Synthesizing multiple data types for biological conservation using integrated population models

Biological Conservation, Journal Year: 2017, Volume and Issue: 217, P. 240 - 250

Published: Nov. 20, 2017

Language: Английский

---

Quantitative evidence for the effects of multiple drivers on continental-scale amphibian declines

Scientific Reports, Journal Year: 2016, Volume and Issue: 6(1)

Published: May 23, 2016

Since amphibian declines were first proposed as a global phenomenon over quarter century ago, the conservation community has made little progress in halting or reversing these trends. The early search for "smoking gun" was replaced with expectation that are caused by multiple drivers. While field observations and experiments have identified factors leading to increased local extinction risk, evidence effects of drivers is lacking at large spatial scales. Here, we use 389 time-series 83 species complexes from 61 study areas across North America test 4 major hypothesized declines. find populations being lost metapopulations an average rate 3.79% per year, not related any particular threat continental scale; likewise effect each stressor variable regional This result - exposure threats varies spatially, vary their response provides generality development strategies. Greater emphasis on solutions this globally shared needed.

Language: Английский

---

Generalized joint attribute modeling for biodiversity analysis: median‐zero, multivariate, multifarious data

Ecological Monographs, Journal Year: 2016, Volume and Issue: 87(1), P. 34 - 56

Published: Nov. 16, 2016

Abstract Probabilistic forecasts of species distribution and abundance require models that accommodate the range ecological data, including a joint multiple based on combinations continuous discrete observations, mostly zeros. We develop generalized attribute model ( GJAM ), probabilistic framework readily applies to data are presence‐absence, ordinal, continuous, discrete, composition, zero‐inflated, censored. It does so as over all providing inference sensitivity input variables, correlations between scale, prediction, analysis, definition community structure, missing imputation. applications illustrate flexibility species‐abundance data. Applications forest inventories demonstrate relationships responding environmental variables. shows environment can be inverse predicted from species. Application microbiome demonstrates how prediction in accelerates variable selection, by isolating effects each variable's influence across

Language: Английский

---

Faster estimation of Bayesian models in ecology using Hamiltonian Monte Carlo

Methods in Ecology and Evolution, Journal Year: 2016, Volume and Issue: 8(3), P. 339 - 348

Published: Oct. 14, 2016

Summary Bayesian inference is a powerful tool to better understand ecological processes across varied subfields in ecology, and often implemented generic flexible software packages such as the widely used BUGS family (BUGS, WinBUGS, OpenBUGS JAGS). However, some models have prohibitively long run times when BUGS. A relatively new platform called Stan uses Hamiltonian Monte Carlo (HMC), of Markov chain (MCMC) algorithms which promise improved efficiency faster relative those by gaining traction many fields an alternative BUGS, but adoption has been slow likely due part complex nature HMC. Here, we provide intuitive illustration principles HMC on set simple models. We then compared using population ecology that vary size complexity. For hierarchical models, also investigated effect parameterization random effects, known non‐centering. small, there little practical difference between two platforms, outperforms model complexity grows. performs well for more sensitive than may be robust biased caused pathologies, because it produces diagnostic warnings where provides none. Disadvantages include inability use discrete parameters, diagnostics greater requirement hands‐on tuning. Given these results, valuable ecologists utilizing inference, particularly problems slow. As such, can extend boundaries feasible applied problems, leading understanding processes. Fields would benefit estimation individual growth rates, meta‐analyses cross‐system comparisons spatiotemporal

Language: Английский

---

The priority of prediction in ecological understanding

Oikos, Journal Year: 2016, Volume and Issue: 126(1), P. 1 - 7

Published: Sept. 1, 2016

The objective of science is to understand the natural world; we argue that prediction only way demonstrate scientific understanding, implying should be a fundamental aspect all disciplines. Reproducibility an essential requirement good and arises from ability develop models make accurate predictions on new data. Ecology, however, with few exceptions, has abandoned as central focus faces its own crisis reproducibility. Models are where ecological understanding stored they source – no possible without model world. can improved in three ways: variables, functional relationships among dependent independent parameter estimates. Ecologists rarely test assess whether have made advances by identifying important elucidating relationships, or improving Without these tests it difficult know if more today than did yesterday. A commitment ecology would lead to, other things, mature (i.e. quantitative) hypotheses, prioritization modeling techniques appropriate for (e.g. using continuous variables rather categorical) and, ultimately, advancement towards general Synthesis therefore understanding. Here address how this inhibited progress explore renewed benefit ecologists. lack emphasis resulted discipline qualitative, imprecise hypotheses little concern results generalizable beyond when data were collected. allow ecologists critical questions about generalizability our making

Language: Английский

---

Guidance for decisions using the Vector Autoregressive Spatio-Temporal (VAST) package in stock, ecosystem, habitat and climate assessments

Fisheries Research, Journal Year: 2018, Volume and Issue: 210, P. 143 - 161

Published: Oct. 26, 2018

Language: Английский

---

A multispecies occupancy model for two or more interacting species

Methods in Ecology and Evolution, Journal Year: 2016, Volume and Issue: 7(10), P. 1164 - 1173

Published: June 29, 2016

Summary Species occurrence is influenced by environmental conditions and the presence of other species. Current approaches for multispecies occupancy modelling are practically limited to two interacting species often require assumption asymmetric interactions. We propose a model that can accommodate or more generalize single‐species assuming latent state multivariate Bernoulli random variable. probability each potential with both multinomial logit probit present details Gibbs sampler latter. As an example, we co‐occurrence probabilities bobcat ( Lynx rufus ), coyote Canis latrans grey fox Urocyon cinereoargenteus ) red Vulpes vulpes as function human disturbance variables throughout 6 Mid‐Atlantic states in eastern United States. found evidence pairwise interactions among most species, some pairs occupying same site varied along gradients; were independent at sites little disturbance, but these likely occur together high disturbance. Ecological communities composed multiple Our proposed method improves our ability draw inference from such permitting detection/non‐detection data arbitrary number without Additionally, permits variables. These advancements represent important improvement community‐level subject imperfect detection.

Language: Английский

---

A guide to Bayesian model checking for ecologists

Ecological Monographs, Journal Year: 2018, Volume and Issue: 88(4), P. 526 - 542

Published: May 15, 2018

Abstract Checking that models adequately represent data is an essential component of applied statistical inference. Ecologists increasingly use hierarchical Bayesian in their research. The appeal this modeling paradigm undeniable, as researchers can build and fit embody complex ecological processes while simultaneously accounting for observation error. However, ecologists tend to be less focused on checking model assumptions assessing potential lack when applying methods than more traditional modes inference such maximum likelihood. There are also multiple ways the models, each which has strengths weaknesses. For instance, P values relatively easy compute, but well known conservative, producing biased toward 0.5. Alternatively, lesser approaches checking, prior predictive checks, cross‐validation probability integral transforms, pivot discrepancy measures may produce accurate characterizations goodness‐of‐fit not ecologists. In addition, a suite visual targeted diagnostics used examine violations different at levels hierarchy, check residual temporal or spatial autocorrelation. review, we synthesize existing literature guide through many available options checking. We illustrate procedures with several case studies including (1) analysis simulated spatiotemporal count data, (2) N‐mixture estimating abundance sea otters from aircraft, (3) hidden Markov describe attendance patterns California lion mothers rookery. find commonly based posterior detect extreme inadequacy, often do subtle cases fit. Tests (including “sampled value”) appear better suited have overall performance. conclude necessary ensure scientific founded. As discovery, it should accompany most analyses presented literature.

Language: Английский

---

Spatial autoregressive models for statistical inference from ecological data

Ecological Monographs, Journal Year: 2017, Volume and Issue: 88(1), P. 36 - 59

Published: Nov. 13, 2017

Abstract Ecological data often exhibit spatial pattern, which can be modeled as autocorrelation. Conditional autoregressive (CAR) and simultaneous (SAR) models are network‐based (also known graphical models) specifically designed to model spatially autocorrelated based on neighborhood relationships. We identify discuss six different types of practical ecological inference using CAR SAR models, including: (1) selection, (2) regression, (3) estimation autocorrelation, (4) other connectivity parameters, (5) prediction, (6) smoothing. compare showing their development connection partial correlations. Special cases, such the intrinsic (IAR), described. depend weight matrices, whose uses definition row‐standardization. Weight matrices also include covariates structures, we emphasize, but have been rarely used. Trends in harbor seals ( Phoca vitulina ) southeastern Alaska from 463 polygons, some with missing data, used illustrate types. develop a variety regression fit maximum likelihood Bayesian methods. Profile graphs for covariance parameters. The same set is both prediction smoothing, relative merits each discussed. show nonstationary variances correlations demonstrate effect several take‐home messages including choosing between IAR modeling effects matrix, appeal how handle isolated neighbors. highlight reasons why ecologists will want make use directly hierarchical not only explicit settings, more general models.

Language: Английский

---

Collapse of a desert bird community over the past century driven by climate change

Proceedings of the National Academy of Sciences, Journal Year: 2018, Volume and Issue: 115(34), P. 8597 - 8602

Published: Aug. 6, 2018

Significance Deserts, already defined by climatic extremes, have warmed and dried more than other regions in the contiguous United States due to climate change. Our resurveys of sites originally visited early 20th century found Mojave Desert birds strongly declined occupancy lost nearly half their species. Declines were associated with change, particularly decreased precipitation. The magnitude decline avian community absence species that local climatological “winners” are exceptional. results provide evidence bird communities collapsed a new, lower baseline. could accelerate future as this region is predicted become drier hotter end century.

Language: Английский

---