Imprimir Republish


A more vulnerable Brazil in the 21st century

Projections point to increased risk of natural disasters such as floods, landslides and extreme drought in the coming decades

Marcello Casal Jr /Agência Brasil Vicinity of the Sobradinho power plant, Bahia: droughts are expected to affect other parts of the country in the coming decadesMarcello Casal Jr /Agência Brasil

Since it lies outside the pathway frequented by major hurricanes and has no active volcanos or inhabited areas subject to strong earthquakes, Brazil is not listed among the countries most susceptible to natural disasters. It is ranked 123rd in a world index of countries most vulnerable to cataclysmic events. But this image of a safe place protected from the whims of Mother Nature and geological shocks should be seen in perspective. Here in Brazil, about 85% of disasters are caused by three types of events: flash floods, landslides and prolonged drought. These phenomena are relatively frequent in tropical areas, and their effects can be largely attenuated by government policies aimed at damage mitigation. Over the past five decades, more than 10,225 Brazilians have died in natural disasters, most of them in floods or hillside collapses. Long-lasting droughts such as those commonly seen in the Northeast, however, are the kind of event that causes the most fatalities in Brazil (see Pesquisa FAPESP nº Issue 241).

Two studies based on climate simulations conducted by Brazilian researchers indicate that the risk of occurrence of these three types of disaster, linked to excessive or too little water, will increase by the end of this century in most of the areas already affected by these phenomena. The researchers also point out that new parts of Brazil, generally adjacent to regions currently affected by such occurrences, are likely to become areas of significant risk for these same problems. “The impacts tend to be greater in the future due to climate change, the growth of city boundaries and populations, and occupation of more high-risk areas,” says José A. Marengo, head of the Division of Integrated Research and Development Products at the Natural Disaster Surveillance & Early Warning Center (CEMADEN), an agency of the Ministry of Science, Technology, Innovation and Communications (MCTIC), who coordinated the climate simulations. Some of the results of these projections have been circulated at conferences and in reports such as the federal document sent in April 2016 to the United Nations Framework Convention on Climate Change (UNFCCC). The report serves as a tool to provide guidelines for the strategies presented in the recently created National Plan for Adaptation to Climate Change. But more-detailed data from the simulations will appear in a scientific paper accepted for publication in the journal Natural Hazards, as well as in papers submitted to other publications.

Expanding droughts
According to the two studies, severe droughts—a public catastrophe today nearly always associated with places in the Northeast—are expected to intensify in the West and part of eastern Amazonia, as well as in the Central West, including around Brasília, parts of the southeastern states and even in the South. “Although part of the Northeast is naturally more arid, climate is not the only cause of drought,” says civil engineer Pedro Ivo Camarinha, a researcher at CEMADEN. “The region’s vulnerability is also due to a number of problems of a socioeconomic nature, related to land use and low capacity for adapting to the impacts of climate change.” The lack of government policies that focus specifically on the drought months, low education levels and scarce resources are some of the determining factors that the authors cite for increasing exposure of significant portions of the Northeast to future droughts.

Vulnerability to heavy rains and flooding will trend 30% higher in the three southern states, southern Mato Grosso and much of the coastal Northeast, according to a scenario projected for 2100 by climate simulations. In São Paulo, Brazil’s most populous state, an uptick in flash floods that occur within minutes of torrential rains is likely to be more modest, at around 10%, but still significant. Flood vulnerability in central Brazil is expected to decrease, in part because the projections indicate lower rainfall (and more droughts) in much of the region. “The models diverge on the future rainfall pattern in western Amazonia,” explains Marengo, who conducted the studies partly under the auspices of a FAPESP thematic project. “One study points to a significant increase in flood frequency, while the other indicates a scenario characterized by stability or a slight increase in floods.”

The pattern of landslides that are associated with heavy or days-long rains is expected to largely follow the same trends seen with floods, although at a more moderate rate of increase. Hillside collapses are expected to increase by 3% to 15% in places already affected by this type of phenomenon. The negative spotlight falls on Brazil’s southernmost reaches. Increasing portions of the states of Rio Grande do Sul, Santa Catarina and Paraná will likely be subject to landslides, encompassing much of the land area in these states by 2100. The mountainous region in the Southeast along the border between the states of São Paulo, Rio de Janeiro and Minas Gerais is expected to become even more vulnerable to this type of disaster. “We urgently need to implement government policies in the regions must vulnerable to flooding and landslides,” says geographer Nathan Debortoli, a coauthor of the studies who is currently doing postdoctoral research at Canada’s McGill University. “Greater exposure to climate change events could make survival unfeasible in some areas of the country.”

Arnaldo Alves / ANPr Flood of 2014 in União da Vitória (Santa Catarina): South expected to see more floodsArnaldo Alves / ANPr

To generate the projections of future disaster risk, the researchers used two global climate models, HadGEM2 ES, developed by the Hadley Centre in England, and Miroc5, created by the Japanese Meteorological Center. Along with these two models, they ran the regional-scale Eta model developed by Brazil’s National Institute for Space Research (INPE). Working in this manner, the authors were able to evaluate the predominant future climate patterns associated with the occurrence of natural disasters in areas measuring at least 400 square kilometers—each square having 20-kilometer sides.

More convergences than divergences
The results provided by the two climate models are similar for about 80% of the Brazilian territory—a finding that lends weight to the projections. The English model has been used for more than 10 years in simulations conducted by Brazilian climatologists, who have gained considerable experience with it. The Japanese model is now entering into more frequent use. There are, however, a few divergences in the long-term simulations generated by the two models. For example, the list of the 100 municipalities most vulnerable to drought episodes in the next three decades provided by the HadGEM2 ES simulations differs from the one generated by Miroc5. The cities at highest risk, according to the Japanese model, are in four states in the Northeast: Rio Grande do Norte, Paraíba, Pernambuco and Alagoas. The English model names cities that are mostly in other northeastern states and in the Central-West region and northern Minas Gerais State. “With the exception of these extreme examples, the projections generated by the two models largely coincide,” says Camarinha. With respect to hydrological phenomena, the most significant discrepancy relates to rainfall patterns in Amazonia, especially in the western states of the North (Acre, Amazonas and Rondônia). HadGEM2 ES projects more rainfall—and thus a higher risk of floods and landslides—while Miroc5 predicts less. “Predicting rainfall in Amazonia is still a challenge for the models,” Marengo comments.

In order to quantify the future risk of natural disasters in an area, the simulations need to include not just climate information, but also a set of local data, such as the economic, social and environmental conditions of the more than 5,500 Brazilian municipalities and their people. Once calculations have been completed, each area is classified into one of five vulnerability levels: very low, low, medium, high and very high. “The model chosen, the quality of the data from each city and the weight given to each variable all affect the final index,” explains Camarinha.

The weight of humans
In addition to natural susceptibility to droughts, floods, landslides and other disasters, human activity holds considerable weight in turning what could be a minor problem into a catastrophe. The researchers estimate that one-third of the impact of landslides and half of flood damage could be avoided by altering human practices related to land occupation and improvements in the socioeconomic conditions of the population in areas at risk.

Substandard housing in ill-suited locations near hillsides or flood-prone areas; poor infrastructure, such as roads or paths that hinder easy access to highly vulnerable areas; lack of a functioning civil defense; overcrowded, impermeable cities that do not channel away rainwater—all of these unnatural factors related to human culture can influence the final outcome of a risk situation. “Even day-to-day habits, such as not littering the street, and the level of solidarity and social cohesion of a population can at least mitigate the impacts of a disaster,” notes geographer Lucí Hidalgo Nunes of the Geosciences Institute at the University of Campinas (IG-Unicamp). “Obviously, there are natural disasters so violent—such as the major earthquakes in Japan—that not even an extremely well-prepared population can avoid them. But recovery in better-organized countries is much faster.”

In their work, the researchers used RCP 8.5, an end-of-century global scenario that is relatively pessimistic but quite plausible. It is used in the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC). This scenario is characterized by large temperature increases, intensified rainfall and extreme droughts. In the case of Brazil, the projections indicate that the country is likely to be at least 3ºC warmer by the end of the century, and that rainfall amounts could increase as much as 30% in the South-Southeast and decrease up to 40% in the North-Northeast. Climate change is expected to cause greater frequency of extreme events, which could be manifested in different ways: prolonged droughts, temperature spikes, stronger storms, rainfall lasting several days, stronger undertows. These occurrences increase the risk of disasters. “To offer an example, it’s not just a question of the amount of rain that falls in a particular place,” Marengo explains. “Sometimes the amount is the same, but the distribution of the rain over time changes, and this can create more disasters.” In a city like São Paulo, a 50-millimeter rainfall over three or four days would likely not cause damage. But if the rainfall is concentrated into just one afternoon, there will probably be flooding.

To test the reliability of the vulnerability index, the Brazilian researchers compared the results obtained by the models against actual disaster records from the recent past (1960 to 1990), as compiled by the Atlas brasileiro de desastres naturais [Brazilian Atlas of Natural Disasters]. This method made it possible to get a good idea of whether the models were indeed useful in predicting the areas where floods, landslides and droughts would occur in Brazil during the last few decades. The atlas data also served as a point of comparison, providing a present-day baseline for quantifying the future increase or decrease in an area’s vulnerability to disasters. For drought, the Miroc5 simulations proved to be generally more reliable in most of the Brazilian territory. In the case of floods and landslides, HadGEM2 ES provided more-accurate predictions for subtropical and mountainous areas in the South and Southeast, and Miroc5 was more accurate for the rest of the country. Amazonia, as indicated earlier, was a focus of disagreement.

A study using methodology similar to that employed in the studies done by Marengo and his colleagues—but focused only on the present, without projections of increased or decreased future risk—was published in the International Journal of Disaster Risk Reduction in April 2016. In collaboration with German researchers, geographer Lutiane Queiroz de Almeida of the Federal University of Rio Grande do Norte (UFRN) calculated a set of indicators to identify the risk of natural disasters in each Brazilian municipality. Known as the DRIB Index (Disaster Risk Indicators in Brazil), it is an adaptation of the work done on a global scale by the United Nations University and European institutions (see map and text on pages 22 and 23). In addition to taking into account data on the risk of droughts, floods and landslides, the DRIB includes exposure of coastal municipalities to rising sea levels. For this type of problem, the cities identified as being at greatest risk were Vila Velha and Vitória in the state of Espírito Santo, Santos (São Paulo) and Salvador (Bahia).

Almeida produced vulnerability indicators for the principal disaster types throughout the national territory and a final number, the DRIB, that would indicate a locale’s overall risk for extreme events. He called attention to the classification of practically the entire territory of Amazonas and Acre and half of the state of Pará as areas at very high risk, exposed to floods and having socially vulnerable populations. Of the 20 municipalities that performed poorly on the DRIB Index, 12 are in the North. The others are in the Northeast (six) and Southeast (two). “These municipalities have small populations of 3,000 to 25,000, high exposure to disasters and low adaptive capacity,” comments the UFRN geographer. “The study points out that only 20% of Brazilian municipalities are adequately prepared to mitigate impacts and react immediately to extreme events.” Generally speaking, this is characteristic of the South and Southeast regions.

Marlene Bergamo / Folhapress Landslide in Nova Friburgo (Rio de Janeiro) in 2011: high vulnerability to disastersMarlene Bergamo / Folhapress

Recurring tragedies
Man’s interest in natural catastrophes was aroused well before the present-day discussions about climate change. Disasters have been a tragic chapter in the history of humankind from time immemorial. The mythical global deluge, cited as divine punishment and said to have ended life on Earth except for the people and animals aboard Noah’s Ark, is a narrative found in Genesis, the first book of the Old Testament, as well as in the Tanakh, Judaism’s set of sacred texts. Alleged gigantic, catastrophic floods before and after the publication of Genesis appear in accounts from several cultures throughout the ages, from the ancient Mesopotamians and Greeks to the Maya of Central America and the Vikings. The ancient Roman cities of Pompeii and Herculaneum were buried under lava that spewed from Mount Vesuvius in the famous eruption of 79 AD, killing an estimated 2,000 people. Seventeen years earlier, that same Campania region in Italy had been hit by a minor-magnitude earthquake. “We usually say that, if a disaster has occurred in a particular place, it will happen again someday,” Nunes says.

Assessment of impacts and vulnerability to climate change in Brazil and strategies for adaptation option (nº 2008/58161-1); Grant Mechanism Research Grant – Research Program on Global Climate Change – Thematic (FAPESP/CNPq – Pronex Agreement); Principal Investigator José A. Marengo (CEMADEN); Investment R$812,135.64.

Scientific articles
DEBORTOLI, N. S et al. An index of Brazil’s vulnerability to expected increases in natural flash flooding and landslide disasters in the context of climate change. Natural Hazards. In production.
ALMEIDA, L. Q. et al. Disaster risk indicators in Brazil: A proposal based on the world risk index. International Journal of Disaster Risk Reduction. April 17, 2016.