Allen Finchum, Assistant Professor
Joseph Seig, Coordinator
Department of Geography
Oklahoma State University
�
Evaluating
Environmental Risks of Petroleum Pipeline Product Spills: The Case of Oklahoma
Lakes and Rivers
�
Acquisition
of Sensitive Points
�
Oklahoma is a crossroads of North America's petroleum
product pipelines. Pipeline operators have recently realized the need to deal
with spills into surface water sources. While Oklahoma has a relatively low
population density, most of the municipal water supplies are from surface
reservoirs. These lakes and rivers also provide habitat for a wide variety of
species. To address the need for a rapid response system, a major pipeline
company in Oklahoma, with the help of the OSU Geography Department, has built a
detailed GIS using ArcView GIS. The data was acquired from public domain,
proprietary, and GPS field survey sources.
�
INTRODUCTION� Top
Oklahoma has been at the center of the oil industry
in North America throughout in 20th century. From the period of the early oil
"booms" in Oklahoma and Texas in the early part of the century
through until the 1990's, pipelines have been used to transport oil and other
petroleum products throughout the region and the country. While petroleum
product pipelines experience fewer spills than trucks or rail transport
systems, even these "stationary" transport mechanisms occasionally
experience product spillage. These pipelines often cross streams, rivers, and
lakes in order to move petroleum products throughout the region. In the event
of a spill into one of these water resources, the spill is not geographically
contained, and potentially impacted sites along the edge of the water body can
be impacted for an extended distance.
The primary focus of this project was the creation of
a GIS based support database to aid planners and emergency clean-up crews in
preparing for and conducting spill clean-up operations in the event of a
product spill into a water resource.
THE REGIONAL SITUATION� Top
The pipeline in question is operated by the Conoco
Pipeline Company, a subsidiary of Conoco, Incorporated, a long standing
Oklahoma based petroleum product producer. The pipeline system has several
major trunk lines crossing central and eastern Oklahoma, and one major trunk
extending northeastward from Tulsa, Oklahoma to East St. Louis, Illinois
(Figure 1). This region is the home of numerous major rivers, lakes, and other
minor streams, many of which are crossed by
the pipeline.
In late 1996, the need for information regarding the
potentially impacted sites along streams and rivers crossed by the pipeline became a major focus of interest
for the pipeline operators of the system outlined above. It was determined that
in order to minimize the potential impact and financial risk from a product
spill, a database of potentially impacted sites along the streams should be
assembled. At this point, Conoco approached the Department of Geography at
Oklahoma State University for assistance in assembling the database and
providing a geographically based search solution for user referencing of the
data. The database and ArcView GIS based search and display systems are
described and shown in this paper.
The first data layer to be used for the planned
acquisition of sensitive point information was the Conoco Pipeline shown in
Figure 1. This geographic layer was developed using Global Positioning System
(GPS) technology, carried by Conoco employees and refined with the assistance
of the Department of Geography at Oklahoma State University (OSU). OSU provided
base station support for post-processing and differential correction of the raw
data collected, and in verifying the accuracy of the data collected in these
efforts. Conoco uses this layer to aid in the management of the pipeline
system, and internally maintains a full attribute dataset for the pipeline
layer, including lengths, capacities, flows, and cutoff valve information. This
layer provided crossing locations and potential spill amounts and product types
used in determining the sections of rivers and streams potentially impacted and
what type of risk the area could be experiencing.
PUBLIC
DOMAIN BASE DATA� Top
Prior to the assembly of the potentially impacted
sites along the selected rivers, streams, and lakes, base-referencing data was
assembled in order to allow for the reasonable display and association of the
mapped points. These data came from numerous information sources, and provide
the user with various referencing information in determining the location and
situation of potentially impacted sites in the event of water based pipeline
product spill. The primary geographic data source was the National
Transportation Atlas Database for 1997 (NTAD), a geographic database produced
and distributed by the Bureau of Transportation Statistics and the Federal
Highway Administration. This set of geographic data was selected because it is
public domain data, allowing Conoco to freely distribute the base data as well
as the data concerning the collected sites without concern for ownership or
other legal restrictions.
Data layers used from the NTAD are outlined in the
Table 1:
TABLE 1
|
State and County Boundaries |
|
Urban Area Boundaries (FHWA Urban Area
Definitions) |
|
National Highway Planning Network (NHPN) |
|
Railroads (BTS) |
|
Navigable Water Channels and Lakes (BTS, Corps of
Engineers) |
|
Airports (BTS) |
|
TOFCCOFC facilities (BTS) |
|
Major Military Facilities (DoD) |
The NHPN was also selected to provide users with the
capability of determining the shortest path between potential emergency
clean-up personnel and the spill control location. The location of railroads,
airports, TOFCCOFC, and military bases also provide emergency response planners
and leaders with information as to the location of other vital facilities in
the region (Figure 2).
A five-state region was selected to allow for a
complete coverage of the entire area from which emergency response crews could be
gathered, as well as all areas potentially impacted by water borne spills such
as those outlined here. These states included Oklahoma, Missouri, Texas,
Arkansas, and Kansas.
In addition to the NTAD Layers described above, one
additional layer included was a DRG Layer of 1 /1 00,000 USGS Topographic
(Topo) Maps to be used for small area analyses of specifically impacted areas.
Figure 3 provides an example of the use of this type of background layer. In
combination, these layers provide the users of the entire database package a
reasonable base for identifying the location of and situation of the
potentially impacted sites along the selected streams.
�
ACQUISITION
OF SENSITIVE POINTS� Top
In May 1997, graduate student assistants of the
Department of Geography at Oklahoma State University undertook the process of
locating, identifying, and cataloging potentially impacted sites. This process
proved to be extensive, requiring the effort of two 'traveling" TA's over
two full summer terms (four months). During this period, 1,998 points were
located and cataloged; including 283 state identified drinking water intakes
and well-heads.
Sites were selected and identified by actual
inspection by OSU Graduate Research Assistants traveling the entire length of
the potentially impacted stream from 1-2 miles upstream of the crossing through
to the nearest dam or other impediment, or the end of the potential damage zone
as calculated by Conoco engineers. Two research assistants were dispatched,
working each side of the selected rivers or the entire coastline of potentially
impacted lakes; each with a Trimble Geo-Explorer GPS Unit pre-loaded with a
specific data-dictionary to allow easy coding of each point. The
data-dictionary included the base classification, on-stream/off-stream coding,
the research/river zone, and film and picture number.
The research assistants also kept detailed notes as
to directions, name and/or ownership of the property, and the county in which
the sensitive point is located. The assistants returned to OSU approximately
once a week, to download point location and attribute data. At this point a
third research assistant processed the downloaded GPS locations, and
differentially corrected the locations to obtain the most accurate location
information possible, using the OSU GPS Base Station located in the Department
of Geography. After correction using Trimble P-Finder, the points were
converted and loaded to ArcInfo 7.1 on a Windows NT platform. Following this
the points were loaded to ArcView GIS 3.1, and the written attribute data
(directions, name, etc.) were also loaded as an attribute table to ArcView and
joined to the data points to allow for querying, as will be shown in a later
section.
Other than the previously identified well-heads and
surface drinking water intakes, data gathered for each site included the exact
location as provided by the Global Positioning System using Trimble
Geo-Explorer receivers, road directions to the site, the county, and the
classification of the site. Such sites were classified as shown in Table 2.
TABLE 2
|
Water Intake |
73 |
|
Medical Facilities |
2 |
|
Fish and Wildlife |
36 |
|
Transportation Facilities |
467 |
|
Recreational Areas |
167 |
|
Other Economic |
13 |
|
Boat Ramps |
203 |
|
Residential Sites |
204 |
|
Water Resources |
389 |
|
Utilities |
43 |
|
Drinking Water |
7 |
|
Businesses |
95 |
|
Staging Area |
13 |
|
School/Educational Facilities |
2 |
�������������������������������������������������������������������������������
Figure 3 shows the entire pipeline region with all of
the potentially impacted points displayed. As can be noted in this Figure, the
majority of the sensitive points follow along the rivers and lakes within the
region. Also notable in this Figure is that the extent of the identified points
at times appears stunted along the downstream path of the river's and lakes.
This is often due to the fact that dams function as a barrier to the downstream
flow of spilled petroleum products. Conoco engineers computed the maximum
downstream impact distance of spilled products during full stream flow and/or
flood conditions. These distances often represented movement further downstream
than the manmade obstacle represented by a dam would allow. Therefore, no
potentially impacted points would exist or were identified beyond such dams.
ANALYSIS
OF SENSITIVE POINTS� Top
As can be noted from Table 2, Transportation
Facilities, Water Resources, Residential Sites, and Boat Ramps are the most
common sites identified as potentially impacted points for consideration for
protection and/or clean-up. In planning for and conducting clean-up exercises,
the pipeline operator reclassified the groups listed above based on the
following broader groupings: Human Impact, Environmental Impact, and Economic
Impact. If clean up or protection resources are limited, those points
classified as having direct human impact are given the highest priority,
environmental resources are second, and economic resources third. However, items
such as bridge abutments (Transportation Facilities) are often given high
priority because damage to such facilities often impacts the ability to provide
clean-up or protection services to other impacted sites. A limited protection
effort could be necessary due multiple or massive spills, where booms, and
other spill protection hardware and personnel becomes limited. Therefore, while
classifying the site might seem a simple process, logic and common sense must
be used in attacking the clean-up and protection problem.
Boat Ramps pose a curious case in understanding their
importance to the potential response to a product spill. These facilities,
especially those which are public, can act as major staging points for response
crews. Boats, buoys, and oil barricades, as well as other clean-up gear can be
moved into position for dealing with spill impacts and preparing protection
efforts for dealing with oncoming downstream flow. Bridges can also provide
sites for clean up and protection crews to reach into a stream without actually
having to place crewmembers into potentially heavily polluted waters.
Figures 4,5, and 6 shows the downtown area of Tulsa,
Oklahoma, a city which could be severely impacted in the event of a major spill
from the Conoco Pipeline Company pipeline which crosses the Arkansas River
approximately 2.5 miles west of downtown Tulsa. The major developed area of
downtown Tulsa lies immediately east of the Arkansas River, with a major park
and numerous storm sewer drains along this edge of the river, as can be noted
in Figure 4. Immediately adjacent to this park in downtown Tulsa are numerous
apartment complexes and businesses, which could also be adversely impacted if a
product spill into the Arkansas River occurred. Figure 5 shows a smaller area,
including the USGS Topo Map layer and a more detailed view of downtown Tulsa.
This layer provides planners and clean-up crews with additional information on
the area around impacted points, and also provides additional street
information to aid in moving clean up and protection crews into place. Figure 6
shows the detail information for a selected point in downtown Tulsa, in this
case the 21st Bridge which crosses the Arkansas River immediately south of
downtown. Using a pre-programmed function button, the user simply selects the
desired points and both the detail information and an available picture are
displayed simultaneously.
In addition to allowing planning and clean-up staff
valuable information in responding to a product spill, after a spill occurs
analysts can also begin the process of assessing damages by areas which were not possible to either clean or protect. Legal
and engineering staff can use the information in this database to determine all
possible impacted points in the damage zone, and initiate the process of
after-spill clean-up and financial consideration for property owners in the
impacted areas. Also, the database can be used to determine the validity of
claims against the company, based on the location of sites in relation to the
known zone of impact along the river or stream suffering the brunt of a product
spill.
�
CONCLUSION� Top
The project described briefly in this paper was
completed in order to allow planning and engineering personnel of the Conoco Pipeline
Company to move as quickly as possible in responding to a potential or actual
oil product spill from a pipeline into a major river or lake. Users of the
database will be able to identify potentially impacted sites along such streams
and lakes, allowing them to make the best possible plans for both protection
efforts during the spills impact period, and in cleaning those sites that were
unprotected for numerous reasons.
Since late 1997 the database has been used in
planning efforts by the Conoco Pipeline
Company and other environmental consulting and engineering companies in
practicing for responding to spills in two major Oklahoma rivers (the North
Canadian in Oklahoma City, and the Arkansas in Tulsa). Additionally, the data
gathered in this process has been used in the production of maps and datasets
distributed to Conoco and consulting firm staff for use in the event of an
actual spill. Fortunately, this oil pipeline operator has not had an actual
spill in this region since the creation of the dataset, but the information has
been used by another contractor in analyzing the area and sites impacted by a
major spill from another pipeline in mid-1998 near Tulsa.
ACKNOWLEDGEMENTS� Top
The authors wish to thank Ms. Shellie Rudd, Mr. Jake
Besterman, Mr. Tim Hayes, and Mr. Mark Carper for their efforts in collecting
and assembling the datasets described in this paper. Special thanks are due to
Ms. Rudd who worked on the development of the database for the full two-year
duration of the project. We also wish to thank Dr. Thomas Wikle of the
Department of Geography at Oklahoma State University and Mr. John Barrett of
Conoco, Incorporated for their support, and the Conoco Pipeline Company for
their financial support during this project.
�
AUTHOR
INFORMATION� Top
Allen Finchum, PhD
Department of Geography
Oklahoma State University
Stillwater, Oklahoma�
74078
E-Mail: finchum@okstate.edu
PhD, The University of Tennessee, 1992
Assistant Professor, Oklahoma State University,
1996-Present
Research Associate, The University of Tennessee,
1995-1996
Systems Specialist, The University of Tennessee,
1984-1993
Joseph Seig
Center for the Application of Remote Sensing
Department of Geography
Oklahoma State University
Stillwater, Oklahoma�
74078
E-Mail: jseig@hotmail.com
BS, University of Oklahoma, 1989
Coordinator, Oklahoma State University, 1991-1999
Phone: 405/744-6250
FAX:�� 405/744-5620
