Three consecutive dry summers in western Europe (2018–2019–2020) had widespread negative impacts on society and ecosystems, and started societal debate on (changing) drought vulnerability and adaptation measures. We investigate the occurrence of multi-year droughts in the Rhine basin, with a focus on event probability in the present and in future warmer climates. Additionally, we investigate the temporally compounding physical drivers of multi-year drought events. A combination of multiple reanalysis datasets and multi-model large ensemble climate model simulations was used to provide a robust analysis of the statistics and physical processes of these rare events. We identify two types of multi-year drought events (consecutive meteorological summer droughts and long-duration hydrological droughts), and show that these occur on average about twice in a 30 year period in the present climate, though natural variability is large (zero to five events can occur in a single 30 year period). Projected decreases in summer precipitation and increases in atmospheric evaporative demand, lead to a doubling of event probability at 1 $$^\circ$$ ∘ C additional global warming relative to present-day and an increase in the average length of events. Consecutive meteorological summer droughts are forced by two, seemingly independent, summers of lower than normal precipitation and higher than normal evaporative demand. The soil moisture response to this temporally compound meteorological forcing has a clear multi-year imprint, resulting in a relatively larger reduction of soil moisture content in the second year of drought, and potentially more severe drought impacts. Long-duration hydrological droughts start with a severe summer drought followed by lingering meteorologically dry conditions. This limits and slows down the hydrological recovery of soil moisture content, leading to long-lasting drought conditions. This initial exploration provides avenues for further investigation of multi-year drought hazard and vulnerability in the region, which is advised given the projected trends and vulnerability of society and ecosystems.

<p>The summer of 2018 in North-Western Europe was exceptionally warm and dry, which negatively impacted many sectors. The drought of 2018 was followed by the dry summer of 2019 and the dry spring of 2020. Such multi-year droughts bring unique challenges to the agricultural sector, water authorities and society, and require different adaptation strategies compared to ‘normal’ single-year droughts. The succession of these dry years raises a question: is it pure coincidence that North-Western Europe experienced such a multi-year drought, or are there physical processes that cause multi-year droughts? Furthermore, in the present era it is obvious to ask whether anthropogenic climate change will amplify multi-year droughts in the region.</p><p>We aim to find drivers of multi-year droughts by using <em>ERA5 reanalysis</em> data and  state-of-the-art <em>Large Ensemble simulations</em> from seven climate models. We select multi-year droughts in these datasets based on the <em>Standardised Precipitation and Evapotranspiration Index</em> and compare drought characteristics in the 1991-2020 reference period with multi-year droughts towards the end of the century. The models show a strong increase in multi-year drought risk from present-day to the end of the century. The frequency of multi-year droughts near doubles and the median duration of selected drought events increases from 16 months to 50 months. Model differences are substantial, mostly due to differences in temperature trends, but all models agree on the increase in multi-year drought risk. Internal variability is large, indicating a large ensemble approach is indeed required to study this problem.</p><p>Next we discuss geophysical drivers of multi-year droughts. Slow-varying ocean processes (through sea surface temperatures) and land processes (through soil moisture) are investigated as potential sources of meteorological conditions that lead to multi-year droughts. We consider the full Earth system, including ocean-land-atmosphere feedbacks, as potential forcing for these events. Summarizing, we will show that anthropogenic warming has potentially large impacts on the frequency, duration and therewith societal risk of multi-year droughts, warranting detailed studies of this topic.</p>

<p>Three consecutive dry summers in western Europe (2018-2019-2020) had widespread negative impacts on society and ecosystems, and started societal debate on (changing) drought vulnerability and needs to revise adaptation measures. To facilitate that discussion, we investigate multi-year droughts in the Rhine basin, with a focus on event probability in the present climate and in future warmer climates. Additionally, we studied the temporally compounding physical processes leading to multi-year drought events. A combination of multiple reanalysis datasets and multi-model large ensemble climate model simulations was used to robustly analyse the statistics and physical processes of these rare events. In these data, we identify two types of multi-year drought events (consecutive meteorological summer droughts and long-duration hydrological droughts), and show that these occur on average about twice in a 30 year period in the present climate, though natural variability is large (zero to five events in a single 30 year period). Projected decreases in summer precipitation and increases in atmospheric evaporative demand, lead to a doubling of event probability in a world 1 °C warmer than present and an increase in the average length of events. Consecutive meteorological summer droughts are forced by two, seemingly independent, summers of lower than normal precipitation and higher than normal evaporative demand. The soil moisture response to this temporally compound meteorological forcing has a clear multi-year imprint, resulting in a relatively larger reduction of soil moisture content in the second summer and potentially more severe drought impacts. Long-duration hydrological droughts start with a severe summer drought followed by lingering meteorologically dry conditions. This limits and slows down the recovery of soil moisture content to normal levels, leading to long-lasting drought conditions. This initial exploration provides avenues for further investigation of multi-year drought hazard and vulnerability in the region, which is advised given the projected trends and vulnerability of society and ecosystems.</p>

<p>In recent years the surface temperature at high latitudes – especially in the Arctic region – warmed faster than the global average. This process called Polar amplification is a key characteristic of anthropogenic global warming. Polar amplification will reduce the equator-to-pole temperature gradient of the lower troposphere, but the consequences to the atmospheric general circulation are not fully understood. Here, we study the effect of the reduced meridional temperature gradient by enhanced Sea Surface Temperatures (SSTs), primarily at the Polar regions, on the atmospheric general circulation by executing idealised aquaplanet simulations with the Open Integrated Forecast System (OpenIFS) from the European Centre for Medium-Range Weather Forecasts (ECMWF).</p><p>Firstly, we show the global atmospheric response to Polar amplification by comparing various atmospheric climatologies of the control and Polar amplification simulations. We found that Polar amplification influences the Hadley cell circulation and jet stream: both key features of the atmospheric general circulation are strongly reduced in strength. Secondly, we studied the influence of Polar amplification on jet stream waviness. In this study, waviness is quantified by the Sinuosity Index. Sinuosity decreases in a statistically significant manner under Polar amplification: the jet stream becomes less wavy and the meridional extend of the zonal flow decreases as well. The perturbed SSTs around the Polar edge of the Hadley cell and in the baroclinic zone found to be key. Our result contradicts results found in other studies on the consequences of Polar amplification: potential reasons for this and the implication of our results will be discussed in this study.</p>