D. G. Preminger

Photometric quantities for solar irradiance modeling

[1]xa0We analyze photometric quantities for the modeling of the total solar irradiance, S. These quantities are derived from full-disk solar images taken at the San Fernando Observatory. We introduce a new quantity, the photometric sum, Σ, which is the sum over an entire image of each pixels contribution to the irradiance in that image. Σ combines both bright and dark features; and because the sum is over the entire image, it will include low contrast features that cannot be identified directly. Specifically, we examine Σr, Σb, and ΣK, the photometric sums over broadband red, broadband blue, and 1-nm bandpass Ca II K images, respectively. Σr and Σb measure the effects of solar features on the variability in S at two different continuum wavelengths. ΣK measures the variability in spectral lines due to solar features. We find that Σr and Σb have no long-term trend. ΣK, however, varies in phase with the solar cycle. We carry out several multiple linear regressions on the value of S from cycle 22; the best fit uses Σr and ΣK and reproduces the observed composite S with a multiple regression coefficient R = 0.96. We conclude that the long-term change in S over the solar cycle can be accounted for by the variability in the spectral lines as measured by ΣK, assuming no change in the quiet Sun; the contribution of the continuum to the variations in S is only on active region timescales.

Solar Feature Identification using Contrasts and Contiguity

We present a new technique for the rapid, automatic identification of solar features on full-disk photometric images. The technique permits the detection of features whose contrasts are only slightly above the noise level. Contrast and contiguity criteria are used to identify pixels belonging to an individual feature. The criteria used are simple and objective, and do not require one to guess at the contrast distribution of the features. Comparison of Cau2009iixa0K images with magnetograms shows excellent agreement between the identified features and observed magnetic features. In addition, we can now reliably identify faculae on continuum images. Since this technique can be rapidly applied to a large set of images, it allows us to compile a database of the physical and photometric properties of individual solar features.

The Contribution of Faculae and Network to Long-Term Changes in the Total Solar Irradiance

A new database of individual solar features has been compiled from the full-disk photometric Ca II K images taken at the San Fernando Observatory (SFO) during solar cycle 22. The distribution of facular region sizes differs at different phases of the solar cycle; the area coverage of large active regions is reduced by a factor of about 20 at solar minimum compared to solar maximum, while the smaller regions cover about half as much area at minimum as at maximum. The irradiance contribution of large features is about 10 times greater at maximum than at minimum, while that of small features is about twice as large. We have used this data set to model the fraction of variation in the total solar irradiance S that is due to solar features of various sizes. The data show that large-scale bright solar features, i.e., faculae, dominate the ~0.1% change in S between solar maximum and solar minimum. Using a variety of data sets, we conclude that large active regions produce about 80% of the total change.

Processing Photometric Full-Disk Solar Images

Daily, photometric, full-disk digital solar images have been taken at the San Fernando Observatory (SFO) at two resolutions and in several wavelengths for more than eleven years. We describe the standard data processing techniques used for these images, including: calibration, limb fitting, geometric correction, and production of a solar contrast map by limb-darkening removal. The resulting contrast maps have a photometric accuracy which is often a few tenths of a percent. We show that the geometric accuracy of our images, as measured by the reproducibility of disk and sunspot areas, is very high as well. The techniques described in this paper should be applicable to any instrument producing full-disk photometric images.

Differences in the Sun's radiative output in cycles 22 and 23

Analysis of the current solar cycle 23 shows a greater increase in total solar irradiance (TSI) for the early phase of this cycle than expected from measurements of the total magnetic flux and traditional solar activity indices, which indicate that cycle 23 is weaker than cycle 22. In contrast, space observations of TSI from the Solar and Heliospheric Observatory/VIRGO and the Upper Atmospheric Research Satellite/ACRIMII show an increase in TSI of about 0.8-1.0 W m-2 from solar minimum in 1996 to the end of 1999. This is comparable to the TSI increase measured by Nimbus 7/ERB from 1986 to 1989 during the previous cycle. Thus, solar radiative output near the maximum of the 11 yr cycle has been relatively constant despite a factor of 2 smaller amplitude increase for cycle 23 in sunspot and facular areas determined from ground-based observations. As a result, empirical models of TSI based on sunspot deficit and facular/network excess in cycle 22 underestimate the TSI measurements in 1999. This suggests either a problem in the observations or a change in the sources of radiative variability on the Sun.

A new model of total solar irradiance based on sunspot areas

[1]xa0We show that daily sunspot area can be used in a simple model to reconstruct daily variations in the total solar irradiance, S. The model assumes that all fluctuations in S are correlated with the emergence of sunspots on the solar disk. Cotemporal data for S and sunspot area are analysed to extract the finite impulse response function that describes the time evolution of S in response to a sunspot. The impulse response function clearly shows the evolution of a dark sunspot into a well-defined bright region which then spreads out and decays over a period of about 400 days. This function can be used to reconstruct S from the Greenwich daily sunspot area database, which extends from the late 1800s to the present. We find that the level of S at solar minimum has no long-term secular trend that is correlated with the level of sunspot activity.

Restoration and Photometry of Full-Disk Solar Images

Daily, photometric, full-disk digital solar images have been taken at the San Fernando Observatory (SFO) in several wavelengths for more than 10 years. This work describes a project to evaluate and remove the effects of scattered light from the images, while preserving the photometry. We model both the solar limb and the point-spread function analytically, and the algorithm uses a least-squares fitting technique that does not require artificial extension of the data. Image restoration is carried out using standard techniques, with the exception of a method for estimating a non-white noise component. We show using artificial solar images with sunspots and faculae that the effects of blurring are greater near the solar limb, but that the restorations recover most of the actual contrast of surface features. Images taken with two different telescopes, after restoration, show surface feature contrasts that are in better agreement than before the restoration. In addition, the measured umbral contrast of approximately -90% on restored images at 6723 A is the expected contrast value for umbral temperatures near 3000 K. Modeling of the solar irradiance using restored images instead of the original images from a 3 month period in 1988 shows no advantage if we use variations in sunspot deficit and facular excess to model irradiance variations. However, a new parameter, the model total irradiance derived from restored broadband red images, can model the actual solar irradiance during this period with an R2 value of 0.96.

Analysis of sunspot area over two solar cycles

We examine changes in sunspots and faculae and their effect on total solar irradiance during solar cycles 22 and 23 using photometric images from the San Fernando Observatory. We find important differences in the very large spots between the two cycles, both in their number and time of appearance. In particular, there is a noticeable lack of very large spots in cycle 23 with areas larger than 700 millionths of a solar hemisphere which corresponds to a decrease of about 40% relative to cycle 22. We do not find large differences in the frequencies of small to medium spots between the two cycles. There is a decrease in the number of pores and very small spots during the maximum phase of cycle 23 which is largely compensated by an increase during other phases of the solar cycle. The decrease of the very large spots, in spite of the fact that they represent only a few percent of all spots in a cycle, is primarily responsible for the observed changes in total sunspot area and total sunspot deficit during cycle 23 maximum. The cumulative effect of the decrease in the very small spots is an order of magnitude smaller than the decrease caused by the lack of large spots. These data demonstrate that the main difference between cycles 22 and 23 was in the frequency of very large spots and not in the very small spots, as previously concluded. Analysis of the USAF/NOAA and Debrecen sunspot areas confirms these findings.

A Statistical Analysis of the Characteristics of Sunspots and Faculae

We present results from a study of sunspots and faculae on continuum and Cau2009iixa0K images taken at the San Fernando Observatory (SFO) during 1989–1992; a total of approximately 800 images in each bandpass were used. About 18u2009000 red sunspots, 147u2009000 red faculae, and 800u2009000 Cau2009iixa0K faculae were identified based on their contrasts. In addition, we computed the contrasts of pixels on the red images cospatial with Cau2009iixa0K faculae. Sunspot contrasts show a strong dependence on size but no dependence on heliocentric angle. There are continuous but systematic differences among facular regions. We find that the contrast of Cau2009iixa0K faculae is relatively insensitive to heliocentric angle, but is a strong function of facular size, in the sense that larger Cau2009iixa0K faculae are always brighter. The contrast of red faculae is a function of both heliocentric angle and size: the contrast functions show that larger regions contain larger flux tubes, contain deeper flux tubes, and have larger filling factors than small facular regions. Comparisons of cospatial pixels on red and Cau2009iixa0K images show a tight correlation between the average contrast of a region in the continuum and its size and heliocentric angle in the Cau2009iixa0K images. The average contrast of all facular regions is positive everywhere on the disk, even though the largest regions contain flux tubes which appear dark at disk center.

On the decay rate of sunspots

[1]xa0We have analyzed the decay of 32 sunspots observed during the years 1988 through 2001 at the San Fernando Observatory (SFO). The data are from digital images obtained in the red (672 nm) with the Cartesian Full Disk Telescope No.1 (CFDT1). We find that the rate of decay is strongly correlated with the total sunspot area and the umbral to total area ratio. The multiple correlation coefficient is 0.93. Thus, the unexplained variance from this simple model is (1–0.87). We find that for the sunspots of this study, the decay rate is not a constant and that there is no significant correlation between the decay rate and the square root of the total spot area.