Equatorward shift of the boreal summer intertropical convergence zone in Maritime Continent and the impacts on surface black carbon concentration and public health

The intertropical convergence zone (ITCZ) is a key regulator of tropical and global climate, governing deep-tropical atmospheric circulation and influencing higher latitudes via the Hadley circulation. Its location generally follows the energy flux equator and migrates seasonally from the Southern to the Northern Hemisphere. Understanding past variability and future changes of the seasonal-mean, zonal-mean ITCZ is critical because ITCZ shifts can reshape circulation patterns, redistribute air pollutants, and affect public health. While many studies suggest a poleward expansion of the tropics in annual means, seasonal-scale ITCZ dynamics remain uncertain, with potential confounding by natural variability and anthropogenic forcing. Zonal contrasts are evident: different basins exhibit distinct ITCZ responses under warming, and methodological choices (e.g., precipitation vs OLR vs vertical motion) add uncertainty to ITCZ diagnostics. The Maritime Continent (MC) is a major heat source for the global circulation and a hotspot of aerosol emissions from frequent wildfires and biomass burning. Black carbon (BC) poses notable health risks, yet comprehensive assessments of how ITCZ shifts modulate BC exposure and associated premature mortality in the MC are lacking. This study investigates long-term seasonal ITCZ shifts over the MC, attributes recent equatorial warming to anthropogenic influences, quantifies impacts on surface BC via circulation pathways, and assesses consequent public health effects.

Prior work highlights that annual-mean metrics can obscure seasonal ITCZ dynamics that better reflect the ascending branch of the Hadley circulation. Global projections often indicate poleward ITCZ shifts with Northern Hemisphere warming linked to drought and albedo changes, but observations and models reveal regionally contrasting shifts (e.g., northward over Eastern Africa and the Indian Ocean, southward over parts of the Pacific and Atlantic). Methodological differences in defining ITCZ (e.g., precipitation maxima, OLR minima, or vertical velocity peaks) contribute to discrepancies. Divergent atmospheric energy transport has been proposed as a zonal control on ITCZ latitude. The Maritime Continent, a strong convective heat source, has experienced frequent biomass burning with substantial carbon emissions (e.g., 1997 Indonesian fires contributed ~11% of global carbon emissions). BC is associated with adverse health outcomes under both short- and long-term exposures. Despite clear links between ITCZ position, circulation, and aerosols, a comprehensive evaluation of how seasonal ITCZ shifts affect BC distribution and health risks in the MC has been lacking. This study addresses these gaps by combining reanalyses and CMIP6 simulations to examine ITCZ shifts, their drivers, and consequences for BC and public health.

Study period and season: 1980–2014 focusing on boreal summer (June–September, JJAS). Region: Maritime Continent with analyses using zonal means across the MC longitudes (e.g., 90°E–160°E) and subregions (maritime area, the Philippines, continental area) as indicated in analysis boxes. ITCZ metrics: Using reanalyses (ERA5, MERRA-2, NCEP-DOE, ERA-Interim), three metrics were computed following Zhou et al.: (1) Location defined as the latitude of peak ascent at 500 hPa (ω@500) averaged over the MC longitude band; (2) Intensity as the peak ω@500 magnitude; (3) Width based on ω@500 < −15 hPa day−1. Because the northern boundary is ill-defined over the MC, width was computed as Width = Area/Length, where Area is the total area of grid cells with ω@500 < −15 hPa day−1 and Length is the zonal length (70° of longitude). Equatorial warming attribution: Equatorial band defined as 5°S–5°N over 90°E–160°E. Observed monthly SST anomalies (COBE) were compared with near-surface air temperature anomalies from 10 CMIP6 models under two historical scenarios: All-Hist (anthropogenic + natural forcings with observed SST/sea ice) and Nat-Hist (natural forcings only). Probability density functions (PDFs) of anomalies were compared to observations using the Kolmogorov–Smirnov test. Anthropogenic contribution is inferred when All-Hist matches observations (p < 0.05 difference vs obs indicating similarity in distribution, with Nat-Hist differing; as described) and Nat-Hist does not match observations (p > 0.05 threshold failing), consistent with the paper’s attribution protocol. BC and circulation datasets: Surface BC concentration from MERRA-2; low-level winds (e.g., 850 hPa) and vertical velocity from ERA5; precipitation from GPCP; additional BC column density diagnostics from MERRA-2. Anomalies were computed as departures from 1980–2014 monthly means. Composite and correlation analyses: Monthly anomalies were composited for periods of equatorward ITCZ shift defined by AITCZ < 0 (AITCZ is monthly ITCZ latitude anomaly). Extreme equatorward shift was defined as ΔITCZ < −2.1°, the recent-decade shift magnitude. Statistical significance assessed using two-tailed Student’s t-tests (e.g., for BC and wind anomalies) and correlation analyses between ΔITCZ, BC, and rainfall. Health impact assessment: Estimated all-age, all-cause premature mortalities attributable to BC using a log-linear concentration-response function (CRF). Relative risk (RR_k) at grid k: RR_k = exp(γ X_k), where X_k is monthly surface BC (MERRA-2) and γ is the risk coefficient from meta-regression (low 0.0023, median 0.0060, upper 0.0098). Excess cases E computed by summing across grids the population-weighted attributable fraction (AF_k = (RR_k − 1)/RR_k) times baseline mortality rate. Population (P_k) from WorldPop (100 m), aggregated to 0.5°×0.5° and interpolated to 0.5°×0.625°. Baseline all-age all-cause mortality rates from GBD (2010 and 2015); monthly rates derived by dividing annual rates by 12. Uncertainty: Monte Carlo with 10,000 γ samples assuming a triangular distribution (low/median/high), reporting median and 95% CI. Country-level and subregional (maritime, continental, Philippines) estimates were derived by spatial aggregation of gridded results.

- ITCZ shift and trends: The JJAS ITCZ over the Maritime Continent is climatologically near ~10°N. The ITCZ shifted equatorward by ~2.1° in 2005–2014 vs 1980–1989, significant across four reanalyses (p < 0.001). A significant linear trend of equatorward shift was detected from 1980–2014 at −0.08° per year (p < 0.05 in each dataset). ITCZ intensity increased by ~0.016 Pa s−1; width showed no significant change. - Attribution and SST: The probability of equatorial (5°S–5°N) warming SST anomalies increased roughly three-fold in the recent decade vs the historical decade. Warmer equatorial SST anomalies are significantly associated with equatorward ITCZ shifts. CMIP6 attribution (All-Hist vs Nat-Hist) indicates anthropogenic activities contributed to recent equatorial warming consistent with observations (KS tests; All-Hist matching observations but Nat-Hist not). - Circulation and BC responses (composites for AITCZ < 0): • Maritime area: Significant BC decrease up to ~74% (−9.88 μg m−3) over Central Sumatra and Borneo; enhanced updrafts and increased rainfall (+28 mm month−1) promoted vertical dispersion and wet deposition; BC column density decreased significantly. • Continental area: Surface BC increased up to +0.11 μg m−3 with suppressed updrafts and rainfall reduction (−20.1 mm month−1); no significant BC column density change, indicating local dispersion suppression rather than advection. • Philippines: BC decreases were significant mainly in the south, driven by anomalous low-level easterly/northeasterly winds weakening the summer monsoon transport from upwind sources; rainfall decreased but transboundary transport suppression dominated. - Quantitative ΔITCZ–BC relationships (scatter analyses): • Maritime: ΔITCZ < 0 associated with mean −0.13 μg m−3 BC; under extreme ΔITCZ < −2.1°, −0.21 μg m−3. • Philippines: ΔITCZ < 0 associated with −0.03 μg m−3 BC on average. • Continental: ΔITCZ < 0 associated with +0.05 μg m−3 BC; under extreme ΔITCZ < −2.1°, +0.10 μg m−3. - Health impacts (2010 baseline population and mortality rates): • Monthly mean BC-associated premature mortalities in JJAS: Indonesia 304 (95% CI: 289–319), Vietnam 98 (93–103), Thailand 57 (54–60). Maritime countries total: 355 (337–372); Continental countries total: 195 (185–205). • During equatorward ITCZ months, maritime countries experienced ~−13% change in BC-attributable premature mortalities (−46; 95% CI: −48 to −43), while continental countries saw ~+5.6% (+11; 95% CI: 11–12). Philippines changes were relatively small (~−4.2%). Overall across the MC: −6.1%. • Under extreme equatorward shift (ΔITCZ < −2.1°): maritime −14.1%; continental +6.7%; overall −6.6%. - Dynamical interpretation: Weakened East Asia Summer Monsoon during equatorward ITCZ periods is linked more to positional shifts of the Western North Pacific Subtropical High than to its strength, consistent with an equatorward shift of the East Asia summer jet and Meiyu–Baiu rainband in early summer.

The study demonstrates that anthropogenically driven equatorial warming has contributed to a significant equatorward shift of the seasonal JJAS ITCZ over the Maritime Continent, addressing a key question about regional ITCZ dynamics under climate change. This shift alters circulation patterns that modulate BC: enhanced ascent and rainfall over maritime regions promotes wet deposition and dispersion, lowering BC and associated health risks, whereas suppressed ascent over continental regions elevates surface BC and health risks. In the Philippines, the primary effect is dynamical suppression of transboundary transport due to anomalous low-level northeasterlies, offsetting local rainfall decreases. These regionally heterogeneous impacts underscore that climate-driven circulation changes can reallocate air pollution burdens and associated health risks even in seasons or regions not dominated by biomass burning. The results also clarify that seasonal ITCZ shifts do not necessarily equate to tropical expansion; they reflect changes in the ascending branch of the Hadley cell, whereas tropical belt width concerns the descending branches. The findings suggest that in the MC, anthropogenic equatorial warming can outweigh PDO-related influences, favoring continued equatorward seasonal ITCZ displacement, with implications for future air quality management and public health planning.

Using multiple reanalyses (1980–2014), the study identifies a significant equatorward shift (~2.1°) and intensification (~0.016 Pa s−1) of the JJAS ITCZ over the Maritime Continent, attributable in part to anthropogenic equatorial warming supported by CMIP6 attribution analyses. Composite diagnostics link this shift to reduced surface BC in maritime regions via enhanced updrafts and wet deposition, increased BC over continental regions via suppressed vertical dispersion, and reduced transboundary BC transport to the southern Philippines via weakened summer circulation. Health assessments indicate notable regional disparities: maritime countries experience ~13–14% reductions in BC-related premature mortalities during equatorward ITCZ periods, while continental countries see ~6–7% increases; overall, a net regional reduction of ~6–7% is found in JJAS. These findings highlight the need to integrate climate-induced circulation changes into air quality and public health strategies. Future work should incorporate higher-resolution and bias-corrected precipitation datasets, explicitly model aerosol–cloud–precipitation processes and daily-scale BC chemistry/aging, and apply numerical simulations to quantify mechanisms and reduce uncertainties.

- ITCZ width over the MC is difficult to define; different metrics can yield differing trends, and the seasonal ITCZ shift does not directly represent tropical expansion. - Precipitation datasets used for long-term detection are relatively coarse (e.g., GPCP), potentially limiting sensitivity to shifts near model resolution. - Climate model biases (e.g., double-ITCZ in CMIP5/6) require further correction to provide robust long-term precipitation and ITCZ diagnostics. - The statistical approach did not explicitly resolve daily-scale BC processes (e.g., aging, local wet deposition) and their perturbations; these may introduce uncertainties. - While monthly-scale composites/correlations were used to mitigate confounding by interannual/decadal variability, comprehensive quantitative mechanism assessments via numerical simulations are still needed. - Health analysis focused on BC; other co-emitted species (PM2.5, OC, sulfate) may contribute additional health risks not quantified here.