| dc.description.abstracteng | The solar activity has been observed to vary at various time scales, of which the 11-year solar cycle is the most prominent one. Since the Sun provides the main external energy to the Earth, knowledge of solar variability is highly crucial for understanding the influence of the Sun on the Earth's climate system. One of the important measurements related to solar variability is the solar irradiance: the total solar irradiance (TSI) and the spectral solar irradiance (SSI). However, space-based TSI/SSI measurements only cover the last four decades, which is unfortunately not sufficient for studying long-term solar variability and its influence on climate. Therefore, reconstructions of the solar irradiance on longer time scales are required. Reconstructions of the solar irradiance require knowledge of proxies of solar magnetic activity. The only directly-observed solar quantity back to 1610 is the group sunspot number, while one has to rely on indirect proxies going further back in time. The commonly used indirect proxy is the concentration of the cosmogenic isotopes (14C and 10Be) retrieved from natural archives. Cosmogenic isotopes are produced in the upper terrestrial atmosphere by the impinging galactic cosmic rays, whose flux is modulated by the heliospheric magnetic field. The signals in the 14C record are well globally-mixed while in the 10Be records they are highly subjected to the local climate. We constructed the first multi-isotope composite based on one global 14C and six local 10Be records, using a new Bayesian approach (Chap. 3). All six 10Be records were first synchronized with respect to the 14C record using a wiggle-matching method. Next a Monte Carlo simulation was performed to search for that solar modulation potential which best fits all the available isotope data sets at any given time. This composite is considered more robust compared to other composites constructed linearly. Hence, it is further used in this thesis as a proxy of solar magnetic activity on a millennial time scale. In this thesis, we use two SATIRE (Spectral And Total Irradiance REconstruction) versions, SATIRE-T and SATIRE-M, to reconstruct the long-term changes in the solar irradiance. The SATIRE-T model uses the sunspot number to deduce the evolution of the solar surface magnetic components and to reconstruct the TSI/SSI back to the beginning of the 17th century. The SATIRE-M model employs the cosmogenic isotopes as proxies of solar activity to reconstruct the TSI/SSI over the Holocene. Since the SATIRE-M model is based partially on the SATIRE-T model, we first re-visited the SATIRE-T model with various modifications (Chap. 4). With these improvements, the free parameters in the SATIRE-T model are constrained and further employed in the SATIRE-M model. Next, we use the SATIRE-M model and the first multi-isotope composite to reconstruct the solar irradiance over the last 9\,000 years. This is the first SSI reconstruction that not only uses physics-based models to describe all involved non-linear physical processes, but also bases on a multi-isotope composite. The TSI/SSI reconstructions have been recommended for studies of long-term climate changes within the Palaeoclimate Modelling Intercomparison Project-phase 4 (PMIP4). Due to the sampling and quality of the cosmogenic isotope data, the reconstructions with the SATIRE-M model have only a resolution of 10 years, which, unfortunately, might cause biases in the climate models. Therefore, we developed a statistical approach to simulate the quasi 11-year solar cycle from the decadally-averaged sunspot numbers (Chap. 5). This is done by characterising the solar cycle properties and finding the linear relationships between these properties and the decadally-averaged sunspot numbers. This simulated sunspot number series with 11-year solar cycles has annual resolution, and is further employed in the SATIRE model to reconstruct the annual values of TSI/SSI over the Holocene. The TSI/SSI reconstructions with simulated cycles are consistent with the reconstructions based on the directly-observed sunspot numbers. This final solar irradiance reconstruction has been provided as a solar forcing input to climate models. This hopefully will help us to better understand the degree of the solar influence on the Earth's climate on long time scales. | de |