In conventional pollen analysis, usually one sediment core per basin is analyzed to reconstruct past environmental conditions. This approach does not consider spatial heterogeneity of pollen assemblages, and assumes that one analyzed location is representative of the whole basin. To improve the spatial resolution of fossil pollen studies, further knowledge of the factors influencing variations in pollen assemblages throughout a basin is needed. We examined the spatial heterogeneity of pollen assemblages from 45 lacustrine surface samples from a lake with relatively simple hydrology and compared this dense network of surface pollen samples with the Lithuanian State Forest Service arboreal vegetation database. Calculations of pollen productivity at different locations across the lake revealed variations in the behavior of a pollen-vegetation relationship model in different parts of the basin. Our findings suggest that the model underestimated pollen contributions from the lakeshore vegetation. We demonstrate that detailed investigations of surface pollen as a step prior to fossil pollen investigations can provide useful insights, including understanding the influence of sedimentation rate on modelling results and spatial variations in pollen composition, thus providing guidance for site selection for fossil pollen studies.