Record estimation and extension for precipitation data

The desire to use data from climate stations installed in the 1980's for streamflow volume regression equations required an estimating procedure to fill in missing data and extend the records back to 1961 to create a complete 30 year record. The past success of using organic regression MOVE 1 and MOVE 2 procedures in extending streamflow records (Hirsch, 1982; Alley and Burns, 1983) led to the development of a similar application for precipitation data. AMOVE 2 estimation procedure was developed for monthly precipitation data which uses : maintenance of variance techniques to minimize variance bias, and preserves seasonal variation in precipitation. Results using this procedure to estimate precipitation for climate stations located in the Bridge River system in British Columbia are presented and compared to estimates made using simple linear regression.