Results from the SBI-1 Flight on September 1, 2003

The following are the preliminary results from the reduction and analysis of the data of a single mosaic recorded during the flight on September 1, 2003. However, a more through analysis will be necessary to reduce artifacts and sources of noise, such as the 5-minute oscillations.

These results were first presented at the 2003 AGU Fall Meeting in San Francisco. Click here to download the poster in pdf [1.5 MB].

For additional informations regarding the implications of these observations and a more detailed discussion on the results see also an article that we published on ApJ Letters: Foukal et al., ApJ Letts. 611, L57, 2004. Click here to download the paper in pdf [155 KB].

The full-disk mosaic

The above image on the left shows the first full disk observation of the Sun recorded in total-light by the SBI on September 1, 2003. The image is a composite of 10 tiles, each of them is the average of 60 individual frames. The heliographic North is towards the upper left corner and West is towards the upper right corner. The image on the right is the same as the one on the right but with the limb-darkeing function removed (as described in step 5 of the data reduction procedure).
The spatial resolution is about 5.6 arcsec. Both images have been contrast enhanced to outline details.

Photometric precision

The figure on the left summarizes the performance of the SBI system. In the top image we show the upper right tile of the mosaic with the full field-of-view of the detector (917 × 687 arcsec). The two lines (labeled "Scan 1" and "Scan 2") indicate the location of the intensity profiles shown in the two plots below. For reference, we have highlighted with capital letters some interesting features that are crossed by the scans. These profiles show that faculae and enhanced network smaller than 10 arcsec in size are clearly resolved.
The spot (marked as "D") exhibits a contrast deficit of about 17% which is significantly lower than what one would expect for a sunspot. However this is a relatively small spot and its contrast value measured with SBI is consistent with the values measured for this same spot with the CFDT2 photometric scanner at the San Fernando Observatory, which has similar angular resolution observing in three 10 nm pass-bands centered at 472, 672, and 780 nm. Furthermore, during our ground based test measurements we have observed other spots of similar size and they also show simmilar contrasts.
The pixel to pixel noise level is around 0.1-0.2%. Features with contrasts as low as 0.3% can be easily discerned. Most of the low contrast features are due to the 5-minute oscillations (p-mode) and in part to granulation. We consider these low intensity features as Sun induced noise and they need to be removed in order to properly determine the facular and network contrasts. As a next step of our data analysis is to remove this solar-induced noise by averaging all the mosaics we recorded during the flight. Reducing further the noise is of paramount importance for achieving the science objective to search for weak, non-magnetic sources of the Total Solar Irradiance, like convection and meridional flows.
Click on the image to view a larger version.

Limb darkening

One of the first results of the post flight data analysis is the very first determination of the limb-darkening function (LD) from actual measurements in total-light. Here LD is plotted as a function of μ = cos(θ).
In Step 5 of the data reduction procedure we have described how it is derived.
In the same figure we compare the SBI measured LD function with estimates from ground-based multi-spectral observations. We obtained the dashed curve (2) by integrating a set of monochromatic LD curves from Neckel and Labs (1994), weighted by the spectrum of the solar radiation. This curve agrees well with the SBI derived one, with the exception for μ < 0.2, which corresponds to the last 20 arcsec before reaching the limb. The discrepancy in the last 20 arcsec can not be reasonably explained when considering imperfections in the SBI's optical system. The point spread correction would require a totally unrealistic large amount of scattering or blurring to notably influence the shape of the observed LD. Also shown in the plot are the total LD function (3) from the Astrophysical Quantities book of Allen 2000, pp. 355-357 (but of unknown origin), and the curve (4) calculated from the gray LD formula (e.g., Foukal 1990, p. 54):

I(μ) = 3 / 5 ⋅ (μ + 2 / 3)

The agreement of these curves to a few percent provides some external validation of the linearity of our instrumental photometric response curve, which was derived in the lab and checked in flight with ND filters.
Click on the plot to view a larger version.

Center-to-limb variation of facular and enhanced network contrast

Another result of the preliminary data analysis is a first estimate of the bolometric center-to-limb variation of the photospheric contrast of faculae and enhanced network. Determination of such a curve has always been elusive because of the lack of truly broad-band images of the solar photosphere. Knowing this curve with good accuracy is of critical importance when trying to model the observed daily and secular variation of the TSI.

We identified faculae and enhanced network on a full-disk CaIIK image from the San Fernando Observatory that was recorded at the same time when the SBI was taking the broad-band mosaics. First we removed the LD function from the CaIIK image by following the same procedure as for the SBI mosaic. Second, to identify the faculae and enhanced network, we applied a threshold to the image by taking into account only areas with normalized intensity higher than 1.13 which would isolate only the bright faculae and the enhanced network (above LEFT image). For this preliminary analysis we choose such a high threshold because a lower intensity would have added too much scatter at low values of μ. Finally we applied the obtained facular mask to the SBI intensity mosaic (above RIGHT image).

By considering all the pixels lying within the mask we have built a scatter plot of normalized facular intensity versus distance from Sun center (μ). The overlay dashed curve represents the average value of the facular contrast when the dataset is divided in bins 0.07 μ wide and the mean contrast value is calculated for each bin. The vertical bars show the standard deviation within each bin. Except very close to the limb, this curve agrees well with a similar plot determined by Foukal et al. (1991) who sampled a variety of monochromatic measurements of facular contrasts at different wavelength. Most of the scatter observed is mainly caused by facula-to-facula differences due to different magnetic flux as emphasized by Ortiz et al. (2002).
Click on the plot to view a larger version.