Use of principal component analysis to identify homogeneous precipitation stations for optimal interpolation
Article first published online: 9 JUL 2010
Copyright 1979 by the American Geophysical Union.
Water Resources Research
Volume 15, Issue 6, pages 1841–1850, December 1979
How to Cite
1979), Use of principal component analysis to identify homogeneous precipitation stations for optimal interpolation, Water Resour. Res., 15(6), 1841–1850, doi:10.1029/WR015i006p01841., , , , and (
- Issue published online: 9 JUL 2010
- Article first published online: 9 JUL 2010
- Manuscript Accepted: 23 APR 1979
- Manuscript Received: 18 DEC 1978
The Eaton River precipitation network was set up as one of the Quebec International Hydrologic Decade projects in 1965. At the end of the decade, in 1975, 30 stations had been in operation for most of the period in this 248-mi2 (643 km2) basin. After a principal components analysis of the data from 14 of the 30 stations on 10-day precipitation totals, it was found that the stations could be divided into three groups, the composition and geographic distribution of which change from season to season. It was also possible to identify which stations were redundant and could be closed should it become necessary. Application of optimal interpolation, season by season, to a 30- and a 5-station network showed that the precision of point interpolation varies more from season to season for the same network than from the 30- to the 5-station network for the same season. In fact, errors of interpolation caused by microclimatic irregularities and by observational errors in the initial data are greater than those resulting from the reduction of the network from 30 to 5 stations.