Waterintellect researches computing methods for the simulation of the hydrological cycle. The core tool is the Representative Elementary Watershed (REW) Hydrological model. The REW model is an integrated hydrological simulation tool that uses spatial information in combination with hydrometerological data, numerical weather forecasts and remote sensing information for a detailed analysis of the hydrological cycle. The research aims at improved rainfall-runoff simulations and studies concerning the impact of global change on water resources in river basins.
Remotely sensed information has gained importance for hydrological studies. Real-time data on land-cover, moisture content of the soil surface and precipitation can be readily assimilated into hydrological models for an improved simulation of terrestrial water balance and the hydrological cycle.
The availability of the 30x30 m Digital Terrain Models (DTM), that have become available since the NASA Space Shuttle Topographic Mission (http://srtm.usgs.gov/) in the year 2000 have brought an accurate spatial analysis of the landscape within reach. Until recently such detailed 3-D information was not available for most regions of the world.
The present scientific challenge consists in an optimal combination of this information for hydrological research. This can be achieved through hydrological models that are able to make maximum use of this wealth of spatial data.
The central philosophy used in the research approach of Waterintellect consists in pursuing a process-based representation of hydrological processes. Process-based modelling relies on using the complete set of physical laws that govern flow and energy transfer in the shallow subsurface and between land surface and atmosphere.
The process-based representation is a tradeoff between using as much physics as possible, whilst focusing on the characteristic spatial scales that are relevant for the description of specific transfer processes within hydrological systems.
The general aim remains to simplify the description of the processes as much as possible by explicitly preserving the physical laws that govern them.
The growing amount of on-line data becoming available from ground-based monitoring systems and satellites needs to be exploited to improve hydrological simulation models.
For example soil moisture or snow cover data in a river basin observed from satellites, or the on-line water level recordet along a river system can be effectively used to reduce model output uncertianty and to set a model "on track". This practice is known as data assimilation.
Process-based models can effectively use spatial data and offer vast opportunities to exploit this information for enhanced modeling of the hydrological cycle. Waterintellect's research is oriented towards combining deterministic and statistical methods for data assimilation in hydrology.