Evaluation of different methods for detecting 'kalte Dunkelflaute' events with respect to climate change projections

Weather-driven periods of low electricity production from renewable energy sources (RES) can result in so-called ‘energy droughts’, sometimes known by the German term ‘Dunkelflaute’. When these weather phenomena occur over a large geographical area for extended periods of time, and coincide with periods of high electricity demand due to cold temperatures (‘kalte Dunkelflaute’), these events pose a risk for maintaining resource adequacy in a future power system relying significantly on RES.

The most robust way of identifying energy droughts is to use hourly electricity market simulations to capture both the demand- and supply-side effects of different climate years, but these simulations are computationally intensive to perform and Transmission System Operators (TSOs) can typically only consider 30-40 historical climate years in resource adequacy studies. However, as vastly more climate data is becoming available from future climate projections, a robust way to identify ‘kalte Dunkelflauten’ from climate data alone is needed in order to identify challenging years to consider in resource adequacy studies. A variety of approaches for defining and analysing energy droughts can be found in the literature such as those which detect singular events by defining a threshold for renewable energy production (production drought) or how much (net) load is covered by RES (supply drought) (e.g. Raynaud et al., 2018), and those which use statistical methods to assess the risk of an energy drought within a predefined timespan (e.g. Ruhnau & Qvist, 2022). However, there is no clear consensus on which is the best method.

In this study we will present an evaluation of different methods to assess the risk of occurrence of such ‘kalte Dunkelflaute’ events, and validate these methods by comparing with results from detailed hourly simulations, with a focus on the Netherlands and Germany. By applying different detection methods to both existing and projected RES capacity, and using both historical and future climate data from a Pan-European Climate Database, we compare past and future risks posed by energy droughts.  As extreme ‘Dunkelflaute’ events are rare but their impact may be severe, comparing different approaches of how to statistically evaluate these events is an important contribution to evaluating resource adequacy, and assessing the resilience of the future energy.

 

References

• Raynaud, B. Hingray, B. François & J.D. Creutin (2018). Energy droughts from variable renewable energy sources in European climates. Renewable Energy, 125, 578-589, https://doi.org/10.1016/j.renene.2018.02.130.
• Ruhnau & S. Qvist (2022). Storage requirements in a 100% renewable electricity system: extreme events and inter-annual variability. Environmental Research Letters, 17(4), 044018, https://doi.org/10.1088/1748-9326/ac4dc8