The Global Oscillation Network Group (GONG) Project is a community-based activity to operate a six-site helioseismic observing network, to do the basic data reduction, provide data and software tools to the community, and to coordinate analysis of the rich data set that is resulting. GONG data is available to any qualilfied invesitgator whose proposal has been accepted, however, active membership in a GONG Scientific Team ecourages early access to the data and the collaborative scientific analysis that the teams are undertaking. Information on the status of the project, the scientific investigations, as well as access to the data, is available on our WWW server whose URL is www.gong.noao.edu.
During the fourth quarter of 1997, the GONG network continued to perform in an outstanding manner. As expected, there was some network down time mostly due to weather as the northern-hemisphere instruments moved into winter. Preventive maintenance (PM) trips occurred on schedule and without fanfare, with visits to Tenerife (October), Udaipur (November), and Cerro Tololo (December).
During October, Cerro Tololo suffered an earthquake of magnitude 6.8, and although the GONG shelter moved 12 mm, no apparent damage occurred to the instrument. The PM team was able to reposition the shelter back to its original location during their December visit. In addition, the team replaced the turret, which required a complete realignment of the optics and additionally ensured that any possible problems introduced by the tremor were rectified. The team also installed new earthquake protection gear, which will hold the electronics rack and UPS in place in the event of another quake.
Mirror fogging, which occurs occasionally at the Big Bear and Tenerife sites, appears to be a function of the drying agent in the air dryer and may be solved by replacing it more frequently. Data-read problems were fixed at Learmonth by swapping drives and the main circuit breaker was replaced at Udaipur. The on-site staff continues to head off most of the Exabyte failures with timely intervention providing normal, continued operation. All in all, we are very pleased with the network duty cycle.
Jay LeBlanc joined the Operations group as an Electronics Technician supporting the GONG facilities. Jay will share the tasks of Duty Responder and Site Checker as well as participate in PM trips. The Project has said good-bye to Dan Bass, the GONG Operations Engineer. Dan joined the Project in 1993 and has been an integral part of the development, deployment, and operations teams. Dan leaves us to join Wyko, a Tucson-based optics firm--we wish him the best.
During the past quarter, month-long (36-day) velocity, time series and power spectra were produced for GONG months 22 and 23 (ending 970811) with respective fill factors of 0.84 and 0.86, and for GONG year 2 (960817 through 970811) with a fill factor of 0.86. The p-mode reprocessing campaign added GONG months 12 and 13. The project is also producing time series and power spectra from the intensity images. These products were generated for GONG month 17.
The Field Tape Reader (the subsystem that receives the raw data cartridges from the observing sites) processed 84 cartridges containing 559 site-days from the seven instruments. 368 site-days were calibrated.
During the past quarter, the Data Storage and Distribution System (DSDS) serviced 10 data distribution requests for 262 files totaling 0.3 Gigabytes of data. Each request was filled within two working days of receiving the request. The DSDS performed 1,545 data cartridge transactions (library check-ins and check-outs) in response to requests from the data reduction pipeline and other internal operations.
The Data Management and Analysis Center staff has found a simple solution for removing much of the noise that is mapped into the velocity images during the calibration process (Figure). The noise originates from dirt and flaws on the optical surfaces of the light-feed assembly. Because these cosmetic blemishes change only slightly from day to day and are typically removed each week when cleaned, the noise can be greatly reduced by averaging. Tests using a 60-day Gaussian weighting function effectively suppress the noise. The artifacts from the interferometer are nearly constant and are preserved during the averaging. It appears that this will be an effective solution for removing the diagonal artifact that is sometimes visible on l-v spectra which originate at the origin and reach the first p-mode ridge at l = 200.
There has been substantial activity in peak finding algorithm development. Rachel Howe joined the project on 1 November. She has implemented a multi-dimensional fitting procedure for low-l modes that treats the leakage matrix explicitly, particularly in the m direction. Her results show a marked improvement in the estimates of low-l rotational splitting from GONG data, along with a more reasonable picture of deep rotation in the solar core.
Stuart Jefferies has been working with several collaborators to develop a physically meaningful model of the asymmetric line shape of the oscillations. They are finding that the solar background mechanism plays an important role in determining the line shape. In particular, the phase relationship between the background and the oscillation signal can flip the sense of the asymmetry, reinforcing the suggestion that a simultaneous fitting of the spectrum observed in velocity and intensity will be required to estimate the mode frequencies accurately.
Stuart is also working with Ed Anderson to install a "simplified" asymmetric line profile model in the GONG fitting. This should be available soon. Finally, along with Cliff Toner and Nadege Meunier (Stanford), Stuart is rewriting a multi-dimensional fitting code for intermediate degree m-averaged spectra. This new code is being written in Fortran90, which will make it portable to a parallel processing super computer.
Ed Anderson has fit the GONG year 2 spectrum up to l = 45 using the standard peak fitter, and is currently fitting two 3-month long time series (months 12-14 and 21-23) up to l = 150 for use by the Inversions Team. The asymmetric profile model will also be used to fit these data.
Rudi Komm continues to develop the multi-taper method. He has now implemented code originally developed by Imola Fodor and Philip Stark (Berkeley) to compute renormalized sine tapers for gapped data. The multi-taper routine will be applied to one of the three-month long GONG time series mentioned above and then rerun through the peak-fitting algorithms.
Frank Hill has been examining the effect of slight variations on the initial guesses used in the peak bagging. It turns out that a few simple revisions to the post-processing examination of the fitting results go a long way towards removing modes that have problems. A combination of H. Antia's (TIFR) strict numerical convergence criterion for the BFGS minimization, the current GONG "bad mode" flag, a reduction of the allowable fitted width to twice the guess width, and a test for locking on to the input guess, does a good job in weeding out modes that are sensitive to the first guess. There is, however, a small remaining set of modes that eludes this armada of tests. These renegade modes are in the spectral region (high frequency, high degree) where spatial leaks are unresolved and have blended into ridges. The fitting model assumes resolved leaks and is inappropriate for this situation, leading to multiple minima in the likelihood function and typically two or three distinct solutions widely separated in frequency. Work is underway to detect this situation and flag the suspicious modes.
Frank has also been working on the full leakage matrix calculation. He has found a method that is well suited to the problem of numerically integrating spherical harmonics. Instead of using an evenly-spaced grid, which requires many tens of thousands of points to converge for high degree (l = 200) near-zonal modes, a stretched grid can be used that has only slightly more points than the highest value of l being computed. The method used to determine the grid is very new (only about 200 years old): the well-known Gauss-Legendre quadrature formalism! This discovery provides further evidence that "two years of research can save you a whole week in the library!" Armed with this advanced numerical technology, the leakage matrix calculation shall resume shortly.
The GONG Instrument Group has been studying retrofitting a higher-resolution (1024 × 1024), square-pixel camera to the existing observing stations. Increasing the detector scale will provide significantly improved helioseismic resolution in the near-surface regions of the Sun that are the home of the intense magnetic fields that seem to cause much of the more dramatic aspects of solar activity, extend all aspects of "local helioseismology" dramatically, as well as enabling many non-helioseismic, diachronic solar measurements. The Project is continuing this new-camera upgrade effort with the development of a proof-of-concept breadboard instrument during FY 1998.
After reviewing cameras from many manufacturers, it was determined that the Silicon Mountain Design camera (SMD 1M60_20) best satisfied the requirements for our application with regard to quantum efficiency, frame rate, spectral response, linearity, susceptibility to interference fringing, etc. The prototype camera is expected to arrive in mid-February, the high-speed electronics are under development, and we look forward to images this spring.