Assimilation of GNSS Zenith Total Delay in NAVGEM

A new data source from ground-based stations that track global navigation satellite system (GNSS) transmitters has been implemented in the Navy Global Environmental Model (NAVGEM) with the NRL Atmospheric Variational Data Assimilation System-Accelerated Representer. The observable is the ground-based zenith total delay (ZTD) at each ground station. This ground-based GNSS ZTD represents the tropospheric delay of the propagation of GNSS radio L-band signal between the transmitters and the ground stations after correcting for ionospheric effects. Here, we present the implementation procedure, quality control, single observation test, bias correction scheme, and data impact assessment. One of the critical elements in the quality control scheme is to correct for the difference between the observing ground station elevation and the numerical model terrain. A single observation test shows comparable impact from this new data source to a single radiosonde observation at a single level. The ZTD biases estimated from a 6-month experiment run are generally small, but have a dependency on the processing center and are computed separately for each station. The forecast sensitivity to observation impact diagnosis from a 3-month experiment demonstrates that a single ZTD observation shows similar impact as the average impact from a single tangent point of a GNSS Radio Occultation profile. The assimilation of ZTD observations also significantly improves the forecast skill by 0.25%-0.75% for wind, temperature, and precipitable water and by 1%-2% for geopotential height in the Southern Hemisphere beyond 3 days. Plain Language Summary Adequate sampling of water vapor in meteorological and climate observing systems is critical to environmental characterization and climate monitoring. Currently, atmospheric humidity observations in both space and time are not adequately sampled. The ground-based global navigation satellite system (GNSS) network offers a cost-effective pathway to retrieve high-quality humidity observations. Observations from the GNSS network can serve as anchoring observations for the data assimilation and numerical weather prediction. These observations complement other satellite measurements to further improve forecast quality. This study outlines the implementation procedure for the ground-based GNSS zenith total delay and assesses its impact on Navy's global numerical weather forecasts. Our results indicate that the data can have a significant positive impact on the Southern Hemisphere's forecasts beyond 3 days. When evaluating each observation's impact on the 24-hr model forecasts, the impact of a single zenith total delay measurement is comparable to that of the average impact of a single GNSS Radio Occultation data at a single tangent point.