Wednesday, 17 January 2001: 1:45 PM
An analysis of global monthly precipitation has been constructed on a 2.5 degree latitude/longitude grid over the global land areas for a 50-year period from 1950 to the present by interpolating gauge observations at over 20,000 stations. The gauge observations used here are collected from two sources: the version 2 data set of the Global Historical Climatology Network (GHCN) of NOAA/NCDC and
the Climate Anomaly Monitoring System (CAMS) of NOAA/CPC. First, gridded fields of monthly climatology of precipitation are defined from observations at over 17,000 gauge stations with 10 years or longer records during the 40-year period from 1951-1990. Gridded fields for each month are then calculated by interpolating all station observations available for that month using the Optimal Interpolation (OI) algorithm with the monthly climatology as the first guess. Cross validation was conducted and the results showed stable and improved performance of the gauge-based analysis for all seasons and over most areas over the global land, due to the inclusion of more gauges as inputs and the application of the OI algorithm. The 50-year gauge-based analysis is applied to investigate the annual and inter-annual variability of large-scale precipitation over the global land areas. In particular, since about half of the 50-year period has identifiable ENSO episodes, inter-ENSO variability is examined using the new data set.
Over the global oceanic areas, the analysis is defined by projecting the historical gauge observations over coastal regions and islands onto EOF patterns determined from satellite-based precipitation estimates for a 20-year period from 1979 to 1998. The construction of the oceanic monthly precipitation analysis is in process. The research work so far has shown encouraging results of the reconstructed fields in representing inter-annual variations of large-scale precipitation over ocean, especially over tropical central and western Pacific.
![](images/flag.png)
![](images/blue_ribbon.png)