A method of comparison of two close batches data: Application to analysis of fog formation causes

Given suitable conditions of air temperature and humidity, the density of a fog and its microphysical properties depend mainly on the availability of cloud condensation nuclei (CCN) and their nature. Fogs become particularly dense near certain industrial plants because of high concentration of hygroscopic combustion particles in the air. Their role in dense fog formation is estimated by comparing the local climates and CCN concentrations at two similar sites, the first being more subject to air pollution and dense fogs than the second. Orthogonal regression is applied to three meteorological variables (air temperature, relative humidity, wind speed) and CCN concentration. As we compare very close variables, bootstrap provides precise confidence intervals independent of Gaussian assumptions. Two sites are compared: they are located in the Meuse valley (Belgium) at a distance of about 15 km. We found that the local climate of the polluted site is not only colder and wetter but also richer in CCN that the control site. These results suggest interactions of natural and anthropogenic causes in dense fog formation at industrial site. This method is useful in various domains of geophysics when correlated time series have to be compared