HS
Hydrological Sciences

Hydrological Sciences

YHS interview Martyn P. Clark: “rainfall-runoff modelling, per se, is dead”

In its “Hallway Conversations” series, the Young Hydrologic Society has recently published an interview with Martyn P. Clark, who is currently professor and the Associate Director of Centre for Hydrology and Canmore Coldwater Lab, at the University of Saskatchewan, Canada. The interview was conducted by Sina Khatami, a PhD student at the University of Melbourne. With their agreement, we reproduce below some short extracts of the interview. For the full interview, visit the YHS Blog (here).

Martyn Clark did his undergraduate degree and his Master degree at the University of Canterbury, in New Zealand, and was awarded a PhD degree by the University of Colorado in Boulder. After working back in New Zealand for a while, he came back to the US in 2010 to work at NCAR. You can check the interview he gave to HEPEX in 2016, while in Boulder (see here). In December 2018, he moved to University of Saskatchewan, where he is currently working with new challenges ahead.

Your research spans across a wide range of domains of hydrology, hydro-climatology and model development. How did you expand your knowledge and expertise so widely?

In the early days, it was more of a random walk. My interests evolved into different areas and I pursued opportunities where they were. I read a lot. Even when I was doing my master thesis, I read and read and read. So, I was able to get a fairly good understanding of the literature and identify what the major science questions are. Later on in my career, I’ve been much more strategic than tactical as I was in early stages of my career: thinking about what the big problems are that we want to solve, and how we can go for the funding opportunities out there that lead more towards this larger vision… more of a proactive approach, than a reactive one.

Over the past few years, you’ve become the Editor-in-Chief of WRR (see the EoS interview), moved from public sector (NCAR) to academia and from Colorado to the Canadian Rockies. Each of these decisions are big enough to be a challenge for a few years. So, first, how’re you holding up [I laugh]? And what motivated such major changes?

Well these were more sequential than simultaneous [we laugh]. So, let’s deal with them sequentially. I was asked to apply for the Editor-in-Chief position for WRR. They had a search committee together and they asked me if I would consider doing it. My initial response was no. Then I thought about it for a while. Two things had happened in that year. First, I was promoted to senior scientist at NCAR, which is the top level there. So, I didn’t have any opportunities for additional promotion. And also, I was elected as Fellow of AGU. So, I thought I have kind of established myself in my career and perhaps now is the time to give back to the community more. And there was this opportunity. I was weighing all of my commitments and then thinking about how I could push the field forward. And I thought, well… what good can I do? I thought if I publish, say, two fewer papers a year and be the WRR Editor-in-Chief instead, I can probably do more good and continue along my current trajectory. I was also keen for a new additional challenge.

NCAR is federally funded research and development centre and received a lot of its funding from government and through NSF (National Science Foundation in the USA). The decision to move to the University of Saskatchewan was in part because I wanted the broader challenges that comes with the university setting. And it was in part because of the funding that they already had in place with the Canadian government as part of the Global Water Futures (https://gwf.usask.ca/) programs. This really provided the opportunity to achieve a lot of my research ambitions that I’ve had for many years.

Reviewing and handling numerous papers as WRR Editor-in-Chief has provided you with a big picture of the research community. How is that is influencing your own approach to defining new questions, particularly for your new career line at University of Saskatchewan.

Yes. For my new career at the University of Saskatchewan at Canmore, a wonderful location by the way, we are building up the research program there (https://uofs-comphyd.github.io/). A lot of the research thrusts and the global water futures program are the things that I have been working on over the past twenty years anyway. It is dominated by two main application questions: (1) improving streamflow forecasting methods, and (2) improving assessments of impacts of climate change on water security. Those are the two applied questions that have guided my research on process understanding, model development, strengthening the link between algorithms and theories, etc. It is not as if I’m going to a new research area; I’m going into an area where I have had an extensive presence for a very long time. So, that part of it is not new. But the part that is forcing me to extend myself a little bit is that the funding available is more than an order of magnitude larger than what I ever had before. So, being able to think more strategically, like build up a large cadre of postdocs to answer these questions, or how to orchestrate a large research program — it is really exciting.

Looking back at your research career, what do you think your major breakthroughs are and why?

I think my major breakthrough is quite broad. But I can list some specific papers if you want. Developing a more structured approach to hydrological model development is something that I’ve been working on for many years. The first paper that I really published in that area was my FUSE paper, working with bucket style models. Then my most recent big modelling paper was my SUMMA paper (paper 1 & paper 2) [both are modelling frameworks that allow a user to analyse the impact of individual modelling decision; such as the choice of model structure, the choice of specific flux equations, and the choice of numerical method with which to solve the model equations].

How do you describe your research style? Or, what are the main elements for you when you’re impressed by a piece of research?

For me, personally, I’m really interested in making a step change in our understanding of modelling capabilities. So, most of the major papers that I’m proud of have had a gestation period of more than five years. And so, if you look at my publication history — it’s kind of interesting — I had no first-author research publications in the time period of 2011 to 2015, when I was developing SUMMA. And that can be a little bit dangerous [he laughs] for people at earlier stages of their career. I really wanted to make a major contribution in the way that we develop models. I was worried that a lot of our model development was somewhat ad hoc and we didn’t have the structure that we needed in order to really understand where and what model weaknesses are. I was worried that model evaluation wasn’t done in a controlled way and that we really needed a new framework in order to push forward in those areas.

What would you identify as the main gaps or big picture questions of hydrological sciences for the coming decades that you think early career scientists can pursue?

I think we really need to evolve towards a more interdisciplinary Earth System Science approach to modelling. For many years, hydrology has been rooted somewhat in what was called rainfall-runoff modelling. That term is not really applicable anymore, because we now are modelling a large number of complex interrelated processes in the terrestrial water cycle. So, multi-process modelling in an Earth System modelling context, not just focusing on the short-term fluxes but also the longer-term evolution of our systems. Understanding the evolution of soils in the catchment, understanding the evolution of vegetation in the catchment and understanding how those slowly varying processes feed back on to the higher frequency variability, which has typically been the domain of hydrologists.

And this goes back to the SUMMA paper that you mentioned?

Well that’s just a part of the bigger picture. SUMMA has a more complete representation of the terrestrial hydrological cycle than many hydrological models. But many models already have that level of complexity. SUMMA doesn’t even begin to get into the issues of bio-geochemistry, catchment co-evolution, etc., which are going to be really important. What SUMMA does is provides a structured template for process-based hydrological models which can be extended into the Earth system modelling framework. But it’s nowhere near complete enough of what we need moving toward. So, what I’m talking about is not something that we can do in the next couple of years but something that we need much more concerted effort over the timescales of several decades.

Are there any papers or books that you would like to recommend on this grand idea of expanding the spectrum of processes within current hydrological models towards Earth system modelling?

The first part of the SUMMA paper provides some beginning thoughts in that area but it doesn’t go as far as it needs to. We wrote a paper on improving the representation of hydrological processes on Earth System models. That’s really just beginning to scratch the surface as well. I think the paper that everybody should read is the one led by Ying Fan on providing the link between hillslope hydrology and Earth system modelling that provides lots of pointers in that direction. But it’s funny that you ask that. There’s something that I’ve been kind of stewing on for a while, which is to put together a coherent commentary paper that emphasizes that as a research direction that’s necessary.

You’ve pointed out many great things so far, is there any other advice you may have for young hydrologists?

I think I’ve covered a lot of it already. Be bold. Think about how you can really make substantial advances in the research frontier. Be strategic. You need the incremental progress. You need the intermediate scale products as you are conducting your research so that you can feed the beast [he smiles] and work effectively through the career track. But those intermediate scale products need to be conducted within the context of a larger scale vision. So, really think about defining that vision. Talk about that with your colleagues and keep refining that. And having an idea how your career contributions will really begin to make a difference.

Some guidance would be to think about three levels of strategic planning or technical planning in some respects: (1) what do you want to accomplish in your career? In terms of always keeping that and the longest timescale. (2) What’s the thing that you’re going to present at the next conference? Most people are thinking about those two or perhaps not giving as much attention to the vision aspects as they should. But the third that often gets neglected based on my interactions with people is (3) what are you going to do tomorrow, and the coming week? So, basically organising your activities on the shorter timescale, so that they are feeding the ambitions that you have on the longer timescales, I think is really important.

This might be a somewhat stupid question. Do you have any measures to evaluate a good PhD or postdoc? Like the number of their publications or good publications in a year, etc.

Yeah, this has been my problem. I don’t like the way that people are being judged in academics. There’s a saying that managers know how to count but they don’t know how to read [we both laugh]… In the sense that people are focused too much on outputs, like how many papers you published, than outcomes. I think that things are going to change. I wrote an editorial in WRR on the citation impact of hydrology journals. There I was talking about the need to shift away from quantitative assessments to more qualitative assessments to really begin to measure how people are making a difference in the community. For me that’s the major thing. So, if we get back to what would help people get a job, I can tell you what I’m looking for. Yeah, you need some papers to get on people’s radar screen. If you have finished your PhD and you don’t have any papers then that’s a red flag. But what you really need, in my mind, is to be known for something. That people look at you and say okay that person has done X, or that person has accomplished Y. So, the number of papers that you’ve written become less important. So, what I’m looking for is what have you done to make a difference in the community. And that’s what a lot of other people are beginning to look for more.

I’m curious to know more about this. How would this qualitative assessment process work, to assess the impact of a person on hydrological sciences or even the broader geosciences?

You should read the Declaration on Research Assessment (DORA), which I also referred to in our 2017 editorial paper. DORA comes up with a set of guidelines for funding agencies, universities, managers, etc. to show how they can move towards research assessment practices that are more fair. It has been picked up by a lot of different institutions and universities. A lot of it is there. It’s more just changing the structure of the research assessment. You know there’s not going to be one size fits all template that people can use, but structuring it in a way that emphasises the contributions rather than the specific papers. It takes more work, but we should value our colleagues and take the time to really make sure people’s efforts are directed in productive ways.

 

Note from the EGU HS Blog Editorial team: In the past EGU 2019 GA in Vienna, EGU organized a debate on “Rewards and recognition in science: what value should we place on contributions that cannot be easily measured”, where DORA was also debated, among others. 

Guest author Sina Khatami is a PhD student at the University of Melbourne (Australia). He is interested in hydrological modelling, uncertainty, and philosophy of science. In his PhD project, he developed a process-based model evaluation method, called Flux Mapping, to gain insights into the internal dynamics of conceptual models. He is also the current Secretary of Young Hydrologic Society and a student member of AGU’s Hydrology Section Hydrological Uncertainty Technical Committee.

Edited by Matthias Sprenger and Maria-Helena Ramos

All models are wrong but…

All models are wrong but…

“All models are wrong but some are useful” is a quote you probably have heard if you work in the field of computational hydrology – or ‘hydroinformatics’ – the science (or craft?) of building computer models of hydrological systems. The idea is that, even if these models cannot (by definition!) be a 1:1 representation of reality, their erroneous predictions can still be useful to support decision-making – for flood protection agencies, urban planners, farmers, water resource managers, etc.

From a modeller perspective, however, the question remains open: how do we decide whether a particular model (the one we are developing!) is ‘wrong within reasonable limits’ and thus suitable for use? How do we decide whether a certain simplifying assumption is acceptable? Or whether adding details is really beneficial? Ironically, one of the most thoughtful – and enjoyable – articles about the matter was published the year I started primary school… and here I am writing about it 30+ years later… (… so maybe it was ‘destiny’ after all?!).

MO’ DATA, MO’ PROBLEMS

One way to test (or ‘validate’) a computer model is by comparing its predictions against observations. If the model is able to reproduce the data collected in past circumstances, then it can be used to make predictions in other circumstances: those we are observing now (forecasting) or some that may occur in the more distant future and are of interest to us (simulation). Underpinning this approach is an implicit assumption: we trust the data – the ones we use for ‘validation’ and those we will feed into the validated model when making new predictions. However, this is a strong assumption in hydrology where, differently from some other sciences, data are not generated through controlled lab experiments but collected ‘out there’ – on hillslopes, in rivers, etc. – where the ‘observed’ processes are affected by myriads of factors, mostly out of our control, often non stationary, sometimes simply unknown (a crucial point I first came across in one of the very first papers I read for my master thesis – and clearly stuck to mind!).

One may think the problem will go away as we get new data products from intelligent sensors, satellites, drones, etc. And yet even these data will not cover all components of the water cycle (think of groundwater); they often require complex pre-processing, which introduces a whole new set of uncertainties; and they cannot completely close the spatial, temporal or conceptual mismatches between the quantities represented in our models and those we actually measure. So, it seems to me that the availability of new data products, rather than solving our problems with computer model development, is mainly giving us a deeper understanding of their deficiencies and limitations.

CODE HARD BUT TEST HARDER

If we cannot simply rely on data for validation, where else should model credibility come from? As pointed out in another great paper on model validation in earth sciences, often the most we can ask of a computer model is, simply, that it “does not contain known or detectable flaws and is internally consistent” – in other words, that it behaves as we expect it to.

Such a validation criterion may sound weak: so models are not meant to reproduce reality but only our understanding of it? Well, I’d say… yes: we use models to reveal how complex systems may behave, given the behaviour we have defined for their individual components. But computer models do not possess more knowledge than we do; they only do the calculations for us – more accurately and more quickly than we would be able to do otherwise (unless we have human computers!).

The criterion may sound obvious: wouldn’t all computer models pass it? Here I’d like to say ‘yes’ but… I am afraid not! In my experience, hydrological models can quickly go ‘out of hand’ if only we combine a few (non-linear) processes and parameters. And when we scrutinise them a bit deeper, we often find that, be it because of plain bugs in the code or more subtle numerical interactions, they may behave in ways that we would have not anticipated, and are inconsistent with what we know about the real system.

So, I think we have to accept that adding complexity to a model does not guarantee, per se, that it will give us a better representation of the real system. Actually, there may even be a trade-off whereby adding complexity reduces our ability to carry out a comprehensive testing of the model behaviour, paradoxically undermining its credibility instead of enhancing it. Ultimately, what we need is not only more sophisticate models but also more sophisticate evaluation tests – using formal, structured approach to analyse propagation of uncertainties and sensitivities. And to accept that sometimes a simpler model with known behaviour and limitations is probably more useful than a more complex one that cannot be comprehensively tested.

Edited by Maria-Helena Ramos

________________________________________

Guest author Dr Francesca Pianosi is a lecturer in Water and Environmental Engineering at the Department of Civil Engineering of the University of Bristol, and an EPSRC Fellow. Her main research interest is the application of mathematical modelling to advance the understanding and support the sustainable management of human-environment systems, and in particular water resource systems. She is also interested in open-source software development and is the lead author of the SAFE Toolbox (safetoolbox.info).

 

Gender balance in the HS division- some personal thoughts

Gender balance in the HS division- some personal thoughts

Gender balance in the HS division- some personal thoughts

On 14 June 2019, there was the Swiss nationwide women strike day, with the main topic of equal pay for equal work (see e.g. here). A good opportunity to share some thoughts about gender balance in the HS division. If you have a look on the HS division composition today, you will see that we have a female president and a female deputy president, in addition 5 female officers out of 11 officers (in charge of the subdivisions) and a female early career scientist representative. Overall, 12 out of the 25 officers are female. This is indeed impressive and a nice achievement. It is without doubt the result of the passionate gender balance debate that took place during the 2014 HS business meeting (see my HEPEX blog post on this).

After that debate, it was clear that something had to change. And the change did happen! Why? Certainly because many colleagues became more proactive when looking for excellent female candidates for division positions. And, of course, because many female colleagues became less reluctant to accept these positions. I was one of them. And while I am extremely happy to see where we are now, I continuously ask myself how to make this change sustainable. Besides nominating female candidates at all levels, the most important task for all of us is certainly to keep the discussion alive, to make that little extra effort while looking for invited speakers or while nominating colleagues for awards and, more importantly, to make change happen at all levels, for example for the next summer school or for the weekly seminar.

The leaky pipeline as reported here.

And where is the link to the Swiss women strike? Back in 2014, I triggered the business meeting debate around gender balance because I had just heard about the wage imbalance in Switzerland. This imbalance continues to persist. It continues to not be explicable (e.g. here a link to a Swiss research project on this topic). And I have experienced it myself during a former position in Switzerland where my male colleague in the same lab and at the same position and with the same age and the same achievements had a considerably higher wage. Why did I not do something against it? Because I did not have the energy to fight. Let’s hope that those times are almost gone.

Talking hydrology: an interview with Hjalmar Laudon on hydrological research at the Krycklan catchment

Talking hydrology: an interview with Hjalmar Laudon on hydrological research at the Krycklan catchment

For our second post of “Talking hydrology”, we interviewed Hjalmar Laudon, professor and chair of forest landscape biogeochemistry at SLU Umeå (Sweden). We talked about past and current research in the Krycklan catchment and the usefulness of long-term datasets.

 

1) You have been conducting hydrological research at the Krycklan catchment (North of Sweden) since 2002. How did you keep yourself motivated and interested in the research in the Krycklan catchment?

I think that what we do is interesting, unique and fun, and a better understanding of the role of water in this catchment is very important in order to understand how climate, forestry and other human induced activities will affect our waters. The work and the focus of the projects themselves have been multidisciplinary. This has taught me many new things and made the work more exciting! I work with hydrologists and biogeochemists from all over the world, but also with forest historians, political scientists, forestry policy makers. Lately we also have seen big political changes and have pursued collaboration with stakeholders. Approaching Krycklan from all these different perspectives is very interesting! In addition, it is really inspiring to see that sometimes changes in the policies are implemented based on our findings. What also motivates me is the interaction with other scientists, especially young ones that come to Krycklan to carry out their research! They bring a different atmosphere compared to the ‘old guys’ that are just proceeding with their projects.

2) How were the long-term observations in the Krycklan catchment superior to the usual project lifespan of 3-5 years?

An obvious benefit of having a long-term monitoring station is that the measurement infrastructure is already installed. Since there is no need to set up everything with the start of a new project it is relatively easy to collect new data. A longer dataset also gives the opportunity to describe and understand long-term processes, such as age structure of the forest or the history of deposition of pollutants, and, all our measurements can be put into a climatic context. Nature is not in a steady-state condition if we assess it on a climatic time scale. So, if you make measurements for just a few years, you do not know on which long-term trajectory you are. In Krycklan, the long-term measurements have revealed a set of trends occurring simultaneously, for instance the increase of dissolved organic carbon (DOC) in stream water, the decrease of nitrate concentrations and pH in stream water, and changes in winter conditions. Of course, there are also negative sides to measuring at one location for a long period. Since we learn more about one specific site, we might be biased to the climate and biophysical characteristics of this particular site. Therefore, we must be aware of the mindset that we take from the catchments that we study, and how the dominant processes that we observed apply to other regions. Finally, resources are finite, and with every project one needs to decide which processes are most important to continue study. We cannot just add new measurements, we also have to close some down. This is the hardest part.

3) What type of funds are used to maintain the hydrological monitoring of the catchment? 

The measurements in Krycklan started with funding by small research grants, and we combined bits and pieces to set up the infrastructure. In the last seven years we have received funding from the Swedish Research Council to maintain 15 monitoring sites on a standard protocol, which we of course hope will continue. Additionally, we search for project-based funding to do additional measurements or sampling and people that want to come to do a project in Krycklan usually take their own funding or grant with them.

4) Based on your experience with field work, what kind of advantages and disadvantages can be encountered while collecting data for a PhD or a post doc project?

One thing that has become clear is that while discharge measurements are fundamental to any hydrologist, they can be extremely difficult to carry out and are also highly uncertain. There is a large uncertainty in the discharge (or stage height) and the methods for calibration that we have are poor. We realized this when we calculated the biogeochemical fluxes in Krycklan and we found that sometimes there were some inaccuracies. So, even at a well-maintained site like Krycklan it is hard to get the discharge right, especially at the shoulder seasons to winter! To make it even more complicated: each stream section can behave differently, so it is difficult to extrapolate from a few measurements to another location. Another problem is that with extreme events the locations are likely to get damaged or require a new stage-discharge relation. This is part of the reason why we measure at 15 sites: at least one of the sites will probably survive during a flood.

Generally, when working with field data, more data show the complexity of the system which inevitably makes it harder to describe. We must also be aware that no data is flawless. That is why it is really good to have the experience of field work when you work with other people’s data.

5) Do you involve students (B.Sc. or M.Sc. level) in the work you do in the Krycklan catchment?

Currently we have about 20 PhD and 20 MSc students involved at the Krycklan catchment, but data are available on the web (open access), so there could be more people that work with the data that we are not aware of. As a result, we have around 100 publications per year that use Krycklan data! We do a lot of teaching both at SLU and at Umea University, and we take a lot of students out to the field site. Students usually work on topics on hydrology, forestry and atmospheric sciences, and currently we are starting two big new experiments: on the restoration of wetlands, and one on forest ditching and ditch network maintenance.

6) What kind of advice would you give to early career scientists approaching field work? How would you encourage them and explain them the importance of experimental hydrology?

One advice I would give is to go to a site where there are good data. It is very valuable to work with good data. On the other hand, the problems and errors in the data are not found until someone starts working with the data. It is very easy to make errors, and they can be carried on for a long time until someone discovers them. That is why only through working with the data you know how to further develop the data collection. I also found that adding a historical, cultural or educational component to the research site can be a great help to further develop and better understand field measurements.

7) Have you ever wondered how experimental hydrology could evolve in the next few years? What kind of hypotheses should be tested? And what kind of groundbreaking instrumentations or methodologies are needed to improve our knowledge of hydrological processes by experimental data?

The comparison of different sites, and the investigation of specific questions at different sites is definitely one direction for future research. Currently, we have measurements and understanding at different locations, but to compare them can bring a new, deeper understanding to the hydrological system. Another challenge is the upscaling from plot to catchment scale. I think that large controlled experiments, such as artificial extreme events or the monitoring of the effects of impactful events such as forest logging on catchment scale (several 10’s of hectares) are both very important in exploring how to tackle this challenge. As for groundbreaking methodologies, I think that high-resolution LiDAR measurements which came in over the last ten years can continue to be very powerful to get a detailed overview of the landscape. Lastly, we need to integrate our understanding of different processes across the landscape; atmosphere, vegetation, soil, groundwater, surface water. This needs a big team effort and it is interdisciplinary in many ways! We need models to validate our ideas and to connect the pieces, and data from experimental sites is of course crucial for this exercise. One of the compartments that still needs attention before we can successfully make this integration is the soil compartment. Aspects as groundwater-surface water interaction across depth and the connection between deep and shallow subsurface waters still need a refined representation.

8) Do you have any other thoughts or inspiring words that you would like to share?

 Come to work in Krycklan, we need you!

Thank you, Hjalmar, for your time and insights in experimental research in a long-term catchment as Krycklan!

 

For more info, the reader can visit the Krycklan catchment website or contact Hjalmar Laudon.

Edited by Matthias Sprenger and Giulia Zuecco

________________________________________

Guest author Leonie Kiewiet is a PhD student at the University of Zurich (Switzerland). She is interested in experimental hydrology and focuses on the hillslope and catchment scale. In her PhD project, she uses a combination of tracer-based approaches and hydrometric measurements to investigate runoff generation processes and to quantify the uncertainty in source-area analyses due to spatial variability in shallow groundwater composition.