Introduction Of Meerkat In South Africa
Radio astronomy is the world's next big scientific advancement. Radio telescopes, unlike optical telescopes, are not hampered by poor weather conditions or light pollution and can observe 24 hours a day. Radio astronomy allows us to view the invisible universe, and thus investigate things like galactic structure, pulsars and quasars, and the early universe. These topics make up a huge portion of science's currently unanswered questions, specifically around cosmology and Einstein's theory of gravity. It is for these reasons, radio astronomy is being cultivated across Africa and the world, not only on an industry level but in the furtherance of academic programs at both the secondary and tertiary level.
The University of Cape Town is one such institution with a leading Astronomy department and a desire to nurture a love of galactic and extra-galactic radio astronomy among its students. In this vein, a visit to the Square Kilometre Array South Africa (SKA SA) once in Pinelands was arranged, to expose the bunch of wide-eyed, star-struck wannabe astrophysicists in the third year class to the behind the scenes of a real scientific project.
The visit consisted of a meeting with SKA SA's Technical Lead for Scientific Computing, Simon Radcliffe, who presented to us the "politician lecture" - the speech he gives to government department members who allocate funding - interspersed with some in-depth information about the computing abilities and power of the SKA system. This was followed by a tour of the SKA SA home base, including the control room, an engineering store room where the 3-D printer and mechanical once prototypes are held, and a data storage room where the information collected from SKA is physically stored.
Simon's talk detailed the specifics of the SKA project and MeerKat's place in it. The SKA is a cross-continental radio interferometer, which will be located in South Africa and Australia. It will comprise thousands of dishes and up to a million low frequency antennae. According to SKA, it will have the capability to "monitor the sky in unprecedented detail and survey the entire sky much faster than any system currently in existence. Its unique configuration will give the SKA unrivalled scope in observations, largely exceeding the image resolution quality of the Hubble Space Telescope. It will also have the ability to image huge areas of sky in parallel, a feat which no survey telescope has ever achieved on this scale with this level of sensitivity." SKA SA will be incorporated into SKA1 mid-frequency range. It will consist of 133 antennae, divided into three spiral arms with a maximum baseline of 150km. Construction is expected to start in late 2019 and be completed and operational around 2024. It is expected to surpass the image resolution of the HubbleSpace Telescope by a factor of 50 times, while also being capable of the parallel imaging of huge areas.
MeerKat is the precursor of SKA SA and will be the infrastructural and technological base around which SKA SA will be built. It began as the Kat-7, seven antennae with a max baseline of 185m situated around90km outside of Carnarvon in the Northern Cape. Kat-7 was intended as a precursor of MeerKat, to prove South Africa's capacity to host the SKA. It was later extended and renamed MeerKat, now consisting of 64antennae with 2000 unique antenna pairs and a max baseline of 8km. This is considerably more than any comparable telescope and allows for incredible high-fidelity images of the radio sky. This max baseline is expected to be extended to around 20km with the construction of seven additional antennae. It is currently operational, having been completed in May 2018 and inaugurated in July. It has been doing science since around April and has already been involved in the creation of the clearest image yet of the supermassive black hole Sagittarius A* at the center of our galaxy.1Since MeerKat is now online and is primed to be the precursor for the SKA, five years of observing time have been allocated to leading astronomers for research purposes, with around half of the time going to South African led projects. These include LADUMA (Looking at the Distant Universe with the MeerKat Array), led by UCT's Dr. Sarah Blyth, which will be an ultra-deep survey of neutral hydrogen gas in the early universe.
There are also projects led by international astronomers investigating topics like Gravity and Gravitational Radiation through pulsar observations, the role of molecular hydrogen in the early universe through CO observations, the fundamental cosmological constants of the early universe through absorption line surveys, galaxy cluster formation and evolution, and the hunt for new pulsars. The engineering project that was the construction of MeerKat was colossal. Each antennae is 19m tall and consists of a main reector, sub reector, electronic equipment including receivers and digitizers, 9 tons of steel and 120m3 of concrete. The radio signal is captured by the reectors (set in an onset Gregorian configuration) and directed to the receiver which converts it to a voltage signal. This is then amplified by cryogenic receivers which are cooled to -203o, more than three times colder than the lowest temperatures recorded in Antarctica, to add as little noise as possible. Four digitizers then convert this voltage signal to a digital signal using an Analogue to Digital Converter (ADC). This signal is sent via buried breoptic cables to the on-site correlator housed at the Losberg site in the Karoo Array Processor Building (KAPB).This process is the computation of the cross-correlation function, and involves a staggering 170km of cable situated 1km underground to preserve thermal stability. An amount of digital processing also takes place at KAPB including correlation of the signals from all the receivers. The resulting data is archived at KAPB and sent on-site again via fibre cables to the Centre of High Performance Computing (CHPC) in Cape Town.
Being involved heavily in the computational side of SKA, Simon's talk also focused on the computing aspects - the challenges and solutions of transferring, combining, reducing, storing and analyzing colossal amounts of data at lightning speed. Each antenna generates more than 4Gb of data every second. This obviously requires incredibly fast data transport and computing algorithms. In order to achieve this, SKA SA has the technological infrastructure which, when it all comes online, will have 20 petaops of computational speed.
When SKA is completed, it is expected to have more than 100 petaops of raw processing power. To put thatinto perspective, the world's fastest super computer as of June 2018 has 200 petaops. In comparison, MeerKat's specs are staggering. This computing power is achieved through partnerships with the CHPC, the Institute for Data Intensive Astronomy (IDIA) and tech companies like IBM and South African-based Peralex. These allow for the building of shared platforms where researchers can process and analyze data, as well as explore future technological solutions like machine learning and cloud computing.
One of the significant challenges of MeerKat, and indeed the nal SKA project, is data storage. The sheer size of the data is the rst issue, followed by the worry that digital drives can be corrupted and data lost.
The solution to this is, astonishingly, tape. Instead of writing onto a disk, the data is stored on tapes. These tapes are housed in a tape library with around 3000 tapes and a retrieval system. These tape libraries retail for around R10-million all told, leading to the second issue with data - cost. Digital data storage like hard-drives are incredibly expensive - tape is considerably cheaper. To further minimize costs, the team at SKA SA found that they could build their own tape library and retrieval soft- and hardware, for around60% of the price, reducing their costs massively. Tape also makes data transport considerably easier. Instead of sending hundreds of gigabytes of data across the world via cables or the internet, tape is capable of being packed into a protective container and physically shipped across continents with little risk of data corruption or loss. According to Thomas Bennett, a member of SKA SA's Science Data Processing team who showed us around their data storage room, the ultimate goal would be to pack a shipping container full of tapes and be able to ship it to the UK. Simon was not too certain of this idea.
All told, MeerKat is expected to bring significant value to South Africa. Not only will it become part of the world's best radio interferometer (already being a world-class telescope in its own right), and allow important South African science to be at the forefront, it is also stimulating the South African economy. MeerKat is largely funded by the South African government, with assistance from the EU and the global SKA. While it does have construction partnerships with international companies in the USA and Germany, it was a requirement that 75% of the funded contract value must be spent in South Africa. As a result, most MeerKat components were manufactured in SA, boosting both our industrial and tech economy. There is also a larger focus on Africa's role in the global furtherance of the SKA and radio astronomy. Eight other African countries are already partnered with SKA (Botswana, Ghana, Kenya, Madagascar, Mauritius, Mozambique, Namibia and Zambia) with the goal of extending the SKA and establishing radio astronomy programs across the continent. This will bring Africa to the forefront of science and technology going into the fourth Industrial Revolution.
The visit to the SKA once was informative, eye-opening and encouraging. By far the most interesting thing to learn was that SKA SA has a host of mechanical engineers who are designing, 3-D printing and building their own infrastructure and tailoring it to the needs and specifications of the project. SKA SA really is a mammoth task, but it seems in such capable hands. It is so exciting to see real science in progress and South Africans at the helm.