PRODUCTION PROCESSING OF THE POSITIVE-ION SOLAR WIND DATA OBTAINED BY THE JPL 0G0 - 5 PLASMA SPECTROMETER Marcia Neugebauer and Barbara A. Weber Jet Prapulsion Laboratory Pasadena, Catifornia 91103 March, 1973 TABLE OF CONTENTS 1. Instrumentation 2. Data Reduction 3. Spectrum-by-Spectrum Data A. Magnetic Tapes B. Listing C. Plots 4. Hourly-Average Data A. Punched Cards B. Listing C. Plots 5. Bibliography A. Papers specifically referred to in text. B. Other papers whose major purpose is the presentation or use of data from this experiment. C. Other papers which make some use of data from this experiment. 1. INSTRUMENTATION The OGO-5 plasma experiment consisted of a Faraday-cup detector and a curved-plate analyzer which both pointed toward the sun at all times and an identical set of sensors which pointed radially away from the Earth. The inatrumentation is described in detail in Reference 1. The solar-wind data collected here were obtained by the solar-oriented set of instruments when the satellite was upstream of the Earth's bow shock. The Faraday cup made high-time-resolution measurements of the total charge flux of those positive ions whose velocity component parallel to the satellite- sun line corresponded to an energy per unit charge (E/Q) between 100 and 11,000 volts; this range of E/Q usually included both the protons and alpha particles in the solar wind. The average direction of positive-ion flow could be computed from the relative currents reaching each of four collectors in the Faraday-cup. At the same time that the ions' total charge flux and direction were measured by the Faraday-cup, a series of voltages was applied to the electrodes of the curved-plate analyzer to obtain the E/Q spectrum of either positive ions or electrons. There were 128 overlapping E/Q channels equally spaced on a logarithmIc scale between 2.54 volts and 16.9 kv; alpha particles should appear 10 channels above protons with the same velocity. The instrument usually cycled through a fixed sequence of taking spectra. The majority of spectra in this sequence were 32 channel sweeps about the positive-ion peak; this type of spectrum is called "proton-narrow-fine". It was also possible to sample all 128 channels consecutively ("wide-fine spectra), to sample only every fourth channel-("wide-coarse" spectra), or to reverse the polarity of the voltage and analyze electrons. See Reference 1 and Appendix 2 of Reference 2 Łor further details. 2. DATA REDUCTION The computation of plasma parameters was based on the assumption that each ion species in the plasma could be adequately described by an isotropic Boltzmann distribution (density Ni and temperature Ti) in a reference frame moving relative to the detectors with vector velocity Vi. The method of computing Vi, Ni, and Ti from the outputs of the Faraday-cup and the curved-plate analyzer is described in detail in Reference 2. Briefly, the process consisted of: a. Computing the direction of the flux vector from the Faraday-cup data, assuming a cold plasma beam, b. Using this direction to find the values of Vi, Ti, and Ni for ion species i which best fit the E/Q spectrum measured by the curved-plate analyzer, c. Correcting the flow direction for the finite temperature Ti of the ions, d. Using this corrected direction to recompute the values of Vi,Ti, and Ni, e. Iterating steps c and d until convergence was obtained, and f. Using the total flux observed by the Faraday-cup, the final direction of the flux vector, and the final velocity value to compute the total charge density Ne of the Plasma. It is known that the distribution of thermal motions in the solar wind is not isotropic. The temperature computed by the method outlined above is thus only a measure of the therma1 motions in the solar-antisolar direction; it may be greater than or less than the average temperature, depending on the instantaneous direction of the interplanetary magnetic field. The angular acceptance cone of the curved-plate analyzer was much narrower than that of the Faraday-cup (5° vs 20° HWHM); thus the curved-plate analyzer cut through only a slice of the distribution function in velocity space while the Faraday-cup accepted almost the entire particle distribution. The total charge density Ne determined in step f of the outline above is considered to be much more reliable than the ion density Ni computed in step d for several reasons: a. The value of Ni is actually the density of a fictitious isotropic ion distribution with a temperature equal to Ti. If the magnetic field were nearly parallel to or nearly perpendicular to the solar direction, the deviation of the computed Ni from the true ion density could be large for many combinations of anisotropy, temperature, and flow direction likely to be encountered. The computation of Ne is relatively insensitive to any anisotropy of the plasma; it would enter the calculations only as a correction to the small thermal correction of the flow direction. b. The value of Ne computed from the Faraday-cup data is much less sensitive to errors or uncertainties in the direction than is the computation of Ni from the curved-plate analyzer data. - c. At large angles of incidence (>~10°), the values of Ni appear to be consistently too high; this effect is probably related to reflection of glancing incidence particles and other edge effects not studied in sufficient detail during instrument calibration. The parameter Ni is not completely useless, however; it can be relied upon for the calculation of the alpha/proton density ratio Na/Np because the alpha and proton anisotropies were probably aligned with each other. The accuracy to which the plasma parameters could be determined is discussed in Reference 2. The OGO data were scanned to determine bow-shock crossing times. All fine-scan positive-ion data upstream of the bow shock were then automatically processed. It was possible, however, for the computer program to reject an ion spectrum for many reasons. The rejection rate over a period of an hour was sometimes zero and sometimes 100%, depending on the properties of the solar wind. Those spectra which survived analysis during periods when the rejection rate was high may have been atypical; an upper limit to the rejection rate over an hour is given by (1 - PCT), where PCT is a parameter given with the hourly averages. The reasons for rejection of a spectral peak were the following. a. The spectral scan (after elimination of data immediately following switching of the electrometer from a less sensitive to a more sensitive scale) did not include data on both sides of the spectral peak. b. Any current measurement in the spectral peak was a full scale reading (i.e., a 9-bit word ~ 511) which could be spurious if the electrometer was read out while in the process of changing scale c. Any spectral peak which could not be adequately corrected for the anomalous "photodip" in the electrometer zero level. This photodip problem is discussed in more detail in Appendix 2 of Reference 2. The effect severely limited the observation of the solar wind when its velocity was in the range 320 to 400 km/sec. d. No angular measurement by the Faraday-cup was available. e. The angle of incidence was greater than 10°. f. The current at the spectral peak was less than 2 x 10**- 12 amp for protons or 4 x 10**-13 amp for alphas. g. The thermal Mach number was apparently outside the range of Mach numbers in the comparison table; i.e., mv**2/2kT was either less than 16 or greater than 1000. h. The variance of the measured data from the curve corresponding to the best-fit parameters Vi, Ni, Ti was anomalously high; in the notation of Reference 2, a spectral peak was rejected if ELSQ > 0.l5. i. The best-fit curve had an unusually strange shape. In the notation of Reference 2, a spectral peak was rejected if |ESKEW| > 0.5 or if |EKURT| > 0.5. j. The ratio of the proton density Np to the total charge density Ne was outside the range 0.25 < Np/Ne < 2. k. The computer program did not try to find an alpha-particle peak if the proton peak was rejected for any reason. l. The alpha-particle peak was also not analyzed if the apparent flow direction changed appreciably (more than a 2° change in either latitude or longitude between the times at which the proton and alpha peaks were observed. Only the first reason for rejection was absolutely necessary. It is possible to give special treatment to limited amounts of data to recover some of the rejected spectra. For example, the very low-temperature, high-velocity plasma observed on Feb 2, 1969 are not included in this collection because of rejection for reason (g); these data have been reprocessed using an expanded comparison table, and can be obtained from M. Neugebauer at JPL (telephone 213-354-5182). Any potential user who is vitally interested in solar wind data for a period of time for which there are little or no data in this collection is invited to discuss the possibility of special processing with M. Neugebauer. Finally, this collection ls limited to positive-ion spectral data in the solar wind. Other types of data which have not yet undergone production processing include: a. Rapid measurements of total charge flux in the solar wind and magnetosheath. The time resolution of these data usually exceeded the time resolution of the spectral data by a factor of 16. b. Electron spectra in the magnetosheath and, occasionally, in the magnetosphere. c. Positive-ion and electron data in the plasmasphere. 3. SPECTRUM BY SPECTRUM DATA This section of original paper document was not converted to ASCII text by 4. HOURLY-AVERAGE DATA A. Punched Cards. Hourly averages of many of the plasma parameters are available on punched cards. All fine time scale spectra were given egua1 weight in forming the averages. The format of the cards is as follows: Word Format Meaning 1 I4 Year - (1968 to 1971) 2 I4 Day of year (1 to 366) 3 I3 Hour of day (0 to 23) 4 I4 Number of proton spectra during the hour 5 F6.3 PCT- ratio of number of proton spectra during hour to.maximum possible number of proton spectra at data rate and width of spectral scan being used. This ratio can be less than 1.0 for many reasons, such . as: data gaps, time used for electron spectra or proton-wide-coarse spectra, unacceptable spectra (because angle too large, or could not correct for photodip, or poor fit, etc.), and/or time spent in magnetosheath or geomagnetic field. 6 F6.0 Proton velocity in km/sec. 7 F8.0 Proton temperature in deg. K 8 F6.1 Proton density in cm-3, as determined by the curved plate analyzer. 9 F6.1 Total charge density Ne in cm-3, as determined by combining Faraday-cup and curved-plate analyzer data. This is more reliable than the proton density given in word 8. See Section 2 for an explanation of the differences. 10 F6.1 Ecliptic north-south angle in degrees Positive for flow from south to north of ecliptic. 11 F6.1 Ecliptic east-west angle in degrees.Positive for f1ow from west to east of the sun, which means the flow has some amount of corotation with the sun. This angle has been corrected for the effect of the satellite velocity. 12 F6.0 CZ in degrees ~ angle between the solar-ecliptic y axis and the center of collector 1. 13 F7.3 (Va - Vp )/Vp =(alpha velocity - proton velocity)/ (proton velocity) 14 F6.1 Ta/Tp = ratio of alpha temperature to proton temperature 15 F6.3 Na/Np = ratio of number density of alphas to number density of protons 16 I4 Number of alpha-particle spectra during hour. 5a. BIBLIOGRAPHY 1) GRAHAM, R. A., AND F. E. VESCELUS OGO E PLASMA SPECJROMEJER PROC. THIRICENTH NAT. INSTRUM. SOC. AMER. AEROSP. INSTRUM. SYMP. 111-153, 1967 2)NEUGEBAUER, M. COMPUTATION OF SOLAR WIND PARAMETERS FROM THE OGO- 5 PLASMA SPECTROMETER DATA USING HERMITE POLYNOMIALS JPL TECHNICAL MEMORANDUM ANOUH 33-519 (DEC. 15, 1971). Other sections of original paper documents were not converted to ASCII text by NSSDC