Droughts in Switzerland have become more frequent and severe in recent years, and this trend is expected to continue. At the same time, increasing water demand and competition between different actors are putting more pressure on existing water resources. Having recognized drought as a significant risk for various economic sectors in Switzerland, a comprehensive national monitoring and forecasting system, to be launched in 2025, is being established through the joint efforts of three different government agencies (Federal Office for the Environment, Federal Office for Meteorology and Climatology, and Federal Office of Topography).

We will present the Swiss national drought project with a particular focus on in-situ, modelled and satellite-based monitoring, the integration of sub-seasonal forecasts, and the drought early-warning and information system. Specifically, this means creation of a national in-situ soil moisture monitoring network with approximately 30 stations, the development of meteorological and ecological drought products and indices derived from satellite and in-situ data, as well as the establishment of near real-time, downscaled, monthly forecasts. The integration of these diverse data streams into seamless products ranging from historical observations to sub-seasonal forecasts, all within a consistent climatological baseline and as open government data, is expected to be a significant step forward of great benefit to downstream user applications. The improved meteorological basis will directly feed into impact-relevant drought indices and hydrological models, with the aim of refining the early warning system to meet the needs of a very diverse user community, such as hydropower production, fluvial navigation, agriculture, forestry, artificial snow production, or ecology.

Overall, the project provides important information on the current and future drought situation on a regional to local scale. Daily updated maps and infographics are accessible through a user-friendly web platform designed to facilitate informed discussion and decision-making. Ultimately, the project aims to increase preparedness by facilitating emergency planning, reducing impacts and enhancing drought resilience across the affected sectors in Switzerland.

The system was designed by actively integrating user needs. The results of a user survey showed that although drought is multidimensional and affects stakeholders in different ways, one of their main needs is still a holistic "combined" drought index that can serve as a common basis for discussion and decision-making. Simple, locally focused designs were found to be the most efficient and useful, while designs, that present nation-wide maps or scientific quantities (SPI, etc.) were judged to be the least meaningful to educated but non-specialist users.

Drought is a water deficit and a persistent and recurring natural hazard that affects ecological and socio-economic systems. This results in a socio-economic drought (agriculture, drinking water, forestry, hydropower, tourism, etc.) where decision-makers need to quickly get an overview of the drought situation in their region. Use cases are e.g. that a community representative has to decide where the use of water should be regulated (e.g. watering gardens), or that a farmer analyses the drought situation of the past years to potentially evaluate a change of crop variety.

Currently, many specialized platforms on e.g. precipitation or soil moisture exist across administrative levels with graphs of varying spatio-temporal resolution. The challenge for the decision-makers is to collect the relevant data from all these platforms to get an overview of the drought situation in their region. This is time-consuming and prone to misinterpretation.

How do we design a public platform on drought (as part of the Swiss national drought project) that covers the requirements on analysis for decision-makers?

Our approach is «User-Centered Design».

1. Research with people from the target group to determine the requirements from a user’s perspective. Their biggest problem is to get an overview of data at different platforms to see the whole picture of drought in their region, and to currently make decisions based on a gut feeling.
2. Conception of the interface into a prototype to get an understanding of the platform among stakeholders.
3. Usability-Test of the prototype to evaluate the concept with the users. Results showed that this platform enables users to efficiently make data-based and not a gut feeling decisions because data streams from different government agencies (e.g. Federal Office for the Environment, MeteoSwiss) are collectively displayed.
4. Visual Design of the concept to ensure an interface based on user-centered GUI standards, incl. accessibility.
5. Specification of the concept to ensure users’ needs during technical development (currently in progress). Along the process, we included data providers to ensure feasibility of data structures.

With User-Centered Design, we designed a platform - for and with the users - that supports decision-making regarding drought in Switzerland.

The Standardized Precipitation Evapotranspiration Index (SPEI) is the WMO recommended drought index. It is computed using precipitation and evapotranspiration data and indicates deviation from a chosen historical mean, i.e. the reference period. MeteoSwiss provides the SPEI index at several measurement stations around Switzerland for 1-,3-,6-, and 12-month accumulation periods. The reference period used to compute these indices is 1961-present (11.08.2024 at time of writing). I hypothesize that this reference period does not account for climate change that occurred in the 20th century and risks under-estimating current SPEI values. Given a non-stationary climate, the first half of the reference period is different to the second half, and especially the decades in the 21st century where all climate records are being broken. It is unclear whether the water balance is stationary, a desired quality for the SPEI estimation methods. While this does not invalidate the model, the practical impact is critical; using a reference period that includes recent climate will reduce SPEI values, under-estimating the recent drought risk! This could impact agriculture, insurance, water management etc.

I compute SPEI for all of Switzerland using different reference periods and verify atmospheric water balance stationarity. The data used is the reanalysis ERA-5 Land (0.1*x0.1*), for monthly precipitation and evapotranspiration. SPEI is computed using the R package "SPEI", with log-logistic distribution fit and 3-month accumulation. The atmospheric water balance(wb) is tested for trends using the MannKendal trend test and the wb data for pre- and post-1991 is compared using t-tests. SPEI values are computed with reference periods starting in 1961 and ending in 2021, with iterative reductions of the period end. The SPEI results are evaluated at the 2012-2022 period SPEI monthly means.

Results of the trend test indicate a significant increasing trend (p<0.05) for the Spring/Autumn period in areas of the Rhone and Rhein valleys, Ticino, and Bernese Alps. These results are corroborated with the t-tests. There is no indication of significant wb trend in the rest of the country. I compared the 2011-2021 SPEI monthly means for 6 reference periods, with control reference period (1961-2021), and 5 periods each ending a decade earlier down to 1961-1971. Normalizing them with the control period, I observe seasonally and spatially variable results. For winter and summer (Figure 1), there is a monotonic increase in SPEI values with reference period reduction. However, spring/autumn results require further inquiry to explain observed trends; it is not monotonic and there is a spatial discontinuity (Figure 2.). Differences for the 1961-2011 period are minor, while the 1961-1981 and 1961-1971 reference periods results are spatially incoherent, indicating bad SPEI distribution fits.

The SPEI reference period must balance data non-stationarity, and model estimation errors. Maximizing these two requirements, a reference period from 1961 to 1991 or 2001 has demonstrated spatially coherent results, with sufficient deviation from the control period. This will incorporate recent climate change and result in higher SPEI intensity for our preceding decade which will reflect in the computed return periods of recent historical drought events, i.e. the drought risk.

Identifying drought indices that effectively predict future drought impacts remains a critical challenge in seasonal forecasting, as these indices provide the necessary actionable information that enables stakeholders to anticipate better and respond to drought-related challenges. This study evaluates how drought indices balance forecast skill and relevance for estimating European impacts. Using ECMWF SEAS5 seasonal predictions and ERA5 reanalysis as benchmarks, we assessed the predictability of drought indices over various accumulation periods and their relevance in estimating drought impacts across Europe to enhance impact-based forecasting (IbF). Our findings reveal higher predictability in Northern and Southern Europe, particularly during winter and summer, with some regions showing extended predictability for up to six months, depending on the season. Focusing on case studies in the UK and Germany, our results highlight regions and seasons where accurate impact predictions are possible. In both countries, high impact predictability was found up to six months ahead, with sectors such as Agriculture, Water Supply, and Tourism in the UK and Agriculture and Water Transportation in Germany, depending on the region and season. This analysis represents a significant step in identifying the most suitable drought indices for predicting European impacts. Our approach introduces a new method for evaluating the relationship between drought indices and effects and addresses the challenge of selecting indices for estimating impacts. This framework advances the development of operational impact-based drought forecasting systems for Europe.