What is Inside

Deployment .................................................................................................. 1

Instrumentation .......................................................................................... 4

Focus ...................................................................................................... 4

Timing ..................................................................................................... 5

Operations .................................................................................................... 6

Data System ................................................................................................. 8

DMAC Activities ..................................................................................... 8

DMAC Readiness Review ..................................................................... 9

10th DMAC Users´ Committee ...............................................................12

11th DMAC Users´ Committee ...............................................................13

Management ................................................................................................17

Science .........................................................................................................18

Publications Policy ................................................................................18

Science Article Outlines .......................................................................18

Theses ....................................................................................................21

Recent Publications ..............................................................................212222222222222222222222222222222222222222222222222222222222222222222211

111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222221

111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111

GONG

33333333333333333333333333GONG Newsletter33333333333333333333333333

Number 26 The Global Oscillation Network Group October, 1995

3333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333The GONG network is completely

deployed and operational, and theDMAC has network data availablefor you!!!

It took 16 months of hard workfrom the first groundbreaking, at Lear-month, to the last camera alignment, atUdaipur, to complete the deploymentof the network; not to mention the de-cade of design, development, and pro-duction that preceded it. Thank you,dear reader, for your patient supportwhile the Project has been working toachieve this, and thank you to the Pro-ject staff and all of our collaboratorsand hosts around the world for all oftheir effort on behalf of GONG. Nowis the time for you to get actively in-volved and help us all work with thedata to pursue the exciting scientificinvestigations that are becoming possi-ble.

The deployment sequence waschosen to achieve a three-site network,with roughly 120 ˚ separation at theearliest possible moment, in order toget data into the hands of the members— to start working with it to helpunderstand the performance as rapidlyas possible. The duty cycle for thefirst GONG month of three stationswas a quite respectable 70%, and thedata have been pushed all the waythrough the DMAC pipeline. All ofthe data products — including ourfirst–cut frequencies — are now avail-able through the DSDS. ( See the boxeson page 10 and page 11. ) If you donot already have an account with theDSDS, the simplest way to set one upis through our WWW server(http://helios.tuc.noao.edu ).

While the Project team is still work-ing on a long list of improvements tomany aspects of the processing, we en-courage you to get your investigations

1111111111111111111111111111111111111111111111111111111

underway. At the same time, pleaseappreciate the challenges associatedwith understanding the characteristicsof such a complicated system, and ini-tiating a number of complex process-ing tasks — such as merging imageddata — that have never been undertak-en before.

The Science Teams are shifting intohigh gear, and they are schedulingworking sessions to work toward threenear-term goals for the presentation oftheir work: i ) a summary presentationat the Winter meeting of the AmericanAstronomical Society in mid-January,ii ) a special issue of the journal “Sci-ence” to be submitted in February, andiii ) a joint meeting with the AAS andits Solar Physics Division in June1996. These “opportunities” are dis-cussed starting on page 18.John Leibacher

11111111111111111111111111

3333333333333333333333333333333333333333333333333333333333333333Deployment

Deployment is behind us at last. Ittook about 10 months from the verybeginning to the very end when FrankHill and the Red Team departed fromNew Delhi in early October. Therewere many anxious moments ( manymore of those than real problems ),many successes, and many good timesalong the way. The short version ofthe story is this: all of the stations ar-rived where we intended them to go,none were damaged beyond repair, andthey were all working when we left.Now, the rest of the story.

The activity officially began in Tuc-son last December when the Teide sta-tion was lifted by crane onto the backof a flat bed truck, beginning its per-sonal odyssey to the Canary Islands.

- 2 -33333333333333333333333333

Calendar222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222

Event Date Location and Contact222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222The Solar Cycles: Perspectives and Future Research April 1 - 4, 1996 Hermosillo ( Mexico )

Antonio Sanchez-Ibarra(antonio@fisica.uson.mx)

IRIS Team Meeting April 1 - 4, 1996 Nice ( France )Eric Fossat(fossat@ayalga.unice.fr)

GONG First Results Workshop June 12 - 15, 1996 Madison ( Wisconsin )John Leibacher(jleibacher@noao.edu)

Helioseismology July 23 - 27, 1996 Brisbane ( Australia )1996 Western Pacific Geophysics Meeting David Cole and John Leibacher

(david@ips.oz.au, jleibacher@noao.edu)

Heliosismology status and asteroseismology prospects September 30 - October 3, 1996 Nice ( France )IAU Symposium 181 Gabrielle Berthomieu

(astro96@rameau.obs-nice.fr)22222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222221

11111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111 22222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222211

1111111111111111111

222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111

The stories it could tell, if only itcould talk! There was the day that oneend of the external cable raceway andthe doorknob were crushed beyondrecognition at the hands of somenameless stevedore somewherebetween Los Angeles, California andSanta Cruz de Tenerife. Then therewas the day or days during that sameperiod when salt water flooded overthe station, corroding every lock andhinge that hadn’t already been crushed,almost ( if not completely ) beyonduse. This, however, was hardlysufficient to discourage either FrankHill and the rest of GONG’s RedTeam, or the truly excellent IAC staffat El Teide. The cable raceway wasrebuilt by the observatory shops, thedoorknob was replaced, and judiciousapplication of bolt cutters, oil, cleanwater, and paint rendered the shelterindistinguishable from the condition itwas in when it left the Arizona desert.The instrument itself went back to-gether just like the manuals said itwould, and the first data was obtainedon February 17. The installation pro-cess took six weeks, which was exact-ly what had been scheduled. Specialthanks go out to Pere Palle, JesusPatron, Otilia de la Rosa, AntonioPimienta, and many others at the IACwho helped us make our first deploy-ment a success.

11111111111111111111111111111111111111111111

Next stop: Learmonth, WesternAustralia. Ed Stover and Rob Hub-bard, the vanguard of the Blue Team,left Tucson on March 15, and caughtup with the shelter in Learmonth onMarch 20. To our surprise the shelterwas in exactly the same condition asthe El Teide station. The cable race-way was crushed to the same unrecog-nizable pulp in exactly the same place,as was the doorknob. Were we deal-ing with the same evil stevedore, sin-gling out the GONG project for thisspecial attention, or were we missingsomething fundamental? While inAustralia, Duane Miller, the blueteam’s instrument maker, saw a briefimage on local television that providedus with the answer. The piece ofequipment which is used world wideto move standard shipping containersaround ports has a hydraulic ram thatconflicts in several places with someof the modifications we had made tothe container to make it into a GONGinstrument shelter. Although we hadasked for special handling, our “bigwhite box” just looked too much like astandard shipping container, and wasroutinely being treated as such.

When the balance of the Blue teamarrived in Learmonth, Master SergeantCoffman, a member of the UnitedStates Air Force contingent at the ob-servatory, gave us our official “inbrief.” The message was simple: drive

11111111111111111111111111111111111111111111

on the left, pedestrians don’t have theright of way, and almost everythingthat crawls or swims is deadly. If wewould just keep our hands in ourpockets and not leave our rooms, wewould be safe. Coming fromrattlesnake country, we were not soeasily frightened. We did adopt thetemporary site password deathadder,however, just as a reminder of ourmortality. The bite of this particularserpent renders its victim dead withinabout 30 minutes in most cases. In thefinal analysis, the only death adder wesaw was a dead death adder. ( Self-inflicted wound? ) The people and thecritters alike were wonderful to us, andwe had the second GONG network siteup and running when the last of theteam returned home on April 28. Herethe acknowledgements have to includeJohn Kennewell, Alex Liu, Major JeffCarson, Jenny House, and the rest ofthe Australian and USAF observatorystaff.

No rest for the wicked, however.After three weeks at home to get toknow our families again, the Blueteam was back in the air, preparing toinstall the Mauna Loa station. By nowwe had learned what pieces of the sta-tion should be removed prior to ship-ping, and the Mauna Loa station ar-rived in Hilo in far better conditionthan the first two. Most of the teamleft together this time, arriving in Hilo

- 3 -33333333333333333333333333

The El Teide Station at the Instituto Astrofisico des Canarias site there,with the vulcano of the same name in the background, was the firstGONG station to come online on March 2.

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222221

111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111

on May 22. We had come to partici-pate in a deployment in a tropicalparadise, but instead we found our-selves — within a very few hours ofmaking landfall — up at an altitude of11,000 feet on a desolate volcano inthe rain and fog. Every day we com-muted two hours from Hilo to MaunaLoa Observatory, and two more hourshome at day’s end. We called ourwork site Mars, because it was slightlymore colorful ( though no less lifeless )than the moon. Our offerings to Pele,the goddess of the volcano, forestalledthe next major eruption of Mauna Loaat least long enough for us to completethe task at hand. As proof that we dolearn from our experiences, this de-ployment was completed in only fourweeks. And just in case anyone outthere is feeling too sorry for the BlueTeam, Hilo may not be the greatestplace in the world to look for beaches,but we found out where to drive toplaces that are. We actually sufferedvery little during our stay on the BigIsland. Our heart-felt thanks in thiscase go out to Charlie Garcia, EricYasukawa, Darryl Koon, Russ Schnelland Judy Pereira for making this one alot of fun.

Next came Cerro Tololo and achance for the Red Team to try theirhand for the second time. The activitybegan with the departure of Frank andmost of the others on June 24. Onceagain, the support afforded our teamby the local staff was nothing short ofamazing. The shelter had arrived inexcellent condition this time, and as-sembly proceeded very smoothly.Even the weather cooperated, yieldingup almost summer-like conditionssome days in the midst of the Chileanwinter. It was reported that an un-named instrument maker associatedwith the Red Team contingent layprostrate on the ground for an extend-ed period one day in an attempt to lurecondors, but this cannot be confirmed.Nor can we categorically deny that hisbunk mates staked him there. The endresult was that the fourth GONG net-work site was left to the CTIO staff

11111111111111111111111111111111111111111111111111111111111111111

and the condors on July 22, anotherfour-week installation. All kinds ofthanks go out to Oscar Saa, RicardoVenegas, and all of the great peopleworking at our sister observatory inthe Chilean Andes.

The last hurrah for the Blue Teamwas Big Bear Observatory in southernCalifornia. Our adventure began onJuly 24. We had had more thanenough of airplanes and transoceanicflights by this time, so the team justloaded up our gear, piled into two carsin Tucson, and headed west on Inter-state 10. Although we were facedwith eight hours of some of the hottestinterstate driving anywhere in the US( it was 50 ˚C when we stopped forlunch in Blyth, California ), the cool,clean air and spectacular sceneryawaiting us at Big Bear Lake made itwell worth the little bit of effort to getthere. The shelter, too, had no boatrides to contend with, traveling thesame interstate that we did about a daybehind us. We saw it go by the win-dow of the cafe at which we were en-joying a leisurely breakfast after amorning of shoveling sand in prepara-tion for its arrival. No paint job this

111111111111111111111111111111111111

time. The shelter looked — notsurprisingly — just like it had when itleft Tucson a day earlier. What a fineplace to work! The locals complainedof the oppressive heat ( it got up to al-most 30 ˚C one day ), but those of usused to the desert in July and Augustin Tucson could hardly relate to theircomplaints. We had the instrument upand running in about three weeks, butseveral of us remained a few extradays to train the Big Bear staff in theoperation of the instrument, and tobreath just a little more pine-scentedair. We grudgingly returned home onAugust 18, ending our deployment ad-ventures. Many thanks for help andfond memories go out to Bill Mar-quette, Jeff Nenow, John Varsik,Karen Carlson, Randy Fear, and Me-linda Hope. It was the perfect endingfor us.

The final test of the Red Team be-gan with the Udaipur deployment onAugust 27. The Udaipur Solar Obser-vatory staff at their new facility justoutside of Udaipur already had theshelter and generator roughly in posi-tion when the team arrived. Therewere some delays initially, as two

- 4 -33333333333333333333333333

The Learmonth station at the Learmonth Solar Observatory of the IPSRadio and Space Services in Exmouth, Western Australia becameoperational on April 28. The SEON radio telescope can be seen in thebackground.

22222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111

22222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222221

1111111111111111111111111111

22222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111

large install kits sent to India by airfreight ultimately lagged behind the in-strument shipment. The team and theUSO staff did a great job of “gettingalong without” until the gear showedup about half way through the deploy-ment.

Although frustrating, these delaysdid allow the team to take a day off toexplore the countryside and do somesite seeing. They visited a beautifulJain Temple near Udaipur, where oneof our number was confronted by anangry monkey. During what seemedlike an innocent photo opportunity, theprimate dropped down from its perchand assumed a fighting stance reminis-cent of an antagonist skilled in themartial arts. Sang Nguyen — himselfhaving earned a black belt in TaeKwon Do — instinctively adopted athreatening posture of his own whichso intimidated the ape that it opted tostand down and live to fight anotherday. Many skills are needed on a wellrounded GONG install team.

Other than colds and flu that mostlikely accompanied the team all theway from Tucson, the balance of the

11111111111111111111111111111111111

deployment went smoothly. On Oc-tober 3, that instrument came on line,and the network was complete. As al-ways, there are many folks to thank,but we should certainly name ArvindBhatnagar, Ashok Ambastha, SushantTripathy, Sudhir Gupta, Naresh Jain,Raj Mal Jain, and the team’s driver,Chaman Lal Tak.

If we learned anything from ourglobe trotting experiences, it would in-clude the following:d The only thing one can be sure

about if ships are involved is thatthey will arrive late, and the peoplein country will know about it beforewe do.

d Any GONG deployment teammember is by nature and trainingmore dangerous to his or her col-leagues than any indigenous insects,reptiles, or sea creatures.

d The weather will always be gooduntil it needs to be.

d When the going gets tough, trust thelocal people. They’ll know what todo! Rob Hubbard

11111111111111111111111111111111111111111111111111111111111111111

3333333333333333333333333333333333333333333333333333333333333333

Instrumentation

33333333333333333333333333333333Where is the best place

to put the focus? ( Part II )

In the last Newsletter we discussedthe issue of how to focus the GONGinstruments to best avoid problemswith spatial aliasing. The article end-ed with the words “Stand by …”.This is a further report about the issue.

All imaging helioseismology instru-ments have to deal with spatial alias-ing. Unlike the well studied problemof spectral contamination by temporalgaps and crosstalk with modes of near-by l and m values, there has been littlestudy of spatial aliasing and its effectsin helioseismology. Stated simply,discretely sampled images of the Suncan produce a result in which a high-lspherical harmonic may mimic a low-lharmonic if the image is undersam-pled. The result is that differentmodes may appear at the same place inan oscillation spectrum.

The usual way of avoiding the prob-lem is to either use a large number ofpixels and/or a small aperture tele-scope such that the highest spatial fre-quencies present in the image arebelow one-half the sampling frequency( known as the Nyquist frequency ).For CCD arrays of a few hundred pix-els on a side, this means a telescopeaperture of less than one cm. Suchsmall apertures have been used with afew moderate l helioseismology obser-vations, but more typically a somewhatlarger aperture is used because of theneeds for accurate guiding and greaterlight flux to obtain high signal-to-noiseratio measurements. There is also anatural desire to try to observe modeswith l values as close as possible to theNyquist frequency. These pressures toviolate the sampling theorem lead tothe spatial aliasing problem.

As explained in the last Newsletter,the GONG instrument was designed toattenuate high spatial frequencies inthe image of the Sun before it is

- 5 -33333333333333333333333333

The Mauna Loa Station was commissioned on June 15. Theinstruments of NOAA’s Climate Monitoring and Diagnostic Laboratorycan be seen in the background. The GONG station is being operatedin collaboration with NCAR’s High Altitude Observatory site there.

22222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111

22222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222221

1111111111111111111111111111

22222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111

sampled by the CCD detector. Theattenuation is done by a combinationof a small amount of spherical aberra-tion and a larger amount of defocus.Atmospheric seeing also attenuateshigh spatial frequencies. When it is infocus, the GONG instrument violatesthe sampling theorem. The result isspatial aliasing. It is possible to de-focus the instrument enough to satisfythe sampling theorem fairly well butthe penalty is significant attenuation ofmodes with l below the Nyquist fre-quency.

Also as explained in the lastNewsletter, the initial plan for theGONG instruments was to defocusenough to attenuate l = 250 modes byabout a factor of six. With the currentCCD cameras used by GONG, thisamount of defocus attenuates spectralfeatures of moderate l relative to theinstrumental background noise levelmore than is desirable.

We have evaluated the options ofstaying with this focus, going to anin-focus setting, and an intermediatefocus setting. Many competing factorsinteract to make the choice difficult.

11111111111111111111111111111111111

The goal is to keep the solar back-ground noise and p modes above theinstrument noise level up to the Ny-quist frequency while attenuating thesefeatures to a level smaller than the in-strumental level above the Nyquist fre-quency. This will not eliminate spatialaliases completely because long in-tegrations will reveal narrow spectralaliases against the smooth spectralbackground. For the record, theamount of defocus chosen is 1% of thecamera lens focal length whichoperates at f/25. This choice increasesthe ratio of the response of the systemat l = 100 to its alias at l ~ 600 by morethan two orders of magnitude in powercompared to an in–focus system.

Since most other helioseismologyimaging instruments also suffer fromspatial aliasing, it is time to explicitlydeal with it in the production andanalysis of p-mode spectra. This is aninteresting challenge. There areseveral characteristics of the aliasesthat may allow them to be “dealiased”.For example, one might produce twospectra from the same data, one favor-ing the disk center where aliasing is

11111111111111111111111111111111111111111111111111111111111111111

strong and one favoring away fromdisk center where aliasing is weak.The difference should allow aliases tobe distinguished from non-aliasedfeatures.

Cliff Toner has made significantprogress in establishing where the spa-tial aliases fall in the m-ν spectra, andwe are optimistic about our prospectsfor dealing with the artifacts until ahigher resolution camera becomesavailable. Jack Harvey33333333333333333333333333333333

Timingof Helioseismic Observations

It may seem to be a simple thing todecide how to time helioseismic obser-vations. For the GONG observations, itwas decided long ago to synchronizethe observations at all the sites and tointegrate observations for exactly oneminute. The main issue remaining iswhen to start the one-minute integra-tion periods. Why not start on minuteticks? When we refer to time casually,we mean the time used in everyday ac-tivities. This is Coordinated UniversalTime (UTC ) plus some number ofhours appropriate for the time zone ofthe user. UTC is widely available andis maintained to about 1 millisecondsynchronization among various coun-tries. Although UTC has the same rateas atomic time, the varying rotation ofthe Earth requires that sometimes leapseconds be added or subtracted to UTCat the end of June or December. Hav-ing a minute of 59 or 61 seconds dura-tion every once in a while can be anobservational and instrumental nui-sance.

A better observational time systemis international atomic time (TAI )which does not have leap seconds andis the current SI time standard. During1995, UTC minute ticks are 29seconds later than TAI minute ticks.The Global Positioning System (GPS )distributes time via 24 satellites that issynchronized to TAI to an accuracybetter than a microsecond. Each

- 6 -33333333333333333333333333

The Cerro Tololo station — dwarfed by the nighttime telescopes of theCerro Tololo Interamerican Observatory and the Chilean Andes —came online on July 20.

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222221

111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111

GONG site has a GPS receiver to ob-tain GPS time. GPS time started atmidnight on 5 January 1980 with a de-lay of 19 seconds relative to TAI ( i.e.in synchronization with UTC on thatdate ). Thus, a solution to when to startGONG minute-long integrations issimply to use GPS minute ticks.

Things get more complicated whenwe want to coordinate observationsmade by different projects. In particu-lar, it seems to be a good idea to try totake both GONG and SOI data in syn-chronization. The SOHO spacecraftwill be five light seconds closer to theSun than the Earth. Therefore, GONGshould start its one-minute integrationsfive seconds later than SOI. TAI willbe available on the SOHO spacecraftand it is planned to center SOI obser-vations on TAI minute ticks. Thus, fora one-minute observation, SOI willstart 30 seconds after a TAI minutetick and GONG should start 5 secondslater, or 35 seconds after the TAIminute tick. Since the GPS tick is 19seconds after the TAI tick, this meansthat GONG should start 16 secondsafter the GPS tick.

111111111111111111111111111111111111

This complicated situation can besummarized as a progression ofevents:0 sec - TAI minute tick ( simultaneous

on SOHO and Earth )19 sec - GPS minute tick29 sec - UTC minute tick ( until 1996

January 1 )30 sec - SOI start of one minute obser-

vation35 sec - GONG start of one minute ob-

servationThis scheme for taking GONG data

started on April 11, 1995.There are several subtle issues of

timing. For example, as Earth movestoward and away from the Sun duringthe year, the observed frequency andphase of a coherent oscillation of theSun will vary by a small amount.Also, an oscillation seen at the limb isseen about 2.3 seconds later than thesame oscillation at disk center due tothe increased light travel time. As theSun’s brightness waxes and wanesduring a day at a GONG site, theintensity-weighted center time of aone-minute integration will varyslightly in such a way as to make theobservations seem to be taken a bit

11111111111111111111111111111111111111111111111111111111111111111

faster than one per minute. This latereffect on frequencies is small, of theorder of a part in 200000. It can shiftthe velocity zero point by up to 2 m/sfor observations made when thebrightness is rapidly changing.

For those interested in matters oftime, the Astronomical Almanac isrecommended. The latest informationabout leap seconds can be found onthe World Wide Web at URLhttp://maia.usno.navy.mil.This URL also points to other sourcesof excruciating detail about time andrelated matters. A discussion aboutsynchronizing GONG and SOI isfound in SOI Technical Note 24 byBogart, Bush and Harvey.

There is a second, less critical issueof synchronization between differenthelioseismology projects. This has todo with the length and start times ofreduction periods. In the GONG pro-ject, data products such as l-ν diagramsare generated on daily and roughlymonthly bases. The GONG day is 24hours long. It starts with the first one-minute integration that falls entirelywithin a new GPS day. In otherwords, the first observation of aGONG day is the one that starts at00:00:16 GPS time or ( in 1995)00:00:06 UTC.

Users of DSDS ancillary and qualityassurance products ( not time series ofvelocity images ) should be aware thatdaily products from single sites arestored using the local date in the namerather than GPS date. This is to avoidthe problem that some sites experiencetwo GPS dates during an observingday. Users of ancillary items in indivi-dual image headers should check withthe project to determine what time theparticular item was measured within aone minute interval.

The GONG ’month’ is 36 GONGdays. The project will reduce a’month’ of data every half ’month’.The first GONG Month started on May7, 1995. Jack Harvey

- 7 -33333333333333333333333333

The Big Bear station in the San Bernardino mountains of southernCalifornia, with the Big Bear Solar Observatory in the background,came online on August 17.

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222221

111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111

3333333333333333333333333333333333333333333333333333333333333333

Operations

Just a few short months ago all sixGONG instruments and the tools need-ed to deploy them were lined up out atthe integration site in Tucson, awaitingtrucks, boats and airplanes to carrythem to their destinations around theworld. Now we’re a network. All sixsites have been deployed, and the oldprototype has been moved from theRincon Vista site to the Farm, where ithas been upgraded to production stan-dard. That station is also operational.

The significant organizationalchanges that have resulted from thischange in our lives are outlined start-ing on page 17. In summary, there isnow a group of people within the pro-ject whose primary job it is to operateand maintain the network.

The highest priority of the new Net-work Operations group is keeping thedata flowing. As one might expect,the most critical people in this regardare the dedicated folks working withus at the host sites. We have provid-ed them with remote status boards,usually located in an area at the hostsite where the technical people spendmost of their daylight hours. For ex-ample, at Big Bear, the status terminalis located in one of the offices in thedome, while in Learmonth, it’s locatedin the radio (RSTN ) building. Whatthe local folks see is a video simula-tion of green, yellow and red lights in-dicating the status of the instrument.Green means the instrument is opera-tional and all monitored parameters arewithin operating limits; yellow meansa parameter has become marginal, butthe bits are still flowing ( or could ifthe Sun were out ); red means thatthe site is no longer operational. Thedisplay also shows an intensity imagederived from the digital data and up-dated every minute, and a magneto-gram derived from the most recenthourly observation. A graphical userinterface allows the on-site personnelto enquire as to the state of about 100monitored instrument and weather

11111111111111111111111111111111111111111111111111111111111111111

parameters, and generally determinethe exact source of trouble shouldproblems arise.

What we have noticed so far innearly six months of operating instru-ments in the field is just how good thepeople at our host sites are. We havehad very few serious problems so far,and in many cases the scientists andtechnicians at the site have been ableto fix things without our help at all!But when serious problems arise, thefirst step is to contact the GONG Net-work Operations Duty Responder( know locally as the “Duty Dude” ).That is an individual within the Net-work Operations group whose job it isto monitor the Internet, the phone, andfax machine for trouble and service re-quests. Only two of our sites are notcurrently on the Internet, and they ex-pect to be soon. Thus, electronic mailhas proved to be the best way to con-tact us and get the group here workingon the problem.

So what does the traffic from thesites look like so far? A review of theactivity over the last 30 days lookssomething like this:

111111111111111111111111111111111111

Issue Service Requests

Routine Correspondence 6Tape problems 4Modem problems 2Network Connections 2Computer hung 1Image Display 1Shelter Thermostat 1

The “Routine Correspondence”items include things like shortages ofsupplies, and other notes and com-ments on the routine operation of theinstrument. We have had some prob-lems with the modems at two of thesites, requiring in both cases some in-tervention by local experts. The socalled “Network Connections” prob-lems involve failure of the local-areanetwork at two of the sites, which tem-porarily kept us from obtaining statusand operations traffic over the Internet.( The modems are the fall back in thissituation. ) The “Hung Computer” wasthe SPARC station ( not actually partof the data path ), and was fixed with areboot; the cause is still unknown.

The biggest problem is data taping,and this one-month score card some-what understates the problems that we

- 8 -33333333333333333333333333

The Udiapur station at the Physical Research Laboratory’s UdaipurSolar Observatory, which came online on 3 October completing thenetwork, is watched over by a Hindu temple in the background.

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222221

111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111

have been having with tapes and tapedrives. Two of the service requestsduring the last month relate to Exabytetape drives that had to be replaced attwo different sites with new unitsshipped from Tucson. Prior to this,we have had a number of other prob-lems with these and other drives, re-quiring operator intervention in mostcases. These were also caused by tapedrive hardware problems or media er-rors. Fortunately each site has four ofthese drives in service, and two moreavailable as additional backups. Forthis reason, we have lost only aboutsix hours of data from a single site asa result of these various failures sincethe first station came on line in March.Still, we’re sure that the Exabytes arethe biggest source of frustration for theon-site personnel, to say nothing of theoperations group here! We are anxi-ous to have a look at our data-tapingsoftware to see if we can’t make itmore robust to hardware failure, andsee if we cannot take better advantageof the redundancy that we have at ourdisposal at each site.

In addition to this sort of “passive”support of the GONG network (that is,sitting by and waiting for the phone toring ), we also have some active pro-cedures in place. Every morning( Tucson time ) someone remotely con-nects to each station, looks at thestatus, and reads any messages sent tous by our hosts, or by the operatingsoftware. Next, we download the last24 operating hours of instrumental en-gineering parameters. These minute-by-minute values, such as power sup-ply voltages and motor currents, areanalyzed daily back in Tucson forunusual behavior, and become part ofa data base allowing longer-termanalysis for trends and signs of fatigueand imminent failure.

Keep an eye on this column in fu-ture GONG Newsletters to see whatworks, what doesn’t, and what we’redoing to fix it. So far, we are verypleased with the reliability of the sta-tions over all. Our goal is to bury youin excellent data! Rob Hubbard

11111111111111111111111111111111111111111111111111111111111111111

3333333333333333333333333333333333333333333333333333333333333333

Data System

333333333333333333333333333Data Management

andAnalysis Center

During the past six months, theDMAC successfully integrated the al-gorithms for refining the limbgeometry and extracting MTFs and formerging multi-site data using a MTF-derived weighting scheme into the datareduction pipeline. The DMAC re-duced network data acquired by sixdeployed network instruments (includ-ing the prototype in Tucson ), test dataacquired to verify the successful instal-lation of six instruments, and test dataacquired as construction of the observ-ing instruments was completed at theUniversity of Arizona farm site beforethe units were disassembled for ship-ment. The project also generated dataproducts from the first GONG month( 36 days ) of network data that was ac-quired by the three-site mini-networkof Teide, Learmonth, and the Big Bear

111111111111111111111111111111111111

station in Tucson. This includedmonth time series, power spectra, andmode frequency information that weremade available to the participants inthe inversion workshop that was heldin Boulder.

With the onset of network opera-tions, the data reduction activity in theDMAC has increased dramatically.For example, the number of site dayscalibrated per calendar month has in-creased from 25 in February to 152 inAugust. The upstream data reductionstages that perform site-dependent pro-cessing (VMICAL, GEOMPIPE,DNSPIPE, and AVER ) have ramped-up more quickly than anticipated. Ex-cept for AVER, these reduction stageshave been maintaining backlogs ofabout 50 site-days.

A key technical hurdle for mergingmulti-site image was the precise deter-mination of the orientation of the cam-eras in the observing stations. A pro-cedure developed and implemented bythe instrument team successfullysatisfied this requirement. This infor-mation was integrated into the datareduction software that registers thevelocity images into heliographic

- 9 -33333333333333333333333333Next Meeting of Inversion Team in Tucson in December

The Inversion Team is planning to have an extended working session in Tucson during an interval startingabout Wednesday December 6, 1995 and ending on about Tuesday December, 19. We are hoping to havethe greatest overlap of team members during the middle of that interval, from about December 10 to 17, butrecognize that various individual schedules may favor coming earlier or leaving later. The primary objective isto have close interaction among the inversion team members and resident GONG project experts indiscussing the scientific interferences for differential rotation and mean structure that can be drawn from boththe multi-station and full network data that has become available by then. A number of inversion teammembers are typically in twice-weekly contact by workstation conferencing to discuss progress related tofrequency fitting and inversion of the resulting data sets, and on developments related to the joint inversionworkbench. If anyone would like to join in those conferencing sessions, please inquire of either SylvainKorzennik or Mike Thompson for details on how to participate. Further, all GONG members are most welcometo join in the general activities of the inversion team, and to join also in the Tucson working session if their timeand interests permit. More details about the December working session, with opportunities to work actively onthe GONG data using the inversion workbench tools on a suite of workstations and x-terminals, will beavailable from either Douglas Gough or Juri Toomre (inversion team co-leaders ) at least by 10 November.( Prior to that Toomre will be on travel in India. )

2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222221

1111111111111111111111

22222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111

coordinates.In general, merging the mode

coefficient time series from multipleobserving sites has proven to be lessproblematic than anticipated. By theend of the exercise that merged thefirst GONG month, this data reductionstage was operating routinely. At thiswriting, all available data through July30 has been merged.

The DSDS is moving the user inter-face for the catalog query and dataproduct request functions fromCURSES to HTML. This developmentwas released for testing and was dis-cussed with the DMAC Users’ Com-mittee at the recent meeting in Bould-er.

The DSDS also recently conducted asurvey of its users regarding varioustechnical aspects of using the projectsfacility for accessing data products.The DSDS received 25 responses fromthe 64 GONG members to whom thesurvey was emailed.

With the ramp-up of network opera-tions, the frequency and volume ofdata requests has also increased.Eleven data distributions totaling 9.7GB were made in July and six distri-butions totaling 19.2 GB were made inAugust. This compares to zero distri-butions in February and two distribu-tions totaling 101 MB in March.

11111111111111111111111111111111111111111

The DMAC Users’ Committee metat HAO in Boulder on August 14.Dave Hathaway’s report on the meet-ing appears elsewhere in thisnewsletter. The meeting focussed onthe first GONG month data productsand the ramp-up of network opera-tions.

Marguerite Rodriguez who had beenworking as a DSDS operator left Tuc-son in March, but we were fortunate tobe able to quickly fill the vacancy withRobert Perry. Rob operates the DSDS,performs various system administrativetasks, and develops and supports data-base applications. In August, TatiaDeKeyser joined the project. Tatiaoperates the data reduction stage thatrefines the geometry of the solar im-ages and extracts the MTF. Jim Pintar

11111111111111111111111111111111111111111

3333333333333333333333333333333333333333333333333333333333333333Data Management and

Analysis Center( DMAC ) Readiness ReviewAn expanded DMAC Users Commit-

tee met in Tucson on March 16th for areadiness review of the DMAC facili-ties and pipeline software. The reviewcommittee consisted of Tim Brown,David Hathaway ( chair ), ToddHoeksema, and Steve Tomczyk. ( Syl-vain Korzennik and Mike Thompsonwere unable to attend. )

The GONG staff members presentedreports on the current status of theGONG pipeline software as well as thestatus of the DMAC itself. The com-mittee listened to these reports, askedquestions for clarification, and provid-ed a list of recommendations. This re-port consists of a brief overview of thepresentations along with the recom-mendations themselves.DMAC Overview: Jim Pintar gave anoverview of the DMAC. This facility isdesigned to develop and support thedata reduction and analysis softwarefor reducing the GONG field data toproduce the Data Products. The facili-ty also supports the GONGcommunity’s access to and use ofthese Data Products. The DMAC iscomposed of several subsystems: the

- 10 -33333333333333333333333333GONG Merged Velocity Mode Frequencies Available

The GONG DMAC announces the first GONG Velocity Merged ModeFrequencies are available for distribution. There is a table for each l-degree, 0-250. Each table will be distributed as a compressed text file.The set of tables is 75 MB, compressed, 376 MB uncompressed. Theentire set of tables can be ordered as as online distribution. TheDSDS will assist any GONG member with repackaging the set ofcompressed files for optimum remote transfer to his/her homeinstitution. To transfer the 251 files using ftp, invoke ftp with the flag toturn off interactive prompting during multiple file transfers. Within theftp protocol, set the transfer type to binary and execute the mgetcommand. Tape distributions are always available. If you have anyquestions or encounter any problems after receiving a mode frequencydistribution, please contact, the DSDS operators, Wendy Erdwurm,(520)-318-8371, Rob Perry, (520)-318-8523, or via e-mail todsds_operators@noao.edu.

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222221

111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222221111111111111111111111

Off-site Copy Facility (OCF ), theData Storage and Distribution System(DSDS ), the Field Tape Reader (FTR ),and the data reduction pipeline with itsown subsystems. The facility current-ly includes 17 workstations ( mostlySun SPARC machines running SUNOS4.1.3 or SOLARIS 2.4 but also includ-ing two DEC5000s, a DEC3100, andan RS6000 ) with 120 GB of diskspace, 32 Exabyte tape drives, and aExabyte Cartridge Library with a capa-city for 10,000 cartridges. There arecurrently 12 full-time employees withplans for four additions as the networkis deployed. Each data processing stepmust be capable of processing tensite-days of data per day in order tokeep cadence with the flow of datafrom the six-site network. This as-sumes that data is not processed onweekends, holidays, and vacation days.Off-site Copy and Storage Facilities(OCF/OSF ): Jean Goodrich reportedon the OCF/OSF. The OCF producesduplicate copies of all archived dataproducts and stores them in the OSFlocated in the basement of the KittPeak Vacuum Telescope building.This activity does not require continu-ous operator attention and can be runas a background activity. One out-standing issue concerns finding a reli-able method for testing the cartridgequality.Data Storage and Distribution Sys-tem (DSDS ): Wendy Erdwurm andMarguerite Rodriguez described theDSDS. The DSDS does the catalogingand archival storage of the data pro-ducts, supports the data reduction pipe-line, and provides the user interface fordata product queries and requests.This system is the central hub for theDMAC. It consists of twoSPARC10/51s and a SPARC20/61 ( thecommunity’s User Machine helios )running SOLARIS 2.4 with ORACLEas the database management tool.GONG member accounts are providedon helios to enable members to querythe GONG data product database andto order GONG data products. Thedata products are distributed via

11111111111111111111111111111111111111111111111111111111111111111

network or tape with the network dis-tributed products being placed in theusers space on helios. An anonymousftp area is also maintained on heliosfor standard datasets and a WWW in-terface is planned. On-line storage iscurrently one GB. Each of the twoDSDS operators can currently process20 new data tapes per hour, make tenon- line transfers per hour, and pro-duce two three-GB data distributionsper day. Future user demand for bothdisk space and data distribution arestill somewhat uncertain.Field Tape Reader (FTR ): Dave Ar-met described the FTR. This system isthe interface between the data acquisi-tion sites and the DMAC. Data tapesare received from the GONG field sta-tions, copied to disk, written to tapefor pipeline analysis, and data charac-teristics from the headers are plotted.The original field tapes are then storedin the OCF. Two one-week field tapescan be processed per day with thecurrent system.

Data Pipeline Processing Steps

Calibration (VMICAL ): Dave Armetalso described the pipeline calibrationstep VMICAL. This step takes datatapes from the FTR and produces cali-brated velocity ( V ), modulation ( M ),and intensity ( I ) images that are thenwritten to tape. These images are also

111111111111111111111111111111111111111111

viewed on screen and bad images arenoted by the operator and flagged assuch. Single image gaps are filled byinterpolation and VMI images takenfrom the magnetograms are rescaled tomatch the standard VMI images.Currently four to five site-days of datacan be processed each day by twooperators.Hankel Geometry (GEOMPIPE ):Jean Goodrich described the GEOM-PIPE pipeline procedure. This pro-cedure takes VMI images from VMI-CAL and produces a more accuratemeasure of the image geometry alongwith a measurement of the imagemodulation transfer function (MTF ).The MTF is then used in the mergingstep further down the pipeline. GEOM-PIPE performs a flat field correction tothe intensity images, calculates the op-timum geometry and updates thegeometry header parameters, and cal-culates the MTF for the intensity im-ages. Current performance on aSPARC 10 allows for the reduction ofone site-day in three hours. Nearlyhalf of this processing time is associat-ed with reading data from tape andthen writing it back to tape afterward.Site-day Mode Coefficients(DNSPIPE ): Ed Anderson outlined theDNSPIPE pipeline step. This steptakes calibrated velocity images fromVMICAL and GEOMPIPE and

- 11 -33333333333333333333333333Query: Order GONG DATA via WWW

The DSDS announces the release of an HTML (HyperText MarkupLanguage) version of mqry and mdata. It provides GONG membersthe capability to query and order GONG data via the World Wide Web.The URL for the system is http://helios.tuc.noao.edu/gong_query.html.

The DSDS has created, in each member’s home directory on helios,the user, group writable directory, www. All files required as inputs to,or created as outputs by, the WWW server software reside in thisdirectory. A new chapter, Data Query and Distribution via WWW, hasbeen added to the DSDS User’s Guide.

22222222222222222222222222222222222222222222222222222222222222222222111111111111111

22222222222222222222222222222222222222222222222222222222222222222222111111111111111 22222222222222222222222222222222222222222222222222222222222222222211

1111111111111

222222222222222222222222222222222222222222222222222222222222222222111111111111111

produces a time series of sphericalharmonic coefficients and an l-ν di-agram for each site-day. The velocityimages are remapped to heliographiccoordinates, filtered to remove the lowfrequency velocity signal, apodized,and then projected onto spherical har-monics. Gaps in the spherical har-monic coefficient time series are filledand then power spectra and l-ν di-agrams are produced. One operatorusing a SPARC 20 can process tensite-days per 8-hour working day.Multi-site Merge (MERGE ): WinifredWilliams described the multi-sitemerging process MERGE. This processtakes the individual site-day timeseries along with the image MTFs toproduce a merged network-day timeseries of mode coefficients. The merg-ing process calculates weights fromthe MTFs and produces a weightedmean of the individual site coefficientsfor the network coefficient. This pro-cess takes about one work-day on aSPARC 20 to produce a network-daytime series.Month Products (Time Series,Power Series, Frequencies ) (MTS,MPS, PEAKFIND ): Ed Andersondescribed the pipeline procedures forthe GONG month products. Thesesteps start with the network-daycoefficient time series from theMERGE process. Every 18 days a 36day GONG month time series (MTS )of the coefficients is produced. Powerspectra (MPS ) of these time series arecalculated and the frequencies of theindividual modes are found (PEAK-FIND ). These three procedures must

11111111111111111111111111111111111111111111111111

be completed in 18 days to keepcadence with the network. It currentlytakes about 2.3 days to produce thetime-series, 3.1 days to produce thepower spectra, and about nine days tofind the peak frequencies. The PEAK-FIND algorithm is still evolving. Theadditional complexity of futureprocesses may override any increasesin efficiency that might save process-ing time on this step.Four-minute Cadence Time Aver-aged Images (AVER ): Jim Pintardescribed the procedures (AVER ) forproducing the time-averaged imagesfor the low frequency modes and thenearly steady flow components. Theseprocedures take VMI images fromVMICAL/GEOMPIPE and apply a 17-minute temporal average to remove thep-mode oscillations. This new timeseries of images is then resampled atfour minute intervals. In addition, thevelocity images have the signal due tothe observer’s motion removed and adetrended velocity image with solarrotation removed is also produced.Currently this step requires one day toprocess eight site-days.Findings and RecommendationsThe review committee was delightedto find that all the necessary pieces forthe GONG pipeline now exist in work-ing form. Considering some of thegloomy predictions of several yearsago and the uncertainties related tosome processes, amazing progress hasbeen made. The committee was par-ticularly happy to see the imagegeometry (GEOMPIPE ) and merging(MERGE ) sections up and running.

11111111111111111111111111111111111111111111111111111111111111111

The calibration procedures (VMICAL )also seem to be well in hand. Wewere all impressed with the attentionto detail shown by the GONG staffmembers in each of the task areas.We found that all of the data process-ing steps now operate at a cadenceclose to that needed for operating thesix-site network. We did find someneed for improvements and offer thefollowing recommendations.1. The GONG project should make

an initial one-time effort to push athree-site network month ( 36days at ~80% ) through the entirepipeline at the earliest reasonabletime. This would exercise theDMAC; provide an important testof the pipeline from end-to-end;and put real data in the hands ofthe GONG community for ana-lyses that might reveal unforeseenproblems. After digesting theresults of these analyses, GONGshould focus on getting thesoftware into final form.

2. GONG should continuehardware/software improvementsto ultimately get the processingsteps done in half the time neededto maintain cadence with the net-work. Many steps seem to be un-comfortably close to being righton cadence. This will make theinevitable need for reprocessingdata very difficult to undertake.Since several steps seem to be I/Obound, hardware upgrades maygo a long way toward reachingthis goal.

3. The time, both human andmachine, consumed in I/O is asignificant part of many process-ing steps. Substantial savings, atleast on the input side, could berealized if the verification of out-put tapes were in some casesseparated from the analysis pipe-line and data were made availableto the following pipeline elementon disk. We urge the project toevaluate this possibility.

4. Pipeline steps VMICAL andGEOMPIPE should be combined.

- 12 -33333333333333333333333333GONG ’95 - Asilomar Proceedings

Advance copies of the Proceedings arrived on October 3 via airmail.Thanks and congratulations to Bruce Battrick, Todd Hoeksema, andMargie Stehle for their expeditious handling of the editing andpublication. The rest should be following along soon.

2222222222222222222222222222222222222222222222222222222222222222222211111111111

2222222222222222222222222222222222222222222222222222222222222222222211111111111 2222222222222222222222222222222222222222222222222222222222222222221

111111111

2222222222222222222222222222222222222222222222222222222222222222221111111111

GEOMPIPE takes VMI imagesfrom VMICAL and outputs VMIimages with additional header in-formation on the image geometry.Combining the two would elim-inate one I/O cycle. Other stepsshould be evaluated for similarsavings.

5. The MERGE step is particularlyI/O bound. We recommend in-vestigating adding disk or tapedrives to alleviate this problem.

6. The image characterizationmethod should be improved. Thecurrent procedures merely flagimages as good or bad. Since aninformed judgement is beingmade to reject images, it wouldbe helpful to record why, perhapsby simply making a selectionfrom a standard list of reasons, sothat others who might want tokeep such images have somemeans of identifying them. Itwould also be good to documentwhat criteria are used in rejectingimages, what characteristics areused to decide whether an l-ν di-agram is good or bad, and todevelop some level of standardi-zation among the evaluators.

7. It would be desirable to havemore data available on-line. Wesuggest that GONG keep track ofwhat datasets are most requestedand then provide more disk spacefor keeping some of these on-line.

8. GONG should survey the com-munity to better determine whatsoftware and data will be request-ed and in what volume. Thisshould include ascertaining howmuch employee time will beneeded to support the GRASPsoftware and provide help to thecommunity in its usage.Volunteers from the communitymight be enlisted to help with thislast point and in establishing andmaintaining a library of contribut-ed software.

9. GONG should “cross-train” indi-viduals so that at least two peopleare capable of running each

11111111111111111111111111111111111111111111111111111111111111111

section of the pipeline. The pipe-line should not come to a standstill when a crucial staff memberis unavailable.

10. GONG should strive to maketheir computing environmentmore uniform. The pipeline ele-ments currently run on severaldifferent makes of machine underseveral different operating sys-tems using different shells and en-vironmental variables. It is cru-cial that data be processed withdocumented and known software.We are concerned that this couldbe a source of problems.Dave Hathaway

3333333333333333333333333333333333333333333333333333333333333333

10th DMACUsers Committee Meeting

The DMAC Users Committee(DUC ) met at the GONG DMAC inTucson on Friday March 17, 1995 fol-lowing the DMAC Readiness Review.In attendance were committeemembers Brown, Hathaway ( Chair ),Hoeksema, and GONG representativesAnderson, Erdwurm, Hill, Leibacher,Pintar, and Toner. Korzennik andThompson were unable to attend butTomczyk sat in for part of the discus-sions.

Hill presented the status of GONG.Comparisons between instruments,both at common sites and at multiplesites, have been performed. The BigBear instrument results have beencompared with those from the Teide,Udaipur, and Learmonth instrumentswhile all were at the farm site. Com-paring time series for mode

111111111111111111111111111111111111111111111111111111

coefficients shows good agreement.The Big Bear instrument results havealso been compared with results fromthe Teide instrument after the latterwas deployed at its site in the CanaryIslands. There was also good agree-ment between these results. The resi-dual differences are attributed to cam-era errors: offsets in camera angle andin the position of the axis of rotation.Current plans are to use drift scans tobetter determine the orientation. Thesecomparisons also reveal velocity scalefactor differences from instrument toinstrument. Additional efforts may beneeded to produce more uniformresults. (Hill’s responsibility. )

Erdwurm described the recent pro-gress on the DSDS. Work is underwayto produce a WWW interface to theDSDS that would be similar in func-tion to the present menu system. Thiswould allow easy access to the DSDSfrom a wide variety of platforms andsystems. Information on Data Eventswill also be more readily availableunder both systems. ( Erdwurm’sresponsibility. )

Hill discussed the GONG/SOI com-parison. This exercise was meant totest the GONG data processing algo-rithms and software by pushing datathrough these two ( almost ) indepen-dent pipelines. SOI data is not avail-able but the SOI project should beready to analyze GONG data withinthe next few months. Meanwhile, theanalysis of the artificial data string isin progress. ( Hill’s responsibility. )

Toner reported on the defocusingexperiments. Spatial aliasing from thehigh l modes was noticed in previousanalysis of GONG data. Althoughsome scientific information is con-tained in this signal, the DUC

- 13 -33333333333333333333333333suggested that the signal be removedto produce a cleaner spectrum. Tonerfound that defocusing the instrumentscan remove the aliasing but it alsoreduces the amplitude of the l = 250modes by a factor of about 5. De-focusing also introduces noise at thelimb which shows up as streaks atconstant ν in an l-ν diagram. Furthertesting is in order. ( Toner’s responsi-bility. )

Toner also reported on the imagerestoration procedures. The imageMTF is determined from the intensityimages and then used in a deconvolu-tion routine to produce clean intensityimages that are then used to produceVMI images. The process is still tooslow to become a part of the pipelinebut might be done elsewhere for in-terested GONG members. Additionalwork is needed to determine the low-lresponse of this procedure. ( Toner’sresponsibility. )

Anderson described his work on thepeak bagging procedure PEAKFIND.This procedure fits a Lorentzian on alinear background to each peak todetermine the frequency and width ofthe peak. Testing is underway withartificial data. Additional work isneeded to characterize errors and forhandling asymmetric peaks. Schou’sa-coefficient code is due to be import-ed and tested as well. (Anderson’sresponsibility. )

Anderson also reported the 16 vs. 32bit data storage. The images are wellrepresented by 16 bits. The 16 bittime series shows only very smalldifferences from the 32 bit time series.The DUC recommends storing data as16 bit FITS files with BZERO=0 andBSCALE set at 6 to 10 standard devia-tions above the signal in the timeseries so that outlying points don’t gooff scale. ( Anderson’s responsibility. )

Pintar presented the current status ofthe Low Frequency/Nearly SteadyFlow pipeline. This pipeline conveys17-minute Gaussian-weighted averagesof the VMI images at 4-minute inter-vals. Outstanding problems with thisdata include the presence of gradients

11111111111111111111111111111111111111111111111111111111111111111

and zero offsets that vary with time.Improvements in the calibration pro-cedures may alleviate these problems.At the moment the variations are wellfit by linear trends that can be removedfrom the data. ( Pintar’s responsibili-ty. )

The next DUC Meeting is set forThursday, August 17, 1995 in Boulder,Colorado. Brown and Hoeksema willbe completing their terms on the com-mittee and will be replaced byTomczyk and Schou. David Hathaway3333333333333333333333333333333333333333333333333333333333333333

11th DMACUsers Committee Meeting

The DMAC Users Committee(DUC ) met at HAO in Boulder onMonday August 14, 1995. In atten-dance were: committee members Ha-thaway (Chair ), Korzennik, Schou,Tomczyk, and Thompson, GONGrepresentatives Anderson, Erdwurm,Hill, Leibacher, Pintar, and Toner, andGONG community members Brown,Gilman, Gough, and Toomre.

Pintar presented the current status ofGONG. The site installations havegone better than expected with five ofthe six sites deployed at the time ofthe meeting. A new strategy for align-ing the telescopes to geographic northhas been developed. It relies on theproven technique of letting the imageof the Sun drift to establish an east-west line in the focal plane. Thisfunction then defines the locus of thedrifting limb, and it is fit with a modelthat allows the east-west direction tobe accurately determined. The DMACis now quite busy processing data fromthe first GONG month. Data requeststo the DSDS are increasing and someactivities of the Offsite Copy Facilityhave been taken up by the DSDS aswell. The Field Tape reader is keepingup with the current data flow. Howev-er, the pipeline processing is ex-periencing backlogs which are

11111111111111111111111111111111111111111111111111111111111111111

expected to grow to be as much as 6months long on some data products.The only staff change since the lastmeeting was hiring Rob Perry to re-place Marguerite Rodriguez. At leasttwo people are now capable of runningany of the processing steps as was sug-gested by the DMAC review ( with theexception of PEAKFIND, which is runby Anderson only, and AVER, which isrun by Pintar only ).

Hill described some of the resultsfrom the processing of the first GONGmonth of data. This first month in-cluded data from three sites (Lear-month, Teide, and Tucson ) with about75% coverage for 95/05/07 through95/06/11. The sidelobes at 1/day areless than about 1%. Several problemshave been identified in the data. Thepower spectra show problems at l = 0-3that have been attributed to changes inthe ephemeris velocity from one imageto the next that are not accounted forin the backward time differencing.There may also be some problems dueto camera nonlinearities. Other ar-tifacts are attributed to A/B cacheproblems (images on the even minutesare loaded in one image cache whileimages on the odd minutes are loadedin the second image cache ). Theremay also be an “alternate row prob-lem” within the camera. The problemthat received the most attention wasrelated to the measured frequencies ofthe modes. Below l = 150 the spatialsidelobes are well defined and eachone is fit to determine mode frequen-cies. Above l = 150 the spatialsidelobes begin to merge and adifferent process is required. An ap-parent consequence of this appears inthe frequency differences between thezonal modes and the tesseral modes.Below l = 150 these splittings show onebehavior. Above l = 150 they show adifferent behavior. Another area ofconcern is the asymmetries in thepeaks. The asymmetries found in thisfirst month of GONG data weredifferent from those found in SouthPole data from 1994.

- 14 -33333333333333333333333333A Virtual Helioseismological World

In order to improve communication amongst Team Members, the Inversion Team has put in place severalmechanisms for network conferencing over the Internet.

A first tool, called helioMOO, (aka ‘a virtual helioseismological world’, or should it be helioSeisMOO?), is anObject Oriented MUD, based on on Xerox Parc’s public domain LambdaMOO server. It was inspired by theAstroVR project, and is currently running at CfA, where I maintain it.

This server allows people to enter a virtual space with rooms (and other objects) that can be explored. Themain feature used by the Inversion Team is its chatting capability, that allows people that are in the same"room" to talk to each other over the net. Bi-weekly, one-hour long meetings are carried out in the InversionTeam Chat room. Logs of each meeting are kept and posted on the Web.

More information on helioMOO can be found at the URL:

http://www-sgk.harvard.edu:1080/~sylvain/helioMOO/

or via anonymous ftp at:

cfa-ftp.harvard.edu:/pub/sylvain/helioMOO

Related to the helioMOO, a public image exchange repository is available at CfA for incoming files, while asecond one for outgoing files is also in place. These are on the anonymous ftp site cfa-ftp.harvard.edu under/incoming/sylvain/helioMOO/images and /outgoing/sylvain/helioMOO/images respectively. By adding the URLs

file://cfa-ftp.harvard.edu/incoming/sylvain/helioMOO/images/

and

file://cfa-ftp.harvard.edu/outgoing/sylvain/helioMOO/images/

to your Web browser hot list ( or by following the hyperlinks in place at http://www-sgk.harvard.edu:1080/~sylvain/helioMOO/ ) you can exchange and view documents using your Web browser,and/or anonymous ftp.

Finally, in order to provide an interactive tool to view and annotate plots, I wrote an X-based whiteboardprogram, called xwb. The programs — available for SUN, SGI and Linux platforms — allow several users tointeractively draw and annotate an image. This tool is still being developed, but a beta version is available viaanonymous ftp in:

cfa-ftp.harvard.edu:pub/sylvain/xwb/

or at URL:

file://cfa-ftp.harvard.edu/pub/sylvain/xwb/

Sylvain G. Korzennik

2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111111111111111111111111 22222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222211

1111111111111111111111111111111111111111111111111

222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111111111111111111111111

Hill also described the results of thetests on the data merging process fromthe first month of data. Some noise isevident at low-l but it is diminishedwhen images with “clouds” are re-moved from the data stream. Evenwithout employing a “cloud killer” the36-day run looks good with no evi-dence of site-to-site transition prob-lems. Further assessments of themerging process should be undertaken.

11111111111111

( Hill & Toner’s responsibility. )Erdwurm described the progress on

the DSDS. The recent survey indicatedthat about 40% of the users connect tothe DSDS via PCs or MACs. Some ofthese users experience compatibilityproblems that need to be addressed.Several responses indicated the needfor data distribution via DAT tapes.The DUC recommended purchasingone of the newer units. A WWW

11111111111111

interface to the DSDS has beendeveloped. Members of the DUCliked the interface but had some prob-lems and suggestions. A reply orresponse feature should be added tosend problem reports to Erdwurm.Graphics bitmaps should be added toshow the product availability at muchfiner increments than daily. Links tothe GONG homepage and documenta-tion should be added. The DUC

- 15 -33333333333333333333333333

DRAT Workshop

GONG’s Data Reduction and Analysis Team (DRAT ) will have a working meeting in Tucson, November 13-15 at NSO. Thepurpose of the meeting is to review the early results from the network, address outstanding data reduction problems andanalyze the quality of the results for the first publication of results. We will, insofar as possible, produce a draft of the firstpublication. Five of the six sites are now operational. The sixth is being installed at Udiapur in September. There is amonth of 3-site data already processed; there should be at least another month of data, based on a greater number ofsites, processed and available by this meeting. The most important reduction issues at this time are spatial aliasing, asmall discontinuity in frequency splittings at l = 150 where the peak-finding algorithm changes, low-l noise stemming fromthe ephemeris correction and camera non-linearity, amplitude attenuation as a function of m, and the “A-B cache problem.”While some of these issues may be settled before the meeting, others are likely to crop up. The current plan calls forDRAT to report the standard GONG results, compare them to earlier work and to produce an error analysis of those resultsin a special series of articles in a May issue of Science. That error analysis is expected to encompass a validation of theinstrument, the calibration and the pipeline reduction and a noise analysis of the instrument and reduction process. TheDRAT article must be in a relatively finished form by February 1996. This three day workshop will be the main opportunityfor DRAT to fulfill its primary function. The first half day will consist of presentations on the status of the data reduction andcurrent issues. Attendees will then identify tasks to be done and break up into working groups to address those actionitems. There will be occasional gatherings to report progress and a final session to assemble results and to set schedulesand work assignments for the following four months. Please indicate your intention to participate by sending the followinginformation to stebbins@jila.colorado.edu: name, email address, address, phone, fax. If you are unable to attend thisworkshop, but wish to participate in this critical DRAT activity, please respond and so note. Authorship on these GONGpapers is determined by participation.

Tuck Stebbins

2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222221

1111111111111111111111111111

22222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222211111111111111111111111111111

agreed that the input history could beomitted and that the limit of 10 hitslists was sufficient ( but should bemonitored ). ( Erdwurm’s responsibili-ty. )

The project’s support for theGRASP software was discussed. Onlya handful of people outside the projectare using it at this time. Discussion ofany expanded support or futurereleases was delayed until the nextDUC meeting.

Hill discussed the GONG/SOI com-parison. GONG data from 94/10/30was processed through the SOI pipe-line. The low frequency end of thespectrum was somewhat better due tothe deterministic detrending used bySOI which does include ephemerismotion. Streaks through the k-ω di-agram at 1/hour frequencies werefound and attributed to the magneto-gram acquisition. Linear interpolationdiminished but did not eradicate thesefeatures. Further comparisons of timeseries and frequencies are needed.( Hill’s responsibility. )

Toner reported on the defocusingexperiments. Aliasing from modes

11111111111111111111111111111111111

with l > 250 indicated the need for somedefocusing of the instruments. Earlierexperiments had over-adjusted for thisproblem. Current procedures place theimaged plane 0.065" from the nominalfocus. This removes most of thealiased signal while reducing the am-plitudes at l = 250 to 1/2 their originalvalues. Some aliasing remains and thelongitude-latitude range had to be re-duced to ±54˚ from the previous ±60˚.This compromise solution was accept-ed by the DUC while noting that theoptimum solution would be to switchto a 10242 square pixel CCD withoutdefocusing.

Toner also reported on the imagerestoration procedures. These pro-cedures take the MTF from the Hankletransform in GEOMPIPE and then re-store the three intensity images to ulti-mately make restored VMI images.The advantage this has, over using theMTF to simply adjust the spectralcoefficients, is that it includes the m-dependent variations. The proceduresrequire about 19 seconds of computetime to restore the VMI images butthey do produce a much more uniform

11111111111111111111111111111111111

image series. The DUC supports a tri-al using the first GONG month of datato determine the usefulness of employ-ing these procedures as part of thepipeline processing. The DUC alsosupported keeping GEOMPIPE andVMICAL separated until these tests arecompleted. (Toner’s responsibility. )

The DUC had further discussions onthe peak bagging procedures.Although no immediate solutions wereidentified, the committee suggestedthat Anderson and Schou collaborateon an effort to identify possible solu-tions. ( Anderson’s and Schou’sresponsibility. )

Anderson reported that no furtherprogress was made on the 16 bit datastorage problems. The DUC hadrecommended storing data as 16 bitFITS files with BZERO = 0 andBSCALE set at 6 to 10 standard devia-tions of the signal so that outlyingpoints don’t go off scale. CurrentlyBZERO and BSCALE are not fixed.( Anderson’s responsibility. )

Pintar described the current status ofthe AVER data products. A set ofmerged and detrended velocity images

- 16 -33333333333333333333333333

Mode Frequencies for GONG Month 1

A first pass has been made at mode fitting on individual m-spectra, as well as m-averaged spectra. It isexpected that revisions to this table will be made in the next few months as the PEAKFIND procedures arerefined.

Total Number of Modes fit ( including spatial leaks ): 4,736,571

Total Number of Modes available: 1,048,713

Number of modes judged to be a good fit: 571,291

The following parameters are included: Radial order, Degree, Azimuthal order, Primary target or spatial leak,Fit frequency, Frequency error, Fit FWHM, Positive FWHM error bar, Negative FWHM error bar, Fit powerspectrum amplitude, Positive peak power error bar, Negative peak power error bar, Mode amplitude, Positivemode amplitude error bar, Negative mode amplitude error bar, Background power at mode frequency,Background power error at mode frequency, Background slope at mode frequency, Background slope error atmode frequency, Merit function value, Number of iterations taken to reach solution, Initial guess for modefrequency, Initial guess for mode width, Good or bad fit.

The following criteria were used for determining whether a fit was considered good or bad. The criteria werechosen to be fairly certain that the obvious bad fits were marked as such at the expense of possibly excludingsome good fits as well.

1)All fits to modes with frequencies above 5000 µHz were considered bad. This is due to problems fittingblended profiles where the spatial leaks are not resolved. The number of modes fit with frequencies ≤ 5000µHz is 768,672 Thus, 26% of these fits were considered suspect by the remaining criteria.

2)All fits to modes whose FWHM exceeded 5. × FWHMguess

were considered bad.

3)All fits to modes whose power was less than the background were marked bad.

4)All fits where the fitted frequency was greater than a FWHM away from the initial guess frequency wereconsidered bad.

5)All fits where the merit function was greater than 1.5 were considered bad.

6)All fits where the frequency error bar was greater than half the fitted FWHM or greater than 5 µHz wereconsidered bad.

Ed Anderson

22222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111111111111111111

22222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111111111111111111 22222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222211

11111111111111111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111111111111111111

is available on helios.tuc.noao.edu

in /usr/users/pintar/mr/mrvda/ forthe first GONG month. The magneto-grams should also be available.

Harvey and Toussaint are investigat-ing changes to the calibration pro-cedures to remove camera nonlineari-ties and to smooth the calibration im-ages. This should better decouple theVMI images and should reduce noiseintroduced by the calibration into theoscillation images. An investigationof the temporal stability of the calibra-tion images indicated that day-to-daychanges do occur in some instancesand that, at best, a set of calibrationimages can only be used within a 11-day interval centered on the day the

1111111111111111111111

calibration sequence was recorded.A pipeline error analysis was dis-

cussed. A formal error analysis hasnot been undertaken but would be ex-tremely useful ( although quite timeconsuming ). Hill outlined the pro-cedure and identified some of thesources of error. The DUC felt thatany large efforts in this direction wereunnecessary as long as the errors werein line with those expected from sim-ple analyses. The systematic errorssuch as the spatial aliasing are prob-ably more important to understand.

A final topic was the plan for futurepublications and the publication poli-cy. Leibacher indicated that therewould be a special session at the

1111111111111111111111

January AAS meeting in San Antoniowith presentations (probably ) byLeibacher and Toomre. A series of ar-ticles are scheduled for a May/June is-sue of Science that would have to beprepared by April 1996. This wouldinclude an overview of the project andteam reports on early results. The ear-ly results would then be presented atthe June AAS/SPD meeting in Madis-on, Wisconsin and published as re-ferred journal articles at about thattime. Until the publication of the Sci-ence articles there will be an embargoon articles using network data.

The final item of business was toprioritize some of the GONG DMACefforts. The DUC felt that the highest

- 17 -33333333333333333333333333Steady Flows Team Meeting and Publications

GONG plans to publish results for early GONG data in a series of articles inScience next March/April. We will be responsible for writing the paper onresults concerning the nearly steady surface flows. I would like to propose thatsome of us meet in Tucson in early December to workout some of the detailsof this paper. The GONG Inversions team is meeting there at that time and theDMAC Users Committee will also be meeting there. Please let me know if thisworks for any of you. I will be writing again in the near future with more details.

David Hathaway22222222222222222222222222222222222222222222222222222222222222222222222222211

11111111111111

2222222222222222222222222222222222222222222222222222222222222222222222222221111111111111111 22222222222222222222222222222222222222222222222222222222222222222222222221

11111111111111

2222222222222222222222222222222222222222222222222222222222222222222222222111111111111111

priority was to put Anderson to workon the peak bagging problems. Inparallel with this, Toner should test hisimage restoration procedures on thefirst month of data. The third prioritywas to continue to plow through theincoming data.

The next DUC Meeting is set forearly December in Tucson in conjunc-tion with the inversions team meeting.David Hathaway3333333333333333333333333333333333333333333333333333333333333333

Project Management

Without a doubt the last few monthshave been the most intense, and excit-ing, of any the project has experiencedto date. Beginning with the Decembershipment of the first station from theTucson Farm to Tenerife, until thepresent, when we have just completedthe Udaipur deployment, we have beenrunning at full speed.

Our annual budget struggle wasresolved favorably in February with a$2.6M figure for FY 95. This has beenadequate to allow us to deploy thefield stations at a rate of about oneevery six weeks, since the first stationarrived on Tenerife in January. Itshould be borne in mind that, sincethen, we have been simultaneouslycompleting the integration and testingof the remaining stations, completingsite land preparations, deploying sta-tions, and operating them as they cameon line. This has been an interestingmanagement challenge since the de-ployments have necessarily meant thatwe were carrying on these many taskswith half or more of our instrumentstaff thousands of miles away fromhome base.

Nevertheless, Frank Hill and RobHubbard, the on-site deployment-teamleaders, oversaw the successful com-pletion of all of the installations, aheadof schedule and under budget. Togeth-er with the deployment teams andthose who stayed at home, they ac-complished many other instrumenttasks as well.

11111111111111111111111111111111111111111111111111111111111111111

Among these other tasks has beenthe development and implementationof the network operations activities.Rob Hubbard, who has served as theAssistant Project Manager during thedevelopment and deployment phasesof the project has now been named asNetwork Operations Manager. Assuch, Rob will be responsible for over-seeing the day-to-day operations of thefield network. The operations grouphas established an effective program toaccess each instrument daily to retrieveand examine health and welfare statusfiles, and to rapidly respond and in-teract with our hosts’ staffs that lookafter the instruments in the field.

Jim Pintar and his DMAC crew havebeen as busy as any of the rest. Theyprovided the significantly increaseddata processing throughput required tocertify the stations in test at the Farm,to do the same for the remote stationsas they were certified on site, and tothen continue to process the data fromthe field stations as they progressivelycame on line. The climax of this ac-tivity was the production of the firstmerged data sets from the mini-network consisting of the first threesites to come on line, El Teide on Ten-erife, Learmonth Western Australia,and the Big Bear station temporarilyoperating in Tucson.

A sadder activity has been sayinggoodbye to some old friends. As theproject has essentially completed itsoriginal development mission and hasbeen making the transition to a net-work of operational observatories, anumber of people on the instrumentteam have had their work come to an

111111111111111111111111111111111111111111111111111

end. In the past few months, MarkTrueblood has joined the U.S. GeminiProject as its Project Engineer, BretGoodrich has also joined Gemini, BobHartlmeier has retired, Don Farris, Ar-den Petri, and Dee Stover have takenpositions in other NOAO programs,Jerry Gonzales, Jeff Vernon, and TomBajerski have all left NOAO for otheremployment. Their contributions tothe GONG program have been essen-tial and we wish them every success.

On the other hand, the DMAC’s realwork is just beginning. This hasmeant the addition of two staff posi-tions and plans for one or two moreover the next few months. More de-tailed reports on these various activi-ties will be found starting on page 8.

As usual, our FY 1996 budget is un-clear. Our original request includedfunds to vigorously pursue a newdevelopment program to produce a1024 × 1024 square-pixel camera thatmight be installed as a retrofit into theexisting field stations to produce evenhigher quality data. However, this re-quest was rejected in favor of a moremodest one which included little morethan the basic funding required tooperate and maintain the network anddata management facilities. Neverthe-less, it amounts to some $2.1M, and ifthis figure survives the Federal pro-cess, and the network is blessed withrelatively trouble-free operation, wemay be able to at least begin a modestexploration of the new camera con-cept. All in all, we are looking for-ward to our first year of “normal”operations with great anticipation.Jim Kennedy

- 18 -33333333333333333333333333Low Frequency Team Meeting

Many of you probably know that a publication of the early GONG results will besubmitted to Science next March. It is expected that the results of each teams’work will be presented at next summer’s AAS meeting. The detection of lowfrequency modes is the responsibility of our team and it would be highlydesirable to participate in the first GONG publication.

I’m writing this note for two purposes:

First, to see if there’s any interest in getting together as a team, perhapsadjacent to the Steady Flows Team meeting, to discuss our progress anddevelop an outline for our team’s contribution to the GONG initial resultspublication.

As I understand it the Flows team will meet in Tucson sometime in midNovember or early December. Feel free to propose and alternate site/time ifyou like.

d If we had such a meeting in mid November, would you come?Second, to encourage you to take a look at the initial GONG low & slow timeseries and to share your thoughts, impressions, and results with the rest ofthe team.

d Has anyone found anything interesting or disturbing the the data?

Please use the GONGLow@solar.stanford.edu alias to communicate with theentire team.

Todd Hoeksema2222222222222222222222222222222222222222222222222222222222222222222222222221

111111111111111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222222222222211

11111111111111111111111111111111

222222222222222222222222222222222222222222222222222222222222222222222222221111111111111111111111111111111111

3333333333333333333333333333333333333333333333333333333333333333Science

33333333333333333333333333333333

Presentations andPublications

The initial phase of the scientific in-vestigation has developed virtually asdiscussed at this year’s Annual Meet-ing in Asilomar. In order to get datainto the hands of the members to startworking with it to help understand theperformance as rapidly as possible —and to pursue the teams’ scientific ob-jectives — the deployment sequencewas chosen to achieve a three-site net-work, with roughly 120 ˚ separation atthe earliest possible moment. Theduty cycle for the first GONG monthof three stations was a quite respect-able 70%, and the data have now beenpushed all the way through the DMACpipeline. All of the data products - in-cluding our first cut frequencies - arenow available through the DSDS. Ifyou do not already have an accountwith the DSDS, the simplest way to setone up is through our WWW server.While you are browsing there, pleasedo check out the scientific programs.

With the availability of preliminarynetwork data, the Teams’ activities arenow shifting into high gear as theywork toward four presentations of theirwork:1) a special session at the January,

1996 AAS meeting in San Antonio,Texas where two overview talkswill be presented, to keep thebroader astronomical community ap-praised of the highlights of GONG’sinitial performance and results,

2) a series of articles are scheduled fora special issue of the journal Sciencein the May-June 1996 timeframe,that will have to be prepared byMarch-April. These will include anoverview of the project and team ar-ticles on early results. This will bethe first publication of GONGresults. Science is a refereed

11111111111111111111111111111111111111111111111111111111111111111

journal, with a very broad circula-tion, and it is a tremendous oppor-tunity to be invited to avail our-selves of this forum. It means a lotof work as well.

3) the AAS and its Solar Physics Divi-sion have invited us to hold GONG’96 in conjunction with their jointmeetings in Madison, Wisconsin inJune 1996. The Teams will presenttheir collaborative results in a spe-cial joint session with the entire so-ciety, to be followed by Team meet-ings.

4) the Teams will publish their work inrefereed journals following theGONG ’96 meeting.To help the coordinated analysis

along, we are establishing facilitieshere in Tucson to support extendedworking sessions of a dozen or so in-vestigators at a time, with worksta-tions, etc…, in close contact with theproject staff. If you haven’t heardfrom your Team leader, don’t hesitateto get in contact with them. The Net-work and the DMAC are up and run-ning. It’s time to join the fun!

1111111111111111111111111111111111

An important aspect of the invita-tion from Science to dedicate a specialissue to the first results from GONG isthat in fact this be the first publication.There will be no publications prior tothis time. The GONG Publication Pol-icy calls for a subsequent, more sub-stantial — in terms of the amount ofdata, work, and presentation — teampublication prior to individual publica-tions.

Individual investigators are remind-ed that their continued membership isbased upon1) providing a brief report on the status

of their investigations every sixmonths, for sharing among the othermembers,

2) submitting for notification to theother members of the appropriateteam oral presentations, seminars, ...

3) submitting for review by the Projectand the cognizant Team all publica-tions utilizing GONG data.Recall that the presentation and pub-

lication clearance forms are availableon the WWW. John Leibacher

- 19 -333333333333333333333333331996 Western Pacific Geophysics Meeting

July 23 - 27, 1996 Brisbane ( Australia )

Session SP01: Helioseismology

The GONG, IRIS, BISON, and TON networks of ground-based helioseismologyinstruments and the SoHO spacecraft’s SOI-MDI, VIRGO, and GOLFexperiments will be providing a tremendous injection of new data about theinternal structure and dynamics of the Sun. This special session will provide theforum for the discussion of these new results, and for their confrontation withtheory. ABSTRACT DEADLINE: March 15, 1996 Conveners: D. Cole, IPS -Radio and Space Services, PO Box 1548, Chatswood NSW 2057, Australia,tel: +61 2 414 8334, e-mail: david@ips.oz.au and J. Leibacher, National SolarObservatory, P O Box 26732, Tucson AZ 85726, USA, tel: +1 520 318 8305,e-mail: jleibacher@noao.edu.

2222222222222222222222222222222222222222222222222222222222222222222222222221111111111111111111111

2222222222222222222222222222222222222222222222222222222222222222222222222221111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222222222222211

11111111111111111111

222222222222222222222222222222222222222222222222222222222222222222222222221111111111111111111111

33333333333333333333333333333333Science article outlines

33333333333333333333333333333333Data Reduction and Analysis

1. PurposeTo describe the initial GONG dataproducts and demonstrate their reliabil-ity.2. ThemeThis is unique data.3. Contentsa) Describe main results.

(i) l, ν diagram(ii) Line profiles: compare to "best

previous" data.(iii)Frequencies, linewidths: com-

pare to "best previous" data.b) Error analysis.

(i) Validation of instrument, cali-bration, analysis.d Artificial data.d Comparison with other results

(accuracy).d Comparison between different

time strings (precision).(ii) Noise analysis.d Instrument and calibration -

systematic and random noise(e.g. photon)

d Reduction - systematic effectslike peak-finding and randomnumerical noise

d Solar noisec) Others

(i) Low frequency limit? or "LowFrequency" Paper

(ii) Low & Slow Pipe Results(iii)Diode as Σ low l Tuck Stebbins

33333333333333333333333333333333Radial Structure of the Sun

determined from GONG Data

IntroductionBrief history plus motherhood state-ments about physics, direct and in-direct inferences, neutrino flux, materi-al redistribution, globular clusters andcosmology.Acknowledgement that knowing thestructure is not very interesting exceptto the extent that it teaches us physics

11111111111111111111111111111111111111111111111111111111111111111

and astronomy.Modelling the solar interiorThe physics of what goes into themodel, and the justification or lack of.The concept of a reference model.Seismic inferencesRelation between structure and oscilla-tion data; what we might learn directlyand what must require additional as-sumptions before a statement can bemade.Inversion techniquesBrief summary of at least one tech-nique, and demonstration of itsefficacy using artificial data withGONG noise estimates and, of course,the GONG data state.Emphasize that frequency differencesno(??) causes 3-4 figures to be discard-ed.Sound-speed variationPresentation of the basic p-mode pri-mary inference from c2-ρ inversionsfor c2. Core speculations perhaps hint-ed at here. Acknowledgement ofstructural issues concerning the base ofthe convection zone.Equation of State ( in convectionzone )Testing EQS, helium abundance.The surface layerWhat little knowledge we can succeedin gleaning about the upper convectiveboundary layer, etc…Speculations, conclusions, and a lookinto the future

111111111111111111111111111111111111111111111

Perhaps we should await for someGONG data before we write this sec-tion. Douglas Gough33333333333333333333333333333333

Internal Rotation andAsphericity of the Sun

IntroductionBrief history. Importance of angularmomentum in astrophysics ( stellar );spindown, instabilities, etc… Thegreat suprise of 1984. So-calleddynamos ( i.e. numerical simulations. )Seismic diagnosticsIntegral constraints, inversion pro-cedures and test on artificial data. 2-D, problems with overcomingsidelobes. Poor a priori ’knowledge’,unlike(?) structure inversion. Em-phasize differences in accuracy ofresult of structures. In that respectthere is an analogy with structuralinversion at the current state of pro-gress. Artificial-data inversions todemonstrate degree of reliabilityPresentation of Ω(r,θ) - not a titled What we can say about it, and how

it relates to previous results.d Influence of annihilator on shapes of

contours, and why it looks sensitive.d Lack of apparent shear layer within

the convection zone that was previ-ously troubled by diurnal sidelobes.

d Shear layer at base of convectionzone, and its torque.

- 20 -33333333333333333333333333d Dynamics of deep interior, if possi-

ble. ( Probably not )Asphericityd Asphericity of surface layers and its

relation to meand Brightness temperature measure-

ments.d Shape of base of convection zone.d Putative magnetic fields, as an in-

direct inferenceDiscussionDynamics - relation between aspherici-ty and Ω variations, if possible ( un-likely with such a short data set )Statement of how things should im-prove with time. Temporal variations,in particular with regard to upperreflection layer. Implications relatingΩ contours and magnetic field lines, ifpossible, and relation ( if there is anyhint from the data ) between rotation ofcore and the helium-abundance varia-tion inferred from the structural varia-tions. Douglas Gough33333333333333333333333333333333Energetics of Solar Oscillations:

Amplitudes, Line Shapesand Phase Relation

d Energetics of Solar Oscillationsd Solar Oscillations - Amplitudes, line

shapes and phase relationd Time dependence of mode ampli-

tudesd Line widths as a function of l and νd Line asymmetrics as a function of l

and νd Relationship between physical vari-

able and GONG observablesd Activity effects on phase relationsRoger Ulrich33333333333333333333333333333333

Magnetic Effects

Large Scale Internal Magnetic Fieldd Toroidal magnetic fields inside the

Sun ( Goode )d Physics of the solar cycle ( Kosovi-

chev )d Global magnetic fields in the Sun’s

deep interior ( Thompson )

11111111111111111111111111111111111111111111111111111111111111111

Effect of Magnetic Fields on P-modes ( Forward Problem )d P-mode propagation in active re-

gions ( Brown )d P-mode interaction with magnetic

fields ( Penn )d Amplitude and phase changes in os-

cillation signals correlated withmagnetic fields ( Ulrich )

Local Area Helioseismology of Mag-netic Field Regions ( Inverse Prob-lem )d Active Region Seismology - Focus-

ing on higher-l and cross comparingwith Stokes polarimeter data ( Bog-dan )

d P-mode scattering from sunspotsand subsurface magnetic fields( Braun )

d First cut at time-distance seismology( Duvall ) Phil Goode

33333333333333333333333333333333Low Frequency

Solar Oscillation Modes

d Probing the Deep Interior of the Sunwith Low Frequency Modes

d Description of the GONG Low-Frequency Pipeline

d Merging Methods for Low-Frequency Data

d Upper Limits of Low Frequency Os-cillation Amplitudes

d Effects of Solar Noise Sources andInstrumental Errors

d Plans for Further Analysis and Pro-gress Todd Hoeksema33333333333333333333333333333333

Nearly SteadyMagnetic Fields and Flows

Rotation Profile (Doppler)d Are higher order terms needed to

represent it:A + B cos2 Lat + C cos4 Lat + ?

d What are the North-South asym-metrics?(Are odd powers of cos (Lat) need-ed?)

11111111111111111111111111111111111111111111111111111111111111111

d Can the torsional oscillation signalbe seen in a short (3-6 mo.) timeseries?

d Does the rotation vary on a short(3-6 mo.) time scale?

d Are the spatial or temporal timevariations associated with magneticstructure

Meridional Circulation Profile(Doppler)d What is the strength and direction of

the meridional flow?d Is there evidence for additional lati-

tudinal structure (sources or sinks atmid-latitudes)?

d Is there evidence for temporal varia-tions?

d Are the structure or temporalchanges associated with changes inthe rotation profile?

d What does this tell us about convec-tion zone dynamics?

Convection Limb Shiftd What is the structure of the quiet

Sun limb shift?d How does the limb shift change as a

function of magnetic field strength?d What does this tell us about the

solar granulation?Non-Axisymmetric Motions(Doppler)d What are the lifetimes of super-

granules and how do they evolve?d What is the spectrum of the photos-

pheric convective velocities out to l= 250?

d How strong are the radial flows insupergranules?

d How much vorticity is associatedwith supergranules?

d How does the ratio ofradial/horizontal flow speed varywith cell size?

d Are Giant Cells evident in theDoppler velocities?

d What is the structure of flows in ac-tive regions?

d What does this tell us about theMHD of active regions?

Correlation Tracking of Super-granulesd Do supergranules move with veloci-

ties? Different from the photospher-ic gases?

- 21 -33333333333333333333333333d Do the velocity differences vary sys-

tematically with cell size?d What does this tell us about the sub-

surface flows?d Are active region flows or giant

cells detectable with this technique?Dave Hathaway33333333333333333333333333333333

Theses

Congratulations to Said Loudagh forhis thesis entitled “Traitement du Sig-nal du Reseau IRIS de SismologieSolaire: Du signal brut a la rotationdu coeur solaire.” (Signal Processingof the IRIS Network for SolarSeismology: From the raw signal tothe rotation of the core) which he re-ceived from the Universite de Nice.33333333333333333333333333333333

Recent Preprints inHelio- and Astero-

seismology

The titles included here are preprintswhich have been sent, as a courtesy, tomembers of the GONG project staff.They are listed for the convenience ofNewsletter readers. Please contact theauthor(s) for additional information. Ifyou would like your preprint titles tobe included in future Newsletters, senda copy of your preprints to one of theGONG staff. If you would prefer thatyour preprint titles do not appear inthe GONG Newsletter, please indicatethat when you send any of us a copy.Otherwise, we shall share it!

Antia, H.M.: “Effects of Surface Layers onHelioseismic Inversions”

Antia, H.M.: “Nonasymptotic HelioseismicInversion: Iterated Seismic Solar Model”

Appourchaux, T.A., Toutain, T.,Telljohann, U., Jimenez, A., Rabello-Soares, M.C., Andersen, B.N., andJones, A.R.: “Frequencies and Splittingsof Low-Degree Solar p Modes: Resultsof the Luminosity Oscillations Imager”

11111111111111111111111111111111111111111111111111111111111111111

Audard, N., Kjeldsen, H., and Frandsen,S.: “Analysis of the AsteroseismologicalData of Six δ Scuti Stars in the OpenCluster NGC 6134”

Audard, N., Provost, J., and Christensen-Dalsgaard, J.: “Seismological Effects ofConvective-Core Overshooting in Starsof Intermediate Mass”

Bahcall, J.N., Krastev, P.I., and Leung,C.N.: “Solar Neutrinos and the Principleof Equivalence”

Banerjee, D., Hasan, S.S., andChristensen-Dalsgaard, J.: “TheInfluence of a Vertical Magnetic Fieldon Oscillations in an IsothermalStratified Atmosphere II”

Basu, S., and Antia, H.M.: “HeliumAbundance in the Solar Envelope”

Basu, S., Thompson, M.J., Christensen-Dalsgaard, J., and Hernandez, F.P.: “ASelf-Consistent Approach to Filteringout Near-Surface Uncertainties fromHelioseismic Inversions”

Bogdan, T.J., Hindman, B.W., Cally, P.S.,and Charbonneau, P.: “Absorption of p-Modes by Slender Magnetic Flux Tubesand p-Mode Lifetimes”

Bogdan, T.J., and Cally, P.S.: “Jacket-Modes: Solar Acoustic OscillationsConfined to Regions SurroundingSunspots and Plage”

Braun, D.C.: “Scattering of p-Modes bySunspots. I. Observations”

Chaboyer, B., Demarque, P., Guenther,D.B., and Pinsonneault, M.H.:“Rotation, Diffusion, and Overshoot inthe Sun: Effects on the OscillationFrequencies and the Neutrino Flux”

Chaboyer, B., Demarque, P., andPinsonneault, M.H.: “Stellar Modelswith Microscopic Diffusion andRotational Mixing I: Application tothe Sun”

Christensen-Dalsgaard, J., Schou, J.,Thompson, M.J., and Toomre, J.: “ThePotential For Resolving Jets andShearing Flows with the GONGHelioseismic Dataset”

Christensen-Dalsgaard, J., Monteiro,M.J.P.F.G., and Thompson, M.J.:“Helioseismic Estimation of ConvectiveOvershoot in the Sun”

Christensen-Dalsgaard, J., Larsen, R.M.,Schou, J., and Thompson, M.J.:“Optimally Localized Kernels for 2DHelioseismic Inversion”

11111111111111111111111111111111111111111111111111111111111111111

Christensen-Dalsgaard, J., and Thompson,M.J.: “SOLA Inversions for the RadialStructure of the Sun”

Christensen-Dalsgaard, J. et al.: “Near-Surface Effects in ModellingOscillations of η Boo”

Christensen-Dalsgaard, J., and Reiter, J.:“A Comparison of Precise Solar Modelswith Simplified Physics”

Christensen-Dalsgaard, J., and Petersen,J.O.: “Pulsation Models of the Double-Mode Cepheids in the Large MagellanicCloud”

Christensen-Dalsgaard, J., Bedding, T.R.,and Kjeldsen, H.: “Modelling Solar-Like Oscillations in η Boo”

Christensen-Dalsgaard, J.: “Effects ofOpacity on Stellar Oscillations”

Christensen-Dalsgaard, J., and Mullan,D.J.: “Accurate Frequencies ofPolytropic Models”

Christensen-Dalsgaard, J., Schou, J.,Thompson, M.J., Toomre, J.: “Huntingfor Azimuthal Jets and Shearing Flowsin the Solar Convection System”

Espagnet, O., Muller, R., Roudier, T.,Mein, P., Mein, N., and Malherbe, J.M.:“Spatial Relations Between the 5-minuteOscillations and Granulation Patterns”

Fan, Y., Braun, D.C., and Chou, D.Y.:“Scattering of p-Modes by Sunspots. II.Calculations of Phase Shifts from aPhenomenological Model”

Frandsen, S. and Viskum, M.:“Asteroseismology in Open Clusters”

Houdek, G., Balmforth, N.J., andChristensen-Dalsgaard, J.: “AmplitudeRatios and Phase Shifts in the SolarAtmosphere”

Houdek, G., Rogl, J., Balmforth, Neil J.,and Christensen-Dalsgaard, J.:“Excitation of Solarlike Oscillations inMain-Sequence Stars”

Monteiro, M.J.P.F.G., Christensen-Dalsgaard, J., and Thompson, M.J.:“Helioseismic Constraints on Theoriesof Convection”

Monteiro, M.J.P.F.G., Christensen-Dalsgaard, J., and Thompson, M.J.:“Seismic Properties of the Sun’sSuperadiabatic Layer I. TheoreticalModelling and Parameterization of theUncertainties”

- 22 -33333333333333333333333333The GONG Newsletter is intended to keep the community abreast ofnews and progress relating to the GONG project and other activitieswithin the field of helioseismology. The current mailing list for theNewsletter includes about 400 individuals who have expressed aninterest in these topics. We welcome contributions from anyonewishing to disseminate information of general interest to thiscommunity. Contributors to this issue include Wendy Erdwurm, PhilGoode, Douglas Gough, Jack Harvey, David Hathaway, ToddHoeksema, Rob Hubbard, Jim Kennedy, Sylvain Korzennik, Jim Pintar,Tuck Stebbins, Juri Toomre, and Roger Ulrich.

John LeibacherGONG Project Director, Editor

Mailing address: National Solar ObservatoryP.O. Box 26732Tucson, Arizona 85726-6732USA

Electronic Mail: Internet ≡ jleibacher@noao.eduNSI-DECnet ≡ 5355::jleibacher

Phone: +1 520 318-8305

Anyone on the GONG project staff can be reached via electronic mailby substituting their first initial and last name for jleibacher in the aboveelectronic mail addresses.

The Project is on the World Wide Web at the URL helios.tuc.noao.edu

222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111111111111111111111111111111

222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111111111111111111111111111111 2222222222222222222222222222222222222222222222222222222222222222222222222211

1111111111111111111111111111111111111111111111111111111

22222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111111111111111111111111111111111111111111

Musielak, Z.E., and Moore, R.L.: “Klein-Gordon Equation and the Local CriticalFrequency for Alfven WavesPropagating in an IsothermalAtmosphere”

Petersen, J.E., Christensen-Dalsgaard, J.:“Models of the Double-Mode Cepheidsin the Large Magellanic Cloud”

Rogers, F.J., Swenson, F.J. and Iglesias,C.A.: “OPAL Equation of State Tablesfor Astrophysical Applications”

Rosenthal, C.S., and Christensen-Dalsgaard, J.: “The Interfacial f Mode ina Spherical Solar Model”

Rosenthal, C.S., Christensen-Dalsgaard, J.,Nordlund,

oA., and Trampedach, R.:

“Convective Perturbations to SolarOscillations: The f Mode”

Rosenthal, C.S., Christensen-Dalsgaard, J.,Houdek, G., Monteiro, M.J.P.F.G.,Nordlund,

oA., and Trampedach, R.:

“Seismology of the Solar SurfaceRegions”

Rouse, C.A.: “Calculation of StellarStructure. IV. Results Using a DetailedEnergy Generation Subroutine”

Schou, J., Christensen-Dalsgaard, J., andThompson, M.J.: “Some Aspects ofHelioseismic Time-Series Analysis”

Tomczyk, S., Schou, J., and Thompson,M.J.: “Measurement of the Rotation inthe Deep Solar Interior”

Tripathy, S.C., and Christensen-Dalsgaard,J.: “The Effect of Opacity Modificationson Solar Structure and Oscillations”

3333333333333333333333333333333333333333333333333333333333333333

11111111111111111111111111111111111111111111111111111111111111111