Models for the representation of the hydrological processes can achieve today a high level of accuracy in the reproduction of the available observations. However, the most complex “physically-based“ models often require equally complex and accurate data availability to be driven to their desired level of accuracy. Moreover, quasi-exact representation of some separate sub-processes not necessarily leads to an equally accurate reproduction of a combination of them, as can be for instance, the case for flood processes.
Many times in practical problems, but also in the attempt of simulation of some processes, the available data is far from being sufficient (both in terms of accuracy and of the spatial/temporal resolution) for the calibration or application of the "ideally best" model. For this reason, in daily practice, tragically simple methods such as the rational formula still demonstrate to be effective for applications in poorly-gauged basins. In situations of data scarcity, bridging the gap between theory and practice requires that the concept of "ideally best" model be adapted to the available possibilities of model validation and of transfer of information from gauged to ungauged sites. All this is sometimes called "parsimony in model building". Model adaptation to different temporal resolution of data by means of process conceptualisation or by using resolution-independent parameters clearly exemplify parsimonious modelling approaches.

This session is intended as a forum for presentations covering the above issues from many viewpoints as, for instance:
- Approaches to model selection and model validation in poorly-gauged basins;
- Data-based modelling approaches;
- Models that integrate data from different sources and at different time resolution
- Model adaptation to changes in the scales of observation (i.e. Downward or Top-Down approach)
- Modelling processes in presence of uncertain or unreliable data
- Model conceptualisation and parameter normalization to reduce data-time-step-dependency
- Use of simple models for detection and reproduction of dominant sub-processes (i.e minimalistic approach)
- Presentation of case studies and of the state-of-the-art in data availability at regional/national level in countries with different levels of organization and development.