SOME LONG-RUN PROPERTIES OF CLIMATIC RECORDS

Fractal analysis allows for a mathematical comparison of climatic changes obtained from a variety of observations and recorded at different time scales. Analysis of the oxygen isotope curve of Pacific core V28-239 indicated fractal geometry of the oxygen isotope record for time scales 5000 to 2 000 000 years, with a fractal dimension of 1.22. On a time scale of some hundred years, tree ring index values yield an average fractal dimension of 1.32 +/- 0.02. For annual precipitation records from between 1817 and 1963 at nine major cities in the United States, a mean fractal dimension of 1.26 +/- 0.03 was found. Analysis of the annual mean surface air temperatures at seven meteorological stations in Hungary for the period of 1901-1991, indicates that the considered temperatures are fractals with a mean fractal dimension of 1.23 +/- 0.01. For the global surface air temperature change estimated from meteorological station records from 1880 to 1985, we derive in the present study a fractal dimension of 1.21. It is reasonable therefore to assume that climatic changes are characterized by one general fractal dimension over the spectral range 10 to 10(6) years. Records with such values of fractal dimension have some long-term persistence: even observations sufficiently distant from each other are not completely independent. If these conclusions become confirmed through analysis of a wider set of climatic records, long-run climatic prediction (in statistical sense) on different time scales will appear feasible.