In this paper, the effects of the assumptions made in the calculation of the Dst index with regard to longitude sampling, hemisphere bias, and latitude correction are explored. The insights gained from this study will allow operational users to better understand the local implications of the Dst index and will lead to future index formulations that are more physically motivated. We recompute the index using 12 longitudinally spaced low-latitude stations, including the traditional 4 (in Honolulu, Kakioka, San Juan, and Hermanus), and compare it to the standard United States Geological Survey definitive Dst. We look at the hemisphere balance by comparing stations at equal geomagnetic latitudes in the Northern and Southern hemispheres. We further separate the 12-station time series into two hemispheric indices and find that there are measurable differences in the traditional Dst formulation due to the undersampling of the Southern Hemisphere in comparison with the Northern Hemisphere. To analyze the effect of latitude correction, we plot latitudinal variation in a disturbance observed during the year 2005 using two separate longitudinal observatory chains. We separate these by activity level and find that while the traditional cosine form fits the latitudinal distributions well for low levels of activity, at higher levels of disturbance the cosine form does not fit the observed variation. This suggests that the traditional latitude scaling is insufficient during active times. The effect of the Northern Hemisphere bias and the inadequate latitude scaling is such that the standard correction underestimates the true disturbance by 10–30 nT for storms of main phase magnitude deviation greater than 150 nT in the traditional Dst index.
