Distribuição e caracterização de microplásticos no sedimento da represa de abastecimento de água alagados

Abstract

The contamination assessment and management in water supply reservoirs is essential under a sanitary aspect, once their presence is associated with environmental impacts and can also compromise human health. In this scenario, microplastics perform a major role, since the microscopic dimensions ease their spread and turn their detection, quantification, and removal into a difficult task. In this manner, microplastic characterization in freshwater environments constitutes a necessity, by retrieving data and diagnosing an outlook of this modern pollution. It also contributes to highlighting the problem, calling attention to the scientific community to the development of an intertwined effort toward effective solutions. This research presents the results of the distribution and characterization of microplastic contamination in the Alagados Reservoir, in Ponta Grossa, south of Brazil, which is responsible for a part of the city’s water supply. This is extremely important for involving not only the protection of the environment but also the public health. It evaluates the temporal and spatial variation of the particles and their correlation to environmental parameters, such as water transparency, depth, and temperature. The sampling was conducted over twelve months, from June 2023 to April 2024, with samples collected bi-monthly from six different sites. Microplastics were found in each of the 108 samples, totaling 3,459 particles. The concentration ranged from 40 to 810 particles.kg-1

, a mean of 320.28 ± 162.83 particles.kg-1 of wet sediment. The predominant shapes were filaments (43.25%) and fibers (34.58%). 18.65% of microplastics were colored, 645 in total. It was performed a two-way variance analysis (ANOVA), Tukey’s honestly significant difference (HSD) test, and the Person test. This way, it was possible to investigate variations of the microplastic concentrations associated with the month and site of each sample, as well as relate the microplastic concentrations to the environmental parameters. Therefore, microplastic particles were identified in all collected samples, exhibiting a diverse range of characteristics. This indicates a widespread presence of microplastics, which poses an environmental problem for the ecosystem. Filaments and fibers were the primary shapes and most of the microplastics observed were uncolored. When microplastics were found with fading color, it could indicate pigment release into the environment, water, and sediment. There was a statistical correlation between the concentration of microplastics in sediment and the time, water clarity, and water temperature; however none between microplastic concentration and water depth. This research also presents a deep learning model for the microplastic and algae automated classification, developed in Python using the Google Colab environment. 1140 images were used, 80% for training and 20% for validation. The model had an 80% accuracy, reaching a recall of 93% and 92% for algae and microplastic filaments, respectively, despite still improvement in other classes and aspects. Overall, the accuracy of the model is encouraging considering the dataset size and all the challenges that involve the automatic identification of microplastics, with all their shape variations and nuances, but the results are promising. This research underscores the importance of enhanced data collection and analysis of microplastics in freshwater and riverbed sediments, especially within water supply reservoirs. Although microplastic contamination in the Alagados Reservoir is relatively low compared to similar locations, further studies and consistent monitoring remain crucial.