Statistical predictability and spectra of air temperature over the northern hemisphere

Parametric models and statistical predictability of mean annual air temperature, both at individual stations (MAAT) and zonally averaged (MZAT), are evaluated. The time series have lengths from 125 to 315 years for MAAT and 86 years (1891-1976) for MZAT. The optimal parametric models of MAAT are sho...

Full description

Saved in:
Bibliographic Details
Published in:Tellus. Series A, Dynamic meteorology and oceanography Dynamic meteorology and oceanography, 1983-01, Vol.35 (1), p.51-59
Main Author: Privalsky, V. E.
Format: Article
Language:English
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Parametric models and statistical predictability of mean annual air temperature, both at individual stations (MAAT) and zonally averaged (MZAT), are evaluated. The time series have lengths from 125 to 315 years for MAAT and 86 years (1891-1976) for MZAT. The optimal parametric models of MAAT are shown to be closely approximated by a first-order Markov sequence with a small characteristic time scale, so that their statistical predictability is minimal. Variations in MZAT should be described with more complicated models; the limits of statistical predictability amount to between 4 and 5 years. The temperature spectra are estimated, and the relatively high statistical predictability is shown to result from the concentration of spectral energy at low frequencies.
ISSN:1600-0870
1600-0870
DOI:10.3402/tellusa.v35i1.11418