Long-term persistence of solar activity

The solar irradiante has been found to change by 0.1% over the recent solar cycle. A change of irradiante of about 0.5% is required to effect the Earth's climate. How frequently can a variation of this size be expected? We examine the question of the persistence of non-periodic variations in solar activity. The Hürst exponent, which characterizes the persistence of a time series (Mandelbrot and Wallis, 1969), is evaluated for the series of¹⁴ C data for the time interval from about 6000 BC to 1950 AD (Stuiver and Pearson, 1986). We find a constant Hürst exponent, suggesting that solar activity in the frequency range of from 100 to 3000 years includes an important continuum component in addition to the well-known periodic variations. The value we calculate,H ≈ 0.8, is significantly larger than the value of 0.5 that would correspond to variations produced by a white-noise process. This value is in good agreement with the results for the monthly sunspot data reported elsewhere, indicating that the physics that produces the continuum is a correlated random process (Ruzmaikin et al., 1992), and that it is the same type of process over a wide range of time interval lengths. We conclude that the time period over which an irradiance change of 0.5% can be expected to occur is significantly shorter than that which would be expected for variations produced by a white-noise process. The full paper has been submitted to Solar Physics. Part of the research decribed here was carried out by JPL, Caltech under a contract with NASA.