Abstract

Linear features are relatively short, non-cultural linear elements detectable on satellite images that mostly reflect topographic features, such as cliffs, slope breaks, and stream valley segments, whose shape and orientation are controlled, in large part, by joints and faults. A linear features map is a statistical sample of existing fractures that can be used to infer regional fracture patterns, including fracture densities and the orientation of major fracture sets. Areas of increased fracturing are related to accumulations of coalbed methane in fracture-enhanced reservoirs in the San Juan Basin region. Linear features analysis may provide a means to focus exploration on the Southern Ute Reservation.

The Landsat Thematic Mapper (TM) satellites acquire digital image data in 6 spectral bands in the visible and near-infrared parts of the spectrum and 1 band in the thermal infrared. Visible and near-infrared digital TM image data of the Southern Ute Reservation and vicinity were computer processed to enhance tonal boundaries (edges) and tonal contrast (contrast stretch). 917 linear features were mapped from 1:280,000-scale film transparencies of the enhanced TM images, digitized, and analyzed for length and orientation characteristics and regional concentration patterns.

The number of linear features increases exponentially with decreasing length until about 500 meters (1,640 ft) where the 3-meter (10 ft) ground resolution of the TM system begins to effect the detectability of linear features. Linear features shorter than 500 meters (1,640 ft) must be exceptionally well topographically expressed or preferentially oriented perpendicular to the solar illumination direction to be detected on the images. Consequently, linear features shorter than-500 meters (1,640 ft) do not reflect the actual distribution of linear features (faults and joints) and cannot be used in the statistical analysis of orientation characteristics.

The orientation characteristics of the linear features data were statistically analyzed for all linear features and only those linear features longer than 500 meters (1,640 ft). In addition, analyses were conducted for the linear feature data weighted according to their individual lengths to test for the existence of a few long, parallel linear features that may be of regional importance and could go undetected in the unweighted analyses. Stake-frequency histograms of the linear feature data consistently show a general sinusoidal shape with a trough in the direction parallel to the solar illumination direction and a peak in the direction perpendicular to the solar illumination direction. These results show a significance bias in the orientation of the linear features as a function of solar illumination angle. Consequently, these linear feature data are considered unsuitable for determining regional fracture orientation characteristics.

In spite of the apparent bias in the orientation of the linear feature data, the relative concentrations of linear features may be useful for detecting areas of greater or lesser fracturing. A contour map of linear feature concentrations was prepared and the concentration patterns were compared to mapped formations and geologic structures. The most conspicuous concentrations of linear features are spatially associated with the distribution of Cretaceous sandstones exposed along and north of the Hogback Monocline. Local linear feature concentrations are also associated with folds in the vicinity of the Hogback Monocline, as well as the Ignacio Anticline and Barker Dome and anticline. Elsewhere south of the Hogback Monocline, local linear feature concentrations are not clearly related with specific formations or geologic structures. These linear feature concentrations may mark areas of enhanced fracture permeability and are worthy of additional investigation, perhaps in conjunction with detailed field fracture and joint studies.