Pressure layer combines all human activities that cause changes to hydrological conditions. The human activities were presented as point data which were given spatial extents (given below). The pressure value was given as the proportion of the grid cell under the pressure. The following human activities were combined into the changes to hydrological conditions layer; - Hydropower dams (a 1km2 grid cell in the river estuary was selected) - Water course modification (1 km) - Wind turbines (operational, 0.3 km, linear decline) - Oil platforms (0.5 km, linear decline) The human activity datasets were first processed separately covering the whole Baltic Sea and then summed together and overlapping areas were dissolved to remove double counting. Attenuation gradients are assigned to each layer as described above. Area effected decreases when distance from avtivity increases. Layer was normalized.

This pressure dataset is derived from three human activities datasets - Urban land use (on land) - Recreational boating and sports (updated layer for 2018 version, please see separate http://metadata.helcom.fi/geonetwork/srv/eng/catalog.search#/metadata/8c30e828-1340-4162-b7f9-254586ae32b6) - Bathing sites These data are described in more detail in separate fact sheets. Urban land use data was first converted to 1 km grid cells and expanded with 1 km. Thus, coastal urban areas extended also to the sea. These areas were given value 1 and other sea areas, value 0. Bathing sites (points) were converted to 1km grid and given value 1, rest of the sea areas were given value 0. Normalized recreational boating data was converted to 1 km grid cells. These three layers were summed to produce the layer (values from 0 to 3), after that the layer was normalized. Hunting and recreational fishing data were excluded from human disturbance layer, as they are mostly reported per country and would have resulted in overestimation of the actual pressure.

The layer depicts the pressure of hazardous substances in the Baltic Sea, based on the data from the HOLAS 3 integrated hazardous substances assessment. The methodology utilizes the integrated status values available for each HELCOM assessment unit on level 3. The results are based on multiple hazardous substances groups integration, done through the CHASE tool. The integrated assessment assess the hazardous substances status in biota, water and sediment, and final result in based on the worst status. As the SPIA is carried out using a 1x1km grid and the Integrated hazardous substances is assessed on vector-based HELCOM assessment units, the vector data is rasterized. First, the vector data is rasterized to 100x100m resolution, and thereafter it is aggregated to 10x10km grid using a mean value. A 10 km grid is used in order to make the gradients between assessment units slightly smoother and finally values are converted to 1x1 km resolution. Please see "lineage" section below for further details.

Introduction of radionuclides is based on HELCOM MORS discharge data (2016-2020) . Annual averages of CO60, CS137 and SR90 from the period 2016-2020 per nuclear power plant. Gradual buffer around outlet to 10km distance (Type B decline). 10 km buffer with linear decline composed of 5 rings from discharges of radioactive substances (Type B decline)12.

This data set on deposition sites of dredged material (points) reported by HELCOM Contracting parties according to http://www.helcom.fi/Recommendations/Rec%2036-2.pdf for the reporting period 2011-2016. The dataset contains data reported by nationally by nominated experts by HELCOM PRESSURE group for Denmark, Germany, Estonia, Finland, Latvia, Lithuania, Poland, Russia and Sweden.

Concentration of nitrogen pressure layer is interpolated from annual seasonal average of total nitrogen concentrations from surface waters (0-10 m) extracted from ICES’s oceanographic database, database of Swedish Meteorological and Hydrological Institute, EEA’s Eionet database and Data from Gulf of Finland year 2014. The points were interpolated to cover the entire Baltic Sea with Spline with barriers interpolation method. Values were log-transformed and normalised (more detailed description below).

This pressure dataset is derived from three human activities datasets Recreational boating and sports: Total fuel consumption of recreational boats modelled directly to 1 km grid cells[1]. Total fuel consumption of recreational boats presented as presence / absence. Rescaled with depth, log-transformed and normalized. Bathing sites, beaches: Point data converted directly to 1 km grid cells. Location of beaches presented as presence (1) / absence (0). Urban land use: Urban land use data was first converted to 1 km grid cells and expanded with 1 km[2]. Urban land use data was first converted to 1 km grid cells and expanded with 1 km. Thus, coastal urban areas extended also to the sea. These areas were given value 1 and other sea areas, value 0. Hunting and recreational fishing data were excluded from human disturbance layer, as they are mostly reported per country and would have resulted in overestimation of the actual pressure.S pecific pressure layers first modified by spatial extents and depth influence. Each of them is considered as of equal importance (same weight). Calculate the sum of the pressure in a cell. Normalized. [1] SHEBA project [2] Estimate of the human disturbance (underwater sound, visual disturbance).

The Baltic Sea Impact Index is an assessment component that describes the potential cumulative burden on the environment in different parts of the Baltic Sea. The BSII is based on georeferenced datasets of human activities (36 datasets), pressures (18 datasets) and ecosystem components (36 datasets), and on sensitivity estimates of ecosystem components (so-called sensitivity scores) that combine the pressure and ecosystem component layers, created in http://www.helcom.fi/helcom-at-work/projects/holas-ii project. Cumulative impacts are calculated for each assessment unit (1 km2 grid cells) by summing all pressures occurring in the unit for each ecosystem component. Highest impacts are found from the cells where both are abundant, but high impacts can be caused also by a single pressure if there are diverse and sensitive habitats in the grid cell. All data sets and methodologies used in the index calculations are approved by all HELCOM Contracting Parties in review and acceptance processes. This data set covers the time period 2011-2016. Please scroll down to "Lineage" and visit http://stateofthebalticsea.helcom.fi/cumulative-impacts/ for more info.

This dataset is built from the following Human activities dataset: • Hunting of seals The number of hunted seals (see separate metadata on hunting of seals) were averaged over 2011-2014 separately for grey seals, ringed seals and harbour seals (e.g. number of hunted grey seals / year). In Sweden the numbers of hunted grey seals in 2011 (74) were reported for the whole Swedish territorial waters), but here the numbers were set only to Swedish Gulf of Bothnia, as corresponding numbers were reported there in 2013 (75) and 2014 (65). The area of the reporting unit was used to calculate the number of hunted seals / km2 and the data was converted to 1km x 1km grid. For the Baltic Sea Impact Index, the values were normalized. Normalized value 0.5 was set to the level of quota for hunting of seal species in the Baltic Sea. The following quotas for hunting were used: Grey seal: 2000, Ringed seal: 350, Harbour seal 230.

This dataset is built from following Human activities datasets: • http://metadata.helcom.fi/geonetwork/srv/eng/catalog.search#/metadata/81c917ea-492d-48e2-9f00-e1bb7fe3e4fc • http://metadata.helcom.fi/geonetwork/srv/eng/catalog.search#/metadata/4fcd51dd-b8be-4e83-8cad-37c566782e8f The game hunting of seabirds data (see separate metadata): The total number of hunted seabirds were averaged over 2011-2015 (number of hunted seabirds / year). The area of the reporting unit was used to calculate the number of hunted seabirds / km2 and the data was converted to 1km x 1km grid. The predator control of seabirds data (see separate metadata): The total number of hunted cormorants were averaged over 2011-2015 (number of hunted cormorants / year). The area of the reporting unit was used to calculate the number of hunted cormorants / km2 and the data was converted to 1km x 1km grid. The two datasets were first separately log transformed and then summed, to get the total value for each grid cell. Zero values were given to all grid cells with no reported seabird hunting activity. The layer was normalized.