Concluding Remarks

Abstract

We conclude this account of the history of the SKA in its formative years with a brief analysis of the major issues and challenges faced by the project against a background of megaprojects in general and science megaprojects in particular. Other projects and mega-project scholars may benefit from seeing how this particular endeavour navigated its way towards becoming a reality.

11.1 Introduction

The major elements of SKA development in its first two decades from 1990–2012 have been explored in some detail in the preceding chapters. Here we attempt to draw back from the details of the historical narrative and sketch the broad issues that occupied the attention of the SKA project, particularly in what we have called the “Transition Era” from 2006 to 2012 (see Fig. 4.1). By 2006, the global radio astronomy community was convinced it was in a position to take on a project of the scale of the SKA and accomplish the goal of constructing the world’s largest radio telescope. This conviction was based on the work already carried out on the project nationally and internationally, as well as previous experience in the community in building large telescopes like the VLA and global collaborations for Very Long Baseline Interferometry (VLBI). Recognition of the SKA as a potential “project of pan-European relevance” by the European Strategy Forum for Research Infrastructures¹ (ESFRI) in 2006 was one of the first steps. But there was much still to learn about how to implement a global project and problems to overcome before the SKA was accepted as one of the landmark astronomical observatories for the twenty-first Century.

For much of the time in the Transition Era, considerable uncertainty remained as to whether the SKA would gain sufficient support from the community and funding agencies to become a reality. It was also a time when it had to weather a number of severe storms including the withdrawal of a major partner.² Several potentially existential issues faced the project including (i) achieving recognition of the SKA as a high priority project in “roadmaps” generated by the wider astronomy community in Europe and the USA as well as recognition by funding agencies and governments,³ (ii) establishing a long-term governance structure centred on a legal entity to create a stable environment for the project,⁴ and (iii) choosing a site for the telescope in an increasingly tense competition that had been politicised and elevated to President/Prime Minister level in the shortlisted candidate countries. In parallel with resolving these issues, the preparatory phase, PrepSKA, contract with the European Commission⁵ expected the project to deliver a costed telescope design and a signature-ready document to start construction by 2011.

Projects of all sizes, and large projects in particular, have their own tempos dictated by internal and external—national and international—influences, many of which are unpredictable. No two projects are alike, and the authors do not plan to set out here a list of prescriptions for the mega-science development process for other projects in similar phases of development.⁶ Nor will we compare the SKA to other individual science mega-projects in order to draw conclusions about better practice that might have been adopted. Rather, in the first part of this chapter we will attempt to sketch the characteristics of the SKA during its phase as an early-stage mega-project in the hope that other projects and mega-project scholars will draw benefit from seeing how this particular endeavour navigated its way towards becoming a reality. In the second part of the chapter we will also reflect on specific issues and decisions that were taken in the SKA project collaboration over the course of the two decades we cover in this book, and then make some general observations.

We start by sketching what defines a mega-project and in particular a science mega-project before going on the describe the challenges faced by the SKA in working at this level.

11.2 Mega-Project Characteristics

Mega-projects are usually defined as large-scale complex ventures, typically having (multi-)billion-dollar budgets, timeframes measured in decades, and attracting a high level of public and political attention. Mega-projects are often (although not always) transformational, and can have social, economic, scientific, and technological impact. “Mega” also implies the size of the task involved in developing, planning, and managing projects of this magnitude. The risks are substantial and cost overruns are common.⁷

As (Flyvbjerg, 2014) notes, “megaprojects are not just magnified versions of smaller projects but are a completely different breed of project in terms of their level of aspiration, lead times, complexity, and stakeholder involvement. Consequently, they are also a very different type of project to manage.”

Mega-projects appear across a range of national and international endeavours including infrastructure, water and energy, defence, information technology and software systems, many fields of science, aerospace projects, industrial processing plants, mining, transport, and large strategic corporate initiatives and upgrade programs to increase the capability of already existing mega-projects. Global megaprojects carry additional areas of complexity in terms of collaboration, logistics and funding compared with national megaprojects.

Mega-projects are constantly growing in scale and cost (Flyvbjerg et al., 2003).^8,9 Budgets of 50–100 billion dollars or euros are now relatively common, and costs above 100 billion dollars not unknown, e.g. the International Space Station, the F-35 Joint Strike Fighter and the UK’s high-speed rail project. A conservative estimate in 2014 by Flyvbjerg (2014) for the global megaproject market was 6–9 trillion dollars per year, or approximately 8% of total global gross domestic product. Funding for mega-projects most often comes from central governments, often appropriated through domestic agency budgets. Philanthropy is also a funding source, and not unusual especially in the USA. Foreign aid is occasionally a source in association with capacity building for developing nations.

11.2.1 Science Mega-Projects

As discussed in Chap. 1 (De Solla Price, 1983) recognised the important role played by the transition in project scale from individual researcher to institute to national facility and, finally, to international facility, each step removing a resource limitation ceiling. He coined the terms ‘little science’ and ‘big science’ to describe the two extremes (De Solla Price, 1963). Institutional facilities are built to enable research on a scale which no individual can afford; national facilities are built to enable research on a scale which no single institute can afford; and likewise international facilities are built to enable research on a scale which no single nation can afford.

Science mega-projects are at the top end of international facilities and have five distinguishing characteristics (Crosby, 2012a)¹⁰: (i) Typical budgets are in the range of 1 to 10 billion euros, with space science and high energy physics occupying the high end of the scale; (ii) Funding is usually derived from national (government) sources, although philanthropic contributions to funding also occur, especially when encouraged by tax incentives. For the most part, the funding for global projects remains under the control of the national funding sources rather than unencumbered cash contributions to a central project office, and is usually made in the form of in-kind contributions to the design efforts and juste retour (fair work return) contracts in the construction phase; (iii) Almost by definition science mega-projects are daring cutting-edge enterprises that leapfrog existing technological capability to deliver new knowledge and understanding. In contrast, “Commercial enterprises are understandably more risk averse in terms of direct financial return to themselves but may well think in terms of issues like building capability. A government research group on the other hand, should arguably have wider and longer-term measures of impact.”¹¹ (iv) Science mega-projects are complex¹² to manage and almost always involve international collaborations between scientific institutions, universities, and industry; and (v) A key difference with institutional infrastructure mega-projects is that the users, who are in a sense customers, are members of the research community and are distinct from the financier(s).

Delving deeper into science mega-project management, Crosby¹³ pointed out that the innovative character of these projects requires that many new technologies and components need to be developed in parallel by academic and industry partners, and this adds to the complexity. A substantial preparatory phase before the beginning of construction is necessary to deliver a competent understanding of technical scope and associated risks. This is sometimes undertaken by academic institutes, often under rather blurry funding arrangements that can complicate later claims that the work done qualifies as in-kind support for the project. Central project authority is aspired to for obvious project management reasons. However, without central financial control on an international scale, major participants naturally maintain their highest allegiance to their own institute or department and the national/regional funders. For SKA in the Preparatory and Pre-Construction phases, this necessitated a different approach to project control, best described as “centrally managed best efforts”.

Crosby also noted that technical project reviews in science mega-projects are also seen in a different light to their industrial equivalents, being less procedural and action oriented, more collegiate and allowing more freedom in terms of options for corrective action especially in the project shaping phase—a characteristic that can sometimes lead to “scope-creep”. In the SKA, this was a consequence of the centrally managed best-efforts basis of project management (see Sect. 6.2.2.6 and hba.skao.int/SKASUP6-3). Science technical project review panels for sub-system elements are usually dominated by specialists—people with domain knowledge—from the project stakeholders, in most cases individuals with experience in other major projects. With stakeholders from around the world, SKA had access to a wide range of, often divergent, views. In contrast, industrial style reviews include a higher proportion of external experts from other types of major projects and industry, and this was seen as bringing worthwhile value from ‘the outside world’. The SKA benefitted from this approach for its International Engineering Advisory Committee (IEAC) formed in 2007 (see Sect. 6.2.2.3) and for the SKA Concept System Design Review in 2010 (see Sect. 4.5.2). The latter had a major positive effect on project direction and focus at a critical time in the project.

Crosby’s PhD research undertaken concurrently with the final years of the period under review in this book included the SKA as a case example. It produced a practical output from the developed theory in the form of an audit (or review) tool, the Checklist for HIgh technology ProjectS (CHiPS).¹⁴ This has been used since then to complement reviews of the type discussed above, or as an independent assessment tool to verify key project indicators for any or all project stages.

Project termination in mega-science is rare although “descopes” are quite routinely used as a coarse instrument to contain cost.¹⁵ This occurred for the SKA in 2014–15 as a result of the imposition of a cap on the budget for capital expenditures by the SKA Organisation’s Board of Directors (see Sect. 8.7). A feature of institutional science mega-projects with relatively weak central project authority is a perceived lack of consequences for the collaborative teams in cases of non-conformance on, for example, unfulfilled delivery promises. Prior to the project entering the centrally controlled construction phase with its contract-based activities, any sanctions on individual institutes for poor performance or delivery are measured more in terms of loss of reputation than any legal or financial penalty. This is an essential difference with industry-based models for mega-projects which should be recognised.

11.3 The SKA as a Science Mega-Project

The initial 1990 concept of the SKA as a radio telescope with a collecting area of one million square metres and one hundred times the sensitivity of the then current state-of-the-art, the US Very Large Array (VLA) (see Sects. 2.4 and 5.5.5) implied a mega-scale project scale even without the label of “mega-project” . The first recorded mention of a link between a large radio telescope like the SKA and mega-science was in May 1993 following a talk given by Ron Ekers on the large telescope ideas circulating in the radio astronomy community at the European Space Agency meeting on Frontiers of Astronomy. Françoise Praderie (then Scientific Head, OECD Mega-Science Forum, MSF) contacted Ekers with the suggestion that SKA could be considered by the MSF.¹⁶ This led to the OECD-sponsored activities on radio astronomy and large telescopes described in Sect. 3.2.5.2 during the period from 1996 to 2004. It also led to the involvement in the SKA of the two international scientific unions most relevant for radio astronomy, the International Union of Radio Science (URSI)¹⁷ in 1993 and the International Astronomical Union (IAU) in 2004. Two years later, the European Strategy Forum for Research Infrastructures (ESFRI, see Sect. 4.3.2.2.1) included SKA as a potential pan-European research infrastructure in the roadmap compiled in 2006. By this time, it was clear that SKA ticked the boxes for a science mega-project—large price tag, long timescale for completion, innovative science, cutting-edge concept, global, and complex project management. The ESFRI action paved the way for SKA to obtain Preparatory Phase funding from the European Commission in 2007, a crucial step on the path to the successful transition to the SKA Organisation in late-2011.

One characteristic that distinguishes the SKA from other science mega-projects is its “grassroots” origin.¹⁸ As described in Chap. 2, it began life in 1993 as a global community-driven mega-scale project¹⁹ that did not originate in an existing organisation or institution or government. One of its major activities during the period we describe in the book was to create a host organisation since no suitable candidate existed in the astronomical world that could accommodate a global project like the SKA (see Chaps. 3 and 4). Key to the SKA’s early development was the involvement of a small number of institute directors and senior staff who actively sought funding.²⁰ The URSI Large Telescope WG²¹ built up the science and engineering case over time, and, as funds became available, relatively light forms of governance (MoUs) were created by the institute directors to coordinate activities. The SKA remained a collaboration governed by MoUs or MoAs for 18 years until December 2011 when a legal entity was established for the first time to manage the activities.

11.3.1 Innovative Concept

As set out in the SKA Science Case in May 1998,²² the aim was to build “the world’s premier astronomical imaging instrument. No other existing or planned instrument in any wavelength regime can provide simultaneously: angular resolution better than the Hubble Space Telescope (< 0.1 arcsec), field of view significantly larger than the full moon (~ 1 square degree), spectral coverage of more than 50% (ν/Δν < 2), spectral resolution sufficient for kinematic studies (ν/dν > 10⁴), and all at a sensitivity about 100 times the VLA.”

The authors of this book note that what is under construction by the SKA Observatory at the time of writing is the first 10% phase of the final telescope which already has a budget and global scope justifying the “mega” designation.

11.3.2 Complex Project Management

The unavoidably wide range of stakeholders makes the SKA a complex project to manage. There are multiple nations involved²³ and multiple players within those nations including large and small research groups in research institutes and universities, government departments and funding agencies, and large and small industrial organisations. In the telescope site candidate countries, Australia and South Africa, radio spectrum management agencies as well as indigenous and farming communities involved in land-use agreements were additional stakeholders.

At an operational level, such a wide range of stakeholders created many challenges and tensions in managing the SKA, as we have discussed in this book. Examples are the different funding cycles in the countries and regions involved, different prior investment histories in radio astronomy leading to differences in experience and maturity level of relevant technology development, different scientific priorities impacting the scientific requirements on the telescope design, different policies towards industry engagement and juste retour on investment, different cultural approaches to science and decision-making at all levels and different science-government interaction cultures. And finally, at a geo-political level, government-level relationships between nations were important. Indeed, mega-projects like SKA are highly relevant to relations at the level of science minister to science minister. Whether the SKA made it up to the next level politically depended on its visibility nationally; it was certainly the case in Australia, South Africa and China and subsequently in the UK. On the other hand, international relations between nations that are on less than friendly terms can lead to restrictions being placed on exchange of information on state-of-the-art technology or cross-border supply of the technology itself—and in a project with such long timescales, who is on friendly terms with whom may itself change. In a global project like the SKA, these differences needed to be understood and managed for a successful collaboration. In all of this, we emphasise open and transparent communication across the project was a key element of success.

As we have noted in earlier chapters, over the 18-year collaboration phase from 1993 to 2011, central project management in the SKA evolved considerably. It took place on four levels: (i) overall project oversight and major decision-making initially (1993–2005) carried out by the scientific steering committees and later in conjunction with the funding agency groups²⁴ and the SKA Founding Board; (ii) coordination of the global effort on the science case and engineering design by the SKA community; (iii) coordination of the three PrepSKA “policy” work packages (governance, procurement & industry engagement and funding) by the funding agencies from 2008–2011; and (iv) coordination of telescope site characterisation by the SKA Project Office in 2004–6 and 2008–2011. In addition, the individual partner institutes were responsible for the management of technical design work-packages.

With no single nation or organisation occupying a dominant position in the SKA project through expertise, magnitude of funding or legal position, an additional layer of management complexity was added compared to the top-down centrally funded and managed projects more common in government and industry. Issues of governance, funding strategy and site decision strategy had to be resolved by “sufficient consensus” within the International SKA Steering Committee (ISSC) and SKA Science and Engineering Committee (SSEC) in conjunction with the Agencies SKA Group (ASG) as discussed in Chaps. 3 and 4. This was successfully carried out for the most part. A balance was found between the scientist-driven aspects of the program and strong agency engagement while avoiding the suggestion that the Funding Agency involvement in the SKA was in any way a formal endorsement of the project²⁵ until the time came to make such a commitment.

As far as oversight of the SKA technical work-packages was concerned, there was a further difference with industrial project development. The latter mostly relies on expert contractors to produce the required physical systems and components to a given set of specifications. In contrast, much of the instrumentation for the SKA was designed and fabricated in-house in the research institutes, drawing on combinations of co-located specialist skills. A notable characteristic of many of the SKA team members, at least in the early stages of the project, was the ability to bridge the science-engineering gap, with scientists well-informed in relation to the engineering challenges (often contributing to the technical design), and many of the engineering staff adept at understanding the science challenges in terms of practical design of experiments and equipment. This characteristic is common to successful science mega-projects and provides the most innovative technology development path.

However, in collaborative enterprises, questions of “authority” on design issues can lead to tensions within the project. In a collaboration without a dominant partner, design authority is gained primarily by domain expertise and track record. Design authority was the role to which the central SKA Project Office - both International SKA Project Office (ISPO) and SKA Program Development Office (SPDO) - aspired but its achievement took longer than expected at the outset.

The role of the SKA Project Office in the development of the SKA design began slowly in 2004 in the early days of the ISPO and initially continued the annual review of technology progress in the institutes by the International Engineering Management Team under Peter Hall’s leadership. In 2005 the ISPO led the Tiger Team developing the Reference Design (see Chaps. 3 and 6), and in 2007 a similar Tiger Team developing the Preliminary Specifications for the SKA²⁶ (see Memo 100 and Sect. 6.2.1.4). In both cases, the Tiger Teams were composed primarily of members from the institutes involved in design work. The new SPDO began to grow in numbers in early-2008 with PrepSKA and institute funding and a mandate to carry out the overall project management and coordination (see Sects. 4.4.2.1 and 6.2.2.4).

It took time to recruit the SPDO staff and for them to achieve a comparable level of expertise, if not experience, to the institute-based engineers. It was difficult to attract leading specialists from the partner institutes, particularly those with system engineering talent, to the central office. Only the relatively unencumbered were prepared to take the gamble of moving to a project not yet funded for construction and no guaranteed long-term future. Many were not interested in moving from the current exciting hands-on activities to a long-term paper design activity. Consequently, it took additional time for SPDO staff to be recognised by the community as the design authority in practice. The SPDO adopted a mode of operation of consulting widely with the community before making decisions, paying attention to the individual concerns of the partner institutes as much as possible in order to move the project forward—the centrally managed best-efforts approach. By the time the SKA Project Execution Plan was generated in 2010, it was judged an opportune moment to formally designate the future SKA Project Office as the “Design Authority” in the forthcoming pre-construction when design consortia would formally report to the Project Office.

One of the issues faced in the SKA at the start of the PrepSKA project in 2008 was the substantial lack of experience in the radio astronomy community of projects larger than could be handled by individual institutes. Only their involvement in ALMA was comparable. The education (and self-learning) of the community about the ramifications of SKA as a science mega-project took some time. One example was the necessity to communicate conflicts in resource priorities between national projects and the international project. The system engineering approach advocated by the SPDO and adopted in 2009 helped the engineers in the institutes come to terms with the distributed nature of the design work and enabled a more coherent approach to the overall design process (see Sect. 6.2.2.2).

Complex Project Management (CPM) as a recognised phenomenon with distinct behaviours, was still in its infancy at the time PrepSKA commenced in 2008 (Geraldi et al., 2011) (Crosby, 2012b). However, CPM as a discipline was not on the radar of the SKA leadership throughout PrepSKA. The focus there was on delivering the required outcomes by 2011 while at the same time resolving the existential questions about the SKA mentioned in the Introduction to this chapter. The issues of uncertainty and dynamic and socio-political complexity then being introduced into CPM were not unknown to the SKA; they had been part of the project management framework from the start. In any case, it is doubtful that the management resources required for CPM would have been regarded as having higher priority than staff to fill the technical domain specialist positions to interact with the institute engineers. The fifteen staff able to be funded in the SPDO²⁷ were well below the original estimate of requirements (28) made in 2006 (see Chap. 9). Both early and later SKA project teams had fundamental project management skills and experience, but true CPM (at the level of the Large Hadron Collider project at CERN) was not part of their toolkits. It is probably fair to say that management of the design process of a global science mega-science project is much more straightforward to establish within an already well-established organisation like CERN or the European Space Agency (ESA) . In 2011, the SKA was just on the point of establishing a stable project environment.

11.3.3 Governance

It is obvious that a governance framework is critical to any science project once it involves collaboration among scientists and engineers in more than a small number of institutes. The ability to revise or change the governance to suit the development phase of the project is also critical. In each major transition that SKA went through (see Chaps. 3 and 4), experience showed that finding a governance structure that ensured mutual advantage for all parties was a key factor underpinning continuing success in the collaboration. Understanding the agendas of the potential member parties in the collaboration was key to ‘sealing the deal’ in each transition. The sequence of changes in governance entities from the Large Telescope Working Group in 1993 to the SKA Organisation’s Board of Directors in 2011 is shown in Table 11.1 (See also Fig. 4.1).

Table 11.1 Governance entities for the SKA

Interesting to note is that the SKA leadership did not regard the global governance aspects from 1993 to 2011 as being particularly challenging in practice, with the exception of the tri-partite governance in operation during the PrepSKA contract period from 2008–2011. The three governance entities in the PrepSKA era were: the SKA Science and Engineering Committee, the Agencies for SKA Group, and the PrepSKA Board, an arrangement that proved onerous at times for the SPDO.²⁸ Years of successful experience in running collaborative global VLBI networks indicated that well-tested governance structures existed in radio astronomy which could be, and were, adopted for the SKA in its collaboration phase up to 2011. In the pre-construction era that followed, it seemed natural that the approach taken in most European inter-governmental research organisations, including the European Southern Observatory (ESO), of a Board with one government representative with voting rights and one scientist per country, would serve for the SKA. and so it did.

11.3.4 Industry Engagement

While it had been obvious to the SKA project from its grassroots days that engaging industry was going to be essential in the design, prototyping and construction phases, there was also an expectation among some in the radio astronomy community that industry involvement—and lobbying—would put additional pressure on governments to provide funding for the SKA. Later, in the transition era from 2006–2012, there was also explicit encouragement from the funding agencies in all countries except the USA to involve local industry in order to facilitate associated industrial spin-off to adjacent markets and ensure national benefits from government investment in the project. In the USA, the approach is less political; the National Science Foundation funds basic research purely on research quality and it is assumed that US industry captures the benefits without making this an explicit requirement. This meant that the SKA project had to target individual companies in the US like IBM and others (see Chap. 10). But this effort was hindered by the lack of experience among US astronomers with the industrial strategies used in other countries.

Early engagement with industry was only partially effective as discussed in Chap. 10. It did allow the development of productive relationships with industry players setting a foundation that both inspired many collaborations with industry and gave confidence to the funding agencies to support the project at the national level. The ambitious science goals of the SKA served as a catalyst for innovative thinking by industry. However, one consequence of engaging industry in the design phase was that protection of intellectual property at institute level became an issue. In some cases, the design progress was impeded by Non-Disclosure Agreements binding some of the participants in design meetings. Inability to obtain protected industrial production methods had a profound effect on cost estimates for dishes, a dominant cost component of the SKA (see Sect. 6.4.6).

The dangers of the radio astronomy community not understanding the different cultures, expectations, and modes of operation of industry were ever present. While the larger companies were well used to R&D phases lasting many years and saw involvement in a state-of-the-art project like the SKA as a means to enhance their capability,²⁹ the majority of smaller companies envisioned a shorter timescale between initial interactions and commercial contracts. In retrospect, the initial projections for the early contractual, and procurement, phases were not only optimistic, but subject to continual slippage. This resulted in much disenchantment in the smaller companies.

Having a well-considered procurement approach  was recognised in the PrepSKA study to be strategically important to the successful completion of the SKA project and would underpin productive and open relationships with suppliers. Prior to 2012, there was no need to implement a formal procurement function, but thought was applied to how procurement might best be implemented and managed through the work of PrepSKA Work Package 5 (WP5, see Sect. 4.4.1 and Chap. 10). As well as producing much in the way of procurement guidelines and preferred approaches, WP5 endorsed the early establishment of a procurement office structure with resources, processes, roles and responsibilities, and information management systems in place. There was also encouragement from WP5 for the SKA leadership to obtain a full understanding of global supplier capability information, and implementation of appropriate contractual instruments with terms and conditions directly referencing and supporting project goals.

11.3.5 Project Plans: Costs and Timelines

Both cost and timescale were severely underestimated in practice in the SKA, a situation not uncommon in mega-projects.^30,31 Many project plans were made for the SKA over the years. They were necessary to provide focus for the project but were less than accurate predictions of the course of events.³² Even with well-developed cost planning tools and techniques, experience with mega-projects, especially those with a large software component, shows cost uncertainties of +100% are not unknown.^33,34 There was much activity and thought given to project costing throughout the Transition Era from 2006 to 2012, where possible using ALMA as a “reference class” radio telescope for the SKA, but little of this activity influenced the publicly announced ‘big numbers’ for which the SKA could be built. A mantra of a “one and half billion Euro” instrument (for SKA Phase 2, the full SKA) prevailed throughout the Transition Era until the first industry-led analysis of the dish costs (the major cost driver for the telescope) made in 2011–12 showed that early estimates of these costs used in the submission to the US Decadal Review Panel in 2009 were an underestimate by a factor of four to five.³⁵ This led to estimates of the total cost for the full SKA being most likely a factor of three too low (see discussion in Sect. 4.4.3.3.1, and 4.5.3.5).

In retrospect, the project was not at a sufficient technical readiness level to make a detailed cost estimate in 2009 as the system-wide Conceptual Design Review had not been held by then and would not be held until a year later. This points to one of the complexities inherent in a global project, that of the project not having control of the timing of major review cycles in the different countries in terms of technical readiness to comply with proposal requirements on budgets. The conclusion here is that the SKA project was naïve to give as much credence as it did to the early cost estimates until there was much better access to the details of the design and the assumptions made. The conundrum is that to obtain accurate costs, irrevocable design decisions needed to be made, cutting off options.

Preliminary cost estimates based on scaling and rough models are in most cases only known to within a factor of two or more. This should be translated into a contingency allocation (see Sect. 6.4.6). But the reality is that very few projects include contingency in their initial estimates as it is not required until formal funding requests are made.

The timeline for SKA has proven to be equally optimistic and governed to some extent in the pre-2006 period by what was the “acceptable” horizon for completion (as seen by the astronomical community). As noted in Sect. 4.6.1, a relatively short projected time to potential funding and, beyond that, to start of construction is almost universally found embedded in large project plans and is a well-known sociological phenomenon.³⁶ The National Audit Office report clearly links over-optimism with the nature of complexity. Truly complex projects are quite different to merely complicated ones. Their behaviour is non-linear, hard to predict and inevitably flawed. To try to predict outcomes by time or cost is not sensible without a multiplier for contingency.

However, an initial project timescale with a distant milestone for funding approval runs the risk of the wider community turning its attention to shorter-term projects with more immediate scientific and reputational returns. A shorter timescale to funding approval gives institutes and individual colleagues the feeling it would be useful to engage with the project sooner rather than later in order to get an inside edge on technology development and so influence design decisions or access to the telescope when operational. It also makes logical sense to quote a technically limited timescale when there are so many unknowns, like political decision-making, outside the project’s ability to manage.

This situation began to change for the SKA when the project started to appear in national and European roadmaps in 2006. Other external influences on the timeline came into play at that time including competing projects like the European ELT, and the PrepSKA deliverable of an Implementation Plan for the Pre-Construction Phase³⁷ starting in 2012.

11.3.6 Measuring Success

Measuring success in any mega-project, as a general class, is ill-defined. Industrial-style reviews focus on measuring progress against rigidly defined critical success factors, with schedule, cost, and scope (or performance) ranked uppermost. The first two of these factors are frequently given less weight in mega-science, and scientific performance itself may be impacted in the light of construction funds available. Project success factors may differ from national success factors such as industrial return or national prestige. The SKA is no exception to this observation.³⁸

Success factors were not explicitly defined for the SKA in the 2006–2012 period. In hindsight, there was a straightforward meta-goal at the outset—to resolve the existential issues noted in the Introduction to this chapter and to ensure the project was well on the way to successful completion by the end of the period. To re-iterate, these were: (i) recognition of the SKA as a high priority project in “roadmaps” generated by the wider astronomy community as well as recognition by funding agencies and governments, (ii) establish a long-term governance structure centred on a legal entity to create a stable environment for the project,³⁹ (iii) select the best site to host the telescope, and (iv) deliver the PrepSKA outcomes of a costed telescope design and a signature-ready document to start construction.

All these goals were achieved with the exception of the PrepSKA outcomes defined in iv) above which were adjusted to match project progress and became the preparation of an implementation plan and business plan for the Pre-Construction Phase. Nevertheless, widely accepted PrepSKA progress on the overall design of the SKA telescopes enabled a baseline design to be defined in 2013,⁴⁰ which has broadly held into the construction phase.

Ultimately, success for the SKA was seen as the fulfilment of the common vision held by partners at all levels in the project to build the world’s largest radio telescope. At the time of writing, the first major step towards achieving this vision, SKA Phase 1 construction, is well on its way.

11.3.7 Project Resilience

Project resilience is defined formally as the building of inherent robustness during project shaping and was shown by Crosby’s study (Crosby, 2012c) to raise the chances of mega-project success. Resilience appeared to be strengthened through key practical factors such as the early setting of project mission and success definitions and clear and consistent structures and processes for reporting and decision-making. For SKA the shared common vision from the outset in 1993 was to complete the construction of the world’s largest radio telescope. This was not only the ultimate measure of success but also a major factor in project resilience that carried the project and its community through the inevitable ups and downs of an undertaking of this scale.⁴¹

Also important for project resilience was the feeling in the SKA scientific community, throughout the 2006–2012 period, that the funding agencies and governments in most countries were broadly supportive of the SKA endeavour (see Chap. 4, Box 4.7). The global adoption of the project generated resilience and was a key factor when striving for a treaty level organisation. This resilience was key to the project remaining coherent and able to adapt in the wake of the US decision not to continue its participation in the project in 2010,⁴² as well as at the time of the telescope site decision in 2012.⁴³

11.3.8 Motivations for Participating in a Project like the SKA

The primary motivation for a project like the SKA comes from the scientists and engineers who create and develop the concept because they want to have the opportunity to do ground-breaking research and make new discoveries with innovative instrumentation. National and regional level motivations follow on from this primary drive. In the case of SKA, the scientist motivations were not uniform at personal and institute level. Many sub-disciplines exist and there are many directions in which to take innovative engineering. The scale of a mega-science project such as the SKA requires international funding and involvement of multiple scientific groups. Different emphases on what constituted the most important science almost inevitably drove the project scope beyond a “simple” experiment to solve a particular scientific question—detecting neutral hydrogen in the distant parts of the universe—and into the realm of an observatory with an instrumental capability to accommodate many different types of investigations.⁴⁴ Perhaps this would not have occurred had there been a dominant partner in the SKA, but that was not the case for a project “born global” .

Institutes and university astronomy departments with a long history of radio astronomy research as well as telescope design and construction were natural partners for a project like the SKA. The prospect of playing a leading role in initiating and then designing and building a global telescope was a strong motivation for the senior radio astronomers in charge of institutes, observatories and university departments around the world,⁴⁵ in particular in Australia, Canada, China, India, the Netherlands, and the USA, and somewhat later in the UK. Elsewhere, there was more interest in the scientific opportunities provided by the SKA than in the design aspects of the telescope itself.

There was an additional motivation at scientist level in China, South Africa and New Zealand. The SKA was seen as an entry point into state-of-the-art radio astronomy and membership of a global community for themselves and their institutes.

The national and regional motivations for the SKA beyond the excellent science case were covered in Chaps. 3 and 4.⁴⁶ Chief among them were the national prestige accruing from international recognition of the country/region playing a significant role in a global science project, developing a new research infrastructure, accruing knowledge and industrial return on investment, maintaining a competitive position with respect to other observatories/institutes, and developing science and engineering talent in the country via education and training opportunities provided by the SKA to build human capacity and stimulate technological innovation. For all countries involved, both large and small, there was an underlying argument for participation, the mutual benefit of membership in an international partnership that has the prospect of expanding the national base of state-of-the-art astronomical techniques and technological approaches to complex problems. In addition, international partnerships like the SKA can enhance global and transcultural collaboration in communicating advances in knowledge to the general public.

There were arguments specific to each country as well. For Australia and the Netherlands, maintaining their positions as leaders in the historical development of world class radio astronomy despite being “small” countries in terms of population was important. For Australia and South Africa their excellent sites for hosting the SKA telescope combined with far better access to the astronomically-rich southern hemisphere sky compared to a northern hemisphere site were drivers of the strong political support received in both countries. In the post-apartheid era, South Africa prioritised astronomy as a means of attracting young people into science and engineering and providing them with exciting and challenging projects, particularly the expensive, multi-national science infrastructure projects. Having such a project located in South Africa was seen as creating a centre of science and engineering that could stimulate technology in local industry and science and technology in universities.

For most of the bigger countries involved in the SKA, a different mix of motivational factors came into play. In Europe as a whole and in the individual countries, the science had to be seen to be excellent by the wider astronomy community. However, an additional motivation for the European Commission was that the SKA was seen as a potential enhancement of the European Research Area,⁴⁷ a single, borderless market for research, innovation and technology across the European Union, able to increase Europe’s impact and prestige on the global stage. In more developed countries the sociological and educational impacts are relatively less important, but the economic payoff from industrial contracts is interesting to all. In contrast, for US funding agencies like the National Science Foundation, national prestige and industrial return were not a direct concern; the economic value of industrial involvement is assumed to flow naturally without making this explicit. Funds are provided to high priority science projects as judged by peers in the science community. Philosophical labels for the SKA like “born global” while attractive to the radio astronomy community in the US, were not relevant to the US funders.

11.4 Reflections on Specific Issues

11.4.1 Project Politics and Funding, 2006–2012

It was important for the project politics and funding of the SKA as a whole that the project leaders in the candidate hosts for the telescope site, Australia and South Africa, could access high government levels not easily available to their counterparts in other countries and enable governmental support early in the project’s life. This high-level access provided a significant impetus to the SKA in the other countries making it visible in a way that might not have been otherwise possible.⁴⁸

11.4.2 Key Science Projects: Who Requires them?

It seems a requirement for aspiring large science projects that Key Science Projects (KSPs) are identified. But who drives that requirement and where does it come from—governments, funding agencies, the astronomy community? And do KSPs generated by committee consensus actually constrain the scientific ambition of the large projects? Is sufficient recognition given to the historical fact that, in astronomy, most discoveries made by new telescopes are in fields that were not included in the project proposal at the time of funding so are not going to be covered by the KSPs?

Governments: as noted above in Sect. 11.3.8, governments are mostly interested in global mega-projects for their impact on the economy, human capacity building, public perceptions, international visibility and science diplomacy.⁴⁹ Governments will prefer simple descriptions of the scientific value, while at the same time wanting there to be potential for big and unanticipated discoveries.

Funding Agencies: are the interface between the government and the scientific communities they support. They need KSPs as evidence of focus in the project and ability to keep the project scope under control. KSPs also simplify the communication process with governments. They may have to advise the government on the relative value of competing projects and provide feedback to the scientific community on how best to promote their projects to the government, industry and public.

Astronomy community: Being able to set out KSPs that answer key questions in the field is seen by project leaders as being a way to get scientific community support which, in turn, is seen as demonstrating maturity and an essential element in the funding process. Unified community support is also needed to prevent fragmentation and competition between competing projects in the same sub-discipline. Occasionally two separate fields, e.g., optical and radio astronomy, can work together with a common set of KSPs to enhance prospects for funding, as happened in Australia in the 2000s. In this case, astronomy was competing for funds with other disciplines such as medicine and biology and astronomy leaders in Australia did not want the astronomy submission fragmented.

Scientists more deeply involved in the project will understand that the KSPs are only a subset of all the science drivers. However, the engineering community strongly prefer a well-focused set of KSPs which can be translated to design specifications as part of a system engineering approach.

Finally, KSPs are a core part of project promotion to the wider community. They help make a project like the SKA easily recognisable to young scientists and to the general public. In this sense, outreach has political value and can influence governments if the public is perceived to be interested and supportive.⁵⁰

To answer the questions posed in the first paragraph of this section: (i) it is clear that KSPs are required by the astronomy community, funding agencies, and governments, for different but complementary reasons; (ii) it is arguable for valid project profile and engineering reasons that KSPs do constrain the scientific vision of the SKA, and (iii) in view of the past history of unforeseen major discoveries resulting from telescopes with the potential for discovery, there is general recognition⁵¹ of the exploration of the unknown as a valid scientific driver for a telescope. However, the translation of this recognition into concrete plans of action for design teams is not straightforward (see discussion in Sect. 6.2.2.8).

11.4.3 Technology Innovation and the SKA

From the earliest days of the project,⁵² it was understood that the concept of a square kilometre of collecting area at a reasonable cost would require an innovative design solution. The cost/m² for different designs was considered a key metric since the cost of the antennas themselves would be the largest contributor to total construction costs. International meetings of the world’s top radio astronomers and engineers generated many innovative proposals to build the SKA. Historical precedent in the VLA and the Australia Telescope Compact Array (ATCA) had shown that some innovative changes had been successfully implemented throughout the design and construction phases. In both cases this involved substantial risks⁵³ and was instrumental in creating the landmark instruments they became.

Three stages of innovation in antenna design can be discerned for the SKA: (1) 1990–2005, (2) 2006–2010, and (3) 2011–2021. Stage 1 included dense aperture arrays with all-sky electronic multi-beaming capability, Luneberg lenses to enable simultaneous all-sky multi-beaming at higher frequencies, long focal length reflectors with airborne focus (Large Aperture Radio-telescope, LAR), the deformable surface reflector implemented in FAST in China, cylindrical paraboloids (later implemented as CHIME⁵⁴), focal plane arrays⁵⁵ (also called phased array feeds, PAFs) to enlarge the field of view for a given collecting area, large number—small diameter (LNSD) dish arrays with relatively cheap elements, and active dipole arrays at low frequencies (later implemented in LOFAR and the Murchison Wide-field Array, MWA) (see Sect. 6.2.1.2, Table 6.2).

Stage 2 began in November 2005 when the International SKA Steering Committee (ISSC) decided (see Sect. 7.3.7.1) that an array of large diameter dishes would not provide the desired image quality for the SKA. This automatically meant that the LAR and FAST-like concepts were no longer in the competition to be part of the SKA design. Following the first encounter with the funding agencies, at Heathrow Airport in June 2005, an antenna technology down-select was carried out to demonstrate project maturity to the funding agencies. In the process, Luneberg lenses were also eliminated and a reference design including the Large N-small D (LNSD) array, mid-frequency “dense” aperture arrays, and wide-band, low-frequency sparse aperture arrays based on dipoles took centre stage for the next 5 years. The LNSD array component included both wide-band single-pixel feeds and PAFs at the dish focus, the so-called dishes + “smart” feeds.

Stage 3 can be described as the phase when the SKA design confronted reality (see Chap. 6), starting with the Concept Design Review in early-2010 and made manifest with the Project Execution Plan (PEP) later in the year. The project perceived an opportunity for continued funding in the next decade as long as a workable plan for the first 10% phase of the SKA was generated by early-2011. This was the timescale needed if the SKA were to complete PrepSKA in December 2011 with the ingredients in place for the pre-construction phase including the establishment of a legal entity.

In the PEP, the project took what can be described as a risk-averse approach using known technologies—the LNSD dish array with narrow band-width single pixel feeds and low frequency dipole aperture arrays. It also assigned the innovative higher-risk solutions—wide-band single pixel feed, phased array feed and dense aperture array concepts—to a new Advanced Instrumentation Program (AIP). At the time when the PEP decisions were made neither the PAFs nor the wideband single pixel feeds had competitive noise performance. However, they would now, at the time of writing, be competitive due to further technology developments, but that could not have been known in 2011. It is interesting to note that it was recognised in 2007⁵⁶ that discoveries of transient radio sources would be better facilitated by optimising a different metric for the telescope design which had a stronger dependence on field of view (see Sect. 6.4.7). However due to the complexity of implementing this metric, it was ignored in technology decisions. Since 2012, most of the discoveries of new transients have been made with SKA precursors and pathfinders with a large field of view. In contrast, the dense aperture arrays did not receive the required continuing funding in the post-2011 period and research on this approach slowly petered out. At the time of writing, it is unclear whether any of the AIP concepts will reach fulfilment in the final design for SKA Phase 2 despite there being an Observatory Development Program in place in the SKA Observatory.

Was this decision correct in retrospect? There is no clear answer. As the 2010s decade proceeded, PAFs were implemented successfully on the ASKAP antennas in Australia which themselves included an innovative three-axis feed rotation arrangement at the focus to enable simple wide-field imaging and excellent polarisation calibration (see Sect. 6.4.4.1). Single pixel, wide-band feeds have been implemented on a small number of telescopes, but the wide bandwidth presents a difficult signal processing challenge. A further consideration is that in 2011 there was a window of opportunity for pre-construction funding, as well as engagement and enthusiasm from the governments and funding agencies to carry the project forward. These factors could easily have disappeared or dissipated if there was perceived to be a substantial technology risk remaining.

In summary, the desire for innovation in the SKA was slowly replaced with increasing emphasis on conventional solutions which reduced risk and had predictable time scales and costs. The adoption of such solutions appears inevitable in global big science projects since risks cannot be shared easily over an international community, and the large committees needed to manage such facilities are necessarily risk averse.⁵⁷

Reflections on technology development in large radio astronomy projects

In 2009, Rick Fisher (National Radio Astronomy Observatory, USA) and colleagues⁵⁸ submitted a white paper to the US Decadal Review (ASTRO2010) Committee on Large Instrument Development for Radio Astronomy. The abstract, quoted below, noted that the desire for an all-purpose solution and continued R&D into the construction phase led to higher risks, cost overruns, schedule delays, and project de-scoping, and made a plea not to put all the radio astronomy R&D “eggs” in one basket. Without referring to the SKA directly, they also made several other relevant observations for a project with the ambitions of the SKA.

This white paper offers cautionary observations about the planning and development of new, large radio astronomy instruments. Complexity is a strong cost driver so every effort should be made to assign differing science requirements to different instruments and probably different sites. The appeal of shared resources is generally not realized in practice and can often be counterproductive. Instrument optimization is much more difficult with longer lists of requirements, and the development process is longer and less efficient. More complex instruments are necessarily further behind the technology state-of-the-art because of longer development times. Including technology R&D in the construction phase of projects is a growing trend that leads to higher risks, cost overruns, schedule delays, and project de-scoping. There are no technology breakthroughs just over the horizon that will suddenly bring down the cost of collecting area. Advances come largely through careful attention to detail in the adoption of new technology provided by industry and the commercial market. Radio astronomy instrumentation has a very bright future, but a vigorous long-term R&D program not tied directly to specific projects needs to be restored, fostered, and preserved.

In a talk given by Ekers in connection with the award of the Grote Reber Prize at the URSI General Assembly in Beijing in 2012, he reflected on how changing technologies had enabled continuing innovation in radio astronomy and how to accommodate the differing timescales of change in both science and technology.

The time scale for science topics to evolve, and even go in and out of fashion can be much less than the lifetimes of major physical elements of the telescopes themselves such as antenna structures and site layout and infrastructure (typically 50 years). Hence it is no surprise that telescopes are not known for the science they were originally built to study, but instead for new unanticipated discoveries which, for a new instrument, emerge on much shorter time scales (see also Chaps. 1 and 5). Flexibility to upgrade may be more important than desired functionality at the outset to solve particular scientific questions⁵⁹ (Wilkinson et al., 2004), (Kellermann & Bouton, 2023). Due to the influence of rapid technology development, particularly digital technology which progresses on a faster time scale (Moore’s Law) than software developments, it is necessary to continually innovate and take risks to push the boundaries of instrument performance. For large projects, however, this has to be balanced with the reality noted by Fisher and colleagues that more complex instruments with longer lists of requirements and are necessarily further behind the technology state-of-the-art because of longer development times. A facet of the SKA design approach was to produce a constrained maximisation of discovery space by retaining flexibility wherever possible (see Sect. 6.1). But we note that instruments designed to be general purpose are not necessarily more complex. It is possible to design simple robust systems which are sufficiently flexible to enable innovative upgrades. In radio astronomy this has generally been true of the large parabolic dishes.

In Chap. 1 (Sects. 1.2, 1.3 and hba.skao.int/SKASUP1-1), we noted that Livingston curve analyses of high energy physics accelerator performance and radio astronomy sensitivity as a function of time showed that there had been an exponential increase in performance for 60 years but that is now flattening out. There are probably two reasons for this: i) no fundamentally new technology has yet emerged⁶⁰ which can maintain the exponential envelope; and ii) when international collaboration is required to provide the financial and skilled human resources for a mega-project, an inevitable delay is introduced. For global science mega-projects there will be limitations set by government priorities in each participating country on the total funding provided and in the rate the funding is made available. As noted in Sect. 1.5, global science mega-projects are particularly complex to deliver, especially in cases where no single partner is dominant as is the case for SKA. This complexity increases as some power of the number of different stakeholders. Delays would be inevitable even if fundamentally new technology was to be implemented. Our conclusion is that for large multi-national projects, it is not feasible to maintain rapid technological innovation as is the case with simpler organisational structures.⁶¹

SKA Phase 1 is under construction. The coming years could provide an opportunity for changes in technology to emerge that enable new parameter space in sensitivity and survey speed to be explored and show the way for a future expansion towards SKA Phase 2, the full SKA originally envisaged. See further discussion in Sect. 11.4.6.

11.4.4 Timing of the Engagement with the Funding Agencies

There was disagreement in the ISSC in 2005 about the timing of the first approach to the funding agencies in the SKA countries (see Sect. 3.4.1). While some were not inclined to expose the project to outside political influences before the major decisions on technology and site had been made on scientific grounds, the majority decided to take up an invitation to provide a position paper on the SKA ahead of these decisions.

The funding agencies had other issues on their minds. In Europe, with other pressures on funding including ALMA and the era of extremely large optical telescopes coming along, it was difficult to see how the SKA was going to be afforded. According to Richard Wade (Science and Technology Facilities Council, UK), co-chair of the group of funding agencies discussing the large telescope projects in 2005, the main problem they saw with the SKA at that time was how they could make it happen. In retrospect in 2019,⁶² Wade’s view overall was:

You can never engage with the funding agencies too early. And [with] something as important as site selection, if you don’t engage with the funding agencies then you might as well pack up and go away.

The SKA project engaged with the funding agencies at the opportune time. It was ready for external scrutiny, building on 12 years of global collaboration in which a case for ground-breaking science had been generated, significant progress on the telescope design had been made, a well-established governance structure was in place, and an agreed formal site selection process was in progress.

11.4.5 Impact of Site Short-Listing vs Definitive Decision in 2006

The ISSC began the process of selecting a single site for the SKA in 2001.⁶³ Its approach was straightforward and followed the example of ALMA: find the best site to optimise the science, free of political interference, and sort out other issues including politics and funding later. With the goal of starting SKA construction in 2010 in mind,⁶⁴ it was also thought essential to complete site selection by 2005 or 2006 to allow the project to focus on the telescope design knowing where it would be located and taking any design requirements imposed by the site location and infrastructure into account. The selection process was well advanced at the time of the ISSC’s first encounter with the funding agencies at Heathrow Airport in June 2005, with multiple responses to a call for proposals expected at the end of 2005 and an outright decision on one site planned for mid-2006.

The funding agencies had other ideas about the process⁶⁵ and preferred an approach that created a short-list of two or more sites meeting agreed minimum requirements. The final down-select would include other factors such as the willingness of the potential hosts to make significant contributions called host country premiums. Creating a competition via the down-select process carried with it the prospect of higher host nation contributions to the project than might otherwise have been expected. Although not articulated at the time, this approach was consistent with ameliorating the funding agencies’ concerns about how to fund the SKA in the crowded project environment then in prospect in 2006 (see Sect. 11.4.4 above). Support of the funders⁶⁶ was needed for both site selection approaches, the difference being the timing of that involvement.

In retrospect, following the funders’ advice to include a short-list stage was the best decision for the ISSC to take despite the (unanticipated) 6 years it took before the final site decision was made. As made clear at the Heathrow Airport meeting, the project as a whole was not sufficiently well supported by funders around the world at that time and, without that broad support, an outright selection of the site may not have been the best way to proceed. More important than persisting with outright selection of the site was for the ISSC to make sure the SKA was included on national and regional (European) roadmaps of future projects so that funding agencies and governments around the world had a formal basis for engagement with the SKA and, with that, a basis to proceed with a site decision.

It is interesting to speculate whether the post-Heathrow ISSC proposal in August 2005 to rank all four sites followed by the “softer” approach of “round-table discussion” would have led to a different outcome for the site selection than the hard competition via the short-list approach adopted. 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 developments was achieved. 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, but also would have been less inclined to accept that a compromise was necessary; we shall never know.

As far as the final site selection was concerned, the competition in its later stages did trigger some unhelpful animosity between countries that had had good relations previously, as related in Sect. 8.6.3. One further 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 decision in Australia to make the post-2006 internal deliberations on their site proposals confidential. This prevented the normal process of community consultation and comment. This was not the case in South Africa.

Consequences of an outright site decision in 2006

Had the definitive site decision been made in 2006, it is most likely Australia—New Zealand would have been selected ahead of Southern Africa.⁶⁷ The former was seen at the time by the ISSC and the International SKA Site Advisory Committee as having the better credentials to host the SKA⁶⁸ although both were acceptable sites.

It is interesting to reflect on what might have happened to the project in that case. The political impetus and profile of the project would have been much lower. Negotiating a host country premium would have been off the table. Without the high political profile and lobbying associated with the competition, the likelihood of substantial funding for the project may have been jeopardised.⁶⁹ There would have been strong arguments in favour of locating the project headquarters in Australia in addition to the telescope. This may have led to the perception of SKA being more of a national rather than global project, possibly resulting in less interest from other potential partners in making major funding contributions. A predominantly Australian SKA with some European involvement was one likely scenario. Were the site decision to have gone to Australia in 2006, the development of radio astronomy in South Africa may have been less vigorous despite MeerKAT. In addition, the intangible benefits for science and training/education in Africa of having a high profile global scientific infrastructure on the continent would have been foregone.

Finally, it is interesting to note that despite their being a central issue in the funders’ minds in 2006, host country premiums played no part in the site decision outcome in 2012. No negotiations took place on premiums due, no doubt, to their being regarded as a step too far in an already fraught site decision process in which the priority was to hold the project together. However, as time went on, it was recognised that, during discussions on overall contribution levels, there were other assets and past investment at the telescope host sites which, when made available, would constitute an additional financial contribution. This was not an active part of the main discussion on the contribution.⁷⁰

Was the dual-site outcome in 2012 inevitable?

In retrospect, the answer to this question is yes, from the moment the SKA Site Advisory Committee (SSAC) submitted its recommendation to the SKA Organisation in early-2012.

The outcome was not expected despite having been discussed briefly at several meetings of the SSEC and Agencies SKA Group/Founding Board as well as by the Science Working Group⁷¹ in the previous 4 years and was included as an option, in somewhat ambiguous terms, in the SSAC mandate. This stated that analysis and evaluation of the site proposals from Australia/New Zealand and Southern Africa was to be “open to a variety of site selection solutions, if the data and information provided to the SSAC support them”.⁷² However, as we describe in Sect. 8.6.1, the SSAC did not consider any alternative solutions for the SKA project in the belief it was charged to make a firm recommendation on which site to choose from the two proposals, not to work out a grand compromise acceptable to all parties.⁷³

Four factors underpinned the dual-site decision made by the Members of the SKAO Organisation. (1) The SSAC had determined that the SKA could be sited in either Australia/New Zealand or in Southern Africa. It was their detailed analysis of the selection factors that “favoured” Southern Africa. Australia had been expected to win decisively but that was clearly not the case. Neither was Southern Africa’s case so strong that it was seen as the decisive winner; (2) The Site Options Working Group (SOWG)⁷⁴ established by the SKAO Board of Directors had concluded that a dual-site implementation with the two sites hosting different technologies operating at different frequencies, was capable of delivering the SKA Phase 2 science case. There was no scientific, technical, significant cost or operational disadvantage compared with a single-site implementation for doing so; (3) The SOWG further concluded that for SKA Phase 1, distinct advantages arose from incorporating the MeerKAT and ASKAP precursors and related infrastructure into the SKA in terms of increased scientific capability and the availability to the SKA project of an estimated total investment in excess of €300 M in the precursors; and (4) The SKAO Board of Directors and Members were conscious of the overriding concern to keep the project together and maximise its financial viability in the longer term through continuation of the current Organisation membership, and ensuring its global character remained attractive to future members.⁷⁵

One might argue that South Africa had more to lose from the dual-site decision than Australia in view of the original SSAC recommendation. However, indications are that the recommendation came as a surprise to South Africa as well as Australia and that the dual-site decision was eventually an acceptable outcome to South Africa (see Sect. 8.6.3.3). It is interesting to note that the South African SKA leadership recognised early on that a split site was a possibility,⁷⁶ even though it was not regarded as a good idea.⁷⁷ At that time, it was important to avoid an “also-ran” outcome from the site short-listing in 2006; their aim was to be one of the short-listed sites. It is interesting to note that, at the time of writing, both sites are relieved that they only have one telescope technology, the mid-frequency or low-frequency array, assigned to them. This has avoided the difficulties, unforeseen at the time, related to having two quite different sets of construction issues to solve at the same time on the same site. It is now clear that the dual-site outcome can be defined as a classic moment of project peripety when the future trajectory suddenly seems clear and possible.

11.4.6 The Influence of the Precursors and Pathfinders on the SKA

The SKA-specific Precursors and Pathfinders—ASKAP, MeerKAT, MWA, LOFAR and FAST—were designed and built for a variety of reasons: to demonstrate technologies of interest to the main SKA project, provide a fallback in case the site decision did not go their way, or to provide a means of maintaining and growing the radio astronomy community nationally and globally while the SKA was in its long design and construction phase. National prestige considerations were important in all cases.

In Australia, ASKAP was originally a 10-element array demonstrating the phased array feeds technology in an array of telescopes and the potential of the remote, RFI-quiet site in Western Australia. It grew into a 36-element array when additional money was found from government to provide innovative observing opportunities for the local astronomy community and a state-of-the-art telescope as a fallback position in case the SKA was not built in Australia.

In South Africa, the MeerKAT project was more than an engineering prototype. It proved South Africa could assemble a team of highly qualified engineers, astronomers and commercial enterprises to build a state-of-the-art radio telescope that would, at the same time, act as a beacon for young scientists from the African continent. A key aspect of the approach was also to build up relationships in the global radio astronomy community and benefit from external expertise in designing, building and operating MeerKAT. South Africa was keen to show it was much more than just a developing-world host country for an international facility, receiving access to the telescope or financial contributions to a development program for astronomy in that country in return for a telescope site. Continental, national and institutional pride was involved. It is probably fair to say that the site fallback argument carried more weight there than in Australia.

In The Netherlands, the development of LOFAR was driven by George Miley’s assessment of the state of radio astronomy in the country in 1997.⁷⁸ He argued that maintaining and growing the radio astronomy community by building a telescope that explored new parameter space and offered new science opportunities on the relatively short term was a better strategy for the long-term health of the community than contributing all the Dutch resources to the main SKA project.

After entering the SKA collaboration with a very innovative big dish concept, China remained enthusiastic to explore this option themselves even though the down-select had excluded the big dish options. It is interesting to note that there are plans to expand FAST to the multiple element array first proposed as KARST (Kilometre-square Area Radio Synthesis Telescope) in the 1990s. This would approach the original sensitivity conceived for the full SKA.

The SKA precursor instruments in both Australia and South Africa were much larger than necessary for the SKA from an engineering demonstrator perspective for the reasons outlined above. However, the larger than necessary scale has had benefits. The resulting telescopes are powerful instruments on the world stage and have shown themselves capable of ground-breaking research and hence attractive to the community. Their size also means that their instrumentation and software systems provide a more useful guide to future design directions for the SKA while enabling innovative design changes on shorter time scales and at lower costs than possible during the SKA construction project.

Would the SKA have made faster progress if the precursors on the sites had been restricted to engineering prototypes? Yes, engineering resources would have been freed up in principle, particularly as the timescales for completing the precursors and pathfinders were considerably underestimated. However, even if all the national resources had been put at the disposal of the SKA in its design and pre-construction phases rather than towards developing the precursors and pathfinders into working state-of-the-art telescopes, it is hard to quantify how much faster overall SKA progress would have been. Moreover, it is unlikely there would have been the same build-up of the user community that has occurred as MeerKAT, ASKAP, MWA, LOFAR, and FAST have come into operation.

There is little doubt that the top-level hardware design of a project the size of SKA Phase 1 had to be frozen at a relatively early stage from which point onwards further innovation was no longer possible without delaying the project. In contrast, the SKA software design was much less well developed for a substantial fraction of the Pre-Construction Phase and is now strongly influenced by principles developed by the precursors. It may be that precursor and pathfinder upgrades and enhancements will be a major segment of the planned Observatory Development Program (ODP) coordinated by SKA Observatory and lead to innovations that are implemented for SKA Phase 2 while the role of SKA Phase 1 will be as the workhorse for regular astronomy. One area of innovation could be exploiting Artificial Intelligence in future software development at the individual institutes involved in the SKA Regional Centre Network (SRCNet) which will eventually handle the SKA data products. Perhaps the SRCNet can serve as an example for a future network of instrumentation innovation centres. And perhaps, totally unexpected developments such as a quantum formulation of imaging theory in radio astronomy may impact the way we carry out our science in some fundamental way.

A deeper analysis of the influence of the precursors and pathfinders is beyond the scope of this book but would certainly be worthwhile.

11.5 General Observations from the SKA Experience

• Science mega-projects need to incorporate the ambitions of both the science community that drive the science goals and the engineering community that provide the underlying practical reality. Ideally, they will include some key (rare) individuals who live in both communities.

• Although a grassroots initiative may be the starting point of a science mega-project, as was the case for the SKA, spokespersons with established reputations are very important, particularly in the early years. The SKA had several of these people including Ron Ekers, Govind Swarup, Harvey Butcher and Ken Kellermann.

• Although the initial aspirations expressed by an international cross-section of scientists are essential, at some stage the profile of the project must grab the attention of government institutions and relevant global forums. In SKA’s case, government buy-in had quite different motivations in Australia, Europe and South Africa. When and how this buy-in is accomplished is a matter of tactics, strategy, and luck.

• To enable the achievement of transformational science goals, initial key objectives must include efforts to overcome traditional technical approaches either by innovation or by smart adoption of newly available technologies. Even if these goals are not ultimately achieved, the interest created by well targeted efforts will generate excitement and interest from a broad community (and governments).

• It is likely that initial schedules and budget estimates will be optimistic.

• As the project proceeds, it will be necessary to make irrevocable design decisions, so that the project as a whole can maintain technical credibility and a believable timescale. This is a very sensitive stage in any project in which there are multiple potential technical approaches, and addressing this issue requires a nuanced decision-making process.

• The SKA has demonstrated that a global mega-project can succeed even without a single dominant national or international entity providing leadership and majority funding. In fact, a widely supported mega-project provides resilience against withdrawal of a single partner.

• As a global project, the SKA was seen as an entry point for participation in a state-of-the-art project by scientists from countries that did not have long-established facilities in the field. This lent support from a wider international base and from governments seeking a means to increase their country’s exposure to leading science, especially the prospect of joining an international treaty organisation at a reasonable cost.

• New science mega-projects are likely to require a new site. The SKA’s experience in selecting the telescope site provides many lessons, but the option of a dual-site, dual nation solution may be unique to the SKA. It was also an unexpected option resulting from the requirement to use two technologies to cover the SKA frequency range. There is no doubt that competition among the partners to host a mega-project is a means to raise its profile.

• Project management techniques which are typically applicable to projects with a single sponsor are unlikely to be adequate for a global mega-project without a dominant partner because textbook project management methods cannot fully allow for the inherent political considerations and sometimes unquantifiable risk. Funding agencies must be willing to accept substantial risk contingencies in cost and timescale in return for the advantages of participation in a global project.

• The SKA experience shows that with sufficient vision, tenacity, creativity and technical ingenuity, the many barriers to implementing a global science project can all be successfully overcome.

11.6 The Future

In 2019, at the meeting on the history of the SKA shortly after the signing ceremony for the Convention leading to the establishment of the SKA Observatory as an Inter-Governmental Organisation, Martin Gallagher⁷⁹(formerly an ASG member from Australia) raised the question of whether the SKA project was likely to move towards an ESO-like Observatory for radio astronomy. His arguments were that, initially, the SKA story was one of building a giant telescope, the largest in the world, to carry out ground-breaking research. Since then, partly as a consequence of the dual-site decision, the SKA has evolved to a concept with multiple sites, multiple instruments, a separate headquarters,⁸⁰ and perhaps the prospect of a longer-term life as a scientific enterprise with an even broader scope of research than originally conceived. An Observatory of this size would be able to influence a scientific discipline and build capacity in the science, technology, engineering and mathematics (STEM) areas over many decades.

Events have gone in this direction with the ratification of the SKA Convention in early-2021 and start of Phase 1 construction in mid-2021. The SKA project has already created a rich legacy for radio astronomy with the precursors and pathfinders. In the process it has regenerated a vibrant radio astronomy community eager to harvest the scientific results and change our view of the universe, as well as to prepare for the second phase of the telescope to fulfil the original vision set out decades before. And perhaps even loftier goals?