Biology subjects

Ergon, T.

Publications and source records attributed to Ergon, T..

3 recordsLinked to original sources

bioRxiv2020

Efficient Estimation of Large-Scale Spatial Capture-Recapture Models

Capture-recapture methods are a common tool in ecological statistics, which have been extended to spatial capture-recapture models for data accompanied by location information. However, standard formulations of these models can be unwieldy and computationally intractable for large spatial scales, many individuals, and/or activity center movement. We provide a cumulative series of methods that yield dramatic improvements in Markov chain Monte Carlo (MCMC) estimation for two examples. These include removing unnecessary computations, integrating out latent states, vectorizing declarations, and restricting calculations to the locality of individuals. Our approaches leverage the flexibility provided by the nimble R package. In our first example, we demonstrate an improvement in MCMC efficiency (the rate of generating effectively independent posterior samples) by a factor of 100. In our second example, we reduce the computing time required to generate 10,000 posterior samples from 4.5 hours down to five minutes, and realize an increase in MCMC efficiency by a factor of 25. We also explain how these approaches can be applied generally to other spatially-indexed hierarchical models. R code is provided for all examples, as well as an executable web-appendix.

ecology↗

bioRxiv2019

Estimating cause- and size- specific mortality hazard rates using mark-recapture-recovery data

O_LIEvidence-based management of natural populations under strong human influence frequently requires not only estimates of survival but also knowledge about how much mortality is due to anthropogenic versus natural causes. This is the case particularly when individuals vary in their vulnerability to different causes of mortality due to traits, life-history stages, or locations. C_LIO_LIHere, we estimated harvest and background (other cause) mortality of a landlocked migratory salmonid over half a century. In doing so, we quantified among-individual variation in vulnerability to cause-specific mortality resulting from differences in body size and spawning location relative to a hydropower dam. C_LIO_LIWe constructed a multistate mark-recapture model to estimate harvest and background mortality hazard rates as functions of a discrete state (spawning location) and an individual time-varying covariate (body size). We further accounted for among-year variation in mortality and migratory behavior and fit the model to a unique 50-year time-series of mark-recapture-recovery data on brown trout (Salmo trutta) in Norway. C_LIO_LIHarvest mortality was highest for intermediate-sized trout, and outweighed background mortality for most of the observed size range. Background mortality decreased with body size for trout spawning below the dam and increased for those spawning above. All vital rates varied substantially over time, but a trend was evident only in estimates of fishers reporting rate, which decreased from over 50% to less than 10% throughout the study period. C_LIO_LIWe highlight the importance of body size for cause-specific mortality and demonstrate how this can be estimated using a novel hazard rate parameterisation for mark-recapture models. Our approach allows estimating effects of individual traits and environment on cause-specific mortality without confounding, and provides an intuitive way to estimate temporal patterns within and correlation among different mortality sources. C_LI

ecology↗

bioRxiv2017

Study mortality with hazard rates, not probabilities

O_LIMortality is a key process in ecology and evolution, and much effort is spent on statistical and theoretical modelling of this process. Mortality hazard rates describe individuals instantaneous ability to survive at different times, whereas survival probabilities are defined for time-intervals of a given length. In this commentary, we argue that it is often more meaningful to model and interpret interval-specific time-averaged mortality hazard rates than survival probabilities, also when using discrete-time models.\nC_LIO_LIWe discuss three topics where the concept of hazard rates is essential for sound biological inference, but nevertheless often not used: (i) modelling of covariate effects on survival probabilities or mortality hazard rates, (ii) modelling of multiple sources of mortality and competing risks, and (iii) elasticity analyses of population growth rate to demographic parameters. To facilitate estimation of cause-specific mortality hazard rates, we provide amendments to R package marked.\nC_LIO_LIBy focusing on modelling mortality hazard rates, one avoids several inherent problems of comparing survival or mortality probabilities. In particular, interpretations about relative differences between mortality hazard rates (measurements on a ratio scale), or effects of relative changes in mortality hazard rates on population growth (elasticities), are more meaningful than similar interpretations involving survival (or mortality) probabilities or odds (measurements on an absolute scale).\nC_LIO_LIThe concept of hazard rates is essential for understanding ecological and evolutionary processes and we give an intuitive explanation for this, using several examples. We provide some practical guidelines and suggestions for further methods developments.\nC_LI

ecology↗