RELIABLE Project

The goal of the RELIABLE project is to develop a geographic public dashboard with real-time and high-spatial-resolution information on health risks to building occupants during extreme weather events - heat and cold waves. This dashboard will improve current warning systems in two dimensions:

1) improve current risk prediction models by integrating new public data sources processed with machine learning algorithms;
2) increase the spatial and temporal resolution of warnings, if possible, to the statistical subsection level (BGRI).

This will be done through the collection, aggregation, processing and modelling of four data sources:

a) Historical information relating to health indicators during extreme weather events. It is important to improve the models that correlate the incidence of these events with events in the health system (people seeking health care, deaths), demographics (age) and sociodemographic indicators (more exposed and vulnerable), as well as characteristics of the building (year of construction, typology). One of the central aspects is to attribute historical information to the BGRI level, improving the existing information layers and distinguishing effects such as the exact location of death (e.g. hospital) and actual residence (e.g. retirement homes) and official/tax residence;

b) Weather forecast with high spatial resolution, where detailed information is calculated on the BRGI scale. Forecasting models will be improved by taking into account historical meteorological data and health data (mortality, morbidity and other indicators);

c) Data from the 2011 national census need to be updated, especially regarding age and sociodemographic indicators (the next census will be carried out in 2021, and data will be accessible in 2022). This will be done by updating INE data with new data sources, such as digital information from municipalities (e.g. property tax information / IMI) and satellite images.

d) The database of more than 1 million building energy performance certificates (EPC) has already been issued by ADENE to estimate indoor thermal comfort conditions. This is the main source of information, as it allows estimating in detail the experience of thermal comfort inside buildings.

e) The joint processing and analysis of all these data flows will allow the development of a new model that calculates a new indicator - similar to the ICARO and FRIESA alert indicators - to the resolution of the BGRI statistical subsection (which divides Portugal into 178 364 geographic areas), which correspond in urban environments to a street or in rural areas to houses/places. The current resolution is at the district level (18 on the mainland and 2 on the archipelagos).

f) The model and data will be available on a public panel that can be embedded on the websites of different public authorities (ADENE, INSA, municipal services) and the information will be updated every 6 hours (the minimum temporal resolution of weather forecast models). The dashboard will have an API that will allow anyone to receive direct alerts about a specific location (for example, nursing home managers or citizens who want to receive specific alerts about their parents/family members' residences).