Radar imaging matched with gravity and bathymetry

Tracking seepage points by satellite microwave

Mary Jo Wagner
Contributing Editor

Artist's rendition of the Radarsat satellite and image-capture path (Illustration by Paul Fjeld) [40,258 bytes].
The eastern Adriatic Sea, illustrated by the Offshore Basin Screening method (bottom to top), offshore satellite gravity (basin shape), Radarsat image (slicks) and Landsat Thematic Mapper image, with a digital elevation model (onshore geology) [26,371 bytes].
Gulf of Mexico Green Canyon area oil slick acquired by RADARSAT on November 10, 1997. Multi-directional, crisscross trails are characteristic of this area [18,280 bytes].
Enlarged portion of image showing details of slick shapes. Similar shapes indicate local influence of identical changing wind and current conditions over a period of several hours preceding image acquisition [28,896 bytes].

Petroleum exploration is all about confidence. Confidence that particular drilling blocks have been chosen wisely. And confidence that they have been chosen wisely because all investigative evidence points to structures with hydrocarbon charges.

However, certitude in offshore exploration, particularly in deepwater, is difficult to obtain. Once companies start moving into water depths of 6,000 ft, their conviction in solid bid strategies and effective drilling operations starts to wane like the pull of an anchor. Radar satellite data can help reduce that level of risk.

By combining radar imagery from the Canadian Radarsat-1 satellite with other data sources such as satellite-derived gravity and bathymetry, companies have been able to clearly detect and map natural oil slicks on the ocean surface, and subsequently relate it to subsurface structures. It's a method that's proven effective in deepwater Angola, Nigeria, Brazil, and the Gulf of Mexico.

"Any methodology or piece of information that will help minimize our risk in exploration, especially in deepwater, is of great value," explains Fernando Pellon de Miranda, Senior Geologist at Petrobras' Center of Excellence in Geochemistry (Cegeq), an internal lab at the Brazilian petroleum company. "Regular grid-sampling is expensive and it's nearly impossible to do effectively in deepwater frontier basins. Having prior knowledge and clear indications of natural seeps allows us to strategically choose areas for field sampling."

Radar's role

The oil and gas industry has relied primarily on data from weather satellites and from optical remote sensing satellites for exploration until the early 1990's. After that, radar satellite imagery was introduced to the industry and ushered in a new era in exploration operations.

The effectiveness of C-band synthetic aperture radar (SAR) as a viable tool

for locating ocean surface oil slicks has been proven in studies since 1992, when the European Space Agency launched the European Remote Sensing Satellite (ERS-1). That radar capability further improved with Canada's Radarsat-1 satellite, launched in 1995 by the Canadian Space Agency.

In addition to its C-Band SAR instrument, Radarsat-1 also offers petroleum companies a wide range of incidence angles (10!-60!), image widths (50-500 km), and resolutions (8 meters to 100 meters). Most importantly, it carries on-board tape recorders (OBRS), ensuring global coverage. Radarsat International (RSI), located in Richmond, British Columbia, Canada, is the commercial worldwide distributor of the imagery.

Optical satellite imagery has been a predominant source of data for locating, identifying and mapping onshore geological structures since the early 1970s. But optical sensors have not proven to be a very effective tool in identifying natural oil slicks on the sea because of their inability to image through clouds and their difficulty in adequately discerning ocean surface textures.

Unlike optical sensors which use visible and infrared wavelengths to collect energy reflected from the Earth's surface, radar satellites are active sensors and emit energy at microwave frequencies. With microwave energy, radar satellites can penetrate darkness, clouds, rain or haze, enabling them to acquire imagery day or night under any atmospheric condition.

Radar sensors transmit a microwave signal to the Earth's surface and then measure the amount of energy that returns to the satellite. They are very sensitive to surface roughness and can distinguish textural differences on both land and water by measuring the variations in the backscatter signal (signal returned to the satellite).

In the case of slicks, oil suppresses the ocean's capillary waves and creates a surface smoother than the surrounding water. Smooth or flat surfaces tend to reflect microwave energy away from the satellite sensor resulting in a low to nil radar backscatter. The amount of backscatter reduction varies with the slick's physical properties, particularly its visco-elastic properties, but the weak return signal from the smooth slick causes it to appear black or dark on a radar image.

"Although slicks can only be clearly identified on radar imagery within a particular wind-speed window, radar satellite data has become the data of choice for natural oil seep detection applications," reports Nigel Press, President of Nigel Press Associates Limited (NPA). The company, located in Edenbridge, England, is a value-added company specializing in satellite applications for 25 years.

"You've got to have radar imagery to successfully map slicks," he says. "Using Radarsat-1, which orbits at an altitude of 798 km, we've identified slicks as small as 100 meters long in varying weather environments such as the cold and rough South Atlantic, and the warm, heavily polluted Mediterranean Sea.

"In the Falkland Islands, subtle, yet clear examples of natural hydrocarbon seepage have been mapped to within a few 100 meters using the radar imagery. That was a significant confidence builder for petroleum companies to commit exploration expenditure in an area where doubts previously existed about the presence of a working petroleum system."

Basin screening

NPA has been providing oil seep detection studies to oil and gas companies such as Exxon, Shell, Mobil and Statoil since 1994 through its Offshore Basin Screening (TM) technique. The OBS combines seepage detection using satellite radar imagery with high-resolution satellite gravity to investigate basin structure, the presence of a working source, its distribution and migration pathways, and to map the location of seepage points.

Once an area of interest is selected, NPA staff consults RSI to search Radarsat's archive to see if imagery is available. If scenes with the appropriate beam mode are not available, the satellite is programmed to acquire the data. Proprietary archives of geological structure data, meteorological data and satellite gravity data are consulted to best choose the dates of acquisition.

Typically, two to three acquisitions of Radarsat Wide 1 (W1), a beam mode with 30-meter resolution and nominal coverage of 150 km by 150 km, data are requested. NPA then checks the scheduled dates against hindcasts of weather conditions to ensure the wind-speed compliance of each scene.

After obtaining the imagery, image-processing experts apply proprietary techniques to process the raw signal of the radar imagery to enhance ocean/sea contrasts and to identify anomalies. Referring to NPA's global database of oil slick characteristics, oceanographers isolate each anomaly on the imagery and compare it to the slick characteristics database.

Based on the shape, backscatter characteristics, and orientation of the features, coupled with the water temperature, water depth, wind speed and geographical location, they classify and "rate" each feature, distinguishing natural slicks from ship dumpings, pollution, or natural film. The slicks are mapped showing their locations and their confidence ratings.

Simultaneous to the image interpretation process, geologists use bathymetry data from the surrounding area of interest to locally enhance and correct the gravity data. The local gravity model is merged with other ancillary geological data to create a 3D representation of the basin structure.

The classified slick distribution map is then overlaid on the gravity model to produce a detailed map of the basin, indicating its geological structure, the number of natural slicks present, and the location of hydrocarbon charges.

To date, NPA has used OBS worldwide to provide petroleum companies with natural seepage conditions in areas such as the UK Atlantic margin, Vietnam, Northwest Australia, Caspian Sea, Persian Gulf, South Africa, Brazil, Nigeria, Gabon, Zaire and Angola.

Although there is no fundamental difference in applying the OBS technique to near shore or deepwater applications, Press says this method provides competitive intelligence at a competitive price.

"The next best way to collect geological information at depths of over 6,000 ft is through an airborne survey. Typically, such surveys equate to about $20 per line-km flown, and they only provide non-continuous narrow strips of data. Using Radarsat imagery, the OBS costs less than $1/sq km and the satellite provides wide and continuous coverage. And when you do acquire a study, you'll have a very unique set of data and a fairly high level of confidence that the data is accurate."

"In Angola, for example, where we're presently using Radarsat imagery in OBS to cover the entire deepwater, we can see hydrocarbon seepage in water depths of 13,000 ft," he continues. "There isn't any way petroleum companies can obtain wide-spread geological information of that area with traditional methods. Coring is very difficult and drilling is certainly not an option because of its expense. So you've got to study the water surface instead and build an analysis from that. Ultimately, we can lead petroleum companies to areas of ocean where they have a good chance of sampling the seeps by boat and being able to type the hydrocarbons with gas chromatography."

Petrobras' approach

The expense and difficulty in adequately studying a deepwater basin floor through conventional means were the motivations that led Petrobras to develop its own satellite imagery methodology to locate potential natural hydrocarbon seeps, says Miranda.

"Traditionally, oil companies use seismic data to derive geological structure information about a frontier basin," he says. "But seismic data studies depict neither the conditions of the sea-state nor features on the seafloor. And in offshore frontier areas, the sole use of regular gridding for geochemical sampling of ocean water and sea floor is simply not feasible because of high operational costs. Radarsat images are helpful in providing reliable and inexpensive data about the surface of the sea-state which subsequently can lead to the location of natural oil seeps."

Although no newcomer to the application of remote sensing in its onshore exploration operations, Petrobras is now broadening that application to offshore exploration for the first time. The trial area is the deepwater mouth of the Amazon River, specifically offshore Para and Amapa states in northern Brazil.

The image processing and interpretation of radar imagery is one of the key elements to successfully identify the presence or absence of a working source. Miranda and his colleagues developed an operational image interpretation methodology, which allows them to highlight smooth water surfaces and determine potential natural oil slicks based on the surface textures.

To investigate the Para/Amapa region, a combination of Radarsat W1 and ScanSAR Narrow 1 imagery was acquired over a one-year period. This beam mode nominally covers an area of 300 km by 300 km, with a resolution of 50 meters. A multi-temporal approach is necessary to study anomaly distribution patterns to determine that a feature is indeed an oil slick produced from a natural hydrocarbon charge.

Petrobras' staff processed the radar imagery with a semi-variogram textural classification algorithm to enhance the ocean's surface features. The algorithm is a deterministic classifier that provides the option of combining both textural and radiometric information to automatically delineate areas of smooth surface texture from rough and intermediate areas.

The smooth surface anomalies were then highlighted and isolated for subsequent interpretation. The interpreted imagery was compared with ancillary geological and geophysical data to integrate the textural features into the tectonic and stratigraphic context of the sedimentary basin. This detailed information serves as an aid to further geochemical sampling.

And the methodology has proven effective, says Miranda. Hundreds of areas related to smooth texture were detected in offshore Para/Amapa. About 12 of those areas were further determined to be potential sites for natural oil seeps.

"The results of this study are extremely valuable," concludes Miranda. "Not only do they prove that this methodology works, but they give us the means to more efficiently and effectively plan our sampling strategies. By performing an image interpretation beforehand, we can optimize our data collecting activities because the imagery indicates where we should focus our exploration efforts. We no longer have to blindly enter offshore frontier basins."

Radar satellite data can help petroleum companies "see" the most important element in the exploration process: the geological structure of a basin. Coupled with gravity, bathymetry and seismic data, radar imagery offers valuable support to exploration decision-makers to plan more cost-effective operation strategies.

Copyright:RADARSAT data ? Canadian Space Agency/ Agence Spatiale Canadienne 1996. Received by the Canadian Centre for Remote Sensing. Processed and distributed by Radarsat International. Image enhancement and interpretation by NPA.

Author

Mary Jo Wagner is a London-based writer specializing in remote sensing, GIS, and GPS (Tel: 44-181-675-7229; Fax: 673-7227; email: [email protected]).