1D-VAR Retrieval of Aerosol Properties from Satellite Spectral Lidar and Radiometric Observations
Abstract
Aerosols play an important role in atmospheric physics and chemistry through their impact on air pollution, actinic fluxes, visibility, acid rain, and climate. Numerous atmospheric models at the urban, regional or global scale include a representation of aerosols aimed at simulating their physical and chemical properties such as their concentration, size distribution, chemical composition and state of mixture. Unfortunately these models are not well constrained by observations, which limits the possibility of accurate operational forecasts of aerosol properties at these different scales. Important progresses have been made in the observation of aerosols (in clear-sky) from space. One can measure with a good accuracy the aerosol optical depth and the Ångström exponent (a signature of aerosol size) from radiometric measurements in the visible over the ocean and to some extent over land. Active remote-sensing by lidar offer an opportunity to measure the aerosol vertical profile. In this work we use a simple 1D radiative transfer model and its adjoint to study the retrieval of the vertical profile of aerosol properties from variational assimilation of spectral lidar and radiometric measurements. The cost function is minimized using a BFGS algorithm forcing the model towards the observations. Results of this one dimensional variational assimilation (1D-VAR) scheme with synthetic satellite measurements will be presented and analyzed. In particular we will focus on the retrieval of aerosol properties such as the extinction coefficient vertical profile and size distribution of the accumulation and coarse modes. Colocated lidar and radiance data from the GLAS and MODIS instruments will be used to test the algorithm.
- Publication:
-
35th COSPAR Scientific Assembly
- Pub Date:
- 2004
- Bibcode:
- 2004cosp...35.2263H