Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - CLAVICHORD (large-scale CLimate And VegetatIon CHanges OveR the last Deglaciation)

Teaser

The last deglaciation (from ~21,000 to 9,000 years ago), during which the huge ice sheets over the North America and Scandinavia melted, is a period of tremendous climate and environmental changes. These changes are documented by physically based paleoenvironmental indicators...

Summary

The last deglaciation (from ~21,000 to 9,000 years ago), during which the huge ice sheets over the North America and Scandinavia melted, is a period of tremendous climate and environmental changes. These changes are documented by physically based paleoenvironmental indicators (such as oxygen or carbon isotopes in ice cores and marine cores), but less by biologically based data (such as paleo-vegetation). Paleo-vegetation data is, however, crucial to document continental-scale climate and environmental changes over the deglaciation. This project aimed, for this period of the last deglaciation, at 1) building a comprehensive documentation of vegetation and climate changes over terrestrial areas from widely available pollen data, 2) assessing the impact of both climate and atmospheric CO2 changes on the vegetation and 3) investigating the changes in large-scale atmosphere circulation and hydrological cycle responsible for these surface climate and vegetation changes. These results should contribute to quantifying the range of possible changes in these circulations in the future. Finally, this project provides new benchmarking data for understanding environmental changes and evaluating climate models which are used for climate projections.

Pollen-based climate reconstructions have generally been implemented with conventional statistical approaches (e.g., modern analog, regression and response-surface techniques), which assume that vegetation responses to climate remain the same through time and that modern data contain all the information necessary to interpret fossil pollen data. However, plant-climate interactions are sensitive to atmospheric CO2 concentration, and present-day relationships between climate and present plant distributions may not be representative of those under much lower atmospheric CO2 concentration. That is the reason why pollen-based climate reconstructions over the last deglaciation had hardly been performed. Therefore, for pollen-based climate reconstruction in this project, we used an “inverse modeling through iterative forward modeling” (IMIFM) approach (Izumi and Bartlein, 2016). The IMIFM approach, which originated as an inverse vegetation-modeling approach, have developed to overcome some debatable issues of conventional statistical approaches for climate reconstruction, as was mentioned earlier. The IMIFM approach is based on a forward modeling approach with equilibrium vegetation models (BIOME4 and BIOME5-beta) that uses input of climate, soil properties, surface air pressure and atmospheric CO2 concentrations to mechanically simulate vegetation under a specific environment. As a result, we consider several factors including atmospheric CO2 concentration for pollen-based climate reconstruction over the world.

Our biomization procedure and IMIFM approach were applied to several regions and validated by our former papers, Izumi and Lézine (2016) for Africa and Izumi and Bartlein (2016) for North America. Moreover, we enlarged the application beyond the last deglaciation period (i.e. vegetation and climate changes over the last 90,000 years at the Lake Bambili, Cameroon).

Work performed

Pollen-based climate reconstructions:

1) We collected the pollen data from pollen databases (for North America and Africa) and from Basil Davis (Université de Lausanne) for Europe, Asia, Siberia.
2) We standardized this data for the biomization approach (i.e. the conversion of pollen to biomes) and then implemented the biomization procedure at each region.
3) We implemented the IMIFM approach at each pollen site for the present state for the validation. Then, we applied the same processes to the past (i.e. the last deglaciation) data. We validated our biomization approach and IMIFM approach (e.g. Izumi and Lézine, 2016 and Izumi and Bartlein, 2016). In this project, we also upgraded the IMIFM approach about comparison of observed biomes with simulated biomes.

In spite of coarse resolution temporal data, our vegetation and climate reconstructions show establishment of forests (from xerophytic shrublands and grasslands) and gradual warming and wetting over the last deglaciation, but our several data do not appropriately show the abrupt climate change events (e.g. Heinrich Event 1, Bølling Warming and Younger Dryas cooling). We still struggle with climate reconstruction over Europe, Asia and Siberia (e.g. the LGM climate at some sites of Europe and Siberia is similar to or warmer than the present). Our climate reconstruction is based on the biomes and their scores, and thus the biomization approach over the regions might cause problems and we need to improve the approach. On the other hand, we also got the new biome data over the Europe and Siberia from Mary Edwards (University of Southampton) and implement the IMIFM approach using the new data now.

Assessing the impact of both climate and atmospheric CO2 changes on the vegetation:

1) In the IMIFM approach, we can assess the impact of atmospheric CO2 and of climate by setting them as constant in separate experiments.
2) We also run a dynamic vegetation model, LPJ-GUESS with the Simulation of the Transient Climate of the Last 21,000 Years (TraCE-21 ka) at each pollen site over the Africa and North America.
3) We are still implementing data-model comparison methods for transient vegetation and climate variables.

Investigating the changes in large-scale atmosphere circulation and hydrological cycle responsible for these surface climate and vegetation changes:

1) According to our vegetation and climate reconstructions over the Africa, we investigate the changes in large-scale atmosphere circulation and hydrological cycle over the equatorial Africa (i.e. east side vs. west side of the Congo basin) during the last deglaciation.
2) We are still investigating the changes in large-scale atmosphere circulation and hydrological cycle over the North America.

Final results

Our approach of pollen-based climate reconstruction allows us to avoid both no-analog and wrong-analog problems (e.g., Jackson and Williams, 2004) and to assess the potential bias in reconstructions that may result from varying atmospheric CO2 concentrations. In particular, because atmospheric CO2 concentration influences plant water use efficiency, with low CO2 concentrations requiring a greater consumption of water to maintain the same rate of photosynthesis (Prentice and Harrison, 2009), conventional statistical approaches could “misread” the generally more xeric vegetation at the Last Glacial Maximum than at present as indicating drier-than-present conditions, even if moisture did not change. As a result, our project shows that pollen-based climate reconstruction beyond the Holocene period works and more large-scale data is available for understanding of vegetation and climate changes beyond the Holocene. This project also provides new benchmarking data for understanding environmental changes and evaluating climate models which are used for climate projections.