampling Date
Legend
■ -W W 1 ®
- Petak
+ - ww 2 a
-MW-1
► - WW 5 E -LC4
▲ - WW 12 •
- Disgusting
X - WW 13 <
- Kelvin Spring
▼ - Tolosckzo - Precipitation
Figure 4-3. Tritium activity and precipitation vs. sampling date
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Rainfall (in)
83
modem precipitation but the water from these wells should be a mixture of modem
precipitation and older water from up gradient. Therefore, these higher tritium
activities probably reflect a component of bomb produced tritium. A plot of tritium
activity vs. CFC-12 partial pressure for each well suggests a negative correlation
(Figure 4-4), expected if both tracers reflect mixing of old and young water.
The tritium activity does display a seasonal variability. Well Disgusting was
sampled in February 2003 and had a measured tritium activity of 1.6 ± 1.0 TU but
increased to 4.0 ± 1.0 TU in April 2003. A similar pattern was seen in WW-2, where
tritium activity increased from 1.3 ± 0.7 TU and 2.4 ± 1.1 TU in April and
September 2003 to 5.8 ± 2.9 TU in November 2003. There were significant rainfall
events in late February and November 2003 which may have played a role in the
dissimilar tritium activities; however, recent rainfall should not increase the tritium
activity to concentrations above natural levels. One explanation for these high
tritium activities is that these wells contain the last remnants of bomb-produced
tritium in the basin. Since these wells are located at the toe of the landslide, they
should contain the oldest water in the basin.
It is possible that well Disgusting is influenced by water flowing upward
through the slide plane. Hill (2000) measured higher temperatures in Disgusting
during 1997-1998 and concluded that this well had a deeper hydrothermal source.
The flow of warm water has been observed elsewhere in the peninsula, particularly
at White’s Point. This warm water, which must travel deeper, longer flow paths,
might contain a portion of bomb-produced tritium and should be low in CFC-12.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8 4
Although temperatures for Disgusting were not significantly higher during this study
period, the tritium activity of this well may indicate pulses of older, deeper water
supplying this portion of the basin. This water could be introduced along fractures
and joints located in the basalt bedrock which directly underlie the landslide surface
in this area of the basin.
A rainfall sample, collected in February 2003, was also measured for tritium
activity. The tritium activity of this sample was 1.2 TU (± 1.0) which is lower than
the calculated tritium activity of rainfall at Santa Maria, California. Hill (2000)
reported that the tritium activity of rainfall was 2.0 TU and 2.8 TU, which is again
lower then the Santa Maria GNIP station. Therefore, the Santa Maria tritium history
is probably higher than actual precipitation on the Palos Verdes peninsula. The low
tritium activity of wells in the basin is further indication of the lower tritium activity
in rainfall on the Palos Verdes peninsula.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
is
O'
£
S
£
Oh
t:
cd
P h
< N
r-H
I
U
PH
u
6 0 0
Rain
500
400
300
200
100
0
5 4 2 3 0 1
Tritium Activity (TU)
Legend
■ - WW 1 ▼ - Tolosckzo
► - WW 5 • -Disgusting
A - WW 12 < - Kelvin Spring
X - WW 13
- Atmospheric CFC-12 (pptv) vs Tritium Activity
on Palos Verdes pennisula through time
Figure 4-4. CFC-12 partial pressure vs. tritium activity of wells. The dashed
lines represents the atmospheric CFC-12 partial pressure (pptv) vs. tritium activity of
rainfall through time. The tritium activity of rainfall on the Palos Verdes pennisula
was calculated to be 50% of the tritium activity of precipitation at the Santa Maria,
Califormia GNIP station. The square represents the measured rainfall during this study
in Abalone Cove. The pattern correlation between samples and the reconstructed curve
suggests that water in the wells is not always well mixed.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
CHAPTER V
DISCUSSION AND SYNTHESIS
A. Introduction
The objective of the MODFLOW and MODPATH models was to constrain
the rate of groundwater flow defined by the water budgets. Several assumptions
were made in the construction and analysis of the computer models. One important
assumption is that the entire landslide complex behaves as an unconfined system.
However, landslide mechanics may have created small areas of confined flow. For
example, if a slumped block was rotated as it moved the area underneath the rotated
block and the landslide surface may behave as a confined system, with the rotation
surface acting as a confining layer. Also, some discontinuous beds may exist within
the slide mass, producing confined slow. On the whole, this assumption of
unconfined flow should be true since no continuous confining bed exists within the
landslide complex.
A second key assumption is the amount of subsurface inflow through the
head scarp of the ancient landslide. Hill (2000) estimated this inflow to be anywhere
between 8 and 118 acre-ft/year. With such a wide range in flow rates, this could be a
possible source of error in the models. In the model, an intermediate value of 39.5
acre-ft/year was chosen. In addition, Hill (2000) predicted that a large portion of
rainfall directly recharges the groundwater basin in Altamira Canyon (146 acre-
ft/year). Since Hill’s study was completed during an El Nino year, this number was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8 7
scaled down to 51 acre-ft/year (see Chapter 2 for details). The amount of surface
runoff directly recharging the groundwater basin should be dependent upon the
amount and rate of precipitation during a rainfall event, but the magnitude of the
head scarp infiltration has some uncertainty. Finally, the MODFLOW model
assumed that no flow occurs across the slide plane.
B. Groundwater Mixing Model
The validity of the MODFLOW model can be tested by using it to predict
tracer distributions and comparing predictions and observations. This was done by
simplifying the MODFLOW output into a 1 -D multibox numerical model. The
objectives were to determine to ( 1 ) estimate the concentration of the hydrologic
tracers in water entering the basin through the head scarp of the ancient landslide, (2 )
determine if the estimated budgets for subsurface inflow and surface percolation
along Altamira canyon are valid and (3) evaluate the possibility of vertical flow
across the lower boundary. This numerical model assumes complete vertical mixing
of water masses in each cell. The MODFLOW model was used as the framework for
the numerical model. Column 18 of the MODFLOW simulation was selected to
represent downslope transport (see Chapter 2 for details).
In transforming to a 1-D model, the assumption was made that tracer
concentration varies only in the downhill direction and not in the direction
perpendicular to the flow axis. A second assumption is that the water in each box is
well mixed. Alternatively this assumption is satisfied if flow velocity is uniform
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
with depth and samples collected by wells have integrates concentrations from all
depths.
The concentration of the tracer in each cell of the model is controlled by the
rate of subsurface inflow from adjoining cells, subsurface outflow and the net
percolation rate (Figure 5-1). As previously mentioned, the tracer concentration
across each row is assumed to be equal so that (Equation 5-1):
C1H = c u = CiJ+l = C, (Equation 5-1)
where C is the concentration of the tracer, i is the row number and j is the column
number. This assumption should be reasonable because wells on the east and west
sides of the basin have similar CFC-12 concentrations.
The change in the concentration of the tracer in one box of the model during
one time step can be calculated (Equation 5-2):
A c , = [ < a v _ , x c , > + (a .,., x c , > + ( q M j x c ,_ i > + yc„ > - « ? „ yc, \
(Equation 5-2)
- 2
where Q, is the rate of subsurface inflow from the adjoining cell (L /t), Qperc is the
- 2
net rate of percolation (L /t), Ca tm is the concentration of the tracer in percolating
rainfall, St is the length of the time step (t), and V , is the volume of water in the cell
(L ). The rate of inflow into the cell must equal the rate of outflow (Equation 5-3).
Qu ~ Q u -1 ~ Qij* = Qt-hj + Q perc (Equation 5-3)
Using this water balance, Equation 5-2 reduces to: (Equation 5-4)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
Top View
/
1
c
r d
P Uj
J A
r
j ►
Cross-sectional View
Net QP e rc
'inflow
i+ 1
‘Outflow1
8y
Legend
QPerc = Rate of percolating distributed recharge (L3/t)
Qoutflow = Rate of subsurface outflow (L3/t)
Qinflow = Rate of subsurface inflow (L3/t)
i - Model row number
j = Model column number
Figure 5-1. Schematic diagram of groundwater flow in the numerical model
illustrating boxes i- 1 to i+ 1 .
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9 0
(.c p e r c -C ,) a (Equation 5-4)
V r i y \ i J
Dropping the subscript j, the rate of inflow from uphill, as calculated by
MODFLOW, is:
( X - ,+ X )
2
lY h: , - h , \ X
— ---- (z,x,) (Equation 5-5)
A #
where Q, is the rate of flow (L3/t), K is hydraulic conductivity (L/t), i is the cell
number, h is the hydraulic head (L), z is the saturated thickness (L), y is longitudinal
distance (L) and x is the latitudinal distance (L). The saturated thickness of each cell
was determined using the groundwater elevations calculated by the transient state
MODFLOW model (Table 5-1). The numerical model was run for a 40-year period
with time steps of 0.1 years. Computations were done in an Excel spreadsheet.
The fraction of each cell that is replaced by autochthonous water in each time
step can be calculated using the net rate of rainfall percolation:
A similar equation can be used to define the fraction of allocthonous water. As
groundwater flows towards the ocean, young water is continually added. In this
discussion, allocthonous water is always considered to be water that entered the cell
through subsurface inflow from the uphill cell. Any water that enters the cell
directly through percolation will always be referred to as autochthonous water.
(Qpercm
(Equation 5-6)
autochonous s i
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5-1. Transient state MODFLOW calculated
groundwater elevations and flow rates
Bottom
Total Flow
Cummulative
Row Elevation Groundwater
Volume Rate (Q)
Residence
No. (ft) Elevation (ft) ( 105 ft3 ) (103 ft3/d) Time (years)
8 541 578.0
9 447 563.2 1.9 1.5 0
1 0 355 522.2 2.7 0.5 2
1 1 293 495.7 3.2 0 . 6 4
1 2 248 464.1 3.4 0 . 8 5
13 207 428.2 3.5 0.9 6
14 172 394.5 3.5 0.9 7
15 145 365.0 3.5 0 . 8 8
16 127 331.2 3.3 0.9 9
17 1 1 2 291.2 2.9 0.9 1 0
18 97 257.4 2 . 6 1 . 2 1 1
19 79 238.4 2.5 1 . 0 1 1
2 0 57 219.5 2 . 6 1.9 1 2
2 1 34 204.7 2.7 1.5 1 2
2 2 13 183.6 2.7 1 . 1 13
23 9 168.8 2.5 0.9 14
24 -28 152.0 2.9 2 . 2 14
25 -41 124.6 2 . 6 1 . 1 15
26 -48 90.8 2 . 2 0 . 8 15
27 -54 71.8 2 . 0 1 . 2 16
28 -60 52.9 1 . 8 1 . 8 16
29 - 6 6 46.5 1 . 8 3.2 16
30 -67 42.3 1.7 3.9 16
31 -65 40.2 1.7 2 . 0 17
32 -67 38.1 1.7 1.9 17
33 -70 36.0 1.7 1.9 17
34 -69 31.8 1 . 6 3.7 17
35 -63 29.7 1.5 1 . 8 17
36 -51 25.4 1 . 2 3.2 18
37 0 2 1 . 2 0.3 2 . 6 18
Cell dimensions x = 138 ft (40 m), y = 165 ft (50 m)
Residence time is defined as the length of time water resides in a
cell before it is replaced.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
Flow in these cells is dominated by longitudinal flow (see Figure 2-15). The
concentration of the tracers in each interior cell was calculated at each time step:
C,), =OW/ocM),XCM)( - C ,_x) + {fauto -C , ),_.) + <:,),_,
(Equation 5-7)
where fauto is the fraction of the autochthonous water from uphill, and falloc is the
fraction of allocthonous water. The concentration of each tracer is also dependent
upon the concentration of old water. The following boundary condition was
established for the uphill contribution to the model:
C,), =(/a/toc,)O (C 0 - C ,),_,) + (fauto,),)(C„. - C ,) W) + C,)M)
(Equation 5-8)
where Co is the concentration of the tracer in subsurface inflow recharging at the
head scarp of the ancient landslide.
The concentration of allocthonous water entering from the basin uphill
through the head scarp of the ancient landslide plays an important role in the
concentration of the tracers in the basin. The calculated concentrations were
compared to the observed concentration of the tracers in nearby wells (WW-1 and
WW-12). If Co is equal to 0 pptv is assumed (making this water last recharged prior
to 1940), the partial pressure of cells 20 and 24 (near wells WW-1 and WW-12) are
too low, indicating that old water must have a CFC-12 partial pressure greater than 0
pptv. If water entering the basin is always in equilibrium with the atmosphere
(having an age of zero years), then the predicted partial pressure is too high (see
Figure 3-1 for CFC-12 atmospheric concentrations). Therefore, the concentration of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
water entering through the head scarp of the ancient landslide must lag atmospheric
concentrations, but by only a few years. The model was run with several different lag
times. The numerical model was run for 50 years, using values of Ca tm defined by
Figure 3-1, assigning a Co concentration equivalent to ages of 1, 4, and 5years.
Calculated and measured values are consistent if the CFC-12 lag time for water
flowing into the basin is 1-5 years (Figure 5-2).
This inherited age for the basin is dependent upon the rate of subsurface
inflow and surface runoff percolation along Altamira canyon. The proportional
contribution of each of these inputs has been estimated in Chapter 2. If all of the
recharge at the head scarp is contributed by subsurface inflow, then its age should be
the median age of water in the uphill groundwater basin, whereas if all the water is
recharged from surface runoff in Altamira canyon then it will have an age of zero
years. The amount of subsurface inflow can be constrained knowing the area of the
uphill groundwater basin. The uphill groundwater basin boundary, which is assumed
to be the same as the surface water boundary, has a total area of 356 acre (1.44 x 106
m2 ). Since the Palos Verdes peninsula is an anticlinal structure, the only water
contributing to this basin must be percolation of rainfall and domestic water (Figure
5-3). The rate of percolation in the uphill basin was calculated assuming input per
unit area is the same as those calculated for the landslide complex (Table 5-2).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(a )
600
500
WW-12
a 400 4 year lag in C(
WW-1
C M
5 year lag inC.
q 300
T J
200
Disgusting
Kelvin Spring
100
9 11 13 15 17 19 21 23 25 27 29 31 33 35
94
Row Number
&
>
I
E
3
••P
5
Disgusting
4
■W W -12
3
ww-i
2
Kelvin Sprinj
1
0
9 11 13 15 17 19 21 23 25 27 29 31 33 35
Row Number
Figure 5-2. Numerical model predicted CFC-12 partial pressures and tritium activity.
The bars represent sample ranges, (a) CFC numerical model results, (b) Tritium
numerical model results. Well WW-12 shows slightly higher CFC-12 and tritium
concentrations than WW-1. This could be the result of differences in the screening
intervals of these two wells.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Q perc Groundwater
Divide
400 ft
560 ft
KCS
■perc
WW-12.
!perc
WW-1
Portuguese Tuff Catalina Schist
Disgusting
Basalt
Figure 5-3. Schematic diagram of hypothesized groundwater recharge and flow in the basin. Question marks indicate that the
well depths are unknown. The geometry of the basalt sill is somewhat uncertain. The Portuguese Tuff is not a continous unit
within the basin.
96
Table 5-2. Estimated rate of input to uphill groundwater basin.
Rate
C I O " 4 ft/d)
Rate
(acre-ft/year)
Inputs
Rainfall Percolation 5.6 72.8
Water Main Losses 0.58 7.5
Sewage 3.15 40.9
Landscape Irrigation 2 . 1 2 27.5
Outputs
Deep Rooted 3.0 38.9
Evapotranspiration
Net Input 8.45 109.8
If all of the water entering the uphill groundwater basin flows directly into
the ancient landslide, then the highest possible flow rate is 110 acre-ft/year. In the
opposite extreme, all of this water may flow deeper and none of it enters the
landslide complex. Therefore, the range in possible subsurface flow across the head
scarp of the ancient landslide is 0-109.8 acre-ft/year. An intermediate value of 40
acre-ft/year was used in the MODFLOW model. The age of this subsurface flow can
be estimated using the total area of the basin, porosity and saturated thickness
(Equation 5-9).
age = (Equation 5-9)
Q p e r c
where A is the basin area (L2 ), r\ is the porosity (always assumed to be 0.07), z is the
saturated depth (L) and Q perc is the rate of percolation (L3/t). The thickness of the
Altamira shale unit above the Portuguese tuff is approximately 400 ft (121.9 m). If
this entire thickness is saturated with water, the age of water would be 91 years old at
the bottom of this basin and zero at the top. Therefore, the average age of water
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
entering the basin would be 45 years. It is unrealistic to have this entire unit
saturated with water, so this provides an upper limit on the age of subsurface inflow
into the landslide complex. Groundwater elevations measured in monitoring wells
just down gradient of the head scarp of the ancient landslide indicate a saturated
thickness of approximately 70 ft (21.3 m). If the uphill basin is similar, the age for
water at the bottom of the uphill basin is 16 years, making the average age of
subsurface inflow 8 years. This calculation was based on a very simplified view of
the geology in the area; however, it provides a general picture of the hydrology in the
area.
The transient numerical model indicates that water recharging the first cell in
the model is 1-5 years old (an average of 3 years). This implies there must be a
substantial contribution of recent water entering the system through surface runoff
along Altamira canyon, consistent with a mixture of 40% inflow and 60% infiltration
from Altamira canyon at the head scarp. This is in good agreement with the ratio of
these inputs used in MODFLOW. Only a small portion of the water from the uphill
basin enters the landslide. The rest of the water percolating in the uphill basin must
flow beneath the landslide and circulate deeper.
The CFC data is evidence that little of this deep circulating water enters the
basin further downslope, although some deep water does enter the basin at Kelvin
Spring, resulting in its low CFC-12 partial pressure and slightly elevated tritium
activity.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
If water takes this deep path and is isolated from rainfall percolation, it
should appear relatively old. The average age of this deep circulating water can be
estimated using Equation 5-9. The depth of Altamira shale beneath the landslide
should be approximately 560 ft (Figure 5-3). If the subsurface inflow rate used in
the MODFLOW and numerical models is valid, then approximately 60 acre-ft/year
of the water from the uphill basin flows into this deeper basin. Therefore, the
average age of this deep water should be 125 years. Water of this age should have
no CFC-12 or tritium; consequently, it is not possible to verify this age with the
tracers used in this study.
The tritium activity of groundwater throughout the basin is similarly
controlled by the activity of allocthonous water (Figure 5-2). The most important
aspect of the concentration of allocthonous water in this case is whether or not it
contains a portion of the bomb-produced tritium spike. In the numerical model, if
the old water never contains bomb-produced tritium, then the model predicts a
tritium activity of 1.5 TU near WW-1 which is close to the measured activity (1.2
TU). This indicates that it is possible to produce some of the observed activities in
the basin without the influence of bomb-produced tritium. This run of the model
predicts tritium activities near the toe of the landslide to be approximately 2.8 TU.
Again, this is consistent with some of the measured activities in wells near the toe.
Well WW-2 was sampled in April and September 2003 and had activities of 1.3 and
2.4 TU, respectively. In February 2003, Disgusting had a measured tritium activity
of 1.4 TU. However, both of these wells have displayed higher concentrations on
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
other sampling dates. Well WW-2 has a measured tritium activity of 5.8 TU in
November 2003, while Disgusting had a measured activity of 4.6 TU in April 2003.
Such high tritium activities must indicate a component of bomb-produced tritium,
since natural activities should be less than 4 TU. The difference in tritium activities
in these wells is perhaps seasonably variable.
C. Conclusions
Merriam (1960) and Ehlig (1992) concluded that water was the short-term
catalyst for mass movement in the Abalone Cove area. The purpose of this study
was to constrain the sources of water to the landslide through the use of hydrologic
computer models, geochemical tracers and numerical modeling. Simulation of flow
in the ancient landslide complex was done using the program MODFLOW.
Assumptions of this model were tested using CFC-12 and tritium as tracers. Key
assumptions of the water budget were ( 1 ) the magnitude of recharge from infiltration
of surface inflow into tension cracks in the head scarp, (2 ) subsurface flow from the
uphill basin at the head of the slide, and (3) the base of the slide acts to prevent
vertical flow. The conclusions of this study are:
• The dewatering system activated to stabilize the landslide has greatly reduced
groundwater elevations in the basin and increased groundwater velocities.
According to the MODFLOW model, groundwater elevations have decreased
100-150 ft (30-46 m) and groundwater velocities have doubled in the toe of
the Abalone Cove landslide.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
• Hydraulic conductivity is highly variable throughout the basin. Dominated
by landslide mechanics, the hydraulic conductivity varies between 0.15 ft/d
(0.04 m/d) and 10 ft/d (3 m/d). The Abalone Cove landslide has high
hydraulic conductivity values, presumably because the toe area is dominated
by radial fractures which act as conduits for groundwater flow. Areas of low
hydraulic conductivity are assumed to be areas that are either intact slump
blocks or areas where fractures have been filled in with fine-grained material.
• The estimated values used in the MODFLOW model for surface runoff along
Altamira canyon and subsurface inflow across the head scarp of the ancient
landslide are valid based on the numerical model.
• MODPATH predicts that water entering the basin near the head scarp of the
ancient landslide takes approximately 28 years to flow through the basin.
Since the hydrologic conductivity is high in the Abalone Cove landslide,
water passes through this portion of the basin in 1 - 2 years.
• Most water pumped from the basin is relatively recent rainfall, with CFC
ages of 8 - 2 0 years.
• The concentrations of hydrologic tracers in the basin are seasonably variable,
with the highest variability in wells directly above the head scarp of the
Abalone Cove landslide.
• The low CFC-12 and tritium activity of Kelvin spring indicate a source of old
water entering the landslide basin. This water is possibly derived from
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
beneath the Portuguese tuff member, indicating some vertical flow may
occur.
• A portion of the water from the uphill basin flows into the landslide through
the head scarp of the ancient landslide. The remaining water flows beneath
the landslide. Even though it is not possible to further constrain the exact
proportions, this deep circulating water can possibly exert a strong hydraulic
pressure on the landslide surface, increasing the potential for mass
movement.
There are some areas of uncertainty in this study. As previously mentioned,
CFCs are degraded in anoxic conditions. Consequently, it is unclear if the low CFC
partial pressures of at Kelvin spring and well Disgusting represent old water or
simply degraded CFC concentrations. The screening interval of the wells in the
basin are also not well constrained. During this study it was assumed that the wells
screened the approximately the same intervals and that water in the wells was the
integrated concentration of water from all depths of the screened interval. Finally,
the MODFLOW model assumed there was no vertical flow from beneath the
landslide. MODPATH predicts that water flows through the basin on the order of 2-
28 years. Water of this age does have large concentrations of tritium. The high
tritium activity of well Disgusting cannot be explained simply by water that entered
the landslide surface through rainfall percolation or inflow along the head scarp of
the ancient landslide. Therefore, there must be some vertical flow through the base
of the landslide along the toe.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
D. Recommendations
All of the dewatering wells located in the Abalone Cove and ancient landslides
only pump water from the slide mass. Few, if any of these wells, are drilled deeper
than the slide plane. According to the water budgets, MODFLOW models and
hydraulic tracers, most of the water contributing to the landslide mass is recent
precipitation. A substantial portion of water recharged in the uphill groundwater
basin must flow beneath the slide plane. This flux of water may exert a large
hydraulic pressure on the slide plane, inducing mass movement. A possible solution
is to develop more dewatering wells within the slide mass and above the head scarp
of the ancient landslide. These new wells should screen the aquifer beneath the
Portuguese tuff in an effort to remove water and reduce the pressure exerted on the
slide plane by this large volume of water.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
REFERENCES
Anderson M.P. and Woessner, W.W. 1992. Applied Groundwater Modeling:
Simulation of Flow and Advective Transport. Academic Press (San Diego).
381 pp.
Beyerle, U., Aeschbach-Hertig, W., Hofer, M., Imboden, D.M., Baur, H., and
Kipfer, R. 1999. Infiltration of river water to a shallow aquifer investigated
with 3H/3He, noble gases and CFCs. Journal of Hydrology, v. 220. no. 3-4.
pp. 169-185.
Brown, R.M. 1961. Hydrology of tritium in the Ottawa valley. Geochimica et
Cosmochimica Acta. v. 21. pp. 199-216.
Brown, R.M., and Grummitt, W.E. 1956. The determination of tritium in natural
waters. Canadian Journal of Chemistry, v. 34. pp. 220-226.
Bullister, J.L. and Weiss, R.F. 1988. Determination of CCI3F and CCI2F2 in
seawater and air. Deep Sea Research, v. 35. no. 5. pp. 839-854.
Busenberg, E. and Plummer, L.N. 1992. Use of chlorofluorocarbons (CC13 F and
CCI2F2) as hydrologic tracers and age-dating tools: the alluvium and terrace
system of central Oklahoma. Water Resources Research, v. 28. no. 9.
pp. 2257-2283.
Carlston, C.W. 1964. Tritium-hydrologic research: some results of the U.S.
Geological Survey research program. Science, v. 143. pp. 804-806.
Clark, J.F., Stute, M., Schlosser, P, Drenkard, S., and Bonani, G. 1997. A tracer
study of the Floridan aquifer in southeastern Georgia: Implications for
groundwater flow and paleoclimate. Water Resources Research, v. 33.
no. 2. pp. 281-289.
Cook, G.T., Passo, C.J., Carter, B. 1998. Environmental liquid scintillation
analysis. In: Handbook of Radioactivity Analysis. Academic Press
(San Diego), pp. 331-386.
Cook, P.G., Solomon, D.K., Plummer, L.N., Busenberg, E., and Schiff, S.L. 1995.
Chlorofluorocarbons as tracers of groundwater transport processed in a
shallow, silty sand aquifer. Water Resources Research, v. 31. no. 3.
pp. 425-434.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
Cotton, Shires & Associates. 2001. Technical review geologic/geotechnical data:
Abalone Cove landslide and Zone 2 areas. Los Gatos, California. 52 pp.
Craig, H. and Lai, D. 1961. The production rate of natural tritium. Tellus.
v. 13. pp. 85-105.
Doney, S. C., Glover, D.M. and Jenkins, W.L. 1992. A model function of the global
bomb tritium distribution in precipitation 1960-1986. Journal of Geophysical
Reseacrh. V. 97. no. C4. 5481-5492.
Ehlig, P.L. 1992. Evolution, mechanics and mitigation of the Portuguese Bend
landslide Palos Verdes peninsula, California. In: Engineering Geology
Practice in Southern California. Pipkin. B.W. and Proctor, R J. (eds). Star
Publishing Company (Belmont, CA). pp. 531-553.
Erikson, E. 1965. An account of the major pulses of tritium and their effects in the
atmosphere. Tellus. v. 17. no. 1. pp. 118-130.
Environmental Simulations, Inc. 1999. Guide to using Groundwater Vistas.
Version 2.4. pp. 251.
Environmental Simulations International. 2000. Groundwater Vista
(Student Version 3.40). Developed by J. Rumbaugh and D. Rumbaugh.
http://www.groundwatermodels.com/software/SoftwareDesc.asp7software_d
esc_id= 19&software_id=6
Faltings V., Harteck P. 1950. Der Tritium Gehalt Der Atmosphere. Z. Naturforsch.
v. 5a. pp. 438-439.
Haskell, E.E., Leventhal, J.S., and Bianchi, W.C. 1966. The use of tritium to
measure the movement of groundwater toward irrigation wells in western
Fresno County, California. Journal of Geophysical research, v. 71. no. 16.
pp. 3849-3859.
Hester, N.E., Stephens, E.R. and Taylor, O.C. 1974. Fluorocarbons in the Los
Angeles basin. Journal of the Air Pollution Control Association, v. 24.
no. 6 . pp. 591-595.
Hill, C.A. 2000. A geochemical and hydrological assessment of groundwater in the
Portuguese Bend Landslide, California. [Ph.D. thesis]: Los Angeles,
University of Southern California. 139 pp.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
Ho, D. T., Schlosser, P., Smethie, W.M. and Simpson, H. J. 1998. Variability in
atmospheric chlorofluorocarbons (CCI3F and CCI2F2 ) near a large urban area:
implications for groundwater dating. Environmental Science and
Technology, v. 32. pp. 2377-2382.
IAEA/WMO. 2001. Global Network of Isotopes in Precipitation. TheGNIP
Database. Accessible at: http://isohis.iaea.org
Johnson, R.L., Pankow, J.F. and Cherry, J.A. 1987. Design of a ground-water
sampler for collecting volatile organics and dissolved gases in small-diameter
wells. Ground Water, v. 25. pp. 448-454.
Kaufman, S. and Libby, W.F. 1954. The natural distribution of tritium. Physical
Review, v. 93. no. 6 . pp. 1337-1344.
Kehew, A.E. 1998. Geology for Engineers and Environmental Scientists. Second
Edition. Prentice Hall (Edgewood Cliffs, NJ). 574 pp.
Kiessling, E. 1963. Field trip to Palos Verdes Hills. Mineral Information Service,
v. 16. n o .ll. pp. 9-14.
L’Annunziata. MichealF. 1998. Nuclear radiation, its interaction with matter and
radioisotope decay. In: Handbook of Radioactivity Analysis. Academic
Press (San Diego), pp. 1 -74.
Libby. W.F. 1946. Atmospheric helium three and radiocarbon from cosmic
radiation. Physical Review, v. 69. pp. 671-672.
Lovely, D.R. and Woodward, J.C. 1992. Consumption of ffeons CFC-11 and
CFC-12 by anaerobic sediments and soils. Environmental Science and
Technology, v. 26. pp. 925-929.
Lucas, L.L. and Unterweger, M.P. 2000. Comprehensive review and critical
evaluation of the half-life of tritium. Journal of Research of the National
Institute of Standards and Technology, v. 105. pp. 541-549.
Merriam, R. 1960. Portuguese Bend landslide, Palos Verdes Hills, California.
Journal of Geology, v. 6 8 . pp. 140-153.
Ostlund, H.G., and Dorsey, H.G. 1977. Rapid electrolytic enrichment and hydrogen
gas proportional counting of tritium. In: Low-Radioactivity Measurements
and Applications. Slovenske Pedagogicke Nakladatelstvo (Bratislava).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
Pipkin, B.W. and Nash, K.W. 1967. Field trip guide to Baldwin Hills and Palos
Verdes Hills, Los Angeles; pre-convention field trip no. 6 in conjunction with
the annual meeting of the AAPG and SEPM. pp. 8-18.
Plummer, L.N. and Busenberg, E. 2000. Chlorofluorocarbons. In: Environmental
Tracers in Subsurface Hydrology. Cook, P.G. and Herczeg, A.L. (eds).
Kluwer Academic Publishers (Boston), pp. 441-478.
Proffer, K. A. 1992. Ground water in the Abalone Cove Landslide, Palos Verdes
Peninsula, southern California. In: Landslides/Landslide Mitigation.
Slosson, J. E., Keene, A. G., Johnson, J. A. (eds). The Geological Society of
America (Boulder), pp. 69-82.
Rozanski, K. and Groning, M. 2004. Quantifying uncertainties of tritium assay in
water samples using electrolytic enrichment and liquid scintillation
spectrometry. IAEA TECHDOC in preparation. Available at:
http://www.iaea.org/programmes/rial/pci/isotopehydrology/docs/intercompari
son/ViennaH3-v 12.htm
Solomon, D.K., and Sudicky, E.A. 1991. Tritium and helium 3 ratios for direct
estimation of spatial variations in groundwater recharge. Water Resources
Research, v. 27. no. 9. pp. 2309-2319.
Spitz, K. and Moreno, J. 1996. A practical guide to groundwater and solute
transport modeling. Oxford Unoversoty Press, pp. 251.
Stencel, J.R., Griesbach, O.A., Ascione, G., Elwood, S.M., and Frankenfield, R.A.
1995. Pratical aspects of environmental analysis for tritium using enrichment
by electrolysis. Radioactivity and Radiochemistry, v. 6 . no. 2. pp. 40-49.
Stewart, G.L. and Farnsworth, R.K. 1968. United States tritium rainout and its
hydrologic implications. Water Resources Research, v. 4. no. 2.
pp. 273-289.
Szabo, Z., Rice, D.E., Plummer, L.N., Busenberg, E., Drenkard, S, and Schlosser, P.
1996. Age dating of shallow groundwater with cholorfluorcarbons,
tritium/helium 3 and flow path anaylsis, southern New Jersey coastal plain.
Water Resources Research, v. 32. no. 4. pp. 1023-1038.
U.S. Geological Survey. 1981. Redondo Beach, California [map], 1:24,000. 7.5
Minute Series. Washington D.C. USGS.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
U.S. Geological Survey. 1981. San Pedro, California [map]. 1:24,000. 7.5 Minute
Series. Washington D.C. USGS.
U.S. Geological Survey. 1981. Torrance, California [map]. 1:24,000. 7.5 Minute
Series. Washington D.C. USGS.
Warner, M.J. and Weiss, R.F. 1985. Solubilities of chlorofluorocarbons 11 and 12
in water and seawater. Deep Sea Research, v. 32. pp. 1485-1497.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
APPENDIX A
SAMPLING PROCEDURE FOR TRITIUM AND CFC
WATER SAMPLES
I. Wells with pumps installed
a. Items needed for tritium samples:
500 ml glass bottles (cleaned)
1 length of 5/8" ID Tygon tubing (approximately 1 foot long)
Masking tape
Black marker
Paper towels
500 ml plastic beaker
Temperature, conductivity, salinity, dissolved oxygen meter
b. Tritium sampling:
1. Turn on pump if the well is not already pumping.
2. Place one end of the Tygon tubing in the faucet outlet of the line
and place the other end at the bottom of a glass bottle.
3. Turn on faucet and allow sample to fill the bottle. Let 2-3 sample
volumes fill the bottle before removing. Make sure the cap is
tightly secured.
4. Dry the outside of the bottle and label sample.
5. After the sample has been taken, place the end of the Tygon
tubing into the plastic beaker.
6 . Allow 2-3 sample volumes to fill the flask before removing.
7. Use the meter to measure and record the temperature, salinity,
conductivity and dissolved oxygen of the sample.
c. Items needed for CFC sampling:
3/8" OD copper tubing cut into approximately 1 meter length
pieces
3/4" to 3/8" OD brass reducing union
3/4" OD brass female hose coupling
3/8" OD brass Swagelock nuts
3/8" OD brass Swagelock ferrules (front and back)
3/8" OD brass Swagelock unions
3/8" OD nylon Swagelock ferrules (front and back)
Aluminum plates (Figure A-l)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
(N
Top View
L l l u j
^ 5 /8"
)o
' U
1/ 2 ”
[-1 3/4"- |
) o o (
1/4"-
1
C M
3"-
■H
Side View 1
V/
00
Side View 2
i — r 3Z~T
1/4"
1/ 2"
1
Figure A -l. Aluminum plates for CFC sampling. The gray area on the top view
indicates the location of the stainless steel clamps.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
Stainless steel clamps (Figure A-2)
Allen wrenches
Socket wrench (1/2" head)
Adjustable wrenches
Teflon tape
Tubing cutter
Black marker
d. CFC sampling:
1. Before entering the field, all copper fittings and tubing must be
cleaned. All pieces must be flushed with hexane and allowed to
diy. Tubing should be flushed with approximately 5-10 ml of
hexane.
2. Once all copper tubing is cleaned and dry, it must be cut into ~1 m
length pieces and straightened. The copper tubing must be as
straight as possible in order to fit into the sample extraction
system. When carrying copper tubing throughout the sampling
day take great care to ensure that all pieces are straight.
3. Prepare the piece of copper tubing that will connect the pup outlet
faucet to the sampling line (referred to as the faucet assembly
from this point on) (Figure A-3). Connect the 3/4" brass hose
coupling to the 3/4" to 3/8" brass reducing union. Connect the
3/4" to 3/8” brass reducing union to a male 3/8" brass union.
Connect a length (approximately 1 m) of copper tubing to the 3/8"
brass union with a 3/8" brass nut and brass ferrule (front and
back). Connect the other end of the copper tubing to another 3/8"
brass union with a 3/8" brass nut and ferrule (front and back).
Make sure that all connections are secured with teflon tape and
tightened with adjustable wrenches. The faucet assembly should
last the entire day.
4. Attach the faucet assembly to faucet. Make sure this connection
is tight.
5. Attach two steel clamps onto an aluminum plate. Prepare two
plates per well for duplicate samples.
6 . Open the clamps.
7. Slide a piece of copper tubing into the plates. Make sure that 3-4
inches of copper tubing extends outside of the clamps on either
end of the plate.
8 . Finger tighten the clamps. Make sure that the clamps are
tightened evenly.
9. Attach a brass nut and nylon ferrule (front and back) to one end of
the sample line.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
I ll
Side View 1
o p e n in g
, , r -i □
t
1/ 2 "
Top View
— 3 3/8" -
1 1/4"
1/16'
Side View 2
T
i
y
t Specs: T
" x 1 1/4" I
Bolt Specs
3/8
3 /8 “
U
B ottom View
1/4"
1 /4 " . l Q ( p )
1/ 2 "
I
( °) o
| - 7 /8 " “
Figure A-2. Stainless steel clamps for CFC sampling.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
112
3/8" Swagelock
brass union
3/4" Female
hose coupling
3/8" Copper
tubing
3/8" Swagelock
brass nut
3/4" to 3/8"
brass reducing
union
Figure A-3. Photograph of the faucet assembly used for CFC sampling.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
113
10. Attach sample line to the faucet assembly. Tighten this
connection with wrenches. (See Figure A-4 for a photograph of
the entire assembly.)
11. Turn on pump (if it is not already on) and open faucet.
12. Flush the assembly for 1-2 minutes before sampling. While
flushing, lightly tap the faucet assembly and sample line in order
to remove any air bubbles that may have accumulated on the
interior walls.
13. Start sampling by tightening the clamp at the outlet end of the
sample line with the socket wrench. Tighten the clamp evenly on
both sides and as fast as possible. Make sure the clamps are
securely tightened. If the clamps are not tight the sample will be
lost.
14. Once the first clamp is tightened, work your way down the sample
line tightening clamps until all clamps are tightened.
15. Turn off faucet and remove assembly.
16. Remove sample line from the faucet assembly and remove the
clamps from the plates.
17. Using the tubing cutter, separate the two samples by cutting the
copper tubing in the middle. Again, make sure that the ends of the
copper tubing are as straight as possible.
18. Dry the copper tubing and label the samples. As a rule of thumb,
always label the sample at the outlet end of the sample line as # 1
and the sample closet to the well as #2 .
19. Store samples with clamps until analysis. Samples can be stored
for approximately 3-4 months in the copper tubes.
II. Wells without pumps installed
a. Items needed for both tritium and CFC sampling:
Deionized water (DIW) (30 ml)
Ultra High Purity (UHP) N 2 gas
3-way valve
Sampler (Figure A-5)
2 plastic syringes (300 ml each)
2 lengths of 1/4" Tygon tubing (approximately 20 cm each)
Batteries
Masking tape
Black marker
Water depth sounder
30 ml ground glass syringe with luerlock metal tip
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure A-4. Photograph of the complete assembly of aluminum plates, stainless
steel clamps and copper tubing for CFC sampling.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Selenoid (See side detail)
115
Stainless steel spring
Stainless steel rod
Tritium syringe (150 ml)
Detail of solenoid before release
Solenoid
|Rod
r
P i/ *
^ Ball
bearings
Guide wires Detail of solenoid after release
^ Tritium syringe (150 ml)
Pin— L I
Ball
bearings
-Rod
CFC syringe (30 ml) Fi8 ure A ' 5 SamPler designed for
non-pumping wells.
The sampler is fixed to an extension cord
( OD) by a Chinese finger. The cord is wired
to the solenoid which is potted.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
116
b. Tritium and CFC sampling:
1. Before entering the field a degassed glass syringe must be
prepared. Attach a 3-way valve to the end of the syringe. Fill the
syringe with deionized water. Bubble UHP N2 into the syringe.
Shake the syringe and coalesce all bubbles into one large bubble.
Expel the bubble from the top of the syringe.
2. Repeat step 1 approximately 3-4 times per syringe. Close the 3-
way valve. Store the syringe in a bucket of water until the time of
sampling. Glass syringes are gas permeable and water samples
can only be stored for 24 - 48 hours.
3. With a water depth sounder, measure and record the depth of
water in the well.
4. Assemble sampler, making sure the plastic syringes are
completely empty and the plungers are at the bottom. Pull the
plungers of the plastic syringes up and down a few times to ensure
that sediment has not accumulated in the syringe and might
prevent the syringes from engaging.
5. Attach the degassed glass syringe to the sampler.
6 . Rig the sampler springs. Move the stainless steel rod down to the
top most syringe, engage the ball bearing by pulling down on the
pin, and attach the spring to the threaded handle of the top most
syringe. Make sure that all guide wires for the syringes are within
the structure of the sampler and they are not snagged on any
screws or twisted.
7. Lower the sampler into the well. You will hear the sampler hit the
water. Make sure that the sampler is lowered well below the
water level in the well.
8 . Connect one lead of the extension cord to the batteries by twisting
the wires together. Connect the other lead of the extension cord to
the batteries by touching the two wires together and holding for a
few seconds. This will activate the solenoid and release the ball
bearing. Once the ball bearing is released, the spring will pull the
plungers to the end of their guide wires and fill the syringes.
9. Allow the syringes to fill with sample. This takes approximately
2-3 minutes.
10. Remove sampler from the well. Be careful not to jostle the
sampler too much when pulling it from the well.
11. Close the 3-way valve on the end of the glass syringe and remove
it from the sampler. Label with masking tape and place in bucket
of water.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
117
12. Attach one length of 1/4" OD Tygon tubing to the end of each
plastic syringe. Place the other end of the Tygon tubing in the
bottom of a glass bottle.
13. Transfer sample from the syringes to the bottle.
14. Dry bottle and label.
III. Sampling from Kelvin Spring
a. Items needed for both tritium and CFC sampling:
24" x 1/4" OD stainless steel push point sampler
1/4" Tygon tubing (approximately 3 cm in length)
3-way valves
30 ml ground glass syringe
Masking tape
Black marker
500 ml glass bottle
500 ml plastic beaker
Shovel
Temperature, conductivity, salinity, dissolved oxygen meter
3/8" OD copper tubing
1/4" OD copper tubing
1/8" OD copper tubing
Deionized water (DIW) (30 ml)
Ultra High Purity (UHP) N2 gas
Aluminum plate (Figure A-l)
Stainless steel clamps (Figure A-2)
1/4" to 3/8" UltraTorr
3/8" Tygon tubing approximately 1-2 m in length
b. Tritium sampling:
1. Find an area where water is flowing from the ground.
2. With the shovel, create a depression in which to place the glass
bottle.
3. Allow flow from the spring to fill bottle.
4. Dry and label.
5. In the same depression, place the plastic beaker and allow water to
fill.
6 . Record the temperature, dissolved oxygen, salinity and
conductivity of the spring water with the meter.
7. If water is flowing from multiple areas, collect samples from all
areas.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
118
c. CFC sampling:
1. Before entering the field, prepare a degassed ground glass syringe
in the same manner as indicated in Section II, Part b, Number 1 of
this appendix.
2. In the same area where the tritium sample was collected, push the
push point sampler into the ground. Push the sampler as deep as
possible.
3. Pull out the internal steel rod from the sampler in order to induce
flow.
4. Place the small piece of 1/4" Tygon tubing onto the end of the
sampler.
5. Attach a 3-way valve to the end of the Tygon tubing (Figure A-6 ).
Use just enough Tygon tubing to attach the 3-way valve. The
sample should have as little contact with the Tygon tubing as
possible since Tygon tubing could be a source of sample
contamination.
6 . If water is not flowing from the sampler, manually induce flow
with a plastic syringe.
7. Attach the 3-way valve at the end of the degassed syringe to the 3-
way valve at the end of the sampler.
8 . Turn the 3-way valve at the top of the sampler to the position
where the syringe is connected directly to the outlet.
9. Open the 3-way valve on the syringe and flush a large volume of
water from the syringe through to the outlet.
10. Turn the 3-way valve at the end of the sampler to the position that
connects flow from the sampler directly to the syringe.
11. Slowly draw water from the sampler into the syringe. Fill the
syringe with approximately 15 ml of water.
12. Repeat steps 8 through 11 approximately 3 times or until no
bubbles are present in the syringe.
13. Slowly draw 20 - 30 ml of water.
14. Turn the 3-way valve on the syringe to the position that closes the
syringe from the sampler and outlet.
15. Remove the syringe from the sampler, dry and label.
16. Place the syringe into the bucket of water.
17. Once sampling is completed for the day, the CFC samples must be
transferred from the syringe into copper tubing for storage (Figure
A-7).
18. Attach a length of 1/8" OD copper tubing to the tank of UHP N2 .
19. With a small length of 1/8" ID Tygon tubing, attach the copper
tubing to a 3-way valve.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
119
Syringe ^
3-way valves
1/4" OD Tygon tubing
Push point
sampler
Screening ^ }
interval ^ i
Figure A-6 . Schematic diagram of CFC sampling system for Kelvin Spring
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
120
Syringe
1/4" OD
Copper tubing
3-way
valves'
1/8" OD
Copper Tubing
1/4" to 3/8" OD
UltraTorr
3/8" OD
Copper tubing
Figure A-7. Schematic diagram of system used to transfer CFC samples from
ground glass syringes to copper tubing. The connection bewteen the copper
tubing and the 3-way valves were made using a small length of Tygon tubing.
When filling thecopper tubing, the sample assembly and syringe are in the
vertical position.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
121
20. With a small length of 1/4" OD Tygon tubing, attach another end
of the 3-way valve to small length (3 - 4") of 1/4" copper tubing.
21. Connect the 1/4" copper tubing to the UltraTorr.
22. Prepare the aluminum plate with clamps and 3/8" copper tubing in
the same manner as indicated in Section I, Part d, Numbers 5-8 of
this appendix.
23. Attach the sample line to the UltraTorr fitting.
24. Attach a length of 3/8" Tygon tubing to the outlet of the sample
assembly.
25. Attach a syringe to the final outlet on the 3-way valve.
26. Turn the 3-way valve to the position that connects the gas tank
directly to the sample assembly.
27. Turn on the gas tank and flush the sample line for approximately 2
minutes.
28. Turn the 3-way to the position that connects the syringe to the
sample assembly.
29. Hold the sample assembly vertical and depress the plunger on the
syringe slowly.
30. Once water is seen in the length of Tygon tubing at the outlet of
the sample assembly, tighten the clamps with the socket wrench
starting with the clamp closest to the outlet.
31. Once both clamps are tightened, remove the sample assembly
from the UltraTorr, remove the clamps and copper tubing from the
plate and label.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
122
APPENDIX B
CFC ANAYLSIS1
I. Items needed:
Adjustable clamp
Socket wrench (1/2" head)
Purge and trap system designed for extraction of water samples from
copper tubing (Figure B-l)
Dewar of dry ice and methanol
Dewar of boiling water
Trap packed with unibeads of 80/100 mesh glass beads (12" x 1/8" OD
stainless steel tube)
UHP N2 gas
CFC-11 and CFC-12 standard gas
Gas chromatograph (Shimadzu 14A)
Precolumn (6 " x 1/8" OD stainless steel column packed with Porapak C)
Capillary Column (6 ' x 1/8" OD x 0.085" ID stainless steel capillary
column packed 1% AT-1000 on Carbograph 1, 60/80 Mesh).
Electron capture detector (ECD)
Integrator (Shimadzu C-R8 -A Chromatopac)
II. Purge and Trap: Basic Purge and Trap Instructions (to be completed for
every system blank, stripper blank, standard and sample):
1. The purge and trap system is run by the basic program in the
integrator. The user will be prompted to perform each task indicated
below. The system will pause until the indicated task has been
performed. After each task has been performed (i.e., the trap has been
immersed in the cold bath), the user will press the enter key on the
keyboard which indicates that system is ready and the purge and trap
operation will continue.
2. Immerse the trap in the cold bath (dewar of dry ice and methanol).
3. Indicate the mode of sample injection (i.e., gas, syringe, copper tube).
System blanks and standards are gas samples.
4. Indicate the loop through which the sample will be transferred (1-5).
System blanks, stripper blanks and samples are run through loop 3.
5. Indicate the type of sample to be run (i.e., system blank, standard,
copper tube).
1 CFC analyses were performed in Jordan Clark’s laboratory at the University o f California, Santa
Barbara, by a procedure similar to that developed by Bullister and Weiss (1988).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Multi Port
iStd Loops
Cu Tube
Std Gas
Vent
Vent
VI V2
V3
Drain
Syringe
a.
Vent
Vent
Vent
V4
V5
Valve Positions
VI A - Standard gas in
B - Stripper in
V2 A - Stripper to dessicant
Cu tube flush
B - Stripper to vent
V3 A - Fill water sample loop
B - Inject water to stripper
V4 A - Stripper to vent
Trap to GC (backflush)
B - Trap isolated
V5 A - Stripper to vent
B - Stripper to trap
Precolumn backflush
Vent
Figure B -l. Schemcatic diagram of the purge and trap system in Jordan's Clark's laboratory at the University of California,
Santa Barbara.
N >
U >
124
6 . The sample fills the loop indicated (approximately 45 seconds).
Carrier gas (UHP N2 ) transfers the sample from the loop, through a
desiccant column (packed with Mg(C1 0 4 )2 ) and into the trap packed
with unibeads (80/100 mesh). Flow rate of carrier gas is 50-60
ml/min.
7. The sample is trapped for 4-5 minutes.
8 . Place the trap in the hot bath (dewar of boiling water). Always make
sure that the trap is submerged farther into the hot bath than it was in
the cold bath to ensure that the trap is not frozen shut.
9. The trap will sit in the hot bath for 4.1 minutes. The sample is then
flushed from the trap and transferred into the pre-column (packed
with Porapak C), into the column, then into the detector. The pre
column and column are located inside the gas chromatograph oven
which is heated to approximately 70° C. The detector is heated to
approximately 300° C.
10. Once the trap is offline (i.e.; no longer connected to the precolumn),
remove the hot bath from the trap.
11. The pre-column is back flushed for approximately 1.5 minutes after
sample transfer.
III. System Blanks and Standards:
1. A system blank runs a gas sample of carrier gas.
2. When running system blanks and gas standards ensure that the purge
and trap is switched to “Inject sample”. Always run a system blank
before running standards and samples.
3. Before running standard samples record the atmospheric pressure and
temperature.
4. Gas standards are run through a series of loops (1-5) with different
volumes. Six standards must be run before running water samples.
The first and last standard should be run through loop 3. The other 4
standards can be run through loops 1,2,4, and 5 in any order.
5. During the middle of a series of samples, run one standard through
loop 3 and record the atmospheric pressure and temperature.
6 . At the end of a series of samples, run another suite of gas standards by
repeating steps 1 -6 .
7. A calibration curve is constructed using the standard volumes and
peak areas (see Appendix D).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
125
IV. Sample Extraction from Copper Tubes:
1. Before samples can be run, cut off any extra length of tubing. With
an extra UltraTorr, ensure that all ends of the copper tubing will fit
into the UltraTorr fitting. If a sample does not fit, it cannot be run.
2. Before opening the first sample, run a stripper blank.
3. Weigh and record the mass of a copper tube sample including clamps.
4. Switch the purge and trap from “Inject sample” to “CV to stripper”.
5. Attach the copper tube to the purge and trap and tighten the
UltraTorrs.
6 . After indicating the type of sample to be run, enter the name of the
sample. This name will be recorded on the chromatograph; however,
it will not be recorded in the electronic file saved on the integrator.
7. Open the copper tube, beginning with the bottom clamp. Once the
clamps are open, with an adjustable clamp, open the crimped ends of
the tubing. The copper tube should be open after 2-3 pinches with the
adjustable clamp. Do not overly deform the copper tube because this
might result in breakage and loss of the sample.
8 . Once the sample has been released from the copper tube it is
transferred to the stripper.
9. The sample is stripped for approximately 4 minutes by bubbling
carrier gas (UHP N2 ) through the sample.
10. After stripping, the sample is passed through loop 3 and the desiccant
column (packed with MgCClO^) before entering the trap.
11. Once the sample has been transferred to the trap, release the water in
the stripper through the vent.
12. Remove the copper tube and clamps from the purge and trap and
weigh it. The difference in mass will be indicative of the volume of
sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
126
APPENDIX C
SAMPLE CALCULATION OF CFC PARTIAL
PRESSURE
Calibration Curve
a. Equations
i. PV = nRT (Equation C-l)
where P is pressure (atm), V is volume (ml), n is the
number of moles of CFC-12, R is the gas constant
(82.06 atm-ml/mole-K), and T is temperature (K)
ii. — = -5 - (Equation C-2)
V RT
where P, is partial pressure of CFC-12 (atm)
iii. P, =(x,XPror) (Equation C-3)
where x, is mole fraction of CFC-12 and P j o t is the
total atmospheric pressure (atm)
iv. n = -X | ^ (Equation C-4)
RT
v. Concentration of the standard ( C s,d)= 2053 pptv = 2.053 x 10' 9
moles CFC-12/mole N2 ; therefore X j = 2.053 x 10"9
b. Calculate the number of moles of CFC-12 per sample loop.
Table C -l - Sample loop volumes
Loop
Number
Volume (ml)
1 0.025
2 0 . 1
3 0.5
4 1
5 2
c. Plot moles of CFC-12 in each standard sample vs. peak area (Figure
C-l).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Moles
127
Standard Calibration Curve
10- 23-03
2 1 0 '1 3
5097e-15 + 9.6726e-19x R = 0.99869
0
2 105 0
Peak Area (uV )
Figure C-l. Example of a standard calibration curve for CFC-12.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
128
d. Curve fit the data with a straight line and calculate the slope (m) and
intercept (b)
i. y = mx+ b (Equation C-5)
where y is the number of moles of CFC-12 and x is the peak
area of the sample
II. Calculation of Sample Concentration (in pptv)
a. Using the slope (m) and intercept (b) of the calibration curve (Figure
C-l), calculate the number of moles of CFC-12 per sample (Equation
C-5).
b. Divide the number of moles for each sample by the sample volume.
c. Calculate the solubility of CFC-12 using the solubility defined by
Warner and Weiss (1985).
In K' = ax+a
100^1
- +a,
T )
(Equation C-6 )
f f rv 'n'N
In
V v io o ,
+s bx +b 7
vlOO,
+ b -i
100
where K' is the solubility of CFC-12 (mole/L-atm), T is
temperature (K), S is salinity (ppt) and o ’s and V s are
constants (ai = -122.3246, a2 = 182.5306, a . 3 = 50.5898,
bi = -0.145633, b2 - 0.092509, b3 = -0.0156627)
d. Calculate the partial pressure (Pi) of CFC-12 in atm for each sample.
i. Pi = (Cw){H) (Equation C-7)
where Cw is the concentration of CFC-12 in water (mole/ml) and
H is Henry’s constant
ii. H=l/K'
f n Y 1 Yl000/w/A
iii. P =
UJU'A 1 l
(Equation C-8 )
(Equation C-9)
e. Calculate the partial pressure (P j) of CFC-12 in pptv
i. Pt (pptv) = (7* )(lxl 01 2 ) (Equation C-10)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
129
III. Compare the concentration of CFC-12 in sample to known CFC-12
atmospheric curves through time to obtain age of sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
130
APPENDIX D
PROCEDURE FOR TRITIUM ENRICHMENT
I. Items needed:
1 0 0 0 ml distillation flask
Allihn condenser
Potassium permanganate (KMn0 4 )
Soxhlet heater mantle and transformer
Fume hood
Glass enrichment cell (Figure D-l)
Electrode pair consisting of one nickel and one iron electrode (Figure D-l)
Constant current power source adjustable from 0 to 3.0 amps at 0.1 amp
Increments (Extech Instruments Model 382213 DC regulated)
Refrigerated bath
Sodium hydroxide (NaOH) in pellet form
Vacuum grease
2 0 ml distillation trap
Liquid nitrogen
Vacuum pump
Nickel/Chromium heating wire
Plastic liquid scintillation counting vial with aluminum foil lined cap (20 ml)
Ultima Gold LLT scintillation fluor
II. Sample preparation
1. Transfer sample (500 ml) to a 1000 ml distillation flask attached to an
Allihn condenser (Figure D-2).
2. Add approximately 0.25 g of KMn0 4 .
3. Set transformer to approximately 55 on dial and start heating the soxhlet
heater. Sample will boil.
4. Distill until dry (approximately 6-7 hours).
5. Store sample in airtight glassware. Typically, a small fraction (<5%) of
the sample is lost in this step.
6 . Clean flask with Alconox. Let the flask soak over night in a bath of
Alconox and deionized water. After soaking, scrub the flask with wire
brush and rinse with distilled water.
III. Enrichment
1. In electrolysis cell, add 400 mg of NaOH (approximately 2 pellets) to 10
ml of sample. Shake until the pellets are completely dissolved.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34/28—►
0.5
< N
■ 'T
i k
i k
131
0.05
0 .6 1
s
3
H— ►
ooo
1.5
Figure D-l. Diagram of tritium enrichment cell and electrode assembly.
Measurements are in centimeters (cm). The electrode plates are
2.8 cm wide x 0.4 cm thick. The electrode leads are 0.2 cm OD. The leads are
spot welded to the plates and soldered to the wire leads
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
132
Glass
joint
24/40
Variac
transformer
Allihn
Condenser
24/40
Soxhlet heater
mantle
1 0 0 0 ml
Glass distillation
flask
24/40
Sample Bottle
Figure D-2. Photograph of first distillation for tritium analysis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
133
2. Insert the electrode assembly and mark the 10 ml water level with
electrical tape. The tips of the electrodes should almost touch the bottom
of the cell. Make sure the electrodes are not touching and the leads are
outside the closure.
3. Add 15 ml of sample to cell and mark the 25 ml water level with
electrical tape.
4. Add 25 ml of sample to make the volume in the cell 50 ml.
5. Apply a small amount of vacuum grease to all fittings.
6 . Place cell in a cold bath (maintained between 5 and 10°C). The entire
electrolysis process should be carried out in a working fume hood.
7. Connect cell to power supply. The red lead is the nickel anode and the
black lead is the iron cathode. Connect the red lead to the positive
terminal of the power supply. Connect to the black lead to the negative
terminal of the power supply. If creating a series of electrolysis cells,
connect the black lead to the red lead of the following cell in the series.
Connect the black lead of the last cell and the red lead of the first cell to
the power supply.
8 . Turn on power supply capable of 30 volts and adjust the current to 3.0 A.
By connecting several cells in a series, up to 10 samples can be done
simultaneously with a 30 volt power supply.
9. Reduce sample volume to 25 ml (approximately 24 hours; electrolysis
proceeds at a rate of ~ 1 ml/hour).
10. Turn off electrodes and add another 25 ml of sample to the cell.
11. Repeat steps 8-10 until 250 ml of the sample has been reduced to 25 ml.
12. Reduce the current to 0.3 A and run until the sample volume is reduced to
about 10 ml (approximately 5-6 days).
IV. Vacuum distillation
1. Prepare and weigh the distillation trap.
2. Remove the cell from the cold bath and remove electrodes from the cell.
3. Assemble vacuum distillation apparatus (Figure D-3).
4. Apply vacuum grease to all joints.
5. Wrap heating wire to the bottom of the cell and connect it to a
transformer.
6 . Cool the trap with a dewar of liquid nitrogen. Adjust the height of the
dewar to ensure that the bottom of the trap is always immersed in liquid
nitrogen. Add more liquid nitrogen as needed.
7. Set transformer to 5.
8 . Apply vacuum.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29/42
See Figure D-l
for specs
Dewar of
Liquid N2
Figure D-3. Diagram of vaccum distillation.
Measurements are in centimeters (cm).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
135
9. When the bottom of the cell is completely dry, move the heating wire to
the top of the cell to ensure that the entire cell dries. Distillation will take
3-5 hours.
10. Disconnect vacuum pump. Thaw the trap, remove and weigh. The
difference in mass will indicate the final volume of sample from the
electrolysis process.
11. Add 9 ml of sample to 20 ml plastic liquid scintillation counting vial.
12. Add 11 ml of Ultima Gold LLT fluor to counting vial.
V. Tritium Beta counting:
1. All enriched samples were counted by Roy Dunker at Idaho State
University using a low background beta counter (Packard Tri-Carb model
3170TR/SL liquid scintillation analyzer).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A geochemical and hydrological assessment of groundwater in the Portuguese Bend landslide, California

PDF

Integrated geochemical and hydrodynamic modeling of San Diego Bay, California

PDF

A proxy for reconstructing histories of carbon oxidation in the Northeast Pacific using the carbon isotopic composition of benthic foraminifera

PDF

Active deformation at Canyonlands National Park: Distribution of displacements across the grabens using spaceborne geodesy

PDF

Dynamic fluvial systems and gravel progradation in the Himalayan foreland

PDF

Early Jurassic reef eclipse: Paleoecology and sclerochronology of the "Lithiotis" facies bivalves

PDF

Barnacles as mudstickers? The paleobiology, paleoecology, and stratigraphic significance of Tamiosoma gregaria in the Pancho Rico Formation, Salinas Valley, California

PDF

Biotic recovery from the end-Permian mass extinction: Analysis of biofabric trends in the Lower Triassic Virgin Limestone, southern Nevada

PDF

Characterization of geochemical and lithologic variations in Milankovitch cycles: Green River Formation, Wyoming

PDF

Helicoplacoid echinoderms: Paleoecology of Cambrian soft substrate immobile suspension feeders

PDF

Attachment, accommodation, and the outcomes of romantic relationships

PDF

Investigating the influence of atmospheric changes on the variability of the North Pacific using paleoproxy data and a fully coupled GCM

PDF

Helmets: The safe choice. A study on the benefits of helmet safety while skiing and/or snowboarding: A public relations program to encourage helmet use

PDF

Determination of paleoearthquake age and slip per event data, and Late Pleistocene-Holocene slip rates on a blind-thrust fault: Application of a new methodology to the Puente Hills thrust fault,...

PDF

Elliptical Ground

PDF

Fractionation of nitrogen isotopes during early diagenesis

PDF

Holocene sedimentation in the southern Gulf of California and its climatic implications

PDF

Grain-size and Fourier grain-shape sorting of ooids from the Lee Stocking Island area, Exuma Cays, Bahamas

PDF

A case study comparing measured and simulated acoustical environments: Acoustical improvements of Verle Annis Gallery

PDF

Geology of the southern Sierra San Francisquito, Baja California, Mexico

Asset Metadata

Creator DiFilippo, Erica Lynne (author)

Core Title Groundwater modeling and geochemical tracer (CFC-12 and tritium) distribution in the Abalone Cove landslide, Palos Verdes, California

School Graduate School

Degree Master of Science

Degree Program Earth Sciences

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag geochemistry,Geology,Hydrology,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Hammond, Douglas (committee chair), Douglas, Robert (committee member), Weh, Johnson (committee member), Williams, Dennis (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-45592

Unique identifier UC11328609

Identifier 1430388.pdf (filename),usctheses-c16-45592 (legacy record id)

Legacy Identifier 1430388.pdf

Dmrecord 45592

Document Type Thesis

Rights DiFilippo, Erica Lynne

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA