Invasive species from the Suez Canal, also named "Lessepsian species", often have an ecological and financial impact on marine life, fisheries, human well-being and health in the Mediterranean Sea. Among these, the silver-cheeked toadfish Lagocephalus sceleratus (Gmelin, 1789) has rapidly colonised the eastern Mediterranean basin and is moving westwards. This pufferfish is highly opportunistic; it attacks fish captured in nets and lines and seriously damages fishing gear and catches. It is a highly toxic species with no immediate economic value for the Mediterranean market, although it currently represents 4% of the weight of the total artisanal catches. Consequently, the possible effects on Mediterranean fisheries and health require enhancing the understanding of the future geographical distribution of this pufferfish in the whole basin.
An overall habitat suitability map and a geographical spread map for L. sceleratus at the Mediterranean scale were produced using the D4Science cloud computing facilities to merge seven machine learning models. The potential impact of the species was estimated across the Mediterranean countries.
The results suggest that, without an intervention, L. sceleratus will continue its rapid spread and likely have a high impact on fisheries. The presented method is generic and can be re-applied to other invasive species. It is based on an Open Science approach conducted through D4Science, and all processes are freely available as Web services through a dedicated Virtual Research Environment of the e-Infrastructure.

Advantages Provided by D4Science
1. Advanced Analytical Tools
D4Science offers analytical tools that enable users to perform complex data analyses and visualisations. In particular, general machine learning models (Artificial Neural Networks, Support Vector Machines, Maximum Entropy) were used to process big data of environmental variables over one decade and in 2050.
2. Collaborative Environment
The platform fosters collaboration among scientists, researchers, and policymakers by providing a shared workspace. Users can share data, discuss findings, and develop joint strategies for fisheries management, enhancing the collective knowledge and efforts towards sustainability. D4Science was used as a collaboration tool for the research team, which was distributed over three countries (Italy, Spain, and UK). The "Alien and Invasive Species" Virtual Research Environment and the D4Science Workspace were used to share data, exchange opinions, update the team about the status of the analysis, and share and comment on the results.
3. Scalability and Reliability
The D4Science robust infrastructure ensures high availability and scalability, accommodating the growing needs of the global fisheries community. The platform's reliability guarantees uninterrupted access to critical data and tools, which is essential for timely and effective decision-making.
For this work, FAIR machine learning data and models provided by other infrastructures were used. Moreover, the entire experiment was conducted following an Open Science-oriented approach. Specifically, the experiment was published as a Workflow, described by a sequence of links to open services (processes) with standardised interfaces (Web Processing Service). This choice enabled the reproducibility and repeatability of each step in the experiment, with the results harmoniously published in the D4Science infrastructure catalogue aligned with the FAIR paradigm. The Open Science approach was also fundamental in convincing FAO decision-makers to include the results in their official guidelines for Mediterranean countries, as the results and methodology were entirely transparent. The described research was the first attempt to quantitatively and spatiotemporally estimate the spread of the silver-cheeked toadfish in the Mediterranean Sea. The results included the future geographical reachability distribution of the fish and an estimation of the invasion time and spread. A risk estimate of the invasion was visually represented as the density of intense invasion locations in different subdivisions of the Mediterranean. A general westward decreasing impact pattern was highlighted, with high-risk zones predicted in the middle and south of the Mediterranean Sea (Sicily, Malta, and Tunisia). The overall depicted scenario was that L. sceleratus would soon be a great risk to fisheries and, consequently, to the health security of many Mediterranean countries. The estimated distribution foresees the invasion by the pufferfish of the Bosporus, which could enable it to spread into the Black Sea. This perspective should foster strategies such as selective fishing to decrease its population in the most likely future-impact areas. The produced maps can support and have supported these strategies and helped fisheries researchers advise managers and decision-makers.
The article describes an innovative application of a machine learning ensemble model to study the environmental characteristics favourable to the spread of L. sceleratus in the Mediterranean and to predict its future impact (density of colonised areas) on Mediterranean nations. After four levels of audit, the results were selected as a virtuous example of Open Science application and presented at the inaugural European Open Science Cloud event in Vienna (November 2019) to the Austrian Presidency and the European Commission. The results were also reported in an article in the prestigious "New Scientist" magazine, which was dedicated to the silver-cheeked toadfish. Thanks to the reproducible approach adopted through D4Science, the generated maps were included in the FAO's official advisories for Mediterranean nations. Over the years, the pufferfish spread maps have become a reliable reference for the countries at risk (especially Turkey), and the related scientific paper has collected 80 citations in 5 years. For more information, we invite you to read: Coro, G., Vilas, L. G., Magliozzi, C., Ellenbroek, A., Scarponi, P., & Pagano, P. (2018). Forecasting the ongoing invasion of Lagocephalus sceleratus in the Mediterranean Sea. Ecological Modelling, 371, 37-49. https://doi.org/10.1016/j.ecolmodel.2018.01.007