Multiple phases of mg-calcite in crustose coralline algae suggest caution for temperature proxy and ocean acidification assessment: lessons from the ultrastructure and biomineralization in Phymatolithon (Rhodophyta, Corallinales)1.

Magnesium content, strongly correlated with temperature, has been developed as a climate archive for the late Holocene without considering anatomical controls on Mg content. In this paper, we explore the ultrastructure and cellular scale Mg-content variations within four species of North Atlantic crust-forming Phymatolithon. The cell wall has radial grains of Mg-calcite, whereas the interfilament (middle lamella) has grains aligned parallel to the filament axis. The proportion of interfilament and cell wall carbonate varies by tissue and species. Three distinct primary phases of Mg-calcite were identified: interfilament Mg-calcite (mean 8.9 mol% MgCO3 ), perithallial cell walls Mg-calcite (mean 13.4 mol% MgCO3 ), and hypothallium Mg-calcite (mean 17.1 mol% MgCO3 ). Magnesium content for the bulk crust, an average of all phases present, showed a strongly correlated (R2  = 0.975) increase of 0.31 mol% MgCO3 per °C. Of concern for climate reconstructions is the potential for false warming signals from undetected postgrazing wound repair carbonate that is substantially enriched in Mg, unrelated to temperature. Within a single crust, Mg-content of component carbonates ranged from 8 to 20 mol% MgCO3 , representing theoretical thermodynamic stabilities from aragonite-equivalent to unstable higher-Mg-calcite. It is unlikely that existing current predictions of ocean acidification impact on coralline algae, based on saturation states calculated using average Mg contents, provide an environmentally relevant estimate.