The hunt for low-frequency modes of oscillation of the Sun: application of statistical

Broomhall, Anne-Marie (2008). The hunt for low-frequency modes of oscillation of the Sun: application of statistical. University of Birmingham. Ph.D.

Preview

Broomhall2008PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (16MB) | Preview

Abstract

The main focus of this thesis is to search for low-frequency solar acoustic and gravity modes of oscillation of the Sun. Low-frequency modes are sensitive to the structure of the solar core and so provide stringent tests for models of the internal structure of the Sun. In this thesis we develop and use statistical techniques that aim to detect low-frequency modes by searching for statistically significant features in frequency-power spectra. Frequency-amplitude spectra are compared for coincident prominent features. The amount of common noise shared by two data sets affects the probability that a coincident prominent feature is noise and so the statistical tests were non-trivial to derive. The optimum condition for comparing frequency-amplitude spectra is when no common noise is present. The power and source of the common noise is dependent on the data that are compared.
Statistical tests are used to search contemporaneous and non-contemporaneous observations made by the Birmingham Solar Oscillations Network (BiSON) and the Global Oscillations at Low Frequencies (GOLF) instrument. We also searched contemporaneous and non-contemporaneous sets of BiSON data. Many different combinations of BiSON data are compared to find contemporaneous BiSON timeseries that contain as little common noise as possible. Detection threshold levels are significantly reduced by searching for coincident prominent features, however, no new mode candidates are detected. To investigate whether the gaps in the list of detected modes are expected, given the reduced threshold levels, predictions on the detectability of low-frequency modes are made, based on estimates of the modes' powers and damping times. The fraction of modes detected in Monte Carlo simulations is small at low frequencies. Significantly more detections would be made if the power signal to noise could be increased by a factor of 2, which could be achieved by lengthening timeseries and by reducing the amount of noise in the data.
The quality of solar observations is affected by instrument stability. We consider three effects that mean unresolved Doppler velocity observations are not homogenous across the solar disc; solar rotation, limb darkening and the position of a detector. The results indicate that observations made by BiSON instruments do not represent a uniform average across the solar disc. Different detectors are weighted towards different regions of the solar disc, and so observe different realizations of the solar noise. The calculated weightings are compared to BiSON observations. Agreement between the calculations and the observations is improved by altering various parameters, such as the observed size of the solar image. We find that the weighting varies with epoch and so can have consequences for the quality of BiSON data.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):