The numbers given are in case of a class-9, or worse case scenario meltdown, and are based on 1982 population data and on 1982 dollars.  This report was mandated by the Nuclear regulatory Commission and carried out by the Sandia Labs of New Mexico. The Calculation of Reactor Accident Consequences (CRAC-2) report was published by Congress November 1, 1982. It was also printed by the Washington Post the same day. Other major media, including the New York Times published it shortly thereafter. 

Some experts claim that the assessment of the dangers inherent in any and all commercial nuclear power plants has several faults. Among the faults -- any accident can spread to the spent fuel pool where huge amounts of radioactive waste are stored.  The authors of the Reactor Safety Study concluded that changing some of the criteria for data gathering would actually increase the number of early fatalities by a factor of 3 to 4 depending upon circumstances [NUREG-0340]. 

"Peak" does not necessarily mean worst case results because the CRAC-2 model is acknowledged by its authors to have uncertainties in its meteorological modeling capability. Since the CRAC-2 model considers only one year's worth of data and does not model precipitation frequency beyond a distance of 30 miles from the reactor, the model cannot adequately characterize the frequency of precipitation events or range of fallout. 

Economic costs, not included are: the cost of providing health care to the affected population; all onsite costs; litigation costs; direct costs of health effects; and indirect costs.

Early fatalities are deaths due to radiation exposure from causes other than cancer occurring within one year of the accident. However, fatalities will continue over hundreds, possibly thousands of years. Fatalities in the populace directly exposed to the accident will take place over a period long after one year, much of these deaths taking decades to occur. Ionizing radiation can cause aberrations in the genetic pool, hence the hundreds or thousands of years over which fatalities will occur. 

It should be noted that the evacuation model for CRAC-2  does NOT account for actual site conditions such as bottlenecks and terrain barriers which can cause major evacuation routes to overlap the area likely to be covered  by the plume once a release of radioactivity occurs. 

You can get a copy of the report from the Nuclear Information Resource Service at: Phone:202-328-0002, Fax:202-462-2183.