Abstract / Summary
Lake-surface pollen assemblage is the ideal modern analogue for past climate and vegetation reconstructions quantitatively based on lacustrine fossil pollen spectra. Soil-surface pollen assemblage, on the other hand, has been more widely applied due to its ready accessibility, yet its application is limited by strong local vegetation signals. To overcome this limitation, in this study, we performed spatial homogenization at various scales (radii: 2, 5, 10, 20, 30, 40, 50, 60, 75, 100, 150, 200 km) to 100 soil-surface pollen assemblages and compared the homogenized datasets with the corresponding 100 lake-surface pollen assemblages covering the Tibetan Plateau, to access the potential of spatial-homogenized soil-surface pollen data to serve as a substitute for lake-surface pollen assemblage. We found that soil-surface pollen data spatially homogenized at scales of 40–60 km show the highest similarity to corresponding lake-surface pollen assemblages, indicating the pollen source area for the regular lake likely ranges between 40 and 60 km on the Tibetan Plateau. In summary, spatial homogenization reduces the influence of local plants on soil-surface pollen data, supporting its suitability as a potential substitute for lake-surface pollen assemblages.