Accurate measurements of solar irradiance started in 1978, but a much longe
r time series is needed in order to uncover a possible influence on the Ear
th's climate. In order to reconstruct the irradiance prior to 1978 we requi
re both an understanding of the underlying causes of solar irradiance varia
bility as well as data describing the state of the Sun (in particular its m
agnetic field) at the relevant epochs.
Evidence is accumulating that on the time-scale of the solar cycle or less,
variations in solar irradiance are produced mainly by changes in the amoun
t and distribution of magnetic flux on the solar surface. The main solar fe
atures contributing to a darkening of the Sun are sunspots, while active-re
gion faculae and the network lead to a brightening. There is also increasin
g evidence for secular changes of the solar magnetic field and the associat
ed of solar brightness variability. In part the behavior of sun-like stars
is used as a guide of such secular changes.
Under the assumption that solar irradiance variations are due to solar surf
ace magnetism on all relevant time scales it is possible to reconstruct the
irradiance with some reliability from today to around 1874, and with lower
accuracy back to the Maunder minimum. One major problem is the decreasing
amount and accuracy of the relevant data with age. In this review the vario
us reconstructions of past solar irradiance are presented and the assumptio
ns underlying them are scrutinized.