A model of oxygen isotope fractionation in body water of large mammals

A model is proposed for oxygen isotope fractionation in body water of terrestrial, herbivorous mammals larger than 1 kg. The goal of this model is to estimate the oxygen isotopic composition (δ 18O) of intake water in order to reconstruct paleoclimate from the δ 18O of fossil biogenic phosphate. The principal oxygen inputs are liquid water, atmospheric O2, and oxygen in food. The principal outputs are water (liquid and vapor) and CO2. Body mass-dependent scaling equations are used to assign O2, H2O, and CO2 fluxes. The model predicts that the δ 18O of body water is higher than the δ 18O of intake water and approaches the δ18O of intake water with increasing body size, as observed in empirical data. This reflects the increasing importance of liquid water flux relative to atmospheric O2, CO2, and water vapor flux at larger size (i.e., water flux increases relatively faster than metabolic rate and surface area with increasing body size). These results suggest that the largest fossil taxa should be used for paleoclimate reconstruction because (1) potential errors are smallest at large body sizes and (2) drinking water forms a larger proportion of the oxygen intake. Paleoclimate reconstruction based on the δ 18O of biogenic phosphates can thus be corrected for body-mass fractionation effects, a significant cause of previously uncharacterized interspecific variation.