Innovative system for monitoring and forecasting hydrological dynamics in semi-arid Ceará, NE-Brazil

The state of Ceará is located in the semi-arid northeast of Brazil and is characterized by strong inter- and intra-annual variability in precipitation. Thus, droughts and an uncertain water supply threaten the people in one of the most densely populated dryland regions in the world. To store and supply water during dry periods, tens of thousands of dams of various sizes have been built, especially since the end of the 19th century. Only 155 of these reservoirs are systematically monitored and managed. For the remaining reservoirs, there is no systematic monitoring and coordinated water resource management so far. In addition to a comprehensive monitoring, it requires an adapted hydrological modeling and forecasting tool to best manage water resources in Ceará and to reduce the impact of future droughts.

In this project, an innovative system for monitoring and forecasting hydrological dynamics in Ceará was developed in collaboration with the Federal Agency of Hydrology and Meteorology (FUNCEME). This system is based on an integrated use of climate modeling, process-based hydrological modeling, remote sensing, and existing databases. Specifically, the following three complementary products have been developed:

• Satellite-based monitoring of stored water volume in reservoirs: Weekly monitoring of water masks of > 30,000 reservoirs is performed by evaluating and classifying Sentinel-1 scenes. The stored water volume can then be inferred from the area-volume relationship derived using high-resolution TanDEM-X CoSSC DEMs for these reservoirs during explicit drought years (i.e. when reservoirs were empty).
• Modeling and seasonal forecasting of hydrological dynamics using WASA-SED: The process-based hydrological model WASA-SED, developed for semi-arid areas, was adapted and calibrated for the state area of Ceará. Information from satellite-based reservoir monitoring is dynamically assimilated in the simulations. Based on an ensemble of ECHAM4.6 climate simulations (updated monthly), the adapted hydrological model is used to generate seasonal forecasts with six months lead time on streamflow and reservoir filling conditions.
• Web-based visualization of monitoring and forecast results: The results of satellite-based monitoring and dynamic hydrological modeling and forecasting are centrally managed in a database and can be retrieved from there by a web application. The corresponding information is visualized online as maps and graphics and made available to different user groups and decision makers.