An analysis of the impact of data inhomogeneities on the snow record

Snow is a significant factor in the economy and water resources of the United States. It is also an important indicator of the climate signal. Unfortunately, the physical properties of snow make it more difficult to measure accurately compared to other variables such as temperature and rainfall. Inconsistencies in observational practices over time can introduce inhomogeneities into the data record which can complicate the interpretation of long-term snow trends and climatologies. Daily snowfall data from Urbana, Illinois were analyzed in a case study to assess the impact of observational inconsistencies on several snowfall indicators. The suite of snowfall indicators includes total snowfall amount, greatest 1-day and 2-day snowfall amount, length of snow season, number of days with snowfall, and median daily snowfall. The analysis indicates that snowfall amount is most sensitive to inhomogeneities in the data, number of days with snow is less sensitive, and snow season length is least affected. Changes in observation time appear to be the most important inhomogeneity, along with exposure changes. Station relocations did not have as large an effect.