General Information

Co-organiser: Dave Stainforth

Public Resource Distributed Modelling (PRDM), and its in-house intranet equivalent, has the potential to provide vast computing resources for certain types of modelling experiments, thus enabling entirely new directions of scientific enquiry. The UK has taken the lead in international research and application of this technology and interest is growing worldwide. There is a window of opportunity to utilise this technology for novel research while providing opportunities to engage the public in a plethora of environmental modelling research activities.

PRDM provides exciting opportunities but also requires imaginative experimental design which may be different but complementary to traditional modelling approaches. For instance, large, complicated, memory intensive models such as mesoscale global circulation models can be used, but there may be trade-offs in terms of the timescales and the need to pre-define data analysis requirements.

Further information about this event and PRDM has been provided by Dave Stainforth.

Aims

The aim of this meeting is to explore the opportunities for PRDM experiments within the UK environmental modelling community. This is an opportunity to brainstorm the types of modelling experiments which researchers need/want to do; the starting point being what is necessary or desirable rather than what is currently perceived as computationally practical. Discussions will not therefore exclude the types of experiment more suited to supercomputing or GRID computing but will focus on distributed modelling which is currently the greatest source of computational capacity for the environmental sciences.

The design and implementation methods for PRDM will be presented with examples from existing projects. Brainstorming sessions will explore opportunities in the range of fields represented by participants, and discussions will address the barriers to participants’ “ideal” experiment and how they may be overcome. If substantial opportunities in a number of fields can be identified, as anticipated by the organisers, then the discussion will progress to funding opportunities to support initial implementation in the range of sub-disciplines identified.

Provisional programme

Substantial time will be devoted to discussion but there will also be a number of presentations to bring participants up to speed on distributed modelling: its potential and practicalities. These will cover:

  1. An overview of volunteer computing projects (also known as Public Resource Distributed Computing) with a particular focus on climateprediction.net.
  2. PRDM software: BOINC.
  3. Models for understanding processes / models as predictive tools.
    4. The value of sensitivity/uncertainty analyses.
  4. Public involvement and engagement. Public involvement .vs. internal resources.
  5. Security and reliability.

Venue and Accommodation

The workshop will take place in Exeter (more information to follow) on the 3rd and 4th July.

Registration

This event has been cancelled .

Contact and further information

For information about registration, and all other types of queries, please contact Therese Williams in the first instance.