Biome-scale vegetation dynamics in North America since the last glacial maximum: Maps and reconstructions from fossil pollen data and the testing of biogeography models.

摘要

The changing biome distributions reconstructed from fossil pollen data illustrate the dynamic nature of the vegetation in boreal and eastern North America since the last glacial maximum, 21,000 years ago. Biomes, defined as mixtures of plant functional types, emphasize structural changes in the vegetation in response to climate change. A primary use of these data-based biome maps is to test simulations produced by atmospheric general circulation models (AGCM's) and vegetation models.; Biome maps for 6000 and 18,000 radiocarbon years B.P. were created as part of an international effort to assemble global paleovegetation maps for model testing. The biome reconstructions for 6000 years B.P. serve as a benchmark for a series of detailed evaluations of the biogeography models BIOME1 and BIOME3. A 2 x 2 factorial set of biome simulations was performed in which climate simulations for 6 rka from CCM1 and CCM3 each were used to run BIOME1 and BIOME3. Only the biome simulation produced by CCM1 and BIOME3 matched the data well. Sensitivity analyses revealed the poor performance of the CCM3/BIOME3 simulation primarily was caused by BIOME3 being overly sensitive to secondary influences such as cloud cover and soil characteristics. This sensitivity builds a greater uncertainty into the BIOME3 simulations, particularly because cloudiness may not be accurately simulated by AGCM's.; Biome distributions and individual pollen taxon abundances were mapped at 1000-year intervals to explore vegetation dynamics in eastern North America. Pollen assemblages without any close modern analog prevailed adjacent to the Laurentide Ice Sheet between 17,000 and 12,000 kcal. These “no-analog” pollen samples indicate that late-glacial plant associations were unlike any today. Distinguished by the intermixing of boreal taxa with temperate taxa and herbs, these no-analog communities were likely a mixed parkland. CCM1 simulations suggest that the no-analog vegetation grew under highly seasonal and dry climates. The temporal and spatial distribution of the no-analog vegetation strongly corresponds to the distribution of no-analog climatic conditions simulated by CCM1. The fastest rates of climatic and vegetation change occurred after the peak periods of no-analogs. Together, these lines of evidence support the hypothesis that vegetation is in equilibrium with climate at millennial timescales.