Contributed by Ilias Pechlivanidis.
Rafael (also known as Rafa) has been a key member of SMHI’s Hydrology R&D group for about 2 years. I personally got to meet him a year earlier at EGU to discuss about hydrological processes in data-sparse mountainous regions and Spanish reservoir regulations which are quite difficult to describe. However what makes Rafa unique is the multi-dimensionality of his character that can jump from serious to funny and from hydrological world explorer to the winner of the HEPEX logo competition.
Rafa is a post-doctoral hydrologist at SMHI since 2016. His research focused on studying snow dynamics at different scales in Mediterranean areas combining remote sensing and hydrological modelling. At SMHI his main topic has been global hydrology, and specifically setting up the HYPE hydrological model at the global scale, which is internally known as World-Wide HYPE (WW-HYPE). He has also been focusing on improving hydrological process understanding through earth observations and data assimilation.
I stopped by his office to report some of his adventures in hydrological science (and more):
Ilias Pechlivanidis (IP): Rafa, you have been “upscaled” from understanding hydrological processes in the small local scale into predicting hydrological responses over the entire globe. What has been more exciting for you?
Rafael Pimentel (RP): Hard to answer, as I find both topics very exciting. The first one is of key scientific importance as one needs to understand the local hydrological system first to be able to model and predict; for example, understanding the complexity of snow processes in semiarid areas. The particular climatic conditions generate a very heterogeneous snow distribution, with several accumulation-snowmelt cycles within the same hydrological year. Therefore, simple temperature-degree models cannot be used, instead physical approaches based on the energy and mass balance equations are needed. Earth Observations play also a crucial role in understanding this process. Not only the conventional remote sensing snow products, but also detail photography. They are able to capture the interaction between the micro-topography and the snowpack, fundamental to understand better these particular accumulation and melting processes.
Regarding global hydrological modelling and specifically with the setup of the WW-HYPE, it has been a real challenge. First, dealing with big data (i.e. the delineation of about 131000 catchments using a high resolution (90 x 90 m) DEM; analysis, quality check and application in model calibration/evaluation of around 11000 river flow stations; application of a new delineation for big lakes and dams; process of 6 different EO world-wide products and comparison against model outputs). Second, making compromises between the detail in process representation and what is really adequate to capture the global hydrological patterns: for example including in the model setup specific routines to capture the behavior of the biggest floodplains.
IP: You are a full time Post-Doc, but your artistic touch has been well accepted by the HEPEX community. For Rafa, in the game of “Science vs Art”, who would win?
RP: (Rafa laughs) I believe that they are complementary. Science is my work, and this is already exciting since I feel I am under constant learning. On the other hand art is one of my hobbies. For me painting is a really good way to disconnect, escape, and see everything from another perspective. I would also argue that there is a tone of art in our science. For instance, we put a lot of effort in our plots, whilst our more creative side appears in all our results. Each of us has a particular “style”.
IP: Is there an artistic interpretation of the HEPEX logo?
RP: My idea was to be a bit different, more artistic and less formal; but still respecting the two main pillars of HEPEX: hydrology and ensembles. I used the handwritten font and different colors to make it more casual, and for example one of the versions had a bright magenta background. I also wanted more chaos in the ensembles, with no clear patterns, to respect the same concept. And finally, I added a simple drop of water to make the link to hydrology. I was very pleased and surprised to see the results of the competition. It was a big shot beyond the generally classical style but I personally believe it is time for a change and a new tone. I also wanted to say that the final version was refined by the graphical designers at ECMWF. Thanks, for this effort!
IP: If there was a key scientific problem (not strictly related to hydrology) that you would like to solve, what would it be?
RP: Thanks for giving me this flexibility but it is easier to go for a hydrologically related scientific challenge. Recently, IAHS has put an effort on gathering unsolved questions in hydrology with a target to be solved in the next decade. My favorite is: “Why do changes in the snowfall regime have a very different impact on stream flow in different catchments?” This is particularly interesting under global warming climate which in some regions has already shifted the snowfall regime, and changed the accumulation and melting phases, and consequently affected streamflow. To me, it is crucial to better understand the role of the evaporation from the snowpack, or the rain on snow events. These processes play a key role in snow dynamics, whilst in most hydrological model setups they are not well represented or even not represented.
IP: Tricky question, but I will go for it. If you had the chance to change something else in HEPEX, what would that be?
RP: I have recently started to be involved in HEPEX, so I feel “untrained” and biased to have a strong opinion on this. In the HEPEX community we have been discussing about forecasting skill, accuracy, benchmarking, sensitivity to initial conditions etc., which are estimated from models set up at the regional or large scale. However to my experience, large-scale hydrological models are still prone to a number of uncertainties resulting into poor performance in some regions (i.e. at karst systems or at regions with large human impacts).
Therefore, I wonder what the real role of the hydrological model could be in relation to the hydrological forecast. Could we get something from the hydrological forecast although the model is performing poorly? Do we (as a community) need to put additional effort in improving hydrological modeling and hence achieving better forecasts? If part of the answer is address by data assimilation, how much could we improve the forecast performance by assimilating in-situ and/or earth observations and what is their optimum combination of variables? Is uncertainty what we represent with our forecasting ensembles or are we introducing also noise? I generally feel that we lack of a framework to guide scientists and users on how forecasted information could be extracted from the forecasting services and further used for better decision-making. I would like to see the HEPEX community being on the front line in this effort.
IP: What is next for you?
RP: There are mixed feelings about this, but from this October I will be working in my former research group at the University of Cordoba, Spain. I have learned a lot at SMHI and managed to extend my scientific network and interests; however it is now the time to go back to my snow research. Yet I will this time give it a world-wide perspective!
Thanks to the HEPEX community for accepting my proposed logo. Personally I am very happy about the outcome, and I hope you are too.
Thank you, Rafa, for this interview. On behalf of the HEPEX members and co-chairs we would like to thank you for proposing this beautiful logo. We will keep an eye on your achievements in the future!