RESUMEN
Ambient gamma dose, radon, and rainfall have been monitored in southern Bucharest, Romania, from 2010 to 2016. The seasonal cycle of background ambient gamma dose peaked between July and October (100-105â¯nSv h-1), with minimum values in February (75-80â¯nSv h-1), the time of maximum snow cover. Based on 10â¯m a.g.l. radon concentrations, the ambient gamma dose increased by around 1â¯nSv h-1 for every 5 Bq m-3 increase in radon. Radon variability attributable to diurnal changes in atmospheric mixing contributed less than 15â¯nSv h-1 to the overall variability in ambient gamma dose, a factor of 4 more than synoptic timescale changes in air mass fetch. By contrast, precipitation-related enhancements of the ambient gamma dose were 15-80â¯nSv h-1. To facilitate routine analysis, and account in part for occasional equipment failure, an automated method for identifying precipitation spikes in the ambient gamma dose was developed. Lastly, a simple model for predicting rainfall-related enhancement of the ambient gamma dose is tested against rainfall observations from events of contrasting duration and intensity. Results are also compared with those from previously published models of simple and complex formulation. Generally, the model performed very well. When simulations underestimated observations the absolute difference was typically less than the natural variability in ambient gamma dose arising from atmospheric mixing influences. Consequently, combined use of the automated event detection method and the simple model of this study could enable the ambient gamma dose "attention limit" (which indicates a potential radiological emergency) to be reduced from 200 to 400% above background to 25-50%.