Variational Single Site Data Assimilation for ARM

Sponsor: DOE Atmospheric Radiation Measurement Program

Abstract:

The main purpose of this project is to provide a continuous analysis of the atmospheric structure over the ARM sites, at a scale comparable to that of typical climate models. This analysis is needed to initialize and validate single column models and parametrization schemes. The data assimilation uses a variational technique that minimizes the difference between the model results and the observation during the analysis period. The adjoint model is used to compute the gradient of a measure of the model errors with respect to the control variables that force the model output closer to the data. This is the most advanced type of data assimilation being developed for operational use. It has the advantage of being able to use the indirect and non-conventional observations that are available at the ARM sites.

The model is based on the AER Local Forecast and Assimilation (ALFA) model. It is a single column model that solves the same equations as a climate model, with complete physics, but uses large scale, three-dimensional analyses or forecasts for all the horizontal derivatives.

An additional goal of the project is to use the adjoint model to perform sensitivity experiments, especially with regard to the radiation scheme and the cloud parametrization.

Information
Major Projects
Publications