GEM, Snowmass, June 16, 1998
The November 1993 storm is a typical solar minimum storm associated with a recurring high speed stream of plasma velocity in the interplanetary medium as shown in Figure 1a. The storm started about 18 UT on Nov 3 with a brief burst of activity, and then at 22 UT began a sustained period of elevated activity as seen in the AE (Fig 1b) and the Dst (Fig 1c) calculated using ground magnetometers. The activity is also seen in the cross-tail potential drop (Fig 1d), the particle heating (Fig 1e), and the Joule heating (Fig 1f), as determined from the Assimila- tive Mapping of Ionospheric Electrodynamics (AMIE) procedure for the winter northern hemisphere (NH) and the summer south- ern hemisphere (SH). The AMIE electric potential, and auroral electron flux and mean energy were put into the Thermosphere- Ionosphere-Electrodynamics General Circulation Model (TIEGCM). Figure 1g shows the Joule heating from the TIEGCM including neutral winds. The effect of the neutral winds is to reduce the Joule heating by 21% over the 10 days, 12% on quiet days and 24% on Nov 4. The neutral winds are more tightly coupled to the ion drifts in summer because of higher electron densities, so the Joule heating is reduced by 18% in the NH and 24% in the SH. Therefore the SH Joule heating compared to the NH in the TIEGCM with winds is 8% lower in general and 13% lower on Nov 4 as seen in Figure 1g. The TIEGCM Joule heating was then multiplied by a correction fac- tor to account for variability in the E field that is largest in winter and increases the temperature, but does not affect the momentum transfer. The factors chosen were 1.5 for the SH and 2.5 for the NH. The resulting Joule heat curves are plotted in Figure 1h with the summer Joule heat now half the winter on Nov 4.
| Figure 1 |
Figure 2 shows the TIEGCM neutral temperature (Figs 2e,g) and winds (Figs 2b,d,i,k,m,o) at 200 km for the SH at 2 UT on Nov 3 during quiet conditions, and at 22 UT on November 5, ~48 hours after the storm began. Also shown are the electric potential and ion drifts (Figs 2a,c). Next to the TIEGCM results are the Mass Spectrometer Incoherent Scatter (MSIS) model temperature (Figs 2f,h) and the Horizontal Wind Model (HWM) winds (Figs 2j,l,n,p) for the same conditions. The quiet-time winds are upwards during the day and downwards at night (Fig 2b), and blow from day to night, with HWM westward (negative) winds (Fig 2j) strongest ~10LT, and TIEGCM (Fig 2i) strongest ~4 LT, just before sunrise where the electron density is smaller. Similarly, the quiet-time HWM meridional winds (Fig 2n) are strongest poleward (negative) at noon while the TIEGCM pole- ward winds (Fig 2m) are strongest ~8 LT. During the storm, the temperature (Figs 2e,g) is ~200 K higher ~6 LT in the positive potential cell (Fig 2c) in the TIEGCM, and ~150 K higher in MSIS (Figs 2f,h). The zonal winds are more westward during the storm (Figs 2i,j,k.l), with the TIEGCM winds (Fig 2k) east- ward at 6 LT where the ion drifts (Fig 2c) are eastward. The TIEGCM meridional winds (Fig 2o) also show strong coupling to the ion drifts (Fig 2c) during the storm, and the TIEGCM ver- tical winds (Fig 2d) are upwards on the dawn side of the auroral zone, with downward winds just equatorwards.
| Figure 2a | Figure 2b | Figure 2c | Figure 2d |
| Figure 2e | Figure 2f | Figure 2g | Figure 2h |
| Figure 2i | Figure 2j | Figure 2k | Figure 2l |
| Figure 2m | Figure 2n | Figure 2o | Figure 2p |
The WIND Imaging Interferometer (WINDII) on the Upper Atmosphere Research Satellite (UARS) measured neutral winds around 200 km in the SH ~8 LT at 20 S to ~20 LT at 60 S. The satellite path on Figure 2 shows that the winds are expected to be mostly westward at this time, increasing in magnitude with the storm. Figure 3 shows the WINDII observations at 200 km in the SH on Nov 3, 5 and 6, compared to the TIEGCM winds and the HWM winds using 3-h ap histories. 2 UT is ~100 E on Nov 3 and 22 UT is ~150 E on Nov 5. The greatest variability is in the WINDII observations and the strongest winds are near the mag- netic pole. The winds are generally westward with the HWM showing some equatoward flow and TIEGCM some poleward flow, with WINDII inbetween. Both WINDII and TIEGCM show the greatest variability in the longitude sectors close to the magnetic pole. The westward winds increase with the magnetic activity.
| Figure 3 |
Figure 4 shows the hmF2 and magnetic meridional winds for Darwin, Australia between Nov 2 and 5. There are obvious gravity wave signatures in the observations, which are well cap- tured in the TIEGCM using the realistic AMIE inputs to deter- mine the Joule heat sources. The HWM winds are almost all equatorwards, as might be expected from Figures 2n and 2p for this latitude. The WINDII observations are triangles and are in excellent agreement with the FLIP winds, although they are of opposite sign for the same LT on Nov 3 and 5.
| Figure 4 |
Figure 5 is a plot of the Fabry-Perot Interferometer (FPI) nighttime doppler temperatures and winds from ~250 km and neutral exospheric temperatures deduced from the incoherent scatter (IS) radar at Millstone Hill at 55 mag N. The model tem- peratures at 300 km increase about 200 K during the day and 300 K at night during the storm and then decrease about 20 K/day and 14 K/night after a sharp drop between Nov 4 and 5, similar to the observations. The observations show a further sharp decrease between Nov 5 and 6. The IS nighttime temperatures are normal on Nov 10, while the model nighttime temperatures are still high at the end of the period. There are large gravity waves on Nov 4 in the FPI winds, similar to the wind variations in the TIEGCM. The equtorward wind is doubled at night on Nov 4 in the model. The TIEGCM westward winds are enhanced and the eastward winds are depressed after the storm. The FPI eastward winds appear to recover from the storm around Nov 9, although the model winds still show the eastward winds increasing at night. The westward shift of the zonal winds is the consequence of high latitude heating during magnetic storms, and is associated with the disturbance dynamo.
| Figure 5 |
Figure 6 shows the differences between the storm time temperature, meridional winds and vertical winds from quiet time values at 300 km on November 4 for 85 W. There are clear gravity wave signatures in all three variables going from the northern and southern hemispheres to the opposite poles at a rate of about 22 degrees/hr or about 700 m/s phase speed. The wave fronts are clearest coming from the NH because more Joule heat- ing is deposited in the winter hemisphere, and because in the summer hemisphere, the wind surges are damped by stronger ion drag. The wave fronts in the vertical velocity (Fig 6b) are short periods of downward velocity, which coincide with the start of periods of enhanced temperatures (Fig 6c) and equatorward winds (Fig 6d). The equatorward winds indicate a polar source. Figure 6a is a plot of the total NH heating from corrected Joule heating and particle sources. The times of the temperature wave fronts in Figure 6c at 60 N clearly correspond to peaks in the Joule heating with the waves arriving at the equator ~3 hours later. Gravity wave phase speeds are about 80% slower at 125 km where the waves arrive at the equator ~45 min later.
| Figure 6a | Figure 6b |
| Figure 6c | Figure 6d |
The differences in the zonal geographic wind between the storm run and the quiet background are plotted at 0 LT for the TIEGCM and the HWM with 3-h ap histories in Figure 7 for selected magnetic latitudes at 300 km and 125 km. The largest wind enhancements are at 200 km in the TIEGCM, and drop 10- 20% at 300 km, but are largest at 300 km from the HWM. The TIEGCM values are also largest in the NH winter where the elec- tron density is lowest at night and the Joule heat correction factor is 2.5 instead of 1.5. The initial westward surges are ~20 UT on Nov 3 at 40 mag N and S, and ~22 UT at the magnetic equator, consistent with gravity wave propagation. The gravity waves originating in the NH are also evident in the SH at 40 mag S. The HWM westward wind enhancements are comparable to those of the TIEGCM at low latitudes, but are smaller at higher latitudes and seem to be recovered in the winter hemisphere by Nov 9. The linear correlation coefficient between the TIEGCM enhanced westward winds at 300 km at 40 mag N in Figure 7a and the Joule heat peaks in Figure 1f is -0.84 with a lag of 1.7 hours. The lag times increase at lower latitudes consistent with gravity wave propagation. At 125 km, the TIEGCM magnitudes are considerably reduced. The westward peaks are related to the peaks at 300 km, but slightly delayed, smoothed, and increasing in magnitude with time. The TIEGCM 24 hour average west- ward enhancements at 0 LT are strongest on Nov 7 between 25 mag S and 35 mag N, although initial responses are evident on Nov 4 only about an hour later at 125 km compared to 300 km. However, the delay in the maximum amplitudes suggests satura- tion time scales of the order of 1 to 3 days, consistent with the longer time constants in the lower thermosphere.
| Figure 7 |
Figure 8 shows the average westward enhancement (dif- ference between storm and quiet geographic zonal winds) over Nov 4 and 7 as a function of magnetic latitude for the TIEGCM and the HWM at 300 km and 125 km. The TIEGCM averages are largest around 40 S and 45 N magnetic latitudes, and least around 15 S at 300 km, shifting 5 degrees higher at 125 km. The HWM averages are largest at 65 mag S and N at 300 km and at 55 mag S and N at 125 km. The TIEGCM winds are also stron- ger in the winter hemisphere because the Joule heating correction term is larger, because the electron densities are lower at night in the winter hemisphere, and because the neutral winds are more tightly coupled to the ion drifts in the summer hemisphere.
| Figure 8 |
The neutral temperature and O/N2 ratio were averaged over the globe at 300 km and compared to MSIS using the 3-h ap time history in Figure 9. The TIEGCM global mean temperature at 300 km increased 188 K in 26 hours and then decayed at a rate of 17 K/day after Nov 5, ending 28 K above initial conditions. The MSIS global mean temperature increased 105 K during the storm, and recovered by the end of the calculation. The O/N2 ratio at 300 km decreased ~40% in the TIEGCM and ~28% in MSIS, recovering completely in MSIS, and nearly in the TIEGCM. The recovery after Nov 5 in the TIEGCM was 0.30/ day or at ~5%/day of the 40% drop. Also plotted are the TIEGCM global averages at 125 and 150 km which look similar to those at 300 km.
| Figure 9 |
Linear correlations were calculated between the tempera- ture and the input potential drop averaged over both hemispheres (Fig 1d) and the official Dst (Fig 1c), which had similar enve- lopes to the neutral responses except for the rise in Dst at the start of the storm. The Dst correlation was thus divided into two parts, between a growth phase starting with the maximum Dst, and a decay phase with a steeper slope.
Figure 10 shows the linear correlations. The best lag was 8 hours for the temperature increase. Further correlations were found between the neutral temperature and the O/N2 ratio. The correlations were largest when there was no lag between the tem- perature and the O/N2 ratio. The lag time of about 8 hours between any of the high latitude inputs and the temperature increase and decay was found to be independant of latitude for the zonal temperature means between 50 S and 50 N, although the initial temperature increases moved from the auroral regions to the equator with the phase speed of the gravity waves. The lag time of 8 hours reflects the saturation time of the cumulative effect of the gravity waves and circulation changes. The lag time was different in height, with the global mean temperature lag at 8 hours at 300 km, at 9 hours at 200 km, and at 7 hours at150 and 125 km, all with correlation coefficients of about 85%. However, a difference of 1 hour at different heights over the 8 days of activ- ity is probably not very significant.
| Figure 10 |
- The AMIE auroral inputs produced gravity waves in the neutral temperature, wind components, and the hmF2 that were largest on Nov 4 and agreed with observations most of the time.
- The latitudinal variation of the dynamo winds show maximum increased westward winds at midnight at 300 km of ~200 m/s in the NH and ~135 m/s in the SH at ~40 degrees magnetic latitude, while the westward wind enhancements at the equator are only ~40 m/s. The maximum westward wind enhancements at mid- night are 15-20% larger at 200 km, unlike HWM where the wind differences increase with height.
- The neutral temperature increased over the storm period, and then decreased with the decay of the input forcing from high lati- tudes, with maximum enhancements at night. The global mean temperature increased by 188 K during the first 26 hours of the storm, and then decreased after November 5 by 17 K/day, corre- lating well with Dst and the potential drop with an 8 hour lag.
There are 2 main time scales involved in the neutral thermo- sphere responses to magnetospheric forcing -- the phase speed of gravity waves and the saturation times which were deduced for this long solar minimum storm case.
- The phase speed of gravity waves from auroral heat sources is ~22 deg/h or 700 m/s at 300 km and ~80% less at 125 km.
- Westward wind enhancements at 300 km are driven by the grav- ity wave phase speed, and the saturation time is about the same.
- The saturation time for the westward wind enhancements at 125 km is between 1-3 days.
- The saturation time for the global and zonal neutral temperature and densities (like the O/N2 ratio) is ~8+/-1 hours between 125 and 300 km from correlations with magnetospheric forcing.
This study made use of the CEDAR Data Base at the National Center for Atmospheric Research which is supported by the National Science Foundation (NSF). We are also grateful to the following individuals and institutions for providing data for the analysis or for the AMIE simulation: L. Frank (GEOTAIL plasma data); A. Lazarus and K. Paularina (IMP-8 plasma data); M. Hairston and R. Heelis (DMSP Ion Drift Meter and Retarding Potential Analyzer data); F. Rich (DMSP J4 precipitation data); R. Frahm (UARS Medium Energy Particle Spectrometer and Atmospheric X-ray Imaging Spec- trometer data); D. Evans (NOAA-12 precipitation data); E. Sanchez (Sondrestrom IS radar conductances); the Sondrestrom IS radar is supported by NSF;; the EISCAT Scien- tific Association is supported by Finland (SA), France (CNRS), the Federal Republic of Germany (MPG), Japan (NIPR), Norway (NFR), Sweden (NFR) and the United Kingdon (PPARC); the Millstone Hill IS radar and Fabry-Perot are supported by NSF; J. Scali and B. Reinisch (Qaanaaq and Sondrestrom digisonde drift data); M. Ruohoniemi (Goose Bay and Kapuskasing HF radar data, which are operated by the Applied Physics Laboratory of The Johns Hopkins University with support from NSF and from NASA for Kapuskasing); G. Sofko (Saskatoon HF radar, which is operated by the University of Saskatchewan with support from the Natural Sciences and Engineering Research Council of Canada); M. Pinnock (Halley HF radar, which is supported by NSF and the British Antarctic Survey). We are grateful for the following providers of magnetometer data: O. Troshichev (Arctic and Antarctic Research Institute); C. Maclennan (AT&T); D. Detrick (University of Maryland); A. Rodger (British Antarctic Survey); G. Rostoker and T. Hughes (CANO- PUS, which is supported by the Canadian Space Agency); E. Friis-Christensen and T. Morretto (Danish Meteorological Institute); R. Sitar (MAGIC, University of Michigan); S. I. Solovyev (Institute of Cosmophysical Research and Aeronomy); H. Luehr (IMAGE); L. Cafarella (ING); A. Green (INTERMAGNET); A. Zaitzev (IZMIRAN); M. Engebret- son and J. Hughes (MACCS); L. Morris (NGDC); D. Milling (SAMNET); V. Mishin (SIBIZMIR); and the 210 Magnetic Meridian magnetometer chain. The 210 chain is managed by K. Yumoto and kept at the Solar-Terrestrial Environment Laboratory (STEL) at Nagoya University. K. Shiokawa assisted in getting the data. The following individuals and institutions supported 210 ground magnetic observatories used in the AMIE study: Electronics Research Laboratory, K. J. W. Lynn, Australia; Birdsville Police Station, Aus- tralia; Kakioka Magnetic Observatory, S. Tsunomura, Japan; Dalby Agriculture College, Australia; University of Newcastle, B. J. Fraser and F. W. Menk, Australia; CSIRO Trop- ical Exosystems Research Center, Australia; STEL, Japan; Tohoku University, T. Taka- hasi, Japan; Institute of Cosmophysical Research and Aeronomy, S. I. Solovyev and G. F. Krymsky, Russia; Institute of the Physics of Earch, V. A. Pilipenko, Russia; and Weipa North State School, Australia.
