Publicações

Brazil has the major area of tropical forest and carbon stocks of the Amazon. There are plenty biomass distribution analysis with great difference between them, that make impossible to determine which is closer to the reality. IPCC (2006) recommends a biomass stratification at different levels of complexities (tiers), tier 1 requires IPCC default assumptions, methods and data; tier 2 requires detailed country specific strata, methods, assumptions and data. Thus, the aim of the study was to compare the carbon map of the Brazilian Amazon at IPCC tier 1 with the tier 2 carbon map used in Brazil`s Second National Communication to the UNFCCC and suggest the improvements that can be made. We started analyzing the forest biomass stratification and the emission factors (carbon estimates) to get tier 1 and tier 2 maps. As the biomass distribution of tier 2 MCT (2010) results in quadrants because of the RADAM Volumes extrapolation, an alternative tier 2 carbon map was elaborated using the mean biomass of the RADAM plots of each forest stratum. Tier 1 map simplified a more complex reality comparing to MCT (2010) tier 2 map with more detail and number of stratums. The alternative tier 2 carbon map had better carbon stocks distribution, even though future studies and an uncertainty analyses are needed to compare, validate and improve the MCT (2010) tier 2 approach and thus get a consistent carbon map for the Brazilian Amazon.

O Brasil lidera o mercado mundial do café, com fatias da ordem de 30 a 40% de toda a produção. Quebras de safras significativas podem afetar tanto o mercado interno quanto o mercado externo. O grão se apresenta como um fator fixador de mão de obra no meio rural e sua cadeia produtiva gera mais de 8 milhões de empregos. O cafeeiro para ter uma boa produtividade necessita ter condições climáticas adequadas. Faixas ótimas de temperatura médias anuais para a espécie Coffea arabica L. (mais cultivada no Brasil e no mundo) ocorrem entre 18oC e 22oC. A suscetibilidade da cultura a grandes variações de temperatura do ar é preocupante quando se leva em conta às projeções climáticas divulgadas pelo Painel Intergovernamental sobre Mudanças Climáticas (IPCC). Alguns estudos realizados para o Brasil relataram que a cafeicultura brasileira seria impactada diretamente pelas mudanças climáticas, com reduções significativas da produção, causando consequentemente prejuízos financeiros e sociais. No entanto, tais estudos consideraram apenas os zoneamentos agroclimáticos das regiões produtoras. Frente à importância da cafeicultura no cenário socioeconômico brasileiro e dos prognósticos futuros de aumentos da temperatura do ar, a proposta desse estudo consiste da avaliação dos impactos das mudanças climáticas na produtividade cafeeira em diferentes cenários climáticos futuros. Para isso serão usadas simulações dos novos cenários climáticos do IPCC (AR5), geradas a partir de downscaling dinâmico do Modelo Regional Eta em alta resolução espacial (5 km). As saídas do Eta serão usadas no Modelo de Superfície INLAND para gerar a produtividade cafeeira futura em áreas do sudeste do Brasil. O INLAD integra a maioria dos processos da superfície terrestre, incluindo robustamente os efeitos físicos e fisiológicos do dossel, fenologia e ciclo de carbono terrestre.

Previsões numéricas de variáveis meteorológicas, tais como: temperatura, umidade, radiação e principalmente precipitação, têm sido aplicadas em diversos setores socioeconômicos, sobretudo nos setores elétrico e agrícola. O Centro de Previsão de Tempo e Estudos Climáticos (CPTEC) do Instituto Nacional de Pesquisas Espaciais (INPE) produz previsões numéricas globais de forma operacional desde 1995 e previsões para a América do Sul com maior detalhamento, através do modelo Eta, desde 1997. O modelo Eta é um modelo numérico atmosférico complexo em sua representação dos processos físicos e dinâmicos da atmosfera. É da categoria de modelos atmosféricos chamados de regionais ou área limitada. Modelos regionais são computacionalmente mais econômicos que os modelos globais na resolução e na região desejada. O modelo Eta na versão sazonal (Eta-sazonal) é rodado mensalmente no CPTEC e as previsões são produzidas para um horizonte de quatro meses, com um domínio que cobre toda a América do Sul, América Central e Oceanos adjacentes na resolução horizontal de 40km. Neste trabalho, foram avaliados 10 anos de previsões climáticas sazonais de precipitação (2001 a 2010) geradas a partir de um segundo downscaling dinâmico do modelo na resolução de 10 km. O aumento da resolução horizontal contribuiu para um ganho no detalhamento da superfície, com uma melhor caracterização das regiões de topografias complexas, como por exemplo na região da Cordilheira dos Andes. As rodadas foram iniciadas a cada mês do ano, cobrindo um horizonte de tempo de quatro meses e meio, com três membros de perturbação, os quais consistiram de datas diferentes de condição inicial (dias 13, 15 e 17). A área de estudo consistiu da Amazônia Ocidental abrangendo também os territórios colombiano, peruano e regiões adjacentes. A precipitação foi avaliada por campos espaciais, tomando como base os dados observacionais do Climate Research Unit – University of East Anglia (CRU). Nas avaliações consideram-se as previsões com um a quatro meses de antecedência, os diferentes membros de perturbação do modelo e a média dos membros. Em geral, as precipitações previstas acompanharam o padrão climático, média do total das 30 previsões válidas para cada mês, apresentando uma boa concordância com a média observada independentemente do prazo de previsão e dos membros de perturbação. A média climatológica fornecida pelo modelo foi capaz de indicar os principais regimes de precipitação nas diferentes épocas do ano, mostrando-se hábil também em prever os períodos de transição entre a estação chuvosa e a seca e vice-versa.

Landslides have frequently occurred in last years, due to the disorderly growth of the cities and the occupation of risk areas by the poor population, causing social, environmental and economic impacts. Urban areas in expansion move to geologically unstable areas and topographically inclined, such as the River Bengalas Basin, located in the city of Nova Friburgo, mountainous region of the State of Rio de Janeiro, Brazil. This article aims to present the model developed and used to evaluate the susceptibility and vulnerability of the River Bengalas Basin to landslides, which in January 2011, with the occurrence of heavy rains, caused landslides that impacted in the death of 429 people in city of Nova Friburgo. For the case study, several investigations have been made related to the areas of the basin, such as slope, soil conditions, lithology, land use and cover, vertical curvature (Very Convex, Convex, Flat, Concave, Very Concave), horizontal curvature (Very Divergent, Divergent, Flat, Convergent, Very Convergent), and precipitation data. With this study it was possible to understand how the natural and anthropics elements of the basin are related to the local dynamics of the disasters regarding to their interferences in the induction of landslides; evaluate the effectiveness of the guidelines of the Plano Diretor Participativo of the city of Nova Friburgo regarding the landslides; identify the susceptible and vulnerable areas to landslides and assess the rates of susceptibility and vulnerability to landslides from new calculation model proposed.

As sucessões secundárias são tipologias importantes para a manutenção da biodiversidade, regime hidrológico e sequestro de carbono. A utilização de métricas texturais GLCM pode colaborar na discriminação dessas classes por extrair a variabilidade espacial do dossel florestal. Assim sendo, também se faz necessária uma técnica como redes neurais artificiais para seleção dos atributos mais relevantes e integração desses dados. O objetivo do presente estudo foi de avaliar e comparar o uso de atributos espectrais ALI/EO-1 e métricas texturais GLCM utilizando a técnica de redes neurais artificiais Multi-Layer Perceptron para mapeamento da cobertura da terra na Floresta Nacional do Tapajós e arredores, com foco na discriminação das tipologias florestais primárias e sucessões secundárias. Observou-se que os atributos texturais mais relevantes foram a textura média das bandas 3, 4, 6, 7 e 8, e textura dissimilaridade da banda 8. Esses atributos, ao serem integrados aos dados espectrais em um conjunto híbrido, proporcionaram uma melhor discriminação entre as classes de NPV e solo, culturas agrícolas e SS1/SS2, SS1 e SS2, SS2 e SS3/FP. Dessa forma, as áreas de SS1, SS2, SS3 e FP puderam ser discriminadas com 89, 63, 62 e 83% de acurácia. Constatou-se exatidão global de 89% para a utilização dos dados híbridos contra 79% para dados somente espectrais. ABSTRACT Secondary successions are important typologies for biodiversity maintenance, hydrological regimen, and carbon sequestration. The use of GLCM textural metrics can collaborate to discriminate these classes due to the extraction of the spatial variability of the forest canopy. Hence, it is also necessary a technique such as artificial neural networks to select the most relevant attributes and to integrate these data. The aim of this study was to evaluate and compare the use of ALI/EO-1 spectral attributes and GLCM textural metrics using the Multi-Layer Perceptron artificial neural networks technique for land cover mapping in the Tapajos National Forest and vicinity, focusing on the discrimination of primary forest and secondary successions. It was observed that the most important textural attributes were the mean texture of bands 3, 4, 6, 7 and 8, and the dissimilarity of band 8. These attributes, when integrated into the spectral data to compose a hybrid dataset, provided better discrimination between the classes of NPV and soil, crops and SS1/SS2, SS1 and SS2, SS2 and SS3/PF. Thereby, the SS1, SS2, SS3 and FP areas could be discriminated with 89, 63, 62 and 83% of classification accuracy. It was observed an overall accuracy of 89% using the hybrid dataset against 79% using only the spectral data.

Studies of upward flashes from tall structures around the world have been reported since 1939. The characteristics of the triggering mechanism and of the upward flashes are of increasing interest to wind power turbines, telecommunication toware and power distribution companies. Observations of upward flashes in Brazil are recent. Since 2012, upward events were observed in all different seasons of year. They initiate on several towers in Sao Paulo, a city in southeast Brazil at an average elevation of around 800 meters above sea level and a flash density of 15 flashes/km².year. Some of the towers are located on top of buildings or hills. In this study, we will present some characteristics of these upward flashes, based on high speed video observation: flash duration, presence of subsequence of return stroke, previous lightning events, initial continuous current pulses and the presence recoil leader.

During the early afternoon of 22 September 2013, severe storms, accompanied by large hail, damaging winds, heavy precipitation and dangerous lightning, devastated a region between 100 and 200 km south-southwest of the Bauru S-band Doppler radar. Several extremely intense storm cells moved at up to 100 km/h east-southeastward and at least one of them spawned a tornado when it reached the town of Taquarituba, in Southeastern Brazil. The synoptic pattern was very favorable for the development of severe weather conditions over southeastern Brazil, due to a cyclogenesis process that generated an MCS (Mesoscale Convective System), with a squall line inserted. The convective cells were tracked throughout their lifetime by IPMet´s Doppler radars. The cells intensified drastically, with towers overshooting into the stratosphere (10 dBZ ≤ 20 km) and shortly before reaching the town of Taquarituba, that particular cell displayed extremely strong radial shear just above the cloud base, which led to the formation of a deep meso-cyclone, from which the tornado spawned and touched down at around 14:30h LT. Before the tornado touch-down relatively few CG strokes were observed near the tornadic cell, but thereafter the frequency increased rapidly. The IC activity was extremely high at most times. WRF model proofed itself extremely efficient to forecast the severe thunderstorm.

High-speed video records of two bipolar cloud-to-ground flashes were analyzed in detail. They were both initiated by single positive return strokes and were followed by more than one subsequent negative stroke. Due to the elevated cloud-base height of its parent thunderstorm, the initiation processes of each subsequent stroke could be registered optically below cloud base. In the first event (Case 1) it was possible to observe that all four subsequent negative strokes were initiated by recoil leaders that retraced one horizontal channel segment previously ionized by the positive leader that preceded the first stroke. Those recoil leaders connected to the original vertical channel segment and propagated towards ground, producing four subsequent strokes that had the same ground contact point of the original positive discharge. The second event (Case 2), on the other hand, presented fifteen subsequent strokes that were initiated by recoil leaders that did not reach the original channel of the positive stroke. They diverged vertically towards ground, making contact approximately 11 kilometers away from the original positive strike point. These are the first optical evidences that both single- and multiple-channel bipolar flashes occur as a consequence of recoil leader activity in the branches of the initial positive return stroke. For both events their total channel length increased progressively with time at a growth rate equivalent to a propagation speed of the order of 104 m/s, comparable to those reported for typical positive leaders.

The objective of this work is the analysis of overall characteristics of 35 cloud-to-ground (CG) strokes that presented two or more contact points to the ground in a millisecond scale. This dataset includes both the classical forked strokes (first observed in streak camera records) and the new class of this type of phenomenon, called upward illumination strokes, introduced in more recent works. The genesis of the latter is very similar to the classical forked strokes, i. e., one branch from the main trunk of a stepped leader produces a second stroke after the first return stroke occurs. The main difference from the classical forked stroke events are: a) UI stroke channels appear to completely disconnect from the main trunk during its development, b) the time between strokes is longer than in the case of classical forked strokes and c) the peak currents of the UI strokes are, usually, very small. From the visual inspection of the 35 selected events, 22 were classified as forked strokes and 13 UI strokes. A new name was given to the broad class of strokes that present two or more contact points in a millisecond and sub-millisecond scale, multi ground contact strokes (MGCS). Subclasses received the names forked strokes and UI strokes, related to the previous findings from the literature. All MGCS flashes were recorded during the summer season of 2013 in Southeastern Brazil as part of a five-day campaign employing a network of high-speed camera sensors known as the RAMMER network. The high-speed video records were obtained by three identical cameras installed in São José dos Campos and a fourth mobile camera (whose only difference from the other three is its capability to record colored images). All four cameras were set up for recording at 1200 x 500 pixels spatial resolution, 2500 frames per second, and 400-μs time interval between frames. A total of 357 negative CG flashes were recorded. The physical processes that generate either forked strokes and UI strokes are discussed in details. RAW data from BrasilDAT network was used to identify and give more information about the MGCS, such as: peak current estimates and precise interstroke intervals in the microsecond range. Plots of time intervals between strokes versus peak current added new information on the physical characteristics and distinctive features of UI and forked strokes.