Prediction of Oil Spills Events at Sea (POSEatSea)
Indian and ASEAN countries are located on global oil takner oceanic superhighway. This route also has sensitive ecosystems such as coral reefs and seagrass meadows in the vicinity. The environmental impact of an oil spill from a tanker ship is well documented. Oil Spills do not occur instataneously. There are events e.g. machinary malfunction that lead to such disasters. Before a spill occurs the crew might be well aware of the challenging situation and often spend time in avoiding such an event. During such times, the ship movement behavior may change from its normal pattern i.e. travelling at a predicted speed towards the port of call. At present, however there is no mechanism that can look into such patterns and predict a potential spill event to mitigate it in timely manner.
While the tankers have AIS and the systems are 'on', only predictive models can help flag unusual behavior that need to be monitored. An open-source based solution (R, QGIS w/ GRASS) could consist of a model that feeds on the tanker AIS data. AI/ML/NN (Artificial Intelligence / Machine Learning / Neural Network) based system can assess the fresh AIS data with the historical data of a) the given ship, b) the given type of cargo menifested and c) the given region to flag any pattern (ship course, speed etc) that is unusual. Such mechanism will help the monitoring agencies by narrowing down the region to utilize SAR (Synthetic Aperture Radar) data (satellite or drone based) and thus, can invoke faster mitigation response. Such a solution would cater to multiple of UN-SDGs and UNDOSSD objectives.