Predictions based on the cumulative curves: Basic principles and nontrivial example

Abstract

In this paper the new prediction method based on analysis of the integrated (cumulative) curves is suggested. This method includes the procedure of the optimal linear smoothing (POLS) for the finding of optimal trends, independent “reading” of relative fluctuations in terms of β-distribution function that are formed after subtraction of the calculated trend and the recognition of the proper fitting hypothesis for the integrated optimal trends by the eigen-coordinates method. The combined noninvasive approach was applied to analysis of temperature data obtained from the site http://data.giss.nasa.gov/gistemp/ related to the global warming (GW) phenomenon. These data are considered as nontrivial examples of verification of new forecasting method. The available data were combined into six files covering the mean/anomalous temperature 1546 month’s points covering the period from the January of 1880 up to October of 2008. Besides the global registered points the combined files included in themselves the north/south data points measured independently for both the Earth’s hemispheres. The combined new method (preliminary verified on mimic data) applied to these files predicts the changing of the GW period by the global cooling (GC) period that will happen during the years 2038–2136. Besides this important result a new method helps to discover the influence of a small but stable oscillating process with a set of self-similar periods Ω n = Ω 0 ξ n , n = 0, ±1, ±2, ±3, ±4 with mean period 〈 T〉 = 12.55 year. This fact should present interest for ecologists and meteorologists working in this field.