The palaeo research of the Climate Variability Research Department aims at developing a methodology for combining dynamical knowledge, based on the analysis of model simulations, and proxy data. This is done in close co-operation with climatologists, glaciologists, geologists and palaeo-ecologists. Research themes are the low-frequency variability of the present climate and the sensitivity of the climate system to changes in the external forcing or in the boundary conditions. Both themes are highly relevant for detection and modelling of anthropogenic climate change. They are fully consistent with CLIVAR, PAGES (Past Global Changes) and the CLIVAR/PAGES intersection.

The project 'Patterns of low-frequency climate variability: a model-palaeodata comparison' ended in 2000. This was a collaboration between the Climate Variability Research Department, the Climate Analysis Department and Prof. Henry Hooghiemstra (Hugo de Vries Laboratory, University of Amsterdam). Focus of this NOP-funded project was on natural climate variability on interdecadal to centennial timescales. As a data base European early-instrumental timeseries and a proxy-based network covering North America and Europe were used. The latter incorporates high-quality climate reconstructions based on various palaeo proxies.

A combination of statistical techniques was applied, including principal component analysis and multivariate singular spectrum analysis (MSSA). On timescales longer than 50 years two statistically significant modes of temperature variability were identified, one on multidecadal and one on centennial timescales. The first mode is oscillatory, with a timescale in the narrow range 60-80 years. The spatial pattern of this mode implies coherent oscillations over Europe and over north-eastern North America, with maximum amplitudes in Europe; over north-western North America this mode is absent. Its geographic shape suggests a connection to the North Atlantic Oscillation. A relation with solar forcing could not be detected. The second mode, which dominates low-frequency variability at high latitudes, describes temperature variations in a wide interval of timescales. The temporal pattern of this mode shows a multiple-phase 'Little Ice Age' and a prolonged 'Medieval Warm Period'. The interpretation of this mode is difficult, given its rather variable large-scale pattern and the uncertainty associated with centennial timescales in the proxy data. It was shown earlier that low-frequency variability patterns in early-instrumental temperature records are season-dependent. Due to a lack of seasonally resolved proxy data it is difficult to address this issue in a rigorous manner for the presently identified low-frequency modes. In Europe, where long seasonal proxy series are available, both modes seem to be season-specific.

An international workshop on combining modelling and palaeodata in the analysis of low-frequency climate variability was co-organised with Dr. H. von Storch (GKSS Research Centre, Geesthacht, Germany), bringing together about 30 scientists from Europe and the US. A short report on the workshop theme appeared in the American Geophysical Union Newsletter, Eos.

Research on this theme continues with another NOP-funded project 'Sea level and climate variability on multi-decadal to centennial timescales'. This is a joint project of the Climate Predictability Research Department and Oceanographic Research Department with Dr. Orson van de Plassche (Faculty of Earth Sciences, Free University Amsterdam). This two-year project started in 1999. The KNMI contribution to this project is to establish, from ECBILT1 output, relations between coherent spatial patterns of North Atlantic sea level variability and prominent modes of the atmospheric circulation and other atmospheric parameters in the North Atlantic. A procedure to compute sea level diagnostically from ECBILT1 was set up. A 1500-year long sea-level record for the North American East Coast, as reconstructed by Van de Plassche on the basis of salt marsh data, will be used in the model-data comparison. Hypotheses regarding the relation between sea level at the North American East Coast and climate in NW Europe, which have been formulated on the basis of reconstructed sea level and various other climate proxies, will be validated within the context of the model experiments.

The separation of solar-induced climate variations from other sources of climatic variability is a fundamental problem in the Earth sciences. By using concepts from dynamical systems theory, a technique has been devised which has the potential to distinguish explicitly between internally generated and externally forced climate variability. This technique was applied to a long early-instrumental temperature record.

The mid-Holocene Optimum has been defined by the international Palaeo Modelling and Intercomparison Project (PMIP) as a key period for assessing the sensitivity of climate models to past changes in external forcing and boundary conditions. The climatic response as simulated by ECBILT1 for the mid-Holocene Optimum was found to be comparable to results of state-of-the-art GCM's. A transient simulation of 10,000 year for the Holocene was recently completed. An off-line glacier model was used to estimate the Holocene evolution of a small number of maritime and continental glaciers. The glacier model generates synthetic glacier length records, which can be validated against known constraints for the Holocene epoch.

A PhD study on 'Modelling of astronomically forced variations in (circum)-Mediterranean climate' started at the end of 1999. This is a joint project with Dr. Frits Hilgen (Faculty of Earth Sciences, University of Utrecht). Mediterranean sedimentary records bear a clear signal of both the obliquity and precession cycle in the astronomical forcing, although the local forcing due to obliquity is very small compared to that due to precession. The aim of this study is to explain the 'observed' climatic response, which is found to be robust over the last 5 Million years. Results of time-slice experiments with ECBILT1 show both a precession and an obliquity signal in the African monsoon and in the winter precipitation in southern Europe.

On geological timescales the Earth's climate may be stabilised by the close coupling between climate and the biota. This is demonstrated mathematically by the Daisyworld model of Watson and Lovelock (1983). It is a highly idealised zero-dimensional Earth System model, based on albedo regulation due to vegetation changes. The steady state solution exhibits homeostasis over a large range of solar luminosities. A detailed analysis of Daisyworld was carried out in order to determine to which extent this model is relevant for the real Earth System. It was shown that the basic assumptions of the Daisymodel imply local temperatures, which are independent of incoming solar radiation. The property of fixed local temperatures and the associated heat transport mechanism do not seem to have parallels in the real climate. At the same time, this property is crucial for the homeostatic behaviour of Daisyworld. An invited lecture on this work was given at the second Chapman Conference on the Gaia hypothesis in Valencia, Spain, in June 2000.

home

Greet de Graaf / Brigitta Kamphuis

Updated on July 10, 2002