Modeling the Evolution of Climate and Sea Level over the third Millenium
MILMO workshop - ABSTRACTS
Brussels, 16-17 November 2004
A large number of modelling studies have addressed the likely effects of greenhouse-gas-induced climate change over the 21st century, but the longer-term effects have received much less attention. Yet these effects could be very significant, as a persistent increase in global mean temperature may ultimately influence the large-scale processes in the Earth system that play a central role in determining the global climate. Investigating the long-term response of climate and sea level to human activities requires the operation of global three-dimensional models which encompass all relevant components of the climate system (i.e., the atmosphere, the oceans, the ice masses, the upper EarthÕs mantle, and the living world). But most importantly, these models must be computationally fast enough so that integrations of longer duration and a larger number of sensitivity experiments can be performed than is usually the case with climate general circulation models. To achieve this requires some simplifications in the most CPU-time demanding model component, which is usually the atmospheric component. Within the MILMO project, we built such a model (LOVECLIM) by coupling an efficient three-dimensional atmosphere-vegetation-sea-ice-ocean model with a model of the oceanic carbon cycle and with thermomechanical models of the Greenland and Antarctic ice sheets. This coupled model was then used to study the evolution of climate and sea-level changes over the period 1500-2000 in response to both natural and anthropogenic forcings. A series of projections of climate and sea-level changes over 2000-2100 were also performed with the model. In these experiments, the model was driven by greenhouse-gas and sulphate-aerosol concentrations evolving with time according to the IPCC SRES scenarios. Finally, projections will be conducted during the last few months of the project over the whole third millennium in order explore the threat of possible rapid climate and sea-level changes during this period.
In its Third Assessment Report (2001), the Intergovernmental Panel on Climate Change (IPCC) reported that in the absence of climate protection policies, continued greenhouse gas emissions would increase global temperature by 1.4 to 5.8C between 1990 and 2100, depending on which scenario and model is used. According to the IPCC, the projected changes in climate would result in significant, often adverse, impacts on many ecological systems and socio-economic sectors, including food supply and water resources, and on human health. In some cases, the impacts are potentially irreversible. Developing countries are the most vulnerable. But the impacts will be far from negligible in Belgium, as shown by a recent synthesis of the expected impacts . Some of the results obtained by MILMO were used in that study to assess possible future scenarios of sea level rise. The United Nations Framework Convention on Climate Change, now reinforced by the Kyoto Protocol, intends to safeguard our climate for present and future generations. But climate is changing faster than the production and consumption patterns which affect it. Some adaptation policies will be necessary. But mitigation at global level cannot be effective without the developing countries participation. They will not move if developed countries donÕt assume their historical responsibilities. The 4th IPCC report, due in 2007 will review the scientific information available about climate processes, impacts, and policies. It will have to be policy-relevant, without being policy-prescriptive. It will be very important to facilitate the implementation of the Kyoto Protocol, and inspire the negotiations about the more ambitious emission cuts needed after 2012 if we want to keep our planet habitable while allowing sustainable development for both North and South.
An ensemble of simulations is performed with a three-dimensional climate model over the last millennium driven by the main natural and anthropogenic forcing. The results are compared to available reconstructions in order to evaluate the relative contribution of internal and forced variability during this period. At hemispheric and nearly hemispheric scale, the impact of the forcing is clear in all the simulations and knowing the forced response provides already a large amount of information about the behaviour of the climate system. Besides, at regional and local scales, the forcing has only a weak contribution in the simulated variability compared to internal variability. This result could be used to refine our conception of 'Medieval Warm Period' and 'Little Ice Age' (MWP and LIA). They were global phenomena, since the temperature averaged over the Northern Hemisphere was generally higher (lower) during those periods because of a stronger (weaker) external forcing at that time. Nevertheless, at local-scale, the sign of the internal temperature variations determines to what extent the forced response will be actually visible or even masked by internal noise. Because of this role of internal variability, synchronous peak temperatures during the MWP or LIA between different locations are unlikely.
A series of climate-change projections are conducted with LOVECLIM, a three-dimensional Earth system model of intermediate complexity that consists of a quasi-geostrophic atmospheric model, an ice-ocean general circulation model, a dynamical model of the terrestrial biomass, a thermomechanical model of the Greenland and Antarctic ice sheets, and a model of the oceanic carbon cycle (which is not interactive in this study). Climate changes during the 21st century are first studied through simulations performed with the model driven by various IPCC's SRES scenarios for greenhouse-gas and sulphate-aerosol concentrations. The model performance is assessed by comparing its results with similar results obtained with climate general circulation models. This analysis reveals that the model response is within the range of current uncertainty, although a bit weak. Idealised experiments are then carried out over the whole third millennium to investigate the possibility of human-induced abrupt climate changes. In the higher greenhouse scenarios considered, the modelled World Ocean's thermohaline circulation collapses, with significant impacts on the North Atlantic climate, and then recovers progressively with time.
Previous simulations by using coupled atmosphere-ocean-biosphere models revealed a strong feedback between atmosphere and vegetation in Northern Africa which appear to enhance changes in Northern African summer monsoon. In the past, such changes were triggered by changes in insolation due to variations in the Earth's orbit. The potential of fast changes during the mid Holocene were predicted in agreement with palaeoclimatic reconstructions. Here, we theoretically explore potential changes in Northern African climate triggered by an increase in atmospheric CO2, and we show that some expansion of grassland into the Sahara is possible if the atmospheric CO2 concentration increases well above pre-industrial values and if vegetations growth is not disturbed by grazing, for example. Vegetation migration into the Sahara could be rapid, but it would not reach early Holocene coverage. Analysis of atmospheric dynamics reveals that the relative role of mechanisms which lead to a reduction of the Sahara differ between past and future.
ARPEGE is a global atmospheric model primarily designed for short- and medium-range weather prediction. A climate version of this model is used for running IPCC scenarios of the 21st century. One of the version is suitable for regional simulation over Europe (grid size 50 to 60 km) thanks to variable horizontal resolution. In the framework of the PRUDENCE European project, simulations of the 1961-1990 and 2071-2100 period (A2 scenario) have been produced. The response of the mean climate for temperature and precipitation is rather credible, since the simulation of present climate is fairly realistic and the climate change very similar to what other partners in the project produce. The direct use of extreme values from the simulation poses a few problems, and we consider that an extreme temperatures produced by the model cannot be directly compared to an observation value, but to the distribution of simulated temperatures at a particular location for a particular season. Thus we assume, for example, that the 99th percentile of the model distribution in present climate simulation corresponds to the 99th percentile of the observed distribution. In the framework of the French IMFREX project, we have thus calculated frequencies of summer heat waves in a possible future climate. This method, which requires a dense network of daily observed temperature over 50 years, shows that the frequency of events like 2003 might be multiplied by 10 at the end of the century, whereas winter cold waves below -5¡ should nearly disappear.
The climate feedbacks caused by increased atmospheric CO2 levels have the potential to significantly affect the uptake of carbon by the ocean and by the continental vegetation (e.g. Joos et al, 1999; Dufresne et al., 2002; Ewer et al., 2004). Using LOVECLIM, a coupled model of intermediate complexity which incorporates a comprehensive global carbon cycle, we investigate the magnitude and sign of different processes and their effects on CO2 sinks over the period 1750-2300. Land cover change, radiative warming, enhanced stratification and circulation changes in the ocean are among the factors having a significant impact on the carbon cycle. We analyze the response of the system either with diagnostic or prognostic CO2 concentrations. In the first case, atmospheric CO2 concentrations are constrained to follow the historical record and future scenarios. In the prognostic case, atmospheric carbon dioxide concentrations results from the anthropogenic emissions and the evolving ocean-atmosphere and land-atmosphere fluxes; several future emissions scenarios are considered. Under prescribed atmospheric CO2 concentrations climate change reduces further carbon uptake by the oceans. Land cover change and climate both impact negatively the uptake by the continental vegetation. On the contrary, in prognostic CO2 simulations, the ocean uptake appears less affected by climate change; any positive feedbacks leading to a larger atmospheric CO2 pressure is quickly compensated by an increase of the ocean invasion rate. We also examine the implications for future CO2 stabilization scenarios.
A complex coupled earth system model, consisting of an atmosphere and ocean general circulation model (ECHAM-LSG), a dynamic terrestrial vegetation model (LPJ), an ocean biogeochemistry model (HAMOCC) and a thermodynamic icesheet model (SICOPOLIS), is being developed. The model will enable studying the processes within as well as between the components of the earth system. The terrestrial carbon cycle is modelled with the dynamic global vegetation model LPJ. Ten plant functional types are defined, and the occurrence of the plant functional types is modelled using bioclimatic limitations. Within the plant functional types, growth and decay of four living biomass carbon pools and three litter carbon pools are calculated. The soil contains two carbon pools. The marine carbon cycle is modelled with the ocean biogeochemistry model HAMOCC. It models 13 geochemical tracers, including the main parameters for the carbon cycle (CO2, CaCO3 and particulate organic carbon), alkalinity, phosphate, oxygen and sillicate. Biogeochemical processes in the ocean include primary production and remineralization. Both the terrestrial and the marine carbon cycle models include 13C and 14C isotopes. Experiments with an increasing CO2 concentration were performed with the coupled model, showing the reaction of the atmosphere, land surface and ocean models on higher CO2 concentrations.
CARAIB (CARbon Assimilation In the Biosphere) is a Dynamic Global Vegetation Model. It was originally designed to study the carbon cycle in the terrestrial biosphere and especially the contribution of land ecosystems to atmospheric CO2 sequestration (Warnant et al., 1994; Nemry et al., 1996; Gérard et al., 1999). The model integrates various modules describing (1) the hydrological cycle, (2) photosynthesis and stomatal regulation, (3) carbon allocation and biomass growth, (4) heterotrophic respiration and litter/soil carbon storage, and (5) plant competition and biogeography. CARAIB is able to describe biomass growth and, hence, can simulate transient changes in carbon reservoirs. The biogeography module calculates the distribution of the model plant functional types (PFTs) as a function of their respective productivity calculated by the model. Model PFT assemblages can then be translated into biomes to produce palaeovegetation maps. This method has been used to reconstruct Last Glacial Maximum (Otto et al., 2002), as well as Late Miocene vegetation (François et al., in press). In parallel, the model has been substantially adapted to be used at the scale of the European continent. These adaptations include the definition of finer PFTs (Bioclimatic Affinity Groups, BAGs), allowing a more precise description of European vegetation. These BAGs have been defined from a statistical analysis of current natural vegetation distribution in Europe (Laurent et al., in press). Some high resolution simulations of this improved model over Europe will be presented for the present and the LGM. A quantitative comparison of the model distribution with vegetation and pollen data is currently undertaken for all BAGs considered in the model. Future projections of European or Mediterranean vegetation are also planned.
The surface ocean is everywhere saturated with respect to calcium carbonate (CaCO3), and this has probably been the case for at least the last 20 million years. Yet increasing atmospheric CO2 reduces ocean pH and carbonate ion concentration and thus the level of saturation. Here we show with ocean data and models that due to this anthropogenic acidification, some surface waters will become undersaturated within decades. When atmospheric CO2 reaches 550 ppmv, in year 2050 under the IS92a business-as-usual scenario, Southern Ocean surface waters begin to become undersaturated with respect to aragonite, a metastable form of CaCO3. By 2100 as atmospheric CO2 reaches 788 ppmv, undersaturation extends throughout the entire Southern Ocean (<60S) and into the surbarctic Pacific. Meanwhile, Weddell Sea surface waters even become slightly undersaturated with respect to calcite, the stable form of CaCO3. These transient changes due to increasing anthropogenic CO2 are much larger than changes due to climate change as well as seasonal, interannual, and decadal variability. This undersaturation severely threatens high-latitude aragonite secreting organisms including cold-water corals, which provide essential fish habitat, and shelled pteropods, i.e., zooplankton that serve as an abundant food source for marine predators.
Volume changes of the Greenland and Antarctic ice sheets have the potential to significantly increase the rate of global sea-level rise in future warmer climates. Crucial aspects are how climatic changes will affect the ice sheet's mass balance and how ice dynamics will react to the imposed environmental forcing. This is in addition to the longer-term background trend from adjustments as far back as the last glacial period. These questions are addressed with 3-D thermomechanical ice sheet/ice shelf models which have been fully interactively coupled with climate models of varying complexity. A first series of experiments considers the coupling between models of the Antarctic and Greenland ice sheets with the LOVECLIM model for the period between 1500 and 3000. Whereas changes in the polar ice sheets are found to be small over historic times, important changes are expected for the next 1000 years. It is found that for SRES scenario A2 kept constant after the 21st century, the Greenland ice sheet almost completely disappears within a period of about 1000 years. Together with significant shrinkage of the Antarctic ice sheet and the contributions from mountain glaciers and thermal expansion of the world's oceans, this may cause a global sea level rise in excess of 10 m by the year 3000. A second series of experiments investigated the issue of reversibility of Greenland ice sheet melting once ice sheet disintegration has been initiated. For that purpose, the Greenland ice sheet model was inserted in the HadCM3 atmosphere-ocean general circulation model. In this experiment, the Greenland ice sheet is found to disintegrate to less than 5% of its current volume within 3000 years under constant 4xCO2 conditions. With the ice sheet removed, the model shows that it would not regrow to its present state for present climate conditions, indicative of hysteresis. At different moments in time, we additionally interrupted the melting process by inserting a 1xCO2 climate and let the ice sheet evolve to a new steady state. It was found that there exists a point-of-no-return once ice sheet disintegration has set in beyond which complete removal of the ice sheet becomes irreversible, even if climatic conditions were to revert to present-day conditions. This point may already be reached after 250 years of ice-sheet melting under a non-extreme greenhouse warming scenario.
Predictions of climate change and its consequences for the 21st century and beyond rely principally on coupled atmosphere-ocean general circulation models. Because these models are computationally too expensive to run for long periods of simulated time with a range of scenarios, and as an aid to interpretation, it is useful to consider simpler ideas which can represent AOGCM results. Climate sensitivity determines the steady-state climate change due to imposed forcing. This final steady state may take millennia to attain, because of the slow processes associated with ocean heat uptake, which limits the rate of change, and is also associated with sea level rise due to thermal expansion. I will discuss the some aspects of the use, relationship and quantification of these concepts.
Projections of sea-level rise from glaciers and ice caps for the next century and beyond should be based on an assessment of the ice available for melting. Projections to date are based on all regions except Greenland and Antarctica, yet no sound estimates for this volume of ice and its potential for sea level rise are evident in the literature. An ice cap data set is compiled allowing the separate treatment of glacier area coverage data. Glacier inventory data are comprehensive enough in some regions to allow the estimation of glacier size distributions. The differences in the distributions are related to a metric of the regional topography, allowing glacier size distributions to be estimated on a 1o latitude longitude grid of glacier containing cells. Appropriate volume-area scaling for glaciers and for ice caps gives global estimates of glacier and ice cap volumes by size class. This leads to an estimate of the total ice volume of 0.087 ± 0.005 106 km3 and a sea-level rise equivalent of 0.240 ± 0.015 m. The glaciers and ice caps contribute 41% and 59% to these estimates respectively. Some aspects of modelling glacier and ice cap melt will be discussed and some idealized projections will be presented.