Assimilation of Surface Meteorological Observations in COAMPS® NAVDAS

Conventional meteorological surface stations typically offer complete observations of temperature, dew point, winds, and pressure with high temporal frequency. The density of these stations is increasing significantly, especially in the United States, as government agencies, private businesses, and individuals continue to build and maintain these stations for various meteorological uses. While these stations can offer important information to numerical models by correcting errors in the lower boundary condition of the initial state, they can also be very difficult to assimilate, due to siting and representativeness, station maintenance and quality control, and instrumentation configuration issues.

This work presents the methods used to assimilate these observations into the NRL Atmospheric Variational Data Assimilation System (NAVDAS), including interpolation, quality control, and bias correction procedures. The Coupled Ocean-Atmosphere Mesoscale Prediction System (COAMPS®) is used to determine impact of surface observations on mesoscale forecasts for the Intermountain West during part of the 2010/2011 winter season. Data withholding experiments are performed comparing forecasts with and without the surface observations. The adjoint of COAMPS and NAVDAS is also used to determine the benefit of surface observations through adjoint sensitivity experiments.