Publicações

While there is an increasing understanding on forest response to severe droughts episodes recently occurred in Amazon region (e.g. tree mortality rates, live biomass, CO2 exchanges and forest resilience), there has been not so many works on the hydrological response of the region to rainfall anomalies. The analysis of river discharges along the main stem and major tributaries during the drought episodes of 1997/98, 2005 and 2010 shows that they were of increasing intensities. Discharges well above the historical levels on the southern tributaries restricted the 1997/98 temporal and spatial drought coverage, lasting around 3 months during the low water period and being spatially restricted to the northern tributaries. The 2005 drought lasted the entire period of low water (4-5 months) and was mainly driven by the low discharges coming from the southern tributaries. The high water season presented discharges slightly above the historical means. Finally, the 2010 drought lasted 4-5 months. The discharges dropped to values well below the historical average during the low water and slightly below the historical average during the high water season, mainly driven by the southern tributaries. Additionally, the northern tributaries showed a strong decrease of the discharge during the preceding high water season, which may have affected the groundwater recharge, fully expanding the impact of decreasing rainfall during the dry season. An analysis of the dynamic of the flooding coverage using MODIS sensor shows that between 2000 and 2010, the smallest flooded area occurred in 2005 (36.000 km2) and 2010 (37.800 km2) years. The reduction in flooded area – and the associated increase in the area of exposed soil has a direct relationship with the lives of the riverine communities during extreme drought episodes, as indicated by the number of municipalities which declared state of emergency, according to the Civil Defense data.

The Amazon floodplain ecosystem links large-scale climatic phenomena to the lives of animals (including people), plants, and other organisms that live within it. The floodplain is subject to seasonal changes in water level that can reach up to 16m, as it aggregates rainfall inputs from the entire Amazon basin. The amount and distribution of such rain is related to annual variations in the Tropical Atlantic and Pacific Oceans sea surface temperatures. On the other hand, the assemblies of species that are found within the floodplain depend to a large extent on flooding dynamics. Moreover, floodplains of the major rivers are home to over 70% of the human population living in the Amazon. Within them people fish, raise cattle, grow crops, transport goods, and have leisure. People depend on its biodiversity, though several species have already had their populations drastically depleted due to over-hunting and over-fishing. To contribute to the understanding of how the predicted increases in the frequency of droughts and other extreme climatic events in the Amazon will impact the floodplains present biodiversity, and its biogeochemistry, it is thus necessary to understand the dynamics of the floodplain ecosystem. Here, we begin such investigation from a large-scale perspective. We consider the land cover characteristics of the floodplain (woody vegetation, annual herbaceous vegetation, and open water), and how its spatial and temporal dynamics are related to the flood cycle. We use multi-scale analysis of remote sensing datasets, geographical information systems, hydrograph analyses, fieldwork and several analytical techniques. Preliminary results show that cover types with faster dynamics, such as macrophyte assemblies, respond quickly to extreme droughts. Impacts on these habitats might be carried over to other components of the system, thus influencing carbon fixation and transformation, animal migration patterns, and fish abundance and availability.

A agricultura desempenha papel fundamental na economia brasileira. Porções significativas do território nacional são dedicadas ao cultivo de culturas que constituem importantes fontes de alimentação para o mercado interno e de recursos financeiros, através das exportações. Nesse contexto, a modelagem da fenologia e fisiologia de culturas agrícolas é uma importante ferramenta para, por exemplo, traçar cenários de risco climático ou segurança alimentar da agricultura brasileira. O presente trabalho, iniciado em setembro de 2011, tem como principal objetivo criar um banco de dados baseado em testes de sensibilidade no modelo de culturas agrícolas AgroIBIS, através da alteração de alguns dos vários parâmetros relacionados à estrutura da vegetação disponíveis no modelo. Um dos benefícios trazidos por esse trabalho será fornecer um referencial para pesquisadores que possuam interesse em pontos chave da calibração do AgroIBIS antes da realização de seus experimentos numéricos. O AgroIBIS simula vários aspectos do ciclo de vida de culturas agrícolas: plantio, enchimento de grãos, irrigação, transporte de nutrientes e colheita. O primeiro passo desse trabalho foi a integração do AgroIBIS com a componente de superfície do Modelo Brasileiro do Sistema Climático Global (MBSCG), denominada Inland, que se encontra atualmente em desenvolvimento no CCST/INPE. Embora o AgroIBIS e o Inland sejam baseados no mesmo modelo global de vegetação dinâmica, o IBIS, as rotinas exclusivas de culturas agrícolas tiveram de ser reescritas e incorporadas ao Inland, o que representou um verdadeiro desafio do ponto de vista computacional. Em seguida, uma revisão de literatura foi realizada para fundamentar a escolha dos parâmetros a serem testados. As variáveis do modelo eleitas para mostrar os efeitos dos testes de sensibilidade foram o índice de área foliar (LAI) e a biomassa. O LAI foi escolhido por ser uma importante estimativa da densidade vegetal, que controla, por exemplo, a absorção de radiação e taxa de transpiração da planta. A biomassa representa a soma de todo o carbono assimilado pela fotossíntese e estocado nas diversas partes da planta, e, portanto, também foi avaliada. O banco de dados obtido por esse trabalho será sumarizado no Seminário de Iniciação Científica através de mapas produzidos a partir da diferença entre as variáveis simuladas (LAI e biomassa) utilizando o parâmetro ajustado e o parâmetro padrão do modelo (a chamada simulação de controle).

Deforestation in Amazonia has decreased significantly in the last five years, from around 27,772 km2 in 2004 to around 6,000 km2 in 2010. This decrease has been attributed to a net of governmental/civil society actions, and commodities markets related factors. In spite of this, there is an enormous uncertainty about the fate of the forest. In the next decades, we can expect an increase in food and biofuel demands associated to global population growth and consumption patterns, which can induce both direct and indirect deforestation. Ambitious infra-structure plans also may induce occupation of unprotected areas. Possible climate change can affect the region through intensification of droughts and vulnerability to forest fires. In order to explore future scenarios for the region, there is a need to understand which institutional, political and economical conditions at multiple levels, from national to local, will be able to act as a counterforce to the commodities market pressure, reducing emissions while bringing social development. We present the preliminary results of a project aiming at discussing sustainability scenarios for the Amazonia, combining a qualitative scenario methodology and two spatially-explicit quantitative modeling frameworks: one for land use change (LuccME) and another for deforestation emission estimates (INPE-EM), both representing the heterogeneity of land change processes in the region. A panel of expert constructed storylines quantified in terms of pasture, annual/perennial crops, and secondary vegetation trajectories. INPE-EM is coupled to LuccME to estimate emission reductions resulting from spatial projections of land use trajectories under three different scenarios: (a) emission reductions without social development; (b) sustainability and social development; (c) return to uncontrolled deforestation. Initial modeling results showed that emission reduction in (a) and (b) can be of the order of 6 GtCO2 or more, depending on socioeconomic incentives to the expansion and permanence of secondary vegetation as sinks of carbon.

We are investigating how multi-scale participatory scenario methods can contribute to the process of consolidation of the mosaic of diverse territorial units which constitute the Brazilian Amazonia nowadays. Today, indigenous lands, conservation units and settlements represent more than 60% of the Amazonia. Such units were created with distinct goals, including: biodiversity conservation; to act as a physical barrier to the deforestation frontier; to assure the territory to indigenous and traditional population; as an answer to land reform demands; to provide resources to the forestry sector in Brazil. The future of the region, in terms of social development and greenhouse emissions, depends a great deal on the long term maintenance and sustainability of such units, intrinsically linked to – and dependent on – the future of the their population. We will discuss the results of two case studies at different types of units, a PAE (type of federal settlement with collective land tenure regime, aiming mostly at assuring territory for traditional populations, which live in the area for centuries) and a PA (a conventional type of settlement, in which land parcels are individual, and population are land reform beneficiaries who came from other places). We choose PAE Lago Grande and PA Moju for this end, both located in Santarém, Pará State. PA Moju is still an on-going work, but we have completed the PAE Lago Grande study. Results at PAE Lago Grande show the potential of the proposed methodology as tool to support discussion about a desirable sustainable future among government, civil society and communities, promoting the visibility and empowering of local populations. The comparison of both case studies also sheds light on the heterogeneity of such units, and the different institutional arrangements which act upon them, and also in the potential for replication the methodology at different units and contexts.

Este trabalho tem como objetivo avaliar a variabilidade espaço-temporal da precipitação sazonal na Região Sul do Brasil entre 1988 e 2007, a partir do cálculo do desvio padrão. Para isto, foram utilizados dados de pluviômetros provenientes do banco de dados da Agência Nacional de Águas e estimativas de precipitação do satélite TRMM (Tropical Rainfall Measuring Mission), algoritmo 3B43. Os resultados apontam que, no verão, os menores desvios de precipitação foram observados nas porções leste de Santa Catarina (SC) e centro-leste do Paraná (PR), onde os valores foram inferiores a 40%, enquanto o sudoeste do Rio Grande do Sul (RS) registrou desvios superiores a 70%. No outono, os desvios variaram entre 50% e 60% em boa parte da região. No inverno, os menores desvios localizam-se no centro-leste do RS e na porção central do Estado de SC, variando entre 35% e 50%; enquanto os desvios mais altos foram registrados no sul e oeste do RS, e no extremo norte do PR onde, neste último setor, os desvios ultrapassam 80%. Na primavera, as áreas com os menores desvios localizam-se no oeste catarinense, com percentuais inferiores a 40%; enquanto os maiores percentuais são observados em grande parte do RS e no extremo norte do PR.

Este artigo buscou quantificar a vulnerabilidade socioambiental e climática da RMSP, como conseqüência dos eventos de extremos de chuva, relacionando indicadores sócio-demográficos, econômicos e ambientais. Desse modo, foi gerado um índice que permitiu visualizar as áreas suscetíveis e vulneráveis em períodos de chuvas intensas e com risco de deslizamento.

The purpose of this study is to elaborate the climatology of extremes indices of rainfall in the Amazon basin. For this purpose, climatic indices (CDD, R95p, RX1day, Rnnmm) recommended by the joint World Meteorological Organization CCL/CLIVAR/JCOMM Expert Team on Climate Change Detection and Indices (ETCCDI) were calculated. One of the most important aspects of this study is to characterize possible change in extreme of rainfall (e.g. floods, drought, etc) over Amazon basin, since these events have strongest impacts on society. Extremes indices were derived from daily precipitation at NOAALs Climate Prediction Center (CPCp) with resolution 1×1, for period 1979-2005. The indices calculated in this study are R50mm and R80mm defined as the number of days per year with precipitation amount . 50 mm and . 80mm respectively. CDD defined as annual maximum number of consecutive dry days and R95p (very wet days exceeding the 95th percentile).These indices can be related to drought (CDD) and floods (R95p) events. The rainfall of Amazon basin exhibits high interannual and interdecadal variability associated with ENSO events and with the positive anomaly of SST in the North Atlantic. The annual variability of extreme precipitation indices, such as, maximum length of dry spells (CDD) and RR> 95p (R95p) were sensitive to identification periods of droughts. Also, it verified that extreme rains . 50mm and . 80 mm (R50mm and R80mm), occur with or without events El Nino. On the other hand, droughts are not necessarily associated with a diminution of extreme rainfall events (R50mm e R80mm). Individual locations in the west of the basin (Santarem, Cachimbo and Manaus stations) show negative trends of CDD. However, Rio Branco station (southeast of basin) showed statistically significant positive trends of CDD, and captured the drought of 2005 and 2010. The future scenarios of ETA_HadCM3 model for the period 2011-2040, show an increase (decrease) of the index CDD (R95p) for the average area of the Amazon basin, and it will be stronger in the 2071-2099 climatology. However, as this result is a projection of model there are still many uncertainties. The fact is that historically these extremes of rain are part of climate variability in the region, and they can happen again.