Open AccessInfluence of nonlocal chemistry on tracer distributions: Inferring the mean age of air from SF 6
Author(s) -
Hall Timothy M.,
Waugh Darryn W.
Publication year - 1998
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.67
H-Index - 298
eISSN - 2156-2202
pISSN - 0148-0227
DOI - 10.1029/98jd00170
Subject(s) - troposphere , stratosphere , atmospheric sciences , tracer , atmosphere (unit) , environmental science , climatology , range (aeronautics) , mixing ratio , meteorology , geology , materials science , physics , nuclear physics , composite material
Sulfur hexafluoride (SF 6 ) is nearly inert in the troposphere and stratosphere and has a documented, steady increase in the troposphere due to industrial sources, making it a useful tracer of atmospheric circulation. Studies using SF 6 to estimate the mean age of stratospheric air have assumed the influence of mesospheric photochemical destruction is negligible. However, the mean age of an air parcel may be sensitive to small fractions of the air that have resided for long times in the upper atmosphere. Here we use two three‐dimensional chemical transport models to estimate the influence of mesospheric SF 6 loss on mean age inferences in the stratosphere. Because the mechanisms of SF 6 loss are uncertain, we perform a number of simulations employing a range of magnitudes of a simple constant mesospheric loss frequency, as well as a range of scenarios for SF 6 time variation in the troposphere. Using loss rates producing plausible global lifetimes of SF 6 (1000 to 3000 years), and tropospheric time variation matching observations, we find that age estimates inferred from SF 6 mixing ratios may be significantly biased and that the bias is increasing in time. For example, the present‐day time lag of SF 6 mixing ratio from the troposphere overestimates the mean age by up to 18% at 68°S and 20 km, and up to 65% at 68°S and 30 km, depending on the loss rate and model. Correcting for this bias would bring recent comparisons of modeled and measured mean age closer in line.