Precipitation Estimation in Canada using Archival Climate Data

Hourly measurements of relative humidity and temperature along with observations of daily precipitation taken during 1977-2002 were used to compute the probability of various precipitation amounts. These data were taken at 75 stations located throughout Canada. The algorithm was tested by randomly dividing the data into two datsets A and B. Dataset A was used to determine the precipitation probabilities. Precipitation for dataset B was then estimated and compared to the observed amounts. The observed and estimated precipitations correlate very closely. This model can be customized for a given season and/or a particular geographic region. It appears that precipitation on the prairies and Arctic is much more sensitive to a temperature increase than is the case for other regions of Canada.