Temporal density of pollen sampling affects age determination of the mid-Holocene hemlock (Tsuga) decline

In establishing the age of a past event, emphasis has been placed on dating accuracy, precision, calibration, and age-model development, while the importance of adequate sampling and accurate reconstruction of the event itself is often neglected. Here we evaluate effects of temporal sampling density on stratigraphic and hence chronological placement of an event horizon, using the mid-Holocene hemlock decline as an example. Results from simulations on theoretical and actual high-resolution pollen sequences suggest that the pollen-sampling density has a critical impact on the stratigraphic placement of the hemlock-decline horizon, and in turn, on age estimation of the event. Also, decline-age frequency distributions from multiple sites were affected by sampling density. Therefore, the hemlock decline should be used with caution as a precise chronostratigraphic marker, and the current dataset of hemlock-decline records may not be sufficient to analyze the spatiotemporal pattern of the hemlock decline. Finally, we propose criteria for obtaining adequately sampled records in studying an event, and estimated the "adequate sampling density" for age estimation of the mid-Holocene hemlock decline. Published by Elsevier Ltd.