A novel method for improved estimates of absolute microfossil abundance: A big step towards a deep-time terrestrial productivity proxy

A holy grail of both palaeoecology and biogeochemistry has been an accurate proxy of past biological productivity. Such a metric would offer a way to identify and quantify Earth’s deep-time ecosystem and carbon cycle function (and dysfunction). Plants have been the principal contributors to the terrestrial carbon cycle for hundreds of millions of years. We hypothesise that their absolute abundances in the fossil record can indicate ecosystem-mediated changes in carbon sequestration rates (='terrestrial net ecosystem productivity').

Many key parameters of biological systems—e.g., productivity, population sizes, biomass—are best expressed as absolute values. Unlike proportional data (e.g., percentages), absolute values provide standardized metrics for comparing the functioning of organisms, species and ecosystems across time and space. Since it is generally impractical to count entire populations, statistically significant abundance estimates require an accurate and precise sampling method. These typically entail more data collection effort (or time) than proportional data.

Firstly, we present a new method for precise estimates of microfossil concentrations: the ‘field-of-view subsampling’ (FOVS) method. It applies ecological quadrat sampling principles to microfossil samples spiked with exotic markers (e.g., Lycopodium spores). We tested the new FOVS method against the traditional ‘linear method’ with two case studies: 1, computer simulations; and 2, observational data of terrestrial organic microfossils from the end-Permian event (EPE; c. 252 Ma) records of eastern Australia. Four output parameters were measured: 1, absolute abundance (measured as specimens per unit sample size [e.g., sediment mass]); 2, accuracy (measured as variance from an idealised data set); 3, precision (measured as statistical error); and 4, data collection effort (measured as time). The FOVS method consistently provided estimates with greater accuracy, and higher precision and/or reduced effort under almost all conditions.

Secondly, we assessed the potential application of this method (and others) for gauging palaeoproductivity. As a result of this review, we: 1, identified the factors that influence the preservation of land-derived organic carbon in the fossil record; 2, adapted and applied a framework of modern ecosystem productivity to prehistoric settings by incorporating post-burial impacts; and 3, explored the conditions under which terrestrial organic microfossil concentrations may provide valid estimates of relative changes in palaeoproductivity.

Lastly, we demonstrate how refined estimates of deep-time terrestrial productivity may be achieved in the future. This would lead to more precise land carbon cycle models since the emergence of large land plants >360 million years ago.

Although we have explored a narrow application of the new method to palaeoproductivity, the range of potential applications is far broader. In the microfossil realm, the method can be immediately applied to any study using exotic markers (e.g., Lycopodium spores) for absolute abundances. Given its demonstrable increased efficiency, we recommend the FOVS method as the new standard for such absolute abundance estimates.