Seasonal streamflow forecasting in Sweden

Further development of the HBV model for alpine areas

Innerthal, Switzerland (2017)

Snow is a major contributing factor of stream runoff in many alpine and high-latitude areas of the world, as well as a significant temporary reservoir of freshwater. For this reason, detailed information and knowledge on the timing, magnitude, and variability of snow processes (i.e. snowfall, accumulation, and melt) is important for successful water resources management and risk reduction in such areas. In this context, and due to the persistent limitations in data availability, hydrological models are widely used for making assessments and predictions on freshwater management and risk reduction. Among the large array of available hydrological models, the HBV model is a widely used model due to its simplicity, flexibility, and robustness. HBV’s design, representing hydrological processes in a simplified way, is particularly valuable in data-scarce areas. Additionally, the simplicity of the model is also beneficial for minimising uncertainties. Previous studies, however, show that an increased realism in the representation of some processes – e.g. snow processes – might improve the performance of hydrological models.

In this study we review and assess alternative conceptualisations to the snow routine of HBV-light, a user-friendly yet powerful version of the HBV model. Additionally, we explore the suitability of using additional input data such as relative humidity and shortwave solar radiation. We then evaluate the most promising model variations for a selected group of catchments representative of the different climatic and geographical areas of Switzerland. We also perform a further evaluation on catchments located in climate zones and landscapes different than Switzerland (e.g. USA) to ensure the robustness of the selected approaches.

Information Needs for Water Resource and Risk Management – Hydro-Meteorological Data Value and Non-Traditional Informaiton

Dresden, Germany (2011)

The research project was carried on within the Centre for Natural Disaster Sciences (CNDS) in Sweden. The research project arised from the need to improve water management in order to mitigate the costs and losses derived from the interaction of society with the natural variability of the hydrological system.

Data availability is a crucial component in any water management system. While many parts of the world continue to strive with chronical data scarcity, high-resolution sensor networks are becoming increasingly available in some regions. For this reason it is important to correctly assess and rationalise the real data needs for ensuring an adequate implementation of water management practices. Recent developments have shown that additional data sources can be used to complement traditional “hard data” (e.g. precipitation, streamflow) to improve how management and response actions are performed. On one side social media data such as Twitter or YouTube can provide valuable near-real-time information about disaster events. This potential has already started to be used for many studies and applications around the world. Finally, data on how people behave in disaster circumstances as well as their perception of the risks is extremely important to be able to exploit the full potential of the technical capabilities and advances in hydrological modelling and forecasting.

Along my PhD I focused on different aspects of hydrological and socio-hydrological data and I explored how these data affect the performance of hydrological models and early warning systems.