Confidence intervals and sample size for estimating the prevalence of plastic debris in seabird nests. (August 2020)

• Record Type: Journal Article
• Authors: Tavares, Davi Castro; Moura, Jailson Fulgêncio; Acevedo-Trejos, Esteban; Crawford, Robert J.M.; Makhado, Azwianewi; Lavers, Jennifer L.; Witteveen, Minke; Ryan, Peter G.; Merico, Agostino
• Is Part Of: Environmental pollution. Volume 263(2020)Supplement Part A
• Publication Date: 2020-08
• Subjects: Confidence interval -- Error estimation -- Nests -- Plastic -- Prevalence -- Sample size -- Sampling design; Pollution -- Periodicals; Pollution -- Environmental aspects -- Periodicals; Environmental Pollution -- Periodicals; Pollution -- Périodiques; Pollution -- Aspect de l'environnement -- Périodiques; Pollution -- Effets physiologiques -- Périodiques; Pollution; Pollution -- Environmental aspects; Periodicals; Electronic journals; 363.73
• DOI: 10.1016/j.envpol.2020.114394
• Languages: English
• ISSNs: 0269-7491

Abstract

Abstract: Evidence is accumulating about the impacts of plastics on marine life. The prevalence of plastics in seabird nests has been used as an indicator of levels of this pollutant in the ocean. However, the lack of a framework for defining sample sizes and errors associated with estimating the prevalence of plastic in nests prevents researchers from optimising time and reducing impacts of fieldwork. We present a method to determine the confidence intervals for the prevalence of debris in seabird nests and provide, for the first time, information on the prevalence of these items in nests of the Hartlaub's gull Larus hartlaubii, the African penguin Spheniscus demersus, the great white pelican Pelecanus onocrotalus, and the white-breasted cormorant Phalacrocorax lucidus in South Africa. The method, based on observations and resampling simulations and tested here for nests of 12 seabird species from 15 locations worldwide, allows for straightforward hypothesis testing. Appropriate sample sizes can be defined by combining this method with a Bayesian approach. We show that precise estimates of prevalence of debris in nests can be obtained by sampling around 250 nests. Smaller sample sizes can be useful for obtaining rough estimates. For the Hartlaub's gull, the African penguin, the great white pelican, and the white-breasted cormorant, debris were present in 0.75%, 3.00%, 6.41%, and 25.62% of the respective nests. Our approach will help researchers to determine errors associated with the prevalence of debris recorded in seabird nests and to optimise time and costs spent collecting data. It can also be applied to estimate confidence intervals and define sample sizes for assessing prevalence of plastic ingestion by any organism. Graphical abstract: Image 1 Highlights: A method to define errors and sample size for the prevalence of debris in nests. We use data from 12 seabird species from 15 locations worldwide. Sampling 250 nests provide precise estimates of prevalence of debris. Our method can be used also for data on prevalence of debris ingestion. The method is freely available as an R package. Abstract : A method for defining confidence intervals and sample size for estimating the prevalence of plastic debris in seabird nests.