LONG-PERIOD VARIATIONS IN SEASONAL SEA-LEVEL PRESSURE OVER THE NORTHERN HEMISPHERE1

Publisher Website
Google Scholar
Abstract
The standardized departures of sea-level pressure from the normal over the Northern Hemisphere for each of the four seasons were examined by means of spectrum analysis. Emphasis was placed on interpreting the geographical distribution of relative spectral power at different periods, rather than the power distribution within the individual spectra. The predominant characteristic of the data was trend, found to represent real meteorological effects and not spurious results of changes in station locations or times of observation. Particularly in subtropical latitudes where the trend was greatest, the data were characterized by large 1- and 2-year-lag autocorrelations and pronounced “red noise” spectra. Exceptionally large 2-year-lag autocorrelation coefficients over several extensive regions of the hemisphere were related to marked departures of the data series from a linear first-order Markov process in those areas. Although not so strong in an absolute sense as the trend, several moderately strong... Abstract The standardized departures of sea-level pressure from the normal over the Northern Hemisphere for each of the four seasons were examined by means of spectrum analysis. Emphasis was placed on interpreting the geographical distribution of relative spectral power at different periods, rather than the power distribution within the individual spectra. The predominant characteristic of the data was trend, found to represent real meteorological effects and not spurious results of changes in station locations or times of observation. Particularly in subtropical latitudes where the trend was greatest, the data were characterized by large 1- and 2-year-lag autocorrelations and pronounced “red noise” spectra. Exceptionally large 2-year-lag autocorrelation coefficients over several extensive regions of the hemisphere were related to marked departures of the data series from a linear first-order Markov process in those areas. Although not so strong in an absolute sense as the trend, several moderately strong...

This publication has 0 references indexed in Scilit: