A simple model to estimate deposition based on a statistical reassessment of global fallout data

Atmospheric testing of nuclear weapons began in 1945 and largely ceased in 1963. Monitoring of the resulting global fallout was carried out globally by the Environmental Measurements Laboratory and the UK Atomic Energy Research Establishment as well as at national level by some countries. A correlation was identified between fallout deposition and precipitation and an uneven distribution with latitude.

In this study, the available data from 1954 to 1976 for 90Sr and 137Cs were reanalysed using analysis of covariance (ANCOVA) and logarithmically transformed values of the monthly deposition density as the response variable. Generalized additive models (GAM) were used to explore the relationship of different variables to the response variable and quantify the explanatory power that could be achieved.

The explanatory variables which consistently explained most of the variability were precipitation at each site, latitude and change with time and a simple linear model was produced with similar explanatory power as the GAM. The estimates improved as the temporal resolution of the precipitation data increased.

A good logelog fit could be obtained if a bias of about 1e6 mm precipitation per month was added, this could be interpreted as dry deposition which is not otherwise accounted for in the model. The deposition rate could then be explained as a simple non-linear power function of the precipitation rate (r0.2e0.6 depending on latitude band).

A similar non-linear power function relationship has been the outcome of some studies linking wash-out and rain-out coefficients with rain intensity. Our results showed that the precipitation rate was an important parameter, not just the total amount. The simple model presented here allows the recreation of the deposition history at a site, allowing comparison with time series of activity concentrations for different environmental compartments, which is important for model validation. © 2012 Elsevier Ltd.

All rights reserved.