[cesarb]

> though I am not sure that any of the results should surprise anyone [...] none of the "clean" sources can easily scale on demand. Sun, wind, hydro and nuclear have either fixed or random output

That does surprise me a lot. Here in Brazil, it's hydro which follows the load. Nuclear has fixed output (always the maximum except when it's offline), non-nuclear thermal is mostly constant (with large changes whenever generators are dispatched on or off), solar and wind do their own thing, and hydro is the one left to fill the gaps between generation and demand. See it for yourself at http://www.ons.org.br/paginas/energia-agora/carga-e-geracao (select "SIN" on the first drop-down for the whole country); selecting "Geração Hidráulica" (hydro) on the second drop-down you can see how well the shape of the generation curve matches the shape of the demand curve at the top, while for instance "Geração Térmica" (non-nuclear thermal) is nearly flat.