South Africa develops a transient-tracking “intelligent observatory”

On 18 August 2017, astronomer Petri Vaisanen was in the right place at the right time—near a giant telescope. The previous day the LIGO and Virgo observatories had detected gravitational waves, and astronomers around the world were desperately seeking access to telescopes so they could observe the remnants of the neutron-star collision that had sent the ripples through time and space (see Physics Today , December 2017, page 19 ).

That night, Vaisanen happened to be the observer on the Southern African Large Telescope (SALT), an 11 m optical telescope near Sutherland, South Africa. Using the coordinates forwarded to him by South African Astronomical Observatory (SAAO) colleagues in the LIGO and Virgo collaboration, he was able to produce one of the first optical spectra clearly showing the residual fireball that resulted from the violent collision.

Vaisanen’s presence in the observing room that night was pure chance, he said at a conference celebrating SAAO’s 200-year anniversary in October. A new initiative at the SAAO aims to take some of the luck out of observing future brief but momentous events. Astronomers and engineers there are developing an “intelligent observatory,” in which networked telescopes will receive and filter discovery alerts from facilities around the world and then automatically point to astronomical objects of interest. To achieve that capability, SAAO is retrofitting its stable of reliable, but also relatively old, optical telescopes.

Transients, such as the neutron-star collision, are a growing area of astrophysics study. Other targets include fast radio bursts (see Physics Today , January 2021, page 15 ), supernovae, and gamma-ray bursts. Although astronomers are working on ways to more rapidly disseminate the details of new transient discoveries, the current process often involves time-consuming chains of phone calls and emails. An automated process offers the advantages of identifying targets and commencing observations with minimal delay.

The SAAO Sutherland site is home to more than two dozen optical telescopes of different sizes. Some of them are fully owned by South Africa; others are hosted in exchange for observing time and data. The goal is to have all the telescopes incorporated into a single network that is controlled by a centralized algorithm. Upon receiving alerts from partner institutions, the intelligent observatory would prioritize particular astronomical targets, determine the best telescopes and instruments for the observation, and then automatically insert the required observation duration and location into the telescopes’ observing queues. “Having that capability to access a suite of instruments at the same time, that’s powerful,” says SALT observatory scientist Lisa Crause.

The project’s engineers have begun modifying three of the locally owned telescopes to allow remote automatic observing. The team has already adapted the SALT telescope, which is the largest optical telescope in the Southern Hemisphere, so that astronomers can both observe and choose instruments remotely.

Automating SAAO’s older instruments is a difficult undertaking. The observatory’s 1.9 m telescope, for example, is about 80 years old. For now, its three observing instruments—for imaging, acquiring spectra, and measuring polarization—have to be swapped in and out manually by technicians.

The formidable feat of retrofitting optical telescopes is being supported by the country’s radio telescope engineers. Having designed and built South Africa’s 64-dish MeerKAT radio telescope, they are waiting for construction of the Square Kilometre Array to start within the next two years. The SKA will be the largest radio observatory in the world, with hundreds of dishes on the African continent and more than a million antennas in Australia . “We are providing resources and direct funding primarily to ensure gainful employment for all our engineers . . . and to assist the optical astronomy discipline,” says Willem Esterhuyse, head of engineering at the South African Radio Astronomy Observatory.

On the software side, SAAO is building on the expertise of a collaborator, Las Cumbres Observatory . LCO tracks transient events with a global network of telescopes, including some at the Sutherland site, that are equipped with uniform instruments. Las Cumbres pioneered open-source software that allows its network to receive alerts from researchers or large survey projects and automatically schedule observations on one or more of the telescopes. The software takes into consideration what needs to be observed and when, and if needed it can bump scheduled observations out of the queue.

What sets SAAO’s intelligent observatory apart from LCO and other projects is its automated access to telescopes of different sizes and a wide variety of component instruments at a single site. “Covering this large scale, from a 1.9-meter up to 11-meter telescope, this is definitely new,” says Mirko Krumpe, an astronomer at the Leibniz Institute for Astrophysics Potsdam who works on data from the space-based eROSITA telescope and collaborates with SAAO.

“What they are attempting to do is tremendous,” LCO project scientist Rachel Street says of SAAO’s aspirations. The intelligent observatory will use an adapted version of LCO’s Target and Observation Manager software, which, among other things, allows astronomers to easily interact with and display their observation data. “SAAO wants to say, ‘Tell us what we need to observe [for you] and we will send you reduced data,’” says Street. “They will deliver that observation across multiple instruments and wavelengths. It cuts down the path to science.”

Unlike purely robotic observatories, SAAO’s telescopes, such as SALT, also have time allocated for observing astronomers. When one telescope is in use, the intelligent observatory will divert incoming observation alerts to other available telescopes on the Sutherland site. Which alerts get preference is still up for discussion, says Vaisanen, who is now the observatory’s director. “That’s why it is important to write down the rules or flowcharts beforehand; otherwise you’ll get into almighty fights.”

The team plans to have half a dozen telescopes upgraded and ready for intelligent observing by late next year, when the NSF Vera C. Rubin Observatory in Chile begins its Legacy Survey of Space and Time (LSST). The observatory’s 8.4 m telescope will scan large tracts of the sky, acting as a discovery engine for transients and other phenomena. The survey will generate some 10 million alerts a night, and it will be the role of other observatories, such as SAAO, to follow up on them as quickly as possible.

Federica Bianco, the project’s science collaborations coordinator, says algorithms and intelligent telescopes, such as those SAAO is planning, are necessary to keep up with the deluge of data coming out of large surveys such as the LSST. “You can’t take a month to decide what is worth following,” she says. “You need automation in the selection of what to observe, and automation in pointing to it, collecting the data optimally. You don’t have enough time for individual human decisions.”