Dr. Hans Graber at the University of Miami was awarded an RFP-II grant at $800,000 to lead the RFP-II project entitled “Monitoring of Oil Spill and Seepage Using Satellite Radars” that consisted of 1 institution and 8 research team members. Dr. Graber and his team worked toward a technological development that will increase situational awareness for agencies tasked with oil spill response. As part of the proposed work, the investigators aimed to develop a new statistics-based oil detection algorithm based upon a complete reanalysis of the Deepwater Horizon (DWH) Synthetic Aperture Radar (SAR) dataset. The DWH dataset consists of over 600 SAR images from 9 different satellites, each with unique metadata and image access protocols. Dr. Graber’s team developed a common metadata system for SAR to facilitate the use of the DWH imagery for algorithm development, as well as aid the application of the algorithm they developed to future satellite platforms. In keeping with the project goal of advancing the use of satellite SAR to improve oil spill response, and GOMRI’s encouragement to its various grantees to work together toward common goals, Dr. Graber’s team also participated in the Surface Current Oil Pathways Experiment (SCOPE) in collaboration with the CARTHE consortium. The forthcoming scientific results from this unexpected opportunity provide an important supplement to the technology development component of the project.
Dr. Graber’s research has to date research resulted in 3 peer-reviewed publications, 5 scientific conference presentations and 2 datasets being submitting to the GoMRI Information and Data Cooperative (GRIIDC), which are/will be available to the public. Dr. Graber engaged 2 students over his award period. Significant outcomes of his research according to GoMRI Research Theme are highlighted below.
Theme 4: Technology Developments
- In an extensive reanalysis, Dr. Hans Graber and colleagues used synthetic-aperture radar (SAR) images to contour the daily (and even hourly) extent of surface oil spreading during the DWH event. The contours, created by the same expert analysts who were active during the DWH event, were saved as shape files and then used to develop automatic detection algorithms. This dataset is the most comprehensive collection of surface oil coverage from DWH ever assembled.
- The dataset was used to develop and train a feed-forward neural network algorithm that can detect marine oil spills in an automated manner with minimal supervision. Inputs, in addition to the satellite images, include satellite radar parameters and ancillary information such as wind speed and direction. Dr. Graber’s team at University of Miami’s CSTARS, the premier satellite ground station in the Southeast US, was responsible for the download, analysis, and distribution of over 90% of the SAR imagery and oil detection products utilized during DWH. The new oil spill detection algorithm will be utilized by the CSTARS team in any future oil spill response scenario.
- Dr. Graber’s group collected multiple satellite SAR and optical images per day during the SCOPE experiment for an unprecedented view of tidally driven exchange processes between an estuary and surrounding coastal waters in the Gulf of Mexico. Freshwater outflow plumes from the Choctawhatchee River, typical of river plumes along the Alabama-Florida Panhandle coast, were observed on 15 days. These plumes are important not only because they are critical to understanding circulation near the coast, but also because they represent a virtual “buoyancy barrier” that might have protected sensitive estuaries and coastlines in this area from the landfall of oil from DWH. Results from this project include several firsts. Physical properties of the plumes such as their propagation speed and frontal thickness were made around the entire perimeter of a plume for the first time. This required the combination of SAR imagery from multiple satellites, and a new high-resolution optical satellite called RapidEye that had never been used for marine applications. A number of internal gravity waves, which are important for both mixing nutrients from the estuary and dispersion of spilled oil from offshore, were observed in shallower water and closer to shore than in any previous satellite study. Comparing the tracks of surface drifting buoys with salinity fronts observed in the satellite imagery, it was discovered that drifters can become embedded in the fronts from not only new tidal freshwater plumes, but from older “remnant” plumes as well. If surface oil acts similarly to drifting buoys, the behavior of these drifters provides evidence that freshwater outflow plumes can act as barriers to the transport of surface oil. The scientific results of the project will assist in the development of computer models to accurately predict these processes. This is a formidable challenge because of their complexity and the large range of spatial scales (from one meter to hundreds of kilometers) that must be resolved. Until that goal is met, Dr. Graber’s results show that high-resolution satellite imagery, if downloaded and processed in a timely manner, can directly assist spill response efforts by illustrating which shorelines may be relatively protected by freshwater outflow plumes, and which shorelines are at higher risk. Furthermore, the satellite images collected for SCOPE clearly showed how important such data is for situational awareness and providing critical context in understanding sparse in-situ measurements of ocean dynamics and processes and hence the behavior of hydrocarbon films in the environment. The importance of satellite data, especially in near-real time, seems to be totally missed by the GoMRI leadership.