The European Space Agency-funded CAREHeat project has made significant progress in marine heatwave (MHW) research through the development of comprehensive global and subsurface sea temperature catalogues, enhancing our understanding of extreme ocean temperature events and their impacts on marine ecosystems and economies, particularly in the context of climate change.
One of the main objectives of CAREHeat deals with the development of 40-year MHW global catalogues from sea surface temperature (SST) data with a daily temporal resolution and a spatial resolution of 1°x1°. Such catalogues cover the period between 1982 and 2022, encapsulating key characteristics of marine temperature extremes, such as location, category, abruptness, and duration.
Atlas comparison of the occurrence of marine heatwaves between the
ESA CCI dataset and US the National Oceanographic and Atmospheric
Administration product over the global oceans (reproduced from the dataset’s technical documentation)
Two catalogues were produced. The MHW Global Catalogue v1 employs the standard detection method proposed by Hobday et al. (2016), with the reference climatology set to 1991-2020. The MHW Global Catalogue v2 is similar to v1, but long-term SST trends were removed. The purpose of this detrending was to separate the slow-varying SST component — including trends and climate phenomena, such as the El Niño Southern Oscillation (ENSO) — from transient changes and extreme events.
Marine heatwaves conditions, based on ESA CCI data and expressed as categories, at the peak of the strong 1997-1998 El Niño event in the tropical Pacific
CAREHeat’s validation procedures reveal that altering the reference climatology from 1991-2020 to earlier periods (such as 1985-2012, as suggested by the US NOAA) results in significant changes to MHW characteristics and, consequently, the definition of extreme events that occurred during the satellite era. As anticipated, v2 demonstrated reduced sensitivity to the choice of reference climatology, thereby providing a more reliable framework for identifying MHWs.
Illustration of the MHW Atlas at the surface with different metrics (intensity, duration and frequency)
2D Atlas: Effect of the removal of the temperature trend in MHW detection (Martinez et al 2023)
CAREHeat prepared weekly subsurface MHW 4D atlases at both global and regional scales. The global atlas, covering the period 1993-2021, features a weekly temporal resolution and a 0.25° spatial resolution, with input data sourced from the ARMOR3D dataset, distributed by the Copernicus Marine Service.
Example of MHW at different depths (0, 20, 50, 100m) for the 24/07/2019: we see regions (ex: Central Pacific, southern branch of the subtropical Atlantic gyre) where surface MHW have a subsurface signature
The regional atlases for the Western Mediterranean Sea, the Tropical Pacific, and Madeira were produced through a two-step approach. First, a Machine Learning algorithm (Multi-Layer Perceptron, MLP) was trained. ARMOR3D temperature and salinity fields served as targeted data for the MLP learning. SST and sea surface height (SSH) from satellite observations were the input data for that process. Afterwards, an optimal interpolation method was used to merge the results of this first step with in-situ observations.
The MHWs were then estimated across depths from the surface to 300 meters, maintaining a weekly resolution and a spatial resolution of 0.25°. Notably, in the Mediterranean region, subsurface warming trends since 2014 introduced biases in the final reconstruction of the MHW data. Two data periods were therefore used: 1993-2015 for MLP learning; 2016-2021 for inference. This allowed to explain those trends. Over the period used for the model training, no significant subsurface warming occurred because that behaviour was not identified by the ML.
The challenge arose because the learning algorithm faced difficulties when the target variable shifted over time, reaching values that had not been previously observed during the training phase. Furthermore, the observability of such events from SST data was limited, with SLA showing very weak signals due to subsurface salinity and temperature compensating for each other.
Subsurface warming since 2014-2015 in the Liguro-Provençal Basin (train period in blue up to 2018 in this example instead of 2015, test period in green)
The outputs from both the 2D and 4D regional atlases underwent rigorous validation against in situ observations. This validation included comparisons with equivalent MHW atlases and ARMOR3D temperature reconstructions, ensuring the robustness of the results.
The outputs from CAREHeat offer considerable value for marine scientists and ocean-based industries alike. The availability of both detrended and non-detrended atlases enables a comprehensive analysis of how low-frequency climate variability affects MHW characteristics. The subsurface atlases allow for a detailed exploration of how MHWs propagate vertically in the ocean, a critical factor in understanding the full ecological impact of these extreme events.
These resources can help stakeholders in aquaculture, fisheries, and marine protected areas better understand the risks posed by MHWs and inform mitigation strategies. The CAREHeat team has utilized these atlases to assess the vulnerability and adaptability of aquaculture farms and marine protected areas to MHWs, enabling targeted strategies for monitoring and management. They will also play a crucial role in ongoing research, helping scientists characterize extreme marine heatwave events, assess their drivers, and evaluate their impacts on biodiversity, biogeochemistry, and compound stressors.
Evolution of meridionally averaged SST anomalies over the Mediterranean as a function of the longitude and time from January 1982 to April 2023 (a). Zoom over the 2003/04 (b), and over the 2022/23 (c) intervals, bounded by the dashed lines in plot (a). (reproduced from Marullo et al., 2023)
The findings and tools developed by the project are expected to have a lasting impact beyond its lifespan. They will be instrumental in future studies aimed at understanding MHWs and their broader implications. The global and subsurface atlases will continue to be used for research on marine biodiversity, the role of MHWs in biogeochemical cycles, and their potential to cause compound events alongside other oceanic stressors.
By providing a comprehensive framework for identifying and analysing MHWs, these resources will enable the identification of new MHW events and provide the scientific foundation for better-informed management of marine ecosystems under climate change pressures.
The CAREHeat project has led to significant research advances, contributing to a deeper understanding of Marine Heatwaves (MHWs) and their impact on ecosystems. Here are some key publications that resulted from the project:
Stay updated with the project by visiting the CAREHeat website or following us on Twitter/X.