Uncertainty of industrial noise measurement at distant locations from the source

Methodologies for industry noise measurement usually include the range of weather conditions
in which the measurements must be taken. The effects of meteorological conditions
on sound propagation are small for short distances, and larger for longer distances at greater
receiver and source heights.
One can find some algorithms in ISO 9613-1,2 for calculation of weather conditions impact
on community noise, so called Cmet, but especially wind correction, is rather poor, limited to
only two cases; moderate downwind and a variety of meteorological conditions as they exist
over months or years. The problem begins in calculating long-term average A-weighted level
using short-term data with unknown detail weather conditions on the path of sound propagation.
The paper deals with some real word data of partial uncertainties of noise prediction and
measurement from large industry and impulse sound sources, taken in different meteorological
conditions. It has been shown that in some cases maximum spread of the data exceeds 20
dB with the same state of industry running and completely similar weather conditions. In case
of the impulse sources it has been shown uncertainty analysis for the impulse sound power
and sound exposure level at reference distance of 1 km.