Submitted to: American Geophysical Union
Publication Type: Proceedings
Publication Acceptance Date: October 1, 1999
Publication Date: October 1, 1999
Interpretive Summary: Historical records of precipitation for a region and at individual stations are used to relate regional and local precipitation characteristics. Such relationships are necessary to utilize regional climate forecasts in local agricultural applications. In a first analysis, the accuracy and shortcomings of regional precipitation due to record discontinuities was established. The findings showed that discontinuities in the historical precipitation records can lead to difference of about 10% of monthly precipitation and 20% of monthly variations. In a second analysis, the size of the spatial variability of monthly precipitation at stations over a large region was also quantified. It was found that the spatial variability over the region is equivalent to 36% of the mean variation in time. This spatial variability explains the observed large localized departures from regional values, and provides an estimate of the expected degradation of information content of regional precipitation forecast as one applies the forecast information to local scales.
Technical Abstract: The utilization of emerging climate forecast information for agricultural applications requires the use of the historical National Climate Data Center (NCDC) divisional precipitation index to establish the link between regional and local distributions of monthly precipitation values within the climate division. To this end the accuracy and shortcomings of the NCDC index due to record discontinuities must first be established. The mean absolute differences between a continuous and a hypothetical discontinuous record vary from 6 to 13% of the mean monthly precipitation and 9 to 24% of their mean temporal variations. These values demonstrate that the discontinuities in station records of the NCDC divisional precipitation index can be relevant in time series analysis and result interpretation. The size of the spatial variability of monthly precipitation values within a climate division was also quantified. Twenty five percent of the spatial ldifferences between the standardized values of stations and index are greater than half a standard deviation of the temporal time series, with extreme values near three standard deviations. The average of the absolute differences is equal to 0.36 which implies that the spatial variability within the climate division is equivalent to 36% of the mean temporal variation. This spatial variability explains the observed large localized departures from divisional values, and provides an estimate of the degradation of informational content as one applies divisional precipitation information to local scales.