Back to Summary


WP5 Modelling of spectral irradiance based on modelled proxies

The main objective of WP5 was to reconstruct a daily time series of total and spectral  irradiance variations covering the last century. This was achieved using magnetograms simulated with a surface flux transport model (Cameron et al. 2010, Jiang et al. 2011). These simulated magnetograms provide the spatial distribution of active regions (sunspot and faculae) and their decay products (network) on the solar surface, and therefore enable a better description of the irradiance variations on time scales of weeks to months than when using the sunspot number alone.

The final data product, total and spectral (115 nm – 160 μm) solar irradiance reconstructions from 1700 to 2014, is provided as a composite that incorporates the detailed disc-image and magnetogram-based SATIRE-S reconstructions by Yeo et al. (2014, see also SATIRE-S total and spectral data page) and the reconstructions based on simulated magnetograms developed as part of  the SOLID FP7 programme, SATIRE-T2. The most important characteristics of the composite are described below. For more information on the reconstructions, please see Yeo et al (2014), and Dasi-Espuig et al (2014, 2016 submitted).

Data Characteristics

The solar irradiance reconstructions are a combination of irradiance reconstructions from two versions of the SATIRE model that use the best input data available for each period of time: sunspot numbers in the 18th and early 19th century,  sunspot areas from 1878 to 1974 and, since 1974, full solar disc images (see Table below). Depending on the input data, the reconstructions show different characteristics, e.g., reconstructions covering the satellite era reproduce daily changes of the irradiance exquisitely well, while reconstructions based on sunspot number only reproduce daily variability in a statistical sense, but still provide reliable monthly mean irradiance values.


Input data

Time span




Magnetograms / continuum images

1974 – 2014


Yeo et al., 2014


Sunspot areas (RGO-SOON)

1878 – 1974


Dasi et al., 2014


Sunspot number (here: Rg)

1700 – 1878

~ month

Dasi et al., submitted

Table 1: SATIRE models used to generate the composite. The fourth column indicates the shortest time span for which the data are reliable. Data are released on a daily cadence and give a good representation of the statistical solar variability even when detailed day-to-day agreement is not expected, as in the case for SATIRE-T2 (see text).

Solar Irradiance since 1974: SATIRE-S

The solar irradiance reconstructions since 1974 have been presented in Yeo et al (2014) and are based on detailed solar images and magnetograms. The reconstructions  reproduce 96% of the variability seen in the VIRGO/PMO6V total solar irradiance variability. Comparisons of the spectral solar irradiance (SSI) for wavelengths where measurements are available and reliable show similarly good agreement between the reconstructions and the observations (Yeo et al. 2014, 2015, Ball et al. 2012, 2014). For wavelengths between approximately 250 nm and 400 nm, reconstructions such as SATIRE-S currently offer the highest accuracy and best agreement with the available data (see Yeo et al. 2015).

Solar Irradiance between 1878 and 1974: SATIRE-T2 (sunspot areas)

For this time period, records of sunspot-group areas and positions are available from observations taken at the Royal Greenwich Observatory (RGO). Facular areas and positions were derived using a surface flux transport model (SFTM). While the spatial resolution of the resulting facular maps is much lower than that of the observed magnetograms used for SATIRE-S, the synthetic magnetograms nevertheless capture the evolution of the active regions well and the resulting TSI and SSI reconstructions are expected to be reliable on rotational timescales.

Solar Irradiance pre-1878:  SATIRE-T2 (sunspot number)

For irradiance reconstructions before 1878 we use semi-synthetic records of sunspot area, latitude and tilt angle. These semi-synthetic records rely on observed statistical relationships between properties of sunspot group emergence, cycle strength and phase. As for the reconstructions based on the observed sunspot areas and distributions, the semi-synthetic records are fed into a flux-transport model to obtain facular maps. The spot and facular maps that are obtained in this way are thus not exact solar surface maps, but hypothetical statistical surface maps that are consistent with what one might expect to see given the observed sunspot number. The resulting irradiances are therefore not intended to be accurate daily irradiances, but their monthly average is expected to be in good agreement with the historic averages. In addition the variability pattern is typical of what is expected for the Sun for the given phase in its activity cycle.


There are three main sources of uncertainties for irradiance models: uncertainties in the input physics, in the input data (e.g., sunspot record) and in the fitting process that is used to fix the model parameters. In the context of SOLID, we have explored the effects of using different sunspot numbers, and those of the fitting process, as described in the following.

The free parameters of the model are fixed by fitting modelled data such as the TSI and (for SATIRE-T2) the total magnetic flux  to observations. Observations of the TSI and the total flux have been available since the mid-1970s only and we fix the free parameter of SATIRE-T2 using data from cycles 21 to 23. Figure 1 below illustrates how uncertainties arising in the fitting process affect the long-term trends seen in the irradiances modelled using SATIRE-T2.

Figure 1: The blue line shows the TSI reconstructed with SATIRE-T2 that corresponds to  the best fit to the observations during cycles 21 to 23. The grey shading encompasses the range of the SATIRE-T2 reconstructions parameter values within approximately 1σ relative to the best fit

Example Irradiance Time Series

The SATIRE-T2 SSI reconstructions agree well with the UARS/SUSIM measurements as well as with the Lyman-α flux. A comparison of the reconstructed Lyman-α flux and LASP composite (Woods et al 2000) is shown in Figure 2. This good agreement is obtained even though none of the free parameters were adjusted to specifically fit the UV irradiance (the free parameters are optimised using TSI and total magnetic flux measurements only).

Figure 2: Lyman-α irradiance smoothed over 3 months. The red and blue curves correspond to the SATIRE-T2 model based on the RGO-SOON and the semi-synthetic sunspot group records, respectively. In black we also show the LASP Lyman-α composite. The correlation coefficient between the 3-month smoothed reconstructions and the composite is 0.96.

Figure 3 shows example SSI time series for a number of climate-relevant wavelength bands. For wavelengths between 180nm and 300nm, the absolute intensity is adjusted to match the Whole Heliospheric Interval (WHI) reference solar spectrum (Woods et al. 2009). For wavelengths between 115 to 180 nm, the reconstruction is scaled and offset according to empirical relationships derived by Yeo et al. (2014). As expected, the 11-year cycle variability, as well as the irradiance increase since the end of the Maunder Minimum, is weaker at  longer wavelengths. The reversed variability with respect to the solar cycle between 1500 nm and 2500 nm can be attributed to the decrease of the facular contrast at these wavelengths.

 Figure 3: Yearly means of the reconstructed spectral solar irradiance variations between 1700 and 2009. The black lines show SATIRE-S reconstructions (Yeo et al 2014); red and blue lines are for reconstructions based on RGO data and the statistical model, respectively. The wavelength intervals shown are (from top to bottom) the Schumann-Runge oxygen continuum (130nm – 175nm) and Schumann-Runge oxygen band, the Herzberg oxygen continuum (200nm – 242nm) and Hartley-Huggins ozone (200nm – 350nm) bands and two near-IR water vapour bands.


The SATIRE-S model provides the most accurate reconstruction of solar total and spectral irradiance variability. Its extension back to 1700, based on sunspot data is less precise on daily timescales; this is a result of the deteriorating quality as well as of the spatial and temporal resolution of the input data. Still, the model captures the measured irradiance changes adequately and also agrees with other available related data, such as the total magnetic flux and the solar open flux. The composite SATIRE record is currently the best available data set providing total and spectral irradiance variations since 1700 and we recommend its use for climate models.


D5.1) Release daily solar full-disk simulated magnetograms

D5.2) Release the final time series of the reconstructed daily total solar irradiance

D5.3) Release of the time series of the reconstructed daily spectral solar irradiance

D5.4) Recommendation of total and spectral irradiance reconstruction for WP6



W T Ball, Y C Unruh, N A Krivova, S K Solanki, T Wenzler, D J Mortlock & A H Jaffe, 2012, A&A
541, A27

W T Ball, N A Krivova, Y C Unruh, J D Haigh & S K Solanki, 2014, J. Atm.Sci. 71, 4086

M Dasi-Espuig, J Jiang, N A Krivova & S K Solanki 2014, A&A 570, A23

M Dasi-Espuig et al 2016, submitted to A&A

J Jiang, R H Cameron, S Schmitt & M Schüssler 2011, A&A 528, A83

T N Woods, W K Tobiska, G J Rottman, J R Worden 2000, JGR 105, 27195

T N Woods, P C Chamberlin, J W Harder, R A Hock, M Snow, F G Eparvier, J Fontenla, W E McClintock & E C Richard 2009, GRL 36, L01101

K L Yeo, N A Krivova, S K Solanki & K H Glassmeier 2014, A&A 570, A85

K L Yeo, W T Ball, N A Krivova, S K Solanki, Y C Unruh & J Morrill 2015, JGR 120, 6055