The ISO 9223 standard indicates that there should be no corrosion at temperatures below 0 degrees celsius. Independent researchers have however proposed to lower the minimun temperature stated in the standard to lower values in order to account for the actual corrosion observed in Nordic climates. What is the explanation for the observed corrosion at temperatures below the freezing point?
As Dan mentioned in the comments, the presence of salt (and other pollutants) can alter the freezing point of water.
The ISO 9223 abstract states:
key factors in the atmospheric corrosion of metals and alloys. These are the temperature-humidity complex, pollution by sulfur dioxide and airborne salinity.
All of which are factored into the "time of wetness", which can increase in the presence of salinity (and other pollutants, acknowledged in the ISO standard itself) as found in observations made in Antarctica published in Uhlig's Corrosion Handbook, (p. 330):
..that in the presence of marine salt, liquid water monolayers could form under ice layers resulting in high corrosion rates in temperatures well below 0C
Sulfur dioxide is also identified in the ISO abstract, this a common pollutant from vehicles, industry etc. which according to can chemically react with water vapour, forming sulfuric acid, the basis of acid rain and acid snow.