SUMMARY REPORT FOR THE GAIM 6000 YR BP
EXPERIMENT AND BIOME 6000


Click here for a summary of "The Global Palaeo-Vegetation Project (BIOME 6000) Workshop", 6-10 October 1998, Jena, Germany

Click here for a summary of "BIOME 6000: Towards a global palaeovegetation data set", 11-13 May 1994 Horby, Sweden

The experiment aims to quantify the importance of biogeophysical feedbacks in the climate system by comparing the performance of coupled and uncoupled climate-biosphere models which are driven by the Earth's orbitally-induced change in seasonal insolation from 6000 yr BP to present. The existing, extensive coverage of paleodata describing the state of the terrestrial biosphere at 6000 yr BP should provide a decisive standard against which to evaluate the model results.

Biogeophysical feedbacks in the climate system arise because energy and water fluxes across the land-atmosphere interface are mediated by vegetation. Recent work on the sensitivity of climate models to changes in vegetation distribution has established that biogeophysical feedbacks can be of major significance for climate. Modelling of Late Quaternary climates allows the possibility of quantifying the importance of these feedbacks by comparing model simulations with palaeoenvironmental data, e.g. on sea-surface temperatures, lake levels, and vegetation. Data onpast vegetation come from pollen and plant macrofossil analysis, mainly in sediment cores from lakes and peatlands. Such data are very extensive, and major efforts are underway to improve coverage in previously understudied regions, such as Siberia, central Asia, the high Arctic, and the wet tropics.

Several earlier data syntheses have focused on the time interval around 6000 yr BP. This interval is also a focus for climate modelling, e.g. in the Paleoclimate Modelling Intercomparison Project (PMIP). 6000 yr BP allows a pure "orbital forcing" model experiment: the seasonal and latitudinal distribution of insolation was disserent from present, while ice sheets and atmospheric CO2 concentration had stabilized at their Holocene states. The GAIM 6000 yr BP experiment aims to use palaeodata for 6000 yr BP as the yardstick by which to compare simulations of the first-order effects of orbital forcing as in PMIP, with simulations that incorporate biogeophysical feedback through two-way coupling to global vegetation models. This aim is to be achieved through a combination of modelling, data synthesis and data analysis.

Key features of the 6000 yr BP climate include the more poleward extent of northern forests, and the greatly expanded African and Asian monsoon regions with vegetation and lakes in areas that are now arid. These features can be simulated by PMIP-type model experiements, but current models apparently underestimate their magnitude. Sensitivity experiments with appropriate changes to the land-surface conditions suggest that such changes may have provided the necessary amplification of the climate system's response to orbital forcing. Further analysis requires coupling of climate and ecosystem process models. Some experiments have already been made, using asynchronous coupling between atmospheric models (run for 5-15 years) and the BIOME model. Experiments by Martin Claussen at the Max Planck Institute for Meteorology, Hamburg, for example, suggest that under present orbital conditions the coupled African vegetation/monsoon system possess two stable states - the relatively arid present state, and an alternative state in which part of the western Sahara is vegetated and the subtropical anticyclone is located far to the west. The alternative state can only be reached from a different (vegetated) initial land-surface condition. On the other hand, under 6000 yr BP orbital condidtions, the coupled system converges on a "green" Sahara consistent with the pattern of vegetation shown by the palaeodata.

Further intriguing insights have been provided by Nathalie de Noblet at the Laboratoire de Modelisation du Climat et de l'Environnement, Paris, who has shown that vegetation feedbacks at the start of the last glacial period, 115000 yr BP, may have greatly increased snow depth and duration in the supposed initiation centres for the mid-latitude ice sheets. Coupled model experiments for 6000 BP and other times are also envisaged with the IBIS (Integrated Biosphere Simulator) integrated vegetation dynamics/land-surface process modelling framework developed at the University of Wisconsin, Madison by Jon Foley, as part of the NSF-supported TEMPO (Testing Earth System Models with Paleoenvironmental Observations) project.

Such efforts create a need for systematically constructed and properly documented global palacodata sets, above all for biome distribution at key times in the past, so that more rigorous palaeodata-model comparisons can be organized. The GAIM 6000 yr BP experiment has therefore spawned a data project, called BIOME 6000. BIOME 6000 is jointly sponsored by GAIM and three other IGBP programme elements, IGBP-DIS, GCTE, and PAGES. The TEMPO project will also contribute to the goals of BIOME 6000. BIOME 6000 aims to assemble the existing palaeovegetation data for 6000 yr BP (and, secondarily, the last glacial maximum) on all of the continents, and to use this data to construct a "palaeobiome map" using standard methods.

A suitable "biomization" method has been developed for use in BIOME 6000. It is based on prior assignments of a taxa to plant functional types (PFTs) followed by application of a fuzzy logic algorithm to assign pollen assemblages to biomes, which are combinations of PFTs. This method has been tested using modern pollen assemblages in Europe, East Africa and West Africa. 6000 yr BP biome maps have now been created using the method for Europe and most of Africa, and are under construction for Australasia and eastern North America. A BIOME 6000 network including experts on all of the continents was inaugurated at a workshop in May 1994 (see a forthcoming GAIM report for further details), and several initiatives are underway to assemble the data from other regions. Plans for 1996 include an August meeting of scientists from Russia/NIS, to develop the data set for northern and central Eurasia. The first global 6000 yr BP "product" should become available in 1997, allowing comparison with the next generation of coupled model experiments.

We plan to explore several new initiatives to help construct biome data sets for 6000 yr BP (top priority) and the last glacial maximum (LGM, second priority). We plan to complete a first global 6000 yr BP biome data set during 1997. The global data set will provide a comprehensive test of preliminary hypotheses already emerging from modelling studies and data-model comparisons. For example:

* Paleoclimate Modelling Intercomparison Project (PMIP) simulations seem to underestimate both the African and Asian monsoon expansions and the poleward spread of forests at 6000 yr BP [Guiot et al., in press; Texier et al., in press; Yu and Harrison, in review]. Additional amplifying factors (ocean and land surface) must be included in order to achieve a realistic simulation of 6000 yr BP conditions.

* Monsoon expansion can be amplified by prescribing appropriate land surface changes in the Sahara (vegetation and soil water-holding capacity) [Kutzbach et al., submitted]. The true poleward spread of forests at 6000 yr BP can be approximated by the inclusion of Arctic sea-ice feedback and appropriate land-surface changes [Bartlein et al., in review; Texier et al., in press].

* At least one coupled atmosphere-biosphere model (ECHAM-BIOME) produces two stable states of the vegetation-climate system over northern Africa: one with an extensive Sahara (as present), and one with a westward-displaced subtropical anticyclone over the Atlantic and much more extensive vegetation, especially in West Africa [Claussen, 1994]. Under 6000 yr BP conditions, however, only the "green Sahara" configuration is stable [Claussen and Gayler, 1995a; Claussen and Gayler, 1995b; Noblet et al., 1995]. We propose to explore the implications of these preliminary results by running a more in-depth simulation with several models.

By running comparable experiments with different coupled models, using the 6000 yr BP lake and biome data as a benchmark, we expect to assess the generality of such results with respect to their dependence on different models and coupling strategies. We also expect to be able to compare the data with more comprehensive model results that include feedbacks from land-surface hydrology and sea-surface temperature changes, thus shifting the emphasis from the problem of quantification of biosphere-atmosphere feedbacks in isolation, to the problem of explaining the observed changes in terms of all of the relevant Earth system processes.

As the 6000 yr BP data set grows, a new data synthesis priority will be the LGM (18000 14C-yr BP). Preliminary modelling results using both the LMCE and CCM1 atmospheric models with the new BIOME3 coupled biogeography and biogeochemistry model [Haxeltine and Prentice, in press] suggest that plant-physiological effects of low atmospheric CO2 content may have contributed substantially to the expansion of tropical grasslands [Foley et al., 1995; Prentice et al., 1995a] and lowering of treelines [Jolly and Haxeltine, submitted] at the LGM. The modelled effects of low CO2 are dramatic and testable using the paleobiome data. As these modelled effects partly depend on altering the competitive balance between C3 and C4 plants, we will also include the increasingly large body of information of d13C shifts in terrigenous organic matter as an additional proxy. Further coupled model experiments using BIOME3 will help to assess the significance of these effects in amplifying vegetation-atmosphere feedbacks.

These activities will provide one of the building blocks for the Paleo trace gas "Challenge", described below. Simulated global fields of variables including vegetation type/structure and net primary productivity under both warm interglacial and full-glacial conditions will be essential in order to model CO2 and CH4 exchanges between the land and atmosphere. Paleobiome data for 6000 yr BP and the LGM will be important to test the realism of these fields. The experience to date with 6000 yr BP strongly suggests that the achievement of realistic fields will require, at least, coupled simulations of the physical climate and the terrestrial biosphere.