Evaluation of SuperMAG auroral electrojet indices as indicators of substorms and auroral power

We use magnetometer chains collaborating with SuperMAG to derive SME, a generalization of the auroral electrojet indices calculated from 100 or more sites instead of the 12 used in the official auroral electrojet indices, AE(12) = AU(12) - AL(12). We investigate how these various indices relate to nightside auroral power by using both particle (DMSP) and image (Polar Ultraviolet Imager (UVI)) data. The best correlation is between SME and total nightside auroral power, namely, r = 0.86. Hence, nearly 3/4 of the minute-by-minute variance in nightside power can be determined by SME alone. Interestingly, although the geophysical meaning of AE(12) has sometimes been challenged, we show that even that index correlates at the r = 0.81 level, or 2/3 of the variance, in nightside power. Most auroral power stems from the diffuse aurora, with a linear relationship between the auroral electrojet indices and nightside diffuse power. Thus, SME has a clear geophysical meaning: It samples the thermal portion of the magnetotail plasma sheet. We study how well SML (the generalization of AL) identifies substorms in two types of tests. The first is a comparison with a set of 1081 substorms determined from Polar UVI in 1997-1998, and the second is the use of DMSP particle precipitation data for superposed epoch analysis. The same algorithm applied to SML is much more likely (about 50% more likely) to identify an onset seen by UVI than is AL(12). Even when both indices can be used to identify onsets, the median delay after imaging onset until the AL indicator is less than half using SML (about 4 min versus 8 min). There are 10,719 onsets in the SML data between 1 January 1997 and 31 December 2002, of which 5084 are isolated. Isolated SML onsets behave almost identically to the onsets determined by global imagers. Specifically, they represent the same sharp spike in auroral power, which is most pronounced in broadband (wave) precipitation, with the same duration and subsequent recovery. However, recurrent substorms (those following less than 2 h after a previous onset) rise from a higher baseline by a smaller percentage but with the same absolute change in auroral power, thus reaching a higher peak power.