Outstanding Problems concerning the Global Electric Circuit

We revisit the problems related to the discrepancy between the observed amplitudes, of responses of the global circuit to inputs related to solar activity, as compared to modeled amplitudes. The solar activity modulates the latitude distribution of the cosmic ray flux. It also modulates the X-ray flux at middle to high latitudes in the stratosphere, that is due to Bremsstrahlung from MeV electrons precipitating from the magnetosphere. Both ionizing inputs modulate the latitude distribution of vertical column conductivity in the global electric circuit. The discrepancies between observations and models increase with increasing content of stratospheric volcanic aerosol. One possibility for resolving the discrepancies is that the column resistance of the stratosphere and upper troposphere is greater than previously calculated, due to the presence of stratospheric aerosol and very thin clouds near the tropopause. The increase would depend on the aerosol content, and would at times bring the stratospheric column conductivity to a level that was not negligible with respect to the tropospheric column conductivity. OVERVIEW OF GLOBAL ELECTRIC CIRCUIT. A schematic diagram of the global circuit is given in Figure 1. The generators in the low latitude troposphere experience much less variation of the cosmic ray flux than the middle and high latitude regions, and are completely separated from MeV electron precipitation and polar cap ionospheric potential variations. The ionosphere is essentially an equipotential above about 50 km altitude from the equator to about ±50° geomagnetic latitude, so that the distance between the surface equipotential and the ionospheric equipotential is small compared to the radius of the earth. The horizontal scale of the polar cap variations are also much larger than 50 km, and so the geometry to a good approximation can be considered plane parallel. The downward ionosphere-surface current density Jz is determined by both the local ionospheric potential Vi and by the local ionosphere surface column resistance R, with Jz = Vi/R, and R given by R =∫0 (1/σ(z))dz, where σ(z) is the conductivity at an altitude z. Because the region above about 50 km has very high conductivity and is essentially at ionospheric potential for any one location, the integral need be evaluated only in the troposphere and stratosphere. In Fig. 1 the tropospheric contributions to the vertical column resistance at given latitudes are designated TE, TL, TH and TP, for equatorial, low, high and polar latitudes, with corresponding designations SE, SL, SH, and SP for the stratospheric contributions. At any latitude the current density is given to a good approximation by Jz = Vi /(T+S). The galactic cosmic ray (GCR) flux modulates T and S continuously. The more intermittent fluxes of MeV electrons, with their associated X-ray bremsstrrahlung modulate SP and SH. The occasional solar proton events modulate SP. The S contributions to R are normally small, on account of the relatively high conductivity in the stratosphere under normal conditions. But their contribution appears to be quite significant when the general stratospheric conductivity has been greatly reduced by H2SO4/H2O liquid droplets and H2SO4 vapor following volcanic eruptions (Tinsley, 2000). Polar stratospheric clouds, and also other sub-visible clouds in the tropopause region, may have the same effect. The flux of MeV electrons in the magnetosphere is strongly correlated with the solar wind velocity, as discussed by Li et al. (2001) who also describe the time and latitude variations of the precipitating flux. It is when the stratospheric conductivity has been lowered by aerosol and cloud particles and vapor (giving relatively large S) that the modulation of SP and SH by MeV electron precipitation seems to affect R and Jz at high and polar latitudes. Measured potential gradients and associated Jz variations responding to MeV electron flux changes are discussed by Tinsley (2000). DECADAL AND LONGER TERM VARIATIONS A large number of balloon measurements of the integrated electric field from 0-12 km (VT ) were made by Fischer and Mühleisen (1980) from Weissenau, southern Germany in 1959-1976. They were compared to proxies for galactic cosmic ray (GCR) flux and stratospheric aerosol content by Meyerott et al. (1983), who noted that a peak in VT in 1964 occurred at the same time as the peak in stratospheric sulphate aerosol concentration, due to the eruption of the Agung volcano. They estimated that it would only require an increase in aerosol concentration of ~50% throughout the troposphere to account for the observed change in ionospheric potential due to the eruption. (We can also note that the GCR flux varied by only 15% at Deep River, corresponding to <10% globally, whereas VT increased by about 50%, so the GCR variation cannot explain the 1964 peak.) Meyerott et al. noted that since the stratospheric aerosol concentration increased by two orders of magnitude, a 50% increase in the tropospheric content by fallout from the stratosphere is a possibility. However, the main loss of such aerosols is by downward transport by tropopause fold and related phenomena at middle latitudes, and the tropospheric aerosol content would have to increase by that factor globally to produce a corresponding increase in Vi. Figure 1. Schematic diagram of the global electric circuit. The equatorial and low latitudes have much smaller conductivity changes, due to energetic particle influx, than high or polar latitudes. The main generator in the tropical troposphere may be considered a constant current source, producing an ionospheric potential of order of 250 kV over most of the globe, except in the polar regions, where superimposed dawn-to-dusk potentials and a pole-to-pole potential are produced by the solar wind-magnetosphere-ionosphere interactions. An alternative interpretation is that the stratospheric aerosols caused a larger increase in resistivity in the stratosphere itself (by attachment of ions to and recombination on aerosol particles, and the reduction of ion mobility by clustering of H2SO4 and H2O molecules onto the ions, and nucleation of ultrafine particles on the ions, as discussed by Yu and Turco (2000)). A uniform increase in resistivity throughout the global stratosphere would increase the true ionospheric potential, but not the measured tropospheric potentials 0-12 km. However, satellite observations show a zone of lower aerosol concentration in the stratosphere at mid-latitudes, where the aerosol is being transported downwards most rapidly, and this would result in an enhanced downward current density there (a smaller S than at other latitudes), and an enhanced VT. The Weissenau data on VT can be looked upon as a response to the Agung eruption together with a solar cycle variation with a higher VT at solar minimum. If so, the solar cycle variation is consistent with that found by Olson, who made balloon measurements in 1966-1977 of Jz over a solar cycle at high geomagnetic latitudes (mostly in northern Minnesota). His results were analyzed by Markson and Muir (1980, their Fig.5) as a function of sunspot number. Olson’s results show a solar cycle variation, with a maximum at solar minimum, and amplitude about 40%. The Weissenau amplitude for the same time period is in the same sense, but less, about 20%. An interesting contrast is found in measurements of the near-surface electric field Ez at low latitudes. These show the opposite solar cycle variation, with a maximum at solar maximum. This is illustrated by Israël (1973, Figure 249). The data from 1926 to 1950 show maxima near the solar maxima in 1928 and 1940 and a minimum near solar minimum in 1933-34 in the parameter Jz /(Λ) , where Λ is the local near-surface conductivity. For these land stations most of the Λ variation would be short term, and due to varying local aerosols, with near-surface ion production dominated by surface radioactivity. Thus the systematic solar cycle variation would be in Jz, averaging about 10%, implying similar or larger variations in Vi. (The changes in R at low latitudes due to GCR variations would be less than 5% and are in the sense of increasing the amplitude of Vi above that of Jz .) While the opposite sense of the solar cycle variation at low latitudes compared to that at high latitudes is consistent with the theoretical expectation described earlier, due to solar cycle variations in the GCR flux, the amplitude of this variation is surprisingly large. Previous modeling has shown that it is not possible obtain 40% variations in Jz with the typical GCR variations of 15-20% at high latitudes and 3-5% at low latitudes. However, the MeV electrons have much larger solar cycle variations, by at least an order of magnitude at L = 3 to L = 4 (55° 60° geomagnetic latitude) as shown for example by Li et al. (2001). The X-ray bremsstrahlung produced by such electrons can penetrate down to 20 25 km (Winkler, 1960; Fram et al., 1997). Thus, if moderate amounts of stratospheric aerosols and vapor enhance stratospheric resistivity and column resistance S, but moderate amounts of MeV electron fluxes produce X-rays and ionization in the stratosphere at middle and high latitudes that reduce it again, then there will be solar cycle effects on the latitude distribution Jz, with opposite changes at high and low latitudes. These would be similar to those due to GCR modulation, but with potentially much larger amplitude. The amplitude will depend on the amount of enhancement of S that is produced by the aerosols and vapor. Even without the MeV electron precipitation, the effect of stratospheric aerosols is to increase the amplitude of Jz changes, because the relative amplitude of GCR flux variations is several times greater in the lower stratosphere than in the troposphere. BALLOON OBSERVATIONS Stratospheric balloon observations by Byrne et al. (1988) measured conductivity variations between 10 and 30 km at three widely separated latitudes in 1973-75 and 1980-81, which were periods of medium to low general stratospheric aerosol content. They inferred the presence of concentrated aerosol layers near 19 km of width a few km, at both high and low latitudes. They found that the high latitude flights yielded conductivities otherwise reasonably consistent with the Hays and Roble (1979) model. However at low latitudes the conductivity above 25 km was a factor of 2-3 lower than the Hays and Roble model, and they attributed this to a global background of aerosols, having a greater concentration at low latitudes. The Hays and Roble (1979) model assumes no stratospheric aerosols, and thus a square law relation (q=αn) between ion production rate q and ion concentration n, rather than a linear relation (q=βZn) for a loss process due to attachment to aerosols, of concentration Z. Byrne et al. (1988) also found a day-to-day variability that they suggested might have been due to variability in aerosol concentrations. Another set of balloon measurements was reported on by Hu and Holzworth (1996). They measured conductivity at 26 km altitude in November 1992 March 1993, which, as they noted, was before volcanic aerosols from Pinatubo had reached the latitude of observation. Thus the concentration of stratospheric aerosols would be quite low, but even so, they found conductivities a factor of 2.5 to 4 lower than the Hays and Roble (1979) model. The found day-to-day variations in the conductivities, of unknown origin, up to 30% of the mean. They made new calculations of conductivity, using updated values for the ion-ion recombination coefficient α, but still their measured conductivities were only 60% of the calculated values, over the whole latitude range from 15° to 80° south geomagnetic latitude. While they did not attribute the day-to-day variations and the discrepancy with theory to the presence of aerosols, these offer an explanation that cannot be ruled out. Another set of balloon measurements indicating aerosol effects is that of Ermakov et al. (1997) , who made extensive observations, from 5-35 km altitude, of ion concentrations and ionizing particle fluxes at low, middle, and high latitudes. Their conclusion was “that at altitudes 7-30 km the relation between the ion production rate and ion density is linear rather than square law as commonly accepted, i.e., q=βn instead of q=αn.” Their measurements were made in 1990-91, which was not a period of high stratospheric aerosol content, but as for the observations of Byrne et al. (1988) and Hu and Holzworth (1996) the aerosol content was significantly higher than it had been in the