Experiment identifies places with higher risk of flooding – 04/18/2023 – Environment

Scientists from INPE (National Institute for Space Research) propose to combine forecasting models for urban expansion, land use change and also hydrodynamics to create a methodology capable of providing geographic information that identifies the places with the highest risk of flooding in cities, including those caused by extreme rainfall.

The pioneering work is based on data from São Caetano do Sul, in the Metropolitan Region of São Paulo. It may be used by other municipalities in the construction of public policies and in decision-making to face the impacts of these phenomena, being able to avoid, in addition to the destruction of buildings and infrastructure, the death of residents.

The preliminary results of the research, funded by FAPESP through two projects (20/09215-3 and 21/11435-4), were published in Water magazine. They are part of the work of doctoral student Elton Vicente Escobar Silva, from INPE, the first author of the article.

In partnership with the federal universities of Paraíba (UFPB) and Rio Grande do Sul (UFRGS) and local bodies, the researchers “tested” the model with data from the Civil Defense of the municipality referring to the flood registered on March 10, 2019, when three people drowned and several streets in São Caetano do Sul were left with almost 2 meters (m) of water.

“I’ve been working with modeling for years, focusing on changes in land use and land cover in urban areas. I wanted to combine it with flood simulation. The opportunity came with the Elton project”, says Cláudia Maria de Almeida, Silva’s advisor to Agência FAPESP. and researcher at Inpe, where she coordinates the Cities Laboratory, focused on theoretical and application research in urban remote sensing.

And he adds: “A differential of the study is, in addition to combining hydrodynamic modeling for the urban area with the complexity of the underground rainwater drainage network, using real data to parameterize and validate the model. We combine images of very high spatial resolution and ‘deep learning’ [aprendizado profundo]. All of this is linked to ‘big data’ and ‘smart cities’ [cidades inteligentes]”.

Discussions around the concept of “smart cities” began in the mid-2010s involving technological issues, such as integrated traffic lights or bus stops with Wi-Fi. Recently, they started to include themes related to sustainability and quality of life for residents.

According to the UN (United Nations), the world population reached 8 billion people last year, with 56% living in urban areas. It is estimated that by 2050 the population will grow to 9.7 billion people, of which 6.6 billion will be in cities (about 68% of the total).

The current rate of expansion of municipalities is twice as fast as population growth. With this, the global area covered by urban settlements is expected to rise to more than 3 million square kilometers in the next three decades, which would be equivalent to the territory of India.

City planning, however, does not move at the same speed. With rampant urbanization there are, for example, changes in land use and land cover, an increase in the impermeable surface area and changes in hydrology. This scenario, combined with the greater frequency of extreme events caused by climate change, exposes municipalities to vulnerabilities, such as floods, floods and landslides during the rainy season.

data crossing

For hydrodynamic modeling, the group of researchers used the HEC-RAS software (River Analysis System of the Center for Hydrological Engineering). It is a computer program that can simulate the flow and elevation of the surface of the water, in addition to the transport of sediments.

In the analysis of the extent of areas subject to flooding, two digital terrain models (DTM) were adopted with different spatial resolutions — from 0.5 m to 5 m. The DTM is a mathematical representation of the ground surface, which can be manipulated by computer programs and is usually represented in the form of a rectangular grid, in which an elevation value is assigned to each pixel. Vegetation, buildings and other features are digitally removed. Furthermore, four different computation intervals (1s, 15s, 30s and 60s) were adopted in order to evaluate the performance of the simulation outputs.

The best results were obtained with the simulations with a spatial resolution of 5 m, which showed the flood maps with the greatest coverage of the flooded points (278 out of a total of 286 points, that is, 97.2%) in the shortest calculation times. They even mapped flooding points that were not observed by the Civil Defense or by citizens of São Caetano do Sul during the flooding.

“Our idea was to create a support methodology for decision makers. We simulated how the change in land use will be in the coming years and also what this impacts on the river flow network. From there, it is possible to make simulations with scenarios An example is to cross the millimeters of rain in a certain time interval to project what it can cause in an area of ​​the municipality. With this, managers could make decisions aiming to avoid both economic damage and lost lives”, says Silva.

The researchers highlight the need for municipalities to have up-to-date databases for this type of work, as is the case of São Caetano do Sul. “The model works and is powered by data. It is important for cities to be able to have up-to-date information banks, including records referring to extreme cases, such as major floods and inundations”, evaluates Almeida.

Intensely connected with the capital and neighboring Santo André and São Bernardo do Campo, the municipality of São Caetano do Sul has a history of floods: there were 29 occurrences between 2000 and 2022, according to the researchers.

On the other hand, it is the most sustainable city among the 5,570 in Brazil, according to the Sustainable Development Index of Cities (IDSC). And, with an estimated population of around 162,000 residents, it has 100% of households with adequate sanitary sewage, 95.4% of urban households on public roads with trees and 37% on roads with adequate urbanization (presence of manholes, sidewalks, pavement and curb), according to the Brazilian Institute of Geography and Statistics (IBGE).

The article “Assessing the Extent of Flood-Prone Areas in a South-American Megacity Using Different High Resolution DTMs” can be read here.