Use of genetic algorithm to identify the source point of seepage slick clusters interpreted from Radarsat-1 images in the Gulf of Mexico

A large multitemporal set of RADARSAT-1 ScanSAR Narrow 1 images obtained in offshore regions of the Gulf of Mexico enabled the identification of a seepage slick cluster, which is considered to share a common geologic origin. The existence of seepage slick clusters is a positive indicator of present-day hydrocarbon generation and migration. Therefore, their correct location reduces the risk of acquiring piston cores with oil at the sea floor for further geochemical studies. A cluster is interpreted as a group of seepage slick polygons which share the same source point in geographic space. The source point can be tentatively defined as the intersection of overlaying polygons or as the intersection of the forward prolongation of closely spaced and converging polygons. The present study aims to identify quantitatively the origin of a seepage slick cluster at the sea surface using a genetic algorithms (GA). The model employ the Euclidian or the Mahalanobis distance function in order to determine the minimum distance among points within one seepage slick cluster and a population of points randomly generated. We used points in UTM coordinates system generated within a grid cell (x,y) of fifty meters defined in the seepage slick polygons that constitute a cluster In addition, one hundred points are generated randomly as the initial population. The fitness function provides the 10 best ranked points in UTM coordinates system that represent the candidates source point of the seepage slick cluster. The best source point identified using GA (considering both the Euclidian and Mahalanobis distance functions) coincided with a salt dome and fault seismically identified at the sea floor. The methodology have been tested elsewhere in the Gulf of Mexico, in order to identify with enhanced precision possible source points of seepage slick clusters detected using RADARSAT-1 images