The determination of water levels of inland waters offers numerous challenges. Observations are usually limited to large rivers or main streams (river widths of many hundreds of meters to kilometers), since radar signals in the case of small water bodies can be heavily distorted by surrounding topography and off-nadir backscatter signals. Furthermore, the spatial and temporal resolution of the observation data depends on the satellite orbits.  

In order to improve the data coverage and produce the most accurate time series possible, DGFI-TUM is working on the development of improved analysis methods for radar echoes perturbed by the influence of land and the combination of different missions with different sensors and orbital characteristics (e.g. missions on repeat-tracks and on drifting orbit configurations). This comprises the classification and retracking of altimeter waveforms, the treatment of inter-mission and retracker biases, the elimination or appropriate use of off-nadir observations (so-called hooking effect) and outliers, and the optimal combination of observations. To provide a comprehensive spatio-temporal description of a river system, statistically robust interpolation methods are developed that take into account the changing flow velocity of the river as well as particular catchment conditions.

Using the developed methods, time series of water levels for lakes, reservoirs and rivers are calculated and thoroughly validated by comparison with in situ gauge data. Depending on the size of the water body and the topography, accuracies of up to one centimeter are achieved.

Data are made available through the web service DAHITI (Database for Hydrological Time Series of Inland Waters).