Validation of the Spire TEC Environment Assimilation Model (STEAM)

Spire Global develops and operates a large constellation of cubeSats which carry, amongst other software defined radio payloads, a dual frequency GNSS receiver (STRATOS). This receiver can be used to sense the Earths ionosphere and can measure large scale features (i.e. synoptic scale variations in the total electron content, TEC), medium scale features (i.e. sporadic E and travelling ionospheric disturbances in the lower ionosphere), and small scale features (i.e ionospheric irregularities causing scintillation). Spire has also developed a data assimilation model (the Spire TEC Environment Assimilation Model, STEAM) that can make use of the TEC data. STEAM uses a 4D Local ensemble transform Kalman Filter (LETKF); this is a Monte Carlo data assimilation method that represents the background error covariance matrix with an ensemble of model states. This paper will describe Spires strategy for measuring the performance of STEAM. Three areas will be discussed: Integrity validation using active (i.e. assimilated) and passive (i.e. not assimilated) data sets in conjunction with the STEAM background and analysis. Desroziers [2005] methods can be used to assess the model and data biases and covariances. External validation where a passive observation data set is used that is tailored for specific user requirements. This may include derived products such as skip distance in an HF system or time to convergence for a precise positioning system. External benchmarking where the validation from point 2 is extended to other models. Given the effort required to curate model outputs and data, it is desirable to move to a process of continuous validation. We will discuss Spire implementation of this approach and examine potential ways it may be more widely applied in the community. References Desroziers G., Berre L., Chapnik B., Poli P. (2005) Diagnosis of observation, background and analysiserror statistics in observation space. Quarterly Journal of the Royal Meteorological Society, 131, 3385-3396.