PARENT SESSION
Oral Session #10: Paleoecology. Presiding: S. Hotchkiss.
Monday, August 6, 2001. 8:00 AM to 12:00 PM. Hall of Ideas J.
The effect of sample size on paleoenvironmental reconstructions using the modern analog technique.
Wahl, Eugene1,2, Lytle, David 1, 1 2
ABSTRACT- The use of modern analog techniques (MAT) to reconstruct paleoenvironmental conditions has become widespread in paleoecology. Sample sizes for these methods traditionally are based on the probability of finding rare types. However, there is no a priori reason to apply these rules when using MAT. Using Monte Carlo simulations derived from fossil pollen assemblages, we created pollen counts of 75 to 1000 grains and investigated the relationship between count size and the reliability of vegetation and climate reconstructions using MAT. Results show: 1) simulated counts of ≥150 grains are very similar to the fossil samples, as measured by squared chord distance; and 2) selection of analogs from a modern reference set is biased to false negatives (not selecting analogs chosen by the fossil samples), rather than to false positives (selecting analogs not chosen by the fossil samples). 95% of simulated counts of 200 grains fall within +0.6 and -0.3 ° C of January and July temperatures reconstructed using the fossil samples, an error range twice that of simulated counts of 1000 grains. Reconstruction quality of annual precipitation is similar, with an error range of ±35 mm at 200 grains. Reconstruction of vegetation is more variable. Pollen counts of 150 grains are sufficient to accurately reconstruct (≥95% success) vegetation types that are well-differentiated. When reconstructing vegetation types with many taxa in common, count sizes of 500 may be necessary.
KEY WORDS: modern analog technique, climate reconstruction, vegetation reconstruction, pollen count size