Dunkelflaute: Dark, Windless Weeks

Dunkelflaute is a German word meaning dark doldrums, and it describes a stretch of weather that is overcast, still and often cold, when solar panels produce little and wind turbines produce almost nothing. Such spells can last from a day to more than a week and can cover much of a continent at once. They matter because they set the hardest test a wind and solar heavy grid has to pass.

The weather behind the word

The usual cause in northern Europe is a persistent high pressure system parked over the region in winter. High pressure brings light winds, which is the problem for wind power. It often brings low cloud or fog that sits for days, which is the problem for solar. In midwinter, days are short anyway, so even bright sunshine would deliver only a few useful hours.

The three conditions arrive together, and they are unhelpfully correlated with demand. The same blocking high that stops the wind often brings cold, clear nights, and cold weather raises electricity demand for heating. In systems where heating has been electrified, the effect is stronger still. So a Dunkelflaute is not just a period of low output. It is a period of low output coinciding with high demand.

The events are also large. A blocking high can span most of western and central Europe, so a country’s neighbours are frequently becalmed at the same time. That is the crucial detail. A system can be very comfortable with low wind in one place, because the wind is blowing somewhere else, and much less comfortable when there is no elsewhere within reach.

Why it is a planning problem and not a daily one

Grids deal with variable output every day. Forecasts are good, reserves are held, plants are started and stopped. A quiet evening is routine.

The difficulty with a Dunkelflaute is duration. It is not about covering an hour with the wrong wind speed, but about covering a week when a large share of the generating fleet contributes very little. Ember’s statistics show that wind and solar together supplied roughly 45% of Germany’s electricity in 2025. During a serious still, dark spell, a system in that position has to meet nearly all of its demand from everything else it owns, for days on end.

That defines the size of the backup fleet. Capacity planning is built around rare, difficult conditions rather than typical ones, in much the same way that flood defences are built for the exceptional year. The plants that cover these events may run for only a handful of days a year, which makes them expensive per unit of energy and awkward to pay for through energy prices alone.

How grids cope

Dispatchable plants held in reserve. Today this mostly means gas turbines, alongside coal in some countries, and reserves of plant kept outside the market for emergencies. They are the workhorses of a Dunkelflaute because they hold energy as fuel, and fuel keeps indefinitely.

Hydro reservoirs. Where they exist, they are the best answer available. A reservoir stores energy for months, discharges for as long as the water lasts and responds within seconds. Countries with large reservoirs ride out these events comfortably and often supply their neighbours while doing so.

Interconnection, with limits. Links to other grids help when the weather differs across the region. A high sitting over the North Sea may leave southern Europe windy, or leave a hydro rich neighbour with water to spare. But when the same high covers everyone, links deliver much less, and a plan that assumes imports will always be available is a plan that has not been tested against the hard case.

Nuclear and other firm capacity. Plants that run regardless of weather reduce the size of the gap that has to be covered, though they do not remove the need for something that can follow the remaining swings.

Demand response. Industrial users who can pause, heat that can be stored in buildings or water tanks, and charging that can be delayed all shave the peak. This works well for hours and less well for days, because most flexible demand is postponed rather than abandoned, and it comes back later.

Batteries, for the edges. Grid batteries are typically built to supply full output for two to four hours. They are excellent at the evening peak inside a difficult week, and cannot by themselves carry a week.

Long duration storage. Hydrogen produced from surplus power and burned or fuel celled later, pumped hydro, compressed air and thermal stores are all aimed at this gap. Pumped hydro is proven and limited by geography. The others are at varying stages of development and cost, and the honest position is that no cheap, scalable answer for multi-day storage is in wide use yet.

Reading claims about Dunkelflaute

Because the word describes a genuine constraint, it turns up in arguments on both sides of the energy debate. Some claims to treat carefully:

Low output does not mean no output. During most such spells, wind output is low rather than absent, and offshore fleets often keep producing when onshore turbines are still. Headlines that describe wind as switched off are usually describing a single low hour.

One event is not a trend. Every winter produces some difficult days, and how a system handles them tells you more than the fact that they happened.

The right question is what covered the gap and what it cost. A system that comes through on gas and imports has passed the test in a way that is workable but neither cheap nor clean. A system that comes through on reservoirs and firm low carbon capacity has passed it differently. Those outcomes look identical if you only look at whether the lights stayed on.

Dunkelflaute is best understood not as an argument against wind and solar but as the specification for what has to sit alongside them. A grid can get most of its energy from weather in most hours and still need a fleet, a fuel store or a reservoir for the weeks when the weather stops cooperating. Designing and paying for that fleet is one of the central problems of the transition, and the word is a useful shorthand for why it exists.