Effects of heterogeneous wind fields and vegetation composition on modeled estimates of pollen source area

Abstract:

Fossil pollen records from lakes, bogs, and small hollows offer the main source of information about vegetation responses to climate change and land use over timescales of decades to millennia. Millions of pollen grains are released from individual trees each year, and are transported by wind before settling out of the atmosphere. Reconstructing past vegetation from sedimentary pollen records, however, requires careful modeling of pollen production, transport, and deposition. The atmosphere is turbulent, and regional wind patterns shift from day to day. In accordance with this, it is necessary for pollen transport models to adequately account for variable, non-uniform wind patterns and vegetation heterogeneity.

Using a simulation approach, with both simulated vegetation patterns and vegetation gradients, as well as simulated wind fields, we show the inconsistency in pollen loading proportions and local vegetation proportions when non-uniform wind patterns are incorporated. Vegetation upwind from the lake is over-represented due to the increased prevalence of winds transporting pollen from that area. The inclusion of North American Regional Reanalysis (NARR) wind records affirms this finding. Of the lake sites explored in this study, none had uniform wind patterns.

The use of a settlement-era gridded vegetation dataset, compiled by the PalEON project and based on Public Land Survey System (PLSS) records allows us to model pollen source area with realistic vegetation heterogeneity. Due to differences in productivity, pollen fall speeds, and neighboring vegetation, there exist patterns of vegetation that may be poorly characterized due to over/under representation of different taxa. Better understanding these differences in representation allows for more accurate reconstruction of historical vegetation, and pollen-vegetation relationships.