The number of reported drought events per year and their impacts have significantly increased in the last two decades. In addition to monitoring drought conditions, forecasting is essential for planning activities. Various Machine Learning (ML) algorithms have experienced a substantial increase in popularity in geoscience applications. This study presents a Systematic Literature Review on drought forecasting utilizing Machine Learning models. Following the PRISMA 2020 protocol (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), the total number of papers was reduced from approximately a thousand to a hundred. The majority of the papers found study areas from Asia and Oceania. Meteorological drought was the most studied event in the articles evaluated due to the greater ease of its estimation using only rainfall data. The Standardized Precipitation Index and the Standardized Precipitation Evapotranspiration Index are the most widely used indices in research relating to drought and Machine Learning. Precipitation is the most commonly used input among the various input data used in ML models. Remote sensing has yet to be widely used in drought forecasting, with less than 20

This study evaluates the performance of subseasonal forecasts for dry spells and heatwaves at a regional scale in Brazil. The forecasts’ verification was designed to provide end-users with relevant information about the forecasts’ quality. The U.K. Met Office model was assessed using a significant sample of weekly forecasts: 552 for dry spells and 240 for heatwaves. The analysis reveals that the overall performance of the forecasts is low, with a chance of detecting an event close to 0.2, indicating that only one out of five observed dry spells is accurately predicted on average. The application of quantile mapping corrections demonstrates improvements in predicting shorter dry spells (up to 5 days) and longer lead times, although the timing of these forecasts often remains inaccurate, leading to increased false alarms. A significant improvement in the forecast quality occurs when categorization by duration is disregarded. The detection chances increase to 0.5−0.7 for dry spells and 0.5 for heatwaves. The Brier Score indicates that the probabilistic forecasts issued by the model are equivalent or less skilful than climatological probabilities. Overall, the findings underscore the challenges in forecasting dry spells and heatwaves in Brazil and highlight the need for ongoing improvements in forecasting methodologies to enhance their reliability and utility for regional decision-making. This research contributes to understanding subseasonal climate forecasting and its implications for managing climate-related risks in Brazil.

The climate modeling techniques of event attribution enable systematic assessments of the extent that anthropogenic climate change may be altering the probability or magnitude of extreme events. In the consecutive years of 2018, 2019, and 2020, rainfalls caused repeated flooding impacts in the lower Parnaíba River in Northeastern Brazil. We studied the effect that alterations in precipitation resulting from human influences on the climate had on the likelihood of flooding using two ensembles of the HadGEM3-GA6 atmospheric model: one driven by both natural and anthropogenic forcings; and the other driven only by natural atmospheric forcings, with anthropogenic changes removed from sea surface temperatures and sea ice patterns. We performed hydrological modeling to base our assessments on the peak annual streamflow. The change in the likelihood of flooding was expressed in terms of the ratio between probabilities of threshold exceedance estimated for each model ensemble. With uncertainty estimates at the 90% confidence level, the median (5% 95%) probability ratio at the threshold for flooding impacts in the historical period (1982–2013) was 1.12 (0.97 1.26), pointing to a marginal contribution of anthropogenic emissions by about 12%. For the 2018, 2019, and 2020 events, the median (5% 95%) probability ratios at the threshold for flooding impacts were higher at 1.25 (1.07 1.46), 1.27 (1.12 1.445), and 1.37 (1.19 1.59), respectively; indicating that precipitation change driven by anthropogenic emissions has contributed to the increase of likelihood of these events by about 30%. However, there are other intricate hydrometeorological and anthropogenic processes undergoing long-term changes that affect the flood hazard in the lower Parnaíba River. Trend and flood frequency analyses performed on observations showed a nonsignificant long-term reduction of annual peak flow, likely due to decreasing precipitation from natural climate variability and increasing evapotranspiration and flow regulation.

The State of São Paulo, Brazil (SSP) was impacted by severe water shortages during the intense austral summer drought of 2013/2014 and 2014/2015 (1415SD). This study seeks to understand the features and physical processes associated with these summer droughts in the context of other droughts over the region during 1961–2010. Thus, this study examines the spatio-temporal characteristics of anomalously low precipitation over SSP and the associated large-scale dynamics at seasonal timescales, using an observation-based dataset from the Climatic Research Unit (CRU) and model simulation outputs from the Met Office Hadley Centre Global Environment Model (HadGEM3-GA6 at N216 resolution). The study analyzes Historical and Natural simulations from the model to examine the role of human-induced climate forcing on droughts over SSP. Composites of large-scale fields associated with droughts are derived from ERA-20C and ERA-Interim reanalysis and the model simulations. HadGEM3-GA6 simulations capture the observed interannual variability of normalized precipitation anomalies over SSP, but with biases. Drought events over SSP are related to subsidence over the region. This is associated with reduced atmospheric moisture over the region as indicated by the analysis of the vertically integrated moisture flux convergence, which is dominated by reduced moisture flux convergence. The Historical simulations simulate the subsidence associated with droughts, but there are magnitude and location biases. The similarities between the circulation features of the severe 1415SD and other drought events over the region show that understanding of the dynamics of the past drought events over SSP could guide assessment of changes in risk of future droughts and improvements of model performance. The study highlights the merits and limitations of the HadGEM3-GA6 simulations. The model possesses the skills in simulating the large-scale atmospheric circulations modulating precipitation variability, leading to drought conditions over SSP.

The strongest El Niño events of the past four decades were associated with large rainfall deficits in North Brazil during the December to February mature phase, leading to substantial societal and ecological impacts and influencing the global carbon cycle. While the teleconnection between El Niño and northern South America is well studied, the small number of El Niño events—and especially high magnitude ‘major’ El Niños—in the recent observational record make a robust characterisation of the response over North Brazil in today’s climate difficult. Here we use a large, initialised ensemble of global climate simulations to provide a much greater sample of North Brazil rainfall responses to recent El Niño events than is available from observations, and use this to form an assessment of the chance of unprecedented dry conditions during El Niño. We find that record low rainfall totals are possible during El Niño events in the current climate, and that as the magnitude of El Niño increases, so too does the chance of unprecedented low rainfall, reaching close to 60% for major El Niños. However, during even the largest El Niños, when the observed North Brazil response has been similar and very dry, we find rainfall rates close to normal are still possible due to internal atmospheric variability. In addition to the predictable influence of the tropical Pacific, an unpredictable influence from the extratropics appears to play a role in modulating the North Brazil rainfall response via an equatorward wave-train that propagates down the western coast of North America and across to the Caribbean. Combining forecasts of El Niño with this improved information on the underlying chance of extremely low rainfall could feed into improved assessments of risk and preparedness for upcoming droughts in Brazil.

Brazil has endured the worst droughts in recorded history over the last decade, resulting in severe socioeconomic and environmental impacts. The country is heavily reliant on water resources, with 77.7% of water consumed for agriculture (irrigation and livestock), 9.7% for the industry, and 11.4% for human supply. Hydropower plants generate about 64% of all electricity consumed. The aim of this study was to improve the current state of knowledge regarding hydrological drought patterns in Brazil, hydrometeorological factors, and their effects on the country’s hydroelectric power plants. The results show that since the drought occurred in 2014/2015 over the Southeast region of Brazil, several basins were sharply impacted and remain in a critical condition until now. Following that event, other regions have experienced droughts, with critical rainfall deficit and high temperatures, causing a pronounced impact on water availability in many of the studied basins. Most of the hydropower plants end the 2020–2021 rainy season by operating at a fraction of their total capacity, and thus the country’s hydropower generation was under critical regime.

Drought poses a major threat to food security and maintenance of rural populations in dry lands, because it leads to losses in agriculture and other productive sectors. Given the intensification and recurrence of drought events and their impacts, it is necessary to deepen our knowledge as a key to subsidize mechanisms for drought preparation and mitigation plans. Therefore, this study aimed to develop and apply methodologies for drought characterization and impact assessment, considering different agro-ecosystems and the impact on smallholder farmers. The new approach developed included as dataset LULC maps, the limits of rural properties, and the Vegetation Supply Water Index (VSWI) time series. The z-score was used in order to identify areas affected by droughts and a zonal statistic was used to compare the results of VSWI in a municipal and rural property scale. Afterwards, the areas identified as significant statistics were used to determine the LCLU and rural property most affected. The results of the analysis showed that Agricultural Lands were the most affected LULC during the study period (about 50%), whereas Micro and Small Properties were the most affected by droughts, over than 65% and 15% respectively. Besides, an important finding is that the analysis at the municipality or rural property levels can affect the drought assessment. This can result in an underestimated or overestimated of drought intensity and in the expansion of impacts. Thus, these results contribute to the local drought mitigation actions and help determine the threshold to better spatialize drought impacts on smallholder farming.

Given the lack of studies about dry spells over Southeastern Brazil (SEB), the present study aims to characterize them considering their duration, geographical incidence, and association with rainfall. Uninterrupted sequences of no-rain days were calculated using the high-resolution (10 km) Multi-Source Weighted-Ensemble Precipitation data set (MSWEP; 1979-2016). The majority of dry spells (80%) last no more than 9 days and those remaining for 10 days or more are relatively rare (top 20%). January and February present a predisposition for the occurrence of dry spells, feature that is not observed in December. The life cycle of dry spells evolves in close association with the subseasonal rainfall tendency. When adding the perspective of dry spells, the effective peak of the monsoon season is December, the month with the highest rainfall rates and the lowest chances for dry spells. Afterward, dry spells modulate a progressive weakening of the rainy season until March, when a slight recovery may happen in the central northern portions of SEB. The relationship between dry spells and rainfall is not straightforward. For dry spells that are most common (up to 9 days), the most likely association is with negative monthly anomalies up to -1.0 SD. However, there are also significant chances of having dry spells associated with positive anomalies. Extreme long dry spells, but extremely rare as well, are unequivocally associated with negative anomalies. The geographic discrimination of dry spells evidences a dipole-like feature, better seen in January and February, similar to the well-known rainfall dipole. The north and northeast portions tend to be drier, more prone to the occurrence of dry spells, both short/frequent and long/infrequent. In the south and southwestern portions, dry spells tend to be shorter. The main result is the indication of a probable lessening in the rainy season, taking place between January and February that might affect a significant portion of SEB.

O aumento na ocorrência e de frequência de secas extremas tem ocasionado o aumento no número de desastres associados a incêndios florestais em todo o planeta. Neste artigo, buscamos contextualizar os incêndios florestais no âmbito de desastres socioambientais propondo uma estruturação de um sistema de gestão e de alerta de risco para este tipo de evento. Sugere-se a estruturação deste sistema baseado em cinco eixos principais, sendo eles: conhecimento do risco, monitoramento, educação e comunicação, capacidade de prevenção e capacidade de resposta. Em seguida, realizamos uma análise diagnóstica sobre as instituições, atribuições, responsabilidades e ações do governo brasileiro, nos níveis federal e estadual, em relação à gestão de riscos de incêndios florestais. Identifica-se a falta de uma regulamentação política e legal sobre as responsabilidades e estratégias para mitigar os riscos e impactos destes eventos. A partir dessa análise sobre as ações atuais, apontam-se alguns desafios à gestão integrada de risco de incêndios florestais no Brasil.

Predicting extreme events is one of the major challenges of sub-seasonal to seasonal (S2S) forecasting due to the high human and financial cost of such disasters. S2S forecasts of high-impact events should help with mitigating actions and putting contingency plans into place. This chapter discusses the S2S prediction of two categories of extreme events: large-scale, individual, long-lasting extreme weather events whose onset, evolution, and decay could be predicted a few weeks in advance; and small-scale, high-impact events that cannot be predicted individually weeks in advance, but statistics for which could be predictable at the S2S timescale due to their interaction with large-scale circulation. Examples of each type of extreme event prediction are presented.

© 2019 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).A supplement to this article is available online (10.1175/BAMS-D-18-0102.2)

Drought-related disasters are among the natural disasters that are able to cause large economic and social losses. In recent years, droughts have affected different regions of Brazil, impacting water, food, and energy security. In this study, we used the Integrated Drought Index (IDI), which combines a meteorological-based drought index and remote sensing-based index, to assess the drought events from 2011 to 2019 over Brazil. During this period, drought events were observed throughout the country, being most severe and widespread between the years 2011 and 2017. In most of the country, the 2014/15 hydrological year stands out due to the higher occurrence of severe and moderate droughts. However, drought intensity and observed impacts were different for each region, which is shown by the different case studies, assessing different types of impacts caused by drought in Brazil. Thus, it is fundamental to evaluate the impacts of droughts in a continental country such as Brazil, where a variety of vegetation, soil, land use, and especially different climate regimes predominate.

The value of weather and climate information, at any timescale, is a function of the availability, comprehensibility, and usability of the information so that decisions and actions can be taken in response to uncertain future events. The uncertainties and available skill of sub-seasonal to seasonal (S2S) forecasts have the potential to make communication and dissemination of these forecasts more challenging, particularly in scenarios where decisions are critical to life and well-being or have significant economic impact on the users. Engagement with user communities, therefore, is essential to ensure that these forecasts provide their anticipated value and to prevent misconceptions or disparities between user expectations and the available science. This chapter describes the current state of the literature on S2S application in a range of sectors (agriculture, energy and water, disaster risk reduction (DRR), and health) and the readily available public products and services. Gaps and trends in current S2S application research are identified, and descriptions of best practice examples are synthesized to provide a set of guiding principles for S2S forecast communication, dissemination, and user engagement.

A seca é considerada o desastre natural que pode causar as maiores perdas econômicas e sociais, com o maior número de pessoas afetadas diretamente dentre todos os tipos de desastres naturais. Na região semiárida do Brasil, é manifestada por meio da redução da produtividade agrícola ou mesmo pecuária, causando sérios problemas sócio-políticos. Diferentes indicadores de seca são conhecidos e utilizados na comunidade científica. O presente trabalho explora a aplicabilidade de um índice híbrido de seca, calculado a partir de dados de NDVI e temperatura da superfície, denominado VSWI. O índice foi aplicado em áreas de pastagens no semiárido do Brasil para a avaliação dos impactos da seca de 2012-2013 na vegetação. De acordo com os resultados 85% da região foi impactada pela seca entre os anos de 2012-2013. De maneira geral, os resultados obtidos por meio do índice VSWI concordaram com aqueles obtidos por meio de interpolação de dados observacionais de precipitação, armazenamento de água no solo (modelo de balanço hídrico) e dados de produção pecuária. Diante disso, a relação empírica LST-NDVI pode ser eficazmente explorada como um indicador das características espaço-temporais de condições de estresse hídrico na vegetação para a região semiárida do Brasil. Além disso, os resultados também pontuaram para a importância da dinâmica da vegetação nas análises dos impactos da seca, uma vez que anos previamente mais secos ou mais úmidos apresentam impactos nos anos subsequentes (efeitos memória e recuperação vegetal).

Este estudo teve por objetivo detalhar aspectos climáticos relacionados ao período de seca que acarretou nos grandes incêndios ocorridos no Acre, durante a estação seca (junho a setembro) de 2005. Um aspecto original deste estudo foi a avaliação da variabilidade climática sub-sazonal. A avaliação dos padrões anômalos de variáveis climáticas relevantes mostrou que associações consistentes entre a superfície e a atmosfera que determinaram a severidade da estiagem de 2005. Primeiro, houve um pré-condicionamento dado que a estação chuvosa imediatamente anterior (2004-2005) foi deficiente. Segundo, os volumes de precipitação durante a estação seca de 2005 foram irrisórios, com a maior parte do estado experimentando menos de 50% da chuva esperada na estação, e algumas regiões, principalmente no leste acreano, com menos de 25%. Em uma perspectiva histórica, a estação seca de 2005 classificou-se como a menos chuvosa na série 1998 a 2014. Considerando-se uma média para toda a estação seca, a umidade relativa do ar apresentou anomalias negativas, devido a um transporte anômalo, contrário ao fluxo climatológico, de ar mais seco das latitudes mais ao sul para o sudoeste da Amazônia. De forma consistente com o déficit de umidade na atmosfera, a umidade do solo (modelo de balanço hídrico) também apresentou anomalias negativas durante a estação seca de 2005. O panorama sazonal de anomalias negativas de umidade relativa, foi resultante de dois ciclos sub-sazonais nos quais a umidade relativa do ar apresentou declínio acentuado e posterior recuperação, intercalados por um período de aproximadamente 20 dias de relativa normalidade. Em cada um destes ciclos foram atingidos valores mínimos extremos (inferiores a 1% dos mínimos da série de 1998 a 2014). Estes mínimos extremos de umidade relativa aparentam estar associados a condições mais persistentes de estiagem, condições estas que são sucedidas por um aumento de queimadas e/ou maior alastramento de incêndios. Os dois ciclos sub-sazonais mostraram associação com os movimentos de subsidência atmosférica de grande escala. Este mecanismo de subsidência mostrou-se associado com pulsos de atividade inibidora intensificada que se propagaram do Oceano Pacífico para a região do estado do Acre, de forma análoga à Oscilação de Madden-Julian.

A Amazônia vem sofrendo o aumento de intensidade e de ocorrência de eventos climáticos extremos. A ocorrência de secas nesta região aumenta a susceptibilidade das florestas a incêndios florestais, com diversas consequências para o meio ambiente, economia e saúde da população. O objetivo deste estudo foi fornecer uma análise espaço-temporal do uso do fogo no Estado do Acre, e assim auxiliar a Sala de Situação do Estado na tomada de decisão para priorizar o monitoramento de áreas com risco de incêndios. Para isso, foram utilizados dados de focos de calor oriundos de múltiplos satélites, dados de unidades fundiárias e análises estatísticas para gerar um ordenamento de áreas prioritárias para monitoramento de incêndios. O satélite AQUA, desde o início de sua operação em 2002, foi responsável por 40% a 75% dos totais de detecção de focos. Com o lançamento do satélite S-NPP em 2013, este vem sendo responsável pela maioria das detecções de focos de calor devido melhores em suas resoluções espaciais e radiométricas. Trinta e nove por cento do total de focos de calor detectados, foram localizados em projetos de assentamentos, 26% em áreas particulares, 10% em unidades de conservação e menos de 2% em terras indígenas. Um mapa de risco de incêndios baseado em análises de tendência e número de ocorrências de focos de calor foi proposto para auxiliar a definição de áreas prioritárias para monitoramento e fiscalização. Conclui-se que as informações geradas com base nos dados históricos de focos de calor podem ser incorporadas aos modelos de risco de incêndios que operam com dados puramente climáticos de forma a melhorar a espacialização do risco e assim apoiar o planejamento e a tomada de decisão.