The Limits of Empirical Electron Density Modeling: Examining the Capacity of E‐CHAIM and the IRI for Modeling Intermediate (1‐ to 30‐Day) Timescales at High Latitudes. Issue 4 (30th March 2020)

• Record Type: Journal Article
• Authors: Themens, David R.; Jayachandran, P. T.; Reid, Benjamin; McCaffrey, Anthony M.
• Is Part Of: Radio science. Volume 55:Issue 4(2020)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-03-30
• Subjects: ionospheric storms -- empirical modeling -- E‐CHAIM -- IRI -- polar cap -- ionosphere; Radio meteorology -- Periodicals; Radio wave propagation -- Periodicals; 621.38405
• DOI: 10.1029/2018RS006763
• Languages: English
• ISSNs: 0048-6604

Abstract

Abstract: The Empirical Canadian High Arctic Ionospheric Model (E‐CHAIM) is a new empirical 3‐D electron density model intended as an alternative to the use of conventional standards, such as the International Reference Ionosphere (IRI), at high latitudes (above 50°N). In this study, we have manually scaled a year of data from two Canadian High Arctic Ionospheric Network (CHAIN) ionosondes. Using this high‐quality data, we examine the behavior of the polar cap ionosphere under disturbed geomagnetic conditions and assess the capacity of E‐CHAIM to model polar cap F2‐peak electron density variability on "weather‐like, " intermediate timescales (1–30 days). This is a particularly challenging environment for monthly median empirical models due to the regular occurrence of variations about the monthly mean of up to 2 MHz. We demonstrate in this study that E‐CHAIM's storm model is capable of explaining 4 to 25% of polar cap foF2 variance at 1‐ to 30‐day timescales and 5 to 50% of the amplitude of that variability, while the IRI's Storm‐Time Ionospheric Correction Model (STORM) only explains 0.2 to 9% of the variance at these timescales and no more than 5% of their amplitude. While the IRI's STORM model provided no measurable improvement over the monthly median, E‐CHAIM's storm parameterization was able to improve overall root‐mean‐square errors by 0.05 to 0.1 MHz over its quiet time model. The overall improvement through the use of storm foF2 parameterizations is found to be limited, but measurable, particularly during storm periods, where an average improvement in root‐mean‐square error of 20% is observed. Key Points: E‐CHAIM outperforms the IRI at 1‐ to 30‐day timescales within the polar cap E‐CHAIM is capable of explaining 4–25% of the foF2 variance at storm timescales at high latitudes Storm models are capable of improving overall model performance beyond the best monthly median representation