Geodata for Environmental Health

Published: 10 October 2018Updated: -Folkhälsomyndigheten


The geodata landscape in Sweden and Europe as a whole is complex, fragmented, and at times duplicative. The commissioning of data, that is, the collection of variables in the field, is rarely undertaken as part of a long-term monitoring program, and rather it is short duration and spatially limited projects that are the norm. At the local level, the data might be acquired to meet specific information needs or in response to the particular interests of environmental health officers or decision makers. At the regional or county level in Sweden, great differences between the capabilities of County Boards means that data availability is incredibly uneven. At the national and trans-national (i.e. EU) level, data collection is dominated by regulatory needs to report particular variables. Examples of this include the European Noise Directive that requires estimation of strategic noise pollution (i.e. noise pollution from major roads, railways, airports and urban areas) on a five-year cycle for cities with more than 100,000 people. The lack of data collection strategy, open data, and transparency in metadata are considerable hindrances to the full and efficient exploitation of geodata for environmental health purposes.

Despite difficulties accessing accurate, timely, spatially distributed, and relevant data, research has shown a range of methods that can provide valuable insights into important issues in environmental health. For example, numerical models of noise pollution dispersal are being challenged by conceptually simple empirical models using spatial analysis within a GIS. These offer faster, high resolution but less accurate alternatives to strategic estimates based on acoustic science. Similarly, index and indicator approaches are being used to map pollution vulnerability for groundwater without the need for complex descriptions of transport pathways within bedrock or sediments.

New technology such as satellite sensors for atmospheric composition monitoring or clip-on passive diffusion monitors for air pollution measurement are changing the way science can track events, threats, and behaviours. Spatial energetics, the science of the geography of activity, is utilising new data sources to explore how humans experience their environment and the health outcomes of such experiences. There is simultaneously a growing awareness that different groups perceive and experience their environment differently. For example, the elderly and the young face different challenges and have different preferences and needs. There are differences too between genders and people of different socio-economic status. This is expressed as different behaviours, limitations, and opportunities and ultimately different health outcomes.

Exciting new technologies and a rigorous scientific basis have led to some exciting new approaches to spatial epidemiology and environmental health. They have also highlighted new directions in which research can pursue insights into how the environment affects human health.

Om publikationen

I Folkhälsomyndighetens uppdrag ingår att identifiera, analysera och förmedla relevant kunskap till kommuner, landsting och andra berörda samhällssektorer. Geodata, dvs. data som kan kopplas till en geografisk plats, har ett potentiellt brett användningsområde men nyttjas inte till fullo idag inom arbetet med miljörelaterad hälsa.

Denna rapport syftar till att ge en överblick av vilken geografisk data som finns tillgänglig, fri eller till en kostnad, för svenska aktörer på lokal, regional och nationell nivå att använda i arbetet med miljörelaterad hälsa. Rapporten innehåller exempel på hur geodata kan användas och vilka analyser som är möjliga. Vidare görs en genomgång av den vetenskapliga litteraturen med exempel från olika länder för att illustrera hur geodata kan komplettera forskning och policyarbete inom olika områden av miljörelaterad hälsa i Sverige. Dessa innefattar användning av geodata inom hälsorelaterat arbete med luftföroreningar, dricksvatten, grönytor, buller, strålning, radon och klimatförändring.

Rapporten har tagits fram av docent Ian Brown vid naturgeografiska institutionen vid Stockholms universitet, på uppdrag av Folkhälsomyndigheten. Karin Ljung Björklund har varit projektledare.


Agneta Falk Filipsson
Enheten för miljöhälsa

Access to geodata

The geodata landscape is complicated by replication in availability and delivery systems, lack of clarity regarding accessibility and user eligibility, a tendency towards technical and often obscure language usage, and by a lack of transparency in pricing of data and services. This report will attempt to clarify the availability of geodata and related services for Swedish government agencies working in environmental health. It is directed towards The Public Health Agency (Sv. Folkhälsomyndigheten) but is in general relevant for other state actors, although with some reservation for some issues, which will be highlighted in this report.

It is important to identify broad classes of data and accessibility that share commonalities in order to facilitate an analysis of this landscape and to enable the reader to quickly identify the most relevant information for their given interest. Firstly, we can divide the landscape into free and paid data and services. Secondly, as apparent in the preceding sentence, we can separate data from services. The former, in this context, includes bundled observations from which information can be extracted for analysis, for example, a satellite image. The latter includes the delivery of a product for analysis or implementation, for example, air quality indicators derived from satellite imagery. National and international sources may also be used to further divide the geodata. Further classifications can be implemented depending on the needs of the individual, and separation into vector or raster data models and classification by spatial resolution, instrument type, or application are all possible. For the purposes of this report, we will maintain the first two classification levels although we will explicitly highlight national and international sourcing.

Swedish Geodata and Services

National Data Sources

Sweden has a national infrastructure for geodata called Vision 2020. The Vision 2020 strategy promotes free data, usability, accessibility, and good collaboration. The drive towards free and open geodata is not a new phenomenon, and public agencies have long had access to free national geodata, traditionally in the form of printed maps. Land surveys were established to support national and regional planning, census operations, taxation, and, of course, military or national security needs. In Sweden some of the earliest geodata with national coverage are the General Staff Maps from the late 1800s (Sv. Nationalstabskartan). These have been scanned by the National Land Survey (Sv. Lantmäteriet, hereafter referred to as LM) who have made them freely available through their website (Fig. 1). Older data enable the extrapolation of models back in time to investigate changes in epidemiology (1). H(2-17)Astelligh resolution versions of these products can be ordered for a relatively small price (150 SEK per map sheet in digital format). The provision of a basic service level and a higher-level paid service (in this case a higher-resolution copy) is indicative of the trend towards multiple service models in the geospatial industry. We can also see that older data, often considered of lower value, are made available at little or no cost whereas up-to-date data, with potential commercial value, are less widely available. A third model is that of data or services freely available for non-commercial use only.

Figure 1. Part of the General Staff map over Solna and Vasastan in Stockholm and Solna municipalities. The map includes surface elevations, bathymetric point data, and highly generalised land cover classes. This example is from 1873.

Figure 1. Part of the General Staff map over Solna and Vasastan in Stockholm and Solna municipalities. The map includes surface elevations, bathymetric point data, and highly generalised land cover classes. This example is from 1873.

LM and partners such as the Swedish Geological Survey (SGU) and the Swedish Meteorological and Hydrological Institute (SMHI) offer a range of data to the public, governmental organisations, and municipalities. The principal delivery and ordering platform is the Swedish geodata portal ( where one can find a wide range of geodata and metadata. classifies data as a dataset, web map service (WMS), or downloadable data. There is no clear definition of ‘dataset’ in this context, but it is assumed to be an acknowledgement that the data are held by an agency but might not be readily available. There are considerably fewer downloadable products compared to WMSs or datasets. For example, for the category ‘Health’ 18 datasets were available from the portal at the time or writing compared to 9 WMSs and only 2 downloadable products (Environmental hazard areas mapped by the County Boards and air quality data mapped from the Swedish Environment Agency). For many products, only metadata are available. These describe the type of resource, the responsible agency, and contact information, and, where available, a link to online access is included. Alternatively, a link to the ordering information may be found in the metadata.

To use the full range of products available through, individual agencies sign data access agreements with the collaborative Geodatasamverkan (a description of the collaboration can be found here). LM alone offers 81 resources through Geodatasamverkan, and additional resources are delivered by partner agencies SMHI, SCB, SGU, and the Swedish Maritime Administration (Sv. Sjöfartsverket). A full product list (in Swedish) is found here. At the time of writing, 32 agencies had signed on to the agreement along with 253 municipalities, 5 regions, and 3 state-owned entities also signed up, but the Public Health Agency is not one of these nor is the Swedish Work Environment Authority (Sv. Arbetsmiljöverket). Stockholm County Council (Sv. Stockholms läns landsting) pays 659,000 SEK/year for data access through the Geodatasamverkan agreement.

Much of the data that are not freely available directly through the portal are nevertheless available to users in governmental organisations such as the Public Health Agency or universities. The latter have access to geodata through a portal hosted by the Swedish University of Agricultural Sciences (Sv. Sveriges Lantbruksuniversitet, SLU). The Geodata Extraction Tool (GET) at SLU is intuitive and offers direct access to a range of vector and raster data, including 2 m digital elevation model (DEM) data, maps, and vector data from LM and Statistics Sweden (Sv. Statistiska centralbyrån, SCB). The situation for non-university government agencies is less clear. Access to geodata collections needs to be licensed on a per-case basis if no bilateral agreement exists and access through research agreements is not negotiated. However, the Vision 2020 strategy promotes free and open data and specifically states that fee-based delivery models must be replaced with other forms of financing. It is therefore possible, or even likely, that more national geodata will be made freely and openly available to users in the future.

Regional and Local Sources

At the regional and local levels, increasing amounts of data are being made available.

Regional data are collected by the County Boards (Sv. Länsstyrelserna) who have together launched a WMS that collates geodata in one portal. The available data vary widely with some counties barely contributing any data while others offer a wide range of products for download and presentation through a WMS. Much of the data in the County Boards’ portal is environmental or cultural. The data layers that are available are often apparently intended to support planning and enforcement decisions related to natural and cultural protection.

Table 1. Geodata available from the County Boards via the WMS portal at:

County BoardNumber of LayersDirect Health-Related Layers†
Blekinge 19 None- cultural and environmental layers only.
Dalarna 53 None- primarily environmental, cultural, and economic layers.
Gotland 28 None- environmental and cultural. Metadata and access limited.
Gävleborg 10 None- environmental and cultural (50% coastal protection).
Halland 25 None- cultural and environmental.
Jämtland 4 None- fisheries and hunting, environment, and culture.
Jönköping 121 Yes- some data on water quality and health risks
Kalmar 116 Yes- noise pollution estimates and sea level rise risk analysis (only).
Kronoborg 11 None- only environmental layers.
Norrbotten 4 None- only environmental and cultural.
Skåne 97 None- only environmental and cultural.
Stockholm 58* Yes - includes environmental health-risk sites and flood risk, noise pollution. *Includes links to other databases
Södermanland 3 None- limnology/hydrology only.
Uppsala 19 None- only environmental and cultural.
Värmland 20 None- only environmental and cultural.
Västerbotten 8 None- only environmental and cultural.
Västernorrland 29 Maybe- some layers on groundwater and water protection areas.
Västmanland 18 None- only environmental and cultural.
Västra Götaland 79 Yes- noise pollution from roads and air traffic, water protection areas, and ground water sampling.
Örebro 62 Yes- water quality/safety and environmental chemistry.
Östergötland 89 Yes- air and water pollution sampling.
County Boards Yes- noise pollution free areas.

†Disclaimer: this is based on a basic analysis of content and does not represent an analysis of all possible uses of the available geodata. For example, sampling of mercury in pike is included as a public health-related layer, but eel fisheries are not.

Some counties have, nevertheless, included data with clear public health applications (Table 1). Noise pollution data exist for at least four counties, and the other subject area with data from multiple counties is water quality and water pollution. However, it should be noted these data are not regularly updated in most cases and appear to have been one-off projects of the respective regions. That said, it is highly likely more geodata are held by the regions and are freely available but are yet to be published in the geodata portal developed the County Boards. Thus, access to such data requires prior knowledge of their existence.

At the local level there is a great deal of variation between the open geodata offered (unsurprising given that the population of Bjurholm is less than 2,500 while Stockholm City has almost 950,000 residents). Many smaller municipalities do not host geodata or WMS services of their own. Bjurholm municipality, for example, links to the Swedish Agency for Marine and Water Management (HaV) website for data on local water quality at swimming sites; they appear to host no data themselves though they do clearly provide contact information. Åsele municipality has a Survey and Mapping unit (Sv. Mätning, beräkning och kartering, MBK) but no online data or services. A small random sample of other municipalities is shown in Table 2.

Table 2. A small random sample of geodata availability based on visits to municipality websites. Both large and small municipalities are represented.

Bjurholm No Link to HaV water quality sampling.
Falun Yes- some health related Includes radon risk maps.
Göteborgs Stad Some- at a cost A range of geodata available to purchase very similar in content to LM’s products. Little or no health content.
Kalmar Some Web-GIS available with radon risk areas; social and cultural data layers. Refer to extensive collections at County level.
Stockholms Stad Yes- some health related Includes noise pollution estimates.
Sölvesborg None Public health information also lacking.
Västerås Yes- including health related Extensive collection of web-GIS map layers including noise pollution, solar radiation, radon gas measurements, and cultural and social layers.
Åsele None available Has a GIS unit within the municipality.

Stockholm City offers a range of products through the Öppna dataportalen, and geodata can be accessed through WMS or download services. The data include model estimates of traffic noise pollution, other environmental geodata, and socio-economic and cultural datasets. Similarly, Kalmar municipality has an Open geodata portal based on an ArcGIS platform. Currently 37 datasets are available, from cycle paths to aerial photos and radon gas emissions. More municipalities can be expected to create open data portals as policies change, technology improves, and demand rises.

Much of the Swedish geodata is generated and reported to meet national and EU reporting requirements to track progress towards environmental goals, and summary data are often delivered to the EEA. These tend not to be geospatial data but rather summary tables. Like the Swedish geodata landscape, that of Europe is fragmented with the EEA hosting data and other platforms, such as IPChem, mirroring or replicating parts of those databases. Neither Sweden nor Europe has succeeded in developing an open portal with comprehensive, up-to-date geodata. The contrast between and the SLU GET service is striking. The latter is considerably more user friendly and its holdings are clear. This contrast is likely a function of the clear responsibility delegated to SLU and the demand generated for GET services from a relatively coherent and knowledgeable user group. This might mean that other geodata portals or databases will face challenges in their operations until there is a clear mandate and a well-defined, knowledgeable user group.

EU Copernicus Data and Services

Copernicus is a programme that aims to give Europe an independent and comprehensive Earth Observation capability. Copernicus aims to not only provide geodata in the form of satellite imagery, but also to develop a range of information services. These value-added services are targeted at end users primarily within the public sphere such as government agencies and regional authorities. The services are broken down into six thematic areas (or streams) sponsored by the EU (Fig. 2).

Figure 2. The six ‘streams’ of thematic services of the Copernicus program (image:

Figure 2. The six ‘streams’ of thematic services of the Copernicus program (image:

Copernicus services are delivered under indirect or direct management structures (the distinction being related to financial regulations of the EC). Currently the only direct management cases are Emergency Management Services (EMS) and elements of the Land Monitoring Service (LMS) that are delegated to the Joint Research Centre of the EU. Other services are delegated to partner organisations that operate under indirect management, for example, other LMS are implemented (managed) by the EEA; Atmosphere Monitoring and Climate Change are operated by the European Centre for Medium Range Weather Forecasts under a delegation agreement.

The following sections describe each service stream in turn with an emphasis on public health. The Copernicus Market Survey (2016) includes ‘health’ in one of the non-space domains in which Copernicus data and services provide added value. Cited examples are: “Monitoring of air quality, mapping of potential outbreaks of epidemics or diseases” (p. 21). Precision farming (LMS) is touted as having fewer risks for human health, forest fire prevention and monitoring (EMS) includes positive outcomes for human health, and urban heat islands are associated with health impact management of urban environments. Air quality is identified as a key sector for public health, and the Copernicus Atmosphere Monitoring Service (CAMS) reports air quality as a key human health domain.

However, no single service targets public health applications.

Copernicus Atmosphere Service (CAMS)

CAMS bridges the gap between climate and atmospheric science and policy and public health. It delivers services that address both the scientific parties interested in atmospheric composition and forecasts and the real-time services that target public health and health policy. The primary example of the latter is the European air quality service that provides four-day predictions of gases such as carbon monoxide, ozone, sulphur dioxide, nitrogen dioxide, and PM2.5 and PM10 particulates.

Figure 3. CAMS PM10 estimates for Stockholm, 30th October-2nd November 2017. Note that the legal threshold under European statutes is reported. PM10 may not exceed 50μg/m3/day for more than 35 days per year.

Figure 3. CAMS PM10 estimates for Stockholm, 30th October-2nd November 2017. Note that the legal threshold under European statutes is reported. PM10 may not exceed 50μg/m3/day for more than 35 days per year.

CAMS provides synoptic map data for Europe north of the southern Mediterranean shore to around 70°N extending from 25°W (west of Iceland) to 40°E (the Caucasus). Data can be accessed online through a WMS, and archive maps are available. City forecasts are available for larger European cities including Stockholm (Fig. 3), Gothenburg, and the Öresund region encapsulating Malmö and Copenhagen. Helsinki and Oslo are the only other Nordic cities served. Additionally, CAMS delivers global maps of air quality through an online WMS. Global archival data are only available upon registration and approval.

CAMS also provides 5-day forecasts of stratospheric ozone and of surface ultra violet (UV) radiation exposure in support of public health. The service delivers global UV index maps and higher-resolution maps over Europe. These products specifically target end users active in (skin) cancer prevention.

Copernicus Marine Environment Monitoring Service (CMEMS)

CMEMS offers a range of services focussed on physical oceanography (including sea ice, wind, waves, and temperature) and biology/biogeochemistry (nutrients, plankton, and turbidity). While products that affect the safety of shipping and recreational use of water bodies may be said to promote public safety, and therefore by extension health, these products do not target end users in the public health domain. The biogeochemical products, such as ‘Baltic Sea chlorophyll-a forecast’, are proxies of the content of potentially harmful cyanobacteria in the water. Cyanobacteria forming algal blooms might result in local levels of cyanobacteria that are a health hazard. The Baltic Sea product is a 2-day forecast updated twice daily. The product has a resolution of one nautical mile, and delivery is by file download or WMS (currently broken at the time of writing). Zooplankton, whose concentrations are also estimated by CMEMS, might act as a disease reservoir. Thus some CMEMS products might provide information in support of public health policy and advice to the public (e.g. the presence of summer algal blooms).

Copernicus Land Monitoring Service (CLMS)

The CLMS is a collection of three land cover mapping and monitoring services operating, respectively, at the global, regional (pan-European), and local scale. Additionally a reference data framework is provided as a WMS that presents the LUCAS land use and cover data collected by Eurostat every three years.

The CLMS global service offers moderate to low resolution products in the thematic areas of vegetation, energy, water, and cryosphere and a high resolution ‘Hot Spot’ on-request product for areas of special interest. Typical products are land cover-related such as the 300 m Burnt Area product (ex post facto wildfire mapping) and 300 m normalised difference vegetation index product (vegetation greenness). The global service therefore does not provide data or services that are directly relevant for actors in the public health sphere. Emphasis is on non-European locations. The CLMS pan-European service provides land cover/land use data, very high-resolution satellite image mosaics, and related products to end users. Data are available for download or as a WMS. While of interest in planning and land cover change investigations and perhaps useful in health-related applications such as insect and rodent breeding potential, these data do not directly address the needs of public health administrations. Furthermore, national datasets at better resolution or with better validation are likely to be available in many cases.

The local CLMS service is probably the most relevant with regard to public health issues in a Swedish context. The local CLMS offers three primary service lines to end users – the Urban Atlas, riparian zone monitoring, and Natura 2000 site monitoring. Of these, the Urban Atlas is probably the most relevant to public health (e.g. this service includes two Atlas iterations (2006 and 2012) and a Street Tree Layer that maps urban trees. These products are available as a WMS or downloads.

Coverage over Sweden includes Umeå, Uppsala, Stockholm, Örebro, Linköping, Jönköping, Gothenburg, and Malmö/Öresund. The product is currently only partially validated and contains clear underestimations of urban vegetation (Fig. 4). The Urban Atlas classifications might also require post-processing to enhance usability. In the case of Stockholm, the difference between ‘Green urban areas’ and ‘Sports and leisure facilities’ seems somewhat arbitrary with parks such as Tantolunden classified as both and Årsta Rugby Centre as a ‘Green urban area’, and the island of Djurgården mostly covered with ‘Green urban areas’ and not forest (a separate class) in the Urban Atlas. Further validation might improve the usability of the service, although perhaps national solutions with better customisation for local requirements and conditions might be attractive.

Figure 4. A screenshot of the WMS Street Tree Layer product over the Karolinska hospital and campus. There is a significant underestimation of the tree population, and tree polygons are mislocated over major arterial roads. Figure 4. A screenshot of the WMS Street Tree Layer product over the Karolinska hospital and campus. There is a significant underestimation of the tree population, and tree polygons are mislocated over major arterial roads.

Copernicus Climate Change Service (C3S)

The relevance of the C3S to public health monitoring and research lies in climate change monitoring that enables actors to better predict climate change impacts and to implement responses. The C3S seasonal forecast product includes a three month temperature and precipitation forecast that might be useful in planning for severe climate events. For example, the temperature product might be useful in planning and preparing for severe summer heat episodes. Other examples of the utility of the service are more likely to be relevant to public safety (i.e. relevant to the Civil Contingencies Agency, MSB). It should be noted that the product is currently experimental and yet to be fully validated.

Copernicus Emergency Management Service (CEMS)

The CEMS is directed at civil defence and public safety. As such, products are more suited to the appropriate agency, MSB. Events such as flooding and forest fires are not a major health hazard to the Swedish public in the same way as they might be elsewhere where wildfire smoke is a significant seasonal health hazard or flooding can act as a major vector for the spread of disease. Hence CEMS is only really relevant under scenarios where wildfires become endemic or climate warming facilitates the spread of disease-carrying insects that breed in standing water. CEMS might of course be relevant for any international activities in the public health sphere (for example as part of foreign aid or international disaster response).

Copernicus Security Service

The security service is active within border surveillance and activities in support of the security of the EU. As such, the service is unlikely to have direct relevance to public health professionals or for public health policy and practise in Sweden. No CSS products are publicly available at the present time.

Copernicus Satellite Data

The EU Copernicus programme has developed a range of geospatial services aimed at supporting European policy, industry, and security through better access to state-of-the-art geoinformation. With a focus on environment and security, Copernicus builds on two European systems of satellites – the Sentinel series of Earth Observation satellites and the Galileo Global Navigation Satellite System (GNSS). The Sentinels are a series of satellites and instruments developed to monitor the Earth and to provide environment, climate, and security observations. The Sentinels are currently comprised of five series, although a sixth, a CO2 mission, might be added.


Sentinel-1 is comprised of two identical platforms, Sentinel-1A and -1B, flying six days apart in low Earth orbit carrying C-band (5.65 cm wavelength) synthetic aperture radars (SARs). Building on experience gained from ESA missions including ERS-1 and -2 and Envisat, Sentinel-1 offers an all-weather imaging capacity largely unaffected by clouds or the polar night. The SAR can acquire data in different acquisition modes that affect the coverage and resolution of the data. In Stripmap mode, the resolution is 9 m for the Ground Range Detected product used by most non-specialists. The Single Look Complex pixel spacing is as high as 1.5 m and 3.6 m in range and azimuth, respectively. In Interferometric Wide (IW), the standard acquisition mode over Europe, a resolution of ~20 m is achieved.

Sentinel-1 is an environmental monitoring mission that best suits missions such as land cover change mapping and climate change impact detection. Sentinel-1 is capable of producing interferometric products that can analyse surface displacement in 3D. Typical applications for interferometric SAR include glacier velocity mapping, ground stability and earthquake investigations, and deforestation detection. The complex processing required and the presence of a type of image noise called speckle make the use of SAR datasets challenging for many non-specialists. Sentinel-1 data might not be a natural choice for applications in environmental and public health, though such data have been widely used in flood mapping.


Europe’s high-resolution optical data needs are met by the Sentinel-2 satellites, of which two are currently in operation. Sentinel-2 carries a Multispectral Instrument (MSI) with a revisit time of 5 days. These satellites acquire data at spatial resolutions of 10–60 m depending on the spectral band. Four bands in the visible blue to near-infrared range are acquired at 10 m spatial resolution, six bands are acquired from the red edge (the boundary between red and near infrared) to the shortwave infrared at 20 m resolution, and three bands are sampled at 60 m resolution to support the atmospheric correction and cloud masking of the data.

With six bands at the red edge and in the near infrared, Sentinel-2 is designed to support applications in vegetation analysis such as precision agriculture and natural resource monitoring. The high resolution of the data makes Sentinel-2 useful for many urban applications including mapping the growth of urban areas and the mapping of urban green spaces. Water quality investigations are also possible with Sentinel-2, though not with the sensitivity of Sentinel-3.

Sentinel-3 Products

Four instruments are carried on the Sentinel-3 platform, of which two might be of interest to public health investigations in a broad sense. The Ocean and Land Colour Imager (OCLI) and the Sea and Land Surface Temperature Radiometer (SLSTR) acquire environmental data that can be used as part of an environmental monitoring package (the other two instruments acquire data for oceanographic and climate-related applications).

The Copernicus Science Hub is a portal for searching for satellite data from Sentinels-1 ,-2 and -3. Sentinel-3 OCLI products are currently available at Level-1B and Level-2 production levels from the Copernicus Science Hub. The former are radiometrically corrected, ortho-rectified, top-of-atmosphere radiances at spatial resolutions of 300 m (Full Resolution, FR) or 1200 m (Reduced Resolution, RR). SLSTR products are also available at Level-1B (brightness temperatures and top-of-atmosphere radiances) and at Level-2. Level-2 products are geophysical quantities estimated using Level-1 data as input into complex processing chains often using models (Table 3).

Table 3. Selected Sentinel-3 Level-2 products.

ProductApplicable to public health?
OCLI Global Vegetation Index No- low resolution vegetation mapping product
OCLI Terrestrial Chlorophyll Index No- low resolution vegetation mapping product
Algal Pigment Concentration Yes- indicator of algal blooms
Total Suspended Matter Concentration Yes- indicator of water quality
Coloured Dissolver Matter Absorption Maybe- indicator of water quality
Sea Surface Temperature No- though may be used to predict algal blooms
Land Surface Temperature Maybe- Near-real time availability: could be useful for mapping extreme temperature events (heat waves)
Fire Radiative Power No- fire intensity product

Sentinel-4 will provide atmospheric products using an Ultraviolet, Visible and Near-Infrared Sounder (UVN) in support of the CAMS. The Sentinel-4 instrument package will be carried on the Meteosat Third Generation Sounder (MTG-S) weather satellites to be launched by EUMETSAT in 2023 and 2031. These are geostationary satellites orbiting in phase with the Earth’s rotation to constantly cover the same ground area. It should be noted spatial coverage is limited to 65°N and will therefore not include northern Sweden.


The Sentinel-5 precursor mission (S5P) is a mission aimed at providing next-generation atmosphere observations. The mission is partly a stop-gap measure until the launch of the full Sentinel-5 instrumentation on Metop-SG 1A and successors in 2021 and later. These polar orbiting satellites will provide advanced weather and atmosphere observations.

S5P data are not yet available – the mission launched on October 13th 2017 and is currently going through commissioning. Data policy states all data and products will be freely available to end users, including the general public. Level-2 products are likely to be most relevant to all but scientific users. The products include aerosol concentrations and atmospheric thickness. However, the primary end user of S5P data outside of the atmospheric sciences community will be CAMS who will use S5P data to develop air quality products that are more appropriate for use in monitoring air quality by public health professionals. Together, Sentinal-4 and S5P support better air quality monitoring in Europe, but their contribution is mostly through CAMS and therefore it is to CAMS that interest should be directed.

Commercial geospatial data and service providers

The commercial market segment for geospatial data is dominated by the providers of very high-resolution satellite imagery. These data providers also typically offer limited mapping products derived from satellite imagery. These are not geospatial services per se, but rather value-added products. A number of large actors are present in this market, which has seen some consolidation over the last decade. In Europe, Airbus is a leading provider of data as are former Land Surveys and spin-off companies such as Lantmäteriet in Sweden and their former daughter company Metria, e-geos of Italy (partially owned by the Italian Space Agency), and in North America the MDA Corporation and Satimaging Corporation are major suppliers of imagery and image-based solutions. New actors, often targeting specific applications or industries, are appearing. For example, Enview offers services that monitor pipelines and power transmission lines in order to detect hazards. Products include vegetation encroachment near power lines and geohazard detection near pipelines.

Service providers exist at the national and regional level delivering commercial geospatial services mainly to government and larger industrial clients. In Sweden, for example, Metria delivers services in the energy, forestry, and telecoms branches for clients such as Vattenfall, Fortum, Skanova, and Skogsstyrelsen. Other actors include Brockmann Geomatics, with customers including Norrköping municipality and the County Boards (länsstyrelserna). Products in these cases tend towards environmental monitoring solutions of limited duration. Some monitoring services are emerging offering clients longer monitoring, for example, monitoring of infrastructure in support of maintenance. An evolution of these services is the bespoke geospatial service that addresses the unique needs of the client. Services available might include the mapping of a resource within a given space, such as green areas within a municipality; a connectivity analysis of habitats; inventories of green infrastructure or ecosystem services; and urban planning and decision support tools.

Non-commercial and Philanthropic Service Providers

At the regional and local level, universities and research institutes have emerged as service providers. Working either in research-industry partnerships, as contract R&D, or in partnerships as part of outreach and community engagement, academia is particularly active in North America. For example, The North Carolina Institute of Public Health (NCIPH) at the University of North Carolina offers a range of services including GIS and Spatial Health:

NCIPH is dedicated to assisting state and local public health agencies, hospitals, and other agencies with using GIS for spatial analysis and data collection”

NCIPH offers training and technical assistance in geospatial methods for public health applications. The products include an online analysis system showing determinants of health as a support service for policy makers and planners. NCIPH is a partner in the Vulnerable and At-Risk Populations resource guide, which is a spatial information resource/tool for health agencies and emergency management. Drexel University’s Urban Health Collaborative (UHC) engages with the local community in Philadelphia to promote community improvement. The UHC provides briefs, offers GIS consultation, and maintains a data repository, and the data held include census and health outcome data. Other universities focus more on teaching, training, and research, for example, Johns Hopkins University has a centre for Spatial Science for Public Health that is mostly engaged in research and teaching, and Oregon State University has a Spatial Health Lab focussing largely on research on urban environments and air quality, but with some outreach. Where the universities excel is in the development of new approaches and the identification of solutions to community issues. All of the above have case studies showing effective and often innovative approaches to the use of geodata for improving health awareness and improving health outcomes. Typically, it is the most vulnerable or disadvantaged who benefit from these interventions.

While many examples of this kind of geodata usage and service delivery are taken from the US, it might be expected that Swedish universities and research organisations will become more active in the future. Given renewed emphasis on outreach as the third pillar of university activities and the maturity of many of the techniques discussed, we might see more activities employing geospatial data for public outreach, for example, through the development and publication of environmental health indicators. The activism and success of US universities in promoting open data and debate in their local communities serves as a good model for their Swedish counterparts. Actions supporting partnerships between academia and the environmental health community, such as workshops, seminars, and networking, might do much to hasten such developments.

Spatial Analysis using Geodata in Public Health

The Environmental Health Report 2017 (18) describes how environmental factors affect the health of the Swedish public, and it details the state of current knowledge regarding exposure, health outcomes, and trends over time. From this report and its predecessors, we can establish a number of environmental factors affecting public health that can be wholly or partially evaluated using geodata and geospatial methods. This chapter presents a review of methods and data needs for the monitoring of exposure to, and the estimation of health outcomes from, air pollution, other environmental pollution, noise pollution, urban green spaces, and green cities. Consideration is also given to aspects of radon exposure estimation, electromagnetic fields, sunlight and skin cancer, and to climate change impacts on health.

Air pollution outdoors

One of the major applications of geospatial methods and geodata in environmental health has been the estimation of outdoor air pollution dispersal and exposure. Early work attributed health outcomes, including mortality, to environmental factors using statistics from which evolved the spatial methods that today are recognised as geographic information science (2). Throughout the 1990s advances in computing and the development of more sophisticated algorithms resulted in the development of modern GIS offering wider access to spatial analysis. The adoption of GIS advanced what was known as spatial or geographical epidemiology partially through inter-disciplinary research and collaboration.

By the early 2000s, GIS was improving the performance and accuracy of air pollution models, and GIS offered efficient calculation of stochastic models based on available geodata (3). GIS approaches enabled model resolution to increase relative to earlier physical dispersal models while improving performance because GIS could reduce the need for tessellation (19). Stochastic and dispersal model approaches each had, and still have, respective advantages and disadvantages, but GIS approaches exploiting newly available geodata increased the availability of spatial analyses at high resolution. GIS components have also improved the performance of physical models, for example, through better representation of dispersal of emissions modelled using a physical (source) model (19, 20).

GIS solutions to modelling air pollution are largely based on stochastic (regression) approaches (1, 3, 6). The availability of land cover data at high resolution supported the development of Land Use Regression models (6, 21). Cadastral data at high resolution (city block level) improved both dispersal modelling by improving descriptions of the cityscape and improving the accuracy of georeferencing for exposure analysis of receptors (4). Buffering offers a simple method to evaluate at-risk populations in proximity to sources such as roads (4-6). These methods require the accurate positioning of addresses through a process called georeferencing or geocoding. Using the accurate positioning of receptors allows GIS to model exposure in large numbers of individuals, thus significantly outperforming interpolation methods by better resolving the scale of pollution variability (21, 22). GIS also facilitates the linking of databases, and in Denmark this has allowed the contribution of residential wood combustion (RWC) to net pollution to be accurately estimated by exploiting data in the Building and Dwelling Register (20). Hence the availability of open cadastral and infrastructure data and geocoded addresses from census data and other sources has enabled the modelling of individual exposure rates for large populations and diverse species and sources.

Validation of dispersal estimates has traditionally been problematic. Few studies have had the resources to measure multiple pollutants across a cityscape or landscape, and the relationship between ambient and personal exposure is often uncertain (23). London has a relatively dense network of measurement sites, and yet (19) were able to use only 30 annual and 4 daily sites to validate their model. Point sensors have sometimes been periodically redeployed to a new site to try to improve spatial coverage, and recent innovations in sensor design have facilitated a wider distribution of sampling to improve the spatial resolution of the observations. The use of cheap passive sensors enabled (24) to measure NOx and O3 simultaneously at 20 sites across two Swedish cities during three spring-summer campaigns. The ability to deploy so many sensors facilitated the analysis of differences between pollutant, site, and time. The results have implications for stochastic modelling and study design and demonstrated the potential of passive sensor deployment.

Local or regional patterns in exposure rates can clearly be modelled using geospatial data and methods and/or using physical models building on spatial data. These provide exposure and risk information that can be transformed into general health outcomes on a statistical basis. They do not, however, provide information on the temporal variability of exposure nor do they resolve differences based on personal behaviours. However, personal exposure histories have been established using a combination of global navigation satellite systems (GNSS, also known as GPS) and portable samplers (7, 25).

Small clip-on passive diffusive sensors allow research into personal exposures based on location and behaviour (23, 25). These innovative sensors enable the accurate logging of personal exposure, which provides a far more realistic spatio-temporal understanding of risk. Results have shown that ambient street-level measurements overestimate exposure (23) and that there are large differences between cities in terms of NO2 exposure (25). Thus a general model is not appropriate and local factors such as urban canyon effects and vegetation coverage are important to include in the models. Volatile organic compounds such as 1,3-butadiene, benzene, and formaldehyde generated by combustion engines, vary more greatly between individuals than cities meaning that for these carcinogenic and highly toxic emissions personal behaviour determines risk.

A method that plots personal exposure during daily activity has been demonstrated to refine estimates of personal exposure to near-surface O3 (7). The spatio-temporal exposure was plotted on a space-time cube revealing the receptors’ exposure during activities such as commuting to and from work and taking a lunch break. The actual exposure rate was refined using estimates of intake based on specific behaviour (cycling, running, sedentary). This coupling of activity logs and space-time mapping can be used to help individuals make informed choices leading to lower personal exposures; for example, a lunchtime run in a high-risk urban setting might be deferred, and alternative cycle routes might be preferred based on exposure data. Hence mapping behaviour and personal exposure is needed to better understand the risk to individuals, workplace exposure, and temporal patterns.

Variations in exposure to natural O3 are known to be related to geography. Klingberg et al. (9) showed that O3 measurements in southern Sweden were strongly related to local topography and proximity to the coast. Night-time temperature inversions lead to lower vertical mixing and lower relative ozone concentrations at low-lying locations as O3 is depleted (because O3 is a product of daytime photochemical production, there is no night-time replenishment). At coastal locations, low deposition over water and low mixing increases the local night-time O3 concentrations. Hence simple topographical modelling might support a better understanding of O3 exposure risks at night, and topographic data and spatial modelling are made easier by (recent) widespread access to high-resolution digital elevation data such as Sweden’s NNH 2m GSD dataset and the near global SRTM digital elevation model.

Similarly, easy access to timely remote sensing imagery improves the mapping of air pollution sources in near real time. Weather services regularly issue air quality warnings based on satellite observations. As noted previously, the CAMS offers air quality projections. For sporadic events, pollution can be tracked using satellite imagery, and seasonal pollution in New Delhi, forest fires originating in Indonesia, and Saharan dust have all been tracked using satellite data and reported in the media. Trans-boundary pollution tracking is important for risk assessment and timely hazard warnings. For example, smoke from Canadian forest fires was shown to result in an almost 50% increase in respiratory hospital admissions among the elderly living thousands of kilometres away (8). There are also indications that socio-economic status (SES) might affect susceptibility to negative health outcomes from wildfire smoke (26). Large-scale agricultural burning has been largely eradicated in Europe; however, a large scale event in 2006 in eastern Europe saw the transport of pollutants such as PM10 and PM2.5 particulates to Scandinavia resulting in up to fivefold increases in the measurement of such particulates as well as 100% increases in background O3 (27). Estimates of sources of particulates are difficult, but long-range transport has been shown to be a major contributor of PM10 and PM2.5 in Swedish cities and might account for as much as 50% of premature deaths attributed to PM2.5 and black carbon (28). These estimations are produced with a range of geodata inputs, including road network, buildings, traffic data, and georeferenced RWC-source data.


Geospatial monitoring and modelling have an important role in the assessment of trans-boundary pollution exposure and risk assessment. GIS modelling is an important tool in the assessment of urban exposure to air pollution with geodata such as cadastre databases or the Swedish fastighetskartan being necessary for accurate reconstruction of dispersion and geocoding of receptors (1, 19). Geodata such as digital elevation models and land cover/land use maps used in a GIS framework can extend exposure modelling from urban to rural settings (9). Improvements in sensors and methods have extended monitoring to the personal exposure level resulting in a better understanding of the role of behaviour in exposure and risk (7, 23, 25). Thus a better understanding of spatial and temporal patterns in health outcomes from air pollution exposure is being made possible through improved models, geospatial data, and methods.

Environmental pollution in drinking water and indoors

The two major challenges to health and health policy regarding exposure to harmful substances in water are probably pollution in drinking water, particularly in private wells, and pollution of water at bathing sites. Both are particularly important in a Swedish context where relatively large numbers of people live, at least partly, in cottages not serviced by municipal water supplies and where bathing, boating, and fishing are very popular during the summer months.

Fishing and regular consumption of fish might expose sectors of the public to elevated levels of heavy metals. Indeed, as noted above, there are irregular programmes to monitor the toxicity of fish in inland waters. However, beyond interpolation of fish sampling, it is difficult to see how geodata can contribute to improved understandings of risk unless better data on point sources of pollution become available. Bathing waters might, however, be regularly monitored using geodata. The remote sensing of chlorophyll as an indicator of toxic algae is operational, and both SMHI and the CMEMS service offer monitoring products for the Baltic Sea. The monitoring of larger lakes has been demonstrated by geospatial consultancies, see, for example, the CyanoAlert project and The science behind these observations is well established and continues to be refined as spatial, spectral, and temporal resolutions of satellite imagery improve (29, 30). Spatial resolution in particular remains a challenge for monitoring inland waters as ocean colour sensors typically operate at spatial resolutions of hundreds of metres (31).

Heavy metals in drinking water

A greater health risk to the Swedish public is the relatively widespread presence of potentially harmful compounds in drinking water from private wells. Contaminants such as heavy metals; agricultural discharges including pesticides, fertilisers, and excreta; and naturally occurring arsenic and radon gas might be present in wells.

Mining and industrial activity are major anthropogenic sources of heavy metal and arsenic pollution, and urbanisation has been a major source in China (32). High arsenic and heavy metal concentrations have even been found in Swedish tarmacadam products, but the lack of contamination of ground water indicated low levels of leaching and little environmental risk (10). Arsenic and heavy metals are also naturally occurring and are associated with particular bedrock and drift formations (33). Rock fracturing can increase water contamination risk (12).

Arsenic contamination of drinking water is a serious issue in many parts of the world, and Bangladesh has been particularly affected by arsenic in drilled wells. A joint study by SGU and SSI found arsenic concentrations above the recommended maximum levels in Swedish wells (33). Around 5% of rock-drilled wells and around 4% of earthen wells exceeded recommended limits, and one site was reported to have arsenic levels of 231 μg/l, which is well over the 10 μg/l threshold.

The estimation of arsenic enrichment or contamination has been performed using statistical models including indicators of spatial association and regression statistics and using spatial analysis in a GIS (32). The identification of independent variables associated with arsenic or heavy metal concentrations is well established. Geological and hydrological factors are the primary determinants of geogenic contamination, and simple models using geodata describing geology and sedimentology, wetness, and landcover or land use have been found to be effective in the modelling of arsenic and metal concentrations. A more advanced regression model classifier found 8 of 16 of the considered proxies to be relevant, and variables describing soil, hydrology, geology, topography, and gravity properties were significant (11). Contamination transportation models driven by relatively simple variables have been around prior to widespread GIS applications (34), but indexed and weighted overlay models are conceptually simple and are easy to implement in a GIS framework. Many modern implementations build on the early work of (34), which described the DRASTIC model (35). Recent improvements include the use of fuzzy logic (36, 37) and land use data from satellite imagery (35). Fuzzy logic methods are readily implemented in commercial and open-source GIS and improve the flexibility of index models. Land use data are used to both model overland flows and to identify sources of pollution such as agriculture (animal effluent, pesticides, and fertilisers).

Radon in drinking water

Radon gas dissolved in water is toxic and is a risk for a large number of people in the Nordic countries who use private wells (Fig. 5). Geological factors are the most important for the generation of radon – granitic rocks and alum shales are associated with high radon concentrations, and glaciofluvial sediments containing rock fragments have been found to result in higher measures of radon gas (13).

Figure 5. Radon levels in drilled wells (Bq/l) across Sweden from data gathered by SGU and SSI 2001-2006 (after (33)).

Figure 5. Radon levels in drilled wells (Bq/l) across Sweden from data gathered by SGU and SSI 2001-2006 (after (33)).

Hydrological controls are important for transport and delivery of radon. Yang et al. (12) have shown that fracturing is important because rock fractures promote the dissolution of uranium and radium. Additional variables such as till or drift overburden, residence time, groundwater chemistry, and flow path are also important, and the spatial association of high radon levels with such variables makes spatial analysis and GIS particularly useful in the prediction of radon exposure risk. A weighted index approach applied to 4,400 wells in Stockholm county achieved correlations of r = 0.87 between modelled and measured radon levels (14). Grid spacings of 50m were achieved, although mapping on an individual well basis is complicated by the complexity of local geological conditions and ground water connectivity (12). It has also been noted that regular usage reduces the concentration of radon (14), which might indicate that behavioural data, or proxies thereof, might provide refinement to risk models.

Indoor radon gas

Indoor radon gas levels are similarly affected by geological variables and can be efficiently estimated in a GIS environment using widely available geodata. Significant variables affecting indoor radon concentrations were established in the 1990s (38, 39). Primarily geological data are used to describe the occurrence or generation of radon gas, and both bedrock geology and overlying sediments (or drift geology) are important to the genesis and emanation of radon gas (13). Building design features are associated with the build-up of concentrations of the gas that might be harmful in the long term. Generally speaking, wooden and most modern materials are preferred over natural stone, and houses built into the subsurface might experience elevated levels of radon while cellars are associated with reduced levels of radon gas due to better ventilation and sealing (13, 38). In Sweden and Norway, the presence of radon in water is, perhaps not surprisingly, associated with higher concentrations of radon gas indoors (16, 39).

Figure 6. Potassium, uranium, and thorium mapped at 1:500000 scale over Stockholm by SGU.

Figure 6. Potassium, uranium, and thorium mapped at 1:500000 scale over Stockholm by SGU.

Geological data are widely available at moderately high resolution in the Nordic countries. Additionally, dwellings data are available describing variables such as the presence of a cellar and the age of a building that are known to affect the concentration of radon gas indoors (13, 39). Where available, gamma ray spectroscopy data might improve the mapping of sources (14, 15). Sweden, for example, has national data gamma ray spectroscopy data captured by SGU. Combinations of dwellings data, remote sensing, and geological maps can result in efficient and accurate estimates of indoor radon at city block or individual building scales (15).


The granite rocks and rock-bearing shales and tills of Sweden and the Nordic countries might contain significant levels of uranium that in turn results in elevated concentrations of radon gas in water and the indoor atmosphere. Even naturally occurring arsenic can be found in drinking water from private wells. Exposure to elevated levels of such toxins is therefore to be expected given the geology of the country. Coupled with a culture that values a connection to the countryside and the maintenance of summer houses or cabins that have been held by families for generations, risk levels might be higher than in many other European countries. Reliance on private wells without effective filters and houses with natural stone walls or foundations increase the risk of exposure to radon in particular. However, the accurate knowledge of determinants of higher concentrations means that simple but effective models of risk can be generated with suitable input data.

Geological data exist to support the evaluation of risk. For example, gamma ray spectroscopy data are available to improve on generalised geological mapping, and fracture maps are available to further improve models. Finally, national data on buildings and dwellings can be used to infer elevated risks associated with building design and age. Such data were sufficiently comprehensive to enable a cohort study of over 600,000 Swedes between 1973 and 2000 (16). Hence, geodata exist to support risk modelling, and point data are available to validate such models.

Urban Green Spaces

Urban green spaces are generally acknowledged to offer a range of services from leisure to health (40). They provide playing spaces for children often with amenities such as benches, playing areas with equipment, and prepared surfaces such as level grass playing fields, cushioned surfaces, and sand boxes that promote safe and healthy use. Urban green spaces might also provide community facilities such as public toilets, outdoor cafes, and pleasant spaces to congregate away from the bustle and stress of city life. Furthermore, larger spaces facilitate active lifestyles among some users, and walking, running, and, in particular, cycling might be facilitated by larger green spaces. Of course urban green spaces might also be badly maintained, unpleasant, and even threatening (e.g. (17)).

Here the term urban green spaces is used in a broad sense and includes local neighbourhood parks and larger urban forests. For example, in Stockholm spaces as diverse as Tegnérlunden and Djurgården are included, and in Malmö it would include spaces as diverse as Kapellplan and Ribbersborgstranden.

Measuring health and utilisation of urban green spaces

The health benefits of urban green spaces have been widely promoted and extensively researched. A review of the scientific literature published in 2010 found 485 articles broadly relevant to the subject (40), and the use of GIS in obesity research alone was found in 121 cases according to a 2017 review (41). However, evidence for direct and quantifiable health benefits was more nuanced, and many studies were found to be flawed or vulnerable to bias. Nevertheless, the authors did find that:

“There is strong evidence of the health benefits of physical activity, [but] the evidence for the link between physical activity levels and green space availability is weaker” ((40), p. 213).

They found a lack of robust evidence for positive mental health benefits of urban green spaces, but acknowledged the difficulty of quantifying intangibles. This finding was replicated by Astell-Burt and co-workers (42) who found no correlation between green space within 1 km of one’s residence, physical activity, and mental health. Contradictory findings were reported from Auckland, New Zealand, where anxiety/mood disorder treatments were associated with different measures of urban green space accessibility (43). Use of treatment numbers might be a better/more objective measure of mental health than self-reporting. Spatial aggregation was used to reduce the data down to the parcel level, which might help reduce the variance introduced at the individual level while retaining spatial detail. In both the Auckland study and a study from Catalonia, Spain (44), the level of surrounding green space was more important than access to green space within a short walking distance (e.g. 300 m). There was, however, an understanding that people feel that urban green spaces offer health benefits, and this is a positive outcome in itself.

In a survey of over 250,000 Dutch citizens’ self-reported general health, a clear linear relationship was found between increased amounts of green space within both 1 km and 3 km of the individual’s residence and decreased percentages of respondents reporting poor health (45). Not just the presence of urban green space is important, but the quality of such spaces also seems to be important. General health and the percentage of urban green spaces (as opposed to agricultural or natural green spaces) were negatively correlated in the study by Maas et al. (45), and the authors attributed this to the co-location of urban green spaces and high density urban development with the latter being perceived as unhealthy. However, in England it was found that respondents living near green spaces, particularly formal green spaces, were more likely to utilise green spaces than those farther removed (46). Furthermore, those utilising the green spaces were more likely to meet physical activity guidelines promoting healthy lifestyles. The differences between the studies by Maas et al. (45) and by Coombes et al. (46) might result from methodological differences. It is possible that the latter’s focus on urban green spaces is representative of behaviour patterns associated with people limited to the utilisation of those spaces, whereas the (45) study includes a combination of rural or peri-urban and urban respondents and includes preferential consideration of non-urban green spaces.

Research in Australia has found that despite perceptions of good access to sidewalks and shops, perceptions of neighbourhood attractiveness were lower in lower SES areas, and parks were 50% less likely to be perceived to be within walking distance in low SES neighbourhoods (47). It has been hypothesised that the prevalence of obesity might be linked to SES and hindrances to physical activity such as low walkability of neighbourhoods and poor access to green spaces (41). A lower body mass index has been associated with higher land-use mix, population density, the presence of major intersections and access to green spaces or recreational spaces in a number of GIS studies (41). In a cohort of 3,173 children in southern California, there were significant inverse associations between body mass index at age 18 and park acreage within 500 m of the home (48).

It has been argued that only the elderly and youth benefit from urban green spaces in very strongly urban areas (45). Research on positive and negative perceptions of spaces in Helsinki, Finland, found marked differences between age groups, and older people preferred locations close to home while children and adolescents favoured sports facilities over urban green and blue spaces (49). This might indicate that generalising facilities into green or non-green spaces oversimplifies perceptions of desirability and function. GPS and accelerometer data have been used to track the behaviour of 180 59–65 year-olds in Ghent, Belgium (50). They found that light and moderate-to-vigorous physical activity was more likely to occur in green spaces than non-green spaces and that sedentary behaviour was more likely to occur in non-green spaces. It was also shown that more sedentary activities were confined to the margins of green spaces and to access locations. Again, this shows that utilisation is complex and non-uniform between groups. The physically active late middle-aged adults were found to be selective in their use of urban space favouring urban green spaces. Similarly, in a study of older Finns (employing self-reporting and not measured behaviour) higher physical activity levels were associated with better environmental factors and better walkability (51).

Perceptions of walkability and environmental quality were identified as possible factors for some contrary behaviours such as lower reported activity despite positive objective evaluations of environmental quality. Similarly, it has been shown that self-reported activity is often overestimated (52), arguing for the implementation of measurement-based studies of behaviour.

The use of GIS methods in research on urban green spaces exploits both the spatial analysis capabilities of GIS software and the attributes of the data needed to explore relations between location and behaviour. Brownson and co-workers (53) identified the following variables as some of the most frequently assessed in GIS studies:

  • population density: exploiting census, buildings, and planning datasets
  • land-use mix: categorised from ownership records, planning data, and remote sensing imagery.
  • access to recreational facilities: simple buffers using geolocated residency data and more complex cost-path measurements requiring infrastructure network layers (roads, sidewalks and paths, and public transport).

Other specific approaches have been developed to address local or regional issues; for example, in the US walkability is considered important because urban planning tends to favour cars thus putting the poor at a disadvantage. Walkability can be calculated using connectivity tools and measures of urban density, land use, and other variables. Crime frequency and traffic safety statistics have also been used to investigate effects on the accessibility of urban green spaces, but no statistical evidence was found linking these factors to inactivity, although it was acknowledged perceived safety might be more important than actual safety (17). Parcel-level SES indicators, crime statistics, and other quantitative data can be easily employed when geolocated. Indeed, reducing complexity by spatial aggregation might improve understanding of the data (e.g. (43)). For Sweden, such data are normally available at the urban block (Sv. kvarter) or district (formerly parish) level. Despite the potential of GIS in exploiting geodata for public health applications, such potential has yet to be fully realised (53). Stratification by gender shows that women tend to report stronger associations with green space and health than men (e.g. (44)), but little research has investigated gender differences despite known behaviour and health variations between genders.

Ecosystem services provided by urban green spaces

Several studies have attempted to quantify the contribution of green spaces, particularly trees, to improved air quality through the deposition of particulates and ozone (e.g. (54-56)). Different levels of model sophistication have been employed from simple interpolation of point data of air quality measures and remote sensing mapping of tree spaces (56) to interpolated numerical simulations of air quality and particulate deposition (54, 55). While urban green spaces can remove relatively large amounts of contaminants from the air, it has been shown that regional vegetation is more important to urban air quality (54).


The availability of relatively cheap accelerometers and small GPS receivers, and their omnipresence in mobile phones, means that more studies might be able to generate objective data on behaviour. Dubbed spatial energetics, these approaches offer “high spatial resolution and temporally linked objective measures of environment and behaviour” offering new insights and perspectives on utilisation of green spaces, health, and physical activity ((57) p. 797). More innovation in quantifying behaviour and health outcomes should be accompanied by careful consideration of study design and targeted investigations. Reviews of the literature show that stratification by gender, SES, and age are needed for the investigation of physical and mental health and physical activity and obesity (e.g. (44, 51)). Consideration of different measures of green space and their quality are required to improve understandings of causal relationships (e.g. (43)).

Mapping noise pollution

Noise pollution mapping and reporting meets a range of policy and planning needs in the field of environmental health. Mapping noise levels enables authorities to design and implement mitigation actions such as building barriers, reducing vehicle speeds, and/or designing quieter road surfaces. Cities with populations of over 100,000 people are required to map noise pollution and to implement noise-reduction measures according to EU directive 2002/49/EG. Reporting is required at 5-year intervals and will, from 2018, need to conform to the Common Noise Assessment Methods in the EU (CNOSSUS-EU) method.

Interest in noise pollution mapping stems largely from epidemiological studies showing that long-term exposure to noise has a range of negative health outcomes, including cardiovascular problems and increased risk of obesity (e.g. (18, 58, 59)). Noise pollution is therefore clearly more than a negative aesthetic and has rightly attracted the attention of the public and environmental health communities and local to trans-national agencies. Stockholm Municipality, Stockholm County, Swedish government agencies such as The Public Health Agency (Sv. Folkhälsomyndigheten), the Swedish Environmental Protection Agency (Sv. Naturvårdsverket), the National Board of Health and Welfare (Sv. Socialstyrelsen), and the EU all engage in activities directed toward noise pollution policy, monitoring, and/or reduction.

Monitoring noise pollution is not trivial. In Stockholm municipality, a city of almost 1 million inhabitants, only two measurement stations are operated offering open data. Currently no online data are available for most cities including for example, Gothenburg, Malmö, Umeå, Norrköping, Örebro, Oslo, or London. However, Swedish cities are excellent at making available online noise pollution maps following the Nordic or EU estimation methods.

The choice of noise prediction strategy is an exercise in trade-offs between ease of implementation and model accuracy. Complex, accurate models require extensive input data describing noise emission, dispersal and reception. Simpler models employ empirical functions to account for ground reflection, attenuation, and absorption (60). Detailed comparisons of models have been published by Garg and Maji (60) and by Murphy and King (61). It has also been noted that different implementations of the same models might lead to different prediction results (61).

The Nordic noise prediction method uses input data describing noise emission (e.g. vehicle type and speed and road characteristics), noise dispersion (terrain, distance to receiver, barriers, and reflections), and the receiver (height, façade descriptors). Look-up tables provide inputs for different variables (62). This model allows users to predict noise levels using relatively simple inputs describing traffic flow and the receiver’s environment. However, the Nordic method does not consider the effects of urban canyon corrections, which have had to be manually estimated (63).

The Nordic method has since been replaced by the CNOSSOS-EU model, which requires more detailed inputs (64, 65). The newer model includes corrections for the spatial layout of the road such as intersections, roundabouts, land cover, terrain, and the use of studded tyres in winter. Such data are available from land surveys or highways authorities, and elsewhere they can be mapped from remote sensing data. However, corrections for some Swedish road surfaces, which are generally louder that their typical European counterparts, are not available (65). Nevertheless, traffic flow data is the most important component of the model, and therefore the other variables can be approximated where necessary to reduce the data burden of modelling. Ideally, traffic flow data should be measured in situ using electronic or pneumatic instruments over relatively long time periods. Alternative methods using remote sensing have also been employed (63).

Strategic noise mapping software implementations have relatively limited spatial tools and typically calculate noise at nodes in a grid and not as a continuous surface:

None of the packages compare positively with the mapping techniques available in commercial Geographical Information Systems (GIS) packages. In particular, the ability of GIS packages to deal with numerous types of spatial data far outweighs that available within commercial noise mapping packages. Indeed, as a reflection of this, some commercial software packages offer import/export functionality in attempt to take advantage of the greater ability of GIS to manipulate spatial data in a more sophisticated and customised manner” ((61) p. 295).

Using look-up tables to estimate noise emissions from traffic is a simple approach that reduces complexity. Using look-up tables and estimates of traffic flow from in situ observations and remote sensing data (66) were able to predict noise and annoyance levels in the city of San Francisco. The approach used in that study had limited input data requirements making model set-up less demanding, but this was at the cost of accuracy. More complex spatial solutions have since been developed, and a GIS implementation of a noise propagation model for ecological applications has been reported (67) and has been adapted to urban settings (68). The team that developed the original toolbox as an add-on to ArcGIS (the leading GIS software) has now expanded the toolset to include a second noise model NMSIMGIS ((69); Fig. 7). The Cadna/A noise model has also been applied in a GIS environment creating prediction maps for Munich (70). However, the noise model was designed to meet European Noise Directive requirements and was found to perform poorly along residential streets. Instead, a simple approach to noise annoyance was implemented using distance to major roads without a loss of accuracy in residential settings (70). This suggests that strategic noise modelling frameworks are inefficient outside of city centres and away from major highways.

Figure 7. Urban noise levels mapped in part of Fort Collins, Colorado, with the Sound Mapping Tools in ArcGIS (69). The levels in the urban park (green) and university campus (blue) exceed comfort levels for humans and birds. Figure used by the kind permission of Alexander Keyel.

Figure 7. Urban noise levels mapped in part of Fort Collins, Colorado, with the Sound Mapping Tools in ArcGIS (69).  The levels in the urban park (green) and university campus (blue) exceed comfort levels for humans and birds. Figure used by the kind permission of Alexander Keyel.


It should be noted that computationally intensive numerical methods might not suit applications of noise prediction such as town planning and urban development where practitioners might lack the skills needed to accurately employ such methods. Alternative empirical approaches that are simpler might meet the needs of many non-expert users.

Nevertheless, given the availability of high-resolution data describing buildings, structures, topography, and land cover at a national level in Sweden, spatial implementation of the noise prediction should be a priority. Garg and Maji (60) conclude that the CNOSSUS-EU framework is a good approach for strategic noise mapping, and a GIS implementation of the CNOSSUS-EU model should not present a major challenge. However, these geodata would also support noise prediction using the freely distributed Sound Mapping Tools that plug in to ArcGIS (67, 71). This would require less effort in development, and a freeware implementation is anticipated that would free users from a platform monopoly. Regardless of the preferred model or framework, it is clear that GIS is a natural platform for noise modelling as expounded by Murphy and King (61).

Solar and electromagnetic radiation exposure

Cancer risk mapping

In a review of spatial epidemiology (72), three type of approach were identified – mapping, geographic correlation, and clustering. An early implementation of GIS in spatial epidemiology used cluster analysis for the detection of childhood leukaemia hotspots at the parish level in Sweden (73). This investigation exploited census data and vector layers defining the units of aggregation (parishes). While no anomalies were found, the authors recognised that such an approach could be used to create a surveillance programme for detecting future leukaemia clusters. A similar approach applied to five Spanish regions found spatial variation in the occurrence of cancer, but no clusters (74). The cluster analysis method using the spatial scan statistic has been widely applied, indicating an important role of GIS and geodata in the identification of epidemiological anomalies (74). Evidence from the Netherlands also suggests a systematic spatial variation in overall cancer incidence with rates slightly higher in large cities (75). Lung cancer was highest in cities, and skin melanoma was highest along the west coast and lowest in cities, although recent increases in incidence rates have reduced the spatial variation. Spatio-temporal patterns in cancer risk and knowledge of causal risk factors led the authors to conclude that there exists a large potential for cancer prevention (75). GIS tools and spatial data clearly have an important role to play in the detection of patterns and the planning of mitigation efforts. Spatial patterns of cancer risk have also been reported for Ireland broken down into 3,401 electoral districts (76).

Skin cancer risk

Rising rates of skin cancer are a public health concern, particularly in Sweden where incidences of malignant melanomas and other skin cancers have increased dramatically since 1970 ((18), see figure 9.4 and 9.5). As noted above, spatial variations in skin cancer are well known, and coastal locations in Europe have exhibited higher incidence rates than neighbouring inland regions (75, 76, 77). Claeson (77) noted that higher cloudiness inland was associated with lower skin cancer incidence, and women in coastal areas of western Sweden had a 15% higher melanoma incidence than women inland while for men the difference was 13%. A comparative study using data from northern and southern Sweden found that there is a latitudinal influence on the number of naevi on children, a proxy for later skin cancer (78). Thus even crude geographic segmentation can be used to estimate global risk. Social class has been associated with relative risk in Sweden, and higher social classes or education might indicate increased foreign travel (e.g. to sunny destinations) and increased relative risk (79). Socio-economic data can therefore contribute to risk mapping, for example, through multivariate models in a GIS environment.

The sensitivity of biological effective solar radiation to ozone attenuation has been established experimentally (80). The action spectra peak is around 300 nm (UV-B) with a secondary peak around 375 nm (UV-A) and very low relative effectiveness at visible-blue wavelengths. Ozone depletion has led to increased exposure to UV-B radiation. However, stratospheric ozone is predicted to recover as a result of global action on ozone-depleting chemicals, and by 2050 harmful UV irradiance is expected to decrease by up to 20% at northern high latitudes and by 2–10% at middle latitudes (71).

Stratospheric ozone levels, cloudiness, and human behaviour will all affect skin cancer incidences in the near- to medium-term. CAMS, as mentioned above, offers a global solar UV index service (Fig. 8), and such data can be combined with near real-time cloudiness data or predictions or with a cloudiness index to predict exposure risk. In addition, radiative transfer models can accurately estimate irradiance for a range of atmospheric conditions (79), and socio-economic data can be used as input for calculating indicators that can be used in mitigation planning and in public education. Thus spatial modelling using geodata is capable of providing exposure risk-prediction products.

Figure 8. Example of the CAMS total sky UV Index (© ECMWF). UV indexes can be used to inform the public, to create policy, and as input into spatial risk models.

Figure 8. Example of the CAMS total sky UV Index (© ECMWF). UV indexes can be used to inform the public, to create policy, and as input into spatial risk models.

Electromagnetic radiation exposure

A great deal of research has investigated the incidence of cancers, including childhood leukaemia, around overhead high-voltage power lines (also referred to as transmission lines). A Norwegian study found no evidence to support the association of cancer or childhood leukaemia with proximity to high-voltage power lines (81), although a study in Sweden found weak support for such an association (82). However, both studies used relatively small sample sizes and did not take advantage of spatial analyses. A more modern Danish study using geodata in a GIS and using approximately three times the number of cases found no association between leukaemia and living close to power lines (83). Similar findings were reported in a large British study (84). The combination of detailed cancer registers, census data, and infrastructure databases means that large-scale studies are possible using registry or spatial/GIS approaches. The advantage of the latter is that manual accuracy checking of locations and distances is made easier. Where data on power lines do not exist, remote sensing methods can be employed (85), but whether such investments are necessary is a policy decision beyond the scope of this report.

Climate change

The impact of climate change on public health is expected to be severe and inequitable, and its prediction is hampered by uncertainties as explained in an extensive review (86). Obstacles to an effective international response to the threat of climate change and the expected negative health outcomes include:

  • scientific uncertainty that hinders determined and effective targeted actions
  • poor communication of risks in language that is accessible to the public
  • unwillingness to withdraw from established investments and to change behaviours (so-called “lock-in”)
  • active promotion of misinformation and wilful obstruction of efforts to prepare for climate impacts

Further extensive reviews have focused on the situation faced by the US (87), explored impacts on air quality (88), and examined the toxicology of climate change (90).

Collaboration between countries, government agencies, and individuals is a major message of the Lancet Commission Report on Health and Climate Change (86), and eliminating fossil fuels, stabilising atmospheric composition, and promoting equity are all central messages in the report. They note that surveillance and monitoring need to be strengthened and that capacity building is required for many countries:

“Information and data collected from public health surveillance or monitoring systems can be used to determine disease burdens and trends, identify vulnerable people and communities, understand disease patterns, and prepare response plans and public health interventions” ((86), p. 1878).

Specific impacts of climate change might require targeted monitoring, mitigation, and/or research actions. Persistent organic pollutants (POPs) are expected to re-volatilise under global warming scenarios leading to re-mobilisation of stored POPs and farther transport of atmospheric POPs (89,90). Clearly this represents a heightened challenge, and the lack of coordinated long-term monitoring amplifies the uncertainty associated with POP response scenarios and complicates mitigation efforts. Similarly, there is a need for more detailed and widespread inventories of pollutants, including volatile organic compounds, and enhanced satellite techniques will go some way to resolving this need (88). Platforms such as Sentinel-5P and the Sentinel-5 missions will go a long way to providing the data needed for emissions surveillance and exposure monitoring.


Many of the indicators identified by (86) and (87) require or can be supported by geodata and geospatial methods. These include pollution dispersal modelling, epidemiological modelling, ecosystem service inventories and surveillance, and weather and climate monitoring. There is also a need for key health indicators, empirical studies on health effects scenario analyses of future effects, and studies facilitating priority setting (87). Indicators are almost by definition spatial analyses and are typically generated from geolocated data in a GIS framework.


Buffering/buffer analysis: Buffers are geometric polygons extending a given radius from an object such as a line or point and are used to demarcate regions to be used in, or excluded from, an analysis. For example, a buffer of 300 m extending from the boundaries of a park might be used to delimit those houses within easy walking distance of the park.

Cluster analysis: The identification of points or samples that exhibit similar properties when projected onto a given space. Typically in GIS this refers to points representing a common sampled variable, or a similar numerical property, that are spatially close to each other.

Geographical Information System (GIS): A computer program, or sometimes a conceptual or software framework, within which geospatial data are stored, visualized, and/or analysed. At its simplest, it is a mapping tool. More advanced uses include spatial modelling, advanced geostatistical analyses, and 3D visualisation and analysis.

Geodata/geospatial data: Data that have a spatial dimension and are georeferenced to a coordinate system or datum.

Georeferencing/geocoding: The process by which data objects lacking a geographical reference such as a set of coordinates are assigned coordinates (a georeference) representing the geography of their origin. For example coordinates representing the centre (or centroid) of a dwelling might be assigned to variables associated with individuals living at that dwelling in order to enable the data to be used in spatial analysis.

Portal: An interface that links databases or web map services to browser-based searching and visualisation or downloading. For example a portal might be an access point for geodata repositories.

Web Map Service (WMS): A protocol or format for the delivery of map images using a web browser. A WMS allows the user to define a geography to be processed and might allow the user to select multiple layers of data to be presented.


  1. Bellander T, Berglind N, Gustavsson P, Jonson T, Nyberg F, Pershagen G, et al. Using geographic information systems to assess individual historical exposure to air pollution from traffic and house heating in Stockholm. Environmental health perspectives. 2001;109(6):633.
  2. Tim US. The application of GIS in environmental health sciences: opportunities and limitations. Environmental Research. 1995;71(2):75-88.
  3. Jensen SS, Berkowicz R, Hansen HS, Hertel O. A Danish decision-support GIS tool for management of urban air quality and human exposures. Transportation Research Part D: Transport and Environment. 2001;6(4):229-41.
  4. Nuvolone D, della Maggiore R, Maio S, Fresco R, Baldacci S, Carrozzi L, et al. Geographical information system and environmental epidemiology: a cross-sectional spatial analysis of the effects of traffic-related air pollution on population respiratory health. Environmental Health. 2011;10(1):12.
  5. Perez L, Lurmann F, Wilson J, Pastor M, Brandt SJ, Künzli N, et al. Near-roadway pollution and childhood asthma: implications for developing “win–win” compact urban development and clean vehicle strategies. Environmental health perspectives. 2012;120(11):1619.
  6. Ranzi A, Sterrantino AF, Forastiere F, Sartini C, Casale G, Cavallini R, et al. Asthmatic symptoms and air pollution: a panel study on children living in the Italian Po Valley. Geospatial health. 2015;10(2).
  7. Lu Y, Fang TB. Examining personal air pollution exposure, intake, and health danger zone using time geography and 3D geovisualization. ISPRS International Journal of Geo-Information. 2014;4(1):32-46.
  8. Le GE, Breysse PN, McDermott A, Eftim SE, Geyh A, Berman JD, et al. Canadian forest fires and the effects of long-range transboundary air pollution on hospitalizations among the elderly. ISPRS International Journal of Geo-Information. 2014;3(2):713-31.
  9. Klingberg J, Karlsson PE, Karlsson GP, Hu Y, Chen D, Pleijel H. Variation in ozone exposure in the landscape of southern Sweden with consideration of topography and coastal climate. Atmospheric Environment. 2012;47:252-60.
  10. Qvarfort U. The high occurrence of arsenic in Macadam products from an iron mine in central Sweden: significance for environmental contamination. Environmental geochemistry and health. 1992;14(3):87-90.
  11. Zhang Q, Rodríguez-Lado L, Johnson CA, Xue H, Shi J, Zheng Q, et al. Predicting the risk of arsenic contaminated groundwater in Shanxi Province, Northern China. Environmental pollution. 2012;165:118-23.
  12. Yang Q, Smitherman P, Hess C, Culbertson CW, Marvinney RG, Smith AE, et al. Uranium and radon in private bedrock well water in Maine: geospatial analysis at two scales. Environmental science & technology. 2014;48(8):4298-306.
  13. Sundal A, Henriksen H, Soldal O, Strand T. The influence of geological factors on indoor radon concentrations in Norway. Science of the Total Environment. 2004;328(1-3):41-53.
  14. Skeppström K, Olofsson B. A prediction method for radon in groundwater using GIS and multivariate statistics. Science of the total environment. 2006;367(2-3):666-80.
  15. Smethurst MA, Strand T, Sundal AV, Rudjord AL. Large-scale radon hazard evaluation in the Oslofjord region of Norway utilizing indoor radon concentrations, airborne gamma ray spectrometry and geological mapping. Science of the total environment. 2008;407(1):379-93.
  16. Tondel M, Lindgren P, Hellström L, Löfman O, Fredrikson M. Risk of malignancies in relation to terrestrial gamma radiation in a Swedish population cohort. Science of the total environment. 2011;409(3):471-7.
  17. Cutts BB, Darby KJ, Boone CG, Brewis A. City structure, obesity, and environmental justice: an integrated analysis of physical and social barriers to walkable streets and park access. Social science & medicine. 2009;69(9):1314-22.
  18. Folkhälsomyndigheten och Karolinska Institutet. Miljöhälsorapport 2017. Solna: Folkhälsomyndigheten, 2017.
  19. Gulliver J, Briggs D. STEMS-Air: A simple GIS-based air pollution dispersion model for city-wide exposure assessment. Science of the total environment. 2011;409(12):2419-29.
  20. Plejdrup MS, Nielsen O-K, Brandt J. Spatial emission modelling for residential wood combustion in Denmark. Atmospheric Environment. 2016;144:389-96.
  21. Craig L, Brook JR, Chiotti Q, Croes B, Gower S, Hedley A, et al. Air pollution and public health: a guidance document for risk managers. Journal of Toxicology and Environmental Health, Part A. 2008;71(9-10):588-698.
  22. Cyrys J, Hochadel M, Gehring U, Hoek G, Diegmann V, Brunekreef B, et al. GIS-based estimation of exposure to particulate matter and NO2 in an urban area: stochastic versus dispersion modeling. Environmental Health Perspectives. 2005;113(8):987.
  23. Bellander T, Wichmann J, Lind T. Individual exposure to NO2 in relation to spatial and temporal exposure indices in Stockholm, Sweden: the INDEX study. PloS one. 2012;7(6):e39536.
  24. Hagenbjörk A, Malmqvist E, Mattisson K, Sommar NJ, Modig L. The spatial variation of O 3, NO, NO 2 and NO x and the relation between them in two Swedish cities. Environmental monitoring and assessment. 2017;189(4):161.
  25. Hagenbjörk-Gustafsson A, Tornevi A, Andersson EM, Johannesson S, Bellander T, Merritt A-S, et al. Determinants of personal exposure to some carcinogenic substances and nitrogen dioxide among the general population in five Swedish cities. Journal of Exposure Science and Environmental Epidemiology. 2014;24(4):437.
  26. Reid CE, Jerrett M, Tager IB, Petersen ML, Mann JK, Balmes JR. Differential respiratory health effects from the 2008 northern California wildfires: A spatiotemporal approach. Environmental research. 2016;150:227-35.
  27. Targino AC, Krecl P, Johansson C, Swietlicki E, Massling A, Coraiola GC, et al. Deterioration of air quality across Sweden due to transboundary agricultural burning emissions. 2013.
  28. Segersson D, Eneroth K, Gidhagen L, Johansson C, Omstedt G, Nylén AE, et al. Health impact of PM10, PM2. 5 and black carbon exposure due to different source sectors in Stockholm, Gothenburg and Umea, Sweden. International journal of environmental research and public health. 2017;14(7):742.
  29. Jin Q, Lyu H, Shi L, Miao S, Wu Z, Li Y, et al. Developing a two-step method for retrieving cyanobacteria abundance from inland eutrophic lakes using MERIS data. Ecological Indicators. 2017;81:543-54.
  30. Toming K, Kutser T, Laas A, Sepp M, Paavel B, Nõges T. First experiences in mapping lake water quality parameters with Sentinel-2 MSI imagery. Remote Sensing. 2016;8(8):640.
  31. Clark JM, Schaeffer BA, Darling JA, Urquhart EA, Johnston JM, Ignatius AR, et al. Satellite monitoring of cyanobacterial harmful algal bloom frequency in recreational waters and drinking water sources. Ecological indicators. 2017;80:84-95.
  32. Li H, Li Y, Lee M-K, Liu Z, Miao C. Spatiotemporal analysis of heavy metal water pollution in transitional China. Sustainability. 2015;7(7):9067-87.
  33. Ek B-M, Thunholm B, Östergren I, Falk R, Mjönes L. Naturligt radioaktiva ämnen, arsenik och andra metaller i dricksvatten från enskilda brunnar. Strålskyddsinstitutet; 2008.
  34. Aller L, Lehr JH, Petty R, Bennett T. DRASTIC: a standardized system to evaluate groundwater pollution potential using hydrogeologic settings. National Water Well Association, Worthington, Ohio, United States of America. 1987.
  35. Brindha K, Elango L. Cross comparison of five popular groundwater pollution vulnerability index approaches. Journal of Hydrology. 2015;524:597-613.
  36. Dixon B. Applicability of neuro-fuzzy techniques in predicting ground-water vulnerability: a GIS-based sensitivity analysis. Journal of hydrology. 2005;309(1-4):17-38.
  37. Dixon B. Groundwater vulnerability mapping: a GIS and fuzzy rule based integrated tool. Applied Geography. 2005;25(4):327-47.
  38. Geiger C, Barnes KB. Indoor radon hazard: a geographical assessment and case study. Applied Geography. 1994;14(4):350-71.
  39. Tell I, Bensryd I, Rylander L, Jönsson G, Daniel E. Geochemistry and ground permeability as determinants of indoor radon concentrations in southernmost Sweden. Applied geochemistry. 1994;9(6):647-55.
  40. Lee AC, Maheswaran R. The health benefits of urban green spaces: a review of the evidence. Journal of public health. 2011;33(2):212-22.
  41. Jia P, Cheng X, Xue H, Wang Y. Applications of geographic information systems (GIS) data and methods in obesity‐related research. Obesity reviews. 2017;18(4):400-11.
  42. Astell-Burt T, Feng X, Kolt GS. Mental health benefits of neighbourhood green space are stronger among physically active adults in middle-to-older age: evidence from 260,061 Australians. Preventive medicine. 2013;57(5):601-6.
  43. Nutsford D, Pearson A, Kingham S. An ecological study investigating the association between access to urban green space and mental health. Public health. 2013;127(11):1005-11.
  44. Triguero-Mas M, Dadvand P, Cirach M, Martínez D, Medina A, Mompart A, et al. Natural outdoor environments and mental and physical health: relationships and mechanisms. Environment international. 2015;77:35-41.
  45. Maas J, Verheij RA, Groenewegen PP, De Vries S, Spreeuwenberg P. Green space, urbanity, and health: how strong is the relation? Journal of Epidemiology & Community Health. 2006;60(7):587-92.
  46. Coombes E, Jones AP, Hillsdon M. The relationship of physical activity and overweight to objectively measured green space accessibility and use. Social science & medicine. 2010;70(6):816-22.
  47. Giles-Corti B, Donovan RJ. Socioeconomic status differences in recreational physical activity levels and real and perceived access to a supportive physical environment. Preventive medicine. 2002;35(6):601-11.
  48. Wolch J, Jerrett M, Reynolds K, McConnell R, Chang R, Dahmann N, et al. Childhood obesity and proximity to urban parks and recreational resources: a longitudinal cohort study. Health & place. 2011;17(1):207-14.
  49. Laatikainen TE, Broberg A, Kyttä M. The physical environment of positive places: Exploring differences between age groups. Preventive medicine. 2017;95:S85-S91.
  50. Dewulf B, Neutens T, Van Dyck D, De Bourdeaudhuij I, Broekx S, Beckx C, et al. Associations between time spent in green areas and physical activity among late middle-aged adults. Geospatial health. 2016;11(3):225-32.
  51. Portegijs E, Keskinen KE, Tsai L-T, Rantanen T, Rantakokko M. Physical limitations, walkability, perceived environmental facilitators and physical activity of older adults in Finland. International journal of environmental research and public health. 2017;14(3):333.
  52. Dewulf B, Neutens T, Van Dyck D, De Bourdeaudhuij I, Van de Weghe N. Correspondence between objective and perceived walking times to urban destinations: Influence of physical activity, neighbourhood walkability, and socio-demographics. International journal of health geographics. 2012;11(1):43.
  53. Brownson RC, Hoehner CM, Day K, Forsyth A, Sallis JF. Measuring the built environment for physical activity: state of the science. American journal of preventive medicine. 2009;36(4):S99-S123. e12.
  54. Baro F, Chaparro L, Gomez-Baggethun E, Langemeyer J, Nowak DJ, Terradas J. Contribution of ecosystem services to air quality and climate change mitigation policies: the case of urban forests in Barcelona, Spain. Ambio. 2014;43(4):466-79. DOI:10.1007/s13280-014-0507-x.
  55. Fusaro L, Marando F, Sebastiani A, Capotorti G, Blasi C, Copiz R, et al. Mapping and assessment of PM10 and O3 removal by woody vegetation at urban and regional level. Remote Sensing. 2017;9(8):791.
  56. Marando F, Salvatori E, Fusaro L, Manes F. Removal of PM10 by forests as a nature-based solution for air quality improvement in the Metropolitan city of Rome. Forests. 2016;7(7):150.
  57. James P, Jankowska M, Marx C, Hart JE, Berrigan D, Kerr J, et al. “Spatial energetics”: integrating data from GPS, accelerometry, and GIS to address obesity and inactivity. American journal of preventive medicine. 2016;51(5):792-800.
  58. Münzel T, Gori T, Babisch W, Basner M. Cardiovascular effects of environmental noise exposure. European heart journal. 2014;35(13):829-36.
  59. Pyko A, Eriksson C, Lind T, Mitkovskaya N, Wallas A, Ögren M, et al. Long-Term Exposure to Transportation Noise in Relation to Development of Obesity—a Cohort Study. Environmental health perspectives. 2017;125(11).
  60. Garg N, Maji S. A critical review of principal traffic noise models: Strategies and implications. Environmental Impact Assessment Review. 2014;46:68-81.
  61. Murphy E, King EA. Strategic environmental noise mapping: Methodological issues concerning the implementation of the EU Environmental Noise Directive and their policy implications. Environment international. 2010;36(3):290-8.
  62. Bendtsen H. The Nordic prediction method for road traffic noise. Science of the total environment. 1999;235(1-3):331-8.
  63. Ögren M, Barregard L. Road traffic noise exposure in Gothenburg 1975–2010. PloS one. 2016;11(5):e0155328.
  64. Kephalopoulos S, Paviotti M, Anfosso-Lédée F, Van Maercke D, Shilton S, Jones N. Advances in the development of common noise assessment methods in Europe: The CNOSSOS-EU framework for strategic environmental noise mapping. Science of the Total Environment. 2014;482:400-10.
  65. Ögren M, Ryberg JB. Bullerberäkningar med CNOSSOS-EU i Sverige. Arbets-och miljömedicin (VMC), Göteborg; 2015.
  66. Seto EYW, Holt A, Rivard T, Bhatia R. Spatial distribution of traffic induced noise exposures in a US city: an analytic tool for assessing the health impacts of urban planning decisions. International journal of health geographics. 2007;6(1):24.
  67. Reed SE, Boggs JL, Mann JP. A GIS tool for modeling anthropogenic noise propagation in natural ecosystems. Environmental Modelling & Software. 2012;37:1-5.
  68. Bilaşco Ş, Govor C, Roşca S, Vescan I, Filip S, Fodorean I. GIS model for identifying urban areas vulnerable to noise pollution: case study. Frontiers of Earth Science. 2017;11(2):214-28.
  69. Keyel AC, Reed SE, McKenna MF, Wittemyer G. Modeling anthropogenic noise propagation using the Sound Mapping Tools ArcGIS toolbox. Environmental Modelling & Software. 2017;97:56-60.
  70. Birk M, Ivina O, von Klot S, Babisch W, Heinrich J. Road traffic noise: self-reported noise annoyance versus GIS modelled road traffic noise exposure. Journal of Environmental Monitoring. 2011;13(11):3237-45.
  71. Williamson CE, Zepp RG, Lucas RM, Madronich S, Austin AT, Ballaré CL, et al. Solar ultraviolet radiation in a changing climate. Nature Climate Change. 2014;4(6):434.
  72. Roquette R, Painho M, Nunes B. Spatial epidemiology of cancer: a review of data sources, methods and risk factors. Geospatial health. 2017;12(1):504.
  73. Hjalmars U, Kulldorff M, GUSTAFSSON G, Nagarwalla N. Childhood leukaemia in Sweden: using GIS and a spatial scan statistic for cluster detection. Statistics in medicine. 1996;15(7‐9):707-15.
  74. Ramis R, Gómez-Barroso D, Tamayo I, García-Pérez J, Morales A, Romaguera EP, et al. Spatial analysis of childhood cancer: a case/control study. PLoS One. 2015;10(5):e0127273.
  75. Siesling S, Van Der Aa MA, Coebergh JW, Pukkala E, Registry WGoTNC. Time‐space trends in cancer incidence in The Netherlands in 1989–2003. International journal of cancer. 2008;122(9):2106-14.
  76. Hegarty AC, Carsin A-E, Comber H. Geographical analysis of cancer incidence in Ireland: A comparison of two Bayesian spatial models. Cancer epidemiology. 2010;34(4):373-81.
  77. Claeson M. Epidemiology of cutaneous malignant melanoma in Western Sweden. 2016.
  78. Karlsson MA, Rodvall Y, Wahlgren C-F, Wiklund K, Lindelöf B. Similar anatomical distributions of childhood naevi and cutaneous melanoma in young adults residing in northern and southern Sweden. European Journal of Cancer. 2015;51(14):2067-75.
  79. Pérez-Gómez B, Aragonés N, Gustavsson P, Lope V, López-Abente G, Pollán M. Socio-economic class, rurality and risk of cutaneous melanoma by site and gender in Sweden. BMC Public Health. 2008;8(1):33.
  80. Horneck G, Rettberg P, Rabbow E, Strauch W, Seckmeyer G, Facius R, et al. Biological dosimetry of solar radiation for different simulated ozone column thicknesses. Journal of Photochemistry and Photobiology B: Biology. 1996;32(3):189-96.
  81. Tynes T, Haldorsen T. Electromagnetic fields and cancer in children residing near Norwegian high-voltage power lines. American journal of epidemiology. 1997;145(3):219-26.
  82. Feychting M, Alhbom M. Magnetic fields and cancer in children residing near Swedish high-voltage power lines. American journal of epidemiology. 1993;138(7):467-81.
  83. Pedersen C, Raaschou-Nielsen O, Rod NH, Frei P, Poulsen AH, Johansen C, et al. Distance from residence to power line and risk of childhood leukemia: a population-based case–control study in Denmark. Cancer Causes & Control. 2014;25(2):171-7.
  84. Bunch K, Keegan TJ, Swanson J, Vincent T, Murphy M. Residential distance at birth from overhead high-voltage powerlines: childhood cancer risk in Britain 1962–2008. British journal of cancer. 2014;110(5):1402.
  85. Matikainen L, Lehtomäki M, Ahokas E, Hyyppä J, Karjalainen M, Jaakkola A, et al. Remote sensing methods for power line corridor surveys. ISPRS Journal of Photogrammetry and Remote Sensing. 2016;119:10-31.
  86. Watts N, Adger WN, Agnolucci P, Blackstock J, Byass P, Cai W, et al. Health and climate change: policy responses to protect public health. The Lancet. 2015;386(10006):1861-914.
  87. Frumkin H, Hess J, Luber G, Malilay J, McGeehin M. Climate change: the public health response. American journal of public health. 2008;98(3):435-45.
  88. Monks P, Granier C, Fuzzi S, Stohl A, Williams M, Akimoto H, et al. Atmospheric composition change–global and regional air quality. Atmospheric environment. 2009;43(33):5268-350.
  89. Noyes PD, McElwee MK, Miller HD, Clark BW, Van Tiem LA, Walcott KC, et al. The toxicology of climate change: environmental contaminants in a warming world. Environment international. 2009;35(6):971-86.
  90. Ma J, Hung H, Macdonald RW. The influence of global climate change on the environmental fate of persistent organic pollutants: A review with emphasis on the Northern Hemisphere and the Arctic as a receptor. Global and Planetary Change. 2016;146:89-108.

Geodata for Environmental Health

Denna rapport ger en överblick av vilken geografisk data som finns tillgänglig, fri eller till en kostnad, för svenska aktörer på lokal, regional och nationell nivå att använda i arbetet med miljörelaterad hälsa. Rapporten innehåller exempel på hur geodata kan användas för miljörelaterad hälsa, vilka analyser som är möjliga samt en genomgång av den vetenskapliga litteraturen. Exemplen innefattar användningen av geodata inom hälsorelaterat arbete med luftföroreningar, dricksvatten, grönytor, buller, strålning, radon och klimatförändring.

Rapporten har tagits fram av docent Ian Brown vid naturgeografiska institutionen vid Stockholms universitet, på uppdrag av Folkhälsomyndigheten.

  • Author: Folkhälsomyndigheten
  • Published: 10 October 2018
  • Updated: -

Open the publication

Read the publication


This publication is not available for ordering.