If you study even high school economics for a single term, you will find that you can't just divide initial cost by operating lifetime to amortize it. There is this little thing called TIME VALUE OF MONEY. A few percent per year, carried over 40-60 years, really adds up.
You also have to count insurance. Even if it is partially or wholly subsidized by the government, SOMEBODY is paying. The worst disaster (leaving aside normal environmental pollution) that could possibly happen at a coal or oil or gas or solar plant is pretty much limited to the plant premises. I'll grant you that hydropower is capable of making vast areas wasteland and killing untold thousands if a big dam bursts. Otherwise, nuclear has a downside potentially thousands of times more devastating than the others. Insurance, fairly accounted, has to cover this.
It would take me hours to decide if the study you reference really accounts for all costs. Just at a glance, it LOOKS like they are properly accounting for amortization, but I see no mention of insurance. I do see the notation for nuclear: "does not reflect decommissioning costs or potential economic impact of federal loan guarantees or other subsidies".