Site Selection Story, 2002–2006: Shortlist

Abstract

The SKA telescope site selection process was long and complex. It went through five distinct stages: (1) separate national initiatives in China and Australia to identify potential sites (1994–2002), (2) centrally coordinated activity to identify and characterise these and two other potential sites in Southern Africa and Argentina-Brazil (2000–2005), (3) the short-listing process and decision (2004–2006), (4) further characterisation of the two short-listed sites, and politicisation of the site competition (2007–2011), and (5) the final site selection process and decision (2009–2012).

This chapter describes and analyses the first three stages of site selection from the early national initiatives to the creation of a shortlist of two potential hosts, Australia and Southern Africa. Initially the International SKA Steering Committee (ISSC) aimed to select the SKA site in 2006 but after advice from the funding agencies interested in the project, the goal of outright selection changed in mid-2005 to a shortlist of acceptable sites.

Chapter 8 continues the story to its culmination in 2012 when the SKA governing body decided to locate the mid-frequency dish array in Southern Africa and the low-frequency array in Australia.

7.1 Introduction

Site selection is often one of the most contentious decisions a project will face as hosting a large international scientific facility is a major prize for any country. It has the potential for severe disruption of the project, and often requires considerable negotiation and compromise to reach an acceptable solution. Decisions need to be driven primarily by considerations of where the best science can be done but political aspects and the obligations involved almost always come into play. In SKA’s case, the site selection process was meticulously prepared with the participation of all contenders but, even so, in the final stages, a compromise was required taking political demands into account to ensure the project survived.

SKA site selection went through five distinct stages: (1) separate national initiatives in China and Australia to identify potential sites (1994–2002), (2) centrally coordinated activity to identify and characterise these and other potential sites (2000–2005), (3) the short-listing process and decision (2004–2006), (4) further characterisation of the two short-listed sites, and politicisation of the site competition (2007–2011), and (5) the final site selection process and decision (2009–2012).

Key characteristics of an acceptable site from the science perspective at the start of the formal process in 2002 were: a wide sky coverage and substantial overlap with other major astronomical facilities, radio quietness in view of the high sensitivity of the SKA, ability to site remote stations at distances up to 3000 km from the core of the array to provide the wide range of angular resolution required by the science case, and ionospheric and tropospheric stability to optimise image quality at low and high frequencies respectively.

This chapter describes and analyses the first three stages of site selection from the early national initiatives to the creation of a shortlist of two potential hosts, Australia and Southern Africa.

7.2 Early (National) Site Investigations

7.2.1 China (1994–2002)

Initial studies in early 1994 of possible sites for a large radio telescope in China focussed on mid-latitude locations with large areas of relatively flat land that were isolated but accessible.¹ However, the best of these possible locations, in Sichuan Province, was not regarded as competitive with other countries and attention began to turn to the unique karst depression region of Guizhou province where multiple Arecibo-like spherical reflectors could be accommodated (see Fig. 7.1 left) in an array called KARST (Kilometre-square Area Radio Synthesis Telescope)). The first SKA Radio Frequency Interference (RFI) measurements in the world were made at eight potential sites in Puding and Pingtang counties in Guizhou by Bo Peng, Yaoping Nie (all from the National Astronomical Observatories of China) and Richard Strom (ASTRON, The Netherlands) in late 1994 (Peng et al., 1995), and resulted in a decision in mid-1995 to focus on this area as a potential site for the SKA in China.

In late-2002, following several years of project development and engineering effort on the telescope design for a KARST prototype called FAST (Five hundred metre Aperture Spherical radio Telescope), a site survey group in the Guizhou University of Technology extended the search to cover the entire province of Guizhou. Some 90 candidate depressions were located, each of 500 m or more in diameter, even up to about 1000 m in a small number of cases (see Fig. 7.1 right). Eventually, in 2005, the Dawodang depression was selected for FAST² in the final stages of the SKA site short-listing process (see Sect. 7.3.7.1).

Fig. 7.1

Left: The Dawodang karst depression in Guizhou Province before construction of the FAST telescope began. (Credit: National Astronomical Observatories of China). Right: Results of site surveying in Guizhou Province, China. Vertical axis is number of depressions vs diameter and depth in metres. (Credit: National Astronomical Observatories of China)

7.2.2 Australia (1996–2002)

As we have seen in Sect. 3.2.6.3, a proposal for Australian participation in the SKA was included in the decadal plan by the National Committee for Astronomy (an Australian Academy of Science sub-committee) in 1995. It was felt that Australia could provide design expertise and a ‘superior site free from interference if sited in the western areas of the continent’. This was taken up by the State Government in Western Australia in 1997, following an interview given by ATNF Director, Ron Ekers, to the local press the previous year. Government support was offered in a letter from the State Premier to Ekers in February 1998.³

Bruce Thomas, an engineer in the CSIRO Radiophysics Division, spear-headed investigations on the suitability of potential sites for the central area of the SKA⁴ (see Fig. 7.2). The main factors affecting the choice were the provision of protection against radio interference, minimising the impact on land users including the indigenous communities, pastoralists, and the mining industry, and how easy would it be for SKA operations staff to travel to the central site and to be accommodated there.

Fig. 7.2

Map of Australia showing the central array at Mileura in Western Australia with remote stations out to 2000 km (Credit: Bruce Thomas, CSIRO Radio Astronomy Image Archive CRAIA-SKA005)

The first area of interest was near Carnarvon (830 km north of Perth), with a focus on properties which were no longer economic to operate as pastoral enterprises. In these cases, the State determined it could purchase the leases for conservation purposes. More interesting opportunities came up further south in the Murchison area north-east of Geraldton, and in 1999–2000 detailed studies⁵ (see Fig. 7.3) were made by Thomas and colleagues of the possible impact of radio communication links and other services in the area. The Mileura Station (~100 km west of Meekatharra) was the main location⁶ and a testing program was carried out in March–April 2001 once the sheep muster⁷ was over! The results showed that the levels of radio communication activity were extremely low. Subsequently, two adjacent areas in the Murchison (one including the Mileura Station) were gazetted under the WA Mining Act for exemption from mineral exploration. More elaborate studies continued in Western Australia to monitor RFI in 2002 and 2003.

Fig. 7.3

left and right: Two of the antennas used to monitor RFI in the Murchison area of Western Australia in 2001 (Credit: Bruce Thomas, CSIRO Radio Astronomy Image Archive CRAIA-SKA003 (left) and CRAIA-SKA004 (right))

Other Australian states were also interested in exploring opportunities to be involved and, by late 2001, South Australia, NSW and Queensland had also expressed interest in hosting the SKA. The Federal Government became involved in discussions that eventually led to WA being given priority in 2004.

By mid-2000, efforts were already underway to define a legal framework for RFI protection for the SKA if it were to be sited in Australia, and a workshop on interference mitigation was held in Sydney in December that year. This was motivated by the US Government’s legal protection for the National Radio Astronomy Observatory’s site at Green Bank, West Virginia which provided a precedent for Australian legislators. The interference environment became a major selection criterion for the SKA site as the process developed and, as we will see later in this chapter and the next, RFI measurement campaigns were carried out for both major stages of the selection.

7.3 Site Short-Listing

7.3.1 Identification of Potential SKA Sites (2000–2004)

The International SKA Steering Committee (ISSC) began to take central control of the site selection process in 2000. The clear understanding was that the site choice would be made by the scientists involved based on where the science goals could be best accomplished. A recent example in people’s minds was ALMA where the high-altitude Atacama Desert in northern Chile was the unanimous choice and was made by the scientific community.

At an ISSC meeting in Aug 2000, following a recommendation from the Five-year Management Plan and Technical Oversight Working Group chaired by Bob Preston, (Jet Propulsion Laboratory, USA) the ISSC decided to bring the site selection activities together under a common umbrella and formed the Site Evaluation and Selection Committee (SESC) chaired by Bruce Thomas (see Box 7.1). The same meeting decided to call for Expressions of Interest (EoI) to host the SKA, initially via the OECD Task Force on Radio Astronomy (see Sect. 3.2.5.2), but this was not pursued. Instead, direct invitations to URSI Commission J (Radio Astronomy) national representatives from around the world were issued by the ISSC Executive Secretary, Russ Taylor, (University of Calgary, Canada) in late-2001.

The Call for EoIs had the desired effect, prompting the US to start examining the possibility of siting the SKA in the Southwestern United States in relation to the Expanded VLA project,⁸ while in South Africa a proposal was submitted to the National Research Foundation in mid-2002 for SKA funding including €200 thousand for a site survey.⁹ In the end, institutes in seven countries—Argentina, Australia, Brazil, China, Ireland, South Africa, and USA—responded to the Call. Ireland withdrew their response after discussion with European consortium leaders since Europe’s position was that the SKA would be best sited in the southern hemisphere.¹⁰

Box 7.1 Site Evaluation and Selection Committee (SESC)

Bruce Thomas (CSIRO Australia) was the first chair. Other members of the initial SESC were Subramaniam Ananthakrishnan (TiFR, India), Willem Baan (ASTRON, Netherlands), Justin Jonas (Rhodes University, South Africa) and Shengyin Wu (NAOC, China). On Thomas’ retirement in 2002, Yervant Terzian (Cornell University, USA) took over as SESC chair and Wim Brouw (CSIRO) replaced Thomas as Australian representative. In 2004, the management structure of the SKA evolved to reflect the growing coordination role of the International SKA Project Office (ISPO) and, as part of this change, the SESC became part of the ISPO with a new name—the Site Evaluation Working Group. Terzian remained as an enthusiastic and effective chair until 2008 when, as part of the PrepSKA re-organisation, the SESC was replaced by the Site Characterisation WG chaired by Rob Millenaar (ASTRON, The Netherlands) as Site Engineer in the SKA Program Development Office.

One of the high priority tasks for the SESC was to generate site selection guidelines to accompany the letters of invitation in 2002. Other tasks were to standardise RFI measurement sets and define the requirements for Radio Quiet Zones in which to site the SKA.

In November 2002, the ISSC invited 20-page initial site analyses from the six remaining candidate countries with the intent of narrowing down the field. Specific topics to be addressed by the candidate countries are shown in Box 7.2. In its preparatory discussion, the ISSC approved a suggestion from Justin Jonas (South Africa) that it would level the playing field for emerging radio astronomy countries if interim site “white papers” could be submitted for non-prejudicial assessment and feedback from the ISSC. Although the South Africans never availed themselves of this opportunity, the results of the site short-listing 4 years later proved they were fast learners!

Box 7.2 Initial Site Analyses: Issues to be Addressed

• general location and impact on science including sky observation range

• general attributes of the proposed Central Site, including the surrounding area out to a distance of about 300 km for access to services and radio-quietness and compatible land use;

• specific site attributes of importance for engineering design issues, impact on the operation and construction of the SKA facility, and service provision;

• radio-quietness determination and frequency band-occupancy measurements;

• willingness to host the SKA including an in-principle agreement with national and local authorities relating to legislation and/or regulation for protection of radio-quietness and land-use, as well as an indication of interest in providing the Central Site as an internationally-recognised Radio-quiet Reserve for compatible scientific facilities and research (in addition to the SKA), and for supporting its acceptance (prior to proclamation) at both national and international forums.

By the deadline of May 2003, the ISSC had received proposals from Australia, China, South Africa and the U.S.¹¹ The proposal from Australia noted three possible central sites and the array distributed throughout Australia. South Africa also suggested three possible central locations and with long baseline stations located in several countries to the North. China proposed two possible sites, one in the karst depressions region in Guizhou for Arecibo-type antennas, and the other, a region in Western China for an array of fully steerable smaller diameter dishes. The U.S. proposed a location near the VLA site in New Mexico with array stations across the U.S., Canada and Mexico.

Argentina, Brazil, and Europe did not submit proposals—lack of resources was the problem in Argentina and Brazil, and Europe’s opinion was that the site should be in the southern hemisphere. However, they reserved the right to submit a northern hemisphere proposal if no suitable site were to be found south of the equator. The ISSC extended the deadline for Argentina and Brazil to the end of March 2004 and separate proposals were submitted by the new date. Argentina identified three possible locations for the core including a high site at 2550 m and the distant array stations in Brazil. Brazil itself also proposed three possible locations for the core of the SKA, all in the eastern part of Brazil, a few hundred kilometres from Brasilia.

Evaluation of the telescope site locations by the SESC was based on three global criteria¹²: (1) the ability to do the optimal science with the instrument, (2) the construction cost at the proposed site, and (3) the operational costs for the telescope at the proposed site. In July 2004, they duly recommended that the four original preliminary proposals from Australia, China, South Africa and the USA be accepted for a full site proposal. The Brazilian proposal was not accepted since its central sites were too close to the magnetic equator and at too low an altitude (900 m). Proximity to the magnetic equator carried the risk that observations would be subject to severe ionospheric disruption at the lower frequencies while low altitude, wet conditions carried the risk that tropospheric phase irregularities would cause problems at high frequencies. The SESC recommended that Argentina and Brazil cooperate in one new proposal, and this is what transpired. Five potential sites were in the competition.

7.3.2 Initial Characterisation of Potential Sites: Radio Frequency Interference (RFI) Monitoring, Radio Quiet Zones, 2002–2005

7.3.2.1 RFI Monitoring

In 2002, the ISSC approved the establishment of a working group on RFI measurements¹³ with Steve Ellingson (Ohio State University, USA) as chair. A report on RFI measurement was produced within a year,¹⁴ and based on this report, the SESC recommended that RFI tests of all sites should be done with the same equipment and the same team. Standardisation of calibration and measurement was critical for a good quality comparison of the RFI situation at the sites, both in a relative and in an absolute sense.

In 2004,¹⁵ the ISSC decided that ASTRON would supply the RFI monitoring equipment and the measurement procedure.¹⁶ Four-week periods of testing were planned at the nominated candidate central sites by an ASTRON team under an MoU with the ISPO. Although providing the set of uniform measurements recommended by the SESC, the 4-week visits to each site by the ASTRON team were recognised as insufficient for full knowledge of the RFI environment and its variability, so each potential site was required to carry out a program of RFI monitoring using the same measurement protocol for 12 months. This was to include overlap with the centrally coordinated measurements, for cross-calibration purposes.

The ISPO (Peter Hall, International SKA Project Engineer) coordinated the international campaign with logistical assistance from a committee comprising members from each site. Measurements by ASTRON engineer, Rob Millenaar, took place at four of the five sites in 2005 over a range of frequencies from 70 MHz to 26.5 GHz. By this time, the US had withdrawn from the competition (see Sect. 7.3.5.1). The required local monitoring campaigns were also carried out in all four countries/regions for about 1 year (see Figs. 7.4 and 7.5).

Fig. 7.4

Samples of RFI measurements carried out at the four candidate sites by ASTRON on behalf of the International SKA Project Office in 2005. The cluster of frequency responses around 260 MHz is from the UFO (Ultra high frequency Follow On) geostationary satellites used by the US Navy. The other responses in the spectrum are mainly due to distant TV transmitters and other communications systems, see the plots for South Africa and Argentina (hba.skao.int/SKAHB-434 SKA Site Spectrum Monitoring (SSSM) Miscellaneous Notes—Sites: South Africa, China, Australia, Argentina-Measurement period: 2005/2006, A. J. Boonstra and R. P. Millenaar, ASTRON Report to ISPO, 19 March 2006. This figure is a re-worked version by Rob Millenaar of figure 4 in Boonstra and Millenaar, 2006. Information on the sources of interference in this figure from Rob Millenaar, email to R. Schilizzi, 12 May 2022) (Credit: A. J. Boonstra and R. P. Millenaar, ASTRON)

Fig. 7.5

Left: Rob Millenaar (foreground) and Bou Schipper (background, seated), ASTRON engineers in charge of the RFI measurement campaign together with Chinese colleagues, on site in July 2005. (Credit: NAOC). Right: The ASTRON RFI monitoring equipment in the foreground on site in China in 2005 (credit: R.P. Millenaar, ASTRON)

7.3.2.2 Radio Quiet Zones

Building on experience with the designated Radio Quiet Zones (RQZ) surrounding the NRAO Green Bank site, it was a requirement from the early days of the SKA that the core should be located in such a designated zone. In 2004, the OECD Global Science Forum Task Force on Radio Astronomy and the Radio Spectrum noted that, for optimum performance, the SKA should have local protection for all frequencies and global protection for specific bands allocated to radio astronomy.¹⁷ A Resolution was passed at the URSI General Assembly in 2002 calling for an investigation of the desirability of, and issues involved in, establishing International Radio Quiet Reserves for future radio astronomy. The OECD Task Force also recommended the establishment of a small number of “…zones on the ground where future radio observatories could be located and within which satellite emissions could be controlled.”.¹⁸ A proposal to establish such internationally protected zones was not approved by the ITU.

Since the establishment of RQZs required governments to pass appropriate laws or regulations, it was essential for the national communities to raise the priority of the SKA project among senior government figures. Discussions on RQZs began in earnest in Australia in 2004, when Ron Ekers made a presentation on the case for Australia to host the SKA to the Prime Minister’s Science Engineering and Innovation Council.¹⁹,²⁰ The Prime Minister supported this concept in principle and it was then given high priority in discussions by Government officials. In South Africa, Justin Jonas, Bernie Fanaroff (SKA South Africa) and colleagues raised the possibility with government officials in 2003 but in connection with a radio quiet area in the country that could be used for SKA technology testing.²¹ Chinese discussions began somewhat later, in 2006, following approval of the FAST project by the Chinese Academy of Sciences.²² In Argentina, the proposed location for the central site for the SKA was in a National Park in the Andes in San Juan Province which had the status of a Protected Astronomical Reserve. There, the Complejo Astronomico El Leoncito (CASLEO) operated two optical telescopes and one solar radio telescope. A provincial law from 1987 protected the site from electro-magnetic interference, lights etc. within an area 15 km across. An extension to the law specifically for the SKA was under consideration in late-2005 at the time of the site proposal submission that would include a Coordination Zone for all transmitters operating between 100 MHz and 25 GHz located in the same area.

7.3.3 Timeline to Site Selection

Bob Preston, as Chair of the Five-year Management Plan and Technical Oversight Working Group, led an ISSC discussion in January 2003²³ in which site selection by the ISSC in late-2005 and Facility Definition (technology, governance) in late-2007 was proposed. This led to questions about whether a simultaneous choice of site and technology might not be a better strategy (see Sect. 7.3.7), but the goal of site selection in late-2005 remained.

7.3.4 Request for Full Proposals, 2004–5

A formal Request for Proposals (RfP)) for siting the SKA²⁴ was issued in September 2004 by the ISSC to the five countries that had submitted acceptable Site Analyses in the earlier round, with a deadline of the end of 2005, and decision on the SKA site in August 2006. It set out the defining characteristics of the SKA for both the LNSD and SNLD concepts, and the criteria by which the site proposals were to be evaluated (see Box 7.3). Each potential site was also required to carry out a 12-month program of RFI monitoring in addition to the month-long international measurement campaign at each individual site (see Sect. 7.3.2.1) and include the results in their response to the RfP. The clear aim of the ISSC was to select one site for the SKA Telescope by September 2006.

The RfP also set out the evaluation process. Proposals would be reviewed by the ISSC separately and an independent, external International SKA Site Advisory Committee (ISSAC), who would report their findings to the ISSC. Expert analysis of the data provided by the proposers would be made by the Site Evaluation Working Group (SEWG) and the Simulations Working Group (SimWG). The ISPO served as the clearing house for all queries from individual site proponents and distributed answers to all candidate sites to maintain an open and transparent process.

The final stage was expected to be a discussion between the ISSC and highly ranked proposers in order to come to the final decision. The ISSC declared itself as the final authority in all aspects of the decision process but, as we will see, the funding agencies intervened in mid-2005 and early 2006 to exert their influence and change the desired outcome to a short list rather than outright selection of a single site.

Box 7.3 Defining Characteristics of the SKA in the Request for Proposals to Site the SKA

• Large Number-Small Diameter (LNSD) concept: 1 km core diameter + 40 stations within 5 km diameter + 60 remote stations out to 3000 km

• Small Number-Large Diameter (SNLD) concept: 1 km core diameter + 20 stations within 5 km diameter + 30 remote stations out to 3000 km

• Central 5 km configuration supplied to proposers

• Maximum baseline >3000 km

• Visible sky—above 30° for all stations for >4 h/day

• Data transport from remote stations at 100 Gbit/s (minimum), 1 Tbit/s (final)

• Link from Facility Support Centre to national and international SKA data centres at 100 Gbit/s (minimum)

• Power

  • Central site 12 MW (peak)

  • Facility Support Centre 2 MW (peak)

  • Remote Station 150 kW (peak)

The global criteria for the evaluation of the proposed telescope sites were set out in the RfP²⁵ as follows:

• 1. the ability of the SKA to maximise the science return of the instrument if located at the proposed site;

• 2. the construction cost to project at the proposed SKA site

• 3. the operational cost to project for the proposed SKA Facility; and

• 4. physical and political issues

The specific criteria within each category (see Box 7.4) were discussed in detail in the RfP, but no weights were assigned at this point. (These were generated in the run-up to the selection in July 2006.)

Box 7.4 Specific Evaluation Criteria for Potential SKA Sites

The quality of science:

• (a) Short- and long-term radio-frequency interference and protection issues.

• (b) Array configuration and performance.

• (c) Ionospheric and tropospheric conditions.

Infrastructure, climatic and costing issues:

• (a) Climatic issues.

• (b) Physical site-characteristics for Stations.

• (c) Impact of land-use and urban centres.

• (d) Existing infrastructure.

• (e) Data interconnects.

• (f) Costs—capital and operating.

National attributes for siting the SKA:

• (a) General issues.

• (b) Government and departmental interaction with SKA community.

• (c) Support for astronomy and the SKA Facility by national and regional governments.

It is safe to say that the “quality of the science” criterion had the highest weight in the minds of ISSC members. Three particular issues were matters of debate in the lead-up to the Request for Proposals:

• 1. the relative advantages and disadvantages of the two main array types in contention—Large Number-Small Diameter (LNSD) and Small Number-Large Diameter (SNLD) (see Chap. 6 on SKA Design). Proposers were instructed to generate one overall layout of the array consisting of the central 5 km diameter area and remote stations on spiral arms and two specific configurations within the overall layout to allow for the LNSD and SNLD concepts. This became an issue for the Chinese site bid near the end of the proposal submission period (see later in this section).

• 2. The RFI environment (see Sect. 7.3.2), and

• 3. The tropospheric environment.^{26 27} In-situ interferometric measurements of tropospheric phase fluctuations were not feasible on the short timescale to the site decision. As a proxy, the ISSC decided that archival meteorological data including satellite data should be assembled by the site proponents to characterise the water vapour content throughout the year. In addition, the Simulations WG was to study the capabilities of self-calibration as a means of removing the effects of differential phase fluctuations, as well as the influence of possible variations on imaging dynamic range.

7.3.5 Responses to the Request for Proposals

7.3.5.1 USA

In early-2005, the chair of the US SKA Consortium, Yervant Terzian, informed the ISSC that the US would not be able to respond to the SKA siting RfP by the end of the year.²⁸ Behind that bland statement, it was clear that there were significant differences of opinion about the desirability of the NRAO taking charge of the site bid process with a core centred on the VLA in New Mexico (see Sect. 3.3.3.8). However, the Consortium noted that the US could provide an outstanding site for high frequencies and it may be possible to take advantage of that later, were the SKA to divide between low- and high- frequency arrays. A decade later, preparation began on proposals for the “new generation VLA” that has been inspired by early ideas for SKA-high.

7.3.5.2 Argentina-Brazil

The proposal²⁹ was submitted by the Argentine and Brazilian SKA Committees led by Marcelo Arnal (Argentina) and Jacques Lepine (Brazil). A radio-quiet high altitude valley located near the Andes Ridge in the Argentinean Province of San Juan was proposed for the central five kilometres of the SKA layout (Fig. 7.6) with remote stations stretching across the north-eastern territories of Argentina and into south Brazil (see Fig. 7.7). As mentioned in Sect. 7.3.2.2, the intended SKA core location was within a National Park that already had been designated a Protected Astronomical Reserve and hosted two optical telescopes and one solar radio telescope.

Fig. 7.6

The proposed central site for the SKA in San Juan Province in Argentina (Credit: R.P. Millenaar, ASTRON, 2005; see also Millenaar, R., 2016)

Fig. 7.7

Proposed array layout for Argentina-Brazil (hba.skao.int/SKAHB-438 A proposal for siting the SKA in the Territories of Argentina and Brazil, 2005) (Credit: the Argentina-Brazil SKA Committee)

Key factors in support of the proposal were:

• legislation was already in force in the Astronomical Protected Reserve to prevent the installation of transmitters detrimental to radio astronomy;

• negligible water vapour content at the altitude of the Central Site (2300 m);

• relatively low land acquisition costs; a major astronomical facility only 12 km away from the selected Central Site able to provide support and initial infrastructure such as roads, communications, and accommodation;

• proximity to some of the front-line twenty-first Century astronomical facilities located in South America (primarily in Chile just across the border with Argentina) allowing large common sky coverage and simultaneous observations with these facilities; and

• strong support from governmental authorities and from scientific federal agencies and major national academic institutions.

A major attraction of the bid (to potential members of the SKA HQ staff reading the proposal) was the promise of “pleasant weather, superb regional wines, excellent natural food and world class ski centres within driving distances, add an extra value assuring the quality of daily life for all SKA headquarters personnel”!

The proposal noted that the relatively close location of the Central Site to the magnetic equator and the possibility of unstable ionospheric conditions caused by energetic electrons that could affect the quality of low frequency observations. A subsequent detailed study provided by the Argentina-Brazil Consortium showed that this was a “disabling” characteristic for the proposed location (see discussion in Sect. 7.3.8.1).

7.3.5.3 Australia and New Zealand

The proposal³⁰ was submitted by Brian Boyle, Australian SKA Director, on behalf of the Australasian SKA Consortium. It envisaged the core site for the SKA located at Mileura Station (see Fig. 7.8) in the State of Western Australia, almost 700 km from Perth and remote SKA array-station sites spanning the continent of Australia (see Fig. 7.9). In addition, SKA array-stations could be located in New Zealand at Warkworth and Ardmore in the North Island near Auckland and Rangiora and Awarua in the South Island to extend the maximum continental east-west baseline of 3200 km to over 5500 km to provide even higher angular resolution observations.

Fig. 7.8

The proposed central site for the SKA core near Mileura Station in Western Australia (Credit: R.P. Millenaar, ASTRON, 2005; see also Millenaar, R., 2016). The location of the central site was changed to Boolardy Station on Wajarri Yamaji Country, approximately 80 km west of Mileura, in 2008 to avoid potential interference from mining activities

Fig. 7.9

Proposed array layout for Australia with the core site centred at Mileura (hba.skao.int/SKAHB-439 Proposal for Siting the SKA in Australia, 2005). The locations of individual array elements are shown as green dots. (Credit: CSIRO Radio Astronomy Image Archive SKA008)

Key factors noted by the Australasian Consortium in support of their proposal were Australia’s radio-quietness across the whole extent of the SKA, state and federal Government commitments to preserve this environment in the long term, the stability of the ionosphere and low water vapour content of the troposphere in the winter months, the high-quality existing infrastructure and political and economic stability, and the international strength of Australia’s radio astronomy community. Both LNSD and SNLD array types could be accommodated. The planned extended New Technology Demonstrator (xNTD) project which transformed into ASKAP (see Sect. 4.3.3.1) a year later was mentioned as an important factor in establishing a site that satisfied geophysical, radio-quietness, environmental, governance and regulatory requirements prior to the SKA.

7.3.5.4 China

The proposal³¹ was submitted by the National Astronomical Observatories of China and proposed a configuration for the SKA centred around the Dawodang karst in Guizhou Province in south-west China (see Fig. 7.1 left). Several hundred possibly suitable depressions were identified for the remote stations within a distance of a few hundred kilometres (see Fig. 7.10). This depression, and the core region around it, lies in a remote mountainous area with a sparse population. The Guizhou Radio Managing Bureau together with the national Radio Regulatory Department had agreed to establish a radio quiet preserve with a diameter of 200 km centred on Dawodang depression.

Fig. 7.10

Proposed array layout for China (hba.skao.int/SKAHB-440 A proposal for siting the SKA in China, 2005). The large-scale distribution of karst depressions. The blue arc is at 3000 km from the central karst depression. (Credit: National Astronomical Observatory of China)

A key supporting factor for the proposal was the RFI environment of the candidate sites with the hills around each depression providing a natural shield against radio interference. This had been shown by a series of on-site monitoring sessions for radio interference to investigate the ability to satisfy the RFI requirements for building the SKA (see Sect. 7.3.2).

The proposal did note that since the depressions are the result of a long geological history and are naturally distributed, there was no possibility to fit their configuration to the arbitrary distribution in the core area set out by the ISPO Configuration and Simulation WG in any exact way. The final choice of the depressions would have to be optimised to produce an acceptable reception pattern for the array. By its very nature, the depression geometry did not support an LNSD configuration, a fact that ultimately led to this proposal not being short-listed as we discuss in the next section. The zenith angle limitations of the spherical structure would also lead to limited tracking duration for southern declination radio sources and, consequently, relatively poorer overlap for simultaneous observations with other major astronomical facilities.

Furthermore, Guizhou Province and the proposed central region are located in a sub-tropical humid monsoon climate that would not trouble an SKA operating at low frequencies but would be a substantial drawback for the higher frequencies in the science requirements. Guizhou also suffered from the same problem as the Argentina-Brazil proposal that it is close to the magnetic equator with the attendant ionospheric stability problem potentially impacting the scheduling of low-frequency observations in particular.

7.3.5.5 South Africa

The South African proposal³² was submitted by Rob Adam on behalf of the Department of Science and Technology and proposed to site the SKA core site in the arid Karoo area in the Northern Cape Province (see Fig. 7.11). Remote stations were proposed in other parts of South Africa as well as in Namibia, Botswana, Mozambique, Mauritius, Madagascar, Kenya and Ghana to achieve the 3000 km baseline requirement (see Fig. 7.12).

Fig. 7.11

Proposed central site for the SKA in the Karoo Desert in the Northern Cape province of South Africa. (Credit: R.P. Millenaar, ASTRON, 2005: see also Millenaar, R., 2016)

Fig. 7.12

Proposed array layout for Southern Africa submitted in December 2005 in response to the Request for Proposals from the International SKA Steering Committee (credit: SKA South Africa) (hba.skao.int/SKAHB-441 South African Bid to Host the Square Kilometre Array, 2005)

Key factors in support of the bid were the quiet RFI environment at the central site and at a representative sample of remote station locations within 150 km where RFI measurements were carried out in support of the site bid. The very low population density and lack of economic drivers in the Karoo area, the terrain screening of the central core area in the direction of Cape Town almost 500 km away and the prospect of further improvement provided by regulatory measures via the Astronomy Geographic Advantage Act, were all factors supporting the claim of RFI quietness.

With an altitude of over 1000 m at the central site and its location in the arid Northern Cape Province, there is a low precipitable water vapour content in the troposphere above the site which makes it suitable for high frequency observations with the SKA. The ionosphere is stable over Southern Africa apart from low elevation observations in the west where the South Atlantic Anomaly comes into play.

As with Australia and China, South Africa realised that designing and building a pathfinder telescope on the proposed site would provide added weight to their proposal in terms of demonstrating an active site. The Karoo Array Telescope (KAT) served that purpose and was the forerunner of their precursor, MeerKAT (see Sect. 4.3.3.2).

7.3.6 Reflections on the 2005 Site Proposals

It is interesting to see which of the selection criteria were emphasised in the individual proposals. Australia put in a confident proposal, no weaknesses, which reflected their view and that of almost all in the community at the time that they were the front-runners. Argentina-Brazil were defensive on the configuration and ionosphere, and China was defensive on the configuration, ionosphere and troposphere. The South African proposal was also confident but did choose to devote the first page of the Executive Summary to a statement from the President of the Republic, Thabo Mbeki, giving his support for projects like the SKA, as well as a list of government entities providing science-support infrastructure and political support for South African participation in the SKA. This suggests they had taken on board the funding agencies’ view (see Sect. 7.3.7) that site selection was not up to the science community alone and that politics would play a role in site selection. It was therefore important to dispel any feeling in the outside world that local political support was a weakness in the proposal.

Also noteworthy is how the attitude and ambition of South Africa evolved in the space of a few years from being one of participation in a global scientific project, to playing a leading role³³, to playing the central role in the project as evidenced by the site proposal itself (see Sect. 7.3.5.5). As the final sentence of the Executive Summary of the Site Proposal said: “South Africa has the will, the capacity and the sites to construct, operate and maintain the SKA over its lifetime in support of the global astronomy partnership that it represents.”

7.3.7 The Funding Agencies Intervention

7.3.7.1 The Heathrow Meeting in June 2005

At the time of the Request for Proposals in Sept 2004 and for nine months thereafter the SKA project was proceeding smoothly towards a definitive selection by the ISSC of a single site for the telescope. The informal view of all ISSC members was that Australia was the obvious choice for the SKA site since it was a country with traditional strength in radio astronomy and vast tracts of RFI-free space to locate the telescope. The US was in second place, also a traditionally strong radio astronomy country, but located in the northern hemisphere and with far less open space than Australia. So, when the US withdrew, the result of the competition appeared a foregone conclusion, with politics playing a minor role.

However, as described in Sect. 3.4, the feeling of “plain sailing” changed fundamentally in June 2005 following the first interaction with funding agency and government ministry representatives who were meeting as a group at Heathrow airport to discuss future large astronomical facilities. It was made clear to the ISSC that SKA site selection in September 2006 (by the ISSC) was premature and that the funding agencies must be involved.

After considerable debate, the ISSC agreed that the outcome of the RFP process should change from a “final decision” on the location of the SKA to a decision on the ranking of the four sites based on scientific, technical and infrastructure cost grounds.³⁴ This was to form the basis of a recommendation on acceptable sites the ISSC would submit to the governments and funding agencies involved in the SKA in September 2006. Thereafter, further characterisation of the physical characteristics and RFI environment of the sites would be carried out in parallel with round-table discussions on scientific and geo-political issues including radio quiet zones, cost implications and cost-sharing possibilities, prior to a final decision on the site by the end of 2008.

A decision with ramifications for the site selection was made at the 14th meeting of the ISSC in Pune, India in November 2005. The LNSD dish concept was chosen as one of the three elements of the Reference Design to take forward to the funding agencies in early 2006.³⁵ Effectively this removed the Chinese site proposal from contention less than two months from proposal submission deadline since it envisaged a small number of large diameter spherical dishes in karst depressions in south-west China. However, during the ISSC meeting, the Chinese delegate, Bo Peng, informed the meeting that not including the FAST concept in the Reference Design was not a problem as long as the Chinese delegate at the planned governments/agencies SKA meeting in The Hague in February 2006 saw a picture of FAST during the meeting. No doubt this relaxed reaction was motivated by Peng’s further announcement that the FAST project had just been approved formally by the Chinese Academy of Sciences with fifteen new positions and an initial budget of 10 million Chinese Yuan. However, it was not the Chinese intention to withdraw their proposal for the SKA site and Peng was keen to ensure it was evaluated in the same way as the other three.

7.3.7.2 The Hague Meeting in February 2006

This was the first funding agency meeting devoted entirely to the SKA. It was notable for the proclamation by the chair, Richard Wade (STFC, UK), at the end of the closed session that the site selection process should result in a short list of acceptable sites since the agencies felt the project would benefit from a period of negotiation ahead of the decision, if needs be with “blood on the floor”.

Accordingly, the ISSC modified its approach from ranking to short-listing the proposals. It did, however, accept that short-listing was a non-confrontational way of selecting scientifically acceptable sites, even if the “blood on the floor” approach to the final negotiations leading to selection was not consistent with the collegiate culture of the SKA project (see Box 7.6). A fuller discussion of the Hague meeting is given in Sect. 3.4.3.

Box 7.6 ISSC Versus Funding Agency (Soft Versus Hard) Approach to Site Selection

The ISSC’s original approach that it would make the final selection of the site independent of higher-level political considerations reflected a certain naiveté that was quickly dispelled at Heathrow. However, it is interesting to speculate whether the “softer” approach of “round-table discussion” proposed by the ISSC in August 2005 would have led to a different outcome for the site selection than the hard competition mandated by the funding agencies. One such outcome of the ISSC approach might have been a compromise between the top-ranked sites, Australia and South Africa, whereby a mutually—acceptable sharing of the spoils was achieved without the rancour of the final stages of the selection process in 2012. With Australia in the pole position in 2007-8, they would have been in the stronger position to fashion the compromise more to their liking.

One negative consequence of the competitive relations between Australia and South Africa engendered by the governmental/funding agency entry into the site selection process was the move in Australia to make the post-2006 internal deliberations on their site proposals confidential. This blocked the normal process of community consultation and stifled innovation.

7.3.8 Evaluation of the Proposals

In parallel with the changing position on the outcome of the site selection, the ISSC and ISPO worked out the details of the evaluation process sketched in the Request for Proposals, and the procedure to be followed to come to a short-list. A sub-committee of the ISSC (Wim Brouw (ASTRON (The Netherlands), Phil Diamond (Jodrell Bank Observatory, UK), Schilizzi (convenor), Jill Tarter (SETI Institute, USA), Yervant Terzian (Cornell University)) re-worded the criteria by which to judge the proposals, established their individual weights in the evaluation process and selected the Analytic Hierarchical Process (AHP) as a quantitative method of comparing the sites for each criterion.³⁶ Mindful that the full ISSC membership, including site representatives, would vote on the short list, an external committee, the International SKA Site Advisory Committee (ISSAC), was formed to carry out an independent evaluation of the proposals and report to the ISSC.

Members appointed to the ISSAC were all well-known figures in international astronomy and included Richard Hills (UK, chair), Jacob Baars (Germany), Jacqueline van Gorkum (USA/Netherlands), James Moran (USA), Ernest Seaquist (Canada), Govind Swarup (India), and Robert Williams (USA). Schilizzi, as SKA Director, acted as Advisor to the ISSAC.

Protocols were agreed for each of the entities involved in the site evaluation—ISSC, ISPO, ISSAC, and the Site Evaluation Working Group (SEWG) and Simulations Working Group Task Forces—to define their roles and work to be done, and to ensure the selection procedure was understood and followed by all parties. The ISPO was again the clearing—house for all queries or issues raised by any of the parties during the evaluation phase.

7.3.8.1 ISPO-Led Evaluations

Detailed evaluations of different aspects of the proposals began early in January 2006. These were carried out³⁷ by a sub-group of the SEWG on ionospheric conditions, the RFI Assessment Task Force on the RFI environment, the Regulatory Task Force on Radio Quiet Zone regulations, the Configuration Simulations Task Force to calculate Figures of Merit, and an external consultant (Parsons-Brinkerhoff) on power generation costs.

These evaluations, and those carried out in parallel by the ISSAC, led to several requests for additional information or questions of clarification for the potential sites.³⁸ All sites had to provide more information on tropospheric opacity, costs of providing and operating access roads, costs of power generation and distribution, and costs of data connectivity to the SKA. Specific questions were asked of Argentina, Australia and China, one of which—information on ionospheric scintillation related to the core site in Argentina—had a profound effect on that country’s prospects of selection for the short list. Ionospheric scintillation is a frequency-dependent effect that causes phase distortions in the incoming cosmic radiation, particularly at the low frequencies envisaged for the Epoch of Reionisation array (which later became SKA-low). Of concern to the SEWG sub-group was that the geomagnetic equator of the Earth swings south of the geographic equator in the vicinity of Brazil and Argentina and this was known to cause enhanced scintillation.

To respond to the request, Marcelo Arnal and colleagues in Argentina commissioned an ionospheric scintillation impact report from NorthWest Research Associates in Tucson, Arizona, USA.³⁹ The report pointed out that the latitude of the proposed location for SKA was just south of the “southern equatorial anomaly”. This is a region of potentially severe ionospheric scintillation that shows a strong diurnal variation as well as a seasonal variation which peaks in intensity near the equinoxes. The scintillation also increases with the 11-year solar cycle, driven by increases in the solar output of extreme ultraviolet, or EUV, radiation.

To back up this general statement, 72 day-of-year versus time contour plots of the expected worst-case S₄ intensity-scintillation index for four frequencies of interest (100, 250, 600, and 1000 MHz) were included at nine viewing geometries (overhead, and 30° and 60° elevation at 0°, 90°, 180°, and 270° azimuth from true north). This was done for both solar minimum (defined as a sunspot number of 10) and solar maximum (150) conditions (see example at 100 MHz in Fig. 7.13). The plots showed there was a considerable risk that severe distortions would occur particularly for observations to the north, and that these distortions would be experienced even at frequencies as high as 1 GHz. In the final analysis by the ISSC (see Sect. 7.3.8.4), this was seen as a disabling characteristic for the Argentina-Brazil bid.

Fig. 7.13

Scintillation coverage map superimposed on South America (credit: NorthWest Research Associates, prepared for the Instituto Argentino de Radioastronomia)

7.3.8.2 ISSC Internal Site Evaluation Process

In proposing that the Analytic Hierarchy Process (AHP) of pair-wise comparison of candidate sites be used by the ISSC, the sub-committee noted that this approach had been used in site selection processes for LOFAR and the Australian SKA core site. In contrast, the South Africans had used a “fatal flaw” analysis to select their core site.⁴⁰ This latter approach also formed part of the final decision process for the SKA site, but with the less confrontational name of “disabling characteristics” .

AHP is a multi-criterion decision support tool⁴¹ that allows groups knowledgeable about the subject to convert their well-informed qualitative judgements into a quantitative structure. It is well-suited to a site selection process where many factors are relevant to each site criterion and judgements on complex comparisons need to be quantified.

There were two phases to the application of AHP to the SKA⁴²: (1) ISSC agreement on the selection criteria and weights proposed by the sub-committee, and (2) independent analysis of the relative merits of the sites against these criteria by the ISSC members in order to rank the four sites.

Despite the adoption by the ISSC of the LNSD concept as part of the Reference Design just before the proposal deadline, and the consequent change in acceptable configurations, the selection criteria were not adjusted to respect the considerable resources and effort spent by proposers in generating their RfP responses which included both the LNSD and SNLD options.⁴³ However, one exception was the category of “political issues” in the RfP (called “National attributes for siting the SKA” in Box 7.4) which were no longer considered since they did not fall under scientific, technical and infrastructure cost, and were deferred to the post-short-list stage of site selection.⁴⁴

The final evaluation criteria for selection of the site short-list and weights are shown in Table 7.1. The set of weights generated by each ISSC sub-committee member independently were the same within the statistical errors.⁴⁵

Table 7.1 Final evaluation criteria and weights for SKA site short-listing

All but one of the 21 ISSC members carried out their own AHP analysis and sent the results independently to the SKA Director by mid-August 2006. All analyses received returned the same result: (1) Australia, (2) Southern Africa, (3) Argentina-Brazil, (4) China.

7.3.8.3 ISSAC Review and Report

The International SKA Site-selection Advisory Committee (ISSAC) provided a ranking on the basis of the selection criteria⁴⁶ based on assessing the same information received by the ISSC as well as interviews with delegates from each site. On advice from the ISSC, the ISSAC did not consider questions relating to political support for the project or to the possibility of financial contributions linked to the choice of site.

The Executive Summary of the ISSAC report stated:

The ISSAC has studied the documentation provided by the four candidate sites and heard presentations from representatives of each of the proposers. The Committee also took account of a set of factual reports and summaries provided by the International SKA Project. The ISSAC assessed the strengths and weaknesses of the sites on each of the criteria set out in the request for proposals. Our conclusion is that two of the sites, those based in Australia and the Republic of South Africa, clearly stand out as the best. They are both excellent sites and the Committee believes that they are both fully capable of meeting the full range of the requirements of the SKA, including scientific, technical and practical issues. There are some discernible differences between these two sites: for example, Australia can accommodate long East-West baselines more easily and currently has somewhat lower levels of interference, while South Africa has a rather more benign climate, and some aspects of the infrastructure are better. If the heaviest weighting is given to the factors which directly affect scientific performance, Australia has a small but measurable advantage over South Africa.

The two other sites, in Argentina-Brazil and in the People’s Republic of China, are also very good sites for radio astronomy. In particular Argentina would be favoured for a project that was focused on the higher radio frequencies but is disadvantaged by ionospheric effects and radio interference at low frequencies. The karst landscape and low levels of interference in the Guizhou Province of China make it ideal for large reflectors like the FAST project, but the terrain presents difficulties for the construction of the SKA, which has stringent requirements for the placing of the antennas, and especially for the configuration of the SKA which is based on a large number of small antennas. The ISSAC therefore concluded that neither Argentina and Brazil nor China represent competitive sites for the SKA.

This report was transmitted to the SKA Director on 1 August 2006 in whose care it was held in confidence until the ISSC internal evaluation had been completed 2 weeks later. After the release of the report to the ISSC and proposers, both Argentina and China took the opportunity to rebut comments made by the ISSAC. Argentina noted that one of the RFI sources at 100 km distance was to be removed later in 2006, while China objected to the speculation that mobile phones would soon be a significant source of RFI for an SKA sited in China. Subsequent experience at the FAST site has shown that the Chinese objection was well-founded.

The ISSAC report, together with the AHP result from the ISSC and the analyses of the submissions by the ISPO working groups and consultants, formed the primary material taken by ISSC members to the 16th ISSC meeting in Dresden, Germany in late-August 2006 where the site short-list decision would be made.

7.3.8.4 ISSC Decision on the Short-List of Acceptable Sites

The order of business for the ISSC meeting in Dresden was carefully choreographed by the ISSC Executive Committee comprising Phil Diamond (chair), Wim Brouw, Brian Boyle, Jill Tarter and Richard Schilizzi in order to create a robust, fair, open and transparent process.

The first agenda point was formal approval of the definition of an acceptable site as ‘An acceptable site is one for which the usable frequency range, configuration, sky coverage, and physical characteristics allow the key scientific goals of the project to be achieved efficiently over the lifetime of the telescope’.

The proceedings began with an open session with all members, site proponents, and the ISSAC Chair, Richard Hills, present (see Fig. 7.14). At each stage, there was an opportunity for questions, discussion and comment. Schilizzi provided an overview of the analyses of the site proposals carried out by the ISPO working Groups and task Forces (see previous section), followed by Hills with a report on the ISSAC process and its outcome (see previous section), and finally Schilizzi with a report on the ISSC AHP outcome (also described in the previous section). No large variations in the conclusions were found as a result of varying the weights of the main selection criteria, in particular by making the cost, ionosphere and troposphere weights equal to zero in a variety of combinations.

Fig. 7.14

The ISSC members and observers, taking a break from deliberating on the site short-listing at its meeting in Dresden, Germany, in August 2006. Front row, left to right: Bo Peng, Bryan Gaensler, Peter Dewdney, Ken Kellermann, Anne Green, Gloria Dubner, Joe Lazio, Justin Jonas, Yervant Terzian, Richard Hills (ISSAC Chair), Phil Diamond (ISSC Chair), Wim Brouw. Back row left to right: Richard Schilizzi, Russ Taylor, Jill Tarter, Brian Boyle, Peter Wilkinson, Bob Preston, Jim Cordes, Wim van Driel, Thijs van der Hulst, Peter Hall, Dave De Boer, and Arnold van Ardenne. (Credit: Franco Mantovani, also ISSC member)

A closed session followed with only ISSC members, ISSC Secretary and SKA Director present in order to generate a short list of sites that were “acceptable” taking into account potential disabling characteristics. An open vote was held on the ranking of the sites, using a two-step procedure:

• 1. Each ISSC member was allocated 12 virtual votes that he/she could distribute to each site with no allocation exceeding 6 votes. A clear order was established: from top down, Australia, South Africa, Argentina—Brazil and China.

• 2. Each member then filled in a matrix of the first six major criteria vs. site. If a site was regarded as unacceptable with respect to a particular criterion, this was noted in the appropriate matrix element. If more than 75% of the ISSC felt a site was unacceptable with respect to the same criterion, that was counted as a disabling characteristic. If less than 25% of the ISSC felt a site was unacceptable on the basis of a particular criterion, that was to be ignored. Any vote between 25% and 75% triggered discussion and a formal majority vote on the acceptability of that site.

As a result, all sites had at least one disabling characteristic indicated by at least one ISSC member. Australia and South Africa had less than 25% of the votes for any potential disabling characteristic and were included in the short-list. China was seen by more than 75% of the vote to have no viable configuration option as an SKA site and was removed from the short-list. The Argentina—Brazil ionospheric properties were seen by between 25% and 75% of the vote as being a disabling characteristic, thereby needing further discussion and a majority vote. Note that the actual percentage of the vote in each case was not recorded in the minutes of the meeting.

Subsequent discussion centred on whether the ionospheric limitations could be bypassed for any of the Key Science Projects, but none were found; in particular, it was felt that no Epoch of Reionisation (EoR) observer would select the Argentina-Brazil site for an EoR telescope. The formal vote confirmed this, which left a short-list of two acceptable sites: Australia and South Africa.

7.4 Site Short-Listing Outcome

The ISSC Chair, Diamond, summarised the outcome as follows:

There seems to be good consistency between all the input information, an indication that we have a fair and robust process overall. There is no question that all four proposed sites were amongst the best sites for radio astronomy in the world. However, following a voting process, the ISSC has ranked the sites in the following order (best first): Australia, South Africa, Argentina- Brazil and China. In the short-list selection the ISSC found no disabling characteristic for Australia and South Africa and declared them short-listed.

For the proposal from China the configuration options were considered to be a disabling characteristic for the SKA as envisaged, and China was removed from the short-list. However, the site is exceptional in its low RFI properties, and should be preserved at all costs as the site for large collecting area, single dish radio astronomy.

For the Argentina—Brazil proposal, the ionosphere was considered a disabling characteristic for important parts of the SKA Key Science Projects, and Argentina—Brazil was not included in the final short-list. The Argentina site is a great site for high-frequency radio astronomy and would get the support of the ISSC for any proposal to put a high-frequency instrument there.

The candidate site representatives all noted that they accepted the outcome and regarded the process as being robust, fair, open and transparent.

The SKA Funding Agencies Working Group approved the short-listing outcome, and the accompanying summary of the strengths and weaknesses of the four sites, at a meeting a month later, in September. The project moved on to the next phase of site selection which was every bit as complicated and difficult as had been predicted by Richard Wade earlier in the year.