Abstract. Aerosol removal processes control global aerosol abundance, but the rate of that removal remains uncertain. A recent study of aerosol-bound radionuclide measurements after the Fukushima Daiichi nuclear power plant accident documents ¹³⁷Cs removal (e-folding) times of 10.0–13.9 days, suggesting that mean aerosol lifetimes in the range of 3–7 days in global models might be too short by a factor of two. In this study, we attribute this discrepancy to differences between the e-folding and mean aerosol lifetimes. We implement a simulation of ¹³⁷Cs and ¹³³Xe into the GEOS-Chem chemical transport model and examine the removal rates for the Fukushima case. We find a general consistency between modelled and measured e-folding times. The simulated ¹³⁷Cs global burden e-folding time is about 14 days. However, the simulated mean lifetime of aerosol-bound ¹³⁷Cs over a 6-month post-accident period is only 1.8 days. We find that the mean lifetime depends strongly on the removal rates in the first few days after emissions, before the aerosols leave the boundary layer and are transported to altitudes and latitudes where lifetimes with respect to wet removal are longer by a few orders of magnitude.

We present sensitivity simulations that demonstrate the influence of differences in altitude and location of the radionuclides on the mean lifetime. Global mean lifetimes are shown to strongly depend on the altitude of injection. The global mean ¹³⁷Cs lifetime is more than one order of magnitude greater for the injection at 7 km than into the boundary layer above the Fukushima site. Instantaneous removal rates are slower during the first few days after the emissions for a free tropospheric versus boundary layer injection and this strongly controls the mean lifetimes. Global mean aerosol lifetimes for the GEOS-Chem model are 3–6 days, which is longer than that for the ¹³⁷Cs injected at the Fukushima site (likely due to precipitation shortly after Fukushima emissions), but similar to the mean lifetime of 3.9 days for the ¹³⁷Cs emissions injected with a uniform spread through the model's Northern Hemisphere boundary layer. Simulated e-folding times were insensitive to emission parameters (altitude, location, and time), suggesting that these measurement-based e-folding times provide arobust constraint on simulated e-folding times.

Despite the reasonable global mean agreement of GEOS-Chem with measurement e-folding times, site by site comparisons yield differences of up to a factor of two, which suggest possible deficiencies in either the model transport, removal processes or the representation of ¹³⁷Cs removal, particularly in the tropics and at high latitudes. There is an ongoing need to develop constraints on aerosol lifetimes, but these measurement-based constraints must be carefully interpreted given the sensitivity of mean lifetimes and e-folding times to both mixing and removal processes.