Impacts of tropical cyclones on the global water budget

The study examines how tropical cyclones (TCs) modify the global hydrological cycle by inducing anomalous moisture transport. TCs are known to redistribute large amounts of water vapour from ocean to land and from the tropics toward higher latitudes, significantly contributing to precipitation in affected regions, though their poleward moisture transport is smaller than that of atmospheric rivers. Moisture availability is essential for TC genesis and intensification, and increased mid-level humidity favors intensification. Prior work shows TC precipitation often derives from nearby moisture sources steered by regional circulation. However, it remains unclear which areas experience anomalous evaporation versus anomalous precipitation during TC events globally, how these anomalies compare to climatology, and whether long-term trends in TC-induced surface freshwater flux (evaporation minus precipitation; E−P) are linked to global warming and rising sea surface temperatures (SSTs). Motivated by projected increases in atmospheric water vapour with SST warming (approximately 6–7% per °C), the study aims to quantify TC-induced anomalies in surface freshwater fluxes at regional and global scales and assess their variability, ENSO modulation, and long-term trends since 1980.

Previous research has addressed components of the TC water cycle but seldom the global anomalous moisture transport induced by TCs. Studies have investigated TC water budgets and found moisture convergence is the main contributor to TC rainfall; TCs contribute substantially to onshore moisture transport (e.g., 14–19% of net onshore moisture to North America) and can account for significant fractions of annual precipitation totals regionally and globally. Lagrangian moisture tracking has been widely used to identify moisture source–sink relationships for TCs and other systems, revealing that TC precipitation primarily draws from nearby sources influenced by large-scale circulation. Work on anomalous moisture transport linked to TCs is comparatively limited; case studies indicate anomalous moisture flow from adjacent ocean basins can drive extreme rainfall during specific hurricanes, and spatial patterns of anomalous moisture convergence align with cyclonic anomalies in some regions. Broader climate context shows atmospheric moisture content scales with SST per Clausius–Clapeyron and that extreme precipitation changes can exceed this scaling due to dynamical intensification. ENSO strongly modulates basin-scale TC frequency and tracks, implying potential ENSO impacts on TC-induced moisture transport. Gaps remain in globally quantifying where, when, and how TC events produce E−P anomalies versus climatology, and how these anomalies have trended with warming.

The analysis uses a global Lagrangian moisture tracking approach with the FLEXPART model to quantify TC-induced surface freshwater flux (E−P) anomalies over 1980–2018. TC best-track data were compiled from HURDAT2 (North Atlantic and NE Pacific) and Joint Typhoon Warning Center (other basins), at 6-hourly resolution; TC outer radii were obtained from the TCSize dataset. FLEXPART v6.2 was driven by ECMWF ERA-Interim reanalysis (6-hourly, 61 vertical levels, 1°×1° grid), a configuration validated in prior source–sink studies and comparable to ERA5-based results for these purposes. FLEXPART advects approximately two million equal-mass parcels, using large-scale meteorology and parameterizations for convection (Emanuel–Živković-Rothman) and boundary-layer turbulence (Langevin). For each 6-hourly TC position, all air parcels within the TC outer radius were backtracked for up to 10 days (240 h), the typical atmospheric residence time of moisture from evaporation to precipitation. Along parcel trajectories, changes in specific humidity dq/dt were converted to parcel freshwater flux (e−p) using the Lagrangian moisture budget (e−p = m dq/dt), and aggregated over 1°×1° grid cells to yield surface freshwater flux (E−P). Positive E−P indicates net evaporation (source), negative indicates net precipitation (sink). For each TC time and location, a climatological E−P field was computed by averaging E−P at the same month/day/hour over 1980–2018 (including TC times), enabling anomaly estimation as TC-related E−P minus climatology. Statistical significance was tested using the Wilcoxon signed-rank test (95%). ENSO modulation was assessed by classifying years as El Niño, La Niña, or neutral based on the Oceanic Niño Index (ONI) and comparing TC-related E−P fields between El Niño and La Niña years, with significance testing as above. Trends in TC-related E−P were computed grid-to-grid for 1980–2018, highlighting significant tendencies (p<0.05), and examined alongside SST trends and correlations. Basin and global annual mean SSTs during TC occurrence were derived from NOAA OISST (0.25°, 1982–2018), averaged over regions of higher TC activity per basin. Relationships between TC-related E−P and SST, TC size, frequency, lifetime, and accumulated cyclone energy (ACE) were analyzed (including Spearman correlations and linear trends). To compare evaporation and precipitation within storms, lifetime accumulated TC-induced precipitation within the TC outer radius was summed from MSWEP v2 (0.1°, 3-hourly) and evaporation from ERA-Interim, for each storm and aggregated by basin and globally. The global annual TC-related E−P time series was constructed from FLEXPART outputs and its trend assessed.

- Spatial anomalies: During TC days, evaporation and precipitation are both enhanced relative to climatology. Positive E−P anomalies (net evaporation) dominate broadly over the tropics, while negative anomalies (net precipitation) are more regionally concentrated along climatological sink regions, notably the ITCZ in both hemispheres, the Western North Pacific monsoon trough, and Southeast Asia. Large positive anomalies reflect strong TC-driven moisture transport; major negative anomaly nuclei occur in the WNP and NEPAC, consistent with their high TC frequencies (~31% and ~20% of global annual TCs, respectively). Southern Asia exhibits positive E−P anomalies (implying reduced precipitation relative to same-date climatology) during TC days. - Seasonal modulation: The TC impact on the water budget is stronger in the Northern Hemisphere and peaks in August–September, aligning with seasonal TC frequency and high SSTs. - ENSO influence: Compared with La Niña years, El Niño years show more negative TC-related E−P (greater net precipitation) over southern Asia and along the Pacific and eastern Indian Ocean ITCZ, and less negative over the Gulf of Mexico/southern US and Atlantic/western Indian ITCZ. Positive E−P (net evaporation) increases during El Niño over large parts of the Pacific, Wharton/Perth basins, Arabian Sea, southern Bay of Bengal, Caribbean, and western North Atlantic, and decreases north and south of the Atlantic ITCZ, the central Indian Ocean south of the ITCZ, and the Somali Basin. - Long-term trends (1980–2018): There is a widespread, statistically significant decrease in TC-related E−P across many basins, especially in the South Indian and Pacific. Sink regions over the eastern tropical Pacific and the Philippine Sea show the largest reductions, up to ~70–90 and ~40–60 mm year−1, respectively. Source regions (e.g., Wharton and Perth basins) also decrease by ~40–90 mm year−1. - Global decline: The global annual TC-related E−P exhibits a statistically significant decreasing linear trend of approximately −40 mm year−1. The decline is linked to increasing SST and slight decreases in global TC frequency and lifetime in the last two decades. - SST relationship: Basin-wide SSTs during TC occurrence increase significantly (e.g., ~0.035–0.04 °C year−1 in NEPAC and SIO), and SST is inversely correlated with TC-related E−P (global r ≈ −0.65). The authors estimate that a 1 °C SST warming corresponds to an 86% reduction of TC-related E−P relative to its 1980 value. Positive correlations occur over sink regions, consistent with decreasing magnitude of negative E−P (i.e., less negative) as SST rises. - TC characteristics: Average TC size increases significantly (~0.84 km year−1), and increases by ~6 km per 1 °C SST warming. TC-related E−P decreases with increasing TC outer radius (~8 mm year−1, p<0.1), indicating increasing TC-related precipitation versus evaporation. ACE shows a significant relationship with TC-related E−P after 1990 (1.18 mm day−1 per 10−4 kt2, p<0.05), supporting the overall reduction in TC-related E−P. Interannual variability of ONI, SST, TC frequency, size, and lifetime-accumulated evaporation and precipitation explains ~80.5% of annual TC-related E−P variability. - Within-storm budgets: Lifetime accumulated (evaporation − precipitation) within TC outer radii increases in the North Atlantic but decreases in most other basins and globally (significant in South Pacific, WNP, and globally), underscoring the role of moisture flux convergence from external sources in supporting TC precipitation.

The findings demonstrate that TCs systematically alter the global water budget by enhancing evaporation over source regions and precipitation over sink regions relative to same-date climatology, with the strongest impacts aligned with regions and seasons of high TC frequency. ENSO modulates these patterns through its control of basin-scale TC activity and associated circulation and convection anomalies, enhancing TC-related moisture transport in the Pacific during El Niño while reducing it in the North Atlantic and parts of the Indian Ocean. The pervasive, significant decline in TC-related surface freshwater flux since 1980 indicates that TC-induced precipitation has increasingly balanced or exceeded TC-induced evaporation, consistent with rising SSTs, slight reductions in TC frequency and lifetime in recent decades, increasing TC size, and the dominant role of moisture flux convergence in TC rainfall generation. The inverse SST–E−P relationship suggests that in a warmer climate, increased low-level moisture availability (scaling near or above Clausius–Clapeyron) and potential dynamical intensification can support greater TC-induced precipitation, including moisture transported from outside the immediate storm environment, while net E−P declines. Regional nuances, such as reduced negative E−P over southern Asia during TC days due to monsoon background conditions, highlight the interplay between TCs and seasonal circulations. Overall, the results clarify where and how TCs contribute anomalous moisture fluxes and how these contributions are changing in a warming climate, while noting complexities and feedbacks in TC–ocean interactions that can modulate these responses.

This work provides a global quantification of TC-induced anomalies in surface freshwater flux using a Lagrangian moisture tracking framework. It shows that TCs increase evaporation over moisture source regions and precipitation over sink regions compared to climatology, with stronger impacts in the Northern Hemisphere and during peak TC season. ENSO substantially modulates these patterns via basin-scale changes in TC activity. Critically, TC-related surface freshwater flux has declined significantly since 1980, at a global rate of about −40 mm year−1, with an estimated 86% reduction per 1 °C SST warming relative to 1980, likely reflecting rising SSTs and slight decreases in TC frequency and lifetime alongside increases in TC size and moisture convergence contributions. These results have implications for anticipating hydrological impacts of TCs under continued global warming. Future research should refine understanding of TC–SST feedbacks, integrate additional ocean–atmosphere coupled processes, and assess projections of TC-induced moisture transport considering expected changes in TC frequency and intensity.

- The Lagrangian method neglects liquid water and ice content, parcel mixing, and evaporation of precipitating hydrometeors, potentially affecting E−P estimates. - ERA-Interim’s coarse resolution may not fully capture TC extremes, although prior validation indicates reasonable agreement with observations for TC-related precipitation using this approach. - The rate of change in TC-related E−P per °C SST warming is referenced to the 1980 value, which may affect interpretation of percentage reductions. - The TC–SST relationship is complex, involving feedbacks (e.g., freshwater flux effects on salinity and stratification, wind-stress-induced cold wakes) not fully represented here. - Basin definitions and averaging over regions of heightened TC activity may mask subregional variability; ENSO classifications are annual and may not resolve intra-seasonal variability.