SAGE II aerosol data validation and initial data use: An introduction and overview

Abstract

Correlative sensors used in the SAGE II aerosol data validation program included two‐ and six‐channel dustsondes; ground‐based and airborne multiwavelength lidars; wire, ribbon, and related impactors; laser spectrometers; filter collectors; spectral polarimeters; and spectral limb cameras. Measurement campaigns were conducted in the northern, southern, eastern, and western hemispheres, and in low, middle, and high latitudes. Measurement conditions were often strongly perturbed by El Chichón and other volcanic eruptions, which posed new challenges because of (1) multimodal size distributions, (2) spatiotemporal inhomogeneity, and (3) large extinction values. New experimental and analytical approaches were developed to meet these challenges. Several of the analytical approaches use SAGE II four‐wavelength aerosol extinction values to derive or check particle size distributions. Extensive comparisons between SAM II and SAGE II were also conducted. The validation results are detailed in the following six papers. They show that when correlative measurements are nearly coincident with the SAGE II measurements in space and time, the SAGE II and correlative values are consistent within their respective uncertainties. This is especially so of the SAGE II 1.02‐μm extinction profiles, for which numerous instances of good coincidence show agreement from heights near the tropopause to ∼28 km. For the shorter wavelengths (0.525, 0.453, and 0.385 μm), several comparisons show agreement within mutual error bars at heights between ∼15 and 26 km. However, above ∼24 km, correlative observations are scarce and the data is of poorer quality because of low aerosol content and relatively strong Rayleigh, NO2, and ozone contributions. Thus for the short wavelengths, especially above ∼24 km, further correlative measurements would be desirable to provide a conclusive validation. Initial data use investigations demonstrate methods for inferring aerosol physical and optical properties from the SAGE II four‐wavelength extinction data. The methods yield values and uncertainties for particle volume, area, effective size, distribution width, model size distribution parameters, number larger than specified sizes, and backscatter coefficient. In many cases, SAGE II‐derived values are compared to measurements by dustsonde and lidar. Results vary, but they are usually consistent within mutual error bars.