"Fatalities per kw" is terrible risk management. I really wish nuclear activists would stop making crappy arguments.
Risk analysis has two axis - likelihood, and impact. The maximum potential impact of solar/wind plant failure is negligible - what's it going to do, fall over on someone? Meanwhile, the maximum potential impact of virtually any civilian nuclear plant is massive - release of toxins and carcinogens over large residential areas, rendering them uninhabitable.
Any respectable risk analysis requires looking at impact, not just likelihood.
The long term data (over decades) shows that nuclear kills fewer people than other forms of power generation. Maybe that will change over time (fewer people falling off of roofs when installing solar).
That data includes "high impact" events, including things like Fukushima, which evidently has not resulted in any radiation-related fatalities.
I'm not just measuring fatalities. I'm measuring economic costs. A nuclear accident can render a city uninhabitable, even if no one dies.
edit: An additional risk is that a nuclear plant is much more complex and expensive than a solar plant. A major accident, if nothing else, wrecks the plant, taking away a key resource that could take a decade and billions of dollars to rebuild (if it's politically possible to rebuild at all). Solar and wind, made up of collections of small units, can't suffer the same kind of unrecoverable catastrophic failure. Again, economic risk.
And yet another risk... because of the potential destructive nature, nuclear plants are obvious targets for terrorists and military action. This ties back to the ability to disable a plant and its impact. Even if it doesn't cause fatalities, taking a nuclear plant offline and forcing evacuations isn't that hard. High, high economic risk.
This is another example of your original mistake. You're narrowly focusing, and not doing a solid risk analysis. Nuclear power is riskier than solar, because the potential impact of failure is much higher.
> release of toxins and carcinogens over large residential areas
Thought experiment: how big of a nature-only walled-off preserve would we need around a nuclear plant in order to protect against that? And if that amount of area were instead covered with solar panels, how much energy would that produce?
My impression given the efficiency of nuclear is that it'd still win, plus you get a sweet nature reserve out of it.
I'm all for investment in nuclear but it would take a really bloody colossal nature preserve. The Chernobyl exclusion zone is only 2,600 km^2 but about 3,900,000 km^2 of Europe - about 40% of it's total land mass - was contaminated with caesium-137 after that disaster and about 2% of Europe was contaminated to 'high' levels. Even as far away as the UK there are restrictions on hundreds of farms related to that radiation.
The fear is irrational, but so is the cost. Nuclear always comes in over budget and with far greater construction delays than any other generation type.
Even if you were to start on new reactors today, it would take a decade to complete them; a whole lot of wind and solar are going to be deployed in a decade.
Measured by fatalities per k/W generated, nuclear is by far the safest form of power. The fear of nuclear is irrational.