Remote sensing assessment of oil spills near a damaged platform in the Gulf of Mexico
An oil platform in the Mississippi Canyon 20 (MC-20) site was damaged by Hurricane Ivan in September 2004. In this study, we use medium- to high-resolution (10–30 m) optical remote sensing imagery to systematically assess oil spills near this site for the period between 2004 and 2016. Image analysis detects no surface oil in 2004, but ~40% of the cloud-free images in 2005 show oil slicks, and this number increases to ~70% in 2006–2011, and >80% since 2012. For all cloud-free images from 2005 through 2016 (including those without oil slicks), delineated oil slicks show an average oil coverage of 14.9 km2/image, with an estimated oil discharge rate of 48 to ~1700 barrels/day, and a cumulative oil-contaminated area of 1900 km2 around the MC-20 site. Additional analysis suggests that the detected oil slick distribution can be largely explained by surface currents, winds, and density fronts.
Detection of Floating Oil Anomalies from the Deepwater Horizon Oil Spill with Synthetic Aperture Radar
Detection of oil floating on the ocean surface, and particularly thick layers of it, is crucial for emergency response to oil spills. While detection of oil on the ocean surface is possible under moderate sea-state conditions using a variety of remote-sensing methods, estimation of oil layer thickness is technically challenging. In this paper, we used synthetic aperture radar (SAR) imagery collected during the Deepwater Horizon oil spill and the Texture Classifier Neural Network Algorithm (TCNNA) to identify the spill’s extent. We then developed an oil emulsion detection algorithm using TCNNA outputs to enhance the contrast of pixels within the oil slick in order to identify SAR image signatures that may correspond to regions of thick, emulsified oil. These locations were found to be largely consistent with ship-based observations and optical and thermal remote-sensing instrument data. The detection method identifies regions of increased radar backscattering within larger areas of oil-covered water. Detection was dependent on SAR incident angles and SAR beam mode configuration. L-band SAR was found to have the largest window of incidence angles (19–38° off nadir) useful for detecting oil emulsions. C-band SAR showed a narrower window (20–32° off nadir) than L-band, while X-band SAR had the narrowest window (20–31° off nadir). The results suggest that in case of future spills in the ocean, SAR data may be used to identify oil emulsions to help make management decisions.
Analysis of Oil-Volume Fluxes of Hydrocarbon-Seep Formations on the Green Canyon and Mississippi Canyon: A Study With 3D-Seismic Attributesin Combination With Satellite and Acoustic Data
Natural hydrocarbon seeps have an important role in the carbon cycle and in the Gulf of Mexico (GOM) ecosystem. The magnitude of these natural oil seeps was analyzed with 3D-seismic attributes in combination with satellite and acoustic data. Hydrocarbon seepage in the deep water of the GOM is associated with deep cutting faults, generated by vertical salt movement, that provide conduits for the upward migration of oil and gas. Seeps transform surface geology and generate prominent geophysical targets that can be identified in 3D-seismic data. Seafloor-amplitude anomalies in plain view correlate with the underlying fault systems. On the basis of 3D-seismic data, detailed mapping of the northern GOM has identified more than 24,000 geophysical anomalies across the basin. In addition to seismic data, synthetic aperture-radar (SAR) images have proved to be a reliable tool for localizing natural seepage of oil. We used a texture-classifier neural-network algorithm (TCNNA) to process more than 1,200 SAR images collected over the GOM. We quantified more than 1,000 individual seep formations distributed along the outer continental shelf and in deep water. Comparison of the geophysical anomalies with the SAR oil-slick targets shows good general agreement between the distributions of the two indicators. However, there are far fewer active oil seeps than geophysical anomalies, probably because of timing constraints during the basin evolution. Studying the size of the oil slicks on the surface (normalized to weather conditions), we found that the average flux rate of oil (per seep) may be affected by the local change in the baroclinic and barotrophic pressures [e.g., warm core eddies (WCEs) and storms]. We found that oil slicks in the Mississippi Canyon (MC) protraction area tend to be more sensitive to pressure changes than Green Canyon (GC) protraction-area seeps.
Detection of thick patches of floating oil emulsions using X, C, and L-band SAR during Deep water Horizon oil spill
In this paper we use examples of Synthetic Aperture Radar (SAR) imagery collected during the Deepwater Horizon (DWH) oil spill and the Texture Classifier Neural Network Algorithm (TCNNA) to identify SAR image signatures that correspond to regions of emulsified (thicker) oil, which were verified by sea level observations and other remote sensing instruments. The method is sensitive to the SAR incident angles. L-band SAR was found to have the largest window of incidence angles (between 16 and 38 degrees off-nadir angle) that were able to detect Oil Emulsions (OE). C-band SAR were found to have a narrower OE detectable window (between 18 to 32 degrees off-nadir angle) than L-band. The X-band SAR had the narrowest OE detectable window (between 20 to 31 degrees off-nadir angle).
Adaptive thresholding algorithm based on SAR images and wind data to segment oil spills along the northwest coast of the Iberian Peninsula.
Satellite Synthetic Aperture Radar (SAR) has been established as a useful tool for detecting hydrocarbon spillage on the ocean’s surface. Several surveillance applications have been developed based on this technology. Environmental variables such as wind speed should be taken into account for better SAR image segmentation. This paper presents an adaptive thresholding algorithm for detecting oil spills based on SAR data and a wind field estimation as well as its implementation as a part of a functional prototype. The algorithm was adapted to an important shipping route off the Galician coast (northwest Iberian Peninsula) and was developed on the basis of confirmed oil spills. Image testing revealed 99.93% pixel labelling accuracy. By taking advantage of multi-core processor architecture, the prototype was optimized to get a nearly 30% improvement in processing time.
Remote-sensing evaluation of geophysical anomaly sites in the outer continental slope, northern Gulf of Mexico
Synthetic Aperture Radar (SAR) images obtained from satellites are a reliable tool for localizing natural hydrocarbon seeps. For this study, we used the Texture Classifier Neural Network Algorithm (TCNNA) to interpret SAR data from the RADARSAT satellite and a geostatistical clustering analysis to compare seeps detected in 579 SAR images covering the northern Gulf of Mexico (GOM). Geostatistical analysis results indicate that, in a typical active seep formation, oil vents would be found within a seep formation ∼2.5 km in diameter. Repeated observations of slicks at a given seep formation indicate that advection of rising oil in the water column causes an offset from the vent depending on water depth. At 500 m, the radial offset is up to 1400 m; at 2000 m, it is up to 3270 m. Observations with submersibles showed that, in 100% of the cases, the calculated seep formations that are matched with active oil seeps correspond to anomalies interpreted from surface amplitude maps and migration pathways in the seismic data. However, episodically, larger releases from persistent seeps occurred, and also some other seep formations showed intermittent releases. Our analysis indicates that active oil seeps detected with SAR represent a subset of the total array of geophysical features generated by hydrocarbon migration on the northern continental slope of the Gulf of Mexico.
Detection of Oil near Shorelines during the Deepwater Horizon Oil Spill Using Synthetic Aperture Radar (SAR)
During any marine oil spill, floating oil slicks that reach shorelines threaten a wide array of coastal habitats. To assess the presence of oil near shorelines during the Deepwater Horizon (DWH) oil spill, we scanned the library of Synthetic Aperture Radar (SAR) imagery collected during the event to determine which images intersected shorelines and appeared to contain oil. In total, 715 SAR images taken during the DWH spill were analyzed and processed, with 188 of the images clearly showing oil. Of these, 156 SAR images showed oil within 10 km of the shoreline with appropriate weather conditions for the detection of oil on SAR data. We found detectable oil in SAR images within 10 km of the shoreline from west Louisiana to west Florida, including near beaches, marshes, and islands. The high number of SAR images collected in Barataria Bay, Louisiana in 2010 allowed for the creation of a nearshore oiling persistence map. This analysis shows that, in some areas inside Barataria Bay, floating oil was detected on as many as 29 different days in 2010. The nearshore areas with persistent floating oil corresponded well with areas where ground survey crews discovered heavy shoreline oiling. We conclude that satellite-based SAR imagery can detect oil slicks near shorelines, even in sheltered areas. These data can help assess potential shoreline oil exposure without requiring boats or aircraft. This method can be particularly helpful when shoreline assessment crews are hampered by difficult access or, in the case of DWH, a particularly large spatial and temporal spill extent.
Chronic, Anthropogenic Hydrocarbon Discharges in the Gulf of Mexico.
Satellite-borne Synthetic Aperture Radar (SAR) was used to obtain more precise estimates of the magnitude of the chronic hydrocarbon discharges described in qualitative pollution reports associated with the production and transportation network of the U.S. coast of the Gulf of Mexico. The National Response Center (NRCen) oil pollution reports were collected and filtered for the period of 2001 to 2012 to determine which of the reports coincided with archived SAR images. Some of the images covered multiple reports and some of the oil discharges described in one report could be observed in more than one image. In all, 177 reports could be investigated from 137 SAR images collected on or near the corresponding report dates. Further analysis found that oil slicks observed in 66 of these SAR images could be attributed to 67 of the reported incidents. Objective measurements indicated that the area of these transient oil slicks visible in SAR images was, on average, significantly larger than what was reported to the NRCen. The only recurring point source for oil slicks was the former site of the Taylor offshore platform. Here chronic, oil slicks were observed that were consistently much larger than other anthropogenic discharges. The SAR images of floating oil discharged from the Taylor site were verified by visual inspection from a boat and aerial photography. For some of the oil slicks discharged from the Taylor site, the accuracy of SAR images for detecting oil slick areas was validated by comparing SAR results to Landsat 7 Enhanced Thematic Mapper Plus (ETMþ) and Moderate Resolution Imaging Spectroradiometer (MODIS) images. These results show that surveillance by SAR would improve accuracy for estimates of chronic anthropogenic oil pollution, particularly where continuous discharges are on-going.
Using SAR Image to Delineate Ocean Oil Slicks with a Texture Classifying Neural Network Algorithm (TCNNA).
Satellite-borne synthetic aperture radar (SAR) data are widely used for detection of hydrocarbon resources, pollution, and oil spills. These applications require recognition of particular spatial patterns in SAR data. We developed a texture-classifying neural network algorithm (TCNNA), which processes SAR data from a wide selection of beam modes, to extract these patterns from SAR imagery in a semisupervised procedure. Our approach uses a combination of edge-detection filters, descriptors of texture, collection information (e.g., beam mode), and environmental data, which are processed with a neural network. Examples of pattern extraction for detecting natural oil seeps in the Gulf of Mexico are provided. The TCNNA was successful at extracting targets and rapidly interpreting images collected under a wide range of environmental conditions. The results allowed us to evaluate the effects of different environmental conditions on the expressions of oil slicks detected by the SAR data. By processing hundreds of SAR images, we have also found that the optimum wind speed range to study surfactant films is from 3.5 to 7.0 m·s–1, and the best incidence angle range for surfactant detection in C-band is from 22° to 40°. Minor postprocessing supervision is required to check TCNNA output. Interpreted images produce binary arrays with imbedded georeference data that are easily stored and manipulated in geographic information system (GIS) data layers.
Oil Spill Mapping and Measurement in the Gulf of Mexico With Textural Classifier Neural Network Algorithm (TCNNA)
We developed a Textural Classifier Neural Network Algorithm (TCNNA) to process Synthetic Aperture Radar (SAR) data to map oil spills. The algorithm processes SAR data and wind model outputs (CMOD5) using a combination of two neural networks. The first neural network filters out areas of the image that do not need to be processed by flagging pixels as oil candidates; the second neural network performs a statistical textural analysis to differentiate between pixels of sea surface with or without floating oil. By combining the two neural networks, we are able to process a full resolution geotiff SAR image (16 bit, ~ 350 MB) in less than one minute on a conventional PC. The algorithm performs efficiently for all radar incidence angles when wind conditions are above 3 m/s. When low wind conditions are present, the performance of the neural network classification is limited, however the algorithm output allows the user to easily discard any elements of the classification and export the final product as a map of the water covered by oil. The results of this algorithm allowed us to process rapidly all of the images collected by Envisat during the Gulf of Mexico (GOM) Deepwater Horizon (DWH) oil spill event. By normalizing oil detections by the frequency that each area was sampled, we estimate that oil covered a mean daily area of 10,750 km 2 (with a total extent of 119,600 km 2 of the GOM surface waters), and approximately 1,300 km of the Northern GOM shoreline was threatened by the presence of drifting oil.
Mapping sea surface oil slicks using RADARSAT‐2 quad‐polarization SAR image
Polarimetric SAR decomposition parameters, average alpha angle (equation image) and entropy (H) are estimated for oil‐slick contaminated sea surfaces and slick‐free conditions using a RADARSAT‐2 quad‐polarization SAR image. The values of H and equation image within oil slick areas are significantly higher than those of the ambient sea surface, indicating the dominance of Bragg scattering for the slick‐free ocean and non‐Bragg scattering for the oil‐slick area. In land classification, the conformity coefficient (μ) is often used to discriminate surface scattering with double‐bounce or volume scattering. Based on this rationale, we also develop a method using μ as a logical scalar descriptor to map oil slicks under low‐to‐moderate wind conditions. The proposed method is assessed using a RADARSAT‐2 quad‐polarization SAR image of oil slicks in the Gulf of Mexico. Analysis shows that when μ is positive the sea surface is slick‐free, whereas μ is negative for oil‐slick areas. This method provides a simple and effective mapping technique for oil slick detection.