Rapid intensification of tropical cyclones in the Gulf of Mexico is more likely during marine heatwaves

The study addresses the challenge of predicting rapid intensification (RI) of tropical cyclones (TCs), a process defined by the National Hurricane Center as an increase of at least 35 mph in maximum sustained winds over 24 hours. The unpredictability of RI, particularly prior to landfall, significantly elevates risks to coastal communities, as intensity forecast errors are 2–3 times larger for rapidly changing storms. The paper highlights recent high-impact events, such as Hurricane Ian (2022), which underwent multiple RI phases over anomalously warm waters of the Caribbean Sea and Gulf of Mexico. RI is influenced by multiple factors, including elevated sea surface temperature (SST), reduced vertical wind shear, abundant moisture, and enhanced ocean heat content. Because prolonged SST anomalies (marine heatwaves, MHWs) can precondition the ocean and sustain elevated enthalpy fluxes, the central research question is whether RI events are more likely during MHWs in the Gulf of Mexico (GoM) and northwestern Caribbean Sea (NWCS). The purpose is to quantify the change in RI likelihood given MHW occurrence using a probabilistic, gridded framework over 1950–2022.

Prior research links warm ocean conditions to TC intensification and variability, with SST > 26 °C favoring cyclogenesis and modest SST increases (~1 °C) substantially boosting air–sea enthalpy fluxes. Studies of compound events emphasize that concurrent or sequential MHWs and RIs can act synergistically, with case evidence from TC Amphan (Bay of Bengal) and Hurricane Michael indicating that strong MHWs can enhance intensity change even under otherwise unfavorable atmospheric conditions (e.g., elevated vertical wind shear). Basin-scale analyses (e.g., Choi et al.) found that intensification rates for TCs encountering MHWs are roughly three times higher than for non-MHW TCs. Other works underscore roles of Loop Current heat content, barrier layers modulated by salinity (e.g., Mississippi River discharge), and downwelling-favorable winds in sustaining upper-ocean warmth. Despite these insights, comprehensive regional to global quantification of MHW contributions specifically to RI phases has been limited, motivating the current probabilistic assessment for the GoM and NWCS.

- Study domain and period: Gulf of Mexico and northwestern Caribbean Sea; TCs from January 1, 1950 to December 31, 2022 using IBTrACS (3-hourly best track). - RI extraction and declustering: RI defined as ≥35 mph increase in 24 h. To ensure independence, a minimum separation time (MST) of 24 h was imposed between successive RI detections along a TC track. - Gridding and baseline probability: A 1° × 1° grid was used. The empirical probability of RI occurrence per grid cell P_RI(i,j) was computed as the count of RI occurrences in that cell divided by the total number of RI events. - MHW identification: Daily-averaged ERA5 SST (1940–2022) processed with heatwaveR following Hobday et al. with adaptations: threshold at the 80th percentile (PC80), minimum duration ≥5 days, maximum gap ≤2 days to merge events, 11-day window for threshold estimation, and 31-day smoothing of climatology/threshold. Two climatology baselines were used to reduce biases from long-term warming: 1940–1980 (cooler) and 1981–2022 (warmer). - Double-threshold linking of MHWs to RI: For each RI start location/time, influential MHWs were those whose event window fell within ±10 days of the RI start and whose location was within 125 miles (~200 km) of the RI start. This accounts for the effective outer-core flux region and practical limits for preconditioning influence. - Conditional probability and amplification metric: The conditional probability P(RI|MHW) was computed as the frequency of RI-MHW co-occurrences (satisfying thresholds) normalized by the total number of TCs, on 1° grids. The multiplication rate (amplification) was defined as P(RI|MHW) divided by the conditional probability in the absence of MHWs, P(RI|¬MHW), to quantify how MHW presence changes RI likelihood. - Environmental context (2013–2022 subset): Tropical Cyclone Heat Potential (TCHP), depth of 26 °C isotherm (D26), and mixed layer depth (MLD) from NOAA’s ocean heat content product (0.25° resolution) were analyzed along with ERA5 latent heat flux (LHF) and vertical wind shear (VWS; 200–850 hPa). Means were computed over the 10 days prior to RI onset to relate subsurface heat and atmospheric conditions to observed RI hotspots. - Statistical significance: Gridded probabilities were evaluated with p ≤ 0.1 (90% confidence), with insignificance indicated where applicable.

- Prevalence of RI and linkage to landfall: In the GoM and NWCS, 119 of 406 unique TCs (29.3%) experienced at least one RI (1950–2022). About 70% of major landfalling TCs in the study area underwent RI at least once along their tracks. RI events predominantly occurred May–November, peaking in September (53 events), with August exhibiting fourfold more RIs than July (43 vs. 12). - MHW exposure and influence: Approximately 70% of hurricanes formed between 1950 and 2022 were influenced by MHWs. Using the double-threshold linking, 75 unique TCs (69.44% of all TCs that underwent RI) had at least one MHW within the defined impact area (≤10 days and ≤125 miles of RI start). - Amplification of RI likelihood: RI is on average 50% more likely during MHWs. Multiplication rates reach up to 5-fold (mean ~1.5-fold) in hotspot regions, indicating substantial amplification of RI likelihood in the presence of MHWs compared to their absence. - Spatial hotspots: Elevated P(RI|MHW) and amplification were found near the Cayman Basin in the NWCS, the Bay of Campeche (Campeche Canyon), the Yucatán Channel, and parts of the Texas–Louisiana Shelf. These align with Loop Current pathways and regions of high ocean heat content. - MHW trends: Between cooler (1950–1980) and warmer (1981–2022) periods, MHWs became more frequent, more intense (mean i_max_rel shift toward higher values), and longer in duration. Mean annual MHW duration rose from 36.5 to 49.5 days regionally, with notably prolonged events along the Loop Current, Yucatán Channel, Straits of Florida, and southwestern GoM. - SST anomalies and recent extremes: In 2023, daily SSTs exceeded the +2σ climatological envelope (1982–2011 baseline) for much of the year, reflecting increasingly anomalous warmth during hurricane season. - Ocean–atmosphere conditions: High TCHP values (e.g., 83–184 kJ cm⁻² near RI points for Harvey 2017, Michael 2018, Ida 2021, Ian 2022) and deeper D26/MLD were associated with RI, indicating substantial subsurface heat reservoirs. However, regional VWS and LHF patterns modulate RI likelihood, underscoring that MHWs are necessary but not sufficient conditions. - Total RI occurrences and distribution: After applying a 24 h MST, 165 RI occurrences were identified across 119 hurricanes: 81 storms with one RI, 31 with two, 6 with three, and 1 with more than three consecutive RIs.

The analysis demonstrates that marine heatwaves significantly elevate the likelihood of rapid intensification in the GoM and NWCS, addressing the central hypothesis that RI is more probable during MHWs. Spatial alignment of RI hotspots with regions of frequent, longer, and more intense MHWs, particularly along the Loop Current and adjacent areas, supports a mechanistic link through enhanced air–sea enthalpy exchange and subsurface heat reservoirs (high TCHP, deeper D26 and MLD). Nonetheless, RI formation depends on multiple co-favorable conditions: low to moderate vertical wind shear, sufficient ambient moisture (latent heat flux), and conducive synoptic environments. The observed patterns where high TCHP does not always coincide with high RI probability reflect the suppressive role of unfavorable atmospheric factors. The findings quantify the amplification effect: on average, RI is 1.5 times more likely during MHWs, with localized enhancements up to fivefold. Given projected increases in MHW frequency, duration, and intensity under continued ocean warming, the compounding MHW–RI risk is expected to rise, posing increasing challenges for forecasting and coastal risk management.

This work introduces a probabilistic framework that quantifies how marine heatwaves amplify the likelihood of tropical cyclone rapid intensification in the Gulf of Mexico and northwestern Caribbean Sea. By linking ERA5-derived MHW events to IBTrACS RI records via a double-threshold in space (≤125 miles) and time (≤10 days), the study shows that RI is on average 50% more likely during MHWs, with regional amplification up to 5-fold and clear hotspots aligned with Loop Current pathways. Increasing MHW frequency, intensity, and duration in recent decades further imply rising RI risk in a warming climate. Future research should: (1) refine MHW definitions for TC applications (including alternative thresholds and baselines), (2) integrate additional environmental controls (e.g., SST anomaly length scales, TC translation speed/direction, salinity-driven barrier layers), (3) expand datasets with synthetic TC scenarios and statistical–deterministic models for robust projections, (4) leverage machine learning (e.g., LSTM) and emerging MHW forecast tools for improved RI predictability, and (5) evaluate interactions among El Niño, MHWs, and RI to resolve regional variability in atmospheric shear impacts.

- Historical-event basis: Inference relies on observed records; limited sample sizes in some grids reduce statistical power, and environmental context analyses (e.g., TCHP, VWS, LHF) were restricted to 2013–2022 due to data availability. - MHW definition sensitivity: Use of PC80, smoothing, and baseline choices influence detection; standard Hobday thresholds were adapted to capture less extreme but frequent warming events, which may affect comparability. - Double-threshold simplifications: Fixed temporal (≤10 days) and spatial (≤125 miles) thresholds may not capture variability in MHW impact scales, SST anomaly length scales, and TC translation speed/direction. - Omitted factors: Some influential parameters (e.g., detailed barrier layer dynamics, river discharge variability, mesoscale eddies, case-specific downwelling) were not explicitly modeled in the probabilistic framework. - Necessary but not sufficient conditions: While MHWs enhance RI likelihood, unfavorable atmospheric conditions (e.g., high vertical wind shear) can suppress RI; thus, MHW presence alone does not guarantee RI.