Variability and intermittency in generation

node

The property that distinguishes wind and solar from every generator that burns something: their output is set by the weather rather than by an operator. It is worth separating three distinct problems that the single word "intermittency" conflates. The first is *variability* — output changes continuously, predictably in the case of the solar day, less predictably in the case of wind — and is handled by forecasting, reserve and flexible plant. The second is *diurnal mismatch*: solar produces at midday and demand peaks in the evening, a gap of hours that storage genuinely solves. The third is *sustained low output*: multi-day, region-wide conditions in which neither wind nor solar produces much, which is a problem of energy rather than power and which four-hour batteries do not address at any price. Capacity factors frame the scale of the task — solar photovoltaics typically deliver something like ten to twenty-five per cent of nameplate over a year depending on latitude and climate, onshore wind roughly twenty-five to forty-five per cent, offshore wind higher again. The cost of variability is therefore not a single number but a function of penetration: negligible at low shares, rising as the cheap flexibility is used up, and dominated at high shares by the rarest and least tractable events rather than by the average day.

A node is not a place. Drawing it on a map would assert something about the world that no stored fact supports.