Jeremy Bennett, Dr.rer.nat
  • Home
  • About
  • Research
  • Blog
  • Links
  • Contact

Good scientific practice

11/19/2015

0 Comments

 
For I have known them all already, known them all:   
Have known the evenings, mornings, afternoons,        
I have measured out my life with coffee spoons;   
I know the voices dying with a dying fall   
Beneath the music from a farther room.   
So how should I presume?*

I recently attended a seminar entitled 'Good Scientific Practice' presented by Prof. Alex Weber, a cell biologist at the University of Tübingen. The aim of the seminar was to encourage doctoral students here to ascribe to the aforementioned GSP in their work. I enjoyed the Prof. Weber's perspectives on how research should proceed, although as a researcher working solely in modelling I was in a minority in the audience.

The crux of good scientific practice is honesty - with respect to our planning, procedures, data interpretation and storage. Prof. Weber emphasised the importance of keeping thorough notes in a lab book and provided good examples of how this can be effective in conducting research. A lab book acts as a record of what work has been completed, and what are the outcomes. It is effectively a journal of your research.

My days are spent, for better or worse, sitting at a computer writing code to run numerical models of porous media. What does a lab book look like for someone who isn't in a lab, who's work is conducted predominantly in programming environments? The answer I came up with in my head is the Version Control System that I was introduced to early on in my doctoral studies.

A Version Control System (VCS) is something that you use to track progress of computer codes and programs. Every time you make a change or add a feature to your code you commit the changes to the VCS; the changes are stored, as well as notes that you make about what you did, and possibly why you did it. A VCS might be a stand-alone software, such as Mercurial which I use, and are often free for  non-commercial use. Using a VCS with something like BitBucket also makes it a lot easier to share coding projects with other collaborators.

Over the last few months as I have been developing codes for my work I have begun to embrace the 'commits' that I have made in my VCS - my commit comments have become more descriptive, and I can see the evolution of my project as I look back through my previous entries. I guess it is my electronic 'dear diary...' and in much the same way we might look back at our teenage diarising with a mixture of wistfulness and embarrassment do I look back on some of my efforts so far.

How are you do you measure out your research life? (The coffee spoons are a given...)

*The love song of J. Alfred Prufrock by T.S. Eliot. In Prufrock and Other Observations. From Poems. New York: A.A. Knopf, 1920; Bartleby.com, 2011. http://www.bartleby.com/198/1.html#46.
0 Comments

Integrated Hydrosystem Modelling conference 2015

4/21/2015

0 Comments

 
From 7-10 April the Center for Applied Geoscience at the University of Tübingen hosted an international conference on integrated hydrosystem modelling. A large number of prominent researchers attended and fourteen keynote lectures were given. Participants were encouraged by IRTG spokesperson Olaf A. Cirpka to ask questions during the sessions, in the manner of a Gordon Research Conference.

Five broad areas of environmental modelling research were addressed during the conference:
  • Integrated modelling philosophies
  • Reactive transport  &  biogeochemistry across scales
  • Geology-hydrology interactions
  • Modelling of land-surface processes
  • Uncertainty quantification in environmental modelling

The sessions from Jef Caers (Stanford), Richelle Allen-King (Buffalo) and Jasper Vrugt (UC Irvine) seemed the most relevant to my proposed course of study.  I found talks from Philippe Van Cappellen (Waterloo), Gabriel Katul (Duke) and Erkan Istanbulluoglu (Washington) to be engaging - they were able to explain the complexities in their work to an audience with a wide range of interests.
Two poster sessions during the conference allowed other participants to present their work, as well as promoted discussion between researchers.

The Gordon Conference-style of presentation allowed for some vigorous discussion, most notably between Jef Caers and Jasper Vrugt. To me it was clear that they bring very different perspectives and motivations to the field of environmental modelling, both of which are necessary.

Jef's role as  professor for energy resources engineering at Stanford includes assisting oil & gas producers to make decisions regarding the location of hydrocarbon reservoirs, and thus where new drilling will occur. He is interested in presenting information to people who may not be familiar with the mathematical tools he is using but need to make decisions based on his work. Lead times in the energy industry are short (around three months) and the end goal is ultimately economic - to produce enough resources to make a profit.

Jasper's work as assistant professor in civil and environmental engineering at UC Irvine focuses more on optimisation and the quantification of uncertainty in environmental modelling. At the IRTG he presented on how the modelling cycle could be improved, particularly the reformation of hypotheses following model-data analysis. It is my opinion that his focus on developing better modelling methods is much more 'fundamental' science than the applied work conducted by Jef, and neither approach can be neglected.

I'm hoping to follow this idea up with a post on the 'fundamental/applied' dichotomy in science that I am starting to discover as delve deeper into the rabbit hole.


0 Comments

    Author

    I am a hydrologist interested in environmental modelling as well as the application of water science in the 'real world'.

    Categories

    All
    Coding
    Conference
    General Science
    Groundwater
    IRTG
    Modelling
    Multiple-point Statistics
    New Zealand
    Sedimentology
    Tübingen
    Water

Powered by Create your own unique website with customizable templates.