Developing an Ensemble-Based Aerosol Assimilation System with JEDI

An assimilation system based on hybrid ensemble-variational approach is being developed at
NOAA/ESRL and NCEP to issue global aerosol forecasts. This system is currently capable of assimilating the most common globally available observations of aerosol optical depths (AOD). The components of the system are:
1) forward model and its adjoint from the Community Radiative Transfer Model (CRTM);
2) variational solvers within JEDI framework;
3) an adaptation of the operational EnKF code to the native cube-sphere grid (until an ensemble filter within JEDI becomes available);
4) FV3-Chem model based on GOCART (Goddard Chemistry Aerosol Radiation and Transport) parameterization coupled via NUOPC framework with a meteorological counterpart of GFS with the new FV3 dynamical core.

Initial meteorology is obtained from operational ensemble analyses and AOD retrievals are derived from measurements of Visible Infrared Imaging Radiometer Suite (VIIRS) instrument.

Difficulty of developing an ensemble-based assimilation system for atmospheric aerosols, and also many other chemical constituents, relies on devising perturbations that reflect large uncertainty of emissions sources as perturbed meteorology alone does not allow sufficient spread across the members. We are currently exploring perturbation strategies to improve functioning of the system.

We will also present plans to extend the current capabilities to encourage a discussion and a critique.