Model-Data Fusion in Hydrology

Hydrological model simulations are prone to errors and uncertainty related to deficiencies in the model structure itself, and to input errors, especially related to parameters and model forcings. It is therefore important to combine model predictions and measurement data, to correct model predictions via adjustments to model parameters, forcings or state variables. Increasing computational power and the increasing availability of (real time) measurement data provides new opportunities and challenges for improved hydrological estimates by model-data fusion. Various methods are used for this model-data fusion process, including inverse modelling and sequential data assimilation, and recent alternative approaches such as deep learning. Furthermore, new data types from satellites, measurement networks and new technology like drones also offer new perspectives on model-data fusion.

This special issue focuses on new developments in the area of hydrological model-data fusion, which can include, but are not limited to, improving understanding of measurement types to support the development of better measurement operators, the evaluation of the value of novel data types in model-data fusion, the design of monitoring networks, new or improved methods for model-data fusion, new or improved methods for the estimation of soil, hydrological, aquifer or vegetation parameters, and case studies which provide precise insights in the performance of hydrological model-data fusion at any spatial scale, from the point to the catchment and global scale.