A Water Availability Prediction System for water management.


Xavier Llort, Álvaro Rodríguez, David Sancho, Rafael Sánchez-Diezma, Marc Berenguer, Daniel Sempere-Torres, Arnau Cangros, Esteban Muñoz

Thursday 2 july 2015

11:15 - 11:30h at North America (level 0)

Themes: (T) Special session, (ST) FP7 ICT and water

Parallel session: 11I. Special session: FP7 ICT and water


A key point in water management is to have a good prediction of its availability (both short-term and long-term) as a primary source of information. In water systems where the superficial water is one of the involved sources, the correct estimation of accumulated rainfall (observed as well as forecasted) over the contributive catchments is crucial. In this framework, within the EU-FP7 project UrbanWater and WatERP a Water Availability Prediction System [WAPS] is both developed and implemented in several pilot sites. Two parts compose the proposed WAPS: - A long-term (seasonal) statistical WAPS based on historical records and current situation measures. - A short-term WAPS that merges different sources of precipitation (both observed and forecasted) in a continuous time series of hourly rainfall accumulation fields and integrates them over the corresponding catchments domains to provide aggregated values at basin levels. In the long term, records of dam levels are used to set volume-change distributions along the year. Those distributions are applied to current volumes to obtain a probabilistic forecast of the next months based on historical seasonal behaviour. In the short-term, different sources of precipitation (measurements and forecasts) are merged to obtain a continuous time series of precipitations fields having the best precipitation estimation available at each time step. For this, the past rain estimates are based on a geostatistical approach to combine radar and raingauge observations (where both instruments are available). In the future, and for leadtimes between 0 hours and 6 hours, radar nowcasting is combined with Numerical Weather Prediction [NWP] models rainfall forecasts taking into account the performance of each source. For leadtimes beyond 6 hours, the rainfall fields are composed by NWP models forecasts. The WAPS is used on the pilot sites of the two projects as a tool to provide information for management purposes, and as a source of information for other modules of the Decision Support Systems integrating the project platforms.