Curriculum Vitae (including links to publications) | Software and Open Source Code | Blog

 

Current Research (National Research Council) 

Chris Darimont

Stable Isotope Analysis Over the last two years, my colleagues and I have been developing and refining Bayesian methods for stable isotope mixing model analyses. In a 2008 issue of Ecology Letters we first described a novel Bayesian framework (MixSIR) for mixing model anlaysis that incorporates uncertainty in the isotopic signatures of both predators (mixtures) and prey (sources). In response to a technical comment, we published a followup article in Ecology Letters at the beginning of 2009. Most recently, we published a hierarchical version of the mixing model formulation that incorporates individual (and group-level) diet variability among predators, and demonstrates how to use information theoretic approaches to evaluate data support for alternate mixing model parameterizations. This paper is freely available through the PLoS One Website.

Time- and Space-Series Analysis My colleagues and I at the NWFSC continue to work on time series (population trajectories) and space series (animal movement paths) analysis methods for corrupted data sets. State space analysis methods offer a robust framework for analyzing  series data with both process error (i.e. error associated with variability in biological processes such as birth and death) and non-process error (i.e. error associated with imperfect observations, spawner-recruit estimation methods, etc.). Below are some of the recent projects I have been involved in related to state-space methods. Hierarchical state-space model of juvenile Chinook salmon movement -- Using high-resolution acoustic tracking equipment in Willapa Bay, WA, I monitored the movement of ~20 juvenile Chinook salmon over estuarine habitat, and subsequently related animal behavior to habitat using a hierarchical state-space model of movement. Findings from the study (published in Canadian Journal of Fisheries and Aquatic Science) provided strong support for the hypothesized importance of native eelgrass in smolt survival. DARTER (Diffusion AppRoximation Tools for Extinction Risk)  -- When I first started working at NWFSC I developed DARTER, an open source visual basic tool in Excel that carries out Bayesian state-space model fitting of time-series data in order to create robust extinction risk metrics. The tool makes use of a massive bank of Kalman filters within a Sampling Improtance Resampling (SIR) framework in order to develop posteriors for error terms and population trends. This method is novel, and outperforms standard maximum likelihood methods for fitting state space models given short time-series' (<~20 time steps). Bayesian Multivariate Autoregressive Models -- First-order multivariate autoregressive (MAR1) models  are used to identify likely interactive effects in communities using both multiple population time-series and environmental covariates. I am currently extending MAR1 techniques into a fully Bayesian framework, an approach that has been discussed in theory but has not yet been put into practice. Similarly, my colleagues and I are extending MAR1 models in order to investigate hypotheses about how communities are spatially structured (for instance, data support for levels of metapopulation connectivity). Finally, because MAR1 framework allows for the estimation of 1) population trends through time, and 2) the influence of interactions on such trends, trophically-mediated population viability analyses (PVA) are a natural extension of the state-space and MAR1 methods I have been working on. My colleagues and I intend to add PVA metrics to existing MAR1 modeling tools published by the NWFSC (e.g. LAMBDA software).

Reef Environmental Education Foundation

Grouper Moon Project  For the last 5 years I have been the principle investigator of Grouper Moon, a multi-institution conservation project aimed at describing the spatial, demographic and numerical impact of spawning-site marine protected areas (MPAs) on endangered Nassau grouper in the Cayman Islands. This project involves (1) monitoring space use through acoustic telemetry, (2) characterizing demographic responses to MPAs through Leslie matrix analytic techniques such as sensitivity analyses, and (3) characterizing the spatial aspects of fishing pressure through mark-recapture analysis. This research project, funded by the Pew Charitable Trusts, was explicitly designed to carry out a comprehensive assessment of the risks Nassau grouper face, and by implication the potential benefits provided by MPAs in the Cayman Islands and elsewhere. My colleagues and I have published aspects of this work in the Journal of Fish Biology (Semmens et al. 2006); we are currently developing several additional manuscripts for peer reviewed publication and a technical report for the Cayman Islands Marine Conservation Board.