A re-evaluation of the palaeoclimatic significance of phosphorus variability in speleothems revealed by high-resolution synchrotron micro XRF mapping

Abstract. Phosphorus (P) is potentially a very important environmental proxy in speleothem palaeoclimate reconstructions. However, the transfer of P to a speleothem seems to vary between cave sites. Therefore, it is important to investigate the source of P and the way it is incorporated into a speleothem on a site-by-site basis before it can be used as a robust palaeoclimate proxy.

In this paper, the distribution of P in one modern and two Early Pliocene speleothems formed in coastal caves on Christmas Island (Indian Ocean) and the Nullarbor Plain (southern Australia) is investigated using microscopy and ultra-high resolution chemical mapping.

Phosphorus has been found to be both incorporated in the lattice and present as diverse P-rich phases. Monitoring data from Christmas Island suggest that co-precipitation of P-rich phases occurs when "prior calcite precipitation" decreases following recharge, even if the drip rate decreases. Microbial mediation may also play a role, which complicates a direct climate relationship between P and hydrology. We find that some P-enriched layers contain dissolution features, with possible involvement of microbial mats which colonise pores during reduced drip rates associated with prolonged dry spells.

In the two Early Pliocene speleothems the relationship between P and microbial laminae is clearer. Both petrographic and chemical data suggest that phosphorus-rich phases in the microbial laminae mark intervals of reduced drip rates, which may indicate dry intervals during the otherwise wet palaeoclimate of the Early Pliocene.

We develop a speleothem distribution coefficient for phosphorus (SK_P) rather than the thermodynamic partition coefficient (K_P) to account for the presence of crystalline phosphate inclusions. SK_P describes P enrichment in speleothems regardless of the process, as similar mechanisms of phosphate co-precipitation may be in operation in biotic and abiotic conditions.

The most important implication of our study is that variability in P concentration may be related to diverse processes which can be recognized through petrographic observations and chemical mapping. In particular, there may not be a direct relation between an increase in P concentration and seasonal infiltration as has been found in some previous studies, especially if the source of this element is not the labile phosphate released through leaching during seasonal vegetation dieback in temperate climates.