A –Streams of Thought– contribution by Wouter Knoben (WK)

Prof. Thorsten Wagener is a hydrologist, currently head of the Water and Environmental Engineering research group at the University of Bristol, UK. He has received various prestigious awards and was recently a Humboldt fellow at the University of Potsdam. I had the pleasure of interviewing Thorsten in his office, where we spent a good hour and half going over his experiences.

WK: Can you tell us a bit about your background and formal education? How were you in school?

I studied Civil Engineering in Siegen (BSc), Delft (MSc), and finally Imperial College London (PhD), but I never really wanted to be a civil engineer when I was in high school. My dream was to become an architect. Unfortunately, I only did the absolute minimum in high school, so my grades where average at best. Entry requirements for architecture were too high for my grades but civil engineering seemed like architecture, and at least I could get into that degree. I had never heard of hydrology then, and programs like the hydrology course in Freiburg (which produced many excellent hydrologists like Jan Seibert, Markus Weiler, Kerstin Stahl, Doerthe Tetzlaff…) would never have taken me anyway due to their entrance requirements. I was also quite bad at computer programming at the time. In fact, I did so poorly during a test that I ended up with a negative score (points were subtracted for mistakes from an initial score of 100) and only got a passing grade if I promised not to take the course the following year. Quite funny really, because now all I do is use computers.

WK: So, what did inspire you to pursue in hydrology then?

During my undergraduate degree in Siegen, I had a professor who worked in Africa and the Middle East. This was quite practical work, and he offered his students final year projects in Ethiopia. That seemed an exciting idea, so I spent 5 months there. I really liked the idea of combining engineering with helping people, so I abandoned my idea to switch from civil engineering to architecture. Instead, at the end of my undergrad degree, I went to Delft for a Master’s in Civil Engineering with a strong focus on hydrology.

WK: Was being an academic an obvious next step for you and did you envision yourself as a professor at any point? If not, which events led to where you are now?

Like I said, my high school start wasn’t particularly promising. After high school, the change for me was that I spent a year in the air force – Germany’s compulsory military service. I got a desk job and was bored out of my mind. I then realized that I should figure out what I’m doing with my life. And so I was extremely motivated when I went back to university and to my own surprise that went really well. Completing my undergrad degree motivated me to go a bit further for a Master’s in Delft. That didn’t start quite so well though. I was there for 6 months and failed my first set of exams. I left for 1 month because I was so frustrated and wanted to quit. My dad and best friend convinced me to at least try until summer and decide after that. I went back, passed my exams by the skin of my teeth, and decided to carry on after all. The second year went a lot better and my final thesis project went really well, which got me excited about the idea of doing research. When I started my PhD in London, again things didn’t start so well because it was a rather different environment from the casual Dutch style. This took me some time to get used to, but again it worked out. So, at no point did I really see myself as a professor in Bristol or anywhere else for that matter. It worked out well in the end, but it really went step by step. I had to build up new confidence again at every step, and never really saw myself all the way at the top. I do think though that overcoming these difficult stretches was a really important growth period for me.

WK: Now you are where you are, what have been the highlights of your career so far?

I have a lot of things I enjoy looking back on. Publishing my PhD as a book, graduating my first PhD student, seeing my former students become academics themselves, etc. I am not sure about “the” highlight but one moment that stuck with me happened during my PhD. I had an idea about how to use a time-varying analysis of parameters to detect missing dynamics in hydrologic models, but I did not seem to be able to make it work. I had read an excellent paper by Bruce Beck from the 80s where he tried to do something similar using a Kalman Filter (Beck 1987), but I wanted to use a Monte Carlo approach so that I could distinguish informative and non-informative periods while making fewer assumptions. At the time of my PhD, running the analysis on a single PC meant that I could go for a coffee every time I was producing a graph, and I had to reboot the PC in between analyses because it would get stuck otherwise, so it was a frustrating experience. Then one day, it suddenly worked. I got the graph I was looking for, and it looked like what I was hoping for. I ran to the next room to get my friend Neil to show him because I was so excited. That was the first Eureka moment in my career!

WK: About publishing your PhD research as a book, how did that happen?

When I finished my PhD, I thought my thesis was quite a nice story that showed clearly where we were with modelling gauged and ungauged catchments. So, we presented the idea to Imperial College Press and they said “yeah that looks great”, and I got to spend the first year of my post-doc mostly working on that book.

WK: In a similar vein, one of your top-cited papers is called “Catchment Classification and hydrologic Similarity” (Wagener et al., 2007). It outlines a way forward for hydrologists to build an overarching system in which catchments can be classified based on their similarities. It’s also one of the top cited articles of the journal (Geography Compass) it is in over the last three years. What’s the story of this paper?

Well, I like organisation and order as you can tell from looking at my office [WK: pretty clean, lots of neatly sorted books, desk empty apart from a screen and a water bottle].     I’m also a big fan of Ven Te Chow whose textbook “Applied Hydrology” is an amazing example of how you can organise hydrology theoretically. Often hydrology seems like a hotchpotch of empirical equations, but he created a consistent approach to derive equations for hydrologic processes from first principles. Our paper was a (or at least our) first approach to do the same with a view towards a main entity we study in hydrology – the catchment. What should a generic framework look like in which we can organise catchments from a hydrological viewpoint? We started with the idea of basic hydrological behavior – the catchment functions. We decided that we need to be able to quantify the effect of these functions and hydrologic signatures seemed a good way to do that. Then we tried to envision where these might take hydrology and why it is so important. I think the fact that this paper is so widely cited – even though it is published in an obscure journal that is not listed in Web of Science – is an indication that we have at least partially succeeded. Retrospectively, we all regret that we submitted it to Geography Compass though, which was a brand new and promising journal at the time, but unfortunately never took off.

WK: What new research projects are you currently involved in?

There are two main themes: the first is how do we do hydrologic analysis and modelling at larger scales? Modelling, hydrologic similarity, models of everywhere, those kinds of things. One starting point was the work with Andreas Hartman. When he came to join me for his postdoc, we said, let’s do large-scale karst modelling but do this with (1) parsimonious models, and (2) by trying to use the few bits of data that do exist efficiently. At larger scales we’re quite limited by the amount of data available and you need to make use of what you have. In the current projects we take a similar approach by integrating data mining to understand complex model behavior, while using perceptual models to develop educated guesses of how our models should behave even in the absence of direct observations of the variables of interest.

The other area I’m increasingly interested in is to understand the relevance of hydrology to other fields. For example, hydrologic services and disservices: what are the services derived from having water in the right place, and how can we better understand and predict water-borne diseases. There is lots of work in this area, but it doesn’t really use hydrologic knowledge or does so very poorly because hydrologists are not often involved. This started through the interdisciplinary Cabot Institute at Bristol University. I found a few other people from outside hydrology (and most importantly a very good and very motivated PhD student) interested in the overlap between the fields of hydrology, biology and infectious diseases. We spend two years just meeting every two weeks or so, trying to figure out where to take this. Previously, models to assess the disease we chose (fasciolosis) were empirical regression type models. We’ve now introduced a mechanistic model, which links the disease to variables that control the spread of the disease (soil moisture in this case) by explicitly representing processes, which lets us look in much higher detail at the spread of the disease and potential intervention methods (Beltrame et al., 2018). It also lets us work in non-stationary systems, where empirical models (based on correlations observed in the past) lose a lot of their usefulness. There are tremendous step changes we can make by bringing hydrology into other fields of study.

WK: Does this relate to what you see as the current needs of hydrology?

Absolutely! One problem I see at the moment is that we have not been very good at demonstrating the relevance of our work. We are in danger of being surpassed by meteorologists, climatologists and others who use hydrology, but do not see it as central; while, in fact, hydrology is central. Hydrology translates climate signals into most variables needed to assess the Sustainable Development Goals, hydrology is key to understanding the spreading of infectious diseases such as Schistosomiasis (a waterborne disease contacted from certain parasites that might live in fresh water), or to understand the changing patterns of flooding. In part I think this is because hydrology gets more complicated the closer you look, which suggests that we need to understand our environment in ever greater detail. While hydrology research focused on small scales is relevant, it should not prevent us from going larger scale – though with quantification of uncertainties. We need to make sure that we demonstrate why hydrology is useful. Using hydrologic understanding to facilitate better understanding in other fields is one way to do this.

WK: What are other essential steps for the future we should take?

Apart from potentially having new ways to observe, what will certainly happen is a step change in our ability to build models and analyse data. Things will happen much quicker. When Google is building floods models, we are really entering another era. This is not bad of course, but in this context, it is crucial not to lose sight of the limitations of models. Currently, our understanding of how hydrology works seems to grow slower than the availability of data and data analysis methods. One of my main interests is finding out how to integrate the two and ensure that knowledge keeps up with data availability. Even in the field of data mining, scientists increasingly realise that domain knowledge plays a crucial role, and that without it we might only increase the problem of reproducibility (Hutton et al., 2016).

WK: What are the biggest challenges and opportunities for hydrologists in the next 10 years or so?

One big challenge is that we need to not lose track of the large societal questions. Hydrologic understanding was and is the foundation for our survival and for sustainable development. Without water, there is no energy, no food, no clothing. The spreading of many infectious diseases is closely coupled to the water cycle. We cannot understand most local climate change impacts without understanding hydrology. There are many opportunities for hydrologic knowledge to contribute to society.

However, hydrology is also complicated because the closer you look at our environment, the more complex it becomes. So, as hydrologists we have long focused on understanding this complexity. Now, we need to start tackling the big problems and clearly separate the work on technical details (regardless of whether this is related to models or measurements) and focus on big societal questions. We are very good on tackling the former; we have historically been rather poor in identifying the latter (including the role of hydrology in understanding them). People like Tom Gleeson or Mark Bierkens have shown that we – as hydrologists – can identify big scale problems and provide answers – though they might be more approximate than what we can say about specific catchments. Don’t get me wrong. I am not suggesting that we focus only on this, but that we simply have to do significantly more of this type of work. Here is where we need to be more creative: in identifying interesting and relevant problems and questions – as well as solutions.

WK: What’s an unusual place where you find inspiration for yourself or the work you do?

Football! Playing (in the past), watching and reading about it. [Points at bookshelf] that’s my football bookshelf. My current role, as Professor and Head of a research group that includes 9 academic staff and many more students and postdocs, is – at least in my head – a bit like being a football coach. Your team is a group of highly talented individuals who have all self-selected to be here as top people in their own right (like a professional football team). The role of the coach is to get them all to somehow work together or at least benefit from each other, without losing the chance to shine individually. So, I have a whole bunch of books written by or about football coaches on my shelf. Including “My Turn” by Johan Cruyff, “Quiet Leadership” by Carlo Ancelotti and “Leading” by Sir Alex Ferguson. Teams follow coaches because they believe that the direction the coach suggests will bring them success. There is a need to be convincing, genuine as well as caring – otherwise players will not believe that you want them to succeed, but that maybe you are only looking for short-term success for yourself. That’s more the Mourinho strategy, which always fails after a couple of years when players get disillusioned. I try not to do that here but try for everybody to benefit from the group.

WK: Could you share any insights on how you approach creativity? Do you think that creativity and success are correlated?

Creativity is a bit under-appreciated in research, I think. I like learning about how creativity is approached in other fields. There is an excellent book called “Creativity Inc.” by the former President of Pixar, Ed Catmull. He discusses how they spend decades trying to optimise their creative process. They decided that everybody should share ideas early, so that bad ideas would fail quickly and not waste time, and that they needed to create continuous opportunities for feedback and involvement of everybody. We, in science, often see creativity as a by-product when it really is the essence of research. We often see it as a gift that one has or does not have, rather than something that can be significantly improved and nurtured. I think that in research, success and creativity are closely related and that everybody can improve their ability to be creative. I think that the most influential hydrologists are also the most creative. People who I admire for their creativity include (but are of course not limited to) Hoshin Gupta, Keith Beven, Chris Duffy or Brian McGlynn. Very different personalities, all of them (apart from maybe Chris Duffy) not the best students during their undergraduate degrees, but very creative people who can think outside the box and in a different direction than the rest. I think the distinction between short-term and long-lasting scientific contributions is often due to differences in creativity, but I should stress that it is not just that. Others succeed through their energy, through their persistence or depth of knowledge.

WK: To conclude, have you got any advice for young hydrologists?

Don’t rush to submit a paper quickly. Write good papers that make significant contributions. It’s not the quicker way, but it tends to take time to produce high-quality work. Learn the basics very well: reading papers, processes, statistics, computing, etc. Keep up to date with new literature by learning how to decide which papers are just another confirmation of what we already know, and which ones represent real steps forward. Most importantly, take the time to think and share your thoughts with others in debate. Spend time debating your ideas with fellow students, your supervisor, etc. This helps tremendously in increasing the depth of your arguments. And learn how to communicate well (through text and figures). You won’t write influential papers if people have to work hard to figure out what you are saying.

About the author

Wouter Knoben is a PhD student at the University of Bristol (UK). He has been involved in AGU’s Hydrology Section Student Subcommittee (H3S) in 2017-2018, and in the Young Hydrologic Society between 2016 and 2019. Correspondence to w.j.m.knoben@bristol.ac.uk

References

Beck, M. B. (1987). Water quality modeling: A review of the analysis of uncertainty. Water Resources Research, 23(8), 1393-1442. https://doi.org/10.1029/WR023i008p01393

Beltrame, L., Dunne, T., Vineer, H. R., Walker, G. J., Morgan, E. R., Vickerman, P., McCann, C. M., Williams, D. J. L. and Wagener, T. (2018). A mechanistic hydro-epidemiological model of liver fluke risk. Journal of the Royal Society Interface, 15(145). https://doi.org/10.1098/rsif.2018.0072

Hutton, C., Wagener, T., Freer, J., Han, D., Duffy, C. and Arheimer, B. (2016). Most computational hydrology is not reproducible, so is it really science? Water Resources Research, 52(10), 7548-7555. https://doi.org/10.1002/2016WR019285

Wagener, T., Sivapalan, M., Troch, P. and Woods, R. (2007). Catchment Classification and Hydrologic Similarity. Geography Compass, 1(4), 901–931, https://doi.org/10.1111/j.1749-8198.2007.00039.x