Uncertainty in biological monitoring: a framework for data collection and analysis to account for multiple sources of sampling bias

Abstract

Summary Biological monitoring programmes are increasingly relying upon large volumes of citizen‐science data to improve the scope and spatial coverage of information, challenging the scientific community to develop design and model‐based approaches to improve inference. Recent statistical models in ecology have been developed to accommodate false‐negative errors, although current work points to false‐positive errors as equally important sources of bias. This is of particular concern for the success of any monitoring programme given that rates as small as 3% could lead to the overestimation of the occurrence of rare events by as much as 50%, and even small false‐positive rates can severely bias estimates of occurrence dynamics. We present an integrated, computationally efficient Bayesian hierarchical model to correct for false‐positive and false‐negative errors in detection/non‐detection data. Our model combines independent, auxiliary data sources with field observations to improve the estimation of false‐positive rates, when a subset of field observations cannot be validated a posteriori or assumed as perfect. We evaluated the performance of the model across a range of occurrence rates, false‐positive and false‐negative errors, and quantity of auxiliary data. The model performed well under all simulated scenarios, and we were able to identify critical auxiliary data characteristics which resulted in improved inference. We applied our false‐positive model to a large‐scale, citizen‐science monitoring programme for anurans in the north‐eastern United States, using auxiliary data from an experiment designed to estimate false‐positive error rates. Not correcting for false‐positive rates resulted in biased estimates of occupancy in 4 of the 10 anuran species we analysed, leading to an overestimation of the average number of occupied survey routes by as much as 70%. The framework we present for data collection and analysis is able to efficiently provide reliable inference for occurrence patterns using data from a citizen‐science monitoring programme. However, our approach is applicable to data generated by any type of research and monitoring programme, independent of skill level or scale, when effort is placed on obtaining auxiliary information on false‐positive rates.

Habitat selection in transformed landscapes and the role of forest remnants and shade coffee in the conservation of resident birds

“…While the ideal way to measure the value of a habitat would be to estimate survival and reproduction parameters based on long‐term demographic and/or movement studies of species showing a representative range of ecological and behavioural traits (Milder et al, ; Sekercioglu, Loarie, Brenes, Ehrlich, & Daily, ), these are rarely feasible in tropical ecosystems. Our results show that occupancy and abundance models generated from the wide‐scale application of simple sampling methodologies (Irizarry et al, ; Ruiz‐Gutiérrez, Hooten, & Campbell Grant, ) may neglect important differences in habitat preference and performance. Capture data allow for the exploration of habitat–species relationships at a resolution intermediate between these two extremes (Ruiz‐Gutiérrez et al, ; Sekercioglu, ), but to increase the reliability of our inference, we need further research into the relationships between individual‐level data and consequences at the population level.…”

The influence of drought on flow‐ecology relationships in Ozark Highland streams

“…Environment–ecology relationships in this study showed substantial differences between years (drought versus flood) and across taxonomic groups. Using biological data to inform policies and models can improve the likelihood of achieving desired results (King et al., ; Poff et al., ), yet there is a great deal of variation inherent in the collection of such data (Hurlbert, ; Ruiz‐Gutierrez, Hooten, & Grant, ). While long‐term monitoring studies designed to assess year‐to‐year temporal variation at the same site would be the ideal way to better understand environmental flow relationships (MacDonald & Cote, ), limitations of time and funding make this an impractical option.…”

Quantifying data quality in a citizen science monitoring program: False negatives, false positives and occupancy trends

“…They found that simultaneously estimating both false-negative and false-positive error rates is computationally challenging, as any set of detection histories can be equally well explained by multiple sets of parameter values (Guillera-Arroita, Lahoz-Monfort, van Rooyen, Weeks, & Tingley, 2017). Other approaches have made use of calibration experiments to experimentally infer false-positive error rates under controlled conditions and use these to inform analysis of survey data (Chambert et al, 2015;Guillera-Arroita et al, 2017;Lahoz-Monfort, Guillera-Arroita, & Tingley, 2016;Ruiz-Gutierrez, Hooten, & Grant, 2016). Subsequent developments have developed alternative solutions to this identifiability issue by utilizing extra information to inform the detection parameters.…”

“…Royle and Link (2006) addressed this issue by forcing a constraint upon the model that the false-positive error rate must be lower than the true detection rate. These approaches have been successfully applied, however performing calibration experiments to inform false-positive error rates in survey data is often impractical (though see McClintock, Bailey, Pollock, and Semons (2010) and Ruiz-Gutierrez et al (2016) for a successful application), and there may be situations where secondary datasets are not available. Typically this involves jointly analyzing the dataset of interest alongside a second, independent dataset at which a subset of sites are monitored using secondary detection methods in which the probability of false-positive observations is considered impossible (Chambert, Miller, & Nichols, 2015;Miller et al, 2011Miller et al, , 2013.…”