Trends in the nonvolcanic component of stratospheric aerosol over the period 1971–2004

The six longest records of stratospheric aerosol (in situ measurements at Laramie, Wyoming, lidar records at: Garmisch‐Partenkirchen, Germany; Hampton, Virginia; Mauna Loa, Hawaii; São José dos Campos, Brazil, and SAGE II measurements) were investigated for trend by (1) comparing measurements in the 3 volcanically quiescent periods since 1970 using standard analysis of variance techniques, and (2) analyzing residuals from a time/volcano dependent empirical model applied to entire data sets. A standard squared‐error residual minimization technique was used to estimate optimum parameters for each measurement set, allowing for first order autocorrelation, which increases standard errors of trends but does not change magnitude. Analysis of variance over the 3 volcanically quiescent periods is controlled by the end points (pre‐El Chichón and post‐Pinatubo), and indicates either no change (Garmisch, Hampton, São José dos Campos, Laramie‐0.15 μm) or a slight, statistically insignificant, decrease (Mauna Loa, Laramie‐0.25 μm), −1 ± 0.5% yr −1 . The empirical model was applied to the same records plus 1020 nm SAGE II data separated into 33 latitude/altitude bins. No trend in stratospheric aerosol was apparent for 31 of 33 SAGE II data sets, 3 of 4 lidar records, and in situ measurements at 0.15 μm. For Hampton and Laramie‐0.25 μm, the results suggest a weak negative trend, −2 ± 0.5% yr −1 , while 2 SAGE II data sets (30–35 km, 30° and 40°N) suggest a positive trend of similar magnitude. Overall we conclude that no long‐term change in background stratospheric aerosol has occurred over the period 1970–2004.