SPACE WEATHER: TIMELY SYNOPTIC MODELLING of the

NORTHERN E and D REGION ELECTRODYNAMICS and HEATING

J. K. Walker, P. Stauning, and 0. Troshichev

88 Starwood Rd., Nepean, ON, Canada

Danish Met. Inst., Copenhagen, Denmark

Arctic & Antarctic Res. Inst., St. Petersburg, Russia

INTRODUCTION

During the past few decades magnetic or space storms and related geospace storms have caused an increasing number of major disruptions of important power and communication services, malfunctions and loss of expensive facilities and satellites, and degradation of numerous geophysical surveys (Bames and Van Dyke, 1990). The next solar maximum is expected in a few years and the risk will again be high. Real-time modeling and improved forecasting of geospace activity are urgently needed to mitigate the effects of magnetic storms on sensitive industrial facilities, satellites, and radio transmissions and on geophysical surveys.

Accurate forecasting of the dynamic magnetic and geospace activity is exceptionally difficult. However, now casting may be possible with (1) L1 (WIND, ACE and SOHO) real-time solar wind observations, (2) in situ magnetospheric observations such as GOES, LANL, POES, GPS, THEMIS and CASSIOPE data and (3) key parameters mapped from real time electrodynamic models of the high latitude E region ionosphere. Most geospace activity leaves a footprint on the ionosphere. Hence the first step in space weather modeling and now casting is the determination of timely and sound modeling of the high latitude D and E region electrodynamics and heating. Such ionospheric and electrodynamic modeling requires, of course, extensive observations of the dynamic D and E region ionospheres in addition to numerous ground-based magnetic observations. From such ionospheric, magnetic and auroral observations models of the ionosphere and its conductances, the electric fields, the ionospheric and field-aligned currents and the Joule heating of the upper atmosphere can be determined and the thermospheric winds estimated. The very energetic precipitating particle flux and related heating in the D region can also be estimated.

The E region parameters are crucial for understanding its electrodynamics and timely models of critical ones can be mapped to the magnetosphere, and with the in situ satellite observations, help provide an indication of its present state for now casting. The magnetic Kp and other indices, which are presently used, do not provide adequate spatial and temporal information about the activity in the ionosphere or that within the magnetosphere. The heating and electrodynamics are also important for understanding the high latitude thermosphere and some associated mesospheric disturbances. This note addresses this important link, namely timely synoptic modeling of the ionosphere E and D regions from several different observational platforms for improved spatial and temporal resolution of the layer. Hence more accurate determination of (1) the ionospheric electrodynamics, (2) upper atmospheric heating and (3) related magnetospheric parameters then become possible for now casting of the geospace activity. The timely models are also important for mitigating storm effects on power lines, communication lines, pipelines and on geophysical surveys.

Magnetic storms and polar cap absorption/solar proton events (PCA/SPE) subject high altitude spacecraft to very energetic particles that cause charging and other effects on satellites (Wadham, 1987-1; Lam and Hruska, 1991; Baker et al., 1994 and see Aerospace study). Such events can lead to the failure of components and satellite systems. The storms also inflame the upper atmosphere so much that it significantly perturbs the stability and reduces the lifetime of low altitude satellites. Major storms also heat the mesosphere (Walker and Bhatnagar, 1989) and, during the winter, may alter its composition (Callis and Natarajan, 1986) and perturb the circulation of the polar vortex. However, four extended periods of solar inactivity during the 1000-1700 AD period were coincident with mini ice ages. The associated disturbances in the ionosphere are the bane of high latitude HF and some satellite radio communication and distant surveillance systems. The enhanced high latitude radiation during storms is also of concern to the new commercial transpolar aircraft flights as the increased exposure can significantly affect passenger and crew safety. The storms disrupt expensive high latitude magnetic surveys and directional drilling, while unsettled conditions significantly degrade the surveys. Storm-induced telluric currents feed into long power lines and can destabilize and sometimes disrupt the power distribution system (Boteler, 1991) and destroy expensive equipment. Hydro Quebec, following a 9 hr outage in March 1989, has installed several expensive (a few billion dollars) series capacitors in most of its long power lines to mitigate such problems and associated outages. However, other utilities do not have such protection and hence are vulnerable to storm induced currents. Super storms are expected to cause induced currents and other disturbances in the lower latitudes such as the U.S. and in southern Europe. The associated large ionospheric and induced currents in these regions will probably cause numerous outages which may extend for some time as the transformers are difficult to replace. The telluric currents can also overwhelm the cathodic (corrosion) protection systems on pipelines as well as disrupt pipeline potential surveys (Boteler, 1992). The following Bell Labs diagram depicts some of these relations. An electromagnetic pulse (EMP) from a nuclear explosion a few hundred kilometers above the earth can induce large currents in power and communication lines that can destroy significant equipment and thereby disrupt or degrade the power and communications services to a large community (Kramer, 2009).

Real-time models and now castings with maps showing the magnetic field disturbances, ionospheric and induced currents and the ionospheric and geospace activity are urgently needed by operators and surveyors to mitigate the effects of magnetic storms on their equipment and on their surveys. Aircraft crews need to be fully aware of radiation from storms so that they can lower their planes altitude to mitigate the effects on passengers and crew. Numerous geospace experiments, such as those from sounding rockets, balloon, active spacecraft and ground-based facilities, often require such timely information with maps and models showing the ionospheric activity and its associated electrodynamics for optimum observations.

TIMELY MODELLING OF THE MAGNETIC and GEOSPACE ACTIVITY

Some real-time observations are becoming available from several networks and facilities and possibly some from INTERMAGNET. Walker developed a facility for synoptic modeling of the magnetic disturbances and the associated ionospheric and induced equivalent currents over Canada (Walker, 1989) and the northern high latitude region ( Walker et al., 1997). The 55 magnetic observatories and the cap, which is centered on the eccentric dipole and used for the spherical cap harmonic analysis, are shown on the following map.

The hourly magnetic observations for 1980 were used to model the vector and total magnetic disturbance over the whole cap and also the ionospheric and induced currents. The total magnetic field disturbance for a storm on Dec 19^th is nearly 1400 nT over Alaska and the Yukon and the ionospheric currents at this time were over 2 A/m. Such timely maps are of interest to those during geophysical surveys so they can optimize their observing period and later possibly correct some of the data with values calculated from such models. The compass deviation can be several degrees in high latitudes during storms and maps showing such variations can also be calculated. These deviations are of interest to navigators and crews who are using magnetic sensors for directional drilling.

The arrows on the map below show the direction and magnitude of the ionospheric and induced currents one hour later when the storm had moved over Siberia. The ionospheric currents show where the action is in space while the induced currents are of interest to operators of power lines, communication lines and pipelines. The maximum induced currents will flow into such grounded utilities when they are aligned with the pipeline or power line.

Walker also developed and tested in 1996 for over 2 years a method for timely modeling of the magnetic activity using real-time magnetic observatory data while a volunteer sometime ago at the GSC. However, support was very limited and indeed oppressive and working conditions were also poor. It is important that this or some other real-time facility be made operational so that industry can see the extent of the activity on the maps and optimize their operations vis-à-vis the regions and periods of storms and substorms. The first expansion phase of this modeling should be to extend it over the northern USA, Alaska and eastern Siberia with the additional real-time data presently relayed by GOES West and GMS. The modeling should also be extended eastward first over Greenland with data relayed by GOES East, and then over northern Europe and eventually northern Russia with timely (<12 min.) data relayed by METOSAT. The activity seemingly moves westward with the rotation of the earth so such data and real-time maps from the models can provide an indication of approaching activity.

The present distribution of magnetic observatories in the high latitude regions is very sparse and more are needed to provide adequate coverage for the delineation of substorms and for better real-time (half-hour) modeling. About 80 well-distributed observatories would be needed for such detailed modeling. It is suggested that a program be initiated to significantly improve the coverage in the high latitude Northern Hemisphere. A second study of the high latitude Southern Hemisphere should also be made but with the objective of about 40 observatories for only basic coverage (hourly models). Note the quiet nighttime undisturbed level (Walker, 1982) should be used for the reference for accurate modeling of the disturbances and separation of the external and internal current sources. Furthermore, only magnetic observatories should be used in and above the auroral zone for reliable determination of the extrapolated undisturbed level during the long summer and other extended periods of activity.

The models determined from ground-based magnetic observations are only the so-called 'equivalent ionospheric currents' and need to be upgraded to the physical current systems. Such currents can be determined in association with ionospheric observations and models of the D and E region plasma density and its associated conductances. By combining available ionosonde, riometer and DMSP imagery and also POLAR auroral observations, basic synoptic models of the E and D ionospheric regions and its conductances can be determined. The electric fields can then be estimated using either of the methods developed by Faemark (1977), Kamide et al., (1981) or Richmond (1992) with these ionospheric conductivity models and the equivalent current determined from the magnetic observations. The Hall, Pedersen and field-aligned currents and Joule heating can then be determined. As some of these ionospheric data are also becoming available in real-time, the models of the D and E region electrodynamics and some related magnetospheric parameters could be made on a timely basis.

The main purpose of this note is to discuss synoptic modeling of the northern D and E region ionosphere from ionosonde, riometer and satellite observations. The determination of the Hall and Pedersen conductivities can then be made and also the energetic particle precipitation can be estimated. These parameters and the equivalent currents determined from the magnetic observations can then be used to determine the associated electrodynamics and the Joule and energetic particle heating. Such modeling techniques could be incorporated into the real-time (phase 2) and definitive modeling procedures of the magnetic activity to provide a more accurate measure of the geospace activity.

TIMELY MODELLING OF THE IONOSPHERIC D and E REGIONS

The accurate modeling of the high latitude ionosphere is difficult because it is generally always disturbed with both local and large scale undulations during unsettled times and orders of magnitude changes during magnetic storms. Furthermore, all the ionospheric observing techniques have significant limitations and the sparse distributions of ground-based stations in the high latitude regions results in spatial aliasing of localized disturbances. However, Walker (1989), using spherical cap harmonic analysis (SCHA; Haines, 1985, 1988) and data from the sparse Canadian observatory network, successfully demonstrated the feasibility of modeling the large scale magnetic disturbances by smoothing the data to reduce the spatial aliasing of short period but local disturbances. Note the magnetic vector observations provide 3 data points at each station while ionospheric observations provide only one data point per station for modeling (SCHA) purposes. Hence at least twice as many ionospheric observations are required for the same model resolution as for that determined from magnetic data.

The soundings from an ionosonde, the ionospheric absorption determined from a riometer and the auroral brightness observations obtained by satellites based observations could all be reduced to Chapman layers for the D and E regions and combined for improved resolution of the ionosphere. A Chapman layer is defined with the peak value and height of the density and the scale height of the layer. The peak E region electron density can be calculated from these three independent Chapman models; combined and used in SCHA to determine the spatial variations of the electron density over the region. Hence, the number of observations can be significantly increased by combining the data from the numerous riometer stations (Ranta et al., 1994) with that from the ionosondes and the DMSP imagery and POLAR auroral observations to reduce the spatial aliasing of such single point observations. The models determined from the DMSP and POLAR imagery can also be calculated at ionosonde, riometer and IS radar sites and compared to check the different modeling procedures.

Ionosonde soundings of the lower ionosphere are available during quiet to moderately disturbed conditions from about 20 stations in northern Europe and Russia and from about 12 in North America and Greenland including 7 new (and inexpensive) CADI ionosondes (Daniell et al., 1990). Note such bottomside soundings are up to only the peak electron density of the D, E and F2 regions. However, there are perhaps 60 or more riometers in the northern region, which can be used to supplement the ionosonde observations. Modeling the D and E region ionosphere over a high latitude region from several different platforms is complicated and requires the following steps.

1. Scale the ionograms and convert the frequency for the E region to height profiles of the electron density and also fit a Chapman layers to the lower half of the layer (Walker and Bhatnagar, 1989). Such modeling also provides a good estimate of the ubiquitous upper half of the E region. However, during large disturbances the ionosonde signals are scattered and another method is necessary to estimate the E region electron density.

2. A complement to the ionosonde during storm conditions is the riometer, which measures the intensity of extraterrestrial HF waves traversing the ionosphere. The absorption of these waves is particularly sensitive to ionization in the D region which, in the auroral zone, is closely related to the D and E region ionization and the auroral brightness. Walker and Bhatnagar (1989) related the absorption of these waves to Chapman layers for the D and E region electron density. Their modeled height profile for a 3 dB event agreed well with that independently determined by Kirkwood and Collis (1987) from an average of several EISCAT IS observations of 3 dB events. See figure below which compares the observations and the modeled height profiles of the electron density.

Determine the absorption by subtracting the quiet day curve for each riometer station and Gaussian smooth the data over the hour-long sample period. Next, determine the peak electron density and its height from the Chapman relation developed by Walker and Bhatnagar, for such reduced absorption periods. Note the reduced riometer observations are relative to the quiet day and hence need to be supplemented with a 'rest-absorption'. For remote sites the rest-absorption can be estimated following the method of Friedrich and Torkar (1995) for the appropriate quiet night (-O.03 dB) or day ionosphere which is dependent on the solar zenith angle. However, during unsettled conditions the height of the layer and also its scale height can be adopted from those determined from nearby ionosonde observations, as the E layer is generally similar throughout a region. Note the magnitude of the Chapman layer is determined from the riometer observations.

Substorms are the primary phenomena, which cause most of the problems in industrial facilities and on satellites and are also the most perplexing disturbance in the magnetosphere. Ionospheric observations must be made at intervals typical of the scale size of moderate substorms, which is, at most, several hundred kilometers.

The DMSP imagery can indicate the extent and brightness of the aurora. It has a resolution of about 2.5 kilometers at auroral altitudes and a swath of ~3000 km, and with 2 satellites in dawn-dusk polar orbits at any one time, can provide a temporal resolution of about 45 minutes. However, the times of the passes are irregular and the period for the synoptic modeling should initially be an hour but eventually be at half hour intervals or less to encompass the dynamics of large substorms.

3. First determine and subtract the background albedo from the DMSP images. Sample the DMSP auroral imagery at fixed locations and moderate resolution for 4 or more consecutive passes and model these data with SCHA using time terms to link up the observations. The dynamics of the aurora can be delineated from such modeling and the position and brightness of the aurora can then be calculated from the coefficients for the time dependent model at any time and place between the first and last satellite pass used in the modeling. Note, unfortunately the DMSP-imagery is presently unfiltered and the calibration step is unknown for some old satellites. However, these observations can be calibrated during stable auroral emission periods with coincident observations from POLAR.

4. The brightness of the aurora at the sampling intervals can be determined from the time dependent auroral models and related to a Chapman layer from which the peak electron density and its height can be inferred (Walker, 1972, Walker et al., 1981 and Meng et al., 1986). The POLAR UV auroral imagery compliments the DMSP observations and can also be used to monitor the auroral activity in sunlight regions. The POESS particle data might also be used to calculate the ionization rate from which an ionosphere can be estimated. The International Reference Ionosphere (IRI; Bilitza, 1990) can be used in remote regions where there are no observations.

5. The peak E region electron densities determined from the ionosonde data and that modeled from the riometer and DMSP imagery and the POLAR auroral observations can be combined and used to model (SCHA) the spatial variations of the density and the scale height of the E layer. A 35 degree (half width) cap for the SCHA that is also centered on the eccentric dipole provides a natural co-ordinate system for the modeling that also just coincides with the region of observations. By using SCHA for the modeling higher resolution can be obtained with models of low order and degree and fewer coefficients and numerical problems as the modeling region just fits that of the observations. A more dynamic and accurate E region can be constructed by using the observed peak and scale height parameters from the different platforms for the Chapman layer and SCHA. Hence reasonably sound synoptic models of the E region could be constructed over the northern polar cap, auroral and subauroral zones.

6. The Chapman functions determined for the different regions (step 5) could be used to calculate the height profiles of the electron density. Then with an appropriate model of the atmosphere, the Hall and Pedersen conductivities can be determined. Plots of the contoured modeled peak density along with the peak values of the observed electron density can be used to check for inconsistencies in the observations and the modeling. Similarly plots of the contoured height integrated Hall and Pedersen conductances can be used to check for inconsistencies in the scale heights of the electron density. These models of the Hall and Pedersen conductances are subsequently used in the solution of Poisson's equation for determination of the electric field (KRM or other methods),

7. The energetic particle precipitation and heating can be estimated from the POESS data or from the auroral models. The combined Joule and energetic particle heating can be used to infer the polar thermospheric winds. Timely plots of the aurora are of interest to those operating low altitude satellites and power distribution systems as they indicate the location of the main activity and the state of the magnetosphere.

8. Reeves et al. (1998) have devised a method to infer the inner magnetosphere temporal and spatial variations of the particle population from only a few satellite observations. The energetic particle precipitation determined from the E region models can be mapped into the magnetosphere to extend these in situ observations and provide better spatial and temporal resolution of the particle population. However, the relation between the ionospheric models and the magnetosphere particle population should be frequently 'calibrated'. It is suggested that a multibeam riometer, of at least 16 beams and possibly an ionosonde be located at each of the conjugate points of the geosynchronous satellites so that a more accurate relationship can be established between the different regimes. The E region electric fields can also be mapped into the magnetosphere to infer the direction and magnitude of the convection.

E REGION ELECTRODYNAMIC MODELLING and COUPLING

A mid-winter storm might be selected for detailed study of the proposed modeling techniques. The first step involves determining the magnetic disturbances for the period from all available observatories and selected variation stations in the northern region and determination of the E region electrodynamics. Walker has developed a procedure and facility for accurately modeling magnetic disturbances and the associated equivalent ionospheric and induced currents. It uses the quiet night-time undisturbed field for the reference level (Walker, 1982); the Dst index to remove the symmetric part of the ring current field from the observations; Gaussian smoothing of the data to reduce the spatial aliasing of sparse observations and harmonic modeling. The harmonic modeling separates the induced internal current sources from the external (ionospheric) sources. The modeling errors of the vector field range from 5 nT during quiet times to ~80 nT during the peaks of major storms (Walker et al., 1997).

The next step involves modeling of the electric field using possibly the KRM, AMIE or possibly the new rtAMIE method and the modeled conductances and the equivalent ionospheric currents. These models should be compared with those determined from the SuperDARN radar observations. The Hall and Pedersen ionospheric currents can be calculated from the conductances and electric fields anywhere and at anytime for comparison with other observations such as those from various types of auroral radars, imagers and interferometers. The currents can also be calculated along the footpath of satellite-based field, particle or imagery observations. Plots of the current vectors are useful for inferring the intensity, direction and spatial extent of the current system (Walker and Papitashvili, 1994). Maps of such vectors are of interest for studying the morphology and dynamics of the high latitude Sq, convection, storm and polar cap current systems (Papitashvili et al., 1990). The ionospheric currents are also of interest for investigating their coupling with the magnetosphere. The field-aligned currents can be estimated from the divergence of the ionospheric currents.

The field-aligned currents, the electric field and the particle flux can be mapped along the earth's magnetic field to the magnetosphere and, with in situ L1 and GOES observations, can be used to infer its status. Timely snapshot models (maps) of the magnetosphere are essential for monitoring its activity and provide an indication of the location of active regions for those operating geosynchronous and other satellites. Finally, with real-time L1 solar wind observations, these magnetospheric snapshots provide the basis for inferring the short-term dynamics (now casting) of the magnetospheric and related auroral, ionospheric and magnetic activity.

The Joule heating can be calculated directly from the Pedersen conductivity and electric fields. It is valuable for monitoring the deposition of the energy of the storms and for comparison with related phenomena in the magnetosphere, ionosphere and in the upper and middle atmospheres. The upper atmospheric temperatures and winds can be inferred from the heating and E fields and compared with observations such as those by WINDII on UARS or on ODIN. The heating in the mesosphere appears to disturb the polar vortex, which builds up each winter and links the middle atmosphere with the polar troposphere.

It would be valuable to have available "definitive" synoptic models of the magnetic activity, the ionospheric electrodynamics and the induced currents in order to investigate the effects of storms on the different commercial facilities. The models could also be used to correct magnetic surveys for temporal disturbances and for geospace research such as those of the International CAWSES program. Walker et al. (1997) have made such hourly models of the magnetic activity and the associated equivalent ionospheric and induced currents for the whole northern region from data from 55 observatories for 1980 and for northern North America for 1989 (17544 models). These simple models have been used to correct magnetic surveys for temporal disturbances (Newitt and Walker, 1990 and 1993), for investigation of induced currents in powerlines (Boteler, 1996) and for space research (Walker and Papitashvili, 1994).

References

Baker, D.N., S. Kanekal, J. B. Blake, B. Klecker and G. Rostoker, Satellite anomalies linked to electron increase in the magnetosphere, EOS Trans. AGU, 75, 410, 1994.

Bames, P.R. and J.W. Van Dyke, Economic consequences of geomagnetic storms (a summary), IEEE power Engineering Review, Nov., 3, 1990.

Bilitza, D. International Reference Ionosphere 1990, URSI-COSPAR, Report NSSDC/WDC-A-R&S 90-22, Greenbelt MD, 1990.

Boteler, D.H., Predicting geomagnetic disturbances on power systems, Eos, Transactions, Amer. Geophys. Union, 72, 14, 159, 1991.

Boteler, D.N., Telluric current effects on pipelines, Presented at International Conference on Pipeline Reliability, Calgary, June 2-5, 1992.

Boteler, D.N. and B. St. Louis, Monitoring and study of the magnetic and electric fields occurring in Quebec during geomagnetic storms, Hydro-Quebec and Geological Survey of Canada Joint Project, 1996.

Callis, L.B. and M. Natarajan, The Antarctic ozone minimum: relationship to odd nitrogen, odd chlorine, the final warming and the 11-year solar cycle, J. Geophys. Res., 91, 771, 1986.

Daniell, R.E., L.D. Brown, W.G. Whartenby, D.N. Anderson, J-A. Whalen, J.J. Sojka and R.W. Schunk, A high latitude ionospheric specification model (HLISM), in Solar-Terrestrial Predictions: Proceedings of a workshop at Leura, Australia Oct, 1989, NOAA/ERL, Boulder, CO, USA, 1990.

Faermark, D. S., A restoration of 3-dimensional current systems in high latitudes by the use of ground-based magnetic observations, Geomag. Aeron. 17, 163, 1977.

Friedrich, M and K.M. Torkar, Typical behavior of the high latitude lower ionosphere, Adv. Space Res., 16, 1 73, 1995.

Haines, G.V., Spherical cap harmonic analysis, J. Geophys. Res., 90, B3, 2583, 1985.

Haines, G.V., Computer programs for spherical cap harmonic analysis of potential and general fields, Computers and Geosciences, 14, No. 4, 413, 1988.

Kamide, Y., A. D. Richmond and S. Matsushita, Estimation of the ionospheric electric fields, ionospheric currents, and field-aligned currents from ground magnetic records, J. Geophys. Res. 86, 801, 1981.

Kirkwood, S. and P.N. Collis, The high-latitude lower ionosphere observed by EISCAT, Adv. Space Res., 7,83, 1987.

Lam, H.-L. and J. Hruska, Magnetic signatures for satellite anomalies, J. Spacecraft and Rockets, 28, 1, 93, 1991.

Meng, C.I., R.E. Hoffhlan, R.A. Skrivanck, D.J. Stficlan and R.E. Daniell, Jr., Remote sensing of the ionosphere using ultraviolet and visible emissions, Soc. Photo Opt. Instrum. Eng., 687,62,1986.

Newitt, L.R. and J.K. Walker, Removing magnetic activity from high latitude repeat station observations, J. Geomag. Geoel., 42, 93 7, 1990.

Newitt, L.R. and J.K. Walker, Correcting magnetic repeat station data with models of the field variations, IAGA, Buenos Aires, 1993.

Papitashvili, V.0., Ya.l. Feldstein, A.E. Levitin, B.A. Belov, L.I. Gromova and T.E. Valchuk, Equivalent ionospheric currents above Antarctica during the austral summer, Antarctic Sci., 2 (3), 267, 1990.

Ranta, H., A. Ranta, P. Stauning, Z.Ts. Rapport, V.M. Sinelnikov, S. Kirkwood, J. Lastovicka and A.K. Knyazev, Morphology of the D and E-region at high latitudes in the northern hemisphere during the DYANA campaign in 1990, J. Atmos. Terr. Phys., 56, 1933-1945, 1994.

Reeves, G., R. Friedel and R. Hayes, Maps could provide space weather forecasts for the inner magnetosphere, EOS, 79, 50, Dec. 15 1998.

Richmond, A. D. Assimilative mapping of ionospheric electrodynamics, Adv. Space Res., 12, 6(59), 1992.

Wadham P.N., The effects of electrostatic discharge phenomena on Telsat's domestic communications satellites, AGARD-CP-406, p25,1987.

Walker, J.K., Auroral ionosphere models, J. Atmosph. Terr. Phys., 34, 1681, 1972.

Walker, J.K., Extremely quiet night-time currents and magnetic baselines at Canadian observatories for the separation of external and internal fields, J. Geophys. Res., 87, A9, 7525, 1982.

Walker, J.K., Spherical cap harmonic modelling of high latitude magnetic activity and equivalent sources with sparse observations, J. Atmosph. Terr. Phys., 51, 2, 67, 1989.

Walker, J.K., J.A. Koehler, F. Creutzberg, A.G. McNamara, A. Valiance Jones and B.A. Whalen, Poststorm convection and auroral arc currents determined from multiple ionospheric, magnetic and electric field observations, J. Geophy. Res., 86, 9975, 1981.

Walker, J.K. and V. Bhatnagar, Ionospheric absorption, typical ionization, conductivity and possible synoptic heating parameters in the upper atmosphere, J. Geophys. Res., 94, A4, 3713, 1989.

Walker, J.K. and V.0. Papitashvili, Models of polar geomagnetic activity and equivalent currents: approach and comparison of techniques, STEP Columbia Proceedings 1992, Planet. Space Sci., 1994.

Walker, J.K., V. Semenov and T.L. Hansen, Synoptic models of high latitude magnetic activity and equivalent ionospheric and induced currents, J. Atmosph. Sol. Terr. Phys., Vol 59, No 12, pp 1435-1452, 1997.

This plan involves modest expansion of some resources so please drop me a note, thanks.