Validation of ion drift assimilation for stormtime ion drift and neutral wind estimation

During geomagnetic storms, the earth’s magnetic field is penetrated by plasma and energy flow from solar activities, and this affects atmospheric behaviors. One of the phenomena being observed through the study of the October 2003 extreme storm is that of the enhanced ionospheric plasma, which corotates persistently with earth during the late-night period (18 LT - 24 LT), initially named the “Florida effect.” This “nighttime ionospheric localized enhancement” (NILE) was observed through the data collected from over 400 ground-based multi-frequencies GNSS receivers. Studies on August 2018 and November 2003 storms [1], based on the implementation of data assimilation tool IDA4D [2] coupled with ionosphere model SAMI3 [3], showed the evolution of NILE in terms of occurrence time and region along with GPS and ionosonde validations [4]. These mid-latitude events were less concentrated in spatial extent and less severe in density than that observed during October 2003, which raised the question of whether the mid-latitude effects observed during these storms could be classified as NILE. To validate the occurrence of NILE, the data assimilation algorithm can be applied to investigate other storm events, e.g. March 2015 storm.

A data assimilation tool, Estimating Model Parameters Reverse Engineering (EMPIRE), has been developed to reduce the gap between background climate models and real measurements in estimating the neutral winds and ion drifts [5]. The EMPIRE algorithm is led by the physics-driven ion continuity equation, which relates the electron density rate to the terms of production and loss rates, and motion terms parallel and perpendicular to the magnetic field. Time-differenced electron density 4D global maps obtained from IDA4D are treated as measurement input. Other climate models feed source terms as background values. Vector spherical harmonics and spherical harmonics basis functions are applied for estimating the state, consisting of neutral winds and electric potential in the global scope. Beyond the ion continuity equation, augmentations to the EMPIRE algorithm that directly ingest neutral wind and ion drift information improve the estimation of physical drivers. After linearizing the system, the 3DVAR Kalman filter is applied for the optimal estimation of state and covariance [6].

In this work, we implement the EMPIRE algorithm to ingest the 4D plasma density rate and ion velocities on the March 2015 storm event and compare the estimated state to a validation data set. The 4D plasma density map is produced from IDA4D algorithm, and the finite-differenced 4D map is fed into the algorithm as the density rate measurements. As for the ion velocity measurement, there are four ISR sites available for data ingestion and validation. Two sites are located at low-to-mid latitudes(Arecibo and Millstone Hill), and the other two datasets are obtained from SuperDARN and Poker Flat. Our primary goal is to ingest the ion drifts from a low-to-mid latitude site and compare them with another low-to-mid latitude site. Then, we compare the EMPIRE ingestion results with the high-latitude datasets to check whether ingesting ion velocities improves the calculations of global physical drivers of neutral winds and electric potential.