Applying Hpo Indices to Empirical Thermospheric Density Models During Geomagnetic Storms

Abstract Accurate atmospheric drag modeling is essential for precise orbit determination and prediction of Low Earth Orbit satellites. A key component is the thermospheric density, typically estimated using empirical models driven by geomagnetic activity indices such as the 3‐hr Kp or ap. However, the coarse temporal resolution and fixed upper bounds of Kp/ap indices limit their ability to reflect rapid geomagnetic variations, especially during storms, leading to underestimated thermospheric density. In this study, we investigated the usefulness of replacing Kp/ap with the higher‐frequency and open‐ended Hpo/apo indices during geomagnetic storms. The correlation between Kp/ap and Hpo/apo indices over 126 geomagnetic storms was analyzed, and strong correlations (correlation coefficients > 0.93) were found. Using satellite‐derived density as a reference, we evaluated the performance of four empirical thermospheric models, including JB2008, NRLMSIS2.1, DTM2020 Operational, and DTM2020 Research, when driven by Hpo/apo. Results show that the performance of the JB2008 and DTM2020 Operational models is improved during moderate and intense geomagnetic storms but deteriorates under super storm conditions. The improvements are approximately 3%–9% for computed to observed (C/O) means and 4%–11% for C/O RMSEs with the JB2008 model, respectively, and less significant with the DTM2020 Operational model. For NRLMSIS2.1, only slight differences (<1% in means and <2% in RMSEs) were found in the model performance after index substitutions. Under super storm conditions, the impacts of index substitutions are statistically insignificant at the 95% confidence level. Additionally, increasing Hpo/apo temporal resolution from 60 to 30 min yields limited benefits (<1% improvement) for all models.