ABSTRACT

Modelling the drift of marine debris in quasi-real time can be of societal relevance. One pertinent example is Malaysia Airlines flight MH370. The aircraft is assumed to have crashed in the Indian Ocean, leaving floating wreckage to drift on the surface. Some of these items were recovered around the western Indian Ocean. We use ocean currents simulated by an operational ocean model in conjunction with surface Stokes drift to determine the possible paths taken by the debris. We consider: (1) How important is the influence of surface waves on the drift? (2) What are the relative benefits of forward- and backward-tracking in time? (3) Does including information from more items refine the most probable crash-site region? Our results highlight a critical contribution of Stokes drift and emphasise the need to know precisely the buoyancy characteristics of the items. The differences between the tracking approaches provide a measure of uncertainty which can be minimised by simulating a sufficiently large number of virtual debris. Given the uncertainties associated with the timings of the debris sightings, we show that at least 5 items are required to achieve an optimal most probable crash-site region. The results have implications for other drift simulation applications.

摘要

准实时模拟海洋垃圾的漂移可能具有社会意义。一个相关的例子是马来西亚航空MH370航班。假定这架飞机在印度洋坠毁，留下残骸漂浮在水面。其中一些物品是在西印度洋周围发现的。我们使用由操作海洋模型模拟的洋流结合表面斯托克斯漂移来确定碎片可能采取的路径。我们认为：（1）表面波对漂移的影响有多重要？（2）时间向前和向后跟踪的相对好处是什么？（3）包含更多项目的信息是否可以优化最可能的崩溃地点区域？我们的结果强调了斯托克斯漂移的关键作用，并强调了需要精确了解这些物品的浮力特性的需要。跟踪方法之间的差异提供了不确定性的度量，可以通过模拟足够数量的虚拟碎片将不确定性最小化。考虑到与发现碎片的时间有关的不确定性，我们表明至少需要5件物品才能实现最佳的最可能发生碰撞的区域。结果对其他漂移仿真应用具有影响。跟踪方法之间的差异提供了不确定性的度量，可以通过模拟足够数量的虚拟碎片将不确定性最小化。考虑到与发现碎片的时间有关的不确定性，我们表明至少需要5件物品才能实现最佳的最可能发生碰撞的区域。结果对其他漂移仿真应用具有影响。跟踪方法之间的差异提供了不确定性的度量，可以通过模拟足够数量的虚拟碎片将不确定性最小化。考虑到与发现碎片的时间有关的不确定性，我们表明至少需要5件物品才能实现最佳的最可能发生碰撞的区域。结果对其他漂移仿真应用具有影响。