Site survey in Gansu
Meeting with local officials, Gansu
Site survey in San Juan, Argentina

Q: Is it really possible to do neutrino astronomy? As neutrinos don’t have a horizon, won’t we just see background neutrinos with no source identification?

A: We demonstrated in Fang et al. (2016) that one can spot clusters of neutrinos and identify sources significantly (> 4 sigma level) with an experiment that can detect hundreds of events with an angular resolution of a fraction of degree. This is precisely the aim of GRAND, which will have the characteristics to really do neutrino astronomy.

Q: What if other experiments already detected (quite a few) UHE neutrinos by the time GRAND was built?

A: In this case, GRAND will be a guaranteed working horse for UHE neutrino astronomy, UHE gamma astronomy, source evolution and particle physics at the highest energies. Note that GRAND10k’s timeline and sensitivity target are already comparable to ARA/ARIANNA. Thus if all goes as planned, GRAND10k should also detect neutrinos if other experiments do at that time. The full GRAND array would then  provide yet unknown discoveries with a sensitivity that is an order of magnitude better than what will be out there. Compare to gravitational waves: Even though first results are in, people are extremely anxious to build more sensitive detectors!

Q: Will GRAND really be able to clean the radio background?  Can the background be cleaned without throwing away too much of the signal?

A: It is not possible to give a definitive answer to this at this time, but this is precisely why we build GP300. However, we may point that air showers have very distinct features (beamed amplitude pattern at ground+Cerenkov cone enhancement, polarization to start with) while a vast majority of the background will cluster in position and time, as previous data (ANITA, TREND, AERA, CODALEMA) has shown already. So we are optimistic we will succeed. Outside Antarctica, TREND already obtained very promising results, discriminating signal from background with a very basic (single polar) and small (1 km²) setup.

Q: What is the range of frequencies observed? Why is GRAND targeting this particular band?

A: GRAND will observe the frequency range between 50 and 200 MHz. Because we would like to detect radio Cherenkov rings – which could help background rejection and signal reconstruction – we set the upper limit of the frequency range to 200 MHz, instead of the 80 MHz or 100 MHz used in existing radio arrays. This is aided by the Galactic radio background dropping significantly above 100~MHz. Going to higher frequencies also allows us to decrease the size of the antenna and the pole holding it. That makes the transport and installation easier.

Q: Why would we want to use GRAND to observe FRBs compared to other dedicated radioastronomy experiments?

A: With GRAND, we get a survey of the whole sky, a complete time and sky coverage. We don’t have the angular resolution, but we get the dispersion measure, spectra, etc. Maybe even with several hotspots, we can have some idea of the direction. This is a totally new way of doing things. We could also find pulsars and other radio transients in spite of the absence of directionality.

Q: How do you plan to do radio-astronomy with GRAND?

A: We are planning to search for radio transient sources. This can be only done with an experiment surveying a large portion of the sky, as will GRAND. The present proposal is to perform an incoherent detection of transient radio sources, such as FRBs, by summing (in the frequency domain) the unphased signals of all antennas in each GRAND 10k hotspot. Incoherent summing implies a signal-to-noise ratio improving in sqrt(number of antennas), yielding detection potential for FRBs of Lorimer-type. FRBs would be detected through the typical dispersion f(t) signal. The detection of a same FRB with different hotspots would additionally allow for triangulation and direction reconstruction.

Q: What is the estimated construction cost of each phase of GRAND?

A: We estimate that each station will cost 3500 euros in the early stage (GRANDProto300) and between 1500 and 2000 euros for the first 10k hotspot. Adding 30% cost for the DAQ and the infrastructure, GRANDProto300 and GRAND10k each amount to 1.3 M and 20 M respectively, excluding the price for renting the land and manpower for deployment and maintenance. This cost per antenna is expected to be reduced to approximately 500 euros per station for GRAND200k, by going to industrial production. The total cost of GRAND200k is estimated to be 200 million euros, again excluding the price for renting the land and manpower.

Q: How will GRAND build 10,000 antennas?  How will they be deployed? How will they be maintained?

A: We will not build 10k antennas, industry will. The real issue is how we define tolerances on performance, etc, and how industry will be able to measure this. The other issue is how we can test 10k antennas in a reasonable amount of time. This needs to be worked out in the GP300 stage. The exact manner of deployment depends on the final mechanical design. However, it will require a crew to prepare the surface, a crew for installation and a crew for (remote) checking of the antenna station. Maintenance is a serious issue. Station failure could arise from lightning or atmospheric E-fields, extreme atmospheric conditions (winds), solar power malfunctioning, battery lifetime, installation mishaps (connectors), vandalism. Goal is to reduce the station failure rate by design considerations (reduce connectors, ease installation, lightning protection, wind load requirements). At this moment, no realistic estimate for the failure rate exists. Again, GP300 will serve as a proper field to investigate this.

Q: What is the failure rate of detection units?  What is the impact on the failure rate of environmental factors, like wind and snow?

A: We do not know right now. GP300 has to provide information. So far, for cosmic rays only engineering arrays have been built where maintenance is rather high. This is different for radio-astronomy where maintenance is low. The main difference is in the fact that air-shower detector units are autonomous, so we have to work on increasing reliability. Auger has shown that with 1600 units it is possible. There, most issues arose from the phototubes, which we do not have.

Q: How will data communication work?

A: We will use a commercially available solution and adapt to our setup. A 5 GHz WiFi system has been measured to work reliably over more than 10 km, and is designed to work over more than 50 km. Such a system could be used when dividing the array into overlapping segments that each talk to a ‘central’ daq where the trigger is created. Alternatively, a gossiping system where each antenna only sees its neighbors requires less power and is more flexible. Event storage may be a bottleneck in this case as each station close to an event sink needs to retransmit a complete event; preferrably without introducing deadtime. Both the ‘classical’ and ‘gossiping’ options will be further investigated. For GP300, a classical centralized system will be used.

Q: Can we find a construction site that is sufficiently radio-quiet and has sufficient ease of access?  Can we find more than one such site?

A: Considering the sheer size of Western China, the answer has to be yes. So far, we have made ~50 measurements in 6 sites from April 2017 to August 2018, and have identified 5 out of 6 that were very good sites for a GRAND 10k hotspot. On the other hand, other considerations, such as political ones, also have to be taken into account.

Q: How can we be sure that the location we deploy our arrays will still be radio quiet in 30 years?

A: We cannot be sure. What we can and have to do is to make agreements with local officials on electromagnetic noise but especially on the communication band used by us! In addition, we should continue to improve trigger algorithms to take care of changing (and likely increasing) backgrounds.