You are here

Semantically Enhanced Model Experiment Evaluation Process (SeMEEP) within the Atmospheric Chemistry Community

NeSC Research Seminar Series
Peter Dew (School of Computing, Leeds University)

This talk reports on an in-depth study of the application of e-Science within a chemical kinematics modelling community which is part of a much larger atmospheric science community. The origin of this study is the CombeChem project. This captured semantic provenance of experimental data from source to data preservation. The data was generated from physical experiments (rather than modeling) conducted by chemists. Jeremy Frey states that one of the challenges is to encourage modelling scientists to annotate a think, try & track process. Our longer term goal is study community support for the whole of the scientific Model-Experiment-Evaluate process providing the community with evaluated semantically-enabled quality data to use in their experiments. This is a collaborative project with Southampton University (Jeremy and Nick Gibbons).

The research has largely been undertaken by Chris Martin, an NERC funded research student. Importantly Chris is embedded in the Leeds Atmospheric Chemical Laboratory ensuring a sound understanding of the modellng process. The main aspect of the research has been to study a user-centric (i.e. not a system oriented) design and evaluation of an Electronic Laboratory Notebook (ELN). The ELN records semantic provenance of the modelling process. This is difficult because the data provenance that modellers provide is much less structured than is the case for physical chemical experiments. In the later scientists are obliged to keep more detailed records to meet safety regulations.

The second part of the talk will focus on the community evaluation of new experimental results by a chemical kinematics community. We refer to this as Semantically Enhanced Model Experiment Evaluation Process (SeMEEP). This involves a group of experts meeting to evaluate experimental and modelling data to determine the data values (in our case chemical rate constants) that can subsequently be used by the community in their experiments. Our longer term is goal is to study e-Science support for all scientists that are capturing both experimental and modeling results in an ELN. Potentially this means that the evaluators have much richer information set in which to undertake their evaluation. Finally the implications to other modelling communities will be discussed.

Date and time: 
Friday, 6 March, 2009 - 11:30
60 minutes