Comparing drought simulation performance from large-scale and locally set up hydrological models for large mountainous rivers in Switzerland

Historically, Switzerland and the nearby alpine countries have not been associated with major droughts. However, in recent years, the European Alpine space has experienced several unprecedented low-flow conditions and drought events. As many economic sectors in the region depend heavily on sufficient water availability, such as hydropower production, navigation and transportation, agriculture, and tourism, it is important for decision-makers to have early warnings of drought tailored to their needs and geographical conditions.

The European Flood Awareness System (EFAS) has been in operation since 2012 providing flood risk overviews for Europe up to 15 days in advance. More recently, it has also run long-range hydrological outlooks for sub-seasonal to seasonal horizons. While EFAS early flood warnings have been extensively evaluated in the past years, less attention has been paid to evaluating the system’s ability to detect upcoming drought conditions. In this study, we turn our focus to this other extreme of the spectrum and on EFAS’ predictability of drought events in large Alpine catchments. Our goal is to investigate how hydrological patterns of skill at a large spatial scale can be combined with local model outputs to more accurately inform decision makers on droughts and their spatio-temporal evolution.

For this, we evaluate the performance of EFAS comparatively to that of a local model in terms of the ability to simulate drought conditions. The Precipitation-Runoff-Evapotranspiration HRU (PREVAH) local model was set up for 59 stations in Switzerland. The PREVAH model is a distributed conceptual hydrological model that accounts for processes such as evapotranspiration, interception, snow- and ice-melt, soil moisture storage, groundwater storage, and runoff generation. We analyse 25 overlapping stations between the local model and the EFAS reporting stations (river network points where EFAS outputs are available to users), and compare the drought simulation performances of the two models. We focus on evaluating the duration, deficit, and magnitude of the drought events, as well as metrics including Nash–Sutcliffe model efficiency coefficient (NSE) and Kling-Gupta efficiency (KGE).

The outcome of this study will lay a foundation for how a large-scale hydrological model like EFAS can complement a local model like PREVAH to improve the predictability of sub-seasonal drought forecasting and provide more reliable early warnings for better water resources management.