Evaluation of sampling methods for fracture network characterization using outcrops

Conny Zeeb*, Enrique Gomez-Rivas, Paul D. Bons, Philipp Blum

*Corresponding author for this work

Research output: Contribution to journalArticle

28 Citations (Scopus)

Abstract

Outcrops provide valuable information for the characterization of fracture networks. Sampling methods such as scanline sampling, window sampling, and circular scanline and window methods are available to measure fracture network characteristics in outcrops and from well cores. These methods vary in their application, the parameters they provide and, therefore, have advantages and limitations. We provide a critical review on the application of these sampling methods and apply them to evaluate two typical natural examples: (1) a large-scale satellite image from the Oman Mountains, Oman (120,000 m(2) [1,291,669 ft(2)]), and (2) a small-scale outcrop at Craghouse Park, United Kingdom (19 m(2) [205 ft(2)]). The differences in the results emphasize the importance to (1) systematically investigate the required minimum number of measurements for each sampling method and (2) quantify the influence of censored fractures on the estimation of fracture network parameters. Hence, a program was developed to analyze 1300 sampling areas from 9 artificial fracture networks with power-law length distributions. For the given settings, the lowest minimum number of measurements to adequately capture the statistical properties of fracture networks was found to be approximately 110 for the window sampling method, followed by the scanline sampling method with approximately 225. These numbers may serve as a guideline for the analyses of fracture populations with similar distributions. Furthermore, the window sampling-method proved to be the method that is least sensitive to censoring bias. Reevaluating our natural examples with the window sampling method showed that the existing percentage of censored fractures significantly influences the accuracy of inferred fracture network parameters.

Original languageEnglish
Pages (from-to)1545-1566
Number of pages22
JournalAAPG Bulletin
Volume97
Issue number9
DOIs
Publication statusPublished - Sep 2013

Keywords

  • opening-mode fractures
  • mean trace length
  • Oman Mountains
  • rock discontinuities
  • mechanical system
  • geological media
  • Jabal Shams
  • flow
  • geometry
  • density

Cite this

Evaluation of sampling methods for fracture network characterization using outcrops. / Zeeb, Conny; Gomez-Rivas, Enrique; Bons, Paul D.; Blum, Philipp.

In: AAPG Bulletin, Vol. 97, No. 9, 09.2013, p. 1545-1566.

Research output: Contribution to journalArticle

Zeeb, Conny ; Gomez-Rivas, Enrique ; Bons, Paul D. ; Blum, Philipp. / Evaluation of sampling methods for fracture network characterization using outcrops. In: AAPG Bulletin. 2013 ; Vol. 97, No. 9. pp. 1545-1566.
@article{90523430a82f48188d4331a9a309665f,
title = "Evaluation of sampling methods for fracture network characterization using outcrops",
abstract = "Outcrops provide valuable information for the characterization of fracture networks. Sampling methods such as scanline sampling, window sampling, and circular scanline and window methods are available to measure fracture network characteristics in outcrops and from well cores. These methods vary in their application, the parameters they provide and, therefore, have advantages and limitations. We provide a critical review on the application of these sampling methods and apply them to evaluate two typical natural examples: (1) a large-scale satellite image from the Oman Mountains, Oman (120,000 m(2) [1,291,669 ft(2)]), and (2) a small-scale outcrop at Craghouse Park, United Kingdom (19 m(2) [205 ft(2)]). The differences in the results emphasize the importance to (1) systematically investigate the required minimum number of measurements for each sampling method and (2) quantify the influence of censored fractures on the estimation of fracture network parameters. Hence, a program was developed to analyze 1300 sampling areas from 9 artificial fracture networks with power-law length distributions. For the given settings, the lowest minimum number of measurements to adequately capture the statistical properties of fracture networks was found to be approximately 110 for the window sampling method, followed by the scanline sampling method with approximately 225. These numbers may serve as a guideline for the analyses of fracture populations with similar distributions. Furthermore, the window sampling-method proved to be the method that is least sensitive to censoring bias. Reevaluating our natural examples with the window sampling method showed that the existing percentage of censored fractures significantly influences the accuracy of inferred fracture network parameters.",
keywords = "opening-mode fractures, mean trace length, Oman Mountains, rock discontinuities, mechanical system, geological media, Jabal Shams, flow, geometry, density",
author = "Conny Zeeb and Enrique Gomez-Rivas and Bons, {Paul D.} and Philipp Blum",
year = "2013",
month = "9",
doi = "10.1306/02131312042",
language = "English",
volume = "97",
pages = "1545--1566",
journal = "AAPG Bulletin",
issn = "0149-1423",
publisher = "AMER ASSOC PETROLEUM GEOLOGIST",
number = "9",

}

TY - JOUR

T1 - Evaluation of sampling methods for fracture network characterization using outcrops

AU - Zeeb, Conny

AU - Gomez-Rivas, Enrique

AU - Bons, Paul D.

AU - Blum, Philipp

PY - 2013/9

Y1 - 2013/9

N2 - Outcrops provide valuable information for the characterization of fracture networks. Sampling methods such as scanline sampling, window sampling, and circular scanline and window methods are available to measure fracture network characteristics in outcrops and from well cores. These methods vary in their application, the parameters they provide and, therefore, have advantages and limitations. We provide a critical review on the application of these sampling methods and apply them to evaluate two typical natural examples: (1) a large-scale satellite image from the Oman Mountains, Oman (120,000 m(2) [1,291,669 ft(2)]), and (2) a small-scale outcrop at Craghouse Park, United Kingdom (19 m(2) [205 ft(2)]). The differences in the results emphasize the importance to (1) systematically investigate the required minimum number of measurements for each sampling method and (2) quantify the influence of censored fractures on the estimation of fracture network parameters. Hence, a program was developed to analyze 1300 sampling areas from 9 artificial fracture networks with power-law length distributions. For the given settings, the lowest minimum number of measurements to adequately capture the statistical properties of fracture networks was found to be approximately 110 for the window sampling method, followed by the scanline sampling method with approximately 225. These numbers may serve as a guideline for the analyses of fracture populations with similar distributions. Furthermore, the window sampling-method proved to be the method that is least sensitive to censoring bias. Reevaluating our natural examples with the window sampling method showed that the existing percentage of censored fractures significantly influences the accuracy of inferred fracture network parameters.

AB - Outcrops provide valuable information for the characterization of fracture networks. Sampling methods such as scanline sampling, window sampling, and circular scanline and window methods are available to measure fracture network characteristics in outcrops and from well cores. These methods vary in their application, the parameters they provide and, therefore, have advantages and limitations. We provide a critical review on the application of these sampling methods and apply them to evaluate two typical natural examples: (1) a large-scale satellite image from the Oman Mountains, Oman (120,000 m(2) [1,291,669 ft(2)]), and (2) a small-scale outcrop at Craghouse Park, United Kingdom (19 m(2) [205 ft(2)]). The differences in the results emphasize the importance to (1) systematically investigate the required minimum number of measurements for each sampling method and (2) quantify the influence of censored fractures on the estimation of fracture network parameters. Hence, a program was developed to analyze 1300 sampling areas from 9 artificial fracture networks with power-law length distributions. For the given settings, the lowest minimum number of measurements to adequately capture the statistical properties of fracture networks was found to be approximately 110 for the window sampling method, followed by the scanline sampling method with approximately 225. These numbers may serve as a guideline for the analyses of fracture populations with similar distributions. Furthermore, the window sampling-method proved to be the method that is least sensitive to censoring bias. Reevaluating our natural examples with the window sampling method showed that the existing percentage of censored fractures significantly influences the accuracy of inferred fracture network parameters.

KW - opening-mode fractures

KW - mean trace length

KW - Oman Mountains

KW - rock discontinuities

KW - mechanical system

KW - geological media

KW - Jabal Shams

KW - flow

KW - geometry

KW - density

U2 - 10.1306/02131312042

DO - 10.1306/02131312042

M3 - Article

VL - 97

SP - 1545

EP - 1566

JO - AAPG Bulletin

JF - AAPG Bulletin

SN - 0149-1423

IS - 9

ER -