ARTIGO 01.pdf
Documento PDF (20.4MB)

Exibir conteúdo Baixar

                    Weather and Climate Extremes 39 (2023) 100545

Contents lists available at ScienceDirect

Weather and Climate Extremes
journal homepage: www.elsevier.com/locate/wace

Flash floods and landslides in the city of Recife, Northeast Brazil after heavy
rain on May 25–28, 2022: Causes, impacts, and disaster preparedness
J.A. Marengo a, b, E. Alcantara b, *, A.P. Cunha a, b, M. Seluchi a, C.A. Nobre b, c, G. Dolif a,
D. Goncalves a, M. Assis Dias a, L.A. Cuartas a, b, F. Bender a, A.M. Ramos d, J.R. Mantovani b,
R.C. Alvalá a, b, O.L. Moraes a, b
a

National Center for Monitoring and Early Warning of Natural Disasters (CEMADEN), São José dos Campos, Brazil
Graduate Program in Natural Disasters, UNESP/CEMADEN, São José dos Campos, Brazil
Institute for Advanced Studies, University of São Paulo (IEA/USP), São Paulo, Brazil
d
National Institute for Meteorology (INMET), Brasília, Brazil
b
c

A R T I C L E I N F O

A B S T R A C T

Keywords:
Heavy rain
Recife
Flash flood
Easterly wave disturbance
Landslides

From late May to early June 2022, 130 people died in catastrophic landslides and flash flood events triggered by
exceptionally heavy rains in the states of Pernambuco, Alagoas, and Paraíba, along the coast of Northeast Brazil.
Total rainfall in the city of Recife on May 25–30 was 551 mm, 140 mm higher than the average of the month of
May. Rain was heaviest on May 25 and 28, with 100–200 mm and 151–250 mm, respectively. This coincided
with easterly wave disturbances. May 28 saw the most rain, due to a significant cold front. Fourteen munici­
palities in the metropolitan region of Recife declared a state of emergency. According to the Civil Defense of
Pernambuco state, the rain impacted 130,000 people there. Most of the heavy precipitation fell over areas with
medium to very high geological vulnerability to landslides and extreme hydrological events. The loss of life and
substantial economic impacts in Recife caused by the heavy precipitation of May 2022 and the disasters induced
by it show that this city, like many others around the world, has limited capacity to cope with climate extremes.
Urbanization has increased population density occupying hills and slopes of the city, contributing to the problem.
To reduce the impact of such disasters, residents must be made aware of the risks of climate-related events, and
they must be encouraged to heed alerts warning of natural disasters issued by state and federal institutions.
Efficient monitoring of risk is also needed. Risk management will be viable only when everyone participates,
which requires education and cultural change.

1. Introduction
Brazil has an estimated 3,000 km2 of areas at risk for prevailing
climate-related disasters. These areas were recently determined to have
high and very high risk of landslides, floods, and flash floods. At least
825 municipalities are critically vulnerable to disasters (Alvalá et al.,
2019; Assis Dias et al., 2020). Of the 8,266,566 people most at risk,
around 75% are in areas susceptible to landslides or combined landslides
and floods (Alvalá et al., 2019; Zachariah et al., 2022). This population
faces considerable vulnerability (IBGE 2010). Recurring events in recent
years underscore how urgent it is to mitigate the disaster risk (UN Office
for Disaster Risk Reduction, 2019a, b).
Between January 1, 2013, and April 5, 2022, natural disasters caused
US$ 67 billion in damages throughout Brazil (CNM 2022). From January

to July 2022 alone, landslides, flash floods, and floods triggered by
heavy precipitation events have left hundreds dead and missing in
Brazil. The National Confederation of Municipalities (CNM 2022) esti­
mates 503 deaths due to disasters related to heavy precipitation as of
July 4, 2022. This is higher than the 290 deaths in all of 2021. From
2014 to 2018, fewer than 100 fatalities were caused by disasters trig­
gered by heavy precipitation annually. After 2019, this number jumped
to over 200 per year, with 2022 being the worst yet (as of July).
The city of Recife, capital of the state of Pernambuco, and 13 other
municipalities, comprise the metropolitan region of Recife (MRR).
Together, they form the fifth-largest population center in Brazil (Fig. 1).
The MRR has an estimated population of 4,047,088 in 2020 (IBGE and
CEMADEN, 2018), spread over a territory of 2,774 km2. The population
density is high, at 1,459 inhabitants/km2 (https://www.citypopulation.

* Corresponding author.
E-mail address: enner.alcantara@unesp.br (E. Alcantara).
https://doi.org/10.1016/j.wace.2022.100545
Received 16 September 2022; Received in revised form 24 December 2022; Accepted 27 December 2022
Available online 2 January 2023
2212-0947/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/).

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

de/en/brazil/metro/RM0003__recife/, last accessed on August 1, 2022).
Geographically, Recife is a coastal city, about 2.5–5 m above sea
level, in a region comprised of the Atlantic Rainforest Biome. The ge­
ography is characterized by its low elevation, the influence of tides, and
a water table very close to the surface. Recife has undergone a process of
urban densification and verticalization, occupying areas of banks and
hills, and channeling rivers. Together, these elements hinder adequate
municipal drainage – which becomes even more problematic with heavy
rains in a short amount of time. Floods, flash floods, and landslides then
impact Recife’s population negatively and severely, causing deaths,
temporary eviction from homes, damaged physical infrastructure, dis­
rupted urban mobility, and diseases linked to contaminated water
(ICLEI, 2020).
According to data from Recife’s Drainage Master Plan (ICLEI, 2020),
the city has 159 critical points of flooding, and 52% of its population
lives in areas that often flood on rainy days. Of all existing channels in
the city, 42% remain in a natural state, but most have had their banks

occupied by irregular constructions. The Municipal Plan of Risk
Reduction (ICLEI, 2020) maps more than 670 risk areas in the city. This
plan proposes to reduce these areas as much as possible in sectors with a
very high risk of landslides and floods. The city’s poorest live on river­
banks and alongside canals. While heavy rainfall can trigger natural
disasters, this driver is not the sole reason for them. A high population
density lives in poorly built housing, in areas of risk, and this is also a key
part of the problem (Assis Dias et al., 2020).
Between the last week of May and early June 2022, 130 people died
in landslides and flash floods in the states of Pernambuco, Alagoas, and
Paraíba, all located along the coast of Northeast Brazil (ENEB). Excep­
tionally heavy rains in the area, combined with inadequate protective
measures, caused these tragedies. As reported in the news and by
meteorological agencies of Brazil and Pernambuco itself, in less than 24
h on May 27 and 28, the MRR received more than 70% of the rain that
usually falls in all of May. Ordinarily, May is the third-rainiest month,
with an average of 311.8 mm (INMET). The impact of this heavy rain

Fig. 1. (a) Location of study site in Brazil; (b) digital elevation model for the study area; (c) land use and land cover for the metropolitan region of Recife (Per­
nambuco State, Brazil) and the CHIRPS sampling spot used to extract rainfall data from 1981 to May 2022; (d) land use and land cover (LULC) area in km2; and (e)
LULC in %.
2

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

was catastrophic. It was Recife’s worst natural disaster since 1966, when
175 people died (Almeida da Silva and Mandú, 2020). This extreme
rainfall accumulation followed a week of very heavy rain that intensified
on May 25 and 28 over several parts of ENEB, including the states of
Pernambuco, Sergipe, Alagoas, Rio Grande do Norte, and Paraíba. The
results were landslides and widespread floods across these regions.
These events displaced 39,285 people in Alagoas, and 4,200 people
across 24 municipalities in Pernambuco, as reported by the state and
federal Civil Defense (Flood List- https://floodlist.com/america/brazil
-floods-july-2022-alagoas-pernambuco-riograndedonorte, last accessed
on August 1, 2022). The densely populated MRR was hit particularly
hard, with impacts mostly in vulnerable areas near hillsides. Outside of
the urban center, another 80 municipalities across Pernambuco and
Alagoas declared a state of emergency. As we investigate in later sec­
tions, the total rainfall in Recife over only five days was 551 mm (May
25–30), 140 mm above May’s monthly climatology. In northern Per­
nambuco and Paraiba states, floods were also reported. In Alagoas, au­
thorities declared a state of emergency after recording the highest
rainfall totals since 1935, when rainfall monitoring began.
The Brazilian Panel on Climate Change (PBMC, 2016) ranks Recife in
16th place, of all the cities in the world most vulnerable to climate
change. Considering these concerns, this study has several objectives.
We describe the rainfall climatology of ENEB and rigorously analyze
meteorological drivers that caused May’s long and intense rains. The
Easterly Wave Disturbances (EWD) events in the last week of May 2022
in Recife are particularly important. We map the extreme precipitation
events and the areas affected in high resolution. These maps of land­
slides and floods show that the highest number of deaths occurred on
hillsides. The maps indicate tendencies in heavy precipitation, and how
the events of 2022 compared to historical records. We simulate and map
the inundations in the study area resulting from the heavy precipitation
events of May 23–29, 2022. We question why more than a hundred
people died, despite precise early warnings of the risk of severe natural
disasters having been issued prior to May 25. Did the population or
municipal governments fail to perceive the risk of natural disasters? Is
this why residents in areas of risk ignored the alerts? In practice, cor­
recting the strategy used to prepare for disaster will require an in-loco
investigation involving civil defense, communities, and local govern­
ments. This is a matter of ongoing work.

2.1. Precipitation data
Station rainfall data comes from various sources: INMET, APAC, and
INPE. For studies on rainfall extremes, we use the Recife-INMET and
Recife-APAC stations. While the former has data up to 2020, the later
shows data up to July 2022.
Additional data comes from the CEMADEN Environmental Obser­
vational Network. Among other facilities, this network includes 3,500
telemetric rain gauges, and 301 hydrological stations equipped with
pluviometers, water level sensors, and cameras for all Brazil. CEMADEN
designed its observational network specifically to monitor conditions in
areas of risk. This is quite different from conventional observational
meteorological networks for weather and hydrological monitoring. Most
of CEMADEN’s rain gauges are located directly in areas of risk, in the
vicinity of unstable slopes or in critical watersheds for flash floods. This
dataset is available from 2013 onward. However, due to maintenance
and logistical problems, there are some observational gaps. This data is
freely available from http://www2.cemaden.gov.br/mapainterativo
and additional information on the CEMADEN network is provided by
Marchezini et al. (2020) and Espinoza et al. (2021).
For more comprehensive analysis of rainfall trends in the state of
Pernambuco, we use the Climate Hazards Group InfraRed Precipitation
with Station data (CHIRPS) dataset (Funk et al., 2015). CHIRPS is a
rainfall product available at daily to annual time scales with a spatial
resolution of 0.05◦x0.05◦ , covering 1981 to the present. The dataset
includes satellite imagery and rain gauge data to create gridded rainfall
time series. The new dataset of CHIRPS, Version 2.0, has stations
worldwide, including more than 11,000 in Brazil alone. This new
dataset is available at https://data.chc.ucsb.edu/products/CHIRPS-2.0/
. Studies of other regions in Brazil have applied the CHIRPS-2.0 dataset
(Marengo et al., 2021, and references quoted therein).
2.2. Atmospheric circulation and SST fields
We explore upper and low-level circulation changes in South
America that lead to the occurrence of Easterly Wave Disturbances
(EWD) and heavy precipitation on the eastern coast of Northeast Brazil.
We focus on the last week of May 2022, when EWDs occurred, and when
heavy precipitation hit the Recife metropolitan region (MRR) and
nearby cities. We analyze the period of May 25–28, 2022 for 850- and
500-hPa vertical-velocity anomalies, and May 25 and 28 for 500- and
850-hPa zonal-wind and specific-humidity anomalies. We select these
parameters to observe lower- and middle-level circulation and moisture
transport along ENEB. All atmospheric fields are from the NCEP/NCAR
reanalysis. We calculate anomalies relative to the 1981–2010 long-term
mean. Sea surface temperature (SST) data is from the NOAA website:
https://psl.noaa.gov.

2. Data and methodology
The study uses a wide range of data sources, from observational
(rainfall data from various networks) as well as gridded data sets; NCEPNCAR reanalysis for atmospheric circulation fields; indices of rainfall
extremes; and sea surface temperatures (SST and GOES-18 imagery) to
observe the propagation of easterly wave disturbances (EWD) during the
rainfall event.
As auxiliary information, and for qualitative analysis that corrobo­
rates the impacts of observed rainfall extremes on floods and landslides,
we include information from the newspapers on the web and other social
media. In addition, we include information from the website of Flood
List (https://floodlist.com); this site provides updated flood-related in­
formation worldwide. Its information is drawn from government and
international agencies, describing the magnitude of disasters,
geographical coverage, and impact information, as well as damages and
casualties. We also review the weather reports from the National Insti­
tute for Meteorology of Brazil (INMET), Pernambuco’s Water and
Climate Agency (APAC), Brazil’s National Institute for Space Research
(INPE), the National Center for Monitoring and Early Warning of Nat­
ural Disasters (CEMADEN), and the National Center for Risk and
Disaster Management (CENAD), for the month of May 2022 to check on
station rainfall amounts and related weather patterns.

2.3. Other datasets
To see the development of convection during EWDs, we use infrared
(IR) images from GOES-16 for May 23 to 28, 2022 at 900 UTC from
CPTEC/INPE. The weather map for May 23 is from the National Centers
for Environmental Prediction/National Weather Service/NOAA/U.S.
Department of Commerce, 2015): NCEP GFS 0.25.
2.4. Indices of rainfall extremes and test of significance
We use the rainfall indices for extremes developed Frich et al. (2002):
R10, R50, R100, and R200. These represent the number of days each
year with rainfall above 10, 50, 100, and 200 mm R5xday represents the
maximum consecutive 5-day precipitation measured each year, an
amount that could lead to flooding and landslides. We calculate these
indices using CHIRPS data for 1981–2022, along with data from the
Recife-INMET (1934–2020) and the Recife-APAC (2017–2022) weather
stations. We use the Mann-Kendall (MK) nonparametric trend test
3

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

(Mann, 1945; Kendall, 1975; Yue et al., 2002) to identify statistical
significance in the indices of rainfall extremes.

converting them to a geometrical (ellipsoidal). The main limitation of
this study’s model parametrization is the lack of sea-level data to cali­
brate it. Hence, the model does not take into consideration the tide effect
during the flash flood.

2.5. Flood inundation modeling
We simulate the flood events of May 24, 25, 26, 27, and 28, 2022
using Brunner’s (2016) HEC-RAS 2D model. We analyze daily rainfall
data covering 1981 to May 2022 from CHIRPS (Funk et al., 2015) using a
box and whisker plot to calibrate the model. HEC-RAS 2D has been used
worldwide (e.g., Costabile et al., 2021). The Manning coefficient, n, is
obtained using the land use and land cover (LULC) data and the n table
from Chow (1959). The LULC comes from MapBiomas for 2020 (htt
ps://mapbiomas.org/en?cama_set_language=en). The digital elevation
model (DEM) is from ALOS PALSAR available at the Alaska Satellite
Facility (ASF). We resample ALOS/PALSAR images from 30 m to 12.5 m
pixel size with orthometric altitude (EGM96 geoid model) before

3. Rainfall in Eastern Northeast Brazil (ENEB)
3.1. Climatology
The city of Recife, located in ENEB, is susceptible to extreme pre­
cipitation events. Intense rainfall episodes have been recorded prior to
its current high annual average (Wanderley et al., 2021). The most rain
ever recorded was 335.8 mm on August 8, 1970. Out of several
rainfall-driving systems, we highlight the Intertropical Convergence
Zone (ITCZ), EWDs, and coastal breezes (Kousky, 1980; Reboita et al.,
2010; Gomes et al., 2019). Of all capital cities of Northeast Brazil, Sousa

Fig. 2. Climatology of 1981–2010 average monthly rainfall in meteorological stations in ENEB: Aracaju-State of Sergipe, Natal-State of Rio Grande do Norte, RecifeState of Pernambuco, Maceió-State of Alagoas, Joao Pessoa-State of Paraíba; source: Brazilian Institute for Meteorology (INMET).
4

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

et al. (2016) note that Recife has had the most significant high daily
rainfall events for the return time of 50 years (203.8 mm) and 100 years
(224.5 mm).
The rainy season in Recife and other cities in ENEB starts in April and
ends in August (Fig. 2), with peaks in June. The region receives 2,263
mm/year (long-term mean 1981–2010). In recent years, extreme rainfall
events over the eastern coast of Northeast Brazil has caused deaths and
homelessness, as along with other social and economic costs to the re­
gion (Ávila et al., 2017; Comin et al., 2021). Extreme events are often
seen when several atmospheric systems interact. Among them, EWDs
occasionally cause heavy rains in this region (Neves et al., 2016). Fer­
reira et al. (1990) identify EWDs in the wind field, which propagates
westward over the tropical South Atlantic Ocean during the austral
autumn and winter, presenting wave configurations in the wind and
pressure fields.
Low-level convergence increases when these disturbances interact
with local circulations, causing intense rainfall on the eastern and
northern Northeast Brazil coasts. EWDs’ outpouring of rain also in­
creases as they move over warmer waters off the western tropical
Atlantic near the coast of ENEB, particularly for SSTs higher than 27 ◦ C.
EWDs often drive the high levels of rainfall that result in flash floods,
landslides, and other environmental problems (Gomes et al., 2015), as
was the case in Recife in May 2022 and Maceió in June 2010. The
Atlantic Ocean warmer than normal over a wide area also produce
temperature increase in the lower atmospheric layers. This effect allows
both greater evaporation from the ocean and, particularly, retention of a
greater amount of moisture in the lower the atmosphere, which is then
transported to the continent by southeasterly winds, favoring the
development of convection and intense rainfall.
In the first half of the last century, the literature already documented
the presence of cloud clusters in the tropical atmosphere moving from
east to west, including in the South Atlantic (Berry et al., 1945). Later
studies specific to the South Atlantic associated such cloud clusters with
EWD and revealed their relationship with high-rainfall episodes during
the rainy season in eastern NEB (Wallace, 1970; Wallace and Chang,
1972; Yamazaki and Rao, 1977).

Pontes da Silva (2011) and Gomes et al. (2019) use the ERA-Interim
reanalysis for 1989–2010 to identify an annual average of 23 EWDs
reaching the coast of ENEB, with small interannual variability. The
precipitation composition indicates that the EWDs account for 60–70%
of total rainfall during the rainy season in ENEB. On average, EWDs’
lifetime is 4–6 days, wavelength is 4307 km, and phase velocity is 9.5 m
s-1. According to these authors, EWDs can originate from the association
with four distinct types of atmospheric systems: cold fronts, convective
conglomerates from the west coast of Africa, ITCZ, and high-level
cyclonic vortices. The annual occurrence of EWDs seems to be lower
(higher) during El Niño (La Niña). EWDs are more active in the months
of the rainy season. Their composite analysis indicates strong relative
vorticity and divergence anomalies at low levels, as well as in the ver­
tical profiles of relative humidity and vertical velocity (omega). In May
2022, two EWD episodes occurred, on May 25 and 28 (CEMADEN,
INMET). According to APAC, EWDs were observed on June 1, 4, 7, 12,
16, and 23.
4. Intense rainfall in the last week of May 2022 in the MRR:
Results and discussion
4.1. Rainfall on the coast of ENEB in May 2022
Between the last week of May and early June 2022, landslides and
flash floods triggered by intense rainfall also affected Recife and the
neighboring states of Alagoas, and Paraíba, all located along the coast of
Northeast Brazil (ENEB). Fig. 3 shows intense rainfall during May
23–28, more intense on May 23 and 28, over the states of Alagoas,
Paraiba and Pernambuco. According to INMET, May 2022 was the
rainiest May since the beginning of records in 1961, with 714 mm
(climatology of 287.7 mm) in João Pessoa, the capital of the state of
Paraíba. The previous record was in May 1964, with 701.6 mm. In
Recife, it rained 783.6 mm in May 2022, and in Maceió, 579 mm
(climatology of 276.3 mm). During the EWD episodes, on May 25,
rainfall was about 180 mm above normal over Alagoas and Pernambuco
on May 28 rainfall was 120 above normal over Pernambuco and Paraiba.

Fig. 3. Observed rainfall on May 25 and 28, and accumulated rainfall for May 23–28 (top row), and rainfall anomalies for May 25 and 28 (bottom row), in mm, for
the ENEB region. Source of data: CHIRPS.
5

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

INMET and APAC reported during the last week of May 2022 the
heavy rainfall that affected the coast of ENEB. As of May 27, 2022, some
areas of the state, including Ipojuca, São Benedito do Sul, Belém de
Maria, and Martial had already seen twice their monthly average rainfall
totals. From May 1–28, 2022, the MRR registered these totals: Corrego
Jenipapo, 487.8 mm; Ibura, 430.2 mm; Alto Mandu, 519.6 mm; Dois
Irmãos, 486.6 mm; Recife/APAC, 414.7 mm; Porto, 419.8 mm; Morro da
Conceição, 428.4 mm; Jaboatão dos Guararapes/Alto Vento, 476 mm;
and Jaboatão dos Guararapes/Curado, 472.6 mm (CEMADEN).
In northern Alagoas state, at the border of Pernambuco, the munic­
ipality of Jacuípe had flooding, as levels of the Jacuípe River rose. On
May 25, the river stood at 5.11 m, below the flood triggering level of 5.3
m but still rising. Authorities of the state of Paraíba reported that the
dam of the Tanques Reservoir broke in Pocinhos on May 25, after
continued heavy rain. Local authorities reported six houses severely
damaged or destroyed, four people injured, and roads flooded in
Jaboatão dos Guararapes, south of Recife, on May 25. A total of 14
municipalities declared a state of emergency: Recife, Olinda, Jaboatão
dos Guararapes, São José da Coroa Grande, Moreno, Nazaré da Mata,
Macaparana, Cabo de Santo Agostinho, São Vicente Férrer, Paudalho,
Paulista, Goiana, Timbaúba, and Camaragibe. The Civil Defense of
Pernambuco counted 119,223 people impacted in the state. Meanwhile,
in neighboring Alagoas, the Civil Defense reported 2,102 people evac­
uated from their homes and moved to relief camps, and another 8,017
sheltered with relatives or friends.
From May 23 to 28, 2022, a considerable accumulation of more than
500 mm was recorded in the city of Recife, culminating in the natural
disaster addressed in the present study (Fig. 4). This rain, however, was
heavier on May 25 and 28, consistent with the occurrence of EWDs on
those days, with a significant contribution from the cold front associated
with the cold air incursion on May 28, 2022.
Olinda and Recife suffered from flooding where the Tejipió river
broke its banks, and heavy rain continued afterwards, particularly in
MRR. As of May 25, Civil Defense had reported damage caused by floods
and landslides in eight municipalities in the state of Pernambuco. The
state of Alagoas reported more than 250 mm of rain in 24 h on May 24,
the average amount seen in the entire month of May. Among the mu­
nicipalities most affected is Penedo, where authorities declared a state of
emergency after recording the highest rainfall since 1935, when moni­
toring began. From late May 24 to May 25, the area received 260 mm of
rain.

experienced rapid growth in urban vulnerability and exposure. Settle­
ments have proliferated in areas with limited adaptive capacity –
especially unplanned and informal settlements. The MRR shows 1,802
areas at risk for natural disasters associated to heavy rainfall, in 12
municipalities. In total, CEMADEN monitors 333 municipalities in
Northeast Brazil. This region has 294 municipalities, with 638,872 of
their residents exposed to such risks. Around 80% of those live in areas
with a high or very high risk of landslides (Assis Dias et al., 2020). The
flooding and landslides that severely impacted these areas were a direct
consequence of extremely heavy rainfall in the week of May 23 and
continuing into June.
State and federal meteorological agencies reported that heavy pre­
cipitation caused flooding and wind damage in parts of Pernambuco
state on May 23, 2022. At a more regional ENEB scale, Fig. 5 shows daily
rainfall from May 1 to Jun 21, in stations from the CEMADEN network in
MRR. As indicated by the arrows, the highest amount of rain fell in
Recife and Maceió during the EWD events of May 25 and 28. Both events
affected Recife, with the first having the biggest effect on Maceió. The
municipality of Maceió reported that from late May 24 to May 25, the
area received 260 mm of rain, higher than the previous record of 200
mm, in 1945.
Fig. 6 shows rainfall on May 25 and 28 (and anomalies) and accu­
mulated rainfall from May 23–28th, 2022. The regions most affected in
both episodes were the southern part of Recife and the northern part of
the municipality of Jaboatão dos Guararapes, which had more than 200
mm above average. Between May 24 and 29, accumulated rainfall was
482 mm in São Lourenço da Mata, 467 mm in Recife, 462 mm in
Jaboatão dos Guararapes, 448 mm in Olinda, and 429 mm in Camar­
agibe, all municipalities in the greater Recife metropolitan area. For the
municipality of Recife, the regions most vulnerable to landslides and
hydrological events near Olinda to the north and Jaboatão dos Guar­
arapes to the south (Assis Dias et al., 2020) are also those with the
highest poverty rate and population density (Souza et al., 2014). The
municipality ranks third among MRR’s most critical, in terms of its
number of landslides.
The maps in Fig. 7 show the accumulated rainfall for May 25 and 28
for the MRR (dataset from weather stations). Red circles show areas with
medium to very high geological vulnerability to landslides and extreme
hydrological events, following Assis Dias et al. (2020). Large amounts of
rain fell in areas with high geological risk: 100–200 mm on May 25 and
151–250 mm on May 28.

4.2. High-resolution mapping of intense precipitation events in Recife and
areas at risk for disasters
Recife, like other cities in Brazil and around the world, has

Fig. 5. Daily rainfall in the Metropolitan Region of Recife for May and June (up
to June 21), 2022 from CEMADEN rainfall gauges. Map inset in upper right
corner shows location of region. Arrows indicate the EWDs on May 25 and 28.
(Source: CEMADEN’s environmental network).

Fig. 4. Daily and accumulated rainfall (mm) from May 23 to 28 at the Córrego
Jenipapo meteorological station in Recife (Source: CEMADEN).
6

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

Fig. 6. Observed rainfall on May 25 and 28, and accumulated rainfall for May 23–28 (top row), and rainfall anomalies for May 25 and 28 (Sidebottom row), in mm,
for the Recife metropolitan area. Source of data: CHIRPS.

4.3. Atmospheric circulation fields

cold fronts contributed to Recife’s extreme rainfall from May 23 to 28,
2022. Measurements of atmospheric pressure fields at sea level, wind,
vorticity, and divergence in different levels of the atmosphere (1000,
925, 850, 700 and 500 hPa), from May 15 to 29 generated by the United
States weather forecast system (Global Forecast System - GFS) and
synoptic charts of the Brazilian Navy (Marinha do Brasil, 2022a) show
perturbations associated with EWDs on May 25 and 28. Fig. 10 shows
sea level pressure and wind field vorticity on May 23 at 700 hPa. A
positive vorticity line (representing an anticyclonic/counterclockwise
wind rotation, red band inside the green circle reaching the coast) east of
the NEB, aligns with an extensive positive vorticity band (indicated by
the yellow line) that extends to near a low-pressure center (indicated by
the letter “B” in red). This extensive band is associated with the rem­
nants of a cold front, and its respective low pressure. Fig. 11 also
highlights other disturbances (black ellipse) over the ocean, along the
coast of ENEB. These propagated eastward, reaching the coast of ENEB
from May 23–28 (figures not shown).
The Hovmöller diagram of vertical velocity (omega) at the 850 hPa
level (Fig. 12a) reveals the eastward propagation of the upward vertical
motion towards the ENEB coast. The rising motion (blue) moves from
about 20oW to Recife precisely on May 25 and 28, the same days that the

Fig. 8 shows satellite images of the infrared channel for May 23 to 28,
2022. On May 25 and 28, the lower temperatures represent intense
convective activity over coastal ENEB, indicative of EWD events. The
sequence of images reveals clusters of shallow clouds over the Atlantic
Ocean that propagate from east to west and develop vertically as they
approach the coast of ENEB due to the observed warm sea surface
temperature (SST).
For this same area, Fig. 9 (left and right) shows an abnormally
intense upward motion at low (850 hPa) and medium (500 hPa) levels.
These features indicate a vigorous and deep convective activity and
intense rainfall accumulation, compatible with EWDs (Gomes et al.,
2019).
In ENEB on May 25 and 28, the days with EWD, Fig. 10 shows the
intense easterly flow at lower and medium levels with the high moisture
content. May 25 shows intense easterly flow up to 500 hPa from the
tropical Atlantic into Northeast Brazil with higher moisture content
along the ENEB coast. On May 28, despite more intense easterly flow,
higher moisture content is mainly concentrated along 13oS.
Analysis of previous meteorological conditions suggests that two
7

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

Fig. 7. Areas of the metropolitan region of Recife showing monitored by CEMADEN (gray color) and levels of vulnerability to landslides and hydrological events,
summarized by red circles (left) [Assis Dias et al. (2020)]. Accumulated rainfall (in mm) for the May 25 and 28, 2022 EWDs in this area (center and right). Rainfall
color scale is shown in the upper left of last two maps. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of
this article.)

Fig. 8. Infrared (IR) images from GOES-16 for May 23 to 28, 2022 for ENEB at 900 UTC for Northeast Brazil (Source: CPTEC/INPE).

EWD and the highest volumes of rain occurred. This propagation also
appears at other levels, and it reveals the origin of the disturbance at low
levels and the vertical development to higher levels as convection

deepens. The phase velocity is 10 m/s for the wave that reached Recife
on May 25. Gomes et al. (2019), Pontes da Silva (2011), Diedhiou et al.
(2010), Yamazaki and Rao (1977), and others find similar phase
8

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

Fig. 9. 850 and 500 hPa Ω (Pa/s) anomaly for May 25–28, 2022 in South America, relative to 1981–2010. Color scale is on the right side of each panel. (For
interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 10. 850 and 500 hPa zonal wind and 850 hPa specific humidity anomaly for May 25 (upper panel) and May 28 (lower panel), 2022 in Northeast Brazil, relative
to 1981–2010. Color scale is on the right side of each panel. Arrows indicate easterly flow and “+” shows regions with higher moisture content. (For interpretation of
the references to color in this figure legend, the reader is referred to the Web version of this article.)

coast of ENEB. Through the vertical gradient of potential equivalent
adiabatic temperature, the flow increases thermodynamic instability,
and therefore the potential for precipitation. Convection is particularly
strong when it passes over water above 28.5 ◦ C. The TSM anomalies
were not statistically significant (below 1 ◦ C), and this may not relate to
onset nor to development of convection. The emergence appears to be
exclusively atmospheric. Analysis shows that EWDs did not appear in the
20o W of the Hovmöller. They appeared to the southeast, with an origin
that seems unrelated to intense SST anomalies. EWDs have increased
convection as they enter a region with warmer waters, with SST above
28.5 ◦ C.

velocity. On the other hand, the disturbance that hit Recife on May 28
had a mean phase velocity of approximately 4.3 m/s (20◦ longitude in 6
days). This speed is below the 6–14 m/s found in the literature (Gomes
et al., 2019). The propagation speed of the disturbance that arrived in
Recife on May 28 is not constant. It is lowest between May 26 and 27
(approximately 2.6 m/s) and highest between May 27 and 28 (approx­
imately 9 m/s). The fact that the upward motion intensifies when it
arrives at the coast may be an effect of the increase in convergence at
low levels (due to breeze circulation and friction on the surface of the
continent).
The Hovmöller figure of SST (Fig. 12b) shows that convection growth
appears to be related to the SST values in the Atlantic Ocean, with a
greater flow of vapor to the lower layers of the atmosphere along the
9

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

Fig. 11. Mean sea level pressure (black lines) and
relative vorticity at 700 hPa level (shaded - multi­
plied by 105), on May 23, 2022 at 1200 UTC. The
green arrow indicates the trajectory of low-pressure
disturbances. The red arrow indicates the direction
of propagation. The black and green ellipses, and the
yellow line, indicate the presence of disturbances.
The letter “B” in red indicates the position of the lowpressure center. Source: National Centers for Envi­
ronmental Prediction/National Weather Service/
NOAA/U.S. Department of Commerce, 2015): NCEP
GFS 0.25 Degree Global Forecast Grids Historical
Archive. Research Data Archive at the National Cen­
ter for Atmospheric Research, Computational and
Information Systems Laboratory, Boulder, CO
[Available online at https://doi.org/10.5065/D65
D8PWK.]. Accessed 10 Jul 2022

Fig. 12. a) Hovmöller diagram of vertical velocity (omega) at the 850 hPa level; blue colors show rising motion and arrows represent the movement of the EWD
episodes of May 25 and 28. b) Hovmöller diagram of SST. Both figures are for May 21 to 29 between 0o and 40o W covering the Atlantic and the coast of Recife. (For
interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

4.4. Rainfall extreme tendencies

Station. However, stations from CEMADEN’s networks show that on
May 28, three fluviometric stations in Jaguatão dos Guararapes showed
daily rainfall of 216.1, 217.6, and 217.6 mm, and another four in the city
of Recife showed 197.8, 212.8, 204.6, and 198.9 mm. Since these sta­
tions have few years of data, we cannot perform any trend analysis, nor
compare with previous decades. On May 25, no occurrences of precip­
itation above 200 mm were detected in the MRR, but 36 stations from
the CEMADEN network measured precipitation between 100 and 150
mm on that day.
We use the RX5day, R50, and R100 mm indices. We also analyze the
trend of extreme rainfall in the study region at a 5% significant level. We
calculate the anomalies of R50 and R100 related to the 2022 extreme
events, considering historical rainfall data. Fig. 14a–d shows the
monthly trends of total rainfall for the rainy months of ENEB. June
showed the most extensive area with a significantly increasing rainfall
trend, especially in the states of Pernambuco and Alagoas. Also, the
maximum five consecutive days of total precipitation (RX5 day) has
been increasing in the region (Fig. 14e) for the last 40 years. The extreme

As seen in previous sections, rainfall in the 2022 rainy season (until
June) was well above average, higher than in 2020 or 2021, in the Recife
meteorological station from APAC. Our analysis extends to the end of
July 2022. Rainfall volumes for May and June 2022 are high (Fig. 13a,
c), compared to 2020 and 2021, and to the climatology. The tendency
detected at INMET’s Recife meteorological station from 1969 to 2020
shows non-statistically significant downward trends in the number of
days with 20 mm in 52 years. The previous record was on August 11,
1970, with 336.7 mm. The number of days with daily rainfall totals of
and above 50 and 100 mm does not change over the decades (Fig. 12b).
In 2022, the rainy season from April–July accumulated 1612 mm of
rain, beating the 1981–2010 climatology of 1347.1 mm. For 2020, 2021,
and 2022, the Recife APAC meteorological station shows an increase in
the number of days with daily rainfall totals above 20 and 50 mm, and
three days with rainfall above 100 mm in 2021 and 2022. Fig. 12d does
not show days with rainfall of 200 mm or above for the Recife APAC
10

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

Fig. 13. a) Monthly rainfall for the Recife APAC meteorological station for 2020, 2021, and 2022 (until July 2022); b) Number of days with daily rainfall totals equal
to or above 20, 50, 100, and 200 mm at the decadal level for Recife INMET meteorological station from 1969 to 2020; c) Accumulated total and rainy season
(April–July) rainfall for 2020, 2021, and 2022 (until July 2022) for the Recife APAC meteorological station; d) Number of days with rainfall equal to or above 20, 50,
100, and 200 mm for the Recife APAC meteorological station for 2020, 2021, and 2022 (until July 2022). Recife INMET meteorological station (8.05oS, 34.95oW),
Recife APAC (8.04oS, 34.87oW).

rainfall event of 2022 contributed to a positive anomaly in the annual
count of days with rainfall exceeding 50 and 100 mm (R50 and R100).
Alagoas and Sergipe had 2–4 days of rainfall exceeding 50 mm than
usual. On the other hand, the trends analysis considering R50 and R100
indices does not indicate an increasing tendency in the last 40 years (not
shown).
The RX5 day represents the intensity (mm/day) of rainfall while the
R50 and R100 indices represent the frequency (number of cases) of
events over predefined defined thresholds (50 mm or 100 mm). Thus,
trend analyzes in Fig. 13 show that although there is no evidence that
the frequency of heavy rainfall events has increased in the ENEB, there
are signs that events have become more intense (greater volume of rain
accumulated in a short time), indicating that although random, the
events may be becoming more destructive, especially if it occurs in
densely populated regions (as the 2022 event in Recife).

slowly, increasing the magnitude of the flood.
Annual rainfall data (Fig. 15a) shows an average value from 1981 to
May 2022 of 187.2 mm, with a minimum and maximum of 15.6 and
1,237.8 mm, respectively. The highest rainfall occurred in 1994, and the
lowest in 1995. The highest average rainfall (293.0 mm) occurred in
1986. The minimum, maximum, and average rainfall from January to
May 2022 were 41.6, 625, and 247.2 mm, respectively. The average
rainfall of 2022 was higher than the average for the analyzed time series.
Taking only the month of May of each year analyzed (Fig. 15b), the
average rainfall was 11 mm, almost 17 times lower than the annual
average. For example, the minimum and maximum rainfall observed
were 3.84 (May 2001) and 266.9 mm (May 2022), respectively. The
average rainfall in May 2022 was 20.2 mm, twice as high as the average
of the entire period. Based on the rainfall time series, we simulated
spatial and temporal variation of the flood inundation in the study area
for the following days: May 24, 25, 26, 27, and 28, 2022 using the HECRAS 2D model (Fig. 16). Results show a water depth (or water height, m)
ranging from 0 to 5 m. The reference for the water depth (or water
height, m) was the DEM used in the simulation. The water height fol­
lowed the rainfall rhythm for those days. The maximum flood inunda­
tion occurred on May 28, 2022. Recife’s simulated area of flooding in
May 2022 includes the following totals: 15.27 (May 24), 40.37 (May 25)
26.01 (May 26), 20.43 (May 27), and 40.65 km2 (May 28).
Neves and Muehe (1995), using data from tide gauge measurements
obtained between 1946 and 1988, indicate a high relative rise in sea
level of 5.6 mm/year. These authors simulate the potential impact of 1 m
of sea level rise. Their results show that 60% of the impacted area would
be mangroves, but the entire business district would be directly affected.
As highlighted by Olbert et al. (2017) coastal flooding occurs when high
water levels due to tides and storm surges takes place in low-lying
coastal areas, in combination with waves. Sa’adi et al. (2022) studied

4.5. Flood inundation mapping
Intense rainfall in MRR in May 2022 increased the chance of floods
when accompanied by (a) larger astronomical tide (spring tide); or (b)
positive weather tide. The two cases may have contributed to the May
28, 2022, flooding, especially. The astronomical tide table shows a
higher tidal range on May 28, 2022, consistent with the “new moon” of
May 30, 2022 (Marinha do Brasil, 2022b). The tide heights measured at
Recife Port (8◦ 3′ 4′′ S, 34◦ 52′ 0.998′′ W) on May 28 were 2.1 m (0200 a.
m.), 0.4 m (0845 a.m.), 2.2 m (0247 p.m.), and 0.4 m (0900 p.m.). Signs
also point to a contribution from the meteorological tide: strong
southeasterly winds blew on the coast of NEB on those days (with the
second cold front contributing), which may have caused the “piling up”
of water on the coast, causing sea level rise, adding to the tidal height
(Ekman transport). Higher sea levels cause rainwater to flow more
11

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

Fig. 14. Pixel-wise Mann–Kendall trends for total monthly rainfall for (A) May, (B) June, (C) July, (D) August, and (E) Rx5day from 1981 to 2022. Positive values
represent increasing trends, and hatched areas indicate trends significant at a 5% confidence level. Anomalies of (F) R50 and (G) R100 for 2022, in number of days.

12

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

Fig. 15. Box and whisker plot for (a) annual rainfall and (b) daily rainfall for May of each year from 1981 to 2022.

the effect of the tide on flood modeling in Tinggi, Malaysia and found a
43% lower than the observed data of flood depth when tidal effect was
not considered and an average similarity of 91.4% when considered. On
the other hand, Yonehara and Kawasaki (2020) analyzed the effect of
tidal effect on flood inundation in a low-lying river basin and found that
depending on the location of the study site the rainfall contribution to
flood occurs was far important then the tide influences. The influence of
tide will be stronger near the coastal zone and weaker into the urban
zone.
From the flood inundation simulations, we estimate the area (%)
affected by flash flooding in each LULC class on May 25, 26, 27, and 28,

2022 (Fig. 17). Classifications are based on the following water depths:
0, 1.25, 2.5, 3.75, and 5 m. On May 24, 54.84 mm of rain fell; the
following day, 109.68 mm (see Fig. 1 for location data supplied by
CHIRPS). Fig. 16a shows which areas were affected by flooding in each
LULC class on May 25. On this day, 7.41% of agriculture and pasture had
water at 0 m and 5.44% by 1.25 m. Forested land was covered 3.65% at
1.25 m. Urban areas had 4.4% with 1.25 m. On May 26 (with no rain,
Fig. 17b), water at least 1.25 m deep reached 3.42%; May 27 (with no
rain, Fig. 17c) saw 4.94% of urban area was at the same level (0 m), and
2.3% at 1.25 m of water. On May 28 (with rain of 266.95 mm, Fig. 17d),
the urban area suffered flooding of 11.05%, 5.03%, 1.25%, 0.12%, and
13

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

Fig. 16. Spatial and temporal flood inundation simulation in the study area. (a) May 24, (b) May 25, (c) May 26, (d) May 27, and (e) May 28, 2022. (e) Flood
inundation occurred on May 25, 2022, Avenida Recife, Ipsep (Source: G1). (f) Flood inundation occurred on May 28, 2022, Adelmar da Costa Carvalho Stadium, Ilha
do Retiro (Source: CNN Brazil).

0.01% for 0, 1.25, 2.5, 3.75, and 5 m, respectively. Rivers and lakes
overflowed, with levels 2.5 m over their banks in 54% of their area.

Recife was 1,661,017 inhabitants. In 1966 and 1975, the deaths during
floods were mostly linked to drowning. In contrast, many flood-related
deaths in May 2022 were due to landslides. The increased population
density occupying hills and slopes of the city contributed to this prob­
lem. Possible solutions in the medium and long term include improving
drainage in sloped areas.
More than half of Recife’s population lives in highly vulnerable areas
(Assis Dias et al., 2020). As in other disasters in Brazil (Marengo and
Alves, 2012; Marengo et al., 2021; Alcântara et al., 2022), heavy pre­
cipitation events trigger natural disasters, with floods— and particularly
landslides— disproportionately affecting vulnerable communities.
Low-income neighborhoods located in vulnerable and exposed areas in
capital cities and rural areas are devastated. Thus, the impacts of such
disasters are exacerbated by the pre-existing structural vulnerability in
the region. This indicates a need to review the protocols for disaster risk
reduction and disaster management. Warnings of extreme
weather-related natural disasters and risk alerts must be stronger. When
such warnings are made, anticipatory action to protect people and
critical structure must be taken. For example, the intense rainfall and

4.6. Disaster risk in MRR: What happened in the last week of May 2022?
The tragedy in the Recife metropolitan region (MRR) in May 2022
claimed more than 100 lives. The area has experienced repeated
flooding in the past. The 1966 floods caused more deaths, and more
people affected overall. In 1975, the flooding was more chaotic, caused
more damage, and inundated MRR completely. The floods of 1966 and
1975 happened in a period when the population of Recife was consid­
erably smaller. From 1940s to 1950s, Recife had its first major de­
mographic expansion, growing from 200,00 to 500,000 inhabitants. The
population continued to grow from 524,000 to 1.4 million inhabitants
between 1950 and 2000. Urbanization occurred more by densification
than by spreading (Leão et al., 2021). The population grew from 1,006,
000 inhabitants in 1966 to 1,249,841 inhabitants in 1975, according to
the Anuário Estatístico do Brasil – 1966, and Anuário Estatístico do
Brasil – 1975 (IBGE 1966, 1975). In 2021, the estimated population of
14

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

Fig. 17. Flood inundation depth (m) in each LULC on (a) May 25, (b) May 26, (c) May 27, and (d) May 28, 2022.

assessing the disaster risk alerts information received by CEMADEN and
weather forecasts from state and federal meteorological agencies,
CENAD delivers alerts of disaster risk to civil defense offices and mu­
nicipalities. Based on these alerts, municipalities inform the population
about possible upcoming events by mobile phone (SMS), radio, TV, or
even by deploying vehicles equipped with loudspeakers. CEMADEN was
created in 2011 and monitors 1,038 Brazilian municipalities. These
cities and towns are prioritized due to their history of disasters and high
concentrations of vulnerable residents in high-risk areas. So, while
CEMADEN works with disaster risk reduction, the civil defense works
with disaster management.
In advance of the event on May 25, 2022, CEMADEN issued alerts of
the risk of hydrological events for MRR (bulletin 1938/22) at the
moderate level at 16:01 LST on May 23, and for the risk of landslides
(bulletin 1941/22) at the moderate level at 16:23 LST. At 20:25 LST, the
landslide alert was upgraded to high level, and later, at 20:25 LST, it was
further upgraded to very high. The same 1941/22 bulletin upgraded the
alert for landslides predicted for May 28 to very high at 06:13 LST. The
alert in bulletin 193/22 for hydrological risk for May 25 was updated to
high at 05:08 LST and then to very high at 13:31 LST. Figs. 5 and 7 show
the rain continuing unabated after May 23. Bulletins 1938/22 and
1941/22 can be obtained at CEMADEN (www.cemaden.gov.br).
Despite the meteorological predictions and alerts of the risk of
disaster, the population and government appear to not have been
convinced to take actions that could potentially have saved lives.
Meteorological centers cannot predict vulnerability and exposure, since
these depend on several non-environmental factors, such as communi­
cation, governance, and cultural attitudes towards messages from sci­
entific agencies. Other additional barriers can interfere with the
decision-making process needed to implement a disaster risk alert.
Regarding communication, the target population may not understand
the nature of the alerts, or the technical language they sometimes use.
For example, news media reported that people evacuated by the civil
defense were reluctant to leave their residences. Some were afraid of
looting. Others returned home to retrieve forgotten documents or pets.

floods in Bahia and Minas Gerais in Brazil during the last week of
December 2021 were adequately monitored, and early warnings issued
by CEMADEN helped minimize related damage and protect human lives
and property (Marengo et al., 2021; WMO, 2018).
Increased vulnerability of the local population to these hazards arises
from unplanned and informal urbanization in flood-prone areas and
steep hillsides. Socioeconomic, demographic, and governance factors
will likely mediate long-term impacts and recovery. In May 2022,
weather forecasts and warnings were issued by meteorological agencies
(INMET, INPE, APAC). However, they underestimated the quantity of
rain expected in MRR. A similar situation had been observed in the
heavy rainfall and related disasters triggered by such rainfall extremes in
Petropolis in February 2022 (Alcântara et al., 2022). While meteoro­
logical services issue forecasts for extreme weather, these do not
represent forecasts or early warnings of risk of climate-related natural
disasters.
Worldwide, some progress has been made in implementing multihazard early warning systems (MHEWS). As defined by the World
Meteorological Organization (WMO) (WMO, 2018), these centers
monitor meteorological and vulnerability/exposure conditions that may
lead to disasters. They can issue risk alerts for natural disasters at the
municipal level. Early warning is essential for disaster risk reduction. It
can prevent loss of life and reduce the economic and material impacts of
hazardous events. To be effective, early warning systems must actively
involve the people and communities at risk from hazards. The public
must be educated and aware of risks, receive messages and warnings
efficiently, and remain constantly prepared to take early action.
In Brazil, the National Center for Monitoring and Early Warning of
Natural Disasters from the Ministry of Science, Technology, and Inno­
vation (CEMADEN) is closest thing to an MHEWS center. This federal
institution is responsible for monitoring and issuing early warnings and
alerts at various levels of risk of landslides, floods, and flash floods
(CEMADEN-www.cemaden.go.br, Assis Dias et al., 2020). These alerts
are delivered in real-time to the Civil Defense Executive Secretary
(SEDEC) and CENAD from the Ministry of Regional Development. After
15

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545

These errors left some people in vulnerable areas when disaster hit and
killed them. These failures highlight the need to improve how alerts are
issued and communicated to the affected population. The combination
of increasing frequent extreme events and a population living in areas or
risk that may be inattentive to alerts is lethal. It results in more deaths
from climate-related disasters. Research involving those who work with
the physical aspects of disasters and social scientists is required to
improve disaster risk reduction.

CRediT author statement
Marengo, J.A.: Conceptualization, Methodology, Formal analysis,
Writing - Original Draft, Writing - Review & Editing. Alcantara, E.:
Conceptualization, Methodology, Formal analysis, Writing - Original
Draft, Writing - Review & Editing. Cunha, A.P.: Methodology, Formal
analysis, Writing - Review & Editing. Seluchi, M.: Formal analysis,
Writing - Review & Editing. Nobre, C.A.: Formal analysis, Writing Review & Editing. Dolif, G.: Formal analysis, Writing - Review &
Editing. Goncalves, D.: Formal analysis, Writing - Review & Editing.
Assis Dias, M.: Formal analysis, Writing - Review & Editing. Cuartas, L.
A.: Formal analysis, Writing - Review & Editing. Bender, F.: Formal
analysis, Writing - Review & Editing. Ramos, A.M.: Formal analysis,
Writing - Review & Editing. Mantovani, J.R.: Formal analysis, Writing Review & Editing. Alvalá, R.C.: Formal analysis, Writing - Review &
Editing. Moraes, O.L.: Formal analysis, Writing - Review & Editing.

5. Conclusions and recommendations
The heavy precipitation events in 2022 prove just how vulnerable
Recife and the other districts in MRR are to extremes of climate vari­
ability. The likelihood of more extreme events due to climate change
cannot be ignored. On May 28, 2022, almost 17% of the entire urban
area of Recife was hit by inundations. Water approximately 5 m deep
covered 0.01% of the city (according to the digital elevation model).
World Weather Attribution (Zachariah et al., 2022) notes that although
these heavy-precipitation events are atypical in ENEB, they are now
more likely to happen, in a climate warmed by human activity. Since
both events broke all previous records, it is impossible to quantify how
much more likely climate change has made them. The warming of the
planet has increased the intensity of rainfall. Rainfall events as rare as
these, had they occurred in a climate 1.2 ◦ C cooler, would likely have
been approximately one-fifth less intense (Zachariah et al., 2022).
To reduce adverse impacts of climate-related disasters, support
resource management decisions, and improve outcomes requires climate
services, end-to-end early-warning systems, and sustainable in­
vestments. The systems in place are not yet adequate.
Therefore, more investments—and more precisely targeted invest­
ments—in climate services are needed to strengthen early-warning
systems and decision support for adaptation in climate-sensitive areas.
In addition, it is essential to educate the population and governments
about the deadly risks of climate-related disasters, and to reinforce
public perception of the need to react to alerts and warnings of natural
disasters issued by the state or federal institutions. A well-implemented
plan, developed in co-production between the academic, public, and
private sectors, and with local and state governments, will certainly
improve disaster risk management. These actions must join the efficient
monitoring of the risk of disasters induced by heavy rainfall to reduce
this risk. In the end, well-coordinated initiatives aim to support
municipal governments in protecting critical urban infrastructure and
vulnerable populations from natural disasters caused by extreme
weather events. Since extreme rainfall has become more intense and
frequent across the globe, and populations are increasingly vulnerable,
the risk of disaster only rises (IPCC, 2021, 2022).
Municipalities need to create an infrastructure of material, financial
and human resources. Workers must be informed, motivated, trained,
qualified, and guided to carry out civil defense actions on a full-time
basis, since disaster can happen any day or time (CNM 2022). These
actions require a great deal of preparation from local management.
Adequate support and integration with the other entities of the Feder­
ation are necessary to maintain the actions at the local level. The many
municipal civil defense protection competencies require uninterrupted
technical and financial support from federal and state governments to
strengthen them.
In practice, risk reduction must work together with risk manage­
ment, starting with prevention and mitigation. Optimizing the prepa­
ration and response actions of the municipal civil defense and protection
systems to disasters at the local level is possible. However, risk man­
agement will only be viable when everyone participates, which requires
cultural change. The entire population, young and old alike, must un­
derstand the risks posed by those natural disasters, and remain vigilant
and attentive to the alerts that may be disseminated, for a true collab­
oration in reducing these risks.

Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Data availability
Data will be made available on request.
Acknowledgments
This work was supported by the National Institute of Science and
Technology for Climate Change Phase 2 under CNPq Grant 465501/
2014-1; FAPESP Grants 2014/50848-9; and the National Coordination
for Higher Education and Training (CAPES) Grants 88887.13640200INCT, 88881.691139/2022-01 and 88881.593660/2020-01. Addi­
tional funds come from the RED-CLIMA (Red española e iberoamericana
sobre variabilidad climática y servicios climáticos en ecosistemas ter­
restres y marinos: RED-CLIMA) Project, under Grant INCCLO0023 from
the Consejo Superior de Investigaciones Científicas LINCGLOBAL CSIC
from Spain.
References
Alcântara, E., Marengo, J.A., Mantovani, J., Londe, L., San, R.L.Y., Park, E., Lin, Y.N.,
Mendes, T., Cunha, A.P., Pampuch, L., Seluchi, M., Simões, S., Cuartas, L.A.,
Massi, K., Alvalá, R., Moraes, O., Filho, C.S., Mendes, R., Nobre, C., 2022. Deadly
disasters in southeastern south America: flash floods and landslides of February 2022
in Petrópolis, Rio de Janeiro, Nat. Hazards Earth Syst. Sci. Discuss. https://doi.org/
10.5194/nhess-2022-163 [preprint], in review, 2022.
Almeida da Silva, E., Mandú, T.B., 2020. Variabilidade da intensidade da precipitação no
período chuvoso em Recife-PE, SILVA. In: Anais do I Congresso Brasileiro
Interdisciplinar em Ciência e Tecnologia. Anais...Diamantina (MG) Online, 2020.
https://doi.org/10.29327/121206.1-35. Available from. https://www.even3.com.
br/anais/icobicet2020/263893-VARIABILIDADE-DA-INTENSIDADE-DA-PRECIPI
TACAO-NO-PERIODO-CHUVOSO-EM-RECIFE-PE.
Alvalá, R.C.S., Assis Dias, Saito, S.M., Stenner, C., Franco, C., Amadeu, P., Ribeiro, J., de
Moraes Santana, Nobre, C.A., 2019. Mapping characteristics of at-risk population to
disasters in the context of Brazilian early warning system. Int. J. Disaster Risk Reduc.
41, 101326. https://doi.org/10.1016/j.ijdrr.2019.101326.
Assis Dias, M., Saito, S.M., Alvalá, R.C.S., Seluchi, M.E., Bernardes, T., Camarinha, P.I.M.,
Stenne, C., Nobre, C.A., 2020. Vulnerability index related to populations at-risk for
landslides in the Brazilian Early Warning System (BEWS). Int. J. Disaster Risk Reduc.
49, 101742 https://doi.org/10.1016/j.ijdrr.2020.101742, 2020.
Ávila, A., Justino, F., Wilson, A., Bromwich, D., Amorim, D., 2017. Recent precipitation
trends, flash floods and landslides in southern Brazil. Environ. Res. Lett. 11, 1–13.
https://doi.org/10.1088/1748-9326/11/11/114029.
Berry, F.A., Bollay, E., Norman, R. Beers, 1945. Handbook of Meteorology. McGraw-Hill
Book Company, p. 1068. https://ui.adsabs.harvard.edu/abs/1945hame.book.....B.
Brunner, G.W., 2016. HEC-RAS River Analysis System 2D Modeling User’s Manual
(Online). State of California. US Army Corps of Engineers—Hydrologic Engineering
Center, pp. 1–171. Available: https://www.hec.usace.army.mil/software/h
ec-ras/documentation/HEC-RAS 5.0 Users Manual.pdf.
Chow, V.T., 1959. Open Channel Hydraulics. McGraw-Hill, New York, p. 680p. https://
doi.org/10.1126/science.131.3408.1215.b.

16

J.A. Marengo et al.

Weather and Climate Extremes 39 (2023) 100545
Capdeville.. (Org.). Direito Ambiental e Geografia, 1ed, pp. 307–354. Rio de Janeiro:
Lumen Juris, 2020, v. 1.
Marengo, J.A., Cunha, A.P., Cuartas, L.A., Deusdará Leal, Broedel, E., Seluchi, M.E.,
Michelin, C.M., De Praga Baião, Chuchón Ângulo, E., Almeida, E.K., Kazmierczak, M.
L., Mateus, N.P.A., Silva, R.C., Bender, F., 2021. Extreme drought in the Brazilian
Pantanal in 2019–2020: characterization, causes, and impacts. Front. Water 3,
639204. https://doi.org/10.3389/frwa.2021.639204.
Neves, D.J.D., Alcântara, C.R., de Souza, E., 2016. Estudo de Caso de um Distúrbio
Ondulatório de Leste sobre o Estado do Rio Grande do Norte – Brasil (Case study of
an easterly wave disturbance over Rio Grande do Norte State – Brazil). Rev. Bras.
Meteorol. 31 (4) https://doi.org/10.1590/0102-778631231420150075.
Neves, C.F., Muehe, D., 1995. Potential impacts of seasea-level rise on the metropolitan
region of Recife, Brazil. J. Coast Res. 14, 116–131. Available: http://www.jstor.org/
stable/25735704.
Olbert, A.I., Comer, J., Nash, S., Hartnett, M., 2017. High-resolution multi-scale
modelling of coastal flooding due to tides, storm surges and rivers inflows. A Cork
City example. Coast. Eng. 121, 278–296. https://doi.org/10.1016/j.
coastaleng.2016.12.006.
Painel Brasileiro de Mudanças Climáticas (PBMC), 2016. Impacto, vulnerabilidade e
adaptação das cidades costeiras brasileiras às mudanças climáticas: Relatório
Especial do Painel Brasileiro de Mudanças Climáticas. Rio de Janeiro-RJ, 184p.
Pontes da Silva, L.F., 2011. Contribuição dos Distúrbios Ondulatórios de Leste para a
chuva no Leste do Nordeste do Brasil: evolução sinótica média e simulações
numéricas. Universidade de São Paulo, Instituto de Astronomia, Geofisica e Ciências
Atmosféricas, São Paulo, p. 123. https://doi.org/10.11606/D.14.2011.tde04102011-221401. M.S. Dissertation.
Reboita, M.S., Gan, M.A., Rocha, P., Ambrizzi, T., 2010. Regimes de precipitação na
América do Sul: Uma Revisão Bibliográfica. Revista Brasileira de Meteorologia 25
(2), 185–204. https://doi.org/10.1590/S0102-77862010000200004.
Sa’adi, Z., Ismail, A.Z., Yusop, Z., et al., 2022. Effect of the tide on flood modeling and
mapping in Kota Tinggi, Johor, Malaysia. Nat. Hazards 112, 2053–2081. https://doi.
org/10.1007/s11069-022-05256-4.
Sousa, F.A.S., Vieira, V.R., Silva, V.P., Melo, V.S., Guedes, W.S., 2016. Estimativas dos
riscos de chuvas extremas nas capitais do Nordeste do Brasil. Revista Brasileira de
Geografia Física 9 (2), 430–439. https://doi.org/10.26848/rbgf.v9.2.p430-439.
Souza, W.M., de Azevedo, P.V., de Assis, J.M., de, de, O., Sobral, M., do, C.M., 2014.
Áreas de Risco mais vulneráveis aos Desastres Decorrentes das Chuvas em Recife-PE.
Brazilian J. Environ. Sci. (Online) 34, 79–94. Retrieved from. https://www.rbciamb.
com.br/Publicacoes_RBCIAMB/article/view/219.
UN Office for Disaster Risk Reduction (UNDRR), 2019a. Part I, 2019: the Sendai
framework’s broadened view of the world’s risk (Chapter 3). In: Global Assessment
Report on Disaster Risk Reduction. Geneva, Switzerland.
UN Office for Disaster Risk Reduction (UNDRR), 2019b. Review of efforts made by
member states to implement the Sendai framework (Chapter 9). In: Global
Assessment Report on Disaster Risk Reduction, 2019.
Wallace, J.M., 1970. Time-longitude Sections of Tropical Cloudiness (December 1966November 1967). US Department of Commerce, Environmental Science Services
Administration, National Environmental Satellite Center.
Wallace, J.M., Chang, L.A., 1972. On the application of satellite data on cloud brightness
to the study of tropical wave disturbances. J. Atmos. Sci. 29 (7), 1400–1403.
Wanderley, L.S.A., Nóbrega, R.S., Duarte, C.C.A., Moreira, A.B., dos Anjos, R.S., 2021.
4513 Weather Types Associated with Daily Intense Rainfall Events in the City of
Recife-PE, Brazil, v33. Sociedade e Natureza, 10.14393/SN-v33-2021-60520.
WMO, 2018. Multi-hazard Early Warning Systems: A Checklist: Outcome of the First
Multi-Hazard Early Warning Conference, 22-23 May 2017, Cancún, Mexico. World
Meteorological Organization (WMO), Geneva, Switzerland, p. 20.
Yamazaki, Y., Rao, V.B., 1977. Tropical cloudiness over the south Atlantic Ocean.
J. Meteorol. Soc. Jpn. Ser. II 5, 205–207.
Yonehara, S., Kawasaki, A., 2020. Assessment of the tidal effect on flood inundation in a
low-lying river basin under composite future scenarios. J. Flood Risk Manag. 1,
e312606 https://doi.org/10.1111/jfr3.12606.
Yue, S., Pilon, P., Cavadias, G., 2002. Power of the Mann-Kendall and Spearman’s Rho
tests for detecting monotonic trends in hydrological series. J. Hydrol 259, 254–271.
https://doi.org/10.1016/S0022-1694(01)00594-7.
Zachariah, M., Vasconcelos Junior, F.C., Silva, T.L.V., Dos Santos, R.P., Coelho, C.A.S.,
Alves, L.M., Martins, E.S.P., Köberle, A.C., Singh, R., Vahlberg, M.,
Marchezini, Heinrichj D., Thalheimer, L., Raju, E., Koren, G., Philip, B., Kew, S.,
Bonnet, R., Li, S., Ynag, W., Sun, U., Vecchi, G., Otto, F.E.L., 2022. Climate change
increased heavy rainfall, hitting vulnerable communities in Eastern Northeast Brazil
(Available from: https://www.worldweatherattribution.org/climate-changeincreased-heavy-rainfall-hitting-vulnerable-communities-in-eastern-northeast-bra
zil/.

Comin, A.N., Justino, F., Pezzi, L., et al., 2021. Extreme rainfall event in the Northeast
coast of Brazil: a numerical sensitivity study. Meteorol. Atmos. Phys. 133, 141–162.
https://doi.org/10.1007/s00703-020-00747-0, 2021.
Confederação Nacional de Municípios, C.N.M., 2022. Danos e Prejuízos Causados por
Desastres no Brasil entre 2013 a 2022, Confederação Nacional de Munícipios.
Estudos Técnicos/Defesa Civil – April 2022, 18 pp. Available: https://www.cnm.org.
br/biblioteca/exibe/15317.
Costabile, P., Costanzo, C., Ferrano, D., Barca, P., 2021. Is HEC-RAS 2D accurate enough
for storm-event hazard assessment? Lessons learnt from a benchmarking study based
on rain-on-grid modelling. J. Hydrol. 603, 126962 https://doi.org/10.1016/j.
jhydrol.2021.126962. Part B, 2021.
Diedhiou, A., Machado, L.A.T., Laurent, H., 2010. Mean kinematic characteristics of
synoptic easterly disturbances over the Atlantic. Adv. Atmos. Sci. 27 (3), 483–499.
Espinoza, N.S., dos Santos, C.A.C., Silva, M.T., Gomes, H.B., Ferreira, R.R., da Silva, M.L.,
Santos e Silva, C.M., de Oliveira, C.P., Medeiros, J., Giovannettone, J., Amaro, V.E.,
Santos, C. A.G., Manoranjan, M., 2021. Landslides triggered by the May 2017
extreme rainfall event in the east coast northeast of Brazil. Atmosphere 12, 1261.
https://doi.org/10.3390/atmos12101261.
Ferreira, N.J., Chou, S.C., Prakki, S., 1990. Analysis of easterly wave disturbances over
south Equatorial Atlantic Ocean (in Portuguese). Proc. XIth Brazilian Cong.
Meteorol. Salvador, Bahia, Brazil, CBMET, 10 pp. [Available online at http://www.
cbmet.com/cbm-files/18-66f87184b3a74625603a5cfb8345f5fe.pdf.].
Frich, P., Alexander, L.V., Della-Marta, P., Gleason, B., Haylock, M., Tank Klein,
Peterson, T., 2002. Observed coherent changes in climatic extremes during the
second half of the twentieth century. Clim. Res. 19 (3), 193–212. https://doi.org/
10.3354/cr019193.
Funk, C., Peterson, P., Landsfeld, M., Pedreros, D., Verdin, J., Shukla, S., Husak, G.,
Rowland, J., Harrison, L., Hoell, A., Michaelsen, J., 2015. The climate hazards
infrared precipitation with stations—a new environmental record for monitoring
extremes. Sci. Data 2, 150066. https://doi:10.1038/sdata.2015.66.
Gomes, H.B., Ambrizzi, T., Herdies, D.L., Hodges, K., Silva, B.F.P., 2015. Easterly wave
disturbances over Northeast Brazil: an observational analysis. Adv. Meteorol. 2015,
1–2. https://doi.org/10.1155/2015/176238. ID 176238).
Gomes, H.B., Ambrizzi, T., da Silva, B.F.P., Hodges, K., Dias, P.L.S., Herdies, D.L.,
Gomes, H.B., 2019. Climatology of easterly wave disturbances over the tropical
South Atlantic. Clim. Dynam. 53, 1393–1411. https://doi.org/10.1007/s00382-01904667-7, 2019.
IBGE, 1966, 7. Anuário Estatístico do Brasil, Rio de Janeiro, pp. 1–548.
IBGE, 1975. Anuário Estatístico Do Brasil, p. 36. Rio de Janeiro.
IBGE, 2010. Instituto Brasileiro de Geografia e Estatística, Censo demográfico 2010.
http://www.ibge.gov.br/home/estatistica/populacao/censo2010/default.
IBGE (Instituto Brasileiro de Geografia e Estatística) and CEMADEN (Centro Nacional de
Monitoramento e Alertas de Desastres Naturais), 2018. População em áreas de risco
no Brasil. https://www.ibge.gov.br/apps/populacaoareasderisco/. accessed: july
2018.
ICLEI, 2020. Plano Local de Ação Climática da Cidade de Recife 2020. Prefeitura de
Recife, URBAN LEEDS. ONU HABITAT, ICLEI. 81 pp. Available: https://americad
osul.iclei.org/documentos/plano-local-de-acao-climatica-do-recife-pe/.
IPCC, 2021. Summary for Policymakers. In: Climate Change 2021: The Physical Science
Basis. Contribution of Working Group I to the Sixth Assessment Report of the
Intergovernmental Panel on Climate Change [Masson-Delmotte. Cambridge
University Press. In Press.
IPCC, 2022. Climate Change 2022: Impacts, Adaptation and Vulnerability. Contribution
of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel
on Climate Change. Cambridge University Press. Cambridge University Press,
Cambridge, UK and New York, NY, USA, p. 3056. https://doi.org/10.1017/
9781009325844.
Kendall, M.G., 1975. Time Series, second ed. Hefner, New York, NY, USA, p. 40p. 1975.
Kousky, V.E., 1980. Diurnal rainfall variation in northeast Brazil. Mon. Weather Rev. 108
(4), 488–498. https://doi.org/10.1175/1520 0493(1980)108%3C0488:DRVINB%
3E2.0.CO, 1980 2.
Leão, E.B.S., Andrade, J.C.S., Nascimento, L.F., 2021. Recife: a climate action profile.
Cities 116, 103270. https://doi.org/10.1016/j.cities.2021.103270.
Mann, H.B., 1945. Nonparametric tests against trend. Econometrica 13, 245–259.
https://doi.org/10.2307/1907187.
Marengo, J.A., Alves, L.M., 2012. The 2011 intense rainfall and floods in Rio de Janeiro.
Sidebar 7.1, State of the Climate in 2011. Bull. Am. Meteorol. Soc. 93 (7), S176.
https://doi.org/10.1175/2012BAMSStateoftheClimate.1.
Marchezini, V., Ferreira, A.M., Mourao, C.E.F., Scofield, G.B., Nery, T.D., Luiz, R.A.F.,
Luz, E.F.P., Ishibashi, R., Ivo, A.A.S., Saito, S.M., 2020. A governança dos dados no
sistema de alerta de risco de desastres associados a inundações e deslizamentos. In:
José Rubens Morato Leite; Larissa Verri Boratti; Fernanda Salles Cavedon-

17