University of Southern California Dissertations and Theses

Evaluating surface casing depths of oil & gas operations in an effort to protect local groundwater: a GIS enabled process

(USC Thesis Other)

Evaluating surface casing depths of oil & gas operations in an effort to protect local groundwater: a GIS enabled process

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Evaluating Surface Casing Depths of Oil & Gas Operations in an Effort to Protect
Local Groundwater: A GIS Enabled Process

by

Mary Elaine Nienkamp

A Thesis Presented to the
Faculty of the USC Graduate School
University of Southern California
In Partial Fulfillment of the
Requirements for the Degree
Master of Science
(Geographic Information Science and Technology)

May 2016

Copyright ® 2015 by Mary Elaine Nienkamp

This thesis and the many hours of hard work it represents is dedicated to my son, Sloan R.
Nienkamp-Glasscock. Your unconditional love and smile kept me motivated the many times I
questioned myself.

iv
Table of Contents
List of Figures ................................................................................................................................ vi
List of Tables ............................................................................................................................... viii
Acknowledgements ........................................................................................................................ ix
List of Abbreviations ...................................................................................................................... x
Abstract .......................................................................................................................................... xi
Chapter 1 ......................................................................................................................................... 1
1.1 Motivation ............................................................................................................................1
1.2 Thesis Organization .............................................................................................................3
Chapter 2 Groundwater Regulation, Aquifers, and Oil and Gas Operations in Texas .................. 4
2.1 Groundwater Regulation in Texas .......................................................................................4
2.2 Carrizo-Wilcox Aquifer .......................................................................................................8
2.2.1. Water Quality of the Carrizo-Wilcox Aquifer .........................................................11
2.2.2. San Antonio Water System Activity ........................................................................12
2.3 Oil and Gas Operations in Texas .......................................................................................13
2.3.1. Baseline Water Sampling .........................................................................................13
2.3.2. Well Casing ..............................................................................................................13
2.3.3. Railroad Commission of Texas Regulation .............................................................14
Chapter 3 Geologic Mapping Practices ........................................................................................ 18
3.1 Interpreting Well Logs .......................................................................................................18
3.2 Structure Maps ...................................................................................................................20
3.3 Interpolation Techniques in GIS ........................................................................................21
3.3.1. Spline Interpolation ..................................................................................................22
3.3.2. Kernel Interpolation .................................................................................................23
3.3.3. Empirical Bayesian Kriging .....................................................................................23
3.3.4. Error .........................................................................................................................26
Chapter 4 Data and Methods......................................................................................................... 28
4.1 Study Area Boundary .........................................................................................................28
4.2 Data Acquisition and Manipulation ...................................................................................29
4.2.1. Well Locations .........................................................................................................30
v
4.2.2. Well Logs .................................................................................................................32
4.2.3. Surface Casing Depths .............................................................................................34
4.2.4. Salinity Determination Lines ...................................................................................37
4.3 Methodology ......................................................................................................................39
4.3.1. Interpolating a Surface .............................................................................................39
4.3.2. Depth Comparison ...................................................................................................48
4.3.3. 3D Representation ....................................................................................................49
Chapter 5 Results .......................................................................................................................... 52
5.1 Is the Carrizo-Wilcox Aquifer Protected? .........................................................................52
5.2 Error ...................................................................................................................................55
5.3 Aquifer protection over time ..............................................................................................59
5.4 Difference between depth of surface casing and predicted aquifer depth .........................61
Chapter 6 Conclusions .................................................................................................................. 64
6.1 Limitations and Observations ............................................................................................65
6.2 Current and Future Work ...................................................................................................65
References ..................................................................................................................................... 67

vi
List of Figures
Figure 1 Eagleford Wells along South Texas ................................................................................. 2
Figure 2 Groundwater Conservation Districts (top) & Groundwater Management
Areas (bottom). ........................................................................................................................ 7
Figure 3 Distribution of Carrizo-Wilcox aquifer across Texas. ..................................................... 9
Figure 4 Geologic stratigraphy of the studdy area. Modified from Dutton, et al 2003. ............... 10
Figure 5 Groundwater pumpage estimates for 2013, in acre-feet. Source: Texas Water
Development Board ............................................................................................................... 11
Figure 6 Example of surface casing. Source: Bonanza Creek 2015. ............................................ 14
Figure 7 Sample of surface casing letter from the groundwater advisory unit. ............................ 16
Figure 8 New surface casing letter after an exception was filed. ................................................. 17
Figure 9: Sample well log. Modified from Camp and Outlaw 1993. ........................................... 19
Figure 10 Simulated Semivariograms created from project data. ................................................. 25
Figure 11 Eagle Ford drilling region in South Texas with study area boundary. ......................... 29
Figure 12 Distribution of well locations from P2 Energy Solutions. ........................................... 31
Figure 13 Screen capture of Petra's cross section module displaying correlated well logs. ......... 32
Figure 14 Carrizo Base Depth Picks from Well Logs .................................................................. 34
Figure 15 Depth of surface casing of Eagle Ford wells within study area. .................................. 37
Figure 16 Carrizo-Wilcox thickness map modified from the Army Corps of Engineers, with
salinity lines highlighted in yellow. Source: Texas Water Development Board ................... 38
Figure 17 Predicted Base of Carrizo-Wilcox Aquifer using Spline Interpolation Technique ...... 40
Figure 18 Predicted Base of Carrizo-Wilcox Aquifer using Kernel Interpolation Technique
with the Constant Function ................................................................................................... 41
Figure 19 Predicted Standard Error Map produced with Kernel Interpolation. ........................... 42
Figure 20 Predicted Base of Carrizo-Wilcox Aquifer using Empirical Bayesian Interpolation
Technique with the Exponential Detrended Semivariogram ................................................ 44
vii
Figure 21 Predicted Standard Error Map produced with Empirical Bayesian Kriging
Interpolation .......................................................................................................................... 45
Figure 22 Side-by-side comparrison of evaluated interpolated predictive surfaces and their
accompanying standard error map. ....................................................................................... 47
Figure 23 Predicted base of aquifer surface showing wells with surface casing that penetrate
the surface (white) and well with surface casing that do not (black). ................................... 49
Figure 24 Screen capture of 3D environment displaying a predicted surface, representing the
base of the aquifer (orange), standard error halos (blue) and surface casings (red). ............. 51
Figure 25 Pecentages of wells with surface casing that penetrate the predicted surface
representing the aquifer base. ................................................................................................ 52
Figure 26 (Reproduced Figure 23) Predicted base of aquifer surface showing wells with surface
casing that penetrate the surface (white) and well with surface casing that do not (black). . 53
Figure 27 Percentages of Eagle Ford Wells whose surface casings penetrate the aquifer
by TDS. ................................................................................................................................. 54
Figure 28 Standard Error Surface: Actual value falls within the range of the predicted value,
plus or minus the standard error with 95% confidence. ........................................................ 56
Figure 29 Histogram of Standard Error Values for the EBK surface ........................................... 57
Figure 30 Close-up of surface casings, the predicted aquifer surface, and the margin of error. .. 58
Figure 31 Percentage of wells for which the surface casings penetrate the shallower error
value (left), and deeper error value (right). ........................................................................... 59
Figure 32 Percentage of wells whose surface casings penetrate the predicted aquifer surface,
by year. .................................................................................................................................. 60
Figure 33 Eagle Ford Wells symbolized by spud year. ................................................................ 61
Figure 34 Difference between surface casing depth and predictd aquifer base depth. ................. 62
Figure 35 Difference between surface casing depth and base of aquifer. .................................... 63

viii
List of Tables
Table 1 Regulatory agencies involved with Texas groundwater. ................................................... 8
Table 2 Comparrisson of Interpolation Method Functionality ..................................................... 27
Table 3 Attribute data for Eagle Ford wells. ................................................................................. 36
Table 4 Diagnostic Comparrison between Kernerl and EBK interpolation methods. .................. 46

ix

Acknowledgements

I would like to thank, first and foremost, my advisor Karen Kemp for providing me valuable
direction and insight. I consider it an honor to have the opportunity of working with her.
I am grateful to Nestor Phillips, Geologic Advisor, who allowed me to use his log interpretations,
and provided much guidance along the way. I would also like to thank my employer, EOG
Resources, for allowing me to use their databases from IHS, TGS and P2. This thesis would not
have been possible without the support of my family, throughout this process.

x
List of Abbreviations
API American Petroleum Institute
ASR Aquifer Storage and Recovery
DFC Desired Future Condition
EBK Empirical Bayesian Kriging
EPA Environmental Protection Agency
FBSL Feet Below Sea Level
GAM Groundwater Availability Model
GAU Groundwater Advisory Unit
GCD Groundwater Conservation District
GIS Geographic Information System
GMA Groundwater Management Area
GPS Global Positioning System
LPI Local Polynomial Interpolation
RRC Railroad Commission of Texas
SAWS San Antonio Water Systems
SP Spontaneous Potential
SSTVD Subsea Total Vertical Depth
TCEQ Texas Commission on Environmental Quality
TDS Total Dissolved Solids
TWDB Texas Water Development Board
xi
Abstract

As groundwater is a vital resource, it is important that oil and gas operations do not jeopardize
water quality. Many consumers, including farmers and municipalities, rely year after year on the
freshwater provided by aquifers. Along South Texas, oil and gas companies are targeting the
Eagle Ford formation containing hydrocarbons. In this same region, the Carrizo-Wilcox aquifer
must be drilled through to reach the Eagle Ford below. To protect the above aquifer, cemented
surface casing is used to seal the Carrizo-Wilcox from contaminants within the well borehole.
This study incorporated Geographic Information Systems (GIS) to evaluate surface casing depths
of oil and gas wells, to verify if they are deep enough to adequately protect the aquifer. To
understand the geologic structure occurring in this region, aquifer depths obtained from well logs
were used to interpolate the base of the Carrizo Sands. After comparing three interpolation
methods, the Empirical Bayesian Kriging (EBK) interpolator, using the Exponential Detrended
semivariogram, was selected to create a predicted surface and a standard error map. Surface
casing depths of Eagle Ford wells were mapped and queried to determine if they are deeper or
shallower than the predicted surface representing the aquifer. Over half of the wells within the
study area had surface casing shallower than the aquifer. However, most of those fell within
areas where groundwater was brackish. Results from this study should motivate regulatory
agencies in tightening up policies and guidelines pertaining to oil and gas operations affecting
aquifers within the State of Texas. In addition, methodologies conducted during the study
provide a viable means to improve the current process of determining surface casing depths.
1
Chapter 1
Groundwater describes water that fills the fractures and pores between rocks and soil below the
earth’s surface. Groundwater is a vital, natural commodity which should be monitored and
protected. Because many consumers rely on groundwater to provide fresh water, it is important
that great care is taken to protect this valuable resource. Many state and federal agencies are
responsible for and involved with the various aquifer systems across the nation, including Texas.
However, in Texas, only one regulatory agency, the Railroad Commission of Texas (RRC), has
jurisdiction when it comes to oil and gas operations affecting groundwater. This thesis evaluates
the protection of groundwater in an active oil and gas producing area, by investigating surface
casing depths regulated by the RRC.
1.1 Motivation
As of 2009, the Eagle Ford Shale Play in South Texas became an active area for oil and
gas production. Figure 1 shows this region where Eagle Ford wells are being drilled. This is also
the area of focus for this study. Oil and gas companies are drilling through the Carrizo-Wilcox
aquifer to extract petroleum from the Eagle Ford formation below. Surface casing, composed of
steel and cement, is used to line the borehole which seals the local aquifer from any
contaminants.
2

Figure 1 Eagleford Wells along South Texas.
In Texas, before an oil or gas well can be drilled, a Groundwater Protection
Determination letter must be received from the Railroad Commission of Texas (RRC). This letter
states the depth to which the cemented surface casing must reach. If this surface casing is not set
deep enough, the base of the aquifer is not fully penetrated, and therefore may be susceptible to
contamination. Currently there is no evidence that the RRC is conducting any actual subsurface
mapping of the Carrizo-Wilcox.
The primary objective of this study is to demonstrate an applicable method for mapping a
subsurface geologic formation in an effort to identify appropriate surface casing depth
recommendations. Information from well logs are used to create a continuous surface
3

representing a geologic horizon, in this case, the base of an aquifer. After this geologic
representation has been created, surface casing depths of already-drilled, Eagle Ford oil and gas
wells can then be compared to the aquifer depth. Thus, the two key goals of this study are:
1) Determine an appropriate interpolation method, available in ArcGIS, for mapping
subsurface geology.
2) Determine if surface casings of drilled wells adequately protect the local aquifer, as
defined by the interpolated surface.
The research described in this document found that over half of Eagle Ford wells within
the study area do not have surface casings to the base of the Carrizo-Wilcox aquifer. As a result
of these findings, this study provides a method of improving the current process of determining
the required surface casing depth by using Geographic Information Systems (GIS) and enhanced
subsurface maps.
1.2 Thesis Organization
The remainder of this thesis is structured into five additional chapters. Chapter Two
provides contextual information regarding groundwater regulation in Texas and the Carrizo-
Wilcox aquifer. Chapter Two also explains oil and gas operations and how groundwater is
vulnerable to these operations. Chapter Three outlines a procedure for creating a continuous
representation of a subsurface geologic formation. Chapter Four describes the methodology
conducted during this thesis. Chapter Five summarizes the outcomes produced and Chapter Six
offers concluding comments and recommendations.

4

Chapter 2 Groundwater Regulation, Aquifers, and
Oil and Gas Operations in Texas
Before introducing geologic mapping, it is important to examine some background information
concerning groundwater, and oil and gas activity in Texas. This chapter provides: (1) a
framework of Texas groundwater and the regulatory agencies involved; (2) context on the
Carrizo-Wilcox aquifer and groundwater quality; and (3) information regarding oil and gas
operations, especially surface casing.
2.1 Groundwater Regulation in Texas
Sixty percent of water used within the state of Texas is supplied by groundwater (Texas
Water Development Board 2015). Over eight billion gallons a day of groundwater is pumped
from aquifers throughout the state (National Groundwater Association 2010). Because of the
heavy use of groundwater for human consumption and activity, this vital resource requires
protection.
In Texas, groundwater is owned by landowners and is, therefore, considered private
property. However, groundwater is managed and regulated through various state organizations.
The Texas Commission on Environmental Quality (TCEQ) is responsible for protecting water
quality while the Texas Water Development Board (TWDB) is responsible for managing and
financing adequate water supplies. The TCEQ is the primary environmental organization for the
State of Texas. The organization has more jurisdiction with water quality and quantity than any
other state agency (Sansom 2008). The TCEQ’s roles and responsibilities are directed by federal
laws and Environmental Protection Agency (EPA) rules. The Texas governor appoints three
commissioners who are accountable for the overall direction of the agency.
5

The TWDB is composed of board members also appointed by the governor. Using
predictive groundwater availability modeling, the TWDB’s chief responsibility is to establish
and publish a state-wide water plan every five years (Sansom 2008). This groundwater
availability model (GAM) predicts future groundwater trends including water levels and
recharge, plus the characterization of geology and aquifer properties for all aquifers within the
state. Initially, the TWDB was developed to offer low-interest loans for water improvement
projects. Since then, the board has taken on additional responsibilities including facilitating water
rights transfers between sellers and buyers, and managing the Water Trust. Water can be
donated, leased or purchased for environmental purposes, through the Water Trust.
In 1949, groundwater conservation districts (GCD) were generated as political
boundaries to manage groundwater at the local level (George, Mace and Petrossian 2011).
Today, there are 100 GCDs that are responsible for managing water well spacing and production,
the permitting of new water wells, and major alterations of existing water wells (Texas Water
Development Board n.d., Porter 2014). When evaluating permits, it is the GCD’s responsibility
to ensure that the water is dedicated to beneficial use, and does not exceed amounts that would
adversely affect groundwater resources. In addition, GCDs work closely with the TWDB by
submitting local GAMs, every five years, for approval (Texas Water Development Board, n.d.).
Groundwater conservation districts cover approximately 66% of the state, with the
remaining areas not being protected by a groundwater conservation district (George, Mace and
Petrossian 2011). Areas not protected by a GCD have no regulation and therefore have no limits
on water pumping. This study overlies five groundwater conservations districts. Of note there is
one small area in the southeast portion of Gonzales County not within a GCD. The top of Figure
6

2 displays the groundwater conservation districts in relation to the study area. The study area
boundary is displayed here for reference and is discussed in further detail in Chapter Four.
Groundwater management areas (GMA) were established in 2005 as a means for regional
planning of groundwater (Porter 2014). These GMAs are displayed on the bottom of Figure 2.
While the GCD boundaries closely follow many of the county borders, GMA boundaries align
better with aquifer borders and therefore cover much larger areas. Groundwater management
areas are responsible for conserving, preserving, protecting, recharging, and preventing waste of
groundwater resources (Texas Water Development Board 2015). Also, GMAs regulate any
subsidence caused by the pumping of groundwater. Another important role of the GMAs is to
create long-term goals which support the desired future conditions (DFC) of the aquifer they are
responsible for. DFCs are defined as “the desired, quantified condition of groundwater resources
(such as water levels, spring flows, or volumes) within a management area at one or more
specified future times as defined by participating groundwater conservation districts within a
groundwater management area as part of the joint planning process” (Texas Water Development
Board 2015). For a DFC to pass, it must receive a two-thirds majority vote of all GCDs within
the GMA, and will assist in guiding those included GCD’s policies. DFCs are submitted every
five years, but may be modified at any time (Porter 2014). There are three groundwater
management areas within this study area.
7

Figure 2 Groundwater Conservation Districts (top) & Groundwater Management Areas (bottom).
8

What is interesting is that these state and federal agencies have no authority over
groundwater when it comes to oil and gas operations. This role belongs to the Railroad
Commission of Texas, which in 1919 was granted jurisdiction over oil and gas operations for the
State of Texas (Railroad Commission of Texas 2015). Table 1 summarizes the responsibilities of
the various agencies involved with Texas groundwater. Further detail regarding the Texas
Railroad Commission is addressed later in the chapter.
Table 1 Regulatory agencies involved with Texas groundwater.
Organization Acronym Area Responsibilities
Texas Commission on
Environmental Quality
TCEQ Statewide Responsible for protecting water quality.
Texas Water
Development Board
TWDB Statewide
Manages & finances adequate water
supplies for the state.
5-year statewide water plan.
Groundwater
Conservation Districts
GCD Local
Local management of groundwater levels.
Water wells permits & regulation.
Groundwater
Management Areas
GMA Regional
Regional groundwater planning.
Creates desired future conditions
Groundwater Advisory
Unit (Railroad
Commission of Texas)
GAU
(RRC)
Statewide
Regulates oil and gas operations,
including those involved with
groundwater.

2.2 Carrizo-Wilcox Aquifer
The Carrizo-Wilcox Aquifer is one of nine major aquifers in Texas, and spans from the
Louisiana border to the Mexico border (Figure 3). The aquifer runs parallel to the Gulf of
Mexico, covering 66 counties in Texas. The Carrizo-Wilcox reaches up to 3,000 feet thick in
some areas, with freshwater thickness averaging 670 feet (George, Mace and Petrossian 2011).
9

Figure 3 Distribution of Carrizo-Wilcox aquifer across Texas.
The aquifer is mostly composed of sand, interbedded with gravel, slit, clay and lignite.
The Carrizo-Wilcox consists of two hydrologically connected aquifers composed of the Carrizo
Sand, located at the base of the Claiborne group, overlying the Wilcox group (Huang, et al.
2012). Figure 4 shows the stratigraphic order of the formations within the study area.
10

Figure 4 Geologic stratigraphy of the study area. Modified from Dutton, et al 2003.
The Carrizo-Wilcox supplies water for multiple uses across Texas, including irrigation,
municipal water supply, manufacturing, steam power and watering livestock. In 2013, almost
415,000 acre-feet of water was pumped from the Carrizo-Wilcox, with irrigation and municipal
water supplies using 89% of the groundwater (Figure 5) (Texas Water Development Board
2013). The Winter Garden area – a farming region - of the Carrizo-Wilcox aquifer occupies
Zavala, Frio, Atascosa, Wilson and Dimmit counties (Figure 3) (Boghici 2009). With respect to
irrigation pumpage, this farming area uses 35% of extracted Carrizo-Wilcox groundwater. The
primary municipalities that depend on the fresh water from this aquifer include Bryan-College
Station, Lufkin-Nacogdoches, Bastrop, Tyler, Pflugerville, and Hutto, while San Antonio is
11

relying on the Carrizo-Wilcox for untraditional water supply, and is further explained in section
2.2.2 (Figure 3).

Figure 5 Groundwater pumpage estimates for 2013, in acre-feet. Source: Texas Water
Development Board.
2.2.1. Water Quality of the Carrizo-Wilcox Aquifer
Total dissolved solids (TDS) describe the amount of mineral content within groundwater
(George, Mace and Petrossian 2011). TDS is measured as milligrams per liter (mg/L) of water,
and if values are too high, the water could be undrinkable, unsuitable for irrigation or watering
livestock, or even toxic. According to the Texas Water Development Board (2011), groundwater
TDS values of less than 1,000 mg/L are considered fresh, and therefore suitable for human
consumption. Groundwater with these low TDS values typically lie near the surface, in the
outcrop region of the Carrizo-Wilcox Aquifer. This is because the aquifer recharges here from
surface runoff. As the formation dips downward toward the Gulf Coast, water collects minerals
and sodium from the rock composition, and groundwater quality degrades. Groundwater with
TDS values up to 1500 mg/L could be used to irrigate crops, and lie in the subsurface region.
176,125
43%
13,567
3%
9,116
2%
13,556
3%
191,982
46%
10,475
3%
2013 Carrizo-Wilcox Pumpage
Municipal
Manufacturing
Mining
Steam Electric Power
Irrigation
Livestocks
12

Groundwater with TDS values of 3,000 mg/L or less, located further down-dip, could be used to
water livestock.
2.2.2. San Antonio Water System Activity
The San Antonio Water System (SAWS) provides water to the City of San Antonio, and
much of Bexar County. Although the Edwards Aquifer provides SAWS with nearly all of their
water, SAWS is turning to alternatives as a way to diversify their water supply. These
alternatives rely on the Carrizo-Wilcox aquifer to provide support. In 2002, SAWS began
constructing an Aquifer Storage and Recovery (ASR) facility, completing Phase I in 2004 (Crow
2012). Because the Edwards aquifer is a karst formation, the aquifer fills and drains rather
quickly compared to a sand aquifer. Therefore when the Edwards aquifer is sufficiently full and
consumption is low, water is pumped out of the Edwards and injected into the Carrizo-Wilcox
ASR for storage. During times of drought and higher consumption, there is additional water in
the ASR available to the community.
Another means of diversifying SAWS’s water supply is through the construction of their
desalination plant. This plant will pump brackish water from the Wilcox portion of the Carrizo-
Wilcox aquifer and utilize reverse osmosis treatment to remove the dissolved solids, making the
water drinkable (San Antonio Water System 2015). Phase I is expected to complete in 2016,
providing twelve million gallons of freshwater per day. This project indicates the necessity of
water in the region and that brackish water should also be protected from contaminants. These
developments demonstrate alternate ways to utilize this plentiful aquifer other than traditional
pumping of fresh groundwater.
13

2.3 Oil and Gas Operations in Texas
Texas is known for its leading role in oil production within the United States. The Eagle
Ford shale play has become the most recent “boom” in Texas, providing more than one million
barrels of oil equivalent per day (boe/d) (Eagle Ford Shale 2015). In addition to oil and gas, the
Eagle Ford has provided jobs, money and growth to the region. While it is important to continue
facilitating this increased economic development, it is also important to insure that the
environment (including groundwater) sees no negative impact.
2.3.1. Baseline Water Sampling
It would be ideal to have water sampling conducted in the region prior to oil and gas
operations to determine the baseline water quality. Although most states do not require this, it
has been recommended by the American Petroleum Institute (Holloway and Rudd 2013).
Groundwater could then be sampled during drilling and fracking operations, and even long after
a well is completed, to ascertain if local groundwater quality has deteriorated. There are some
historical records of water wells in the region. However, a systematic method of water sampling
was not conducted prior to oil and gas operations.
2.3.2. Well Casing
When drilling oil and gas wells, shallow aquifers are penetrated to reach a formation
below filled with hydrocarbons. Therefore, the first step of drilling an oil or gas well is to set the
surface casing. Surface casing is intended to run from the top of the borehole to below the base
of the aquifer (Figure 6). Surface casing consists of a steel, hollow pipe set in cement between it
and the borehole wall (Holloway and Rudd 2013). The initial cementation of the casing is critical
as it seals the annular space before further operations occur. This hydraulic barrier isolates fresh
14

groundwater from the inside of the well. If the surface casing is not deep enough or is not
properly installed, contaminant leakage into the groundwater could occur.

Figure 6 Example of surface casing. Source: (Bonanza Creek 2015).
2.3.3. Railroad Commission of Texas Regulation
The Railroad Commission of Texas (RRC) is responsible for regulating oil and gas
operations for the state of Texas, including water quality aspects related to those operations. The
Groundwater Advisory Unit (GAU) is a division of the Texas RRC which provides guidance on
issues such as saltwater disposal wells, and groundwater contamination and protection, including
casing depths (Gearhart 2014).
Prior to January 2014, the Texas Administrative Code Rule 3.13, which addresses well
casing, cementing, drilling, well control and completions, had not been revised since 2003
(Gronewold 2014). Rules from the new revision increase regulation guidelines which protect
15

groundwater as well as safety procedures. Rule 3.13 outlines several requirements for surface
casing. For example, surface casing must be at least one and one-half inches less than the
wellbore diameter and must be pressure tested (Texas Administrative Code 2014).
After the upper portion of the borehole is drilled, the steel casing is set with cement down
past the protection depth. According to Part 1, Chapter 3, Rule 3.13 of the Texas Administrative
Code, the Texas RRC defines the protection depth as,
Depth to which usable-quality water must be protected, as determined by the
Groundwater Advisory Unit of the Oil and Gas Division, which may include
zones that contain brackish or saltwater if such zones are correlative and/or
hydrologically connected to zones that contain usable-quality water. (Texas
Administrative Code 2014).
The difficulty with this definition is the word “usable” because it is a not a quantitative unit of
measure. Therefore, there are some instances where groundwater with TDS values of up to 7,000
mg/L are protected with surface casing, and other instances where groundwater with less than
3,000 mg/L are not. In addition, being hydrologically connected could mean anything between
two impermeable formations, or aquitards.
To find this usable depth, operators are referred to the “Surface Casing Estimator”
website hosted by the Bureau of Economic Geology and located at
http://coastal.beg.utexas.edu/surfacecasing/. This site displays a map with satellite imagery and
several oil and gas wells. Here, the operator can view casing information and well logs, if
available, of nearby, pre-existing wells. The results of the well logs identify various formation
characteristics such as lithology and fluid content. This will assist in finding an estimated depth
for surface casing. Although this method is functional, it is certainly not ideal because the aquifer
base nor salinity information is not shown. Additionally, the site provides a limited number of
wells, where the closest well to the area of interest may be several miles away. A map displaying
Carrizo-Wilcox depth would be a more informative approach to this part of the process.
16

After estimating a depth by using a nearby well, the operator must then submit a
Groundwater Protection Determination Request form (GW-1) to the groundwater advisory unit
(GAU). After review, the GAU will issue a surface casing letter which outlines where fresh and
usable water is located (Figure 7).

Figure 7 Sample of surface casing letter from the groundwater advisory unit.
Once an operator receives this letter, this is the depth they are to protect with surface
casing. No shallower, and no deeper. If an operator feels that the casing depth should be deeper
because they have better information, they are instructed to submit an exception request to the
GAU, and present this evidence (Gearhart, Email message to Railroad Commission of Texas
2015). After further review, the GAU will then issue a new letter (Figure 8) and use the newly
gathered information for future Groundwater Protection Determination letters, according to
Gearhart (2015).
17

Figure 8 New surface casing letter after an exception was filed.
The drawback to this process is that some less-responsible operators might see this as a
nuisance, costing too much time. The operator might set the casing depth to that which was
recommended by the RRC, whether it protects the aquifer or not.
It is important to understand the framework of oil and gas operations, and groundwater
protection and regulation. Having described this in Chapter Two, the next chapter introduces
subsurface geologic mapping and the use of Geographic Information Systems (GIS) in that
context.

18

Chapter 3 Geologic Mapping Practices
Mapping geologic subsurface structures and features is not entirely easy because it cannot be
directly observed. Therefore, locations where the formation (or aquifer) has been identified are
used to interpolate a predicted surface where direct observations cannot be made. This chapter
highlights how well logs are used as observation points, summarizes traditional geologic
mapping techniques, and explains interpolation methods for subsurface mapping available in
GIS.
3.1 Interpreting Well Logs
Often geophysical tools are sent down the borehole of a well to record geophysical
properties of the subsurface (Evenick 2008). Sometimes, these tools are sent downhole after a
well has been drilled, and other times these tools are on the same assembly as the drill bit and
records data while drilling. The results of these recordings are called well logs and are typically
provided on paper, as well as in digital files. By recording geophysical and physical properties,
well logs provide insight on subsurface geologic formations and conditions.
The first page of a log contains general information such as well name and number, API
(American Petroleum Institute) number, well operator, logging company, and elevation. This
first page is called the well header. The following pages, or body, contain the recorded data
separated into tracks, or columns. Three primary logs, which are implemented most often,
include gamma ray logs, resistivity logs and spontaneous potential (SP) logs. However, there are
numerous other types of logs available which provide further detail of geophysical properties.
Gamma ray logs record the radioactivity produced naturally by a formation (Evenick
2008). These logs are useful in identifying clays and shales because these materials emit high
gamma ray values (Camp and Outlaw Jr 1993). Additionally, gamma ray logs are valuable
19

because they are fairly inexpensive, simple to interpret and provide great vertical resolution,
allowing for easy stratigraphic correlations. Resistivity logs record the resistance of electrical
flow through a formation. Resistivity logs aid in identifying fluid type and formation porosity.
Like resistivity logs, SP logs are also electrical logs, but record the electrical current caused by
the mixing of two fluids with contrasting salinities. Drilling mud will typically have different
salinity than the formation fluids producing certain measurements (Hyne 2001). SP logs are also
useful in identifying formation permeability (Evenick 2008). However, SP logs do not offer great
vertical resolution, and can therefore lose detailed information on formations which are very
thin. Figure 9 demonstrates a simplified example of these three logs curves.
When identifying fresh water bearing formations such as aquifers, the log should reflect
low gamma ray values, high resistivity values and high SP values. However, as an aquifer
becomes more brackish, resistivity and SP will decrease.

Figure 9: Sample well log. Modified from Camp and Outlaw 1993.
20

To effectively map the subsurface, it is important to logically pick horizontal changes in
formations, or horizons. This is not easy as there are many variables that contribute to the output
of a well log, and therefore require the technical experience of a geologist.
Well logs provide the “observed” information needed to begin correlating and
interpreting a regional formation. Correlating well logs describes finding matching patterns
between multiple wells. Recognizing these matching patterns is important in finding these same
formation boundaries, or picks, across an area of interest. Accurate well correlations are
necessary to map subsurface structures.
3.2 Structure Maps
By correlating well logs from multiple wells, the structure of a formation can be
interpreted and ultimately mapped. When mapping the subsurface, geologists are interested in
understanding geologic structure and stratigraphy. Often areas are complicated where the
structure can be folded or faulted (Tearpock and Bischke 2002). These complications can make
mapping difficult and require a good understanding of basic geologic principles. Geologists
create many different types of maps in an effort to understand what is occurring deep in the
earth, including structure maps, thickness maps, fault maps, facies maps, hydrologic maps and
cross sections.
The most common map generated is one that represents a stratigraphic horizon of interest
and is called a structure map. Structure maps represent the depth of a formation, including its
structural features, and are created from correlated well log picks (Evenick 2008). If all of the
overburden were removed, a subsurface elevation map, or topographic map, representing a single
horizon would be left. Structure maps, or depth maps, can be represented by either contour lines,
or a continuous surface, or both. Whichever method is selected to represent the horizon of
21

interest, the outcome should still be reviewed with geologic intuition to be sure that it makes
sense within the region.
Traditionally, depth maps are created by hand by generating contour lines that connect
points of equal depth (Evenick 2008). Because there are many rules applied to contouring that
must be followed to create mechanically correct maps, this process requires technical knowledge
of an experienced geologist (Tearpock and Bischke 2002). However, the advancement of
computer technology and mapping software applications have made this process simpler and
automated. In addition to contour maps, software applications can use interpolation methods to
create a continuous surface that represents a geologic horizon.
Interpolation describes a technique where measured values at known locations are used to
predict values at locations where measurements have not been directly observed. For subsurface
features, this creates a digital model of a continuous surface representing a phenomenon such as
the base of an aquifer.
3.3 Interpolation Techniques in GIS
Well logs provide the control points needed to interpret a viable structure map of the
subsurface. Geographic Information Systems (GIS) use these control points to interpolate or
predict values between them. In GIS, the result is usually a raster dataset.
There are two categories of interpolation methods: (1) deterministic; and (2) geostatistical
(Esri 2015). Deterministic interpolators use predefined, smoothing, mathematical functions of
distance from known points to assign values to a predicted location. Deterministic interpolators
have the option of considering all of the points in the dataset globally, or only considering points
in smaller areas, locally (Esri 2015). Deterministic interpolation methods tried during this thesis
include Spline and Kernel.
22

Geostatistical interpolators not only use the values of surrounding locations directly, but
also the statistical spatial relationship between the values at those locations. Geostatistical
interpolators have the added benefit of providing a measure of uncertainty of the predicted
values. Empirical Bayesian Kriging is a geostatistical interpolation method tested during this
thesis.
Utilizing the most appropriate interpolation method for creating a raster depth surface is a
key aspect of this project. Often, in the field of geology, the traditional minimum curvature
technique is applied when interpolating a geologic surface (Zoraster 2003). In ArcGIS, the spline
interpolation tool is representative of this function.
3.3.1. Spline Interpolation
The Spline interpolation method creates a smooth, gently varying, predicted surface by
using a minimum curvature mathematical function (Esri 2015). The interpolator passes exactly
through all control points, meaning that at the location of a control point, the observed value will
be the same as the predicted value. There are two Spline types available with the ArcGIS tool:
(1) regularized; and (2) tension. The regularized option creates a smoother output and allows
values to extend beyond the data range of the control points. The tension option creates a slightly
more uneven surface by constraining the predicted values to the data range of the control points.
One advantage of the Spline interpolator is that it has the option to honor barriers, such as faults.
However, this project did not require this as the Carrizo-Wilcox fractures are minimal within the
study area, providing slight displacement of rock, and no influence on the movement of fluid
(Phillips 2015). The biggest disadvantage of the Spline interpolation method is the lack of any
uncertainty measurements associated with the predicted surface.
23

3.3.2. Kernel Interpolation
The Kernel interpolation method utilizes a moving window method to predict values at
locations where values are unknown. This local interpolator is similar to the Local Polynomial
Interpolation (LPI) method, but varies by estimating regression coefficients which reduces model
instability (Esri 2015). By altering the ridge parameter, bias in the model can be increased to
allow for a more stable model. Because bias is added to the model, the ridge parameter should be
kept as small as possible. As there are six Kernel functions available in ArcGIS, cross-validation
and validation diagnostics can help determine which Kernel function is most appropriate for the
sample dataset. The Kernel interpolator in ArcGIS has the ability to honor barriers, to a certain
extent. Values can still be interpolated around barriers if the shortest distance between points still
fall within the searching neighborhood specifications.
Unlike the Spline method, the Kernel method is an inexact interpolator, meaning that the
predicted surface does not have to pass through the values of the control points. This feature is
acceptable within this study because of the type of phenomena being modeled. This study is
focused on modeling subsurface geology- the base of the Carrizo-Wilcox aquifer. When geologic
formations are transitioning from one to another, in most cases, there is no hard, defining line
between them. There will be a transition zone where one ends and another begins. Therefore,
some flexibility is acceptable.
3.3.3. Empirical Bayesian Kriging
Kriging is a geostatistical form of interpolation that uses statistics to model spatial
autocorrelation between observed values (Esri 2015). By using statistics, there is probability
associated with the predicted values. Therefore, uncertainty can be quantified by the standard
errors produced (Krivoruchko 2012). Kriging is considered a robust interpolator because there is
24

less prediction uncertainty and error is minimized. Kriging quantifies the spatial dependence
within the data by using semivariogram and covariance functions. A semivariogram is calculated
by comparing the values at each pair of points, and dividing the squared difference of values in
half. Each halved, squared difference is plotted against the distance in points. Once the values for
every pair are plotted, a best-fit model is estimated, and predictions are made using generalized
linear regression techniques (Esri 2015). Within ArcGIS, there are several different kriging
methods available and many parameters which can be tweaked to find the best model. This
requires considerable interaction from the user. Because there can be uncertainty within a
semivariogram model, prediction standard errors are sometimes underestimated.
Empirical Bayesian Kriging (EBK) is one form of Kriging offered within Geostatistical
Analyst of ArcGIS, which automates many of the difficult parameters by using simulated models
of subset data (Esri 2015). The EBK method creates several semivariograms within these subset
neighborhoods (defaulted to 100 points). To accomplish this, a semivariogram is first estimated
using a subset of values, new values are estimated at the input locations based off of the original
semivariogram. Next a new semivariogram is created from the new estimated values. The
original semivariogram is used to repeat the process several times creating many
semivariograms. By averaging these semivariograms, a suitable model is determined without
reliance on the user making arbitrary parameter choices.
Figure 10 shows an example of the many simulated semivariograms plotted together,
using project data of wells with the base of the aquifer interpreted from well logs. Here, the
average value is represented by the solid red line (Esri 2015). Quantile lines represent the 25
th

and 75
th
percentiles and are symbolized by dashed red lines.
25

Figure 10 Simulated Semivariograms created from project data.
Because the regional geology within the study area has a physical trend, where the
formations dip downward to the Southeast, utilizing an EBK model which detrends the data will
provide a surface with less error. When a dataset is nonstationary, meaning that average values
vary over space, EBK offers the ability of removing this large scale variation, by applying a first-
order trend removal (Krivoruchko 2011, Esri 2015). EBK provides several kriging models which
apply transformations to the simple kriging model, three of which will remove data trends. Using
validation and cross-validation diagnostics help determine the best model.
The EBK interpolation method does not have means of employing barriers, such as
faults. Therefore, when interpolating a geologic formation with faults, this would not be an ideal
interpolator. However, in this study, as stated above, faults are not significant. Additionally, like
the Kernel method, the EBK method does not pass exactly through the measured values of the
control points, which is acceptable within this study.
26

3.3.4. Error
Error is described as the difference between an observed value and a predicted value.
Both Kernel and Empirical Bayesian Kriging (EBK) interpolation methods provide predicted
error statistics to aid in selecting a suitable interpolation method. Validation and cross-validation
techniques are used to generate error values and asses the quality of the model (Esri 2015). The
validation process removes a sample of input data points and utilizes the remaining input data
points to create predictions at the omitted point locations. Cross-validation removes one point at
a time and uses the remaining dataset to predict a value at the omitted point. Predicted values are
compared with the control point values to determine the amount of error in the prediction.
One way ArcGIS illustrates the errors produced during the interpolation is by using
scatterplots. Here the predicted value is plotted against the measured value. Another way ArcGIS
presents this data is by providing statistics of the prediction errors. The best model will have a
standardized root-mean-squared prediction error near one and a standardized mean prediction
error near zero (Esri 2015). In addition to the statistical output, the Kernel and EBK interpolators
also provide standard error maps. These maps show the margin of error from the actual value
with a 95% confidence (Krivoruchko 2011). In other words, by adding and subtracting the error
values from the predicated value, there is a 95% chance that the actual value falls within this
range. These error statistics and error maps provide great insight as to how well the selected
model performs.
Table 2 compares the interpolation methods investigated during this project and the
available functionality they offer.
27

Table 2 Comparison of Interpolation Method Functionality.
Functionality Spline Kernel EBK
Exact interpolator Yes  No No
Honors barriers Yes  Yes  No
Provides standard error map No Yes  Yes 
Model spatial autocorrelation No No Yes 
Ability to detrend data No No Yes 

This study uses ArcGIS and the interpolation methods available within it. The Spline
interpolation method was selected to evaluate because it is a traditional algorithm used by
geologists. The Kernel and empirical Bayesian Kriging interpolators were chosen to examine
because they provide quantitative error analysis. Chapter Four provides detail of the
methodologies conducted within the project to create an interpolated surface, and compare
surface casing depths to that surface.

28

Chapter 4 Data and Methods
To determine if the surface casings of oil and gas wells drilled into the Eagle Ford shale are set
deep enough to adequately protect the shallower Carrizo-Wilcox aquifer, geologic and associated
data were manipulated within a Geographic Information System (GIS). GIS is the ideal toolset
for analyzing and managing spatial information in a project such as this. The following sections
first delineate a study area boundary to conduct analysis within. Next, an explanation of the data
needed is provided. The methods discussed in Chapter Three are then conducted within GIS to
complete this analysis. In addition, evaluating the data in a 3D environment brings a true sense of
where features lie in relation to each other, spatially.
4.1 Study Area Boundary
To define the working area of this project, a boundary was created to limit the
investigation spatially. The recently discovered Eagle Ford Shale region was selected because
this play is still very active, with 88 rigs operating in December 2015 (Alford 2015). Because of
this, applicable mapping techniques are needed to provide a good basis for determining surface
casing depths. However, the study area does not cover the entire Eagle Ford drilling region, and
a boundary was created to encompass wells where picks of the Carrizo-Wilcox aquifer has been
completed. Therefore, the study area boundary covers almost half of the Eagle Ford region. The
method demonstrated in this study can easily be extended to cover the entire Eagle Ford area.
Figure 11 displays the study area of this project in relation to the Eagle Ford drilling region. The
red stars represent wells in which the Carrizo sand base has been picked from well logs. These
points were used to create an interpolated surface. The grey dots represent all Eagle Ford wells
drilled along South Texas, as of September 2015.
29

Figure 11 Eagle Ford drilling region in South Texas with study area boundary.

4.2 Data Acquisition and Manipulation
To pursue this project, three primary datasets were needed: (1) well locations of all oil
and gas wells within the study region; (2) purchased well logs of older oil and gas wells; and (3)
surface casing depths of recently drilled Eagle Ford wells. These were all obtained from various
data management companies involved in the oil and gas industry which market such datasets for
a fee. Also, salinity lines provide guides as to where water quality is freshest, and where it
deteriorates.
30

4.2.1. Well Locations
All facts pertaining to the drilling and completion of oil or gas wells in the State of Texas
are recorded with the Railroad Commission of Texas (RRC). The RRC makes much of this data
available for free through their website on a well by well basis through a query interface.
However, it can be tedious querying the data and then combining data from multiple queries. For
a fee, companies such as IHS Energy (IHS) and P2 Energy Solutions (P2) offer well attributes
such as coordinates, well name, drilling details, completion information, etc. in various formats
such as access databases, excel spreadsheets, .csv files for download, or also live direct
connections. These datasets are continually updated, with some attributes updating daily and
others weekly.
X,Y locations of both IHS Energy (IHS) and P2 Energy Solutions (P2), were evaluated.
IHS offers information focused on the drilling and completing of an oil or gas well. Well
locations and various identification data were downloaded through IHS’s data portal, as an Excel
spreadsheet. P2 Energy Solutions provides well locations, in addition to other types of spatial
data, such as abstract lines (original Texas survey lines), ownership lines, lease polygons, etc. P2
provides their data within a file geodatabase, and it was also downloaded for this project through
a data portal. Both sources include pertinent well header information, including well name,
operator, and most importantly, the US Well Number. This US Well Number, previously called
API number, is the unique well identifier (UWI) of a well, and is provided by the Railroad
Commission of Texas.
Preliminary comparisons of well locations provided by the two data sources confirmed
that many well locations, especially those drilled in the 1980s or later, were spatially coincident.
This is likely due to modern surveying technology. However, several older well locations varied
up to eight miles between the two sources. A sample of locations with great variances were
31

further investigated by reviewing the header of the well logs. In the well log header, these older
wells identified spatial locations by using metes and bounds. This is a classic surveying method
which uses a known landmark as a point of beginning and then calls distance and directions from
there. ArcMap was used to measure these distances. Well locations from P2 aligned better with
the metes and bounds recorded on the well log headers. As a result, this dataset was selected to
be used during analysis throughout the remainder of the study.
Figure 12 shows the spatial distribution of oil and gas wells provided by P2. Attributes
used in the table consist of well name and number, operator, X, Y, and US Well Number, which
will later be used for joining to other tables.

Figure 12 Distribution of well locations from P2 Energy Solutions.
32

4.2.2. Well Logs
Geophysical tools are sent downhole in a well to gather borehole data in the form of well
logs. Well logs offer various information about rock characteristics, such as lithology and fluid
content. By identifying the depth at which sandstone composition occurs, the base of the Carrizo
sand can be determined from these logs. Logs used within this study were purchased from a
petroleum data management company, TGS, and are stored in a geological software application
called Petra.
Selecting formation depths, or “picks”, from a well log is an interpretive process,
requiring direct analysis by a geologist or petraphysicist. Figure 13 exhibits a screen capture of
Petra’s cross-section module displaying several logs with correlated depth picks of different
sections of the Carrizo-Wilcox group.

Figure 13 Screen capture of Petra's cross section module displaying vertical correlated well logs
with depth being in feet.
33

Within the study area, there are over 5,000 oil and gas wells with either raster or digital
logs. Raster logs are scanned-in paper logs, like those shown in Figure 13, and digital logs are
provided in a columnar digital file such as a .txt or .csv. However, not all of these wells are
logged through the entire Carrizo-Wilcox, and many of the older raster logs are of poor or
inadequate quality. As a result, a total of 520 wells were selected that have the base of the
Carrizo sands identified on their logs.
These well logs are stored and interpreted within the cross-section module of Petra, a
geological software application. When a geologist identifies a horizon of interest, they create a
new pick entry and this data is stored within the Petra software. These depth picks are recorded
in feet below sea level (FBSL), also known as subsea total vertical depth (SSTVD). In Petra,
these depth picks were queried to create a subset of wells with Carrizo sands base picks. To
create a dataset that ArcMap can read, the depths and identifying well information, such as US
Well Number and well name, were exported as an ASCII file and imported into Excel. In Excel,
the column headers were cleaned up to remove unsupported characters such as spaces, dashes,
etc. After preparing the Excel spreadsheet, it was joined to the well locations in ArcMap, only
keeping matching records. This means that well locations that do not have an aquifer depth
picked from a log were dropped from the layer. The joined dataset was then exported into its
own feature class to use as the control points for the interpolation process. Figure 14 shows the
520 selected wells with Carrizo sand depth picks, symbolized by depth.
34

Figure 14 Carrizo Base Depth Picks from Well Logs in feet.
4.2.3. Surface Casing Depths
Although it was not used for the well locations, the IHS database provides detailed
information pertaining to the drilling and completion process of oil and gas wells. Therefore, it
was used to acquire surface casing depths, recorded with the Railroad Commission. This data, as
well as additional attributes, including well name and number, US well number, operator and
ground elevation were downloaded from the IHS internet data portal in Excel format. Ground
elevation is obtained using Global Positioning System (GPS) equipment, and was useful in
converting the casing depths from measured depth values into subsea values.
35

Because there are many older wells within the study area which were drilled into other
formations, this study is only focused on evaluating the surface casing depths of oil and gas wells
producing from the Eagle Ford formation. The Eagle Ford lies below the Carrizo-Wilcox aquifer,
meaning that the aquifer must be penetrated. Also, this shale play is a fairly recent discovery,
with wells drilled since 2009. This formation was selected to keep the dataset current. Because of
this, regulatory agencies should already have processes in place to accurately identify aquifer
depths and therefore it is assumed would recommend casing depths accordingly. Although there
are many other older non-Eagle Ford wells within the region, they were filtered out of the dataset
in Excel.
The standard size of surface casing in this region is 9.625 inches. Therefore, the wider
conductor casing and the narrower intermediate and production casings, were also filtered out of
the dataset. The “base depth” field gives the depth of the surface casing in measured depth. This
is a positive number, but measures how many feet the casing is from ground elevation. Because
the base depth value is not in subsea format, it needed to be converted. A simple formula which
subtracts the ground elevation from the base depth and then multiplied by -1 was applied within
Excel. Therefore, when displaying data in a 3D environment, a digital elevation model was not
needed to acquire base heights and extrusion values, as these were already within the table.
This spreadsheet is also joined to the well locations in ArcMap, only keeping the records
that match. All other wells were excluded from the study. The joined dataset was then exported
into its own feature class. The final attributes in the feature class include US Well Number, well
name and number, operator, spud date, producing formation, ground elevation, casing diameter,
base depth and casing depth SSTVD. Table 3 displays the attribute data associated with a portion
of the Eagle Ford wells within the study area.
36

Table 3 Attribute data for Eagle Ford wells.

Figure 15 shows the distribution of Eagle Ford wells throughout the study area,
symbolized by surface casing depth.
37

Figure 15 Depth of surface casing of Eagle Ford wells within study area.
4.2.4. Salinity Determination Lines
To understand where water within the Carrizo-Wilcox is freshest and where it becomes
more brackish, identifying these areas spatially, is helpful. A map produced by the U.S. Army
Corp of Engineers (USACE) in 1976, delineates the boundaries between water with less than
1000 mg/L of total dissolved solids (TDS), and water with more than 3,000 mg/L of TDS (Figure
16) (Klemt, et al. 1976). Although the Carrizo-Wilcox is considered a slow moving aquifer, it is
likely that these areas have migrated since 1976. However, more current data was not found, and
it still provides a general reference of salinity. This map was scanned into a .tif image. The image
38

was georeferenced in ArcMap, and the lines digitized into a shape file. Although these lines are
approximate, they provide a good idea of where groundwater salinity varies.

Figure 16 Carrizo-Wilcox thickness map modified from the Army Corps of Engineers, with
salinity lines highlighted in yellow. Source: Klemt et al. 1976.
39

4.3 Methodology
The two primary datasets needed to conduct this study, (1) wells with Carrizo-Wilcox
aquifer depths, picked from logs, and (2) Eagle Ford wells with recorded surface casing depths,
provide the basis for further analysis. This section shows the results of evaluated interpolation
methods, and reveals how surface casing depths of Eagle Ford wells compare to the aquifer
digital representation.
4.3.1. Interpolating a Surface
To recognize the geologic structure of the Carrizo-Wilcox, the next step was to create a
continuous surface representing the base of the aquifer within the study area. Three interpolation
methods were considered and evaluated to determine the most suitable digital representation.
4.3.1.1. Spline
The Spline technique is a deterministic method and creates a smooth minimum curvature
output. If geostatistical tools are not available, this is generally considered to be an ideal
interpolation method for mapping subsurface geology as it has been traditionally used in the field
of earth modeling and geophysics. This exact interpolator method precisely honors the measured
values at the control points and has the option to honor barriers such as faults. Figure 17 shows
the results of the Spline interpolation method. When compared to other interpolators, it is
apparent that the Spline method does not handle the known northwest to southeast trend of the
formation as well. Also, because the map passes through all the measured control points, the
boundaries of the data ranges undulate.

40

Figure 17 Predicted Base of Carrizo-Wilcox Aquifer using Spline Interpolation Technique.
4.3.1.2. Kernel
Unlike the commonly used spline method, the Kernel and Kriging interpolation methods
provide quantitative assessments of their predictions, as well as standard error maps. They are
inexact interpolators, meaning that the resulting surface is not required to precisely pass through
all the measured values at the control points. This is acceptable in this study because there is not
a hard, defined boundary, or depth, between geologic transitions. This section discusses the
Kernel interpolation and the next section focuses on the Kriging method used.
41

Although considered a deterministic interpolator, the Kernel interpolation method is
provided with Geostatistical Analyst in ArcGIS. This interpolator was evaluated and the different
Kernel functions were compared. The Constant function provided a prediction surface with the
best root-mean-square standard error of all Kernel functions tested. Figure 18 shows a map of the
results of the Kernel interpolation method. Near the control points, this output appears to follow
the geological trend of dipping downward to the southeast. However, further away from the
control points, the interpolator does not predict as well. In many areas, it does not predict all the
way out to the study area boundary.

Figure 18 Predicted Base of Carrizo-Wilcox Aquifer using Kernel Interpolation Technique with
the Constant Function.
42

Because it is provided with Geostatistical Analyst in ArcGIS, the Kernel interpolation
method does create a standard error map. Figure 19 shows the standard error map produced from
the constant function Kernel interpolation method, and shows how quickly error increases as you
move away from the control points. Error here is displayed in feet, the same unit of measure as
the predicted values. It is important to take note of the scales in the error maps, as they are
symbolized differently by geometric interval. These default symbology classes were kept to
show the full range of standard error values on each map.

Figure 19 Predicted Standard Error Map produced with Kernel Interpolation.
43

4.3.1.3. Empirical Bayesian Kriging
Empirical Bayesian Kriging (EBK) is a geostatistical interpolator and offers the robust
functionality of kriging, while automating many of the parameter choices required to create
optimal results. One great advantage of the technique, particularly in this study area, is the ability
to detrend the data. Because the geologic trend in the region is for the formations to dip
downward toward the southeast, this option is very powerful. EBK offers the ability to account
for this trend and still provide good quantitative error results. Many semivariogram models of the
EBK method were compared. The Exponential Detrended model worked best with this dataset,
with the root-mean-square standard error closest to one, at 0.9958. Figure 20 shows the results of
the exponential detrended EBK interpolation method. The first thing to note is how well the
interpolator appears to predict past the control points due the detrended model. Recall that the
Kernel interpolator did not predict out to the edges of the study area boundary. In addition, by
applying a detrending function and not strictly honoring the control points, the resulting map
shows smoother boundaries of ranges than the map generated by the Spline interpolation method.
44

Figure 20 Predicted Base of Carrizo-Wilcox Aquifer using Empirical Bayesian Interpolation
Technique with the Exponential Detrended Semivariogram.
The error map generated by the EBK interpolation method shows a smaller range in error
than the Kernel interpolator. The error produced by the EBK method has a minimum value of 8
feet and a maximum value of 450 feet. The error produced by the Kernel interpolator has a
minimum value of 8 feet and a maximum value of 732 feet.
45

Figure 21 Predicted Standard Error Map produced with Empirical Bayesian Kriging
Interpolation.
Table 4 displays a comparison of the error diagnostics generated using the Kernel and
EBK interpolators. Because the Empirical Bayesian Kriging interpolation method provided
better error diagnostics, the resulting predicted surface was selected for further analysis.
46

Table 4 Diagnostic Comparrison between Kernerl and EBK interpolation methods.
Error Diagnostics Kernel EBK Objective
Root-mean-square standardized
prediction error
1.0148 0.9958 Closest to 1
Mean Standardized -0.1410 -0.0143 Closest to 0
Root-mean-square prediction error 66.512 54.144 Lowest value
Average Standard Error 65.006 53.259 Closest to RMS prediction error

Figure 22 provides a side-by-side comparison of the evaluated interpolated surfaces with
their standard error maps. All prediction maps are symbolized with the same ranges as they were
in previous figures. Both EBK and Kernel standard error maps have had their symbolized ranges
adjusted to match each other. This provides a better sense of the true difference in error. The
Kernel error map looks similar to the previous figure (Figure 19). However, in the EBK standard
error map, looks different than previously (Figure 21). The dark red areas no longer appear
because with the new symbology those darker colors now fall in the 500 to 800 foot range.

47

Figure 22 Side-by-side comparison of evaluated interpolated predictive depth surfaces (feet) and
their accompanying standard error map (feet).
48

Overall, the surfaces produced by the three interpolators were very similar. Because the
Empirical Bayesian Kriging method has the best root-mean-square standard error, predicts well
past the control points and models the southeast downward dip of the regional geology the best,
it was selected to be used for continued analysis.
4.3.2. Depth Comparison
After selecting a final interpolated surface, the next step was to compare the depths of the
surface casings against the predicted depths from the predicted aquifer. To accomplish this, the
GA to Points tool was used to extract predicted and error values from the digital surfaces to each
point, in this case the Eagle Ford wells. It was then simple to determine which wells do not fully
protect the aquifer by subtracting the surface casing depth from the predicted value of the aquifer
base. Any results that are a negative value do not penetrate the base of the aquifer. Figure 23
shows which wells penetrate the predicted surface (white) and which wells do not (black). Also
shown are the lines representing the boundaries of water quality zones obtained from the Army
Corps of Engineers. These lines separate the freshwater area to the northwest at 1,000 mg/L or
less of total dissolved solids (TDS), and the brackish area at 1,000 – 3,000 mg/L of TDs. This is
important when determining if “usable” water is protected and is investigated further in the next
chapter.
49

Figure 23 Predicted base of aquifer surface showing wells with surface casing that penetrate the
surface (white) and well with surface casing that do not (black).
Additional fields were calculated to help understand error. Each point record in the table
contains the predicted value and standard error value. By adding and subtracting the standard
error values from the predicted values, this creates a range of bounding minimum and maximum
values in which the actual value falls within with 95% confidence.
4.3.3. 3D Representation
ArcScene provides a 3D environment to view datasets containing elevation information.
By viewing in 3D, a user can get a true sense of where subsurface features lie in relation to each
other. The interpolated surface representing the Carrizo-Wilcox aquifer is an elevation surface
50

where the units are in sub-sea total vertical depth (SSTVD). Therefore, it was easily brought into
an ArcScene project. The point feature class representing Eagle Ford wells contain ground
elevations and surface casing depths. Before this layer can be displayed properly in ArcScene,
the Feature to 3D by Attribute tool was used to assign the ground elevation field as Z values of
the layer. After this was completed, the point layer was displayed correctly in vertical space. To
create a vertical line representing the depth of the surface casing, the point layer was extruded to
the SSTVD depth. Another copy of the Eagle Ford well layer was added to the ArcScene project.
This time the points were extruded to a common value of -15,000. This layer was displayed in a
different color and represents the remaining vertical portion of the well.
The same Eagle Ford well layer was added a third time. On this layer, the base height
was set to the Minimum field, which represents the shallower margin of error of the predicted
surface. This layer was extruded to the Maximum field, which represents the deeper margin of
error. By doing this, the layer was symbolized to create an error halo above and below the
predicted surface. Figure 24 is a capture of the ArcScene project containing these symbolized
layers. Here, one can quickly see if a well’s surface casing is shy of penetrating, and thus
protecting, the aquifer, and by how much.
51

Figure 24 Screen capture of 3D environment displaying a predicted surface, representing the
base of the aquifer (blue), standard error halos (yellow) and surface casings (red).

Chapter Five provides quantitative analysis of the results produced while conducting the
methodologies outlined in Chapter Four.

52

Chapter 5 Results
As mentioned previously, two goals of this project were: (1) to find an appropriate interpolation
method to create a digital model of the base of Carrizo-Wilcox aquifer; and (2) to determine if
regional oil and gas wells penetrate this surface. Chapter Five summarizes the results found by
the methods defined in Chapter Four.
5.1 Is the Carrizo-Wilcox Aquifer Protected?
After analysis, it was determined that the groundwater within the Carrizo-Wilcox is not
completely protected from oil and gas operations. Of the Eagle Ford wells drilled within the
study area, 56.5% do not have surface casings that penetrate the surface representing the base of
the aquifer. Figure 25 shows the overall percentages of aquifer protection.

Figure 25 Pecentages of wells with surface casing that penetrate the predicted surface
representing the aquifer base.

53

Figure 26 is reproduced from the previous chapter for ease of analysis here, and provides
a spatial representation of Eagle Fords wells within the study area. Wells in white have surface
casings that are deeper than the predicted aquifer, and would therefore protect it. Wells in black
have surface casings shallower than the predicted aquifer and therefore leave it susceptible to
contamination.

Figure 26 (Reproduced Figure 23) Predicted base of aquifer surface showing wells with surface
casing that penetrate the surface (white) and well with surface casing that do not (black).
It is interesting to see the overall spatial distribution of aquifer protection. As the
formation dips downward to the Southeast, the amount of total dissolved solids (TDS) in the
water increases, causing the water within the aquifer to become more brackish. The lines in
54

yellow provide a rough boundary where TDS values become higher than 1,000 mg/L, and higher
than 3,000 mg/L. For the most part the aquifer is protected where water is freshest at 1,000 mg/L
or less. However, where water has between 1,000 mg/L and 3,000 mg/L, the Southwest region
has many wells whose surface casings do not penetrate the aquifer. Where water is the most
brackish at 3,000 mg/L TDS or more, there are very few wells that actually penetrate the base of
the Carrizo sands. Figure 27 shows the percentages of wells within the study area which
penetrate the base of the aquifer, grouped by TDS.

Figure 27 Percentages of Eagle Ford Wells whose surface casings penetrate the aquifer by TDS.
Where water is freshest at 1,000 mg/L or less, and used for human consumption, 80.6%
of Eagle Ford wells within the study area had surface casings which penetrated the Carrizo-
Wilcox aquifer. However, this percentage decreases immensely to 22% where groundwater
within the aquifer is brackish to saline at 3,000 mg/L or more.
55

5.2 Error
There is added value in obtaining the standard error values from the interpolators that
provide them. This offers a quantitative assessment of how close the predicted values are to the
actual values. In this study the Empirical Bayesian Kriging (EBK) was selected to create a
surface representing the base of the aquifer. In addition to predicted values, this interpolator also
provides standard error values. The standard error values can be used to calculate a range by
adding and subtracting the standard error from the predicted value. The actual value is likely to
fall into this range with 95% confidence. Figure 28 exhibits the EBK surface of standard error
values along with the distribution of Eagle Ford wells. Areas near the control points, or well logs
with depth picks, have less error than areas far from the control points. The areas with dark blue
circles are where the original control points lie. One item to point out is that nearly all standard
error falls within the range of 200 feet or less.

56

Figure 28 Standard Error Surface: Actual value falls within the range of the predicted value, plus
or minus the standard error with 95% confidence.
By using the GA Layer to Points tool in ArcGIS, the standard error values from the error
surface is extracted to the Eagle Ford well points. The bar chart below (Figure 29) represents the
number of wells within each error range. This bar chart uses the same ranges as the previous
map. Most of the error values fall within 30 to 60 feet, with over 900 wells in the study area
falling in these categories. Again, nearly all error values are 200 feet or less. If surface casing
continues to 200 feet deeper than the predicted value, that should be sufficient to protect the
aquifer.

57

Figure 29 Histogram of Standard Error Values for the EBK surface.
By adding and subtracting the standard error from the predicted value, a range can be
created in which there is 95% confidence that the actual value will fall. Using the minimum error
value and the maximum error value, these ranges are shown as error halos within the 3D
environment (Figure 30).
58

Figure 30 Close-up of surface casings, the predicted aquifer surface, and the margin of error.
By including the margin of error, this somewhat alters the final percentages of how many
Eagle Ford wells with surface casings penetrate the base of the aquifer. If we use the shallowest
error value to represent the predicted surface, the number of Eagle Ford wells in the study area
with surface casings going below that increases 1.5%, from 43.5% to 45%. This is still not half
of the wells within the study area. By using the deepest error value, the number of wells which
have surface casing deeper than the base of the aquifer decreases 2%, from 43.5% to 41.5%.
59

Figure 31 Percentage of wells for which the surface casings penetrate the shallower error value
(left), and deeper error value (right). Blue represents surface casings that protects the aquifer and
red represents surface casings that do not.
5.3 Aquifer protection over time
To identify any temporal patterns, Eagle Ford wells were grouped by the year they were
drilled, to see if surface casing depths have improved or declined over the years. Figure 32 shows
a bar chart of the percentage of wells within the study area which have surface casings that
penetrate the aquifer, by year. This chart demonstrates that there has not been a steady trend
throughout the years. Surface casings which penetrate the aquifer generally improved through
2013. After 2013, the number of wells protecting the aquifer decreased until the time of this
study. This outcome was somewhat surprising as it was believed that maps and regulations
would improve over the years, committing operators to set deeper surface casings. However,
these results show otherwise. Figure 33 shows the spatial distribution of Eagle Ford wells
symbolized by spud year, or year that the well was drilled. It is evident that there is no clear
spatial trend of when wells were drilled.

Shallow
Deep
Shall
De
60

Figure 32 Percentage of wells whose surface casings penetrate the predicted aquifer surface, by
year. Blue represents surface casings that protect the aquifer and red represents surface casings
that do not.
61

Figure 33 Eagle Ford Wells symbolized by year it was drilled.
5.4 Difference between depth of surface casing and predicted aquifer depth
It has been established that the Carrizo-Wilcox is not completely protected within the
study area. The next question to address is “By how much?” Figure 34 displays a bar chart which
reflects the difference between the surface casing depth and the predicted aquifer depth. One
important item to point out is that most wells with surface casings that do not penetrate the base
of the aquifer are short by over 2,000 feet. This leads one to conclude that when groundwater is
considered “unusable” (and therefore protection is not needed), the casing is set well above the
depth of the formation.
62

Figure 34 Difference between surface casing depth and predictd aquifer base depth.
The map below (Figure 35) displays the spatial distribution of the difference in surface
casings across the region. The ranges used in the map are the same as the ranges in the previous
bar chart. The surface casings of the wells in blue penetrate the base of the aquifer and wells in
orange do not. As the colors become darker, the distance between the surface casing depth and
the aquifer depth increases. The map clearly shows that when water is considered brackish to
saline (or above 3,000 mg/L TDS), surface casing is set far above the base of the aquifer.

63

Figure 35 Difference between surface casing depth and base of aquifer in feet.

64

Chapter 6 Conclusions
Due to the Eagle Ford shale, South Texas has experienced a recent increase of drilling activity.
Because the Carrizo-Wilcox lies above the Eagle Ford formation, it is being compromised every
time the surface casing is not set below the base of the aquifer.
To understand the regional structure of the Carrizo-Wilcox, well logs were used to
identify formation depth within a study area. Empirical Bayesian Kriging was selected as a
suitable method to interpolate a continuous surface representing the base of the aquifer. The
standard error values and accompanying map provide a quantitative assessment of how well the
interpolator performs. These errors provide a range within which there is a 95% chance that the
actual value falls.
Comparison of the surface casing depths of 3700 Eagle Ford wells against the predicted
surface representing the base of the aquifer, produced unexpected results. Although it was
expected that there would be some wells with surface casings which do not fully penetrate the
base of the aquifer, it was surprising to discover that over half of Eagle Ford wells drilled within
the study area fall into this category, even when considering the base depth at the shallowest
value of the standard error range. These results provide justification for further extensive
investigations.
In addition, the area examined during this study covers only a portion of the Eagle Ford
drilling region. The same methods should be continued throughout the remainder of the region to
determine if similar trends apply.

65

6.1 Limitations and Observations
This study was conducted on the premise that no previous groundwater contamination
had occurred due to oil and gas activity. However, when considering the number of wells drilled
in the region throughout the past, this seems improbable (refer to Chapter Four, Figure 12). It is
likely that many older wells did not have standards regarding surface casings in place, as they are
today. However, when dealing with old wells, especially prior to 1970, determining which wells
penetrated the aquifer is a much more difficult task, as most drilling records are not complete.
Often information such as wellbore depth and casing depths are not readily available, especially
in a database form. Nonetheless, it would be worth conducting research and similar procedures
on historic wells, if available.
As mentioned in Chapter Two, having groundwater quality sampled before oil and gas
operations, would be an ideal situation. This would provide a baseline, and would better indicate
if oil and gas operations were deteriorating groundwater quality. This could be accomplished
during the exploration phase of a newly discovered, petroleum producing formation. Regulatory
agencies could require operators to drill and sample a water well within a specified distance of
newly drilled oil or gas well. After a number of wells were drilled, sampling at these locations
could then be conducted on a regular basis by one of Texas’s regulatory agencies.
6.2 Current and Future Work
As aquifers are an important topic, research is increasingly being conducted in an effort
to model, map and protect this vital resource. The Texas Water Development Board recently
announced a new project to be conducted by the Bureau of Economic Geology and INTERA,
Inc. to map the fresh, brackish, and saline groundwater in the Carrizo-Wilcox (Texas Water
Development Board 2015). The project is scheduled to be completed in 2017 and will contribute
66

to the desired future condition of GMA 13. As lines delineating areas of salinity are investigated
and perhaps modified since those created in 1976, it will be interesting to see how that could
affect the results determined within this study.
Future work should also include modeling groundwater flow within the aquifer. Although
the Carrizo-Wilcox groundwater moves slowly within the aquifer- it still moves. Increased
pumping of fresh groundwater will affect this movement. Good groundwater flow modeling
techniques could help determine if and by how much these regions are hydrologically connected.
Even if protecting brackish water from oil and gas operations is not currently in the interest of
the RRC because of usability, it should still be considered in case contamination could move
through the aquifer to fresher “usable” areas.
One final thought regarding regulatory agencies involved with groundwater, is that they
currently appear disconnected from each other. As new ideas and projects continue to develop
and mature, communication amongst the agencies is vital for groundwater protection.

67

References
Alford, Elizabeth. 2015. Eagle Ford Shale. December 5. Accessed December 9, 2015.
http://eaglefordshale.com/news/eagle-ford-rig-count-remains-steady-2/.
Boghici, Rudu. 2009. Water Quality in the Carrizo-Wilcox Aquifer, 1990-2006. Austin: Texas
Water Development Board.
Bonanza Creek. 2015. Technology & Innovation. Accessed Nov 2015.
http://www.bonanzacrk.com/about-bonanza/technology-innovation/.
Camp, C V, and J E Outlaw Jr. 1993. "Constructing Subsurface Profiles Using GIS." Advances
in Engineering Software 18: 211-218.
Caumon, G, P Collon-Drouaillet, C Le Carlier de Veslud, S Viseur, and J Sausse. 2009.
"Surface-Based 3D Modeling of Geological Structures." International Association for
Mathematical Geosciences 41: 927-945.
Childs, Colin. 2004. "Interpolating Surfaces in ArcGIS Spatial Analyst." ArcUser.
Crow, Cassi L. 2012. "Water-Level Altitudes and Continuous and Discrete Groundwater Quality
at and near an Aquifer Storage and Recovery Site, Bexar, Atascosa, and Wilson Counties,
Texas, June 2004-September 2011."
Dutton, Allan R, Bob Harden, Jean-Philippe Nicot, and David O'Rourke. 2003. Groundwater
Availability Model for the Central Part of the Carrizo Wilcox Aquifer. Technical Report,
Austin: The University of Texas at Austin.
Eagle Ford Shale. 2015. Eagle Ford Production. October 27. http://eaglefordshale.com/eagle-
ford-production/.
Esri. 2015. An Overview of the Raster Interpolation Toolset. October 23.
http://desktop.arcgis.com/en/desktop/latest/tools/3d-analyst-toolbox/an-overview-of-the-
interpolation-tools.htm.
—. 2015. Deterministic Methods for Spatial Interpolation. October 23.
http://desktop.arcgis.com/en/desktop/latest/guide-books/extensions/geostatistical-
analyst/deterministic-methods-for-spatial-interpolation.htm.
—. 2015. How Kernel Interpolation with Barriers Works. October 23.
http://desktop.arcgis.com/en/desktop/latest/guide-books/extensions/geostatistical-
analyst/how-kernel-interpolation-with-barriers-works.htm.
—. 2015. How Spline Works. October 25. http://desktop.arcgis.com/en/desktop/latest/tools/3d-
analyst-toolbox/how-spline-works.htm.
—. 2015. "Performing Cross-Validation and Validation." ArcGIS for Desktop. October 26.
http://desktop.arcgis.com/en/desktop/latest/guide-books/extensions/geostatistical-
analyst/performing-cross-validation-and-validation.htm.
—. 2015. "Random Processes with Dependence." ArcGIS for Desktop. October 26.
http://desktop.arcgis.com/en/desktop/latest/guide-books/extensions/geostatistical-
analyst/random-processes-with-dependence.htm.
—. 2015. "Using Cross-Validation to Assess Parameter Values." ArcGIS for Desktop. October
26. http://desktop.arcgis.com/en/desktop/latest/guide-books/extensions/geostatistical-
analyst/using-cross-validation-to-assess-parameter-values.htm.
—. 2015. What are the Different Kriging Models. October 23.
http://desktop.arcgis.com/en/desktop/latest/guide-books/extensions/geostatistical-
analyst/what-are-the-different-kriging-models-.htm.
68

—. 2015. What is Empirical Bayesian Kriging. October 24.
http://desktop.arcgis.com/en/desktop/latest/guide-books/extensions/geostatistical-
analyst/what-is-empirical-bayesian-kriging-.htm.
Esri: Help Files. n.d. "Introduction to Interpolation Methods."
—. n.d. "Surface Creation and Analysis."
Evenick, Ph.D., Jonathan C. 2008. Introduction to Well Logs and Subsurface Maps. Tulsa:
Pennwell Corporation.
FracFocus. 2015. Well Construction & Groundwater Protection. Accessed June 15, 2015.
http://fracfocus.org/hydraulic-fracturing-how-it-works/casing.
Gearhart, Norman. 2015. "Email message to Railroad Commission of Texas." July 24.
—. 2014. "Online Information and Completing Groundwater (GAU) Applications." Railroad
Commission of Texas. Accessed June 28, 2015. www.rrc.state.tx.us/media/22886/9-gau-
online-information-and-completing-applications-2014.pptx.
George, Peter G, Robert E Mace, and R Petrossian. 2011. Aquifers of Texas. Austin: Texas
Water Development Board.
Gronewold, Nathanial. 2014. "Tighter Well-integrity Rules Take Effect in Texas." The Texas
Environmental News. February 7. Accessed June 28, 2015.
http://txenvironews.org/articles/fossil-fuels-governmentpolitics-waterriversgulf/tighter-
well-integrity-rules-take-effect.
Holloway, Michael D, and Oliver Rudd. 2013. Fracking: The Operations and Environmental
Consequences of Hydraulic Fracturing. Beverly: Scrivener Publishing.
Huang, Yun, Bridget R Scanlon, Nicot Jean-Phillipe, Robert C Reedy, Alan R Dutton, Van A
Kelley, and Neil E Deeds. 2012. "Sources of Groundwater Pumpage in a Layered Aquifer
System in the Upper Gulf Coastal Plain, USA." Hydrogeology Journal (20): 783-796.
Hyne, Ph.D., Norman J. 2001. Nontechnical Guide to Petroleum Geology, Exploration, Drilling
and Production. Tulsa: PennWell Corporation.
Klemt, William B, Gail L Duffin, Glenward R Elder, and TWDB Geologists. 1976. Ground-
water Resources of the Carrizo Aquifer in the Winter Garden Area of Texas; Volume 1.
Report 210, Austin: Texas Water Development Board.
Krivoruchko, Konstantin. 2012. "Empirical Bayesian Kriging Implemented in ArcGIS
Geostatistical Analyst." ArcUser 6-10.
—. 2011. Spatial Statistical Data Analysis for GIS Users. Redlands, CA: ESRI Press.
Lesikar, Bruce, Ronald Kaiser, and Valeen Silvy. 2002. Questions About Groundwater
Conservation Districts. College Station: Texas A&M University.
National Groundwater Association. 2010. National Groundwater Association. October 18.
http://www.ngwa.org/fundamentals/use/pages/groundwater-facts.aspx.
Phillips, Nestor. 2015. Geologic Advisor (October).
Porter, Charles R. 2014. Sharing the Common Pool; Water Rights in the Everyday Lives of
Texans. College Station: Texas A&M University Press.
Railroad Commission of Texas. 2015. Oil and Gas. July 20. Accessed July 26, 2015.
http://www.rrc.state.tx.us/about-us/history/informal-history-toc/oil-and-gas/.
San Antonio Water System. 2015. Brackish Groundwater Desalination. October 25.
http://www.saws.org/your_water/waterresources/projects/desal.cfm.
Sansom, Andrew. 2008. Water in Texas: An Introduction. Austin: University of Texas Press.
Tearpock, Daniel J., and Richard E. Bischke. 2002. Applied Subsurface Geological Mapping.
New Jersey: Prentice-Hall PTR.
69

Texas Administrative Code. 2014. "Part 1, Chapter 3, Rule 3.13, Casing, Cementing, Drilling,
Well Control, and Completion Requirements." Title 16: Economic Regulation. January 1.
Texas Water Development Board. 2015. "Desired Future Conditions." Texas Water Development
Board. October 26.
http://www.twdb.texas.gov/groundwater/management_areas/DFC.asp.
—. 2015. Groundwater. November 9. https://www.twdb.texas.gov/groundwater/.
—. n.d. Groundwater Conservation District (GCD) FAQs. Accessed June 26, 2015.
http://www.twdb.texas.gov/groundwater/faq/index.asp.
—. 2015. Mapping fresh, brackish, and saline groundwater in the Carrizo-Wilcox, Queen City,
and Sparta aquifers. Accessed November 23, 2015.
http://www.twdb.texas.gov/groundwater/models/research/czwx_qcsp_s_mapping/czwx_q
csp_s_mapping.asp.
Texas Water Development Board. 2013. "Water Survey Groundwater Pumpage Estimates." Data
Table.
Zoraster, Steven. 2003. "A Surface Modeling Algorithm Designed for Speed and Ease of Use
with All Petroleum Industry Data." Computers & Geoscience 1175-1182.

Abstract (if available)

Abstract As groundwater is a vital resource, it is important that oil and gas operations do not jeopardize water quality. Many consumers, including farmers and municipalities, rely year after year on the freshwater provided by aquifers. Along South Texas, oil and gas companies are targeting the Eagle Ford formation containing hydrocarbons. In this same region, the Carrizo-Wilcox aquifer must be drilled through to reach the Eagle Ford below. To protect the above aquifer, cemented surface casing is used to seal the Carrizo-Wilcox from contaminants within the well borehole. This study incorporated Geographic Information Systems (GIS) to evaluate surface casing depths of oil and gas wells, to verify if they are deep enough to adequately protect the aquifer. To understand the geologic structure occurring in this region, aquifer depths obtained from well logs were used to interpolate the base of the Carrizo Sands. After comparing three interpolation methods, the Empirical Bayesian Kriging (EBK) interpolator, using the Exponential Detrended semivariogram, was selected to create a predicted surface and a standard error map. Surface casing depths of Eagle Ford wells were mapped and queried to determine if they are deeper or shallower than the predicted surface representing the aquifer. Over half of the wells within the study area had surface casing shallower than the aquifer. However, most of those fell within areas where groundwater was brackish. Results from this study should motivate regulatory agencies in tightening up policies and guidelines pertaining to oil and gas operations affecting aquifers within the State of Texas. In addition, methodologies conducted during the study provide a viable means to improve the current process of determining surface casing depths.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Filling in the gaps: 3D mapping Arizona's Basin and Range aquifer in the Prescott Active Management Area

PDF

Modeling nitrate contamination of groundwater in Mountain Home, Idaho using the DRASTIC method

PDF

Integration of topographic and bathymetric digital elevation model using ArcGIS interpolation methods: a case study of the Klamath River Estuary

PDF

Soil lead contamination from the Exide battery smelter: the role of spatial scale in cleanup efforts

PDF

Design and implementation of an enterprise spatial raster management system

PDF

Surface representations of rainfall at small extents: a study of rainfall mapping based on volunteered geographic information in Kona, Hawaii

PDF

Creating a well database and web mapping application: using geographic information systems to manage and monitor groundwater resources in Sonoma County, California

PDF

Using GIS to identify potential dynamic marine protected areas: a case study using shortfin mako shark tagging data in New Zealand

PDF

Out-of-school suspensions by home neighborhood: a spatial analysis of student suspensions in the San Bernardino City Unified School District

PDF

Demonstrating GIS spatial analysis techniques in a prehistoric mortuary analysis: a case study in the Napa Valley, California

PDF

Development and evaluation of a USV based mapping system for remote sensing of eelgrass extent in southern California

PDF

Geological modeling in GIS for petroleum reservoir characterization and engineering: a 3D GIS-assisted geostatistics approach

PDF

A spatial narrative of alternative fueled vehicles in California: a GIS story map

PDF

The role of GIS in asset management: integration at the Otay Water District

PDF

Precision agriculture and GIS: evaluating the use of yield maps combined with LiDAR data

PDF

Using pattern oriented modeling to design and validate spatial models: a case study in agent-based modeling

PDF

A comparison of GLM, GAM, and GWR modeling of fish distribution and abundance in Lake Ontario

PDF

Development of a Web GIS application to aid marathon runners in the race selection and planning process

PDF

Exploring the impact of the natural gas leak on home values in Porter Ranch, California

PDF

Satellite derived bathymetry in the Canadian Archipelago using multispectral and LiDAR space-based remote sensing

Asset Metadata

Creator Nienkamp, Mary Elaine (author)

Core Title Evaluating surface casing depths of oil & gas operations in an effort to protect local groundwater: a GIS enabled process

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Publication Date 02/24/2016

Defense Date 12/03/2015

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

Tag aquifer,Carrizo,GIS,groundwater,interpolation,kernel,kriging,OAI-PMH Harvest,oil and gas,spatial,spline,surface casing,Texas,Wilcox

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Kemp, Karen (committee chair), Fleming, Steven (committee member), Swift, Jennifer (committee member)

Creator Email maryen.tx@gmail.com,nienkamp@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-214856

Unique identifier UC11278874

Identifier etd-NienkampMa-4157.pdf (filename),usctheses-c40-214856 (legacy record id)

Legacy Identifier etd-NienkampMa-4157.pdf

Dmrecord 214856

Document Type Thesis

Format application/pdf (imt)

Rights Nienkamp, Mary Elaine

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 a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA