Session Chairs: Marcus Eriksen, 5 Gyres Institute; Win Cowger, UC Riverside
This session looks at merging the many data-sets of plastic debris on land and on the sea as a monumental challenge. In doing so, the data-sets allow for richer questions about source, sink, distribution and trends. This session will appeal to an audience interested in citizen science, as well as modelers of distribution patterns of debris on land and sea.
In this session will bring together marine debris practitioners to discuss their data collection process and lessons learned to improve comparability and utility on regional and international scales. We will focus on new advances and data synchronization. Here we aim to bring together the many data sets on land and sea that document plastic pollution. Globally the public, private, nonprofit and education sectors overlap in their efforts to document observations of plastic pollution, from nano to macro, in marine environments and on land, using varied methods.
At sea we use nets, filtered pumps, and sea surface or seafloor visual observations. On land, surveys of beaches, roads, streams, as well as waste characterizations of litter conducted by municipalities, create rich data sets. There are many organizations, universities and government agencies, like NOAA, The Great Nurdle Hunt, SEA Education, Living Lands and Waters, SCCWRP, EOA, and Keep America Beautiful that have developed litter surveys. The Ocean Conservancy, Litterati, and Marine Debris Tracker have created mobile apps to engage the public in data collection. Many environmental NGOs, like Surfrider, Coastkeeper and 5 Gyres, have created their own data sets. The sum of these efforts is tremendous data, but there are challenges in content and comparability. How do we put it all together and make sense of it? To answer, a group from the Alfred-Wegener-Institute, created LITTERBASE, a pool of academic research on plastic pollution.
We looked on land in 2016 to create the Better Alternatives Now List, which combined all land litter data from cities, mobile apps and beach cleanups for California only. In 2017 we are looking at the same for the entire U.S. These data show greater spatial resolution when lumped together, despite challenges to find common methods of categorizing types of products and packaging.
In the sea we combined all known historical data about sea surface floating plastic with our most recent data from our 2013-2017 Travel Trawl citizen science program, reported in multiyear increments from 1980 to present. With greater consistency in data reporting (particle count (#/km2) and (weight kg/km2) we were able to produce a new global budget for floating debris.
These two datasets are useful to show trends in debris accumulation, the natural systems that remove plastic from the sea, and the efficiency of mitigations on land. They also point to the need to continue engaging citizens in meaningful data collection that’s comparable to other efforts. Future monitoring and mitigation require this engagement and consistency, but also a transparent discussion about how these data are utilized.
PANEL (following presentations):
- – Abigail Barrows, Adventure Scientists
- – Win Cowger, University of California, Riverside
- – Chris Wilcox, Commonwealth Scientific and Industrial Research Organization (CSIRO)
- – Marcus Eriksen, 5 Gyres Institute
Anthropogenic microparticle distribution in global marine surface waters: results of an extensive citizen science study
Authors: Abigail Barrows (Adventure Scientists, United States), Sara Cathey (Adventure Scientists), Chris Petersen (College of the Atlantic)
Plastic is a major pollutant throughout the world. It is one of the most prolific materials manufactured globally, with over 322 million tons produced annually. Plastics are cheap, light-weight, and durable—characteristics that have made it an ever-more attractive packaging material and led to its high volume in solid waste streams. Microplastic (plastic particles less than 5 mm in size) residence time and movement along the coast and sea surface outside of the gyres is still not well researched. This five-year project utilized global citizen scientists to collect 1,628 1-liter surface grab samples in every major ocean. Across all ocean basins, open ocean samples (further than 12 nm from land) contained a higher particle average (17.9 L-1) than coastal samples (5.9 L-1). Particles were predominantly microfibers (91%), 100 µm- 1.5 mm in length (77%), smaller than what has been captured in the majority of surface studies. Through µFT-IR we determined material type of 113 pieces, with 59% classified as synthetic and 41% as non-synthetic. Non-synthetic microfibers may pose a new and mostly unconsidered negative environmental and biological impact. Samples came from understudied ocean regions, some of which are emerging as areas of concentrated floating plastic and anthropogenic debris, influenced by distant waste mismanagement and/or airborne particles. Incorporation of smaller-sized microfibers in oceanographic models, which has previously been lacking, will help us to better understand potential fate and transformation of synthetic and non-synthetic microparticles in the marine environment.
Merging Big Data
Authors: Win Cowger (University of California, Riverside, United States)
Data quality, structure, and hygiene vary greatly in marine debris datasets. Quality ranges from uninformed data collectors with a high amount of random variation, to academic where measurements are made precisely with peer review. Structure can be the sum of everything observed in one unit or a fine grained classification and sampling scheme with multiple data dimensions such as weight, volume, number, shape, color, manufacturer, and size. Intradata hygiene typically has a well thought out scheme. However, in the case of data from Marine Debris Tracker and Clean Swell, intradata merging is complex because they use multiple different types of forms. Interdata hygiene is frequently complex, data is being collected with different devices, in different dimensions or units, is preconditioned with highly specialized models or assumptions, and prepared in complex tabular schemes. Simply adding an “Other” column can make or break a successful data merge. The most difficult thing about merging marine debris data is that a lot of data is not publically available. Sometimes organizations decide not to share it for proprietary reasons, institutions keep it until they publish (this can take years), and occasionally even after publishing individuals decide they will only share their data if they are added as a coauthor on the paper it is used in. The questions we want to answer in marine debris require big data, the only way to move forward is through collaboration to create higher quality data, an emphasis toward more data classifications, homogenization of data sheets and types, and open data.
Analyzing large scale marine debris monitoring data – challeges, solutions, and patterns
Authors: Chris Wilcox (CSIRO Oceans and Atmosphere Business Unit, Australia), Qamar Schuyler (CSIRO Oceans and Atmosphere Business Unit), Denise Hardesty (CSIRO Oceans and Atmosphere Business Unit)
The study of plastic pollution on land and in the oceans has stimulated a significant amount of monitoring effort, from at-sea trawls to thousands of volunteer cleanups. Researchers are literally at sea in a wash of data. However, many of these data were collected for reasons other than marine debris monitoring. After global efforts to analyze more than 10,000 trawl records, and large scale analysis of data from tens of thousands of surveys and cleanups in the US and Australia, it is clear that there are a number of major issues with the data we have at had to monitor marine debris. Using this data we have been able to identify sources, estimate trends, quantify distributions, and suggest solutions. However, there are a number of fundamental uncertainties that challenge the use of much of the marine debris data that is available. We will address some of the key issues in citizen science and other data types, and conclude with suggestions for how those interested in monitoring plastic pollution might modify their methods to increase their rigor and usefulness.
Big data as a source of policy to address plastic marine pollution
Authors: Marcus Eriksen (5 Gyres, United States), Jamie Rhodes (Upstream Policy), Matt Prindiville (Upstream Policy), Jeff Kirschner (Litterati)
After many years of surveying the subtropical gyres to document the distribution of plastics of all sizes, we concede that the greatest mitigations are upstream, where the plastic is identifiable, especially to brand. But what of ocean plastic data, trends, characterization of what’s there? Is it meaningful toward productive policy? Can we apply the same “Big Data” waste characterization to trash on land? In collaboration with other NGOs, we looked for the most common types of plastic waste in the United States during 2016 to create the Better Alternatives Now list (BAN list). We will use the BAN list as a baseline for measuring the effectiveness of future mitigation efforts, ranging from public awareness campaigns, to brand-engagement initiatives. In taking on this tactic, can big data be an effective tool for change?
Litterati is a mobile application that allows anyone to identify, map, and collect the world’s litter. With a community that has grown worldwide, and collected over 1M data points, we are able to provide the public, corporations, NGO and governmental organizations with useful data to locate litter hotspots, identify brands and products, and understand seasonal trends. With improvements in image recognition software, machine and deep learning, and an ever-growing Litterati community, we are poised to tackle the global litter pandemic.
Big data sets provide ample evidence to support policy-driven campaigns to mitigate the problem illuminated by that data. It is essential that big data be validated, and preferably peer-reviewed. Our recent example of the BAN list outlines the top 20 product categories found as trash on the ground across the United States. By assigning brand data to these items, we are able to inform source-reduction campaigns targeted at those companies themselves.