Skip to Main Content
Book Chapter

Chapter 12: Relationship of Subsurface Reservoir Properties and Hydrocarbon Sea-surface Slicks in the Northern Gulf of Mexico

By
Margaret M. Dalthorp
Margaret M. Dalthorp
Formerly Texas A&M University; presently Murex Petroleum Corp., Houston, Texas, U.S.A. E-mail: mdalthorp@sbcglobal.net.
Search for other works by this author on:
Thomas H. Naehr
Thomas H. Naehr
Texas A&M University, Harte Research Institute for Gulf of Mexico Studies, Corpus Christi, Texas, U.S.A.
Search for other works by this author on:
Philippe Tissot
Philippe Tissot
Texas A&M University, Conrad Blucher Institute, Corpus Christi, Texas, U.S.A.
Search for other works by this author on:
Oscar Garcia-Pineda
Oscar Garcia-Pineda
Florida State University, MacDonald Lab, Tallahassee, Florida, U.S.A.
Search for other works by this author on:
Published:
January 01, 2013

Abstract

In the northern Gulf of Mexico (GOM), numerous seafloor hydrocarbon seeps are well documented through seismic, seafloor, and chemical studies. In some areas, the seafloor seeps expend sufficient hydrocarbons to pass through the water column and present as sea-surface oil slicks, which with careful review and certain conditions can be identified on satellite images. Some areas with recurring potential sea-surface oil slicks as identified by oil-criteria screened satellite images are located above or near proven subsurface hydrocarbon-producing areas. However, other areas with established production have no sea-surface hydrocarbon slicks, and several areas with potential sea-surface slicks currently have no subsurface production. A publicly available data set of reservoir information for the northern GOM was correlated to a data set of sea-surface slick sites as identified on satellite data. Thirteen variables were selected from the database to evaluate depth relationships, fluid properties, and reservoir properties in relation to slick presence. Because initial multivariate analysis indicated the slick-to-reservoir-variables relationships were nonlinear, a random forest (RF) classification-tree analysis was performed to identify which variables are more important for slick development. The RF analysis used a collection of binary decision trees to determine the relative importance of the reservoir variables in identifying sea-surface slicks. The RF model was able to correctly classify or predict the presence of a slick and identified water depth, gas-oil ratio, and reservoir chronozone as the three most important variables when predicting the formation of a sea-surface slick.

You do not currently have access to this article.

Figures & Tables

Contents

Geophysical Developments Series

Hydrocarbon Seepage: From Source to Surface

Fred Aminzadeh
Fred Aminzadeh
Search for other works by this author on:
Timothy B. Berge
Timothy B. Berge
Search for other works by this author on:
David L. Connolly
David L. Connolly
Search for other works by this author on:
Society of Exploration Geophysicists
Volume
16
ISBN electronic:
9781560803119
Publication date:
January 01, 2013

GeoRef

References

Related

Citing Books via

Close Modal
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close Modal
Close Modal