Estimating the size distribution of plastics ingested by animals

Plastic pollution is a significant environmental problem, impacting terrestrial, freshwater, and marine ecosystems. Numerous species across diverse taxa—including over 690 marine and 50 freshwater species—are documented to ingest plastic debris. The consequences can be severe, ranging from physical damage and toxicity from additives to physiological effects and disruption of energy and nutrient flow within ecosystems. Quantifying global primary plastic ingestion and its subsequent trophic transfer is critical for understanding the ecological impacts of plastic pollution. Current knowledge is limited by a lack of understanding of 'ingestibility'—the propensity of an animal to ingest a given plastic particle. While models exist to predict the spatial and temporal co-occurrence of plastics and animals, these often assume all plastics are equally ingestible, which is unrealistic. Several factors influence plastic ingestibility, including feeding behavior, prey size distribution (for predators), plastic color, degradation level, and the release of odorants. However, these factors are complex and difficult to incorporate into general predictions. Body size, on the other hand, is a readily available and simple metric that could potentially predict ingestibility across diverse species. Allometric studies demonstrate the utility of body size in predicting complex biological traits. This study leverages a large dataset of gut content analyses to establish an allometric relationship between animal body size and the maximum size of plastic that animal can ingest. This relationship will improve global plastic pollution risk assessments by incorporating the biological reality of plastic ingestibility.

The existing literature highlights the widespread problem of plastic ingestion by animals. Studies have documented the presence of plastic in the guts of various species, leading to concerns about the potential ecological and health consequences. Several papers have focused on the distribution and abundance of plastic pollution in different environments, including the oceans, rivers, and terrestrial ecosystems. However, the relationship between animal size and the size of plastic particles they ingest has not been adequately explored. Previous research alluded to the potential role of body size in influencing plastic-biota interactions, but this study aims to quantify this relationship and demonstrate its utility in ecological risk assessments. Studies on the impact of plastic debris on marine life, trophic transfer of microplastics, and the classification of plastic waste as hazardous material are reviewed, along with studies on the catchment-scale perspective of plastic pollution, and impacts of discarded plastic bags on marine assemblages.

This study conducted a systematic review of peer-reviewed literature published between 1900 and 2018 using Web of Science, searching for articles on plastic ingestion by organisms. The search string included keywords related to plastics and ingestion, excluding studies on human consumption and reviews. Articles were screened for relevance based on title and abstract, and then full-text screening was conducted. Inclusion criteria required that articles report on field-based studies with naturally occurring plastic concentrations and size distributions, excluding lab studies. Data were collected on the longest axis of ingested plastic and the body length of animal taxa or individuals. Where specific measurements were not available, weighted means of mid-ranges were calculated. Various types of length measurements were standardized as 'total body length', using judgment to classify reported measurements. Data on individual animals were summarized to the lowest possible taxonomic rank (usually species level) to avoid pseudoreplication and produce values closer to true population values. A mixture of methods were used to detect plastic ingestion, including whole body digestion and analysis of specific organs using chemical agents. Data were collected from both reported values and from images using ImageJ, measuring the longest straight axis of the plastic. Location data was also approximated using reported coordinates or site descriptions. The species' habitat (marine, brackish, freshwater) and depth range were obtained from FishBase and SeaLifeBase. Statistical analysis employed linear regression on log10-log10 scales in Microsoft Excel and R to model the allometric relationship between animal and plastic size. Model validation involved repeated random selection of 10% of the data for validation against the remaining 90% for parameterization. The root mean square of errors (RMSE) was used to evaluate predictive power. To illustrate the utility of the model, the study created risk maps of plastic ingestion by zooplankton. This involved combining ingestible plastic density data from Eriksen et al. (2014) with zooplankton density data from Strömberg et al. (2009) in ArcGIS, creating maps for both ingestible plastics and total plastics. The data collated are available on GitHub.

The analysis revealed a significant allometric relationship (log10-log10 linear regression; R² = 0.42, p = 4.7e-09) between animal body length and the maximum length of ingested plastic, with a rough 20:1 ratio. The dataset included over 2000 gut content analyses from animals ranging in size from 9 mm (dragonet fish larvae) to 10.34 m (humpback whale), representing three orders of magnitude. Data were predominantly from fish (75%), followed by mammals (9%), invertebrates (11%), and reptiles (5%). Species-level data were available for 91% of records. The study noted a tendency to underpredict plastic size for large animals and overpredict it for small animals. A validation analysis of the allometric relationship using a 10% validation dataset and 90% parameterisation dataset (repeated 1000 times) found that approximately 30.45% of observed values fell within 95% confidence intervals, showing reasonable explanatory power (RMSE=0.68, R²=0.38, p<0.001). The study also found a weak correlation between animal length and the size of the smallest ingested plastic fragment (R²=0.10, p=0.008), suggesting that a wide range of plastic sizes is ingested. The application of this allometric relationship to zooplankton communities revealed that areas of high risk of plastic ingestion include the East and South China Seas, Bay of Bengal, Black Sea, Mediterranean Sea, Sargasso Sea, and European coasts of the North Atlantic Ocean. Maps illustrating the risk of plastic ingestion by zooplankton revealed higher accuracy when considering ingestible plastic sizes compared to maps considering all plastics.

This study successfully established a simple yet powerful allometric relationship between animal body length and the maximum size of ingested plastic. This relationship explains over 40% of the variance, significantly advancing current global plastic pollution risk assessments which lack this biological component of ingestibility. The use of body size as a predictive metric is attractive due to its broad applicability. Future research should develop more sophisticated models that incorporate additional life history traits such as feeding mode, mouthpart morphology, ontogeny, and habitat preferences. The study highlights a lack of data from terrestrial ecosystems, which should be a priority for future research. The findings emphasize the importance of plastic size relative to animal size, revealing high-risk areas for plastic entry into aquatic food webs. The significant flux of plastic into this basal layer has considerable environmental and human health implications. The generality of this relationship provides a foundation for scaling up these estimates to entire food webs.

This research provides a novel allometric relationship linking animal body size to the maximum size of plastic ingested, improving global plastic pollution risk assessments. The simplicity and broad applicability of the model make it valuable for predicting plastic ingestion across diverse species and ecosystems. Future work should focus on integrating additional ecological information into the model, addressing data gaps in terrestrial systems, and refining estimates of plastic distribution at various water depths.

The study's reliance on existing literature might introduce biases due to variations in methodologies used across studies. The weak correlation between animal length and the smallest ingested plastic fragment highlights challenges in accurately detecting smaller plastic particles. The absence of data from terrestrial animals limits the generalizability of the allometric relationship to all ecosystems. Further limitations result from the reliance on existing global plastic distribution models, which are currently confined to the surface mixing zone of oceans, whereas some of the animals known to contain plastics in this study are found at depths far greater than this surface mixing zone.