Multivariate stochastic modelling of grass fluoride and airborne fluorides

Stochastic modelling was used to investigate the influence of airborne fluorides (in the case of grass fluoride) and meteorological factors on both grass and airborne fluoride levels, taking into consideration lagged dependence. Using rigorous grass sampling methods, there were indications of a negative relationship between volume of rainfall and grass fluoride concentrations, but no apparent relationship between the time which the leaf was visibly wet and grass fluoride levels. No evidence was found that rainfall washed airborne fluoride from the atmosphere or that wind variables influenced grass fluoride. For one particular experimental site, a multivariate stochastic model, which accounts for 81% of the total variation of the grass fluoride series, is developed.