GLOBAL TO MICROSCALE EVOLUTION OF THE PINATUBO VOLCANIC AEROSOL DERIVED FROM DIVERSE MEASUREMENTS AND ANALYSES

We assemble data on the Pinatubo aerosol from space, air, and ground m easurements, develop a composite picture, and assess the consistency a nd uncertainties of measurement and retrieval techniques. Satellite in frared spectroscopy, particle morphology, and evaporation temperature measurements agree with theoretical calculations in showing a dominant composition of H2SO4-H2O mixture, with H2SO4 weight fraction of 65-80 % for most stratospheric temperatures and humidities. Important except ions are (i) volcanic ash, present at all heights initially and just a bove the tropopause until at least March 1992, and (2) much smaller H2 SO4 fractions at the low temperatures of high-latitude winters and the tropical tropopause. Laboratory spectroscopy and calculations yield w avelength- and temperature-dependent refractive indices for the H2SO4- H2O droplets. These permit derivation of particle size information fro m measured optical depth spectra, for comparison to impactor and optic al-counter measurements. All three techniques paint a generally consis tent picture of the evolution of R(eff), the effective radius. In the first month after the eruption, although particle numbers increased gr eatly, R(eff) outside the tropical core was similar to preeruption val ues of similar to 0.1 to 0.2 mu m, because numbers of both small (r<0. 2 mu m) and large (r>0.6 mu m) particles increased. In the next 3-6 mo nths, extracore R(eff) increased to similar to 0.5 mu m, reflecting pa rticle growth through condensation and coagulation. Most data show tha t R(eff) continued to increase for similar to 1 year after the eruptio n. R(eff) values up to 0.6-0.8 mu m or more are consistent with 0.38-1 mu m optical depth spectra in middle to late 1992 and even later. How ever, in this period, values from in situ measurements are somewhat le ss. The difference might reflect in situ undersampling of the very few largest particles, insensitivity of optical depth spectra to the smal lest particles, or the inability of flat spectra to place an upper lim it on particle size. Optical depth spectra extending to wavelengths la mbda>1 mu m are required to better constrain R(eff), especially for R( eff)>0.4 mu m. Extinction spectra computed from in situ size distribut ions are consistent with optical depth measurements; both show initial spectra with lambda(max)less than or equal to 0.42 mu m, thereafter i ncreasing to 0.78 less than or equal to lambda(max)less than or equal to 1 mu m. Not until 1993 do spectra begin to show a clear return to t he preeruption signature of lambda(max)less than or equal to 0.42 mu m . The twin signatures of large R(eff) (>0.3 mu m) and relatively flat extinction spectra (0.4-1 mu m) are among the longest-lived indicators of Pinatubo volcanic influence. They persist for years after the peak s in number, mass, surface area, and optical depth at all wavelengths less than or equal to 1 mu m. This coupled evolution in particle size distribution and optical depth spectra helps explain the relationship between global maps of 0.5- and 1.0-mu m optical depth derived from th e Advanced Very High Resolution Radiometer (AVHRR) and Stratospheric A erosol and Gas Experiment (SAGE) satellite sensors. However, there are important differences between the AVHRR and SAGE midvisible optical t hickness products. We discuss possible reasons for these differences a nd how they might be resolved.