What would happen if extreme rains as heavy as those that fell in Rio Grande do Sul in 2024, on the coast of São Paulo in 2023, or in the mountainous region of the state of Rio de Janeiro in 2022 and 2011, hit the capital of Rio de Janeiro, a city with a population of over 6 million, at least 475,000 of whom live in areas at risk of flooding or landslides?
This is the question that guided the design and development of the project Rio 60ºC - How the city is preparing for extreme weather events. With the support of the Pulitzer Center and the Serrapilheira Institute, a multidisciplinary team made up of collaborators from Ambiental Media and the RioNowcast+Green research group (Institute of Computing/Fluminense Federal University) interviewed scientists, residents and representatives of government agencies, and delved into data, studies and reports to understand the risks that climate change presents for Rio's residents. In this first phase of the project, the focus was on one of the consequences of global warming, especially in a coastal city with a geography full of hills, rivers and lakes like Rio: increasingly intense and frequent extreme rainfall events.
To understand and warn about the risks posed by heavy rainfall, the Rio 60ºC project team worked on three different initiatives. The first was an extensive data analysis to map the regions in the city that are most vulnerable to heavy rainfall. The second was to identify and illustrate a series of measures to be taken by communities and, above all, by public authorities to reduce risks of disaster. Finally, contributing reporters went into the field and interviewed residents and experts to find out how climate change is already impacting people's lives, particularly working class people.
In the following paragraphs, we detail the techniques and methods used in this first initiative to map vulnerability to heavy rainfall in the municipality of Rio de Janeiro. It's worth noting that the aim, for now, is not to make meteorological forecasts, but to identify the potential impact of rainfall.
The methodology was based on a bibliographical survey of rates of susceptibility and vulnerability to extreme weather events in Rio de Janeiro and other parts of the world. This survey took into account work produced by government agencies - such as the Brazilian government's National Center for Monitoring and Alerts of Natural Disasters (CEMADEN) and the Pereira Passos Institute, connected to Rio de Janeiro's municipal government - and research published by universities and institutes around the world.
We then searched for raw data on different environmental and social indicators for municipal Rio de Janeiro available from scientific and governmental databases. The index uses the most up-to-date data available from the Brazilian Institute of Geography and Statistics (IBGE), the Geological Survey of Brazil and Data.rio, the open data portal of the city of Rio de Janeiro.
The data was organized into two layers of main indicators: environmental susceptibility and socioeconomic vulnerability.
To standardize the format of these two resulting layers, the geospatial data was converted into a raster format, based on the Digital Elevation Model from the Aster project. This process was conducted using a script run on the Google Earth Engine. The results were then vectorized and made available as maps in Mapbox for spatial data visualization.
The layer of socio-environmental susceptibility was built from the Charts of Susceptibility to Gravitational Mass Movements and Flooding, developed and made available by the Brazilian Geological Service (SGB). These charts map the Brazilian municipalities most at risk and define areas of the territory with high, medium or low susceptibility to flooding or mass movements (landslides). Among the data used by the SGB are “hypsometry, slope, relief patterns, hydrological data and, occasionally, lithologies.”
The maps developed by Ambiental cover three different degrees (high, medium and low risk) on a color scale that ranges from gray to lilac.
The socioeconomic vulnerability layer was built from social and census data made available by the Pereira Passos Institute on the Data.rio website, connected to Rio de Janeiro's municipal government, and the IBGE. The indicators analyzed in order to make up this layer were:
The indicators had to be normalized in order to establish heterogeneity between the different scales of data. This calculation took into account similar calculations published in scientific journals (Santos et al., Sena et al. and Marinho et al.). Then a correlation analysis was conducted: the indicators that correlated with other indicators were eliminated and the rest were added together.
The layer has a color scale that ranges from the same gray used for susceptibility to orange. The darker the orange, the higher the level of socioeconomic vulnerability of the population living in the area.
Disasters resulting from extreme rainfall occur when there is a population living in an area that is environmentally susceptible to landslides or floods. Therefore, once the layers of socioeconomic vulnerability and environmental susceptibility had been produced, they were digitally merged using the multiplication mode.
The result of this blend is a bivariate color scale. In this case, the color red is the result of multiplying lilac and orange. It indicates where the people who are most vulnerable to disasters live.
With the layers laid out on the map, we began to analyze the distribution of the population in these areas. To do so, we considered the geographical location of private households as provided by the IBGE's National Register of Addresses for Statistical Purposes for the 2022 Census.
According to location, each address was assigned a value for the socioeconomic vulnerability index and the susceptibility to flooding or landslides index. In the case of vulnerability, the addresses were distributed into three brackets of vulnerability: high (3), medium (2) and low (1) and the values for each bracket were calibrated so that each contained the same number of addresses. For susceptibility to flooding or landslides, the SGB's original classifications were respected: high (3), medium (2) and low (1).
Finally, in order to reach at the index of vulnerability to extreme rainfall, the values of the susceptibility ranges were added to those of vulnerability, so that each address necessarily fell into one of five ranges: very low (2), low (3), medium (4), high (5) or extreme (6). For example: a household in an area of high (3) socioeconomic vulnerability and high (3) susceptibility to landslides will be classified as having extreme (6) vulnerability to heavy rainfall.
Some of the private households listed in the National Register of Addresses for Statistical Purposes (around 57,000, or 1.9%) do not correspond to any value in the socioeconomic vulnerability index. This is because the index has some gaps due to data. Therefore, the figures are underestimated.
Ambiental is dedicated to investigative journalism based on science and data. We dive into the best available science on the most relevant topics to bring you the data and evidence that make a difference in people's lives. We believe journalism is a tool for alerting and raising awareness about the climate emergency, as well as for defending Brazilian biomes and their forms of life, including human life.
Ambiental's unique vision has become a signature recognized by major awards. Among our outstanding reports is Aquazônia, through which we created the Amazon Water Impact Index (IIAA), a report that was awarded six prizes in journalism and voted the best data visualization in Brazil (Cláudio Weber Abramo Award) and Latin America, as well as one of the best in the world (by Wan-Ifra, the World Association of News Publishers).
Due to its proximity to science, Ambiental's work also serves as a source of information for other journalists and news outlets, having been replicated by some of the main press outlets in Brazil and the world, including The Guardian, Folha de São Paulo, Estadão, O Globo, Valor, National Geographic Brasil, BBC Brasil, Mongabay International and others.
Follow us on social media, subscribe to our newsletter and be sure to support Ambiental's journalism. Donate and contribute so that scientific data and evidence can reach as many people as possible in an accessible, appealing way. Our work requires especially high operating costs and we really need your support.
The RNC+Green research group is coordinated by Prof. D.Sc. Mariza Ferro of the Institute of Computing at Fluminense Federal University and is dedicated to the study and application of Artificial Intelligence (AI) models to mitigate the impacts of climate change in the city of Rio de Janeiro. One of the group's main focuses is the development of sustainable AI models for predicting extreme rainfall, using data from various sources, such as telemetry stations, rain gauges, radars and radiosondes.
In addition to seeking greater accuracy in predicting heavy rainfall a few hours in advance (nowcasting), the project prioritizes the use of computationally efficient models that require fewer energy resources and consequently reduce equivalent CO₂ emissions and water consumption, thus minimizing environmental impact. The group is also committed to disclosing the costs of training the models, promoting transparency and awareness of the importance of evaluating these aspects.
Another front of RNC+Green's research is the definition of what constitutes an extreme weather event for Rio de Janeiro, taking into account the particularities of its different regions and neighborhoods. Among the project's objectives is the development of an Extreme Rainfall Vulnerability Index (IVCE), which will integrate social, economic, urban infrastructure and environmental factors. This index will synthesize various components in order to identify, characterize and analyze the population's vulnerabilities to extreme rainfall events, thus contributing to the creation of more effective public policies.”
PRODUCTION
SUPPORT
