Global analysis of local driving mechanisms of marine heatwaves

In the past decade, marine heatwaves (MHWs) have attracted more attention in the research community due to their disastrous consequences on marine ecosystems around the globe. Consequent scientific efforts have led to considerable progress in defining and understanding drivers of MHWs. Most of the research has focused on the characteristics and dynamics of individual events in some high impact regions. However, according to the percentile definition, MHWs can occur anywhere in the world ocean, so there remains a need to understand global properties and mechanisms of MHWs. Although MHWs are globally increasing in frequency and duration, little is known about their behavior in coastal regions. Moreover, there is a lack of global study of the local physical processes controlling the evolution and long-term changes of MHWs. Such drivers are expected to differ for events at a different time or at a different place. Here, we first aim to provide a global analysis of MHWs in coastal regions, identifying hotspots and quantifying the amplitude and drivers of their long-term changes. Secondly, we investigate the difference of MHW long-term changes between coastal and offshore regions. Finally, we seek to understand the local physical drivers of extreme MHWs and assess how these mechanisms differ across the globe.

This analysis goes beyond the current approach in MHW research by combining data from multiple Sea Surface Temperature (SST) satellite products and using an ensemble approach to assess the influence of known SST product biases on MHW investigations and increase confidence in findings. To investigate coastal MHW properties, we select coastal pixels from 4 SST products and reveal a global increase of coastal MHW metrics. The increase in MHW threat is enhanced in regions such as mid latitudinal enclosed seas and parts of the north-western Atlantic coast, where average MHWs during the last decades were amongst the longest and most intense. Results are consistent across all satellite products, indicating that these MHW characteristics are not influenced by biases associated with satellite data retrieval algorithms and interpolation techniques.

However, we find that along more than 2/3 of coastlines, these trends are lower than those observed offshore, across all SST products. Consistent with recent literature, coastal trends of MHWs are largely attributed to long-term changes in mean SST. The cross-shore differences of MHW trends are also attributed to differences in SST trends. Surprisingly, neither air-sea heat fluxes nor wind driven upwelling derived from reanalysis data are found to be consistent drivers of the depressed coastal trends. We assess two global ocean circulation models for their performance in capturing the recent difference of MHW changes between coastal and offshore pixels and find they have a limited ability to simulate this difference except in a few regions. One such region, where models agree well with observations, is along the Chilean coast. A heat budget case study along the Chilean coast highlights the complexity of regional-scale coastal ocean-atmosphere feedback. An increased longwave radiative cooling onshore is found to be the main contributor to the different rates of long-term temperature change between onshore and offshore, rather than an increase in upwelling, as is commonly suggested in the literature. Long-term trend differences of surface heat fluxes derived from coarse resolution reanalysis data differ from our model heat budget analysis, which suggests that driving mechanisms of the slower onshore trends are regionally dependent.

Lastly, we conduct an upper ocean heat budget analysis at every pixel in a global 1/10^th degree ocean model. With this analysis, we investigate the drivers of anomalous upper ocean temperature changes associated with the onset and decay of upper ocean MHW events on a local scale. We defined an upper ocean MHW to be an event associated with depth-averaged ocean temperature above the winter mixed layer depth. Events, therefore, better represent the extreme state of the upper ocean where most marine ecosystems are present. Upper ocean MHWs can also be precursors of SST MHWs. We find that the most extreme upper ocean MHWs were driven by horizontal heat convergence in 78% of the global ocean, while MHWs dominantly driven by air-sea heat flux anomalies occur mostly in tropical regions. In contrast, the contribution of air-sea heat flux plays a larger role during MHW decay, mostly driven by latent heat fluxes. This suggests that an increased latent heat loss feedback response to the anomalous upper ocean heat content acts alongside advective cooling in controlling MHW decay. Most extreme upper ocean events have expressions at the surface, although the surface expression is not always as extreme due to an enhanced influence of lower atmospheric processes, demonstrating the need to differentiate upper ocean MHWs with near-surface MHWs characterized by SST.

Overall, this thesis makes a substantial contribution towards the on-going effort to increase our understanding of extreme events in the ocean. We demonstrate that in the context of climate change, coastal environments are likely to respond differently to increases in MHW related threats. Despite evident and prominent hotspots, coastal regions may globally act as thermal refugia for marine ecosystems from increases in long-term and pulsative MHW changes. We further highlight the complexity of local processes involved with both MHW formation and long-term changes, stressing the need to improve the representation of these mechanisms in climate models. In turn, improved model predictions will provide better insights to help inform various stakeholders about appropriate management decisions needed to alleviate some of the impacts of MHWs.