Monitoring and predicting marine heatwaves in Vanuatu

Abstract Marine Heatwaves (MHWs) have disastrous impacts on ecosystems and communities in the south-west Pacific but there is limited research investigating their onset and evolution in this region. In collaboration with local fisheries and marine sectors, this study applied a MHW framework to define and categorize MHW events in the waters around Vanuatu. A range of events amongst the most intense and longest were investigated, as well as an anecdotally notable event from February 2016. This event was neither in the top five longest, nor the most intense of events, highlighting how impacts in marine ecosystems are non-linear and have cascading interactions with a region’s exposure and vulnerability. In analysing these events, we also explore how sub-seasonal to seasonal (S2S) prediction can be used to forecast MHW events. Hindcasts from the Bureau of Meteorology’s operational coupled ocean-atmosphere model were used to create weekly and monthly forecasts for each MHW event in the period from 1980 to 2018. We found that chance above 90th percentile hindcasts had promising accuracy with average hit rates highest for lead 0 weekly and monthly hindcast but conclude that hindcast accuracy is not always indicative of real time forecast accuracy as real time forecasts use a larger set of model ensembles. We also investigated an event outside the hindcast study period (May 2022) due to its notably ‘off the charts’ impacts. This event was the longest and most intense event on record, surpassing the previously longest and most intense event by 144 days and 0.56 °C. As climate change intensifies, such extreme events will become more frequent and will likely compound with other extremes, making the use and uptake of monitoring and prediction services critical to the long-term resilience of marine-reliant communities and sectors.