Working Group 5: Data Assimilation Across Disciplines

Development of a whole atmosphere-ionosphere data assimilation model

The development of whole atmosphere models, such as the NCAR WACCMX, enables an improved representation of upper atmosphere variability that is driven by disturbances in the lower atmosphere. In order to study real events, it is necessary to constrain the model dynamics, which can be done using data assimilation. In collaboration with the CISL Data Assimilation Research Section (DAReS), HAO scientists have completed the initial development of a data assimilation version of WACCMX. 

WACCMX+DART forecast image
The total electron content at 75°W longitude and 1800 local time. Results are shown for the WACCMX+DART forecast initialized on January 15 (top panel), WACCMX+DART analysis fields (middle panel), and ground-based GPS observations (bottom panel).

Data assimilation was incorporated into WACCMX using the Data Assimilation Research Testbed (DART) ensemble Kalman filter, and currently assimilates meteorological observations up to ~100 km. The initial capability of WACCMX+DART to generate high-fidelity reanalysis fields that are useful for scientific investigations and also to forecast upper atmosphere variability was evaluated by performing WACCMX+DART simulations focused on early 2009, when a major sudden stratospheric warming (SSW) occurred. Though they occur in the polar stratosphere (~30 km), SSWs are known to have wide-ranging impact on the entire atmosphere, including the ionosphere. They thus represent an ideal case study for WACCMX+DART.

The middle and bottom panels of the figure show a comparison between the WACCMX+DART analysis total electron content (TEC) and observations based on ground-based global positioning system (GPS) receivers. The temporal variability, especially the notable decrease in TEC that occurs after the SSW (indicated by dashed line), are reproduced in WACCMX+DART. This demonstrates that despite only assimilating observations up to ~100 km, WACCMX+DART is able to capture the ionosphere variability, which is dominated by changes occurring near ~300-400 km. The good agreement between the observations and WACCMX+DART analysis fields indicates that WACCMX+DART can be a powerful tool for generating high-quality representations of the upper atmosphere variability that is driven by the lower atmosphere. 

It is historically difficult to forecast ionosphere-thermosphere variability beyond several days. This is due to the fact that extended range forecasting requires a good forecast of the sources of ionosphere-thermosphere variability. However, through accurately forecasting the effects driven by the lower atmosphere, WACCMX+DART can potentially improve the useful forecast range of the ionosphere. The top panel of the figure shows the WACCMX+DART TEC forecast initialized on January 15. It is clear that the TEC variability can be qualitatively forecast for at least the next ~20 days. These results illustrate that, at least during certain time periods, there is the opportunity for extended range (10–20 day) forecasts of ionosphere-thermosphere variability using WACCMX+DART.