Increase in Cape Verde hurricanes during Atlantic Niño

Seasonal outlooks for Atlantic tropical cyclone (TC) activity typically rely on known relationships between ocean–atmosphere states, especially ENSO and the Atlantic Meridional Mode (AMM), which modulate atmospheric static stability and vertical wind shear over the main development region. The West African summer monsoon and associated African Easterly Waves (AEWs) also influence Atlantic TC genesis. Yet, ENSO and AMM tend to peak in boreal winter–spring and weaken during the hurricane season (June–November). The leading SST variability mode during the season itself is Atlantic Niño/Niña (warm/cold anomalies in the eastern equatorial Atlantic), but its connection to Atlantic hurricane activity has not been systematically examined. Motivated by case seasons (1992, 2003, 2008) and established links between Atlantic Niño/Niña, the ITCZ, and West African rainfall, the study hypothesizes that Atlantic Niño/Niña modulates Atlantic TC activity via impacts on deep-tropical convection, regional circulations, and AEW activity, particularly affecting Cape Verde hurricanes. The objective is to quantify these links and elucidate mechanisms using observations and reanalyses.

Prior work shows ENSO and AMM strongly modulate Atlantic TC activity by altering vertical wind shear and static stability over the MDR: La Niña and positive AMM enhance genesis, while El Niño and negative AMM suppress it. Atlantic TC activity is also tied to the West African monsoon, which promotes AEWs and the West African westerly jet, generating low-level cyclonic vorticity favorable for deep-tropical cyclogenesis. Atlantic Niño/Niña modulates the ITCZ and West African rainfall, with enhanced ITCZ/rainfall during Atlantic Niño and suppression during Atlantic Niña. Some Atlantic Niño/Niña events relate to AMM in boreal spring and can interact with ENSO across seasons, but during JJASON the ATL3 index is largely uncorrelated with AMM (r = −0.03) and only weakly, insignificantly correlated with NIÑO3.4 (r = −0.15), motivating a separate assessment of its role.

Data and period: Atlantic TC genesis and track density derived from IBTrACS, considering only systems with wind ≥34 kt (tropical depressions excluded). Atmospheric circulation from NCEP/NCAR Reanalysis v1 (NCEP1). Monthly precipitation from NOAA’s Gridded Precipitation Reconstruction dataset. SST from HadISST1. Analysis focuses on JJASON (June–November) for 1948–2021 (74 years). All data are linearly detrended to minimize anthropogenic influence. Robustness is assessed for the post-satellite era (1979–2021).

Event definitions: Atlantic Niño/Niña identified when area-mean SST anomalies over ATL3 (3°S–3°N, 20°W–0°) exceed +1σ or are below −1σ during JJASON (σ = 0.32 K), yielding 19 Atlantic Niño and 14 Atlantic Niña years. ENSO events defined similarly using NIÑO3.4 (5°S–5°N, 170°E–120°W; σ = 0.50 K), yielding 17 El Niño and 21 La Niña events. Positive/negative AMM years identified when normalized AMM index exceeds ±1σ during JJASON (15 positive, 18 negative years).

Partial regression framework: To separate impacts of Atlantic Niño/Niña and ENSO on TC activity, partial regressions of TC genesis and track density onto normalized ATL3 and NIÑO3.4 indices were computed: TC_ENSO_ATL3(t,x,y) = β1(x,y)·NIÑO3.4(t) + β2(x,y)·ATL3(t) TC_ENSO(t,x,y) = β1(x,y)·NIÑO3.4(t) TC_ATL3(t,x,y) = β2(x,y)·ATL3(t) These reconstructions are correlated with observed fields to assess skill. Statistical significance is evaluated with two-tailed Student’s t-tests.

Diagnostics: AEW activity measured by eddy kinetic energy (EKE) at 700 hPa, using a wavenumber–frequency filter retaining 2–10 day periods and zonal wavenumbers corresponding to 10°–40° longitude (∼1000–4000 km). EKE = 0.5·(u'^2 + v'^2), where u' and v' are eddy components. Low-level relative vorticity at 850 hPa (RV850) and vertical wind shear (850–200 hPa) are regressed onto indices. Spatial masks applied over the Americas/Pacific for certain fields. Composite analyses (Atlantic Niño/neutral/Niña) complement regressions. Track density and genesis fields are spatially smoothed for visualization.

• Composite analysis (1948–2021) shows significantly higher TC genesis over the tropical North Atlantic during Atlantic Niño and lower during Atlantic Niña. Over 60°W–10°W, 5°N–20°N, mean TCs per season are 5.1 ± 0.58 (Atlantic Niño) vs 2.9 ± 0.61 (Atlantic Niña), significant at 95% by two-tailed t-test.
• Partial regressions indicate Atlantic Niño increases TC genesis almost exclusively in the eastern tropical North Atlantic (east of 40°W) and raises track density toward the Caribbean islands and around Florida. In contrast, La Niña increases genesis in the southern Gulf of Mexico and Caribbean Sea, with enhanced track density across Central America, the Caribbean, and the southern/southeastern U.S.
• Mechanisms during Atlantic Niño: strengthened ITCZ around 0°–10°N and 30°W–10°E enhances sub-Sahel rainfall and AEW activity; low-level westerly anomalies strengthen the West African westerly jet, increasing positive low-level relative vorticity between 5°–15°N; vertical wind shear is reduced west of 40°W—conditions conducive to Cape Verde hurricane formation and intensification.
• Major hurricanes (Category 3–5) over 60°W–10°W, 5°N–20°N are significantly more frequent during Atlantic Niño (2.16 ± 0.56 per year) than during Atlantic Niña (0.92 ± 0.51 per year). The contrast is stronger in the post-satellite period (Atlantic Niño: 2.5 ± 0.71; Atlantic Niña: 0.2 ± 0.57 per year).
• ENSO phases show no statistically significant difference in major hurricane counts over the same region (El Niño: 1.22 ± 0.52; La Niña: 1.85 ± 0.58 per year).
• Reconstruction skill: NIÑO3.4-based reconstruction captures major hurricane genesis in the Caribbean and central tropical North Atlantic but poorly in the deep eastern tropical North Atlantic. ATL3-based reconstruction better represents eastern deep-tropical major genesis; combining NIÑO3.4 and ATL3 provides the best overall representation.
• Independence: During JJASON, ATL3 is nearly uncorrelated with AMM (r = −0.03) and has an insignificant correlation with NIÑO3.4 (r = −0.15), supporting its distinct role during the hurricane season.
• Case seasons: 2003 and 2008 (Atlantic Niño with neutral ENSO/AMM) were active with multiple major hurricanes, including Hurricanes Isabel (2003) and Ike (2008). In contrast, 1992 (Atlantic Niña with neutral ENSO/AMM) was notably quiet, suggesting a link between Atlantic Niño/Niña and TC activity.

The results demonstrate that Atlantic Niño/Niña, a leading in-season SST variability mode, exerts a significant control on the location and likelihood of TC genesis in the deep tropics near the Cape Verde islands, thereby modulating the probability of major hurricanes. Strengthening of the ITCZ and sub-Sahel rainfall during Atlantic Niño enhances AEWs, increases low-level cyclonic vorticity, and reduces vertical wind shear in key regions, creating a favorable environment for Cape Verde hurricanes that can traverse long warm-water tracks and intensify before approaching the Caribbean and the U.S. This addresses the previously unexplored linkage between Atlantic Niño/Niña and Atlantic hurricane activity, particularly for major hurricanes. While Atlantic Niño/Niña modulates genesis locations and tracks, overall seasonal Atlantic TC activity remains predominantly governed by ENSO and AMM. Interactions among these modes matter: Atlantic TC activity can be amplified or suppressed when Atlantic Niño/Niña and ENSO or AMM are out of phase or in phase, respectively. Because Atlantic Niño/Niña is statistically independent from ENSO and AMM during JJASON and is potentially predictable, incorporating it alongside ENSO and AMM can improve seasonal outlooks, especially when ENSO and AMM are near neutral.

This study establishes Atlantic Niño/Niña as a key in-season driver of Cape Verde hurricane genesis and a significant modulator of major hurricane occurrence in the tropical North Atlantic. Through observational analyses and partial regressions separating ENSO effects, the work shows that Atlantic Niño strengthens the ITCZ, AEWs, low-level vorticity, and reduces shear, leading to enhanced TC genesis in the eastern deep tropics and more major hurricanes. Although ENSO and AMM still dominate overall basinwide seasonal activity, Atlantic Niño/Niña provides complementary predictive information and should be considered as an additional predictor to refine seasonal Atlantic hurricane outlooks, particularly when ENSO and AMM are neutral.

The relationship between Atlantic Niño/Niña and major hurricanes is stronger in the post-satellite era, suggesting undercounts in pre-satellite IBTrACS records and potential observational inhomogeneities. The study is based on observational composites and partial regressions; while robust across the full and post-satellite periods, residual uncertainties remain, and overall seasonal activity continues to be influenced predominantly by ENSO and AMM.