The CRAC2 compute^ cpde is g revised version of CRAC (Calculation of Reactor gL,ccident Consequences) which was developed for the Reactor Safety Study. This document provides an overview of the CRAC2 code and ^ description of each of the models used. Significant improvements incorporated into CRAC2 include ap improved weather sequence sampling technique, a new evacuation model, and new output; capabilities. In addition, refinements have beep made to tl^e atmospheric transport and deposition model. Details of the modeling differences between CRAC2 and CRAC are emphasized in the model descriptions.

the results of such use, of any information, apparatus product or process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights.

A model of public evacuation is developed for use in evaluating the efficacy of evacuation as a protective measure in response to atmospheric releases of radioactive material. Differences between this model and the model of public evacu ation previously developed for the Reactor Safety Study p^e described. Based on an analysis of available EPA evacuation data, ranges are suggested for the temporal parameters in the new model. The relative importance of the model paramett.s is also discussed.

Reactor accident consequence models have been developed (for example, the CRAC model of the Reactor Safety Study (RSS), WASH-1400) to predict the offsite health and economic consequences of severe accidents at a reactor site with generic demographic and meteorological characteristics. Application of a revised RSS accident consequence model, CRAC2, to 91 existing sites results in a band of risk curves around the earlier WASH-1400 average reactor/site predictions. This paper examines these calculations and important model assump tions such as population distribution, emergency response, and meteorological data with respect to their effects on site risk extremes-that is, the combination of high consequence/low probability events.