Aim

Quantitative linkage between surface pollen relative abundance (RA) and source-plant distribution is fundamental for fossil-based vegetation and climate reconstructions. However, how phytogeographical patterns are translated into sedimentary pollen records under the extreme high-altitude environments of the Tibetan Plateau (TP) remains a critical yet poorly understood issue. Here we integrate 3352 unique surface pollen spectra with MaxEnt-modelled plant distribution ranges of 36 common taxa to quantify their spatial coupling across the TP and surroundings.

Location

Tibetan Plateau and surroundings.

Taxon

36 common plant taxa (including 16 tree taxa and 20 herbaceous/shrub taxa).

Method

We integrated 3352 surface pollen samples with MaxEnt-modelled plant distribution ranges to statistically evaluate the spatial representativeness and pollen thresholds. Furthermore, we applied the newly established tree pollen thresholds to fossil pollen records from the Lunpola Basin to refine its palaeoelevation estimates for the latest Oligocene-earliest Miocene interval.

Results

Quantitative correlation analysis shows that for 16 tree taxa, including Pinus (threshold of 18%), Quercus (4.2%), Abies (1.9%), and Picea (2.6%), RA inside plant distribution ranges is significantly higher than outside, enabling robust presence/absence discrimination. In contrast, 20 herbaceous or shrub taxa (e.g., Artemisia, Cyperaceae, Ephedra) display overlapping RA distributions regardless of plant occurrence; their pollen is therefore unreliable as a local vegetation proxy. The tree thresholds are generally lower than earlier estimates derived from forest-cover proxies, highlighting the importance of using occurrence-based vegetation data and extra-regional pollen dispersal control on the TP. Moreover, the relationship between plant distribution and the sedimentary record is taxon-specific: while tree taxa exhibit robust spatial coupling, the geographic boundaries of herbs are blurred in the pollen record. We applied a tree pollen threshold to refine the fossil pollen-based altitude reconstruction, resulting in a palaeoelevation estimate of 3690 ± 200 m for the latest Oligocene-earliest Miocene interval in the Lunpola Basin, with this estimate being closer to the reconstructed palaeoelevation via carbonate oxygen isotopes or other evidence.

Main Conclusions

This study established spatial coupling relationships for 36 plant taxa on the Tibetan Plateau using plant occurrence and surface pollen data. Results show that 16 tree taxa (e.g., Pinus, Quercus) have clear pollen relative abundance thresholds enabling reliable inference of local plant presence, while 20 herbaceous and shrub taxa lack such thresholds due to complex environmental influences, and it demonstrates that reliable reconstruction of past vegetation and associated elevation/climate changes should focus on the well-calibrated arboreal taxa while cautiously interpreting herb/shrub signals.