GNSS Differential Code Bias Determination Using Rao‐Blackwellized Particle Filtering

The Assimilative Canadian High Arctic Ionospheric Model (A‐CHAIM) is a near‐real‐time data assimilation model of the high latitude ionosphere, incorporating measurements from many instruments, including slant Total Electron Content measurements from ground‐based Global Navigation Satellite System (GNSS) receivers. These measurements have receiver‐specific Differential Code Biases (DCB) which must be resolved to produce an absolute measurement, which are resolved simultaneously with the ionospheric state using Rao‐Blackwellized particle filtering. These DCBs are compared to published values and to DCBs determined using eight different Global Ionospheric Maps (GIM), which show small but consistent systematic differences. The potential cause of these systematic biases is investigated using multiple experimental A‐CHAIM test runs, including the effect of plasmaspheric electron content. By running tests using the GIM‐derived DCBs, it is shown that using A‐CHAIM DCBs produces the lowest overall error, and that using GIM DCBs causes an overestimation of the topside electron density which can exceed 100% when compared to in situ measurements from DMSP. Plain Language Summary The Assimilative Canadian High Arctic Ionospheric Model (A‐CHAIM) is a near‐real‐time space weather model of the high latitude ionosphere. A‐CHAIM combines measurements from many different kinds of instruments, including from Global Navigation Satellite System (GNSS) receivers. These GNSS receivers require calibration in order to produce useful data, and a poor calibration can cause A‐CHAIM to produce incorrect results. A‐CHAIM uses a unique technique to calibrate the GNSS receivers self‐consistently without needing outside references. This new technique results in significantly improved performance in the weather model, but produces different calibration results than other GNSS calibration techniques. It is shown that if the other common calibration techniques were used, the weather model would produce large errors when compared to satellite measurements. Key Points Rao‐Blackwellized particle filtering is used to solve for GNSS Differential Code Biases (DCBs) in a near‐real‐time data assimilation model This method produces DCBs with systematic differences when compared to Global Ionospheric Maps, due in part to plasmaspheric effects DCBs determined using Global Ionospheric Maps can cause significant errors in reconstructed electron density when used in data assimilation