Presenting 28 years of Sun-as-a-star extreme ultraviolet light curves from SOHO EIT

Context. The Solar and Heliospheric Observatory (SOHO) Extreme-ultraviolet Imaging Telescope (EIT) has been taking images of the solar disk and corona in four narrow extreme ultraviolet (EUV) bandpasses (171 Å, 195 Å, 284 Å, and 304 Å) at a minimum cadence of once per day since early 1996. The time series of fully calibrated EIT images now spans approximately 28 years, from early 1996 to early 2024, covering solar cycles 23 and 24 in their entirety, as well as the beginning of cycle 25. Aims. We aim to convert this extensive EIT image archive into a set of “Sun-as-a-star” light curves in EIT's four bandpasses, observing the Sun as if it were a distant point source viewed from a fixed perspective. Methods. To construct the light curves, we summed the flux in each EIT image into one flux value, with an uncertainty accounting for both the background noise in the image and the potential spillover of flux beyond the bounds of the image (which is especially important for the bands with significant coronal emission). We corrected for long-term instrumental systematic trends in the light curves by comparing our 304 Å light curve to the ultraviolet light curve taken by SOHO's CELIAS/SEM solar wind monitoring experiment, which has a very similar bandpass to the EIT 304 Å channel. We corrected for SOHO's viewing angle by fitting a trend to the flux values with respect to SOHO's heliocentric latitude at the time of each observation. Results. We produced two sets of Sun-as-a-star light curves with different uncertainty characteristics, available for download from Zenodo, either of which might be preferred for different types of future analyses. In version (1), we treated the EIT instrumental systematics consistently across the entire SOHO mission lifetime, producing a light curve with approximately homoscedastic uncertainties. In version (2), we only divided out the EIT instrumental systematics from November 12, 2008, onward; this is the point at which these systematics start to have a noticeable deleterious effect on the data. Therefore, version (2) has heteroscedastic uncertainties, but these uncertainties are much smaller than the version (1) uncertainties over the first half of the mission. Conclusions. We find that our EUV light curves trace the Sun's ∼11-year solar activity cycle and ∼27-day rotation period much better than comparable optical observations. In particular, we can accurately recover the solar rotation period from our 284 Å light curve for 26 out of 28 calendar years of EIT observations (93% of the time), compared to only 3 out of 29 calendar years (10% of the time) for the VIRGO total solar irradiance time series, which is dominated by optical light. Our EIT light curves, in conjunction with Sun-as-a-star light curves at optical wavelengths, will be valuable to those interested in inferring the EUV/UV character of stars with long optical light curves, but no intensive UV observations, as well as to those interested in long-term records of solar and space weather.