Box-jenkins univariate modelling for climatological time series analysis: An application to the monthly activity of temperature inversions
✍ Scribed by T. D. Davies; A. E. Milionis
- Publisher
- John Wiley and Sons
- Year
- 1994
- Tongue
- English
- Weight
- 783 KB
- Volume
- 14
- Category
- Article
- ISSN
- 0899-8418
No coin nor oath required. For personal study only.
✦ Synopsis
Abstract
Box‐Jenkins modelling has some advantages over other techniques for the analysis of time series of climatological variables. Not only does it provide more information than other methods of anlysis, in a more elegant way, but it is also perfectly acceptable from the mathematical point of view. Other methods may not be immediately applicable because of the problem of autocorrelation in time series. The method of Box‐Jenkins univariate modelling is briefly discussed. As an example of its application to climatological time series analysis, and as an illustration of its usefulness, we examined the monthly activity of temperature inversions over Hemsby (eastern England) over a 14‐year period. The results showed that the monthly activity series, for both surface and elevated inversions, are stationary. However. the series for surface midnight inversions has a seasonal non‐stationarity of lag 12. There is a 1Zmonth seasonality for surface inversions and a weaker 6‐month seasonality for elevated inversions. The monthly activity of surface inversions exhibits less variation than the monthly activity of elevated inversions. This simply reflects the fact that the physical processes responsible for the formation of surface midnight inversions have a more regular evolution overtime than those responsible for the formation of elevated inversions. These results are in accordance with those obtained by using standard statistical techniques.