Publicações

Disturbance of vegetation is a critical component of land cover, but is generally poorly constrained in land surface and carbon cycle models. In particular, land-use change and fire can be treated as large-scale disturbances without full representation of their underlying complexities and interactions. Here we describe developments to the land surface model JULES (Joint UK Land Environment Simulator) to represent land-use change and fire as distinct processes which interact with simulated vegetation dynamics. We couple the fire model INFERNO (INteractive Fire and Emission algoRithm for Natural envirOnments) to dynamic vegetation within JULES and use the HYDE (History Database of the Global Environment) land cover dataset to analyse the impact of land-use change on the simulation of present day vegetation. We evaluate the inclusion of land use and fire disturbance against standard benchmarks. Using the Manhattan metric, results show improved simulation of vegetation cover across all observed datasets. Overall, disturbance improves the simulation of vegetation cover by 35% compared to vegetation continuous field (VCF) observations from MODIS and 13% compared to the Climate Change Initiative (CCI) from the ESA. Biases in grass extent are reduced from -66% to 13 %. Total woody cover improves by 55% compared to VCF and 20% compared to CCI from a reduction in forest extent in the tropics, although simulated tree cover is now too sparse in some areas. Explicitly modelling fire and land use generally decreases tree and shrub cover and increases grasses. The results show that the disturbances provide important contributions to the realistic modelling of vegetation on a global scale, although in some areas fire and land use together result in too much disturbance. This work provides a substantial contribution towards representing the full complexity and interactions between land-use change and fire that could be used in Earth system models.

Aim: Human-driven impacts constantly threat amphibians, even in largely protected regions such as the Amazon. The Brazilian Amazon is home to a great diversity of amphibians, several of them currently threatened with extinction. We investigated how climate change, deforestation and establishment of hydroelectric dams could affect the geographic distribution of Amazonian amphibians by 2030 and midcentury. Location: The Brazilian Amazon. Methods: We overlapped the geographic distribution of 255 species with the location of hydroelectric dams, models of deforestation and climate change scenarios for the future. Results: We found that nearly 67% of all species and 54% of species with high degree of endemism within the Legal Brazilian Amazon would lose habitats due to the hydroelectric overlapping. In addition, deforestation is also a potential threat to amphibians, but had a smaller impact compared to the likely changes in climate. The largest potential range loss would be caused by the likely increase in temperature. We found that five amphibian families would have at least half of the species with over 50% of potential distribution range within the Legal Brazilian Amazon limits threatened by climate change between 2030 and 2050. Main conclusions: Amphibians in the Amazon are highly vulnerable to climate change, which may cause, directly or indirectly, deleterious biological changes for the group. Under modelled scenarios, the Brazilian Government needs to plan for the development of the Amazon prioritizing landscape changes of low environmental impact and economic development to ensure that such changes do not cause major impacts on amphibian species while reducing the emission of greenhouse gases.

Fisheries and aquaculture are important sources of food for hundreds of millions of people around the world. World fish production is projected to increase by 15% in the next 10 years, reaching around 200 million tonnes per year. The main driver of this increase will be based on fish farming management in developing countries. In Brazil, fish farming is increasing due to the climate conditions and large supply of water resources, with the production system based on Nile tilapia (Oreochromis niloticus) farming in reservoirs. Inland waters like reservoirs are a natural source of methane (CH4) to the atmosphere. However, knowledge of the impact from intensive fish production in net cages on CH4 fluxes is not well known. This paper presents in situ measurements of CH4 fluxes and dissolved CH4 (DM) in the Furnas Hydroelectric Reservoir in order to evaluate the impact of fish farming on methane emissions. Measurements were taken in a control area without fish production and three areas with fish farming. The overall mean of diffusive methane flux (DMF) (5.9 ± 4.5 mg CH4 m−2 day−1) was significantly lower when compared to the overall mean of bubble methane flux (BMF) (552.9 ± 1003.9 mg CH4 m−2 day−1). The DMF and DM were significantly higher in the two areas with fish farming, whereas the BMF was not significantly different. The DMF and DM were correlated to depth and chlorophyll-a. However, the low production of BMF did not allow the comparison with the limnological parameters measured. This case study shows that CH4 emissions are influenced more by reservoir characteristics than fish production. Further investigation is necessary to assess the impact of fish farming on the greenhouse gas emissions.

We present a characterization of the chemical composition of the atmosphere of the Brazilian Amazon rainforest based on trace gas measurements carried out during the South AMerican Biomass Burning Analysis (SAMBBA) airborne experiment in September 2012. We analyzed the observations of primary biomass burning emission tracers, i.e., carbon monoxide (CO), nitrogen oxides (NOx), ozone (O-3), isoprene, and its main oxidation products, methyl vinyl ketone (MVK), methacrolein (MACR), and isoprene hydroxy hydroperoxide (ISOPOOH). The focus of SAMBBA was primarily on biomass burning emissions, but there were also several flights in areas of the Amazon forest not directly affected by biomass burning, revealing a background with a signature of biomass burning in the chemical composition due to long-range transport of biomass burning tracers from both Africa and the eastern part of Amazonia. We used the [MVK + MACR + ISOPOOH] / [isoprene] ratio and the hydroxyl radical (OH) indirect calculation to assess the oxidative capacity of the Amazon forest atmosphere. We compared the background regions (CO < 150 ppbv), fresh and aged smoke plumes classified according to their photochemical age ([O-3] / [CO]), to evaluate the impact of biomass burning emissions on the oxidative capacity of the Amazon forest atmosphere. We observed that biomass burning emissions disturb the isoprene oxidation reactions, especially for fresh plumes ([MVK + MACR + ISOPOOH] / [isoprene] = 7) downwind. The oxidation of isoprene is higher in fresh smoke plumes at lower altitudes (similar to 500 m) than in aged smoke plumes, anticipating near the surface a complex chain of oxidation reactions which may be related to secondary organic aerosol (SOA) formation. We proposed a refinement of the OH calculation based on the sequential reaction model, which considers vertical and horizontal transport for both biomass burning regimes and background environment. Our approach for the [OH] estimation resulted in values on the same order of magnitude of a recent observation in the Amazon rainforest [OH] congruent to 10(6) (molecules cm(-3)). During the fresh plume regime, the vertical profile of [OH] and the [MVK + MACR + ISOPOOH] / [isoprene] ratio showed evidence of an increase in the oxidizing power in the transition from planetary boundary layer to cloud layer (1000-1500 m). These high values of [OH] (1.5 x 10(6) molecules cm(-3)) and [MVK + MACR + ISOPOOH] / [isoprene] (7.5) indicate a significant change above and inside the cloud decks due to cloud edge effects on photolysis rates, which have a major impact on OH production rates.

The State of São Paulo in Southeast Brazil experienced during the 2012/2013 summer one of the most severe electrical storm events in recent years, causing various impacts to society, e.g. fires and fatalities. Thus, the objective of this work is to understand which mechanisms and climatic variability modes are associated with the lightning incidence during this summer in this region. The study includes a statistical-temporal assessment of cloud-to-ground (CG) lightning occurrence comprising the 16 years period from 1999 to 2014. The results showed that, for this period of analysis, the lightning incidence was associated with climatic patterns connecting the tropical and extratropical region, through a wave train, from the Indian Ocean to South America, favouring the formation and development of convective storms over Southeast Brazil.

Linking the distribution of plant species to geology has generally been biased by the over-simplification of landscape evolution and the lack of understanding of complex geological processes. The Amazonian lowlands have forests in different successional stages, and a growing perception is that such heterogeneity results from long-term environmental changes. This hypothesis is investigated by designing an analytical model based on past and present-day vegetation and successions of the plant communities, combined with an advanced understanding of geological history. An area of southwestern Amazonia was selected for floristic inventories, and we interpreted the paleovegetation based on C/N and δ13C analyses of sedimentary organic matter. These data were examined in the context of the geological evolution on the basis of new sedimentological and chronological data. The topographically high Late Pleistocene deposits had continuous and highly diversified late-successional terra firme forests as well as local fluvial paleolandforms of younger ages with less diversified campinarana forests. Late PleistoceneHolocene terrains in intermediate elevations had terra firme forests, but shorter trees with lower basal areas and ecotonal forests appeared near the confines of the forestsavanna, while Holocene deposits recorded only seasonally flooded varzea forests. Several deposits of Late Pleistocene and Late PleistoceneHolocene age recorded an expansion of C4 terrestrial plants before the establishment of the forest from ~20,000 cal yr BP to 7578 cal yr BP, which is not related to past arid episodes. We recorded forests with onsets at 61303533 cal yr BP, 34022800 cal yr BP, and 1624964 cal yr BP to terra firme, varzea, and ecotonal forests, respectively. However, not all forests have reached maturity stages due to their location on terrains with a diverse history of terrace downcutting and deposition, which had a direct impact on local hydrology with the interaction of topographic gradients. The hydrology of the study area was also controlled by the distance from the main river valley. Capturing long-term disturbances over this region of still pristine forests may help elucidate the potential mechanisms that also determine trends in tree growth and forest diversity in other Neotropical regions.

A Bayesian method known as the Forecast Assimilation (FA) was used to calibrate and combine monthly near-surface temperature and precipitation outputs from seasonal dynamical forecast systems. The simple multimodel (SMM), a method that combines predictions with equal weights, was used as a benchmark. This research focuses on Europe and adjacent regions for predictions initialized in May and November, covering the boreal summer and winter months. The forecast quality of the FA and SMM as well as the single seasonal dynamical forecast systems was assessed using deterministic and probabilistic measures. A non-parametric bootstrap method was used to account for the sampling uncertainty of the forecast quality measures. We show that the FA performs as well as or better than the SMM in regions where the dynamical forecast systems were able to represent the main modes of climate covariability. An illustration with the near-surface temperature over North Atlantic, the Mediterranean Sea and Middle-East in summer months associated with the well predicted first mode of climate covariability is offered. However, the main modes of climate covariability are not well represented in most situations discussed in this study as the seasonal dynamical forecast systems have limited skill when predicting the European climate. In these situations, the SMM performs better more often.

The Paraíba Valley in São Paulo, Brazil, is marked by great transformations. The Brazilian economic cycles changed this region and transformed it into a great mosaic of pastures, forest fragments, and urban areas. Temporal studies on land use and cover have given information on these changes and evaluated their impacts. The conversion of extensive areas of the Atlantic Forest into eucalyptus crops is noticeable; this monoculture developed greatly and is expanding steadily in this valley. This land-use conversion has consequences, such as forest fragmentation, and losses in biodiversity, and water and soil quality. Developing future scenarios (2010 to 2050) that simulate changes in land use and cover is essential for an efficient and sustainable public management. Thus, the objective of the present study was to develop future scenarios while considering different social, economic, and environmental issues, based on analyses of the dynamics of forestry expansion in the Paraíba Valley in São Paulo, Brazil. The scenarios developed were classified as Trend, maintaining the same change patterns of the observed period (2000 to 2005); Optimistic, decreasing conversion rates of natural vegetation into forestry and pasture areas; and Pessimist, increasing deforestation rates and increasing eucalyptus and pasture areas. Future scenarios for the period of 2010 to 2050 were developed in the Dinamica Ego software, and it proved to be effective, presenting trends and predispositions of changes in land use. These results can be used for environmental planning focused on sustainability, better management of natural resources, and development of economic activities. RESUMO: O Vale do Paraíba paulista é marcado por grandes transformações vinculadas aos ciclos econômicos brasileiros que mudaram a região transformando-a em um grande mosaico de pastos, fragmentos florestais e área urbana. Estudos temporais de uso e cobertura permitem observar essas mudanças e avaliar os impactos gerados por elas. É notória a substituição de extensas áreas de Mata Atlântica para o monocultivo de eucaliptos que possui alto desenvolvimento e constante expansão no Vale. Tal conversão de uso traz consequências como a fragmentação florestal, perda de biodiversidade, da qualidade das águas e do solo. A possibilidade de desenvolver cenários de simulação de mudanças de uso e cobertura da terra é fundamental para uma gestão pública eficiente e sustentável. Visando essas questões o presente estudo busca projetar cenários futuros levando em consideração diferentes questões sociais, econômicas e ambientais, a partir de análises sobre a dinâmica de expansão da silvicultura no Vale do Paraíba paulista. Os cenários criados podem ser classificados como Tendencial que mantém os mesmos padrões de mudanças do período observado (2000-2005); Otimista que diminui as taxas de conversão para silvicultura e pastagem, aumentando as áreas vegetadas; e Pessimista que alavanca as taxas de desmatamento a partir do crescimento dos cultivos de eucaliptos e pastagem. O desenvolvimento de cenários futuros de 2010 a 2050 no DinamicaEgo se mostrou eficaz, apresentando tendências e pré-disposições de transições de uso do solo sendo possível a utilização dos resultados no planejamento ambiental visando à sustentabilidade e melhor gestão dos recursos naturais e desenvolvimento de atividades econômicas.

Land Use and Land Cover (LULC) classes defined by subjective criteria can diminish the significance of a study, hindering the reproducibility and the comparison of results with other studies. Having a standard legend for a given study area and objective could benefit a group of researchers focused on long-term or multidisciplinary studies in a given area, in the sense that they would be able to maintain class definition among different works, done by different teams. To allow for reproducibility, it is important that the classes in this legend are described using quantifiable elements of land cover, which can be measured on the ground, as is recommended by Land Cover Meta Language (LCML). The present study aims to propose LCML formalized hierarchical legends for LULC classes, focusing on the southwest of Pará state, within the Brazilian Amazon. In order to illustrate the potential of these legends, a secondary objective of the current study is to analyze classification results using legends derived from a particular Remote Sensing dataset and compare these results with the classification obtained using the LCML hierarchical legend proposed. To perform this analysis, firstly, we proposed a conceptual class model based on existing classification systems for the upland Brazilian Amazon Biome. From this model, 16 LULC classes were described in LCML, using quantifiable and easily recognizable physiognomic characteristics of land cover classes measured on the lower Tapajós river, in Pará state. These classes were grouped into legends with different levels of detail (number of classes), based on our model or on the image and clustering algorithms. All legends were used in supervised classification of a Landsat5/TM image. Results indicate that it is necessary to incorporate multi-temporal knowledge for class definition as well as the proposed thresholds (height and cover proportion of soil, litter, herbaceous vegetation, shrubs, and trees) in order to properly describe classes. However, the thresholds are useful to delimit classes that happen in a successive way. Classification results revealed that classes formed by the same elements of land cover with similar thresholds present high confusion. Additionally, classifications obtained using legends based on the class separability in a given Remote Sensing image tend to be more accurate but not always useful because they can hide or mix important classes. It was observed that the more generalized the legend (those with few details and number of classes), the more accurate the classifications results are for all types of legends.

A necessidade de informações acerca das dinâmicas de uso da terra tornou estudos relacionados à classificação de uso e cobertura da terra na região Amazônica relevantes no campo científico, com destaque ao uso de classificadores supervisionados. No entanto, o desempenho de um classificador supervisionado é dependente dos dados de entrada utilizados (imagem a ser classificada, definição das classes e coleta de amostras de treinamento), que em muitos estudos não são propriamente avaliados. Para contribuir com o entendimento de como variações nos dados de entrada afetam os resultados da classificação supervisionada, este estudo analisou os resultados de classificações de cobertura da terra em uma região da Amazônia brasileira, considerando diferentes métodos de classificação baseados em pixel, legendas com distintos níveis de detalhe (número de classes) e a sensibilidade dos classificadores à variação das amostras de treinamento (Incerteza) com o uso de uma imagem ótica de média resolução. Para tanto, foram geradas 100 imagens classificadas para três legendas com diferentes níveis de detalhe e diferentes configurações de quatro algoritmos de classificação supervisionados: Máxima Verossimilhança (ML), Máquina de Vetores Suporte (SVM), Árvore de Decisão (J48) e k-Vizinhos mais Próximos (K-NN). No total, foram obtidos 111 (3 níveis de legenda x 37 configurações de classificadores) grupos de 100 imagens classificadas, que foram utilizados para gerar imagens combinadas em que cada pixel recebe o rótulo mais frequente nesses grupos de imagens (moda) e mapas da Incerteza associada. De forma geral, a combinação de imagens retornou uma imagem classificada de igual ou maior exatidão que imagens que a compõem, com diferenças menos expressivas em função dos classificador utilizado que ao se considerar os grupos de 100 imagens classificadas, embora a diferença no valor de exatidão seja pequena. A maioria das imagens classificadas mais acuradas foram obtidas utilizando alguma configuração do algoritmo K-NN. No entanto, esse algoritmo apresentou maiores valores de Incerteza quando comparado a ML e SVM. Essa maior variabilidade (maiores valores de Incerteza), em compensação, colabora para a obtenção de melhores resultados da operação da moda, pois permite reclassificar adequadamente pixels de classificação instável. O algoritmo ML apresentou resultados mais consistentes (variam pouco em relação ao conjunto de amostras de treinamento), adequadamente acurados e possui utilização mais simples, já que a etapa de ajuste de parâmetros não é necessária. Quanto menor o nível de detalhe da legenda, maior a exatidão das imagens classificadas e menor sensibilidade das imagens classificadas à variação das amostras de treinamento (Incerteza). A diminuição do nível de detalhe da legenda resulta também em diferenças menos sutis na exatidão das imagens classificadas por diferentes algoritmos na configuração ótima.

O presente artigo tem por objetivo revelar a importância do monitoramento e da avaliação das políticas públicas de educação ambiental, bem como apresentar um processo participativo de construção de indicadores que dialogam diretamente com o Tratado de Educação Ambiental para Sociedades Sustentáveis e Responsabilidade Global. As políticas públicas de educação ambiental, neste contexto, são compreendidas a partir de uma abordagem plural e de enfrentamento coletivo dos problemas socioambientais. Os primeiros resultados obtidos são oito dimensões articuladas de indicadores de monitoramento e avaliação de políticas públicas de educação ambiental. A partir desses resultados iniciais foram realizadas análises que conectam as referidas dimensões ao Tratado e à Política Nacional de Educação Ambiental, apontando para a criação de um sistema de monitoramento nacional, com o potencial de disparar processos adaptados localmente.

Comparing model output and observed data is an important step for assessing model performance and quality of simulation results. However, such comparisons are often hampered by differences in spatial scales between local point observations and large-scale simulations of grid cells or pixels. In this study, we propose a generic approach for a pixel-to-point comparison and provide statistical measures accounting for the uncertainty resulting from landscape variability and measurement errors in ecosystem variables. The basic concept of our approach is to determine the statistical properties of small-scale (within-pixel) variability and observational errors, and to use this information to correct for their effect when large-scale area averages (pixel) are compared to small-scale point estimates. We demonstrate our approach by comparing simulated values of aboveground biomass, woody productivity (woody net primary productivity, NPP) and residence time of woody biomass from four dynamic global vegetation models (DGVMs) with measured inventory data from permanent plots in the Amazon rainforest, a region with the typical problem of low data availability, potential scale mismatch and thus high model uncertainty. We find that the DGVMs under- and overestimate aboveground biomass by 25 % and up to 60 %, respectively. Our comparison metrics provide a quantitative measure for modeldata agreement and show moderate to good agreement with the region-wide spatial biomass pattern detected by plot observations. However, all four DGVMs overestimate woody productivity and underestimate residence time of woody biomass even when accounting for the large uncertainty range of the observational data. This is because DGVMs do not represent the relation between productivity and residence time of woody biomass correctly. Thus, the DGVMs may simulate the correct large-scale patterns of biomass but for the wrong reasons. We conclude that more information about the underlying processes driving biomass distribution are necessary to improve DGVMs. Our approach provides robust statistical measures for any pixel-to-point comparison, which is applicable for evaluation of models and remote-sensing products.