Revision of Panel Discussion 2: Future Consortia from Fri, 2013-09-27 16:36

The second panel discussion conference was formed of Vince Smith, Guy Cochrane, Alex Hardisty, Renzo Kottmann, Johannes Petersen, Koos Biesmeijer and chaired by Wouter Los.

Chair's intro was his talk before the coffee break. The Commission have not yet revealed details of the first H2020 calls yet, but they will to cover only the first two years of a 7 year programme, so discussions at this and similar meetings may form part of future calls. The first calls are expected to be published on 11th December 2013 but the precise timing of the calls is not yet finalised. Calls from DG Research are expected to close in April 2014. Of particular interest to this community a call related to Virtual Research Environments (VREs, from DG Connect) will be supported with about €40M and close in September 2014. DG Connect will also issue a call to support, manage & analyse big data (including computing) with €50M support. Successful projects will need to focus on societal challenges, not on biodiversity per se. Proposals are expected to be interdisciplinary, IT acting as the connective conduit. There is a preference for supporting service-intensive environment, so business models and sustainability will be critical from day 1. The funding will be 100%, no co-funding. Figures for SME involvement will be averaged over all calls, so not every project will need to have high levels of SME involvement.

The Overall Goal

There was universal support for the goal, sine qua non, to deliver predictive modelling of the biosphere. This single goal can effectively unite those active in understanding the biosphere, in any capacity whatever, and is implicit in all that follows here. Significantly, this conference has a significant buzz and we need to capitalise on it to move the field forward in a unified, not fragmented way.

Ocean Sampling Day (OSD) is a good pilot, but how could it be extended into a global SD? It is not clear how we can separate data from infrastructure, or indeed whether we should. OSD began as a European idea and quickly spread to being global. The genomic observatories are an extension of this idea, also acting as de facto training. We're not discussing how we train the new generation in the skills of software and data management. No one can be expert in all these things, but they need to know how to interact successfully. Other global projects are emerging. As they do they are revealing the need for new infrastructure that is being built as we go along.

We, as a community, have delivered some remarkable products and assembled some large data sets in a variety of ways. It is time that we stopped being modest and started to make use of the data and tools we have created. Attempting to answer real questions will reveal those data that are missing and provide justification for the digitisation effort required to mobilise those data. Such attempts will also reveal the gap between what the infrastructures we've built can be delivered and what the user-base actually wants to do. It is highly likely that we will not have the appropriate infrastructures for large-scale data-handling and processing in a manner that is reasonably effective. How are we, for example, to link large data repositories, such as molecular, remote-sensing and meteorological, located in different centres, so that elements from each can form part of an analytical process?

Raising Awareness and Unifying the Community

The conference has recognised the importance of clarity of vision, with greater focus on coordinated end-goals. This does not mean that we have to agree on priorities, but that we agree that our individual priorities should link to the overall goal (predictive modelling) and products should be linkable. The diversity of those involved is striking, as is the variety of techniques in use, and only a few were represented by the delegates at the conference. In fact, most users of biodiversity information are not aware of this community, so reaching out to make connections is paramount.

It is common to underestimate the amount of effort that outreach requires. We operate at the cutting edge and its tremendously exciting, but we don't communicate it outside our own group. Nevertheless, given a clear vision, clear communication to a range of audiences becomes feasible.

Linking communities depends on knowing the people involved. Its easier for some than others. Networks (of people) are crucial for this type of interaction. How the data are structured is the central issue. It is important to establish how we answer the big questions of our time, such as sustainability, managing climate change, maintaining ecosystem services such as clean water, etc. These are foci, not a guide to organise the data, so the organising principle is surely to provide a service to make science all around us work better. It doesn't fit the traditional funding models, but it should.

We have many islands of capacity and we need some kind of glue to link them together. A significant amount of activity has been driven by H2020, including the community White Paper (Hardisty et al., 2013), the EU's Roadmap for biodiversity ( European Commission, 2013), the GBIC report (Hobern et al, 2013),geo-BON, and individual project reports, such as that from Creative-B ( Hardisty, 2012).

Community Identity

We need a representative body to link these islands. While there is some inertia from this meeting it should be used productively, but how such a body might be constituted remains to be worked out. It is significant that LifeWatch has not stepped up and taken this role.

We need to manage this system not simply try and sustain it. A better name to rally behind (the "Hubble telescope" for biodiversity, perhaps "bioscope") would be a powerful tool in forging community identity and the immediate priority of H2020 funding requires a focus on Europe.

The role of the private sector

The relationship with SMEs and other private sector (PS) players will be a big challenge and well outside our comfort zone. Almost everything we have achieved so far is reliant on Open Data or Open Source software development. Few PS players have business models that will integrate well with such openness, although some do exist, such as Vizzuality for example. Education of the PS is desirable, but expansion of the project goals to work with existing PS players and forming new PS organisations is also possible. Some things we now regard as functions of our own laboratories can become commodities and traded in the PS. Sequencing and some types of computing, for example.

PS organisations are often better placed to communicate with policymakers, including processing and packaging the data in a manner best suited to their requirements.

Sustainability is very important to lend stability and stability can lead to a market. Sustainability is going to be central to H2020 projects from the outset.

For many PS players our projects are pre-competitive. We need to keep an open mind and keep talking.

Potential projects

The Aichi targets require Governments to stop biodiversity loss, but what was it? What tools can the community provide to achieve this straightforward-sounding aspiration?Automated identification, perhaps? Museums are well placed to contribute to many such goals, as well as providing historical context, but the Museum's need guidance to prioritise mobilisation of their data.

Information systems have been put in place that are beginning to reveal microbial processes in the oceans. With such a system we can address questions that are beyond the capacity of any single laboratory. We can't anticipate the impact of climate change unless we understand how it works now. So projects will need to work out how to extend those models to understand the impact of larger organisms. We will also need tools to visualise the molecular pathways being revealed. Such tools will intrinsically require a significant amount of inference.

Workflows are another area that offer potential for extension across multiple domains (cf BioVeL) particularly connecting biodiversity scientists with information technologists.

Another project with strong integrative potential is ViBRANT, offering a standardised VRE that people are using now to mobilise a variety of small data sets. It should be remembered that more data exists on the hard drives of individual's computers than in the global aggregations and repositories.

The overall goal of predictive modelling of the biosphere should start with a global plan to explore what is needed to couple models. This will need cooperative partners from around the world. This would also involve creating a list of dependencies and identifying potential points of failure. The first steps would be a design study to build the human networks capable of realising the challenges. What would the models mean, at what scale and what data are required? This bears comparison with digital weather forecasting that has a 40 year history. What are we going to need for ecological prediction? We are certainly going to need a suite of new observational systems. GEO-BON aims to have a prototype of the "essential biodiversity variables" in 2 or 3 years and should form part of this planning.

Summary

3 key words: integration, cooperation, promotion.

Integration: we have a number of directions like the virtual lab, and we need to give attention to data other than species data and the tools that can utilise them.

Cooperation: if you want to model the whole biosphere you need data from over the whole world, not just Europe. The gap areas (Russia, China, Brazil, India) must be closed.

Promotion: Europe should cat as the leader and pull others into a system of mutual benefit. There are only two main categories of users: scientists and decision makers. The latter are really only CBD and IPBES, all others can absorb the same information structures.

Biodiversity, in the broad sense, is our life support system. It is absolutely essential and more important than almost everything else, but is typically taken for granted. The conference agreed that this conference should be repeated annually, but perhaps in the Spring to make scheduling less of a challenge in the busy Autumn conference season. We need to identify those who will carry this forward. The EU-BON project offered to nucleate the next conference.

Scratchpads developed and conceived by (alphabetical): Ed Baker, Katherine Bouton Alice Heaton Dimitris Koureas, Laurence Livermore, Dave Roberts, Simon Rycroft, Ben Scott, Vince Smith