Research on environmental health is increasingly using satellite data. In addition to studying environmental and climatic changes on Earth, satellite observations have recently been used to track and forecast health consequences and disease outbreaks as well. Satellites broadcast information, for instance, on the level of urban development, air pollution, vegetation, and water quality.
Geospatial data can provide a significantly improved way of gathering environmental exposure data and help us understand the complicated exposome. The data offers a “big picture” of the environment and when translated into policies and practices, another way to improve health.
HEDIMED has established a specific working group for the usage of Satellite data in studying the exposome’s effect in the development of immune-mediated diseases. This Satellite Working Group aims to characterize the nature of the environment where the child has lived during the in-utero period and during the first years of life using earth observation data and derived land cover and atmospheric variables. Together with top-notch partners, HEDIMED takes the advantage to build new knowledge by developing and utilizing latest technology and high-quality optical satellite data provided by the European Space Agency and other satellite operators. This data will be used to quantify the environmental exposure related to different land cover types and green areas. Quantification allows exposomic profiling and the exploration of potential associations between the dwelling environment and the risk of immune-mediated diseases. To conduct this, many steps are taken behind the scenes.
We have asked some of our experts to explain some of the steps proceeding the exploitation of satellite imagery:
Q: What is the role of Terramonitor in the project?
A: Terramonitor is a Finnish company specialized in the efficient processing of large satellite image datasets. We support medical researchers in accessing and analyzing different types of environmental data collected from space. With the help of satellite images, it is possible to study large-scale environmental exposures in urban, agricultural and natural areas. For example, the effects of urbanization and other changes in land use can be studied using remote sensing data. Satellite image datasets are often large and processing the data requires expertise in distributed cloud computing, scalable cloud storage and application programming interfaces (API).
Q: What kind of technology is used and needed for satellite pictures and data?
A: We utilize and develop new open-source software to access and process the satellite images. Most of our software is written in Python programming language and we use PostGIS database technology for storing geospatial information. Large datasets are stored in a scalable cloud storage and virtual machines are used for distributed cloud computing. The Geospatial Data Abstraction Library (GDAL) for handling different types of geospatial data is a good example of open-source software that has become a widely used standard in satellite image processing. In research it is important to use open-source software so that any process and study can be repeated and further developed by the scientific community.
Q: What actions are taken before using data in research?
A: Satellite image datasets are published in various formats and the data format required by the researcher depends on the specific study at hand. In the most common use case, geographical coordinates or street addresses are included in the cohort data and our task is to extract specific information (for example, land cover or vegetation index) from satellite data around these given locations. First, the street addresses need to be converted into geographical coordinates using a process called geocoding. Then, a time series of satellite images is acquired programmatically from these locations using the time period of the study. Depending on the data source, images may be available daily, weekly, monthly or even yearly. Finally, the requested data is extracted from the satellite images by sampling a specific radius around the given coordinate points. The extracted information is stored in a CSV file which can be easily combined with the cohort data.
Q: What is the topic of your research in the project?
A: We are studying how urbanity influences the risk of immune-mediated diseases (IMDs). Our interest stems from the fact that IMDs have increased over the last few decades. At the same time, there is a shift in living environments, with urban areas rapidly expanding. Previous studies have shown that living in urban areas carries a higher risk of asthma and allergy. Vice versa, rural living seems to have a protective effect. Referencing the biodiversity hypothesis, this might be due to people encountering a more enriched microbial environment in rural areas, which promotes proper immune system maturation during childhood.
Our project focuses on early life exposure, i.e., the birth address. First, we identify all individuals born in Denmark in the last few decades. Next, we characterise the living environment around address points using satellite data. Then, we follow them through time utilising the nationwide Danish health registries to see if they develop IMDs.
Q: Which results do you expect?
A: Our preliminary results confirm the association between urbanity and increased risk of asthma and allergy. We are currently running analyses on a population of more than one million individuals and looking toward more disease endpoints, such as type 1 diabetes and celiac disease. We expect some diseases to follow the same trend as asthma and allergy. However, we might also see differences in urbanity and disease patterns; this can be due to, in part, regional differences of how the healthcare system is utilised. Furthermore, certain IMDs, such as type 1 diabetes, are mediated through different immunological pathways than, for example, asthma. Therefore, the relationship between city dwelling and other IMDs might not be as apparent.
Q: What kind of implications could your research have?
A: Our study is observational. It cannot tell us about a causal relationship between exposure and outcome. However, our project supports the biodiversity hypothesis, stating that growing up in natural environments has beneficial effects. Therefore, this project has direct potential for policy changes, i.e., city planning, that can mitigate biodiversity loss and, at the same time, benefit public health.
Q: You have used satellite data in your recent study. What is it about and how was it conducted?
A: We analyzed land-cover types around the homes of children who participated in a large birth cohort study in Finland. Part of the children develop type 1 diabetes during the study. We observed that exposure to agricultural environment during the first year of life was associated with lower risk of developing diabetes, and lower risk of initiation of the autoimmune process that eventually leads to clinical type 1 diabetes. This was the first study of this kind in type 1 diabetes, and we are quite excited about the result. It is in line with previous studies on allergic diseases and with the idea that IMDs share some common immunological mechanisms. We also discovered that snow cover blocks the transmission of environmental microbial diversity inside homes. All this guides us now to go deeper to study which elements in the environment could mediate this effect. Our hypothesis is that exposure to environmental microbiome is important in early children since it can improve the regulation of the immune system. Satellite data together with intelligent data analysis tools will help tremendously in this mission.