Previous Article Next Article Table of Contents
Global Oscillation Network Group (1Mar94) (from GONG, NOAO Newsletter No. 37, 1 March 1994) Sites Eight site survey instruments continue to operate at the six GONG sites, plus downtown Tucson and Sacramento Peak. Another instrument will be installed at Kitt Peak in the near future. The network continues to see the Sun about 93.5% of the time. The two major papers reporting the site survey methods and results have been submitted to Solar Physics. A paper reporting on observations of the dust generated by the Mt. Pinatubo eruption has been drafted, and will be submitted in the near future. Work continues to progress on the preparation of the field sites to receive the GONG instruments. At the Big Bear Solar Observatory (BBSO), a small land fill adjacent to the causeway out into the lake will be required to seat the station. Members of the GONG and BBSO staff met with the Army Corps of Engineers in January to discuss the applications for the various necessary permits. In all, four different permits are required, including the Corps of Engineers, the State water quality and wildlife agencies, and the local water district. In a parallel activity, requests for bids for engineering services to design the landfill have been issued. Work will probably begin this summer. A final location for the site at NOAA's Mauna Loa Observatory has been selected and agreed to by all parties. The site use agreement has been forwarded to NOAA for signature. NOAA intends to have other construction underway on the site in the May to July time frame and GONG expects to coordinate its preparation work with those activities. The High Altitude Observatory, which operates a solar observatory at the same site, will be providing on-site support for the GONG observation program. The GONG Project Manager, Jim Kennedy, will be traveling to the sites in the Canaries, India, and Australia, in February for detailed discussions with the engineering and technical staffs and potential contractors at those locations regarding the land work there. It is expected that administrative and technical details will be worked out that will clear the way for work to begin at those sites in the next few months. Similarly, there are ongoing discussions with CTIO in Chile to those same ends. Meetings with CTIO are tentatively planned for late March or early April. Everyone is anxious to see which of the sites will be the first to break ground. Instrument At the end of December, the GONG project received three interferometers from the manufacturer, Interoptics. Two of the units were immediately tested in our optics lab. The tests were promising so the least perfect unit was installed in the GONG prototype instrument. Solar observations confirmed good performance. The second unit was also installed and performed well. Tests of the third unit indicated that it would also function properly. As a result, the project accepted the three interferometers. Two more are nearly finished and all ten units should be delivered by the end of February. This will complete a long quest for the principal optical component of the GONG instrument. The wavelength transmitted by the interferometers was found to vary less than 1/20 wave across their apertures. This is equivalent to less than 600 m/s in measured velocity. Temperature sensitivity proved to be almost exactly the same as calculated or 550 m/s per degree C. The interferometers are mounted inside ovens which maintain temperature constant to 50 micro degrees. Therefore the basic GONG signal will be stable at the 3 cm/s level before decomposition into spherical harmonics. Earlier realizations of the interferometer produced disturbing ghosts from multi-order quarter-wave plates in the arms of the interferometer. Thanks to the skill of the manufacturer, it was possible to eliminate these ghosts by fabricating very thin quartz wave plates on the mirror surfaces of the interferometer. The major disappointment is associated with the polarizing beamsplitter coating. It introduces a phase shift with angle of incidence that diminishes the angular field of view in one dimension. The effect of this is to introduce a velocity gradient across the image of the Sun in the GONG instrument. It is typically about 1000 m/s across the solar diameter or about 1/4 of the solar rotation gradient. In practice it should not be a problem. We now have interferometers for Big Bear, Mauna Loa, and Tenerife in hand, with the balance due us by the end of February. We are presently awaiting the delivery of entrance windows and the remaining narrow-band interference filters to complete the complement of critical optical elements. [Figure not included] This picture shows two of the production GONG Michelson interferometers manufactured by Interoptics. The coin is 25 mm in diameter. Progress was also made on adding an integrated image velocity signal to the output of the GONG instruments. A port originally designed to use a laser reference signal can now be used to provide a velocity signal integrated across the solar image. Tests to date indicate excellent performance. GONG's three instrument makers, supplemented part time by two members of NOAO's instrument shop, are continuing to work overtime to complete mechanical parts in time for field station first light this spring. About 85% of the 450 mechanical parts that make up the GONG instruments are now complete. Where possible, sets of drawings are being prepared for bid so that the less critical items can be fabricated outside of our shops. At the present time most of our effort in house is going into fabrication of the camera rotators - the last major mechanical element of the instruments and into component assembly. These latter items now appear to be on the critical path to field instrument integration. Excellent progress is also being made with electronics production. We have completed enough production chassis to populate all six field instruments. The main electronics racks are also complete, including intra-rack cabling and harnessing. By mid to late February these racks will be transported to the integration site so that external cable manufacturing can begin. Bringing up the rear for electronics is circuit board production. Only a few relatively simple boards still remain in the printed circuit routing queue. More than 100 populated and wave-soldered data boards have already arrived and are undergoing visual inspection prior to beginning burn-in. Data Management and Analysis The main thrust of algorithm development for the GONG DMAC Pipeline has been along the lines of temporal filtering and merging. The project has adopted a baseline temporal filter of a weighted 21-point running mean that is re-normalized over gappy data, and independently applied to individual site days. This filter is quite effective in removing the leakage of strong low-frequency power from solar velocity fields (primarily rotation and supergranulation) and instrumental residuals into the p-mode band. However, it has two drawbacks: daily data strings with a low duty cycle must be discarded, degrading the overall performance of the network; and the filter results in a complicated - and so far unknown - response in the merged power spectrum. The project has tested an alternative "deterministic detrending" (DD) method in which the known velocity fields of observer motion, solar differential rotation, and limb shift are modeled and subtracted. This method does not discard any data, and has no power spectrum signature, but cannot easily remove supergranulation or instrumental residuals. The tests showed that the DD method worked reasonably well on days free from clouds, but poorly otherwise. Due to a lack of resources, the Project cannot address either improving the DD method, or the power spectrum signature of the baseline filter. Instead, we have decided to retain the baseline filter, and proceed with using it in the course of applying our baseline merging algorithm to a 15-day time series of artificial data which is now available. This will allows us to both develop the production code, and to gain practical experience for the task of merging the GONG data. No attempt will be made to remove the filter signature from the merged spectrum. A scientific visitor, Shushant Tripathy (Udaipur Solar Obs.), has recently arrived and will be investigating ways to improve the DD method. In the context of the merging algorithm, the scheme of correcting the images with their derived MTF as developed by Cliff Toner has been shown to be very effective in reducing atmospheric effects in the spectrum. A dramatic demonstration was provided by the ability of the method to remove the strong periodic features inserted into the spectrum by the hourly magnetograms which have a different focus from the routine observations. Indeed, the NSO South Pole group has found that the method can restore intensity images obtained through cirrus, allowing the use of substantial numbers of images previously though to be useless. The GONG DMAC Users' Committee (DUC) held its sixth meeting in Tucson on 7 January. Topics discussed included peak finding, 16 vs 32-bit storage of the time series, DSDS developments, and a processing history log. Suzanne Forgach left the project on 31 December. Suzanne had been working with the Site Survey data and calibrating data from the GONG prototype instrument. GONG's anonymous ftp disk storage area has moved from the VAX cluster to a DECstation in the GONG/DMAC building. The network address has changed from robur.tuc.noao.edu (140.252.1.10) to helios.tuc.noao.edu (140.252.8.105). helios is the DSDS user machine. The disks on robur are no longer readable by an anonymous user. A significant system upgrade of the Data Storage and Distribution System (DSDS) is underway. When completed in June, the DSDS will be on two Sun SPARC10's with SOLARIS 2.3 using ORACLE's data base management system and with a new design for the file catalog. Despite various problems which were encountered with ORACLE's DBMS and with SOLARIS, the conversion which began during the fall, is on schedule. Initial tests of data base intensive DSDS operator functions indicate that the new file catalog design will result in an order of magnitude improvement in performance. This alleviates concerns about the capacity of the DSDS during the network phase of the project. During the last quarter, the DMAC calibrated, produced site-day l-nu spectra and 4-minute averages for 17 prototype data days: 9, 23-24, 29-31 October; 5-7, 25-27 November; and 10, 12, 16-18 December. The prototype instrument recorded data from 23 December through 4 January which the DMAC will reduce in the near future. Project Management The project has been advised that it will not receive the $2.85M FY 1994 funding requested to begin the planned network deployment in July. It appears that, once again, the NSF will be unable to provide GONG with the budget called for in the NOAO Program Plan. Now that deployment is imminent, this recurring problem may have some particularly unsavory impacts on the project. This will be even more acute if next year's budget continues this pattern. In these stringent times, it is most important to accomplish the GONG research program with the least overall cost. Delays in beginning the scientific observations are very expensive. The project's overhead and personnel costs continue marching on, whether the science is being done or not. The project must have all of the sites in its six-station network operating to meet its fundamental scientific requirement for virtually continuous data. Economy dictates that all of the stations be deployed in as brief a time as possible (six months). Moreover, there is a clear scientific requirement for several years of observations, in order to make the frequency measurements with the required precision, and to observe cyclic and secular changes. Combined with the gigabyte/day data acquisition rate, and the need to verify proper operation of the six instruments, this leads to the need to be able to process data at the observing cadence. Anything less will lead to an unmanageable backlog of data that would seriously threaten the integrity of the program. Unfortunately, the minimalist approach to funding the project over the past four years has forced repeated deferral of the purchase of the main production processing systems for the DMAC. It appears that this will occur again this year, leaving the project's data reduction capacity in serious doubt in 1995. Since 1990, budget projections have shown that this chronic underfunding approach will lead to the need for a one- or two-year funding bump to secure the DMAC systems and cover the necessary one-year overlap in full staffing for both the instrument and DMAC groups during deployment. If the FY 1995 budget is at current levels, not only will the DMAC be compromised, but the network deployment may have to be stretched out over a period of 18 months or more. Such a delay would increase the overall cost of the project by more than $1.5M. As of this writing, the final amount of the available FY 1994 funding is still unclear. Nevertheless, steady - if painfully slow - progress is being made on all fronts. Depending on the actual budget figure, it is still possible that deployment could begin as early as November 1994. Other deployment options are also under consideration and will be reviewed by the project and the GONG Scientific Advisory Committee. John Leibacher and the GONG Team
Previous Article Next Article Table of Contents