Methodology for developing operational hydrological prediction systems
Event details
| Date | 26.11.2013 |
| Hour | 16:15 › 17:15 |
| Speaker |
Dr Massimiliano Zappa, Mountain Hydrology and Mass Movements, Swiss Federal Institute for Forest, Snow and Landscape Research WSL, Dübendorf (CH) |
| Location | |
| Category | Conferences - Seminars |
Abstract:
The cook-book for developing a “new (hydrological) model” to describe a “real process” is well understood and widely accepted. It begins with the observation of the system and ends with testing the (hydrological) model against observations. In case of success the model can be used, in case of failure additional observations and more process knowledge needs to be accounted in the (hydrological) model. Once (hydrological) models are validated practical applications can be envisaged.
In this seminar I will present the way we selected in order to setup innovative operational hydrological prediction systems for practitioners and research. Similarly as for model development a cook-book can be followed. In our experience we replace the “real process” with “something useful you can compute in real-time” and we replace the “new model” with a “solution for supporting decision making”. Following this principle we replace “the observations” with the “dialog with end-users”, so when in the end our solution doesn’t fulfils the end-user expectations, we restart the dialog and try to improve our real-time application. This requires that end-users are confronted with prototypes and scientific novelty.
I will present 4 operational systems developed in this way:
• The real-time hydrological ensemble prediction systems (HEPS) for the city of Zürich, operational since 2008
• The operational use of rainfall radar uncertainty for flash-flood forecasting in the Verzasca river, operational since 2007
• The new platform “www.drought.ch” for early identification of drought in Switzerland, operational since June 2013
• The early warning tool for flash-floods based on IDF-Curves operational for Southern Switzerland since July 2013.
My talk should demonstrate that it is possible to make scientific research using real-time data. After some years of experience it is not easy to give a final word on daily routine with the operation of real-time hydrological prediction systems. If you are so lucky to have one or more end-users being interested in it, then you should be ready to have a 365 days-per-year commitment to your experiment. However, being part of real-time forecast systems yields in many moments a much larger gratification and excitement than any off-line experiment leading to high-impact contributions to scientific literature.
Keywords: Floods, Hydrological droughts, real-time operations, end-users, HEPS
Dr Massimiliano Zappa is hydrologist. He works on operational flood forecast systems and process related catchment hydrology. His research interests focus on spatially distributed modelling of hydrological processes at different spatial and temporal scales. He completed in 2002 a Phd at the Institute of Atmospheric and Climate Sciences (ETH Zurich). Part of the Ph.D were financed by the EU-VOLGAFOREST project. In 2003 he joined the Swiss Federal Research Institute WSL (head of the Hydrological Forecasts group since 2006) where he focussed on applied research for operational flood warning systems including warning systems for the Three Gorges Dam Reservoir (PR China), the Swiss Canton of Glarus and the city of Zürich. He has been included as a national delegate in the Management Committee of COST-731 (‘Propagation of Uncertainty in Advanced Meteo-Hydrological Forecast Systems’) and is also National delegate in the Northern European FRIEND (Flow Regimes from International Experimental and Network Data) Low-Flow working group. He has been recently involved in the FP7-Project IMPRINTS (FP7-ENV-2008-1-226555) and was a co-PI in a sub-project of the NRP61 program on sustainable water use.
The cook-book for developing a “new (hydrological) model” to describe a “real process” is well understood and widely accepted. It begins with the observation of the system and ends with testing the (hydrological) model against observations. In case of success the model can be used, in case of failure additional observations and more process knowledge needs to be accounted in the (hydrological) model. Once (hydrological) models are validated practical applications can be envisaged.
In this seminar I will present the way we selected in order to setup innovative operational hydrological prediction systems for practitioners and research. Similarly as for model development a cook-book can be followed. In our experience we replace the “real process” with “something useful you can compute in real-time” and we replace the “new model” with a “solution for supporting decision making”. Following this principle we replace “the observations” with the “dialog with end-users”, so when in the end our solution doesn’t fulfils the end-user expectations, we restart the dialog and try to improve our real-time application. This requires that end-users are confronted with prototypes and scientific novelty.
I will present 4 operational systems developed in this way:
• The real-time hydrological ensemble prediction systems (HEPS) for the city of Zürich, operational since 2008
• The operational use of rainfall radar uncertainty for flash-flood forecasting in the Verzasca river, operational since 2007
• The new platform “www.drought.ch” for early identification of drought in Switzerland, operational since June 2013
• The early warning tool for flash-floods based on IDF-Curves operational for Southern Switzerland since July 2013.
My talk should demonstrate that it is possible to make scientific research using real-time data. After some years of experience it is not easy to give a final word on daily routine with the operation of real-time hydrological prediction systems. If you are so lucky to have one or more end-users being interested in it, then you should be ready to have a 365 days-per-year commitment to your experiment. However, being part of real-time forecast systems yields in many moments a much larger gratification and excitement than any off-line experiment leading to high-impact contributions to scientific literature.
Keywords: Floods, Hydrological droughts, real-time operations, end-users, HEPS
Dr Massimiliano Zappa is hydrologist. He works on operational flood forecast systems and process related catchment hydrology. His research interests focus on spatially distributed modelling of hydrological processes at different spatial and temporal scales. He completed in 2002 a Phd at the Institute of Atmospheric and Climate Sciences (ETH Zurich). Part of the Ph.D were financed by the EU-VOLGAFOREST project. In 2003 he joined the Swiss Federal Research Institute WSL (head of the Hydrological Forecasts group since 2006) where he focussed on applied research for operational flood warning systems including warning systems for the Three Gorges Dam Reservoir (PR China), the Swiss Canton of Glarus and the city of Zürich. He has been included as a national delegate in the Management Committee of COST-731 (‘Propagation of Uncertainty in Advanced Meteo-Hydrological Forecast Systems’) and is also National delegate in the Northern European FRIEND (Flow Regimes from International Experimental and Network Data) Low-Flow working group. He has been recently involved in the FP7-Project IMPRINTS (FP7-ENV-2008-1-226555) and was a co-PI in a sub-project of the NRP61 program on sustainable water use.
Practical information
- General public
- Free
- This event is internal
Organizer
- EESS - IIE
Contact
- Dr Bettina Schaefli, ECHO