Solar/cosmic ray forcing of global climate change is controversial among physicists and climatologists. Attempting to explain a physical link on the basis of the relationship "solar wind - magnetosphere - ionosphere - atmosphere" is difficult because of a very large difference of the solar wind energy and the energy of the atmospheric processes. We inferred paleo-evidence for this relationship from the records of the cosmogenic isotopes 14C (peat, dendrochronologically dated wood) and 10Be (GISP2), as these data can be considered as proxies for variation in the solar radiant output. When solar activity is high, the extended solar magnetic field more effectively shields the Earth from cosmic rays and reduces the production of 14C and 10Be. In contrast, a low solar activity yields more 14C and 10Be.
We found evidence for solar forcing as the factor causing abrupt global climate change at the Subboreal/Subatlantic transition. This abrupt climate change occurred simultaneously with a sharp rise in Delta 14C, from ca -3 per mil around 850 BC to ca 20 per mil around 760 BC. Millennial-scale climate changes during the last glacial period (Dansgaard-Oeschger cycles) are evident from Greenland ice-cores and North Atlantic ocean cores. The cooling phases seem to be part of a climatic cycle, operating independently of the glacial-interglacial cycles forced by orbital variations. Until now, the cause of these climate changes remained a matter of debate. Bond et al. conclude that a solar forcing of this climatic cycle is unlikely and they suggest that the driving mechanism is to be found inside the atmosphere-ocean system. In contrast, we argue that variations in solar activity may have played a significant role in forcing these climate changes. Fluctuations in the 10Be influx values during the last Glacial evidently parallel the Dansgaard-Oeschger warm/cold cycles in the d18O record from the same ice core: relatively warm phases show low 10Be values and peaks in the 10Be record occur during the cold phases. Although other workers suggest that such variations in 10Be reflect changes in snow accumulation rates and in production and long distance transport, we suppose that these variations are the effect of solar variability in the first place, and thus show the cause of the Dansgaard-Oeschger climatic cycles reflected in the d18O record.
Relatively small changes in solar activity may
induce relatively large changes in global climate. Two positive
feedback mechanisms may explain this phenomenon:
1.Variation in solar ultraviolet radiation alters the stratospheric ozone production, which probably triggers climate changes. Haigh performed simulations with climate models to study the relation between the 11-year solar activity cycles, ozone production and climate change. A chemical atmospheric model showed that a 1% increase in UV radiation at the maximum of a solar activity cycle generated 1-2% more ozone in the stratosphere. This increase in the stratospheric ozone content was used as input in a climate model experiment. In the experiment of Haigh this increase resulted in warming of the lower stratosphere by the absorption of more sunlight. In addition, the stratospheric winds were also strengthened and the tropospheric westerly jet streams were displaced poleward. The position of these jets determines the latitudinal extent of the Hadley cells and, therefore, the poleward shift of the jets resulted in a similar displacement of the descending parts of the Hadley Cells. This change in circulation ultimately caused a poleward relocation of the mid latitude storm tracks. The opposite effect as described by Haigh may have played a role in the climate changes around 850 BC. A reduced solar activity, as indicated by the observed strong increases of atmospheric 14C, could have resulted in a decrease in the stratospheric ozone content. A decrease of the latitudinal extent of the Hadley Cell circulation and an equatorward relocation of the mid latitude storm tracks would follow, with climate change as a consequence.
2. Changes of the cosmic ray flux may directly lead to changes in global cloud cover. This direct link may work through ionisation by cosmic rays, as this positively affects aerosol formation and cloud nucleation. An indication for the importance of this process was found by Svensmark and Friis-Christensen. They reported about a correlation between the variation in cosmic ray flux and the observed global cloud cover for the most recent solar cycle. An increase in the global cloud cover is believed to cause a cooling of the Earth, especially when low altitude clouds are involved, because more incoming radiation is reflected. Earlier, Friis-Christensen and Lassen analysed for the period 1861-1989 the similarity between the northern hemisphere temperature record and the length of the solar cycle (as an indicator of solar activity), and found a close match. A direct increase in cloudiness and accompanying cooling would be in agreement with the reconstructed wetter and cooler conditions at middle latitudes around 850 BC.
Accepting the idea of solar forcing of holocene and earlier climatic shifts has major implications for our view of present and future climate. It implies that the climate system is far more sensitive to small variations in solar activity than generally believed.
The climate variability in polar regions is closely linked to the one at lower latitudes. A particular example, which has received recently great attention, is the interaction between the Arctic and the North Atlantic in the framework of the North Atlantic Oscillation. Polar regions are also the areas where the deep waters of the World Ocean are formed. As a consequence, modification of the conditions in those zones could impact on the ocean thermohaline circulation and thus the global climate. Those processes will be investigated by means of a coupled atmospheric-sea-ice-ocean model of intermediate complexity. The atmospheric component is the ECBILT model developed at the KNMI, which is a spectral T21, 3-level quasi-geostrophic model that includes a representation of the horizontal and vertical heat transfers as well as of the hydrological cycle. It is coupled to the CLIO model of the Université Catholique de Louvain, which is made up of a primitive-equation, free-surface OGCM and of a dynamic-thermodynamic sea-ice model. The horizontal resolution of CLIO is 3 degrees by 3 degrees and there are 20 unequally spaced vertical levels in the ocean. The model is still under test, but the preliminary results show that it is able to reproduce reasonably well the climate of the high latitudes, at least in the Northern Hemisphere.
A 40000 year integration with a coupled atmosphere/ocean/sea-ice model of intermediate complexity (ECBilt) has been performed. The atmosphere model of ECBilt is a global three level quasi-geostrophic model with a spectral T21 horizontal resolution and simple parameterizations for the diabatic processes. It is coupled to a coarse resolution ocean model with 12 layers in the vertical and comparable horizontal resolution as the atmosphere model. Sea-ice coverage is simulated with a zero layer thermodynamic model. The climate of ECBilt displays quasi-periodical behaviour with a period of approximately 13000 year. The quasi-periodical behaviour is characterized by large changes in the overturning cell in the Southern Ocean. The southern cell fluctuates between two quasi-stationary states, with accompanying changes in the atmospheric circulation in the Southern Hemisphere. The transition between these states is rapid and resembles the flushes as observed in ocean general circulation models. The changes in the Southern Ocean affect the strength of the North Atlantic overturning cell displaying variations in the order of 5 Sv.
A prerequisite for the understanding of rapid climatic change is a good chronological correlation of the records derived from the ice caps, the terrestrial and the palaeo-oceanographic records. Several climate records in Africa show abrupt changes during the last deglaciation, but not many are well enough dated to correlate them with records from elsewhere.
Core T89-16 is taken in front of the Congo River mouth, where actually about 10 % of the water are river water. This core offers the opportunity to connect signals of the continental and marine climate 20 to 8 ka BP, with a time resolution of about 100 y for the greater part of the core. We measured 18O in Globigerinoides ruber, a planktic foraminifer species living in the low-salinity plume. A correction for the regional marine signal gives a record depending on palaeotemperature and palaeosalinity only. The palaeotemperature is reconstructed with alkenone data. This leaves us with the palaeosalinity values that are directly coupled with the Congo outflow.
At the time of writing the measurements were not finished. The first results will be presented at the meeting.
Cold polar ice sheets are near perfect archives of past climate states and climate indicators. Next to the more or less direct temperature proxies such as d18O or dD we can extract past accumulation rates (or precipitation rates) as well as determine past atmospheric composition and aerosol load with resolutions down to seasonal. From ice cores drilled in North East Greenland examples of high resolution records will be shown exemplifying the quality and confidence limits of such data sets in order to demonstrate their suitability as boundary values for paleoclimate modelling.
Climatic conditions during the last interglacial (125,000 years before present) are investigated with the atmosphere-ocean general circulation model ECHAM-1/LSG and with the climate system model CLIMBER-2. Comparison of the results reveals broad agreement in most large-scale features, but certain discrepancies are also detected. The fast turnaround time of CLIMBER-2 permits to carry out a number of sensitivity experiments to try to explain the possible reasons for these differences. Contrary to the general expectation, neither model reproduces a significant increase of the globally averaged temperature at the last interglacial. Reconciliation with this assumption might be provided by the inclusion of an interactive vegetation component, which translates into a significant amplification of the warming in response to the enhanced summer insolation in the Northern Hemisphere.
d18O records from Greenland ice cores indicate numerous rapid climatic fluctuations during the last glacial. North Atlantic marine sediment cores record many fluctuations of sea surface temperature and of ice-rafted debris. In contrast, very few continental last glacial records provide the temporal resolution and environmental sensitivity to reveal the geographical extent and nature of the impacts of these environmental fluctuations upon continental regions. We have demonstrated previously that Lago Grande di Monticchio, S. Italy, not only can provide such a record, but also one with an independent chronology. Here we present new data from Monticchio extending to >100 ka. These data reflect numerous rapid fluctuations, with major vegetation changes occurring in <200 yr. We also present new, high resolution d18O data from a Mediterranean sediment core linked directly to our lacustrine record by tephrochronology. Whereas overall our record correlates well with ice-cores, beyond ca 76 ka Monticchio reveals climatic fluctuations not seen in ice-core data. We hypothesise that this reflects limitations of ice cores at greater depth. The Monticchio record indicates that rapid climatic fluctuations characterise not only the last glacial but also the preceding marine isotope stage 5, extending to beyond 100 ka BP.
Detailed lithologic, microfaunal, and radiocarbon analysis of peaty deposits from two salt marshes in Connecticut (USA) has yielded similar records of small-amplitude (10-25 cm) sea-level variations on a century time scale for the past 1500 year. The timing of these variations is in overall correspondence with the Medieval Warm Period, Little Ice Age and period of Modern Global Warming as defined by a recent Sargasso Sea surface temperature record and a North Fennoscandian summer temperature time series. Generally, sea level lags temperature by c. 100 yr, suggesting that steric and ice melt effects are involved. In some instances, however, no obvious lag is observed.
The comparison of climate model results with reconstructions based on paleodata is an important tool in paleoclimate research. This model-data comparison may produce discrepancies that are related to the omission of important boundary conditions. One example is the study on the Younger Dryas (YD, 12.9-11.6 ka cal BP), a marked cooling event that interrupted the general warming trend during the last deglaciation. YD climate reconstructions based on paleobotanical data show that summer temperatures in Europe were a few degrees lower than today. In contrast, climate model experiments on the YD produced a warming over continents, generated by a higher insolation (July insolation was 40 W/m2 more than today at 50N). This data-model discrepancy poses a problem, as it suggests the existence of an unknown forcing factor that works to temper the orbitally forced warming. A possible factor is the effect of permafrost, as there is ample field evidence that permafrost existed in W Europe during the YD. We present a new experiment, performed with the ECHAM4-T42 model, in which the effect of YD-time permafrost is included. We introduced in the model a frozen and waterlogged soil in those regions for which YD-time permafrost exists. This experiment resulted in summer temperatures in Europe that are close to reconstructed values. Moreover, over the N Atlantic Ocean a distinctively different atmospheric circulation is simulated that acts to reinforce these cool YD summer conditions in Europe. This result suggests that during the last deglaciation permafrost played a critical role within the climate system.
Paleo proxy data obtained from valley glaciers, ice core measurements and dendrochronology provide important information on the evolution of the regional or local climate. General Circulation Models integrated over a long period of time could help to understand forcing mechanisms underlying the climate system behaviour. For a systematic interpretation of in situ paleo proxy records, a combined method of dynamical and statistical modeling is proposed. Local 'paleo records' can be simulated from GCM output by first undertaking a model-consistent statistical downscaling and then using a process-based forward modeling approach to obtain e.g. the behavior of valley glaciers and the growth of trees under specific conditions. The simulated records can be compared to actual proxy records in order to investigate whether e.g. the response of glaciers to climatic change is reproduced by models and to what extent climate variability obtained from proxy records (with the main focus on the last millennium) is represented. For statistical downscaling to local weather conditions, a multiple linear forward regression model is used. Daily sets of observed weather station data and various large-scale predictors at seven pressure levels obtained from ECMWF re-analyses are used for development of the model. Daily data give the closest and most robust relationships due to the strong dependence on individual synoptic-scale patterns. For some local variables, the performance of the model can be further increased by developing seasonal specific statistical relationships. The model is validated using both independent and restricted predictor data sets. It is then applied to a long integration of an ECHAM4 mixed layer ocean GCM experiment and to the ECHAM4/OPYC3 coupled GCM. The dynamical-statistical local GCM output within a region around Nigardsbreen glacier, Norway is compared to nearby observed historical station data for the period 1868-1993. Patterns of observed variability are realistically simulated for this location. The local output produced by the described method is used to force a process-based mass balance and a dynamic glacier model for Nigardsbreen glacier. Realistic mass balance variations are obtained compared to observations for the period 1962-1995. A 600 year control integration demonstrating glacier length variations exclusively due to internal variations in the climate system as well as transient runs for 1860-2050 are performed and compared to proxy data for the period 1710-1995.
Air-sea interactions in the ECHAM3/LSG integration are studied based on prototypes of interaction processes, which represent the often used simple atmosphere models (SAMs) in ocean-only studies. The results suggest minimum conditions, which have to be satisfied by a SAM, if it is supposed to mimic an atmospheric GCM. A way how to empirically design such a SAM is proposed.
MoBidiC is a zonally averaged climate model of intermediate complexity in which a quasi-geostrophic atmosphere model is coupled to a three-basin (latitude-depth) ocean-sea ice model. The atmospheric component of MoBidiC contains an interactive representation of the hydrological cycle based on meridional transport of moisture and zonal redistribution of precipitations. The sea ice model included enables to represent explicitly the climatic effect of ice formation and ablation and their consequences on the ocean circulation. The interconnected ocean basins allow this model to simulate the main thermohaline circulation processes. MoBidiC is designed for long term climate simulations and has a low computational cost. The model validation on present-day climate indicates that realistic temperature, precipitation and ocean circulation patterns are simulated, shortcomings appear on the atmosphere dynamics representation and on the ocean salinity distributions. Recent results on present climate and work perspectives on paleoclimatic application will be discussed in this presentation.
Climate modelling can be used in sensitivity analyses of past climates, both for short, abrupt oscillations and for longer term climate conditions. For example, temperature and precipitation effects on land climate caused by the sudden cooling of ocean surface water, e.g. during Heinrich events, can be estimated by model experiments. Similarly, the climatic inferences induced by volcanic eruptions can be numerically simulated. Examples of longer lasting environmental characteristics which may have affected climate circulation patterns can also be modelled. The presence of permafrost, for instance, induces humid soil conditions, and therefore possible changes in precipitation and temperature (Renssen et al., in prep.). Also vegetation may play an important role on the modelled temperature by its effects on surface roughness, albedo and evapotranspiration (Renssen et al., submitted).
A second category of applications is the testing of hypotheses on past climates. Examples are the derivation of circulation patterns during the Late Glacial and during the Last Glacial Maximum over Europe under prescribed boundary conditions. For example, such climatic simulations may give indications of past wind directions (Isarin et al., 1997). More particularly, when dissimilarities appear between models and reconstructions the need for reappraisal of old hypotheses or new interpretations may come out.
The atmosphere-ocean general circulation model ECHAM3/LSG has been employed in a number of climate change and climate variability studies. The main points of interest are the possible future climate change due to an increase of greenhouse gases and the investigation of past climate states. In order to simulate a large number of experiments and/or a period of several hundred years with the AOGCM, the periodically synchronous coupling technique has been applied in some of these experiments. The method, which reduces the consumption of computer resources, will be described. Examples for 'paleo' simulations of the synchronously coupled ECHAM3/LSG as well as for the periodically synchronously coupled one will be presented.
A 1000-year control experiment with the coupled atmosphere/ocean/sea-ice model ECBilt is analysed with respect to low-frequency variability in Surface Air Temperature. Spatial patterns, described in terms of the centers of action and possible teleconnections, are found to depend on the timescale (multi-decadal to centennial) and the season. Model results are compared with earlier derived empirical results (Shabalova and Weber, 1999). Due to the limitations of the used paleo dataset (e.g. limited spatial coverage), it is not possible to validate all features of variability seen in the model. Focus of the intercomparison will be on those characteristics of variability which seem robust in the data (e.g. spatial patterns in well-covered regions, dominant timescales, the seasonality of the signal). The mechanisms generating low-frequency variability in the model will be discussed.