Distribution of Furcellaria lumbricalis based on data submission by HELCOM contracting parties. Mainly pointwise occurrences of Furcellaria were submitted, originally gathered in national mapping and monitoring campaigns, or for scientific research purposes. From Estonian waters, a predictive model was used (200m resolution), that was converted to presence/absence using minimized difference threshold (MDT) criteria. For Poland, only confirmed occurrence of Furcellaria were included (Slupsk bansk, Rowy reef and reef at Orlowo cliff). All data (Furcellaria points and the raster presenting predicted presence of Furcellaria) were generalized to 5km x 5km grid cells.

Data shows the extent of land claim (permanent or temporary establishments of the sea) and the type of the construction. The data was made available by HELCOM Contracting Parties in response to data request. The data was received from Denmark, Finland, Sweden and Poland. The activity was declared as not relevant in Germany, Estonia, Latvia and Lithuanian. From Russia no data was reported. Attribute specification and units: Country: Country Type: Type of construction (land claim) Type_spec: More specified information about the type of land claim Year: Year of construction Estimated: Estimated year of construction from the identification information (environmental permit) given by the country in question Length: Length of the land reclamation (m) Area: Area (km2) of the land claim X_lon: Original Longitude coordinate point (for the data that has been transformed from points into lines) Y_Lat: Original latitude coordinate point (for the data that has been transformed from points into lines)

This dataset contains modelled small vessel fuel consumption. This describes the geographical distribution of the fuel used by small boats. The total fuel consumption was modelled in SHEBA project to study emissions from pleasure boats. The model is based on locations and berths in marinas and leisure harbours, AIS information, statistics on fuel sale and extensive survey. For 2018 version the layer is weighted with depth, log-transformed and normalised (please see below). This dataset was also used on HOLAS 3.

Input of impulsive anthropogenic sound includes impulsive events from 2011-2016 • Seismic surveys (HELCOM-OSPAR Registry; national data call submissions as lines in the folder of data) • Explosions (HELCOM-OSPAR Registry) • Pile driving (HELCOM-OSPAR Registry) • Airguns (HELCOM-OSPAR Registry) For the different event types, numeric intensity value was used to represent the pressure as categorized in HELCOM-OSPAR Impulsive noise registry. All nationally reported seismic surveys were given intensity values “Very low” (0.25) - Very low (0.25) - Low (0.5) - Medium (0.75) - High (1) The impact distance has not been taken into account due to the different nature of separate datasets used for the pressure layer. We acknowledge that e.g. pile driving and airguns may impact up to 20 km from the source event. The spread of the sound wave depends on the sound frequency, water salinity, temperature and density.

The extraction of cod pressure layer is based on two datasets: 1. http://metadata.helcom.fi/geonetwork/srv/eng/catalog.search#/metadata/7a1389b3-382a-487f-8888-ac45c94c5a97 for years 2011-2016 reported per ICES statistical rectangles (tonnes / ICES rectangle). 2. http://metadata.helcom.fi/geonetwork/srv/eng/catalog.search#/metadata/debeafcd-948b-4455-88ae-7a3d1618f5a8 from ICES recreational fisheries reports for 2011-2016, reported per country (only coastal areas included). Landing values were redistributed within each ICES rectangle by the c-square fishing effort data provided by ICES (all gears, 2011-2013). Tonnes / km² were calculated for both data sets and the results were converted to 1 km x 1 km grid cells. The layers were summed together, log-transformed and normalised to produce the final pressure layer on extraction of cod. Please see "lineage" section below for further details on attributes, data source, data processing, etc.

This dataset shows the sub-basins of the Baltic Sea which are used for Baltic Sea Pollution Load Compilation 6.

The dataset contains total landings of sprat for years 2011-2016 reported per ICES statistical rectangles (tonnes / ICES rectangle) under EU Joint Research Centre’s data collection framework for fisheries data. Russian data extracted from ICES annual reports.

Data set represents dredging activities around the Baltic Sea. The dataset contains information about the dredging activity and for some the type (capital, maintenance) and the year of activity as reported by HELCOM Contracting Parties in response to data request. The dredging data is missing from Denmark.

Input of heat pressure dataset contains delta heat values from warm water discharge of - Discharge of warm water from nuclear power plants - Fossil fuel energy production. The Discharge of warm water from nuclear power plants was requested from HELCOM Contracting Parties. Average temperature change of the cooling water (°C) and amount of cooling water (m3) was used to calculate the heat load in TWh. No data on heat load was available for the Leningrad nuclear power plant; therefore the average heat load (TWh) of discharge of warm water from nuclear power plants was given. No heat load information was available for fossil fuel energy production sites. Heat load of 1 (TWh) was given to all production sites, based on average heat load of individual production site in Helsinki from recent years. 1 km buffer was used for both datasets with steep decline around the outlet. Overlapping areas were summed. Heat load from both layers were summed and the layer was normalized.

'Availability of deep water habitat, based on occurrence of H2S' layer describes the suitability of the bottom areas for the Baltic Sea biota, with regard to oxygen conditions of the near bottom waters. The data used to produce the layer was received from Leibniz-Institut für Ostseeforschung Warnemünde (IOW): - areas (polygons) with hydrogen sulfide (H2S) based on point measurements and modelling. Five time periods / year, for years 2011-2016 (altogether 30 layers). The polygons were converted to raster layers in a way, that for each time period (6 years, 5 time periods each year), areas with H2S got a value 0, other areas got the value 1. All layers were summed, (representing 6 years, 5 time periods each year, maximum value 30) and data was normalised. For more detailed information on the data used, please see Feistel et al. 2016.