Observed and predicted ratios of the horizontal and vertical components ofthe solar p-mode velocity eigenfunctions

We present evidence that the observed ratios of the horizontal and vertical components of the solar intermediate-degree p-mode velocity eigenfunctions closely match theoretical predictions of these ratios. This evidence comes from estimates of the observed eigenfunction component ratios that were ob tained from the fitting of the p-mode oscillation peaks in low- and interme diate-degree (l less than or equal to 200) m-averaged power spectra compute d from two different 60.75 day time series of Global Oscillation Network Gr oup (GONG) project Dopplergrams obtained in late 1996 and early 1998. These fits were carried out using a peak-fitting method in which we fitted each observed p-mode multiplet with a model profile that included both the targe t mode and its six nearest spatial sidelobes and which incorporated the eff ects of the incomplete observational time series through the convolution of the fitted profiles with the temporal window functions, which were compute d using the two actual GONG observing histories. The fitted profile also in cluded the effects of the spatial leakage of the modes of differing degrees into the target spectrum through the use of different sets of m-averaged s patial leakage matrices. In order to study the sensitivity of the estimated component ratios to the details of the computation of the m-averaged power spectra and of the image-masking schemes employed by the GONG project, we generated a total of 22 different sets of modal fits. We found that the bes t agreement between the predicted and inferred ratios came from the use of unweighted averaged power spectra that were computed using so-called n-aver aged frequency-splitting coefficients, which had been computed by cross-cor relating the 2l + 1 zonal, tesseral, and sectoral power spectra at each l o ver a wide range of frequencies. This comparison yielded a total of 1906 pa irs of predicted c(t,theory) and fitted c(t,fit) eigenfunction component ra tios. A linear regression analysis of these pairs of ratios resulted in the following regression equation: c(t,fit) = (0.0088 +/- 0.0013) +/- (0.9940 +/- 0.0044)c(t,theory). The resulting correlation coefficient was 0.9817. T his agreement between the predicted and inferred ratios suggests that the p redicted ratios should be used in the fitting of high-degree power spectra where the ratios cannot be inferred because of the blending together of ind ividual modal peaks into broad ridges of power.