Workshop Utrecht 2014

Workshop on hyper-resolution global hydrological modelling: the next step

held on February 13-14 2014 in Utrecht, the Netherlands
        
Organization
Marc Bierkens (Utrecht University), Eric Wood (Princeton University), Nathanial Chaney (Princeton University)
  
Sponsored by
   
Information:
  
Objectives
The main objective of this workshop was to bring together groups involved in large-scale hyper-resolution hydrological modelling in order to:
  1. exchange experiences thus far and learn from each other’s problems, failures, successes and solutions;
  2. investigate possible community building: discuss in what form groups involved in large-scale hyper-resolution hydrological modelling could work together in the future. Would it be a community of models doing comparison experiments, a community striving for one community model, or one that builds model components that can be exchanged between models (e.g. in a Python framework)?;
  3. set up a joint research program where problems/issues will be identified that concern most groups and assign a limited number of groups working on these problems together and making the results available to the community.

Meeting Program

Day 1: Thursday 13 February

9:30-10:00     Arrival, coffee and welcome
10:00-10:30   Eric Wood: Vision for establishing a collaborative network on hyper resolution global land surface modeling, and goals for the workshop - What would the network look like? -  What type of issues would we work on? -  What do we want to achieve this year? -  What do we want to achieve in the following years.
10:30-12:30    Presentations (15 minutes including focused discussion) by groups: progress, challenges, failures, unresolved issues: groups are requested to distribute the most important sheets  with m 3-4 pages, prior to the workshop.)
12:30-13:45    Lunch
13:45-16:00    Presentation by groups (continued)
16:00-17:30    General discussion of the activities by the groups: identification of areas of progress and main challenges by category; areas of collaboration. To guide the discussion we already propose the following categories: 1) computational issues; 2) parameterization and scale issues;
17:30              Drinks at the Basket (on Campus)
19:30              Dinner at Restaurant Goesting (in Town)

Day 2: Friday 14 February

9:00-9:15       Arrival and  coffee
9:15-10:00     Introduction by Nathanial Chaney: Computation Challenges: Modularization and Parallelization
10:00-10:15    Niels Drost (eScienceCenter): short presentation on AMUSE – Python framework now used to couple multiphysics modules (examples from astrophysics)
10:15-12:00    Discussion of main challenges by category: possible solutions/collaboration and subsequent actions.  Identification of groups willing to work on these (commitments). Three break-out groups: 1) computational issues; 2) parameterization and scale issues; 3) setting up a the collaborative network.
12:00-13:30    Lunch
13:30-15:00    Writing up results of morning’s discussion (challenges and action items) by category, and proposing structure of the research network/collaboration, modes of communication (website/wiki; AGU/EGU sessions) and development of a white paper (Contribution to EOS Transactions AGU). Timelines for action items (coming years, years to follow).
15:00-15:30  Wrap up and closure

Detailed list of short talks

Speaker 1 Speaker 2 Model/project university / organisation presentation time
Nathanial Chaney Eric Wood VIC: Talk Princeton University 10:30-10:45
John T, Reager Cédric David RAPID & CLM: Talk University of California Center for Hydrologic Modeling 10:45-11:00
Vicky Bell Grid-to-Grid (G2G) model: Talk Centre for Ecology and Hydrology, Wallingford 11:00-11:15
Edwin Sutanudjaja Marc Bierkens PCR-GLOBWB 2.0 /eWaterCycle: Talk dept. Physical Geography: Utrecht University 11:15-11:30
Rolf Hut Nick van de Giesen eWaterCycle: Talk TU Delft – CITG 11:30-11:45
Hessel Winsemius GLOFRIS: Talk Deltares – Delft 11:45-12:00
Ed Sudicky Hydrogeosphere: Talk University of Waterloo 12:00-12:15
Peter Burek Ad de Roo Lisflood JRC Ispra 14:00-14:15
Laura Condon Reed Maxwell CONUS Parflow: talk Parflow: Colorado School of Mines 14:15-14:30
Stephan Kollet Parflow University of Bonn 14:30-14:45
Luis Samaniego mHM: Talk UFZ Helmholtz Centre for Environmental Research 14:45-15:00
Paul Houser Hyper-resolution forcing: Talk George Mason University 15:00-15:15
Martina Flörke WaterGap: Talk WaterGap: University of Kassel 15:15-15:30
Petra Döll WaterGap: Talk University of Frankfurt 15:30-15:45
Dave Gochis NCAR HR regional modeling: Talk NCAR 15:45-16:00

List of Participants

1 Vicky Bell CEH, Wallingford
2 Marc Bierkens PCR-GLOBWB: Utrecht University
3 Peter Burek Lisflood: JRC
4 Nathaniel Chaney VIC: Princeton University
5 Laura Condon Parflow: Colorado School of Mines
6 Cédric David RAPID: University of California Center for Hydrologic Modeling
7 Ad de Roo Lisflood: JRC
8 Petra Döll University of Frankfurt
9 Niels Drost eScienceCenter Amsterdam
10 Martina Flörke WaterGap: University of Kassel
11 Dave Gochis NCAR HR regional modeling
12 Paul Houser George Mason University
13 Rolf Hut TU Delft – CITG
14 Stephan Kollet University of Bonn
15 Reed Maxwell Parflow: Colorado School of Mines
16 John T. Reager CLM:University of California Center for Hydrologic Modeling
17 Luis Samaniego UFZ Helmholtz Centre for Environmental Research
18 Ed Sudicky Hydrogeopshere: University of Waterloo
19 Edwin Sutanudjaja dept. Physical Geography: Utrecht University
20 Nick van der Giessen TU Delft
21 Hessel Winsemius Deltares – Delft
22 Eric Wood VIC: Princeton University
   
 Main outcomes workshop
  1. Creating an open network: Water innovation partnership on hyper-resolution modelling
    With the main objective: to strive for being able simulate terrestrial hydrology and water resources at hyper-resolution at acceptable accuracy.
    Why:
    -   Society: To be locally relevant for water resources management where needed.
    -   Equity: Equity of information about the globe.
    -   Science: By attempting this we will be able to learn at what scales which processes matter and how to transform information from one scale to another

This poses the question: is it possible to be locally relevant with globally available information?
This generates a number of scienctic questions (See also Wood et al., 2012)

  • Questions of scale: can we seamlessly up- and downscale our models and get consistent water balance and energy balance components?
  • Can we extract the information needed, in particular human impacts, to be able to reach sufficient accuracy at the local scale?
  • …..

2. The establishment of three working groups:

  • WG1: Setting up a testbed for comparing different large-scale models at different resolutions.
    • Global at 5 minutes (long-term goals: global at 1 km)
    • Domains: CONUS, CORDEX Europe at 1 km
    • Rhine + Illinois + California at 100 m
    • Common datasets: provided at the finest scale possible at fixed format.
    • Run the various models at decreasing resolution
    • Run it with local and globally available information
    • Include LIS scaling tools
  • WG2: Around computational challenges, including parallel computing and model component coupling.
    • Common platform to perform comparisons and analyze results: you submit your data and it automatically generate the comparison statistics
    • Common framework for coupling modules: common i/o system (AMUSE, CSDMS).
    • Commont tools for solving computational issues (parallelization)
  • WG3: With the goal to think about delivering the information needed to achieve hyper-resolution (< 1 km) globally:
    • Compile high-resolution global datasets using auxiliary information
    • Devise new model concepts that are replace subgrid-parameterizations with computationally frugal spatially explicit physics
    • Create hyper-resolution global forcing data.

3. Organize regular meetings (EGU and AGU).

4. Writing of an opinion paper.