There have been several attempts to correlate environmental and site-specific variables with atmospheric corrosivity, including the PACER LIME algorithm and the ISO 9223 classification of corrosive atmospheres. However, seasonal variations in corrosivity at the locations tested in this study cannot be readily explained by either of these approaches. The monthly or seasonal corrosivity variations do not track well with monthly variations in time-of-wetness (TOW) or SO2 concentration. It is possible that the approach of averaging TOW, SO2, or chloride deposition rate over a year- or monthlong period hides shorter-term events that are truly influential. Therefore, while seasonal or month-long measurements of corrosivity provide more insight than yearlong measurements, it seems necessary to reduce the time scale down to days or hours before clear correlations can be made between corrosivity and corrosion drivers. Also, in urban settings where roads are de-iced, it is necessary to measure chloride deposition rates since it is a significant corrosion driver. Another conclusion is that accounting for the effects of insolation from the sun and convection on the surface temperature of an exposed metal provides more insights than time-of-wetness calculations, which relies on relative humidity alone.
Keywords: atmospheric corrosivity, seasonal changes, pollution, time-of-wetness