Written Evidence Submitted by the NERC SENSE EO Centre for Doctoral Training (CDT)

(SPA0013)

Dr Anna E. Hogg (1), Dr Ruth Amey (1), Prof Edward Mitchard (2), Prof Christine Gommenginger (3), Dr Andrew Fleming (4), Dr Leif Denby (1)

 

1. SENSE, University of Leeds

2. SENSE, University of Edinburgh

3. SENSE, National Oceanography Centre (NOC)

4. SENSE, British Antarctic Survey (BAS)

 

 

NERC SENSE Background

The Centre for Satellite Data in Environmental Science (SENSE) is a Centre for Doctoral Training (CDT) funded in 2019 by the Natural Environment Research Council (NERC), UK Space Agency (UKSA), with co-funding from Universities of Edinburgh and Leeds. SENSE is training 50 PhD students recruited over 3-years of intake, to tackle cross-disciplinary environmental problems by applying state-of-the-art data analytics to the deluge of satellite data collected each day. SENSE students are supervised by a consortium of world-leading UK scientists (~150 EO experts in our institutes), with space industry engaged from the start (> 40 CASE and industrial partners). All students are hosted at the University of Edinburgh, the University of Edinburgh, British Antarctic Survey (BAS) and the National Oceanography Centre (NOC), and complete an 8-week long comprehensive training programme in Earth Observation and Artificial Intelligence methods in the first year of their PhD. 

 

What should be the aims and focus of a new UK Space Strategy, including considerations of skills and diversity?

Skills: In 2021 the UK government identified Space as one of the UK’s fastest growing sectors, trebling in size since 2010, employing 42,000 people, and generating income of £14.8 billion each year. However, the 2021 UKSA Space skills survey showed that growth of the sector is limited by a severe skills shortage which risks missed scientific insights, slow development of emerging technology and an inability to scale up and grow commercial companies in the UK. SENSE is addressing this problem by i) recruiting 50 PhD students on EO and advanced computing projects; ii) bridging the gap between academia and industry; and iii) improving the diversity of our scientific community by developing PhD recruitment guidelines in line with EDI best practice. As currently funded, SENSE will conduct its 3rd and final round of PhD studentship recruitment in late 2021, leaving a significant training gap in an area of growing national priority in subsequent years.

Recommendation: Fund a programme of ‘Centres for Doctoral Training’ to train 100 to 150 PhD students in Space skills, including satellite data processing, advanced computer techniques (AI and ML), Space instrument design, launch and in orbit engineering, telecommunications. Over the next 5 to 10 years this will guarantee a pipeline of talent into the UK national Space industry (civil, defence and commercial) where we expect to see significant industry growth. Any UKRI investment will be leveraged 100% by match funding from university and industry investment.

 

Diversity: BAME (black, Asian, and minority ethnic) students are underrepresented in all STEM subjects. SENSE was the first UK PhD centre to set out guidelines for PhD recruitment in line with EDI best practice and have shared this widely with the PhD centre community. From SENSE’s EDI discussions, under-represented groups reported that they are more likely to undertake advanced postgraduate training (i.e. MSc or PhD level) if there is a clear link to industry with a viable career path leading from it.

Recommendation: These recruitment EDI best practices should be discussed cross-research council and stakeholders to further refine and improve them, and these should be translated to all hiring decisions across career stages. Taking positive action such as ringfencing studentships, presents legal and logistical challenges, but PhD enrichment awards issued in collaboration with either charitable organisations focussing on this issue or companies that want to improve diversity of their hires, could be a useful tool to make advanced training in the Space sector even more attractive. Investment in summer internship programmes for undergraduate and other training pathways (e.g. apprenticeships) is essential.

Recommendation: Create closer links between industry and academia in the Space sector to encourage minority groups to undertake higher education programmes such as PhDs. The UK should participate in international graduate trainee programmes in technical Space agencies, by investing in a UK stream for the ESA National Trainee Programme (France, Germany, Spain and others already subscribe), and through setting up international trainee exchange programmes with agencies beyond Europe (E.g. Canada, Norway, UAE, Japan etc).

 

What should be the aims and focus of a new UK Space Strategy, including considerations of research funding, investment and economic growth?

The UK is a world-leader in the development of new Earth Observation instrumentation and missions, the production and exploitation of large satellite datasets for Earth science and climate research, and the transformation of science into public good and new commercial services. This capability depends on a creative and dynamic research community with the skills and capacity to respond quickly and innovate across multiple disciplines (environmental science, engineering, data analytics, social sciences, etc).

The long lead-in timescales (typically 10 years from idea to mission) and cross-disciplinary nature of space are not well served by today’s research funding structures in the UK. In contrast to other leading space nations (USA, France, Germany, Japan), national funding for EO in the UK lacks coherence and strategic direction, and is characterised by fragmented, short-term, opportunistic investments that are poorly coordinated across research councils. The apparent wealth of funding opportunities for space research is often undermined by complex restrictions (eligibility, IP, overheads, PV contribution, etc) that put up barriers across different disciplines and sectors of the research community in universities, NGOs, industry, government bodies, etc. We know that currently UK government investments in Space return approx. £6 for every £1 invested.

Recommendation: To consolidate and simplify Earth Observation research funding structures, managed under one umbrella like the UK Space Agency, equipped with the technical knowledge to coordinate nationally, identify strategic priorities and represent UK interests internationally. The UK should invest in national Space technology development and research programmes at the same level it invests in international programmes such as ESA. This will maximise the UKs substantial technical capability, and will enable our national Space sector (civil, defence and commercial) to grow at its maximum potential rate.

 

What should be the aims and focus of a new UK Space Strategy, including considerations of civil and defence partnerships?

Defence operations and international security interests rely heavily on access to space systems, including space weather, GNSS encrypted services, protection of space assets, satellite communications and Earth observation. Space activities are frequently dual use in nature, largely using the same technology and often the same infrastructure to meet both civil and defence goals. Considerations of developing UK independent space capabilities (e.g. in GNSS and EO) should extend beyond the space assets to the expertise required to exploit these satellite datasets for defence and civilian purposes.

Centres for Doctoral Training, such as the SENSE CDT, train the next generation of Earth observation experts with the necessary satellite data processing and advanced data analytics to widen the UK skills base. These skills are equally applicable to civil and defence applications, and training opportunities should be aligned with the required skills and consider the strategic priorities of both uses.

The defence sector will need ‘ears’ in Space, provided by secure quantum telecommunications, and ‘eyes’, provided by the all-weather day-night imaging capability of Synthetic Aperture Radar (SAR) instruments. The early career and PhD space research community will be responsible for delivering innovation and national capability in this technology area in the next 10 years.

Recommendation: To ensure there are skilled workers in the civil and defence sectors, the joint priorities of the defence and civil space strategies must be clearly communicated to support a coherent plan for training of students at all levels. At the advanced skills level this can be delivered through dedicated, Space-specific ‘Centres for Doctoral Training’, encouraging partnerships and sharing of skills across both sectors. This will ensure that the pathway of talent between civil and defence sectors is joined up, with training provided by universities in the priority areas for the defence sector. An exceptionally small investment (~£1mill) in an equipment pool for defence related technology and instruments (e.g in situ and drone mounted SAR’s) that can be used for (civil) PhD research and training, will allow our next generation of UK Space leaders to have hands on experience with lower TRL Space technology, allowing them to innovate and develop the next generation of future satellites.

 

What should be the aims and focus of a new UK Space Strategy, including considerations of international considerations and partnerships?

International partnerships can be built through skills training. SENSE has over 40 national and international partners, including ESA (European Space Agency), NASA (U.S. National Aeronautics and Space Administration), DLR (German Aerospace Centre), DMI (Danish Meteorological Institute).

Recommendation: Facilitate student exchanges with other Space related countries through international partnerships. Facilitate the exchange of skills training from UK universities to other countries, possibly through virtual talent networks.

 

What should be the aims and focus of a new UK Space Strategy, including considerations of technology?

Both existing and emerging technologies in machine learning (specifically neural networks) have a proven track-record in data-driven scientific discovery, but these techniques have not been applied comprehensively to study the Earth system from Space. The UK needs to fund research that will enable Earth sciences domain specialists to use and develop novel machine learning network architectures on Earth observation data. Specifically, techniques such as Physics-Informed Neural Networks where models and model parameters are learnt directly from data (enabled by automatic differentiation), Transformers which can form context-aware predictions and work non-gridded data, and Bayesian Neural Networks which produce well-formed uncertainties on predictions are all recent advances that are already producing breakthroughs in computer vision, natural language processing, protein-folding, epidemic modelling, etc.

The real strength in applying machine learning to the diverse earth observation data available is that these techniques can discover connections between observations that humans may not even conceive to look for. This is already vastly accelerating discovery in other scientific fields.

 

What needs to be done to ensure the UK has appropriate, resilient and future-proofed Space and satellite infrastructure for applications, including earth observation including climate change?

The UK needs to invest in national and international projects that fund research and development in all parts of the Space chain: from building and launching satellites, to collecting and storing data and use of new methods such as machine learning and artificial intelligence to exploit large data driven by EO.

Research programmes need to be supported by digital backbone, which will secure the national data archive of research, climate and operational quality of raw satellite data, and also storage space for storage of new high data. This must include CPUs and GPUs to rapidly process the vast amount of data collected by satellites. The commercial value of the raw satellite datasets and the data processing algorithms and associated training datasets is clear given investments from data driven companies such as amazon and google. UK access to these services is not guaranteed, and while lower cost now, costs are already increasing and long-term continuity and security is not guaranteed. The UK will require a defence data backbone as investment in Space increases, therefore there is an opportunity to create a complimentary civil component that will advance the rate of algorithm development and low TRL idea testing.

Recommendation: The UK must continue to invest in national High-Performance Computer (HPC) facilities (HPC’s) with sufficient CPU and GPU capability. Processing power must be matched with large storage space and national satellite data archives that contain all historical and current mission data. This requires sufficiently large workspace storage areas which should be funded nationally and free to use for researchers. This digital backbone could be joint with the UK’s defence digital infrastructure, which is likely to also hold many of the same civil, raw Space datasets.

 

(June 2021)