Tropical storms trigger phytoplankton blooms in the deserts of north Indian Ocean

The study investigates how tropical cyclones (TCs) modulate ocean productivity in the North Indian Ocean (NIO) through cyclone-induced phytoplankton blooms, quantified via chlorophyll-a (Chl-a). TCs form over warm waters with low wind shear, and the NIO (Bay of Bengal, BoB, and Arabian Sea, AS) is a highly favorable region for such activity. Despite similarities, BoB and AS differ in freshwater input, salinity, stratification, and monsoon wind reversals, which shape their biogeochemistry. BoB is typically low in productivity due to strong stratification from river runoff and rainfall, while AS is comparatively productive. Cyclone passage can induce upwelling and entrainment that inject nutrients into the euphotic zone, potentially enhancing Chl-a. Prior work has largely focused on individual storms, leaving gaps in understanding basin-wide, multi-decadal patterns, drivers (translational speed, eddies, barrier layer), seasonal contrasts, and the roles of ENSO and IOD. This study addresses these gaps using long-term satellite observations to assess frequency, timing, magnitude, and mechanisms of cyclone-induced blooms across BoB and AS.

Previous studies documented Chl-a enhancement associated with specific cyclones using various sensors (SeaWiFS, MODIS, IRS-P4 OCM, OCM) and models, often attributing blooms to wind-induced upwelling and entrainment. Examples include Orissa Super Cyclone (1999) off the coast (Nayak et al.), BOB05/BOB06 (1999–2000) contrasts (Patra et al.), Thane (2011) and Phailin (2013) contrasting responses (Vidya et al.), Hudhud (2014) post-cyclone doubling (Chacko), and ARB01 (2001), Gonu (2007), Phyan (2009) in AS (Subrahmanyam et al.; Chakraborty et al.). Studies highlighted the importance of cyclone characteristics (intensity, TS) and ocean pre-conditions (stratification, barrier layer, eddies) in determining bloom responses. Work has also examined ENSO/IOD influences on cyclone activity in the NIO. However, comprehensive analyses encompassing all cyclones over the satellite era across both basins, seasons, and remote forcing phases remained limited, motivating the present systematic, multi-decadal assessment.

Study domain: NIO (3°–25° N, 50°–98° E), analyzed separately for BoB (7°–23° N, 78°–98° E; subdivided by landfall quadrants) and AS (3°–25° N, 50°–78° E). Period: 1997–2019. Data: Satellite ocean colour OC-CCI v4.2 (merged SeaWiFS, MERIS, MODIS-Aqua) at ~4 km resolution with ±0.15 mg/m³ uncertainty; cyclone best tracks from JTWC (6-hourly) and IMD (2019); ECMWF reanalysis 10 m winds (ERA5 daily, ~12.5–25 km) for Ekman Pumping Velocity (EPV); Copernicus CMEMS sea level anomaly/sea surface height anomaly (SLA/SSHA, ~25 km) for eddy context; Bio-Argo floats (WMO IDs 2902086, 2902114, 2902120) for subsurface Chl-a and hydrography (MLD, ILD, BLT) at 5-day intervals. Definitions and processing: For each cyclone, Chl-a is averaged over four windows: 5 days before, during, 0–5 days after, and 6–10 days after passage. Open-ocean background assessed to set thresholds: pre-cyclone background ~0.2 mg/m³; bloom defined as Chl-a > 0.5 mg/m³. For each event, a 4°×4° region along track where post-cyclone Chl-a > 0.2 mg/m³ is selected; within it, peak bloom is the average over a 1°×1° box. Time lag of peak bloom determined from the temporal sequence. Diagnostics: EPV computed at track points at peak intensity from ERA5 winds; positive EPV indicates upwelling. Translational speed (TS) computed between 6-hourly track points using the Haversine formula (distance divided by time; m s⁻¹). Eddy context inferred from 7–10 day composites of SSHA before and during cyclone passage; negative SSHA denotes cold-core (cyclonic) eddies; positive SSHA warm-core (anticyclonic) eddies. ENSO and IOD phase classification based on Ocean Niño Index and Dipole Mode Index on dates after cyclone passage. Change metric: Percentage bloom change = (peak post-cyclone Chl-a − pre-cyclone Chl-a) / pre-cyclone Chl-a × 100, with additional comparisons relative to fixed baselines (0.2 and 0.5 mg/m³). Bio-Argo analyses: Time–depth sections (to 200 m) of temperature and Chl-a with MLD, ILD, BLT, and 23°C isotherm (D23) to examine subsurface responses and eddy–storm interactions near float locations.

- Frequency and detection: In BoB, 30 of 51 storms produced blooms (Chl-a > 0.2 mg/m³) across seasons; in AS, 18 blooms out of 33 cyclones during 1997–2019. - Magnitudes: Average bloom-associated Chl-a ≈ 1.65 mg/m³. Increases typically 20–3000% relative to open-ocean or pre-cyclone levels, depending on event and baseline; AS percentage changes about five times larger than BoB when referenced to basin averages. - Exemplars: Pre-monsoon BoB largest bloom with BOB01 (2004) ≈ 3.28 mg/m³; post-monsoon BoB notable blooms include Madi (2013) ≈ 2.8 mg/m³ and Vardah (2016) ≈ 1.92 mg/m³; AS highest bloom with Gonu (2007) ≈ 11.06 mg/m³. - Translational speed (TS): Bloom amplitude inversely related to TS (correlation r ≈ −0.30, significant at 95%). Slow-moving storms spend more time over the ocean, enhancing upwelling and bloom magnitude. - Upwelling (EPV): Higher EPV aligns with stronger blooms (e.g., BOB01 2003 EPV ≈ 1.8×10⁻⁴ m/s with notable bloom; Mala 2006 EPV ≈ 1×10⁻⁴ m/s with ~1.8–1.87 mg/m³). Wind forcing and EPV are key drivers. - Eddies: Cold-core (negative SSHA) eddies along tracks enhance and sustain blooms via raised thermocline/nutricline and mixing; warm-core eddies suppress mixing and diminish response. Bio-Argo and SSHA composites corroborate contrasting responses (e.g., Hudhud near cold-core vs Vardah near warm-core). - Seasonality and duration: Post-monsoon exhibits higher cyclone frequency and longer bloom duration (≈10–14 days). Pre-monsoon blooms are spatially smaller and often subside within ~10 days. - Event sequencing: Consecutive storms over the same region can compound blooms (e.g., Lehar preceding Madi; Nada preceding Vardah). - Intensity category: Most blooms are associated with tropical storms and Category 1 cyclones. In BoB, 80% of pre-monsoon and 80% of post-monsoon bloom events are TS/Cat-1; in AS, 8 of 18 (44.4%) are TS/Cat-1. Sustained low-to-moderate winds with slow TS can yield intense blooms. - Timing: Typical time lag for peak bloom is 4–12 days after cyclone passage. - ENSO/IOD impacts: BoB blooms larger during La Niña than El Niño or normal years; PIOD years show higher bloom amplitudes in BoB, while NIOD years show higher amplitudes in AS. In AS, more cyclones occur in El Niño and PIOD years, but bloom magnitudes are generally higher in La Niña/NIOD years. - Barrier layer effects: Northern BoB barrier layer favors cyclone intensification but can inhibit mixing-driven cooling; absence/shallow BL in south BoB favors upwelling and blooms. - Percent-change extremes: Using a 0.2 mg/m³ basin baseline, AS Gonu exhibits ≈6000% change (≈3500% relative to 0.5 mg/m³ threshold); BoB examples include BOB01 (2003) up to ≈1385% increase.

The findings demonstrate that cyclone-induced mixing and upwelling are effective mechanisms for enhancing surface Chl-a in the NIO, particularly when storms move slowly and interact with cold-core eddies that shoal the thermocline and nutricline. The observed inverse relationship between translational speed and bloom magnitude, and the key role of EPV, directly address the study’s objective to identify physical controls on bloom intensity and timing. Seasonal contrasts reflect monsoon-driven stratification and barrier layer dynamics, with post-monsoon storms generating longer-lived and often broader blooms, while pre-monsoon responses are more localized and shorter-lived. Interannual variability linked to ENSO and IOD modulates both cyclone occurrence and biological response: La Niña/NIOD conditions favor stronger blooms, consistent with changes in background ocean state and eddy fields. These results are relevant for understanding the ocean’s role in the carbon cycle, as cyclone-induced primary production can represent significant and rapid perturbations in open-ocean “desert” regions. The documented time lags, durations, and sensitivities to TS and eddy interactions provide actionable constraints for ocean–biogeochemical models, which often under-represent post-storm productivity. The implications extend to climate-change contexts, including a reported global slowdown in cyclone translational speeds, which could amplify biological responses in future climates.

This work provides a comprehensive, two-decade, basin-wide assessment of cyclone-induced phytoplankton blooms in the North Indian Ocean, quantifying their occurrence, timing, magnitude, and controls. Key contributions include: (1) establishing the inverse dependence of bloom amplitude on cyclone translational speed; (2) highlighting the amplifying role of cold-core eddies and EPV-driven upwelling; (3) delineating seasonal contrasts and barrier-layer influences; and (4) clarifying the modulation by ENSO and IOD phases, with stronger blooms in La Niña and PIOD (BoB) or NIOD (AS) years. The study underscores that tropical storms and Category 1 cyclones, particularly slow-moving ones, often produce the strongest blooms, with typical lags of 4–12 days and durations up to ~10–14 days post-passage. Future research could focus on: improved eddy detection and quantification of eddy–cyclone synergy; assimilation of Bio-Argo and high-frequency satellite data to resolve rapid bloom evolution; refined attribution of carbon export following storm-induced production; and coupled ocean–atmosphere–biogeochemical modeling to project responses under changing cyclone characteristics and climate regimes.

- Satellite ocean colour data are sparse during cyclone passage due to clouds, high winds, and heavy rainfall, potentially under-sampling peak or immediate post-storm signals. - Eddy identification relied on SSHA composites without a dedicated eddy-tracking algorithm, introducing uncertainty in eddy characterization. - Limited number of events in some ENSO/IOD categories (e.g., few El Niño-period storms in BoB) constrain statistical robustness; conclusions are conditioned on available cases. - Use of fixed spatial (1°×1°, 4°×4°) and temporal (5-day) averaging windows may smooth fine-scale variability and peak magnitudes. - Track datasets differ for 2019 (IMD vs JTWC), and EPV estimates depend on reanalysis winds and parameter choices; OC-CCI Chl-a carries an uncertainty of ±0.15 mg/m³.