Contributed by Ilias Pechlivanidis and Marie-Amélie Boucher.
Marie-Amélie Boucher has been a HEPEXer since 2007, with a strong involvement in HEPEX-related events, i.e. organizing committee member of the HEPEX 2016 workshop in Québec, and lecturer of the hydrological forecasting course for the Early Career Scientists (ECS) at EGU2017 in Vienna and IAHS 2019 in Montréal, and others. On a scientific level, she is one of the members that brings knowledge to the community on Nordic hydrological processes and forecasting practices in Canada.
Marie-Amélie is an Associate Professor at the Université de Sherbrooke with research interests in, among others, data assimilation, risk and decision-making under uncertainty, and socio-economic and environmental value of hydrological forecasts. These are scientific topics of high interest to the HEPEX members. Particularly for the HEPEX community, Marie-Amélie is the newest Co-Chair, and I therefore took the opportunity to ask her a few questions relevant to her experience.
Ilias Pechlivanidis (IP): Lecturing, researching, user engagement, disseminating, funding… you very well balance them and in moderation; as Kleovoulos o Lindios in the 6th century B.C. said “Παν μέτρον άριστον”. What would be your advice to the Early Career Scientists, pursuing a similar academic path as yours?
Marie-Amélie Boucher (MAB):
Well first, thank you very much! I am extremely tempted to answer only with one of my favorite quotes:
« On dit que la toile, selon son étendue, sa forme, sa solidité, ses leurres, sa beauté, au tout dernier moment tisse l’araignée qui lui est nécéssaire. Les oeuvres inventent l’auteur qu’il leur faut et construisent la biographie qui convient ». (Quignard, 2006)
« It is said that the canvas, according to its extent, its shape, its solidity, its lures, its beauty, at the very last moment weaves the spider which is necessary for it. The works invent the author they need and build the appropriate biography. » (Guignard, 2006)
I am happy with my path, but I was overall very lucky from the start and still am. I suspect that what I did best was to be open to all opportunities and to say « yes » when others said « no ».
One advice would be… Not to listen to advice too much. Do what feels right for you at every moment. Being in academia is great and suits me well, but I suspect many other paths (scientific or not) are fun too!
Also, think carefully before using the word « impossible ». Do not invent limits that don’t really exist. Talk to strangers. Try to get over your fears. Fail and get up. Say yes if it looks fun, you’ll figure it out along the way.
IP: Quite a lot of academic research is not implemented in operational services, due to lack of understanding of user needs. Your research has been linked to industry, whilst your methods are refined to be better communicated to the users. How straightforward is it really to tailor research to user requirements and background, and how easily did users engage with your research?
It is indeed not straightforward to tailor research to user requirements, and I’m not sure one should. Academics are not consulting engineers. It has to be a discussion, and people have to listen to each other carefully.
I find that governmental institutions are especially fun to work with. They have broad mandates and cover large territories. I think the best situation is when a potential research partner comes to you with a very general problem (e.g. we want to modify part X of the forecasting system) but is then open to any possible solution.
For me, one thing that has been beneficial so far is to use independent funding (for instance my start-up fund from the university) to perform research on a topic of my own choosing, but using data and models from governmental or industrial partners and sending them updates and results regularly. It is a nice way to get acquainted with their system and it allows potential partners to gradually gain trust in you. In my case, it led to some partners later providing actual funding for further collaborative research. It can be quite a long shot, though.
IP: You have been interacting with students and early career scientists through a modern pedagogic approach, including open software and educational games, interactive presentations. Has this approach improved your collaboration with the different scientific groups? How did your interaction with students and young scientific groups contribute to your scientific ideas?
At the university, I teach only at the undergraduate level. So far they have taught me a lot in terms of communication skills, patience and empathy. I do not discuss my research a lot with undergrads, though. In my opinion, our education system places them in a situation of constant stress that somewhat limits the type of interactions you can have with them.
I try to use games and interactive presentations to make things more fun, to try to reduce that stress. For that you have to be creative, which is perhaps the closest link with research. It is a fine line though, and I have made mistakes. Complete active learning has been found more efficient (in terms of learning outcome), but makes a lot of students feel insecure (Deslauriers et al., 2019).
Graduate students, however, contribute immensely to my scientific ideas, throughout the discussions we have together. When I was first hired at my previous university in 2011, I was feeling very lonely, with no one nearby to discuss ideas with. It made a huge difference when my first MSc student arrived!
Another group of people that I find very interesting are end-users of forecasts. Unfortunately, I still have very little direct contact with them. The governmental and industrial partners mentioned above are not really end-users for the most part. They produce operational forecasts, but they do not take decisions. End-users are the ones who can really help you understand the value of a forecast and provide insight to what is required to improve this value.
IP: Uncertainty is present in all steps of the forecasting chain. Uncertainty and its sources are something you have tried to identify and quantify. Where does most of the emphasis be given in Canadian river systems? Do you believe that despite the source and magnitude, the practices on how we communicate uncertain forecasts will result into better decision-making?
One important source of uncertainty here is related to observed precipitation, especially in the north, where we have very few ground stations. This is improving with the increasing use of satellites and merged products like CaPA (Mahfouf et al., 2007, Fortin et al., 2015), but it is still present.
Snow is a special case of the above. Snow water equivalent varies immensely even across small spatial scales and it is one of the most important variable in nordic hydrology.
Regarding communication, I would say that we are not there yet. «How should we communicate forecasts? » it is still very much an open question, especially for ensemble water level forecasts, which are so multidimensional. Also, I am not convinced that good communication would necessarily result in better decision-making. As I said earlier, I do not yet have many contacts with end-users. However, one thing that struck me in the few conversations that I had so far with them is that many seemed to trust expert judgement more than « formal », model-based, forecasts. I don’t know yet if this is really a wide spread belief, but if it is, we must figure out why. I don’t think it is only a communication issue.
IP: Thank you, Marie-Amélie, for this insightful interview. On behalf of all the HEPEX members, I will like to thank you for this contribution, and welcome you to the group of Co-Chairs. I am confident that your contribution will be very valuable.
Deslauriers L., McCarthy L.S., Miller K., Callaghan, K. And Greg Kestin (2019) Measuring actual learning versus feeling of learning in response to being actively engaged in the classroom, Proceedings of the National Academy of Sciences of the United States of America, 116 (39) 19251-19257
Fortin V., Roy G., N. Donaldson, and A. Mahidjiba, 2015: Assimilation of radar quantitative precipitation estimations in the Canadian Precipitation Analysis (CaPA). Journal of Hydrology, 531, 296–307, doi:10.1016/j.jhydrol.2015.08.003.
Mahfouf, J. F., B. Brasnett, and S. Gagnon, 2007: A Canadian Precipitation Analysis (CaPA) project: Description and preliminary results. Atmos.–Ocean, 45, 1–17, doi:10.3137/ ao.v450101.
Odry J., Boucher M-A, Lachance-Cloutier S., Turcotte R. And Cantet P. (2019) Estimating snow water equivalent from snow depth and climate data using artificial neural networks, Geophysical Research Abstracts, 21, EGU2019-4442
Quignard P. (2006) Villa Amalia, Gallimard, 304 pages