| SOMETCUBA Bulletin |
Volume 6 Number 1 |
January 2000 |
|
|
THE CHANGE-POINT INSTABILITY OF CLIMATOLOGICAL TIME-SERIES AS ALTERNATIVE TO RANDOMNESS. THE EXAMPLE OF ANNUAL TEMPERATURE AVERAGES 1908 - 1995 AT CASABLANCA (CUBA).
| by |
Raymond Sneyers |
Institute Royal Météorologique
de Belgique. E-mail: Raymond.Sneyers@kmi-irm.oma.be |
| and |
Lourdes Alvarez Escudero |
Institute
of Meteorology Fax: (537) 33 80 10 E-mail: lulu@.met.inf.cu |
Key words: randomness, ranks, change-points, time-series analysis, ties, test efficiency, rank and parametric tests, testing normality, likelihood-ratio tests, general circulation model, observation inhomogeneities, men made effects.
ABSTRACT.
Referring to the definition of randomness, the rank use for testing this property has been justified, while appropriate rank tests are derived for verifying the two basic properties of randomness: stability and independence. Moreover, the non-linearity of the differential equations modelling the general atmospheric circulation leading to situations of indetermination and, thus to the existence of change-points in the climate evolution, the methodology is managed at first on the full determination of those change-points. The further steps of this methodology are closely described by its application to the case of the Casablanca (Cuba) time- series of annual temperature averages. In the conclusions, the link with statistical logics of test efficiency, the exclusion of statistics depending of assumptions unknown in the definition of the tested properties and the place of men made effects accounting for the physics of the measured phenomenon, are discussed.
|
|
