GEOPHYSICAL RESEARCH LETTERS, VOL. 18, NO. 8, PAGES 1493-1496, AUGUST 1991 CHARACTERISTICS OF AKR SOURCES; A STATISTICAL DESCRIPTION A. Hilgers 1 Swedish Institute of Space physics, Uppsala division, Sweden A. Roux CNET-CNRS/ CRPE, Issy_les_Moulineaux, France R. Lundin Swedish Institute of Space physics, Kiruna division,Sweden Abstract. A description of plasma properties within the sources of the Auroral KilometricRadiation (AKR) is given. It is based on data.collected during ,-, 50 AKR source crossings in the altitude range between 4000 and 9000 kin by the Swedish spacecraft Viking. The following results are obtained. (i) The frequency of the lowest fi'e- quency peak of the AKR .œpe•k is found to be very close to f•e, the electron gyroh'equency ((fp•,,k - f•)/f•e <- 0.08), on the average, (ii) The lower cutott frequency f•,c is on the average at f• ((f•:c -./;•)/œ• •- 0), (iii)in the sources the densityis typically lessthan 1..5 cm -a, which is of the order of the density of hot electrons and (iv) the source is located within an acceleration region, as evidenced by electrons accelerated above and ions accelerated below. 1. Introduction Benediktov et al. [1965], Dunckel et al. [1970], and Gur- nett[1974] have discovered a new non thermal radio source: the Ea.rth. This radio emission, the Auroral Kilometric Ra- diation(AKR hereafter) cannot propagate across the dense layersof the ionosphere and cannot therefore be observed on the ground; hence the need for space borneinstruments to observe it. Sincethe discovery of the AKR, the mech- anism that relates it to the electrons that are accelerated over the auroral regionhas remaineda challenging prob- lem for the space physicist. Until recently most of the information we had about the AKR came from observa- tions conductedfar away from the AKR source. These remote observations led to generalimportant conclusions regarding (i) the location of the AKR source [Kurth et al., 1975; Kaiser and Stone, 1975; Kaiser and Alexander,1976; Alexander and Kaiser, 1976], the polarization of the AKR [Green et al., 1977; Guvnett and Green, 1978; Kaiser et al., 1978; Benson and Calvert, 1979; Shawhah and Gur- nett, 1982], its relation to the auroral activity[Benediktov et al., 1968; Dun&el et al., 1970; Gurnett, 1974; Voots et al., 1977; Movioka et al., 1981; Benson and Akasofu, 1984] and (iv) the efficiency of the energy conversion from auroral electrons to the radiated waves[Gurnett, 1974]. Severaltheories have been proposed for the interpreta- tion of the AKR. Some of them were ruled out when the sense of polarization, predominantly RH, wasdeduced from the cutoff frequency measured by Hawkeye [Gurnettand Green, 1978] or from directpolarization measurement by Voyager [Kaiser et al., 1978] andDE 1 [Shawhah and Gut- neff, 1982]. ion leave from CNET-CNRS/CRPE, Issyles_Mx, France Copyright 1991 by the American Geophysical Union. Paper number 91GL01332 0094-8534 / 91 / 91GL-01332 $ 3 . 00 The first in situ measurements published by Bensonand Calvert [1979], have played a decisive role in demonstrating that AKR is generated in regions where the ratio fp/f•e of the plasmafrequency to the electron gyrofi'equency is much lessthan unity. This discovery has focussed the in- terest on the Cyclotron Maser Instability (CMI thereafter) proposed by Wu andLee[1979] and developed by several authors. Given the low energyof auroral electrons, the applicability of the CMI is indeed restricted to plasma de- pleted regions (where fp/f• isless than about theratio of the root-mean-square of the velocity of the energetic elec- tronsto the velocity of light). The trajectory of the Swedish spacecraft Viking was de- signed to cross the centralregions where AKR is expected to be generated. IndeedViking, in spite of its short life- time, did cross about 50 times AKR sources at altitudes between 4000 and 9000 km that corresponds to the peak in the AKR spectrum. Viking carried a complementary set of wave and particle measurements that can be usedto characterize the signatures of AKR source crossings. Bahnsen et al. [1989] haveshown that an AKR source crossing is characterized by (i) an enhanced AKR ampli- tude (ii) a minimum in the lower frequency envelop of the AKR, and (iii) this envelop becomes close to f•,. These characteristics wereconfirmed by U. ngstvup et al. [1990] who stressed that at AKR source crossings (i) low energy electrons (< I keV) are depleted and (ii) up flowing field aligned ionsare regularly observed. From the latter obser- vation, Ungstrup et al. concluded that an AKR sources are located above an acceleration region. Louavn et al. [1990] have extendedthis conclusion by interpretingthe signa- tures of AKR source crossings in the electron distribution function. They concluded that the observed trapped distri- bution function of electrons andthe lack of lowenergy elec- tronsare consistent with electrons beingaccelerated above the spacecraft and trappedbetween an electric and a mag- netic mirror underthe effect of an upward directed parallel electric field that varies in the guiding center frame of the electrons. Hence Louarn et al. concluded fi'om the ob- served electron and ion signatures, that AKR is generated within an acceleration region with ,-• 3 kV both belowand abovethe spacecraft. A detailed analysis of the densityof the various components of the plasma (cold-cool-energetic) hasled Louarnet al. [1990] and Pertaut et al. [1990] to conclude that AKR source coincides with a densitydeple- tion; in the source there are very few, if any coldelectrons. The present study aims at a.ssessinlg the generality of the aboveresultswhich are based on a very few case studies. 2. Wave characteristics of AKR source crossings Figure i shows the dynamic spectrum of oneelectric com- ponent measured bythe V4 H wave experiment IH stands for high frequency) on Viking. An AKR source is crossed at ,,• 16:44UT, at an invariantlatitude A _• 67degrees and 1493