This list seems to include some incidents that are only tangentially related to wind power:
"Second of three public fatalities at the same location put down to driver distraction. Accident spot is where turbines become visible to drivers"
"34 year old was killed when his car collided with a truck carrying large wind turbine components"
"47-yr old man killed when the snowmobile he was driving stuck a fence around a wind farm construction site"
I wonder how they extrapolated from a list of accidents into deaths per 1,000TWh? Do you divide the estimated world output by the number of deaths? Do you 'weight' deaths by the year in which they happened and the output at that time? There's lots of ways to mangle those figures...
No idea, but the other issue is are the same criteria being applied for each source of power? I've got no problem accepting that industrial accidents happen with wind turbine manufacture, installation and maintenance. Also that it's reasonable to include all those in the figures, after all if a particular form of power generation involves a lot of industrial accidents, that's a problem. There's just not enough info to see if these figures are truly comparable. Maybe they are, I just can't tell and on a highly charged issue like this you really need to be able to trust the figures you're looking at.
http://www.caithnesswindfarms.co.uk/fullaccidents.pdf