From Bacteria to Whales: Using Functional Size Spectra to Model Marine Ecosystems

Size-based ecosystem models have proliferated in the past 10 years. They are a general and powerful approach to modeling ecosystem structure and function Great progress has been made toward modeling ecosystems from bacteria to whales. Unifying models across scales and confronting models with data are now the key needs Size-based ecosystem modeling is emerging as a powerful way to assess ecosystem-level impacts of human- and environment-driven changes from individual-level processes. These models have evolved as mechanistic explanations for observed regular patterns of abundance across the marine size spectrum hypothesized to hold from bacteria to whales. Fifty years since the first size spectrum measurements, we ask how far have we come? Although recent modeling studies capture an impressive range of sizes, complexity, and real-world applications, ecosystem coverage is still only partial. We describe how this can be overcome by unifying functional traits with size spectra (which we call functional size spectra) and highlight the key knowledge gaps that need to be filled to model ecosystems from bacteria to whales. Size-based ecosystem modeling is emerging as a powerful way to assess ecosystem-level impacts of human- and environment-driven changes from individual-level processes. These models have evolved as mechanistic explanations for observed regular patterns of abundance across the marine size spectrum hypothesized to hold from bacteria to whales. Fifty years since the first size spectrum measurements, we ask how far have we come? Although recent modeling studies capture an impressive range of sizes, complexity, and real-world applications, ecosystem coverage is still only partial. We describe how this can be overcome by unifying functional traits with size spectra (which we call functional size spectra) and highlight the key knowledge gaps that need to be filled to model ecosystems from bacteria to whales. graphs of relative abundance or biomass across trophic levels in ecological communities [1]. a measurable property of an individual organism, sometimes aggregated to the species-level, that influences its ecological role or performance. continuous first-order partial differential equation that models changes in abundance at size or age through time. Originally implemented for age-structured dynamics, it was later applied to the size spectrum and subsequently adapted to model fluxes of food-dependent growth and mortality by representing size-based predator and prey interactions through the use of a distribution of preferred predator-prey mass ratios. ratio of predator to prey mass measured at the individual level. At the community level, PPMR is the average mass of predators at trophic level n divided by the average mass of their prey at trophic level n–1. Observed PPMR from dietary data is referred to as realized. In contrast, models are parameterized with preferred PPMRs, which are used in combination with prey size availability and predator search rates to predict realized diets. the size distribution of all individuals in a community or ecosystem according to numerical abundance (abundance size spectrum) or biomass (biomass size spectrum) (y axis) across size classes (x axis) typically on log axes. A normalized size spectrum converts the biomasses or abundances to densities by dividing them by the width of the size classes. deterministic continuous time models that explicitly predict changes in the size spectrum through time, starting with size-based, individual-level mechanistic processes. Different levels of complexity can be represented in size spectrum models. Community SSMs only consider differences in size, ignoring differences between species. The community size spectrum can be derived from the trait-based and food-web models by integrating over all trait classes or summing over species. Trait-based SSMs represent differences among species through the strict use of functional traits (e.g., offspring, maturation, and asymptotic sizes) rather than species groups. the slope of a straight line fitted through the size spectrum on a log–log plot, which is also the exponent from a power law fit to the size spectrum on linear axes. A slope of b = 0 of the biomass size spectrum conforms to Sheldon’s [11] original conjecture that biomass is distributed equally across logarithmically binned size classes which equates to a slope of b−1 for the biomass density size spectrum (normalized by dividing by size class widths). Similarly, the abundance size spectrum slope is b−1 and b−2 when it is an abundance density (by dividing by size class widths).