Empirical data from a 6-month mesocosms experiment were used to assess the ability and performance of two machine learning (ML) models, including artificial neural network (NN) and random forest (RF), to predict temporal bioavailability changes of complex chemical mixtures in contaminated soils amended with compost or biochar. From the predicted bioavailability data, toxicity response for relevant ecological receptors was then forecasted to establish environmental risk implications and determine acceptable end-point remediation. The dataset corresponds to replicate samples collected over 180 days and analysed for total and bioavailable petroleum hydrocarbons and heavy metals/metalloids content. Further to this, a range of biological indicators including bacteria count, soil respiration, microbial community fingerprint, seeds germination, earthworm's lethality, and bioluminescent bacteria were evaluated to inform the environmental risk assessment. Parameters such as soil type, amendment (biochar and compost), initial concentration of individual compounds, and incubation time were used as inputs of the ML models. The relative importance of the input variables was also analysed to better understand the drivers of temporal changes in bioavailability and toxicity. It showed that toxicity changes can be driven by multiple factors (combined effects), which may not be accounted for in classical linear regression analysis (correlation). The use of ML models could improve our understanding of rate-limiting processes affecting the freely available fraction (bioavailable) of contaminants in soil, therefore contributing to mitigate potential risks and to inform appropriate response and recovery methods.

Integration of methods based on satellite remote sensing into current maritime monitoring strategies could help tackle the problem of global overfishing. Operational software is now available to perform vessel detection on satellite imagery, but research on vessel classification has mainly focused on bulk carriers, container ships, and oil tankers, using high-resolution commercial Synthetic Aperture Radar (SAR) imagery. Here, we present a method based on Random Forest (RF) to distinguish fishing and non-fishing vessels, and apply it to an area in the North Sea. The RF classifier takes as input the vessel’s length, longitude, and latitude, its distance to the nearest shore, and the time of the measurement (am or pm). The classifier is trained and tested on data from the Automatic Identification System (AIS). The overall classification accuracy is 91%, but the precision for the fishing class is only 58% because of specific regions in the study area where activities of fishing and non-fishing vessels overlap. We then apply the classifier to a collection of vessel detections obtained by applying the Search for Unidentified Maritime Objects (SUMO) vessel detector to the 2017 Sentinel-1 SAR images of the North Sea. The trend in our monthly fishing-vessel count agrees with data from Global Fishing Watch on fishing-vessel presence. These initial results suggest that our approach could help monitor intensification or reduction of fishing activity, which is critical in the context of the global overfishing problem.

A six-month laboratory scale study was carried out to investigate the effect of biochar and compost amendments on complex chemical mixtures of tar, heavy metals and metalloids in two genuine contaminated soils. An integrated approach, where organic and inorganic contaminants bioavailability and distribution changes, along with a range of microbiological indicators and ecotoxicological bioassays, was used to provide multiple lines of evidence to support the risk characterisation and assess the remediation end-point. Both compost and biochar amendment (p = 0.005) as well as incubation time (p = 0.001) significantly affected the total and bioavailable concentrations of the total petroleum hydrocarbons (TPH) in the two soils. Specifically, TPH concentration decreased by 46% and 30% in Soil 1 and Soil 2 amended with compost. These decreases were accompanied by a reduction of 78% (Soil 1) and 6% (Soil 2) of the bioavailable hydrocarbons and the most significant decrease was observed for the medium to long chain aliphatic compounds (EC_16-35) and medium molecular weight aromatic compounds (EC_16-21). Compost amendment enhanced the degradation of both the aliphatic and aromatic fractions in the two soils, while biochar contributed to lock the hydrocarbons in the contaminated soils. Neither compost nor biochar affected the distribution and behaviour of the heavy metals (HM) and metalloids in the different soil phases, suggesting that the co-presence of heavy metals and metalloids posed a low risk. Strong negative correlations were observed between the bioavailable hydrocarbon fractions and the ecotoxicological assays suggesting that when bioavailable concentrations decreased, the toxicity also decreased. This study showed that adopting a combined diagnostic approach can significantly help to identify optimal remediation strategies and contribute to change the over-conservative nature of the current risk assessments thus reducing the costs associated with remediation endpoint.