the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Overcoming Challenges in Coastal Marine Heatwave Detection: Integrating In Situ and Satellite Data in Complex Coastal Environment
Abstract. Marine Heatwaves (MHWs) and Marine Cold Spells (MCSs) are extreme sea temperature events characterised respectively by unusually warm and cold conditions, that can persist from days to months. Both events occur in every ocean, but MHWs are becoming more frequent and intense due to global warming, to the detriment of MCSs. They are generally studied using satellite data with long temporal resolution to build a long-term climatology of the sea temperature. While MHWs and MCSs occurring in the open ocean have been well studied and documented, their occurrence and dynamics in coastal environment remain poorly understood, mainly due to the lack of data available. Indeed, coastal regions exhibit complex characteristics due to intricate coastal dynamics, the presence geographical features such as of gulfs, islands, channels or fjords, posing significant challenges for satellite observation.
In this study, we investigate the development and characteristics of MHWs and MCSs in Chilean Northern Patagonia over the period 2003–2023, as well as their seasonality and trends. Chilean Northern Patagonia is characterised by its complex geography, with the presence of thousands of islands, fjords, channels and gulfs. We present here a new methodology for MHWs and MCSs detection, using a combination of in situ and satellite data. Since the 1990s, Northern Patagonia has been well sampled, resulting in approximatively 3 million of samples across different depths. We interpolated these data using the DIVAnd algorithm to obtain a daily climatology of the sea temperature at 32 different depths (from the surface to 400 m) with a spatial resolution of 900 m that does resolve all the channels and fjords of the study area. Satellite data were used for threshold determination and to compare daily temperature to the climatology. The combination of the in situ-based dataset and the satellite-based do enable the detection of MHWs at very high resolution.
Our findings reveal that MHWs tend to be more frequent across most of the study area, whereas MCSs are becoming less common. MHWs and MCSs intensity is generally much higher in the most enclosed basins of the study area, with an average intensity of 2.5 °C for MHWs and -1.5 °C for MCSs in those places. However, MCSs intensity tends to increase over time whereas MHWs intensity tends to decrease in most basins. Some case studies are also described, including a succession of MCSs and MHWs in 2007–2008, and prolonged MHWs conditions during 2016–2017.
Competing interests: Aida Alvera-Azcárate is editor of the Special ISsue on Ocean Extremes (55th International Liège Colloquium)
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(4913 KB) - Metadata XML
-
Supplement
(7830 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-1421', Anonymous Referee #1, 22 May 2025
This manuscript presents a study of marine heatwaves (MHWs) and cold spells (MCSs) off the coast of Chile, a region characterized by a complex coastline, fjords, and challenging bathymetry. As a result, satellite data are often biased due to land interference, and in situ data are sparse, making the study of extreme events particularly difficult.
Unfortunately, while being well written, the manuscript suffers from several critical issues, and the methodology lacks both clarity and rigor—especially regarding the merging of satellite SST and in situ data. Line 205 states that “the detection of MHWs and MCSs is based on the comparison between daily climatology and daily temperature exceeding a threshold derived from the climatology. Here the climatology was calculated using in situ data, while the threshold a daily temperature were derived from satellite data”. Why do you need a climatology from in situ data? MHWs are simply events when daily temperature exceeds a climatological threshold (typically the 90th percentile) for a few days. How exactly are the satellite and in situ datasets integrated (which is in the title) and why? "Climatologies" sometimes refer to daily, then monthly means, and it is not clear if they are about the 50th or 90th percentiles. Also, please note that, when combining datasets, careful attention must be paid to differences in baseline periods (potentially requiring removal of temporal trends) and to the spatial variability each dataset resolves.
Regarding satellite SST, the authors correctly note limitations due to cloud cover and coastal complexity, but then proceed to use these same data to derive daily 90th percentile thresholds. The use of DINEOF to fill satellite gaps is reasonable in principle, but it is unclear why the method was applied separately for each year. Indeed, the EOFs shown in Figure 7 appear to display unrealistic small-scale offshore patterns.
The monthly climatology from in situ data is created by interpolating "3 million" measurements between 0 and 400 m onto a 900 m grid. However, no information is provided on the spatial or temporal distribution of these observations. Interpolating a large number of measurements into a high-resolution grid does not ensure that the resulting dataset is representative. For example, 3 million measurements could easily be generated from a thermistor with 5-minute resolution over a few sites. The observational dataset itself is not adequately described or visualized, which is a major omission.
It is also unclear what analysis was performed on the in situ data at depth, since all reported MHW statistics appear to be derived from satellite data (although this is not specified in the figure captions). At least the authors do not incorrectly apply surface-based thresholds to subsurface data, which would be inappropriate given the stratification evident in Figure 5.
Sections 3.5 to 3.7 present statistics on MHWs and MCSs, but these are largely descriptive and difficult to follow due to the extensive use of local fjord names. This section reads more like a regional technical report than a manuscript suitable for an international scientific audience, and I worry that the spatial variability is due to the interpolation of too sparse data (which is evident in Figure 6). Again, interpolation is not magic and should not be abused when the initial dataset it too sparse.
Section 3.8 discusses several case studies of extreme events and includes potentially interesting insights into the evolution of atmospheric drivers. However, these claims are unsupported by key evidence, such as a surface heat budget or time series of heat fluxes, wind speed, and sea level pressure. Moreover, the so-called "high-resolution" results appear to be based on ¼° reanalysis data, which is unlikely to resolve the processes of interest in such a complex coastal setting.
In summary, I do not believe this manuscript is suitable for publication. Substantial revisions are needed, beginning with a clearer and more rigorous methodological design and a more careful alignment of datasets with the research questions.
Citation: https://doi.org/10.5194/egusphere-2025-1421-RC1 - AC1: 'Reply on RC1', Cécile Pujol, 21 Jun 2025
-
RC2: 'Comment on egusphere-2025-1421', Anonymous Referee #2, 23 Jun 2025
Title: Overcoming Challenges in Coastal Marine Heatwave Detection: Integrating In Situ and Satellite Data in Complex Coastal Environment
Author(s): Cécile Pujol, Alexander Barth, Iván Pérez-Santos, Pamela Muñoz-Linford, and Aida Alvera-Azcárate
MS No.: egusphere-2025-1421
MS type: Research article
Iteration: Initial submission
Special issue: Special issue on ocean extremes (55th International Liège Colloquium)
The paper proposes a new method for detecting MHWs by combining satellite data and in-situ data. This helps in places where either data is sparse in space or time. The paper proposes using in situ data, combined with DIVAnd interpolating method, to build a climatology and satellite data to set thresholds. This is then used to analyse MHW occurrence in a fjord in northern Patagonia.
The paper is well written and the figures are clear. There are some minor remarks that should be addressed prior to publication.
My general remark is that this analysis would benefit from analysing salinity profiles as well. Fjords tend to have a well defined stratification not only due to temperature but also due to freshwater discharge. Riverine and precipitation preconditioning on monthly to seasonal timescales could play some role in the observed MHW seasonality and intensity but this is not discussed.
Specific remarks: (L=line)
L83: why 900 meters specifically?
L145: how exactly was seasonal interpolation performed to create a seasonal climatology? Mean / running average over JFM, AMJ, etc?
How was the second interpolation “realized” to generate a monthly climatology?
L166: I am not sure I understand how do you perform a 90-day moving average over a monthly climatology? Monthly climatology means you have 12 values. How do you do 90-day rolling mean?
Section 2.3, L201: I think the readers would benefit not only from calculated bias, but from a scatter plot SAT DATA vs IN SITU DATA. This would allow an estimate where you have over(under) estimations…
Section 2.5. I would suggest to either write
Qi = Qs + Qb + Qe + Qc,
where the last three fluxes can be negative or positive. Also define when a specific term is positive or negative. For example, is Qb negative when ocean is losing heat to the atmosphere? If so, then Qs-Qb = Qs + |Qb|? Otherwise this can lead to a very usual confusion which could easily be avoided.
Table 1. Again, I would suggest to add a scatter plot CLIM vs IN SITU.
Figure 8. Is this all computed at the surface? You have 3D data – did you make any analyses of deep marine heatwaves at other depths?
L380-390: Here I am wondering about the riverine input into the fjord? Freshwater induced stratification could boost a surface MHW by inhibiting vertical mixing through the pycnocline. This could explain why the most enclosed regions of the fjord exhibit the highest intensity. I think it would be very welcome to check salinity fields here.
The same comment goes to Figure 9: is there a clear summer halocline? What are the density profiles in the fjord? This could yield a clear salinity-based mechanism of why summer exhibits more MHWs than other seasons.
Citation: https://doi.org/10.5194/egusphere-2025-1421-RC2 - AC2: 'Reply on RC2', Cécile Pujol, 01 Jul 2025
-
EC1: 'Comment on egusphere-2025-1421', Mélanie Juza, 18 Jul 2025
Dear Cécile Pujol and co-authors,
Based on the answers, the reviewer 1 considers that there are still major issues:
- The way to calculate daily climatologies from in situ data is basically a series of interpolating (seasonal, then monthly, then daily). This sounds dodgy, and they did not provide enough explanation and analysis to show that it is reliable.
- The MHW part is just at the surface using MODIS data, again with very strong interpolation. While DINeof is usually a good tool, I feel like the gaps are too large, especially for a calculation of the 90th percentile (the example map they have is far from convincing).
- I think the title is very misleading, they don’t overcome anything.
The reviewer would encourage the authors to write a data paper on the climatology (both the mean and 90th percentile) but spending more time to show the method and uncertainties. The in in situ data they have is very valuable, but we feel like it is not clear that it is used properly. The MHW part is underwhelming.
Although the paper shows serious work and it is well-written, I encourage the authors to reorientate the manuscript following the recommendations of the reviewer and resubmit the manuscript in a more suitable journal/issue.
Best regards.
Citation: https://doi.org/10.5194/egusphere-2025-1421-EC1
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
444 | 95 | 21 | 560 | 28 | 21 | 34 |
- HTML: 444
- PDF: 95
- XML: 21
- Total: 560
- Supplement: 28
- BibTeX: 21
- EndNote: 34
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1