Again on Modelling. On modelling and data.

I was invited to talk to Ph.D students at the 2019 Warredoc School on data rich Hydrology. This follows other more or less opinionated talks I give at the GEOframe Winter School, and complete them. In hindsight I actually scratched the surface of the topic. I would have made a more substantial contribution if I had added examples with GEOframe for clarifying "where do datasets integrate best in models".  Never mind: there is then still matter for the next reflections and investigations.

Notwithstanding, I hope you will find my contribution interesting and useful. Click on the Figure to access the presentation.

Material for the GEOframe Winter School - Opinions on (hydrological) modelling

At the end day of the school there is the place for some opinions o modelling. These were expressed many other times in different posts but collected here in three different parts.

• Simplifying (and viceversa) Hydrological model physics  - YouTube video I part and part II
• How can we make models more physically based - YouTube video 
• Another Post on the same topic
• Which is the best hydrological model ? YouTube video
• Same topic,  another post, and a second post

Some references

• Addor, N., & Melsen, L. (2018). Legacy, rather than adequacy, drive the selection of hydrological models. Water Resources Res., 1–34. http://doi.org/10.1029/2018WR022958
• Kirchner, J. W. (2006). Getting the right answers for the right reasons: Linking measurements, analyses, and models to advance the science of hydrology. Water Resources Research, 42(3), 1–5. http://doi.org/10.1029/2005WR004362
• Klemeš, V. (1986). Dilettantism in Hydrology: Transition or Destiny ? Water Resources Research, 22(9), 177S–188S.
• Savenije, H. H. G., & Hrachowitz, M. (2017). HESS Opinions: Catchments as meta-organisms - a new blue print for hydrological modelling. Hydrology and Earth System Sciences, 21(2), 1107–1116. http://doi.org/10.5194/hess-21-1107-2017

Previous Winter School page

All Winter School pages

Next Winter School page 

Material for the post Winter-School-on-GEOframe

Dear All students of the Winter School on GEOframe. Was a pleasure to have you in Trento and a pleasure to see you work so hard for two weeks. On the organizer side, I have to thank my Ph.D. students Michele Bottazzi, Niccolò Tubini, Giovanna dal Piaz for their work and support. I have also to thank my former Ph.D. students Marialaura Bancheri and Francesco Serafin (almost former) for their support in the background. I also thank my former student Giuseppe Formetta and those who contributed in the past to this successful story (see here). For various reasons they were not present at this school, but they are absolutely part of our community.

It is important for us to have feedbacks from you and critique too, if they are constructive. Please address them to the mailing list of the School.

Now I hope that you complete your exercises, and let me to reuse your results for educational purposes. Let me know your opinions. With someone more interested, we can investigate a way to work around some topic and produce a paper on what you are doing.
If you want to go beyond in your knowledge, when you have assimilated well the GEOframe philosophy and tools, with your exercises,you can try Java. To start you can give a look to the Java 101 course for hydrologists. Is by far incomplete but in the next months I will work to add material (by me and collaborators) to bring people to contribute profitably to GEOframe. Our codes remans free and inspected at Github.
I indicated various book for self-instructions on Java that you can find here. I have other on OSF here that you can access upon request. We will certainly help you either at distance or in Trento, if you want to spend a period working with us.

So, again, was a pleasure to meet you.

Go to the first page of the School

Pictures from the Winter School on GEOframe

The course for doctoral students, post docs and young researchers in Hydrology, Forestry, and related disciplines will cover the simulation of the hydrological cycle of catchments of various sizes with the GEOframe system.  They say that all models are wrong but useful. However, with better tools you forecast and decide better.

The course has been completed and, please you can find below all the material, video, papers cited by clicking on the links.

The  topics treated has been:

• Tuesday, January 8: Introduction to the course - What is OMS - What is GEOframe - Using Python and Jupyter for visualising Inputs and Outputs 
• Wednesday, January 9- Catchments and Hydrologic Response Units delineation 
• Thursday and Friday, January, 10 - 11 - Treatment of spatial data and Calibration  
• Monday January 14 - Estimation of radiation components 
• Tuesday, January 15 - Evaporation and Transpiration 
• Wednesday, January 16 - Rainfall-Runoff modelling with various NewAGE components 
• Thursday, January 17 - Rainfall-Runoff modelling with various NewAGE components 
• Friday, January 18 - Opinions on hydrological modelling. Exercises on assignments.
• Epilogue and thanks

Go to the first day - Go to the last page of the School

Instructors

• Michele Bottazzi
• Riccardo Rigon
• Niccolò Tubini

with material prepared by

• Marialaura Bancheri
• Giovanna Dalpiaz
• Francesco Serafin

honours also to Giuseppe Formetta  and Hydrologis from whom we inherited a  lot.

Please find below some pictures of the School. A class always busy until late. But this did not prevent some social life!

Go to the last page of the School

Why choosing GEOframe over other models/platforms ? I would say for:

• Flexibility: GEOframe is not a model but a system of components that interact at run-time. You can chose among various components options for any of the processes.
• Expandability: If you like to program, with a little investment in Java you can write your own component and make them to interact with the others without having to reinvent the wheel.
• Parallelism. Components work in parallel when their tasks do not interact, but this is transparent for you (we call it implicit parallelism). 
• Spatial discretisation. A catchment is subdivided in parts (HRU) which can be modeled separately and are computed in parallel. The network structure is used to achieve the spatial parallelism. Its spatial modularity can be used to add/cut part of the basins without having to redo the spatial analysis, for doing multisite calibration, to progress the analysis of a larger basin in parts that are assembled together eventually.
• Beyond-state-of art components.  Besides traditional approach to processes, we implemented a few new ideas for all the processes we covered.
• Reliability.  GEOframe is currently used for the flood forecasting in real time by Regione Basilicata. It is not just a system for research that does not work in real cases. 
• Tracers studies. Not treated in the school are present tools for doing tracers studies,
• Process based modelling.  Not treated in the school, we have tools for integrating Richards equation in 1D, and we are developing tools for integrating it in 2d and 3d coupling it with the energy budget. These components will be able to interact with the other. We also started new developments on freezing soil and snow modelling.

Representation of Hydrological Dynamical Systems Using the extended Petri Nets

Finally we came out with the submission of the paper on Petri Nets. This was the topic of various posts collected under the name of reservoirology.  You  can see the submitted paper by clicking on figure below.

The paper deals with the graphical representation of lumped-parameter hydrological models or, as we called them, Hydrological Dynamical Systems (HDSys). We were not satisfied with previous representations of such models and we thought that figures in literature do not convey the right information to the readers, usually being not enough explicative. At the same time, we streamlined the process to document appropriately the models for reproducibility. Insufficiently explained HDSys are not reproducible and this is bad for hydrology. Then we setup a one-to-one relationship between graphics and mathematics and this could help the passage from the initial ideas about processes to their representation in formulas. At that point, we asked ourselves if our graphic tools were suitable to represent models with the ambition to explore the interactions between hydrology and ecosystems, and therefore account for their co-evolution. We obtained a positive answer adding a graphical feature to visualize how state variables regulate the models' parameters through quantities called controllers. Once the representation was complete, we could observe the analogy of our representation with those used in other sciences, as, for instance, theoretical biology. This open the way to connect hydrological work to the graphical methods in the non-linear systems theory.

Who wants to browse the history of the review, they can find it here:

• The first version of the paper
• The Associate editor observations
• The reviewers observations (major) and our rebuttal
• The second version of the paper
• The supplementary material
• The reviewers observations (minor) and our rebuttal
• The third version of the paper
• The revised supplementary material

Now accepted for publication in WAter Resources Research.

Material for the GEOframe Winter School - Rainfall-Runoff

Here we are introducing some modules for rainfall runoff modelling present in GEOframe. Some of them where actually refined for the Civil Protection of the Basilicata Region.

Schedule

• Deconstructing the catchment (YouTube video)
• The Embedded Reservoir Model (YouTube video). On this I was maybe not clearly enough. ERM in itself is not clear on the fact that ERM is not a model in the traditional sense. It is actually the composition of four components. For instance, the runoff one is a simple non linear reservoir. Working with sim files, we can rearrange them also in different ways and obtain different "modelling solutions"
• TheNet3 Infrastructure (YoutTube Video)
• LUCA (YouTube video)
• Examples of Application
• The Posina Catchment (Abera et al., 2017a)(YouTube video)
• The Blue Nile (Abera et al. 2017b) (YouTube video)
• The Basilicata Civil Protection (YouTube video)

Exercises

• The set of sim files and the Jupyter notebook are here
• The Python script by Christian Massari to create automatically the required subfolders. It is here.

General references to Rainfall-Runoff


Beven, K. (2012), Ranfall Runoff, the primer, Wiley-Blackwell

Rigon, R., Bancheri, M., Formetta, G., & de Lavenne, A. (2015). The geomorphological unit hydrograph from a historical-critical perspective. Earth Surface Processes and Landforms, http://doi.org/10.1002/esp.3855

References besides the one already used

For seeing how to represent lumped hydrological models (you can give a look to this paper here)

Abera, W.W. (2016), Modelling water budget at a basin scale using JGrass-NewAge system. PhD thesis, University of Trento

Bancheri, Marialaura (2017) A flexible approach to the estimation of water budgets and its connection to the travel time theory. PhD thesis, University of Trento.

Formetta, Giuseppe (2013) Hydrological modelling with components: the OMS3 NewAge-JGrass system. PhD thesis, University of Trento.

Formetta, G., Antonello, A., Franceschi, S., David, O., & Rigon, R. (2014). Hydrological modelling with components: A GIS-based open source framework, 55(C), 190–200. http://doi.org/10.1016/j.envsoft.2014.01.019

Patta, C, Costruzione di un modello idrologico di stima della disponibilità idrica in area pedemontana, Tesi di laurea (in Italian), Politecnico di Torino, 2018

For open questions about rainfall-runoff see also the Meledrio Posts.

Previous Winter School page

All Winter School pages

Next Winter School page 

Material for the GEOframe Winter School - Evaporation and Transpiration

Evapotranspiration accounts for most of fifty percent of the terrestrial hydrological cycle. We illustrate here some ways to estimate it with the tools offered by the GEOframe system. 

• Momentum and water vapor transport in atmosphere (YouTube video)
• Penman-Monteith according to Shymanski and Or (PMSO- YouTube video)
• PMSO with surfaces temperatures feedbacks (YouTube Video)
• Evaporation from soil and Traspiration mechanisms (YouTube Video)
• From leaves to canopies (YouTube Video)
• Documentation of GEOframe-NewAge component
• Jupyter notebook with exercises
• Live YouTube video on the sim files and Jupyter notebooks

Out of schedule (for Chhay)

• Budyko curves derivation

Exercises

• You can follow this Jupyter notebook

References

Primarily for historic papers browse to the list by Dennis Baldocchi

• Haghighi, E., Shahraeeni, E., Lehmann, P., & Or, D. (2013). Evaporation rates across a convective air boundary layer are dominated by diffusion. Water Resources Research, 49(3), 1602–1610. http://doi.org/10.1002/wrcr.20166
• Lehmann, P., & Or, D. (2013). Effect of wetness patchiness on evaporation dynamics from drying porous surfaces. Water Resources Research. http://doi.org/10.1002/2013WR013737
• Or, D., Lehmann, P., & Shahraeeni, E. (2013). Advances in soil evaporation Physics - A review. Vadose Zone Journal.
• Schymanski, S. J., & Or, D. (2017). Leaf-scale experiments reveal an important omission in the Penman–Monteith equation. Hydrology and Earth System Sciences, 21(2), 685–706. http://doi.org/10.5194/hess-21-685-2017
• Schymanski, S. J., Or, D., & Zwieniecki, M. (2013). Stomatal Control and Leaf Thermal and Hydraulic Capacitances under Rapid Environmental Fluctuations. PLoS ONE, 8(1), e54231–16. http://doi.org/10.1371/journal.pone.0054231
Stroock, A. D., Pagay, V. V., Zwieniecki, M. A., & Michele Holbrook, N. (2014). The Physicochemical Hydrodynamics of Vascular Plants. Annual Rev. Fluid Mech., 46(1), 615–642. http://doi.org/10.1146/annurev-fluid-010313-141411

See also the discussions here:

• The vapour budget
• Soil-Plants fluxes

And the mini-series: 

• The man who planted trees.

Previous Winter School page

All Winter School pages

Next Winter School page