One of the outstanding problems in space physics today is the understanding of the mechanism that is responsible for the transfer of energy from the Sun to the auroral ionosphere. Single spacecraft have provided, and continue to provide a wealth of information on the microphysics in many parts of the Solar Wind-Magnetospheric-Ionospheric system. Thus, an understanding of the basic phenomena that occur at a single point in each part of the system and an average view of the system has been gained from classical magnetospheric missions thus far.

The global correlation between point measurements in vastly separate locations (the solar wind, the magnetotail, the ionosphere, and the polar cap) are being obtained by ISTP, putting to test our current theoretical and computer models of magnetospheric evolution on a gross spatial scale. As we arrive at deeper levels of understanding of each and every region, we realize that spatio-temporal ambiguities riddle our interpretations of single-point measurements. Only fortuitous simultaneous measurements from the same magnetospheric region exist from ISTP spacecraft, and those observations are limited to measurements from (at most) three spacecraft at locations that were not designed to maximize multi-point measurement operations. ISTP spacecraft are still subject to the interpretational ambiguities common to single-spacecraft missions.

The unrealized ESA-NASA mission CLUSTER and similar currently proposed missions,aspire to provide a much improved perspective of the microphysics of different local magnetospheric phenomena. They promise to measure magnetospheric currents routinely, but locally. They promise to resolve spatio-temporal ambiguities around one point of the system. They are primarily designed to study sharp boundaries at intersatellite separations of the order of ~2 RE.

Presently the only approved mission to globally monitor the magnetosphere is the IMAGE mission. Imaging the tenuous magnetotail plasma with current techniques is not possible at the short time scales of approximately a minute during which the system reconfigures. Thus, although IMAGE will provide unprecedented global information at the day side magnetosphere and near-Earth environment, similar information from the magnetotail will be lacking. In the magnetotail in situ measurements are necessary to measure the local currents, particles, and fields, but correlated measurements from many parts of the system are necessary to resolve the spatio-temporal ambiguities. NASA studies over the last ten years have repeatedly incorporated such a goal as a crucial element of space objectives. For example:

(1) The 1988 report of the National Research Council's (NRC) Task Group on Solar and Space Physics suggested that "there is no way to measure these [electromagnetic] currents and fields remotely; in situ observation programs that probe the volume of interest are required. Possible configurations for these probes have been examined and it is estimated that on the order of 300 probes are necessary for this investigation".

(2) In 1995, the Space Studies Board of the NRC, in the magnetospheric physics section of its "Science Strategy for Space Physics," stated explicitly that, after completing ISTP, "Global magnetospheric imaging is the highest priority for magnetospheric investigation". It is the contention of this writeup that multi-spacecraft observations in the Earth's magnetosphere is the only possible way to achieve a comprehensive understanding of the fundamental space physics questions. Such a network of spacecraft (pixels/probes), strategically placed in density and location according to the complexity of the magnetospheric topology and scientific questions to be resolved, would produce a 3D reconstruction of the system with adequate temporal resolution that would allow us to view the magnetosphere like a giant laboratory. This goal is directly aligned with NASA's first stated goal to achieve "Global magnetospheric imaging." It is also aligned with NASA's second stated goal to "explore, use, and enable the development of space for human enterprise", as it will enable a comprehensive understanding of the closest astrophysical laboratory to human activity. Understanding the interaction of micro- and macro-scale processes, resolving the spatio-temporal ambiguities in that interaction is the only way that we can prepare for a predictive algorithm of space weather. The exploration phase of this laboratory, which has entered its maturity with ISTP, requires now an order-of-magnitude improvement in understanding, a "leap forward" that will not only quantify the interrelationships but hand-in-hand with IMAGE can transform the very questions asked about the magnetospheric system and its interaction with the Sun. Finally, the third theme under NASA's mission statement, "to research, develop, verify and transfer advanced aeronautics, space and related technologies" is also part of such an endeavor.

The large number of pixels requires weight, power, and launch costs that are not easy to meet in today's fiscal reality. Technology development that will have the highest science return per dollar and can result in overall mission cost reduction must be identified under a holistic mission design and analysis approach. Mission objectives and science return must be compared against total mission costs in an iterative process. Although the ultimate goal of the field is to pixelize the entire magnetospheric system in a science-driven fashion, the approach ought to be incremental and iterative:

Start by identifying the essential questions that need to be answered in the field of magnetospheric research.

Perform a feasibility analysis of a strawman mission concept.

Identify and resolve the elements that can render the minimal mission possible.

Perform a cost-benefit analysis of a mission upgrade if necessary, to further the goals stated in NASA's science strategy.

2. ZERO LEVEL SCIENCE GOALS

Of all the problems in magnetospheric physics, one has accompanied the historical evolution of the field over the last 30 years, yet still remains largely unresolved: That of conversion of magnetic to plasma kinetic and thermal energy during the course of magnetospheric substorms [Dungey, 1961; Hones, 1976; Lui, 1991;Spence et al, 1996;Stern, 1996]. Knowledge accumulated to date strongly indicates that the difficulty lies in the localization of the phenomena both in space (to scale sizes of 1-2 RE) and in time (1-2 minutes) despite their large-scale consequences over longer time scales. A single spacecraft within the vast magnetospheric system has little chance of being at the key region at the time of substorm onset. More importantly, the lack of simultaneous measurements from adjacent "pixels" of this vast system prohibits placing the observed phenomenon in the context of the global magnetotail evolution.

In the following, we present the zero level questions and document the necessity for a multi-probe investigation of the magnetotail. The most probable state of the magnetotail and the conditions arising during magnetospheric substorms are the zero level mission scientific objectives.

2.1 Substorm Chronology: Substorms are defined on the basis of their ionospheric signatures [Akasofu, 1964;Rostoker et al., 1980;McPherron 1991], but they have repeatable near-Earth magnetospheric signatures. These are: the energization of particles at geosynchronous altitude [McIlwain, 1974] and the elevation increase of the magnetic field in the near-Earth nightside magnetosphere [McPherron et al., 1973]. Contrary to the ionosphere, where a global picture can be obtained with all-sky cameras, radars, and magnetometer networks, in space we depend on statistical analyses to compose a time history of events. Such analyses and a few fortuitous multi-spacecraft conjunctions, we have learned that the field dipolarization starts at a localized region of the order of 1-2 RE [Ohtani et al., 1992] and propagates longitudinally [Nagai, 1982] at a speed of an hour of local time every 1-3 minutes and radially tailward [Jacquey et al., 1993; Ohtani et al., 1991 and 1992] with a speed of ~250 km/s, or 2.5 RE/min. Tailward flows are also observed during substorms, most often at distances greater than 19 RE, while Earthward flows are observed closer to Earth [Baumjohann et al., 1990;Angelopoulos et al., 1994]. These flows have been interpreted as signatures of magnetic reconnection. The reconnection site also moves downtail during the course of a substorm [Sergeev et al., 1995] at speeds comparable to the speed of the current disruption.

The time and location of the first magnetospheric signature of substorm onset and its relationship with the ground observables still remains a mystery. A careful resolution of this point, in turn, inhibits theoretical progress in the understanding of the physics that governs the phenomenon. For example, Coroniti [1985] and Baker and McPherron [1990] suggest that, on the basis of the near-Earth neutral line model, tailward flows should start before ground substorm onset. However, Lui [1991] suggests on the basis of the current disruption model that tailward flows are the consequence of near-Earth current disruption and thus reconnection should then start after substorm expansion phase onset. The "strong version" of the Kiruna conjecture [Kennel, 1993] has been that not only this remains an open question but that there is also the possibility that the two phenomena are entirely uncorrelated. The resolution of this issue requires sufficient (less than 1 minute, the Alfven bounce time) temporal resolution and sufficient (~1-3 RE, the scale size of the initial active region) spatial plasma sheet coverage on the X-Y plane.

The maximum altitude of the spacecraft necessary to capture the reconnection process is determined from previous statistical studies of fast plasma sheet flows. Such studies [Hayakawa et al., 1982; Nakamura et al., 1994; Hones and Schindler, 1979] indicate that at a distance of ~35 RE anywhere between 25% and 80% of fast flows observed were tailward, the variability being due to the selection criteria and methodology. Recent GEOTAIL results place the most likely position of observation of the near-Earth neutral line at around X=-35 RE [Nishida et al., 1996] while they confirm that on the average the flows are tailward at distances of 30-60 RE [Frank and Paterson, 1994]. Since tailward motion of the neutral line occurs, the neutral line is likely to start most often Earthward of X=-40 RE at substorm onset. Thus it is in the region between 7-40 RE that onset and evolution of current disruption and near-Earth neutral line can be intercompared during the time of expansion phase onset of magnetospheric substorms. Many tens of spacecraft at cross-tail separations ranging between 1 and 20 RE and at downtail position between 7 to 40 RE are a necessary and sufficient condition for the study of the magnetospheric evolution during the course of a substorm. Quantities necessary and sufficient are high time resolution (1-5 s) magnetic field and 3D ion moment measurements.

2.2 Substorm Current, Energy and Magnetic Flux Budgets: Substorms are a primary mechanism of energization of the nightside ionosphere and the ring current. Although the ultimate source of energy during substorms is the solar wind, the transfer of that energy to the plasma sheet particles is not well understood. Modeling of the geosynchronous injections during substorms as fronts of Earthward-collapsing particles [Mauk, 1986] based on measured electric fields observed in the near-Earth plasma sheet [Aggson et al., 1983;Maynard et al., 1996] has produced fairly good agreement between theory [Quinn and Southwood, 1982] and observations. In that scenario it is the braking of the Earthward flows [Haerendel, 1992] resulting from either the current disruption or the reconnection process that causes the near-Earth signatures of substorms. In the magnetotail tailward of X=-15 RE, and close to the neutral sheet, fast plasma sheet flows are responsible for most of the plasma sheet energy, particle and flux transport [Angelopoulos et al., 1994]. Such flows are most often observed within |Y|<10 RE from the midtail axis [Baumjohann et al., 1990]. Are the fast flows that are observed at substorm onset of sufficient energy to account for substorm energization of the near-Earth environment and the ionosphere? The answer depends on the 3D evolution of the magnetotail flows. Given the extreme localization in Y of some of the fast flow events observed so far [Sergeev et al., 1996;Angelopoulos et al., 1996;Angelopoulos et al., 1997], simultaneous multi-point measurements over an extent of dY ~ 20 RE and at a probe density of 1 measurement per 1-3 RE in Y are necessary.

The longitudinal spread of the substorm current wedge (SCW) [Nagai, 1982] is coupled to the evolution of the westward traveling surge in the ionosphere [Roux et al., 1991; Robert et al., 1984; Maynard et al., 1996]. The extent in Y of the SCW at distances far from geosynchronous is neither known, nor possible to document with single spacecraft because it is a three dimensional system.

Lopez and Lui [1990] used AMPTE/IRM and CCE observations to show that the SCW is irregular in space. In addition, substorms take place most often as a series of consecutive, isolated activations. The near-Earth and ground response is the integral response to the resulting consecutive elemental current wedges. Such individual activations have been proposed to be the elemental processes [Lui et al. 1991a; Sergeev et al., 1986a, 1986b] that compose substorms. Multi-spacecraft observations will result in resolving the ambiguities associated with the number, location and extent of such activations. In fact, ISTP studies [Angelopoulos et al., 1997; Slavin et al., 1997; iFairfield et al., 1997] are already providing direct measurements of the extreme localization of the elemental activations and their relationship to the ionosphere. Despite the limited number of tail near-by conjunctions of ISTP spacecraft, is it possible to get with the above studies a glimpse of the complexity and richness of the magnetotail plasma when studied at small spatial scales. Perhaps, though, the most important lesson learned from ISTP is the scarsity of fortuitous conjunctions and the need for a distributed satellite system even when the goal is the study of the microstructures. The subliminal, and prevailing, understanding is that the current wedge maps to the entire width of the aurora and that its outermost boundary is the location where the current feeds from the magnetosphere to the ionosphere. Thus, it is a very important key in the chain of energy from the plasma sheet plasma to the ionospheric Joule heating. Understanding the three-dimensional evolution of the substorm current wedge system promises to solidify estimates of the magnetospheric-ionospheric energy coupling during a substorm and in addition probe simultaneously its component activations and their ionospheric mapping.

2.3 Current Sheet Thickness and Evolution: Present theories that attempt to address the substorm onset question on the basis of current sheet instabilities require direct measurements of the current sheet thickness and intensity [e.g., Lui 1991b, Lui et al. 1991, Wang et al. 1990]. Attempts to model the current sheet thickness and density by inverting the observed magnetic field measured at more than one locations during fortuitous conjunctions of spacecraft has lead to some success [Pulkkinen et al., 1991]. In addition, current measurement techniques in the tail using two spacecraft measurements [McComas et al., 1986] have given incontrovertible evidence for the existence of thin current sheets and their importance during substorms [Lin et al. 1991]. However, the only possible way to ensure that the tail current system is adequately monitored at the time and place where it actually becomes unstable is to simultaneously monitor its thickness at many different downtail distances. This necessitates a CLUSTER-like system at many different distances. One way to achieve that is to provide some orbits with enough separation in Z (~1-2 RE) at different downtail distances. For current sheet densities of the order of 1 nA/m2 which are typical of the cross tail current, and intersatellite separations of 1 RE such measurements can be made comfortably with 0.5 nT accuracy i.e. well within the capabilities of conventional magnetometer designs. Vertical separations of the order of 1 RE are necessary but also sufficient for such observations.

2.4 Modeling and Variability of Magnetospheric Currents: Global magnetospheric models, with the exception of global MHD are static [Tsyganenko, 1989; 1995]. The large magnetic field variability in the databases that are used as input to those models [Stern and Tsyganenko, 1992] suggest that a significant part of the physics that determines the most probable magnetospheric state is missed. Most often, severe modification of the model currents is necessary away from expected state [i.e., for the measured Kp or AE parameter] when actual time-dependent situations are considered [Pulkkinen 1991b, Baker et al. 1993]. It is impossible for a static model parametrized on global indices to capture the complexity, the richness and the physics of the instantaneous magnetospheric configuration. However, a large array of measurements from magnetospheric probes can produce, when inverted using techniques available from statistical [Tsyganenko and Usmanov 1982], simulation [Berchem et al. 1995b] and atmospheric [Ghil and Malanotte-Rizzoli] models, an instantaneous "image" of currents and fields that is consistent with the average tail structure but representative of the instantaneous measured fields. This can produce a far clearer picture of the magnetospheric current systems, their sources and sinks and their temporal evolution. Parametrization versus substorm phase and upstream conditions can produce the "typical" substorm sequence; i.e., a time history of the magnetospheric system's response to qualitatively different drivers.

2.5 Fate of Accelerated Particles: The three-dimensional circulation of the substorm-accelerated particles is not clearly understood. Huang et al. [1992] have shown that plasma sheet temperature increases after substorm onset represent an important plasma sheet response to geomagnetic activity that does not always correlate with flow enhancements. Although single particle approaches [Spence 1993a, 1993b; Ashour-Abdalla et al., 1994] have led to certain predictive pictures of the plasma sheet pressure, temperature and density spatial profiles, such profiles have not been reported so far except in energetic particles [Krimigis and Sarris, 1979]. Thus there is a need to observe the system and its evolution at many points simultaneously for a given thermodynamic state. This entails fine spatial resolution observations of the instantaneous flow pattern, density and temperature to observe the source and propagation of the resulting heating. The reason for the fine spatial resolution is twofold: First, the acceleration regions are localized in space [Krimigis and Sarris, 1979; Sergeev et al. 1986, 1996;Angelopoulos et al., 1997]. Second, the plasma sheet flow and thus particle trajectories are variable at inferred spatial scales of ~2 RE (see section 2.6), such that it is unlikely that two spacecraft will have many chances of crossing the same flow line at a given time. A large enough probe density and high-time (~5 s) resolution distribution functions are necessary to resolve this problem. Thus burst-mode operations at times of substorm onset must be provided for in the zero level mission.

2.6 The Most Common State of the Plasma Sheet: Although the most dramatic plasma sheet phenomena occur during short periods, at a good statistical correlation with substorm activity, the most probable state of the plasma sheet (90-95% of the time) is the non-flowing state [Baumjohann et al., 1989, 1990]. That state exhibits flows in any direction with equal probability and with a peak-to-peak amplitude may times larger than the average flow itself [Angelopoulos et al. 1993]. Despite the large flow variability, the average flow pattern exhibits behavior that can be understood in simple 2-fluid theory [Angelopoulos et al. 1993] or single particle orbits [Spence et al. 1993a, 1993b]. Significant variability about the magnetic field average also exists [Stern and Tsyganenko 1992], rendering orbit integration schemes vulnerable to criticism [Cattell et al. 1995]. Is the variability the consequence of nearby high speed flows [Angelopoulos et al. 1994, Borovsky et al. 1996] or the result of complex ion trajectories under steady external conditions [Ashour-Abdalla et al 1995]? Velocity autocorrelation times are of the order of 2-3 minutes and mixing lengths of the order of 1-2 RE. This suggests that although large scale eddies are occasionally observed in the plasma sheet [Hones et al., 1978, 1981] underlying effective mixing also takes place at shorter scale lengths. In order to characterize this state of the plasma sheet correlative measurements at many points of the system are necessary in order to measure the wavelength spectrum of the plasma sheet flows. This will tell us if the variability is generated by the substorm process of it is an inherent property of the convecting plasma sheet. Distribution functions are necessary in order to ascertain whether the flow variability is due to particle bunching or a velocity shift of the entire particle distribution. As the time scale of propagation of an ion of 1 keV energy is ~4 RE/min we can monitor the propagation of such ions between satellites separated by a distance of 10 RE within 2 minutes. Thus, it is important to operate at a burst mode of collection of distribution functions from all spacecraft during periods of 2-10 minutes, but not necessarily at times of substorm onset. Burst mode times should be selected on the basis of maximum plasma sheet coverage.

2.7 Plasma Sheet Thermodynamics. Attempts to characterize the thermodynamics of the plasma sheet (such as measuring its polytropic index) have been so far inconclusive [Baumjohann and Paschmann 1989, Huang et al. 1989, Goertz and Baumjohann 1991]. Room for controversy exists because with single spacecraft we cannot observe the evolution of a single flux tube. Multi-probe studies that utilize two (or more) spacecraft traversing plasma streamlines will be able to answer the above question. However, discoveries from such a multi-probe mission may result in a complete change of mindset regarding this issue, since if the plasma flow is non-laminar the concept of a streamline may be invalid.

3. THE ZERO LEVEL MISSION

3.1 Strawman Mission Concept: Overview The zero level science goals can be met with a fleet of 60 low-cost, low data rate satellites (magnetospheric probes). The number of probes was chosen based on the average density of plasma sheet coverage of ~3X3 RE2 per probe in an area between -10<X<-40 RE and -10<Y<10 RE and a near-random distribution of probes in their final orbits. Each magnetospheric probe can be instrumented with a vector magnetic field and a three-dimensional thermal ion distribution sensor and can be spin stabilized at a period of 1-5 s. Low data accumulation rates will result in low transmission power requirements at perigee (storage and dump).

Figure 1. GEI system projections of six selected zero level mission orbits on March 21, 1997.0000 UT. Perigee can be between 2.5 and 3.5 RE, to ensure an acceptable transmission rate, adequate visibility from two ground tracking stations and a relatively benign radiation environment. Two sets of orbits, at two different sets of orbital inclinations and major axis directions were chosen to meet the science requirements. The two sets of probes can be launched in their final orbit by two ion engines. Each ion engine can take a stack of 30 probes from the same elliptical transfer orbit of 185X45000 km achievable by a Taurus with two Castor 4B strap-ons (Med-Lite configuration) at a cost of 25 M. Each ion engine can release the probes in its ascend to the final orbit of highest apogee within 80 days. The orbital periods are sufficiently different that the orbits will randomize in phase within one month from their placement in final orbit. Each probe can have its own command/telemetry system, power subsystem and DPU. Significant cost and weight reductions result from (1) Flexible mission requirements on orbit parameters, (2) minimal ground operations costs, (3) minimal requirements on probe spin and attitude determination, (4) use of already proven instrument technology.

3.2 Mission Architecture

3.2.1 Orbit Characteristics. The orbits were chosen for the purposes of this strawman mission on the basis of the following requirements: 1) The orbits should have a maximal probability of being inside the plasma sheet, if not at the neutral sheet. Such orbits require a small distance of the model neutral sheet (ZNS) from the X-Y GSM plane, otherwise a significant part of the orbit can remain far from the expected neutral sheet position. Since ZNS is a function of dipole tilt, we demand that the apogees be at midnight at equinox, i.e., an epoch which is associated with minimum daily average tilts. 2) The orbital planes should cross the neutral sheet at times when ground magnetometer and radar coverage is maximal, i.e. when the US sector is near midnight (i.e., at 0630 UT). At that time and at vernal equinox the dipole tilt is approximately -10 degrees. 3) The two sets of orbits should have a separation in the direction of their major axes of about 30 degrees to allow adequate simultaneous cross-tail coverage 4) The two sets of orbits should have a Z-separation of 1-2 RE. 5) The orbits should avoid Earth's shadow for periods longer than 1-2 hours to avoid complex thermal and electrostatic designs. This entails tilts larger than about 5 degrees.

Figure 1 shows six orbits in the X-Y and X-Z GEI coordinate system. The rest of the orbits (54) have intermediate orbital parameters between those in the figure. Table 1 shows the orbital elements of the six orbits of Figure 1. It is evident that both sufficient longitudinal coverage and adequate intersatellite separation of the plasma sheet is feasible, in accordance with the science requirements. Figure 2 shows 60 probes, 3 months after launch assuming (for demonstration purposes) that they all have perigees of 3 Re, apogees between 12 and 41 Re, increasing in 1 Re steps and inclinations varying from 13 to 11 degrees and 7 to 9 degrees in the two stacks respectively. The probes of Figure 2 were all placed for demonstration purposes at their perigee at the same time (Winter Solstice, 1996) and then were propagated to their positions 3 months later. A complete mixing of the orbits due to the different orbital periods is evident. This results in even plasma sheet sampling. The Tsyganenko model for Kp=+3 on March 21, 1997 at 0630 UT is superimposed on the same plot. It is evident that adequate neutral sheet coverage is also possible by these orbits twice per day.

Table 1. Orbital Information for orbits of Figure 1

Apogee (RE)	Inclination (deg)	Ascending Node	Argument of Perigee	Period (days)
12	7	285	90	1.205
12	13	255	90	1.205
25	3	285	90	3.074
25	7	255	90	3.074
42	2	285	90	6.262
42	4	255	90	6.262

3.2.2 Operational Lifetime. Three months of data collection in the magnetotail will have completed the basic mission science objectives within 8 months from launch, including 5 months for achieving the final orbit configuration (3.2.10). However, significant science returns can be achieved at minimal additional cost as the flanks and dayside magnetopause are being monitored. It is highly desirable to have a total mission duration of about 2 years for the following reasons: (1) Mission operations are a relatively low cost of the total mission (section 3.2.8.3). (2) Total radiation dosage does not increase significantly due to the high perigee of the orbits, and (3) It is important to maximize the regions covered by this unique multi-probe investigation. Magnetopause and low-latitude boundary layer science objectives are provided below (section 3.5). (4) Sampling of the magnetotail again during the second year of operations can double the amount of data related to the zero level science.

Figure 2. GSM projections of 60 probes on March 21, 1997, 06:30 UT, along with Tsyganenko [1989] field lines for Kp=+3. Plasma sheet coverage is maximal when US is at midnight.

3.2.3 Instruments.

3.2.3.1 Magnetometer. Magnetic field measurements are necessary at 1-5 s resolution with spin axis and angle knowledge of ~0.1-0.2 degrees, sensitivity of ~0.1 nT and range from 0-2000nT. The strawman mission magnetometer consists of an orthogonal triad of sensors on a fixed boom (30 cm). The electronic design is similar to that used on the FAST Small Explorer mission. The three vector components can be measured to 16 bit accuracy in two ranges: A 2000 nT range with a resolution of 0.03 nT and 200 nT range with a resolution of 0.003 nT. The mass of the main electronics is 0.4 kg and the sensor triad plus cabling is 0.1 kg. The electronics unit draws 0.5 W of refined power. The noise of the system is less than 2x10-5 nT2/Hz at 1 Hz. By minimizing range changes and using digital filters, the magnetometer is both simple to calibrate on the ground and easy to operate in space. The operation and data reduction is also optimized by the high linearity of the system to better than one part in 105. On board spin averaging can be performed with a simple electronic addition and subtraction operation using the sun pulse as reference at a negligible power cost. A data collection rate of 50 bits per spin is required for magnetometer measurements. The magnetometer boom structure, built out of graphite epoxy, weighs less than 100 grams including wiring.

3.2.3.2 Ion Electrostatic Analyzer A top-hat ion electrostatic analyzer with a heritage from AMPTE/IRM, GIOTTO, WIND, CLUSTER and FAST can perform measurements of ions in the range 3 eV-40 keV. The detector geometry is split in half to reduce weight and power. This configuration has a 180X14 degrees field of view allowing measurement of a full 3D distribution function once per spin. Ions entering the analyzer are selected in energy and imaged onto a microchannel plate (MCP) at 8 separate angles. The MCP consumes 200 mW of power. The ions are post accelerated by a -2500 V potential applied between the front of the microchannel plate (MCP) and a grid located 1mm above the MCPs, to increase efficiency. High voltage sweeping consumes ~100 mW. The preamplifiers consume 25 mW each for 8 angles. The power supply conversion efficiency can consume another 100 mW. Using a 20% margin on total power we get power consumption less than 750 mW. Measurements are performed at 16 energy steps with width dE/E=0.78 and a geometric factor of 4x10-2 cm2-sr-eV. The dynamic range of this instrument is 10-109 eV/(cm2-sr-eV-s). Raw data in full distribution function is 16384 bits. This is fed to the data processing unit which performs the moment computations. The basic quantities of density, velocity, and pressure tensor give us ~104 bits per measurement.

Table 2. Science payload and data accumulation for orbits of Figure 1

INSTRUMENT	TYPE	WEIGHT	POWER	HERITAGE	data rate +	Max Data per Orbit ++
Magnetometer	Triaxial fluxgate, boom mounted	0.5 kg (sensor + electronics) 0.1 kg boom	0.5 W	ATS-6, ISEE2, GALILEO, FAST	17 bps	1.0X10^7
ION SENSOR: *Moments: ........... **3D Distribution every 9 minutes for validation ***Burst mode: A: Twice per day for 10 min every 6s. B: Command upon ground event to collect for 20 min, or trigger on local properties. C: Trigger on local properties routinely near boundary	Top hat, 180 degree field of view, 22.5 degree angular resolution, 16 energies (3 eV - 40 keV).	1.0 kg	0.750 W	AMPTE/IRM, WIND, CLUSTER, FAST	*35 bps **30 bps ***3.2X10^6 bits/day	*2.0X10^7 **1.7X10^7 ***2.1X10^7
TOTAL		1.6 kg	1.250 W			68Mbits/orbit

+ Assuming a 3s spin period.
++ Assuming the highest perigee orbit with period ~6.5 days.

3.2.4 Data Acquisition Following the science requirements from the previous section we assume magnetic field and ion moments collection at 3s resolution (1 spin). Routine transmission of ion distributions is necessary every ~5 minutes for ascertaining the validity of the on-board ion moment computations, and studying individual events in detail. Burst mode storage of ion distributions is necessary when most satellites are positioned in the plasma sheet at a high rate (6 s) for a period of ~10 minutes. This mode can be routinely performed at times that the expected neutral sheet crossings occur on the basis of a model. We call the above mode "Burst Mode Type A". A different type of burst mode ("Type B") can be designed on the basis of real-time ground activity monitoring. Burst mode data can be stored on board for a total of 20 minutes during which interval there is enough time to command the spacecraft from the ground with minimal telemetry requirements to start taking data surrounding the first "active" plasma sheet encounter. "Active" is defined on the basis of an on-board trigger (such as magnetic variability or flow magnitude. For a given ground campaign interval, the moment that a substorm takes place the probes can be commanded to keep the burst mode data that bracket the "trigger". Finally, a "Burst Mode Type C" can take burst data for 5 minutes near an expected region encounter on the basis of on-board triggering. The latter mode can be especially useful at the magnetotail boundaries.

The total data budget for the mission is shown in Table 2. A total of 68 Mbits can be stored in the highest apogee orbit, (6.5 days) including burst mode data. Each day has two burst mode type A collections (two neutral sheet crossings) or one type B collection, or a series of type C collections. Lower spin rates and data compression can reduce the total amount of data significantly. We have considered data compression using the Rice compression scheme [Rice, 1979; Rice and Lee, 1983]. This is currently being used for the TIMAS instrument on the POLAR satellite (W. Peterson, private communication) and can achieve a factor of 4 compression on data from that instrument. The memory allocation that becomes available through compression can be used either for higher sampling rates or to decrease the total data transmission time.

3.2.5 Data Storage. Data storage requires a maximum of 80 solid state memory chips (1 Mbit each) at a weight of 2 grams each and a total of 160 grams including a hardware error correcting code. Moment computations, data compression and signal conditioning can be done on the Data Processing Unit (DPU). This entails 2 programmable gate array chips at a total weight of 10 grams and a central computational unit of 50 grams. Additional computations performed are instrument control, command and data transfers to sensor electronics, data packetizing, telemetry formatting, command decoding and power control. An additional 100 grams for the board results in a total memory plus DPU weight of less than 250 grams. The power required for the memory is 0.2 W based on FAST technology, while DPU power is 0.1 W. The above numbers are an overestimate, as new technology memory chips, already available can result in further mass and power reduction.

3.2.6 Data Transmission. Data transmission can be achieved at perigee (3 RE perigee, or ~13000 km) using an omnidirectional transmitter in the S-band (2.3 GHz) with output power 2 Watts, and collected from a 10 m diameter ground station at a rate of 50 kbps. The telemetry system envisioned comprises a turnstile antenna with a hinge mechanism approximately 10 cm above the main body of the probe and a rigid coaxial cable providing the link with the transmitter. Alternatively a microstrip antenna such as the one flown on Firewheel can be deployed from the main body. In either case the total antenna weight is approximately 50 grams. The transmitter can operate at 28V voltage and drawing 0.8A maximum current with 2 Watts output power. We have identified a company that produces such a transmitter at a weight of 113 grams; more than one companies are currently being considered. Even in the absence of data compression the total time required for data transmission is 25 minutes. This requires a total energy of ~10 Wh which has to be supplied by the battery. The commercial receiver envisioned operates using a similar antenna (additional 50 grams) at a frequency of 500 MHz and weighs 315 grams. Multiple command receiver vendors are currently being considered.

3.2.7 Power. Powering transmitters with a nominal 10% efficiency using conventional Cadmium batteries would be exceedingly heavy for the mission. We seek alternative solutions in Li-Metal battery technologies, which are driven by the cellular telephony industry. The rechargeable AA size Lithium battery type TLR-7103 manufactured by Tadiran can operate nominally at 2.8 V and draw 250 mA of current. It can operate in the range -30 to 55 degrees Celsius and provide 500 cycles of charge/discharge operations at 50% depth of discharge (DOD). It weighs 17 grams. As we need 28 Volts and 0.8 Amps maximum current for nominal transmitter operation we need to stack 10 such batteries in series and have three stacks (30 cells total) of a total weight of 510 grams. The nominal battery capacity is 0.75 Ah. This gives us a total of 63 Wh. Operating at DOD of 16% we can achieve the transmitter requirement of 10 Wh at a comfortable margin to allow for battery degradation or partial cell failure. Given one battery discharge (data transmission) per ~6 days this allows a lifetime of 3000 days, i.e. much longer than the mission lifetime. Although stacking the batteries in a 28 V configuration has not been tested yet, the company plans to complete the tests within the next year. Simultaneous efforts to push the performance envelope of Lithium batteries have been undertaken by companies under NASA funding (see for example the NASA funded project by Lithium Energy Associates, Inc. at http://sbir.gsfc.nasa.gov/95abstracts/10.02/950062.html). A part of the tasks than ought to be accomplished before this strawman mission is considered viable is the study of the Lithium battery performance.

According to the BITSY micro-satellite numbers (built by AeroAstro), with GaAs cells we can achieve power to mass ratios of 66 W/kg. Most of the solar cell weight is in the coating, and this number depends heavily on total radiation dosage. BITSY is a low altitude satellite with mission duration 6 months. For a high perigee mission like the one we are proposing we expect that the total radiation will be smaller than on a low altitude orbit. We assume a uniform coverage of solar cells in a spherical configuration. This theoretically entails 4 times the mass for a given power output since the sun only illuminates part of the surface. Assuming solar cell performance degradation of 40% at the end of the mission, the factor of 4 in reduction due to the overhead from uniform coverage, and an additional 25% reduction in solar power resulting from instrument, antenna and radiation openings, we can get 7.92 W/kg out of BITSY-type solar cells. For our power needs of 2.834 W we require 0.358 kg of solar cells.

The conversion efficiency of GaAs cells is 130 W/m2 at the end of mission. To get 2.834 Watts, and assuming a factor of 4 overhead in area due to incomplete solar illumination of the solar cells we require a surface area of 5x218 cm2 = 1090 cm2. This is the surface area of a spherical shell of radius 9.3 cm. The volume of the structure is therefore commensurate to the scientific payload.

The solar cells should be mounted on a graphite epoxy substrate, spherical in shape for stiffness, and less than or equal to 0.5 mm in thickness. For the kind of structure we envision this entails a total mass less than 75 grams, including mounting posts on the spacecraft body.

3.2.8 Mission Operations Requirements

3.2.8.1 Spacecraft operations. Spacecraft operations are minimal. They are limited to simple instrument control commands, uplink of burst mode operations, simple power, thermal and telemetry, and uplink of a Burst Type B bit, in the case of a campaign mode operation. However, work on this area is necessary because never before such a mission has been attempted. It is imperative that a detailed description of the mission operations requirements of this strawman mission be detailed, and a good estimate of the expected cost be derived. Mission operations can kill such a mission through cost overuns, if care is not taken from the design phase to accurately describe and minimize them.

Figure 3. Ground Tracking Schedule from 10M Antenna Stations: Stanford (Open Boxes) Malinidi (Filled Boxes).

3.2.8.2 Ground Station Visibility. The question whether a single ground station can perform the major part of the tracking can be answered with the following experiment: Assuming a random set of orbits with perigee at 3 RE and apogees ranging from 12 to 42 RE, how many can we track within a given period of ~6.5 days (longest period orbit)? We performed a simulation of the above question after launching the spacecraft from perigee on March 21, 1997, allowing two months for orbit randomization. Figure 3 shows in open boxes the times that a probe was tracked by the 10 m antenna at Stanford versus probe number (increases with apogee). What is shown is essentially a ground station tracking schedule for a given 6.5 day-period.

A successful probe contact in the above figure constitues 30 minutes of tracking (probe visibility 10 degrees above the horizon) and 1 additional hour for ground station functions. With a single 10 m dish at Stanford we were able to track 46 probes during any 6.5 day interval tested (succesful contacts are marked by open boxes in Figure 3). Probes that were not tracked by the Stanford station were predominantly those that did not make ground contact over the 6.5 day period considered (mostly highest apogee). Those can be tracked from a station approximately 180 degrees geographic latitude away from Stanford, such as Malindi, Kenya (currently operated by ESA). By using station Malindi to track the remaining probes (filled boxes in Figure 3) we were able to retrieve nearly 100% of the data (small losses still exist at the rate of 6% per year for the highest appogee orbits). We conclude that 2 ground stations can achieve almost complete mission data retrieval.

3.2.8.3 Ground Operations. Simple probes do not necessitate complex ground operations. These operations are: schedule the tracking by allocating priority to probes with highest memory storage between previous and next data dumps; turn-on two way communications; verify spacecraft health and status; download and validate data, uplink burst mode schedule and housekeeping commands, turn-off two-way communications, and update orbital elements, spin and attitude information.

The above operations can be performed out of automated ground operations facilities with a maximum of 2 persons present at a given time, and would be insensitive to personnel absence or turnover. Either fully automatic TOTS antennas or existing ground facilities can be used at a cost of ~1M/groundstation/year.

3.2.9 Spacecraft Mass and Power. Table 3 presents the strawman mass and power estimates of the strawman payload. The command receiver can be turned off in the data collection phase but in the event of malfunction it will be in the "ON" mode. We thus include it in the power budget with a pulsed, 10% duty cycle.

Sun sensors required for attitude determination (0.1-0.2 degree accuracy) weigh around 100 grams and consume around 10 mW of power. Total power consumption is (using a 30% margin) 2.834 Watts.

Harnessing required is wires (0.5 lb/ft), and connectors (20 gr/connector). Wiring will weigh <80 gr (320 ft) and connectors 120 gr (6 connectors). Harnessing will be kept to a minimum by using a backplane for most interconnections.

Care should be taken to ensure thermal isolation of the battery. Active heating can allow survival of the battery during eclipses. Because of the orbit design specifically to avoid prolonged eclipses, this is not expected to be a severe problem in the mission's thermal design. Battery and transmitter can be mounted near the probe's skin, for thermal isolation and stability. Heat dissipation from the 23 W transmitter is not expected to present significant design problems. A total of 100 grams per probe is thus allocated to thermal blankets.

Assuming 100 gr per probe for a release spring and a solid rocket for spinup and adding a 30% margin, we get a weight of ~5.2 kg per probe.

3.2.10 Launch Weight and Orbit Attainability Possible launch configuration is to launch two stacks of 30 probes, with a Taurus rocket and then use another propulsion system (one per stack) to attain the final orbits. A preliminary analysis indicates that a solid rocket would require doubling the total mass of the payload to achieve the required delta V. However, an ion engine would weigh significantly less. At the same time an ion engine can achieve the desired gradual increase of the apogee and the release of the probes one at a time at different apogees. The difference in apogees will randomize the orbits. A dispenser system appropriate for the probe release has to be designed. Each probe can be released to its final orbit by a spring and be spun at a frequency of 1-3 rev/s by a small solid rocket. Although it is necessary to know the spin axis orientation and spin period (to allow 0.1-0.2 degrees accuracy in knowledge of probe orientation), it is not necessary to control it, resulting in no need for active attitude control. Omnidirectional antenna for transmission and nearly uniform solar cell coverage of the satellite would allow any satellite orientation. Below we describe the components of the system envisioned.

Table 3. Strawman Probe Weight and Power Estimates

ITEM	WEIGHT	POWER (Duty Cycle)
POWER SUBSYSTEM: Battery (28 V, 0.750 A, 63 Whs, Li-Metal) Battery mount/radiator/heater Battery charging/conditioning; Low voltage inverter	0.510 kg 0.100 kg 0.200 kg	0.200 W when charging 1.000 W (at 1% duty cycle drawing from battery)
ATTITUDE SUBSYSTEM Sun Sensor/Mounting	0.05 kg	0.010 W
COMMAND/TELEMETRY Transmitter Command Receiver Antenna	0.428 kg	20. W [10% efficiency, duration: 30 min/6.5 days) 0.420 W [4.2 W, 10% duty cycle]
SCIENCE PAYLOAD	1.600 kg	1.250 W
DPU MEMORY SUBSYSTEM	0.250 kg	0.200 W
MECHANICAL SUBSYSTEMS Solar Cells (Body mounted, Uniform coverage) Solar Cell Substrate Thermal Blankets Harnessing	0.358 kg 0.075 kg 0.2 kg 0.2 kg
Solid rocket for spinup; Release spring	0.1 kg
TOTAL	3.971 kg	2.180 W
TOTAL+30% MARGIN	5.162 kg	2.834 W

The Taurus XL launch vehicle can comfortably achieve a 185x45000 km orbit at 10 degrees inclination if launch takes place from Kourou or Brazil for a payload of 480 kg. The Taurus XLS can achieve a similar orbit if launch takes place from Cape Canaveral for a mass of 340 kg. Larger mass can be achieved from Cape on Taurus XLS if the fairing size is reduced or the orbit inclination is closer to 28 degrees. How far away from an equatorial orbit can we be in order to not sacrifice the science objectives and be able to launch on a fairly inexpensive vehicle is a matter that requires further study. Alternative orbits and maximization of ion engine capabilities to achieve a low cost launch vehicle should be part of the study beyond this current writeup.

According to NASA/Lewis engineers the NSTAR SEP ion engine thrust depends linearly on input power. Input of 2.5 mW results in ~90 mN of thrust. For our purposes 1.25 W are sufficient. We also assume a non-spinning vehicle with active attitude control for good solar panel illumination. BMDO and NASA are funding development of a new type of solar cells that can achieve higher power over weight performance specifically for the NSTAR ion engine. The system is based on triple junction arrays with concentrators. One wing can produce 1.25 kW of power and weighs (including mounting) 28 kg. Adding 15 kg for the ion engine and 10 kg of Xenon fuel (computed out of the rocket equation for the deltaV budget required) we get a total ion engine weight of 53 kg for this propulsion system.

The probe dispenser system can be a six-sided honeycomb structure supporting 6 columns of 5 probes each. Engineer Dave Pankow (SSL) has designed and built successful dispensers of up to 4 payloads in the past (Firewheel, Porcupine). Defense Systems Incorporated has designed and built a successful 7 probe dispenser and their heritage can be used in the design of the dispenser for this strawman mission. After consultation with the parties involved at the above designs we have determined that the expected weight of such a system is approximately 20 kg.

Stack attitude control can be achieved through inertial stabilization and solar pointing of the arrays with a relatively crude, 10 degree sun-pointing of the panels. This requires three momentum wheels and torque rods of a total weight of approximately 5 kg. Thus the total weight of each stack adds up to ~238 kg. The question arises how long does the ion engine take to raise the first probe from an elliptical orbit of 185X45000 km, 10 degree inclination to an elliptical orbit of 3X12 RE, 7 or 13 degree inclination. NASA/Lewis and UCB independently computed a requirement of ~68 days duration for orbit raise with continuous solar illumination and 1.25 kW of input power. The highest apogee orbit can be reached within less than twenty days after the 3x12 RE orbit has been reached. Thus the probes can be raised and released by the ion engine within three months.

The angle between the major axis of the two types of orbits can be attained from the precession of the low perigee, 185X45000 RE orbit. If the ion engine of the second stack delays its ascend to the final orbit by 1 month, it will stay longer in the low perigee orbit and precess relative to the raised perigee (3X12 RE) orbit at a rate of 15 degrees per month. A two month delay will result in the required separation. Thus total probe deployment can be achieved within 5 months of launch.

3.3 MISSION DATA HANDLING

3.3.1 Data Receipt and Validation Data are received at the ground station at a rate of 68 Mbits per probe contact. This is less than 80 Mbytes per day counting all probes. Note that this data rate is quite low: It is equivalent to the data rate of a single instrument from current ISTP missions, e.g., the EPIC instrument on board the GEOTAIL satellite. Housekeeping and instrument health data should be automatically transmitted to engineering teams over the network immediately after receipt. Data validation programs should be generated to check the validity of the data upon receipt in an automatic fashion. Most of the instrument health validation by the principal investigators can be achieved during the first 5 months of constellation deployment. The automatic data validation routines will reflect the experience gained from that initial manual instrument and data validation period. The automatic data validation programs can be run at the receiving site. If a problem exists the data could be stored temporarily for access by engineering teams for further action. After successful automatic validation the data can be stored on disks (on-line) and CDs (juke-box) for automatic data retrieval by the science community over the network while a copy will be immediately sent to NSSDC and CDHF. An entire year's data can fit on 45 CDs, which is well within the capabilities of a single juke-box at a cost of less than $50 K including programming labor.

3.3.2 Reduction and Dissemination to Scientific Community. Mirror sites supporting similar systems (i.e., a hard disk for daily updates and possibly a CD juke-box if not CDs with data) could be supported at various educational and government institutions (nodes) to allow widespread dissemination of data over the network. The common data handling (CDF) format can be used for data files, while ascii output for limited intervals also supported. The ultimate validation of the data is its use in scientific investigations by the community via the interaction with instrument teams. The dataset should be open to NASA guest investigator proposals immediately after healthy data retrieval is ascertained. All data, including distribution functions should be available to the community. Analysis ought to take place primarily in a guest investigator program fashion. This will increase the participation of a large number or scientists with diverse backgrounds (theory, simulation, modelling, phenomenology).

3.3.3 Analysis. Data visualization and analysis tools already exist at various institutions, in support of ion sensor and fields instruments on other missions (WIND, POLAR, FAST, CLUSTER). However, the proposed mission transcends traditional data analysis operations associated with single spacecraft. Visualization of multipoint datasets from many different points in space will be necessary. We can utilize existing inversion techniques such as the approach used by Tsyganenko and Usmanov [1982] to fit sparsely sampled data to a prescribed magnetosphere model. In addition the community can borrow tools already developed for similar operations in meteorology and computer simulations [Berchem 1995a, Ghil and Malanotte-Rizzoli 1991]. Such science communication tools developed between now and a future mission carry the potential of helping define better the questions asked from the mission.

Three types of data products can be distributed:

1) Synoptic, three dimensional images of extrapolated pressure, magnetic field strength and plasma flow field profiles. These can provide a first look at the dataset. A first operational magnetospheric model to be used for magnetic field inversion can be the Tsyganenko [1995] model that requires Solar Wind and Dst input and has specified magnetopause currents. The output can be a time-dependent 3-D model not parametrized by AE but fit to the probe data. The superimposed location of the probes in the images provided to the community can give instantaneous information on data availability and information on the extent to which the synoptic images are restricted by measurements. A three dimensional walk through space in a "Virtual Reality" fashion is possible today. Much like the HTML language which has recently revolutionized data exchange and communications of two-dimensional information, the VRML language is now on the verge of revolutionizing the exchange of three-dimensional images. A glimpse of that technology and its use for information exchange can be found at http://vrml.wired.com/.

2) Selection of a probe in a hypertext fashion can spawn a separate window with time series overview information on a single probe. The user can then request the data from that period, plot the data locally or click on another probe to intercompare data from two points.

3) Raw data should be sent to the user upon request in a fashion employed by the CDHF facility. Furthermore, the user should be able to select events in order to satisfy their analysis needs based on probe position and density of coverage in the synoptic images. Furthermore, event selection may take place by querying the orbit database for specific occurrences, e.g. maximum number of probe conjunctions with a ground station. In that design the community can follow the experience gained from both the satellite situation center (http://sscol1.gsfc.nasa.gov/) and from the CLUSTER-ground based campaign coordinations (http://wdcc1.bnsc.rl.ac.uk/gbdc).

3.4 PRELIMINARY COST ESTIMATE

Table 4 summarizes the cost estimates of the mission components. The probe bus estimate is based on the BITSY micro-satellite built by AeroAstro, but also on independent UCB component cost estimate. Instrument fabrication cost estimates are based on UCLA and UCB experience. Instrument testing can be minimized by the careful initial design of the units, and with the use of an automated testing facility (such facilities already exist at UCLA and UCB). Given a person month per probe for instrument calibration (UCLA/UCB experience) such costs are ~0.6 M. Another 0.6 M (one man-month per probe) will be expended on instrument and bus integration (UCB experience on FAST).

Non recurring engineering costs for the design of instruments and bus are estimated at the 5 M level. The ion engine price is based on the NSTAR engine cost estimate. The solar panel price reflects the cost of the SCARLET II array developed by AEC-ABLE Engineering, Inc. as quoted by that company. The probe dispenser cost is an estimate based on UCB and consultants' experience. This includes the integration of the probes and the ion engine on a stack that meets the fairing specifications of the launch vehicle. Ground operations and in flight calibrations require ~0.45 M for the first 5 months, i.e. simultaneous with probe deployment (1 week for each instrument and 1 week for bus health). Probe tracking from one ground station for the first 5 months requires 0.45 M. A person month cost is 10 K in all of the above estimates. The total mission cost assuming a 20% contingency is then 46.3 M, which classifies it in the MIDEX category.

Table 4. Strawman Mission Cost Estimate

ITEM	COST
Bus components (purchasing/fabrication)	100 K X 60=6.0 M
Instrument fabrication	100 K X 60=6.0 M
Non-recurring engineering cost	5.0 M
Instrument-bus calibration/integration/testing	1.2M
Ion engines	6 M X 2=12.0 M
Solar panels	1.2 M X 2=2.4 M
Probe dispenser design/fab./integration	5.0 M
Probe tracking and inflight calibration (5 mo)	0.9 M
TOTAL+20%	46.3 M

3.5 ADDITIONAL SCIENCE AT NO ADDITIONAL COST: The Magnetopause, The Magnetosheath And The Low Latitude Boundary Layer

The magnetopause is the site of mass, momentum and energy transfer from the shocked solar wind into the magnetosphere. The properties of the magnetopause have been studied observationally over the past 30 years, and are being investigated with the current missions such as Geotail, Interball, WIND, Polar and also with the upcoming Cluster mission. While these missions reveal many local features and consequences of various magnetopause processes such as magnetic reconnection [e.g., Paschmann et al., 1979; Gosling et al., 1990], they do not address the crucial relationship between the local processes and the global interaction between the solar wind and the magnetosphere. This is due to the fact that all the previous and current missions involve single spacecraft, or in the case of Cluster, four spacecraft flying in close formation. Multipoint measurements at the magnetopause are necessary and sufficient to answer the following outstanding questions in space physics today:

(1) What is the state of openness of the large-scale magnetopause for certain upstream solar wind conditions? From the estimates of the reconnection rate measured at various magnetopause sites, one can get a good understanding of the role of reconnection in the global flux transfer process.

(2) Where are the "flux transfer events" (FTE) [Russell and Elphic, 1979] generated and what is their extent? By observing FTEs at many different local times simultaneously, one can deduce the scale size of the FTEs, in the dawn-dusk dimension. Their scale size in the north-south dimension can be inferred from their flow velocity in that direction. Due to the increase of the magnetosheath flow away from the subsolar region, one would expect the FTEs to grow in size as they convect downstream from the region of generation. Thus one can use information on the evolution of FTE size to locate the source region. Such information is crucial in the determination of the role of FTEs in the global interaction between the shocked solar wind and the magnetosphere.

(3) How does the magnetosheath magnetic field drape around the magnetopause ? The amount of draping of the magnetosheath field is a strong function of the magnetopause processes: reconnection is likely to reduce the extent of draping [Zwan and Wolf, 1976; Anderson and Fuselier, 1993; Phan et al., 1994]. Thus one could deduce the state of magnetopause and estimate the rate of flux transfer across the globally by utilizing multi-point observations of the magnetic field draping.

(4) What is the evolution of the low-latitude boundary layer thickness and structure with distance from the subsolar point ? The structure of the boundary layer region immediately Earthward of the magnetopause should reveal the processes responsible for its formation. Previous observations from single spacecraft have found the boundary layer thickness to be highly variable from crossing to crossing [e.g., Berchem and Russell, 1982; Mitchell et al., 1987; Phan and Paschmann, 1996]. This is possibly due to the variability of the solar wind conditions. Only simultaneous multi-point measurements over a wide range of local time and latitude could reveal the spatial evolution of the boundary layer under a given state of upstream conditions.

4. EXTENDED GOALS

It is possible that the power and weight estimates given above can be improved due to better use of existing software (compression) schemes, hardware components, but also technological improvements in battery, solar cell efficiency, as well as the availability of lighter mission components. It is then possible to envision an extended mission for little additional cost (as launch cost, and possibly power/transmission requirements, i.e., the dominant cost drivers of the mission will remain the same). One may consider adding more scientific instruments or adding more probes. Below we explain these two different scenarios.

4.1 Electrons: Although electrons do not contribute much to the total pressure (Te~Ti/8) and their density/flow is not easy to compute from distribution functions due to the presence of photoelectrons/high electron thermal speed, electrons may contribute significantly to the magnetotail dynamics. Thus it makes sense to consider measuring and transmitting electron distribution functions during burst mode collection. How important is this adition to the overall objectives of the mission? At what cost should electrons be perceived as necessary to add to the mission?

4.2 Heavy Ions: Occationally a significant contributor to the overall ion energy density and ion total density. Although hot heavy ions are expected to follow the protons, cold ions from the ionosphere may influence the plasma behavior and their ommission may be deemed a downfall of the strawman mission. It is necessary to reach a consensus in the community on the importance of heavy ions on the structure and dynamics in Earth's magnetotail with respect to resolving the primary goals of the field.

4.3 Energetic particles: An energetic electron detector channel (between 20 and 100 keV) can give important information on whether field lines are open or closed. This gives global topological information about the flux tubes in which the probes reside.

A limited set of energetic ion energy channels can also give important information on the magnetotail and magnetopause boundary velocity at times of boundary crossings using remote sensing technique [Kettmann et al. 1990]. The anisotropy of the energetic particle distribution can provide an independent measure of the distance to the last closed field line [Sergeev et al. 1995] and of the current sheet thickness [Mitchell et al. 1990].

Five discrete energy ranges could suffice since the particle spectrum in its superthermal range as that can be fit reasonably well with a kappa distribution under diverse geomagnetic activity conditions [Christon et al. 1989, 1991]. Differential speed measurements between the energy channels can give important non-local information on the location and onset time of the energization in the magnetotail and magnetopause [Richardson et al. 1987, Williams et al. 1981].

Earthward of X=-10 RE, the ion temperature increases such that a significant part of the ion distribution is beyond the upper limit of the ion instrument. Furthermore, the probes will spend a significant amount of their time in the ring current, the largest reservoir of energy in the course of substorms and storms. If equipped with energetic particle measurements, the probes can provide the necessary correction to the ion moment computations as well as important time history profiles of the ring current particles' evolution during the course of geomagnetic storms and substorms.

4.4 Solar Wind Detector. The high altitude of the outermost probes render them perfect monitors of solar wind structures and their propagation through the bow shock, magnetosheath and magnetosphere. The inclusion of a solar wind detector in the complement of instruments can open up new possibilities for physics of interplanetary shocks, solar wind structures, upstream particles, accurate shock front determination and accurate timing of solar wind triggers of magnetospheric substorms and sudden impulses. We will assess the benefits and cost of including dual thermal/solar wind detectors based on WIND heritage [Lin et al. 1995] in some of the probes.

4.5 Towards a magnetospheric laboratory. An alternative to increasing the number of instruments on the current probes is to increase the number of probes at different orbital configurations, in order to increase the number of pixels with the ultimate goal to sense all magnetospheric regions simultaneously. The number of probes required for in-situ measurements of the system does not increase as N3, where N is the number of pixels in one dimension, because the largest volume in the magnetotail is occupied by the magnetotail lobes, which have fairly uniform magnetic field across them, and thus require only one probe per given X-distance for adequate monitoring. The emphasis and the interesting physics in the magnetosphere takes place at the magnetospheric boundaries and within the plasma sheet. The number of probes required to adequately do this in conjunction with modeling will be one of the objectives of our study.

5. STRAWMAN MISSION DEPENDENCE ON FUTURE TECHNOLOGIES

Since this mission is 3-5 years before final design we may leave room for realistic technology development in cases where significant improvements are expected. These are:

(1) Solar Arrays: AEC-ABLE Engineering has been contracted by BMDO and NASA to develop solar panels with high radiation tolerance. This mission can be the beneficiary of that development.

(2) Battery development: Companies are improving Li-metal battery capabilities, and this project can be a direct beneficiary of this consumer-driven development. We expect the battery configuration required for this application to be tested and sold commercially in the next 6-12 months. Similarly memory chips are becoming more efficient in terms of power and more lightweight.

(3) Mission design: The strawman mission views the satellite as a unified experiment, rather than a bus on which instruments will fit. A holistic approach of satellite development ought to lead to a mission that can be flexible in making acceptable science compromises against mission development cost. Advances in the area of satellite constellation design are taking place today, in particular in the field of cellular telephony. By chosing to keep abreast of the strides in technology in those areas we can be the beneficiary of a rapid technological evolution supported by commercial drivers.

6. TOWARDS THE REALIZATION OF A MULTIPROBE MISSION

The ideas expressed in this writeup cannot materialize without the help of interested scientists and engineers who envision to participate in it. It is open to comments and criticism from the community. Below, a few ideas on tasks to be accomplished in order to further such a mission are listed. The list is undoubtetly a subset of the things that are necessary to be done, but all comments on additional topics and any help is welcome.

A virtual (=on the net/phone-line) study team may be formed that will communicate via e-mail, WEB and phone and meet two times per year at AGU meetings, to evaluate the progress of the different parts of the strawman design. Topic to be discussed in the first meeting may be information from vendors and sources necessary for a full overview of the zero level mission, results on further specifications of the mission science requirements and orbit selection, first design of a dispenser with the advice/participation of experienced industrial partners/consultants, integrated design of zero level strawman mission probe. In the meanwhile such a team may:

(1) Use the Tsyganenko model for selecting orbits with highest science return. The satellite apogee will be chosen on the basis of maximizing the longitudinal and radial coverage of the plasma sheet at times when the US sector is in the midnight sector. We will identify alternative mission plans with minimal impact on science return, in order to reduce the launch cost of the zero level mission.

(2) Use a global MHD code [Raeder et al., 1995] for flying different orbit configurations through the simulation. Such a code has been previously used in conjunction with spacecraft data from the ISTP mission [Raeder, 1994, 1995; Berchem et al., 1995a]. The output of the code can be analyzed in a single probe and a multi-probe fashion.

Objectives of this essential excersice are: First to assess the ability of the selected orbital configuration to capture the critical elements of the magnetotail development during substorms. Second, to assess the ability of the probes to study most spatial scales of the quiet plasma sheet and perform inter-probe comparisons. The k-spectrum of magnetic and plasma fluctuations in the plasma sheet can be studied using simulation-provided data under a variety of solar wind input conditions. Third, to study the ability of the probes to monitor the essential regions of the magnetopause and the low-latitude boundary layer.

(3) Use the robust inversion techniques of Tsyganenko and Usmanov [1982] and techniques from meteorology and oceanography [Ghil and Malanotte-Rizzoli, 1991] in order to invert the probe measurements of the simulated magnetosphere. For magnetic field inversion one can use the Tsyganenko [1995] model. It is possible to develop a time-dependent Tsyganenko model of the magnetosphere that does not depend on the AE index (when the probes sample the magnetotail), but requires the solar wind input, the Dst index, and the probes' instantaneous measurements. It is essential to visualize the output using already developed simulated visualization techniques [Berchem et al., 1995b] and compare the results with the simulation data that was the original input, to assess the strengths and limitations of the technique.

(4) Orbit selection can be tested for orbit attainability in conjunction with NASA/Lewis and OSC. One can assess the ability of a single station to track all orbits using existing orbit integrators. Such orbit integrators/tracking packages are the Berkeley-developed routine "orbgen" and the Bester tracking systems commercial software package. The former was developped for the FAST satellite, while the latter was developed by Manfred Bester (currently also at UCB); information on it can be obtained at: http://www.primenet.com/~bester. These computations will complement the NASA/Lewis ion engine group's computations, but will also allow quick assessment of compromises that need to be made on mission components.

(5) Perform further instrument design analysis, using the heritage of existing designs but with strong emphasis on power and weight savings. Identify components that can achieve these requirements, such as improved memory and processor chips.

(7) Significant mass reduction can result from solar panel improvements. Identify companies that can provide a better power to mass ratio, even at the expense of higher cost (e.g., GaAs panels). An important cost driver in the mass of solar panels is the total radiation to which the solar panels are exposed. Radiation calculations will be performed using the code developed at NASA/GSFC by E. G. Stassinipoulos' group.

(8) Tangible products envisioned are (1) A report on: mission requirements, science maximization, feasibility study of probe construction, orbit attainability, launch vehicle and mission alternatives. This can be communicated to the wider community via EOS publications. (2) Visualization tools for data analysis the proposed mission. (3) A time-dependent Tsyganenko model of the magnetosphere on the basis of probe measurements from a simulated magnetosphere. Parametrization according to substorm phase can be attempted. (4) 3-D reconstructed images of the magnetosphere in terms of magnetic field, flow field profiles, as well as energy, and magnetic flux transport. The above products can be made available to the scientific community via technical publications and the World Wide Web.