Novel approach for the validation of Riskaware’s oil spill dispersion modelling capability

点击翻译按钮获取中文摘要

"In 2017, Riskaware contributed to a UKSA funded initiative known as EASOS (Earth and Sea Observation System). EASOS was led by the Satellite Application Catapult and aimed to help the Malaysian government with key environmental issues. One of the most serious of these challenges was that of marine pollution from oil slicks. As part of the consortium Riskaware developed its oil spill modelling capability and accompanying user interface, known as Marine Watch within the EASOS project, and Clear Ocean outside of it.The purpose of the model is to predict and help visualise the most likely path of oil in a spill and to determine the risk of that oil reaching the coast. Based on the information provided by the modelling, mitigation strategies can be planned, such as intercepting the oil at sea using oil booms and dispersant, and deploying land-based coastal clean-up teams. The system also has a number of powerful capabilities which are built onto its core model including probabilistic ensemble modelling and source term estimation.In order for the model's predictions to be relied upon, it is not just important for those predictions to be as accurate as possible, but also for the limitations of the model to be understood. With this in mind, this project will develop a set of methods and algorithms for validating the model, providing metrics that convey to potential customers the accuracy of the model, quantifying any uncertainties and aiding the deployment of the model in new areas of the world.Verification has been performed on the model already in order to prepare for a trial starting in Malaysia shortly, but a universal and unbiased method for validating the model is still required and this is what working with the STFC's Hartree Centre will provide. Hartree will use their role as an independent third-party, to develop a new set of robust and justifiable methodologies for quantifying the performance of the model, and by applying these methodologies, they will be able to carry out an unbiased and thorough validation of the model.The new methods developed under the project will help demonstrate to customers the full capability of the model. This will allow Riskaware to take the capability to market, driving strong export sales and company growth. We also plan to publish a joint paper with Hartree to report the validation results and the new methods we have developed."