| 14 Aug 2026
Marine heatwaves and cold spells in the Philippine seas: evolution and drivers of recent extremes
Abstract. Marine heatwaves (MHWs) and marine cold spells (MCSs) are critical ocean thermal extremes with substantial ecological and socioeconomic impacts, yet their regional characteristics and underlying drivers remain poorly understood. The Philippines, a region of high marine biodiversity and exposed to rapidly changing ocean conditions, is particularly vulnerable to these extremes, but recent changes during exceptionally warm years remain insufficiently documented. Here, we investigate the spatiotemporal variability of MHWs and MCSs in the Philippine seas from 1993–2025 using a high-resolution (0.083° × 0.083°) sea surface temperature dataset. Results reveal pronounced spatial and temporal contrasts in the frequency, intensity, and duration of MHWs and MCSs, reflecting the combined influence of local ocean–atmosphere interactions and large-scale climate variability. We find that MHWs have become increasingly frequent, persistent, and intense in recent years, whereas MCSs have substantially declined, consistent with a shift toward a warmer regional mean state. The use of higher-resolution data enables the detection of fine-scale coastal and shelf processes and identifies localized MHW hotspots, particularly along the western coast of Luzon Island, which are not clearly resolved in coarser datasets. The contrasting trends in MHWs and MCSs are associated with a pronounced shift in the SSTA distribution toward warmer conditions, beginning around 2013. Case studies of major events during 2022–2025 further demonstrate that recent MHWs were shaped by the interaction between long-term warming and ENSO-related atmospheric circulation variability. La Niña and El Niño conditions contributed to distinct warming patterns through changes in regional ocean–atmosphere coupling, including persistent anomalous anticyclonic circulation, reduced cloud cover, enhanced shortwave radiation, and weakened ocean cooling processes. These findings highlight how basin-scale climate variability is translated into regional marine thermal extremes in a tropical archipelagic environment and provide important implications for monitoring, prediction, and risk assessment of ocean warming extremes in vulnerable marine regions.
Review of the manuscript “Marine heatwaves and cold spells in the Philippine seas: evolution and drivers of recent extremes” by Concolis et al.
The central message of this manuscript is that Philippine Sea MHWs have increased, MCSs have declined, and recent MHWs were driven by ENSO/WPSH-related anticyclonic circulation and enhanced shortwave radiation. While the topic is relevant, I have serious concerns regarding novelty, event attribution, methodology, and unsupported claims. I do not consider the manuscript suitable for publication in a journal of this standing.
Recommendation: Reject
1) The title and stated objectives promise an assessment of the drivers of both MHWs and MCSs, but the manuscript investigates drivers only for MHWs. Because no >21-day MCS occurs during 2022–2025, all the detailed process figures concern MHWs. Despite this, the title and Introduction, including the stated objectives, inappropriately oversell the scope of the work. There is a serious mismatch between the stated scope and the actual analysis. The literature also shows that MHWs and MCSs can arise from substantially different surface and subsurface mechanisms.
Zhang, N., Lan, J., & Dong, C. (2025). Subsurface heatwaves and cold spells in the South China Sea regulated by ENSO: Role of the South China Sea throughflow. Geophysical Research Letters, 52(12), e2025GL114692.
Li, C., Sun, W., Ji, J., & Zhu, Y. (2024). Historical marine cold spells in the South China Sea: Characteristics and trends. Remote Sensing, 16(7), 1171.
2) Throughout the manuscript, numerous methodological and internal inconsistencies raise serious concerns about reproducibility and potentially about the event-detection implementation itself. The manuscript also contradicts its own methodology in several places. For example, it states that an event must last at least five days to be considered an MHW, yet it reports event durations as short as four days. The Methods mention an 11-day moving window but do not clearly document the standard 31-day climatological smoothing and the percentile threshold used in the Hobday framework. The Methods state 800 hPa, whereas the figures show 850 hPa. Cumulative intensity is defined as the sum of daily intensity but is reported in °C rather than °C days.
3) The MCS component has a serious novelty issue. A Philippines-specific study already provides a systematic 1982–2021 assessment of MCS frequency, duration, intensity, trends, MHW–MCS asymmetry, and attribution to the long-term trend. It is not clear what fundamentally new understanding of Philippine MCSs is provided by the present manuscript. Simply changing the SST product, climatology, and analysis period is not sufficient for a new process-oriented paper unless those changes lead to demonstrably new conclusions.
https://www.researchsquare.com/article/rs-7801575/v1
4) The manuscript claims to identify physical “drivers,” but it does not actually perform a physical attribution of the MHWs or a heat-budget analysis. The central mechanism is inferred from maps of geopotential height, shortwave radiation, latent heat flux, and currents occurring at the same time as SST anomalies. Spatial concurrence is not a heat budget and does not establish causality. Nevertheless, the Discussion repeatedly states that the events were “driven by” atmospheric and oceanic processes. These claims are unsubstantiated.
5) The interpretation of the 2023–2024 MHW is not convincing considering a 2026 study of an essentially identical regional event. The manuscript attributes the event primarily to ENSO-related WPSH/anticyclonic circulation, suppressed convection, and enhanced shortwave radiation. However, Lu et al. (2026) performed an actual heat-budget analysis for the record-breaking 2023–2024 South China Sea MHW and found that reduced latent-heat loss accounted for ~64.7% of peak warming; long-term warming, positive IOD conditions, and tropical North Atlantic warming also contributed substantially, whereas the direct El Niño contribution was limited. I therefore find the ENSO-centered interpretation in this manuscript incomplete and potentially misleading.
Lu, X., Xiao, F., Wu, Z., Huang, J., Wang, Q., & Wang, D. (2026). A weakened East Asian winter monsoon triggered record-breaking marine heatwaves in the South China Sea during 2023–2024. Journal of Geophysical Research: Oceans, 131(7), e2026JC024234.
6) The novelty of this manuscript is not established because of a poor literature review and is substantially weakened by directly overlapping a recent work. A major omission is Gulakaram et al. (2026), which already provides a regional assessment of MHW frequency, duration, seasonal variability, physical drivers, and subsurface structure across Southeast Asia, explicitly including the Philippine Sea and South China Sea. The present manuscript directly overlaps with this existing understanding and fails to demonstrate sufficiently new scientific insight.
https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2025JC023614
7) The atmospheric fields used here are inadequate for diagnosing the evolution and drivers of daily MHW events. MHWs are identified from daily SST, yet the atmospheric variables are monthly means. The authors then discard partial months depending on whether an event begins or ends before or after the 15th of the month. This is an arbitrary procedure that can remove the synoptic-to-intraseasonal processes responsible for MHW onset and termination. Therefore, I consider the methodology used for this part of the manuscript inappropriate and potentially misleading.
8) The surface energy budget is incomplete, so the conclusions regarding atmospheric heating are quantitatively unsupported. Longwave radiation and sensible heat flux are omitted, and no net surface heat flux is calculated. An anomalous increase in downward shortwave radiation does not establish net ocean heat gain if turbulent and longwave heat losses compensate for it. Consequently, the physical interpretations are questionable and may be misleading.
9) The interpretation of ERA5 latent heat flux appears potentially incorrect in sign, which could invalidate a central physical conclusion. The manuscript interprets negative LHF anomalies, for example approximately −60 W m⁻², as “reduced evaporation.” ECMWF uses the convention that vertical surface fluxes are positive downward; therefore, a negative LHF anomaly indicates a smaller downward flux or a larger upward flux, i.e., potentially enhanced oceanic evaporative heat loss. No sign transformation is documented in the Methods.
10) Current velocity magnitude is incorrectly treated as evidence for horizontal heat advection. Claims that weak currents “inhibited horizontal heat transport” and that a stronger Kuroshio reduced MHW intensity are not physically demonstrated and are unsupported. Heat advection depends on not on current speed alone. The authors need to revisit this analysis and exercise much greater caution in making these claims.
Wang, Q., Zhang, B., Zeng, L., He, Y., Wu, Z., & Chen, J. (2022). Properties and drivers of marine heat waves in the northern South China Sea. Journal of Physical Oceanography, 52(5), 917–927.
11) I noted that the authors claim “prolonged thermal stratification” and “weak ocean mixing” without presenting any MLD analysis or results. These claims are therefore unsupported.
12) The claimed contribution of a warm-core anticyclonic eddy in the Sulu Sea is speculative and is not demonstrated by the results presented.
13) The principal MHW–MCS asymmetry is strongly conditioned by the methodological choice to retain the warming trend. The authors state that they did not detrend SST before event detection. They then identify increasing MHWs and disappearing MCSs as evidence of a warming-driven shift. However, this result is partly inherent when a fixed percentile threshold is applied to a warming time series. An important sensitivity analysis is therefore missing.
Chiswell, S. M. (2022). Global trends in marine heatwaves and cold spells: The impacts of fixed versus changing baselines. Journal of Geophysical Research: Oceans, 127(10), e2022JC018757.
14) The manuscript attributes the asymmetry almost entirely to a shift in mean SST without separating changes in the mean from changes in variance or the shape of the distribution. This is an inappropriate interpretation of changes in temperature extremes. MCS changes cannot simply be treated as the mirror image of MHW changes, and previous work has demonstrated distinct contributions from changes in both the mean and variability of SST.
Wang, Y., Kajtar, J. B., Alexander, L. V., Pilo, G. S., & Holbrook, N. J. (2022). Understanding the changing nature of marine cold-spells. Geophysical Research Letters, 49(6), e2021GL097002.
15) The claimed regime shift “beginning around 2013” is not convincing and is circular because 2013 is imposed by the analysis rather than detected statistically. The record is arbitrarily divided into 1993–2002, 2003–2012, and 2013–2025, after which the final period is found to be warmer, and the manuscript describes a shift beginning around 2013. I consider this interpretation inappropriate in the absence of a changepoint test or another objective detection method.
16) The domain-averaged event-detection approach is poorly suited to such a heterogeneous archipelagic region and may create artificial “regional events.” The authors first average SST over the entire 115–130°E, 4–21.5°N domain and then detect events in that time series. The domain includes dynamically distinct portions of the South China Sea, Sulu Sea, Celebes Sea, and western Pacific Ocean. A 100-day domain-mean MHW may therefore consist of different hotspots occurring sequentially in different locations rather than a single spatially coherent event. This is particularly problematic when that domain-mean event is subsequently assigned a single atmospheric mechanism.
17) The claimed “higher-resolution” novelty is factually overstated and, for the MCS comparison, incorrect. The manuscript uses 0.083° GLORYS and argues that its higher resolution improves upon previous studies. However, the Philippine MCS study already used 0.05° OSTIA SST, and the manuscript itself acknowledges that the earlier study used 0.05° data. The results may also change depending on the climatological baseline. Therefore, the claim that the present approach results in “reduced overestimation” is unsupported.
18) The definition and calculation of MHW/MCS intensity are internally inconsistent and may not follow the standard Hobday metrics. The methods state that MHW/MCS intensities are calculated as SST minus 90th/10th percentile thresholds, and Fig. 1 likewise defines the intensity relative to the percentile threshold. However, Table 1 defines mean and max SST anomalies. These are not the same quantities. In Hobday et al. (2016), these are relative to seasonal climatology. The authors’ methods mentioned in the present manuscript are inconsistent and could have impacted all the results in the manuscript, and thereby may alter claims.
Additional comments on figures: