WP4 aims to collect, align, and standardize global-scale climatic parameters in time (link to WP3) and re-use them in ecological models. In this WP, statistical models and machine-learning models (both supervised and unsupervised) will be used for Big Data analysis operations. The WP includes the development of an inference system for vessel-transmitted information that estimates fishing effort and other features related to patterns and processes of fishing activities per fleet and area. All data and models will feed an overarching model that will assess the impact of climate change and fisheries on the ecosystem, and will produce a vulnerability and a risk assessment index. With the completion of WP4, a global- scale index on the impact of climate and human activities on biodiversity, ecosystems and fisheries will be delivered that will be incorporated in the EcoScope Toolbox (WP5).
WP4 will evaluate the effect of climate change and variability on species distributions and abundance using high- resolution maps, ecological niche models and machine learning techniques for (common and rare) commercial and non- commercial species and will perform a vulnerability and risk assessment of the effects of climate change and fisheries on ecosystem components. The outputs of WP4 will be utilized in the framework of WP5 (energyscapes and ecosystem assessments), WP6 (Ecosystem models) and WP7 (Economic indicators). WP4 is structured to standardize global- scale climatic parameters in time and use them in ecosystem models, using data from open repositories of climatic, oceanographic and geophysical parameters that collected in WP3. All the tasks are interlinked and the output of WP4 will be available via the EcoScope Platform (WP3) and will form one of the EcoScope Toolbox (WP5) metrics.
We will use climatic and geophysical parameters from open repositories collected in WP3 and available through the EcoScope Platform, and align them in time and space. Further, we will make them available as standardised and open- access data under the NetCDF-CF format and multiple resolutions. In particular, we will include at least data from the NASA Earth Exchange platform, Copernicus, EMODnet and AquaMaps (from WP3), updated and processed to be used in ecosystem models. These Big Data include forecasts for several global-scale environmental parameters between 1950 and 2100, at different spatial and temporal resolutions and under different socio-economic and greenhouse gases emission scenarios (IPCC SRES A2, RCP 4.5, and RCP 8.5). This collection will be the basis of qualitative and quantitative models, and will be used to evaluate the effects of climate change on human activities and species distributions under different emission scenarios. Several parameters will be examined including sea surface temperature, primary production, bottom temperature, sea surface salinity, bottom salinity, ice cover, air temperature and precipitation (see also WP3 for detailed data collection and homogenisation). The mapping output of each scenario will be available in the EcoScope Platform (WP3) and will be used in the ecosystem models and management simulators (WP6). An index of climate effect will be delivered that will be later used as one of the metrics in the development of EcoScope Toolbox (WP5).
In this Task we will use vessel transmitted information (VTI) in real time to detect fishing and exploitation patterns, illegal activities, and gear types in strategic study regions (e.g. the northern Mediterranean Sea coastline) at 0.1° to 0.5° resolution. Specifically, we will use vessel trajectories to: (i) determine the distribution of fishing hours across the area per vessel type, (ii) identify highest fishing-pressure areas, (iii) specify the involved stocks, and (iv) and track incidental catch species. This analysis will be fleet and area specific. To this aim, we will use Big Data of historical and up-to- date trajectories from public (e.g. the Global Fishing Watch, AISHub.net) and private repositories owned by CNR and other partners. We will intersect these data with species observation records (from OBIS, FishBase, AquaMaps), stock information (from the FAO Global Record of Stocks and Fisheries) and the Sea Around Us Project (catch and effort data) through statistical and Artificial Intelligence-based inference systems. These data will be validated by fishery independent sources where available in the study areas. These inference systems will also estimate how fishing patterns have changed in time, by using time series forecasting techniques, which study the deviation of the patterns detected at a certain time with respect to the current status. The estimated fishing effort in each case study will inform stock assessment models, including multi-species models and fisheries management models (WP5) as well as ecosystem models and simulators (WP6) for improving the engagement with users (WP8), and will eventually be included as fishing effort maps in the EcoScope Platform (WP3) and as fishing impact on fish stocks in the EcoScope Toolbox (WP5).
Ecological niche modelling (ENM) refers to a set of computer-based approaches that predict the actual or potential distribution of a species across a geographic area and time based on environmental and geophysical data. We will use several Artificial Intelligence (AI)-based ecological niche models to determine the native areas of over 11,000 rare and commercial fish and invertebrate species, marine mammals and reptiles included in AquaMaps (aquamaps.org). We will produce the maps according to the different environmental scenarios and times of the data collected in Task 4.1. Ecological Niche models will include Maximum Entropy, AquaMaps, Artificial Neural Networks, and Support Vector Machines. These models will be combined to best use their complementary features and to increase niche-prediction accuracy. The resulting species distribution maps will be available through the EcoScope Platform (WP3) and used in assessing the ecosystem components and energyscapes (WP5) and in ecosystem models and simulators (WP6).
Climate change influences species habitat distribution over time and the productivity of stocks with consequences for many ecosystem services needed for human well-being. In the Levant Basin, for example, indigenous species of Atlantic origin inhabit deeper areas than in the Western Basin, which is often attributed to active warm water avoidance. Ecological niche models can be used on different snapshots of environmental parameters in time to evaluate changes in species habitat distributions due to important climatic or anthropogenic changes. In this Task, we will use the Task 4.1 forecasts, the fishing effort and patterns of Task 4.2, and the ecological niche models of Task 4.3 as input to a pattern recognition model that will detect trends of potential ecological change due to climate change. In particular, the forecasted parameters will be used to analyse the different responses of ocean areas and seas to climate change and fishing effort. These will be used within an automatic categorization process of a large number of niche models, which will produce global patterns of habitat shift and biodiversity change. Fishing activity patterns will be categorised to detect the regions exposed to highest changes in fishing activity. All these categorisations will be combined, to produce a global-scale impact index of climate change and fisheries effects, on biodiversity and ecosystems. This index will be used to produce a vulnerability assessment and an overall risk assessment of the effects of climate change and fisheries on biodiversity and ecosystems that will be available in the EcoScopium public portal. This impact index will be one of the EcoScope Toolbox (WP5) metrics.
D4.1 - Report on climatic parameters collections
Due Month: 20
D4.2 - Report on vessel transmitted information analysis
Due Month: 26
D4.3 - Report on ecological niche modelling
Due Month: 38
D4.4 - Vulnerability and risk assessment Index report
Due Month: 48