Dataset Information

Identifying species at coextinction risk when detection is imperfect: Model evaluation and case study.

ABSTRACT: Losing a species from a community can cause further extinctions, a process also known as coextinction. The risk of being extirpated with an interaction partner is commonly inferred from a species' host-breadth, derived from observing interactions between species. But observational data suffers from imperfect detection, making coextinction estimates highly unreliable. To address this issue and to account for data uncertainty, we fit a hierarchical N-mixture model to individual-level interaction data from a mutualistic network. We predict (1) with how many interaction partners each species interacts (to indicate their coextinction risk) and (2) how completely the community was sampled. We fit the model to simulated interactions to investigate how variation in sampling effort, interaction probability, and animal abundances influence model accuracy and apply it to an empirical dataset of flowering plants and their insect visitors. The model performed well in predicting the number of interaction partners for scenarios with high abundances, but indicated high parameter uncertainty for networks with many rare species. Yet, model predictions were generally closer to the true value than the observations. Our mutualistic plant-insect community most closely resembled the scenario of high interaction rates with low abundances. Median estimates of interaction partners were frequently much higher than the empirical data indicate, but uncertainty was high. Our analysis suggested that we only detected 14-59% of the flower-visiting insect species, indicating that our study design, which is common for pollinator studies, was inadequate to detect many species. Imperfect detection strongly affects the inferences from observed interaction networks and hence, host specificity, specialisation estimates and network metrics from observational data may be highly misleading for assessing a species' coextinction risks. Our study shows how models can help to estimate coextinction risk, but also indicates the need for better data (i.e., intensified sampling and individual-level observations) to reduce uncertainty.

SUBMITTER: Plein M

PROVIDER: S-EPMC5573280 | biostudies-literature | 2017

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Identifying species at coextinction risk when detection is imperfect: Model evaluation and case study.

Plein Michaela M Morris William K WK Moir Melinda L ML Vesk Peter A PA

PloS one 20170828 8

Losing a species from a community can cause further extinctions, a process also known as coextinction. The risk of being extirpated with an interaction partner is commonly inferred from a species' host-breadth, derived from observing interactions between species. But observational data suffers from imperfect detection, making coextinction estimates highly unreliable. To address this issue and to account for data uncertainty, we fit a hierarchical N-mixture model to individual-level interaction d ...[more]

PMID: 28846734

Similar Datasets

Project description:Understanding species distributions and the factors limiting them is an important topic in ecology and conservation, including in nature reserve selection and predicting climate change impacts. While Species Distribution Models (SDM) are the main tool used for these purposes, choosing the best SDM algorithm is not straightforward as these are plentiful and can be applied in many different ways. SDM are used mainly to gain insight in 1) overall species distributions, 2) their past-present-future probability of occurrence and/or 3) to understand their ecological niche limits (also referred to as ecological niche modelling). The fact that these three aims may require different models and outputs is, however, rarely considered and has not been evaluated consistently. Here we use data from a systematically sampled set of species occurrences to specifically test the performance of Species Distribution Models across several commonly used algorithms. Species range in distribution patterns from rare to common and from local to widespread. We compare overall model fit (representing species distribution), the accuracy of the predictions at multiple spatial scales, and the consistency in selection of environmental correlations all across multiple modelling runs. As expected, the choice of modelling algorithm determines model outcome. However, model quality depends not only on the algorithm, but also on the measure of model fit used and the scale at which it is used. Although model fit was higher for the consensus approach and Maxent, Maxent and GAM models were more consistent in estimating local occurrence, while RF and GBM showed higher consistency in environmental variables selection. Model outcomes diverged more for narrowly distributed species than for widespread species. We suggest that matching study aims with modelling approach is essential in Species Distribution Models, and provide suggestions how to do this for different modelling aims and species' data characteristics (i.e. sample size, spatial distribution).

Project description:Risk factors are key epidemiological concepts that are used to explain disease distributions. Identifying disease risk factors is generally done by comparing the characteristics of diseased and non-diseased populations. However, imperfect disease detectability generates disease observations that do not necessarily represent accurately the true disease situation. In this study, we conducted an extensive simulation exercise to emphasize the impact of imperfect disease detection on the outcomes of logistic models when case reports are aggregated at a larger scale (e.g., diseased animals aggregated at farm level). We used a probabilistic framework to simulate both the disease distribution in herds and imperfect detectability of the infected animals in these herds. These simulations show that, under logistic models, true herd-level risk factors are generally correctly identified but their associated odds ratio are heavily underestimated as soon as the sensitivity of the detection is less than one. If the detectability of infected animals is not only imperfect but also heterogeneous between herds, the variables associated with the detection heterogeneity are likely to be incorrectly identified as risk factors. This probability of type I error increases with increasing heterogeneity of the detectability, and with decreasing sensitivity. Finally, the simulations highlighted that, when count data is available (e.g., number of infected animals in herds), they should not be reduced to a presence/absence dataset at the herd level (e.g., presence or not of at least one infected animal) but rather modeled directly using zero-inflated count models which are shown to be much less sensitive to imperfect detectability issues. In light of these simulations, we revisited the analysis of the French bovine abortion surveillance data, which has already been shown to be characterized by imperfect and heterogeneous abortion detectability. As expected, we found substantial differences between the quantitative outputs of the logistic model and those of the zero-inflated Poisson model. We conclude by strongly recommending that efforts should be made to account for, or at the very least discuss, imperfect disease detectability when assessing associations between putative risk factors and observed disease distributions, and advocate the use of zero-inflated count models if count data is available.

Project description:BackgroundImperfect diagnostic testing reduces the power to detect significant predictors in classical cross-sectional studies. Assuming that the misclassification in diagnosis is random this can be dealt with by increasing the sample size of a study. However, the effects of imperfect tests in longitudinal data analyses are not as straightforward to anticipate, especially if the outcome of the test influences behaviour. The aim of this paper is to investigate the impact of imperfect test sensitivity on the determination of predictor variables in a longitudinal study.Methodology/principal findingsTo deal with imperfect test sensitivity affecting the response variable, we transformed the observed response variable into a set of possible temporal patterns of true disease status, whose prior probability was a function of the test sensitivity. We fitted a Bayesian discrete time survival model using an MCMC algorithm that treats the true response patterns as unknown parameters in the model. We applied our approach to epidemiological data of bovine tuberculosis outbreaks in England and investigated the effect of reduced test sensitivity in the determination of risk factors for the disease. We found that reduced test sensitivity led to changes to the collection of risk factors associated with the probability of an outbreak that were chosen in the 'best' model and to an increase in the uncertainty surrounding the parameter estimates for a model with a fixed set of risk factors that were associated with the response variable.Conclusions/significanceWe propose a novel algorithm to fit discrete survival models for longitudinal data where values of the response variable are uncertain. When analysing longitudinal data, uncertainty surrounding the response variable will affect the significance of the predictors and should therefore be accounted for either at the design stage by increasing the sample size or at the post analysis stage by conducting appropriate sensitivity analyses.

Project description:Environmental risk assessment is a key process for the authorization of pesticides, and is subjected to continuous challenges and updates. Current approaches are based on standard scenarios and independent substance-crop assessments. This arrangement does not address the complexity of agricultural ecosystems with mammals feeding on different crops. This work presents a simplified model for regulatory use addressing landscape variability, co-exposure to several pesticides, and predicting the effect on population abundance. The focus is on terrestrial vertebrates and the aim is the identification of the key risk drivers impacting on mid-term population dynamics. The model is parameterized for EU assessments according to the European Food Safety Authority (EFSA) Guidance Document, but can be adapted to other regulatory schemes. The conceptual approach includes two modules: (a) the species population dynamics, and (b) the population impact of pesticide exposure. Population dynamics is modelled through daily survival and seasonal reproductions rates; which are modified in case of pesticide exposure. All variables, parameters, and functions can be modified. The model has been calibrated with ecological data for wild rabbits and brown hares and tested for two herbicides, glyphosate and bromoxynil, using validated toxicity data extracted from EFSA assessments. Results demonstrate that the information available for a regulatory assessment, according to current EU information requirements, is sufficient for predicting the impact and possible consequences at population dynamic levels. The model confirms that agroecological parameters play a key role when assessing the effect of pesticide exposure on population abundance. The integration of laboratory toxicity studies with this simplified landscape model allows for the identification of conditions leading to population vulnerability or resilience. An Annex includes a detailed assessment of the model characteristics according to the EFSA scheme on Good Modelling Practice.

Dataset Information

Identifying species at coextinction risk when detection is imperfect: Model evaluation and case study.

Publications

Identifying species at coextinction risk when detection is imperfect: Model evaluation and case study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets