Quantitative reconstructions of vegetation abundance from sediment-derived pollen systems provide unique insights into past ecological conditions. Recently, the use of pollen accumulation rates (PAR, grains cm−2 year−1) has shown promise as a bioproxy for plant abundance. However, successfully reconstructing region-specific vegetation dynamics using PAR requires that accurate assessments of pollen deposition processes be quantitatively linked to spatially-explicit measures of plant abundance. Our study addressed these methodological challenges. Modern PAR and vegetation data were obtained from seven lakes in the western Klamath Mountains, California. To determine how to best calibrate our PAR-biomass model, we first calculated the spatial area of vegetation where vegetation composition and patterning is recorded by changes in the pollen signal using two metrics. These metrics were an assemblage-level relevant source area of pollen (aRSAP) derived from extended R-value analysis ( sensu Sugita, 1993) and a taxon-specific relevant source area of pollen (tRSAP) derived from PAR regression ( sensu Jackson, 1990). To the best of our knowledge, aRSAP and tRSAP have not been directly compared. We found that the tRSAP estimated a smaller area for some taxa (e.g. a circular area with a 225 m radius for Pinus) than the aRSAP (a circular area with a 625 m radius). We fit linear models to relate PAR values from modern lake sediments with empirical, distance-weighted estimates of aboveground live biomass (AGLdw) for both the aRSAP and tRSAP distances. In both cases, we found that the PARs of major tree taxa – Pseudotsuga, Pinus, Notholithocarpus, and TCT (Taxodiaceae, Cupressaceae, and Taxaceae families) – were statistically significant and reasonably precise estimators of contemporary AGLdw. However, predictions weighted by the distance defined by aRSAP tended to be more precise. The relative root-mean squared error for the aRSAP biomass estimates was 9% compared to 12% for tRSAP. Our results demonstrate that calibrated PAR-biomass relationships provide a robust method to infer changes in past plant biomass.
Computation of Circular Area and Spherical Volume Invariants via Boundary Integrals
