Bottom marine heatwaves along the continental shelves of North America

Marine heatwaves (MHWs), periods of extreme warm ocean temperatures, significantly impact marine ecosystems globally. Research has primarily focused on surface MHWs (SMHWs), which affect species distribution, primary productivity, and human-wildlife interactions. However, extreme warming at the seafloor (bottom marine heatwaves, BMHWs) can also have substantial biological consequences. While some studies have investigated subsurface MHWs using limited data, a comprehensive analysis of BMHWs along continental shelves is lacking. This study addresses this gap by using high-resolution ocean reanalysis data to characterize BMHWs across North American Large Marine Ecosystems (LMEs). Understanding BMHWs is crucial because intense bottom temperature changes uniquely impact demersal species. Bottom water temperature anomalies (BWTAs) have been linked to declines in fish abundance, shifts in species occurrence, redistribution of invasive species, altered recruitment dynamics, changes in predation pressure, and shifts in disease patterns. Furthermore, it is unclear if SMHWs, measured by sea surface temperature anomalies (SSTAs), are suitable proxies for BMHW conditions. While SSTAs and BWTAs can be correlated in shallow regions, subsurface currents and complex bathymetry can lead to BWTAs that differ in depth, intensity, and persistence from SSTAs. This study aims to comprehensively assess the intensity, duration, and spatiotemporal structure of BMHWs as unique events, providing crucial information for understanding their ecological impacts and informing marine resource management strategies.

Existing literature extensively documents the impacts of surface marine heatwaves (SMHWs) on marine ecosystems, including changes in species distribution, primary productivity, and increased risks of negative human-wildlife interactions. While there has been growing interest in the subsurface characteristics of MHWs, utilizing limited subsurface data from Argo profiles, moored buoys, underwater gliders, and gridded observations, research specifically focused on bottom marine heatwaves (BMHWs) along continental shelves has been limited. Previous studies have explored the effects of bottom water temperature anomalies (BWTAs) on various marine species and ecosystems, demonstrating their significant impacts on demersal communities and fisheries. However, the lack of a comprehensive, large-scale assessment of BMHW characteristics, including their intensity, duration, spatial extent, and relationship with SMHWs, necessitates the current research.

This study utilizes the Global Ocean Reanalysis and Simulations 12v1 product (GLORYS) from the Copernicus Marine Environmental Monitoring Service (CMEMS). GLORYS provides monthly mean ocean bottom temperatures at a high resolution of 1/12° (~8 km) with 50 vertical levels, spanning the period 1993-2019. The reanalysis uses the NEMO ocean model, forced by the ECMWF ERA-Interim atmospheric reanalysis and assimilates various data sources including satellite altimetry, SST, sea ice, and in situ temperature and salinity profiles. The study focuses on nine North American LMEs, analyzing bottom temperatures in depth intervals from the surface to 400 m. BMHW events were defined using a modified approach from Hobday et al. (2016), considering monthly mean bottom water temperature anomalies (BWTAs) exceeding a threshold defined by the 90th percentile of the climatological distribution for each grid cell and month, with persistence criteria. Similar methods were used to define SMHW events using SSTAs. Spatial extent was calculated by summing the area of all grid cells exhibiting BMHW/SMHW conditions, and the average intensity/duration calculated for each event. Statistical relationships between BMHW/SMHW intensity and depth were investigated using scatter plots and probability histograms, while Spearman correlations were calculated to assess the relationships. Synchrony between BMHWs and SMHWs was quantified by calculating the fraction of months where both events co-occurred at each location. The mixed layer depth (MLD) was incorporated to examine the link between surface and bottom events, calculating the ratio of MLD to ocean bottom depth to assess its association with synchrony. Comparisons were made between BMHW and SMHW intensity and duration, and the spatial extent of both events over time was analyzed, including instances of widespread BMHW and SMHW occurrences. The study also includes a validation of GLORYS bottom temperature data by comparing it to independent in situ observations at various coastal locations around North America, assessing agreement in overall variability and BMHW characteristics. The impact of linear detrending of the temperature data before MHW analysis was examined, comparing results with and without detrending.

The study reveals significant spatial variations in BMHW intensity and duration across different LMEs and depth intervals. Average BMHW intensity ranged from 0.5°C (at depths exceeding 200 m) to 5°C (in the Gulf of California at ~100 m depth). A generally negative linear relationship was found between BMHW intensity and depth in the East Bering Sea and Gulf of Alaska, while a more complex, non-linear relationship was observed in the Gulf of California and California Current, with peak intensity occurring at intermediate depths (50-100 m). In the Gulf of Mexico and along the North American east coast, the relationship between BMHW intensity and depth was less clear. The duration of BMHW events also varied significantly across LMEs, with longer durations often associated with deeper shelf regions. Analysis of the spatial extent of BMHWs revealed periods of widespread and prolonged events in several LMEs, notably during the 1997/1998 and 2015/2016 El Niño events and the 2012 Northwest Atlantic MHW. Comparing BMHW and SMHW events, the study found that BMHWs often persisted longer than SMHWs, and in many regions, BMHW intensity exceeded SMHW intensity. While widespread BMHW conditions frequently coincided with widespread SMHW conditions, significant BMHW activity occurred without a clear surface signature, highlighting the independent nature of BMHW events. The synchrony of BMHWs and SMHWs was higher in shallow regions, where the mixed layer depth (MLD) was likely to reach the ocean floor, linking surface and bottom water conditions. This relationship between synchrony and the MLD/bathymetry ratio is statistically significant across all LMEs. Validation against in-situ data showed that GLORYS accurately captures the observed bottom temperature variability and BMHW characteristics.

The findings demonstrate that BMHWs are significant events distinct from SMHWs, exhibiting considerable spatial variation in intensity and duration influenced by bathymetric features and regional oceanographic processes. The study suggests potential links between BMHWs and large-scale climate forcings, particularly ENSO, as evident in the West Coast LMEs. The observed time lags between widespread SMHWs and BMHWs in the Gulf of Alaska and California Current, coupled with the propagation of coastally trapped waves, indicate that subsurface processes play a crucial role in shaping BMHW characteristics. The different physical mechanisms governing SMHWs (atmospheric forcing) and BMHWs (subsurface currents, coastally trapped waves) explain differences in their spatial extent and timing. The observed higher synchrony of BMHWs and SMHWs in shallow areas is consistent with MLD reaching the ocean floor, linking surface and bottom water temperatures. The study's emphasis on BMHWs' independence from SMHWs underscores the need for improved subsurface monitoring to better understand and manage the impacts on marine ecosystems and fisheries.

This study provides a comprehensive assessment of BMHWs along North America's continental shelves, revealing their unique characteristics and importance for marine ecosystems. BMHWs frequently surpass SMHWs in intensity and duration, and can occur independently, particularly in deeper waters. The findings highlight the need for enhanced subsurface monitoring, beyond traditional surface-focused observing systems, to effectively manage marine resources and understand the full impact of extreme ocean warming events. Future research should focus on mechanistic studies examining the specific physical processes driving BMHWs in different LMEs, further exploring the relationship with large-scale climate patterns and improving the accuracy of predictive models for bottom temperature extremes.

The study's reliance on the GLORYS reanalysis, although validated against independent in situ data, introduces potential biases associated with model limitations and uncertainties in data assimilation. The relatively short observational period (27 years) may also limit the statistical power to detect long-term trends and variability. Finally, while the study addresses large-scale BMHW patterns, finer-scale investigations are necessary to fully understand the local processes and ecological implications of these events.