Video summarizing a Ph.D. dissertation for a non-specialist audience. / Emergency Management / Emergency Management / College of Arts, Humanities and Social Sciences

A avaliação da previsão de chuva a curtíssimo prazo com até 3 horas de antecedência na área de cobertura do RSP para diferentes tipos de sistemas precipitantes, principalmente os associados às enchentes e deslizamentos na RMSP, foi realizada por meio de um modelo advectivo a partir do campo de vento 2D médio e da velocidade dos campos das taxas de precipitação estimados com o radar e um Esquema Numérico de Terceira Ordem Corrente Acima (ENTOCA). O ENTOCA utiliza um vetor com deslocamento mantido constante. O desempenho da previsão para precipitação acumulada num determinado intervalo de tempo foi avaliado pelo Índice de Sucesso Crítico (CSI), Probabilidade de Detecção (POD) e Razão de Falsos Alarmes (FAR). Quantitativamente, a acurácia da previsão foi avaliada por meio do Erro Quadrático Médio (EQM). O coeficiente de correlação mostrou que a qualidade da previsão decresce ao longo do tempo, com maior previsibilidade para os sistemas estratiformes do que para os convectivos. O ENTOCA não considera a evolução espaço-temporal dos sistemas precipitantes durante a extrapolação do campo das taxas de precipitação. Em geral, constatou-se uma subestimativa da precipitação acumulada. As previsões também apresentaram maior desempenho para até 90 minutos e menor, a partir de 120 minutos de extrapolação. O desempenho médio da previsão pelo índice CSI para o limiar de 0.2 mm ao final de 60 minutos de precipitação acumulada foi: FFs (77%), Lis (67,5%), BDs (58%), CIs (56,4%) e BMs (47%). Em geral, a partir de 90 minutos de advecção (sistemas convectivos) e 120 minutos (sistemas estratiformes), o desempenho da previsão diminui exponencialmente. / The evaluation of the rainfall short-term forecast up to 3 hours in advance within the surveillance area of São Paulo weather radar (RSP) for different types of precipitating systems, mainly the are associated to floods and landslides in Metropolitan Area of São Paulo (RMSP), was carried out with an 2D wind advective scheme and rainfall rates estimated with the RSP. The third-order upstream numerical scheme (ENTOCA) was used with a uniform wind vector. The rainfall forecast skill for a given time interval was evaluated by the Critical Success Index (CSI), Probability of Detention (POD) and False Alarm Rate (FAR). Quantitatively, the accuracy of the forecast was evaluated with the mean-square error (mse). The correlation coefficient showed that the quality of the forecast decrease with time, with better skill for the stratiform systems than for convective ones, given that the ENTOCA do not take into account spatial-temporal evolution of the rainfall systems. In general, the precipitation accumulation was underestimated. The forecasts had better skill up to 90 minutes. The average skill based on CSI for the thresholds of 0.2 mm at 60 minutes the precipitation accumulation are: FFs (77%), Lis (67,5%), BDs (58%), CIs (56,4%) and BMs (47%). In general, from 90 minutes of advection (convective systems) and 120 minutes (stratiform systems), the skill of the forecast decreases.

Proposal submitted in partial fulfilment for the award of MTech in Environmental Management / It is widely believed that experts often have a more rational approach towards risks. This is because they are known to use algorithms, formal logic, risk assessments and normative rules to make decisions about risks. The central tenet of this research is that communication based on an understanding of how people conceptualise and evaluate risk communication efforts is critical for translating risk management knowledge into effective risk practices necessary for value generation in flood risk mitigation. Rational decisionEmaking requires both analytic and intuitive systems to operate on a parallel level. Therefore, this research proposes a Flood Risk Communication Model that takes cognisance of lay perceptions. The model emphasises on how risk communication efforts are evaluated by the lay using a combination of descriptive psychological and social construction theories. In particular, the prospect theory, heuristics and biases, cultural theory and trust theory are used to provide explanatory sketches on how flood risk communication efforts are perceived in highly vulnerable environmental contexts such as informal settlements. The challenge in this research however, lies in verifying the model empirically. The associative group analysis technique will be used to generate empirical data from a case study population. Two basic analytic methods will be employed to measure psychological dispositions of respondents. Firstly, word associations are scored and weighted based on frequency of occurrence to generate a dominance score. The higher the dominance score, the greater the interpretation and the more meaningful the theme is for that particular group. Secondly, the different theories of the model are factored into a questionnaire to measure priorities. All the responses are then compared to the proposed model and also used to evaluate actual lay perceptions and feelings towards the current risk communication interventions. The results showed a high level of consistency with the FRCM and hence with the descriptive psychological models of Kahneman and Tversky. However, we conclude that what is has been proposed to be biases are intuitive tendencies to adapt and make sustainable decisions in the face of applicable contextual influences. Thus, these contextual hierarchies determine the reference point and status quo of the recipient in decision making. Therefore, these influences and hierarchies need to be factored in the designing of a risk communication.

A capacidade adaptativa às enchentes diz respeito à capacidade inerente de indivíduos ou de um sistema de se ajustar aos efeitos desse evento e lidar com ele, de modo a moderar seus danos potenciais. A cidade de São Paulo é particularmente vulnerável às enchentes devido ao seu histórico de uso e ocupação do solo. O objetivo deste trabalho é analisar a capacidade adaptativa a partir da realidade local de moradores do Jardim Pantanal, localizado na zona leste do município de São Paulo às várzeas do rio Tietê, a fim de propor ações que possam contribuir na construção dessa capacidade. A pesquisa foi desenvolvida por meio de levantamento documental e bibliográfico, entrevistas semiestruturadas, análise das transcrições, codificação, e categorização dos dados. As capacidades adaptativas genérica e específica nos níveis organizacionais individual e de sistema são baixas, e entre os determinantes da capacidade adaptativa às enchentes os recursos financeiros, a vulnerabilidade urbana e as estratégias de enfrentamento foram considerados os mais importantes, em nível individual. A falta de recursos, a irregularidade de rendimentos e a ausência de diversificação na fonte de renda limitam as opções disponíveis de moradia em áreas regulares e dificultam a mobilização de recursos para a adoção de medidas preventivas e de recuperação pós-evento. A vulnerabilidade urbana expressa-se pela ocupação em área irregular, onde não são realizados investimentos em medidas de infraestrutura por parte dos moradores, que poderiam reduzir a exposição aos impactos das enchentes, pois não se sabe até quando poderão permanecer na área. As estratégias de enfrentamento demonstram ter caráter apenas reativo sem qualquer planejamento, sendo decididas e tomadas reativamente quando a água sobe. Tendo em vista os aspectos observados, a construção da capacidade adaptativa às enchentes no Jardim Pantanal requer: a) entrosamento entre as medidas de adaptação autônomas (do indivíduo) e as planejadas (do sistema); b) ações de adaptação antecipatórias, mais do que responsivas; e c) medidas de adaptação de curto e longo prazos que considerem as vulnerabilidades que surgiram durante o período de adaptação. / Adaptive capacity to floods means the inherent capacity of individuals or systems to adjust and cope with its effects, in order to moderate potential damages. The city of São Paulo is particularly vulnerable to floods due to its history of occupation and land use. The goal of this research was to analyze the adaptive capacity in the local reality of the Jardim Pantanal residents, located in the east zone of the municipality of São Paulo, at the bed of Tietê River, in order to propose actions that can contribute in building this capacity. The research was developed through bibliographical and documental research, semi-structured interviews, transcript analysis, coding and categorization of data. Generic and specific adaptative capacities in the individual and system organizational levels are low, and among the determinants of adaptive capacity to floods, financial resources, urban vulnerability and coping strategies are considered the most important at the individual level. The lack of resources and the instability and lack of diversification of income sources limit the options available in regular housing areas and make it difficult to mobilize resources for the adoption of preventive and post-event recovery measures. Urban vulnerability is due to illegal land occupation, where investments in infrastructure, which could reduce the exposure to the impacts of floods, are not made by residents, because they dont know for how long they will be able to stay in the area. Coping strategies were found to have only a reactive character, without any planning, being decided and taken reactively when the water rises. Given the observed aspects, building adaptive capacity to floods in Jardim Pantanal requires a) understanding between autonomous (individual) and planned (system) adaptation measures; b) anticipatory rather than responsive adaptation actions; and c) short and long-term adaptation measures to consider the vulnerabilities arising during the adaptation period.

No description available.

The recent extreme hydrological extremes over the globe highlight the importance of understanding the role of atmospheric dynamics and climate variability on the occurrence of these extreme events and the associated temporal and spatial characteristics of sequences of the precipitation events. Most of the studies have been focusing on overall average impacts of long-term global climate change on extremes. Majority are driven largely by considering the changes of the moisture holding capacity as a function of temperature, as indicated by the Clausius-Clapeyron equation. Given the complex dynamical structure of the atmosphere, one needs to also consider the attendant atmospheric circulation and moisture transport mechanisms that lead to extreme precipitation and subsequent floods as evidenced in the recent major floods. This study first develops insights into the causative climatic factors associated with precipitation induced regional floods events and understand the roles of Atmospheric Rivers (AR) or Tropical Moisture Exports (TME) and atmospheric circulation patterns associated with the frequency and/or persistency of such events in the midlatitudes. The second part explores the spatiotemporal relationship between climate variability and global extreme precipitation occurrence using a graph based approach based upon the concept of reciprocity to investigated the linkages and influences of the slowly changing boundary conditions on the development or propagation of atmospheric circulations, to assess the predictability of global precipitation extremes given the leading modes of identified climate dipole networks. A multi-timescale statistical, climate informed, stochastic streamflow forecast model serves as the bridge linking the first two parts to the application in the third part: application on water resources management by developing a multi-timescale climate informed stochastic hybrid stimulation-optimization model for multi-purpose reservoir systems, which enables the utilization of the streamflow forecast. The novel reservoir operation model attempts to change the game of water resources management from its conservative, rigid rule-following scheme to a robust, market-based, reliable water allocation strategy. Part I. Tropical Moisture Exports, Extreme Precipitation and Major Flood Atmospheric Rivers are being increasingly identified as associated with some extreme floods. More generally, such floods may be associated with tropical moisture exports that exhibit relatively robust teleconnections between moisture source regions and flood regions. First, a large-scale flood event that persisted over Western Europe in January 1995 is studied. During the last ten days of the month, two rare flooding events, associated with heaviest rainfall in 150 years, occurred in two places, one over Brittany (West of France), and the second in the France-Germany border region and parts of neighboring countries. In this study, we explore the month-long evolution of tropical moisture exports (TME) and their connection to the precipitation events that led to the Brittany event. The persistent large-scale atmospheric circulation patterns that led to the birth, death and evolution of these TME as atmospheric rivers with landfalls in Western Europe are identified, and the relationship of daily extreme precipitation to these patterns is examined. Singular value decomposition (SVD) analysis and a generalized linear model (GLM) are used to assess whether knowledge of the atmospheric circulation patterns from the prior record is useful for explaining the occurrence of their rare events. The analysis establishes the importance of both global and regional atmospheric circulation modes for the occurrence of such persistent events and the hydrologic importance of diagnosing global atmospheric moisture pathways. Part II. Seasonal to Interannual Variability of Tropical Moisture Exports, Extremes and ENSO A statistically and physically based framework is put forward that investigates the relationship between Tropical Moisture Exports (TMEs), Extreme Precipitation and Floods. TMEs is the more general phenomena than Atmospheric Rivers (ARs) in terms of (1) facilitates the poleward transport of warm and moist air masses from low latitudes, primarily tropical oceanic areas, to higher latitudes; (2) contributes to the global climatology precipitation and its extremes; (3) closely relates to floods events, especially in the midlatitudes. The TMEs itself has seasonal and interannual variability that is regulated by slowly changing boundary conditions and climate variability, such El Niño Southern Oscillation (ENSO), while the trajectories and movements are presumably led by atmospheric circulations patterns driven by the balance of global energy and water budgets. In this study, we take Northwest US (NE US) to show how the TMEs is related to extreme precipitation and then floods, and the results of the variability of TMEs, coupled with atmospheric circulation patterns, on the extremes. Historical large floods events in NE US in different seasons are studied for their link to the TMEs. Major moisture sources of TMEs that contributes to precipitation, extremes and floods in NE US are identified, together with the sources' seasonally and interannually varying characterizes in terms of both birth and entrance to the NE US, with the consideration of large scale climate regulations and atmospheric circulation patterns. Part III. Correlation Networks for Identifying Predictors for Extended Range Forecasts for Extreme Precipitation Correlation networks identified from financial, genomic, ecological, epidemiological, social and climate data are being used to provide useful topological insights into the structure of high dimensional data. Strong convection over the oceans and the atmospheric moisture transport and flow convergence indicated by atmospheric pressure fields may determine where and when extreme precipitation occurs. Here, the spatiotemporal relationship between climate and extreme global precipitation is explored using a graph based approach that uses the concept of reciprocity to generate cluster pairs of locations with similar spatiotemporal patterns at any time lag. A global time-lagged relationship between pentad sea surface temperatures (SST) anomalies and pentad sea level pressure (SLP) anomalies is investigated to understand the linkages and influence of the slowly changing oceanic boundary conditions on the development of the global atmospheric circulation. We explore the use of this correlation network to predict extreme precipitation globally over the next 30 days, using a Principal Component logistic regression on the strong global dipoles found between SST and SLP. Unprecedented success of the predictive skill under cross validation for 30 days precipitation higher than the 90th percentile is indicated for selected global regions for each wet season considered. Part IV. Applications of Climate Informed Streamflow Forecasts for Water Management Streamflow forecasts at multiple time scales (e.g., season and year ahead) provide a new opportunity for reservoir management to address competing objectives. Market instruments such as forward contracts with specified reliability are considered as a tool that may help address the perceived risk associated with the use of such instruments in lieu of a traditional operation and allocation. A water allocation process that enables multiple contracts with different durations, to facilitate participatory management of the reservoir by users and system operators, is presented here. Since these contracts are based on a verifiable reliability they may in turn be insurable. A Multi-timescale climate informed Stochastic Hybrid Simulation - Optimization Model (McISH) is developed, featuring (1) dynamic flood control storage allocation at a specified risk level; (2) multiple duration energy/water contracts with user specified reliability and prices; and (3) contract sizing and updating to reflect changes in both demands and supplies. The model incorporates multi-timescale (annual and seasonal) streamflow forecasts, and addresses uncertainties across both timescales. The intended use is as part of an interaction between users and water operators to arrive at a set of short-term and long term contracts through disclosure of demand or needs and the value placed on reliability and contract duration. An application is considered using data for the Bhakra Dam, India. The issues of forecast skill and contract performance given a set of parameters are examined to illustrate the approach. Prospects for the application in a general setting are discussed.

A dissertação apresenta um estudo das chuvas extremas relacionadas a quatro desastres naturais ocorridos no Brasil: as inundações do Vale do Itajaí SC, em novembro de 2008, a inundação histórica de São Luís do Paraitinga - SP, em janeiro de 2010, as inundações ocorridas no Vale do Mundaú AL, em junho de 2010 e as inundações e escorregamentos da Região Serrana - RJ, em janeiro de 2011. As chuvas catastróficas foram analisadas através de estatísticas básicas dos dados dos pluviômetros localizados nas regiões. No Vale do Itajaí, as chuvas registradas nos dias 23 e 24 de novembro foram elevadas, atingindo valores acima de 250 mm. Na estação Blumenau, choveu 243,5 mm e 250,9 mm nesses. Na estação localizada em São Luís do Paraitinga, choveu apenas 64,7 mm no dia 1 de janeiro de 2010, quando ocorreu a inundação. Porém, foram observados 205,7 mm em uma das estações localizadas em Cunha. Nesse caso, o elevado volume precipitado na cabeceira da bacia deflagrou as inundações observadas nos dois municípios. No Vale do Mundaú e Paraíba, choveu cerca de 200 mm no dia 5 de junho, em duas das seis estações analisadas. O elevado volume precipitado no dia 5, combinado com as chuvas ocorridas no período de 17 a 19, pode ter causado as inundações observadas no dia 19 nessas bacias. Os dados indicaram que, na Região Serrana do RJ, as inundações e escorregamentos foram causados pela chuva extrema ocorrida nos dias 11 e 12 de janeiro de 2011, que ultrapassou 270 mm no intervalo de 24 h em uma das estações. As chuvas acumuladas nos meses que antecedem os eventos e a alta declividade contribuíram para a saturação do solo e posteriores escorregamentos. Os eventos pluviométricos, classificados através do SPI resultaram, na maioria das estações, chuvas severas ou chuvas extremas A vulnerabilidade das regiões, onde inúmeras habitações estão localizadas em áreas de risco, também foi determinante para que os desastres acontecessem. Outros eventos de magnitude elevada foram observados anteriormente, o que indica que estes eventos são característicos das regiões estudadas. Constatou-se que as regiões analisadas estão sujeitas a chuvas extremas com frequência relativamente alta, muito embora tenha sido observado, em alguns casos, certo grau de raridade nesses eventos. Portanto, nessas áreas devem ser adotadas medidas regionais no sentido de disciplinar o uso e ocupação do solo e reduzir os riscos dos desastres. É fundamental buscar medidas de adaptação da ocupação dessas áreas, considerando o regime hidrológico dessas regiões. / The paper presents a study of extreme rainfall related to four natural disasters occurring in Brazil: the floods in Itajaí Valley, state of Santa Catarina, in November, 2008; the historic flood in São Luís do Paraitinga, state of São Paulo, in January, 2010; the floods in Mundaú Valley, state of Alagoas, in June, 2010; and the floods and landslides in the mountainous region of Rio de Janeiro in January, 2011. The catastrophic rains were analyzed through basic statistical data collected from rain gauges located in those regions. In Itajaí Valley, extremely high rainfalls exceeding 250 mm were recorded on November 23 and 24. In Blumenau, it rained 243.5 mm and 250.9 mm on the same days. At the station located in São Luís do Paraitinga, it rained just 64.7 mm on January 1, 2010, when the flood occurred. However, 205.7 mm were observed in one of the stations located in Cunha. In this case, the high volume of rainfall at the headwater of the basin triggered flooding observed in these two cities. In the valleys of Mundaú and Paraíba, it rained nearly 200 mm on June 5, in two of the six stations analyzed. The high volume of rainfall on June 5, combined with the rains from the 17th to the 19th, may have led to the floods in these basins on June 19. The data indicated that, in the mountainous region of Rio de Janeiro, the flooding and landslides were caused by massive rainfall on January 11 and 12, 2011, which exceeded 270 mm within a period of 24 hours in one of the stations. The rainfall accumulated in the months prior to the events and the high sloping land contributed to soil saturation and subsequent landslides. The rainfall events, sorted through the SPI, resulted in severe or extreme rains in most of the stations. The vulnerability of the regions, which include many homes located in hazardous areas, was also crucial for the disasters to happen. Other major events were previously observed, which indicates that these events are characteristic of the studied regions. It was noted that the analyzed regions are subject to extreme rains with a relatively high frequency, although in some cases these events have demonstrated to be somewhat rare. Therefore, in these areas, region-based measures should be adopted with a view to regulating the use and occupation of the soil and reducing risk of disasters. It is essential to seek adaptation measures of occupation of these areas, considering their hydrological regime.

Em função das perspectivas trazidas pela descoberta de minério de ferro na Serra dos Carajás, no sul do Pará, o Governo Federal promoveu a partir de 1970, uma série de intervenções na cidade de Marabá. Uma vez verificada a localização inadequada do sítio urbano, vulnerável às enchentes periódicas dos rios Tocantins e Itacaiúnas, foi planejada a transferência da cidade para outro local, onde hoje se encontra a Nova Marabá. A população da cidade antiga não viu com simpatia a idéia, tendo já estabelecido um convívio com as enchentes que praticamente todos os anos afetam a cidade. A persistência da Marabá Pioneira e de seus moradores que ainda estabelecem um modo de vida típico das populações ribeirinhas da Amazônia revela também o fracasso da política de planejamento governamental feita de fora da realidade local e de uma acomodação dos moradores com um problema que acaba gerando custos elevados para o Poder Público em função dos prejuízos provocados todos os anos pelas enchentes. / With newfound prospects brought on by the discovery of iron ore in the Serra dos Carajás (Carajás Range), in the south of the State of Pará, the Brazilian Fedral Government undertook a series of interventions in the city of Marabá, starting in 1970. Once the location of the city was deemed inadequate, as it was vulnerable to flooding from the Tocantins and Itacaiúnas Rivers, plans were made to move it to another site, now known as Nova Marabá. The population of the old city did not take to this idea, sinde they had already learned to live with the floods that affected them almost every year. The persistence of the pioneer Marabá and its inhabitants, who still lead the life of typical river dwelling populations throughout the Amazon region, also reveals the failure of a governmental planning policy out of touch with local reality and the accommodation reached by the inhabitants with a problem that ends up generating exorbitant costs for the State every year, the result of damage caused by the flooding.

The Government of Malawi, through delegates from the Department of Disaster Management Affairs, has on several occasions advised smallholder farmers who live in flood prone areas to relocate to upland areas that are safe from floods. Smallholder farmers have refused to do so and continue to live in the flood prone areas despite experiencing on-going flooding. Smallholder farmers living in flood prone areas in Malawi insist that flash floods bring fertile soils from upland areas that enhance crop production in the flood prone areas. These fertile soils allow smallholder farmers to grow a variety of crops, fruits and vegetables throughout the year, some of which they sell. Within this context, my research critically explores how smallholder farmers perceive their vulnerability to floods and seeks to understand the factors and processes that motivate them to live in the flood prone areas. It also examines the realities and dynamics of local adaptation in the flood prone areas in Malawi through opportunities, challenges, barriers and limitations. The research uses 57 in-depth interviews, a household survey involving 227 households, participant observations and 12 focus group discussions with smallholder farmers. Findings show that firstly, smallholder farmers are not ready to abandon their land and relocate upland because floods are part of their lives and livelihood strategies. Secondly, that power dynamics at household and community levels based on gender roles and culture need to be understood and accounted for in local adaptation strategies in order to effectively enhance local adaptive capacity. Thirdly, that social networks and interdependence between the smallholder farmers living in flood prone areas and those living in upland areas play a significant role in the adoption of local adaptation strategies and adaptation to floods and droughts through temporary migration. This thesis reveals that the perception and extent of vulnerability to floods is dynamic and differentiated based on several factors. The thesis also reveals that local adaption is a complex process such that in some cases, the realities of power dynamics at both the household and community level affects local adaptive capacity to floods. Transformational adaptation that incorporates specific and contextual adaptation strategies is therefore recommended as one of the best approaches towards achieving successful adaptation to climate variability and resilience.

A presente dissertação tem como objetivo elaborar modelos em diferentes escalas para previsão, espacialização e análise das áreas inundáveis na bacia hidrográfica do rio Caí, RS. Este trabalho está estruturado em três módulos, de acordo com o grau de detalhamento dos modelos. O Módulo I está relacionado aos modelos de escala média (1:100.000) para o mapeamento e a análise das inundações na bacia hidrográfica do rio Caí. O Módulo II compreende um modelo aplicado à previsão, à espacialização e à análise das inundações na área urbana de São Sebastião do Caí, em escala grande (1:10.000). O Módulo III é semelhante ao segundo, no entanto, em escala maior (1:5.000), na área urbana e entornos de Montenegro. No Módulo I foram utilizados dados de três estações fluviométricas para analisar as características hidrológicas da bacia (tempo de retorno das cheias, distribuição sazonal dos eventos, etc.), e dados SRTM e da DSG para modelar a superfície do terreno e espacializar as áreas inundáveis. No Módulo II foram utilizados dados de duas estações fluviométricas para elaborar um modelo matemático de previsão das inundações. Para a espacialização e a análise das áreas inundáveis, utilizaram-se dados altimétricos de razoável precisão (nivelamento municipal) e uma imagem de satélite de alta resolução espacial. O modelo matemático de previsão de inundações do Módulo III é semelhante ao do Módulo II, alterando apenas as estações utilizadas. A modelagem da superfície foi realizada com curvas de nível com equidistância de 1m e pontos cotados, compatíveis com escala de 1:2.000. O modelo aplicado à espacialização das inundações em Montenegro considerou o desnível existente na lâmina d’água durante os eventos, diferentemente do Módulo II, o que aumentou a precisão e a qualidade da modelagem. Verificou-se que mais de 75% dos eventos de grande magnitude ocorreram entre os meses de junho e outubro. De modo geral, os dados SRTM e da DSG superestimaram as áreas atingidas, mas representaram satisfatoriamente os locais de maior avanço das águas. Os modelos matemáticos de previsão de inundações, em São Sebastião do Caí e Montenegro, apresentaram, em geral, precisão submétrica para a estimativa de cotas, obtendo resíduo médio de 0,68m e 0,25m, respectivamente. Quanto aos modelos de espacialização das inundações dos Módulos II e III, os erros planimétricos verificados foram de 19m (São Sebastião do Caí) e 9m (Montenegro). O erro altimétrico médio (diferença entre as lâminas d’água medida e modelada), nos pontos avaliados, foi de apenas 0,30m em São Sebastião do Caí e de 0,15m em Montenegro. / This dissertation aims to develop models at different scales for prediction, spatialization and analysis of the floods in the Caí River Basin, RS. This work is structured in three modules, according to the level of detail of the models. The Module I is related to models of medium scale (1:100,000) for mapping and analysis of floods in the Caí River Basin. The Module II includes a model applied to prediction, to spatialization and to analysis of floods in the urban area of São Sebastião do Caí, in large scale (1:10,000). The Module III is similar to the second, however, on a larger scale (1:5,000), in the urban area and surroundings of Montenegro. In Module I we used data from three gauging stations to analyze the hydrological characteristics of the basin (recurrence interval of floods; seasonal distribution of events), and the SRTM and DSG data’s to model the terrain surface and spatialize flooded areas. In Module II used data from two gauging stations to develop a mathematical model for flood prediction. For the spatialization and the analysis of the wetlands were used altimetry data of reasonable accuracy (town topography leveling) and a satellite image of high spatial resolution. The mathematical model for flood prediction of the Module III is similar to Module II, changing only the gauged stations used. The modeling of the surface was performed with contour lines in intervals of 1 meters and measured points, consistent with scale of 1:2,000. The model applied to the spatialization of flooding in Montenegro found the gap in water depth during the events, unlike the Module II, which increased the accuracy and quality of modeling. It was found that over 75% of events of great magnitude occurred between June and October. In general, the SRTM and DSG data’s overestimated the affected areas, but accounted satisfactorily the locals of the biggest advance for water. The mathematical models for flood prediction in São Sebastião do Caí and Montenegro, were, in general, sub-meter accuracy for the estimation of quotas, obtaining average residual of 0.68 m and 0.25 m, respectively. As for the models of flood spatialization of Modules II and III, the errors planimetric were 19m (São Sebastião do Caí) and 9m (Montenegro). The average altimetric error (difference between the water depths measured and modeled), in the measured points, was only 0.30 m in São Sebastião do Caí and 0.15 m in Montenegro.