Biology subjects

Wesner, J. S.

Publications and source records attributed to Wesner, J. S..

3 recordsLinked to original sources

bioRxiv2023

Detecting differences in Size Spectra

O_LIThe distribution of body size in communities is remarkably consistent across habitats and taxa and can be represented by size spectra, which are described by a power law. The focus of size spectra analysis is to estimate the exponent ({lambda}) of the power law. C_LIO_LIMany methods have been proposed for estimating{lambda} most of which involve binning the data, summing abundance within bins, and then fitting a ordinary least squares (OLS) regression in log-log space. However, recent work has shown that binning procedures may return biased estimates of size spectra exponents compared to procedures that directly estimate{lambda} using maximum likelihood estimation (MLE). Despite this variability in estimates, it is unclear if the relative change across environmental gradients is consistent across methodologies. Here, we used simulation to compare the ability of two binning methods (NAS, ELBn) and MLE to 1) recapture known values of{lambda} , and 2) recapture parameters in a linear regression measuring the change in{lambda} across a hypothetical environmental gradient. We also compared the methods using two previously published body size datasets across a pollution gradient and a temperature gradient C_LIO_LIMaximum likelihood methods always performed better than common binning methods, which demonstrated consistent bias depending on the simulated values of{lambda} . This bias carried over to the regressions, which were more accurate when{lambda} was estimated using MLE compared to the binning procedures. Additionally, the variance in estimates using MLE methods is markedly reduced when compared to binning methods. C_LIO_LIThe uncertainty and variation in estimates when using binning methods is often greater than or equal to the variation previously published in experimental and observational studies, bringing into question the effect size of previously published results. However, while the methods produced different slope estimates from previously published datasets, they were in qualitative agreement on the sign of those slopes (i.e., all negative or all positive). Our results provide further support for the direct estimation of{lambda} using MLE (or similar procedures) over the more common methods of binning. C_LI

ecology↗

bioRxiv2023

Bayesian hierarchical modeling of size spectra

O_LIA fundamental pattern in ecology is that smaller organisms are more abundant than larger organisms. This pattern is known as the individual size distribution (ISD), which is the frequency of all individual body sizes in an ecosystem. C_LIO_LIThe ISD is described by a power law and a major goal of size spectra analyses is to estimate the exponent of the power law, {lambda}. However, while numerous methods have been developed to do this, they have focused almost exclusively on estimating {lambda} from single samples. C_LIO_LIHere, we develop an extension of the truncated Pareto distribution within the probabilistic modeling language Stan. We use it to estimate multiple {lambda}s simultaneously in a hierarchical modeling approach. C_LIO_LIThe most important result is the ability to examine hypotheses related to size spectra, including the assessment of fixed and random effects, within a single Bayesian generalized (non)-linear mixed model. While the example here uses size spectra, the technique can also be generalized to any data that follows a power law distribution. C_LI

ecology↗

bioRxiv2020

Choosing priors in Bayesian ecological models by simulating from the prior predictive distribution

Bayesian data analysis is increasingly used in ecology, but prior specification remains focused on choosing non-informative priors (e.g., flat or vague priors). One barrier to choosing more informative priors is that priors must be specified on model parameters (e.g., intercepts, slopes, sigmas), but prior knowledge often exists on the level of the response variable. This is particularly true for common models in ecology, like generalized linear mixed models, which may have a link function and dozens of parameters, each of which needs a prior distribution. We suggest that this difficulty can be overcome by simulating from the prior predictive distribution and visualizing the results on the scale of the response variable. In doing so, some common choices for non-informative priors on parameters can easily be seen to produce biologically impossible values of response variables. Such implications of prior choices are difficult to foresee without visualization. We demonstrate a workflow for prior selection using simulation and visualization with two ecological examples (predator-prey body sizes and spider responses to food competition). This approach is not new, but its adoption by ecologists will help to better incorporate prior information in ecological models, thereby maximizing one of the benefits of Bayesian data analysis.

ecology↗