A better understanding of the quantitative relationships between pollen assemblages from lakes and the composition and structure of the vegetation in the surrounding landscape will allow a more detailed interpretation of late Holocene pollen diagrams, and thus hopefully lead to a better understanding of the history of the cultural landscape. Some studies have used modern pollen vegetation calibration data sets, but in Denmark extensive plantation, drainage, fertilisation and use of pesticides have greatly altered the vegetation during the last century. Historical maps are available from around AD 1800 that show areas of different land cover such as arable fields, forest, meadow and heath. The AD 1800 landscape was more comparable to the older ones reflected by the fossil pollen diagrams, so a method using historical analogues was used.

The aim was to improve the quantitative interpretation of fossil pollen diagrams from small lakes by estimating the size of the pollen source area and by testing methods of pollen-landscape calibration. A dataset of pollen assemblages from ca. AD 1800 from 30 small (3-30 ha.) Danish lakes, and landcover data from digitised historical maps surrounding the lakes has been collected and analysed using the Extended R-Value model and the Prentice-Sugita model of pollen dispersal, which are implemented in the POLLSCAPE simulation model.

The size of the Relevant Source Area of Pollen (RSAP) for the lakes is estimated. Wind speed and pollen dispersal characteristics are shown to have little effect on the RSAP, but there is a difference between the RSAP in eastern and western Denmark. Simulations using POLLSCAPE shows that this can be explained by differences in the patch size of the vegetation in the two regions.

Pollen assemblages simulated by the POLLSCAPE model using the historical landcover data and pollen productivity estimates from neighbouring countries are compared to the observed pollen assemblages, in an attempt to validate the model. This is complicated by uncertainties about the species composition of the different landcover types around AD 1800, which is the main disadvantage of using historical analogues in stead of a modern dataset. However, the POLLSCAPE model can reproduce the main trends in the observed data.