Abstract. The interannual variability of the greenhouse gases methane (CH4) and tropospheric ozone (O3) is largely driven by natural variations in global emissions and meteorology. The El Niño–Southern Oscillation (ENSO) is known to influence fire occurrence, wetland emission and atmospheric circulation, affecting sources and sinks of CH4 and tropospheric O3, but there are still important uncertainties associated with the exact mechanism and magnitude of this effect. Here we use a modelling approach to investigate how fires and meteorology control the interannual variability of global carbon monoxide (CO), CH4 and O3 concentrations, particularly during large El Niño events. Using a three-dimensional chemical transport model (TOMCAT) coupled to a sophisticated aerosol microphysics scheme (GLOMAP) we simulate changes to CO, hydroxyl radical (OH) and O3 for the period 1997–2014. We then use an offline radiative transfer model to quantify the climate impact of changes to atmospheric composition as a result of specific drivers. During the El Niño event of 1997–1998, there were increased emissions from biomass burning globally, causing global CO concentrations to increase by more than 40 %. This resulted in decreased global mass-weighted tropospheric OH concentrations of up to 9 % and a consequent 4 % increase in the CH4 atmospheric lifetime. The change in CH4 lifetime led to a 7.5 ppb yr−1 increase in the global mean CH4 growth rate in 1998. Therefore, biomass burning emission of CO could account for 72 % of the total effect of fire emissions on CH4 growth rate in 1998. Our simulations indicate that variations in fire emissions and meteorology associated with El Niño have opposing impacts on tropospheric O3 burden. El Niño-related changes in atmospheric transport and humidity decrease global tropospheric O3 concentrations leading to a −0.03 W m−2 change in the O3 radiative effect (RE). However, enhanced fire emission of precursors such as nitrogen oxides (NOx) and CO increase O3 and lead to an O3 RE of 0.03 W m−2. While globally the two mechanisms nearly cancel out, causing only a small change in global mean O3 RE, the regional changes are large – up to −0.33 W m−2 with potentially important consequences for atmospheric heating and dynamics.