A codimension multifractal methodology was used to analyze and to model sca
lar concentration fluctuations within sulfur hexafluoride tracer gas plumes
from a line source in atmospheric surface-layer flows. Correspondence was
exhibited between the double trace moments parameters a and C-1 of the codi
mension methodology and the experimentally measured plume concentration cha
racteristics of peak-to-mean ratio and concentration fluctuation intensity.
Data series were generated using an extremal Levy, stochastic multifractal
model, with the experimental a and C-1 as inputs. Uncertainties in experim
entally determined plume characteristic values overlapped the uncertainties
in model-simulated values. The utility of the procedure includes 1) charac
terizing the state of scalar turbulent mixing, 2) helping to evaluate and t
o model hazardous plume concentrations, and 3) being able to estimate the p
robability of realizing extreme events at timescales of observation beyond
or at magnitudes in excess of those present in the actual observations.