Marine heatwaves in the ocean analysis and seasonal predictions

Eric de Boisséson, Magdalena Balmaseda


Marine heatwaves (MHW) are defined as periods of usually five days or more of warmer sea-surface temperatures (SST) than the 90th percentile. The importance of marine heatwaves was highlighted by ‘the Blob’, a long-lasting event off the west coast of North America and Alaska over the 2014–16 period that decimated populations of Pacific cod, seabirds, salmon and other species while toxic algae prospered. Another major marine heatwave hit the northeast Pacific in the summer and autumn of 2019 (see ECMWF Newsletter 162). This strong event led the US federal cod fishery in the Gulf of Alaska to close for the 2020 season as a precautionary measure as the number of cod in the area was deemed too low. Advance planning of the fishing season is an example of the value that forecasting marine heatwaves could have for a better management of fish stocks and ecosystems in general. A better understanding of marine heatwaves (origin, characteristics, etc.) and their predictability is the objective of the marine heatwave task force of the Horizon 2020 EuroSea project. Within the project, ECMWF’s role is to investigate the representation of such events in ocean real-time monitoring products such as OCEAN5, and to investigate their predictability using the ECMWF seasonal forecasting system 5 (SEAS5).

Analyses and predictions

OCEAN5 is ECMWF’s current ocean and sea-ice analysis system. The OCEAN5 monitoring page has been online since 2019. The high-frequency real-time ocean monitoring capabilities provide an easy way to detect at a glance if any major anomaly is going on in the global ocean with respect to the long term ocean climatology ( Meanwhile, the behind real-time OCEAN5 monitoring offers a record of anomalies on ocean variables from 1975 to the present, with a focus on low frequency variability ( Large values of SST anomalies provide only a qualitative indication of heatwaves. Work is ongoing to enhance these monitoring tools for a more quantitative detection of marine heatwaves, e.g. events exceeding the 90th or 95th percentile of the climatological values. In 2020, strong SST anomalies were detected in OCEAN5, and values exceeding the 90% threshold were tagged as marine heatwaves, as shown in the figure.

OCEAN5 and SEAS5 marine heatwave days in the summer of 2020.
OCEAN5 and SEAS5 marine heatwave days in the summer of 2020. The charts show the number of days in June–July–August 2020 when SST anomalies exceeded the 90% threshold of the climatological value in OCEAN5 (left) and on average in the SEAS5 ensemble prediction system (right). Areas where such anomalies persisted for less than five consecutive days have been left blank. The black contour in the right-hand panel indicates areas where the probability of SEAS5 predicting a MHW is over 90%.

SEAS5 provides a 50‑member ensemble of 7‑month-long forecasts every month. They are based on the ECMWF Earth system model that couples the atmosphere, land, waves and the ocean. SEAS5 cannot predict the exact observed weather features but can provide indications on weather statistics on a monthly to seasonal basis. Here we explore if SEAS5 can predict the occurrence of marine heatwaves over the next few months. The forecast starting on 1 May 2020 was chosen to assess the ability of SEAS5 to predict marine heatwave events in the second half of 2020, like those captured in the OCEAN5 analysis. For each member of the forecast, we compute the number of days per month and season in which the SST anomalies are above the 90th percentile, and we compare the result to OCEAN5 monitoring (see the figure). 

The outcome is that, on average, SEAS5 can predict marine heatwaves in the right areas. The average number of extreme days is lower than observed, but the range of the ensemble encompasses the observations. These extreme values are predicted with more than 90% probability over many of the areas of occurrence.

Next step

These first diagnostics show there is potential for ECMWF SEAS5 to predict marine heatwave events in the same areas as seen in OCEAN5 in the summer 2020. The next step is to diagnose several years of seasonal forecasts to get more robust statistics on marine heatwave predictions that would provide indications on the possibility to make these predictions part of the ECMWF catalogue.

This work is supported by the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 862626, project EuroSea (Improving and Integrating European Ocean Observing and Forecasting Systems for Sustainable use of the Oceans).