Having to also predict solar yields increases overall prediction error. Shading from dust, smoke, clouds, squall..etc can cause substantial temporary and local changes in solar output not realistically captured by weather forecasting.
These kind of problems are far smaller than the changes to the grid that can happen through other seemingly random events like some particular TV show coming on; some particular group having a festival and wanting to cook more; an announcement of a special offer that makes many people drive their electric car to a shop.
There just are fluctuations in both demand and supply of the grid. On the scale of an individual solar panel, the fluctuations from a cloud can be huge. As much as 80% of power output. On the scale of a neighborhood, on the other hand, that becomes much less of a problem - individual panels come on and off but the overall generation remains almost the same and overall change is much slower. Scale up to a country and you just aren't worried. On the European grid this is scaled over the entire continent, with, for example, wind power gathered both in Scotland and Morocco and many places in between so that when the North Sea is calm, that's the time when Morocco will be generate more and also the opposite.
This article, from 2018 was exactly what I was ribbing you about with my 1990s joke. 2018 is honestly closer to the 1990s than today. Consider that, at the time, there were no renewable systems delivering synthetic inertia and that no grid scale batteries were installed at all. Synthetic inertia is now standard on all new commercial wind turbines and grid batteries are everywhere, often in the GWh scale.
Nuclear is a far more realistic option for decarbonization than solar. We don't even have the technology to support an economically feasible grid that runs entirely on renewables. Cherry picking temporary supply issues with whatever you don't like doesn't speak to anything meaningful.
In the 1990s we didn't have that technology. Solar and wind generated electricity is now so much cheaper than fossil fuel power, let alone the most expensive type of power, nuclear, that the only economic way to generate power for the future is to use them, even when that power needs multi-day storage. To even approach a reasonable economic level, nuclear plants must run continually at their full output, something that not only requires even more storage, but also means that you need a huge amount of controllable (rapidly and flexibly dispatchable, with many changes per day) renewable energy to compensate for the lack of flexibility. This means that even the standard "levelized" prices, often well over €100 per MWh, are in fact massively understated.
This is a fundamental misunderstanding of what decarbonization means. It's not about meeting current demand with renewables it is about shifting off-grid consumption of hydrocarbons to the grid. This requires massive reductions in cost and massive increases in grid capacity.
Here I do actually fully understand what decarbonization means and agree and support your goal. That means that there is absolutely a need for understanding and that the recent massive cost reductions in renewable energy are exactly what is needed. Nuclear is not useless, but it is hugely more costly (levelized cost, which includes use of storage and capacity factors, around $100, ten or more times as much as solar)
You can do a lot on the margins with solar but from the modeling I've seen the benefits are self-limiting. / As solar capacity increases you saturate supply and then you saturate ESS. There is no long term / seasonal technology akin to filling natural gas storage tanks for winter consumption and without it there is no realistic economically feasible avenue for decarbonization via renewables.
Neither solar, nor wind, nor hydro nor geothermal switch off completely in winter. Your capacity may definitely be reduced but you will still get some generation. What that means firstly is that you should install your solar panels at the level needed to generate most of your winter electricity and secondly that you should have diverse supply available. Increasing the installation theoretically takes costs per MWh from, say, €5 to €10 (that's approximately the levelized solar cost, which already includes this). Diversifying means that in almost all locations, in Winter you will have much more wind and geothermal power in your mix and less solar.
There is no long term / seasonal technology akin to filling natural gas storage tanks for winter consumption and without it there is no realistic economically feasible avenue for decarbonization via renewables.
Again, this is the 1990s view. There are a bunch of different deployed technologies at this point. Large scale heat stores which heat inert material like rocks or sand and keep that heat for months. Deep underwater or even simple on surface high pressure air storage facilities that can store energy almost indefinitely. Conversion into hydrogen which is getting more and more economic and can have annual losses at far less than 1% and round trip energy losses of only around 30%. Now there are other alternatives also coming into use. Especially more interesting technologies like sodium batteries and flow batteries look as if they will soon become economical for storing energy for weeks or months. Gravity stores, moving solid weights, can keep energy indefinitely. To some extent, it's a chicken and egg problem. The long term storage will only be installed once there is available cheap renewable energy.
The move to nuclear power itself would never have been possible without the simultaneous addition of technologies such as gas "peaker" plants and pumped hydro storage. Nuclear is extremely expensive, very inflexible and needs to be taken off the grid in huge (many MegaWatts) chunks either for maintenance or in case of unexpected and impossible to avoid failures, such as the jelly fish, but also hot weather, earthquakes, floods, terrorist attacks, coolant leaks and many many other causes. If a reactor SCRAM happens, such as happened in the Spanish incident, it can take weeks to bring a reactor back on line, so the requirement for storage or alternatives in the grid is much higher than for well designed renewables.
The key difference, however, is the overall marginal cost of electricity. If the average sale price of electricity will be say €4, typically because it is only storage that has enough demand to use it, it makes absolute sense to run renewables because the only real cost is a small amount of additional maintenance. On the other hand, for nuclear, this would absolutely not make sense. The plant needs a price of, say €60, just to cover fuel costs (more if long termdisposal is fully included). Sure, nuclear plants do run with such temporary prices, but only when they can make it up later. This makes them very unreliable as a source for filling storage.