Baseline and power analyses for the assessment of beach litter reductions in the European OSPAR region

Marine litter pollution is a global environmental problem. Beach litter is a part of this problem, and is widely monitored in Europe. The European Marine Strategy Framework Directive (MSFD) requires a reduction of beach litter. A reduction of 30% has been proposed in the European Plastics Strategy....

Full description

Saved in:
Bibliographic Details
Published in:Environmental pollution (1987) 2019-05, Vol.248, p.555-564
Main Authors: Schulz, Marcus, Walvoort, Dennis J.J., Barry, Jon, Fleet, David M., van Loon, Willem M.G.M.
Format: Article
Language:English
Subjects:
Baseline value
Beach litter
litteR
Non-parametrical power analysis
Reduction target
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Marine litter pollution is a global environmental problem. Beach litter is a part of this problem, and is widely monitored in Europe. The European Marine Strategy Framework Directive (MSFD) requires a reduction of beach litter. A reduction of 30% has been proposed in the European Plastics Strategy. The aims of this study are to develop (a) a method to calculate sufficiently stable and precise baseline values for beach litter, and (b) to derive a method of power analysis to estimate the number of beach litter surveys, necessary to detect a given reduction, using these baseline values. Beach litter data from the OSPAR (Oslo Paris Convention) region were used, and tailor-made statistical methods were implemented in open source software, litteR. Descriptive statistics and Theil-Sen and Mann-Kendall trend analyses were calculated for the most abundant beach litter types, for 14 survey sites. The length of a baseline period necessary to obtain a specified precision of the mean baseline value, expressed as Coefficient of Variation (CV), was calculated. Power analyses were performed using Monte Carlo simulations combined with Wilcoxon tests to determine significant deviations of the simulated datasets from the baseline mean values. For most survey sites, the mean length of monitoring periods necessary to achieve a CV < 10% amounts to four to five years with four surveys a year. The mean number of surveys necessary to detect a statistically significant reduction of 30% with 80% power ranges from 14 to 20. Power analyses show that a reduction of 10% is difficult to detect, because more than 24 surveys are needed. In contrary, a reduction of 40–50% can be detected easily with a small (
ISSN:0269-7491
1873-6424
DOI:10.1016/j.envpol.2019.02.030