Optimisation of water management by using ensemble forecasts

Principal Investigator

Dr Markus Pahlow
Ruhr-University Bochum
Institute of Hydrology , Water Resources
Management and Environmenatl Engineering
44780 Bochum
GERMANY

markus.pahlow@rub.de

Other researchers: Corinna Möhrlen and Jess Jørgensen (SERG, Department of Civil and Environmental Engineering, University College Cork, Ireland)

Project description

Flooding events in The Netherlands, Germany, Austria, Switzerland, and France throughout recent years have shown that new ways of managing water reservoirs are necessary. This is especially relevant in our societies, where a growing amount of the natural ground is sealed by buildings and transportation, which complicates the assignment of large enough polder areas for flood prevention.

Additionally, the restructuring of today's electricity supply system throughout Europe with a widening focus on the use of renewable energy with mainly hydro power, biomass and wind energy prompts for the adaptation of strategies in water management. The economic value of the potential energy of the water is increased by releasing the hydro energy at times with high prices on electricity with the potential of increasing the damages caused by sudden excessive rain during flooding events.

Renewables and in particularly wind power cause increasing price fluctuations and with the plans of increased wind power in mind, there is a need to develop strategies to reduce the risk of flooding caused by the combination of full water reservoirs and sudden excessive rain.

Seen from a CO2 perspective, hydro energy is best used as reserve for wind power. The optimal strategy is to release hydro energy in periods with low wind power production and keep the hydro plants as spinning reserve in windy periods. However this strategy will occasionally fill the reservoirs and thereby increase the danger level of flooding. To estimate the optimal daily enforcement of hydro generation, considering also safety requirements, there will essentially be a need to predict the degree of hydro and wind generated energy up to one month in advance.

Large scale model systems are becoming more important in that respect. These can account for all reservoirs along contributing rivers, which might cross several countries. The output of such a model system will in the end enable authorities to force the release of hydro energy regardless of the current electricity price to prevent damages. In addition to the flooding risk, the communities are wasting CO2-free energy, if the water reservoirs can not host the incoming water and convert it to energy at the required rate. Because of the liberalisation of the electricity markets, a generally accepted prediction mechanism is required to enforce a control mechanism of the water reservoirs by authorities. The project will investigate, if the seasonal and medium range forecasts in combination with state-of-the-art hydrological models can be used to setup such a management system.

The goal of this project is therefore to use the high resolution ensemble forecasts from ECMWF to define the uncertainty of precipitation events for an improved management of water reservoirs with a balanced focus on the economics, CO2 benefits and safety.

The project will focus on evaluating the applicability of ensemble data of precipitation as input for hydrological modelling.

We anticipate to solve this task in two steps: (i) Statistical analysis of precipitation data throughout Europe together with ensemble predictions (ii) Investigation of the applicability of measurements and ensemble forecasts as input parameter for hydrological simulations.

In part (i) of this project we will identify important areas within Europe for both water management and hydro based electricity generation. A detailed statistical analysis with a probabilistic multi-trend filter will be carried out for these areas and events. These probability distributions are used to thoroughly investigate the accuracy of precipitation forecasts when used on a large scale (in hydrological terms). The ensemble predictions will be verified against a high resolution deterministic forecast and possibly with short-range ensembles to develop new methods of precipitation forecasts to reduce hydrological modelling uncertainty.

For more details, please refer to the latest progress report.

Additional information

Project started in 2006.

Would accept support for 1 year only, if necessary.