Paper detail

Solar total irradiance in cycle 23

The apparently unusual behaviour of the TSI during the most recent minimum of solar activity has been interpreted as evidence against solar surface magnetism as the main driver of the secular change in the TSI. We test claims that the evolution of the solar surface magnetic field does not reproduce the observed TSI in cycle 23. We use sensitive, 60-minute averaged MDI magnetograms and quasi-simultaneous continuum images as an input to our SATIRE-S model and calculate the TSI variation over cycle 23, sampled roughly twice-monthly. The computed TSI is then compared to the PMOD composite of TSI measurements and to the data from two individual instruments, SORCE/TIM and UARS/ACRIM II, that monitored the TSI during the declining phase of cycle 23 and over the previous minimum in 1996, respectively. Excellent agreement is found between the trends shown by the model and almost all sets of measurements. The only exception is the early, i.e. 1996 to 1998, PMOD data. Whereas the agreement between the model and the PMOD composite over the period 1999-2009 is almost perfect, the modelled TSI shows a steeper increase between 1996 and 1999 than implied by the PMOD composite. On the other hand, the steeper trend in the model agrees remarkably well with the ACRIM II data. A closer look at the VIRGO data, that make the basis of the PMOD composite after 1996, reveals that only one of the two VIRGO instruments, the PMO6V, shows the shallower trend present in the composite, whereas the DIARAD measurements indicate a steeper trend. We conclude that (1) the sensitivity changes of the PMO6V radiometers within VIRGO during the first two years have very likely not been correctly evaluated, and that (2) the TSI variations over cycle 23 and the change in the TSI levels between the minima in 1996 and 2008 are consistent with the solar surface magnetism mechanism.