IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group I: The Physical Science Basis Direct Observations of Solar Irradiance Satellite measurements of total solar irradiance

Four independent space-based instruments directly measure total solar irradiance at present, contributing to a database extant since November 1978 (Fröhlich and Lean, 2004). The Variability of Irradiance and Gravity Oscillations (VIRGO) experiment on the Solar Heliospheric Observatory (SOHO) has been operating since 1996, the ACRIM III on the Active Cavity Radiometer Irradiance Monitor Satellite (ACRIMSAT) since 1999 and the Earth Radiation Budget Satellite (ERBS) (intermittently) since 1984. Most recent are the measurements made by the Total Solar Irradiance Monitor (TIM) on the Solar Radiation and Climate Experiment (SORCE) since 2003 (Rottman, 2005). Observed decadal trends and variability

Different composite records of total solar irradiance have been constructed from different combinations of the direct radiometric measurements. The Physikalisch-Meteorologisches Observatorium Davos (PMOD) composite (Fröhlich and Lean, 2004), shown in Figure 2.16, combines the observations by the ACRIM I on the Solar Maximum Mission (SMM), the Hickey-Friedan radiometer on Nimbus 7, ACRIM II on the Upper Atmosphere Research Satellite (UARS) and VIRGO on SOHO by analysing the sensitivity drifts in each radiometer prior to determining radiometric offsets. In contrast, the ACRIM composite (Willson and Mordvinov, 2003), also shown in Figure 2.16, utilises ACRIMSAT rather than VIRGO observations in recent times and cross calibrates the reported data assuming that radiometric sensitivity drifts have already been fully accounted for. A third composite, the Space Absolute Radiometric Reference (SARR) composite, uses individual absolute irradiance measurements from the shuttle to cross calibrate satellite records (Dewitte et al., 2005). The gross temporal features of the composite irradiance records are very similar, each showing day-to-week variations associated with the Sun’s rotation on its axis, and decadal fluctuations arising from the 11-year solar activity cycle. But the linear slopes differ among the three different composite records, as do levels at solar activity minima (1986 and 1996). These differences are the result of different cross calibrations and drift adjustments applied to individual radiometric sensitivities when constructing the composites (Fröhlich and Lean, 2004).


Figure 2.16. Percentage change in monthly values of the total solar irradiance composites of Willson and Mordvinov (2003; WM2003, violet symbols and line) and Fröhlich and Lean (2004; FL2004, green solid line). 

Solar irradiance levels are comparable in the two most recent cycle minima when absolute uncertainties and sensitivity drifts in the measurements are assessed (Fröhlich and Lean, 2004 and references therein). The increase in excess of 0.04% over the 27-year period of the ACRIM irradiance composite (Willson and Mordvinov, 2003), although incompletely understood, is thought to be more of instrumental rather than solar origin (Fröhlich and Lean, 2004). The irradiance increase in the ACRIM composite is indicative of an episodic increase between 1989 and 1992 that is present in the Nimbus 7 data (Lee et al., 1995; Chapman et al., 1996). Independent, overlapping ERBS observations do not show this increase; nor do they suggest a significant secular trend (Lee et al., 1995). Such a trend is not present in the PMOD composite, in which total irradiance between successive solar minima is nearly constant, to better than 0.01% (Fröhlich and Lean, 2004). Although a long-term trend of order 0.01% is present in the SARR composite between successive solar activity minima (in 1986 and 1996), it is not statistically significant because the estimated uncertainty is ±0.026% (Dewitte et al., 2005).

Current understanding of solar activity and the known sources of irradiance variability suggests comparable irradiance levels during the past two solar minima. The primary known cause of contemporary irradiance variability is the presence on the Sun’s disk of sunspots (compact, dark features where radiation is locally depleted) and faculae (extended bright features where radiation is locally enhanced). Models that combine records of the global sunspot darkening calculated directly from white light images and the magnesium (Mg) irradiance index as a proxy for the facular signal do not exhibit a significant secular trend during activity minima (Fröhlich and Lean, 2004; Preminger and Walton, 2005). Nor do the modern instrumental measurements of galactic cosmic rays, 10.7 cm flux and the aa geomagnetic index since the 1950s (Benestad, 2005) indicate this feature. While changes in surface emissivity by magnetic sunspot and facular regions are, from a theoretical view, the most effective in altering irradiance (Spruit, 2000), other mechanisms have also been proposed that may cause additional, possibly secular, irradiance changes. Of these, changes in solar diameter have been considered a likely candidate (e.g., Sofia and Li, 2001). But recent analysis of solar imagery, primarily from the Michelson Doppler Imager (MDI) instrument on SOHO, indicates that solar diameter changes are no more than a few kilometres per year during the solar cycle (Dziembowski et al., 2001), for which associated irradiance changes are 0.001%, two orders of magnitude less than the measured solar irradiance cycle. Measurements of solar spectral irradiance

The solar UV spectrum from 120 to 400 nm continues to be monitored from space, with SORCE observations extending those made since 1991 by two instruments on the UARS (Woods et al., 1996). SORCE also monitors, for the first time from space, solar spectral irradiance in the visible and near-infrared spectrum, providing unprecedented spectral coverage that affords a detailed characterisation of solar spectral irradiance variability. Initial results (Harder et al., 2005; Lean et al., 2005) indicate that, as expected, variations occur at all wavelengths, primarily in response to changes in sunspots and faculae. Ultraviolet spectral irradiance variability in the extended database is consistent with that seen in the UARS observations since 1991, as described in the TAR.

Radiation in the visible and infrared spectrum has a notably different temporal character than the spectrum below 300 nm. Maximum energy changes occur at wavelengths from 400 to 500 nm. Fractional changes are greatest at UV wavelengths but the actual energy change is considerably smaller than in the visible spectrum. Over the time scale of the 11-year solar cycle, bolometric facular brightness exceeds sunspot blocking by about a factor of two, and there is an increase in spectral irradiance at most, if not all, wavelengths from the minimum to the maximum of the solar cycle. Estimated solar cycle changes are 0.08% in the total solar irradiance. Broken down by wavelength range these irradiance changes are 1.3% at 200 to 300 nm, 0.2% at 315 to 400 nm, 0.08% at 400 to 700 nm, 0.04% at 700 to 1,000 nm and 0.025% at 1,000 to 1,600 nm.

However, during episodes of strong solar activity, sunspot blocking can dominate facular brightening, causing decreased irradiance at most wavelengths. Spectral irradiance changes on these shorter time scales now being measured by SORCE provide tests of the wavelength-dependent sunspot and facular parametrizations in solar irradiance variability models. The modelled spectral irradiance changes are in good overall agreement with initial SORCE observations but as yet the SORCE observations are too short to provide definitive information about the amplitude of solar spectral irradiance changes during the solar cycle.