UNIVERSITY OF OSLO
OwC— 86-tB P R 0 C E i DISCS of i hu 13TII ANNUAL MEETING ON UPPER ATMOSPHERE STUDIES BY OPTICAL METHODS Lysebu, Oslo, Norway 19.-23. August, 1985
Edited by Karl M .1 s e i d e Report 86-28 ISSS-0J32-5571 December 1986
DEPARTMENT OF PHYSICS REPORT SERIES
I! 1LJI Jil
U M 1UJ^* *
JJLIJ .ili:mljj(:.iU U PROCEI. D'NCS of the I3TH ANNUAL MEETING ON UPPER ATMOSPHERE STUDIES BY OPTICAL METHOD Lysebu, Oslo, Norway 19.-23. August, 1983
Edited by Karl Må seide Report 86-28 ISSN-0332-5571 December 1986
Department of Physics, University of Oslo, P.O.Box 1038 Blindern, K-03 15 Oslo j, Norway 13TH ANNUAL MEETING ON UPPER ATMOSPHERE STUDIES BY OPTICAL METHODS Lysabu, Oslo, Norway, 19.-23. August, 1985 Participants in front of the Lysebu Conference Center. Ul
TABLE OF CONTENTS
SESSION I: OPENING AND INTRODUCTORY TALKS Chairman: K.Måseide WORDS OF WELCOME 1 T. Amundsen AURORAL RESEARCH IN NORWAY UP TO THE SPACE AGE 3 A.Egeland and A.Brekke UPPER ATMOSPHERE RESEARCH IN NORWAY AT THE PRESENT AND FUTURE PLANS' E.V.Thrane
SESSION II: ATMOSPHERIC EMISSIONS Chairman: A.Vallance Jones
EXCITATION MECHANISMS AND SPECTRAL EMISSIONS FROM THE UPPER ATMOSPHERE 24 A.Vallance Jones
ANALYSIS OF AURORAL Oa ' FIRST NEGATIVE BANDS 43 K.Henriksen and L.Veseth OPTICAL STUDIES FROM SPACE SHUTTLE - A PRELIMINARY REPORT OF MEASUREMENTS ON THE STS 41-G MISSION 66 E.J.Llewellyn, I.C.McDade, D.J.W.Kendall, R.L.Gattinger, S.B.Mende, G.R.Swenson, W.A.Gault, G.G.Shepherd, B.H.Sclheim, L.L.Cogger, and M.Garneau
SESSION III: AURORAL FEATURES AND DYNAMICS Chairman: M.H.Rees THERMOSPHERIC STRUCTURE AND AURORAL MODELING 76 M.H.Rees (Abstract only) THE EFFECT OF LIGHT SCATTERING AND DIFFUSE REFLEC TION ON ATMOSPHERIC SPECTRAL MEASUREMENTS 78 K.Stamnes and G.Witt COORDINATED ROCKET MEASUREMENTS OF AURORAL X-RAYS AND OPTICAL EMISSIONS 91 J.Stadsnes, K.R.Topphol, and K.Måseide
'Papers presented, but not submitted for repro duction here. IV
DAYSIDE AURORAS IN RELATION TO THE INTERPLANETARY MAGNETIC FIELD: A CASE STUDY 105 P.E.Sandholt, A.Egeland, B.Lybekk, and C.S.Deehr
MIDDAY ELECTROJET AND RELATED EXPENSION OF DAYSIDE AURORA1 T.Oguti NEW IDEAS OF SUBSTORM DEVELOPMENT AS OBSERVED BY GROUND-BASED OPTICAL INSTRUMENTS 129 R.J.Pellinen
SESSION IV: AURORAL PULSATIONS Chairman: T.Oguti
RECENT DEVELOPMENTS IN PULSASTING AURORA STUDIES 141 I.Sandahl AURORAL AND CONCURRENT GEOMAGNETIC PULSATIONS 160 T.Oguti PULSATING AURORA AND HYDROGEN EMISSIONS 185 R.A.Viereck and H.C.Stenbaek-Nielsen
ROCKET EXPERIMENTS TO STUDY PULSATING AURORA1 F.Seraas
SSSSION V: AIRGLOW AND ATMOSPHERIC PARAMETERS Chairman: G.Witt
SOME CURRENT PROBLEMS IN AIRGLOW RESEARCH1 G.Witt
ALTITUDE DISTRIBUTION OF SOME AIRGLOW FEATURES 199 J.J.Lopez-Moreno, A.Molina, M.L6pez-Puertas, F.Moreno, and R.Rodrigo
CONTAMINATION IN ROCKET-BORNE IR MEASUREMENTS 210 D.Smith and A.Ratkowski GROUND-BASED OPTICAL OBSERVATIONS OF THERMOSPHERIC DYNAMICS AT THE GEOMAGNETIC POLE, AT THE GEOMAGNETIC EQUATOR, AND AT STATIONS IN BETWEEN1 J.W.Meriwether, Jr.
NEUTRAL WIND MEASUREMENTS BY FABRY-PEROT. INTERFERO- METRY IN ANTARCTICA 229 R.D.Stewart, J.R.Dudeney, A.S.Rodger, R.W.Smith, and D.Rees v
UPPER ATMOSPHERE TEMPERATURES OBTAINED BY INFRA- REO SPECTROSCOPY 240 F.Gerndt, D.Offer-mann, and T.Blix THE UNIVERSITY OF BONN LIDAR EXPERIMENT AT THE ANDØYA ROCKET RANGE 252 U.von Zahn
S E SSI ON VI: ATMOSPHERIC CONSTITUENTS Chairman: G.Gustafsson SOLAR MESOSPHERE EXPLORER SATELLITE OBSERVATIONS OF POLAR MESOSPHERIC CLOUDS AND OF OZONE AT HIGH ALTITUDES' J.J.Olivero
OPTICAL MEASUREMENTS OF ATMOSPHERIC OZONE AT HIGH LATITUDES 260 S.H.H.Larsen
OBSERVATIONS OF THE GEOCORONAL HYDROGEN1 J.W.Meriwether, Jr.
LYMAN-a OBSERVATIONS FROM A HIGH ALTITUDE ROCKET 266 H.Lauche CHAIRMAN'S SUMMARY OF SESSION VI 276 G.Gustafsson
SESSION VII: INSTRUMENTATION AND DATA HANDLING Chairman: H.Lauche LOW LIGHT LEVEL MEASUREMENTS WITH PHOTOMULTIPLIER TUBES - PHOTON COUNTING VERSUS ELECTROMETER OPERA TION 278 A.G.Wright
A PULSE AMPLIFIER/DISCRIMINATOR (PAD) FOR SINGLE PHOTON COUNTING 300 L.Alexander
GROUND-BASED SPECTROSCOPIC MEASUREMENTS WITH IMAG ING PHOTON DETECTORS' J.Stegman A ROCKET PHOTOMETER EXPERIMENT TO STUDY OPTICAL EMISSIONS FROM ARTIFICIAL ELECTRON BEAMS IN THE IONOSPHERE 311 K.Måseide VI
,\ BRIEF DESCRIPTION OF THE ELECTRONICS OF A FOUR CHANNEL ROCKET PHOTOMETER AND THE CHECK-OUT EQUIP MENT 328 T.A.Sten A REAL-TIME SYSTEM FOR MEASURING THE ALTITUDE OF AURORAL STRUCTURES 335 A.Steen
ALTITUDE OF THE RED LOWER BORDER IN AURORAL FORMS 347 K.Kaila
IMAGE PROCESSING OF ALL SKY AURORAL PHOTOGRAPHS 353 B.Lybekk (Abstract only) CALIBRATION OF OPTICAL INSTRUMENTS FOR ABSOLUTE INTENSITY MEASUREMENTS OF UPPER ATMOSPHERIC EMIS SIONS 354 J.Stegman and D.P.Murtagh
A CALIBRATION PHOTOMETER FOR LOW BRIGHTNESS SOURCES 364 H. Lauene and W.Barke
S ESS I ON VII_I :_0,B SER VATION PROGRAMS Chairman: R.Pellinen A SHORT HISTORY AND SUMMARY OF OBJECTIVES FOR THE GROUND-BASED OPTICAL AERONOMY (GBOA) PROGRAM IN THE UNITED STATES 371 C.S.Deehr, G.J.Romick, J.Meriwether, and M.H.Rees A CANADIAN MERIDIAN PHOTOMETER ARRAY (CONOPUS MPS) 384 D.J.McEwen, L.Cogger, F.Creutzberg, R.Gattinger, F.Harris, A.Vallance Jones, R.A.Koehler, and J.Wolfe (Abstract only) GROUND-BASED PROGRAMS RELATED WITH VIKING 385 G.Gustafsson
THE MAPSTAR PROGRAM - GENESIS AND PHILOSOPHY1 J.W.Meriwether, Jr. UPPER ATMOSPHERE OPTICAL STUDIES FROM SVALBARD AT THE PRESENT AND FUTURE PLANS' A.Egeland.
CHAIRMAN'S SUMMARY OF SESSION VIII 391 R.Pellinen 1
WORDS OF WELCOME Tore Amundsen, Dean of the Faculty of Mathematics and Natural Sciences. University of Oslo.
It is a pleasure to take part in the opening of this conference and welcome you all to Oslo and to this 13th Annual Meeting on Upper Atmosphere Studies by Dptical Me thods. Me are nappy to see you all here, and it is encourag ing for the cosmic physics group that so many have found an opportunity to attend this meeting. It is now 10 years since a similar meeting was held in Norway, and at that time the meeting was arranged by The Norwegian Institute of Cosmic Physics. Since then, this institute has been reorganized, it has disappeared as it were, into the Universities of Oslo and Tromsai. So you might say that the former Oslo division of The Norwegian Institute of Cosmic Physics now appears ir. a new disguise; as the group of cosmic physics in our Institute of physics. It is this group which is now hosting this conference. I don't know whether the group was happier as a small and fairly independent "division" rather than just one group among many others in a fairly large institute - I suspect they have had some nostalgic feelings about the good old days from time to time - but on behalf of the faculty I would like to say that we are very happy to have this active groi_p with us at the Uni verslty.
Norway and the University of Oslo have a long and rich tradition in cosmic physics with outstanding scientists like Kristian Birkeland, Carl Stelrmer, Lars Vegard, and many others, and we Are pleased to have a group which continue to do high quality research in this traditional field. I suppose it must be fair to say that wc live in a very exciting time with a great deal happening in science, not least in astrophysics and cosmic physics, and I am sure you have a lot to talk about and are eager to get on with the program. Let me nevertheless very briefly slip in a few facts about the Faculty of Science here in Oslo before I 2
again leave you in the hands of the chairman. Official statistics tell that around 1/3 of all basic research in the mathematical and natural sciences in Norway is being done here at our faculty. So we Are rather large by Norwegian standards, and I think also by European standards. Close to 1000 people are in some way connected with our institutes, and at the moment we have a little more than 3B00 students altogether. The number of students has increa sed through the last 10 years, and on Wednesday, the 1st of September - when the term is officially opened - we will be welcoming more than 900 new students. One of the reasons for this increase is probably that we now have a choice of a more technologically oriented education. We can even offer the students a degree called "ei vi I-eng*neer" for the first time this year, which in itself is a remarkable event in the history of our universi ty. A student working on measuring techniques and electro nics - for example applied to problems in the upper atmo sphere - can certainly qualify for this degree if he wants to. I think the University has a lot to offer also in the direction of a more applied education, and it is encouraging that the response from the students seems to be positive. The preparation of these new alternatives in our curricula has been one of our main efforts - and also our main subject for discussions - for quite some time. Well, these things are on my mind. I am sure you have different problems on yours. With the prospect of Norway entering the European Space Agency (ESA) and our government hopefully allocating funds to ESA-related research, the outlook for the Norwegian activities in cosmic physics should be good. 1 will end on this optimistic note by wish ing you a fruitful conference and a pleasant stay here at Lysebu in Oslo. 3
AURORAL RESEARCH IN NORWAY UP TO THE SPACE A5E
A. Eye!and, Inst i lut r» a* Physics, University of Oslo.
A. Breike, Auroral Observatory, University of Tromsci.
INTPOCUCTIQW Be-fore the? 19th century very little was f nown about tht* -har act en sties of aurora , except that it occur -_-_ at high 1 at i tudes l-c* . e.g. BreUe and Egel and , 1967:) . F r cm the middle of the last century and up to the ~p-?ce age Norwegian scientists placed a dominating role in ^uror^l research. In this Dacer ^he t-j.npnas i s will be on persons who significantly "ontribLited bet h to the interest and I- noi«l edge- of t h'3 nor th ~ E-r n 1 ; ;u i: '. ->or mal 1 y cal 1 ed " aur or a " out si de Sc and : navi a) . Christopher Hansteen < 1784-187-3) was the pioneer in Norwegian natural sc i ence , e = pet. i ally in geomagnetic studies •'Fig. 1 ) . His most i mpor t an t con tnbutiun wa= hi 5 hypothesis that auroral arcs form a ring abound the magnetic pole. o3.r-steen sti-ongl* believed that the- northern light had .-* :3"1 ing i= *' f "• z t on the i a^er atmosphtjr e. \ le al so cl ,ti med ti IJ t he h 3d heard t H? aurora. (Fur f^rUier detaiis aboui_ Hansteen, sr-- ?rel I ^- ^-d Ei---! Ann\ 19^,,) 'h-v fJr,' /^g;3i -Htrunuir-t-r Carl Frederik Fearnley (1B13 *>0'> per f ori'isd ^u.--if aur oral height aieasur «ment s at d arrived a t r n.e^n *. ~*1 _i-~ of 2 •'.•''• I- m . An r .;r .-• 1 • it-i g'i t wa-, * .nui. }. G: =>put^cl qi_ .ttion in t1,... ; jji. t.emi.iry, -*nd Pear-nie/ s valuer w^re 3;'-=: of th-- f . - ? r >-) :ar-;. • f-ti^l-m. Sophus Tromhol t (1ST." 1 r?,L; hor n i r; De?.-imar i - was t he f: ". *• • o . e:- r : i *. : n Scandina-'iJ who - fnr several year s- wc I ed •«.ill 11 ri- -.r, the auroral problem (Pig. 21. He organi
c zed v t = • • -'. ^'..-..': ril ob-e? va* itjiib ffuiii ae\r-ral stations in Northern Po- ope . Pir.rj The r i r st I nt *-»r nat i onal Polar Year ' ! BB2-PT - \ -.-. . ' •-1 =• 1 1 pel an j operated an aur nr a I stat i on i n F i nnnar 1 , No-" the. n Nr.-' H«a / . 7 r omn ol t was particularly iiterps-
+ for-i in an :icc ;r- e correlation jetwe-?.i auroral ccurrenre, r, ->onagnet i r set: • i ty, and t! -J : ! • ,ear suf isp.-- c y^ l^. He ais1: system at i cal 1 y ir-.-eftig^tfd mt^rrnpti rjns on telegraph l i nes A
- particular ly in the auroral zone - due to auroral activity. Tromholt's auroral data catalog - covering almost the same period as the data collected by Rubenson - has, unfortunate ly, not been extensively used. Today TromhDlt is probably best remembered tor his auroral descriptions. His main work "Under the Rays of Aurora Borealis" -from 1885 is both sober and entertaining. He was only 45 years of age when he died. Tromhalt's imposing con tributions have unfortunately been overlooked in the history of aurora. (Far mare details, cf. e.g. Brekke and Egeland, 19B3; Eathsr, l^BO).
KRISTIAN OLAV BIRKELAND <1867-1917) Birkeland's (Fig. 3) outstanding contributions to au roral theory and to the foundation of modern magnetaspheric physics have recently been discussed (Egeland, 1984). In this review therefore only Birkeland's key discoveries and personal characteristics will be briefly discussed. In 1895, while experimenting with cathode rays near a magnet, Birkeland noticed that electrons were guided toward thrf magnetic pole. "Maybe the aurora is produced in a si mi lar way", he reasoned- In 1896 he extended this to an auroral hypothesis, in a psper publi shed by the French Academy. The main point of this theory was that energetic electrons are ejected from sunspots on the solar surface. These particles were later captured by the Earth's magnetic field and were guided to the nightside polar regions, where they produced the vi si ble aurora. BirPeland's theory is the first realistic explanation of haw an aurora was created. In order to veri fy this hypothesis, Birkeland performed extensive laboratory experiments in which the Sun—Earth systsem was - for the first time - simulated in a vacuum chamber. His model experiments were simple, but at the same time ingenious. In a large (up to 1.5 m3) evacuated tank he attached a little sphere, a model of the Earth's globe whir-h he called a terrel la. Inside the terrel la he mounted an electric coil which simulated the Earth's magnetic field, suitably scaled to produce similar particle orbits. He dir ected beams of electrons toward this simulated earth, and by coating the terrella with fluorescent paint, he was able to 5
see where they impinged nn it. Artificial "aurora" was pro duced by this terrella experiment for the first time in 1896 auroral and geomagnetic research of that time. Because of Bi r Lei and'5 unconventional personal i tyf a great part of the scientific cammunity ignored his pioneer ing discover les. Only now, in the age of space exploration, can we proper!y appreciate the prophetic nature 0+ hose discoveries. For a detailed review of Birkeland's Polar Elementary Storms and how he deduced the "Birkeland currents", the reader 13 referred to his main books (Birkeland, 190B and 1913). GLIMPSES OF BIRKELAND AS A HUMAN BEIN6 In 1B93 Birkeland became a ful 1-time research assistant at the University of Kristiania (now Oslo) and was appointed professor in 1B98. In a short, b^u very hectic period, Birke land built an active research environment- He was tireless, extremely energetic, enthusiastic, and constant1y involved in several major projects simultaneously- When he could not receive the financial support he requested from the Govern ment f he used his own money both for salaries and to organize expeditions. He never spared himself. Birkeland's work was extensive and of high caliber- He introduced innumerable ideas and theories - some fundamental and revolutionary, others utterly worthless. Birkeland followed several such ideas and tested them out in detail, and a surprising number o-f them turned out to be correct. In a.1 1 his wor fc , he had a creative and lively 1 magi nat 1 on and an enormous faculty f or intelligently putting pieces of information together to obtain correct solutions to problems. Eirleland was extremely enthusiastic, and he worked hard both days and nights. To all his work, he brought a creative and lively imagination along with a good sense of humor and self-irony. His direct and indirect influence was g. -at. Birkeland was intensely interested in auroral and geo magnetic studies, and he devoted great efforts to increase aur knowledge about these phenomena. He employed several young scientists to work for him, among whom were Vegard, '•'rogness, Sæland, Skol em, and Devik. Through his many pro jects, Birkeland awakened an interest in auroral research in these young cc-leagues, and he inspired and stimulated them 7 to do their wry best. The idea of exercising as a means of staying in good physical shape was strange to Birkeland. He worked far too hard and was careless concerning regular meals. Birkeland was in his 30' s when he married, but he lived with hi-, wife -Tor only one year, and they had no children. His scientific experiments in the 1890s caused him serious hearing defects, and his general health deteriorated rapidly in his 40's. Birkeland died June 15, 1917, at a time when a working com mittee was in the process of nominating him for the Nobel Prize in Physics. At the University commemoration of Birkeland's death, Vice-Chancellor Sæland characterized Birkeland as. a "scien tific explorer by the grace of God". In the eyes of many Norwegians he had an ideal life: he had become both rich and famous in his lifetime. In fact, Birkeland had many problems both in his scientific and private life. He had few close friends, though those he had really loved him. He had many antagonists and was often attacked in papers and by collea gues . CARL FREDRIK MiiLERZ STØRMER (1874-1957) Størmer Inf. Fig. 5) was born on September 3rd, 1874, and died August 13th, 1957, nearly 83 years old. Stdrmer graduated in mathematics from the University of Kristiania (now Oslo) in 1897 as candidatus realium, and the year after he gained a f i* *? year research fellowship. He continued his mathematical studies at Sorbonne (Paris) from 1898 to 1900, and in Gdttingen in 1902. He was married in 1900, and they had five children. In 1902, at the age of 28, he was appointed professor of pure mathematics in Oslo - a position he occupied for 43 years, retiring in 1946 at the age of 72. He had a good economic foundation and brought fame to himself and Norway for his work. BIRKELAND'S EXPERIMENTS INSPIRED STØRMER TO CALCULATE THE PATHS OF AURORAL PARTICLES After his introduction to Birkeland's terrella experi ment in 1903, Stormer began a comprehensive mathematical calculation to determine how auroral particles can propagate 8 from the Sun, through spa», and into the Earth's atmosphere. In a short passage from Stermer s -first paper in 1904 on the particle trajectories, he outlined this work in the following way: "firkeland has described an artificial aurora produced in the laboratory. From a theoretical viewpoint there e::ists an especially interesting problem - namely solv ing the equation of motion of "an electron" in a magne tic field. It is very clear what importance the solution to this problem will have to Birkelands theory." The main problem was that these equations describing the paths of auroral particles did not have a general analytic solution. Therefore, Stormer was forced to use a numerical method and to follow each electron path step by step with tedious calculations (tz-f. Fig. 6). Stormer's mathematical treatment of thr* auroral problem was a natural results of his education and talent as a mathe matician. Altogether, Stormer and his assistants spent 18.000 hours on these calculations. Through a series of treatises (at least 4B) he gave a survey of the possible solutions to this problem. These papers - which were pioneer works - Are now classics. It was shown how auroral particles from the Sun are sucked towards the Earth's magnetic poles ii> two circular zones - the auroral zones - one in each hemisphere. STØRMER'S CALCULATIONS AND INTRODUCTION OF THE RING CURRENT Strirmer believed that his work on particle trajectories explained many features of auroras. Still it was impassible to obtain a good fit with the actual location of the auroral zone, in particular, during geomagnetic storms. In order to remove this difficulty he introduced "a stream of moving corpuscles around the earth" as early as 1911/12. He stated that solar electrons in the Earth's dipole equator tend "to bend around the magnetic globe an the evening and night side - and even to encircle it like a ring". This ring was located some Earth radii above the Earth's surface. Thus, Størmer estimated the north-south movements of the northern light during increased solar activity. This was an important and correct step toward a solution of the auroral problem, which has been completely neglected by later workers in this field. 9 STØRMER*S PHOTOGRAPHIC AURORAL STUDIES Of all the characteristic properties of the northern 1lght, none had been studied and discussed more than its height. Knowledge about the height and geographic location o-f the northern 1 ight was necessary in order to understand and formulate a realistic auroral theory. At the same time a wel 1-determined height of the northern 1ight would yield vital information about properties of the atmosphere at much grafter heights than ground measurements previously, direct or indirect, had been able to accomplish (Fig. 7). \ All reported auroral heights before 1910 were based exclusively an visual observations and consequently there were large i ^accuracies in the measurements. The precise, quantitative auroral studies made by Stbrmer exempli fy the method which he used in his auroral investigations. The manner in which he calculated auroral heights and positions in the atmosphere is described by him as follows: "In 1909 I began systematic experiments to photograph the aurora. By testing and comparing I was able to find a smal 1 , extremely "fast" camera lens which gave excel lent results. Using this lens and the most vi olet—sensi — tive photographic plate available, I succeeded in taking pictures of a bright northern light with an exposure ti me o-f onl y 1 s or less. When the problem of auroral photography was satisfactorily solved, I went to photo graph the northern light and determine its height simul taneously from two stations separated by an appropriate distance - the two stations also being in contact with each other by telephone. On the photographic plates, star patterns were also recorded and from the different positions among the stars as seen from the two stati ons, the hight could be calculated from measurements on the photographic plates." Størmer was the pioneer who introduced a methodical and logical scientific method into auroral height measurements. He began this task in his customary energetic manner and drew up his plan of attack i n record t i me. Strirmer establi — shed one station in his home at Bygdriy (near the waterfront in Oslo) and was allowed to use the old observatory in Oslo as a second station. In addi ti an to these two stations, Steirmer also established others at suitable distances -from Oslo: Kongsberg, Lillehammer, TiSmte in Hurdal, Dombås, Askim, etc. These stations were kept in operation by di-fferent people until the 50'ies. Tt was Strirmer's habit to carry with him a packet spec— 10 troseope with which he examined the night sky, to seek for spectral evidence of the presence of aurora (even when the sly was cloudy), and to observe the main spectral features of coloured auroras. He also kept by him the diagrams (prepared by J. Bart- els) of the 3-hour K-indices of geomagnetic disturbance, as a guide (based on the 27-day recurrence tendency) to the likelihood of auroral appearance on each evening. Chapman (1960) wrote: "This part of his research has earned for him an undying fame in the history of auroral science". His last paper on this subject appeared in the year of his death, summarizing 1956 auroral observations. Stdrmer and his colleagues took more than 40.000 auroral photographs and derived heights and location of more than 18.000 auroral points. Both obtaining all these photographs, not to mention their reduction, was an incredible labour. This work illu strates Stdrmer's perseverance and spirit. Obviously, Stairmer received many good ideas - particularly in reference to data—reduction from his assistants Vegard and Krogness, Harang and Tonsberg, Herlofsen and Egeberg. SUNLIT AURORAS AND SPECIAL AURORAL FORMS Stdrmer's height-range distributions of auroras still form the basis of our knowledge in this field. Already in the 1920'ies he discovered the remarkable properties of sunlit auroras. The solar illuminated auroral rays generally reach greater heights than those which are sit.iated in the darkness. The realization that these northern lights could occur as high as 1.000 km (Fig. 7) stimulated exploration of the outermost parts of the atmosphere and ionosphere to a considerable degree. All the atmospheric models dealing with composition and density at an altitude around 1.000 km had to be completely revised. Observations of sunlit auroral rays became i basic part of StiSrmer's research program. Stormer's auroral work is described in great details in his magnificent book "The Polar Aurora" (1955). 11 PERSONAL CHARACTERISTICS I the proof of all his manuscripts; everything neatly put together. Unfortunately, the main part of this unique collec tion was di strayed a few years after his death. However , we still have a large part of his photographic col lection. He was always wel1 prepared, both for his regular uni vers i ty lectures, as wel1 as for more official duties. He was no doubt a clear and good lecturer, but probably not very enthusiastic and inspiring- He was chairman and presi dent of several international unions, both in mathematics and geophysics. Thus, he was very well respected by his col leagues. As the opening address for a big conference in Oslo in 1948, where he was president, I have found 3 different ver sions; one for nice weather, one for rain, one for an outdoor ceremony with the King present, and al1 of them typed both in French and English. He used both these languages regular ly. In all his work he was very accurate, and he hated care less work and mi stakes. It should be exactly the right value, •far.v or shade. Steirmer became a well known scientist, and was much in demand as a speci al guest 1ectur3r at big events. He recei ved several honourary doctorate degrees from well known univer sities, and was a member of several academies of science alI over Europe. He never fini shed his own Ph.d. In Norway he was an active newspaper-writer and also gave several talks on the radio, almost always about his auroral work, but he also publi shed articles and books in related fields. Dr. Lei v Harang wrote, in a letter to Professor Chapman: "Stormer always reminded me of the most loveable char acter in English literature: Mr. Pickwick - and when he lectured on his researches in auroras I always had to think of the scene in the Pickwick Papers where Mr Pickwick is presenting his report on his expedition to the ponds of London. I am thankful for havi ng 1 earned to know a man like Strirmer; in our streamlined century a personality like his with all his peculiarities cer tainly never will develop". And Professor Chapman wrote about SteSrmer: "In the rooms of the Royal Astronomi cal Soci ety in London there is a portrait of Gauss. Under the portrait there Are the foilowinq words, put by Shakespeare in the mouth of the bastard Edmond in King Lear: " Thou Nature art my goddess, To thy service I am bound". 13 These words are as applicable to Strirmer as to Gauss. Throughout his 1 ong life Størmer pursued his chaser» course of observing nature and applying his fine intellect to the unravel 1 ing of some of its mysteries. In the concluding paragraph of the preface to The polar 3Kircr3i he says: "M/ work has given me infinite pleasure and satisfaction, but I regard it as no more than a pioneer effort." LARS VEGARD <19B0-1(?63) Vegard (Fig. 8), Birkeland's assistant, was appointed professor in physics at the University of Oslo in 191B. He was the first person who fully concentrated on auroral spec troscopy. Prior to 1910t the only emission in the aurora known to a fairly accurate wavelength was the auroral green line. This was the situation when Vegard began his pioneering work on the auroral spectrum. In 1912-13 Vegard measured the wavelength of the green line much more accurately than Ang- st^25rn and obtained a value of 557.69 nm. The atmospheric gas which produces the auroral green line, however, remained a mystery up to 1924. Vegard "got his feet wet" in this discussion. His hypothec is was that at high altitude the:- green line was emitted from nitrogen in ice crystal form. Unfortunately, the laboratory experiments he carried out in Leiden to demonstrate? his hypothesis did not have ^pec tr otr.eter s of high enough quality. Thus, Vegard did not notice that he was off by more than 0,1 nm, *nd his h'.'^-thE=is caused a lot of dispute. It took a long time bn^x? he admitted that he had made a big mistake. Vegard was an active observer. After less than ten years he had accurately camp i led about 40 di fferent auroral em i s*>i ons. He made new di scover i es literally one £f ter the other (Fig. 9). In particular, he noticed and correctly interpreted a lot of emissions due to nitrogen. Due to his important discoveries in the auroral spectra, his name is kept far the Vegard-Kaplan bands, which he was the first co observe. 14 AURORAL EMISSIONS - GAS TEMPERATURE It Wr.s professor Lars Vegard who pointed out that some auroral emissions contain information about the gas tempera ture from which the light originates. These emissions will change with temperature variations in the gas. Vegard also proposed a method to measure the gas temperature. Before rockets and satellites were used in atmospheric research, this was the only method to obtain gas temperatures in the height region 100-300 km. VEGARD DISCOVERED PRDTON AURORA IN 1939 The discovery which made Vegard famous was that of the proton aurora. Already before 1920 Vegard argued that auro ral electrons must be accompanied by positive nucleus, pre ferentially protons - if not, a polarized electric field would stop the electrons. In 1939 in Oslo, Vegard and his assistant Kvifte, dis covered for the first time hydrogen emisssions in their auroral spectrograms (Fig. 10). They found H—emissions both in Oslo and Trorso the following winter. In a popular article from 1952 Vegard described these previous discoveries as fol lows: "In 1939 in Oslo we recorded a series of Auroral spec trograms where hydrogen spectral lines appeared with considerable strength. This showed that there åre times when the Sun sends out positive ions (protons). In the following 2 years we took several spectrograms of strong hydrogen lines. At the Auroral Observatory in Tromsri during both winters of 1939-1940, we made a series of strong spectrograph!c exposures, with a spectrograph having considerably better dispersion than the one we had used in Oslo. With the aid of this instrument, one could establish for the first time with absolute cer tainty that we were dealing with the hydrogen-line and in addition, that on some of these spectrograms the hydrogen line had spread out into the form of a small band and had also strongly shifted towards shorter wavelengths. This broadening and shifting toward shorter wavelengths could not be attributed to defects in the picture because other nearby lines were very sharp. Therefore, the only explanation was that the hydrogen atoms, which produced the light, were moving at high speed and preferentially in the direction toward the observer, or at any rate down toward the Earth." This discovery opened a new window into space. From measurements of Doppler broadening and Doppler shifts, the 15 particle speed i nco/ni ng protons». PERSONAL CHARACTERISTICS Professor L. Vegard was the -founder and the chairman of The Norwegian Institute of Cosmic Physics (N1CP) in the late 20's. In February 1925 he applied for funding from the Rocke feller Foundation for an auroral observatory to be located in the vicinity of TromseS. In May 1927, the Roc kef el ler Foundation approved the plans and granted $• 75.000,- to be used far building and equipment. The first director was Leiv Marius Harang (1902-1970) who started his work at the Auroral Observatory in TromsaS in the summer of 192B. In 1932 the German engineer, Willy Stoffregen, became associated with the staff. Vegard continued as chairman of NICP until 1954- Already in 1915 Vegard bui It some advanced, sensitive pr ism-spectrographs for auroral studies, but he still needed an exposure time of at least 1 hour. Vegard also played an important role in moving the Un i ver5i ty from the center of Dslo up to Blindern where it is located today. His pl ans were carri ed out in the 30's, and it was the Faculty of Science and Mathematics which first moved to the new campus. Vegard was also active in local politics, representing the liberal party, Venstre, for one period (i.e. 4 years). He was active in the debate in reference to technical prob lems within Oslo. Thus, he suggested tunnel constructions in connection with the new subway in Oslo (Holmenkollbanen) in the 20's. In particular, Professor L. Vegard -far several years worked part time an "ice-problems" in rivers for plan ning of new hydroelectric power stations. He was considered as the Norwegian expert on icing in fresh water streams and lakes. Vegard was also an active hunter and fly-fisher, par ticularly in the areas in Southern Norway where he was born. Mrs. Anna-Sophie Andresen, Vegard's last secretary- has upon request - written the following personal views: 16 "Professor L. Vegard was a man of tradition, and it was through tradition that I met him for the first time. I had just finished my first affice-jab when my mother got a telephone call from Vegard wondering "If anyone in the family was free to take a secretarial job with him". My mother had worked for him earlier and so had two of my aunts. After an "interregnum" with an unknown secretary "he wanted to come back to his old family". He did not ask for my qualifications as a secretary at all, belonging to "his old family" was testimony enough, so I was literally hired over the telephone. When I first came to the Institute of Physics, he was 63 years aid and I worked for him until he died, still working, at the age of 80. He lived very close to the University, and every day he took his car and drove very carefully the 4—500 meters to his office. He was most precise and correct to the very last. Towards me, who started as a young secretrerial assi- tant, he was always correct and amiable. He was, howev er, very demanding in his way. If he h^d worked hard on some thesis, and I was not there for some reason, he would telephone my home and ask, and I mean explicitly ask me to come to the office. Another thing that in my opinion was very characteristic •f him, was his enormous ability to concentrate. You could go in and out of his office without him even noticing. I remember very well once I was going to serve him tea in his office, which I did every day. When I put the tray on his desk, the cup dropped to the floor with a loud noise. I picked up the shards, went for a new cup and put that in front of him, whereupon he exclaimed: "Is the tea ready already ?" Vegard was, as far as I can judge, a very hardworking man. There were not many pauses at the office during the day. He took his work very seriously and was invol ved in many different activities. He wrote very neatly everything down. He collected all he had written, from the first draft, through several improved editions, to the final results. His office bulged with manuscripts in every form. But he kept order, and knew were he had his things. Some times during his last years I came to his house to take dictation. He came down from his bedroom dressed in night-shirt and dressing-gown and did his dictations. He was very thankful and very touching. He had his absolute likes and dislikes, and he expressed them. To be sure, he was not an easy man, a truly patri- arcal figure, but as already mentioned, he was always thoughtful and nice to me. He required the work to be properly done, and when he got it, he was considerate and gentle." 17 CONCLUSION'S Since Norway is located in and r.lub^ tu the belt of minimum auroral occurrence, this may explain why Norwegian =.u i t.»r. L i st "5 htwf made? -iui_h a s i qn i f i c ~ut t car.tnbui ions in r'*, ii- -„tl r .„'_.-?-.-r ...'i. Thi'=, howe ,-er , Joe i not explain why Mcr w<^ • j . 'ib i -i - w t-s;?.! •mz-rs jcti /e in thi5 field than scienhiti in :••. " n~? L.J;bu;'aiy •- aun tries. The ba.ne i 5 even true m netsa"'^- ! rg.. . Thau, i : seetns th=*t Morwe.'jidns h^ivs con-:entr ated on *' • ";.; t'Zfoe-r " n ^i.:.»ral 5C i ences. F--c.il *:h: -i b-i.= f review it should be c;~£tr that Pirfe- 1 -.ri-j, Sttfrmer , ard Vegard were the first to apply precise methods ta study aurora and as 5oci at ed phenomena. They were a I sa tiLi= -first to proo&'Bs a realistic theory and to calculate the motion of tast electrons from the Sun to the Earth s polar atmosphere- Through their research, these pioneers discovered many new effects and laid the foundation of our or ^s=n t-day e;;pl or a t i on of aurora from space. REFERENCES Sir' slsr-.d, k". U90S, 1913): The Norwegian Aurora Polaris Expad i tien 1902-1903. Aschehoug, Christiania. Brekke, A. and A. Egeland, 1983: The Northern Light, Springer Veri ag, Hei del berg. Brekke, A. and A. Egeland t 19eå.; Christopher Hansteen < 1764- 1S73 - A pioneer in the study of terrestrial magnetism, EOS (in press). Chapman, 5. 1960: Biographical Memories of the Fellows of the Royal Society, London 4, 257. Eather, R., 1930: Majestic Lights. Am. Geophys. Union, Wash- i ngton DC. Egeland, A., 1984: Kristian Birkeland: The man and the sci entist, p. 1 in Magnetospheri c Currents (Ed. J.A. Potem— ra). Geophys. Monograph 28, Am- Geophys. Union, Washing ton DC. Tromholt, S., 1SS3: Under the Rays of the Aurora Borealls, Houghton-Mi ttlin, Boston. FIGURE 1: Christopher Hansteen (1784-1873), an astronomer and geophysicist professor at the University of Oslo, the founder of the astro nomical observatory at the university and its first director. He devoted himself to a study of geomagnetism and travelled for this reason (1818-30) through Siberia to China. The geomagnetic data collection obtained by Hansteen was an important element in Gauss's epoch- making theory of geomagne tism. FIGURE 2: Sophus Tromholt (1851-96) born in Denmark and edu cated to be a school teacher. He was immensely interested in science, particularly in astronomy and meteorology. In 1875, he came as a teacher to Bergen, Norway, a position he gave up in 1882 to devote himself completely to auroral research. In order to complement his studies with personal observations, he spent the winter 1882-83 in Kauto keino, Finnmark and 1883- 84 in Reykjavik, Iceland. 19 FIGURE 3: Kristian Olaf Bernhard Birkeland (1867-1917), Norwegian professor in physics. He was originally devoted to mathematics but later became interes ted in physics. He was appointed professor at 31 years of age and introduced modern experimental physics to Norway. FIGURE 4: Birkeland made his own "space" in the laboratory to create artificially northern light. He is shown here with one of his most devoted assis tants, 0. Devik. The Earth is represented by the "terrella" which is surrounded by a glow. To the right is shown the "auroral zones" as produced in Birkeland's experiments. FIGURE 5: Carl Fredrik Mulerz Størmer (1874-1957) was ap pointed professor in pure mathematics in 1903 at the University of Oslo. Professor Birkeland introduced him to the problems in auroral physics in 1902, a subject to which Størmer devoted most of his life. FIGURE 6: This figure shows one of Størmer's wire models displaying his calculated particle trajectories through the magnetosphere. LM WK kx» "i *00 ;*00 i I t FIGURE 7: The height distribution of 12 330 auroral points as measured by Størmer and his assistants from Southern Norway 1911-44 by photographic triangu- lation. To the left auroras in the dare atmo sphere, and to the right sunlit auroras. 22 FIGURE 8: Lars Vegard (1880-1963), a Norwegian professor in physics, standing beside his spectrograph. Vegard worked with Birkeland from 1905 to 1913 and was appointed professor in physics at the University of Oslo in 1918. Vegard was one of the modern pioneers within auroral research and he contribu ted strongly to the understanding of the auroral spectrin. 2. P. 0. i i i i - z\ • mmm 111 .- v G ....'r -- FIGURE 9: Auroral spectra as measured by Vegard in the 1940-ies. 23 FIGURE 10: Some of the original observations made by Vegard showing the Doppler broadening of the HB line in a proton aurora. HB lines without Doppler broad ening and Doppler shift are shown at the bottom for comparison. EXCITATION MECHANISMS AND SPECTRAL EMISSIONS FROM THE UPPER ATMOSPHERE A. Vallance Jones •Herzberg Institute of Astrophysics, National Research Council of Canada, Ottawa, Ontario, Canada ABSTRACT This review provides a summary of the principal auroral and alrglow emission features and the processes responsible for their excitation. Processes for which our current understanding is incomplete are discussed in more detail. 1. INTRODUCTION This review is Intended to be a very brief introduction to the central topic of optical observations of aurora and airglow. Many of the aspects of the material were reviewed later In the meeting and consequently it is not necessary or possible to cover everything in this initial discussion. Here I shall try to present a broad summary with some discussion of sped .1 points of Interest. This material will also supplement a review of the subject by VALLANCE JONES et al. (1985) which has appeared very recently. References for the latter publication are denoted with an asterisk after the year and are not listed below. A few older references from VALLANCE JONES (1974) are likewise not listed and are marked with a double asterisk. 2. OVERVIEW Ten years ago, detailed knowledge of the airglow and auroral spectrum was confined to the region accessible from the ground and for wavelengths shorter than 1 um together with some pioneering knowledge of the far UV and the infrared regions. Today rocket and satellite techniques have provided very good data through the ultraviolet region to wavelengths as short as 400A and out to 15 |im in the IR. Table 1. Summary of Upper Atmosphere Emissions. Emitter Emission A/System Nightglow Oayglow Twilight Aurora [01] 6300, 5577, 2972 X X X X 01 1304, 1356, 3947, 7774, 7990, 8446, etc. x,a X X X [NI] 1.04u, 5200, 3466 X X X X NI 1200, 1493, 8216, 8629, 8680, 10115 X X X [Oil] 7320-30, 3727-9 X X X Oil 538, 834, 3749, 4415, 4649 X X X [Nil] 6584, 5755 X X X Nil 746, 4044, 4242, 5001, 5860 X X X b N IP, 2P, VK, LBH, BH, GH X 2 ? * * «2 + IN, M X X X X X X 02 Hzl, HzII, Ch, At, IRAt X °2 + IN X X X NO Y, 6, e 2.7u, 5.3u X X X X Na.K.Li X X X Ca+,Mg+ X X X N02 Visible continuum, 4.5u X X X X CO 4.7u X X N0+ X X X co2 4.3u, 15u X 9.6|i, 14.4u X X 03 OH - 1 * 1 X Notes: not 8446; LBH? Table 1 provides a very brief summary of most of the important emission features of the spectra and their occurence. Most of the features in the table are identified in the energy level diagrams of molecular and atomic oxygen and nitrogen and of NO which are reproduced in Figures 1 to 5. In general, the auroral emissions are those which are excited by electron impact while nightglow features arise from chemical reactions. 26 •f V «"»• *>t «t* Figure 1. Energy levels of atomic oxygen and transitions observed in aurora. DOUBLETS M' *P Figure 2. Energy levels of atomic nitrogen and transitions observed in aurora. 27 —i—«-i; 1 «-i; |>tl MftATIVI f ZZTnZ ^«'ij j r: ' i s «;•%=! fl3z; - ei: HERZBERG I CHAMBERLAIN NOXON ATMOSPHERIC / lyt ZE ATMOSPHERIC Figure 4a. Energy levels and transitions of molecular oxygen observed in aurora. Figure 3. Energy levels and transitions of molecular nitrogen observed in aurora. _ ' _i - WO -1 -'-, - DZI* -A»I; - -1 A'I* ! ID _ \ -e"4„ - Q(3p) • Q(3p) ! o _ 1 a 8 { e'T^ • r \-ia . -b'r; - i -.'A, JO - -x»i; INTERATOMIC DISTANCE - Figure 4b. Potential curves of Figure 5. Energy levels and 0, states (including hypo transitions in NO. thetical 5JI state). 28 Dayglow and twilight emissions Include both electron Impact excitation by photoelectrons as well as direct photolonlzation, photodissociatlon and fluorescent processes driven by solar radiation. The following sections discuss a number of topics of current Interest for which problems in interpretation or observation persist. 3. THE EXCITATION OF THF 0(lS) STATE 3.1 Excitation processes The important candidates are listed in Tables 2 and 3. Table 2. Excitation processes for OC'S) In the nightglow. 0 + 0 + 0 »- 02 + 0(*S) CHAPMAN 0 + 0 + M - 02 + M | +* BARTH 0, + 0 • O^S) + 02 j „+* 0(lD) + OC'S) denotes electronic excitation; denotes vibrational excitation. Table 3, Excitation processes for 0(*S) in the aurora. 0 + e (fast) 0('s) + e Oj"1* + e * 0(1S) + 0(lD) 1 N2(A) + 0 +• N2 + 0( S) 02 + e (fast) » 02(c?) + e LLEWELLYN-YAU 0** + 0 P- 02 + 0('S) 1 N+ + 02 p. 0( S) + NO 3.2 Nightglow The Barth mechanism is generally accepted to be the only important E-region process. There are good arguments that the intermediate state is the c state, (see, however, TORR et al., 1985). However, KRASNOPOLSKY (1985) has recently argued that the suggestion of WRAIGHT (1982*) that the intermediate could be the 5n state has some merit. The potential curves involved are shown in Figure 4b. However,it is not known if the latter state really has the necessary potential minimum. 29 0, dissociative recombination Is Important in the F-region component. GUBERMAN (1985) has recently improved the quantum mechanical calculations which Indicate that the yield of 0(lS) Is only significant for v>2. 3.3 Aurora and dayglow The mechanisms which have been thought to be important In aurora (and dayglow) are listed in Table 3. Direct electron impact is less important in recent models because of the low concentration of atomic 0 favoured In current work coupled with decreased estimates of the low energy secondary electron flux. The work of BURNS and REID (1984*) also supports the view that a major portion of the excitation Is through an indirect mechanism with a short (O.ls) time constant. The N2(A)/0 energy transfer process appears adequate to explain observations while laboratory measurements indicate that it provides an adequate yield of 1 0( S). The 02(c)/0 energy transfer process suggested by SOLHEIM and LLEWELLYN (1979*) would also be an effective mechanism if indeed the stable levels of the c state can be excited by electron impact. Electron loss spectra, e.g. those measured by TRAJMAR et al. (1972) and reproduced in Figure 6, do not show appreciable losses below 5 eV and consequently it is difficult to see how the c state excited by electron impact can be stable. However, some drift tube experiments of Phelps (e.g. YAMABE and PHELPS, 1983*) suggest that stable c state molecules can be so excited and consequently the mystery remains. The (T^/Oj mechanism suggested In 1983 by REES (1984*) is now believed (LANGFORD et al., 1985) to produce 0(1D). This suggestion had the advantage of explaining the frequently observed constant ratio with height between 1(5577) and KN^). + The 0-/e dissociative recombination mechanism which has long been considered of significance for E-region aurora will be suppressed if the result of Guberman noted above is correct since vibrationally excited 0, would be too rapidly quenched. Modelling studies also suggest that any significant contribution from thle reaction leads to longer than observed time constants In rapidly fluctuating auroral lower borders (GATTINGER et al., 1985*). 30 WAVELENGTH ill Figure 7. FUV spectrum of low latitude nightglow after CHAKRABARTI et al. (1984*). i 5 6 7 8 9 10 ENERGY LOSS (.v) Figure 6. Electron impact energy loss spectra for 0 after TRAJMAR et al. (1972). •S° *D 'D"F •S0,P 16 14 - IP Is' 2s* 2p" (4S°> 3p'—<- V 1361 ov 3l' - 12 K) 3»—r- / > 8 J 1 l _ ' 01 ABOUND STATE 6 u'a'zp'V, EXCITED STATES - 4 bfirøtVlto \ / / 2 lrø&WKnp; . 0 VzS Figure 8. Energy levels involved in dielectronic recombination process for production of the 989A line of Figure 7. after ABREU et al. (1984*). 31 4. EXCITATION OF O('D) Excitation processes are listed In Table 4. In the nightglov, dissociative recombination of 0, remains the major mechanisms. Table 4. 0('D) alrglow production mechanism. + + 0 + 02 02 + 0 + 02 + e OC'D) + 0 0 + e (fast) » 0(1D) + e 3 1 02 + hv » 0( P) + 0( D) 2 1 N( D) + 02 " NO + 0( D) + l N+ + 02 N0 + 0(0) 2 In aurora the N( D)/02 mechanism still appears to be a strong candidate to supplement dissociative recombination even though the analysis of rocket measurements by LINK (1983*) led to the opposite conclusion. REES (1985) has recently redone theoretical calculations of 2 the 6300A profile with the N( D)/02 reaction Included. Older results for the primary electron energy vs the ratio I(6300)/I(4278) are not greatly altered in the new results. The importance of the excitation of \63O0 by high temperature thermal electrons has recently been demonstrated both in Type-A red aurora and in cleft studies by WICKWAR and KOFMAN (1984*) and by ROBINSON et al. (1985). 5. HIGHER LEVEL ATOMIC 0 LINES Some processes leading to excitation of higher levels of atomic 0 are listed in Table 5. Transitions from these states Involve some unusual features because some of the transitions to the ground state, (see Figure 1), are strongly allowed and exhibit radiation trapping effects. 5.1 Nightglow It has been believed for some time that the excitation of higher levels is possible as a result of two-body recombination of 0+ with either electrons or negative ions. Satellite spectra, (see Figure 7), 32 Table 5. Excitation of pernitted lines. 0+ + e » 0 0 + e (fast) »• 0 + e 02 + e (fast) » 0 + 0 + e N + e (fast) • N* + e N2 + e (fast) N* + N + e N2 + hv •- N + N* obtained by CHAKRABARTI et al. (1984*) of the 911A continuum edge confirm the usual supposition that electrons are involved in the recombination. This process produces the UV nightglow lines as well as the 7774A feature. Chakrabarti et al. have also observed the 989A line in the same spectrum of the nightglow. This is remarkable because ordinary recombination would involve recombination with an excited ion. Since this is an unlikely process a different mechanism is required. The emission has been explained by ABREU et al. (1984*) as an Instance of dielectronic recombination. The steps involved in dielectronlc recombination are illustrated in Figure 8. 5.2 Dayglow and Aurora The dayglow and auroral spectra involving the permitted 0 transitions show Interesting resonance trapping effects. There have been three important recent papers on dayglow theory and observations (EASTES et al., 1985; CONWAY and CHRISTENSEN, 1985; MEIER et al., 1985). There are still difficulties In getting a consistent interpretation of available rocket spectra In the far UV. There is also a related mystery concerning the 989A and 7990A 01 transitions in aurora. Their relation is shown in Fig. 1. The 989Å transition is optically thick and consequently the 7990Å transition, labelled in Figure I as 7987-95, should be enhanced since for every resonant reexcltation to the upper level there is a chance of branching to the 7990A transition. However the available laboratory data predict that the 7990A line should be much stronger than observed. 33 Another interesting question is the different behaviour with height of the 8446A and 7774A 01 lines. The former has a much more extended height distribution than the latter. It has been speculated that the reason is that 7774A is strongly excited by dissociative exci.atlon of O2. This has recently been confirmed by HECHT et al. (1985), (see Figure 9), who have shown that the 7774A line profile has a broad pedestal indicating excess kinetic energy produced in the dissociation. The 8446A line does not show this behaviour. It may be an excellent candidate on which to base an energy sensitive spectral ratio since it may be excited in a straightforward way from atomic 0. Another interesting transition is the 1641A line of 01 recently discovered by MEIER and CONWAY (1985). This leads from the upper state of \1304 to 0(1D). It could be useful in estimating atomic oxygen abundance. 5.3 Forbidden lines of Ionized Oil The 7320-7330A multiplet Is of Interest because It is excited directly In aurora by electron Impact and because it permits the determination of Ion drift velocities from the Doppler profile (SMITH et al. 1982*). It falls within the (5,3) N2 First Positive band and consequently Is seen easily only in aurora excited by soft electrons such as cleft or type-A aurora, or with high resolution. 6. MOLECULAR OXYGEN BAND SYSTEMS Excitation processes for molecular oxygen are summarized in Table 6. 6.1 Nightglow Several band systems arise through three-body recombination. In the terrestrial nlghtglow the atmospheric, the IR atmospheric, the Herzberg I and the Chamberlain systems have long been known. SLANGER and HUESTIS (1981*, 1983*) have detected the Herzberg II system which is the most prominent 0^ system in the Venus nightglow. The c state is believed to be the precursor of both the b state and 0(1S) although the evidence is Indirect. The vibrational distributions of the excited states and the effects of quenching pose interesting problems. 34 80 100 120 140 160 180 200 BIN NO. Figure 9. Line profile of 777AA line in aurora after HECHT et al. (1985). I^Vil mo sooo MO 2040 WAVELENGTH (A) Figure 10. Resolution of 2143A Nil line from NO, 1,0 band in the dayglow after TORR and TORR (1985). 35 Table 6. Excitation of 02 bands. Processes In nlghtglow 0 + 0 + M • 02 + M ** * o2 + o2 » o2 + o2 1 0 + OH(v>l) •- 02( A> + H 02 + O('D) . 02(b)v>2 + 0 Additional process In dayglow and aurora 02 + e (fast) " 02 + e 6.2 Aurora The atmospheric and XR atmospheric bands are excited In aurora while the other transitions have not been detected. The atmospheric bands are probably excited by the 0(1D)/0, reaction as well as by electron Impact although the relative contributions of the two mechanisms are uncertain. These mechanisms should only excite vibrational levels up to 2. It is consequently an Interesting problem that bands with v<5 are observed. Moreover the v«l progression shows a strong relative enhancement with height whereas the higher progressions do not (GATTINGER and VALLANCE JONES, 1976). This implies that the v-1 level is produced from 02 and vibratlonally quenched by 02. The higher levels are either not so quenched or are dominantly quenched by a species which increases rapidly in relative concentration with height such as 0. SLANGER and TAHERIAN (1985) noted that the reverse reaction to produce 0(^D) may be fast for v>2. The production of the levels with v>3 is also a mystery. It has been suggested that the source could be "hot" 0('D) produced by 0j+/e reaction or vibrationally excited 02(c). There is an apparent problem with estimates of the system Intensity from rockets based on observations of the 0,0 atmospheric band and from the ground based on the 0,1 band. The former are generally about 10 less than the latter (see e.g. McDADE et ul., 1985). It could be a height and composition effect in that the ground meas'irements are generally made with stable strong aurora whereas rocket studies may have been made with more normal forms at greater heights. This needs to be properly resolved. 36 There nay be a problem with the excitation of the IR atmospheric bands in aurora. The long radiative lifetime of the a-state (60 mins) complicates observation but careful measurement from a fixed ground location suggests that .approximately one a-state molecule is produced for every ion. Theoretical calculation with a REES (i964**) type model Indicated that this is Just possible with direct electron impact excitations of Oj (GATTINGER and VALLANCE JONES, 1973c**), However, this type of model is suspected of overestimating the low-energy secondary electron flux and if this is so, and the laboratory cross sections are correct, then some other mechanism to excite the a-state may be needed. 7. MOLECULAR NITROGEN BAND SYSTEMS 7.1 Nightglow Observations of the LBH bands in the mldlatitude nightglow by satellite and shuttle (HUFFMAN et al., 1980a*; MEIER and CONWAY, 1983*; and TORR et al., 1985) are puzzling since it is difficult to imagine a nightglow process which could excite the upper state. It is also puzzling that the vibrational distribution maximizes at v=0. Some other good satellite measurements (CHAKRABARTI et al., 1984*) did not find these bands. It is not impossible that local glow effects associated with lower altitude spacecraft could be Involved. 7.2 Dayglow and Aurora Various N, band systems extend from the IR to the EUV. Observed transitions are shown by the solid lines in Figure 3. The Birge-Hopfield i and c. -X bands are strong In the laboratory but only remnants occur in aurora. This is probably due to predissociatlon of the molecule under i optically thick conditions. The c. -a, Gaydon-Herman, transition is responsible for a few features in the 3000-4000Å region which were long unidentified. The LBH system is prominent in FUV; there may be some problems with its vibrational population distribution (CONWAY 1983*). 7.3 No+ band systems The First Negative and Meinel systems are among the strongest auroral emissions. In normal aurora the rotational structure of these bands corresponds to kinetic temperatures in the region where the bands are excited. The IN system shows extended rotational and vibrational 17 distributions in sunlit, type-A and cleft aurora (see e.g. HENRIKSEN, 1984) which occur at greater heights and may also involve pumping effects. Extended rotational distributions have also been reported in equatorial aurora (excited by ions leaking out of the ring current by charge exchange) (TINSLEY et al., 1984*). There is some evidence that the ions involved include species heavier than protons. 8. NITROGEN ATOMIC FEATURES The permitted atomic features, summarized in Figure 2, are strong in the far UV with numerous less prominent lines appearing from the near UV to the near IR. They are presumably excited by dissociative excitation of N2 and as such their profiles should be broadened. There Is a problem with the 1200A line which is apparently optically thick despite the expected non-thermal broadening (MEIER et al. 1980*). The forbidden 520QS feature is always Interesting because of the very long lifetime of the 2D state and its chemical activity. It is known to react with 0, to produce NO and it has been invoked as a L possible precursor of O^S), 0( D), 02(a), 02(b) and 02(c). The 2143A forbidden transition of Nil appears certainly to be responsible for the "2150A mystery feature" in aurora, once ascribed to NO. This feature has been clearly separated from the NO band in Spacelab dayglow spectra by TORR and TORR (1985). Some of these spectra are reproduced in Figure 10. 9. OH AIRGLOW EMISSIONS The excitation mechanism for OH vibration-rotation bands has been shown to be mostly due to the Oj/H reaction. There are however still problems In a quantitative understanding of the role of quenching and cascading in relation to the observed vibrational populations which are themselves not entirely established because of uncertainties in the transition probabilities. The bands are of course extremely uBeful as indicatorj of the mean atmospheric temperature in the emission region. 10. METALLIC EMISSION FEATURES The Na, K, Li, Ca+ and Mg+ emissions occur in the twilight and 38 dayglow. Sodium Is, of course, prominent In both twilight and nlghtglov spectra. Recent treatments (KIRCHHOFF and CLEMESHA, 1983a*; THOMAS et al., 1983*; JEGOU et al., 1985, GRANIER et al., 198S) appear to account for the height profiles of free sodium (most accurately determined by lldar techniques). 11. RjO* TWILIGHT EMISSION The occurrence of features In twilight spectra near 0.614 and 0.619 |ii reported by KRASSOVSKY and VINIUKOV (1979) led HERZBERG (1980) to point out that there was a remarkable correspondence between the features reported and the HjO* lines found in Comet Kahoutek and produced in the laboratory by LEW (1976). The coincidences appeared remarkably good. More recent work by MATVEEVA (1983*) and S0B0LEV (1983*) provides evi dence that lines of these bands are present In the twilight but at a much + lower intensity than originally reported. The strongest real H20 lines appear to lie at 6168 and 6175A coinciding with the two strongest lines of the 080-000 band at 500 K. The 6175Å line is the most free from interference and its intensity is estimated to be about 3R giving a total band intensity of perhaps 30R: "his is much less than some of the earlier values of at least 200R mentioned by KRASSOVSKY et al. (1982). According to Matveeva the original features at .614 and .619 \tm are mostly due to high-pressure gas discharge lamps and it was a coincidence that the wavelengths were close to the cometary features which are peculiar to H20+ at temperatures less than 50 K. The spectrum showing these features published by MATVEEVA (1983*) is shown in Figure 11. KRASSOVSKY et al. (1982) have provided a theoretical model of the transport of H«0 into the thermosphere and excitation of the bands. This is an interesting problem and independent observation of the emission would be very desirable. 12. INFRARED MOLECULAR EMISSIONS + Infrared emissions of NO, N0 , 03, CO, from the upper atmosphere in aurora and airglow have been extensively observed and modelled. This topic has been reviewed by GORDIETS et al. (1978) and more recently by GORDIETS (1985). The observations require rocket techniques and cooled detectors. CONCLUSIONS In this introductory review I have attempted to highlight some new 39 MHjS-Qjm -. •-..-•- Figure 11. H,0+ rotational lines after HATVEEVA (1983*). The diffuse high-pressure arc lines are clearly visible. 40 and continuing problems In the field of auroral and airglow spectroscopy and excitation processes. The extension of the observable region into the ultraviolet region has been particularly fruitful in recent years and has stimulated more detailed examination of some features observable from the ground. The other Important development has been the systematic studies of cleft aurora which have yielded interesting illustrations of the effects of excitation by much lower energy electrons than those of night time oval aurora. REFERENCES Conway R.R. and Christensen, A.B. The Ultraviolet dayglow at solar maximum: 2. Photometer observations of N2 Second Positive (0,0) band emission, J. Geophys. Res. ^0, 6601, (1985). Eastes, R.W., Feldman, P.D., Gentieu, E.P. and Christensen, A.B. The Ultraviolet dayglow at solar maximum: I. Far UV spectroscopy at 3.5A resolution, J. Geophys. Res. 90, 6594, (1985). Gattinger, R.L. and Vallance Jones, A. The vibrational development of 1 + 3 the 02(b ^ -X J") system in auroras, J. Geophys. Res. jU, 4789 (1976). Gordlets, B.F., Markov, M.N. and Shelepln, L.A., IR radiation of the upper atmosphere. Planet. Space Scl. 2b, 933, (1978). Gordiets, B.F. Excitation of vibrational emission bands in aurora and airglow, Symposium on Alrglow and Auroral Excitation and Models, IAGA, (1985). Granier, C., Jegou, J.P., Chanln, M.-L. and Megie G., General theory of the alkali metals present in the earth's upper atmosphere. III. Diurnal variations, Ann. Geophys. _3> 445, (1985). + Guberman, S.L. 0(1 S) from dissociative recombination of 02 , Symposium on Airglow and Auroral Excitation and Models, IAGA, (1985). Heche, J.B., Christensen, A.B. and Pranke, J.B. High resolution auroral observations of the 01 (7774) and 01(8446) multlplets, Geophys. Res. Lett. J_2, 605, (1985). Henriksen, K. Nj4- emissions in sunlit cusp and nightside aurora, Ann. Geophys. 2_, 457, (1984). Herzberg, G. Hj0+ ions in the upper atmosphere, Ann. Geophys. _36_, 605, (1980). Jegou J.P., Granier, C., Chanin, M.-L. and Megle G., General theory of the alkali metals present in the Earth's upper atmosphere. I. Flux 41 model: chemical and dynamic processes, Ann. Geophys. 3, 163, (1985). Krasnopolsky, V.A. Oxygen emissions in the night alrglrv of the earth, Venus and Mars, Symposium on Airglow and Auroral Excitation and Models, IAGA (1985). Krassovsky, V.I., Rapoport, Z. Ts. and Semenov, A.E. New emissions of the upper atmosphere as a consequence of anthropogenic Influences on the Ionosphere, Cosmic Invest. ^0, 237, (1982). Krassovsky, V.I. and Vlniukov, K.I., New emissions in the earth's atmosphere during twilight, Ann. Géophys _3£t 109, (1979). Langford, A.O., Burbaum, V.H. and Leone, S.R., Auroral implications of recent measurements on 0(1S) and 0(JD) formation in the reaction of N+ with 02. Planet. Space Sci. Q 1225, (1985). Lew, H. , Electronic Spectrum of HjO"1", Can. J. Phys. 5^, 2028, (1976). McDade, I.e., Llewellyn E.J. and Harris F.R., A rocket measurement of the 1 + 3 O2(b J] -X J~)(0-0) Atmospheric Band in a pulsating aurora, Can. J. Phys., 63_, 1322 (1985). Meier, R.R. and Conway, R.R., The 1D-3S transition in atomic oxygen: a new method of measuring the 0 abundance in planetary thermospheres, Geophys. Res. Lett. _1£, 601, (1985). Meier R.R., Conway, R.R., Anderson, D.E. Jr., Feldman P.D., Eastes R.W., Gentleu, E.P. and Christensen, A.B. The ultraviolet dayglow at solar maximum: 3. Photoelectron-exclted emissions of N_ and 0., J. Geophys. Res. _9£« 6608, (1985). Rees M.H. The auroral (01) 6300A red line revisited, Symposium on Airglow and Auroral Excitation and Models, IAGA (1985), also to be submittted to Reviews of Geophysics. Robinson, R.M., Mende, S.B., Vondrak, R.R., Kozyra, J.U. and Nagy, A.F., Radar and photometric measurements of an intense Type A red aurora, J. Geophys. Res. 90, 457, (1985). Slanger T.G. and M.R. Taherian, 03 photodissociation at 157.6 nm, IAGA (1985). Torr M.R., and Torr, D.G., The Nil 2143A dayglow from Spacelab I., J. Geophys. Res. 90_, 6679, (1985). Torr, M.R., Torr, D.G. and Eun, J.W. A Bpectral search for Lyman-Birge-Hopfield band nightglow from Spacelab 1, J. Geophys. Res. 90, 4427, (1985). 42 Trajmar, S., Williams W. and Kuppermann A., Angular Dependence of Electron Impact Excitation Cross Sections of 0,, J. Chem. Phys. 56, 3759 (1972). Vallance Jones A., Meier, R.R., and Shefov, N.N., Atmospheric quantal emissions: a review of recent results. J. Atmos. Terr. Phys. 47, 623 (1985). Vallance Jones, A., Aurora, Reidel Publishing Co., Dordrecht, Holland (1974). 43 ANALYSIS OF AURORAL O* FIRST NEGATIVE BANDS K. Henriksen Institute of Mathematical and Physical Sciences, University o-f Tromsø. N—9001 Tromsc, Norway and L. Veseth Institute of Physics, University of Oslo., 0315 Oslo 3, Norway. Abstract An e: observed auroral 0n 1NG- bands by use of a least squares method. In this way we determine the population densities of the vibrational levels of the upper 0 electronic state, and in addition an average rotational temperature. Our results give clear evidence that the auroral 0„ IN jdnd-a Are mainly generated by particle impact on neutral 0^, molecules in their electronic and vibrational ground states, and that the bands sre produced within the E- region. 44 1. Introduction Several 0^ INtbands can be clearly seen in spectrome try measurements of bright auroral displays. It is generally agreed that these bands are produced by particle impact on neutral 0 molecules in their electronic- and z 1 vibrational ground states, as outlined by Val lance Johes . however, the different bands overlap, and the rotational structure of each band is rather complex, since a total of 48 oranches ar^ involved. To gain information from such complex overlapped band spectra we have used a numerical technique to calculate the rotational part of the line strengths (Hbnl-London factors), and to compute synthetic band spectra. The synthetic spectr um is then in turn fitted to the observed spectrum in digitized form by use of a least squares technique, and in this way we finally obtain the population densities of the vibrational levels of the excited electronic state. In the present study we present tie spectrometric observations, carry out the analysis as outlined above, and our results are in good agreement with the idea that the 0o INfi bands are produced by direct particle impact on neutral 0„. The CL, ions play a significant role in the various physical processes of the ionosphere, and with our method we obtain the production rates into the vibrational levels + 4 - of the 0„ , b E electronic state. The populations of the 4 - g b E vibrational levels yield a key to estimate the total g production rate, and accordingly the part played by dissociat i ve recombination as a source for the auroral green line. Work on this problem is in progress, but will, howevp?r, not be discussed in the present paper. 45 2. Transition Probability and Intensity of Individual Rotational Lines The transition probability (transition rate pr. particle) for a spontaneous radiative transition from an upper rotational level J* of a molecular system to a lower rotational level J" is given by= fl^ vS v^" T^V . (1) UJ-J- denotes the energy difference Ej- - Ea- in wave numbers (cm-1), and the denominator 2J' + 1 refers to the statistical weight of the upper level. The line strength Sj-j- is defined by « 3 3 Alii" In eq. (2) Ta-n- and Wo"i-f denote the total molecular wave function for upper and lower states respectively, and the quantum numbers M" and M" are associated with the space fixed E-camponent of the total angular momentum J. The remaining quantum numbers needed to completely specify the molecular state are suppressed so far. Finally, M is the total transition operator /J- = £/** + iM^ * h/*i , ^ and only electric dipole transitions will be considered in the present work. The radiation described by eqs. C1J and (2) is isotropic and unpolari zed (Condon and Snortiey3, p. 99), and we have /*! = *£*; . Ha denotes the space-fi::ed z-component of the transition operator. The expression for the intensity corresponding to the spontaneous transition is accordingly where Nj- denotes the population (particles m~3) of the upper level J'. 46 The important task is now to compute the line strength SJ-J- given by eq. (4). In the present problen the upper - state is a *E, electronic state, and the lower state a *T1U state. With standard definitions of the quantum numbers A and E both states may be represented as* >W,, |*aMl>»^(|.a»rASS3A«>il-0 |«''-*S-2 3/v,>) ((,) in the Hunds coupling case (a) limit, n is a label to identify different electronic states of the same symmetry, - and the phase factor s0 equals one for E states, and zero in all other cases. The case (a) states of eq. (6) are also parity eigenstates, with the upper sign corresponding to parity +1, and the lower sign to parity -1. From eqs. <4) and (6) it now follows that ;(-l)3""S*>*::<-V,/sL,a,«Va«'nV"-^*-JE"J,,/,">. and eq. (7) also reflects the fact that electric dipole transitions on!y occur between states of opposite parity. The nent step is to transform the space—f i ::ed z- component of the transition operator to the molecule-fixed coordinate system. This transformation is roost conveniently carried out by use of the rotation matrices from the theory of the full rotation group. The general result for Hunds case (a) basis states is* , , , , = L<«VA 4r |/*-1|«V"V4i'> Here M<* represents the molecule-fined spherical compo nents o-f the transition operator, defined by AM « + -k (A **/*0 («0 where ::, y, and z now refer to a molecule-fixed 47 coordinate system (z-axis along the molecular axis). In eq. Sl-***~SL a A + X-**i -A - i and M have to be fulfilled for a non-zero result*. With E' = E" eq. (10) leads to -- A- A" J (17) which means that for given values of A* and A" only one term on the right-hand side of eq. (8) will be different from zero. Eq. (10) also informs us that only diagonal elements in the M-quantum number need to be considered. Actual molecular eigenstates are, however, not coupling case (a) states, but are rather obtained as linear combinations of the case (a> basis states'* X (J %) *(.l)a'***"|*u-/lS-2:iA1>). Here k is a label introduced to identify the different eigenstates and eigenvalues of the molecular Hamiltonian, corresponding to the standard term value designations F„, k = 1,2,3.... The coefficients Uc(k) are the columns (column heading k) of the unitar, matrix U which 48 diagonalizes the Hamiltonian matrix H according to the relation OL1H U. * D . (U) Numerical diagonalization of the Hamiltonian matrix is currently a standard technique for computation of molecular energies and transition frequencies*»T. 3. Ifag fe-lSol - a»n^ (IN) Transition in 0a- Our present problem is to compute the line strengths for the rotational lines of the b*£0- - a*n0 vibronic transitions in 03~. In this case we have S = 3/2, A' = 0, A" = 1, and from eqs. (7), (8), and (12) we obtain for the crucial matrix element* This rather complex expression now has to be squared, multiplied by three, and summed over all M-values to yield the line strength for the transition between the rotational levels J' and J" (cf. eq. (4)) . The eigenvectors (coefficients Uc(k') and Uc(k"> are in the present approach obtained from an exact numerical diagonalization of the Hamiltonian matrices of the upper and lower vibronic states, respectively. Hence, exact numerical values are obtained for the line strengths, contrary to the approximate ones that would follow from 49 analytic expression»" The first term on the right-hand side of eq. (14) is the electronic and vibrational part of the matrix element of the transition operator. This term is normally approxi mated by a factorization in an electronic and a vibrational part (Franck-Condon approximation). We shrill also rely on the Franck-Condon approximation, hence we anticipate "- *„•„• K«,'»'**S'*//«-II'»,,A"-I *•*>'*» where q«.„.. is the well-known Franck-Condon factor *,,.,," •' where we have introduced the Hbnl-London factor *k-j-i."j" to account for the rotational dependence of the line strength. The Hbnl-London factor is obtained in a straight forward way from eq. (14), and the complex expression is not reproduced here. The Hbnl-London factors of a vibronic transition in a diatomic molecule obey simple sum rules*''. For the present *E -*TI transition we have (IS) <»-&$ k3l,J V 1*3 U3* kU"3 Vc-K^^^'^lAil^"^^ *" y'>la ^v/as From eq. (19) it is clear that the relative intensities o-f the rotational lines within a vibrational band are determined by the population N, ,,,, of the upper level, the transition frequency, the vibronic part of the transition matrix element, and in particular by the Honl- London factor. 4. Synthetic Spectrum for the 0„ IN Bands Synthetic electronic spectra can now be computed from eq- (19). The electronic matrix element of eq. (19) is a constant for a given electronic transition, and its value is actually irrelevant for the present study, as we will be dealing with relative intensities only. Computed values of the Franck-Condon factors q , „ taken from the literatu— v v 10 re will be used, and the Honl-London factors will be calculated numerically, as outlined in the preceeding sect i ons. For each vibrational leval the populations are given by the Boltzmann distribution 'Va' «" * ZU3fl)e E*i'* ground state by impact of auroral electrons and protons. According to this direct excitation model the 0„ vibratio nal population should be proportional to the Franck-Condon •factors for the transitions -from the lowest 0„ vibrational level to the relevent 0,, level as indicated by Fig. 1. Since the electronic transition is fast compared with rotational redistribution, the rotational distribution from the ground state of 0_ is preserved when the molecules are ionized. Hence, the rotational temperature derived from the 0 IN bands should be representative for the neutral temperature of the emission region. As the 0.., IN bands are allowed transitions, deexcita- tion of the upper levels is mainly due to emission at auroral heights. Then the excitation rate is equal to the emission rate which is proportional to the population of the upper level multiplied by the Franck-Condon factor for emi ssion. The synthetic spectrum is constructed to give a theoretical spectrogram matching the observed spectrum. Using a scanning spectrometer the rotational lines of the observed spectrum Are always broadened due to a finite width of the entrance and exit slits. The motion of the image of the entrance slit relative to the exit slit defines a triangular instrumental function (or a trapezoi dal function if the image of the entrance and the exit slit do not have the same width). When the synthetic spectrum has to match an observed spectrum, the discrete rotational lines calculated bv eq. (19) have to be convolved by a triangular instrumental function having a half width equal to the resolution of the observing spectrometer. Denoting the total synthetic spectrum including all bands of the electronic transition by I(X ,T), we mav wr i te where R , „(fc. ,T) is a quantity that is completely v' v' i 52 determined by the computations at each wavelength V and temperature T. Since all information related to the individual rotational levels is at hand either -from laboratory measurements or -from computations, we have presented the total intensity in eg. <21> as a summation over the individual vibration bands contributing to the intensity at the wavelength X.. The convolution process described above is also inherent in the computation o-f the quantity R , „(V,T), and its dependence on the vibrational band arises -from the vibrational dependence of the molecular parameters, and -from the frequency dependence of the intensity (cf. eq. <19)). Since the Franck-Condan factors q , „ are also assumed to be known, we may v'v introduce the new quantities K. = l< n' A'-- 0 5 -- % 1 /"-i I•*" A"s ' S = V* >l* Nr i and write the expression for the synthetic spectrum in its final compact form as v' recallinq that the quantities Q , (K. ,T) are completely determined from the computations. In eqs. (21)-<23) we have given the intensities at discrete wavelengths V, which is most relevant since the theoretical spectrum will be compared with an observed one in digitized form. From eq. (23) we see that a small number of unknown parameters determine the synthetic electronic spectrum, i.e. the populations of the upper state vibrational levels, and the rotational temperature T, assumed to be common to all the upper state vibrational levels and an average for the emission volume. Adjusting the few unknown parameters of eq. (23) so that the computed spectrum fits the observed one, we may get valuable information about the excitation processes of the 0^ bands, and the average emission height. 53 The mean temperature of the lower thermosphere from 90 to 150 km increases in a fairly linear manner from 150 to 700 K (CIRA, 1972). In Fig. 2 the synthetic spectrum of the 0* • 1NG (1,0) band is shown for three representative rotational temperatures of the lower thermosphere. From the synthetic spectra we see that the rotational line intensity distribu tion is shifted towards shorter wavelengths as the tempera ture increases. These synthetic spectra have been produced with an instrumental function equivalent to a 5-function, causing no broadening of the individual rotational lines. In Fig. 3 similar synthetic spectra are reproduced, but now the triangular instrumental function of 4A is set equal to the resolution of the spectrometer used in the present work. 5. Determination of Adjustable Parameters by the Least Squares Method In eq. (23) there is a linear dependence on the adjust able parameters N'v'. Hence, a simple linear least squares fit12 would be appropriate for a determination of the parameters N'v- from observed values of I(A^,T> and compu ted values of Qv • (*'/T) at a series of wavelengths "(, . The complex non-linear dependence on the temperature T, how ever, requires a more refined technique (non-linear least squares fit). To fit the temperature we have to make a reasonable guess of a starting temperature To so that the relation I(A-T) * IU.,1.) + (—) (T-T,) (24) *• 5T T=Tft will apply. In this way we have linearized also the depend ence on the temperature, and the correction AT = T-To is determined by a standard linear least squares fit, together with the parameters N'v- inherent in I( A^ (To). Then a new "guessed" temperature T'o = £.T + To is fixed, and the whole process iterated until convergence. The derivatives (6I/6T)T=TO are best determined numerically by using fixed 54 values of the parameters N'v' from the preceding run. With a reasonable starting temperature To the iteration process will converge after a few cycles, leaving final adjusted values of the population parameters N'v , as well as the temperature. The linear least squares method is best formulated in terms of matrices12, and yields standard deviations (error bars) for the adjusted parameters as well as the parameters themselves. 6. Spectrometric Observations, Derived Population Densi ties, and Rotational Temperature for the Oz' IMG System The measurements used in this study were carried out near solstice during December 16, 1984, in a relatively bright nightside aurora, covering most of the sky in the form of arcs and patches for more than one hour. We used an lm Ebert-Fastie spectrometer working in the digital-coupling mode, the second order and with 4.0 A resolution13. The spectrometer was calibrated towards a tungsten filament lamp to measure the intensity in absolute units. Each wavelength scan covered the wavelength region from 4900 to 5700 A with a scanning period of 16 s and a sampling time of 25 ms. Satisfactory signal-to-noise ratios were obtai ned by averaging single 16s scans for 15 min, and one of these spectrograms is shown in Fig. 4. The measurements were taken towards magnetic zenith. The auroral green line at 5577 A is the dominant feature as shown in Fig. 4. When the scale is expanded by a factor of 40, the O2* ING bands and other features appear as shown in Fig. 5. The 02* ING (1,0) and (2,1) bands are merged together around the green line, whereas the 02* ING (2,0), (3,1), and (4,2) bands are merged together between 5200 and 5300 A. The (4,2) band is so weak that it was not included in the analysis of the data. The parameters N'v' and T of eq. (23) were determined by a least squares fit to the observed (digitized) intensities at the discrete wavelengths A^. The wavelength region con taining the 5577 A line was deleted. The spectrogram of 55 Fig. 4 together with the fitted 02' , N2* , and atomic line synthetic spectra are shown in Fig. 5. The observational period from 1845 to 2000 UT is divided into 5 data sets, and each data set contains about 150 digitized intensities which are used as input data to the least squares fitting program. The relative population densities N'v' obtained for each data set are presented in Table 2, together with the fitted rotational temperatures. The rotational temperature varies from 470 to 640 K, ans such temperatures of the neutral atmosphere are found in the height range from 130 to 150 km (CIRA, 1972), assuming an exospheric temperature of 1000 K. The uncertainties of each data set are probably mostly related to the averaging of intensity contributions from different heights with different temperatures. Only in cases when the aurora originates from a thin layer can the temperature determined be ascribed to a specific height. During this aurora no heating of the neutral atmosphere could be detected, and the rotational temperatures found will be representative of the centroid height of the emission volume1* . The relative population densities derived from the least squares fit are for the v' = 1 and v' = 2 vibrational levels of the b«I-s state in good agreement with the ratios predicted from the Franck-Condon factors. For the v' =3 level the fitted population is low by about 25% compared with the ratios obtained from the Franck-Condon factors. However, in the observed spectrum (cf. Fig. 5) the 02* 1NG (3,1) band is rather weak and badly overlapped by the N2* 1NG (0,3) band. Hence, we cannot rule out the possibility that the low v' = 3 population might be an artifact of our analysis. Anyway, the consistency of the low population densities of the v' = 3 level over a long period of time, warrants further investigations. At present we don't see any physical reason why the v' = 3 population should devi ate from the density predicted by the Franck-Condon fac tors . Radiation is expected to be the dominant, depopulating mechanism for the auroral 02' b^I-g state, as the quenching 56 height of this system is found to be close to 60 km1 . As long as the emission occurs from the E-region, there is no need for any indirect excitation mechanism to explain the observed vibrational population of the O2* 1NG system. Hence, the population densities of the 02* b*I-0 vibratio nal levels are in fair accordance with the current expecta tion that the Os• 1NG bands are produced by particle impact on neutral O2 molecules in their electronic and vibrational ground states. Acknowledgement. The authors are indebted to K. Måseide for critical reading of the manuscript. The Norwegian Research Council for Science and the Humanities is acknowledged for supporting this project through Grant No. D.20.14.234, and the Air Force Office of Scientific Research is acknowledged for support to the data processing of this project through Grant No. AFOSR-86-0327. 57 References 1. A. Vallance Jones, "Aurora". D. Reidel, Dordrecht (1974) . 2. M. Larsson, Astron. Astrophys. 128, 291 (1983). 3. E.U. Condon and G.H. Shortley, "The Theory of Atomic Spectra", Cambridge University Press 1964. 4. L. Veseth, "Symmetries and Rotational Line Intensities in Diatomic Molecules", Institute of Physics Report Series, University of Oslo, Oslo 1986. 5. D.M. Brink and G.R. Satchler, "Angular Momentum". Clarendon Press, Oxford 1968. 6. R.N. Zare, A.L. Schmeltekopf, W.J. Harrop, and D.L. Albritton, J. Mol. Spectrosc. 46_, 37 (1973).' 7. J.T. Hougen, "The Calculation of Rotational Energy Levels and Rotational Line Intensities in Diatomic Molecules". NBS monograph 115, 1970. 8. I. Kovåcs, "Rotational Structure in the Spectra of Diatomic Molecules". Adam Hilger, London 1969. 9. E.E. Whiting and R.W. Nicholls. Astrophys. J. Supple ment Series 27, 1 (1974). 10. P.H. Krupenie, J. Phys. Chem. Ref. Data, Vol. 1, No 2 (1972) . 11. G. Herzberg, "Spectra of Diatomic Molecules". Van Nostrand, New York 1965. 12. B.R. Martin, "Statistics for Physicists", Academic Press, London and New York (1971). 13. C.S Deehr, G.G. Sivjee, A. Egeland, K. Henriksen, P.E. sandholt, P. Sweeney, C. Duncan, and J. Gilmer, J. Geophys. Res. 85, 2185 (1980). 14. K. Henriksen, C.S. Deehr, and R.W. Smith, To appear in proceedings from the European Optical Conference, Cambridge, 1986. 38 J' S 4(2J' + 1) 3.5 32.00000 32.00000 7.5 64.00000 64.00000 11.5 96.00000 96.00000 15.5 128.00001 128.00000 19.5 160.00001 160.00000 23.5 192.00001 192.00000 27.5 224.00002 224.00000 31.5 256.00002 256.00000 35.5 288.00001 288.00000 Table 1. Test of the sum rule eq. (18) for the Honl-London 4—4 + •factors of the b E - a II transition in 0_ . The second g u 2 column gives the sum of the computed individual Hani —London factors. Note that for a homonuclear diatomic molecule like 0 with nuclear spin equal to zero, the appropriate theoretical sum rule is 4(2J' + 1) (cf. eq. (18)), since all rotational levels of parity -1 are missing for the 4 - 4 b Eg state, and all of parity +1 are missing for the a IT u state. Obs.period 1845 - 1900 1900 - 1915 1915 - 1930 1930 - 1945 1945 - 2000 v' N'. N'./N\ , N\ N',/N', N\ N',/N\ N\ N'./N', N', N',/N , v v v «i V v V ml V VV«| V VV»| V,0 Wl.O v v v »f 1 0.331(8) 1 0.121(4) 1 0.164(5) 1 0.127(5) 1 0.336 1 0,290(8) 1 2 0.137(7) 0.47(3) 0.162(7) 0.49(2) 0.056(4) 0.46(4 0.07^(5) 0.45(4) 0.054(5) 0.43(4) 0.164 0.49 3 0.038(4) 0.13(1) 0.053(4) 0.16(1) 0.016(2) 0.13(2 0.021(3) 0.13(2) 0.017(3) 0.13(2) 0.063 0.19 T (K) 535(16) 474(15) 534(18) 601(17) 641(19) Table 2. Fitted values of the population parameters N', (cf. eq. (22), unspecified units). Jg q ,„ represents the Franck-Condon factors for the transition 0 X £ v = 0 -0_ , b I , Mv'0 H 2 g 2 g v' (cf. fig. 1). The numbers in parenthesis give the uncertainty (standard deviation) in the last digit(s). Population densities and ratios are derived for every 15 min of the observational per iod toget her with the rotational temperature T and the standard deviation given in parenthesis. 60 Figure captions Fig. 1. Vibrational levels (not to scale) of O2 and O2* showing excitation transitions (dashed lines) and emission transitions (full lines) for O2• 1NG bands. Fig. 2. Rotational structure of the 02* 1NG (1,0) band at 150, 300, and 700 K. Each rotational line is displayed as a line with its relative intensity given by Eqs. (19-20). In each case the total intensity of the lines is normalized to the same value. Fig. 3. Synthetic spectra of the 02* 1NG (1,0) band calculated with a 4.0 A traingular instrument function. The intensity of each spectrum is normalized to the same value. Fig. 4. Spectrogram of the wavelength region from 4900 to 5700 A with 4.0 A resolution in nightside aurora. The 01(5577) green line is the dominant emission. When the scale is expanded by a factor of 40, molecular band emissions become distinct, see Fig. 5. Fig. 5. Observed and synthetic spectrum containing the O2* 1NG (1,0), (2,1), (2,0), and (3,1) bands, and in addition the N2* 1NG (0,3) band and the NI (5200) doublet are identified in the figure. In the least squares fit even the N2* 1NG (0,3) band and the NI (5200) doublet were included together with the O2 ' 1NG bands. The contribu tion from each spectral feature is separately plotted in the lower part of the figure, and the resulting synthetic spectrum is superposed on the observed one. The agreement between observed and synthetic spectrum is so close that the synthetic spectrum is barely visible. 61 / / b'z; / / / i/ / / / / / / / / '// /'/ a"n, XJE; F,J. 1 62 RGTATiONA^ L.NE OiSTRiBb'.ON OF THE Z- :>:'/, 3AKD AT 150 rt, 300 K, ANO 700 K 12 '<• lli •.I.'.-!,' I ,.i;(i;;)iil;;;;';,i::i!ji!;j!! • 5540 5580 562G . ri.-i'f 4 f J!^ 55OL 5540 55B0 562C 2r [I] ;=;.,• .:i.r Mili;' r-rl-l-T'l-fT- 5500 5540 5580 5620 WAVELENGTH [Å] F;V Z . 63 SYNTHETIC SPECTRUM AT 02* 1N(1.0) BAND WIT.- UK RESOLUTION AT 1b0 K, 300 K, AND 700 A 5.- :=0 K I 3 h/W, 2 lf- 1 !v 1 I ' ' ' | I ' ' I I I ' 'T ! I I l i I I ' i I I ' i ' i I • ' i ' i 5500 5540 5580 5620 5660 OT 2 2 i i | I I I l j T I l l | i < l l , . I l ! , I l i ,••••>,. 55C0 5540 5580 562C ciO WAVELENGTH [; Fn«. 3. 9-j AVERAGED SPECTROGRAM OF 01 5577 NIGHTSIDE AURORA 8- LONGYEARBYEN DEC. 16, 1984 1845-1900 7- 6- >- 5- I/) 4- 3- 2- + N2 1N(0,3) 1- NI 5200 Hgl 5461 4800 4900 5000 5100 5200 5300 5400 5500 5600 5700 5800 WAVELENGTH [A] V H. 140-1 + 02 1NG ~~i—I 120- (3.1)' (2,1) ' (2.0) 11.0) + N2 1NG(0.3) 100 NI 5200 4800 4900 5000 5100 5200 5300 5400 5500 5600 5700 5800 WAVELENGTH [A] F*V S 66 OPTICAL STUDIES FROM SPACE SHUTTLE - A PRELIMINARY REPORT OF MEASUREMENTS ON THE STS 41-6 MISSION E.J. Llewellyn, I.C. McDade Institute of Space and Atmospheric Studies University of Saskatchewan, Saskatoon, Canada D.J.U. Kendall, R.L. Gattinger Herzberg Institute for Astrophysics National Research Council of Canada, Ottawa, Canada 5.B. Mende, G.R. Suenson Lockheed Palo Alto Research Laboratory Palo Alto, California, U.S.A U.A. Eault, G.G. Shepherd, B.H. Solheim Centre for Experimental Space Science York University, Downsview, Canada L.L. Cogger Physics Department University of Calgary, Calgary, Canada and M. Garneau Canadian Astronaut Program Offics National Research Council of Canada, Ottawa, Canada ABSTRACT The spatial scanning camera used on the STS 41-G mission is briefly described and some initial results from the images are presented, It is concluded that the ram-glow on the 41-G mission was significantly less than that recorded on previous Shuttle flights. The airglow images suggest that it should be possible to derive atmospheric temperatures which can, in turn, be used to truth more sophisticated instrumentation planned far future Shuttle and other satellite flights. INTRODUCTION During the early flights of the Space Shuttle the presence of an intense glow on the ram side of the vertical stabilizer was first reported rBanks et al . , 19B3; fiende et 67 al.. 1993:>. This optical contamination had m Fact been inferred by both rocket i"'est igators ''Wallis and Anger, 19SS: Sresr et al . . 1983' and e\- perimenters associated with lc- altitude cbssr-atic^s on the »E spacecraft 'Terr. 13B3?, but it required the actual visual Dbservatics of the astronauts to recognize the full significance of the contamination. In an attempt to identify the source and magnitude cf this Shuttle glow Mende, and co-workers, undertook a series of experiments to measure the spectrum and intensity of the glow. These culminated in a measurement of the glow at 30A resolution which clearly revealed that, at this resolution, the glow is a continuum with a peak intensity of some 600R/A CSwenson et al., 19B53. Obviously such an intense glow could have a serious effect on measurements of atmospheric emissions from low Earth orbit. Measurements of this type are planned for WQriDII ^Wide Angle lichelson Coppler Imaging Interferometer), an instrument which is being built by SED Systems, in Saskatoon, for the National Research Council of Canada and is to fly on the Shuttle later in the decade. A modified versiO" of the instrument rWINDIi:> is to be part of the UAES pay load. These instruments uill provide data en temperature a-d uind, i- the e-itting region, by measuring the Doppler line iL'idth and t*-e absolute wavelength of the emission, As the actual determination of these parameters requires observations c~ a four-point algorithm, to deduce the i-terferogram envelope ^Shepherd et al . . 19B4'1, any variable continuum could seriously compromise the quality of the UA1DM results. Thus there uas an urgent need for 68 measurements cf the Shuttle glcu' at high rescljt :o-. and at C j.ave:e-sths te 1,-e .sei bb U "D!:. The inci-s:u-' of the Ca=-sdi3- -a-l=3i specialist :•- the :re* for the .mission STS -rl-2 effzrded = _-:^'je opportunity to make the necessary measurements. CPZIZ&L I'jSTP'J^E'-.'TeTIQN The optical system ccnsisted essentially of an intensified Nit.cn 35mm camera with a narrou' band interference filter positioned directly in front of the objective lens. Thus the image is, in principle, modulated by a set: of Fabry- Perct rings. However, the interference is blocked so as to transmit only a single order and the image is, therefore, simply spatially scanned in wavelength 'Shepherd et al . , 13S5"1. This is illustrated schematically in Figure 1. For the experiment c" STS 41-G, acronym QGLOW, the filters were selected to have a center wavelength Just longward Df the atmospheric emission feature, and a bandwidth of 4fl. This afforded the advantage of always including the desired emission feature, even if the filter shifted due to aging or a temperature change within the spacecraft cabin. For the 35mm image cf the camera the spatial scan corresponded to 30O, for the untilted filter, and 130 A when the filter was tilted sc that the unshifted wavelength was transmitted at one edge of the frame. Thus the system offered a significant improvement m resolution, from that previously flown, and, due to the effic:e-cy of the system, there is an improved throughput over the spectrometers previously used. SH'JTTLE-GLQU PESJLTS Dre cF the ~ajor difficulties of providing valid glow 69 intensity data, particularly where photographic recording is employed, is tc ensure that each frame is accurately calibrated. The Nikon camera, in its flight configuration, was carefully calibrated, both before and after the mission, using a secondary standard low brightness source. For the OGLOld experiment it was decided to use the limb airgicw as an in-flight calibration standard. This was achieved by orienting the Shuttle so that the vertical stabilizer was bisected by the airglow limb at 100 km, with the port side of the stabilizer placed in the ram direction. This orientation afforded the advantage that it provided a wide range of apparent airgicw intensities and the intensity of the ram glow could lie measured with some degree of confidence. The results indicated that the glow intensity was at the limit of detection, and that there were no obvious emission features at the wavelengths of the WflrDII filters. This reduced intensity for the glow, less than 150R/A at the peak of the previously measured spectrum, was unexpected and lead to a detailed examination of the atmospheric images obtained during the mission i:Kendall et al. , 1386? . The automatic calibration provided by the airglow confirmed the low intensity of the glow, and reinforces the suggestion that scrre type of in-flight calibration should be included whenever possible. Subsequent analysis of the glow measurements, from a number of Shuttle missions, has provided an interpretation of these low intensities in terms of the temperature 3f ^ne glowing surface '.Swenson et al . , 19B6). Th;s mterpretat:c nay also explain the different glcu i-tensities that have been recorded for different materials 70 as the emissivity, and absorbtivity, will necessarily affect the temperature of the surface. Thus the glow measurements on STS Hl-G have provided important new information on the factors that apparently influence the intensity of the ram g 1 oui. ATMOSPHERIC EMISSION IMAGES The images of the atmospheric emissions are extremely impressive and provide direct evidence of many optical aeroncmic phenomena. The images at 5300A indicate the presence of an extended, i" altitude, oxygen red line emission and the ''S-31 band of the OH Hemel emission near SO.'"-, Figure 2. The variation of the OH emission across the image is due to both spatial homogeneities m the emission a"d the spatial scanning in wavelength; for a molecular band this latter contains information on the atmospheric temperat-re. For the oxygen atmospheric band, at "76B0A, was, image mterpretaticn/complicated by the optically thick -ature of the emission. This precludes a simple analysis of the image in order to derive a spatial map of temperature. To obtain a true limb view the image must be corrected for the slant path absorption. In their original analysis of the atmospheric ba"d Wallace and Hunten f!19BB^ used a band equivalent Todel but recognized that a more detailed model -i~ht be needed for certain problems. The limb observation, at this iB-'elength. is certainly such a case and we have used a detailed band absorption. We have considered both a line equivalent width technique and a full Doppler emission line- -cppler absorption 1;ne technique. For tangent ray heights below "'Okm there is little difference between the results of 71 the two methods and the economy of computation afforded by the equivalent uidth technique would recommend its adcption. This image, Figure 3, clearly indicates the sharp nature of the upper boundary to the airglow emission, this is in line with the extensive model work of (IcDade et al . C19B6). The derivation of the rotational temperature cf the emission is somewhat complicated but we believe that the ring pattern can be analyzed tc gi^e temperature. This technique has the advantage that it will oFFer the opportunity For simple truthing cf the WAflDII technique. For some of the measurements a Fabry-Perat etalon was used i" series with the interference filter to provide high resolution, ;o, atmospheric images. In this case the individual lines m the oxygen atmospheric band are resolved and so give a deconvolved spectrum that is more appropriate to temperature analysis. This work is still in its prelimi~ary stages, although WE are optimistic cf detecting the airgicw SPISSIC due tc isotcpic oxygen molecules. This may also have some application tc temperature probing oF the atmosphere down to very lew altitudes. CONCLUSION While the ST5 41-G mission failed to detect a strong Shuttle ram gicu it did offer a unique opportunity to obtain images cf the oxygen airglou which can be used to prepare Far future satellite experiments that will use imaging techniques. These images are frequently diFFerent From that initially excected and so provide an excellent data base For instrument Tn. 72 Acknowledgement. This work has been supported by grants From the Natural Sciences and Engineering Research Council of Canada, the National Research Council of Canada and Lockheed independent research Funds. Ule are deeply indebted to the many people who have made the DGLGlii experiment a reality. REFERENCES Banks, P.tl., P.R. Williamson and W.J. Raitt, Space Shuttle j!au observations. Geophys, Res. Letts., 10, 11B, 19B3. Greer, R.G.H., D.P. (lurtagh, G. Witt and J. Stegman, Photometric Observations DF local Rocket-Atmosphere interactions. ESA SP-1B3, 341, 19B3. Kendall, D.J.W., R.L. Gattinger, E.J. Llewellyn, I.C. HcDade, S.B. Hende and G.R. Swenson, Observations of Drbiter Glow on Mission STS 41-G. Planet. Space Sci., paper submitted 19BB. PIcDade, I.C.. D.P. tlurtagh, R.G.H. Greer, P.H.G. Dickinson, G. Witt, J. Stegman, E.J. Llewellyn, L. Thomas and D.B. Jenkins, Quenching parameters For the proposed precursors cF 0;.(b»Za*) and 0C'S3 in the terrestrial nightgiCi.. Planet. Space Sci . , paper in press 19BB. Hende, S.B., O.K. Garroitt and P.n. Banks, Observations of Optical E^!5sic~s on STS-4. GeDphys. Res. Letts., 10, 122, 19B3. Shepherd, G.G.. C.W. Lake, J.R. niller and L.L. Cogger, A spatial scanning technique For the Fabry-Perot spectrometer. App. Optics, 4, 267, 1965. Shepherd, G.G., W.A. Gault, R.A. Koehler, J.C. ricConnell, K.V. Paulso", E.J. Llewellyn, CD. Anger, L.L. Cogger, J.ltf. Hewlett, D.R. Mocrcroft and R.L. Gattinger, Geophys. Res. Letts.. 11. 1003, 1984. Swenson, G.R., S.B. Hende an_ <~- Clifton, Ram Uehicle Glou. Spectrum; Implication of NO-.- ambmation Continuum. Secphys. Pes. Letts., 12, S'7, 19Bb. Sue°scn, G.R., S.B. lende and E.J. Llewellyn, Shuttle glox - the possible efFect OF surface temperature, Nature CLcndo"v, paper submitted 19B6. Torr. 1., Optical Emissions induced by Spacecraft- Atmcsphere interactions. Geophys. Pes. Letts., 10, 114, 19B3. Wallace, L. and D.rt. Hunten, Dayglow of the oxygen A-band. J. Geophys. Pes., "3, 4913. 19BB. Wallis, 0.2. a"d C.D. Anger, High-altitude observations oF a luJn;r-2.5 ua'r s behmd tun Black Brant II rackets. Can. J. Phys., 46, 2753 1969. Objective Lens Coupling Lens IWFWWfWP Image Intensifier Film Plane Interference Filter Fis_rp 1. The cpt:cal systen for ths QGLO'JJ camera. Figure 3. The airglou image obtained throuqh th' 6300A filter. ia a 76 Thermospheric Structure and Auroral Modeling M. H. Rees Department of Physics and Geophysical Institute University of Alaska-Fairbanks Fairbanks, Alaska, USA SUMMARY Thermospheric structure refers to the distribution in altitude of neutral atoms and molecules, of positive and negative ions, and of the neutra1, ion and electron temperatures. Solar ultraviolet radiation Is the principal source which governs the global composition and temperature characteristics of the thermosphere. At high latitudes, however, energetic particle bombardment and convective electric fields, both associated with aurora, provide additional input Into the atmosphere and ionosphere that substantially influences thermospheric structure. In this talk, the observational evidence that demonstrates aurorally induced perturbations is reviewed and the physical processes that account for the variety of measurements acquired by rocket and satellite-borne Instruments, as well as by ground-based remote sensing, are discussed in some detail. Turning first to energetic auroral particles, it is necessary to compute their penetration into the atmosphere and their energy degradation. Energy loss goes into production of ion pairs, dissociation of molecules, electronic, vibrational and rotational excitation, and heating of the plasma and neutral gas. A multitude of chemical-ionic reactions accounts for a highly variable ion composition and production of minor neutral species, such as atomic 77 nitrogen and nitric oxide, A small fraction of the energy carried by auroral energetic particles goes into heating the ambient electron gas. raising the electron temperature. Almost half of the energy goes into heating the neutral gas, directly and indirectly, accounting for an increased neutral temperature in regions under substantial auroral particle bombardment. Convective electric fields force ions to move from one region of the polar cap to another. In the process, the neutrals may be dragged along, resulting in perturbations of both ion and neutral composition. Frictional interaction between ions and neutrals results in heating of both species and a possible enhancement in the ion and neutral temperatures. The continuity equations (or conservation of mass) govern the composition of the ion and neutraL gas while the energy equations are used to derive the temperatures. The aurora Is a temporally (and spatially) varying phenomenon and the time dependent equations must be used. Moreover, the continuity equations are all coupled to each other as well as to the energy equations. To a first approximation, thermospheric structure under auroral influences may be formulated as a one-dimensional problem, along a magnetic line of force or along the vertical. This simplification has made it possible to study the auroral consequences in considerable detail. Influences of horizontal dynamics (winds and neutral temperature) must be introduced in an ad hoc manner, and a self-consistent solution is not possible in one dimension. A three-dimensional time dependent treatment of the coupled thermosphere and ionosphere is a goal of current research. The spatial inhomogene!ty of the aurora and the high temporal variability combine to add considerable difficulty to the solution of a self-consistent model. 78 THE EFFECT OF LIGHT SCATTERING AND DIFFUSE REFLECTION ON ATMOSPHERIC SPECTRAL MEASUREMENTS Knut Stamnes Auroral Observatory, University of Tromsø N-9001 Tromsø, Norway and Georg Witt Meteorological Institute, University of Stockholm Stockholm, Sweden ABSTRACT In optical remote sensing of airglow and auroral emissions one must account for the effect of atmospheric scattering and absorption between the emission region and the observing site. Reflections from a snow-covered ground may also contribute significantly to the measured radiation. We use a comprehensive radiation model to quantify these effects for realistic atmospheric models. In addition to Rayleigh- scattering we include the influence of ozone, and cirrus clouds. Sky-brightness computations show that atmospheric scattering and ground-reflections significantly modify the emitted radiation and must be corrected for in quantitative studies of the remotely sensed emissions. 79 1. INTRODUCTION Optical remote sensing is one of the chief means available to monitor atmospheric and ionospheric behaviour. In fact, ground-based optical observations of airqlow and aurorae have a long tradition and have yielded much valuable information about chemical and physical processes occurring in the upper atmosphere. During the last two decades similar information obtained by rocket- and satellite-borne optical instruments has become increasingly important. Thus, in addition to a well-established network of ground-based stations, there are now many optical "observatoriess" in space carrying sophisticated instruments of steadi ly improving capabilities. A vexing concern in optical remote sensi ng has always been the realization that part of the measured signal may not be due to the true atmospheric emission that we want to monitor, but may rather be caused by scattering in the atmosphere (cf. e.g. Ashburn, 1954). The purpose of the present paper is to estimate the influence of atmospheric scattering on spectral emissions measured from the ground and from space. 2. FOHMULATION OF THE PROBLEM When dealing with atmospheric emissions it is convenient to distinguish between the "scalar problem" in which one neglects the state of polarization of the radiation and the "vector problem" in which the complete Stokes' vector is considered. In "scalar problems" attention is restricted to the intensity of the radiation which is the first component of the Stokes' vector. In this context the intensity refers to -2 -1 -1 the radiance measured in units of W m sr nm To obtain information about the polarization properties of the radiation all four components must be accounted for. Below we mention briefly some current problems that are addressed by this study. In connection with satellite and rocket observations of atmospheric emissions it may frequently be required to correct for possible contamination by backscattered sunlight. Satellite imaging experiments of the aurora and airglow are 60 specific examples of current observations that must be corrected for such contamination. Auroral emissions contain information about the spectral hardness and the energy content of the exciting particles (electrons and protons). Measurements of spectral intensity ratios can be used to infer the energy characteristics of the auroral precipitation using instruments on the ground or in space. The question then arises: to what extent could scattering of radiation from bright features outside the field of view of the instrument contaminate these ratios? A similar problem occurs in connection with airglow studies. Ground-based observations must be corrected for attenuation bv the intervening atmosphere and measurements from space foi additional radiation due to atmospheric backscattering and reflections from the ground. It is then important to assess the extent to which these spectrally varying airglow emission features are affected by a Rayleigh-scattering atmosphere containing clouds and aerosol. Airglow studies of molecular oxygen emissions are examples of studies requiring such assessments. To resolve the uncertainties raised in the preceding paragraphs it is sufficient to consider the "scalar problem" since we are only interested in the intensity of the radiation. In other situations polarization measurements may yield valuable additional information. This has certainly been the case for daytime observations of the polar mesospheric and noctilucent clouds. Thus, although it would be desirable to include polarization effects, we will limit our present study to investigate how a scattering atmosphere may contaminate intensity measurements made from ground and from space. Rayleigh scattering is a strong function of wavelength, -4 \ (the scattering cross section varies as A ), and thus becomes increasingly important for shorter wavelengths. Ozone is the chief atmospheric absorber in the near UV and visible part of the spectrum. The ozone molecule absorbs strongly in the Hartley band (200-300 nm) with a peak cross section of -17 2 about 10 cm at 260 nm. In the Huggins band (310-350 nm) -19 2 the ozone absorption cross section decreases from 10 cm at 81 310 nm to 10~ cm at 340 nm. Ozone also absorbs in the visible part of the spectrum between 450 nm and 850 nm (the Chappuis band) with a peak cross section of about 5 x -21 2 10 cm at 600 nm (cf. e.g. Brasseur and Solomon, 1984). Ozone is only one of several molecules that absorb atmospheric radiation, but it is by far the most important one in the near Uv and visible part of the spectrum. (Water vapour and molecular oxygen must be considered in the near frared part of the spectrum and several other molecules for longer wave lengths.) 3. COMPUTATIONAL PROCEDURE To estimate the effect of atmospheric scattering and absorption, and ground reflection on auroral and airglow intensity measurements we proceed as follows. We assume that the emission comes from an extended layer at the top of the atmosphere that emits radiation isotropically as illustrated in Fig. 1. Thus, in the absence of ground reflection and atmospheric scattering and absorption, an observer on the ground looking up, or in space looking down, would measure the same amount of radiation in all directions. For simplicity we assume that the ground is a Lambertian surface reflecting radiation isotropically. The intensity, T(T,U), of the radiation in a stratified atmosphere (i.e. we assume plane-parallel geometry) is described by the radiative transfer equation y dn^ui = I(TfU) _ yj_o / p(TrlJ,u')i,T,lJ')du' (i, where u = cosine of polar angle and t is the vertical optical depth, equal to the column density times the extinction (absorption + scattering) cross section. ou(i) is the single scattering albedo defined as the ratio between the scattering and the extinction cross sections, whereas P(T,U,U') describes the scattering in angle, and is a normalized scattering cross section commonly referred to as the phase function. Thus, UJ(T) varies between zero (no scattering) and unity (no absorption). In realistic atmospheres - consisting of a mixture of radiatively active gases and aerosols having a 82 DETECTOR - Y DETECTOR ON GROUND wto— PARTIALLY REFLECTING GROUND Figure 1. Schematic illustration of the geometry. non-constant mixing ratio - the single scattering albedo and the phase function depend on altitude or optical depth. To resolve this vertical structure we divide the atmosphere into an appropriate number of layers as indicated in Fig. 1. The integral term in Eq. (1) is due to multiple scattering in the atmosphere. In the absence of scattering (to = 0) Eq. (1) can be solved immediately to yield (u > 0) -T/U I(T,-U> = I (0,-u)e (2) where I (0,-y) is the radiation incident at the top of the atmosphere and the exponential factor describes the attenuation due to atmospheric absorption. To investigate the effect of both atmospheric scattering and absorption we solve Eq. (1) including multiple scattering by air molecules {Rayleigh scattering) and absorption by ozone, assuming unit intensity to be incident uniformly at the top of the atmosphere (I (0,-u) » 1). The underlying ground is assumed to reflect radiation uniformly with albedo A . 83 A variety of methods exist for solving linear transport equations like Eq. (1) (cf. Stamnes, 1986, for a recent review of computational techniques). We use the discrete-ordinate method (cf. Stamnes and Swanson, 1981; Stamnes and Dale, 1981; and Stamnes and Conklin, 1984) to solve Eq. (1) for a realistic, multi-layered atmosphere. We divide the atmosphere into N adjacent, homogeneous layers allowing the scattering and absorption coefficients to vary from layer to layer. The intensity transmitted through the atmosphere, I(T ,-u), and the intensity emerging from the top of the atmosphere, 1(0,+u)f can then be expressed as KV-u> N r n , -[k, T (T - )/u] = \[.i w(-u) e r r+ r = 1 L] = -n J J l k T +( T )/lJ (3) _e -C jr r-1 V r-1 h] 1(0,+ u) N r n , - (4) where 2n is the number of discrete ordinates utilized. The outer sum (index r) in Eqs. (3) and (4) is over atmospheric layers, while the inner sum (index j, i = 0 not included) is over discrete ordinate eigensolutions. The kjr and g,r(u) are discrete ordinate eigenvalues and eigenvectors, and the L- are 2n x N constants of integration determined by requiring the radiation to satisfy boundary conditions (as described above) and continuity conditions at layer interfaces (cf. Stamnes and Conklin, 1984; Stamnes, 1986; and Stamnes et al., 1986, for details). Equations (3) and (4) represent the complete solution to the mathematical problem we have formulated including all orders of multiple scattering in the atmosphere as well as reflection from the ground including all orders of multiple retro-reflections between the ground and the atmosphere. 84 4. RESULTS Below we present some representative results of our computations for emission in the near-ultraviolet and visible part of the spectrum. We use the term "transmission" to refer to the (percentage) fraction of the emitted radiation field penetrating to the ground (modified by atmospheric absorption and scattering as well as ground reflection). Similarly, we use the term "backscatter enhancement" to describe the (percentage) increase in radiation escaping to space due to the presence of the atmosphere and the underlying Lambertian reflecting ground surface. Thus, in the absence of a LOOKING UP LOOKING DOWN ZENITH 120° HORIZON HORIZON 60° NADIR 1 I ' ' 1 i ©0 1 1 1 1 1 1 1 i i ~ 100 1 1 1 1 1 1 i . a) . b) t— 80 (•/. ) - £ 80 - - - CEM i § 60 - § 60 EN H ISSI ' tr 40 l/?l *° - - UJ - H- z _ 1- < ? in -—-._. 20 >- 20 --*. a - »-,—,— • ' 1 1 1 i 1 j 1— BA C ' • ' • • I— i -1.0 -0.5 0 5 10 COSINE (POLAR ANGLE) COSINE (POLAR ANGLE) Figure 2. Transmission a) and backscatter enhancement b) at 310 nm (see text for definition). In this and the followinq figures zenith refers to an observer looking up from the ground and nadir to an observer looking down from space. Ground albedo » 0.8 Ground albedo • 0.05. 85 scattering/absorbing atmosphere and a reflecting ground the transmission would be 100% and the backscatter enhancement 0%. In Fig. 2a,b we show transmission and backscatter enhancement at 310 nm for surface albedo A = 0.05 (dashed line) and A = 0.8 (solid line). At this wavelength absorption by ozone is much stronger than molecular (Rayleigh) scattering resulting in strong attenuation (small transmission) and little backscatter enhancement. The ground albedo plays a minor role since only a small fraction of the emitted radiation reaches the ground. At 350 nm (Fig. 3) the effect of ozone absorption is weak compared to the influence of molecular scattering. The transmission varies between 70% and 30% depending on observing angle for A = 0.05, but increases to more than 90% for A = 0.8 (observing in zenith). The backscatter enhancement exhibits a very similar behaviour in this case. LOOKING UP LOOKING DOWN ZENITH 120° HORIZON HORIZON 60° NADIR f 1 J \ nJU 1 1 r—r 1 1 1 1 1— ~ 100 i i i i 1 i I 1 >- 80 . ^^^ 5 80 " - "*""**"""'--.. wz *"-•*. u _ z 60 ~***"s. < 60 "*~^ - g i ^ N in >. " Figure 3. Transmission a) and backscatter enhancement b) at 350 nm. Ground albedo = 0.8 Ground albedo = 0.05. 86 ln the next two figures we simulate the effect of the earth-atmosphere system on emissions close to the auroral green and red lines. At 555 nm the Rayleigh scattering optical depth is about 0.1 and ozone absorption is relatively weak (Chappuis band). Consequently, as shown in Fig. 4 the transmission is little influenced by atmospheric effects and depends only weakly on surface albedo because the atmosphere is almost transparent except for observing angles close to the horizon for which the effective optical thickness becomes larger. The backscatter enhancement on the other hand is only about 10% (for nadir observations from space) for A = 0.05, but increases to more than 70% for A = 0.8 as one would g expect for small atmospheric optical thicknesses. The results for the auroral red line shown in Fig. 5 are very similar to those for X = 555 nm and need no separate discussion. LOOKING UP LOOKING DOWN ZENITH 120° HORIZON HORIZON 60° NADIR >1 \ ' ' \ i too i —i r i i 1 1 1 1— roo i • 1 1 1 1 1 1 1 80 . o) ~~~^r^^v z 80 . b) s UJ S* \ 5 . ^~* \ \ Uf § 60 \ u z TE R 4 *- Figure 4. Transmission a) and backscatter enhancement b) at 555 nm. Ground albedo - 0.8 Ground albedo - 0.05. 87 LOOKING DOWN HORIZON HORIZON 60° NADIR .J i \ •: 5 80 b) - f ENHANC E cc 40 UJ < \ or 20- i- 3 20 -I i 1 1 1 L_ < > ' ' • -1.0 -0.5 CO 1 0.5 1.0 COSINE (POLAR ANGLE) COSINE (POLAR ANGLE) Figure 5. Transmission a) and backscatter enhancement b) at 630 nm. Ground albedo = 0.8 Ground albedo = 0.05. LOOKING DOWN HORIZON HORIZON 60° NADIR 1 ' 1 1 -j 100 t- uj 80 - - U2J u 6) i 60 - X UzJ - cc 40 UJ - < tt 20 - o* • I I I J I L_J 1 I C Figure 6. Transmission a) and backscatter enhancement b) at 630 nm for qround albedo = 0.3. Cloudy atmosphere (see text) _ _ Clear atmosphere. 88 Whereas the preceding results refer to a clear, cloudless atmosphere, the effect of a moderately thick cloud (optical thickness a 1) is illustrated in Fig. 6. The cloud is assumed to be non-absorbing and may simulate the effect of cirrus clouds on the radiation field. The main effect of the cloud is to decrease transmission and increase backscatter enhancement. 5. SUMMARY The findings reported here may be summarized as follows: 1) At 310 nm strong absorption by ozone leads to substantial attenuation of the transmitted radiation. The radiation emitted to space suffers little contamination by back- scattered light since most of the radiation emitted downwards is absorbed rather than scattered by the atmosphere at this wavelength. 2) At 350 nm absorption by ozone is weak compared to Rayleigh scattering. This situation leads to strong backscatter enhancement and weak attenuation, but strong dependence on surface albedo. 3) In the visible region (including the auroral green and red lines) the absorption by ozone is moderate (Chappuis bands) and the Rayleigh scattering is weak compared to that in the near UV. This implies little atmospheric backscatter enhancement and small modification of the transmission, but strong backscatter dependence on surface albedo. We conclude that atmospheric absorption and scattering as well as ground reflections siqnificantly modify the emitted airglow and auroral radiation and must be corrected for in quantitative studies of remotely sensed optical emissions. Except for difficult geometries associated with broken clouds and spatial and temporal changes in the aurora, it is now feasible to make such corrections with relative ease utilizing 89 a computer code based on the discrete ordinate method briefly outlined above and described in detail by Stamnes et al. (1986). In addition to clouds the radiation field is modified by the presence of aerosols in the atmosphere. Thus, in order to make meaningful corrections it is necessary to measure also the atmospheric turbidity simultaneously with the airglow/ auroral observations. An estimate of atmospheric turbidity can conveniently be obtained by measuring stellar extinction in two colours. Alternatively, it is possible to monitor intensity variations of the ubiquitous mercury line at 546.1 nm resulting from tropospheric scattering of light from illumination systems. The latter possibility is promising but needs further investigation. Finally, we should mention that for studies involving polarization measurements (e.g. studies of polar mesospheric and noctilucent clouds), the theory presented here must be extended to include all four components of the Stokes vector. REFERENCES Ashburn, E.V., The effect of Rayleigh scattering and ground reflection upon the determination of the height of the night airglow, J. Atmos. Terr. Phys., 5, 83-91, 1954. Brasseur, G. and S. Solomon, Aeronomy of the Middle Atmosphere, D. Reidel Publishing Company, 1984. Stamnes, K., The theory of multiple scattering of radiation in plane parallel atmospheres. Rev. Geophys., in press, 1986. Stamnes, K. and H. Dale, A new look at the discrete ordinate method for radiative transfer calculations in anisotropi- cally scattering atmospheres. II. Intensity computations, J. Atmos. Sci., 28. 2696-2706, 1981. Stamnes, K. and R.A. Swanson, A new look at the discrete ordinate method for radiative transfer calculations in anisotropically scattering atmospheres, J. Atmos. Sci., 38, 387-399, 1981. 90 Stamnes, K. and P. Conklin, A new multi-layer discrete ordinate approach to radiative transfer in vertically inhomogeneous atmospheres, J. Quant. Spcctrosc. Rad lat. Transfer, 21» 273-282, 1984. Stamnes, K., W.J. Wiscombe, S.-C. Tsay, and R. Jayaweera, On the numerical implementation of the discrete ordinate method for radiative transfer in vertically inhomogeneous atmospheres, to appear as NASA tech. report, 1986. 91 COORDINATED ROCKET MEASUREMENTS OF AURORAL X-RAYS AMD OPTICAL EMISSIONS J. Stadsnes and K.R. Topphol Department of Physics, University of Bergen, Norway and K. Miseide Department of Physics, University of Oslo, Norway ABSTRACT Coordinated aeasureaents of bremsstrahlung X-rays and the optical auroral emissions N '(427,8 nm) and HB(486.1 nm) have been nade froa a high altitude sounding rocket. The rocket was flown across an extended aurora. By using narrow field detectors looking backwards at slant angles to the rocket axis, fairly good aappings of the auroral source regions could be obtained over wide areas as the rocket was spinning and coning. The eaphasis in this paper is on the comparison of these source regions which reflects the spatial distribution of the energetic electron and proton precipitation, it was found that the distribution of the low energy X-rays (> 2.9 keV) was fairly siailar to the distribution of the auroral N*(427,8 na) eaission. At higher X-ray energies the difference was greater. These measurements are the first we know about of this type. INTRODUCTION The energetic particle precipitation into the upper atmosphere cannot be completely monitored by optical auroral imaging alone, as the optical emissions do not reflect the particle fluxes and energies unambiguously. Particle spectrometers have thus been extensively flown on rockets and satellites to measure the precipitating particles directly in situ. But only inforaation about the fluxes in the immediate vicinity of the spacecraft can be obtained in that way. Breasstrahlung X-rays, however, are also eaitted when energetic electrons are stopped in the auroral 92 regions. By X-ray Mppings perforaed above the absorbing part of the ataosphere coabined with spectral deconvolution techniques, it is thus >ssible to deduce the intensities and energy spectra of the energetic electron precipitation over wide spatial areas. For aany years balloon-borne X-ray aeasureaents have been used to infer the spatial and teaporal distribution of the energetic (> 20 keV) electron precipitation (Barcus and Rosenberg 1966, Kreaser et al. 1982). Froa such aeasureaents one can aonitor the teaporal developaent of the precipitation over long tine intervals (hours to days). The field of view of a balloon borne X-ray detector is, however, fairly snail (<200 ka across). Even though Coapton scattering prevents high resolution aeasureaents, soae results on the small scale spatial distribution have been obtained froa balloons (Bjordal et al. 1976, Hauk and Parks 1981). Colliaated X-ray detectors flown on rockets have been successfully used to aap the electron precipitation (Vij et al. 1975, Aarsnes et al. 1976, Goldberg et al. 1982). By such aeasureaents fron a high altitude rocket looking down on the X-ray production layer, one has negligible scattering effects, and aeasurenents can usually be done down to energies of a few keV without significant ataospheric absorption. Auroral photoaeters have previously been flown on several rockets primarily to aeasure the height profile of the auroral eaissions (Håseide 1967, Jespersen et al. 1969, S^raas et al. 1974). In the present experiment we intended to Measure both electron and proton generated optical eaissions over an aurora with great latitudinal extent, and to aake a detailed coaparison between such eaissions and the breasstrahlung X-rays generated in the saae area. THE EXPERIMENT Instrumentation The aeasureaents reported below were Bade as a part of a more extensive rocket experiment to study an auroral poleward leap. This type of event is a poleward expansion of the oval near the start of the recovery phase of a aagnetospheric substora, and seeas to be closely 93 related to the tailward aoveaent of the reconnection line foraed during the substora (Pytte et al. 197S). A fairly large rocket (Terrier Haleaute) had to be used to obtain a great latitudinal range above the aurora. The payload included 10 separate experiaents, and aaong those were particle spectroaeters which could aeasure both electrons and ions froa 10 eV on and up to 300 keV. Two X-ray detectors were flown. They were colliaated to 12 degrees full width half aaxiaua field of view (FOV) and were pointing at 115 and 135 degrees, respectively, to the rocket axis. In this way they would aainly be looking backwards and downwards during the flight, and they would scan over a fairly wide area of the X-ray production layer froa the upper part of the trajectory when the rocket was spinning and coning (see Figure 3). The detectors were equipped with brooa maonets and baffle systeas to prevent direct entering of electrons into the detecting eleaents, and X-rays down to 2,9 keV could be measured. Such X-rays are eaitted as breasstrahlung when energetic auroral electrons are stopped at altitudes of about 110 kiloaeters. Two photoaeters for optical auroral measurements were aounted in parallel with the X-ray detector looking downward at 135 degrees to the rocket axis. The field of view was 9 degrees circular full angle, and thus comparable with the FOV of the X-ray detectors. The photoaeters were measuring the well known auroral emissions H2*(427,8 na) and HB(486.1 na). In a dark night-tine atmosphere the N2* emission is mainly produced by simultaneous ionization and excitation of N molecules by primary auroral electrons in the keV range' and by their secondaries of energies above 19 eV. Its intensity is thus assuaed to be proportional to the energy flux of the primary electrons, and the maximum photon emission rate coaes fron an altitude of 110-120 kiloaeters (see e.g. Vallance Jones 1974). The auroral HB emission is aainly produced when energetic protons in the 10-100 keV range penetrate into the atmosphere and are being neutralized to hydrogen atoms by charge exchange with aabient atoms and molecules. This process also occurs aainly at the 110-120 ka altitude range (Vallance Jones 1974, Søraas et al. 1974). Supporting measurements froa the ground were made by all sky film- and TV-cameras and by a 4-channel auroral photometer located at the rocket range and scanning back and forth along the rocket trajectory plane. 94 Flight conditions The rocket was flown froa Andøya Rocket Range (ARR), Norway (geogr. 69°17'N, 16°orE, L=6.3), at 210839 UT, on 11. November 19S3, at an aziauth of 338 , i.e. transverse to the Magnetic L-contours. An apogee of 454 ka was obtained, and the horizontal range was S70 ka above the 110 ka altitude level, for simplicity assumed to be the altitude of the production layer of the X-rays and the optical aurora . The flight occurred into the later part of a fairly big auroral substora as shown in Figure 1, where the time of the rocket flight is aarked by the hatched area. From the Z-trace of the magnetogram we see that a poleward aotion of the electrojat started at about 2050 0T and was still in progress at launch. It looks, however, as if the 'poleward leap" had already occurred at that time and that the substora was near to its recovery phase during the flight. Clear sky provided good conditions for optical observations froa the launch site during the flight, and the aurora appeared mainly as a fairly stable patchy glow covering aost of the sky. Figure 2 shows an all sky TV picture taken from the launch site in the middle of the flight. Note the decrease in auroral intensity towards the North, and the alaost lack of aurora in North-East. Observing method The geometry for the optical and the X-ray measurements from the upper part of the rocket trajectory is illustrated in Figure 3. Vertical lines froa the rocket positions and down to the photon production layer, assumed to be ac 110 ka altitude, are drawn for every 100 s. The rocket was spinning with a 1.8 s period and each detector's F0V was aoving along a path in the production layer, assumed to be at h=110 km both for the x-rays and for the optical aurora. In this figure are illustrated typical scanning paths for the central axis of the F0V of the detectors. Near apogee the size of the F0V in the production layer is approximately 70 km across looking straight down, and larger for slant directions. The F0V of the photometers corresponds to some 80 percent of this width. As the rocket had a slow coning notion (period 122 s), the path seen by the detectors changed gradually from one spin to the next. By 95 combining the< effects of the spin and the coning Motions, it is possible to construct a two dimensional image of the sources of the X-rays and the optical emissions. The spatial resolution that could be obtained in this type of an auroral mapping obviously varies during the flight due to the varying geometry. The temporal resolution is mainly limited by the rocket spinning and coning rates combined with the sampling rates allowed by the telemetry. In Figure 4 are shown the magnetic local tines and the L-shells covered by these observations. The large circle has 1000 km radius from the nadir point of the rocket at apogee. The electron precipitation within this circle can be remotely mapped from the rocket with good observing geometry by observing the bremsstrahlung X-rays and the optical emissions. Also shown is the FOV of the all sky TV camera at ARR. Notice the limitations in the FOV in the southern sky caused by mountains near ARR. We will give a short explanation of the method used in producing the maps of the spatial distributions of the X-rays and the aurora presented in Figures 5 to 16. Based on the rocket attitude we find the coordinates of the point where the central axis of the detector's FOV hits the photon production layer. The curved production layer is then folded up into a plane being tangent to the production layer at the rockets nadir point. A time resolution of 18 samples per s is used. Each gyration period of the rocket was 1.6 s so we get 32 points in the figure for each gyration. Colored line segments are drawn between these points, with a color code corresponding to a logarithmic intensity scale. This scale constitutes six colors and is chosen so as to utilize the whole color spectrum for the actual intensities of our observations. Blue is low and red is high intensities. For the X-ray figures the intensity is given in counts per sampling interval, and for the photometer figures the intensity is given in Rayleighs uncorrected for slant angle view through the emission layer. In the figures geographic North is up and East is to the right. ARR is in the center. The figures cover a range of 1500 km to each side of ARR. No corrections are made either for the distance from the rocket to the source region or for slant viewing direction of the source region. These corrections are dependent on the source geometry which we don't know a priori. 96 The same aethod for generating pictures of the source distribution is used both for the X-ray and the photometer data. Since the fields of view of the different detectors are approximately of the saae size, the pictures of the X-ray sources and the pictures of the optical sources can be directly coapared. RESULTS In Figure 5 is shown the source distribution of the greater than 2.9 kev X-rays as measured by the detector at 135 degrees to the rocket axis. The figure covers the flight time interval 314 to 436 s, corresponding to one coning period when the rocket was near apogee. The saae tiae interval is also used in the following figures. Figure 5 shows a region of low intensities to the North, as can also be seen on the all-sky TV picture shown in Figure 4. In South-East there is a region with high X-ray intensities . This region is nearly hidden behind the aountains at ARR and is barely visible on the TV record. Froa about 200 s flight tiae, the rocket is above the blue region of weak electron precipitation. Measurements aade by the electron spectrometers onboard (flown by the Max Planck Institute fiir Aerononie, FRG, and by the University of Bergen) are consistent with the X-ray picture. Significant fluxes of electrons above a few keV were recorded until about 200 s flight tiae, and only very weak fluxes later on, except for a spike near the end of the flight at around 550 s. In Figure 6 we show the aeasureaents of greater than 4.4 keV X-rays as aade by the 115 degrees detector. Because of its orientation with respect to the rocket axis, this detector is scanning over a wider area in the source layer. This figure also shows clearly the lack of electron precipitation in North to North-East. Figure 7 show? the spatial distribution at higher X-ray energies, greater than 12 keV, for the 135 degrees detector. (The intensity scale is the saae as for the greater than 2.9 keV figure). He recognize the saae spatial structures as seen at the lower energies. The ratio between the intensity of the greater than 2.9 keV to the greater than 12 keV X-rays is presented in Figure 8. Red aeans soft spectrum and blue hard spectrin». It seems like the spectrum is harder to the East than to the West. We have a hard spectrum both in NE where the 97 electron precipitation flux is very weak, and in SE where the precipitation is relatively strong. Figure 9 shows the H*(427.8 na) intensity. This emission is generated mainly by electron precipitation over a wide energy range fro» 19 eV on and upwards to several keV, i.e. including energies lower than those generating the X-rays we measure. We see that the similarity between the spatial distribution of the N emissions (hereafter called electron aurura) and the X-ray emissions is striking although the optical emission may also be contaminated jy ground albedo and spread light from the lower part of the atmosphere. The spatial distribution of the 486.1 no H-beta line intensity is shown in Figure 10. This line is mainly generated by proton precipitation. We see also here great similarity with the previous picture of the electron aurora, although in the North we find sone differences. There we see a region with low intensity, but the intensity is increasing again further North. This was not seen either in the X-rays or in the electron aurori. (Varying Doppler profile with looking direction and the van Rhijn effect may however cause problems). The ratio between the intensity of the electron and the proton aurora is shown in Figure 11. We see an East-West oriented band, where the electron intensity - is relatively high with respect to the proton intensity, whereas we have relatively more protons North and South of this band. Also note the blue region in NE indicating proton precipitation in accordance with the picture of the proton aurora. In Figure 12 we show the ratio between the electron aurora intensity and the greater than 12 keV X-rays. We also here see a harder electron spectrum to the East than to the West. The difference is more striking in this figure than it was in Figure 8 displaying the ratio between the two X-ray energies. This indicates that there is a lot of low energy electrons, less than approx. 3 keV, precipitating to the West and contributing to the 427.8 nm emissions, and more energetic electrons contributing to the X-ray fluxes to the East of ARR. Finally, we show four figures (No 13 to 16) where data from nearly the whole flight are used to look for time variations in the electron precipitation. Figure 13 shows the greater than 2.9 keV X-rays as measured by the 135 degrees detector. We display the source mappings generated from the four coning periods »hen the rocket was well above the production layer at 98 110 ka height. In the upper left of the figure the distribution for the flight tiae interval 70 to 192 s is shown. Upper light corresponds to 192 to 314 s, lower left 314 to 436 a and lower right 436 to 558 s. We see no striking tiae variations in the spatial distribution of the X-ray sources during the eight ainutes covered by these observations. In Figure 14 we display the saae kind of Mappings for the >4.4 keV X-rays as seen by the 115 degrees detectoi. This figure confiras the results froa the 135 degrees detecto. in the previous figure. In Figure 15 we have the intensity of the N '427.8 na eaission for the saae tiae intervals as the X-ray figures. Here also the tiae variations are weak. In the last period, however, a localized region of higher intensity appears in the North. This indicates that there is an isolated region of stronger electron precipitation. This region was not seen in the X-ray picture, indicating that the precipitation is liaited to electrons of less than approxiaately 3 keV. In Figure 16 we show siailar aappings for the H-beta line intensity. Soae difference between the proton aurora and the electron aurora is seen. The 'hole' in the aurora on the Northern sky is stretching farther to the Hest in the proton aurora. But generally there is a surprisingly great siailarity between the spatial distribution of the X-rays, the electron aurora and the proton aurora. CONCLUSIONS AND SUMHARY This study of the first rocket based simultaneous aappings of X-rays and optical aurora by four independent detectors gives a very consistent picture of the spatial ditribution of the particle precipitation in a wide area beneath the rocket position. During this rocket flight there was a "hole* in the electron precipitation to the North of the launch site where the rocket spent aost of its flight tiae. This is consistent with the electron spectrometer •easuieaents on the rocket. The H-beta mappings indicate soae proton precipitation in the northern part of this region void of electrons. These reaote mappings give important information on the electron precipitation in wide areas below the rocket, opening for better interpretation of the other in situ rocket measurements. 99 We plan to improve the presentation of the spatial distributions by filling areas with color coded intensities instead of only coloring lines as used now. This will improve the general picture quality and also give better intensity resolution. For the times when the X-ray and auroral detectors are looking down along the geomagnetic field lines through the rocket, we will also compare the photon measurements with the particle measurements from the rocket. This give us a unique possibility to cheque theoretically calculated X-ray spectra and auroral intensities with the measured quantities. From the measured spatial and energy distribution of the X-ray sources, we will try to unfold the spatial and spectral distribution of the electron precipitation. ACKNOWLEDGEMENTS We acknowledge the support from Finn Søraas who was project scientist for the Poleward Leap Rocket, Kjell Aarsnes who participated on the X-ray experiment, and Jon Bjordal and Magne Håvig on the data reduction. We also acknowledge good assistance from Kåre Slettebakken on the mechanical work and Svein Njåstad and Arne Solberg on the electronics for the X-ray detectors. Likewise we would express our gratitude to Bjørn Fjeld for mechanical construction of the photometers, and to Torleif Abell Sten who was in charge of the electronics part of these instruments. Financial support to the project was given by The Royal Norwegian Council for Scientific and Industrial Research (NTNF), Space Activity Division. 100 REFERENCES Aarsnes, K., J. Stadsnes and F. Søraas, Rocket aeasureaents of X-rays and energetic electrons through an auroral arc, In ESA Report SP 115, p 241, 1976. Barcus, J.R., and T.J. Rosenberg, Energy spectrua for auroral zone X-rays 1. Diurnal and type effects, J. Geophys. Res. 11, 803, 1966. Bjordal, J., K. Brunstad, T.E. Hoe, J. Stadsnes, S. Dllaland, A. Egeland, N. Karlsen and E. Thrane, Measurements of auroral-zone phenomena by a versatile balloonborne payload, ESA Report SP 115, 371, 1976. Goldberg, R.A., J.R. Barcus, L.A. Treinish and R.R. rtmdrak, Happing of auroral X-rays froa rocket overflights, J. Geophys. Res. 87., 2509, 1982. Jespersen, H., B. Landaark and K. Kaseide, Coaporison of auroral light eaission and electron density, J. Ataos. Terr. Phys. 11, 1251, 1969. Kreaser, G., J. Bjordal, L.P. Block, K. Brønstad, H. Håvåg, I.B. Iversen, J. Kangas, A. Korth, H.H. Hadsen, J. Niskanen, W. Riedler, J. Stadsnes, P. Tanskanen, K.H. Torkar, S.L. Uilaland, Coordinated balloon-satellite observations of energetic particles at the onset of a nagnetospheric substora, J. Geophys. Res. £7, 4445, 1982. Hauk, B.H. and G.K. Parks, X-ray iaages of an auroral breakup In Physics of auroral arc foraation, Eds. S.I. Akasofu and J.R. Kan, AGO, Washington, D.C., 1981. Håseide, K., Rocket aeasureaents of the voluae eaission profile for auroral glow, planet. Space Sci. 15_, 899, 1967. Pytte, T., R.L. HcPherron, H.G. Kivelson, H.I. West and E.W. Hones, Multiple-satellite studies of magnetospheric substoras: Plasaa sheet recovery and the poleward leap of auroral zone activity, J. Geophys. Res. 8_2, 5256, 1978. Søraas, F., H.R. Lindalen, K. Måseide, A. Egeland, T.A. Sten and D.S. Evans, Proton precipitation and the H beta eaission in a postbreakup auroral glow, J. Geophys. Res. 22, 1851, 1974. vallance Jones, A., Aurora, D. Reidel Publ. Co. Dordrecht-Holijnd, 1974. Vij, K.K., D. Venkatesan and CD. Anger, Investigation of electron precipitation during an auroral substorn by rocket borne detectors, J. Geophys. Res. 80, 3205, 1975. 101 MAGNETOMETER ANO0YA ROCKET RANGE ; i l\ ! i ! i i 1 j \ r'\ ••- '"1 V 4 %n^Lj--J\ q 1 ! i » ^ i ! ' »j! • W i : v. : I i !i \i 1nil\ nm iw i- , 1 1 1 11 1M!I I M l it* hl\m 22— ?3M ?4M Nov 1l - 83 Figure 1. Magnetogram from Andtya Rocket Range (ARR) Figure 2. All sky TV picture from ARR in the middle of the rocket flight. North is up and East to the right. Mountains limit the field of view in the Southern sky. 102 Figure 4. The area covered by the X-ray and auroral mappings from the Poleward Leap rocket. The ground track of the rocket trajectory is indicated by the double line. Dashed grid gives geographic coordinates. Grid of fully drawn lines: Magnetic local time and L-contours. The large circle shows a horizontal distance of 1000 km from the nadir point of the rocket at apogee. The small circle indicates approximate FOV of the all sky TV camera at ARR, and the fully drawn curve shows limitations of FOV on the Southern sky by mountains. 103 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 104 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 105 DAYSIDE AURORAS IN RELATION TO THE INTERPLANETARY MAGNETIC FIElDi A CASE STUDY P.E. Sandholt, A. Egaland, B. Lybekk, Institute of Physics, University of Oslo, Norway and C.S. Deehr, Geophysical Institute, University of Alaska, Fairbanks ABSTRACT. Dynamics of dayside auroras, including cusp emis sions, and their relation to the interplanetary magnetic field (IMF) have been investigated with optical ground-based observations from Svalbard, Norway, and IMF data from various satellites. Combined with the Svalbard program, simultaneous nightside observations from Alaska provide information on the large-scale behaviour of the auroral oval. Drift charac teristics, spatial scale, time of duration and repetition frequency of auroral structures on the dayside, occurring at the time of large-scale oval expansions (IMF Bz < O), are observed to be consistent with expected optical signatures of plasma entry from the magnetosheath, in association with flux transfer events. One case is presented, showing magnetic signatures of the IMF Bv—related ionospheric convection currents (the DPY component) and their latitudinal location relative to polar cap and cusp emissions, for IMF Bz positive and negative values. 1. INTRODUCTION This raport presents a case study of ground based obser vations of dayside auroral emissions in relation to the interplanetary magnetic field (IMF). The study can be separa ted in three specific topics: i) Large-scale dynamics The term large—scale refers to both the time duration (tens of minutes) and the spatial dimension (the entire belt of luminosity associated with the polar cusp and/or the dayside section of the auroral oval). Three possible influences an the location and dynamics of the dayside aurora have been studied extensively during the last 10 years. At present no general agreement exists on the relative importance of the IMF Bz component, the magne- tospheric substorm current system and the storm current (ring current). This is due to difficulties associated with the monitoring of the parameters in question and the fact that they are not entirely independent variables. Me maintain that the infuence of the IMF Bx and Bv components should be taken more seriously than they have been up to now. According 106 to the antiparallel marging hypothesis (Crooker, 1979) non zero B« and Bv values introduce north-south and dawn-dusk asymmetries in the merging geometry. In this study we focus on the midday and midnight sec tors of the auroral oval and thair responses to IMF varia tions, based on simultaneous optical measurements fram Sval bard (Norway) and Alaska. Furthermore, the nightside aurora can be used as a sensor of the state of the magnetosphere, which is important to the study cf the dynamics of the day- side aurora (e.g. Eather et al., 1979). ii> Small-scale dynamics The term "small-scale" refers to individual luminosity structures with limited spatial extent and with lifetimes of the order of minutes. Observed features in the dayside aurora with such scales may be related to flux transfer events (FTEs), which are observed by satellite near the dayside magnetopause (e.g. Russel and Elphic, 1979; Rijnbeek et al., 1984; Lee and Fu, 19B5). According to the present standard model, FTEs are characterized by a rather short-lived and local magnetic connection across the dayside magnetopause, which is initiated by a transient reconnect!on between mag- netosheath and magnetospheric field lines. During such events magnetosheath plasma elements are expected to be injected along the interconnected flux tube (cf. Cowley, 1984) with the possible effect of producing certain luminosity signatu res in the upper atmosphere of the cusp region. iii) Optical amissions and convection currants in the polar cusp and cap ionospheres A meridian chain of magnetometer stations rotating with the earth (Akasofu et al. , 1983; Friis-Christensen et al., 19B4; Friis-Christensen, 1985) may be used to obtain the average diurnal variation of equivalent currents over the northern hemisphere. The diurnal pattern of the IMF—related plasma convection in the polar ionosphere is then obtained by solving the differential equation relating the electric potential and the equivalent current function, derived from the ground-based measurements (cf. Kamide et al., 19B1). Vennerstrdm et al. (19B4) studied the relationship between 107 the latitudinal location of the DPY convection current and the F-region electron temperature enhancement usually asso ciated with the cusp, Measured by the Sondrestrom radar. In this report the relationship between the DPV signature and optical emissions in the polar cusp and cap is investigated -for positive and negative values of IMF Bz. 2. OBSERVATIONS The dayside aurora was recorded by two sets of meridian scanning photometers operated at two stations on Svalbard (Spitzbergen), (cf. Sandholt et al., 1985). The geomagnetic coordinates of these stations, Ny Ålesund (NYA) and Long yearbyen (LYR) are: 75.4°, 131.4° (NYA); 74.4-, 130.0° (LYR). By this technique the dayside auroras can be observed within the range 69° - 80° geomagnetic latitude, at midwinter. Local magnetic noon and solar noon at the recording sites occur at 0830 and 11 UT, respectively. An all-sky imaging photometer was operated at Ny Åle sund. This instrument has a 155° field of view (spanning 1200 km for F-region emissions) and a threshold sensitivity of 30 R at 630 nm (cf. Carlson, 1984). The instrument provided important supplementary information relative to the meridian profiles recorded by the scanning photometers. Dayside geomagnetic disturbances were recorded by stan dard magnetometers at four stations on the Svalbard island group: Ny Ålesund (75.4° gm.lat.), Longyearbyen (74.4° gm. lat.), Hornsund (73.5° gm.lat.), and Bjtirndya (71.1° gm. lat.). On the nightside, the Canada-Alaska chain was used (cf. Akasofu et al., 19B3). A set of meridian-scanning photometers was operated at Poker Flat, Alaska (65° gm.lat). This site is displaced 11 hours in local time relative to Svalbard, providing simulta neous observations of the midnight and midday sectors of the auroral oval, when combined with the Svalbard program. i) The Jan. 04, 1984 case The dayside aurora observed from Svalbard was located at approx. 70° geomagnetic latitude between 06 and 08 UT, corresponding in time to a large negative IMF Bz component (-10 nT). In the period 0745-1000 UT three major changes in 108 the IMF state took place. A strong disturbance was detected by the ISEE-2 spacecraft outside the bow-shock at approx. 0745 UT (cf. Fig. 1>. The total field was suddenly enhanced by a factor of 3, from 10 to 30 nT, then decreasing after a few minutes to a stable level of appr. 20 nT. Associated with this change a switch of Bv from +7 to -17 nT occurred. Bx was changed from —10 to +3 nT, before returning to nearly the initial value. The three left panels in Fig. 1 show B«„e_i, Bv, and Br. These IMF traces have been shifted by IS minutes relative to the ground data in order to take into account the time delay between the arrival of the disturbance at the two sites of observation. A factor of 3 increase in the red luminosity (630.0 nm) occurred just before 08 UT, some 15 minutes after the IMF compression was detected outside the bow-shock. A subsequent decay of the red line started at 0800 UT and reached a stable level at 0815 UT. The blue band at 427.8 nm (Na* 1. negative) did not show as large an initial enhancement as the red and green oxygen lines. Zenith angle profiles from the scanning photometer system are shown in Fig. 2. The red line intensity versus time and zenith angle is also presented in the middle panel of Fig. 1. The right panels show the H-component mag netic deflections at the stations Ny Ålesund (NYA), Hornsund (HSD) , and Bjorndya (BJA). Notice the large negative distui— bance between 0800 and 0840 UT at the southernmost station, located close to the latitude of the aurora. The next major IMF transition occurred at 0818 UT, when Bz turned positive. B„ changed from -6 to +3 nT at the same time. B«o«_i and Bv did not change. After this change the poleward boundary of the luminosity ..-wed towards north.The H—component deflection also shifted northward, increasing (mor" negative) at Hornsund and Ny Ålesund, while decreasing at BjorntSya. A rather narrow, east-west aligned discrete arc at the poleward boundary of the mure diffuse red belt is a characteristic feature during this period (cf. Fig. 2, 0845 and 0900 UT panels). This arc is located close to the minimum in the H-component profile and the zero-point in the Z-com- ponent profile (cf. Fig. 3). At approx. 0925 UT the IMF vector turned farther north ward. This change was followed by a rapid poleward expansion 109 in the equatorward boundary of th» r>d bait (cf. Fig. 1 middle panel and Fig. 2 last panel). From approx. 0900 UT auroral luminosity occurred near the southern edge of the field of view, separated from the aurora farther north and characterized by a relatively intense green line. ii) The Dec. 27, 19B4 case Figure 4 shows simultaneous multi-channel observations of day- and night-side auroras. A stable, quiet cusp arc was located to the north of Longyearbyen for several hours around magnetic noon (0830 UT). Between 1130 and 1210 UT the lumi nosity moved rapidly towards the southern horizon. Associa ted with this auroral movement the H-camponent magnetic -field (Ny Ålesund) went negative. A southward expansion of the midnight aurora occurred in the same period, indicating a large radial expansion of the auroral oval, following a southward turning of the IMF vector (left panel). The IMF observations were recorded by the AMPTE—UKS satellite, loca ted close to the bow-shock, in the dawn sector. At appr. 1210 UT a major substorm intensification occurred in the Alaska sector. The H-component at Arctic Village, shown in the right panel of Fig. 4, is representative of the nightside westward electrajet as observed by the Alaska chain during the substorm. Between 1300 and 1330 UT (appr. 16 MLT) northward and southward turnings of the IMF are observed to cause poleward and subsequent equatorward movements af an east—west aligned polar cap auroral arc (Fig. 4). The event is shown mare clearly in Fig. 5. In the two left panels are plotted the latitude and longitude angles of the IMF vector. Before the IMF transition at appr. 1250 UT, Bz = Bx— 0 and Bv = IO nT. After the change, at 1255 UT, B* = 9.5 nT, B-, = 0 and Bx = -2.6 nT. Thus, associated with the northward movement, Bv also changed from large positive towards zero, i.e. anti clockwise. At 1301 UT, IMF turned back towards a state with positive V and negative Z components. The spacecraft crossed the bow-shock at 1304 (inbound) and at 1319 UT (outbound). A perturbation at 1234 UT is due to a Barium release from a companion satellite (AMPTE IRM). When comparing the quiet, midday aurora (cf. Fig. 4 110 appr. 1030 UT) with the afternoon polar cap arc (1310 - 1330 LIT), we notice a significant difference in the spectral ratio 1(630.0 nm>/I(557.7 nm). The later one shows a stronger green line emission and a lower value of this ratio than the midday aurora. iii) The Dec. 10, 1993 case Figure 6 shows another example of simultaneous photome ter observations of the midday aurora from Svalbard and the midnight aurora from Alaska. The north-south component of the magnetosheath magnetic field, recorded by spacecraft ISEE-2, is shown in the left panel of the figure. ISEE-2 crossed the bow-shock at 1009 UT. A major southward shift of the field vector occurred at 0735 UT. During the following 1 1/2 hr magnetosheath B was rather stable: 8» = 0, Bv - 20 and Bi = -30 nT. Between approx. 0750 and 0820 UT both the day- and night-side luminosities expanded southward. Then, at 0820 UT, a major substarm onset occurred on the nightside, which is clearly reflected in both the optical and geomagne tic recordings (right panel). The auroral break-up was fol lowed by a distinct poleward expansion. Between 0B20 and 0B50 UT, 4 successive poleward moving, transient structures were observed above Svalbard. Details are shown in Fig. 7. At 0855 UT the external magnetic field turned northward. Another southward transition was detected at 1003 UT. Between 0910 and 1020 UT, corresponding to northward external field, the luminosity was observed to extend farther north than in the preceeding period when magnetosheath Bz was negative. At 1020 UT another southward shift took place, some 15 min after the southward field transition. In the interval 09-10 UT the magnetosheath total field was approx. 50 nT, with Bv~ 40 and B*=20 nT. 3. DISCUSSION OF OBSERVATIONS i) The Jan. 04, 1984 case This case can be separated into 3 periods with different IMF conditions. These periods are indicated on Figure 1. Per i od 1. The magnetic signature indicates a southwestward 111 DPV current in a sunlit cusp ionosphere with the amplitude modulated by IMF Bz Period 2. The transition in IMF Bi from negative to positive at 0818 UT (Bv not changing) was followed by marked changes in the optical aurora and the DPY signature. The location and motion of the aurora and the DPY perturbation are illu strated in Fig. 3. At 0900 LIT the center of the westward DPY current is above the auroral observing cite at Longyearbyen (A H max. negative and & Z = 0). At this time the cusp aurora peaks in intensity somewhere to the south of the latitude of Hornsund. The DPY equivalent current center is located at the poleward boundary of the cusp aurora. This shows that the DPY current is extending into the polar cap from approx imately the center of the optical cusp. This is in agreement with Vennerstrom et al. (1984), who found that the equator— ward boundary of the DPY current is colocated with the lati tude of maximum ionospheric electron temperature, usually associated with the center of the cusp region. The DPY cur— rent observed in time interval 2 is probably related to the polar cap lobe cell convection postulated for Bz > 0 (cf. Crooker, 1979; Reiff and Burch, 1985) and inferred by use of the magnetometer chain along the westcoast of Greenland 1984). Period 3. During thi» period, from 0940 to 100O UT as ubmei— ved from the ground, the IMF vector is pointing due north. This state should produce a net magnetic flux transfer from open lobe tubes to closed dayside tubes (cf. Cowley, 1981) with a resulting poleward motion of the equatorward boundary of the cusp. This expected effect is observed at 0740 UT. A weak cusp-like aurora is located to the north of Ny Ålesund. ii) The Dec. 27, 1984 case A major oval expansion and enhanced DP2—type magnetic disturbance started appr„ 40 min. prior to the substorm onset at appr. 1210 UT. Also notice the equatorward movement of the 6. H = 0 line, separating zones of positive and nega tive deflections, in this period IMF Bz is quite clear. From these observations we may infer that the dimension of the polar cap increased towards a maximum approximately at the time of the substorm onset an the nightside. The size of the polar cap is considered to be a measure of the open magnetic flux and thus of magnetic energy in the magnetotail (e.g. Makita et al., 1985). Follo wing this interpretation, the optical and geomagnetic data recorded in the period 1130-1210 UT seem like signatures of the growth phase iii) Th* OK. 10, 1983 cae* A most remarkable feature is the relationship between the location of the poleward border of the midday luminosity and the magnetosheath Bz component. Me notice clear responses to the IMF northward and southward transitions at 0855 and 1003 UT, respectively pattern for IMF Bv > 0 (e.g. Heelis, 1984). In the actual case magnetosheath Bv = 20 nT. A more extensive discussion of similar events is given in Sandholt et al. (1985 and 1 '/86) . 4. SUMMARY AND CONCLUSIONS i) The middag auroral luminosity often shows latitudinal differences in the spectral properties indicating dif ferent sources of the associated particle precipitation. Different particle source regions in the dayside magnetos- 114 phere, projecting to the ionosphere within the -field of view of the scanning photometers at Svalbard, arc: a) The polar cap b) The polar cusp c) The dayside extension of the plasma-sheet boundary layer. The corresponding diversity o-f the dayside auroral emission? must be taken into account when discussing the cusp response to the IMF, based on optical data. ii) The reported observations show that the IMF Y and Z components exert major influences on the polar cusp and cap auroras. Concerning the possible influence from the magnetospheric substarm on the location of the luminosity a reservation is expressed, due to lacking AE—index infor— mati 01 at the time of publication. However, the Dec. 27, 1984 (Fig.4) and Dec. 10, 1983 (Fig. A) cases do not indicate such relationship. iii) A simultaneous equatorward motion of the day- and nightside auroras in response to southward turning of the IMF, prior to substorm onset, is observed. The related increase of the polar cap area, reaching a maximum at the time of substarm expansion phase onset, seems to support the growth phase concept, involving energy storage in the tail before release at expansion onset. This interpretation of the data is based on our identification of the onset time. We realize that data from more than one longitudinal sector on the nightside are necessary in order to resolve this quer.tion. The expansive piiase onset at apprax. 1210 UT on Dec. 27, 1984, could be triggered by the nearly simultaneous northward turning of the IMF vector res i. Opening and connection with magnetosheath fie!d lines (B« < 0, Bv f O) of initially closed flux tube at the equatorward side of the cusp. 2. Injection of magnetosheath plasma along the flux tube and subsequent penetration down to the upper atmosphere. 3. Reconnected flux tube and associated auroral lumi nosity convecting north—westward or north—eastward (Bv < 0) across the cusp, into the polar cap, at speeds of approx. 500 ms_". 4. Auroral luminosity associated with plasma injection disappearing typically after 5-10 min. v) The location along the magnetic meridian of the optical cusp aurora and the DPY convection signature in the winter hemisphere have been determined for IMF Bx negative and positive conditions. Some preliminary results are: a) IMF Bz < 0: The DPY signature maximizes at the lati tude of the cusp aurora. This is consistent with existing models and observations of the merging cell, which convect magnetic flux from the closed dayside into the open polar cap, across the midday cusp. IMF Bv < 0 corresponds to south- westward DPY current (northeastward convection) in the cusp (cf. Fig. 8). The local DPY amplitude is not strongly modu lated by intensity variations in the cusp aurora, durir;-> the observed case. On the other hand, DPY is highly sensitive to IMF Bz variations. b) IMF Bi > 0: The DPY signature of a westward current (Bv < 0), probably associated with the iobe cell convection, peaks at the poleward boundary of the red—dominated, diffuse cusp aurora. The cusp boundary and the DPY peak arm colocated with a discrete arc showing a lower value of the spectral ratio I(630.0)nm/I(557.7)nm (indicating higher energy parti cle precipitation), compared with the central cusp. vi) Plasma transfer across the dayside magnetopause: During IMF Bz < 0 conditions the intensity of the cusp aurora lib is highly sensitive to solar Mind irregularities, indicating efficient plasaa transfer. When IMF Bz turns due north the auroral intensity decreases towards a minimum, indicating that the plasma transport into the «agnetospheric cusp region is markedly reduced. Fig. 8 (left panels) shows the magnetic field configu ration at the dayside magnetopause for IMF Bz < 0 and diffe rent Bv polarities, based on the antiparallel merging hypo thesis (cf. Crooker, 1979) and subsolar reconnection ce, depending on the sign of IMF B>. (Bv 4= 0). Negative (po sitive) Bv means FTE appearance in the pre-noon (post-noon) sector and subsequent eastward (westward) convection, in agreement with our optical observations (cf. Sandholt et al., 1986). ACKNOWLEDGEMENT It is a great pleasure to thank S.W.H. Cowley (Imperial College, London) for valuable comments on expected optical signatures of flux transfer events, as well as R.C Elphic (University of California) and R. Rijnbeek (Imperial College, London) for providing IMF data from respectively the satel lites ISEE-2 (principal investigator: C.T. Russell) and AMPTE-UKS (p.i.: D. Southwood). Magnetograms from Ny Ålesund and Bjornfiya were provided by S. Berger (University of Trom sø). Magnetic recordings from Hornsund, obtained from the Palish Academy of Sciences, Warszawa, Poland (A.W. Wernik), are also gratefully acknowledged. We express our thanks ta K. Maseide (University of Oslo) for valuable comments on an early version of this manuscript. 117 REFERENCES Akasofu, S.-I., B.-H. Ahn, and G.J. Romicfc, A study of the polar currant system using the IMS meridian chain of magnetometers. I: Alaska Meridian Chain. Soac» Sei. Rav.. 24, 337, 1983. Bakar, O.N., E.M. Hones, Jr., P.R. Higbia, R.D. Belian, and P.Stauning, Global properties of the magnetosphere during a substorm growth phase: A case study, J. Geophvs. Res. , B6, 8941, 19B1. Banks, P.M., T. Arak i, CR. Clauer, J.P. St. Maurice, and J.G. Foster, The interplanetary electric field, cleft currents and plasma convection in the polar caps, Planet. Space Sci.. 32. 1551, 19B4. Carlson, H., The HILAT ground-based program, Johns Hopkins APL Technical Digest. 5, p. 143, 1984. Cowley, S.W.H., Magnetospheric and ionospheric flow and the interplanetary magnetic field, in The Physical Basis of the Ionosphere in the Solai—Terrestrial System, Conf. Proe. AGARD-CP-29S. p. 4-1, NATO, Neuilly-Sui—Seine, 1981. Cowley, S.W.H., Substorms and growth phase problem, NATURE. 295. 365, 1982. Cowley, S.W.H., Solar wind control of magnetospheric convec tion, in Proc. Conf. Achievements of the IMS. ESA SP-217. p. 483. 1984. Crooker, N.U., Dayside merging and cusp geometry, J. Geophvs. Res.. §4» 951, 1979. Crooker, N.U., A split separator line merging model of the dayside magnetopause, J. Geophys. Res. 90. 12104, 1985. Eather, R.H., S.B. Mende, and E.J. Weber, Dayside aurora and substorm current systems, J. Geophys. Res.. 84. 3339, 1979. Frank, L.A., J.D. Craven, and R.L. Rairden, Images of the earth's aurora and geocorona from the dynamics explorer mission, Adv. Space Res., 5, 55, 19B5. Friis—Christensen, E., Solar wind control of the polar cusp, Danish Met. Inst.. Geophysical Papers. R-72, 1985. Friis-Christensen, E., Y. Kamide, A.D. Richmond, and S. Matsushita, Interplanetary magnetic field control of high-latitude electric fields and currents determined from Greenland magnetometer data, J. Geophvs. Res. 90. 1325, 1935. Heelis, R.A., The effect of interplanetary magnetic field orientation on dayside high-latitude ionospheric convection, J. Geoohvs. Res.. 89. 2873, 1984. Kareide, V., A.D. Richmond, and S. Matsushita, Estimation of ionospheric electric fields, ionospheric cur— rents, and field-aligned currents from ground magnetic records, J. Geophvs. Res.. 86. 801, 1981. Lee, L.C. and Z.F. Fu, A theory of magnetic flux transfer at the earth's magnetopausr?, Geophys. Res. Lett. . 12, 105, 1985. Makita, K., C.-I. Meng, and S.-I. Akasofu, Temporal and spa tial variations of the polar cap dimension and its relation to the energy input rate and AE index, J. Geophys. Res.. 90. 2744, 1985. Reiff, P.H. and J.L. Burch, IMF Bv-dependent plasma flow and Birkeland currents in the dayside magnetosphere. 2. 118 A global Model for northward and southward IMF, J. Beonhvs. Re».• 90, 1S9S, 1985. Rijnbeefc, R.P., S.M.H. Cowley, D.J. Southwood, and C.T. Russell, A survey of day side -flux transfer events observed by ISEE 1 and 2 magnetometers, J. Geoohy». Res., B9, 786, 1984. Rostoker, 6., S.-I. Afcasofu, J. Foster, R.A. Breenmald, Y. «amide, K. Kawasaki, A.T.Y. Lui, R.L. McPherron, and C.T. Russell, Magnetospheric substorms - definitions and signatures, J. Beoohvs. Res.. 85. 1663, 1980. Rostoker, G., Triggering of expansive phase intensifications of magnetospher i.c substorms by northward turnings of the interplanetary magnetic field, J. Beophvs. Res., 88, 6981, 1983. Russell, C.T. and R.C. Elphic, ISEE observations of flux transfer events at the dayside magnetopause, Geo- phvs.Res•Lett., 6, 33, 1979. Sandholt, P.E., A. Egeland, J.A. Holtet, B. Lybekk, K. Sve nes, S. Åsheim, and i.S. Deehr, Large- and small- scale dynamics of tt*. polar cusp, J. Geophvs. Res. 90, 4407, 19B5. Sandholt, P.E., A. Egeland, C.S. Deehr, B. Lybekk, R. Viereck, and G.J. Romick, Signatures in the dayside aurora of plasma transfer from the magnetosheath. Submitted to J. Beophvs. Res. Vennerstrom, S., E. Friis-Christensen, T.S. Jdrgensen, 0. Rasmussen, CR. Clauer, and V.B. Mickwar, Iono spheric currents and F-region plasma boundaries near the dayside cusp, Beophvs. Res. Lett., 11. 903, 1984. FIGURE CAPTIONS Fig. 1. Left three panels: IMF measurements from the ISEE-2 satellite, i.e. total field, east-west (Bv) and north—south (Bz> components in the geocentric sun-ecliptic (GSE> coordinate system. Middle panel: Intensity versus zenith angle and universal time of the red oxygen line at 630.O nm, observed from Longyearbyen (74.4° geom.lat). Right three panels: H-component magnetic distui— bance detected at the three stations Ny Ålesund (NYA), Hornsund (HSD), and BJDrnrtya (BJA). Nega tive deflection towards left. The data recorded on the ground have been shifted by apprax. IS min relative to the IMF traces. Fig. 2: North-south magnetic meridian scan profiles for the three wavelengths 630.0 nm <0I>, 427.8 nm (Nz~*"> and 557.7 nm (01) observed at five selected times during the period shown in Fig. 1. Intensity scales are marked to the left. Fig. 3: Latitude profiles of the Z- and H-components of the magnetic disturbance based on recordings from the four Svalbard stations located as shown on the horizontal axis. The scale on the vertical 119 axis is in nanoteslas. The latitudinal location of auroral luminosity is estimated from the pho tometer profiles in Fig. 2. Left panel: Latitude angle of the IMF vector. North is positive. Data from the AMPTE-UKS satel lite, located close to the bow-shock in the dawn sector. with Bz < 0. Compare with groundbased observations in the midnight sector. Panel 2: H component disturbance field at Ny Ålesund, Svalbard (75.4™ gm. lat.). Positive deflection towards left. Panels 3—5: Meridian scanning photometer (MSP) traces from Longyearbyen (74.4«» gm. lat.). The AH = 0 line is the boundary between zones of positive and negative H-component deflections. Panels 6-10: MSP traces from Poker Flat, Alaska, (65= gm. lat.) (1130 UT = 24 MLT). Day and night- side intensity scales are different. Right panel: Magnetic disturbance field at Arctic Village (69= gm. lat.), near the center Df the westward electrojet at 1230 UT. Left two panels: Latitude and longitude angles of the IMF vector, respectively. 0"» longitude is towards the sun and 90= is towards east (dusk). Middle section: Intensity versus zenith angle and time representation of the red oxygen line (simi lar to Fig. 1). Right three panels: H—component magnetic deflec tions at Ny Ålesund (NYA), Hornsund (HSD), and Bjdrnoya (BJA). The ground data have been shifted by appr. 15 min. relative to the satellite record ings. Left panel: Magnetosheath magnetic field Z—compon ent in nanoteslas- Time is unversal time. The spacecraft reached the bow-shock (from inside) at 1009 UT. Panel 2: North-south meridian scan profiles of the red oxygen line (630.0 nm) in the midday aurora, recorded at Ny Ålesund, Svalbard (75.4= gm. lat.). Panels 3 and 4: Similar meridian scans of the nightside aurora recorded at Poker Flat, Alaska (65= gm. lat.). Day- and nightside intensity scales are different. Right panel: Magnetometer recording from Arctic Village (69° gm. lat.). The ground data have been shifted by 15 min. relative to the satellite data. North-south meridian scan profiles of the red oxygen line (left panel) and the blue nitrogen band at 427.8 nm (right panel) in the midday aurora above Svalbard. Details from the overview 120 plot in Fig. 6. Fig. St Left panelms Schematic view from the sun of the magnetic field configuration at the dayside mag- netopause resulting from the superposition of the Chapman-Ferraro field and the IMF. For Bv + 0 merging occurs at the outer merging lines (heavy solid). The merging process operating in the subsolar region when Bv 4 0 gives rise to flux transfer event signatures as described by Lee and Fu (1985). For Bv " O antiparallel merging occurs along the equatorial magnetopause. (After Crook- er, 19B5.) Right panels: Convection patterns in the northern polar ionosphere for different IMF Bv polarities: (IMF Bz < O). Streamlines crossing the polar cap boundary (dashed circle) effect a transfer of flux from closed dayside field lines to open tail lobe field lines. These streamlines are constitu ting the merging convection cells. For Bv +• 0 a convection cell entirely confined to the polar cap is also predicted, called the lobe cell, due to the reconnection between tail lobe field lines and the IMF at the flanks of the magnetopause. Solid dots mark the possible locations of tha initial appearance of ionospheric signatures of magnetopause reconnection events (QSRs and FTEs). Expected directions of motion of these signatures are indicated by arrows from these dots. JAN.04,1984 INTERPLANETARY MAGNETIC FIELD CUSP AURORA H-COMP MAGNETOGRAMS (nanoteslas) SVALBARD SVALBARD (630.0nm) r O7.0OU 10 0 10 20 BJA NYA HSD BJA (7111 NORTH ZENITH ANGLE SOUTH FIG.1 122 MIDDAY AURORA SVALBARD JAN.04,1984 N 40 0 40 S ZENITH ANGLE 123 SVALBARD JAN.04,1984 200 A2\ 100 0845 UT 0 -100 -200 -300 L CUSP AURORA I i 1 DISCRETE 100 ARC AZN 0900 0 UT -100 -200 CUSP AURORA i- H DISCR.ARC NYÅ LYR HSD BJA • • ±__J ± L. 76 75 74 73 72 71 GEOMAGNETIC LATITUDE FIG.3 DEC. 27,1984 SVALBARD ALASKA ALASKA MAGNETOMETERS PHOTOMETERS PHOTOMETERS MAGNETOMETER SOLAR WIND SVALBARD UT 11001- 90 0 -90 GSE LATITUDE (DEG.) AMPTE 4278nm 5577nm 630.0nm 55Z7nm 4278nm 486.1nm 630.0nm 732nm UKS H-COMR NORTH-SOUTH MERIDIAN SCANS H-COMP FIG.4 DEC.271984 INTERPLANETARY MAGNETIC FIELD POST-NOON SECTOR AURORA H-COMP MAGNETOGRAMS SVALBARD SVALBARD GSE LAT. GSE LONG. (630nm) (nanoteslas) O 90 180 270 360 -100 -50 -100 -50 -100 -SO 0 -45 0 45 90 AMPTE-UKS NYA HSD BJA 71.1 •ill <1 SO /-.Il 754 735 NYÅ HSD BJA NORTH ZENITH ANGLE SOUTH FIG.5 126 DECEMBER 10,1983 ISEE-2 SVALBARD ALASKA ALASKA MAGNETOMETER PHOTOMETER PHOTOMETERS MAGNETOMETER 630.0 nm MAGNETOSHEATH Bz 630.0nm 5S77nm H-COMPONENT 40 0 -40 N S N Z SN Z S FIG.6 127 SVALBARD DEC.10,1983 CUSP AURORA 630.0nm 427.8nm A\ 08.1125 oa i.? De n» • i 3» •1 i i !« Ol 13 Sl Dl IM ?| 1kR 8.5kR 0. S 03 fTfl 15 *)J - Df s ta as ib ib • • b Ib ai b 52 PB >ti se. UU I 1 JQ- an 1 3D. •• • 05. OB It Sl - ^fe: 0a2Q3O - OB ei ab • QB Bl 13 .__. f'^_ '•" / \Vv 'JR Sl •*». OB SE 1' .--. ai ee *b--- 00 ?3 1? . . TIM E 9fl ?t 41 - - ul i-S E£ ^ OB ?S i' - .- . . .. f 7 -V OB eb 33„ . -••ai •?:•- OB ei -a UNIVERSA L ft co ?> ti . _ ze en ee •B es si v. . ••-"^^ ae ?•» js.- •• n» •*• is . •" ~ ' . k ... .._/ •••...••MF- • ; ' '/i' A.- 0&30.12 . 0a30.12 '• - .. •-.-.,{ .. /""• OB ia -IB.. '*-- •• '•• f-ri. OB 1U HI . _ —-•- ''' •./••'' W'' . -,''- RH 3i ?H ,."- : • H Si S*. .. 7 • -.\'~" '•j'-J - '-"',' ne 3e ui • •• as »? oi •••- - Ofl 1? 31 . OB fi* 11 i \ nn 3 3 i.. - '• OB »i H . • "' '""' ';* nn >) 5! • Dl i) Si - "• '• ''IV.- IIS H" /I. •'- 31 1" fu - /• •' [•H i', ae JI ib ne -- "V UB 1', !<*.. • u6 'S f». . NU Ik. I<. - OB lb S N 40 0 40 S N 40 0 40 s ZENITH ANGLE ZENITH ANGLE FIG.7 128 IMF%<0 tOi.u*p (oU ctiC M mr convection. / ciflWW dusk Ob %y<0 IMF & i-nr c&u.w* . c(u IMF toouspl ^ Js)rM IHF tf 24 fofaroy ,boundary faMin. dusk, Ob if IS ByO 11/ / FIG.8 SiQ natures 129 NEW IDEAS OF SUBSTORM DEVELOPMENT AS 0B5ERVE0 BY GROUNO-BASED OPTICAL INSTRUMENTS Risto J. Pel 1inen Department o-f Geophysics, Finnish Meteorological Institute, P, Box 503, SF-001O1 HelsinKi, Finland Abstract Observat ions made with auroral all-sKy camera, TV, photometers and some other ground-based instruments hAK/e revealed several new features in auroral behaviour in connection with substorms that have not earlier been discussed in the literature. The systematics o-f these observations hint that in order to understand more about the substorm mechan ism one has to add these neu facts to the commonly accepted substorm picture. Southward turnings of the interplanetary magnet ic field as ue.ll as storm sudden commencements occur ing during quiet magnetospheric conditions lead to an immediate onset of particle precipitation lasting for more than 30 min before auroral breaK-up, Substorm onset starts with formation of an auroral bulge in the midnight sector. On the evening-side edge of the bulge a hi3h-speed (up to 30 Km/s) brightening propagates along the poleward edge of the pre-onset precipitation band. It is followed by a westward traueling surge hav ing an order of magnitude lower speed. The auroras on the morning side of the bulge behave similarly but are more irregular, sometimes containing north-south directed structures. Immediately after auroral breaK-up the downward flowing field-aligned current on the morn ir.g-s ide edge of the auroral bulge may partially be carried by upward flowing cold ionospheric electrons producing a weave (a few hundred Raylaighs) 630.0 nm emission. These results will be discussed in the light of the present Knowledge of the magnetosphere-ionosphere coupling and auroral electrodynamics. 1 . Introduct ion Substorm onsets are observed most distinctly in regions close to the magnetic midnight. Studies mada during the International Magnetospheric Study C1MS) have revealed that the scale size of the triggering area in the ionosphere is 500-1000 Km (see e.g. Opgenoorth et al-, 1930) . Fig.l Cleft panel) gives an example of an auroral breaK-up reccoded by the DMSP-2 satellite on 3. Dec. 1977. The circle around the bright area indicates the field of view of An all-sKy camera. The picture demonstrates that all auroral observations made during the few first minutes after 1 10 Fig.l Auroral stibstorm development over Northern Europe recorded on 3. Dec. 1377 by the DMSP-2 satellite during two subsequent Pisses separated by løa min. The circles indicate the field of view of an all =Ky camera. s u b s tor m onsets can be covered by a single, su itably located, al I -s v.v camera. In the right-hand panel of Fig.l,- recorded 100 min 1 ater, it is evident that several ASC's are needed in order to ob serve the entire auroral bulge. He^e images obtained from satel 1 itas are more useful to est imate the spatial extent but, until the temporal resolution has not been as good as uith the B 3C 's . Hence, ground-based optical observations on substorm onset featuras in the midnight sector can safely be done euen i.i i t h a single imag ing instrument. The good temporal resolution and h igh sensitivity Cphotometers> of these instruments may r eve a 1 some new features that cannot be resolved by any other roe ans Beloij thrsa different topics, based on the above observation philosophy, uill be discussed. These are! 1) Soft auroral precipitation appearing before substorm onset; 2) High-speed auroral motions along the oval after breaK-up; 3> Optical observations in the region of ununuard flowing BirKeland currents. 131 Our -first example is a SSC event appearing during fairly stable magnetospher ic conditions on 23. Jan. 1979. Pin Interplanetary shocK wave recorded by four different satellites hit the magnetosphere at 0140 UT. ft substorm followed S3 min later initiated near the magnetic midnight meridian Fig.3 shous that magnetic (convection) disturbances arc observed to start immediately <9lobally> on the ground. The onset in LRM at 0209 is sharp uhile those at FCC and OIK are delayed. Also at KiJpisjarvi, in Fig.2, the enhancement in auroral emission development is delayed for S min. If we assume a spatial distance of 70 degrees to the central meridian the speed of the eastward propagation is more than 10 Km/». H COL 25 JAN 1979 0200 UT 01 03 05 UT 12 00 MLT H » LRV A / l! H°W / 01 03 05 UT Fig.3 Development c-f global magnetic disturbance in four diHerent t ime sectors After a SSC event. 132 2. Pre-onset soft auroral precipitation Conditions leading to auroral breaK-up have been of uida interest for more than ten years. Changes in the direction of the z component of the interplanetary magnetic field (IMF) or storm sudden commencements (SSC) are believed to trigger auroral and magnetospheric substorms (Caan et al.,1977, RostoKer, 1883, KoKubun et al . , 1977). However, the magnetospheric response of these events seems to depend on the level of the preceding magnetospheric disturbance. RostoKer (1983) has shoun that a northward turning of the IMF after an interval of a sustained southuard IMF can trigger the onset of a major substorm. In all his events the southuard turning occured during (or nearly) quiet geomagnetic activity. KoKubun et al. (1977) have shoun that uhen a SSC appears in a quiet magnetosphere (AE less than 100 nT) there is a -jery lou probability <8X) that a substorm follows immediately. In this Section ue uill demonstrate that soft auroral precipitation leading to increased ionospheric Csderseni conductivity appears before the final triggering. X-COMPONENT OF B SODONKYLA 25JSN1979 v "•-'., i> \ z V -1 ••: o V. * - -i \ y 200. - -•n *^\ D 400. - IMF AT IMP-8 J< \J\ p, Ji** B z > 0 ' B2 < 0 600. - V^ 1 p, . p- V TIME/HOUR UT 1 EMISSION ROTE 630 NH KILPISJftRVI 25J«N19?9 0T3 JO X<> Dfae- EXPANSION r : 5- m ofistfT A CI PFtEClP |V\ I : .0 - .5- TIME/HOUR UT ' SSC ' BU ' Fig.S Development of magnetic disturbance, IMF and auroral precipitation after a SSC event. 133 When looking at Fig.2 ona can see an obvious dangar in us',ng tha AE indax alone as an indicator of substorm development. It is raally hard to say uhan tha substorm currant uadga marges on tha bacKground convection currant systam. Tha only raaans to solwa this problem is to IOOK into optical auroral data together uith the global magnetic data. The optical data tails also more about the conductivity development uhich is important to understand the reasons for enhanced convection. 19^6 NOV 11 COMPARISON OF IMF ANO POLAR ELECTROJET VARIATIONS .SOLAR WIND ENEHGV COUPLING \FUNCTION ( i JOULE HEAT o o I a. L 10'-109W •/ 10» v Pi3.<» Comparison of the polarity Ctheta) of the interplanetary magnetic field Bz component and the solar uind energy coupling function Fig.2 also shows that the IMF turned southwards before "the SSC tooK place. This situation might have assisted in the final triggering. Pellinen et al. <1982) showed an example where a sudden southward tarn ing of the IMF lad to a world-wide convection enhancement -each ing a level of 500 nT in the AE index without any signatures o-f substorm act iv ity. Fig. 4 gives details of this development. An aurora] arc appeared in the northern SKy over Finland at the moment o-f southward turning. Be-fore the breaK-up the arc moved some £50 Km equatoruards, stopped and broKe up at the moment when the IMF made a brlef northward excursion some 30 minutes after the southward turn ing. The substorm that fol lowed was weaK, local and isolated, obviously controlled by the short negative bay in the IMF polar ity. Dur ing the "growth phase" the level of the solar wind-magnetosphere dynamo power was above the substorm boundary, i.e., mere than 10t 1 1 UI. In the literature there are several similar examples. Already in 1977 Caan et al. showed several cases where a substorm followed automaticly after about 1 h of southward directed IMF. During that period there u*.& a considerable enhancement in the RE activity. Recently Sauvaud et al. <1385) have reported on a similar event as the one in Pellinen et al. <1982). Satel lite data both from the geostationary orbit and far tail <22 Re) support the idea that a significant enhanced precipitation process is going on in the magnetosphere during the growth phase. ftKasofu and Kan <1982> explained these type of observations as due to an imperfect magnetosphere-ionosphere coupling 3. High-speed auroral mot ions after breaK-up ftt the noment cf substor IT» triggering the eq,uatorwards moving diffuse auroril precipitation band stops. The auroral breaK-up appears or the poleward edg» of the band in the evening part of the magnetic midnight sector. In Fig.l the local nature of siibstorm onset is clearly demonstrated. Also the channel with the propagating auroral brightening is visible towards the even ing sector . £)t a ur or a 1 observatories located near midnight, both in the evening and morning sectors, the arc brightening is observed to propagate away from the triggering area with a speed of 10 Km/s or higher. In the evening sector the brightening is followed by a westward trawelling surge KEVO 25 JAN 1979 2016:43- 2017:43- 2018:43- 2017:23 2018:23 2019:23 UT Fig.3 f)n example of auror.al breaK-up recorded by the rll-sKy c Arne r a at Kevo. Rap idly mo ving auroral brightening is foil owed bv a WTS. 136 Similar results have been reported in Opganoorth «t *1 . (1983) and Yahnin et al. <1983). Recently Samson <198S) interpreted the behaviour of mid-latitude Pi2's as being due to a rapid <2ø-5ø Km/s) westward expansion of arc brightening in the auroral zone. There are other such examples. Why have this type of observation not besn rsported be-fore and how are they related to the present understanding of magnetospharic substorm development? In ths literature appears frequently the statement: "Substorm onset is morphologically defined as the moment when an auroral arc brightens suddenly and begins to advance poleuards in the midnight sector". In the original paper by AKasofu <1964) the length of the brightening arc is defined as a feu thousand Kilometers and it is stated that at the same time, or a LITTLE LATER, quiet arcs in the evening sector may brighten. Hence, a reservation for a propagating brightening is made but the matter is not explicitely stated. All-sKy cameras are commonly run at a speed of 1 frame/min and with -fairly long exposure times '<10 s or more) which maKes it hard to distinguish speeds 1i«e 10 Km/s or more. Many of our observations have been made uith a speed of 3 frames/min and 2 s exposure times on color film. Some of the observations have been supported by real-time video recordings. With . these techniques we have been able to identify cases with speeds more than 20 Km/s . It is generally believed that the (boundary) plasma sheet is the source region of auroral arc particles, and that a substorm starts in a localized region of the plasma sheet and spreads towards all directions ma inly parallel to the equator ial pl ane. Hence, it seems natural that also the arc brightening propagates. The fact that there is no considerable shifting of altitude and latitude compared to the quiet diffuse background band < e.g. Opgenoorth et al . , 19S3) hints that the quiet-arc acceleration mechan ism proposed by Pellinen and HeiKKila <1334) is enhanced and propagates. The mechanism bases on an adiabatic process which continuesly feeds particles into the loss-cone due to a transverse energization of the curvature drifting particles in the neutral sheet. This mechanism is effective in changing the number flux of particles rather than the total energy. Substorm-rslated near-Earth X-type neutral line starts to form in the region where the neutral sheet current starts to decrease. Decreasing current leads to enhanced cross-tail E-field that more effectively feeds particles into the loss-cone. According to Hones <1S83) a full-size, £0Ra wide, neutral line is formed in less than 4.5 min uhich corresponds to an average speed of 10-20 Km/s in the ionosphere. Hence, we have good reasons to believe that the high-speed propagation of auroral arc brightening is associated uith the format ion of a substorm-related neutral 1ine in the magnetotail. U7 4. Downward-f 1owing BirKeland currents The pattern of magnetospheric substorm current system most often referred to is a disruption of the neutral sheet current, flowing across the magnetotail, and its short-circuiting into the ionosphere where the westward closure current is superimposed on the large-scale westward connective electrojet (CUuer and McPherron, 1374). At present there is no generally accepted mechanism to explain this current disruption. The f 11 arnentary field-aligned currents DIFFERENTIAL HORIZONTAL EQUIVALENT CURRENT VECTORS FEB 15 1977 21.20 00 - 21.18*30 UT 21.2200 - 21.18 30 138 approaching positive charge cloud of energized protons from the magnetotail. Optically this uould mean that ueaK high-altitude 633.0 nm emission, caused by the upuard flouing cold ionospheric electrons colliding with oxygen atoms, should be detectable at the onset moment before the protsns arrive a feu minutes later. In Fig.6 we shou an example of the location of Ff=IC centers some minutes after substorm onset (Pellinen and HeiKKila, 1384). Ground-based magnetic data and the so-called differential equivalent current techniques (DEC) The intensity and the amount of current carriers in the uestuard traveling surge rise another question: Is the classical picture of substorrn current uedge correct or do ue have to add some more regions of dounuard FAC in the system to balance the substorm-related BirKeland current loop. Recently Roux (1985) and HuusKonen et al. <19S5> have presented some arguments that support the above hypothesis.Roux presents a model for WTS that claims that in front of the surge there has to be a region of proton precipitation or upuard flouing ionospheric electrons. HuusKonen et al . reported observations made with the EISCAT radar. In the cases they had analyzed there was a dramatic decrease in the ionospheric electron density PHOTOMETER KIRUNA (ELEVATION 75" N) 630.0 nm I red I Background > J I L. 22:02 22:10 22:18 22:26 2231 22-42 22:50 22:58UT Fig.7 Photometric recordings of auroral emissions in the center of dounuard flouing FflC '. around 22.22 UT ). 139 Conclus ions Ir. this short report ye have presented several new views that have not earlier been included in the common picture of substorm development. We have reached the following conclusions: a> In order to enhance the Pedersen conductivity of a quiet ionosphere before substorm onset soft auroral precipitation is required. This means that the ionosphere plays an active role in substcrm triggering. fcO Since substorm processes propagate in the tail in the dawn-dusK direction/ it seems natural that also the arc brightening propagates. The speed of th« brightening mapped into the maflnetotail corresponds ts the formation speed of the neutral line (Hones, 1985). c) Part of the downward flowing BirKeland current is carried by cold ionospheric electrons producing a weaK 630.0 nm auroral emission. Part of downward FAC in the substorm current wedge may flow as a sheet current in front of the WTS. References AKasofu , S.-I., 1984: The development of the auroral substorm. Planet. Space Sc i., 30, 273...£8£. AKasofu , S.-I. and J. R. Kan, 1982: Importance of initial ionospheric conductivity on substorm onset. Planet. Space Sc i., 30, 1315...1316. Caan, M. N. , R. L. McPherron, and C. T. Russel , 1S77: Characteristics of the association between the interplanetary magnetic field and substorms. J. Geophys. Res., 82, 4837...4843. Clauer , C. R. and R. L. McPherrcn, 1974 i Mapp ing the local time-un iversal time development of magnetospheric substorms using mid-latitude magnetic observations. J. Geophys. Res., 79, 281 1. . .282?. Hones, E. W. , Jr., T. A. Fr itr , J. Birn, J. Cooney, and S. J. Bsrrie, 1985: Detailed observations of the plasma sheet during a «ybstorm on April £4, 1979. J. Geophys. Res., submitted. Huuskonen, J. Silen, and T. Turunen, 1985: A special Finnish ETSCAT-exper i ment on Nou. 17/13, 1983, preliminary results. Proceedings of the first Sov iet-Finn ish Auroral UlorKshop , October 1-6, 1984 in Leningrad, USSR < R. J. Pel 1 inert and M. V. 'JspensKy , editors >. The SodanKyla Geophysical Observatory Peport Series, Mo. 44, £7...32. 140 Kamide, Y. and S.-I. AKasofu, 1975: The auroral electrojet and global auroral -features. J. Geophys . Res., 80, 3385... 3601. Kan, J., 1982: Towards a unified -theory o-f discrete auroras. Space Sc i. Rev. , 31, 71...117. KoKubun, S., R. L. McPherron, and C. T. Russel, 1977: Triggering o-f substorms by solar u ind discontinuities. J. Geophys. Res., 82 . 74...as. Opgenoorth , H. J., R. J. Pellinen, H. Maurer, F. Kuppers, Ul. J. HeiKKila, K.. U. Kaila, and P. TansKanen, 1980: Ground-based observations c-f an onset of localized f ie 1 d -al igrted currents during auroral breaKup around magnetic midnight. J. Geophys., 48, 101 ... 115. Opgenoorth, H, J., R. J. Pellinen, UJ. Baumjohann, E. Nielsen, G. MarKlund, and L. Eliassen, 1983: Three-dimensional current floy and particle precipitation in a uestuard travelling surge C observed during the barium GEOS rocKet experiment). J. Geophys. Pes., 88, 3138...3153. Pellinen, R. J. and W. J. HeiKKila, 1978: Energization o-f charged particles to high energies by an induced substorm electric -field uithin the magnetotail. J. Geophys. Res., 83, 1544...1550. Pell iren, R. J., Ul. Baumjohann, W. J. HeiKKila, A. G. Yahnin, G. Marklund, and A. O. MelniKov, 1982: Event study on pre-substorm phases and their relation to the energy coupling betueen solar wind and magnetosphere. Planet. Space Sc i., 30, 371...388. Pellinen, R. J. and W. J. HeiKKila, 1984: Inductive electric fields in the magnetotail and their relation to auroral and substorm phenomena. Space Sc i. Rev., 37, 1...B1. RostoKer , G.. 1933: Triggering of expansive phase intesificat ions of magnetospher ic substorms by northward turnings of the interplanetary magnetic field. J. Geophys. Res., 88. S981...6993. Reux , A., 198": Generation of fie1d-aligned current structures at substorm onsets. Published in Proe. ESA UtorKshop on Future Missions in Solar, Heliospheric and Space Plasma Physics, Garmisch-PartenKirchen, Germany, 30 April-3 May 1985, ESA SP-E35 < June 1985), 151...159. Sauvaud, J, A., J. P. Treilhuo, A. Saint-Marc, J. Oandouras, H. Reme , A. Korth, G. Kremser, G. K. ParKS, A. N. Zaitzev, V. Petrev, L. Lazutin, and R. Pellinen, 1985: Large scale response of the magnetosphere to a southward turning of the IMF. J. Geophys. Pes., submitted. Samson, J. C., 1985: Large-scale studifs cf Pi-2's associated with auroral breaKups. J. Geophys., 56, 133...145. Yahnin, A. G., V. A. Sergeev, R. J. Pellinen, W. Bauirgorinn, K. U. Kaila. H. Ranta, J. Kangas , and 0. M. Raspopov, 1983: Substirm time sequence and micros-tructure on 11. Movember 1976. J. Geophys., 53, 182...197. 141 RECENT DEVELOPMENTS IN PULSATING AURORA STUDIES Ingrid Sandahl Kiruna Geophysical Institute P.O. Box 704, S-981 27 KIRUNA, SWEDEN Abstract The field of pulsating aurora studies is reviewed. The paper begins with a short description of the characteristics of pul sating auroras and the theoretical ideas which, in view of existing experimental results, seem most important. A selec tion of new theoretical results and experimental results from both ground based instruments and instruments on rockets and satellites is then presented. There is now convincing evidence that the luminosity modulation is caused by a modulated flux of electrons. The electron flux modulation seems to arise from a modulated resonant interaction between electrons and whis tler mode waves in the equatorial plane, but the reason for the modulation is not known. Measurements concerning the drift and location of patches and the creation of Pil micropulsa- tions are also discussed. Finally some suggestions for future research work are out lined. Optical measurements, especially with low light level TV, have proven to be of great importance in experimental studies of pulsating auroras. 142 1. Introduction It is almost impossible to study pulsating auroras experi mentally without the support of optical instruments. The opti cal intruments are needed to identify temporal and spatial variations and to provide a good description of the auroral situation. Low light level TV has proven to be particularly useful. The combination of optical data and other kinds of data has been very fruitful and will be so even more in the future. This paper is divided into four parts. The first one is a short description of the appearance of pulsating auroras and their most important characteristics. Then theory will be dis cussed briefly. The theoretical discussion will be restricted to theories which seem to be important in view of recent ex perimental results. In the third part of the paper a selection of important scien tific results obtained during recent years will be de.' •'bed and put into the context of the previously discussed the^ cal suggestions. Finally, some views on the direction of future research will be presented. 2. Description of pulsating auroras An example of a pulsating aurora is shown in Figure 1. Pulsat ing auroras consist of patches that blink on and off with periods of 1-20 seconds. The blinking is often irregular and different patches follow different rythms. The nine photo graphs were chosen to show maxima and minima of a particular patch which was crossed by the Swedish sounding rocket S23-L2. The rocket position is marked with dots. The maximum size of the patch was about 50 km. The photographs were made from a low light level TV-recording by Bengt Holback at Uppsala Iono spheric Observatory. There is no general agreement on how to define pulsating auro ra. Røyrvik and Davis (1976) have suggested the following definition: A pulsating aurora is an aurora which does not exceed 10 kR in + luminosity in the N2 1 NG 427.8 nm band. It undergoes at least one full cycle in which there is first a rapid increase and then a rapid decrease. An additional restriction is that the horizontal motion of the modulated form must be simlar to that of nearby modulated forms. Figure 1 Sequence of photographs of a pulsating aurora over Esrange on January 27, 1979. These photographs were made from a low light level TV-recording by Bengt Holback, Uppsala Ionospheric Observatory. The dots show the position of the sounding rocket S23-L2. From Sandahl, 1984. Ul 144 It should be noted that this definition only requires one cycle. Pulsating auroras are mainly found in the midnight to morning sector in the equatorward part of the auroral oval. This is the location of the region 2 upward field aligned current. Pulsating auroras are on closed field lines. Pulsating auroras are typical of the recovery phase of sub- storms. The onset of pulsations usually takes place 10-15 minutes after the passage of the westward travelling surge. It is a very common phenomenon, indeed one of the most common types of aurora. One of the most important things to remember about pulsating auroras is that they are very variable in appearance;. This is essential when data sets from different events are compared. It is quite possible that different physical mechanisms are responsible in different cases. It is advisable to give a de tailed description of the auroral situation. s 79-01-27 21.58 UT Figure 2 10 Energy spectra obtained in a pulsation maximum (open symbols) and a pul a i sation minimum (filled symbols) by the S23-L2 rocket. Also included is a spectrum measured well outside the loss cone by GEOS-2 in the equatorial plane (line). Note the similarity between the * 105 GEOS-2 spectrum and the S23L2 spectrum in a pul sation maximum, especial ly above 3 keV. From Sandahl, 1984. 10= 6£0S-2 - 17-32° S23-L2 •} 0-40° T = M6» • 70-110° 238 km 104 1} °-«r 1 = 356» • • ENERGY (keV) 145 The pulsation period of patches may vary between 1 and 20 seconds and each patch follows its own individual rythm. The period is sometimes remarkably constant but it may also vary from one pulsation cycle to the next. About half the time there is also an intensity variation of about 3 Hz superim posed on the pulsation maxima. Patch sizes vary from about 10 km to a few hundred kilometers. They can have almost any shape imaginable: arcs, arc segments, irregular forms. Some patches reappear several times without much change in shape but other ones only light up once. There is also a systematic drift of the pulsating patches. This drift is westward in the evening and eastward in the morning. Electron spectra in pulsating auroras look rather undramatic with no large energy gradients. An example taken from the Swedish sounding rocket S23-L2 is shown in Figure 2. The pul sation maximum was characterized by an increase of the elec tron flux in the energy range between 4 and 40 keV. In this case neither the spectrum in the pulsation maximum nor in the pulsation minimum is well described by a Maxwellian distribu- Figure 3 Energy spectra in a pulsation maximum (open symbols) and a pulsa tion minimum (filled symbols) measured by a sounding rocket during the Canadian pulsating aurora campaign. From _*.' McEwen et al., 1981. E0-1.87 E t u (0 10 * 10* r 001 0.1 I 10 100 Energy (keV) 146 tion. On the other hand the spectra measured in a morning side pulsating aurora by McEwen et al. (1981) during the Canadian pulsating aurora campaign are very close to Maxwellian, as can be seen in Figure 3. Also in this case the electron flux in crease occurred in a limited energy range, about 3-20 keV. There exists a very good review about pulsating auroras writ ten by Johnstone (1978). More details about recent develop ments and question marks are given in another paper by the same author (Johnstone, 1983). 3. Theories concerning pulsating auroras. It is not understood today how pulsating auroras are caused, but several suggestions have been given in the literature. The following discussion will be concentrated to those ideas that seem most likely. A few things are fairly well known. Pulsating aurora is a modulation in the auroral emission intensity. It is now quite certain that this light modulation is directly related to modulated electron fluxes. Many experimental results have shown that the electron flux modulation is imposed far out in the magnetosphere in the equatorial plane region. There is now a lot of evidence indicating that during pulsat ing auroras there is resonant interaction between electrons and whistler mode waves. Through this interaction some elec trons alter their pitch angles in such a way that they can reach the atmosphere and be precipitated. That such a process is responsible for the precipitation during pulsating auroras was suggested 15 years ago by Coroniti and Kennel (1970). But at this stage important questionmarks appear. It is not known why the wave-particle interaction is modulated, if the whistler mode waves are coherent, incoherent or both, and how the electrons in the equatorial plane are supplied. The pre cipitation mechanism for the low energy electrons that do not pulsate is not either understood. Let us first examine the problem of how the 1-20 second period modulation is caused. Essentially three mechanisms have been suggested. The first one was proposed by Coroniti and Kennel in 1970. They showed theoretically that micropulsations in the equatorial plane would be able to modulate the growth rate of whistler mode waves and thus also the rate of pitch angle scattering. It is well known from ground based measurements that pulsating aurora and micropulsations of the Pi1 type are correlated. If the micropulsations had been present in the equatorial plane, however, they would have been detected by the GEOS-2 satellite and this was not done. For that reason we may rule this alternative out. The second mechanism is the relaxation oscillator. This is actually a common name for two similar mechanisms, one pro posed by Trefall et al. (1975) and Trefall and Williams (1979) and the other one by Davidson (1979, 1985). These are the sug gestions that seem to meet the least serious contradictions from existing data. The basic principles of the relaxation oscillator mechanism will be explained in section 4.5. 147 The third suggestion is still at a very preliminary stage. Chiu et al. (1983) treated a spatial variation often seen in satellite images of the morning side aurors and suggested that it is caused by striations in the magnetic field intensity and warm plasma density created by the mirror instability driven by inward convecting ions. They suggested that it may be pos sible to extend their theory for spatial variations to explain also temporal variations. Not only the temporal variations with periods of 1-20 seconds in the pulsating auroras call for theoretical explanations. Other important problems are the identification of the elec tron source in the equatorial plane, the factors responsible for the location of the individual pulsating patches and the cause of the 3±1 Hz modulation. It is more or less generally assumed that the electrons are injected in the midnight sector and that they then drift to the morning side, but this has not been proven by measure ments. It is not obvious that this is the whole truth. Back- scattered electrons may also be very important. A promising suggestion concerning the patch location has been given by Oguti (1976). He proposed that the patches appear in regions of increased cold plasma density. Very little theoretical work has been carried out concerning the cause of the 3±1 Hz modulation. There is, in fact, only one published paper. In 1978 Røyrvik presented a model calcu lation using an extremely simplified relaxation oscillator mechanism. 4. Recent results In this section a selection of recent results will be pres ented and put into the context of the theory review in section 3. 4.1 Agreement between auroral light and electrons. Auroral pulsations are directly correlated to precipitating electrons. Whenever a comparison has been possible this result has been obtained. In Figure 4 data from the Canadian pulsat ing aurora campaign is shown. The photometer recording exhib its extremely regular pulsations. Each luminosity maximum cor responds to a maximum in the electron energy flux and number flux measured by a sounding rocket. The rocket was launched in such a direction that the upleg approximately followed a mag netic field line and the photometer viewing direction was along the same field line. A comparison between auroral TV-data and electron measurements on a sounding rocket is shown in Figure 5. This example was taken from the Swedish sounding rocket S23-L2. The top panel represents the auroral light intensity on the same magnetic field line as the rocket. The middle panel shows the electron energy flux and the agreement is obvious. During this event the pulsations were very irregular. 148 Altitude (Km) „a 100 120 140 156 140 120 IQO ,nH> n 1 r 1 1 1 1 i n 10 \ 9 # 10 AA^^UV] 10 — v a E w ioai- 4fo" 600 4278 N2 cr 200- 1145 1146 I 1148 Time (UT) Figure 4 Comparison of the electron energy and number flux measured by a sounding rocket with the auroral emission intensity during a pulsating aurora. From McEwen et al., 1981. The good agreement between light and electrons also exists for the 3±1 Hz modulation. This is seen in Figure 6 which shows a comparison between the flux of >60 keV electrons measured by the sounding rocket S23-L2 and the auroral light recorded by auroral TV. A video integrator was used to obtain the light in the patch where the rocket was measuring. In this figure the time runs from right to left. 4.2 Source location in the equatorial plane The modulation of the electron flux seems to originate in a region close to the equatorial plane. This has been known for many years. All measurements obtained in proper pulsating auroras give this result. It has been found that pulsation maxima in electrons of higher energies appear earlier than the maxima in electrons of lower energies. If it is assumed that the pulsation maxima are imposed simultaneously the difference in arrival time gives the distance to the source. 149 ROCKET POSITION RELATIVE TO PATCH MIOOLE EDGE MIN i mill KT* - ENERG» FLUX ELECTRONS >0.» k«V (J ni' «') Iff* ELECTRONS >*» MV (count») 100- 21.5220 2153 21.54 21.55 21.56 21.57 21.58 TIME (UTI Figure 5 Comparison of auroral light from a low light level TV recording and integrated electron energy flux obtained by the sounding rocket S23-L2. From Sandahl, 1984. More recently a dispersion study of the same kind has shown that also the source for the 3±1 Hz modulation is located in the equatorial plane. Figure 7 shows electron fluxes at three different energies measured by a British sounding rocket by Lepine et al. (1980). During this flight there were oscilla tions of about 2.2 Hz in the flux of electrons. 4.3 Waves in the equatorial plane Data from the geostationary GEOS-2 satellite have proven very valuable in the study of pulsating auroras. An example of whistler mode waves measured in the equatorial plane during a pulsating aurora is shown in Figure 8. This is data from the S-300 experiment provided by Paul Gough, University of Sussex.There was a pulsation in the wave intensity betwen 21.57.08 and 21.57.14 and the wave intensity was modulated with a frequency of about 3 Hz. This is clearly seen in the panel showing the gain for the wave magnetic field measure ments. Similar pulsations occurred before and after the one shown. 150 CUCIMNSHtWr t r M W*».... ^~ - •••• •* • • I • '^ Li-J- VICCO INTEGRATOR OUTPUT 215340 215330 uf ELECTRONS HO VIDEO IMTEORATOR OUIPU1 J"*^W!*rf^^ _J 1 1 L- -1 I 1_ Figure 6 Comparison between flux of >60 keV electrons meas ured by a sounding rocket and the auroral light intensity obtained from a video recording. The correlation between electrons and light exists also for the 3±1 Hz modulation. From Sandahl, 1984. 4.4 Resonant wave-particle interaction Thus* there are examples of pulsating aurora in the auroral oval and whistler mode waves with the same type of temporal variation in the equatorial plane. Can a connection between the two phenomena be proven? Figure 9 shows the configuration of the S23 (Substorm-GEOS) rocket experiment. GEOS-2 was located in the equatorial plane on a magnetic field line which had its foot point very close to Esrange where three sounding rockets were launched. One of themf S23-L2, went into a pulsating aurora approximately at the time of magnetic midnight. Electron energy spectra obtained during this flight were shown in Figure 2. This figure also contains a spectrum measured simultaneously in the equatorial plane. This spectrum is very 151 Interval A Interval B .0.01 I 020- 0.025- 225 226 227 22ST242243 244 245 Flight time (s) Figure 7 Electron fluxes at three different energies measured by a British sounding rocket. Oscillations with a frequency of 2.2 Hz appear at all energy levels. The difference in arrival time show that the source of the oscillations was located in the equatorial plane. From Lepine et al., 1980. similar to the spectrum obtained in the ionosphere in a pulsa tion maximum. The pulsation maximum was characterized by an increase of the flux of electrons between 4 and 40 keV. The wave spectrum measured in the equatorial plane is shown in Figure 10. The peak at 0.7 kHz is of the right frequency for resonance with the precipitating electrons in a pulsation maximum. The details of this study are given in Sandahl (1984), and it strongly indicates that the interaction between electrons and whistler mode waves indeed took place. 4.5 The relaxation oscillator The principle of the relaxation oscillator mechanism for pro duction of auroral pulsations is explained in Figure 11, which was taken from a paper by Davidson (1965). In the lower right hand corner a pitch angle distribution with an empty loss cone is shown. The loss cone is empty because of precipitation. The existence of a loss cone leads to growth of whistler mode waves. These waves interact with the electrons and this inter action changes the pitch angle of the electrons. Some elec trons move into the loss cone as is shown in the upper pitch angle distributions. 152 21.57.09 10 11 12 13 14 TIME (UT) Figure 8 Data from the S-300 experiment on GEOS-2 showing the waves in the equatorial plane. Courtesy Paul Gough, University of Sussex. Then the driving force for the wave generation disappears. But after a while the particles in the loss cone reach the iono sphere and are lost through precipitation. The process can start again. There are objections to this theory, but the basic ideas are promising. 4.6 Cause of micropulsations As Pi1 micropuj.^ations and pulsating auroras are correlated it is essential to find out which one is causing the other. The solution to this problem is one of the important achievements of the recent years. Figure 12 shows an exmple of the relation between optical pulsations measured by a photometer and the variation in the magnetic x-component. Peaks in the two traces occur almost simultaneously. Work by Arnoldy et al. (1982), Engebretson et al. (1983,1984), Oguti et al. (1984), Oguti and Hayashi (1984) and Åsheim and Aarsnes (1984) has shown that the Pi1 micropulsations are secondary effects caused by the precipitating electrons. 153 .PARTICLE TRAJECTORY Figure 3 Configuration of the Substorm-GEOS experiment. The S23 rocket payload and the GEOS-2 satellite per formed measurements on approximately the same geo magnetic field line. 4.7 Location of patches The drift direction and velocity of pulsating aurora patches are shown in Figure 13. The drift was westward in the evening and eastward in the morning. Such observations led to the sug gestion that the drift is equal to the drift of the cold plas ma and that the patches are in fact located in regions of in creased cold plasma density (Oguti, 1976). A more precise com parison was carried out by Scourfield et al. (1983). They com pared the drift velocity of auroral forms to the electron drift velocity derived from STARE-data. An example of their results is shown in Figure 14. The general agreement in direc tion and magnitude was good, at least in the beginning of the event. The definite confirmation of whether the cold plasma density inside the patches is larger than outside must be obtained by direct measurements. Figure 15 shows small scale electron den sity enhancements in the morning sector measured by the Chatanika incoherent scatter radar by Senior et al. (1982). The authors have not stated if there was pulsating aurora during this event, but it seems very likely that this was the 154 fe FREQUENCY (kHz) Figure 10 The wave spectrum in the equatorial plane contained a peak at 0.7 KHz. This agrtes well with the resonant frequency for electrons producing a pulsa tion maximum. Courtesy Paul Gough, University of Sussex. STRONG DIFFUSION, RAPID PARTICLE LOSS Figure 11 Explanation of the re laxation oscillator WAVE GROWTH, mechanism for LOSS CONE FILLS production of auroral pul sations. From Davidson, 1985. 155 „«S_ >*{>t 13jM J^S-,0 PHOT 5 0-29- MM, • I | K „,„ -y—p B M1 06L, -r~r -I OS 0A- -lOÆ S"11»! . | ' Y-AXIS-' -0.7» 0J8cJW,/#*V*V- ;V^#I --0.25 g i --0.8 IJK 13*1 U.T., JUNE 13, 1980 Figure 12 Correlated magnetic field and photometer data. From Engebretson et al., 1983. NORTHWARD ' • SOUlHWARO •,ff m / /a MCSIWtW kVv^M^. TT-7V DRIFTS OF PULSATING AURORAL PATCHES * MAC ImoNrc." ' J L j Li I | | I 04 05 06 07 08 09 11 12 13 UT FEB 16 1980 Figure 13 Drift direction and drift velocity of pulsating auroral patches. From Oguti, 1981. 156 SCALE S STARE ELECTRON Figure 14 DRIFT VELOCITY NORTH.0G Average drift veloci A AURORAL FORM ties of auroral forms 200% DRIFT VELOCITY and average STARE 031112-0315 21 03 22«-IB2550 plasma drift veloci NO STARE data NO STARE data ties. From Scourfield Ibockscatttr abuntl Ibackicatttr absent) et al., 1983. Auroral lor» drift Auroral for* drift velocity «150S velocity* ISO^t 033007-03 31 Ot 03 3027-g; 3101 03 3132-03 3211 x- 03tQ3S-«3tOH 03H22-O3WH O3l9fc2-03S0t1 S /^* /-» Z~< 035im-03S21t 03S5S«-0356St 035» 32-035939 „ S , S 23 JANUARY I9tl 1231 TO 1242 UT Figure 15 Latitudinal varia tions of ionization measured by the Chatanika radar and field aligned cur rents measured by the Triad satellite. It is suggested that the small scale electron density enhancements were connected to pulsating patches. From Senior et al., 82 «3 M 65 W 87 INVARIANT LATITUDE 1982. 157 case. At the time of the measurements the Triad satellite passed overhead showing that each one of the electron density enhancements was connected to a localized field aligned cur rent. The regions of increased plasma density thus seem to exist, but it remains to explain how they are created. 5. Some suggestions for future studies In this section a few points of importance for future research concerning pulsating auroras will be brought up. A complete catalogue of unanswered questions is considered outside the scope of this report. Whenever pulsating auroras are studied optical instruments should be in operation. It is necessary to be able to separate between temporal and spatial variations and find similarities and dissimilarities between different events. Low light level TV has proven particularly useful. In coordinated studies it is important to have common timing. This seems self-evident but is not quite trival in practice. A problem of great interest to the auroral particle measuring community is the development of a method for ground based monitoring of the precipitating particle distribution. It is likely that the shape of the electron spectra depends on for example local time and type of event. The poor understanding of the 557.7 nm emission makes the intensity ratio 557.7/427.8 less suitable for this purpose. A promising instrument is the ionosonde (Mc Dougall et al., 1981). Another thing, which should be studied, is whether the size and temporal development of the region with pulsating aurora really agrees with an injection of keV electrons in the mid night sector and subsequent drift. A study of this could be carried out using an array of low light level TV systems, and since many scientific groups now own such systems this project seems relatively straightforward. More sounding rocket launches are important. The Swedish sounding rocket program contains one pulsating aurora project called Aureld-VIP-II. In this project cooperation with the ground based experiment community is welcomed. 158 References Arnoldy, R.L., K. Dragoon, L.J. Cahill, Jr., S.B. Hende and T.J. Rosenberg, Detailed Correlations of Magnetic Field and Riometer Observations at L=4.2 with Pulsating Aurora, J. Geophys. Res. 87, 10449-10456, 1982. Chiu, y.T., M. Schulz, J.F. Fennell, and A.M. Kishi, Mirror instability and the origin of morningside auroral structure, J. Geophys. Res., 88, 4041-4054, 1983. Coroniti, F.V., and C.F. Kennel, Electron Precipitation Pulsations, J. Geophys. Res., 75, 1279-1289, 1970. Davidson, G.T., Self-Modulated VLF Wave-Electron Interactions in the Magnetosphere: A Cause of Auroral Pulsations, J. Geophys. Res., 84, 6517-6523, 1979. Davidson, G.T., Pitch angle diffusion in the morningside aurorae: 2. The formation of repetitive auroral pulsations, Preprint, Lockheed Palo Alto Research Laboratories, 1985. Sngebretson, M.J., L.J. Cahill, Jr., R.L. Arnoldy, S.B. Mende, and T.J. Rosenberg, Correlated irregular magnetic pulsa tions and optical emissions observed at Siple Station, Antarctica, J. Geophys. Res., j38, 4841-4852, 1983. Engebretson, M.J., L.J. Cahill, Jr., T.A. Potemra, L.J. Zanetti, R.L. Arnoldy, S.B. Mende, and T.J. Rosenberg, On the relationship between morning sector irregular magnetic pulsations and field aligned currents, J. Geophys. Res., 89_, 1602-1612, 1984. Johnstone, A.D., Pulsating aurora, Nature, 274, 119-126, 1978. Johnstone, A.D., The mechanism of pulsating aurora, Annales Geophys i cae, J_, 397-410, 1983. Lepine, D.R., D.A. Bryant, and D.S. Hall, A 2.2 Hz modulation of auroral electrons imposed at the geomagnetic equator, Nature, 286, 496-471, 1980. Mac Dougall, J.W., J.A. Koehler, J. Hofstee, and D.J. Mc Ewen, Ionosonde observations of pulsating auroras, Can. J. Phys., ^9, 1049-1055, 1981. Mc Ewen D.J., E. Yee, B.A. Whalen and A.W. Yau, Electron ener gy measurements in pulsating auroras, Can. J. Phys., 59, 1106-1115, 1981. Oguti, T., Recurrent Auroral Patterns, J. Geophys. Res. jy\_, 1782, 1976. Oguti, T., TV observations of auroral arcs, in Physics of au roral arc formation, Geophys. Monogr., A.G.U. 2!5_, 31-41, 1981. Oguti, T., and K. Hayashi, Multiple correlation between auro ral and magnetic pulsations 2. Determination of electric currents and electric fields around a pulsating auroral patch, J. Geophys. Res., _89, 7467-7481, 1984. Oguti, T., J.H. Meek, and K. Hayashi, Multiple correlation be tween auroral and magnetic pulsations, J. Geophys. Res., 8£, 2295-2303, 1984. Røyrvik, 0., Instabilities in pitch angle diffusion and their possible relation to a 3 Hz modulation in pulsating aurora, J. Atmos. Terr. Phys. 40, 1309-1321, 1978. Røyrvik, 0., and T.N. Dav7s\ Pulsating aurora: Local and glob al morphology, J. Geophys. Res., B£, 4720-4740, 1977. 159 Sandahl, I., Pitch angle scattering and particle precipitation in a pulsating aurora - an experimental study, KGI report 185, October 1984. Scourfield, M.W.J., J.G. Keys, E. Nielsen, and C.K. Goertz, Evidence for the ExB drift of pulsating auroras, J. Geo phys. Res. 8j», 7983-7988, 1983. Senior, C, R.M. Robinson, and T.A. Potemra, Relationship be tween field-aligned currents, diffuse auroral precipitation and the westward electrojet in the early morning sector, J. Geophys. Res. 87, 10469-10477, 1982. Trefall, H., S. UllåTand, J. Stadsnes, I. Singstad, T. Pytte, K. Branstad, J. Bjordal, R.H. Karas, R.R. Brown and J. Munch, Morphology and fine time structure of an early- morning electron precipitation event, J. Atmos. Terr. Phys., il_, 83-105, 1975. Trefall, H., and D.J. Williams, Time Structure of Postmidnight Energetic Electron Precipitation and the Limit of Stable Trapping, J. Geophys. Res. jS£, 2725-2735, 1979. Åsheim, S., and K. Aarsnes, Optical and magnetic pulsations, Planet. Space Sci., 32, 735-744, 1984. 160 AURORAL AND CONCURRENT GEOMAGNETIC PULSATIONS Takasi Oguti Geophysical Research Laboratory, University of Tokyo, Tokyo 113, Japan ABSTRACT Studies of the relationship between auroral and concurrent geomagnetic pulsations, PiC, in the auroral zone are reviewed concentrating on the most recent developments. The conclusion is reached that magnetic pul sations observed on the ground(PiC) are caused by temporal and spatial enhancemnents of the electric conductivity produced by pulsating precipi tation of auroral electrons in the ionosphere. The relationship between auroral and concurrent magnetic pulsations has now been firmly established, not only for the magnetic pulsations on the ground, but also for those observed by a low altitude satellite above the ionosphere. 1. INTRODUCTION It has long been known that pulsating aurora in the morning hours is usually observed at the same time as irregular geomagnetic pulsations, PiC, occur. The study of the relationship between auroral and concurrent magnetic pulsations began as early as 1943(Vestine, 1943). Since then, Luth broad spectrum auroral luminosity pulsations and rioraeter fluctu ations have been compared with geomagnetic pulsations(Campbell and Rees, 1961; Campbell and Leinbach, 1961; Campbell, 1970; Heacock and Hunsucker, 1977). Wave forms of magnetic pulsations observed on the ground have been compared with luminosity variations of concurrent auroral pulsations measured by photometers(Victor, 1965; Paulson et al., 1967; Campbell, 1970; Arnoldy et al., 1982). X-ray pulsations and concurrent geomag netic pulsations(McPherron et al., 1968; Cahill et al., 1981), as well as magnetic impulses associated with riometer impulses(Reid, 1976; Arnoldy et al., 1982; Engebretson et al., 1983; Lanzerotti and Rosenberg, 1983), magnetic impulses associated with VLF chorus bursts(Kokubun et al., 1981), VLF chorus risers associated with impulsive brightening of a pulsating auroral patch(Tsuruda et al., 1981), and pulsating hiss emissions related to auroral pulsations(Ward et al.,1982; Scourfield et 161 al., 1984) have also been studied. Sometimes an excellent relationship was observed, with peak to peak correspondence, but at other times there was no relationship. The question of why the relationship is so erratic has long been a basic problem. Many studies(Reid, 1976; Wilhelm et al., 1977; Lanzerotti et al., 1978; Rosenberg et al., 1981; Arnoldy et al., 1982; Engebretson et al; 1983) show similarity between the wave forms of magnetic pulsations and riometer fluctuations. The results strongly indicate that the magnetic pulsations PiC are of ionospheric origin. Probable ionosheric electric currents responsible for the magnetic pulsations have also been discussed (e.g. Campbell and Matsushita, 1962; Oguti and Watanabe, 1976; Reid, 1976; Wilhelm et al., 1977; Chao and Heacock, 1980; Arnoldy et al., 1982; Åsheim and Aarsnes, 1984). All of these studies show relationships between magnetic pulsations and electron precipitation fluctuations. However, the spatial coverage of the photometric, riometric, X-ray and VLF equipments were usually quite different and inadequate for observing pulsating auroras. As Davis (1966) pointed out, pulsating aurora, when observed by a low-light-level TV camera, consists of many patches which pulsate independently of each other. The horizontal size of these individual patches varies from a few km to a hundred km across. This naturally leads us to question the value of photometric observations of a fixed area of the sky for the study of pulsations, even when multiple photometers are used. Magnetic variations at the ground result from a spatial integration of the fluctuations of all the ionospheric electric currents in the range(Johnstone, 1978). Auroral pulsations seen in a fixed location of the sky must not neces sarily be "representative" but, perhaps, come from just one of several independently pulsating patches. A poor relationship is therefore to be expected unless the photometer field happens to cover a representative patch. A riometer, as well as an X-ray detector (VLF wave receiver as well), usually covers a fairly large area of the sky. Both give spatially integrated values and there is no way to recognize the difference in temporal variations of the precipitation of auroral electrons in various parts of the sky. Therefore, such measurements are not very efficient for determining the relationship between the electron precipitation and 162 the magnetic deflections on the ground. In relation to this geometric observational problem, a suggestive fact has been that the spectral features of magnetic, X-ray and auroral pul sations are usually quite similar even when the correlation is poor(e.g. Victor, 1965; Paulson and Shepherd, 1966; Paulson et al., 1967; McPherron et al., 1968; Campbell, 1970). This indicates that the average period of such pulsations is nearly the same, but the phase as well as the detailed wave form, of the auroral pulsations varies in various parts of the sky. More accurately speaking, since the auroral pulsations are a sequence of impulses(Yamamoto, 1983) with varying repetition time and spatial extent, it is very likely that they show a broad spectral peak, and that the magnetic and the X-ray pulsations, spatially integrated as they are, also give rise to a similar broad spectrum of variations in the same frequency range. Thus, one of the difficulties in reaching a definite conclusion on the relationship has been that a pulsating aurora usually consists of many individually pulsating patches, while the magnetic variation is a spatially integrated effect of all the electric currents around the auroral patches within the range of the magnetic station. Studies of the comparison of wave forms of auroral and magnetic fluctuations have thus naturally been limited to a correlation only in simple cases(Victor, 1965; Paulson et al., 1967; Campbell, 1970; Oguti and Watanabe, 1976; Reid, 1976). A comparison of the wave forms, however, could include an ambiguity, because the wave forms of the magnetic variation components H, D and Z are usually different from each other, so that the observed relation ships, as well as the lead-lag time between auroral and magnetic pul sations largely depends on the magnetic component used for deriving the correlation. The results would be meaningless unless the auroral con ditions are specified, especially the distribution and migration of the pulsating patches relative to the magnetic station. Thus, the general understanding by 1983, was that the ionospheric origin of the magnetc pulsations, PiC, was highly possible because of the relationship between the wave forms of riometric and magnetic pulsations, and due to a small time lag of the magnetic variations behind the auroral pulsations in simple auroral cases(e.g. Campbell, 1970; Oguti and 163 Watanabe, 1976; Arnoldy et al., 1982). Other support for this view also came from radar measurements of ionospheric parameters(Leinonen et al., 1983) and the relationship between ionospheric parameters and magnetic measurements made by a satellite(Senior et al., 1982). Senior et al. (1982) found that a small scale field-aligned current is related to a local enhancement of electron density in the ionosphere, and Leinonen et al. (1983) showed that PiC magnetic pulsations on the ground appear to be related to conductivity modulations, as well as small fluctuations in the ionospheric electric field. Definite evidence of this view, however, was still lacking at that time. This was mostly due to the difficulties in observationally determining the electric currents related to pulsating auroral patches and estimating their magnetic effects on the ground. Therefore, there remained five problems to be solved on the relation ship between auroral and magnetic pulsations at that phase of the study: (1) How could the possiblity that PiC is due to conductivity fluctu ations be confirmed? (2) How could the temporal relationship between auroral and magnetic pulsations be understood, including the difference in the waveform of magnetic H, D and Z components? (3) How are the electric currents responsible for the magnetic pul sations on the ground? (4) Why is the relationship between auroral and magnetic pulsations so erratic, sometimes excellent and sometimes not good at all? (5) What are the mechanisms leading to "one-sided" impulses of magnetic pulsations, and the polarization of the magnetic pulsations, PiC ? In this paper, a review of the solutions of these problems will be given, leading to the conclusion that magnetic pulsations, PiC, are due to conductivity fluctuations in the ionosphere, produced by the pulsating precipitation of auroral electrons. In addition, it will be shown that the magnetic fluctuations above a pulsating aurora measured by satellite are also accounted for by electric currents associated with pulsating enhancements of the ionospheric conductivity. 2. POSSIBLE MECHANISMS Two models for explaining pulsating precipitation of electrons, causing pulsating auroral displays, have been proposed(Coroniti and Kennel, 1970; 164 Davidson, 1979). In the first, the quasi-periodic nature of the pul sating precipitation is attributed to whistler mode wave turbulence, which, modulated by compressional hydromagnetic(hm) waves, scatters energetic electrons into the pitch angle loss cone near the magneto- spheric equatorial plane. In the second, this is accounted for in terms of a relaxation oscillation of the turbulence system itself, involving whistler mode waves and energetic electrons without the cooperation of any hydromagnetic waves. Assuming that the pulsating precipitation of electrons occurs in either model, there are four likely mechanisms which may explain the correlation between auroral and magnetic pulsations. In Coroniti and Kennel's model, the hm waves coupled with the shear Alfven mode and propagating downward after modulating the precipitation flux, cause magnetic pulsations on reaching the ionosphere. The second is that the magnetic pulsation recorded at the ground is mostly due to fluctuations in the ionospheric electric currents caused by temporal and spatial changes in the electric conductivity produced by the pulsating precipitation of auroral electrons. If the modulated precipitating electrons carry a current detectable on the ground, this may be a third source, and if an electric potential depression occurs, associated with the precipitating electrons, this could cause a fluctuating local Hall current as a fourth mechanism. All four mechanisms would lead to a good correlation between auroral and magnetic pulsations. In any case, the magnetic variations in the ground must be due to electric currents in and around the ionosphere, because the ionosphere is the source of the currents whose effects are observed as magnetic dis turbances on the ground. The electric currents, j, in the ionosphere can be expressed as Ji= ( I )?i J„» <«i»Ji. (1) where ( :•: ) and 6 represent the conductivity tensor and the electric field, respectively. If small fluctuations in the conductivity and in the electric field are considered as perturbations ft and 6E, a linear approximation leads to the relationship 165 «Ji- ( E 0)6^w + (^ O^p + (°^0 6j„= div^Æji (2) where I Q and Eg are stationary components of the conductivity and the electric field. &%, and denote electric field fluctuations due to wave arrival and to electron precipitation, respectively. If the hm wave is an essential part of the magnetic fluctuations on the ground, the first term of eq.(2), with fiEy as a wave field, must dominate. However, if the currents due to the conductivity fluctuations are important for the magnetic variations,the third term must prevail. Electric currents carried by the precipitating electrons and the subsequent Hall currents can be expressed by the second term. Each term of eq.(2) gives rise to a different relationship between auroral and magnetic pulsations. Both types of pulsations in the dawn auroral zone are so irregular that they are unlikely to tp a manifes tation of standing oscillations of a magnetic flux tube. They are more likely due to transient local magnetic effects, even if they are related to magnetic variations near the raagnetospheric equatorial plane. Thus, if the first mechanism(hm wave effect) is the dominant cause of the correlation between auroral and magnetic pulsations(that is, the first term of eq.(2) dominates), we can expect the magnetic variations to follow the auroral pulsations by some tens of seconds. This is also the propagation time of hra waves, Tnm, from the magnetospheric equatorial plane to the auroral ionosphere after having scattered the auroral electrons into the loss cone. If this is the case, it is difficult to say whether the time derivative of the magnetic variations or the mag netic deflections themselves whould be best related to the auroral pul sations, because the coupling between the compressional hm waves and the shear hm waves which propagates down to the auroral ionosphere could be highly variable depending on the occasion. On the other hand, if the second mechanism (conductivity effect) operates, the third term must show a better correlation between the time derivative of the magnetic de flections and the auroral pulsations with only a small time difference( depression) operate, the second term must result in a good correlation between magnetic deflections(not time derivative) and auroral pulsations, again with a small time difference. Thus, the crucial point of the problem was: Are the temporal variations in the magnetic field directly related to the auroral pulsations, or is the time derivative of the magnetic variations best related to the auroral pulsations? 3. TEMPORAL RELATIONSHIP BETWEEN AURORAL AND MAGNETIC PULSATIONS In order to reach a definite conclusion on the temporal relationship between auroral and magnetic pulsations, we must be careful in measuring the time difference between them. If a single patch pulsates, then the observationally determined time difference between the auroral and the magnetic variation will be meaningful. However, if there is magnetic "contamination" from any other patches which is not involved in the field of the auroral observation, the time difference could be meaningless. Since a pulsating aurora usually consists of multiple, individually pulsating patches, the possible effect of this should not be overlooked. In a typical situation, many pulsating patches appear at the same time, and we need to divide the magnetic variations observed at a certain station into contributions from the individual auroral patches. Further more, since there are differences between the wave forms of the magnetic components, it is also essential to determine the temporal relationship between the auroral and the magnetic pulsations for H, D and Z sepa rately, and its dependence on the relative locations of the respective auroral patches and the point where the magnetic variations are measured. Oguti et al.(1984) applied a multiple correlation method to study this problem. They divided the sky into 35 domains and the auroral luminosity in each domain was spatially integrated. Then, the temporal variations in the spatially integrated auroral luminosity from all the 35 domains were linearly related to the magnetic variations, H and D components, obtained at a single station. To deterriine the possible lead-lag time, T, between the auroral and the magnetic pulsations, multiple correlations at each time step of T were compared for T varying from -120 to +120 seconds. The multiple corre lations thus examined always showed a maximum value at T=0 (within +1 167 060600-08 HOP UT FEB 16 'BO PS 1 11500-112S00 UT FEB IS '80 LR w^^wh^ x^vvifVA A/wVWH/M/WV /VNVvAAfY^V WV^W^/YY™ ^w^i\j^\^\fy\^ .2 .2 HPC LEAD HUG LUG -?20 -610 0 IrTsEC 120 -V (B) (E) Fig.l An example of time-shifted multiple correlation between mag netic pulsationsttime derivative) and an ensemble of auroral pulsations. The correlation coefficient always has a maximum value at Ofl second time shlfts(Oguti et al., 1984). Similar examination between magnetic de- flections(non time derivative) and auroral pulsations always shews a much lower correlation coefficient. second). It was also found that the multiple correlation coefficient for the time derivative of the magnetic variations and the auroral luminosity pulsations is much larger than that for the non-time derivative magnetic variations and the auroral variations. These results, although already noted in earlier studies using simple auroral pulsations(e.g. Campbell, 1970; Arnoldy et al., 1982), were definitely confirmed here in a more general case where multiple auroral patches pulsated individually. It will be shown in the next section that the analysis here shows no ambi guity on on the magnetic "contaminaton", since it could break down the magnetic variations at a single station into contributions from every relevant auroral domain. An example of the time(T) shifted multiple correlation between the ensemble of auroral pulsations and time dnr' tive of the magnetic vari ations is reproduced in Figure 1. It is „.i.dent from this figure that both the H and D components are related to the auroral variations at T=0 within +1 second. How well the regression model represents the obser vations of the magnetic variations is shown in Figure 2, also reproduced from Oguti et al.(198A). 168 080600-081400 UT FEB 16 '80 P5IT4I 111500-112300 UT FEB 15 '80 LR iW^yVW***»!* 120 240 360 SEC 400 120 240 360 SEC 480 RURORRL RND MAGNETIC VRRIRTIONS RURDRRL RND HRGNETIC VRRIRTIONS CB) CE) Fig.2 Two examples shoving the validity of the regression model in which magnetic pulsations are expressed by a linear combination of lumi nosity variations in various parts of the sky(0guti et al., 1984). The calculated magnetic variations coincide well with the observations. The small lead-lag time( vational fact that the time derivative has a much higher correlation with auroral variations. In relation to this, Oguti and Hayashi(1984) showed that the auroral luminosity should be compared with the time derivative of the electric conductivity variations, and subsequently with the time derivative of the magnetic deflections, vhen mechanise 2 operates under the condition that the temporal variation is comparable to, or shorter than the relaxation time of the ionization in the ionosphere. Here, this was the case(see Figure 3). At this phase of the study we reached the conclusion that mechanism 2 only, that is the currents induced in a local enhancement of conductivity produce magnetic pulsations on the ground, operates by giving rise to a good correlation between auroral and concurrent magnetic pulsations. 4. ELECTRIC CURRENTS AROUND A PULSATING AURORAL PATCH Then next problem was to investigate the electric currents around a pulsating auroral patch. Could electric currents around a pulsating auroral patch be observationally determined? 0 5 10 15 SEC 20 0 2 4 6 a SEC 10 Fig.3 Temporal variation of ion density in the ionosphere produced by pulsating precipitation of auroral electrons(OgUti and Hayashi, 1984). This is proportional to the conductivity, and, hence, approximately to the electric current and the resulting magnetic deflection. This pro vides a reason why a magnetic pulsation below a pulsating aurora is often impulsive. 170 Oguti and Hayashi(1984) did 0806-0B14 UT FEB 16 '80 PS this by use of the regression i ' Si.5 .£" +!?-»-li" -f." coefficients in the equations /' which relate the auroral 2 2 1! 1? 21 27 luminosity in the respective f — 3 B IS IB ZJ 10 domains of the sky to the magnetic variations (time J» derivative) below the pul sating aurora. When the coef- ss are ficients /multiplied by the M '* LONGITUDE IKBC EXC1 average fluctuations in the auroral luminosity in each HRLL/PEOEHSEN CONDUCTIVITY RSTIO - 2 CONDUCTIVITY ENHANCEMENT FACTOR * I.S domain, this gives the ex FIELD-PUGNEO OISCHPRGE ROTE » 55 X pected magnetic def •.:::. iors, ncrthward and eastward re spectively, at the magnetic station when the respective domain brightens. Thus, the current contribution to the magnetic variation from auroral brightening in each domain could be estimated separately. Taking the sys tematic distribution of the Fig.4 The top panel shows an example of expected magnetic deflections observatlonally determined distribution of into consideration, the horizontal magnetic deflection vectors below a distribution on the ground of pulsating patch when it brightens(Oguti and the magnetic deflections Kayashi, 1984). The eletric field was esti below a pulsating auroral mated to be southward for the time by the patch as it brightens was drift of auroral patches. Note that there is determined on the basis of a pair of convergence and divergence of mag netic horizontal deflections. The bottom observations of auroral and panel indicates the sane distribution of the concurrent magnetic pul vectors calculated on the basis of the induced sations. currents as shown in Figure 5. The agreement An example of the magnetic between observation and calculation is deflections below a pulsating excellent. 171 auroral patch is reproduced in Figure 4 from the results of Oguti and Hayashi(1984). The result indicated that the magnetic deflection on the ground belov a pulsating auroral patch is characterized by a pair of divergence and convergence of the horizontal deflection, both 120 km apart from the center of the brightened auroral domain and, in addition the directions of the convergence and the divergence are relatively fixed in the direction of the general drift of the pulsating patches. Postulating that the drift of the auroral patch is due to an ambient electric field, it was also shown that two electric currents, a field- aligned pair current and a twin-vortex current as schematically illus trated in Figure 5, both theoretically expected to be induced in and around a local enhancement of conductivity, fully account for the mag netic deflection on the ground. Thus, the electric currents in and around a pulsating auroral patch were observationally determined. Although Leinonen et al.(1983) noted that the magnetic pulsations, PiC, could partially be due to the observed fluctuation in an electric field, in addition to the conductivity fluctuation, the result here indicated that the fluctuation in the electric field is most likely due to the secondary effect brought forth by the local fluctuation in conductivity. The result also clarified why the relationship between auroral and mag netic pulsations is so erratic, sometimes excellent and sometimes not good at all. This became clear when two facts were taken into consider ation. The first is the poor spatial coverage of the past auroral obser vations wnen some times the "repre sentative" pul sating patch was in the field of the photometer, while at other FIELD-BL1GNED TVIN-V0RTEX tiroes the signifi PBIR CURRENT CURRENT cant part of the Fig.5 Two electric currents, a field-aligned pair pulsating aurora current and a twin-vortex current, both of which are was not. The expected to be induced in a local enhancement of conduc other is that the tivity, account for the magnetic deflection on the magnetic de- ground shown In Figure 4(0guti and Hayashi, 1984). 172 flection when the horizontal vector alone is examined, is not related at all to the auroral pulsations when the magnetic station happens to be located near the convergence and the divergence areas of the horizontal deflection vector. 5. WAVE FORMS AND POLARIZATION OF MAGNETIC PULSATIONS BELOW A PULSATING AURORA The next problem was the wave form called "one-sided" pulse by Engebretson et al.(1983), or magnetic impulse by Kokubun et al.(1981) and by Lanzerotti and Rosenberg (1983). The magnetic pulsations below a pulsating aurora often show an irregular "one-sided" pulse nature. How does this characteristic wave form occur? As already mentioned in a previous paper (Oguti and Hayashi, 1984), one of the reasons for this is seen in Figure 3, which shows the temporal variation of ion density(conductivity, and electric current). When im pulsive ionization(precipitation and auroral luminosity enhancement) occurs, the ion density quickly responds to the increase, while the decay is much slower due to the relaxation time(e.g. Reid, 1976) which usually is longer than the repetition time of the impulsive precipitation. The wave form of the corresponding magnetic deflection must be approxiraately proportional to that of the time derivative of the ion density variation in Figure 3. The pulsating precipitation of electrons causing a pulsating aurora occurs in a fairly broad range of energy. Temporal variations of the modulation of the precipitation depend on the energy, i.e., the earlier arrival of the higher energy component(e.g. Bryant et al., 1969; Smith et al., 1980; McEwen et al., 1981). We can expect that the currents, mentioned in the previous section, horizontally rotate during the cycle of a pulsation causing a rotation of the magnetic deflection vector. Changes in the energy of the precipitating electrons result in changes in the altitude of the ionization and, therefore in the Hall/Pedersen conductivity ratio during a cycle of pulsation. The rotation of the mag netic deflection vector gives rise to a phase difference between the H and D components causing a certain polarization of the magnetic vari- ations(0guti and Hayashi, 1984). Another cause of the specific wave form and polarization of the mag- 173 netic pulsations, PiC, was also shown by Oguti and HayashK 1985). This was due to streaming(longitudinal migration of brightening along an auroral structure) and propagation(lateral migration of an elongated structure) of an auroral patch. Any migration of an auroral patch indi cates migration of the ionization agents(Oguti and Watanabe, 1976). Therefore, a migrating auroral patch is followed by a tail of increased ionization due to the relaxation time of the ion density in the iono sphere. The currents induced in the ionization tail also migrate following the patch migration, causing a rotational variation of the magnetic deflection on the ground. Fea is isse tei RONGE Using three examples of such a dH/dt obi pulsating(migrating) aurora which consisted of a small number of patches, and the magnetic vari ations obtained from 3 stations below, Oguti and Hayashi(1985) showed that the specific wave form, "one-sided pulse" below the aurora, as well as the polari zation of the magnetic pulsations observed not only directly below but also remotely from the aurora, 1120 11TI I 122 Ul were fully explained by the Fig. 6 An example of comparison concept of migrating electric between observed and calculated mag currents. An example of obser netic variations below a streaming vations and calculations of mag (longitudinally migrating) aurora. The netic pulsation is shown in Figure calculation is on the basis of electric 6, reproduced from Oguti and currents Induced in an enhancement of Hayashi(1985). conductivity due to a migrating ioni There was little room to doubt zation agent that corresponds to a migrating auroral patch(Ogutl and that the magnetic pulsations, PiC, Hajashi, 1985). The agreement between were produced by local and observation and calculation is temporal enhancements of conduc excellent. Note that a specific tivity due to the pulsating pre impulsive wave form is reproduced by cipitation of electrons at this the calculation. stage of the study. Strictly 174 speaking, however, this is not provable by ground magnetic variations alone(Tamao, 1964; Fukushima, 1969), but in principle can only be verified by magnetic measurements above the ionosphere. Thus, for a final conclusion on this we needed proof of the presence of a field- aligned pair current above a pulsating auroral patch using satellite data. This proof will be given in the next section. 6. FIELD-ALIGNED PAIR CURRENTS MEASURED BY MAGSAT Field-aligned currents measured by TRIAD have been related to local enhancements of conductivity measured by the Chatanika radar(Senior, et al., 1982). The location of the region 2 field-aligned current was related to the occurrence of PiC pulsations at Siple(Engebretson et al., 1984). These results indicate a close connection between local enhance ments of conductivity and field-aligned currents, and between field- aligned currents and PiC pulsations. However, the physical connection between precipitation of auroral electrons, enhancement of conductivity, electric currents, and PiC EBHS. NORm magnetic pulsations was still not known at that time. In order to find the physi cal relationship, Oguti et al.(1985) thoroughly examined magnetic data obtained by the MAGSAT satellite over a pul sating aurora which occurred over Steen River(geogr. lat. 59.7°N, long.117.2°W; cor rected geomag. lat. 66.6°, long. 293.7°) and Rabbit Lake (geogr. lat. 58.2°, long. Fig.7 All sky picture of aurora at 103.7°; corrected geomag. lat. 13:50:00 UT on January 23, 1980, obtained 68.1°, long. 311.9°) for the at Steen River, Canada. The line indicates time period of 13:49 to 13:51 the trajectory of the magnetic footpoint UT on January 23, 1980. The of MAGSAT at the auroral level(100 km). aurora was recorded by a low- The outermost circle represents 5° in light-level TV camera at Steen elevation (Oguti et al., 1985). 175 River, and the magnetic pulsations on the ground were observed at Rabbit Lake. An example of the all-sky image of the aurora at 13:50:00 UT is repro duced in Figure 7. Here the projection is on the basis of the magnetic field model MGST 6/80 (Langel et al., 1980). The maximum elevation of the magnetic footpoint of MAGSAT at the auroral level(100 km) as seen from Steen River was approximately 15.5° o.^r the eastern horizon. The magnetic footpoint of MAGSAT traversed the puisating aurora from the north to the south about 360 km east of Steen River for the time 13:49 to 13:51 UT approximately along the 301° geomagnetic meridian. The aurora at that time consisted of four large clusters of pulsating patches highly elongated in the east-west direction, and two of the patches had significant luminosity below the MAGSAT trajectory. The aurora, which could affect the magnetic field at MAGSAT, was thus approximated by tw. onal structures within the electron precipitation region. The luminosity variations of the aurora are shown in a latitude- time display in Figure 8-a, under the approximation that the patches were uniform in the east-west direction. The slant line in the figure indicates the trajectory of the magnetic footpoint of MAGSAT at the auroral level (100 km). Therefore, the profile along the slant line represents luminosity vari I34WO 1H000 UilOO UT ations, including contri Fig.8-a Temporal variations in the lumi butions both from temporal nosity profile of aurora. The abscissa is variations and from spatial tine, the ordinate is location along the mag structures. netic footpoint of MAGSAT(geomagnetic latitude approximately along the 301° geomagnetic Next, a similar latitude- meridian line) at the auroral level(100 km), time distribution of the and the darkness indicates the auroral lumi electric conductivity was nosity. The slant line shows- the MAGSAT estimated by relating the trajectory at the auroral level projected auroral luminosity vari- along the magnetic field(Oguti et al.. 1985). 176 ations to the ion pair pro duction rate by 2 dn/dt = kj - aeffn •••(3) where k, J and aeff repre sent a proportionality MMH&tti constant, auroral luminosity, and an effective recombi nation coefficient, re spectively. The resulting distribution of conductivity 1 — T is reproduced in Figure mm nsooo .JSIOO IFT Fig.8-b Temporal variations in the conduc 8-b in the same format as tivity profile calculated on the basis of the Figure 8-a. auroral luminosity in Fig. 8-a(0guti et al,. Then, it was shown that 1985). the westward(eastward) mag netic variations above the ionosphere were proportional to the conductivity enhance ments at the magnetic foot- j/oint if the ionospheric part of the field-algned pair currents had a south- ward(northward) component t3*W0 115000 IJJIOO UT (under the assumption that Fig.S-c Temporal variations of a profile the structure was zonal in of magnetic east-west deflection at MAGSAT the east-west direction). level expected to be due to the preclpitaticn Therefore, the temporal- current. This is obtained by integration of spatial distribution of the auroral luminosity in Figure 8-a along the conductivity in Figure 8-b, vertical(geomagnetic meridian) line(Oguti et was shown to be read as the al., 1985). temporal-spatial distribution of the magnetic E-W deflections expected to be produced by the field-aligned pair currents due to the conductivity variations. The possible contribution of a current carried by the precipitating i'icctrons(called precipitation current hereafter), which is not detect able on the ground but could be significant above the ionosphere, was 177 also estimated by as MAGNETIC E-H COMPONENT AT MAGSAT suming that the auroral luminosity was pro 18 r>T portional to the precipi / X tation current and that the precipitation region was zonal. It was then shown that the luminosity MAG DEF DUE TO CONDUCT IV I TV when integrated along the meridian line, should be proportional to the east- west component of the magnetic deflection pro duced by the precipi tation current. The _L latitude-time distri 13*926 13-49-46 l^SØOØ UT bution of the east-west Fig.9 Temporal-spatial variations of the mag component is reproduced netic field observed by MAGSAT(flrst trace), in Figure 8-c. The auroral luminosity(second trace), expected mag netic field to be produced by the field-aligned profile along the slant pair currents due to the conductivity enhancement line in the figure indi (third trace), expected magnetic fluctuation due cates the variations of to the precipitation current(fourth trace). The the east-west magnetic observed magnetic fluctuations are evidently deflection produced by related with those expected from the conductivity the precipitation current enhancements(Oguti et al., 1985). at the instantaneous location of MAGSAT(again including both temporal and spatial contributions). These profiles of the auroral luminosity and the ionospheric conduc tivity along the footpoints of MAGSAT, and the possible magnetic effects of the precipitation current are reproduced in Figure 9 along with the magnetic variations observed by MAGSAT. Here, only the fluctuation components, from 3 to 20 seconds, are shown. Note that the magnetic deflections expected fron the field-aligned pair currents associated with the conductivity enhancements in the third trace are identical to those of the conductivity but are plotted as the conductivity increases down ward for convenience. It is evident that the westward magnetic de- 178 flections at MAGSAT(downward in trace 1) corre An spond to conduc W tivity enhance r ~»-\ / ments below(down- / ' w^-^ ward in trace 3), V which is equiva w-y -\^. lent to the west ward magnetic de 19BB flection expected from the conduc tivity-induced Fig.10 Auroral luminosity variations and con field-aligned pair current magnetic pulsations observed at Rabbit Lake currents. below a pulsating aurora. Enhai cement of the lumi Finally, by nosity is found to be related to the southward(and a using a normalize'! little eastward) deflection of the ground magnetic multiple re field(08Uti et al,. 1985). gression analysis between the E-W magnetic variations observed at MAGSAT and those estimated from the field-aligned pair currents(traces 1 and 3 of Figure 9), and between the magnetic variations at MAGSAT and those estimated from the precipitation current(traces 1 and 4), it was shown that more than 80 % of the observed magnetic variations at MAGSAT were due to the field-aligned pair currents, and about 20 % could be attri buted to the precipitation currents. Thus, the presence of the field- aligned pair currents induced in a local enhancement of conductivity contributed the most to the E-W component in the MAGSAT data. This also made clear the relationship between small scale field-aligned currents and the local enhancements of electron density in the ionosphere, found by Senior et al.(1982) using TRIAD magnetic data and the Chatanica radar data. It was also shown that the concurrent magnetic deflection on the ground measured at Rabbit Lake below the pulsating aurora, as reproduced in Figure 10, was approximately southward as the auroral luminosity rose, and that the double amplitude of the pulsations was 1-1.5 nT during that time In contrast to the much larger westward magnetic deflection at 179 MAGSAT amounting to 10 nT. 7. A MODEL AURORA AND ELECTRIC CURRENTS Oguti et al.(1985) further confirmed the validity of the amplitude and direction of the magnetic variations as measured at MAGSAT and on the ground by a model calculation. A local enhancement of the conductivity below MAGSAT was approximated by an elliptic domain(0guti and Hayashi, 1985). The ionosphere was assumed to be a thin sheet with height integrated Pedersen and Hall conductivities. The ambient magnetic field was assumed to be vertical, and the ambient electric fieM was assumed to be two-dimensional and uniform within and above the ionosphere. The space charge which accumu lated at the boundary, due to an enhancement of conductivity in the elliptic domain, was postulated to be released partially through the ionosphere by Pedersen conductivity and partially by field-aligned currents towards the magnetosphere. An example of the distribution of the magnetic field due to the field- aligned pair currents and the twin-vortex currents induced in the local MOfflZO«l.ii NtMMCtIC tcriCCriON Ml MMO«MI LCVCL MC»|*9MrML «M-IMCTIC BCrt-CC f IOM Otl fHC «HOUND HUIMM» ILICWK 'Ull ffMICM «IIt-M«ha» IMh*. MNiL'KNHfh.1.1 COMMJCTIVIIV KHMMMCSMCMI-I.2 Fig.11 An eiaaple of calculated aagnetlc deflections produced by electric currents Induced In a unlfora enhancement of conductivity In an elliptic doaaln. Left panel shows horizontal «agnetlc deflections at MAGSAT altitude (380 kai), and right panel shows that on the ground. Sepa ration of grid points la 100 ka(0gutl et sl., 1985). 180 enhancements of conductivity at the MAGSAT level and on the ground, is reproduced in Figure 11. Here, the electric field was assumed to be southward. The east-west and north-south extents of the domain were assumed to be 800 km and 100 km, respectively, in accordance with the observation. The enhancement in conductivity was assumed to be 20 percent, and the Hall/Pedersen ratio was assumed to be 1.5. The result clearly showed that the magnetic deflection on the ground below the domain was approximately southward, and above the domain it was approximately westward, which is consistent with the observations. Thus, the westward deflection at MAGSAT is mostly due to the field-aligned pair currents and the southward deflection on the ground mainly comes from the westward component of the ionospheric part of the pair currents. The westward deflection at MAGSAT is about 10 times larger than the southward deflection on the ground, which is also consistent with the observations. Thus, we are now able to conclude that the magnetic field fluctuations, both above and below a pulsating aurora, are due to the field-aligned pair currents and the twin-vortex currents, induced in a local enhance ment of conductivity produced by the pulsating precipitation of auroral electrons, wjth a possible small(less than 20 %) contribution from the precipitation current above the ionosphere. CONCLUSION The observed relationship between auroral and magnetic pulsations(PiC) at the ground, which was first introduced by Vestine (1943) and further developed by Campbell and Rees (1961) and by Campbell and Leinbach (1961) a quarter century ago, is now shown to be due to conductivity fluctu ations in the ionosphere which are directly related to the auroral particle precipitation. Conductivity-induced currents account for not only ground magnetic pulsations(PiC) but also concurrent magnetic fluctu ations observed at the same time by a low-altitude satellite above the ionosphere. The result provides us not only with a better understanding of auroral ionospheric processes, but also offers a solid basis for studying the magnetospherlc processes leading to the. pulsating precipitation of electrons which cause pulsating aurora. 181 ACKNOWLEDGEMENTS The author thanks all the other campaign members, National Research Council of Canada and the Institute of Space and Atmospheric Studies, University of Saskatchewan. Thanks are also due to NASA and the Japanese MAGSAT team (Chairman: Dr. N. Fukushima) who made available MAGSAT data and to Drs. T. Tamao, S. Kokubun, K. Hayashi and T. Yamamoto, Geophysical Research Laboratory, University of Tokyo, for their valuable discussions on the auroral and magnetic pulsations. REFERENCES Arnoldy, R. L., K. Dragoon, L. J. Cahill, Jr., S. B. Mende, T. J. Rosenberg and L. J. Lanzerotti, Detailed correlations of magnetic field and riometer observations at L=4.2 with pulsating aurora, J. Geophys. Res., 87, 10449, 1982. Åsheim, S. and K. Aarsnes , Optical and magnetic pulsations, Planet. Space Sci., 32, 735, 1984. Bryant, D. A., G. M. Courtier and A. D. Johnstone, Modulation of auroral electrons at large distances from the Earth, J. Atraos. Terr. Phys., 31., 579, 1969. Cahill, L. J. Jr., R. I.. Arnoldy, M. J. Engebretson and T. J. Rosenberg, Irregular magnetic pulsations observed at Siple, Antarctica, Eos Trans. AGU, 62, 1014, 1981. Campbell, W. H., Rapid auroral luminosity fluctuations and geomagnetic field pulsations, J. Geophys. Res., 75, 6182, 1970. Campbell, V. H. and H. Leinbach, Ionospheric absorption at times of auroral and magnetic pulsations, J. Geophys. Res., 66, 25, 1961. Campbell, W. H. and M. H. Rees, A study of auroral coruscations, J. Geophys. Res., 66, 41, 1961. Campbell, W. H. and S. Matsushita, Auroral-zone geomagnetic micropul- satlons with periods of 5 to 30 seconds, J. Geophys. Res., 67^ 555, 1962. Chao, J. K. and R. R. Heacock, Modulation of type PI waves by temporal variations in ionospheric conductivity In a three-dimensional magneto- sphere-ionoaphere current system. Planet. Space Sci., 28, 475, 1980. Coronltl, F. V. and C. F. Kennel, Electron precipitation pulsations, J. Geophya. Rea., 75, 1279, 1970. 182 Davidson, G. T., Self-modulated VLF wave-electron interactions in the magnetosphere: a cause of auroral pulsations, J. Geophys. Res., 84_, 6517, 1979. Davis, T, N., The application of image orthicon technique to auroral observation, Space Sci. Rev,, j>, 222, 1966. Engebretson, M. J., L. J. Cahill, Jr., R. L. Arnoldy, S. B. Mende and T. J. Rosenberg, Correlated irregular magnetic pulsations and optical emissions observed at Siple Station, Antarctica, J. Gephys. Res., 813, 4841, 1983. Engebretson, M. J., L. J. Cahill, Jr., T. A. Potemra, L. J. Zanetti, R. L. Arnoldy, S. B. Mende and T. J. Rosenberg, On the relationship between morning sector irregular magnetic pulsations and field aligned current, J. Geophys. Res., 89, 1602, 1984. Fukushima, N., Equivalence in ground geomagnetic effect of Chapman- Vesine's and Birkeland-Alfvens's electric current systems for polar magnetic storms, Rept. lonosph. Space Res. Jpn, 23_, 219, 1969. Heacock, R. R. and R. D. Hunsucker, A study of concurrent magnetic field and particle precipitation pulsations, 0.005 to 0.5 Hz, recorded near College, Alaska, J. Atmos. Terr. Phys., 39, 487, 1977. Johnstone, A. D., Pulsating aurora, Nature, 274. 119-126. 1978. Kokubun, S., K. Hayashi, T. Oguti, K. Tsuruda, S. Machida, T. Kitamura, 0. Saka and T. Watanabe, Correlations between very low frequency chorus bursts and impulsive magnetic variations at L -'4.5, Can. J. Phys. 59, 1034, 1981. Langel, R. A., R. H. Estes, G. D. Head, E. B. Fabino and E. R. Lancaster, Initial geoaagnetic field model from MAGSAT vector data, Geophys. Res. Lett., 2. 793, 1980. Lanzerotti, L. J., C. G. Kaclennan and C. Evans, Association of ULF mag netic variations and changes in ionospheric conductivity during sub- storms, J. Geophys. Res., 83, 2525, 1978. Lanzerotti, L. J. and T. J. Rosenberg, Impulsive particle precipitation and concurrent magnetic field changes observed in conjugate areas near L-4, J. Geophys. Res., 88, 9115, 1983. Leinonen, J., J. Kangas, A. V. Kustov, G. S. Stepanov and M. V. Upsensky, PIC magnetic pulsations and variations of the Ionospheric electric field and conductivity, J. Ataos. Terr. Phys., 45, 579, 1983. 183 McEwen, D. J., E. Yee, B. A. Whalen and A. W Yau, Electron energy measurements in pulsating auroras, Can. J. Phys., j>9_, 1106, 1981. McPherron, R. L., G. K. Parks, F. V. Coroniti and S. H. Ward, Studies of the magnetic substorms, 2. Correlated magnetic micropulsations and electron precipitation occurring during auroral substorms, J. Geophys. Res., 73, 1697, 1968. Oguti, T. and T. Watanabe, Quasi-periodic poleward propagation of on-off switching aurora and associated magnetic pulsations in the dawn, J. Atmos. Terr. Phys., 38, 543, 1976. Oguti, T., J. H. Meek and K. Hayashi, Multiple correlation between auroral and magnetic pulsations, J. Geophys. Res., 89, 2295, 1984. Oguti, T. and K. Hayashi, Multiple correlation between auroral and mag netic pulsations, 2 - Determination of electric currents and electric fields around a pulsating auroral patch, J. Geophys. Res., J39, 7467, 1984. Oguti, T. and K. Hayashi, Polarization and wave form of magnetic pul sations below pulsating auroras: magnetic effects of electric currents induced in an ionization tail of a moving auroral patch, J. Geomg. Geoelectr. 3_7, 65, 1985. Oguti, T., T. Yamamoto, K. Hayashi and R. Fujii, Proof of ionospheric origin of PiC pulsation - magnetic pulsation as observed by MAGSAT above and on the ground below a pulsating aurora, Proc. Symp. on prospect and retospect in the study of geomagnetic field pertur bations, January, Tokyo, pp 180-195, 1985. Paulson, K. V. and G. G. Shepherd, Short lived brightness oscillations in active aurora, Can. J. Phys., 44, 921, 1966. Paulson, K. V., G. G. Shepherd and P. Graystone, A note on "auroral type" fluctuations in the Earth's electromagnetic field. Can. J. Phys., 45_, 2813, 1967. Reid, J. S., An ionospheric origin for Pi 1 micropulsations, Planet. Space Sci., 24, 705, 1976. Rosenberg, T. J., J. C. Siren, D. L. Matthews, K. Marthinsen, J. A. Holtet, A. Egeland, D. L. Carpenter and R. A. Helliwell, Conjugacy of electron microbursts and VLF chorus, J. Geophys. Res., 86, 5819, 1981. Senior, C, R. M. Robinson and T. A. Potemra, Kelatioship between FAC.'i, diffuse auroral precipitation and the westward electrojet In the early 184 morning sector, J. Geophys. Res., 8_7, 10469, 1982. Scourfield, M. W. J., J. P. S. Kash and . D. Duthie, Auroral pulsations - televison image and VLF hiss correlation, Planet. Space Sci., J32^ 809, 1984. Smith, M. J., D. A. Bryant and T. Edwards, Pulsations in auroral electrons and positive ions, J. Atmos. Terr.Phys., 42_, 167, 1980. Tamao, T., The structure of three-dimensional hydromagnetic waves in a uniform cold plasma, J. Geomag. Geoelectr., 14, 89, 1964. Tsuruda, K., S. Machida, T. Oguti, S. Kokubun, K. Hayashi, T. Kitamura, 0. daka and T. Watanabe, Correlations between the very low frequency chorus and pulsating aurora observed by low-light-level television at L = 4.4, Can. J. Phys., 59, 1042, 1981. Ward, I. A., M. Lester and R. W. Thomas, Pulsating hiss, pulsating aurora and micropulsations, J. Atmos. Terr. Phys., 44, 931-938, 1982. Wilhelm, K., J. W. Munch and G. Kremser, Fluctuations of the auroral zone current system and geomagnetic pulsations, J. Geophys. Res., 8_2_, 2705, 1977. Vestine, E. H., Remarkable auroral forms, Meanook Observatory, Polar Year, 1932-1933, Terr. Mag. Atmos. Electr., 48, 233, 1943. Victor, L, J., Correlated auroral and geomagnetic micropulsations in the period range 5 to 40 seconds, J. Geophys. Res., 70, 3123, 1965. Yamamoto, T., 0n-off characteristics of luminosity fluctuations of pul sating auroras, Mem. Nat'l Inst. Polar Res., Tokyo, No.26, 124, 1983. w -r \%C «Z? IO X2. AD gS II 185 Pulsating Aurora and Hydrogen Emissions R. A. Vlereck and H. C. Stenbaek-Nlelsen Geophysical Institute University of Alaska-Fairbanks Fairbanks, Alaska 99775-0800 Abstract Analysis of hydrogen emissions In pulsating aurora Indicates that the periods of the pulsations are correlated with the H(3 Intensity. The pulsation periods are short, a few seconds, at low H(3 Intensity and Increase with increasing Hp emissions. The change In the period appears to be due mainly to a variation In the time between Individual pulses. In most of the data, recorded at Poker Flat Research Range, Alaska (geomagnetic latitude 64.5°), the pulsating aurora was south of the main region of the proton aurora. However, on two nights pulsations were also observed within the proton aurora main region, and these pulsations were of short period and the period varied less with Hg Intensity. 186 Introduction Pulsating aurora occur predominantly In the aornlng sector of the auroral oval. They are coaaonly associated with the post-breakup recovery phase of the auroral substora (Akasofu, 1968; Ooholt, 1971; VaUmce-Jones, 1974), but also during nights with little substora activity, pulsating aurora may be present in the morning sector (Duncan et al., 1981). Visually, they appear as Irregular patches of luminosity turning on and off with periods from a few seconds to a few tens of seconds (R#yrvik and Davis, 1977; Ogutl, \""f>); their brightness Is normally a few kR In the auroral gieen line at 5577Å and the luminosity Is produced by energetic electrons precipitating Into the atmosphere from above. The energy spectrum of the precipitating electrons is nearly Maxwellian, but the characteristic energy varies with reported values between 1 and 12 keV (Whalen et al., 1971; Johnstone, 1971; Bryant et al., 1975; Smith et al. , 1980; Sandahl et al., 1980; HcEwen er. al., 1981; Yau et al., 1981; Armstrong et al., 1985; Davidson et al., 1985). The source of energy for the electrons causing the pulsating aurora is most likely located in the Equatorial region of the magnetosphere. Several theories have been advanced based on wave-particle interactions (Coroniti and Kennel, 1970; Davidson, 1979; Trefall and Williams, 1979; Davidson et al., 1985), but they all have serious shortcomings (Stenbaek~Nielsen, 1980; Lepine et al., 1983; Thomas, 1983; Sandahl, 1984). The theories have only Involved electrons since the main region of the proton precipitation is normally poleward of the pulsating aurora (Montbriand, 1971; Johnstone, 1971; Fukunishl, 1975; Creutzberg et al., 1981) and the proton precipitation, as observed optically from the ground, does not pulsate (Esther, 1967, 1968). However, the work reported here indicates that proton precipitation and pulsating aurora are not mutually exclusive, as hypothesized by Creutzberg 187 et al., 1981, and further that the proton precipitation appears to affect the period of the pulsations. This surprising observation suggests that a successful theory for pulsating aurora must ln sone fashion involve protons as well as electrons. Data and Results The latitudinal distribution of the relatively weak hydrogen emission (the proton aurora) was obtained from the H0 channel of a 5-channel meridian scanning photometer (MSP). The distribution and periods of the pulsating aurora were obtained from all-sky TV data. Both instruments were operated at Poker Flat Research Range (geomagnetic latitude 64.5*N) from November 15, 1982, to February 15, 1983, and from the total data set we selected for analysis all nights for which 1) data coverage was complete for both instru ments, 2) data quality was not compromised by clouds or moonlight, and 3) there occurred at least one auroral breakup followed by pulsations. These selection criteria resulted lu a total of 12 nights of usable data.A more detailed description of the analysis and the data have been given elsewhere by Viereck and Stenbaek-Nielsen (1985). During 10 of the 12 nights analyzed, the pulsating auroras occurred near the equatorward boundary of the proton aurora. The period of the pulsations appears to be positively correlated with the Intensity of the HB emission in the area of the pulsations. Further, the Intensity of the optical emissions in the pulsations appear to decrease with increasing HB intensity, and no pulsations were observed when the Hfl Intensity was greater than 80 R. An example of the data from these 10 nights is January 4, 1983, shown in Figures 1 and 2. Figure 1 displays the KB emissions across the meridian and the presence of pulsating aurora. The expansive phase occurred shortly after 188 0' UT; pulsating auroras appeared at 09:25 UT and a few alnutes later covered the entire sky. The proton aurora was located poleward of the «aIn region of the pulsating aurora. The pulsating auroral patches near the zenith were analyzed for duration of each individual pulse (time-on), time between individual pulses (time-off) and average period of a pulsatloa series. Figure 2 shows the time-off versus time-on (left) and time-off versus corresponding Hfi Intensity (right). The left panel clearly shows that the period is determined largely by the time- off, as has already been noted by Yamamoto (1983). The right panel shows that the time between pulses appears to be correlated with the Hp intensity. Thus, while the duration of the individual pulses vary only little over the night, the periods of the pulsations are determined largely by the time between individual pulses in a series and the time-off appears to be linked to the level of proton precipitation. The data from the other 9 nights are fairly similar. In Figure 3 we have combined that data with the data plotted in Figure 2 to give pulsation period versus Hfj intensity for all 10 nights when the pulsating aurora was observed at the equator edge of the proton aurora. No data points were obtained atove 55 R in Hp. Pulsation were at times present up to 30 R in Hfj, but they were too faint or sporadic for analysis. Charge exchange processes between the protons and the ambient atmosphere will redistribute the hydrogen emission over a fairly large area (Davidson, 1965; Iglesias and Vondrak, 1974). Thus the HØ emissions within the pulsating aurora could be from proton precipitation farther north. We reject this possibility because the Latitudinal distribution of the H(3 emissions does not appear consistent with such a model. 189 During the ren*lnIng two nights» the pulsating auroras were observed within the mein region of the proton aurora. The data for one of these nights are shown in Figures 4 and 5. Figure U shows the distribution of Hf3 emissions and presence of pulsating aurora in the meridian,and Figure 5 shows the corresponding distribution of pulsation periods versus H0 intensity. The pulsation periods were relatively short for the amount of hydrogen emissions present, and the periods vary only little with Hp Intensity. Discussion In the presentation of our data and results, the two nights during which pulsating aurora was observed Imbedded in the main region of the proton aurora appear to form a separate group. Our data set is not very large, so we do not know how typical t^e.'je two nights are; to our knowledge, similar observations have not been reported previously, which may indicate they are relatively rare. In any case, because of their different morphology and insensitlvlty to proton precipitation, we raise the possibility that these pulsating auroras form a distinct subset of pulsating aurora, indicating a somewhat different mechanism. Clearly, more observational data would be desirable. The main body of the data, obtained in pulsating aurora near the equator- ward edge of the proton aurora, indicate a relation between proton precipita tion and pulsation period. This Is a surprising result since it has been generally assumed that pulsating auroras involve only electrons, an assumption on which all current theoretical work is based. While we can offer no physical mechanism to explain our observations, our data suggest that protons may play a role in creating pulsating auroras. In particular, the proton precipitation appears to be associated with the time between individual pulses (Figure 2). A possible scenario would be a "leaky-bucket" type process in 190 which the electrons are precipitating out of the flux tub* relatively fait by one process, while the refilling involves another nor* variable process Is which the protons have some role. The plots of period and, In particular, time-off versus HJ3 Intensity (Figures 2 and 3) appear to indicate that there would be no pulsations without proton precipitation. However, due to the instrumental characteristics and analysis technique (Romick, 1976), the origin of the H(3 scale is not well defined, and thus the behavior of the pulsations at low H(3 levels Is uncertain. Observationally, we do note that we see most of the pulsations at very low hydrogen emission levels and that the period of these pulsations tends to be short. It wouio be desirable to be able to convert the observed Hf3 intensity to incident proton flux. This, however, is not easy to do because charge exchange processes between the incident protons and the ambient atmosphere will redistribute the hydrogen emissions to cover a fairly large area (Davidson, 1965; Igleslas and Vondrak, 1974). The charge exchange processes will tend to average out any smaller-scale spatial structure in the incident proton flux. In addition to the spatial averaging, velocity dispersion would tend to eliminate temporal structure In the proton flux If the proton precipitation is caused by processes in the equatorial region of the raagnetosphere. Thus the H(3 intensity observed from the ground is a spatial, and possibly also a temporal, average of the source proton energy flux. Relating to the possibility of velocity dispersion, Smith et al. (1980) reported rocket observations indicating pulsations in positive Ions; the data were difficult to analyze but appeared consistent with a modulation in the equatorial plane. It would be valuable to obtain more rocket and satellite observations, both above the atmosphere and in the equatorial region to see if spatial and temporal structure is present in the precipitating protons. 191 Summary of Observations A large majority of the pulsating auroras analyzed were located near the equatorward boundary of the proton aurora. Our observations on these pulsations may be summarized as follows: 1. At low H8 Intensities the period of the pulsations tend to be short, I.e., a few seconds. As the Hp intensity Increases, the pulsation period becomes longer. 2. At high H8 Intensities the video data indicate the pulsations to become temporarily more irregular and they also appear to become fainter; no pulsations were obvious in the data above 80 R in HB. 3. The change in periods appears to be mainly due to a variation in the time between Individual pulses. On two nights, pulsating auroras were also observed within the main region of the proton aurora. These pulsations tended to be fast and appeared to be less correlated with Hp. Acknowledgements We thank Dra. C. S. Deehr, T. J. Halllnan, J. Olson and G. J. Romi etc for many helpful discussions during this project and D. Osborne and J. Baldridge foT help with the data acquisition and data reduction. This work was supported by Grant ATM-8025621 from the National Science Foundation. 192 References Akasofu, S.-I., Polar and Hagnetospherlc Substorms. D. Reldel Publishing Company, Dordrecht-Holland, 1968, Armstrong, U. C, J. H. Doolittle, S. B. Hende and D. S. Evans, Simultaneous ground-based and satellite observations of a pulsating aurora, (Abstract) EOS, £6_, 1047, 1985. Bryant, D. A., H. J. Smith and G. H. Courtier, Distant modulation of electron intensity during the expansion phase of an auroral substorm, Planet. Space Scl., 23, 867, 1975. Coroniti, F. V. and C. F. Kennel, Electron precipitation pulsations, J. Geophys. Res.. 75, 1279, 1970. Creutzberg, F., R. L. Gattinger, F. R. Harris and A. Vallance Jones, Pulsating aurora in relation to proton and electron auroras, Can. J. Phys., 59, 1124, 1981. Davidson, G. T., Expected spatial distribution of low-energy protons precipi tated in the auroral zone, J. Geophys. Res., 70, 1061, 1965. Davidson, G. T., Self-modulated VLF wave-electron Interactions in the magneto- sphere: A cause of auroral pulsations, J. Geophys. Res., 84, 6517, 1979. Davidson, G. T., H. D. Voss, U. L. Imhof and Y. T. Chiu, Interpretation of large-scale spatial structure and electron spectra in morningside pulsat ing aurora, (Abstract) EOS, _66_, 1047, 1985. Duncan, C. H., F. Creutzberg, R. L. Gattinger, F. R. Harris and A. Vallance Jones, Latitudinal and temporal characteristics of pulsating auroras, Can. J. Phys.. 59, 1063, 1981. Eatner, R. K., Auroral proton precipitation and hydrogen emissions, Reviews of Geophys., J5_, 207, 1967. Eather, R. H., Hydrogen emissions In pulsating auroras, Ann. Geoph., 24, 5235, 1968. Fukunishi, H., Dynamic relationship between proton and electron substorms, J. Geophys. Res.. 80, 553, 1975. Igleslaa, G. E. and R. R. Vondrak, Atmospheric spreading of protons in auroral arcs, J. Ceophys. Res., 79, 280, 1974. Johnstone, A. D., Correlation between electron and proton fluxes In post breakup aurora, J. Geophys. Res., 76, 5259, 1971. Lepine, D. R., D. A. Bryant and D. S. Hall, The control of pulsating auroral electron intensities by plasma waves, Rutherford Appleton Laboratory Report RL-83-111, 1983. McEwen, D. J., C. M. Duncan and R. Kontalbettl, Auroral electron anergies: comparison of In situ measurements with spectrcscopically inferred energies, Can. J. Phya., 59, 1106, 1981. 19J Montrland, L. E., The proton aurora and auroral substorra, In The Radiating Atmosphere, B. H. McCorniac (Ed.) D. Reldel Publ. Co., Dordrecht-Holland, 1971. Oguti, T. S., Recurrent auroral patterns, J. Geophya. Res., 81, 1782, 1976. Omholt, A., The Optical Aurora, Sprlnger-Verlag Heidelberg, 1971. Romick, G. J., The detection and study of the visible spectrum of the aurora and alrglow, SPIE, ^i., 1976. Rtfyrvlk, 0. and T. N. Davis, Pulsating aurora; local and global morphology, J. Geophys. Res., 82, 4720, 1977. Sandahl, I., Pitch angle scattering and particle precipitation in a pulsating aurora - An Experimental Study, Kiruna Geophysical Institute Report 185, 1984. Sandahl, I., L. Eliasson and R. Lundin, Rocket observations of precipitating electrons over a pulsating aurora, Geophys. Res. Lett.. 7, 309, 1980. Smith, M. J., D. A. Bryant and T. Edwards, Pulsations In auroral electrons and positive ions, J. Atro. Terr. Phy., 42, 167, 1980. Stenbaek-Nielsen, H. C., Pulsating aurora: the importance of the ionosphere, Geophys. Res. Lett., _7_> 353, 1980. ThomasF R. W., Electron pitch-angle scattering by low frequency waves at the geomagnetic Equator, Nature, 303, 785, 1983. Trefall, H. and D. J. Williams, Time structure of post midnight energetic electron precipitation and the limit of stable trapping, J. Geophys. Res., 84, 2725, 1979. Vallance-Jones, A., Aurora, Geophysics and Astrophysics Monographs, Vol. 9, D. Reidel Publ.~Co., Dordrecht-Holland, 1974. Viereck, R. A. and H. C. Stenbaek-Nielsen, Pulsating aurora and hydrogen emissions, J. Geophys. Res., 90, 11035, 1985. Whalen, B. A., J. R. Killer and I. B, McDiarmid, Energetic particle measure ments in the pulsating aurora, J. Geophys. Res., 76, 978, 1971. Yamamoto, T., 0n-off characteristics of luminosity fluctuations of pulsating auroras, Proceedings of the fifth symposium on coordinated observations of ionosphere and magnetosphere in polar regions, Memoirs of National Institute of Polar Research, Special Issue, No. 26, pp. 124-134, National Institute of Polar Research, Tokyo, March 1933. Yau, A. W., B. A. Whalen and D. J. McEwen, Rocketborn measurements of particle pulsations in pulsating aurora, J. Geophys. Res., 86, 5673, 1981. co —i o z a> Zr- < 111 O N O 3" O CO 9:00 10:00 11:00 UNIVERSAL TIME (hh:mm) H/J INTENSITY DAY 004-1983 (counts) (Jan.4) 1 Count = 8.7R Figure 1. Hp emission In the magnetic meridian aa function of UT tine. The presence of pulsating auroras was obtained from the all-sky TV data and Its extent in the meridian Is delineated by the solid vhlte line. Note the pulsations are equatorward of the main region of proton precipitation. PULSE OFF TIME (sec) XT c •- B> Ell >- f a a O fit Hi C 5T » ft P> PULSE OFF TIME (sec) 961 a a a 20 a • I ee a a a • a • a a • • • • • aa aa • • • • • • a aa • • • a a * a a • a a • a a aa aaaaaa aa a a a a a • • ea aa a a • aa a a aa • aaaa a a * a aaaa a ea a aaaaa PULSATIO N PERIO D • ea aaaa • • aaaaaaae a u n o •• • • •a • aaaa «a a • • . ; 10 20 30 40 50 Hp INTENSITY (R) Figure 3. Pulsation period versus HS Intensity. The plot contains all data from 10 of the 12 nights analyzed when the pulsations were observed at the equatorward edge of the region with hydrogen emission (the proton aurora). CC —, o U z O LU O N I Z> o —I—i—i—i—i—i—|—i—r T-T-T* rø 11:00 12:00 13:00 UNIVERSAL TIME (hh:mm) 05 10 15 Hø INTENSITY DAY 037-1983 20 (counts) (Feb.6) 25 30 35 1 Count = 4.4R 40 45 Figure *+. Hp intensity in the magnetic meridian as function of time. The pulsating aurora as seen in the all-sky TV data was on this night embedded In the main region of the hydrogen emissions. The periods of the pulsations were low as seen in Figure 5. 10 20 30 40 50 Up INTENSITY (R) Figure 5. February 6, L9S3: Pulsation period versus HB intensity. 199 ALTITUDE DISTRIBUTION OF SOME AIRGLOH FEATURES J J. L6pez-Moreno, A. Molina. M. Ldpez -Puerta s. F. Moreno and R. Rodrigo Instituto de Astroffsica de Andalucia. Apartado 2144, 18080 Granada, Spain. Departamento de Fisica Fundamental, Facultad de Ciencias, Universidad de Granada, Spain. ABSTRACT The measurements of the vertical distribution in the nightglow of some transitions in the infrared region of the OH Meinel system and the 0-0 band of the Infrared Atmospheric System (IAS) of the 02 are presented. We have derived the concentration profiles of the vibrationally excited OH in levels v=2 to v=7 that best fit the observed emission profiles. The results point out that the maxima of the concentration of each vibrational level are not at the same altitude in disagreement with the photochemical models dealing with the OH emissions. The concentration profiles obtained are used to evaluate the reaction rate of the process: Q li + H 0H(v < 4) + 0 + 2( g' that results to be less than 4.x10n cm3 s~l for v = 1 to 4. By using a theoretical atomic oxygen profile, it has been derived the efficiency for the production of 0, ('A ) by the three-body recombination of atomic oxygen that results in a yield of 0.7 of the global value for the reaction rate of 4.7x10"33 (300/TP . INTRODUCTION. In December 19.of 1981 it was flown a four channel infrared photometer from El Arenosillo Sounding Rocket Range. Huelva, Spain (37°5' N, 6° 44' W). The preliminary results (Lopez-Moreno et a I ., 1984) showed a strange profile of the OH Meinel bands that was interpreted as a double layer of vi brat ional ly excited OH. In a later paper (L<5pez -Moreno et al., 1985) this double layer was interpreted as a local contamination due to the interaction of the atmosphere with the payload. and it was obtained the fraction of the total measurement that was due to the contamination. 200 Once substracted this contaminating glow, the remaining can be considered as emission coming from the undisturbed atmosphere and being representative of the vertical distribution of atmospheric OH. The contamination was not seen in the channel centered at 1.27 microns, either because of its narrow half power band with (HPBW) (13.5 nm) in comparison with the HPBw corresponding to the channels devoted to the OH (HPBW'v 100 nm), or because the spectral distribution of the glow in this region was less bright than in longer wavelengths where is the AV=2 sequence of the OH. In table 1 are presented the main characteristics of the channels including the fractional transmissions of the features . Table 1. Main characteristics of the photometers Central Ha 1 f-power Fractional Chan. W. 1 ength bandwidth Band Transmission^) (nm) (nm) 3 1 1267.5 13.5 O.(a'flg) - 0, (X Zq) V = 0 v •' = 0 2 1551 104 OHU^JIi ) - 0H(X2ni ) v'= 2 v" = 0 9 v'= 3 v" = 1 69 v = 4 v" = 2 55 v ' = 5 v" = 3 2 1632 1 10 0H(Xar.i ) -> 0H(X2Ili ) v'= 3 v" = 1 5 v'= 4 v" = 2 61 v' ; 5 v" = 3 63 v ' = 6 v"=4 5 1698 124 0H(X2m ) -> 0H(X2ITi ) v'= 4 v" = 2 10 v' = 5 v" = 3 80 v = 6 v"=4 46 v'= 7 v" •• 5 2 201 VERTICAL DISTRIBUTION OF OH. In figure 1 are shown the altitude profiles obtained in channels 2, 3 and 4, that correspond to emissions from the atmospheric OH. Each channel accepts radiation coming from different transitions in which the upper level varies from 2 to 7, in known proportions. The concentrations of the different states, from v = 2 to 7, can be obtained from the measurements by solving a system of 3 equations, one for each channel, for each altitude, and 6 unknown quantities, one for each OH* upper level. It is necessary, however, to assume some physical constraint. Three additional equations have been used from the ratios between the concentration of the states. So that the states that mainly contribute to each channel are mantained as the unknown quantities. The values of the ratios relating state v' with state v'' have been taken by averaging the values of the ratios published by different authors (Krassovsky et al., 1962; Gush and Buijs, 1964; Evans and Llewellyn, 1973; Harrison and Kendall, 1973; Llewellyn et al., 1978; Battaner and Lopez-Moreno, 1979; and Turnbull and Lowe, 1983). The procedure has been carried out for 6 different sets of ratios and the final concentration in the corresponding altitude has been taken as the average, for each altitude of the 6 corresponding solutions. In figure 2 are represented these solutions for the vibrational levels 2 to 7. This altitude distribution is in some disagreement with the results obtained by modelling the OH emissions in which the maximum of concentration for each level is placed at the same altitude and the profiles have the same shape (e.g., Llewellyn and Solheim, 1978; Battaner and Ldpez -Moreno, 1979). Our results show that the vibrational development is altitude dependent. This induces to think in alternative production and loss mechanisms more dependent on the vibrational levels. VERTICAL DISTRIBUTION OF THE (0-0) BAND OF THE INFRARED ATMOSPHERIC SYSTEM OF 02. The channel 1 measured the profile of the IAS corresponding to the transition: 3 Ojfa'Ag) -» 02(x E~) + hv (1.27 microns). The measurements were taken 6900 seconds after the local sunset at 90 km and, as the radiative lifetime of the'Ag state is very long, ^ 3900 s, (Badger et al., 1965) it is expected some remaii. of the very high concentration of this state during the day. This remain is important at altitudes higher than 80 km where the deactivation by quenching is small and the decay is mainly due to spontaneous emission. 202 To evaluate the fraction of the measured profile that is representative of the diurnal remain we have applied a simple decay model: lA [Ojfa'A^ = E°2 is the concentration of the quencher (0, 02, 03, N2), and É? is the reaction rate for quencher "q". The sub-index 0 represents the time of local sunset, t is time in seconds measured from the local s:msetj and the super-index i means altitude. The initial profile before sunset has been derived theoretically with a single photochemical model, and is in reasonable agreement with other profiles obtained experimentally at twilight (e.g., Llewellyn and Witt, 1977). The result is plotted in figure 3 as the dashed line. In table 2 are shown the mechanisms proposed as the most important related with the IAS. Other processes, as transfer from other excited states of molecular oxygen, have been evaluated in the light of the existing values of fractional populations and rate constants, and have been found to be negligible in the contribution. Table 2. Photochemical scheme N. Reaction Reference 0 + 0 + 0 02('Aq) + 0 Slanger and Black (1976) 0 + 0 + M 02('Ag ) + M Campbell and Gray (1973) 0H(vS4) + 0 - 02('Ag ) + H Llewellyn and Solheim (1978) Clarke and Wayne (1969) 02('Ag ) + 0 - Findlay and Snelling (1971b) 02(>Ag) + 02 * + 0, l 02( Ag) + 03 - 202 + 0 Findlay and Snei I ing (197 1a) Becker et al. (1971) 02('Ag ) + N2 - 02 + N2 0 + hv (1.27 um) Badger et al. (1965) «2'% 2 203 In order to evaluate the main sources of the nocturnal Oj('ig) we have used this scheme together with the altitude distribution of OH* derived from simultaneous measurements. The atomic oxygen concentration profile was taken from the model of Rodrigj et al. (1985) that can be considered as representative for mid-latitude. This theoretical profile presents a maximum of the concentration of 5.76x10 u cm"3 at 96 km. The concentration profiles of the other atmospheric compounds used in the calculations were also taken from the same model. The values of the efficiencies for reactions 2 and 3 have not been measured up to date. For reaction 2 the proposed efficiency varies from 0.05 to more than the unit when the third body (M) is molecular oxygen (Sharp, 1985), but its actual valoe remains still unknown. For reaction 3 there is an estimation of the efficiencies e3 , obtained by Llewellyn and Solheim (1978) which ranges from 1, when OH* is in level 1. to 0 for levels higher than 4. We tried to reproduce the observed O-Cig) profile using this photochemical scheme with a value of 0.25 for the efficiency for reaction 2 and the values proposed by Llewellyn and Solheim (1978) for reaction 3. The resulting theoretical profile (see figure 3) is not very close to that measured. A reduction in the value of the efficiency for reaction 2 down to 0 results in a better agreement in the integrated intensity, but the upper part of the profile (over 87 km) is not reproduced at all, The above results point out that the suggested value for the reaction rate of the process 0 + OH*, at least when the final molecular oxygen is in the'i state, is overestimated. To obtain a good agreement with the experimental profile the values of the efficiencies have to be divided by 10, indicating that a very small fraction of the nocturnal emission of the [AS is due to this reaction. With the values of the efficiencies for reaction 3 divided by 10 we have tried to find a better value of the efficiency for reaction 2 that reproduces the measured profile. A value of 0.7 has been obtained. Figure 4 shows the measured profile, the reproduced profile by using the above mentioned values of the efficiencies, and the contributions of each mechanism to the tota 1 emi ss i on. CONCLUSION. Simultaneous measurement1- of different bands of Meinel System of the atmospheric OH are useful to derive the vertical distribution of the concentration of the individual vibrational states of the OH levels. The obtained distribution shows a variation in the altitude of the maxima 204 of concentration indicating a strong dependence of the mechanisms giving rise to the atmospheric OH* on the vibrational level. l The simultaneous measurements of the 02( A_) and OH Meinel bands profiles provide a method to study the responsible mechanisms for the production of nocturnal 02('Ag). The best fitting to the measured profile is obtained when the rate constant of the reaction 0 + OH* is 10 times less than the value proposed by Llewellyn and Solheim (1978), and the quantum yield of the '&_ state in the three body recombination of atomic oxygen (reaction 2) is 0.7 of the total rate given by Campbell and Gray (1973). In consequence, the three body recombination of atomic oxygen is the main responsible of the nocturnal emission of the Infrared Atmospheric System of the 02. and the oxidation of the vibrationally excited OH in levels less or equal 4 represents a small contribution to the observed emission. REFERENCES Badger, R.M., A.C. Wright, and R.F. Whitlock, Absolute intensities of the discrete and continuous absorption bands of oxygen gas at 1.26 and 1.065 y and the radiative lifetime of the lA„ state of oxygen, J. chem. Phys., 4)3, 4345, 1965. 9 Battaner, E., and J.J. L6pez -Moreno, Time and altitude variations of vibrat iona 1ly excited states of atmospheric hydroxyl, Planet. Space Sci-, 27, 1421, 1979. Becker, K.H., W. Groth, and U. Schurath, The quenching of 1 metastable 02('Ag) and Oof !") molecules, Chem. Phys. Lett., 8, 259, 1971 . y Campbell, I.M., and C.N. Gray, Rate constant for 0(3P) recombination and association with N("S), Chem. Phys. Lett., 18, 607, 1973. Clarke, I.D., and R.P. Wayne, The reaction of 02('Ag) with atomic nitrogen and with atomic oxygen, Chem. Phys. Lett., 3, 405, 1969. Evans, W.F.J., and E.J. Llewellyn, Atomic hydrogen concentrations in the mesosphere and the hydroxyl emissions, J. geophys. Res., 78, 323, 1973. Findlay, F.D., and D.R. Snelling, Temperature dependence of the rate constant for the reaction 02('A„) + 03 ->- 202 + 0, J. chem. Phys., 54, 2750, 1971a. 205 Findlay, F.O., and D.R. Snelling, Co 11 i siona1 deactivation of 02('Ag). J- chem. Phys., 55, 545. 1971b. Gush. H.P. , and H.L. Buijs, Near infrared spectrum of night airglow observed from high altitudes. Can. J. Phys., 42, 1037, 1964. Harrison, A.W., and D.J.W. Kendall, Airglow hydroxyl intensity measurements 0.6 -2.3p , Planet. Space Sc i., 21, 1731. 1973. Krassovsky. V.I., N.N. Shefov, and V.I. Yarin, Atlas of the airglow spectrum 300 A - 12,400 A, Planet. Space Sc i ., 9), 883, 1962. Llewellyn, E.J., 8.H. Long, and B.H. Solheim, The quenching of OH* in the atmosphere, Planet. Space Sci., 26, 525, 1978. Llewellyn, E.O., and B.H. Solheim. The excitation of the infrared atmospheric oxygen bands in the nightglow, Planet. Space Sci., 26, 533, 1978. Llewellyn, E.J., and G. Witt, The measurement of ozone concentrations at high latitude during the twilight. Planet. Space Sci., 25. 165, 1977. L6pez-Moreno, J.J., R. Rodrigo, and S. Vidal, Radiative contamination in rocket-borne infrared photometric measurements, J. geophys. Res., 90, 6617, 1985. Lopez-Moreno, J.J., S. Vidal, R. Rodrigo, and E.J. Llewellyn, Rocket-borne photometric measurements of 02('Ag), green line and OH Meinel bands in the nightglow, Ann. Geophysicae, 2, 61, 1984. Rodrigo, P., J J. L6pez-Moreno, M. Lopez -Puerta s, F. Moreno, and A. Molina, Neutral atmospheric composition between 60 and 220 km: A theoretical model for mid-latitudes, Planet. Space Sci., submitted to, 1985. Sharp, W.E., Upper limits to [01 in the lower thermosphere from airglow, Planet. Space" Sci., 33, 571, 1985. Slanger, T.G., and G. Black. 0('S) production from oxygen atom recombination, J. chem. Phys., $4, 3767, 1976. Turnbull, D.N., and R.P. Lowe, Vibrational population distribution in the hydroxyl night airglow, Can. J. Phys., 61, 244, 1983. CHANNELS t 2.3.4 C. W. "1.55 micron* C. W. • 1.63 microns C. W. - 1.70 micron» 2 4 6 8 VOLUME EMISSION RATE in ^ o in o \'t O in O E O a> a» F i p.. 3. Measured profile of the Infrared Atmospheric System of the 02 at 1.27 up) (solid line), together with the theoretical reproduction hy usinp the effi ciencies specified in the figure. The contrihution of each source to the total profile is also shown. TOTAL OF CONTRIBUTIONS 0 + 0 + M Ef i 0. 70 — — — ODaHy + Remanen0 Eft i 0. 10 4 6 8 10 12 14 16 18 20 22 24 VOLUME EMISSION RATE CkR/km) Tip,. 4. Same as figure 3 but using the value?: of the efficiencies thai best fit the observed profile. 210 CONTAMINATION IN ROCKET-BORNE IR MEASUREMENTS D. Smith and A. Ratkowski, AFGL/LSP, Hanscom AFB, MA 017 31 Contamination in rocket-borne experiments due to debris and outgassing from the booster and from the experiment payload has been a recurrent problem since the inception of measurements with rocket-borne sensors. This is par ticularly true of probe measurements in the LWIR spectral region for which high-sensitivity, cryogenic sensors are required. Contamination control activities to date have been of limited success since they have concentrated primarily upon the reduction of particulates. In this review, we examine the data from fourteen LWIR rocket experiments conducted over the past decade and conclude that gaseous sources, not particulates, were the dominant source of LWIR background contamination observed in these experi ments. Spectral measurements indicate that outgassing of water and organo-silicon compounds are likely sources of LWIR backgrounds. Table I lists the twelve LWIR programs which contribu ted data to this study. Also listed are the type of boo ster, the apogee of the payload, and whether or not the booster was separated from the instrument payload before the IR measurements were performed. Separation from the hot, contaminant-laden booster is clearly desirable; how ever, it does not always guarantee a contamination-free measurement as was demonstrated by an experiment performed by AFGL/LC several years ago (Ref. 1). During this experi ment displacement reels connecting the booster (in this case a Black Brant 5C) to the instrument payload documen ted the fact that, in spite of a nominally normal payload separation, the booster actually overtook the payload approximately 37.3 sec after separation (Figure 1). At 400 sec after separation, the booster was approximately 122 m above the payload. 211 Table II lists the IR sensor sensitivity at 10 microns for the 12 programs considered and the contamination back ground levels at 6, 9.3, and 15 microns. One observes that in most cases the contamination background level» exceeded the sensitivity limit of the sensor, thus preventing the sensor from attaining the measurement sensitivity of which it was capable. This is of some concern to the authors inasmuch as we are planning to launch an LWIR (2-24 microns) probe in the near future, SPIRIT, which has an instrument sensitivity vs. wavelength which is shown on Figure 2. Also shown are those of some other AFGL probe experiments. The reason for our concern is clear; in order that SPIRIT achi eve its full capability we will have to reduce the contami nation background one to two orders of magnitude below those of most earlier flights for which considerable effort was expended in cleaning the payload and in maintaining clean conditions prior to launch. It was this realization which provided the impetus for us to examine the data from earlier flights to try to identify the source(s) of contamination in the hope of devising effective anti-contamination procedures during design, fabrication, integration, and alignment of the payload, as well as during maintenance of the booster and payload on the launch pad. Figure 3 shows two spectra recorded with a CVF spec trometer during AFGL's ICECAP '73 campaign (Ref. 2). In addition to the CO2 emission at 15 microns, a prominent feature was observed at S.3 microns, with weaker emissions at 6.9, 8, and 12 microns. This strong emission was a persi stent spectral feature, as shown by the upper part of Figure 4 in which the average of 34 spectral scans near apogee is plotted, and the lower part of Figure 4, in which the 9.3 micron feature is seen to decrease much less rapidly with time than either the 9.6um Oa peak (seen in the scan near 100 km) or the 15|jm CO2 peak. Note in both Figures 3 and 4, one has the impression of an underlying continuum indicated by the hand-drawn solid curves sketched in to guide the eye. The same spectral region was examined in ICECAP '74 with improved spectral resolution (Ref. 3). Figure 5 shows the 212 prominent feature at 9.3 microns but it is now seen to be a doublet. Subsequently, the HAVE SLED program returned spec tra over essentially the same wavelength region. Figure 6 shows a spectrum from HAVE SLED flight No 2. The prominent feature at 9.3 microns is present, though unresolved, with minor peaks at 6.7, 8, and 12.4 microns. HAVE SLED flight No 3 measured spurious emissions at 6.8/6.9, 8, 9.2/9.3, and 12.3 microns as illustrated by Figure 7. Host recently, the Energy Budget Campaign CVF spectra were found to contain unidentified emissions at 6.8, 8, and 9.2 microns as shown in Figure 8. Even examination of spectra from the AFGL SPIRE experiment, often characterized as one of AFGL's "cleaner" experiments, revealed the presence of a spurious doublet at 9.0/9.6 microns and possible emissions at longer wave lengths. Figure 9. Since the available measurement time in a rocket probe experiment is typically less than ten minutes, we examined the radiance-time histories of several of the spurious emitters in a number of programs. Samples are shown in Figures 10-13. Figure 10 from ICECAP '73 shows the spurious radiance in the wavelength band from 8.5 to 10.9 microns. The radiance rises to a maximum at approximately 100 sec after launch and requires approximately 3 min to fall to 1/e of the maximum value. Figure 11 shows emissions in other wavelength bands from the same flight. Figure 12 shows data from ICECAP '74 in the 8.5-10.9 micron band, and it can be seen to closely mimic the behaviour observed in ICECAP "73, Figure 10. Data from the ZIP experiment, for which consider able care was devoted to cleaning the sensor payload, is shown in Figure 13. This baseline model contains emissions unexplained in terms of possible atmospheric emitters and, although the first measurement was made at 100 sec after launch (thus precluding determination of the early time history of the radiance) , the radiance is seen to also decay with a time to fall to 1/e of the earliest measured value of approximately 3 minutes and to persist for some 8 minutes into the flight. In some programs, the spectral resolution permitted the 213 identification of spurious emissions. An example is shown in Figure 14 from the 1979 Field-Widened Interferometer flight (Ref. 4). In this case the emitter is clearly water at 6.3 microns, presumably outgassing from the payload surfaces. It is most likely also the source of the spurious emitter causing the emissions of Figure 13. From analysis of experiments in which the spectral resolution was lower than in the Field-Widened experiment, as in the ICECAP '73/'74, HAVE SLED I/II, SPIRE, and Energy Budget experiments, we have examined the literature for IR spectra (mostly absorption spectra) for compounds with spectral features at approximately 7,8, and 12 microns with a strong doublet at 9/9.6 microns. One class oi' compounds satisfying this requirement is the organo-silicons, as supported by the data in Figure 15. Figure 16 shows the absorption spectrum from a ubiquitous silicone grease in which all four of the spectral features sought are present. Another possible candidate is a silicon pump oil whose spectrum is shown in Figure 17. Silicone adhesives are also widely used in the aerospace industry and the absorption spectra of trfo different types are shown in Figure 18. A table summarizing the wavelengths of contamination species observed in five rocket-borne IR probe programs is compared with the strong spectral features of potential organo-silicon contaminants in Figure Id. The agreement is striking, even including the presence of a doublet centered at 9.3 microns, and strongly supports the contention that the spurious sources observed in several programs are due to vapor-phase contaminants in the IR sensor line of sight due to outgassing surfaces most probably on the sensor payload proper. While these compounds are generally categorized as having low vapor pressures, their vapor pressures may be elevated due to aerodynamic heating during the boost phase followed by a decrease in outgassing as the payload surfa ces cool or as the material on the surfaces is depleted. The former explanation may explain the notable consistency in the decay time of approximately 3 minutes noted earlier. In divising an anti-contamination procedure for SPIRIT I 214 we have formulated the guidelines listed in Table III. The payload exterior has been redesigned to reduce to a minimum the number of skin ports and will be polished to reduce the possibility of adhesion by potential contaminants. All operations on the payload during buildup and alignment will be performed in a specially-modified Class 100 clean room at Poker Flat Research Range. On the launch pad the payload will he surrounded by an insulated clamshell shroud con structed with a double wall through which heated air will be circulated (Figure 20) and which will contain sleeves for the minimum number of feedthrus: umbilicals and cryogen vent and fill lines. At launch the cla. shell will deploy as shown in Figure 21. Consequently, by st ged venting of the payload on ascent and sealing via baroswitch-controlled vent valves, we hope to reduce the in-flight concerns to (1) contaminants from the hot booster and (2) gaseous contaminants from the payload ACS used to control the mission scan pattern. To address the first problem, we will rely upon a sealed airbag separation system to obtain the maximum 'motor/payload sepa ration velocity consistent with payload safety margins. The second concern is being addressed in a study of design, fabrication, and cleaning of the ACS components, not the least of which is the choice and cleaning of the ACS gas itself. We feel confident that this can be managed by virtue of the experience gained during the DOT program (Ref. 5). As seen from the data in Figure 22, reductions in spurious emissions were observed whenever the ACS cold helium gas thruster firing above and parallel to the IR sensor line of sight was activated, thus providing a momentary purging of the instrument field of view. Because rocket probe measurements have become more complex and sensitive and, consequently, more expensive we have attempted to consolidate and analyze some observations of contamination in rocket-borne IR measurements over the past decade. If we have made a contribution to future pro grams, it is due in large part to the openness with which our colleagues have reported their findings. 215 ACKNOWLEGDEMENT We appreciate the assistance of Prof. C. Brown and Dr. M. Maris of the University of Rhode Island for providing the spectrum of Figure 17. Dr. Maris is currently conducting the study of the ACS referred to in the text and will be the SPIRIT I Contamination Control Manager in the field. REFERENCES 1. E. McKenna (AFGL/LC), private communication. 2. AFGL TR-76-0274. 3. AFGL TR-77-0113. 4. R. Straka (AFGL/LSP), private communication. 5. T.F. Greene (Boeing), private communication. 216 Table I. LWIR EXPERIMENT ROCKET BOOSTER SUMMARY BOOSTER EXPERIMENT ROCKET SEPARATION APOOEE(KM) TIME(SEC|/ALT(KM) ICECAP-CVF<73) BLACK BRANT SC YES 185 KM ICECAP-CVF<74| BLACK BRANT SC YES 200 KM HIRISr.78) SERGEANT YES 126 KM SPIREI77) TALOS-CASTOR YES 285 KM FWM(79} ASTROBEE F YES 158 KM FW1-2(63) SERGEANT 63/67 139 KM EBC-CVF(SO) TAURUS-ORION NO 190 KM EXCEDE-CVF(79) TALOS-CASTOR YES 128 KM HAVE-SLED-CVF(SI) SERGEANT-HVDAC NO 320 KM ZIPI83) ARIES YES 500 KM ELS(76) ARIES YES 500 KM OOT YES Table II. LWIR ROCKET CONTAMINATION SUMMARY SENSITIVITY @ 10|jm CONTAMINATION BACKGROUND (w/cm* sr urn) 6pm 9.3ym 15pm ICECAP-CVF - 2x10-» 3x10"' 5x10' 4x10"" HAVE-SLED-CVF - 2«10-" 1x10-* 3xl0-' 5x10-' SPIRE-CVF 1x10-" IxlO-1 6x10-* 1x10-' EXCEDE-CVF 2x10-" 6x10-' - 3x10-' EBC-CVF - 2xl0"« 1x10-* 1X10"* 7 OOT 2xlO-'° 2x10-" - 5x10" ZIP U10-' 2x10-" 3x10-" 2x10-" ELS - 1x10"" - 1x10-' -- HIRIS 5x10-" <2x10" -=5x10-" <1x10-" FWI-1 4x10-'V) 6x10" -- FWI-2 5x10-"C> >5x10-" - SPIRIT 2x10-" . . - ' AT 5pm 217 o z o 5 E 11<1 20- TIME FROM SEPARATION-SEC Figure i. PAYLOAD/BOOSTER SEPARATION PROFILE FHOM TEST FLIGHT OF LOW VELOCITY SYSTEM SENSITIVITIES OF LWIR ROCKET-BORNE SENSORS ^ "6 OBSERVED CONTAMINATION LIMITS 2 5 11 1* 17 23 WAVELENGTH (um) Figure 2 218 UNIDENTIFIED CMIMIOHS ICECAP 73 LAUNCHED 22 MAR 73 POKt? FLAT, ALASKA IBC-II AURORA CRYO. LWIR OVF SPECTROMETER SCANS-4 50-253 TAL ' 106.6-106.0 SECS IUPLEGI ALT - 129.2-130.7KM 0.0 6 8 10 1Z 14 IS 16 20 ZZ Z4 WAVELENGTH IN MICROMETERS Figure 3. SCANS 259-2» (5/9) 142.9-149.» SECS 1(9.0-171.0 KM 2 ICECAP 74 O 5 LAUNCHED 14 FEB 74 s POKER FLAT, ALASKA X va> OUIET.NIGHT.NO AURORA u CRYO. LWIR CVF SPECTROMETER d 1 s 10 12 14 IS 19 20 Z2 24 WAVELENGTH IN MICROMETERS Figure 5. 219 ICECAP 73 2.5 LAUNCHED 22 MAR 73 POKER FLAT, ALASKA Es 2.0 o e IBC-II AURORA o CRYO. LWIR CVF SPECTROMETER I 1.5 aI (A u 1.0 d (0 co 0.5 0.0 ^T**************^. . 10 11 12 13 14 IS 16 17 se18s 19 20 21 22 2D 2* WAVELENGTH IN MICROMETERS 12 13 1» 15 1G 17 18 WAVELENGTH IN MICROMETERS Figure 4. Altitude Dependence of the Spectral Radiance During Ascent from 100 to 185 km. Spectral scans are averages over 2.5 km increments 220 SPECTRAL SCAN FROM CVF SPECTROMETER FROM BMO HAVE-SLEO I ROCKET PROBE SERIES 1.0C-11 1I.M.4T.M 1.0E-12 1.0E-13 II _L J_ J_ S.O 7.S 10.0 12.5 15.0 17.5 20.0 22.5 Figure 6. WAVELENGTH (MICRON) SPECTRAL SCAN FROM CVF SPECTROMETER FROM BMO HAVE-SLED I ROCKET PROBE SERIES 1.0E-10 TT i 1 r SPECTRUM 2 4.M.M.87 >~ CONTAMINATION BANOS —j 1.0E-11 HAVE-SLED, FLT. 3 2x10 LAUNCHED 12 MAR (1 POKER FLAT, ALASKA - •A i»io s 6 5.0 7,5 10.0 12.5 15.0 17.5 20.0 22.5 Figure 7. WAVELENGTH (MICRON) 221 ENERGY - BUDGET CAMPAIGN inni i ii[ n il 111111111 H 111111111 u 11111111111111111 u 111111111111111 1111111n11111 III IJ CVITF SCAN-237 3 ALTITUD.TITUOE-NE M KM 1 NO CONTAMINANTS co, 10 liiilli lijiiiiimiiiiiiifii.iL i, *...,—.t A 1 * * • • • • 4 G 8 10 12 14 16 18 20 22 24 Figure 8. WAVELENGTH-JJM SPECTRAL SCAN FROM CVF SPECTROMETER ON ElO VERTICAL PROBE EXPERIMENT LAUNCHED FROM ANDOYA, NORWAY 16NOV 80 ASPART OF ENERGY BUDGET CAMPAIGN HIGH ALTITUOE SPECTRAL SCAN FROM AFGL SPIRE EXPERIMENT-SEPT 1177 FROM POKER FUT RES. RANGE, ALASKA CVF SPECTROMETER- 3% RESOLUTION ALTITUDE - 245KM [UNIDENTIFIED EMISSIONS! « «-7 10 ti 111111 ii 111111111111111111111 111111111 ||i 1111 ili 111111111111111111111 7 9 11 13 Figure 9. WAVELENGTH UM SCAN 1025 222 FlCfUre 10. Zrntih lUdUntt cf thr VnldenlWrd 5 CmlmoA rtilurf Ml.llim n « Function of — Ti/nr Alter L'uneh < K W" 103 "C* (TALJ * ^^r^^ m IB IH ••« Hi, ifc; ;u, Jn, ;.t ;CC f*U JOG »0 J4? TIME IN SECONDS Figure 11. 9 Zenith Radiance Values of Weaker Unidentified Emlisiom ai Functions of Time After Launch MM IN UCMt Figure 12. T- " 120 sec. Figure 13. 223 FIELD WIDENED INTERFEROMETER LAUNCHED 3 AUGUST 1979 0330 (LOCAL TIME) WSMR, NEW MEXICO RESOLUTION A a = 3 CM ' NO (^-l) ROCKET ALTITUDE 87KM H20 Cj) C0 (» > JiW/II W 2 3 WKNWW^**^ ' e j 25^^*^^* 5.3 Vn Figure 14. SILICON COMPOUND VIBRATIONS ASSIGNMENT ABS. FREQ.(cm') WAVELENGTH (urn) Si-H STRETCH 2200 -100 4.5 =0.2 C-Si 1423 - 3 7.0 C-Si SYM BENOING 1250 = 10 8.0=0.1 C-Si 1110 =20 9.05=0.15 Sl-O-C STRETCH 1155=35 9.5 = 0.3 C-Si ROCKING 825 = 25 12.2 = 0.4 Figure 15. 224 100 90 80- 70- S 60 • I 50- m c K 40- 30 20- 12.3M Wavanumbtr CM Figure 16. ABSORPTION SPECTRUM OF DOW CORNING SILICONE VACUUM GREASE 100 90 80 - 70 S 60 i so 9.2u 2000 iiaa Wivtnumbli CM"1 Figure 17. INFRARED ABSORPTION SPECTRUM OF SILICONE OIL 225 INFRARED ABSORPTION SPECTRA OF SILICONE ADHESIVES AS CAPILLARY FILMS c o «00 HOO 3000 (cm) 3300 1000 1*00 1(00 1400 1300 1000 (on) MO COO 655 LTV Figure 18. LWIR ROCKET CONTAMINATION SUMMARY EXPERIMENT WAVELENGTHS OF CONTAMINATION SPECIES OBSERVED ICECAP-CVF (73 & 74) 6.9 7.3 8.0 9.3D 11.1 12.3 — HAVE-SLED (FLT. 2 & 3) 6.8 — 8.0 9.2 ... 12.3 14.5 EBC-CVF (82) 6.8 ... 8.0 9.2 — ... „. SPIRE-CVF (77) ND ND ND 9.3 11.1 12.3 — POTENTIAL SOURCES SILICONE GREASE 7.0 7.9 9.3U 11.4 12.3 14.4 SILICONE ADHESIVE 4.5 7.1 8.0 9.3u 11.5 12.6 14.3 NOTE: "D" indicates doublet Figure 19. 227 SPIRIT CLEANLINESS PAYLOAD LAUNCHER ENVIRONMENT Figure 20 SPIRIT CLEANLINESS CLAM SHELL DEPLOYMENT AT LAUNCH Figure 21. 228 Table in. SPIRIT 1 CONTAMINATION CONTROL • PAYLOAO SKIN REDESIGNED • REDUCEO SKIN OPENINGS. SCREWS. PENETRATIONS... • SKIN SURFACE • NICKEL PLATED ALUM. SKINS • HIGHLY POLISHED TO REDUCE TRANSPORTED CONTAMINANTS • ENVIRONMENT CLEANSHELL ENCLOSURE • TRANSPORTS PAYLOAD FROM CLEAN ROOM TO LAUNCHER • ASSURES TEMPERATURE AND CLEANLINESS CONTROL ON LAUNCHER • DRY NITROGEN PURGE WHILE ON LAUNCH PAD • STAGED PAYLOAD VENTING • PAYLOAD SECTION SEALED DURING DATA TAKING • PAYLOAD VENTED ON ASCENT AND DESCENT VIA BAROSWITCHES • PAYLOAD/BOOSTER SEPARATION VELOCITY INCREASED BY AIRBAG SYSTEM • REDUCED PROXIMITY OF MOTOR CONTAMINANTS • ACS • HIGH-PURITY GAS • PURGED SS SYSTEM LWIR RADIOMETER DATA FROM BMO-OOT PROBE EXPERIMENT (FLT. 5) UNCUIHIFIED POSITIVE PITCH THRUSTER FIRINOS Itrtlll DEQ \nmnnrr. 12.; —r-1—i—i | i i i—i—m—i—i—i—i i i i -I—I—I—I—I—r—i—r- o LONG BAND T- I I 3 s «.h- INTERMEDIATE BAND -1 •". ' ' ' •-•• • •~ T-T—n—r—r- T—I—I—I—I—I 1—|—I—I SHORT BAND 242.S 213.0 243.5 Figure 22. FLIGHT TIME (SECONDS) UNCLASSIFIED 229 NEUTRAL WIND MEASUREMENTS BY FABRY-PEROT INTERPEROMETRY IN ANTARCTICA R D Stewart*, J R Dudeney & A S Rodger British Antarctic Survey, NERC, Madlngley Road Cambridge, CB3 OET, UK R W Smith* Geophysical Institute, University of Alaska, Fairbanks, Alaska 9970I, USA D Rees Department of Physics and Astronomy, University College London Gower Street, London, WC1E 6BT, U K ABSTRACT A large-aperture (150 mm), spatially scanned Fabry-Perot Interferometer (FPI) has been deployed ai Ualiey (75.5°S, 26.8°W; L-4.2), Antarctica. Thermospheric neutral wind measurements are made by finding the Doppler shift of the OI(3P2 - 'D2) 630.0 nm emission. This has allowed the first comparison to be made between southern hemisphere ground-based thermospheric wind measurements and the predictions of a three-dimensional, time-dependent thermospheric global circulation model. Geomagnetic and geographic latitude, are well separated at Halley and so we may expect a distinct contrast to the dynamic behaviour observed in the more frequently studied northern polar thermosphere. Although the initial results from the experiment are in general agreement with the model, some consistent and significant differences between the observed wind field and that predicted are evident in the morning sector. These may be related to uncertainties in mapping oagnetospherlc boundaries to ionospheric heights in the southern hemisphere. The intensity of the 630 nu emission has been examined with respect to the maximum plasma frequency of the Es layer using data fron the Advanced Ionospheric Sounder at Halley. * Previously with Ulster Polytechnic, Shore Roa4, Newtonabbey, Co. Antrim, BT37 OQB, UK. 230 INTRODUCTION The technique of high-resolution ground-based lnterferometry has been used for several years now, in the northern hemisphere polar cap, to derive the neutral thermospheric wind1!2. This is achieved by finding the Doppler shift of the 630 nm oxygen emission line (which has a peak intensity at around 240 km altitude) from its zero-velocity wavelength. Line of sight velocities can be calculated in each of the cardinal directions and, using the assumptions outlined by Smith , converted to a horizontal flow vector. In many cases these observations have assisted the development of thermospheric models but, until now, only northern hemisphere ground-based data have been available. The new southern hemisphere measurements have been made at Halley, a station operated by British Antarctic Survey. The importance of providing a test for southern hemisphere computer simulations of thermospheric behaviour is revealed when we consider the large differences that exist between geographic and geomagnetic frames. The relatively low geomagnetic latitude of Halley, 65.8°S (see Table 1), is the result of the asymmetric nature of the Earth's magnetic field. Table 1. Coordinates of Halley station. Geographic coordinates 75.5°S 26.8°W Geomagnetic coordinates 65.8°S 24.3°E Invariant latitude 60.8°S (L=4.19) Dip latitude 46.4°S Thus, compared with a northern hemisphere station of similar geographic latitude, different magnetospheric related effects, such as the configuration of energy deposition, are observed. Since there is a geomagnetic latitude dependence of some terms in the thermospheric momentum equation, such effects will cause hemispheric differences in the neutral wind. Data from the austral winters of both 1983 and 1984 are here compared with thermospheric wind simulations of the three-dimensional, time-dependent global circulation model of Fuller-Rowel 1 and Rees1*. 231 With an L-value of only 4.2, Halley is usually equatorward of the auroral oval. However, the diurnal variation in the oval's position brings it closest at around 0200 UT and overhead at times of enhanced geomagnetic activity. The major contribution to observations is therefore from the background airglow of subvisual Intensity, auroral emission providing a somewhat irregular, brighter contribution. THE INTERFEROMETER The instrument deployed at Halley, shown in Figure 1, is similar to that described in reference 5 and incorporates a 150 mm, spatially-scanned, Fabry-Perot etalon. The etalon plates are coated to obtain a reflectance of 843! at 630 no and are optically contacted to three, 10 mm long Zerodur spacers. Spring pressure adjusters, similar to the type used in the Dynamics Explorer-2 FPI*, give the etalon a high degree of temperature stability. The overall compactness of the instrument is partially due to the use of a Cassegrain telescope for focusing. The scanning function of an FPI can be achieved by either altering the optical path difference between the Fabry-Perot plates (ie. piezo-electric or pressure scanning) or, as in this case, spatially scanning a fixed wavefront. Imaging of this wavefront is performed by a D. Rees-type Imaging Photon Dectector (IPD). Functionally, it is composed of three main parts: the photocathode (S20); a V-type microchannel plate pair (image-intensifier) and a proximity-focused linear resistive anode. Spatial information of the interferogram is resolved by a charge amplification process, executed on the signal from each of four anode outputs. The subsequent processing gives an (x,y) value to each detected photon event, which is then recorded and analysed by a microprocessor. Preflltering is brought about by the selection of one of the three Interference filters, mounted in a filter wheel with tilt-tunable holders. A shutter position is also available to assist in making thermionic background measurements. At present, observations have been restricted to 630 nm only, but the facility exists to measure 0I('D2 - 's0) 557.7 nm and the 0II(2D° - 2P°) 732.0 nm emissions. A scanning mirror system allows observations to be made in the four cardinal directions, at 30° elevation, the mirror tilting vertically to permit zenith measurements. The instrument is calibrated using the 630.4 nm Neon line, derived from an RF-excited electrodeless lamp. Light from this source is scattered into the etalon from 232 the white rear surface of the observation mirror» Since perspex cumes are easily scratched, which would result In the introduction of scattered light, glass windows are used in the observation dome. RESULTS AND DISCUSSION A guide to the type of results expected can be obtained by examining the predictions of the Fuller-Rowell and Rees model. Model simulations indicate that, despite a high geographic latitude, many of the predicted effects will be due to the lower geomagnetic latitude. For example, the neutral wind at very high geomagnetic latitudes is driven largely by momentum transfer from cross-polar ion convection. This driving force, although present in the Halley ionosphere, is very much weaker than at a higher geomagnetic latitude. It does, however, contribute to the highest wind speeds predicted at around 0300-0700 UT (see Figure 2b) when forces due to the pressure gradient formed by thermal expansion of the dayslde thermosphere combine in a similar direction. The period of lowest wind velocity predicted by the model occurs in the pre-midnight sector, around 2000 UT, when Halley is at the edge of the weaker dusk ion convection cell. At this time the ion drag force opposes the diurnal pressure gradient, resulting in a region of near stagnation. Even for a simulation of moderately disturbed geomagnetic conditions, the velocity remains below about 70 ms-1 for the period between 1000 UT and 2000 UT. The ilalley FFI now provides the first test of the model with ground-based southern hemisphere experimental data. Similar characteristics are evident in ouch of the data, in particular through the evening and early morning sector. However, significant differences have been found at around 0700-0900 UT when the observed wind is often poleward and of moderate strength, the model calculations indicating a weak equatorward velocity. An example of this can be seen in Figure 2a where data from 12-13 July 1963 are compared with the modelled wind vectors. This simulation is designed to represent the moderately disturbed conditions present throughout this period (Kp vas between 4- and 5). The same feature is evident on many occasions and at different levels of geomagnetic activity. This reversal from equatorward to poleward occurs quite frequently betwen 0700 and 0730 UT, as can clearly be seen In Figure 3 where the meridional component only is plotted for a different day. 233 A possible explanation for the discrepancy between model and observations lie6 in the difficulty of mapping magnetospherlcally-controlled phenomena, such as the convection electric field and particle precipitation boundaries, accurately in the southern hemisphere. Since these are dependent on the Earth's magnetic field, the relationship between the geographic and geomagnetic frames must be of vital importance. At present, the model uses the invariant latitude offset of 16° at the pole, but it is suggested that a figure closer to the 23° offset that exists between geographical and dip poles may be more appropriate7. In addition to the wind data produced by the FPI, information on the relative intensity of the 630 nm emission in the different directions can be determined. Useful comparisons can be made with ionospheric parameters as measured by the Advanced Ionospheric Sounder8, also located at Halley. In particular, the relationship between the 630 nm intensity and the maximum plasma frequency of the sporadic E-layer (foEs) has been investigated. On most days studied, there is not a well defined relationship but occasionally a good agreement does occur. In Figure 4, the 630 nm intensity in the south and east and foEs are plotted for the same day (12-13 July 1983). Clearly, on this day an excellent agreement is present between the times and amplitudes of observed variations. (The north and west intensities have the same general form, but with a poorer amplitude relationship). The presence of the three distinct peaks can be explained in terms of auroral oval movements9. As the auroral oval either expands or moves closer during its diurnal variation, a band of diffuse aurora first passes overhead. This is followed by a region of discrete arcs. As the oval returns to its normally poleward position, the diffuse band once again crosses the station. This interpretation is consistent with all-sky-camera data taken during the observation period. One possible reason for the good agreement between the two data sets may lie in the nature of the energy spectrum of the precipitating electrons. For Instance, if high energy (5-10kev) electrons were causing the Es layer, the low energy tall in the distribution may be responsible for the F-region enhancements. However, it may be possible for lower energy electrons to cause both the Es and the oxygen emission at low altitude (180-200 km), such that an Intensity increase occurs despite the very strong quenching10. In this case the layer of emission would descend and a better coupling between the two parameters may result. 234 CONCLUSION 1. Calculations of the neutral thermospheric wind have been made at Halley, a high latitude, sub-auroral, research station. The FPI observations described have provided the first test of a thermospheric model with ground-based southern hemisphere data. 2. Data from the first and second seasons are in fairly good agreement with the model predictions, however a significant and consistent departure exists in the morning sector. 3. A comparison of the optical data with that from the Advanced Ionospheric Sounder has shown, on some occasions an excellent relationship between the 630 nm intensity and the maximum plasma frequency of the sporadlc-E layer. 235 REFERENCES 1. Hernandez, G. & Roble, R.G., J. Geophys. Res., 81, 2065-2074 (1976). 2. Smith R.W. & Sweeney, P.J., in Proceedings, Seventh Annual Meeting of Upper Atmosphere Studies by Optical Methods, Troms*, ISSN 0373-4854, 57 (1979). 3. Smith, R.W., Neutral winds in the polar cap, in Exploration of the Polar Upper Atmosphere, (eds., C.S. Deehr & J.A. Holtet), 189-198 (Reidel Publishing Co., Holland, 1980). 4. Fuller-Rowell, T.J. & Rees, D., J. Atmos. Sci. 37, 2545-2567 (1980). 5. Rees, D., Rounce, P.A., Charleton, P., Fuller-Rowell, T.J., McWhirter, I. & Smith, K., J. Geophys., 50, 202-211 (1982). 6. Killeen, T.L., Hays, P.B., Kennedy, B.C. & Rees, D., Stable and rugged etalon for the Dynamics Explorer Fabry-Perot interferometer. 2: Performance, Appl. Optics, 21, 3903-3912 (1982). 7. Dudeney, J.R. & Piggott, W.R., Antarctic Ionospheric Research, in Upper Atmospheric Research in Antarctica, (eds., L.J. Lanzerotti & C.G. Park), 200-235, (American Geophysical Union, Washington, 1978). 8. Grubb, R.N., The NOAA SEL HF radar system (ionospheric sounder), N0AA Tech. Memo. ERL SEL 55 NOAA, Environ. Res. Lab., Bouder, Colo., 1979. 9. Akasofu, S.-I., Recent progress in Antarctic auroral studies, in Upper Atmospheric Research in Antarctica, (eds., L.J. Lanzerotti & C.G. Park), 157-199, (American Geophysical Union, Washington, 1978). 10. Chamberlain, J.W., Physics of the Aurora and Airglow, Academic Press, New York, 1961. 236 Dome Secondary mirror Cassegrain telescope Primary mirror Filter wheel Imaging photon detector Fig. 1. Cut-away diagram of the Fabry-Perot Interferometer Installed at Halley. Fig. 2.(a) Neutral wind data from 12-13 July 1983 (183O-09A5 UT). vectors interpolated at 15-ninute Intervals are shown, thel' tails lying on a circle describing the locus of Halley through a 24-h period, (b) Model simulation of winds over a 24-h period for Halley. Wind vectors are shown at hourly intervals in Universal Time. The solar and magnetospheric parameters used in the model are chosen to represent the conditions prevailing during the selected observation period. MERIDIONAL NEUTRAL WIND VELOCITY 04-05 MAY 1984 250 -i EQUATORWARD POLEWARD Fig. 3. Meridional neutral wind component- on 4-5 May 1984, showing a typical posc-raldnlght equatorward wind followed by reversal to a poleward wind around 0730 UT. SOUTH 12-13 JULY 1983 -l 10 30 EAST o o—o—o foEs o co N 03 20 a. lirJÅi ^0 1 X Q/ / \ ' - 5 ^ CO CO /' / \ * - LU c O o / ri 4-» / J* o 10 \\ Si v X ^ X CL \ x^ \ \. <^' -__ _. • 1 1 l_ 1 1 1 I—,__1 - ..I i i i i i -J— 20 22 00 02 04 06 08 10 Universal Time (h) Fig. 4. Maximum plasma frequency of sporadic E (foEs) plotted with the simultaneous 630 nm optical emission intensity observed from Halley In the south and east directions on 12-13 July 1983. A 3-point running mean smoothing routine is used on both data sets. 240 Upper Atmosphere Temperatures obtained by Infrared Spectroscopy R. Cerndt, D. Offervann (Ber^ische Universltat - CesamthuclischuLe - Wuppertal, KB Physik, Post fad'. M0127, 56UU Wuppertal 1, FKG) T. Rlix (Norwegian Defence Research listaulishraent, H.O. Box 25, 2007 Kjeller, Norway) Abstract Spectrometric ana pnotometric techniques fur the infrared part of the spectrum are used at the University of Uuppertal to obtain several parameters of the upper atmosphere. Tie ljnu term behaviour of nesopausc temperatures were studied with t^o Oround based spectrometers. The temperatures were calculated fron Ci.*.* intensity-ratios of the 0u(3,1)-P-branch lines at about 1.5 ura. 'Jsin^ spectral analysis techniques, tne variations of tiie upper nesosphere temperatures were annLysed. Significant variations were found with periods of 4 to 6 days and of 10 to 20 days. The correlation of teaperatur^ variations in tt,e stratosphere and upper mesosphere was examined. Latitudinal brauients of temperatures were .iiso considered and compared to a reference atmosphere model. i. Introduction By use ol" rocket-oorne, balloon-borne and jrouiui based spectrometers and photometers at the University of Wuppertal several parameters of the upper atrcospnere are examined. by ni0ht observation of the vitratioual-rotationaL spectrum of excited Oil molecules it is possible tu deduce me.^upause temperatures. A layer of these OH molecules is centered around 116 km with n half widtti of about 8 km (1, 2). Two ground based spectrometers were used to monitor simultaneously the temperatures of the mesopause over Wu^pertal, KRO (51 fl, 7 E), and Oslo, Norway (60 >',', 11 £) betweea December 1982 and April L934. First results were reported in (3). Mesopause temperatures have been measured since I960 with the spectrometer later used in Oslo (4)» Tae instrument which was operated at Wupportal is a Czcrny-Turner-type spectrometer with a focal length of 0,3 m and a full I'icLd of view of auout 13 decrees. The spectral region covered by this spectrometer reaches from 1.2 to 1.7 um with a resolution of 35 Anjstrom. At Oslo an tibert-l'astie-type spectrometer was located witli the same parameters as the one in Wuppertal except the full field of view of about o decrees and tue focal lengta of 0.5 m. Tlie detectors are cryocooled 241 intrinsic germanium cristals. both instruments ara used to measure the three p-branch-lines of OhO.l) around 1.55 um shown In Figure 1. bach of 'jhese scans yields one temperature value within less than one minute in case of Che Uuppertal spectrometer and within less Chan two minutes in case of Che Oslo instrument. The temperature accuracy is about +/- 4 K. The standard deviation for the mean nijht temperacure is about +/- .5 to 1.5 k. A detailed description of the experiment is ^iven in (I» 5, 6). 2. Results 2.1 Intercaiibratlon In order to compare the mesopause temperatures at different latitudes \t is necessary to exclude differences due tu the experimental set up. Therefore both spectrometers were operated at the same time and at the same location for a short time. The measured deviations of the temperatures were uithin the margins of error uf each of the two instruments. T.tus instrumental differences may be excludes as an explanation tor the uirrerances in t.ie mesopnuse temperatures measured with t lie so two spectrometers. 2.2 Latitudinal Dependence of Hesopause Temperatures Figure 2 snows the annual variation of tue mesopause temperatures at 60 N and 51 H between November 1982 and November 1983. The ^raph in the bottom of the figure gives the latitudinal temperature difference T(f>U !«) minus T(5L N), so that negative values reflect higher tempera- Cures at 51 N" tnan at 6U l\ latitude. The individual points represent :nean ni^ht temperatures at tne mesopause and the solid lines arc the mean temporal variations with a smoothing interval of approximately 12 days . Durin_ a rather extended period in winter ard early spring there were obviously no sijnificant latitudinal temperature diffurences over longer periods of time, although considerable day-to-day fluctuations were found. This behaviour has changed after the short spring transition from the winter to the summer mesopause. During summer a i»reat temperature gradient of around -30 decrees Detween 60 N and 51 N appeared. As measurements at Oslo are not possible during June the summer latitudinaL differences of the mesopause temperatures could not be measured more accurately. This annual variability of mesopause temperatures and the latitudinal differences can be compared with those predicted by empirical reference atmospheres. There arc no such models which reproduce the presented annual variation of the measured temperatures accurately. The best correspondence was found with a modified profile of the model of Cole and Kantor (7), which is yiven as dashed lines in Fiyure 2. To come to a reasonable good representation of the measured data the Cole and Kantor profile for 60 N is lifted by b K and shifted by about 5 days to earlier times. The Cole and Kantor profile for 51 N is lifted by 12 N and shifted by 5 days. The difference of tnese modified model 242 profiles is plotted in Che bottun panel of Figure 2. To ouC.lin beCCer agreomenC wich Che daCa Che shape of Che Cole and iCantor profile would have Co be changed, especially during Che spring and aucumn transi tions. This Is also seen in Che laticudinal temperature differences where Che differences beCweeu measured and model values are considerable. The short term variability of the temperatures mentioned above is further Illustrated in Figure 3. The solid line represents the smoothed data set measured at 60 N and the dashed line shows the smoothed temperature plot obtained at Wuppertal. The temperacures are effectively averaged over a time span of abouc 8 days. Note Che diffe rences in Che decline of Cempcratures at 60 y and 51 11 during spring and in Che rise in autumn which preceed die laCiCudinal difference in summer and the agreement in winter. During winter 1983/34 the spectrometer located at 60 N was not operating continuously. Therefore it U nut possible to compare die temperatures at 60 N with those measured by the second spectrometer on this time scale and for that period. Until November 1983 a considerable coincidence of the day-to day fluctuations over Wuppertal and Oslo was ooserved. The period of these temperature variations is anout 15 to 18 days, but most often around 16 days during winter 1982/83. At icast the presence of a lb-days-variation durinf, summer is apparent, although the amplitude is not as hijjh as during winter. 2.3 Variations at the Sumner Mesopause lo study the summer mesopause the Sbert-Kastie-t/pu spectrometer was set up at Stockholm 1982 during the CAi-IP-Campai^n anu tiie Czerny- Turner-type spectrometer was operated in lielfast 1983 to do joint measurements with the Meteor Wind Uadar at Sheffield, Lindland. De tailed results of this cooperation will be published elsewhere. Tabic 1: Variations at Che Summer Mesopause Parameter Observed Periods of the Variations 1982 1983 12-h-tlde 12-li-tide Zor.al winds 2.1 d 2.1 a (.-ii£ht mean at 95 km) 7.1 J 8 d 3.4 d 16 d (t" d) 12-h-tlde 12-h-tide .leridioual winds 2.4 d 2.9 d (nijjne lean at 95 km) 2.7 d 2.1 d 6.4 d 4.5 d, 5.3 d 21.4 d 3d Dti(3,l)- 16 d 7.1 d Temperatures 8 d 3.6 d at 3D km 3.5 d (3 d) 243 Table 1 illustrates the presence of periodical variations of the sunner mesopause winds and temperatures in the order of their spectral amplitudes (raw data of the meteor wind facility provided by Ur. MuLLer, Sheffield). The 7-days-variation observed in inesosp'neric temperatures during 19U3 is shown in Figure 4. 2.4 Results of the MAP/WINE-Campaign From N'ovember 19U3 to April 19d4 the Caerny-Turner-type spectrometer was operated at Andoya Rocket Range, Andenes, Norway (69 N, 16 E), to participate in the MAP/WINE-Campaign. A combination of data of the meteorological rockets with the 0H(3,1)-temperatures has been reported by (ti). The comparison of mesopause temperatures over Andenes with stratospheric radiances as measured by channel 27 of the SSU is shown in Figure 5 (SSU = stratospheric sounding unit, maximum of weighting function at 1.7 mbar (42 - 45 km), courtesy Meteorological Office Bracknell, UK, in: (9)). The dashed lines are smoothed curves with an effective average over about 8 days. At trie mesopausc level a strung variacion ot temperatures with a period of about 12 days is apparent whicti was still present after several stratospheric warnings and related mesospheric coolings. The 12-days-period is in the order of periods of the planetary wave number I (1-j). From the temperature variations alone it is, nowever, not possible to conclude tnat this period was caused by planetary waves. Intense warmings in the upper stratosphere typically develop after an anomalous amplification of height wave 1 (11). Tne large amplitude of the 12-days-variation of tne inesopause temperatures which preceeds the strong cooling events at tue mesopause is possibly connected with this amplification. This 12-days-fluctuation disappeared after the final warming of the stratosphere and tne break down of the polar vortes nas occured. At that time at the mesopause a late winter warming took place and afterwards the fluctuations of the mesopause temperatures were t;uite moderate. Tnis could be explained by the reversed wind direction at high latitudes which prohibits the propagation of waves to higuer altitudes (11, 12). During tne whole time span of the observation the radiances at 1.7 mbar were considerably anticorrelated to the mesopause tempera tures as described by other authors, tou (11). This was most pronounced during times of stratospheric warmings in early and late Feoruary. obviously the mesosphuric coolings were not quite as large as the stratospheric warmings. This might be explained by a westward slope with height of the upward ifioving disturbance which is movitig from east to west. A quantitative analysis of the anticorreiation of the stratosphere and the raesosphere is illustrated in Figure 6. The upper panel shows the relative spectral power of mesopause temperature fluctuations while the middle panel shows the relative spectral power of the fluctuations of stratospheric radiances. The solid line in the bottom panel gives the squared coherence between the periodical variations of the meso- spnere and the stratosphere, whereas the broken line shows the phase between them. The spectral analysis technique used is described by (13), Tne oscillations witn periods of lb days and about 41 days are correlated with a significance of about 95 %. both data sets analyzed snow a periodical variation of about 12 days duration, which are nevertheless slightly different. Therefore the significance of roughly 2kk :i5 .'! for the correlation at this period is somewhat lower than that for the lo-days-period. The phase difference between the mesospheric and stratospheric Fourier components at 16 and 12 days is about Iwl. The oscillation of the mesopause temperatures with a period of ajout 6 days was less important at the stratospheric level. 3. Future Activities To study short term temperature variations in the order of minutes at the aesopause level the accuracy of each temperature value and the tine resolution of about one minute has to be improved. The short terra variability observed until now was strongly affected by the si^nal-to- noiso-ratio. Therefore a new improved version of the computerized spectrometer was tested in a prototyp assembly. Using three entrance slits, each chopped with a different frequency, together with one exit slit and a fixed gratia- it is possible to obtain one temperature value every 15 seconds with th« same margins of error as with the spectrometer assembly used so far. An example of the capability of this new system is shown in Figure 7. Each of the plottet data points is the mean of about 9 individual temperature values with a respective error of about +/- I K. Tne presence of a two-hours-variation durin- that ni^ht is apparent. A new spectrometer set up like the mentioned prototyp thus may be used to examine effects like gravity waves in the upper atmosphere. The Czeruy-Turner spectrometer with r:ovin^ ^ratin^- will be installed at A.idoya Rocket Range in autumn 1W5 for joint measurements with the :;a-LUAi! of the University of Sonn. In this case the ooou vertical hoi0ht resolution of the L1UAH and the better time resolution of the 0.1(3,1 )-spectronoter will be combined tor further studies of the .ridSos^here. 4. Summary Two fJ:i(3, l)-spectrometers './ere used to study the mesopause tempera tures and their seasonal and latitudinal behaviour. The latitudinal decadence of the mesopause temperature is dependent on the seasons. This and other studies have shown that there are no lar^e temperature differences at the mesopause between 5U U and 80 N latitude over Europe during winter. Xist data obtained by meteor wind radar and by Oti( j, 1 )-spectroscopy indicate significant summer variations in the mesopause which may be caused uy wave phenomena. Uurinj; the MAP/U'LML-Catnpaign significant correlations between the stratosphere and the mesosphere were observed. Main periods of the variations were 12 days, 16 days and around h days. A correlation of a M-days-varlation is indicated. The 16-days-variation has been tdent i f ied by other authors as a plant>tary-wa vc-number-1-variat ion. 245 Acknowledgement We thank Prof. E. Thrane, NUKii, Kjeller, Dr. G. Witt, Meteorological Institute of Stockiiolm University, Ur. K. Smith, Ulster Polytechnic, Belfast and their staff for clieir support with funding and operation of this experiment. References I. Lan^e, G., Doctoral Thesis, Physics Ucpt., University of Wuppertal, KKC (1932) >. Thoraas, K.J., and R. A. Voun,;, J. Geophys. Res., 86, 7389 (1981) 3. iilix, T.,'K. Gerndt, G. Lan^e, Proc. bth ESA Sym. Europ. Rock. a. Ball.. Proyr., ESA SP-lb3, pp 181-136 (1984) 4. Jffermann, l>., i<. Gerndt, G. Lan^e and H. Trinks, Adv. Space Kes. Vol. 3, :io. 1, pp 21-23 (1983) 5. Lanje, C., k. Gerndt, H. Trinks and D. Oifermann, iu: Project riner^y budget Campaign 1930, Buiidesninlsterium fiir Forschunu urid Veclmoloaie, Bonn, Sci. Kep. BMFT-Fii-W-b 1-052 (19.31) u. Gerndt, K., Oiplomarboit, Physics Oept., University of Wuppertal, FRU (1932) 7. Cole, E., ^nd A.J. Kantor, Air Force Geophys. Lab., danscom AF3, AFCL-TX-7U-0051 (197:;) :i. .-.eyer, W., K. Cermit, t.R. Philbrici;, and F.J. Schnidlin, Proc. 7ct. KSA Symp. o. Europ. Rock. a. ball. Projr., ESA SP-229, pp 41-47, (1935) 9. Ldblczke, K., 6. Xaujokat, K. Lenschow, K. Petzold, and A. O'Neill, Heilage zur lier line r Uetterkarte 5b/84, SO 15/34 (1934) III. .-ladden, *.A., J. Atmos. Sci., 33, pp 1605-lulS (1973) 11. Labitzke, K., J. Ccopi-./s. Kes., 3d, C1U, pp 9665-90-73 (1981) 12. Ciiarney, J.G., and P.G. Draziu, J. Gcopliys. Res., 66, pp 83-109 (19b!) 13. dlooufielj, P., Fourier Analysis of Time Series, John Wiley & Sons, New York (197o) 246 -,. s §il x cc ex d O-Q. o ib cccc O ao.' = ~l o _, _ tli rs ' 1-1 - ctad no. —X — - O (_ ff: oro: cs o- ix. CD ora C3 a. n- ^P=T r tt daa. A X I S N 3 1 N 1 3 A I 1 V "I 3 « Kiy. 1: lK-spectrun around J .5 |im 247 220 / 200 rf 160 621106 631111 / GRIPS 1 M<6K-5D IBO X- - •¥. I I 1 I t I ] I < I I 1 I I I I 1 I I I I I I I 1 I 1 I I 1 I I I 1 I I f I I ' 220 ^mji; ^ .Ljfcj •4' IBO GRIPS 2 M*12K-5D IBO i * «> > t' ««' i '' * ' '' •« ' i«' ' * \' *<«t ««* 0 100 200 300 20 GRIPS 1-2 Å/- —— 60 N/50 N COLL/KANTOR i i i i | i < i i 1 i ••• | • • • i 1 i i i i | i i i i I i i i i | i i i 100 200 300 T 1 rt E - I D A Y S ) Fi£. 2: Seasonal variation of latitudinal temperature differences 248 ^%4 £2) ICE fr-;0717 ee KK Gl SOLID. G2 DASHED i • • > •' 11111 ?00 3PP 400 600 T I fl E - I D A Y S 1 Ki^. J: Smoothed ncsopduse terap^ratures obtained at 6U N (solid line) and at 5) N (dasiicd line) I X 1 220 - | 7d ] Td j 16 1 20C- 180- R A T U E - a. 160- *— 1 I I I I 1 • 1 1 1 1 1 1 1 1 1 1 1 1 ' 21 31 2 10 20 JULY 1983 AUGUST 1983 Ki^. 4 : Su::i.:u-r InesojKnnje icr.ipcral ure variation uotuined at Ik-lfa 249 831124 840331 230 ANDENES 66 KM 69 N 220 210 200 ri [i i i 111111111111111111 i 11111111111111 n111111 11111 j i n111111 11111111 :„ ANDENES SSU CH 27 100 80 60 ri I i i i i I i 11 11111 11 i.i i i I i i 11 I i i i i I i i i i 1111 i I 11 i i I i i i i I 11 1111 i i i I i i i i I i i 1 DEC 1 JAN 1 FEB 1 MAR 1 1983 1984 7i£. 5: Mesopause temperatures and stratospheric radiances over Andoya, Norway COHERENCE RELATIVE SPECTRAL DENSITY ' i ' i * i ' i ' t '—r- I I I ' I "iH •< MiS-mW-mH-, iitii'| inlii^ inlif I IIIIIIII| i nliii'l ruling i TTII."| lilting lliln't 11 i' I • n ' I " " I " " | 11" I'' " I''' 11'" P H A S E - [ P I 1 251 l)M-3iiniVH3dU31 Fig. 7: Mesopause temperatures probably influenced by a gravity vave (time resolution: 2 minutes) 252 THE UNIVERSITY OF BONN LIDAR EXPERIMENT AT THE ANDØYA ROCKET RANGE U. von Zahn * Physikalisches Instltut, UniversitSt Bonn Nussallee 12, 5300 Bonn 1, Fed. Rep. of Germany Abstract. Parameters and expected performance figures of an improved sodium LIDAR experiment are reported which is being installed at the Andøya Rocket Range, Norway (69.3°N, lé.O^S). INTRODUCTION The middle atmosphere at high latitudes constitutes a rather lively, however, complex medium. Its temperature structure and dynamics still pose ..'.'.allenging problems for our understanding. A number of the most pressing questions will only be answered after obtaining new and more extensive ob servational data on the structural and dynamic parameters of the middle at mosphere. To this end new ground-based methods for remote-sensing of middle atmosphere parameters are developed and brought into use for routine obser vation of the stratosphere, mesosphere, and lower therraosphere. Furthermore, it is recognized more and more that it is advantageous to measure as many as possible of the atmospheric properties in the same volume and at the same time. Following this line of thought the Univer.. y of Bonn, in cooperation with the Royal Norwegian Council for Scientific and Industrial Research {NTNF), is currently setting up a perminenr- observatory at the Andøya Rocket Range (69.3°N, 16.0°E). Initially, the observatory building will house a LIDAR experiment of the University of Bonn and an OH*-spectrometer of the University of Wuppertal, the latter being operated/the Norwegian Defence Research Establishment (NDRE). This brief report shall give an overview of the planned capabilities of the lidar experiment and is to invite other scientists to join with their own measurements the atmospheric studies at the Andøya observatory. * The paper was presented by Ulf Peter Hoppe, Norwegian Defence Research Establishment, N-2007 Kjeller, Norway. 253 SCIENTIFIC OBJECTIVES The scientific objectives of our LIDAR experiment are: (1) to study density and temperature of the 30 to 80 km altitude region from Rayleigh scatter measurements, (2) to study density and temporal behaviour of the sodium layer, which ex tends typically from 80 to 100 km altitude, from Na resonance fluore scence measurements, (3) to study the temperature structure of the 80 to 100 km region by mea surement of the Doppler width of the Na resonance line, (4) to study polar mesospheric clouds, including noctilucent clouds from Mie scatter measurements, and (5) to study the occurrence and altitude distribution of stratospheric dust and/or aerosol layers from Mie scatter measurements. Observations of the type (l)-(3) have been obtained In January through April 1984 (during the MAP/WINE campaign) by an initial set-up of our ex periment near the village of Bleik at Andoya. Scientific results from this experiment have been reported by Fricke and von Zahn (1985) and Klein and Hoppe (1985). THE LIDAR INSTRUMENT A block diagram of the improved University of Bonn sodium LIDAR in strument is given in Figure 1. Its centerpieces are a narrowband, powerful, and tunable laser, a wavelength meter for the precision measurement of each laser pulse, and aim diameter receiving telescope. Performance parameters of the dye laser are given in Table 1. 254 Table 1. Parameters of Dye Laser wavelength 589 nm llnewidth (FWHM) < 0.15 pm pulse duration 10 ns repetition rate 15 Hz energy per pulse 20 mJ beam pointing vertical (adjustable) beam divergence (full width) 0.4 mrad The dye laser is pumped by an excimer laser working at 308 nra, which has an energy of 450 mJ per pulse. The wavelength meter measures the wavelength of each laser pulse to a precision of ± 0.02 pm (or about 1 X of the wavelength distance between the sodium D2a- and D2D~hyperflnestructure components). The heart of the wavelength meter Is a Fabry-Perot etalon of 5.8 pm FSR. Its interference rings are Imaged onto a linear array of 100 photodiodes. After each laser pulse the charge states of the 100 diodes are read out into a single-board computer which calculates in less than 40 ras the desired wavelength information. The design of the receiving telescope is driven by the requirement to achieve a maximum of light-gathering capability with a minimum in cost. Astronomical quality imaging is not required for a LIDAR telescope. In stead, we have specified that the energy distribution of a point source at the focal plane should be such that 90 % of the energy is collected within a circle of no more than 0.1 mrad diameter. Furthermore, selecting a long focal length and hence large radius of curvature for the primary mirror also acts to substantially decrease the manufacturing cost of the mirror. As a further cost saving measure the primary mirror is made from aluminium. The reflecting surface is covered with nickel and polished to a smoothness of better than 5 nm. The parameters of the telescope are given in Table 2. An outlay of the telescope Is shown in Figure 2. 255 Table 2. Parameters of Telescope geometry 2 quasi-spherical mirrors in Cassegraln mounting diameter of primary mirror 1.0 m focal length of primary mirror 5.0 m material of primary mirror Al alloy diameter of secondary mirror 0.25 m focal length of secondary mirror - 1.49 m material of secondary mirror ZERODUR focal length of telescope 20.8 m nominal field of view (f.o.v.) 0.75 mrad diameter of f,o.v. in focal plane 15 mm viewing direction vertical (fixed) The light from the telescope is fed to a horizontal optical bench which so far contains the following components: a mechanical light chop per, which blocks any backscattered light from altitudes below 20 km, an Interference filter of 0.73 ran (FWHM) bandwidth, a Fabry-Perot etalon with 30 pm (FWHM) bandwidth, three imaging lenses and a cooled, photon-counting multiplier. The free aperture of the etalon is 60 mm diameter. The counts received by the multiplier are collected and processed in a fast 100 MHz pulse counter and memory assembly with separation into 0.5 km altitude channels. The entire system is operated by a NOVA 4/C computer. We expect to receive with this LIDAR system about 100 photocounts/ laser pulse from the entire sodium layer, or an average of 5 counts/km. For example, this signal strength should suffice to obtain for example a complete temperature profile of the 80 to 100 km region at 1 km resolution with an integration time of about 5 rain. 256 THE OBSERVATORY The observatory building Is located close to the Andtfya Rocket Range, a few 100 m towards the village of Andenes. Its first floor comprises 8 x 8 m^ area, the second floor 4x8 m^, and the basement 4 x A m%. The second floor contains a 4 x 4 m^ laboratory and observation windows for the OH*-spectrometer of the University of Wuppertal. The building Is con nected via telephone and Intercom lines to the Andrfya Rocket Range, to permit coordination of remote sensing measurements with potential rocket launches. CURRENT STATUS As of December 1, 1985 the observatory building has been completed and the LIDAR instrument been installed (including the new telescope). During the night of November 24/25, 1985 the instrument collected data on the sodium layer for the first time. INVITATION FOR COOPERATION As indicated In the introduction we would consider it very beneficial to attract additional remote sensing or In situ observations of middle at mosphere or lower ionosphere parameters to Anddya. In this spirit we would welcome any proposal for joint measurements even to the extent of perhaps joint use of our telescope for other purposes than our own LIDAR measure ments. But cooperative studies could certainly also consist of network mea surements in which spatially separated stations participate. ACKNOWLEDGMENTS This experiment benefits greatly from good cooperation between the NTNF and the University of Bonn. This research is supported by grant Ho 858 of the Deutsche Forschungsgemelnschaft, Bonn-Bad Godesberg. 257 References Frlcke, K.H., and U. von Zahn, Hesopause temperatures derived from probing the hyperflne structure of the Dj resonance line of sodium by lidar, J. Atmos. Terr. Phys., 47, 499-512, 1985. Klein, V., and U.-P. Hoppe, The sodium layer above Andøya, Norway, as observed by lidar from January through April 1984, In "Proceedings of the 7th ESA Symposium on European Rocket and Balloon Programmes and Related Research", edited by T.D. Guyenne and J. Hunt, ESA SP-229, 1985. Figure Legends Fig. 1. Block diagram of the sodium LIDAR experiment of the University of Bonn. Fig. 2. Outlay of receiving telescope and detector assembly: (1) primary mirror, (2) secondary mirror, (3) two baffles, (4) 90° deflection mirror, (5) chopper motor, (6) Fabry-Perot etalon, (7) photomultiplier assembly, (8) basement of observatory building, (9) first floor, and (10) second floor. 258 LL OH Ul X Q_ LO O r PHOTO- CHOPPER FILTER MULT. DISCRI MINATOR DYE EXCIMER %-* LASER LASER p(N ) CONTROL PULSE SYNCHRO GENERATOR NIZER WAVELENGTH METER Na WAVELENGTH vapour cell DISPLAY RE MEMORY , DATA , FAST ADC v LAY 8xl6K *• ~ SEPARATOR SCALER CAMAC - NOVA INTERFACE NOVA U/C COMPUTER Figure l Figure 2 260 OPTICAL MEASUREMENTS OF ATMOSPHERIC OZONE AT HIGH LATITUDES 5 H H Larsen University of Oslo, Institute of Physics, Blindern, Oslo 3 ABSTRACT. During a period of 17 years, ozone measurements with a Dob- son spectrophotometer have been made in Troms* (70 deg. N). The total ozone amount is evaluated from an observed log in tensity ratio in the direct sunlight or in the light scatte red from zenith at wavelengths in the near ultraviolet regi on. The total ozone amount means the ozone amount in a ver tical column of the atmosphere of unit cross section.The va lue of the light intensity ratio depends on the ozone amount in the atmosphere,and thus the atmospheric ozone can be eva luated quantitatively by this type of ground based observa- vations. Information about the ozone distribution above the observing station can also be achieved from observations of the intensity ratio in the scattered sunlight from a clear zenith sky. INTRODUCTION. Dynamic and radiative interaction between the lower and the upper atmosphere depends strongly on the thermal structure of the stratosphere and the mesosphere which,in term is lar gely a function of the ozone concentration in those regions. This has increased the interest in the ozone distribution and its variability. The fact that variations in the total ozone amount, for whatever reason, natural or antropogenic, may have deleterious effects on the bio-systems and the eco -systems of the earth,troubles many people. This has expan ded the interest for atmosperic ozone, especially for an o- zone monitoring program. A worldwide ozone monitoring pro gram has thus been established. In 1777 W.M.O. (World Mete orological Organisation ) was designated as the leading a- gency in the work connected with the monitoring of the ozo ne layer and in the research into the overall problem of in advertent modification of atmospheric ozone by man. 261 The ozone observation stations in Oslo,Tromsø and at Spitzbergen contribute to the global ozone monitoring program with data derived -from Dobson spectrophotometer readings. The ozone data -from the global network are ga thered by the Atmospheric Environment Service in Canada and ars in cooperation with W.M.O., published every second month. The global zonal trend can now be -followed , and the data also provide a "ground truth" for satellite ozone ob servations. In Troms» ozone observations have been carried out regulary for many years. Due to the high latitude of this station,one has to depend on zenith light observations,al though these observations are uncertain.Moonlight observa tions have also been taken in Tromsø and at Spitzbergen, but uncertainties in the determination of the instrumental constants make such measurements less reliable. However, one needs informations about the ozone amount and its di stribution at high latitudes, and therefore the ozone data from our three stations ars considered as important data. INSTRUMENTATION AND OBSERVATIONS. A Dobson spectrophotometer measures an intensity ratio of sunlight at wavelengths in the near ultraviolet region. In this region the absorption by ozone increases very rapid ly towards shorter wavelengths. The standard pair of wave lengths are 311 and 332 nanometers. If I denotes the in tensity of light at the short wavelength and I'the intensi ty at the long wavelength, then we can express the loga rithms of the intensities of light reaching the instrument as follows: logI = Jogl0- ax logI'= logIa- a'.x (secz) -s* IQ and I0 denote the light intensities outside the atmo sphere, a and a' denote the ozone absorption coefficients, and s and %' denote the scattering coefficients for the atmosphere at short and long wavelength. The angle z is 262 the zenith distance o-f the sun, and x denotes the thickness o-f the ozone layer at NPT. A subtraction gives the follow ing expression: logd'/I) = log(I0/I0> + x x = <(log(I'/I)-log(I0/I0)J/Ca-a'>secz> - ( modi-fied. Further,the value of log ( IQ/I0) .. the extraterres trial constant, has to be determined.This has been done by extrapolation o-f data achieved -from long series of measu rements made by selected Dobson instruments. The value o-f log ( IQ/IQ) is then tras-fered to other Dobson instruments by intercompari son. Thus,a calibrated Dobson instrument which measures the log intensity ratio in direct sunlight ,the logarithm o-f the ratio (I'/I), provides data -for ozone me asurements. Ox-ford was -for many years the place -for cali bration and intercompari son of Dobson spectrophotometers. The Dobson instrument is a double monocromator with very low internal scattering . But with low sun the light intensity in the UV is so low that the internal scattering at longer wavelengths now will influence the measurements relatively more and make them useless -for evaluating ozone amounts. However,the value o-f the intensity ratio in sun light scattered -from a clear sky in zenith depend also on the ozone amount in the atmosphere, and there-fore zenith observations provide readings -from which ozone amounts can be evaluated. But one has to be aware of the -fact that the vertical distribution o-f the ozone also influences the va lue o-f the zenith light intensity ratio. For this reason one has to assume a standard ozone distribution to rely u- pon in the evaluation of the total ozone amount. An uncer— tainty is thus introduced, which increases with increasing zenith distance of the sun. Internal scattering will in fluence the- neasurements after sunset. 263 MODEL CALCULATIONS. A light scattering model of the atmosphere has been evalua ted where mass scattering points are used, and where diffe rent ozone disributions have been employed.First and second order of scattering are used to calculate the values of the intensity ratio in the zenith light wich should be compared with the instrumental readings. The values are calculated for different total ozone amounts and for selected values of z ,the zenith distance of sun. Figure 1 shows results from our calculations. In a diagram the logarithm of the calculated values of the inten sity ratio in the zenith light are plotted versus the angle z.The plotted log values are assembled in curves correspon ding to the selected total ozone amount. The ozone ditribu- tion will influence the shape of the curve,and the total a- mount of ozone will influence the level of the curve. The total ozone amounts are indicated by the thickness of a hy pothetical ozone layer of standard pressure and temperature located at 22 kms altitude . Large ozone amount gives high log intensity ratios. A few important details concerning the log values in the diagram should be mentioned. The calculated value of an intensity ratio of scattered light as received at the ground,will include the unknown value of the light intensi ty ratio of the sunlight outside the atmosphere. Me may say that this intensity ratio has been found by the long series of observations and statistical ir» tments,since it is nee ded in the evaluation of the tot». ..zone amount from a Dob- son instrument reading. However, the value that has been found is the value a standard Dobson instrument would have read outside the atmosphere. This is an instrumental con stant as mentioned before,which is transfered to other Dob- son instruments by intercomparison. When we observe a light intensity ratio with our ca librated Dobson instrument, we get a reading R. The cali bration tables transfer this R-number to an N-value, where 264 N is the following difference: N = log light, and light, subtract log value of (I0/I0) is cancelled by the subtraction. We> denote the N-value of scattered light by N(S) . It is given by the following expression: N(S) = log is ne where N(S) 135 - 130 • 60 + 125 / W + + 2B0.10~3cm03 120 354 115- \ 110+ 80 85 90 92 94 Z in deg. Fig.l. Calculated values of N(S>. N CONCLUDING REMARKS. Atmospheric circulation usually take away or bring in ozo ne in the lower part of the stratosphere. A change in the log intensity ratio in the scattered 1ight,N(S),wi11 result This is indicated in the diagram in Fig.l, where the diffe rent curve labels correspond to the total ozone amounts. If the ozone density at higher altitudes, say at 40km,is redu- sed,then the shape of the N(S)-curve will change drastical ly .(Fig.l) Such changes in the vertical ozone distributi on may sometimes be caused by solar proton events. This is due to production of N and NO during the event. The reduced amount of ozone in the higher regions may last for several days, and in this period lead to a wrong evaluation of the ozone amount based on zenith light observations. The influence on the log intensity ratio which a small change in the high altitude ozone gives, might be u- tilized to give informations about changes in the concen tration of ozone at these high altitudes,and if so.it would be of great imporance. The heights around 40km are probably the most sensitive part of the ozone layer for partial de struction by industrially produced clorofluoro - methanes. Our calculations will therefore be a starting point for further analysis of our long series of observations from Troms». 266 LYMAN-a OBSERVATIONS FROM A HIGH ALTITUDE ROCKET H. Lauene, Max-Planck-Institut fur Aeronomie, 3411 Katlenburg-Lindau, FRG. INRTERZODIAK is the name of a sc-.entif ic program in astronomy to study the ultraviolet emissions of helium and hydrogen by high altitude rockets. An important part of this light comes from interplanetary dust and is well known by the name of zodiacal light. The project scientist for this inter planetary dust and zodiacal light project is Hans J. Fahr from the Astronomical Institute of University of Bonn, FRG. The earth'r atmosphere is not transparent for these UV- emissions, and geocorona emits light of the same kind, there fore it is necessary to observe from very high altitudes (H > 500 km) in order to reduce the strong contribution of the geocorona. The payload of this Skylark 12 rocket contained two helium resonance cells aligned 6° from the rocket axis, one helium reference sensor, and one hydrogen resonance cell at 40" from the axis. A mirror system made it possible to use the main baffle at about 6° from the axis. The scientific payload, including the three helium sen sors, was built by Dornier System, Friedrichshafen, FRG. An attitude control system (ACS) built by DFVLR oriented the payload axis with respect to the ecliptic plane and the sun at different fixed angles between 15° west and 15° east from the sun. This paper deals only with the Lyman-cc observations ob tained by the hydrogen resonance cell built by the Max-Planck- Institute in Lindau, FRG. Since Morton and Purcell detected an extraterrestrial component of the Lyman-o light in the high atmosphere, many attempts have been made to separate the different sources (Morton and Purcell, 1962). An excellent 'view of the re search on hydrogen inside the solar system is given by Bertaux and Lallement, 1984. They describe instrumentation, methods and difficulties in data interpretation. One special source is 267 the scattering of solar Lyman-oc by interplanetary dust parti< les, see Fahr, Ripken and Lay, 1981. photondetector hydrogen resonance cell puis amplifier electronic Lyman-a la telemetry 121.6nm control temperature of filament MgF2 window filament for dissociation of H2 Fig. 1. Schematic diagram of the instrument. Figure 1 shows the schematic diagram of thf instrument. Mechanical dimensions are given in Fig. 5. Technical data for the resonance cell and the detector are compiled in Table 1. Electrical interface to the power supply and the telemetry of the rocket is included in a separate electronic box (795g and P.»* = 7W) . TABLE 1: TECHNICAL DATA FOR THE "HYDROGEN" EXPERIMENT Weight, cell+detector: 520g Size, cell+detector: 20.3 x 7.5 X 7.5 (cm) Size, cell only: 11.5 x 0 2.5 (cm) Power max, cell only 1.22W (4.29 V x 285 mA) Field of view: Plus and minus 2 degrees Sensitivity: 6 counts per second and Rayleigh count rate max.: 130 000 counts per second Aper cure: 1.76 cm* Transmission: 0.3 268 EXPERIMENTAL OUTLAY The resonance cell and the photon detector are the main parts of the photometer. The resonance cell is a cylindrical vessel made by stainless steel. The inner walls are coated with polytetrafluor-ethylene. Both ends of the cell are closed by magnesium fluoride windows. This material is transparent for wavelengths longer than 110 nm. For r\ = 121.6 nm the transmission is about 55%. This cell contains pure molecular hydrogen at a pressure of 1 mbar, which is transparent for Lyman-a photons. By means of a tungsten filament it is pos sible to dissociate a fraction of the hydrogen gas. Hydrogen atoms are now able to absorb Lyman-a photons by resonance. The density of atomic hydrogen is controlled by the temperature of the filament. This temperature itself is related to the elec tric resistance of the filament and can be selected between room temperature and 2000 K by electric commands. In this way the resonance cell works as a selective modulator for a photon flux of Lyman-a, which is detected by a channel electron multiplier. The funnel of this CEM acts as a photocathode. The material for this funnel was selected to have a sharp cut off in spectral response at a wavelength of 130 nm. Therefore the photometer is sensitive between 110 and 130 nm. In order to use the total flux of photons, a new type of channeltron was designed and tested. This detector was machi ned from glas ceramic and coated with active electron emitting glas. The cermic body was sealed by solder glas into a steel tube and welded on to a flange. The special shape of the funnel makes it possible to collect photoelectrons all over the whole diameter. This improved the total sensitivity. The ceramic body combined mechanical stability and good thermal conductivity to form a detector suitable for a rough rocket launch. The spectral response of this detector is shown in Fig. 2. The same diagram shows the transmission of magnesium fluor ide at wavelengths between 100 and 150 nm for a 1 mm thick polished window. The absorption seems to take place at the surface of the crystalline material. Different windows, 1 and 3 mm thick, have almost the same transmission. 269 The hydrogen atoms are recombined mainly by a three body collision process. Therefore, the chemical properties of the third body control the loss rate of hydrogen atoms. Our aim was to create a high density and a homogeneous distribution of atoms inside the absorption cell. The Teflon layer on the inner walls does not react with the hydrogen. The atoms are reflected about 5000 times by the walls before they recombine. In this way they will get the temperature of the wall. Behind the Teflon layer is the metallic surface of the vessel. Metals support the recombination by a cathalytic process. The diffu sion of hydrogen through metal takes place mainly in form of protons. Since we want to keep the gas in the absorption cell for a long time, we recombine the atomic gas in a small gap between the Teflon layer and the metal. Through a small hole the molecular gas returns back into the main volume. The total loss of hydrogen from this cell is very small. After a storage time of 280 days the instrument is still of the same quality as it was in the first week. ov is: - . _.._..._ . .. . . : : : : i '-• fr :- f\ ^-^-^ 60 ^ ^ - ... }><^...... at >* 10 - c u •Nv^ / ' c a; .. j . — 40 £ u E . i ./. T^sJ. .... U) *+- i •/... |N^. —.-, a UJ 5 _ • -i • -'*• |---J — |r •• ^*XL ' c_ ii - • ' •-•:/- • -H ->V^ 20 V CM o(_ 1 /i 1 1 i i o 100 110 120| 130 140 15 0 121,6 wavelen gth (nm) Fig. 2 1 CEM efficiency for UV-emission as a function of wavelength. 2 Transmittance of polished Ma Fa-windows 1 mm thick, as a function of wavelength. 270 /fv Modulation 100 2 filaments i i i i i i i i I u 1 1—i 1 M-5> 150 200 I 250 I I 228 259 282 JZelle[mA] Fig. 3. Modulation "m" as a function of filament current. (See also Table 2.) 271 CALIBRATION The sensitivity and the spectral rtrponse of the instru ment were calibrated by different methods. We used an EUV diode calibrated by NBS a'J a reference for the sensitivity, which is 6 counts per Rayleigh. The spectral response of the resonance cell was tested by means of a resonance lamp. L «J U N Diss. T° (mA) (V) (W) 1 170 1.47 0.25 0.511E-05 0.1 2 180 1.67 0.30 0.162E-04 0.3 3 190 1.90 0.36 0.366E-04 0.7 4 -"lg 2,14 0.43 0.777E-04 1.5 5 210 2.37 0.50 0.190E-03 3.8 6 220 2.64 0.58 0.506E-03 10.1 7 230 2.84 0.65 0.109E-02 21.8 8 240 3.11 0.75 0.234E-02 47.0 9 250 3.35 0.84 0.352E-02 70.7 10 260 3.60 0.94 0.489E-02 98.0 11 270 3.87 1.04 0.734E-02 147.0 12 280 4.12 1.15 0.895E-02 180.0 13 285 4.27 1.22 Table 2: Optical depth as a function of consumed electric power for a cell-pressure of 4.8x10- » mbar. i < . A radio frequency oscillator of about 3000 MHz, 2W output, excites the primary lamp. Light from this lamp enters the resonance lamp which is very "narrow" in order to avoid self-absorption. Inside this volume part of the light is absorbed and a small fraction is re-emitted into the resonance absorption photometer (10.000 counts per second in a parallel beam of 2 cm diameter). There are two identical filaments for dissociation of hydrogen inside the cell. Using both filaments we produce twice the density of atomic hydrogen. See Fig. 3 and Table 2. The DFVLR team Horaba launched the Skylark 12 rocket at llh<» LT * 14h«» UT, 3rd of March 1985 from Natal, Brazil (5.5" south 35* west), almost vertical into the direction of the sun. The flights of the first and the second stage were -1 ( 1 -r 1 r 1 1 f Tis] 600 640 680 720 120 : 160 200 240 280 320 360 400 440: 480 520 560 760 800 840 -r — 1 ,T — - h —i— H[km] - 200 300 tOO 500 600 700 700 600 500 400! 300 200 100 t Vv [km/s] Level of optical 2.5 i,5: 05 -1 -2 depth Cv 0X •• *•<• • -• s "•% \- s y S y y > Fig. 4 Observed emission of La and countrates versus time, height and vertical velocity. The actual pointing directions with respect to the ecliptic plane and the sun are shown in the bottom line. 2 S pinch off tube 3 & MgF2 windows 4 £ baffle 15 & CEM Fig. 5 Mechanical design of the hydrogen absorption cell. 274 Fig. 6 Channel electron multiplier with big apperture; high quantum efficiency and gain; shock proof, glas ceramic version sealed into a stainless steel tube (Germ. Pat. Nr. P 33 29 885.8). 275 nominal. But during the separation phase of the second stage and the ignition of the third stage some technical problems occurred. The rocket reached 706 km instead of 850 km alti tude. The direction and the orientation were changed, and the altitude control system was influenced and started to drift at a rate of 0.6° per minute towards north. This fact made it difficult to calculate the actual looking angles by which we tried to separate the geocoronal from the extraterrestrial emissions at different al'itudes. The actual pointing directions of the hydrogen instru ment during differe.it phases of the flight (corrected for the drift), 3re shown in the bottom line of Fig. 4. The different levels of dissociation, which correspond to different optical depths is approximately for A, t = 20, for B, T = 90, and for C. r = 180. The upper line shows the observed count rates taken in intergration times of 0.41 seconds. The strong peaks at 120s, 360s, and just before the impact at 845s, are caured by glow discharge. There are two components of Lyman-a emissions in this record: 1. A strong emission of 1.5 kR, at a temperature of about 700 K, appearing mainly below 500 km altitude. 2. A weaker emission of 270 R, at a temperature of about 10.000 K, which appears above 600 km altitude. REFERENCES J.L. Bertaux and R. Lallement, Astronomy and Astrophysics 140, 230-242 (1984). H.1. Fahr, H.H. Ripken, and G. Lay, Astronomy and Astrophysics 102. 359-370 (1981). D.C. Morton and J.D. Purcell. Planet. Space Sci. 9. 455-458 (1962). AC KNOWLEDCEMBNTS I «rant to thank Wilhelm Barke for his help to build and test the instrument. H. Fahr and C. Lay made the theoretical modeling and the mathematical treatment of the calibration procedure. Finally. I want to thank J.L. Bertaux for fruit ful discussions. 276 Sumnary of Session VI: Atmospheric Constituents Cieorg CustafsRon Uppsala Ionospheric Obs. Sweden John J Olivero talked about Solar Mesosphere Explorer satellite observations of polar mesospheric clouds and ozone at high alti tudes. UV-spectrometers at 265 and 290 nm were used for the cloud stud ies. Two quantities were used to describe the polar mesospheric clouds, namely: the occurrence frequency and the average radiance. The data covered the time interval 1982-1983. A number of characteristics of the clouds were given. Among the more important ones were that the clouds are approximately one km thick and reproducible over a long time which implies that the polar atmosphere is extremely stable. He also demonstrated that the polar mesospheric clouds do not show the same characteristics as the noctilucent clouds. The ozone profiles in the height range 50- 90 kn show two states of concentration that differs by a factor of two. This was explained by high mixing rates that destroy the ozone. The mixing is controlled by gravity waves that become blocked and the result is a high concentration of ozone. Siren H H Larsen discussed ozone measurements at high latitudes by Dobson spectrometers. In these instruments the ratio of the intensities at 311.0 and 331,0 nm are used to determine the ozone concentration. In order to get the concentration certain assumptions have to be made. Larsen made a critical review of these and pointed out that in 277 particular the assumption that the intensity versus zenith angle has a certain shape may introduce large errors in some cases. The example he showed was a solar proton event with an unsymmetrical intensity versus zenith angle function which introduced a drastic error in the ozone content with the standard method. John W Meriwether discussed ground-based optical observations 01* thermospheric dynamics. The study was based on Balmer- alpha measurements with a Fabry- Perot interferometer and supported with incoherent scatter radar measurements. Two escape mechanisms for Hydrogen were discussed: escape due to temperature, and escape due to charge exchange. A fourier method was used to study the line profiles, and it was pointed out that the filtering of the fourier coefficients is of great importance particularly in the wings of the profile. A number of temperature versus local time diagrams were shown for different magnetic activity. They showed that thermal escape is dominant during solar maximum, charge exchange is dominant during solar minimum, and red wing enhancement of the hydrogen profile was found for increased magnetic activity. The last point was attributed to higher ion temperatures leading to greater charge exchange. Han. Lauene reported about Observations of Zodiacal Light and Geo orona from a rocket launched from Natal in Brazil. A Hydrogen cell to measure Lyman- alpha had been developed for this rocket program. The cell required only one watt of power. They had also found a solution to contamination from the walls. The data from the rocket were difficult to evaluate due to failure of other instruments on the rocket, and has sn far not been done. So far It has been proved that the photometer la functional and in suitable for further measurements. 278 LOW LIGHT LEVEL MEASUREMENTS WITH PHOTOMULTIPLIER TUBES - PHOTON COUNTING VERSUS ELECTROMETER OPERATION A G WRIGHT Thorn EMI Electron Tubes pic, Ruislip, Middlesex HA4 7TA, England Abstract The intensity of light falling on a photomultiplier may be estimated by using either a current measuring technique, or by employing the photon counting method. In both cases, photomultiplier noise-in-signal dictates the ultimate accuracy attainable. Where very low light levels are involved, photon counting offers optimum sensitivity. Considerations relevant to the selection of suitable photo- multipliers are: spectral response, dark current, dark count and single electron resolution capability. Guidance on setting up a system for low light level measurements Is given. 1 Introduction The photomultlpller, by virtue of its high gain and low noise performance, Is still the only practical detector that will provide a discrete pulse for each detected photon. Over the last few years avalanche photodlodes have been shown to have this capability but their suitability as part of a photon detection system has yet to be demonstrated. In low light level measurements photomultlpllers represent the obvious choice of detector because of the following important attributes. 279 high gain capability, up 10° relatively noiseless amplification high signal/background ratio fast time response, rise times ~ 1 - 10 ns extensive «-ange of photocathode types and area (Fig. 1) long term stability and lifetime VUV to IR wavelengths coverage (Fig.2) The intensity of light falling on a photomultiplier may be estimated or monitored by using either a current measuring technique (electrometer operation) or by detecting individual photons (the photon counting method). The former is suitable for slowly varying light intensities but where fast transient phenomena are concerned photon counting has the capability of resolving pulses separated by as little as 5 ns. This particular attribute is utilised for example in laser light scattering instrumentation using fast digital correlators. In the next section basic theoretical arguments arc given which show why photon counting is the preferred detection method at low light levels. This is followed by a consideration of the sources which give rise to non-ideal performance,comparing their relative importance for the tvo detection modes. Practical guidance on setting up and optimising detection hysti-m.s i one ludes this lee Lure . 2 Ideal photodetection statistics Optical radiation in detected at the photocathode by the emission of photoelectrons following the absorption of photons. The photoelectric effect is a quantum process» governed by Poisson statistics. This means that if M phocoelectrons are produced in a time interval T, then the relative fluctuation in H is cHM> \_ ...(1) 280 The relative variance, by definition [ cr(rl) I , Is X [ The signal-to-noise ratio, S/N, is simply 41 - \-^\ - M ....(2) INI lor(rl): 1 Equation (2) sets the lower limit to the accuracy in measuring any signal. Let M refer to a time interval of one second, then the mean cathode current 1^ is rk = e H ...,(3) where e is the electronic charge, and ^'"fe The more traditional way of specifying the noise in an electrical circuit, where a mean current 1^ is defined, is through the shot noise formula. Here time enters via the system bandwidth, Af. T^* = 2e Ik Af (5) where i^2 is the rms fluctuation in the cathode current. The S/N ratio follows from equation (5) and equation (3) fi'l2 = k2 = 1 .M •••(6) LNJ Tk» (2Af) Equation (2) states the accuracy attainable by the photon counting method, while (5) defines the signal fluctuations inherent in current measuring techniques. The two noise formulations, (1) and (5) are reconciled in equation (6) which illustrates 28t their equivalence at a bandwidth of 0.5 Hz. This docs not imply that the two signal detection methods offer the same accuracy in practice. To show this, It is necessary to examine the nature of the photomultiplier output taking account of other sources of fluctuation. Consider a fictitious photomultiplier which produces a charge q = eg at the anode for each photoelectron leaving the cathode, g is the gain of the multiplier, assumed for Che moment not to fluctuate from pulse to pulse. The total charge Q collected at the anode over a time T is Q = m T, where m is the mean rate of emission of photoelectrons. e relative variance of Q is given by equation (1) as var(Q) = _^1_ = ^ (7) The gain of tho multiplier section is a fluctuating process wh i ch acts on each phoLoelee L ron independentJ y. This arises because of the statistical spread in the secondary emission coefficient of each dynode about a mean value £.. It can be shown, (1), that equation (7) should be modified to allow for this, as follows var(Q) _ _J_ f 1 + var(g) L ..(8) 1 .(9) m T In photon counting we can all but eliminate the effects of varying pulse heights by using a discriminator which produces a standard pulse for all photomultiplier pulses above a fixed charge threshold. This implies, in equation (8), that varCg) •*• o and ideal statistical detection as predicted by equation (1) is achieved. Returning now to the dc detection method. The equivalent 2B2 circuit of the photomultiplier and Associated electronics can be reduced to a parallel combination of resistance and capacitance. The noise bandwidth is simply fcf - 1/ARC (10) From equation (6) and including the effect of fluctuating gain, g MT(Q) = __|! {f I I + + var(gvar(g)) [ ] Where ^ = RC is the integration time constant. Clearly the signal-to-noise ratio can be increased by choosing a long time constant. However a long time constant can effect the ability of the measuring instrument to follow a time varying output current. In other words, a measurement at time T will be correlated to that at T + t or T -t . It is common in the use of a ratemeter (5), to select a time constant giving a 2% exponential residue, in which case ^ = T/4 and equation (11) becomes var(Q) = _J_ ( 1 + var(g) 1 (12) The re lat ive variance is doubled with capacitive integration (equation (12) when compared with equation (9)) which implies twice the experimental time to achieve the same degree of accuracy. Note further that with capac itive integration, the expression 1 + var(g)/ To summarise: equations (1) and (5) refer to ideal detection of photoelectrons. For low light level measurements it is necessary to amplify the photoelectron signal. The secondary emission process in the photomultiplier is noisy and introduces a factor 1 + var(g)/ S/N. In che case of photon counting, it is possible to obtain near ideal detection statistics by eliminating the effects of fluctuating gain but for direct current measure ments this is not possible. 3. Sources of non ideal performance 3.1 Single Electron Response (SER) The theoretical arguments of section 2 show the fundamental importance of the SER in both detection modes. The variance of the SER is also of considerable practical relevance in photon counting and is therefore .in important men sure of a par Iicular phot omultiplier's suitability for photon counting. The SER of a photomultiplier may be obtained by recording the output pulse height distribution when the cathode is illumin ated by a very weak source of light. A light source may be classified as single photon, if the time interval between photoelectrons is much larger than the analysis or dead-time of the recording electronics. In general, a photoelectron rate of < 10^ s"* satisfies this criterion. The distribution of figure 3(a) refers to a linear focused multiplier incor porating a high gain, first stage; 3(b) is the SER for a Venetian blind tube. The scale on the abscissa may be variously expressed as photoelectrons equivalent, gain or anode charge. The optimum discriminator threshold for photon counting is a compromise between maximising the count rate and operating in a sufficiently flat region of the distribution. The valley of figure 3(a) is a satisfactory choice, accepting a small loss in counting efficiency. Distributions such as figure 3(b) do not offer such an obvious choice, but operation in the region indicated is usual practice. 284 It was stated earlier, that for photon counting, the effects of varying pulse height arc eliminated by standardising the photomultIpLier output an a digital pulse. However, no discriminator is infinitely sharp with regard to its detection on. threshold: there is always an area of uncertainty/either side of the set discrimination level. Pulses falling in the shaded bands of figure 3 are counted with less than 1007« efficiency. The shaded area of figure 3(b) clearly represents more noise than the corresponding shaded area of 3(a). 3.2 The plateau characteristic If the photomultiplier is coupled to an amplifier-discrirainator circuit, then a standardised (TTL or EC1.) pulse will be generated every t imc Q exceeds Q', the circuit threshold leve 1. Recording the output rate derived from a steady source of photons as the applitjH voltage, and hence gain, is increased wi 11 provide the plateau characteristic. An example of such a curve is given in figure 4. The plateau characteristic is invariable plotted with overall voltage as the abscissa, and for most practical purposes this is sufficient. However, for investigating system stability it is useful to represent the plateau characteristic in terms of multiplier gain. This is not an easy photomultiplier parameter to measure in absolute terms t for example, is straightforward. The curves of figure 4(b) are derived from the same data as 4(a) but the abscissa now has the dimensions of gain. The general shape oi the characteristics is easily explained. When the photomultiplier gain is low, only a small proportion of the output pulses rxceed Q'. At high gain most output pulses exceed Q' and operation on the plateau is obtained. Any further increase in gain results in only a small increase in count rate until the gain capability of the photomultiplier is exceeded. At this stage the counts increase rapidly with voltage indicating the onset of unstable 285 tube operation. The point 'w the curves marked SEP indicated at half the plateau counts corresponds to the peak of the SER. Amplifier-discriminators are inexpensive compared with multi channel analysers and consequently most photon count ing applications use the former. The question concerning the optimum operating point on the plateau characteristic cannot bo answered without f irst discussing drift and fat igue effects in photomultipliers. 3.3 Drift, fatigue av»d hysteresis effects By drift is meant the short term change in gai n observed each time the photomultiplier is switched on. The effect is at the few percent level and may be positive or negative persisting for minutes or even hours. Tube fatigue or loss of anode sensitivity is a funct ion of out put current level, dynode materia Is and previous operat ing history. The level of output current I hat a t\ i ven phot omul I i |>1 i ei tan ma i nL;i i n varies wide 1 y, even amongst tubes of the same type. As a rough guide, at a steady anode current of 1 pA, the gain will fatigue at a rate of ^1% per 1000 hours; the fatigue rate scales approximately as the anode current. Changes in gain result whenever light levels alter - a common situation in practice. Fortunately the effect is reversible, takes place over a short time interval and is small in magnitude. In practice the effects of drift can be minimised by allowing the photoraultiplier sufficient time to stabilize. Fatigue is more serious, especially in dc detection where the output level varies directly with the gain. Based on the figure of 1% gain loss per 10-* hours, it is a straightforward calculation to arrive at the stability curves for the two modes of detection given in figure 5. The curves relating to photon counting are derived by reference to figure 4(b), 286 assuming an initial operating point on the flat region of ihe curve. A gain reduction by a factor 10 changes the count rate by only —10% and explains why photon counting is relatively immune to gain change in the photomultiplier. The preceding discussion has been concerned with gain changes of the multiplier with no mention of changes in cathode sensitivity. Long term tests done by Thorn EMI and other manufacturers indicates that, except under very high levels of illumination, most photocathodes are stable with time. 3.4 Temperature coefficients The effect of cooling a photomultiplier is three-fold: the dark current reduces, the spec tral sensitivity of the cathode alters and the multiplier gain increases. Cooling the photomultiplier is advantageous when the signal level is comparable with or less than the ambient dark count rate. Photomultipliers with Gallium Arsenide, SI and S20 photo cathodes are Invariably cooled because of their high thermionic emission at room temperature. The variation of dark currert/ counts is shown in figure 6. The spectral response and absolute sensitivity of a photo- cathode is sensitive to temperatur2 changes. Figure 7 shows representative curves for three photocathode types. The loss of sensitivity at wavelengths approaching the red threshold indicates that there ia a trade off between red sensitivity and dark current. In general, the recommendat ion is: cool no more than really necessary. At wavelengths below 500 nm, the temperature coefficient is very small. The temperature coefficient for multipliers is ** - 0.2 X C~*, for SbCs and BeO. The same arguments as discussed in section 3.3 in favour of photon counting for best stability apply here. Naturally changes in photo sensitivity affect both detection modes in exactly the same way. 287 3. "i Gain linearity If the anode current is proportional to the magnitude of the radiant flux, then the photomultiplier is said to be linear. Deviations from linear amplification occur at high levels of anode current where space charge effects are evident. Non linearity, related strictly to the photomultiplier performance, does not normally occur at anode current leve Is -,< 100 uA- Clearly gain non linearity is of no consequence in photon counting - the pulses are still counted even if distorted. Furthermore, the mean anode current levels that apply in photon counting applications arc usually in the range InA - luA, The most likely cause of non linearity in electroraeter applications arises from poor voltage divider design. This is a topic which deserves more attention than it is usual ly gi,ven - designing a divider that will provide linear amplification for anode currents up to 10 u A is not a trivial exerciser Readers are referred to the Thorn EMI reprint R/P 069 on this subject. 3.6 Dead-time losses If a pair of photon detection events occur within an interval outside the resolving time of the photomultiplier or the photomultlplier/electronics system, then only a single count is recorded. The pulse pai:- resolution of a photomult i pl ier depends on the type and on the operating conditions - it ranges from 5-30 ns. Most commercial amplifier/discriminatorc have variable dead-time, from 20 ns - 1 ^s. In practice correcting for dead-time is only approximate because theory refers either to paralysable or non paralysable counters. Real circuits fall somewhere in between these limits and 288 the author's recommendation is to avoid making a dead-time correction of more than 107.. It is preferable to reduce the input signal using a neutral density filter (which can always be calibrated in situ). Dead time is of relevance to photon counting only and the need for such a correction should suggest the use of electrometer methods to the user. 3. 7 Pho tomu11 ip1ie r bac kground In sec t ion 2 we were concerned with noise in s ignal which is an inescapable manifestation of the quantum nature of radiation and charge. It is essential to distinguish between noise and background in photomultipliers. Background refers to the measured photornultiplier output in the absence of cathode illumination. Where pulse counting is of concern, then the output count rate and pulse height distribution are of importance. The background current (usually referred to as dark current) is the significant photomu1tiplier parameter where tie detect ion is under consideration. It is clear from all the previous theoretical considerations that a photocathode with the highest possible quantum efficiency, at the wavelength or wavelength band of interest, should be selected. However, the (S/N) ratio pre dicted by equations (2) and (6) varies only as M^ and hence as^. The dark current and dark counts vary considerably amongst tubes of the same photocathode type. Generally there is an increase in both tl.est parameters, the more red sensitive is the photocatohde. The signal/background ratio<\/B is undoubtedly a critical selection guide. For a group of similar tubeSjfl/B will vary more by virtue of B rather than by *\ ; and B is therefore the more relevant parameter in tube selection. Typical background pulse height distributions for bialkali tubes are shown in figure 8; (a) refers to a linear focused multiplier with a firsL stage of high gain,and (b) represents 289 a Venetian blind photomu1 tiplier. Comparing signal and back ground distributions we note that there are higher proportions of both raultiphotoelectron pulses and fractional photoelectron pulses in the background than in the signal. In photon counting, all pulses above threshold are given equal weight; in dc measurements, pulses are weighted in proportion to their charge and hence the multiphotoelectron pulses can contribute significantly to the dark current while making a negligible contribution to the dark count rate. With the fraetional photoeleetron pulses, the situation is reversed and the lower threshold must be chosen with some care. A more detailed treatment of background is given in reference 3. There is often poor correlation between dark counts and dark current; that is, a tube may have high dark current and yet low count rate or vice-versa. This is partly for reasons already given and partly because dark current has two components I (dark current) - n(q)qdq + 1^ ....(13) where 1, represents the sum of the leakage currents flowing into the anode. Verv little is known about the nature of this component of dark current, but there is evidence to suggest that this together with the cosmic ray component of background, contributes to 1/f noise. (To be consistent it should be referred to as 1/f background). A plot of the dark current frequency distribution generally has a form similar to that shown in figure 9. The corner frequency fc. is rarely below I Hz and may be as much as 1 kHz. The presence of low frequency noise imposes a lower limit to the sampling time T or time constant which should both be << fc. 3.8 External Magnetic and Electrostatic Fields Photomultipliers are sensitive to the presence of magnetic 290 and electrical fields. These fields may deflect electrons from their normal paths between the cathode and the first dynode causing a loss of sensitivity or cause å loss of multiplier gain through interdynode losses . Changing magnetic fields of the order of the earth's field cause noticeable gain changes and mu metal shields arc recommended whenever an application requires rotation or movement of the photomultiplier. An electrostatic shield maintained at cathode potential is usually sufficient to screen the photomultiplier against all but the most severe electrical interference. 4. Basic count ing system Low level light measurements require extreme care in the electrical wiring. Electrical and magnetic shielding and good high frequency earthing techniques must be employed. Any material in contact with, or in close proximity to, the photocathode should be maintained at cathode potential. Failure to do so can result in unstable tube ope rat ion. With the "anode earthed" configuration it is advisable to connect the cathode pin to a conducting coating or shield around the cylindrical surface of the tube. For low level photon counting, the earthed cathode configuration offers the best immun i ty to electrical interference and usually t he most stable photomul tipl i er operjt ion. High and var i able count rates can, however, dictate the need for direct coupling between anode and the associated electronic circuits. Phnon counting electronics need not be sophisticated. The basic requirements call for a stable amplifier /discrii.iinator combination capable of handl i ng pulses over a dynamic range of 100:1. The user is faced with a problem in choosing how to proportion the overall gain of the system between the photomultiplier, the amplifier and the discriminator. This is resolved by recording a family of integral count rate curves as shown in figure 10, for a range of sensitivity 291 settings. A plot of S/B will show the range of sensitivities over which acceptable performance obtains. For the photomultiplier/electronics combination shown, it is clear that the sensitivity of the counter must be > 1 roV. The enhanced background, evident at high sensitivity, was due to »• lee t r ic.il piik up in this particular system. There is considerable advantage to be gained by reducing the background rate by cooling. With the 520 type photocathodes, background reduction by a factor of 100 can be achieved by cooling the tube from 20°C to - 20°C; the bialkali photo- cathodes show a reduction by a factor of 3 over this tem perature range. In those instances where the light can be focused, the "se of a photomultiplier vith reduced cathode area can obviate the need for cooling. The EMI 9893 range of photon count i nj; lubes, for' ex.implr, prnvid.-s ( lir opt inn of .'i 2.5 mm or 10 mm diameter cathode. 5. Range of electrometer and photon counting methods Consider the measurement of a source producing 10^ photons per second incident on a photocathode with i"l - 107». The sensitivity of a typical, commercially available, ampliiier/ discriminator requires la = M e Given a bialkali photomultiplier selected for low dark current (<0.5 nA) and dark counts <<100 s"*) then photon counting offers an efficient means of measuring the signal. The same tube, operated at already oentioned. The lock-in amplifier technique, employing a Light chopper, overcomes Che 1/f limitation by successively measuring S + B and B and provides a more elegant solution. The advantage of photon counting becomes more obvious, the lower the photoelectron rate. Measuring anode currents below 0.1 nA, corresponding to a photoclectron rate of 10 s"* for the example cited, presents a considerable challenge. The ability to recover a signal, one tenth that of the background, is shown in figure 11 taken from (2). Synchronous detection, that is the interleaving of S + B measurements, was used to extract a count rate of 0.04j s~* from a cooled dark count of 0.459 s"*. Although the experiment required several hours of progressively updating S and B estimators, there is no alternative technique capable of this degree of s ignal recovery A complete photon detection system which combines photon counting and dc detection in the same package is offer ed by Thorn EMI. The Photon Dection System, designed by Tothill (reference 5) uses a microcomputer to optimise the operating conditions and provides simultaneously anode current and anode counts. The user has the choice to select whichever is the most appropriate parameter for his experimental conditions. To summarise, photon counting is undoubtedly the preferred method for count rates up to about 10*> - 10*> s"1. At higher count rates the arguments for photon counting based on S/N anu background discrimination assume less importance and the need for dead-time correction now favours elec trome ter operation. 293 Reference» referred to in the text and further reading 1) Determination of the multiplier gain of a photomu1tiplier. A.G. Wright. J Physics E : Sci. Inst rum li±, 851-5. 2) Voltage divider design. THORN EMI Reprint Scries RP/069. 3) An investigation of photomultiplier background. A.C. Wright. J Physics E : Sci. Instrum lb, 1983, 300-307. 4) A comparison of current measurement with photon counting in the use of photomultiplier tubes. C.J. Oliver. THORN EMI Reprint Series RP/066. 5) Photon Counting Techniques. H.A.W. Tothill. Light Measure ment 81, SPIE Vol. M£, 22-26. 6) Photomu11i plicr tube selection and housing design for wide band photon counting. M.R. Z.itzick. Application note 71021, SSR Instruments Co., 1001 Colorado Avenue, Santa Monica, California 90404. 7) Applying digital techniques co photon counting. M.R. Zatzick Research/Development, Nov. 1970, Zl, No. 11, 16-22. 8) Measuring the light infinitesimal. S. Arnold. The Spex Speaker, Spex Industries Inc., 3880 Park Avenue, - Metuchen, NJ. 08840. 9) Photon counting - notes on a basic system. F. Weekes. Electro-Optical Systems Design, June 1977, 30-34. 294 Figure Captions Figure 1 Illustrating the range of qeomr>tr>«s available in photomultipliers. Figure 2 Spectral response curves for photomultipliers covering the range VUV to 1R. Figure .1 The SF.K using a charqi.'-scnsi t i ve multichannel analyser, a) linear focused tube with high gain first stage, b) a Venetian blind tube. Figure 4 Signal and background plateau characteristics plotted a) as a function of applied voltage and b) relative to gain. This highlights the size of the pulses counted at various operating points. Figure 5 The data represent the expected stability in output for the same photomultiplier operated under photon counting and dc modes. The curves are derived from figure (4)b assuming g/g = -1* per 10 hours at 1 uA. Figure 6 The variation of dark count rate with temperature for typical 2" photomultipliers. Figure 7 Temperature sensitivity o£ some typical photocathodes as a function of wavelength. Note the loss of red sensitivity as all cathode types are cooled. Figure 8 Signal and background distributions illustrating the excess of small and large pulse height signals in the background. a) linear focused photomultiplier b) Venetian blind type. Figure 9 A dark current frequency distribution showing an excess in the low frequency components compared with white noise. Figure 10 The use of family of integral count rate distributions to optimize the performance of a photon countig system. Figure 11 The technique described by Oliver (RP/066) for the recovery of a feeble signal in the presence of background. 295 # 0 w m ° rar ior mBTm THjir »0r ll^ll w^pr' Figure 1 QE % Q.E.% •*____ 10 \ \ 1(1 ffT \ \ \ \ 1^ \S2 1 ', Ga/u \ \ \ 1 \ 1 1 l K '' | i • IS KBr Cs I ICsT! Sl \GaIn, Bi alk i 1 0.1 \ 0-1 \ 1 : i | i i i i 200 400 600 800 1000' nm K)0 ZOO 300 400 SIC 600 X nm Fiqure 2 296 •tntctlictrofit «tuWilint 1 •jS U J 0 Mil • • MS» • Figure 3 photoelecrons 1 0;1 0;01 ! I I type 9863/100 signal signal md 10 ?EP / SEP / / in3 ba kgrau nd 10' / 1-3 « 1-7 W M 2-3 applied voltage (kV) (a) Figure 4 297 photon * 'counting * _ OB. Ll_ 10u 10" Figure 5 Figure 6 296 •/. •r* T t -0-3 \ «02 KtSbCs * \RbCs \S20 •o-; \ • • 06 \- •0-8 - • 10 • • —i 1— 400 500 600 700 800 \ rnn Figure 7 »h*to«1tctro is equivbltnt 1 I » -photoelectrom equivalent . tlU#t«7t 1-2. 3 • • • signal • • 9635*33335 • • • bothgr ountf • • o signal • • * background • ::: : • • • • ... . . <+ • • ^"•"So • • 1 ••• •• * • o. * •• "'*. •••• ••?. \ c •. r • • •• • .... ^ *oo»« • • • •.. J q. or g (a) (b) Figure 8 299 Figure 9 M 1-2 U W 18 20 22 voltage Figure 10 Figure 11 300 A PULSE AMFLiriER/DISCETMINATO» (PAD) F01 SIHGLE PHOTON COUNTING Lennart Alexander Department of Meteorology, University of Stockholm, Arrhenius Laboratory, S-106 91 Stockholm, Sweden Abstract The LeCroy MVL 100 pulse amplifier/discriminator (PAD) was evaluated for use in single photon counting. The following properties were studied: -Count rate (N) as a function of photon flux ($) «hen used with an EMI 9924B photomultipller. -Pulse pair resolution time -Power requirements -Necessary additional components The pulse amplifier was found to be reliable, fast, and small. The circuit diagram and the printed circuit board layout are Included. 301 1. INTRODUCTION The LeCroy MVL 100 PAD Is a high speed moiolitlc pulse amplifier and discriminator intended for low-level pulse counting. Threshold and output pulse duration are controlled by external components or by external voltages. The PAD described in the following was designed for single photon counting with the 0 30mm, 11 dynode EMI 9924B photomultiplier. This PMT is often used by our group on rocket borne radiometers. The outputs of the MVL 100 are standard ECL-levels. Rocket telemetry, however, uses TTL or Cmos levels. Hence an ECL to TTL converter is included on the printed circuit board.(Appendix Al). 2. COUNT RATE VERSUS PHOTON FLUX 2.1 Equipnent Count rate versus photon flux at the PMT cathode was measured as shown in the figure below. WoU Wee Figure 1: Equipment used in count rate measurement. 302 The following equipment was used: a halogen lamp, a 558 nm interference filter, six Wratten neutral density filters, a fused silica dlffusoc and finaly a photomultiplier. The interference filter selects a narrow wavelength band (10 nm FWHM), were the neutral density filters are well defined. The neutral density filters control the irradlance at the diffm-or.The diffusor is used to obtain an evenly distributed photon flux at the PMT cathode. All measurements were made,as shown in figure 2, with the PMT cathode at negative potential and the anode at ground potential. This allows both photon counting and anode current measurement. PAD WTO r- I.Ztln. ;2.".1JI '.2 MÆ I.2M2. H'3« VotfrtoqC Figure 2: PMT operation. In photon counting measurements the input impedance of the PAD is the anode load of the photomultiplier. In current mode the picoampere meter input impedance serves this purpose. 2.2 Measurements Count rates and anode currents were measured at well defined steps of photon fluxes ranging over 5 decades. Furthermore, these measurements were repeted at four different PMT dynode voltages. J03 The HVL 100 is a nonparalyzable PAD. This.implies that when two Input pulses are separated by less than the FAD's dead time, the second pulse will be ignored. The PAD will react as if only the first pulse appeard at its input. Unlike the nonparalyzable PAD a paralyzable PAD will be retiggered if a second pulse Is recieved within the dead time. This will, however, only result in an extended output pulse. These two different types of PAD'S require different corrections for the lost counts. For a nonparalyzable PAD the "lost count" corrections are: N N „ (1) i -r-Nn N is the corrected count rate, N is the measured count rate and N . m dc is the corrected dark count rate. Usually we can assume N ~N . dc dc 304 IOOOO' tooo • 100 10 !.o • ik [V,>d« uoLU^e. iarov 0.1 O.Ol J--••10- ->l*i(%) -S -H -I -Z -' O Fig 3: Anode current and corrected count rates versus photon flux §. The PAD outputs were set at 50 ns pulse width which results In a dead time of 60 ns. 305 3. POLSK PAIS RESOLUTION Pulse pair resolution eime is the minimum time required between two consecutive pulses at the PAD input that can be resolved as two separate output pulses. Shorter time intervalls fall within the dead time of the PAD and only one output pulse will result. 3.1 Equipaent A pulse generator, an attenuator and a capacitor are used to AC couple pulses to the PAD input. |_i_n_ GOdE 1 Os c1 Ho scoce Figure A: Pulse pair resolution measurement equipment. The pulse generator was used in the pulse pair mode. In this mode the delay betwen two pulses can be continuously varied while retaining a constant repetition frecuency. Input and output pulses are studied with an oscilloscope as shown in figure 4. This constitutes a total load on the PAD output of llOpF and lH(l(a 70 cm coaxial cable and the oscilloscope input). 306 3.2 Measurement Pulse pair resolution measurements were made at 100 kHz frequency with 45 ns FWHM input pulses. The charge of each pulse was approximately -12 6 1.1*10 C, this equals 7*10 electrons. U» (Pulst ^wer-ator volUje) • so™/*. r-> u.M • PAD oolpui vc.(l W"=/d.v -i-p^ffi•1st' Je'.nu (IS «a! Figure 5: Pulse pair resolution measurement on the MVL 100 PAD For a PAD with 50 ns output pulse duration the dead time was found to be 60ns. This dead time T is used in eq. 1 to correct the measured count rates for lost counts. When the input pulse was increased to 10*10? electrons, the dead time decresed slightly to 58 ns. 4. POWER REQUIREMENTS As the MVLIOO PAD and the MC10125 ECL to TTL converter both contain ECL logic they are rather power consuming. The circuitry requires 95mA regulated +5V and -5V. Two regulators to provide these voltages are included on the printed circuit board. The regulators are designed for input voltages in the range of +_ 7 to 11 V. 307 5. G0KLOSIOB The HVL 100 Is found to be suitable for single photon counting when used with the EHI 9924B photoaultipller tube. The discriminator dead time Is found to be Insensitive to variations in PHT pulse amplitude. Due to the PAD's high power consumtlon and compact design, care must be taken to conduct away excessive heat. IC used with a heat sink, the small size of the PAD makes It suitable for rocket payload experiments. 308 APPENDIX A Conipofteyii. lauQu-é 0 i^lde-v^ c^i O © Holt <-o- •iL.^-j. • if cm i c/iac £= r= S/Op 309 APPENDIX B PC board lauou{ &m PAD fioX c_ /o ø M3 1:1 310 6. G0MFOMWT LIST MV100 LeCroy PAD MC10125 MECL to TTL converter 3.3 kri- Discriminator pulse duration set (55 ns) 100 pF Discriminator pulse duration set (55 ns) 1.2 kn. or 5 kC pot. for threshold control 3.9 ka 1 nF Threshold control voltage decoupling 5 x 10 nF Decoupling 3 x 56ii. Input and output resistors 2 x 100 nF AC-coupling 390r- AC-coupllng filter 15 pF AC-coupling filter 2.2 uH AC-coupling filter LM337 L Negative voltage regulator LM317 L Positive voltage regulator 2 x 220.fi- Supply voltage set resistors 600.T.- Supply voltage set resistor 640-^- Supply voltage set resistor 4 x 1 pT Supply voltage decoupling Conhex 51-049-000 Rf connector for PAD input 9 p D-type Output connector and voltage suply Bitnbox 5001/11 (50 x 50 x 32 imn) 311 A ROCKET PHOTOMETER EXPERIMENT TO STUDY OPTICAL EMISSIONS FROM ARTIFICIAL ELECTRON BEAMS IN THE IONOSPHERE Karl Maseide, Department of Physics, University of Oslo, Norway ABSTRACT. A rocket photometer experiment to study the opti cal emissions generated by intense electron beams introduced into the ionosphere is described. An electron accelerator will eject 8 keV electrons in bursts of different fluxes into the ambient, and four photometers will measure the intensities of the Ns2P(3805A), the Nz • 1NU278A) , and the Hp(4861A) emissions generated in the near surroundings of the accelerator together with the rotational temperatures of the Nj * ions. Both the accelerator and the photometers will be located in a small "daughter" payload which will be separated from the main vehicle on the upleg, and is expec ted to obtain an altitude of some 380 kilometers. This experiment was successfully flown from Andøya Rocket Range, Norway, on 10. November 1985. INTRODUCTION Several electron accelerator experiments have been performed in the upper atmosphere since 1969 to study atmo spheric and magnetospheric parameters and processes (Winck- ler, 1980). The results of such experiments, as well as of related laboratory experiments, have shown that non-linear effects, such as beam-plasma discharges (BPD), may be impor tant in transferring energy from intense electron beams to the ambient plasma. The very nature of the beam-plasma interactions is, however, complex, and is still far from understood (Bernstein et al., 1982; Mæhlum, 1983; Szuszcze- wicz, 1985). Optical measurements have frequently been an impor tant part of such experiments, either to study the artifici ally produced auroral-like emissions per se (Grandal et al., 1980) or in using the electron beams as probes for distant magnetospheric research (Winckler, 1982). Ground-based measurements by TV's have thus been quite useful for geome trical tracing of the electron beam dissemination in the ionosphere. Intensity measurements of the optical emissions, however, are not easily made by ground-based techniques, as 312 the light emitting regions are only some tens of meters across, and the beam pulses are usually only of milliseconds duration and are not visible by eye from the ground. Rocket-borne photometer experiments have shown that electron beams of moderate energy (T>2.5 keV) and fairly low currents (<^ 6 mA) have generated optical emissions in accor dance with expected intensity values up to 220 km altitude, as reviewed by Grandal et al. , 1980. Beams of higher ener gies (8-43 keV) and fluxes (12-100 mA) have, however, appar ently produced more intense emissions than should be expec ted at altitudes above some 150 kilometers. Grandal et al. (1980) found that a 90 mA beam of 8 keV electrons generated a local emission of N2'(3914A) that was proportional with the ambient Nz density up to some 130 km. Above that altitude the intensity was higher than expec ted, and a fairly constant level was recorded from 150 km on and up to the rocket apogee at 220 kilometers. A similar finding had previously been made by Israel- son and Winckler (1979), who believed that a leakage of nitrogen from the rocket attitude control system could be the reason for the unexpected high intensity of N2'(3914A) measured above 165 km in their case. Grandal et al. (1980), who had their experiment on a small "daughter" payload separated from the main rocket, proposed that the extraordinary high photon emission rates was due to a heavy energy loss from the beam by a beam- plasma discharge near the payload and a subsequent electron acceleration giving rise to further ionization and excita tion of the ambient Nz . They could, however, not explain the lack of variations w?ch altitude of the observed emission above 150 kilometers. In the same rocket experiment, Polar 5, MSseide (1979) found that also Hp(4861A) was apparently emitted from the near surroundings of the accelerator payload. The reason for this unexpected finding, however, could not be establi shed as only a fairly short sequence of data ( -45 s) could be obtained in that case. Similar experiments in laboratories have now shown 313 that rather complex beam-plasma interactions may occur in the near surroundings of an active electron accelerator in the ionospheric plasma (Bernstein et al., 1983a and b; Konradi et al., 1983; Hallinan et al., 1984). Further exper iments, however, are needed to obtain a detailed understand ing of the observed phenomena. One main problem to study is the criteria for BPD ignition in the ionosphere (Szuszcze- wicz et al., 1982), and further to stuCy what happens during such conditions (Jost et al., 1980; Sharp, 1982; Arnoldy et al., 1985). In such experiments the beam parameters have to be varied, and the several resulting effects of the beam, such as differently excited optical emissions, have to be measured. The presently described photometer experiment is constructed as a part of a comprehensive rocket experiment, MAIMIK, to study the several effects of energetic electron beams introduced into the ionosphere over a wide range of altitudes. THE MAIMIK EXPERIMENT IN GENERAL The general aim of the MAIMIK experiment is twofold: 1) to study the energy transfer from a fast electron beam to the ionospheric plasma for various beam currents and plasma conditions, and 2) to study the extent and intensity of the variety of plasma processts created by the beam and how these effect the neutralization of the electron accelerator. Project scientists for the experiment are: Bernt N. Mæhlum, Norwegian Defence Research Establishment (NDRE), and Nelson C. Maynard, Air Force Geophysics Laboratory (AFGL), USA. The payload is of a "mother and daughter" type which will be separated at the upleg part of the trajectory. The "daughter" payload carries the electron accelerator and the photometers together with a few other instruments to study the local effects of the beam. The "mother" is instrumented for more distant observations of the beam effects. Both payloads are eqipped with aspect sensors for attitude infor mation, and high speed telemetry will provide a fairly high 314 sampling rate of the instruments. The total scientific instrumentation is listed in Table 1. A Terrier/Black Brant VC rocket will take the pay- loads up to an altitude of some 380 kilometers from Andøya Rocket Range, Norway (geogr. 69.3°N; 16.0°E), and the flight will occur into a quiet ionosphere at clear sky and new moon conditions to provide ground-based observations of the sky. The accelerator is programmed to eject 8 keV elec trons in 13 ms bursts, once every second, from 160 km alti tude on and throughout the flight. The ejections will be at 90° to the payload axis resulting in different pitch angles for the electrons as the payload is spinning and coning. The beam current will be varied at 6 levels from 20 to 800 mA, fixed at each pulse, to study the non-linearity expected in the pulse effects. (The electron energy is selected to be similar to the energy of auroral primary electrons, and the beam currents are varied from high auroral flux levels and upwards.) Jan Trøim, NDRE, is the responsible scientist for the electron accelerator experiment. THE PHOTOMETERS The photometers will measure the following atmosphe ric emissions expected to be produced by the electron beam in the near surroundings of the accelerator: N22P(0.2) at 3805A, N2*1N(0.1) at 4278A, and Hp at 4861A. Two photome ters, being separated in wavelength by 20A, are allocated to the study of the N2* (4278A) emission to measure the local rotational temperatures of the N2 molecules versus altitude in addition to the intensity measurements. The main data for the filters to be used are: Emission Filter data N-2P(3805A): r-. = 3802A, Hw = 24A, T = 41% 2 o max N.+ 1N(4278A)1: t. = 4272A, Hw = 20A, T = 42% 2 max + I:i: T = 31% N.2 lN(4278A) : I = 4252A, Hw = 21A may T Hp(4861A): t = 4861A, Hw = 13A, max = 49% 315 The four photometers are mounted in one package, Figure 1, located just beneath the electron accelerator and facing the same azimuthal direction as the electron beam outlet. The optics, however, are individually tilted a few degrees to focus on a minor part of the beam at about 1 m distance outside the payload skin. The field of view is 8° circular full width for all channels. The photometers are calibrated for intensity measure ments in Rayleigh units in a similar way as for auroral studies, i.e. assuming that the field of view is homogen eously filled by the light emitting region. This assumption might, however, not be perfectly fulfilled in the present experiment as the electron beam is expected to be modified more or less by several mechanisms depending on the beam intensity, the rocket altitude, and the pitch angles of the electrons, see e.g. Grandal et al., 1980. An undisturbed narrow beam of 8 keV electrons will form a helix around the magnetic field with a radius of approximately 6.5 m or less depending on the pitch angle of the ejected electrons. A major problem is thus foreseen in the data analysis for this experiment, a_ the geometrical extent and spatial distribu tions of the several optical emissions might be different and difficult to estimate. The sensitivity settings of the instruments was a difficult task as only a few calibrated photometers have been flown in similar experiments up to now, and unexpected photon emission rates have been detected in all cases (O'Neil et al., 1978; Måseide, 1979; Grandal et al., 1980; Bernstein et al., 1983b). In ordinary auroras the selected Nz and Nz ' emissions froir. a certain volume are expected to be proportional to the number density of Nz and to the electron energy dissipation in that volume (see e.g. Vallance Jones, 1974). In the present experiment the electron flux will be varied from 20 to 800 mA at the different pulses, and the number density of Na decreases by 6 orders of magnitude from 100 km altitude to the rocket apogee at 380 km. When BPD's occur, the energy dissipation near the accelerator is furthermore expected to 316 increase by an unknown factor. The mechanism for production of the Hfl like emission seen in the previous experiment of this type, Polar 5, is still not known, and the intensity which might be expected in the present case is thus almost unpredictable. For these reasons the dynamic ranges of the instru ments should be as wide as possible, and they should ideally cover at least 8 decades. This huge dynamic range, however, is not obtainable in a simple way, as we also wanted to have instruments with fairly fast and linear responses to be able to study the "shape" of the different light pulses. The dynamic ranges which actually could be obtained with these requirements fulfilled, were 4 to 5 decades. The photomultipliers are run in pulse-counting modes, and the telemetry provides 16 bits words to be used for the signals at a sampling rate of 1221 s-' per channel. At this rate 16 samples can be obtained per electron pulse of 13.2 ms duration, which should provide an adequate temporal study of the resulting optical emissions to be done. Further details of the photometer electronics are given by Sten (1986) in the next following paper in this issue. SELECTION OF EMISSIONS The First negative system of N_ , B 53 ~ X £ , is generated by simultaneous ionization and excitation of N2 molecules by fast particle impact, when it is produced in a dare atmosphere, i.e. in the present case: 1 + + 2 + M2(X Eg ) + e - N2 (B 53U ) + e + e. The cross section for this reaction is fairly well known and covers a broad range of electron energies from 18.7 eV on and well beyond the 8 keV of the electrons from the MAIMIK accelerator. See Figure 2. Since the emission is fully allowed, the photon emission rates of the system is also proportional to the excitation rate (see e.g. Vallance Jones, 19741 . 317 The strongest bands of this system, the (0.0) at 3914A and the (0.1) at 4278A, are both fairly free from contaminations in auroral spectra and can easily be isolated by filters. For these reasons they are also widely accepted as intensity references in auroral research (Vallance Jones, 1974). Their relative intensities are constantly 1.00:0.30. I.i electron accelerator experiments of similar kind as the present, the (0.0) band at 3914A, has usually been measured. We felt, however, that the 4278A emission could be quite as convenient to study as good filtering is more easily made in this region, and the intensity is expected to be high enough. The intensity distribution on the rotational lines of molecular bands depends upon the temperatures of the emit ting molecules (Herzberg, 1950). The selected N2'IN band consists of only two branches and is further "clean" enough to be used for temperature studies in a fairly simple way. By using two photometers with properly selected filters, one at the P- and the other at the R-branch of the band, it is possible to derive the rotational temperature of the emit ting molecules simply from the ratio of the two signals. See Figures 3 and 4. This technique has been used in several experiments to study auroral temperatures, see e.g. Hunten et al., 1963, and has also been used to study atmospheric temperatures by electron beam injections (deLeeuw and Da- vies, 1972). The assumptions are then that the Nz molecules are in thermal equilibrium with the ambients in their ground state, and that the excitation process does not disturb the rotational distribution of the molecules. These assumptions are assumed to be fairly well fulfilled for the N2 ' IN emiss ions at night-time conditions in the upper atmosphere (Om holt, 1971; Vallance Jones, 1974). In the present experiment it should thus be possible to measure the rotational temperature of N* locally along the rocket trajectory over a range of altitudes from approx imately 100 to 380 km in an undisturbed atmosphere as illu strated in the Figures 3 and 4. The Second positive system of N2 , C3 ll„ - B3 Ilg , is 318 permitted, and only direct excitations from the electronic ground state of Hi. X'r0* , is important as for the N2 • IN system, see e.g. Vallance Jones, 1974. The limit for exci tation is 11 eV and the cross section is strongly peaked at 14 eV. Electrons of energies above 100 eV have almost no effect in producing the Nz2P system unlike in the production of the IN system of N2 ' . See Figure 2. That means that in the present experiment only secondary produced electrons in the beam surroundings, or electrons belonging to the return currents expected to be set up to neutralize the payload during or after ejection of electrons from the accelerator should be able to produce the N22P emissions. It would be of great interest to study the populations and effects of these low energetic particles produced in the near surroundings of the payload. We decided to measure the N22PI0.2) band at 3805A for this purpose, as it is located in a spectral region where it can be fairly well isolated by an interfer ence filter. (Contamination from the N2VK (2.12) band at 3767A should not be serious in the present case as the time constant of that emissions is *»2 s.) The selection of Hfi(4861A) to be studied by the fourth photometer was mainly based upon the observations made by the author in the Polar 5 rocket experiment as mentioned above (Måseide, 1979). In that rocket, however, the Hp photometer was located in the "mother" payload and could only see the light emitting region around the "daugh ter" during a small fraction of the flight. The reason for the Hy-like emission from this region could thus not be established. It was suggested that Hf3 could either be due to direct excitation of atmospheric hydrogen atoms by the energetic electrons, or that it could be produced by disso ciative exc-.tation of H2 and other hydrogen containing molecules in the near surroundings of the accelerator pay- load. One candidate here could be water vapour outgassing from the payload itself as was assumed to be the case during the EXCEDE rocket experiment (Kofsky et al., 1983), i.e.: H2 0 + e - H« + OH + e 319 This reaction is rather efficient in producing Hfj, but one should not expect to have much outgassing of HiO from a small "daughter" payload compared to that reported from the fairly big rocket EXCEDE. We included an Hp photometer with a fairly narrow filter in the MAIMIK "daughter" payload to try to find out of these problems. POSTSCRIPT The MAIMIK rocket was flown from Andøya Rocket Range into an undisturbed atmosphere on 10. November 1985 at 1856 UT. The rocket reached an altitude of 381 km, and a lot of interesting data was obtained. The photometers did work properly, and gave fairly good responses to the accelerator pulses. The data obtained will, however, be presented elsewhere. ACKNOWLEDGEMENTS The author is greatly indepted to Torleif A. Sten who was in charge of the electronics part of the experiment, and to Bjørn Fjeld and Finn Hostad who made the photometer housing. 320 REFERENCES Arnoldy, R.L., C. Pollock, and J.R. Winckler, The energiza tion of electrons and ions by electron beam-: injected in the ionosphere. J. Geophys. Res. 90, 5197-5210, 1985. Bernstein, W. , P.J. Kellogg, S.J. Monson, R.H. Holtzworth, ana B.A. Whalen, Recent observations of beam plasma interactions in the ionosphere and a comparison with laboratory studies of the beam plasma discharge, in Artificial Particle Beams in Space Plasma Studies, ed. B. Grandal, pp. 35-64, Plenum Press, Mew York, 1982. Bernstein, W., G. Mantjoukis, F.H. Leinbach, and T. Halli- nan. Optical measurements of a large-scale labo ratory BPD, in Active experiments in space, ESA Report SP-195, ed. W.R. Burke, pp. 177-180, Noordwijk, Netherlands, 1983a; Bernstein. W., J.O. McGarity, and A. Konradi, Electron beam injection experiments: Replication of flight observations in a laboratory beam plasma dischar ge. Geophys. Res. Lett., 10, 1124-1127, 1983b. Borst, W.L. and E.C. Zipf, Cross section for electron impact excitation of the (0,0) First negative band of N2 • from threshold to 3 keV, Phys. Rev. A, 1, 834-840, 1970. deLeeuw, J.H. and W.E.R. Davies, Measurement of temperature and density in the upper atmosphere using an electron beam. Can. J. Phys. 5J3, 1044-1052, 1972. Grandal, B., E.V. Thrane, and J. Trøim, Polar 5 - An elec tron accelerator experiment within an aurora. 4. Measurements of the 391.4 nm light produced by an artificial electron beam in the upper atmosphere. Planet. Space Sci., 28_, 309-319, 1980. Hallin?:;, T.J., H. Leinback, G. Mantjoukis, and W. Bern stein, Measurements of the optical emission produced during the laboratory beam plasma dis charge. J. Geophys. Res., 8_9, 2335-2347, 1984. Herzberg, G., Molecular Spectra and Molecular Structure. I. Spectra of P:-.-.tomic Molecules. D. van Nostrand Co., New York, 1950. Hunten, D.M., E.G. Rawson, and J.K. Walker, Rapid measure ment of N2* rotational temperatures in aurora. Can. J. Phys. 41, 258-270, 1963. 321 Ir.iami, M. and W.L. Borst, Electron excitation of the (0,0) second positive band of nitrogen front threshold to 1000 eV. J. Chem. Phys. 61, 1115-1117, 1974. Israelson, G.A. and J.R. Winkler, Effect of a neutral tit cloud on the electron charging of an electron beam - emitting rocket in the ionosphere: Echo IV. J. Geophys. Res. 84. 1442-1452, 1979. Jost, R.J., H.R. Anderson, and J.O. McGarity, Electron energi distributions measured during electron beam/plasma interactions. Geophys. Res. Lett., 7, 509-512, 1980. Kofsky, I.L., D.P. Villanucci, J.L. Barrett, M.T. Chamber lain, and R.B. Sluder, Experiments on interaction of keV particle beams with the ionosphere, AFGL Technical Report -83-0316, Appendix I, 1983. Konradi, A., W. Bernstein, D.L. Bulgher, J.O. McGarity, and J.L. Winckler, Jr., Initial experimental results from a laboratory size beam-plasma discharge device, in Active experiments in space, ESA Report SP-195, ed. W.R. Burke, pp. 185-188, Noordwijk, Netherlands, 1983. Mæhlum, B.N., Ionospheric modification by electron beams: results from electron accelerator experiments on rockets in Norway, in Active experiments in space, ESA Report SP-195, ed. W.R. Burke, pp. 121-127, Noordwijk, Netherlands, 1983. Måseide, K., In situ measurements of optical emissions generated by an electron accelerator in the ionosphere. In Proceedings of the 7th Annual Meeting on Upper Atmosphere Studies by Optical Methods, eds. 0. Harang and K. Henriksen, pp. 19- 26, TromsB, Norway, 1979. Omholt, A., The Optical Aurora, Springer-Verlag, Berlin- Heidelberg, 1971. 0'Neil, R.R., F. Bien, D. Burt, J.A. Sandock, and A.T.Stair, Jr., Summarized results of the artificial auroral experiment, Precede, J. Geophys. Res. fL3, 3273- 3280, 1978. Shemansky, D.E. and A.L. Broadfoot, Excitation of N* and N2' systems by electrons - I. Absolute transition probabilities, J. Quant. Spectrosc. Radiat. Transfer, 11, 1385-1400, 1971. Sharp, W.E., Suprathermal electrons produced by beam-plasma- discharge, Geophys. Res. Lett., 9, 869-872, 1982. 322 Sten, T.A., A brief description of the electronics of a four channel rocket photometer and the check-out equipment. This issue, pp. 328-334, 1986. Szuszczewicz, E.P., Controlled electron beam experiments in space and supporting laboratory experiments: a review. J. Atmos. Terr. Phys., 4J7, 1189-1210, 1985. Szuszczewicz, E.P., K. Papadopoulos, H. Bernstein, C.S. Lin, and D.N. Walker, Threshold criterion for a space simulation beam-plasma discharge. J. Geophys. Res., 87, 1565-1573, 1982. Vallance Jones, A., Aurora. D. Reidel Publishing. Co., Dor drecht, Holland, 1974. Winckler, J.R., The application of artificial electron beams to magnetospheric research, Rev. Geophys. Space Phys., 11, 659-682, 1980. Winckler, J.R., The use of artificial electron beams as probes of the distant magnetosphere, in Artifi cial Particle Beams in Space Plasma Studies, ed. B. Grandal, pp. 3-33, Plenum Press, New York, 1982. 32 i Table 1. ROCKET INSTRUMENTATION Instruments Responsibility "Daughter" payload Electron accelerator J.Trøim, NDRE Multi frequency transmitter M.Friedrich/W.Riedler. TUG Retarding potential analyzers E.V.Thrane/J.Trøim, NDRE Photometers, 4 channels K.Måseide, UiO "Mother-Daughter" connection wire B.N. Mæhlum, NDRE "Mother" pay load Solid state TV B.N.Mæhlum/B.T.Narheim, NDRE Capacitance probe M.Friedrich/W.Riedler, TUG Ion density probe G. Holmgren, UJO Electron temperature probe J.Trøim, NDRE Fast particle spectrometer D.Winningham, SwRl Wave receivers A.Egeland/J.A.Holtet, UiO DC electric field N.C. Maynard, AFGL Abbreviations: AFGL - Air Force Geophysics Laboratory, USA NDRE - Norwegian Defence Research Establishment SwRI - Southwest Research Institute, USA TUG - Technische Universitåt Graz, Austria UiO - University of Oslo, Norway UJO - Uppsala Ionospheric Observatory, Sweden. ni* Figure 1: The four channel photometer constructed for the MAIMIK "daughter" payload. The dimensions are approximately: 10x10x31 cm. 325 ioV T—r ITT T—I I I I I II 1 1—I IIIIU l\L+(4278Å) NJ3805Å) i • • •" j 11111M J 111111 101 10a 103 10< Electron energy (eV) Figure 2: Emission cross sections for electron impact excita tion of the Nz'lN(O.l) band at 4278A and the NJ2P(0.2I band at 3805A. The cross sections are derived from the values obtained by Borst and Zipf (1970) and by Imami and Borst (1974) for the re spective (0.0) bands, by using the transition probabilities of Shemansky and Broadfoot (1971). 326 Figure 3: Rotational intensity distribution of the N2M4278A> band at 200 and 1200K, :..'. = 1A, together with the filter transmission profiles. Contributions from the H2'1N(4237A) band is also taken into account, and the wavelengths are adjusted to vacuum condi tions . 327 400 A Tmean,CIRA,1972 ?300- 0 y = 200- ^_^y < 100- 0- 1 1 1 i 1 r __... J 200 400 600 800 1000 1200 200 400 600 800 1000 1200 T«mptratur« (K) Figure 1: A - Reference atmosphtre temperatures, CIRA 1972. B - Fractions of the N2 '(4278A) emission transmitt ed by the two filters vs. temperature. Their ratio defines the rotational temperature of N2 dirctly. 328 fl BRIEF DESCRIPTION OF THE ELECTRONICS OF A FOUR CHANNEL ROCKET PHOTOMETER ANP THE ChTD--OU7 EOUfPMEM T.fi. Sten, Dec ^rtiT.en t o* Ft.^ics, University at Usla, Norway 1 . INTRODUCTION The photometer design c iten Hi gl t i .1-? " ETiol ut i on i, i- me» i ni y a matter of t el emetry rrapasity but it also implies the use o-f a photon count i ny 5,-ite'T). IJ'I this case each photo.nei.er channel was sampled 1221 times pr second, giving 16 samples/channel during the gL• n pulse being i ""•. 2 ms wide. Concerning the dynamic conge each photometer channel ~oS given a iå-bit word in the telemetry -format, which was iTH'e t^ ^ ". adequate. The In-n tation lies in the pulse amp li- r; er sr,^ discriminator circuits being paralyzed at output co.mt * ate above appro-; i mat el y 1.5-10* conn ts/seconds. B / -j sing ECL circuits the dynamic range coi-id easily '-•-<=: bee,J s:--:tendeJ to 107 .;ouiits/-, hut at the e.:pence o-f .intolerable increase in power consump t i ori. The presented r-jlse amplifier and diser i mi. nat or (se^ Fig. •-•) operates at 5ppro- i mat el / 80 ir.W, whereas ECL circuits would i-equ.re aba -j t 1 wat * 'channel . Regar Jl ess of power restrictions, another -four watts would cause -_m waited i r.c t eaf.e c-f temper a.- ture in the photornul t lpl i er utt'S with the present mechani- . ?1 desi gn. 2. ru0T0r1E"rFo F'. FCTr:0N ! OF Ce = cr .pticti ;:•? the phc t-ijieter --?! ec t ^or i L - ha* bet- divided in'r.. tu:. parts: 2. 1 The ^-titt- section v* i th th- p^I^e- amp: >-f ; er end d i '= c f : -T- L i * c or s < See Fig. i < . 329 2.2 The puls counting system with sequence generator for data read-out and house-keeping (see Fi g. 2). 2.1 Ths PH-tubes 'EMI 9^24) were operated with the cathode at ground potential and with voltage dividers having S20 VSi between the dynodes. Dynode No t was connected to ground via a 150 volt zener diode. Two DC -DC c on ver ter s (Tecne t i c s 95583-115) oper ~h;r!d. The shi-^Idin^ proved to give sufficient protection a.Toi-.ct ncir-e from the electron aceel 1 erators being located c^r,-™? to the photoms ter. The pulse amplifiers .-and discriminators (Fig. 3) were mounted or. two circular boards (42 mm diam.) end fitted to the ^-tubs socket nitti two screws. Thi ^ assembly was not potted but simpl -7 gi ued to the wal 1 s of the FM-tube housi ng b . RTV 102, whi c h proved to give adequate en'vi ronmental pro- te-tion. The circuit diagram in Fig. 2 should be considered as -p idea for further refinements rather than a final solu tion. Even if it has proved to work satisfactorily, great care should be taken in the discriminator design (LM 306) and tte coupling to the pulse counters. The goal is to get well der i ned out-put pulses of full size <0 to VDD> especa- 3lly at high count-rates. In this experiment a Schmitt- tr:gger <4093) and a pui1-up resi stor prevent pulses, terminating around VDD'2,, from laterally Ml ling the pul =e counters \4404>. A light emitting diode was mounted near the base of each PM-tube for generation of "calibration pulses", for check-out of system gain. 2.2 The pulse counting system is straight forward, using two 8-bit counters data in each channel. Pulses -for counter reset, shiftregister parallell to ser i al control and clacking were provided by the sequence generator in accordance with the telemetry -format, organized as fal 1ows: - Frame length: 12S words o-f B bits (305.2 frames/s). - Format length: 32 frames (9.5 formats/s). Each photometer channel was super commutated 1:4 within one frame, giving a samp ling rate of 1221 samp 1es/s/channel. The sequence genrator was syncronized to the PCM encoder, and interconnected via optocoup1ers, by means of: - 2 CP the double system clock frequency -'625 kbi ts/s< . - FP frame pulse, appearing at word 00 in each frame. - SS sub-frame sync, every 32 frame at word 00. Apart from control 1 ing the pulse counting system the sequence generator provided pulses far a temperature moni tor CO to + 32°CJ and in-flight calibration circuits CLED's) as well. These house-keeping data occurred for 2 frames every 64 format (6.7? s) and were lacked to the PCM format for three reasons: - ensure that they would not interfere with data related to the electron gun pulses. - enable separation of house-keeping data from scien tific data. simplify and check-out of the photometer. A control bit, C-bit in Fig. 2, was used to change the counter reset interval from normal 1y being 16 frames to 2 frames during calibration of the photometer. T. THE CHECK-OUT EQUIPMENT Thi s part of the exper i ment was designed to capture the calibration pul ses, 2 frames wide (13.2 ms) at a repeti tion rate of 64 formats (6.75 s>, store the information, and display it on an oscilloscope. During the flight the same equipment was used for quick-loot-, displaying data related ta the gun-pulses as wel1. In principal, data ar ound one gun- or calibration- pulse was stored and resirculated until the next pulse occurred, or one could simpIy "freeze" the display by giving HOLD command. 331 Storage of data was initiated by a pre-cursor pulse and terminated after 128 frames. This window probably cavers the mast interesting events around a gun-pulse, at least for qui cl-1 ook purposes. Only two of the four channels could be selected at a tiiTe. Two memories <2K :< 8) have been used, one for storage of data in real time whi1st the other, containing data from the previous pulse, was re-circulated in READ mode. The 16- bit data words were transferred from the READ memory to bit- shift circuits which enabled selection (thumb-wheel switch es) of the ei ght most relevant bits. A pair of B-bit D/A converters provided two analog outputs together with a sync pulse for scope display. In WRITE mode the memary pointer was incremented for every selected byte, i.e. 2 channels of 2 bytes each sampled 4 times in one frame equals 16 bytes/frame. In READ mode the memory pointer was incremented at the system word-rate (129 bytes/frame), i.e. eight times faster. The pre-cursors for the gun-pulses occur at intervalla of 256 frames where only half of the data have been recorded. This means that the memary was re-circulated sixteen times between each up-dating and a flicker-free display was ob tained. The memory and read-Dut card described above was used together with: - a PCM encoder simulator card for bench-test and cali bration of the photometer, - a PCM decoder card for check-out through the rocket telemetry system. This card received a split-phase signal from the tele metry and contained bit- and format-syneronizers for regene ration of control signals and data. A precursor for the gun- pulse was not included in the PC:1-f ormat. However, a flag- bit for the gun-pulse was used to reset a frame-counter which produced a pre-cursor 9 frames in advance. In this way thp first gun-pulse was lost but the rest of the f 1lght was wel1 covered. FIELD STOP HIGH VOLTAGE DC-DC CONV. LENS INTERFERENCE FILTER PHOTOMULT|PLIER EMI 9924 PULSE AMPLIFIERS AND ©"+ DISCRIMINATORS SILICON MAGNETIC RUBBER SHIELDS +&- I HIGH VOLTAGE LED X DC-DC CONV. CALIBRATION -+ PULSES Fig.1. PM-tube assembly. £L£L *r COUNTER RESET —7^ DE-MUX^/* m R PULSE COUNTERS GATED CLOCKS. CLK P/s SHIFT REGISTERS TEMPERATURE -J MONITOR / / / K SEQUENCE CALIBRATION PULSES GENERATOR C-BIT NRZ OUT • 15V"*- VOLTAGE + 8V «« OPTOCOUPLERS REGULATORS -12V M Fig. 2. •H8vl -18vl Data and control system. PCM-ENCODEu R 334 > 5 >*-$åJ)$ o (0 o (0 c (0 a E (9 0 (0 CO ei) u. 335 A real-time system for measuring the altitude of auroral structures Ake Steen Kiruna Geophysical Institute P.O. Box 704, S-981 27 Kiruna, Sweden Abstract The altitude and the altitude distribution of auroral emissions are one of the results of processes in the iono sphere and in the magnetosphere during auroral activity. A bistatic system has been developed to determine the altitude and the altitude distribution of different auroral emissions in the magnetic meridian plane through EISCAT's transmitting station in Norway. The system makes use of intensified one- dimensional photodiode arrays. In April 1985 the first measurements were made. Initial results, system characteris tics and system configuration are reported. 336 1. Introduction The altitude and altitude distribution of auroral emissions provide important information on magnetospheric processes causing the optical aurora in the ionosphere. Near the turn of the century a measurement period began which utilized the parallactic photographic technique. The distri bution of the 12,330 auroral points obtained by Carl Stdrmer [stormer, 1955] represents even today an impressive amount of data. The first documented visual determination of the height of an aurora originates from the period 1726 to 1730 [de Marian, 1726]. A lot of work in this field has also been carried out since the Stormer period, e.g. Brandy and Hill [1964]. They used all-sky cameras for the height determination of the auroral luminosity. The usefulness of the all-sky camera for a three- dimensional mapping of the aurora has recently been demon strated by Kaila [1981; also paper presented at this confe rence]. A pair of meridian scanning photometers (MSP) was used by e.g. Komick and Belon [1967] for triangulation of the + 391.4 nm N2 emission and a stereo TV-system has been used by Stenbaek-Nielsen and Hallinan [1979] to determine the vertical extent of pulsating aurora. The height measuring system (HMS) we report here has been developed with the aim to combine the advantages of the previous methods, and also to add a real-time recording feature. The objectives of HMS are to obtain the altitude and altitude distribution of auroral emissions with 1 second time resolution and 1 km height resolution. In contrast to the pre viously used imaging techniques HMS provides a one-dimensional monochromatic imaging capability. Although the development work of the system has not yet been completed we report some preliminary measurements and give a technical outline of the system. 337 2. Description of HMS The development work started with the firm belief that a sub stantial part of the incoherent scatter radar measurements made by EISCAT would take place in the plane through Tromsø and Kiruna. The Kiruna-Tromsa plane coincides rather well with the magnetic meridian plane through Kiruna, as defined by Gustafsson [1970]. Figure 1 shows the three EISCAT stations: Tromsø (transmitter at Ramfjordmoen), Kiruna, Sodankylå and the location of the two HMS stations. For comparison the electron density profiles obtained by the Tromsø station in the common program mode are especially important. Figure 2 illustrates the principle of triangulation for HMS. In the ideal case a narrow isolated auroral arc is located within the intersecting fields of view of the two stations. Station 2 is located in Kurravaara at a distance of approx. 13 km from station 1 at the Kiruna Geophysical Institute. The instruments record primarily the distribution of an auroral emission along the magnetic meridian. For the ideal case two angles (aj and oi2 ) are obtained from an identifiable point in the arc, such as the lower border. Approximations to the alti tude (h) and the distance from Kiruna (1) can be derived with the formulas given in figure 2. The triangulation procedure will be implemented in a real-time computer program. For several types of aurora this technique will obviously not work, since an optical measurement is an integration along the line of sight. Broad diffuse aurora and multiple arcs will create interpretational difficulties. It is therefore intended to supplement the bistatic measurements with data ftom other imaging instruments operated in northern Scandinavia. However, other types of aurora will be suitable to study with this technique, such as stable relativly narrow single arcs. Break up aurora also, which are often thin, will be suitable. The height and thickness of pulsating aurora will probably also be possible to study. During the spring of 1985 the bistatic HMS was operated in the configuration shown in figure 3. The stations were equipped with optics for 50° field of view (HMS2) and 90P field of view 338 (HMS1). The 90° field of view permits observations in the magnetic zenith and gives a reasonable coverage along the magnetic meridian. The overlapping field of view is a little less than 50°. In figure 3 the angle differences (a2_ai) are given for a nominal aurora at 100 km height for five distances (1) from Kiruna. It is anticipated than an angle difference of approx. 1° is needed. HMS will therefore obtain height infor mation from a distance interval of about 200 km or 1.7 degrees of latitude. The detector system used for the HMS was described at the 12th annual optical meeting in Stockholm, 1984 (see also Steen, 1985). The light detecting unit is an intensified one-dimen sional photodiode array of 1024 pixels (Reticon RL 1024 SF). So far only HMS1 has been supplied with a filter wheel. Change of filters in HMS2 has to be made manually. Permanent telephone lines connect the remote station to HMS1 (see figure 4). Locally each station is organized through an IEEE-488 bus. The controlling computer synchronizes the read outs from the arrays by sending start and stop commands simul taneously (to within a few ms) to HMS1 and HMS2. Units on the HMS2 bus are as easily accessed as units on the HMS1 bus through the 9600 baud link. The present storage medium for IMS is a start/stop streamer with a 67 Mbyte formatted capacity. With a cycle time of 1.5 s one cartridge tape lasts for about 2.5 nights. Tne 1 s time resolution has not yet been achieved mainly due to software limitations. 3. Initial results and future plans The first observations with HMS were made in March/April, 1985. Only four nights of bistatic operations were accom plished due to an unrealiable start up sequence in an asyn chronous/synchronous converter belonging to the modem connec tion. Figure 5 represents an example of raw data from HMS. Two profiles are shown obtained with 90° and 50° fields of view. The aurora was a diffuse auroral structure above the northern horizon. The profiles are not corrected for the different opening angles and flat field. No scientific conclusion can be made from the raw data. 339 For the observing season 1985/86 the modem connection will be made reliable. Several safety systems have been incorporated which permit a computerized start and stop operation. Software improvements should produce a reduced cycle time reaching the 1 s target and a real-time algorithm for the height calcula tion. Of primary scientific interest is the comparison with electron density profiles obtained by EISCAT for a stable discrete auroral arc. That comparison will be a test of the value of the technique. Scientific topics of interest are also the westward travelling surge, pulsating aurora and optical emissions produced by the heating facility at Ramfjordmoen, Tromsø. 340 References brandy, J.H., and J.E. Hill, Rapid determination of auroral heights. Can. J. Phys., 42, 1813-1819, 1964. de Mairan, Traité physique et historique de 1'aurore boreale, 1726-1730. Gustafsson, G., A revised corrected geomagnetic coordinate system, Ark. Geofys,. 5, 595-617. 1970. Kaila, K., Three-dimensional mapping of the aurora from digi tized all-sky pictures, Tech. Rep. No 25, 1981, Finn. Met. Inst. Helsinki, Finland. Romick, G.J., and A.E. Belon, The spatial variation of auroral luminosity - II, Planet. Space Sci., 15, 1695-1716, 1967. Steen, A., Intensified photodiode array applied to one-dimen sional auroral imaging. Rev. Sci. Instrum., 56, 116-120, 1985. Stenbaek-Nielsen, H.C., and T.J. Hallinan, Pulsating auroras: Evidence for noncollisional thermalization of precipitating electrons, J. Geophys. Res., 84, 325-327, 1979. Størmer, C, The Polar Aurora, Clarendon Press, Oxford, 1955. 341 Figure Captions Figure 1 The magnetic meridian plane through Kiruna coincides almost with a plane through the EISCAT stations in Tromsø and Kiruna. HM3 is a bistatic system obtaining optical measurements along the same meri dian. HMS1 is the station at the Kiruna Geophysical Institute. HMS2 is the remote station in Kurravaara. Figure 2 The principle of measurement for the Height Measuring System (HMS). Triangulation from Kiruna is favourable when auroral structures are located above Ramfjordmoen, the EISCAT station near Tromsø. For mula 1 and 2 are only approximate and are not corrected for the earth's curvature. Figure 3 Mode of operation during late spring 1985. The nominal angle difference is the difference between a2 and al using the formulas in figure 2 for an altitude of 100 km. The latitude coverage of HMS is about 1.7 degrees of latitude. Figure 4 Block diagram of the electronic units of HMS. Most of the units belonging to HMS1 are represented. HMS2-units are only represented as one box. A detailed description of HMS1 is given by Steen [1985]. Figure 5 An example of HMS raw data. An auroral event on April 3, 1985, observed with HMS. The upper panel represents the distribution of the auroral emission 427.8 nm along a magnetic meridian, measured by HMS1 with 90° field of view. The lower panel illustrates the same auroral structure out measured by HMS2 with 50° field of view. The curves are obtained with 1.55 s integration time. The figure represents raw data and no scientific conclusions should be drawn. 342 MAG. MERIDIAN GEOGRAPHIC NORTH W TROMS6< KIRUNA SOOANKYLA 50 km Finurc 1 343 . FiaO LINE 200 / RAH: RAMFJOROMOEN. NORWAY f KUR KURRAVAARA KIR : KIRUNA 3 150 W lon ai ton a2 ,. , h"' tona, «.na, l . d . ^4l~, (km) (2) tonax-tonai 1:13 (km) (3) o • IM (km) W Fiaure 2 6.3* Nominal angi* 1 dif fortne* 200 13 0 TROMSO HMS2 HMSI Geographic latitude 69 6* N 67.9'N Figure J 345 BUS CONTTKXIER TEMP SENSORS f Fiourp 4 346 a393 9 155 HMS1 .....-••" 11B?" U IBB 5 - IB? ...—-'• ... •-. (4 -J35 ..'"' 90° FoV 1 _-—- """ 41136 11549 ,-—""~ *"""'"-— B5BUU—'- ' ^~—^'~' ""~--.., - B 9399 ...^ 155 HMS2 ..-'" 3?? i I B - ' ! IBB •v'1""'-- 5 ; 50° FoV .35114 «i Fiqure b 347 ALTITUDE OF THE RED LOWER BORDER IN AURORAL FORMS Kari Kaila Department of Physics, University of Oulu 90570 Oulu, Finland Abstract In very active forms the lower border of aurora can be red. It is from Hp IPG emissions. Earlier authors have obtained an altitude of 60...100 km for the border where the green colour turns to the red one. In this work, three substorms of Jan. 30, 1979 were studied to deter mine the altitude of the red lower border. A value of 85 ± 2 km was obtained from colour film data of four different all-sky camera stations in Northern Finland. 1. INTRODUCTION Red colour is sometimes observed in the lower border of intense and active auroral forms. There are three different explanations for this colour which originates from N2 1PG bands. The best explanation seems to be the diminuation of 01 557.7 nm line by the quenching of S state which has a lifetime of about 0.7 s, or because of the decrease in the densities of 0( S) precursor species with dep in the atmosphere. The red No 1PG emission dominates at low altitudes giving rise to the visible red lower border of auroral forms. Stbrmer (1955) and Chamberlain (1961) have described this kind of aurora as B-type aurora. Stbrmer made also altitude determinations of these auroral forms and obtained a result of 60- 70 km, or at least lower than 80 km. In a more recent paper Gattinoer and Vallance Jones (1979) obtained an altitude of 30- 100 km on the basis of photometric and spectrometry observations. They report about red lower borders not much below 100 km and deep crimson lower borders at 80 km or even below. Later Gattinger et al. (1985) used two TV-cameras and a TV-spectrograph. From 12 alti tude values, which were from 89.6 to 94.9 km,they obtained a mean value of 91.4 ±1.1 km for the red lower border. 348 Benesch (1981) compared spectra of N2 1PG Av = 3 bands obtained in laboratory at a pressure of 1 torr and those obtained from normal aurora by Gattinger and Val lance Jones (1974). The Av = 3 bands are in the wavelength interval of 620-690 nm. In the laboratory spectra bands (7-4) to (11-8) were much more intense and bands (3-0) to (5-2) less intense than in the normal auroral spectra. The laboratory spectra were measured at 1 torr and the pressure at normal auroral altitude of 105 km -4 is of the order of 10 torr. At 80 km, where red lower border auroras exist, the pressure is about 10 torr. The intensity distribution could be explained according to Benesch by the increasing collision frequency which increases the intersystem collisional transfer of excitation between the N, states W and B. This changes the distribution of W and B states and shifts the observable intensities toward shorter wavelengths. The red colour intensity detected by a human eye can increase almost two fold. Benesch concludes that the colour turns red at 85 km. If the changing collision frequency modifies the intensity distribu tion of N2 1PG AV=3 bands as suggested by Benesch (1981) the effect should be possible to be measured in auroras. However Gattinger et al. (1985) did not find any significant difference in the spectra obtained from 93 and 122 km. In this work the altitude of the red lower border has been deter mined using all-sky colour films taken simultaneously at several stations. 2. OBSERVATIONS Three auroral activations occurred in Northern Scandinavia between 20.35 and 21.25 UT on January 30, 1979. Four all-sky cameras were in operation in Northern Finland (Fig. 1) and during this interval the auroral forms were discrete enough for altitude determinations. The lower borders of auroras were digitized at about every second minute from the Ivalo and Kevo camera data and their altitudes were calculated by the iterational method described by Kaila (1981). Several additional altitude determinations were made by the Muonio and Kilpisjarvi camera pair. The cameras used 16 mm Video News Film from Kodak (VNF 7239 Daylight) and the exposure time was 2 s. The sensitivity of the film in the red part corresponds rather well to the sensitivity of the eye. 349 FM I asc-neiwork 1979 DEG EAST Fig. 1. The locations of the Finnish Meteorological Institute all-sky camera stations Ivalo, Kevo, Muonio and Kilpisjarvi. 3. RESULTS The three intensifications of substorro activity occurred at 20.35, 21.07 and 21.18 UT as was indicated also by the Pi bursts recorded at Sodankyla (Bbsinger et al., 1981). After the onsets, the auroras ex panded northward and their altitude decreased indicating hardening of the precipitating electron flux. From Ivalo-Kevo pair the calculated altitude and latitude values at longitude 27.2° E are shown in Fig. 2. Between 20.35 and 21.07 UT the altitudes of auroral forms were from 113 to 97 km and no red lower border was visible. The red colour was identified for the first time on the Kevo all-sky film at 21.09 UT. Its altitude was then 89 km at 350 IVALO-KEVO 1979 030 Fia. 2. Calculated altitude and latitude values of auroras for the time interval from 20.30 to 21.30 UT. The values are from 27.2° geo- graph. long, which is almost along the connection line between Ivalo and Kevo stations. In the bottom panel is shown the esti mated brightness of the aurora using the IBC classification. 27.2° E. It varied from 89 to 95 km in the longitude range from 26 to 29 degrees. After the second substorm the altitude was mainly at 83-90 km until 21.13 UT. From 21.08 to 21.12 UT the expansion of the auroral region took place and at 21.12 UT the northward movement stopped. The arcs be gan to move again southwards and the altitude increased to about 95 km. The third substorm began at 21.18 UT and again the altitude decreased to less than 90 km. On several occasions the red lower border was visible on Kevo all-sky film and at those times the altitude of auroras was from 84 to 89 km. Ivalo camera did not show clear red lower border situations. The reason for that is not known. From Kilpisjarvi and Muonio all-sky films the red borders were visible more frequently. The first indications of the red lower border 351 were at 21.08. From both Muonio and Kilpisjarvi stations red colour was seen from 21.08 to 21.14 UT and from 21.19 to 21.24 UT. The altitude of these emissions varied from 75 to 88 km. Several single altitude determinations were made using the Kilpis jarvi - Muonio camera pair. At 21.09.43 UT strong red aurora was at 76-84 km (Fig. 3). One minute later the colour was weaker and at an altitude of 80-83 km. At 21.12.03 UT a red border was seen in a limi ted longitude range of two degrees. It was below 80 km. Elsewhere the altitude of the aurora was about 85-90 km. MUONIO KILPISJ 1919 030 2 109^3 Fig. 3. The altitude of an auroral arc at 21.09.43 UT from Muonio and Kiipisjarvi data. Strong red emissions were seen between 18°- 23° of geographical longitude. At 21.16.03 UT the arc was at 87-97 km but no red emissions were seen (Fig. 4). Also at 21.19.03 UT there was an arc at 90-96 km with out the red border. 20 seconds later red emissions appeared according to the Kilpisjarvi film. MUONIO KILPISJ 1979 050 2M603 LCMGI nix e Fig. 4. An example of an auroral arc at 21.16.03 UT, where no red lower border was seen although the altitude was 87-97 km. 352 According to the present data analysis the altitude of the red lower border of auroras is 85+2 km. It is our conclusion that the red colour appears typically below 90 km. Some limitations of our method may occur during very active auroral situations as the 2 second exposure time can be too long in such dynamical events. REFERENCES Benesch, W., Mechanism for the auroral red lower border, J. Geophys. Res. 86, 9065-9072, 1981. Bbsinger, T., K. Alanko, J. Kangas, H. Opgenoorth and W. Baumjohann, Correlations between PiB type magnetic micropulsations, auroras and equivalent current structures during two isolated substorms, J. Atm. Terr. Phys., 43, 933-945, 1981. Chamberlain, O.K., Physics of the Aurora and Airglow, Academic, New York, 1961. Gattinger, R.L. and A. Vallance Jones, Observations and interpretation of spectra and rapid time variations of type-B aurora, Planet. Space Sci., 27, 169-181, 1979. Gattinger, R.L., F.R. Harris and A. Vallance Jones, The height, spec trum and mechanism of type-B red aurora and its bearing on the excita tion of 0(1S) in aurora, Planet. Space Sci., 33, 207-221, 1985. Kaila, K.U., Three-dimensional mapping of aurora from digitized all- sky pictures, Finn. Met. Inst. Techn. Rep. No. 25, 38 p., 1981. Stbrmer, C, The Polar Aurora, Clarendon, Oxford, 1955. Gattinger,R.L. and A.Vallance Jones, Qualitative spectroscopy of the aurora,2, The spectrum of medium intensity aurora be tween 4500 and 8900 Å, Can. J. Phys., 5_2 , 2343-2356, 1974. 353 IMAGE PROCESSING OF ALL SKY AURORAL PHOTOGRAPHS B. Lybekk, Institute of Physics, University of Oslo, Oslo, Norway Abstract. The 630.0 nm emission in the dayside cusp region is usually subvisual. Until now the main investigation instrument of large and small sky dynamics in the dayside aurora at Svalbard (Spitsbergen) has been the meridian scanning photometer. The data coming from this instru ment is well suited for computer handling. However, the information outside the scanning meridian is lost. During two seasons an all sky image intensified camera owned by A.F.G.L. Boston, has been operated by our group at Svalbard. The oamera uses optical filters at 630.0 and 427.8 nm and the data are recorded on 35 mm black and white film. Selected events from the 35 mm film data are digitized. The data are stored on tape. Several image processing programs are run on a micro computer system. The quality of the all sky images are remarkably im proved when using different enhancement procedures. Mapping of the all sky images into a geomagnetic coordinate system is performed. The camera and the image processing system are suitable when investigat ing dynamics of the dayside aurora. 354 CALIBRATION OF OPTICAL INSTRUMENTS FOR ABSOLUTE INTENSITY MEASUREMENTS OF UPPER ATMOSPHERIC EMISSIONS Jacek Stegman and Donal P.Murtagh* Institute of Meteorology University of Stockholm Arrhenius Laboratory S-106 91 Stockholm, SWEDEN ABSTRACT: Procedures for the calibration of an optical instrument for measurements of atmospheric quantal emissions are briefly outlined. The different techniques that are necessary in different spectral regions are discussed. As an example the calibration of a 0.6 m Czerny-Turner spectrometer equipped with an imaging photon detector is presented. 1. INTRODUCTION. The purpose of this paper is to present some experience gained while working with the calibration of different optical instruments for rocket-borne and ground-based measurements of airglow and aurora. The wavelength range of our work extends From near IR to VUV. The instruments Are filter photometers, polychromators, scanning spectrometers and recently imaging devices. Only a few points are discussed in detail, most of the subject is only briefly reviewed and references to published sources are given. 2. CALIBRATION PROCEDURES. 2.1. ABSOLUTE CALIBRATION. The purpose of an absolute calibration of an optical instrument is to relate the output signal of the instrument to the photon flux entering the system. This relationship has to be determined over the entire dynamic range of the instrument, which implies a knowledge of the linearity of the system. Furthermore, the spectral responsivit.y of the system has to be measured. These measurements are usually performed as separate steps, for instance the measurement of the relative linearity of the system is normally obtained by using a set of neutral density filters or the inverse square law. Then, in the next step the absolute scale is established. The specific techniques and different possible approaches to obtain the absolute number of photons entering the system are discussed later. * DPM acknowledges support from the European Space Agency in the form of a post doctoral fellowship. 355 2.2. SYSTEM LINEARITY. As mentioned above, the system linearity may be measured, at least within an "extended" visible range, using a set of neutral density filters. Within the range of 420 to 700 nm an absorption type of gelatine neutral density filters may be used. These filters may be quite safely combined with each other as the resulting transmission is simply obtained by multiplication of the transmissions of separate filters. For applications in the range 200 nm to 2500 nm metallic filters have to be used. Greater care has to be taken in that case as the filters reflect the light which is not transmitted easily creating resonance cavities when used in tandem. Another way to obtain the system linearity is to utilize the fact that the irradiance from a point source of light follows the inverse square law CI], The instrument characteristic is established over a part of its dynamic range only, depending on the length of the optical bench available for the measurement. Another, overlapping part of the instrument characteristic is established in a consecutive measurement by adjusting the intensity of the light source. By repeating this procedure the instrument's response is determined stepwise over its full dynamic range. It should be noted that this measurement provides only the relative dependence of the output signal on light input. For photon counting systems however, the linearity calibration may often be eliminated if the response time of the detectors and associated electronics is known and if the character of the response is known C23,C33. Two limiting cases are easy to describe analytically: - NONPARALYZABLE. Every counted event causes the system to be "dead" for some specific deadtime t. Subsequent pulses entering the system are lost and they do not influence the system. The relationship between the measured count rate Rm and the true count rate R is expressed by a simple formula: R» = R/(1+Rt) or R = R„/(1-R„t) - PARALYZABLE. Pulses entering the system during the deadtime t are not counted but they extend the deadtime by another t. Therefore,for very high true count rates the system will count 1 pulse and paralyze. The measured count rate may be expressed as: R„ = R exp(-Rt) Although in practice the behavior of a counting system might be an intermediate case, the systems we have experience of could be described satisfactorily by one or other response. As an example four different pulse amplifiers are listed below with their nominal output pulsewidth and their measured behaviour C43: 356 SPACOM (250 ns) nonparalyzable for 70 ns paralyzable -for following 180 ns SPACOM (50 ns) nonparalyzable for 70 ns AMPTEK (200 ns) nonparalyzable for 200 to 600 ns LeCroy (60 ns) nonparalyzable for 60 ns 2.3 WAVELENGTH RESPONSIVITY. The calibration of an optical instrument would not be complete without a measurement of wavelength responsivity of the system. Although the technique of this measurement seems to be quite simple there are some precautions to be taken: -Interference filters have to be measured with the optics they are intended to be used with and certainly at the temperatures they are designed for. -Great care has to be taken to eliminate spurious light in monochromators. Considerable background radiation is usually produced by emissions at different wavelengths, multiple dispersion and by diffuse reflection of zeroth order £51. -The absolute responsivity of an instrument to a specific molecular feature requires that the spectral sensitivity curve of the instrument is numerically integrated with synthetic spectra of the feature. As such a calculation often requires some assumptions (e.g. rotational temperature) the filter- characteristics should be chosen to minimize the importance of such effects. 2.4. FIELD OF VIEW. The next important step in the procedure is a determination of the field of view. This can be obtained by rotation of the instrument in a parallel light beam with the centre of rotation at the entry pupil. In some applications this step may be omitted as a knowledge of the field of view is not necessary for interpretation of the measurements such as airglow measurements by a photometer calibrated against an extended source of light. However, an exact knowledge of the field of view is required when such an instrument is used to measure light from point sources, e.g. a check of the absolute calibration against standard stars that happen to enter field of view during flight. The field of view of an instrument is defined by its optics. In this connection it should be pointed out that design of the optics should follow some well-known principles. The most important of these is the principle of imaging aperture on aperture and lens on lens. It usually results in a suppression of stray light in the system and compensates for possible non-uniformities of reflecting surfaces or detector photocathode. Consequently, the filter photometers which are still the most versatile devices for measuring the atmospheric emissions should always be equipped with a field lens that images the entry pupil on the 357 photocathode C63. 3. ABSOLUTE CALIBRATION TECHNIQUES. The absolute calibration of an instrument may be accomplished by using the instrument against a known (calibrated) source of light or by a comparison of the instrument with a detector of known spectral responsivity (Table I). The sources and detectors may be divided into primary standards - whose spectral radiant power or spectral responsivity may be calculated, and secondary standards - transfer sources or detectors which are stable enough and can be controlled by a few parameters like current, voltage or pressure. This subject is quite wel1-documented in the literature, especially for the UV-region [7],[8],[9]. Table I I I SOURCES DETECTORS I I I Quasi-blackbody radiators [ Thermal detectors [15] I (Tungsten strip) [10] I [ Ionization chambers [16] p I Arc plasmas with optically R I thick emission layersCll] I Si(Li) detectors [17] I I M I Continuum emission of Ge(Li) detectors A I hydrogen plasma C12] R I Y I Branching ratios [13] I I Electron storage rings [14] I I I S I E I Halogen lamps [181 Vacuum diodes [22] C I 0 I Deuterium lamps Cl?],C20] 1 N I D I Wal1-stabi1ized high- 1 A I pressure mini-arcs C21] R I Y I Radioactive sources I I Presently, the most promising absolute standard is of course the electron storage ring that can generate spectral radiant power from IR the soft X-ray region with remarkably small uncertainties 'ever, the matching of this powerful light source to the uments that have to be calibrated is not a trivial probl /nchrotron radiation is geometrically 358 limited to a relatively narrow beam which is also highly polarized and exibits coherence effects. Which calibration standard is to be used depends of course on the spectral region of interest. The natural lines of division lie between the VUV and the visible on one side; and the visible and IR on the other. Conventional black-body radiators of temperatures about 3000 K can be used in the IR, visible and near UV. In stationary noble gas arc plasmas sufficiently close to local thermodynamic equilibrium temperatures of 15000 K can be reached. At this temperature the wavelength of maximum emission lies around 200 nil. All transfer source standards for VUV and all primary source standards are sources of spectral radiance as radiance is the only calculable physical quantity. As radiance sources they require quite different calibration procedure than that for spectral i.rradi_ance sources. The spectral radiance of a source is normally calibrated only for a rather limited portion of the source. Therefore it is necessary to limit the field of view of the calibrated instrument in order to measure only the calibrated portion of the source (see e.g.Clll). Consequently, the measurement provides an absolute calibration only for a limited solid angle,and to obtain the absolute res jnsivity of the instrument numerical integration over the entire field of view is necessary. This kind of measurement requires a considerably larger experimental effort but in return it is a more reliable method if the field of view can be measured precisely. In the VUV region it is the only available method unless a comparison with a standard detector is used. The absolute calibration of an instrument for measurements of upper atmospheric emissions is much simplified if a standard source of spectral irradiance can be used. The source is then used to illuminate a screen which is assumed to be a lambertian di f fuser with a known albedo and which is large enough to fill the entire field of view of the instrument. Any measured radiance of an extended source like an airglow emission may then be directly related to the radiance of the screen. Through the years numerous white, diffusely reflecting materials and coatings have been devb*. -d. Among the most commonly used are [233-C273: -smoked MgO, -pressed MgO powder, -pressed BaSO» powder, -sprayed BaSCU with different binders, -pressed PTFE powder. The measurements of the total reflectance and diffusing properties of a diffusing screen require high experimental effort and are usually omitted. At the same time differences in coating techniques and thickness make it difficult to compare the work of various authors and virtually impossible to rely on published data. Furthermore, aging, chemical composition and purity affect properties of the screens. Generally it can be stated that, as mentioned above, for high reliability a calibration against a standard of radiance 359 should be used. When a calibration against an irradiance standard is performed,a freshly smoked MgO screen gives still best results in the visible and near IR while pressed BaSCU (regent powder, Merck) exibits higher albedo in the UV region. It still remains to be seen how reproducible the results obtained with PTFE screens will be. 4. THE ABSOLUTE CALIBRATION OF 0.6 m CZERNY-TURNER SPECTROMETER WITH AN IMAGING PHOTON DETECTOR. As a practical example, the procedures adopted for the calibration of 0.6 m spectrometer equipped with an imaging detector are outlined. Only the general approach is presented as final results of the calibration are not yet available. 4.1 INSTRUMENT SPECIFICATIONS. The grating spectrometer for calibration was originally designed at the University College London as a scanning spectrometer for studies of various chemical releases at thermospheric altitudes from the payloads of high altitude sounding rockets. The instrument is equipped with a grating of 1200 grooves/mm and size 120 mm by 150 mm mounted in a Czerny-Turner configuration. Due to mechanical constrains the instrument may only be used in first order. The spectrometer was recently modified to accommodate an .Imaging Photon Detector (IPD). The original exit slit was replaced with a circular aperture and scanning mechanism by an adjustable locking screw assembly to allow a choice of any region of the spectrum between 320 nm and 750 nm. The IPD is a device that consists of a photocathode, triple multichannel plates and a resistive anode C2B3. The electron cloud arriving to the anode gives rise to a current in four' electrodes attached to the anode. The position of impact of the cloud at the anode can be obtained by either the time of arrival or charge ratio methods. The paired x and y coordinates for each event are taken into a computer via a parallel interface and used to build up an image of the spectrum. The actual system has a capability of taking a full image of 256 it 256 pixels, each pixel being allocated 16 bits of memory. In practice, before the image of a spectrum is saved it is reduced by integration to a size of 16 rows of 256 pixels. Then the image is further integrated to obtain a one-dimensional spectrum as in present application the second dimension was not carrying any information. The spectral range coverd in each image is some 26 nm, with resolution of the order of 0.4 nm. 4.2 LINEARITY CALIBRATION. The IPD itself is capable of processing some 250 000 detected photons/s, thus the linearity of the system is determined by much slower associated electronics (preamplifiers, A-D converters) and by the interface to the computer. The total, linearity of the system was obtained by measuring the signal with and without a single neutral 360 density -filter at different intensity levels. The dead time and filter transmission factor were then fitted in a least square sense for a given deadtime response regime. The function minimized Mas the deviation between the measured signal with the filter present and that computed from the signal without the filter at the same lamp intensity. The results obtained for the two nominal response cases were: NONPARALYZABLE PARALYZABLE tCfisl 46.6 i 0.8 23.7 ~ 0.2 f 3.42 i 0.04 3.00 i 0.02 chi* 0.069 14.96 This indicates that the system was behaving in a nonparalyzable manner and in addition the filter transmission factor for a nonparalyzable system agrees very well with the value measured by other means. The above measurement applies to the integrated response of the system. As we are dealing with a two-dimensional detector the question arises as to how the signal level on one section of the detector affects the signal in another section. Thus, the "local" linearity of the system was also examined using two spectral lamps; one spectral line could be attenuated while that from the other lamp was left unchanged. According to this measurement and also our expectations,the "local" linearity followed the total time response of the system, that is to say that it is only the count rate integrated over the detector that influences the linearity of the system. 4.3 SPECTRAL RESPONSIVITY. To establish the spectral responsivity of the system the absolute spectral calibration of each pixel in the image has to be performed for the entire spectral range of the instrument. However, in the present case the image of the spectrum is integrated along the direction of the entrance slit and in this way reduced to a one dimensional problem. Nevertheless it is similar to the problem of calibration of 256 spectrometer-detector systems over the range " "*0 nm to 600 n-n which is a considerable task. Accordingly .e calibration procedure adopted by us used the following steps: §£§£ !• The wavelength scale for each image is established using spectral lamps. At least two lines are required in every image. A compromise has to be reached between the use of a lamp rich in spectra] lines on one hand, and the possibility Df easy identification of lines actually in the image on the other. We pr eferr >d to use a number of different lamps with relatively fewer lines. The lamps could be used simultaneously and attenuated separatly to facilitate the identification of lines. §iSE 2- Spectral lamps were replaced by a spectral irradiance standard illuminating a freshly smoked MgO screen. The absolute responsivity of each pixel at a corresponding 361 wavelength was established in that way. rt calibrated deuterium lamp and a halogen lamp were used as the standards. Steg 3. The position of the spectrometer grating was changed and steps 1 and 2 were repeated throughout the entire spectral range of interest. Ideally the movement of t.ie grating should shift the wavelength in steps corresponding to the wavelength difference between two consecutive pixels. However, the properties of the detector suggest that the spectral responsivity of the system should not exibit any severe discontinuities and therefore it was decided to use steps of approximately 5 nm in order to keep the quantity of data within managable limits. n—i—i—i—r-1—i—r • i ~i—rT—1—l—I—I—r—1—|—I—I—I—T—I—r—i— I I 1.* MXLAM = 4000. _ • • * * • • «• MXPIX = 1.446 - + 1.2 - . * « * * 1.0 / # 0.8 - * + • * *• 0.6 * _ * 0.* 0-1 IPIX = 44 • •^ ...... i ... .**.*» 4000 tovcUnfth (A) Figure 1. Spectral Responsivity of a single pixel §tee 4. When the measurements throughout the entire spectral range are complete,the collected data have to be reorganized. The spectri.1 responsivity and corresponding wavelengths for a given pixel have to be collected from the entire set of measurements and rearranged as a speet, al responsivity curve for this pixel. In order to facilitate this procedure a polynomial of appropriate degree (in this case 6th) was fitted to each spectral responsivity image so that the sensitivity of a given pixel at its corresponding wavelength in each image could be easily computed. A very preliminary plot of the response is shown in figure 1. 362 5. REFERENCES [13 J.Stegman, P.-E.Tuninger, N.Wilhelm and G.Witt (1978) Spectrophotometry Measuring Techniques -for Atmospheric Quantal Emissions, ESA SP-135,217 [23 R.L.Klobuchar, J.J.Ahumada, J.V.Michael and P.J.Karol (1974) Details o-f Deadtime Losses in Scaling and Multiscaling, Rev.Sei.Instr. 45,8,1073 [33 R.D.Evans (1955) in The Atomic Nucleus, McGraw-Hill, New York 1955 [43 L.Alexander (19B6) A Pulse Ampi i f i er /Discriminator (PAD) -f Dr Single Photon Counting, in these proceedings [53 J.K.Pribram and C.M.Penchina (1968) Stray Light in Czerny-Turner Spectrometers, Appl.Optics 7,2005 [63 J.Michlovic (1972) Fabry Lens, Appl.Opt 11,490 [73 K.Henriksen (1974) Absolute Calibration of Auroral Photometers and Ground-Based Measurements in the 3200- 7500 A Region, Appl.Optics 13,1196 [83 B.S.Dandekar and D.J. Davis, Jr. (1973) Calibration o-f the Airglow Photometers and Spectrometers, Appl.Optics 12,1196 [93 M.R.Torr, P.Espy and P.Wraight (1981) Intercalibration o-f Instrumentation Used in the Observation of Atmosphercic Emissions: Second Progress Report, 19B1, CASS-101 Utah State University [103 R.D.Larrabee (1959) Spectral Emissivity o-f Tungsten, Jour.Opt.Soc.Am. 49,619 Cll] H.Magdeburg and U.Schley (1966) SpektralphDtometrische Eigenschaften des Niederstrom-Kohlbogens, Z.Angew.(Math) 20,465 [12: W.R.Ott, K.Behringer and G.Gieres (1975) Vacuum ultraviolet radiometry with hydrogen arcs. 2:Tne high power arc as an absolute stsndard o-f spectral radiance from 124 nm to 360 nm, Appl.Optics 14,2121 [133 M.J.Mumma and E.C.Zipf (1971) Calibration o-f Vacuum - Ultraviolet Monochromators by the Molecular Branching - Ratio Technique, J.Opt.Soc.Am. 61,83 [143 K.Codling and R.P.Madden (1965) Characteristics of the "Synchrotron Light" from the NBS 180-MeV Machine, J.Opt.Soc.Am.36,380 [153 R.L.Canf ield, R.G.Johnston, K.Codling and R.P.Madden (1967) Comparison of Ionization chamber and a thermopile as absolute detectors in the extreme ultraviolet, Appl.Optics 6,18B6 [163 J.A.R.Samson (1964) Absolute Intensity Measurements in the Vacuum Ultraviolet, J.Opt.Soc.Am. 54,6 C173 J.S.Hansen, J.C.McGeorge, D.Nix, W.D.Schmidt-Ott, I.Unus and R.W.Fink (1973) Accurate efficiency calibration and properties of semiconductor detectors for low-energy photons, Nucl.Instrum.Methods 106,365 [183 R.Stair,W.E.Schneider and J.K.Jackson (1963) A new standard of spectral irradiance, Appl.Opt. 2,1151 [193 R.D.Saunders, W.R.Ott and J.M.Bridges (1978) Spectral irradiance standard for the ultraviolet: the deuterium lamp, Appl.Opt. 17,593 [203 P.J.Key and R.C.Preston (1980) Magnesium fluoride 363 windowed deuterium lamps as radiance transfer standards between 115 and 370 nm, J.Phys.E:Sci.Instrum 12,866 C2U J.M.Bridges and W.R.Ott (1977) Vacuum ultraviolet radiometry. 3: The argon mini-arc as a new secondary standard of spectral radiance, Appl.Opt. .1,6,367 C22: L.R.Canfield, R.G.Johnston and R.P.Madden U973) NBS Detector Standards for the Far Ultraviolet, Appl.Opt. 12,1611 C23] F.Grum and G.W.Luckey (1968) Optical Sphere Paint and a Working Standard of Reflectance, Appl.Opt.7,2289 C24] W.Erb (1975) Requirements for Reflection Standard and the Measurement of their Reflection Values, Appl.Opt. 14 ,493 C253 E.R.YDung, K.C.Clark, R.B.Bennett and T.L.Houk (1980) Measurements and parametrization of the bidirectional reflectance factor of BaSD< paint, Appl.Opt. 19,3500 C263 J.B.Schutt, B.N.Holben, C.M.Shai and J.H.Henninger <19B1) Reflectivity of TFE - a washable surface - compared with that of BaSO*, Appl.Opt. 20,2033 1271 V.R.Weidner and J.J.Hsia (19B1) Reflection properties of pressed polytetrafluoroethylene powder, J.Opt.Soc.Am. 71,856 C2B3 D.Rees, J.Conboy, W.Heinz, G.Witt, J.Stegman and D.P.Murtagh (19B5) Auroral and Airglow Observations Using a Grating Spectrometer with an Imaging Photon Detector, to be published 364 A CALIBRATION PHOTOMETER FOR LOW BRIGHTNESS SOURCES H. Lauche and W. Barke, Max-Planck-Institut fur Aeronomie, 3211 Katlenburg-Lindau, FRG Auroral and airglow observations are usually done by cali brated photometers. It is not convenient to calibrate a photometer under normal conditions of operation by comparing an uncalibrated and a calibrated photometer looking at the same aurora or airglow in the sky. For many reasons it is easier to look onto a calibrated light source of known intensity. For the exchange of data of different airglow and auroral observations it is necessary to compare or re-cali brate the individual secondary sources. This calibration of secondary sources is done by comparison with another light source which has been calibrated in absolute units before. One single photometer of known spectral response, sensitive to low brightness, linear over a wide range of brightnesses, and having low dark current and low temperature coefficients for all these parameters, is used for calibration purposes. Since it is very difficult to keep a photometer constant for a longer period, it is recommended to repeat the measure ments of the primary source as often as possible. In many cases we will find the following situation. The source may have a homogeneous distribution of light output. The size of the luminous plate is larger than the field of view of the photometer at a fixed distance. If this photo meter is linear over a wide range of expected intensities and has a negligible dark current, we will get the simpli fied relation: N = E • B; where N is the count rate, E is the total wavelength dependent efficiency, and B is the surface brightness of the source (including transmission of filter and cathode response). This efficiency is assumed to be just the same when looking at the source to be calibrated and at the reference source. The above mentioned assumptions may be too simple,and it was necessary to test the equipment for the most important parameters: mechanical stability, stray light, and tempera- 365 ture coefficients. Calibrations of light sources have been done for many years by M. Torr. We want to support her work here in the European sector. This is the reason why this new photometer got almost the same physical properties as M. Torr's photo meter.1 Some mechanical modifications make it now easier to =ot a source into a fixed position in front of the entrance of the photometer. - Sources of 2 cm and up to 15 cm in diameter can be centered exactly at the optical axis. A Peltier-cooler helps to reduce the dark current. Figure 1 shows the block diagram and Fig. 2 shows the optical and mechanical configuration of the photometer. The set of filters is compiled in Table 1. The Tables 2, 3, and 4 show the results of the calibration of different sources done during the Lysebu Optical Meeting. We want to thank M. Gadsden for supporting this activity. He also supplied the well calibrated reference source. Table 1 Wheel Position 0 12 3 4 5„.. _J 7 8_ _ Filter Peak r (A) Black 3918 4280 4866 5573 _5882 6299 _6562 _6707 Filter BW(A) 41 27 25 16 13 12. __ lY. 6 Final note: Most phosphor sources show an enhanced emis sion with a decay time of some minutes to some hours when they have been exposed to external light. Even the weak light produced by the phosphor sources themselves may act as external light when it is reflected on a white surface. This unwanted effect, however, is relatively easy to avoid by storing the sources in boxes which are black inside. A simple lamp which combines the properties of a "line" source and a continuum source is shown in the Figs. 3 and 4. This source is called the Sodankyla calibration lamp in Table 2. •Marsha R. Torr. Intercalibration of instrumentation used in the observation of atmospheric emissions: A progress Report 1976-1979, Utah State University, USA, March 1980. stabilized 110/220V 1t50-60Hz 0-4\, stabilized h. 0-3KV C- DC r 1 converter 3- 0 - 6V DC J J Clh -n. i filter v puis housing H(- amplif. • Q- 01 • 5v 2= s centering JllIL device L. & •fc) T^l output Collimator 'i mirror photomultiplier housing w. peltier cooler BLOCK DIAGRAM • ELEKTRONIK Figure. 1 light source filter wheel Photomultiplier centering device \ ^ I 3 \ —j \ \ zt:^~~--m^^JB ^ \ i XH \ \ \ \ \ N \WNN \\\V\\\\\X\\\\\S\ K\VV Q WiC deflecting Collimator objective field stop mirror lens OPTICAL CONFIGURATION Figure 2 368 ibKSl M-TKiili 220V/- 5 0- Bzy/97C 48i& 30V ERG 2 0- blue fluoresz. 5V © 30- ZPU/180V N *f 0- 6V/2.4W «Hi' IEC 7004-30 Figure 3 os ram glow discharge lamp tungsten filament lamp ^ i / Amphenol \ / H opal glass -) field stop ate opal glass L . U •-J *430 B5 Figure 4 AIRGLOW CALIBRATION LAMP 369 Table 2 Optical Calibration, Lysebu, August 1985 Brightness (R/A) Source 3918 4280 4866 5573 5882 6299 6562 Filament Lamps McEwen 1.72 5.7 22.6 71.9 110.5 179.9 228.8 DMI-1 0.05 0.27 2.0 15.3 32.0 65.0 180.0 DMI-2 0.05 0.27 2.6 19.0 38.0 85.0 200.0 Sodankyla 0.12 0.71 5.4 31.8 62.8 125.8 137.0 (tungsten) Sodankyla 2.0 6.0 10.1 9.9 15.3 4.0 0.66 (glow discharge) Phosphors Uppsala Y275 0.03 0.3 3.8 251.0 378.0 217.0 113.0 I Uppsala 920B 5.1 126.0 61.5 18.6 10.5 6.7 8.1 Uppsala L1614 0.07 0.73 32.5 27.7 8.7 2.5 4.3 j Oslo R411 0.04 0.07 0.79 20.7 68.2 98.4 151.3 I i| Oslo 920B 15.3 198.0 88.0 14.2 6.8 6.8 4.1 . Oslo L1614 0.79 11.9 185.0 245.0 73.1 10.4 11.5 j Oslo Y275 1.2 23.0 26.0 204.0 300.0 262.0 195.0 | Oslo 920B/ 10B.0 583.0 462.0 216.0 163.0 216.0 144.0 j KR#2 Andenes 0.78 15.7 17.2 129.1 181.6 182.0 261.2 (yellow) Andenes (blue) 5.1 90.1 47.7 9.7 6.9 10.3 20.0 HPI 2 0.02 0.16 2.43 193.7 290.7 191.8 98.2 370 Table 3 Ratios Previous -- Lysebu/Lindau 1985/1983 Source 4280 4866 5573 5882 6299 McEwen 0.93 0.91 0,87 0.90 0.95 DMI-1 - 0.87 0.81 0.83 0.76 DMI-2 - 0.93 0.88 0.84 0.86 Uppsala Y275 - 0.97 0.73 0.79 0.81 Uppsala 920B 0.70 0.81 0.94 1.15 0.93 Uppsala L1614 0.87 0.72 0.79 0.91 - Andenes 0.84 0.79 0.75 0.73 0.85 (yellow) Andene3 0.71 0.86 1.07 0.96 1.18 (blue) MPI 2 - 1.16 0.92 0.91 0.94 Table 4 Ratios Previous - Lysebu/Aberdeen 1985/1981 Source 4280 4866 5573 5882 6299 McEwen 1.05 1.0 0.96 0.95 0.99 Uppsala Y275 - 1.10 0.88 0.95 0.91 Uppsala 920B 0.84 0.95 1.0 1.08 0.99 Uppsala L1614 0.91 0.79 0.81 0.90 - MPI 2 1.20 0.99 0.97 1.0 . - - 371 A SHORT HISTORY AND SUMMARY OF OBJECTIVES FOR THE GROUND-BASED OPTICAL AERONOMY (GBOA) PROBRAM IN THE UNITED STATES C.S.Deehr1, G.J.Romick1, J.Meriwether2, and M.H.Rees1 1. INTRODUCTION Ground-based optical aeronomers in the USA have orga nized -for the purpose of furthering the work of optical agronomy by establishing a forum -for large—scale coopera tive studies and national observational and analytical facilities. The purpose of the fallowing relatively detailed history of the development of this organization is to document the characteristic phases of what amounts to a revolution in U.S. optical aeronomy. The Aeronomy program at the National Science Founda tion (NSF) is a major source of funds for ground-based observations of the atmosphere by optical methods in the USA. In 1982-S3 the program director for aeronomy, S.G. Sivjee, recognized that a large portion of the aeronomy budget was being directed towards radio-wave aeronomy. In particular, a well coordinated approach for the establish ment and operation of a meridian chain of incoherent- scatter radars was begun at NSF after some years of radio- wave emphasis in the Aeronomy Program. This program, coupled with a singular (and character st i c) lack of signi ficant coordinated effort among optical aeronomers, was leading, in effect, to a budget reduction in optical aerc- nomy. Adr: lGvophys. Inst., Univ. of Alaska, Fairbanks, Alaska. *Spac» Physics RnMrch Laboratory, University of Michigan, Ann Arbor, MI. 372 1°32. G.G. Sivjee called on optical aeranomers to address the prabl err-, and an or gani z at i anal meeting was cal led by Lvlc Broad foot at the 1982 Fal 1 Meeting o-f the American Genphysi cal Un i on in San Franc i sco. The need -far devel op- ment of advanced instrumentation into large user-friendly facilities was stressed and the first of the annual optical aeronomy meetings was planned for August 1983. 1983. The Ground-Bc^sed Airglaw and Aurora Optical Faci1 i ti es Workshop was held August 1-4, 1983 hosted by Marsha Torr and Utah State Uni versi ty. Sixty scientists from 31 different U.S. and Canadian universities and government agenc i es represent ing perhaps three-quarter's of all the active U.S. scientists working in this atreå gathered to discuss the direction of scientific research in gerund-based optical aeronomy. The workshop reviewed the goals of solar-terrestrial and atmospheric research as outlined in the National Academy of Science (NAS) report "Sol ar Terrestr i al Research in the 19B0 s" and discussed the ground-based facilities required to attain these goa.s. It became apparent that the introduction ot advanced instrumentation and data handling systems combined with large-scale model ing could approach the NAS goal=, but it /jor.'d require a cooperative effort, unprecedented in the histo-y of observational optical jeronomy. In order to prepare a pl,?n of scientfic direction, the '^11Jrting sections were appointed to draft a report ot the •7. The report dealt not only with the review 3f aeronomy, but ale? with a proposed manager tal str r ture wi thifi the cj-nmurit/ * or the program. A provisional "Science Steering C nmmi t *. *e" 'SSC * «a-» appointed • c beg» n t Me p*for t jut 1 i ned f. th--> report, i . «s. iubmi t a proposal to NS* request J ny '..n*J> 'or t- *' "d, t eadiny t o *»r. Imp I pnwnt at ion rl«n. The 373 provi si cnal SSC consisted of five sections with t,.«o members each: fcurora and airql DW physics C.J- Ro/ni ck C. S.Deehr Atmospheric Dynamic~ T. Kii 1een J.MfcT i wether Space PIasma Physics R. Eather T . Hal ?. i nati ntmospheric Scat ter inq F.Roesler C.Sechrist Theoretical Models R.Rob1e S. Sol omar.. It should be noted that the discussion of the driving corce in the program vac i 1]ated between scisnce and instrumentation f r on-. - h* beginning. The division of the wor I procp^*1f:d f'.r^t according to thE- availability jf IU-W dptscturs and thence to different processes in the attro- = pre>rs and bacf again. Th??p changes a.-~e reflected ; the na.Te= of the sections. Nute that "Atmospheric scattering ' w-i"i real I . bv3ej on lidar obser\ations arid probably should *- ave t=€?r labeled me-f ospré * i c -' 1 ower t her mo ^.p her IC dynamics. The ot jt=rL •. - - 5 of the G&QA program as laid out in L ho 1 .:g?r. wrrfshjp report we' t- : Stud/ t \e interaction between the at riasphere and incident solar radiaLior. via a nationally coordi nated program of optical observations and modeling, •jji.itruct at 1 east thr .?e optical facilities uti - ' i: i ng adv^nrfd instrumentation for datr» c D^ . t*l .tp a • ompr ehensi ve model ing capability. Cjripile c* databcis*3 in con _, une 11 on with cmpi-ted r.Ljdel? tor r*at-H an^l.-i^ and th»oretirV. in-j r - pr et at i or-. F-^Mish v- s,,*te.n t, p i *n end coordinate data acquisition an J mode.11 i ny activities. '. °S 1. The • >•; r,r t of the 1 ^^ Logan mee- mc; -*r.d pr apojrtl t >r tht i npl emti". t a t i on p 1 an uere i ntraduced >£ ( he second ^.ijTirtier r.g^f i ny Kost ^d b • iohn Me' i wet hi?r r*nd r he University t)f Mil hi^-iri 11, Am» Arbor . T'f ^ *•.•)' -J J*." t adopt, r ri i the Michigan mpet rig Hf, to requfja'. USF th'ough ttw program director * or aeronnmy, i*. rSrii.tcrnei, to empanel j j'ir.- to t»vai uate. th«? prctpcts.il « 374 and, in e-ffect, appoint a science steering committee and -five subcommi t teas (Table 1). The subcommittees were listed according to a (rajor instrument type. vThe concern with i nstrumentatian '-hus again had a major e-f-fect on the dsvalopn-.ertt c? the program.': The charge to the Science Steering Committee w-\s to develop a set of scientific goals with the help T»f the modeling committee. This set of goals was then to be examined by the subcommiitees for the purpose of designing an appropriate prcgram of ubserva- \7FJ5• The steering comir.ittee dtveloped a program called CEDAR (Coupling, Energetics, and Dynamics o*' Atmospheric F'pgionsi . Tr.is wa= distributed to the subcommittees which developed approa-J.e^ to achieving the goals stt by th^- S5C pr^g.'a.n. These r e-iearch proposals by the sutcommi ttees were pr ==ented t J the com-nuni t y du; ing t'ie ~rd sumr.er meeting hosted by l.'.C.Ci-rl and Tne University of Washington. 5o/eral round'-. cf discussions of both the scientific and fiscal ji otedi/ f5 lo'i lin? SBC with the j'lb of combining th = se i -it - a r-'crent scientific and fiscal plan. The f :1 iGrtirg se._tion is a summary of the approach to this plan ir.d i t =, .a.ior point-b. 375 Tabic 1. GBOA co littee membership 1985 Steering Committee Imagery and Photometry Gerald J. Romick, Chairman Robert Eather, Chairman Geophysical Institute Department of Physics University of Alaska Boston College Fairbanks, AK 99701. Chestnut Hill, MA 02167 Charles Deehr A. Lyle Broadfoot Manfred Biondi Thomas Hallinan John Foster Stephen Mende Timothy Killeen Gerald Romick Robert Schunk Gordon Shepherd Chalmers Sechrist Brian Tinsley Brian Tinsley Douglas Torr Interferometry John Meriwether, Chairman Spectroscopy Space Physics Research Lab. G.C. Sivjee, Chairman The University of Michigan Physics Department Ann Arbor, MI 48109 University of Alaska Fairbanks, AK 99701 Fred Biondi Paul Hays A. Lyle Broadfoot James Hecht Supriya Chakrabarti Gonzalo Hernandez Andrew Christensen Fred Roesler Richard Gattinger Gordon Shepherd Gerald Romick Robert Sica Marsha Torr Roger Smith Craig Tepley Modelling Douglas Torr, Chairman Lidar Center for Atmospheric Chalmers Sechrist, Chairman and Space Sciences Department of Electrical/ UMC 41 Computer Engineering Utah State University University of Illinois Logan, UT 84322 1406 W Green St. Urbana, IL 61801 David Andersen Timothey Killeen Sidney Bowhill David Fritts Charles Deehr Manfred Rees David Fritts Arthur Richmond Chester Gardner Raymond Roble Gerald Grams Robert Schunk Fred Roesler Chalmers Sechrist Vince Wickwar Susan Soloman Brian Tinsley 376 2. GBOA PROGRAM SUMMARV The study of the solar-terrestrial system is conside red to be a program of importance to the United States. The list of nationally organized solar-terrestrial programs includes many NASA and DOD sponsored satellite missions as well as NSF programs such as the incoherent scatter radar network. These studies have led to models of the following solar-ten estrial regions: Troposhere, Stratosphere, Mesosphere, Thermosphere, Ionosphere, Exosphere, Plasmasphere, Magnetosphere, Interplanetary Medium, Sun. Previous research programs have been region-specific. For example, the NIMBUS, SABE and HALOE satellites focussed on the stratosphere. SME studied the mesosphere. The Atmosphere Explorer C, D, and E series focussed on the aeronomy of the thermosphere and ionosphere, and the Dynamic Explorer on the coupled electrodynamics of this region. Numerous Explorer satellites, including the ISEE series focussed on the magnetosphere and the interplanetary medium, and SMM studied the sun. Since data taken by these satellites were collected at different times, models were constrained to study the specific regions and periods covered by these various programs. This situation requires that relatively poorly defined boundary conditions must be imposed to represent coupling with adjacent regions. Since no real demarcations exist in nature which might physically isolate the terrestrial regions, the effects of interregion coupling (which must exist to some degree) are almost completely neglected. In some cases the coupling may play a crucial role, for example in the case of the inter—region transfer of solar energy. The principal scientific challenge before us today is 377 to understand the whole coupled Sun Earth system. It is for this reason that the «rational Academy of Sciences report! "National Solar Terrestrial Research Program" lays so much emphasis on coupling. Missing from the current observational database are simultaneous measurements of key parameters relevant to coupling throughout the solar terrestrial system. The proposed NASA International Solar—Terrestrial Physics Program (ISTP) is designed to study primarily coupling of the sun and interplanetary medium with the magnetosphere and the auroral ionosphere. The Upper Atmosphere Research Satellite extends from the mesoephere to well into the topside ionosphere. A latitudinal chain of such observatories exists today and the data derived from these facilities has information with potentially global scale coverage. Detailed spectral observations also provide information on energy sources as a function of altitude. Simultaneous photometric imaging, in addition to providing valuable data on small-scale structure and associated dynamical effects, could serve to define the spatial characteristics of the emissions monitored by the narrow fields of view of interferometers. The advent of array detector technology has greatly enhanced our capabil ities in the area of all-sky photometric imaging. In addition, the global network of some 120 ionosonde stations could provide needed constraints on global ionospheric models and thermospheric circulation models, the former via measurements of electron density and the latter via the height of the peak of the layer which is strongly sensitive to the thermospheric meridional wind. The information derived on convection electric fields from Fabry-Perot measurements of winds at high latitudes could be supplemented by current systems derived from the existing high latitude magnetometer chain. These data would constrain the description of the cross-tail magnetospheric potential which has an important controlling influence on the electrical coupling between the solar wind/magneto- sphere and the ionosphere. Similarly, description of the global high latitude particle energy influx would be constrained by a network of suitably located photometric all-sky imagers. This combination of techniques could serve to monitor the corpuscular or energy coupling between ionosphere/thermosphere and the magnetosphere/ solar wind system. This scheme is a practical measurement program which could test the capability of both global and small scale models to predict short and long term variability of the coupled solai—terrestrial system without the requirement for supporting satellite measurements. Naturally, the availability of simultaneous satellite observations would greatly enhance the program, and it is hoped that the UARS and I3TP programs will provide this additional capability. The least understood region is the lower thermos- phere/upper mesosphere. for which mature models do not 380 e*! i st. Here the word "mature" means model s whi ch have been \erified hy extensive comparisons wz th datasets, from t hi? perspective of dyr_»inics, energetics and photochemistry. Th_i s region therefore must be regarded as a primary target f o"- further concentrated study. Although a comprehensive satellite database is not available to study the details of the micraprocesses which control this region, this shortco ming is offset by two factors: 1) The evolution of the facility-class lidar with the potential capability to probe the photochemistry and dynamics to heights extending beyond the E region with good al titude resolution. Existing 1 idar installations already have the capability to generate detailed altitude information on density, temperature, winds, tides and wave motions through the mesosphere. In addition to providing detailed information on the photoche mistry and dynami cs, these instruments could also def ine the lower boundary conditions, and hence the coupling to the stratosphere and troposphere. 2> In addition, the upper mesosphere is rich in atmospheric emissions which 3. BBOA OBJECTIVES The main purpose o-f the GBOA program ? s to study the interaction between incident solar radiation and the atmosphere via a nationally coordinated program of optical observati ons and model i ng. The program would be coordi nated with ongoing radar-, rocket-, satellite- and other pro grams. Emphasi s wculd be laid on the regi ons bounded by the iT.agnetosphere and mesosphere. The proposed objectives may be summarized in the broadest sense as follows: To study The Coupled Solar-Terrestrial System via - the formation of an improved theoretical framework leading ultimately toward global models of the solar— 381 terrestrial system, laying emphasis on inter-region coupling in order to understand the remarkable natural variability in the system. - the acquisition of data on a global scale to test the growing integrated theoretical understanding. - the design of new critical experimental tests and the validation of models. - the continuation Df work on improving components of global models including the quantification of micro- processes. The means of achieving the objectives set forth are: - To construct approximately eight optical facilities utilizing advanced instrumentation for data acquisit ion purposes. - To develop a comprehensive modeling capability to couple the exosphere, ionosphere, thermosphere and upper mesosphere. - To compile a database in conjunction with computerized models for data analysis and reduction and theoretical interpretation. - To establish a system to plan and coordinate data acquisition and modeling activities. 4. FIRST GBOA INITIATIVE In order to arrive at a practical plan for the study of the coupled solar—terrestrial system, the GBDA Steering Committee identified the three topics for study listed below. Coupling. Energetics and Dynamics of the Atmospheric Regions. "CEDAR" TOPIC I: Dynamics and energetics of the atmosphere from the upper mesosphere to the exobase. TOPIC II: Dynamics and energetics of the atmosphere from the upper thermosphere through the exosphere/ piasmasphere. TOPIC III Dynamics and energetics of the atmosphere from the stratosphere to the lower thermosphere. Topic I is intended to study primarily the dynamics and energetics of the upper mesosphere and thermosphere/ia- 382 nosphere on a global basis laying emphasis on the region between 85-150 km and concentrating on inter-regian coupling. Topic II concentrates on the coupling between the plasmasphere, topside ionosphere and exosphere. Topic III concentrates on the detailed dynamical and photochemical coupling in the upper mesosphere and lower thermosphere. The emphasis is not an the global character, but on the little understood microprocesses and small-scale wave structure which dominate this region. The primary ob jective is the development of mature models o-f this region. Lidars with enhanced capability will play a major role in this study. 5. EPILOGUE The plan outlined in the previous section was presen ted to the National Science Foundation during the winter o-f 19B5—86 in the -form of a proposed 7 year implementation program. Prospects for its adoption by NSF as a national program are good. The current list of accomplishments by those involved in the BBOA effort is large and growing, ft program of synthetic emission spectra by V.Degen (Univ. of Alaska) is available at telephone computer terminals for interested users. A coordinated program of neutral and ionized wind observatories (R.W.Smith, coordinator) has begun along with a coordinated OH and Qa nightglow observa tional program It is obvious that the program can only gain •from the e-f-forts o-f those observers in other countries who are interested in joining any part of the program. For further information and a quarterly (approximately) update via the newsletter "CEDAR POST", please send a request to Dr. Gerald Romick, The Geophysical Institute, University o-f Alaska, Fairbanks, Alaska 99701, U9A. WD »S0 MO 120 CO 00 CO «0 20 0 20 «O «0 TO W0 120 M0 ISO IB0 Figure 1: Stations which are in some way involved in the GB0A program, i.e. either a direct involvement or a cooperative program. (Those interested in the program should contact the appropriate program leader or committee chairman and help increase the coverage.> 384 A Canadian Meridian Photometer Array (CANOPUS MPS) * D.J. McEwen, L. Cogger, F. Creutzberg, R. Gattinger, F. Harris, A. Vallance Jones, R.A. Koehler, J. Wolfe Two meridian chains of photometers are to be installed in Western Canada to provide continuous nighttime monitoring of auroral precipitation along the meridians from 55 to 80 \. The first instrument, built by Bomem Inc., is currently undergoing field trials and will be installed at Gillam, Manitoba in the 1985 autumn before the Viking launch. The instrument will record auroral emissions at 470.9, 486.1, 557.7 and 630 nm with automatic background subtraction. Design calls for data transmission to a central data centre in near real time with 1-minute time resolution. Some of the operating features of the instrument and anticipated uses of the data will be discussed. Institute of Space and Atmospheric Studies, University of Saskatchewan, Saskatoon, SASK S7N OWO, Canada. 385 Ground-Based Programs Related with Viking Georg Gustafsson Uppsala Ionospheric Observatory S-755 90 Uppsala, Sweden Description of the Satellite. Viking is a Swedish satellite with a scientific payload to study magnetospheric phenomena over the auroral zone and the polar cap. It will be launched with an Ariane rocket from Kourou in French Guiana. It has a polar orbit (incl. 98.7 ) which is rather eccentric in the range 800-14000 km. Apogee will be over the northern polar region at the time of launch (argument of perigee 330 ) and it moves in the orbital plane by -0.5 per day. At the time of launch the local time when the satellite passes the equator in the northward direction is 2230 LT and decreases by one hour in 18.5 days. The orbital period is 270 min, but data from the satellite is recorded at only one ground station, Esrange, near Kiruna, which means that the datataking period is less than 150 min per orbit. Launch is expected to be at the beginning of 1986. The payload of the satellite consists of six experiments. The ex periments and principal investigators are: 1. Electric field vector experiment (called VI) IJ Block,Royal Institute of Technology, Stockholm, Sweden 2. Magnetic field vector experiment (V2) T A Potemra,Applied Physics Laboratory, Laurel, Md., USA 3. Hot plasma experiment (V3) R Lundin,Kiruna Geophysical Institute, Kiruna, Sweoen 386 4. Wave experiment, high frequency part (V4H) A Rahnsen,Danish Space Research Institute, I.yngby, Denmark 5. Wave experiment, low frequency part (V4I.) G Gustafsson,Uppsala Ionospheric Observatory, Uppsala, Sweden 6. Ultraviolet auroral imager (V5) C D Anger, University of Calgary, Calgary, Alberta, Canada For the coordination of ground-based experiments with Viking it is of particular interest that the satellite has two cameras for auroral observations. That makes it easier to compare ionospheric observa tions from the ground with those on the satellite. Viking Campaign Concept During the nominal lifetime of the Viking satellite (8 months) spe cific periods of enhanced scientific activity will be organized at the Viking Scientific Center (VISC) at Esrange. Each period of aboui 14 days is called a Viking campaign. The purpose of organizing the campaigns is to obtain an effective way of information exchange between the scientific groups involved, and thereby increasing the scientific output of the project. The plans are not only to coordinate the data recording but also to carry out or at least initiate scientific studies during a campaign. For each campaign scientists representing the six experiments on the satellite and the ground-based community will be at Esrange. All the seven groups have computer facilities at Esrange to make real (or nearly real) time analysis of their data. Ouicklook plots ind a data summary file are also produced by the Space Corporation and made available to the scientists with only a few hours delay. Therefore, short periods or events of interest can be selected and analyzed during a campaign which could lead to a scientific study. The Quicklook plots can be obtained by the scientific community at large on a subscription basis. 387 Efforts are made to organize the campaigns to become as well coordi nated BS possible Including not only Viking but also other space projects and ground-based activities. Georg Gustafsson is the general campaign manager responsible for the coordination of the scientific studies of the campaigns. There Is a campaign leader for each campaign, G Gustafsson (day 20-35 after launch of the satel lite) , C Anger (35-50), R Lundin (50-65), T Potemra (95-100), A Bahnsen (125-140), H Koskinen (155-170) and L Block (180-195). This is the present schedule which may be changed dependent on launch date and the functioning of the satellite. Risto Pellinen is the coordinator for the ground-based observations, organizing the participation from the ground-based community at Esrange. Ground-Based Observations A number of ground-based programs are coordinated or related with Viking. Below, the international programs are listed. The contact person with the Viking project has also been included to assist those who are interested in information or joining a certain project. EISCAT A special program has been planned to cover approximately 100 Viking passes. Those that are closest to the radar will be selected. 270 hours of observing time will be used for this project. Three different operation modes are planned: - a short scan, covers Kiruna to Tromsd - a long scan, along the magnetic meridian through Tromsø covers 60-77 degr. geographic latitude - a static mode, observation along the Troms* field-line. Real time (or nearly real time) data of the electric field and the electron density will be transfered to Esrange. Contact person: Hermann Opgenoorth. 388 STARE-SABRE These two pairs of radars will be operated during ths Viking mission to measure electron drift velocities in the ionosphere over a large aren. Efforts arc made to transfer data in real time to Esrange. Contact person: Erling Nielsen. HF- radars in Canada The radar at Goose Bay operated by R Greenwald API. of John Hopkins University and the radar at Shefferville operated by C Hanuise LSEET in Toulon, will provide maps of the convective flow over a large part of the polar cap (2.5 hours of local time and 10 degr. of latitude). These two radars are important for cusp studies and Viking with an inclination of 98 degr. will pass near the cusp during almost every pass. Contact person: Catherine Senior. GISMOS This is a program for Global Incoherent Scatter Measurements Of Substorms. During incoherent scatter " World Days " all six incoherent scatter radars will operate for 24 to 72 hours conti nuously. Observations by many ground-based as well as several space borne instruments are coordinated with these " World Days". The schedule for 1986 is 14-17 January, 5-6 March, 1-4 April, 6-7 May, 4-5 June, 9-10 July, 27-28 August, 23-26 September, 29-30 October and 10-11 December. The experiments begin at 1600 UT. Contact person: Odile de la Beaujardiere. Sindre Strømfjord The incoherent scatter radar in Greenland will have an exten sive program that is coordinated with the Viking measurements. Contact person: Walter Heikkila. 389 Global Auroral Dynamics Campaign In this program there are 9 key stations with all-sky TV-cam era, a fluxgate magnetometer, a VI,F receiver, and an induction magnetometer. There are also a number of stations with only an induction magnetometer. The observations cover the northern auroral oval and the polar cap. Contact person: Takasi Oguti CANOPUS The Canadian CANOPUS program will make coordinated obser vations with the Viking project. The scientific studies include: the electrojet in relation to substorm development, the electro dynamics of the magnetosphere-ionosphere system, and the study of the dynamic morphology of proton and electron auroral sub- storms. Observations are made by radars, magnetometers, riometers, meridian photometers, and an all-sky imager. Contact person: A Vallance Jones Those are the ground-based programs that at present are coordi nated or related with Viking and are of global or international nature. In addition there are also national programs that are related with Viking. It may also be of interest to the ground-based community that a comprehensive coordinated space program including several space craft is planned for March-May 1986. During that period IMP-8 will be in the solar wind part of the time, ISEE in the tail region, DE can cover the south and Viking the north polar regions. Ground- based support in relation to this magnetospheric-ionospheric study is of course very valuable. 390 Global Wind and Temperature Measurements This is a program including 12 stations and it is linked to the GISMOS project. Contact person: David Rees. Global Magnetic Field and Riometer Measurements Data from 25 magnetic and 25 riometer stations will be compiled and analyzed in Finland. The magnetic stations cover the polar cap, auroral oval as well as lower latitudes. The riometer stations cover only the auroral oval. The final data bank will be available on magnetic tape. Contact person: Risto Pellinen. Magnetometer-ULF Recorder Network A Finnish-German program for magnetic measurements has been initiated which will be run during the Viking mission. Contact person: Jorma Kangas. UT-LT Magnetic Data Based on a number of AE-magnetic stations plots will be pre pared in universal time-local time plots to study the develop ment of auroral substorms. Contact person: Alex Zaitzev. Spitzbergen Observations A program including optical, magnetometer and riometer measure ments is planned for Spitzbergen. During certain cusp studies plans are to transfer data from Spitzbergen to Esrange via Telefax. Contact person: C Deehr. 391 SESSION Villi Observation programs Chairman i Risto Ptllintn (Finnish Met. Inst., Helsinki) Chairman's summary Large multi-instrumental and multi-national observation programs, that cover the emissions and dynamics of almost the entire atmosphere, are in progress. The increased activity both in the European and U.S. sectors proves that the field of GBR is far from dying;. The instrumental methods and data handling techniques have improved considerably uithin the last decade which also enhances the high potential of this field. Only the question of manpouer seems to be the only limiting factor. Fred Rees reviewed on the coordinated GBR optical programs in the U.S.. The start of this activity took place after the 10. Optica] Meeting in Grasse. The program has been named as CEDAR C Coupling Energetics, Dynamics of Atmospheric Regions ). It is a joint financial and scientific program that covers both conventional methods < interferometry, spectroscopy, imaging, photometry and 1idars > and modelling. The uorKing principles are f cluster of instruments operated in one location, cooperative campaigns including radars, instrument development and testing. Three annual meetings have been held since 1983. Tuo issues of a newsletter have been published < G. RomicK, Univ. AlasKa, editor). John Meriwether, chairman of the organization, continued on the highlights of the last CEDAR meeting in Seattle and on the uorKing group meeting in Prague/IAGA in 1985. A global campaign to study plasma convection and uinds in . the global scale together uith incoherent scatter rAd&rs is planned for 14-16 Jan 1988. Also a summer campaign is planned. Improvements to FPI's were proposed in order to meet the demand of a feu Rayleighs in the auroral zone. CCD and other totally neu techniques uill be considered. South Pole and Hauaii uill be included to the netuorK as totally neu sites. A proposal ( 3 million U.S. dollars > for the management and coordination of FPI data has been submitted. Dr. V. Jones reported on the Canadian CANDPUS program that contains recordings of E-fields, magnetic field < currents > and optical emissions. It consists of tuo chains of meridian scanning photometers connected to a central data base uith a remote access capability for individual scientists. The netuorK is operated continuously automaticly. Cooperation is planned uith ViKing, Polar Bear, EXOS-D, UARS, ISTP, and GARS < 30 MHz ). AI 1-sKy imager uses intensified CCD techniques to record separately N(SO, 337.7, 630.0 tin and background. The data are recorded on magnetic tape. MPA records N<£+?, hKbeta), background, 357.7, 630.0 nm, background, field of vieu 4 deg., spectral uidth 2.0 nm. Telemetry requirement is 52 Mbytes/ueeK. The program has costed 10 million Canadian dollars. 392 Dr 0. Gustafston reported on the GBR program associated with the ViKing satellite operation. Some -facts of the VlKing mission uere repeated: Maximum data coveragforbIt Mill be ISO min, the apogee will drift 0.5 dig/day, the equatorial crossing in ascending node ulll be 82.36 LT at launch. The launch is on IS Nou 1985 (nou on 18 Jan 1986). The operation is partially based on scientific campaigns organized as CDAW-type workshops each lasting for tuo ueeKs . QuicK-looK plots, S pages (A4> per orbit, uill be distributed to all scientists that have subscribed for them. Ground-based data uill be received in real time to the ViKing Scientific Center during the campaigns. Cooperation has been established uith the following networks and persons: EISCAT, STARE, CANOPUS, GISMOS (de la Beaujardiere>, global wind and thermosphere measurements John Meriwether told about MAPSTAR (Middle Atmosphere Structure Associated Radiation), which is a four-year program started in 1983. Instruments are located in five different places and run during tuo campaigns per year. The program aims for studying infrared str-ucture sources uith experimental, modelling and laboratory techniques. The instruments involved are radars, photometers, radiometers, interferometers, cameras, FPI's, al1-sky TV's, and Vidjcon II TV's. First results uere reported in the spring 1985 AGU meeting. Alv Egeland described the polar cusp studies carried out at Spitzbergen, uhich is the only place on the northern hemisphere where such observations can be made. Much of the uorK has been done in close cooperation uith the Univ. of AlasKa. At present the instrument coverage is fair, only an ionosonde is lacking. The scientific program is as follows: 1) morphology of the dayside aurora, a) spectrum (also IR), 3) solar spectrum, temperature of the mesosphere, 4) ULF/ELF/VLF micropulsations, emissions, cusp mapping, 5) termospheric circulation, 6) correlation uith satellites and the observations at Barentsburg (PG1 station), 7) correlation uith rocKet and balloon programs, 8) new proposal from the USSR to equip Heiss Island, Novaja Zemlja, Spitzbergen, and BJornoya uith similar instruments. FYSISK INSTITUTTS DEPARTMENT OF 8 FORSKNINGS- PHYSICS GRUPPER RESEARCH SECTIONS Allmennfysikk og didaktikk General Physics Biofysikk Biophysics Elektronikk Electronics Elementærpartikkelfysikk Experimental Elementary Particle physics Faste stoffers fysikk Condensed Matter physics Kjernefysikk Nuclear physics Plasma-, molekylar- og Plasma-, Molecular and kosmisk fysikk Cosmic physics Teoretisk fysikk Theoretical physics ISSN-0332-5571 3 m T M
2 m T '
2 m T [