Region-specific time lags in the response of black carbon to ENSO from 46 years of MERRA-2 reanalysis

Abstract. The relationship between El Niño-Southern Oscillation (ENSO) and global black carbon (BC) loading is not described by a single timescale. Using 46 years (1980 to 2025) of MERRA-2 reanalysis, we regress BC column mass anomalies against the Oceanic Niño Index at lags of 0 to 12 months at every grid point, accounting for autocorrelation and applying false discovery rate (FDR) correction. BC responds to ENSO on region-specific lags: near zero over the Maritime Continent, about five months over Central America, and seven to eight months over the Amazon. These lags are reproduced in 1980 to 2009 and 2010 to 2025, periods driven by different biomass burning emission inventories, so the timing is not an artifact of the inventory or of the 1997 to 1998 El Niño, and it is consistent with the pan-tropical cascade of fire anomalies documented in satellite fire records. No relationship is detectable over the continental United States, central Africa, or Siberia in either period; over the continental United States, the largest full-record median correlation (0.03) is about one fifth of the smallest detectable correlation, partly because weak correlations of opposite sign in the two periods cancel. The fraction of grid points passing FDR correction changes abruptly between adjacent lags while correlation strength varies smoothly, so response timing is better read from correlation magnitude than from significance counts. Anticipating ENSO-driven aerosol loading therefore requires resolving when the relationship emerges in each region, not only whether it exists.