Probabilistic consideration of consequences of long-duration accidental releases based on complete weather data.

This paper presents a method based on chronological weather data for estimating how variations in weather conditions during a radioactive release affect the consequences of the release. The modifications of the method made in the computer-based applications owing to the limited calculation resources are also studied. Additionally, the length of the chronological weather data observation period needed to satisfactorily support the assessments was tested. The results of the different duration times for the normalized concentration--which is the concentration divided by the release rate, and called here the dispersion factor--are presented and compared with the commonly used sector average model and the generic wind meandering method. The new method can be used for a variety of release durations and it makes realistic estimates, managing to avoid the weaknesses of the methods compared. The necessary length of the weather data is on the order of 1 y to get statistically satisfactory results.