We produced a collection of environmental, geophysical, and other marine-related data for marine ecological models and ecological-niche models. It consists of 2132 raster data for 58 distinct parameters at regional and global scales in the ESRI-GRID ASCII format. Most data originally belonged to open data but resided on heterogeneous repositories with different formats and resolutions. Other data were specifically created for the collection. The collection includes 565 data with global scale range; 154 at 0.5° resolution and 411 at 0.1° resolution; 196 data with annual temporal aggregation over ~10 key years between 1950 and 2100; 369 data with monthly aggregation at 0.1° resolution from January 2017 to ~May 2021 continuously. Data were also cut out on 8 European marine regions. The collection also includes forecasts to different future scenarios such as the Representative Concentration Pathways 2.6 (63 data), 4.5 (162 data), and 8.5 (162 data), and the A2 scenario of the Intergovernmental Panel on Climate Change (180 data).
Advantages Provided by D4Science
1. Data workflow and Accessibility
D4Science offered a space not only to host the data but also to make the workflow, through which the raster files of the collection were obtained, available to scientific communities. The description of the process, the applications and scripts used, and the intermediate results obtained at each step were published on D4Science. The final results were described in Coro, G., Bove, P., & Kesner-Reyes, K. (2023). Global-scale parameters for ecological models. Scientific Data, 10(1), 7. www.nature.com/articles/s41597-022-01904-3
2. Advanced Analytical Tools and Reusability
The description of the workflow and the initial, intermediate, and final data linked to the experiment that generated the data collection allow for the repeatability of the experiment. D4Science also provided an RStudio environment with access to workspace data, where it was possible to execute the workflow scripts. This supported the reusability of the methods used to produce the collection.
3. Collaborative Environment
The collaborative environment offered by the D4Science infrastructure allowed us to work efficiently with our co-author from QQuatics Inc. in the Philippines. It was possible to quickly and effectively exchange and validate the data produced by our colleague. Since corrections were sometimes necessary, having a mean to share and exchange data rapidly greatly helped our work.
From a scientific point of view, our data collection was completed by correlation analyses, and extended with forecasts to 2050 and 2100 according to different RCP scenarios, and through Habitat Representativeness Score (HRS) assessments to measure habitat similarities. Researchers can utilize our data collection to develop models that predict the future evolution of marine ecosystems. These models are crucial for assessing the sustainability of current fishing practices and formulating strategies to mitigate overfishing and ensure the long-term health of marine ecosystems in contrast to climate change and growing anthropogenic pressures.
Our work served the scientific community and the researchers involved in the field of ecological modelling in Marine Science. It provided the basis for analysis through Artificial Neural Networks, Support Vector Machines, implementations of the Maximum Entropy model, Random Forests, Cluster Analysis and many other techniques normally used in this field.
The study was conducted according to the dictates of Open Science also thanks to the support of the D4Science infrastructure. During the development, the data were shared through to the D4Science workspace, through private and public links between the authors. D4Science was useful to make the raster files in the collection interoperable because harmonized and optimized for modelling. Overall, D4Science helped to produce a reproducible collection for all potential stakeholders.
