navigation image mapnext pagetable of contentsprevious page

Landsat and other remote sensor systems (SPOT; JERS; radar, etc.) have been heavily used in searching for surface indicators of "leaking" subsurface oil and gas. The general approach to petroleum exploration is described. One line of investigation looks at structural analysis of space imagery in search of subsurface traps. Another, of infrequent success, seeks alteration at the surface caused by chemical changes related to surface-reaching oil or gas. In the early days of Landsat-1, a study in the Anadarko Basin of Oklahoma sought to demonstrate how alteration anomalies and lineaments analysis can aid in finding new petroleum by showing a relationship to already known fields. This pioneering program led to ambiguous but interesting results. Other examples are also considered on this page.


Finding Oil and Gas from Space


If precious metals are not your forte, then try the petroleum industry. Exploration for oil and gas has always depended on surface maps of rock types and structures that point directly to, or at least hint at, subsurface conditions favorable to accumulating oil and gas.

Oil and gas result from the decay of organisms - mostly marine plants (especially microscopic algae and similar free-floating vegetation) and small animals such as fish - that are buried in muds that convert to shale. Heating through burial and pressure from the overlying later sediments helps in the process. (Coal forms from decay of buried plants that occur mainly in swamps and lagoons which are eventually buried by younger sediments.). The decaying liquids and gases from petroleum source beds, dominantly shales after muds convert to hard rock, migrate from their sources to become trapped in a variety of structural or stratigraphic conditions shown in this illustration:

Types of Oil and gas traps
From Physical Geology: Earth Revealed by McGeary and Plummer, First Ed., W.C. Brown Publ.

The anticlinal trap, among the most common, is nicely revealed in a real world setting in this old photograph:

Oil well tapping an anticline, exposed here in a manmade cut; the oil actually occurs in a trapped zone well below the present surface.

The oil and gas must migrate from deeper source beds into suitable reservoir rocks. These are usually porous sandstones, but limestones with solution cavities and even fractured igneous or metamorphic rocks can contain openings into which the petroleum products accumulate. An essential condition: the reservoir rocks must be surrounded (at least above) by impermeable (refers to minimal ability to allow flow through any openings - pores or fractures) rock, most commonly shales. The oil and gas, generally confined under some pressure, will escape to the surface - either naturally when the trap is intersected by downward moving erosional surfaces or by being penetrated by a drill. If pressure is high the oil and/or gas moves of its own accord to the surface but if pressure is initially low or drops over time, pumping is required

Exploration for new petroleum sources begins with a search for surface manifestations of suitable traps (but many times these are hidden by burial and other factors govern the decision to explore. Mapping of surface conditions begins with reconnaissance, and if that indicates the presence of hydrocarbons, then detailed mapping begins. Originally, both of these maps required field work. Often, the mapping job became easier by using aerial photos.

After the mapping, much of the more intensive exploration depends on geophysical methods (principally, seismic) that can give 3-D constructions of subsurface structural and stratigraphic traps for the hydrocarbons. Then, they are sampled by drilling and their properties measured.

Landsat, and other space imaging systems, serve as mega-photos that depict large areas, within which clues to subsurface conditions may be evident. In general, most of the obvious structures that have surface expression had been discovered and mapped (to varying extents) over much of the world. Some regions, however, were not adequately mapped even in the 1970s, so that the advent of higher-resolution space imagery proved a boon to energy companies seeking new sources of fossil fuels. Sometimes the imagery proved especially sensitive to subtle indications of interior structures. For instance, fractures around structures in known oil/gas fields may extend further, as seen in the coherent space images, than suspected from ground work. Also, drainage patterns at broader scales may reflect control by underlying rocks involved in suitable traps. And even vegetation distribution may disclose signs of structure. These and other indicators discernible in space imagery appealed to exploration geologists as another means to survey large areas.

We illustrate these ideas by examining and evaluating one of the first case studies using Landsat-1 to demonstrate the feasibility of direct exploration from space. This pilot study, conducted jointly by the Eason Oil Corp. and the Earth Satellite Corp. of Rockville, MD, shed considerable light on effective criteria for recognizing conditions that might relate to buried hydrocarbons. In addition, some of the pitfalls associated with the space approach were also discovered by carefully assessing the results reported by these investigators.

The strategy behind the study was to look at Landsat imagery of a region already established as a petroleum province, giving special attention to telltale surface indications of the presence of known fields. They used standard-processed and computer-enhanced versions. Rather than test capabilities in a region where there is obvious structural control and other clear-cut evidence, they selected producing areas, in which the surface does not give clear indication of subsurface conditions. If they could succeed in detecting hydrocarbons under such difficult circumstances, then Landsat would increase in stature as an oil/gas discriminator .

The Anadarko Basin of south-central Oklahoma fits this requirement well. Located in the eastern Great Plains, with most of the land used for farming and ranching, the Basin is one of the great producers of the mid-continent petroleum province, which also includes much of Texas, as well.

Map of the Anadarko Basin Petroleum Province (Shaded Pink)

The Basin is a down-sag in the crust that has allowed up to 15,200 m (50,000 ft) of Paleozoic sedimentary rock to accumulate. Structurally, the Basin is an asymmetrical geosyncline (a regional-scale downfold), with the deepest part near the south edge. Oil and gas are present in porous rocks associated with structural (anticlines; fault blocks) and stratigraphic traps. Large gas fields occur mainly along the Basin's western half, whereas oil is more common in the eastern half. Wells as deep as 7,600 m (25,000 ft) have recovered both hydrocarbons, although most pay zones are between 2,750-5,250 m (9,000-15,000 ft).

Generally, surface expression of underlying oil or gas traps in the Basin is meager, because first, there are few structural indicators in the flat-lying sediments atop older folded units and second,there is overprinting of geologic features by vegetation and land use (grasslands; hilly sage-covered terrain; and wheat farmlands). The Eason Oil/Earthsat investigators decided to focus on two search elements: previously undiscovered fractures and subtle chemical alterations of surface rocks by escaping hydrocarbons.

Lineaments analysis was conducted by Eason Oil using Landsat image transparencies backlighted on a light table. The linear features they picked are shown by lightweight black lines on the map below. Superposed as brown and green-black heavier lines are faults that had previously been discovered and mapped. As a geographic reference, note the meander bends (curved segments) of the Canadian River, traced in blue. The majority of the Landsat-mapped linear features are inconspicuous in the imagery. Many of them are suspect, i.e., they could be non-geological or some type of lighting artifacts.

Map of Eason Oil Lineations

A group of four geologists, including this writer, in Code 923 at Goddard Space Flight Center decided to check on the reproducibility of these map results, using the same April, 1973 Landsat MSS full scene (see below). Each person used the same transparencies (mostly winter images) as Eason Oil and worked independently of one another to minimize bias. When done, we registered the tracings to a base map, on which the Eason Oil lineaments were also plotted, as seen below. The comparison disclosed rather startling discrepancies in terms of variance between the two groups. We found only about 20% of the total linear features in common. Eason Oil chose approximately 35% of the questionable features, exclusively, while Goddard geologists chose the remaining 45%, which represented those "missed" by Eason Oil. We immediately suspected that this kind of result is partially due to considerable subjectivity in deciding whether a given linear feature a) really exists, b) is geological in nature, and c) means anything.

Comparison of GSFC and Eason Oil Linears Selections diagram

This suspicion was reinforced by comparing the linear features selected by the four Goddard geologists. Here are the results - a mishmash that requires the following interpretation:

Map of the number of times (color-coded) the same lineament was identified by the 4 Goddard geologists involved in this comparison study.

Of the 785 linear features identified by all four combined, only 4 (0.5%) were noted by every operator. From the remainder, 3 operators mutually selected 37 (4.7%), two operators agreed on 140 (17.8%), and the rest, 604 (77%), each operator found exclusively. This type of result has been reported in similar studies, although the above scores were particularly discouraging. Each geologist had ample experience in photointerpretation and special skills in analyzing Landsat imagery. Their choices were justifiable but overall, our results were questionable.

5-10: In this experiment, and in the technique of picking linear features in space imagery, what do you think was really going on behind the end result of some many linears being found but not consistently by multiple interpreters? ANSWER

The bottom line here is that there often is a strong tendency towards overkill in choosing features that appear to be meaningful lineaments. So many are drawn that it would take a monumental field effort to check them out. If plotted as rose diagrams (see page 2-9), they may reveal valid trends for the orientations of regional fractures, because statistically lineaments of non-geological nature should be in the minority. A study of obvious lineaments in the Adirondacks confirmed this result. Of the 200+ prominent ones that were field-checked, geological fractures directly or indirectly controlled most of them, but about 20% related to human factors, such as fence lines, roads, etc. Thus, we conclude that we should combine lineaments analysis with other indicators of mineralization or hydrocarbons. This combining would encourage geologists to field-check particular sites to verify the lineament presence and nature and their possible correlation with these indicators.

The Eason Oil study sought to recognize such indicators. Their interpreters delineated certain geomorphic anomalies, such as circular patterns and unusual drainage. In the course of their image appraisals, they noticed unexpected tonal patterns that looked a bit like light-colored smudges on the images, such as the April, 1973 full Landsat MSS scene that became the reference base for the study. These they called "hazy" features, as seen here:

Hazy Features (whitish tones at A, B, C and elsewhere) on the Eason Oil Landsat study image.

We labeled three typical hazy patterns A, B, and C. The one at A, at a bend in the Canadian River, is especially prominent, and occurs over a known oil field.

A standard false color subscene (computer-enhanced) around A shows the hazy to have a bluish-white color similar to soils in barren fields. Note the road pattern and white blotches which are accesses to producing wellheads. The yellowish areas coincide with unaltered Permian (late Paleozoic) red beds.

Color composite in quasi-true color of Hazy Feature at Location A.


When we process this April Multispectral Scanner image into three ratio bands that we then combine into a color image (4/5 = Blue; 5/6 = Green; 6/7 = Red), the hazy feature at A takes on a unique yellow-green, and the red beds become orangish.

Ratio Composite of B = 4/5; G = 5/6; R = 6/7 (MSS bands) for the Hazy Feature at A.


From the multiseasonal data sets, only those scenes imaged in late winter to early spring show hazies. At other times of the year, vegetation masks the phenomenon. To understand their explanation of the features, we look now at this photograph of two rock types:

 Rocks collected at outcrops within the Anadarko Basin. From left to right: Fresh Permian Red bed sandstone; altered sandstone; altered gypsum bed; fresh gypsum.

The rock on the far left is a sample from the red beds (sandstones) of Permian age. Next to it is the same material that has been color bleached to yellow-brown by converting iron oxide cement into hydrated iron oxides (analogous to rust). The gray rock on the far right is a limestone (calcium carbonate). To its left is a gypsum rock (hydrated calcium sulphate). Both interior rocks appear to be altered equivalents of the primary exterior rocks. In the field, comparable altered rocks can occupy many square miles.

To account for these hazy features, the Eason Oil people postulated that chemical reactions affected the iron cement or transformed the carbonates into sulphates, when sulphur-laden gases or fluids leaked out of petroleum traps and rose towards the surface, interacted with susceptible rocks, and brought about compositional changes. About the time of their conclusion, evidence for such changes was reported as the Doctoral thesis of Terrence Donovan (later of the U.S.G.S.), in which escaping hydrocarbons drastically altered rocks above the Cement Field, at the southeast edge of the Anadarko Basin. Dr. Donovan found a pronounced set of anomalous values of the ratio of C13 to C12 in samples collected over both producing zones in the field, shown as contoured areas below:

Map showing the spatial variations of C-13/C-12 ratio in the Cement Field in the Anadarko Basin.

These values represent some of the highest departures from normal ratios known anywhere in the world. He attributed them to the effects of chemical action by carbon-rich fluids on the rocks which, as a consequence, appear bleached. Oddly, while evident on the ground, this alteration is not detectable in the Landsat imagery, possibly because it is not strongly expressed in derivative soils.

Accepting this alteration hypothesis, the Eason Oil group looked for at least partial coincidence between these hazies and the surface projections of subsurface oil or gas fields. Of the 57 anomalies they mapped in a control segment of the imagery, they claimed an association with 42 producing fields. Another six occurred above or near non-producing structures, and only 9 showed no coincidence. If this observation remained true, then detecting hazies, sometimes correlative with lineament concentrations, could promise a powerful new way to hunt for oil and gas using space imagery.

The present writer (NMS), being skeptical in habit, decided to challenge these findings. He traced the outlines of the Eason Oil hazy features (in a hachured pattern) on a transparency and then overlaid and registered it to the oil (pinks) and gas (blues) map of Oklahoma. The resulting combination is shown here:

Map of Oil and Gas fields in the Anadarko Basin overlaid by Eason Oil Hazy Features (hachured pattern)

Visually, the coincidence between hazies and fields does not appear strong. This was supported by a spatial correlation analysis, which demonstrated there is no statistical significance to the pattern distribution, i.e., the coincidence is random rather than associative. In practical terms, there would be at least as much chance of striking oil by drilling into points selected by throwing darts at the map, as there would be in drilling into the centers of hazies. Based on a quick field trip to the A hazy, the writer (NMS) believes hazy features are areas where wind has blown away much of the soil fines, leaving reflective quartz grains behind.

However, at one locality designated as a hazy feature, the writer did find convincing evidence of what appears to be distinct color difference attributable to hydrocarbon alteration of red beds. In a dirt road, the reddish-orange of unaltered Permian rocks gives way to a yellow-white color representing hydrocarbon "bleaching" as proposed by EarthSat/Eason Oil. Here is a photo taken at that point:

Conversion of red beds to yellowish altered iron oxide in an exposure made along a dirt road in the Anadarko Basin.

The Goddard geologists didn't perform these studies to discredit the Eason Oil study, which provided some valuable insights into the discerning power of space imagery for petroleum exploration and the potential shortcomings of the apparent results. We did them to independently evaluate this approach and to inject caution into any beliefs that this technique might become a panacea for finding petroleum.

5-11: Critique the Eason Oil study, devising if appropriate a defense of their approach. In general, what do you believe to be the most effective use of remote sensing in exploring for hydrocarbons. ANSWER

At the time of the Anadarko study, several other investigators claimed to have found similar evidence that appeared to indicate that leaking oil and gas reservoirs could indeed be altering surface rock and soil. One that seemed to confirm this was the Beaver Creek Oil Field in the Wind River Basin of Wyoming. Dr. Robert Vincent presented this evidence, an MSS Band 5 (red) to 4 (green) ratio image in which a prominent oval shaped anomaly (shown here in tan) coincided very closely with the outline of the field as determined by subsurface drilling:

The Beaver Creek anomaly, rendered in a tan color in an MSS 5/4 ratio image.

The writer (NMS) visited this field while engaged in his Wyoming investigation work. The area consisted of Lower Tertiary sedimentary rocks that were strongly dissected into gullies. Many of these beds were reddish and could in themselves account for some of the anomaly. A rather quick search for obvious signs of alteration by escaping gases or fluids failed to find any convincing evidence. But the remarkable co-incidence of the 5/4 anomaly with the outline of the Beaver Creek field suggest that this may be a valid example of the concept of alteration by petroleum compounds.

Landsat results in geological applications excited many in the petroleum and mining industries. Various companies banded together as a consortium, starting in 1976, in what became known as The Geosat Committee. Their avowed aims were along three lines: 1) to share information and conduct studies using space imagery to search for petroleum and minerals (mainly metallic ores), and 2) to "lobby" NASA and Congress for a continuation and expansion of the Earth-Observing Satellite program, and 3) to provide inputs in determining and improving sensors in future satellites. One of their principal study sites was the Patrick Draw oil field near the Beaver Creek field in Wyoming. In the writer's opinion, the results from this detailed study did not prove that there was any distinct anomaly(ies) at Patrick Draw which were detectable by the space imagery they used.

Earth Satellite Corp., and another group, Earth Search Sciences have continued to validate data obtained from sensors on satellites and aircraft as potentially decisive indicators of subsurface oil/gas fields. This next diagram summarizes current thinking:

Schematic diagram showing the types of materials that escape from oil and gas reservoirs and the types of alteration detectable at the surface.

Airborne Hyperspectral sensors that were flown over known hydrocarbon leaks have found that an absorption feature near 2.31 µm is very sensitive to the amount of a specific component of the hydrocarbons. A ratio of two reflectance values on either side of that absorption feature divided by the value of the decreased reflectance in the spectral curve at the feature low point enhances the detectability of the hydrocarbon and quantifies its magnitude.

Hyperspectral curves for bitumen compounds in a tar sand.

This next image display shows an actual field case conducted jointly by the HJW GeoSpatial, Inc, the Geosat Committee and Earth Search Sciences in which an oil seep that corresponds to a specific pixel (red) in the Probe-1 hyperspectral scanner image shows the 2.31 µm diagnostic anomaly (strong absorption bands at 1.4 and 2.0 µm are related to other materials):

Spectral curve corresponding to a single pixel associated with a petroleum seep in an oil field.

Leaks from fields below the ocean can serve both as an exploration indicator and as a source of environmental damage. Prospecting for oil beneath the open ocean requires some different techniques as well as use of some of the conventional land methods. Oil seeps and slicks can remain intact on the surface and may be detectable in Vis/NIR and radar imagery. The Earth Satellite Corporation has developed SEP - the Seep Enhancement Algorithm - to bring out an oil signature using radar imagery. Here is an example:

Black oil seeps in a radar image of the ocean surface.

Oil slicks can be both natural or due to manmade oil spills. This Earthsat image shows a slick off the coast from Kuwait as rendered in a natural color Landsat image.

Oil slick near Kuwait City.

Specialized remote sensing can monitor another aspect of petroleum withdrawal not necessarily expressed as leaks. In time, as the oil is removed from pores leaving a partially filled void, the rock units bearing the oil start to contract or crush inward into the voids as support diminishes. This is commonly expressed by all the overlying units pushing downward on the now compressed reservoir rocks, giving rise to progressive surface subsidence. This lowering of elevation can be monitored by radar interferometry (see page 11-10 This next illustration, made from ESA radar data, shows interferometry rings, which can be computed into elevations, at the Lost Hills oil field in the San Joaquin Valley of California. The field is subsiding now at a rate of about 3 cm (1.2 inch) per month, with a cumulative drop since 1989 of 3 meters (10 ft). Subsidence is greater at the two ends of this 1.5 x 6 km (~1 x 4 miles) elongate field.

Radar interferometric patterns from subsidence at the Lost Hills oil field in southern California.

Suffice to close with the remark that since the launch of ERTS-1, the petroleum industry has found new oil and gas fields with the aid of space data and has developed criteria from the images that continue to prove worthwhile in planning and conducting exploration programs, which are leading to payoffs. Most of the successes have come by using space imagery (as has been done before with aerial photography) in the tried-and-true (conventional) way of using the pictures as base maps on which to analyze and plot structural patterns and trends, often supplemented by recognition of stratigraphic units. (Detection of surface alteration, while it happens sometimes, remains a rather rare event. As evident from the preceding paragraphs, one company that has provided a comprehensive exploration package pegged to image utilization is the Earth Satellite Corp., as described under their Petroleum link in their Web Site. Here is one of their interpreted Landsat images, with structural features (anticlines and synclines; faults and lineations) superimposed in black. The Landsat scene covers part of Myamar (Burma) where the Mizo mountains form from a series of fingerlike anticlines that trend almost at right angles to the Himalayas (top).

Anticlinal structures in the Mizo Hills of Burma; black lines record structural features interpreted by Earth Satellite Corp. geologists in their contracted investigation looking for potential areas of petroleum occurrence.

navigation image mapnext pageprevious page


Primary Author: Nicholas M. Short, Sr. email: [email protected]