AskProfs Number One - Modelling in Water

This blog post is article number one in the new AskProfs series being launched by the Students of the Water Institute Graduate Section (SWIGS), where we choose a theme, put together some questions, and see what University of Waterloo (UW) professors, researchers, and postdocs think. For the first one we have chosen modelling in water and posed eight questions to UW faculty. The topic was unilaterally selected by me, but that is the benefit of starting the series off. I promise that future topics will not just be about modelling :) .

The survey allowed respondents to answer as few or as many questions as they liked, and allowed them to remain anonymous if they wished. The survey encouraged satirical and humourous answers, and was also aimed to help young researchers in modelling-related work to bask in the wisdom of the more experienced UW faculty. So hopefully some Master's and PhD students will read this, but if nothing else I got a good chuckle (and a lot of agreement with the frustrations noted).

Without further preamble, here are the answers we received (in no particular order).

What advice would you give to new modellers?

Focus on integrated surface/subsurface flow and solute transport modelling. I started this field of modelling research almost 25 years ago. I also led the formation of a UW spin-off company (see about 5 years ago I I will argue that the Aquanty team of geoscientists, developers and R&D people are the strongest group of hydrosystem analysts amongst any university, national lab or industry on the planet. ~ E. A. Sudicky

Modelling is about logic and structure. Before writing code, draw a schematic flowchart of the model you want to develop. The flowchart should show all the logic steps, where what input is needed, and which calculations are performed and in what order. If your flowchart makes sense, your code will also, but not otherwise. ~ Philippe Van Cappellen

Realize you learn a great deal more by getting things wrong than when you get things right. Realize that a single model run means very little. Don't waste your time with models that take a month to run. Realize your model will rarely be as good as you want it to be. ~ Anonymous

When there is a difference between the hills and a DEM, the hills are probably right. ~ Ric Soulis

Strive to understand the underlying concepts/principles approximated by the math, including how initial and boundary conditions, and source terms influence the solution. ~ Anonymous

How did you get into modelling?

I enjoyed programming and wanted to do something in water resources. It's a natural fit. ~ Anonymous

I got tired of trying to do it all on a hand calculator. ~ Philippe Van Cappellen

Interest -- spent way too much time on engineering undergraduate coursework writing code to solve problems; carried this on to grad school and became an expert on spending too much time on interesting problems. ~ Anonymous

I thought I could model the world. I'm still working on the Mackenzie. Also, I pushed Zoolander down the stairs. ~ Ric Soulis

Strong interest in theory and the quest of understanding processes. ~ E. A. Sudicky


What is the best thing about being a modeller?

I don't need to worry about having other hobbies.  Also, Blue Steel. ~ Ric Soulis

Solving problems too complicated to do by hand; the potentially very cool graphical output of complicated problems. ~ Anonymous

Discovery of why our natural system functions and how we can assess anthropogenic stresses. ~ E. A. Sudicky

The satisfaction when things start to click in place or you fix a bug in the code; avoiding the logistics and costs of the field; math. ~ Anonymous

You can explore endlessly with a model. You are in charge. That's much harder in the lab or in the field. ~ Philippe Van Cappellen

What is the worst thing about being a modeller?

I don't need to worry about having other hobbies.  Also, turning left. ~ Ric Soulis

Listening to researchers who think they understand modelling, or blindly use models, with a lack of serious training. Parsimonious modelling is old school - we have spent decades developing sophisticated process-based models, and actually use them to solve real-world problems at a variety of scales. This is the 21st century. ~ E. A. Sudicky

The frustration of when things don't work after 200 different attempts to fix them. Bad software that doesn't do what it was supposed to. Numerical instabilities. ~ Anonymous

Becoming convinced that models, especially your own, are always the best way to analyze a problem; finding bugs/debugging; setting up large input files. ~ Anonymous

Hearing other people state they are not a modeller. We are all modellers, whether we use a computer code or not.  Modelling is about concepts, logical reasoning and integration of knowledge. As soon as you embark on unraveling how the world works (humans included) you start using models, whether they are quantitative or not. ~ Philippe Van Cappellen

What are the most important skills and/or qualities of a modeller?

Persistence, an ability to scrutinize the details, a willingness to see what's going on under the hood, and a bit of detective skills. Being able to come at the same problem from 6 different ways. An ability to back away and see the big picture even as you're awash in the details. ~ Anonymous

Always question the outcome of your model. ~ Philippe Van Cappellen

Ego.  Also, Blue Steel + turning left. ~ Ric Soulis

Knowing what you don't know; patience; coding; logical thinking. ~ Anonymous

Knowledge of math, physics, hydrologic/hydrogeologic processes, geochemistry, appreciation of field data and how it is collected, etc. ~ E. A. Sudicky

What is the greatest challenge in modelling?

Delineating the assumptions you make and verifying the logic of the model. ~ Philippe Van Cappellen

NaN. ~ Ric Soulis

Use intelligence, high-level training and experience. ~ E. A. Sudicky

Using them responsibly and appropriately with other methods, viewing them a only an approximate tool not reality, developing engineering judgement with complicated models. ~ Anonymous

For environmental modelling, realizing that (ever-present) tremendous uncertainty often squashes the effectiveness of the most elegant models. ~ Anonymous

If you could go back in time to your early career, would you go back into modelling?

Likely since the interest was there ... now not much time to model. ~ Anonymous

Yup. ~ Anonymous

As I said before, once you accept that rationality and logic are the best way to understand the world you are a modeller. So, yes, I don't see how I would not go back into modelling. ~ Philippe Van Cappellen

Is that a question? Of course I would. ~ Ric Soulis

Dumb question, of course I would. ~ E. A. Sudicky

Modeling or modelling?

100% replied 'modelling'

Modelling is Canadian spelling, modeling is US. Trust me. ~ E. A. Sudicky

Hope you enjoyed reading this blog post. If you have comments or suggestions for future topics, or want to submit your own blog post, please drop us a line