摘要：Microfossil records from ice archives allow vegetation, fire and land-use activity reconstructions on broad spatial scales. Samples typically contain low microfossil concentrations. Therefore, large ice volumes are often needed for palynology. Hence, it is crucial to extract maximum microfossil numbers through appropriate physical-chemical treatments. We compare six methods covering the main water reduction procedures: evaporation, filtration and centrifugation with snow samples. Adding a known number of Lycopodium marker spores prior to sample treatment and a second marker ( Eucalyptus ) after laboratory processing allows a quantitative microfossil loss assessment during pollen extraction. We applied the best-performing method (average loss of 22%) to high-alpine firn cores from Colle Gnifetti glacier for validation with a natural archive containing extremely low microfossil concentrations. We conclude that samples processed with different microfossil extraction protocols may give different results for pollen concentrations, percentages and ratios between different pollen types, especially if vesiculate conifer pollen is an important pollen assemblage component. We recommend a new evaporation-based method which delivers the smallest and least variable losses among the tested approaches. Since microfossil losses are inevitable during laboratory procedure, adding markers prior to sample processing is mandatory to achieve reliable microfossil concentration and influx estimates.
