Publicações

Deforestation and drought are among the greatest environmental pressures on the Amazon rainforest, possibly destabilizing the forest-climate system. Deforestation in the Amazon reduces rainfall regionally, while this deforestation itself has been reported to be facilitated by droughts. Here we quantify the interactions between drought and deforestation spatially across the Amazon during the early 21st century. First, we relate observed fluctuations in deforestation rates to dry-season intensity; second, we determine the effect of conversion of forest to cropland on evapotranspiration; and third, we simulate the subsequent downwind reductions in rainfall due to decreased atmospheric water input. We find large variability in the response of deforestation to dry-season intensity, with a significant but small average increase in deforestation rates with a more intense dry season: with every mm of water deficit, deforestation tends to increase by 0.13% per year. Deforestation, in turn, has caused an estimated 4% of the recent observed drying, with the south-western part of the Amazon being most strongly affected. Combining both effects, we quantify a reinforcing drought-deforestation feedback that is currently small, but becomes gradually stronger with cumulative deforestation. Our results suggest that global climate change, not deforestation, is the main driver of recent drying in the Amazon. However, a feedback between drought and deforestation implies that increases in either of them will impede efforts to curb both.

Tropical forests regulate the global water and carbon cycles and also host most of the worlds biodiversity. Despite their importance, they are hard to survey due to their location, extent, and particularly, their cloud coverage. Clouds hinder the spatial and radiometric correction of satellite imagery and also diminishing the useful area on each image, making it difficult to monitor land change. For this reason, our purpose is to identify the cloud detection algorithm best suited for the Amazon rainforest on Sentinel2 images. To achieve this, we tested four cloud detection algorithms on Sentinel2 images spread in five areas of the Amazonia. Using more than eight thousand validation points, we compared four cloud detection methods: Fmask 4, MAJA, Sen2Cor, and s2cloudless. Our results point out that FMask 4 has the best overall accuracy on images of the Amazon region (90%), followed by Sen2Cors (79%), MAJA (69%), and S2cloudless (52%). We note the choice of method depends on the intended use. Since MAJA reduces the number of false positives by design, users that aim to improve the producers accuracy should consider its use.

Brazil hosts the largest expanse of tropical ecosystems within protected areas (PAs), which shelter biodiversity and support traditional human populations. We assessed the vulnerability to climate change of 993 terrestrial and coastal-marine Brazilian PAs by combining indicators of climatic-change hazard with indicators of PA resilience (size, native vegetation cover, and probability of climate-driven vegetation transition). This combination of indicators allows the identification of broad climate-change adaptation pathways. Seventeen PAs (20,611 km2) were highly vulnerable and located mainly in the Atlantic Forest (7 PAs), Cerrado (6), and the Amazon (4). Two hundred fifty-eight PAs (756,569 km2), located primarily in Amazonia, had a medium vulnerability. In the Amazon and western Cerrado, the projected severe climatic change and probability of climate-driven vegetation transition drove vulnerability up, despite the generally good conservation status of PAs. Over 80% of PAs of high or moderate vulnerability are managed by indigenous populations. Hence, besides the potential risks to biodiversity, the traditional knowledge and livelihoods of the people inhabiting these PAs may be threatened. In at least 870 PAs, primarily in the Atlantic Forest and Amazon, adaptation could happen with little or no intervention due to low climate-change hazard, high resilience status, or both. At least 20 PAs in the Atlantic Forest, Cerrado, and Amazonia should be targeted for stronger interventions (e.g., improvement of ecological connectivity), given their low resilience status. Despite being a first attempt to link vulnerability and adaptation in Brazilian PAs, we suggest that some of the PAs identified as highly or moderately vulnerable should be prioritized for testing potential adaptation strategies in the near future.

Deforestation and drought are among the greatest environmental pressures on the Amazon rainforest, possibly destabilizing the forest-climate system. Deforestation in the Amazon reduces rainfall regionally, while this deforestation itself has been reported to be facilitated by droughts. Here we quantify the interactions between drought and deforestation spatially across the Amazon during the early 21st century. First, we relate observed fluctuations in deforestation rates to dry-season intensity; second, we determine the effect of conversion of forest to cropland on evapotranspiration; and third, we simulate the subsequent downwind reductions in rainfall due to decreased atmospheric water input. We find large variability in the response of deforestation to dry-season intensity, with a significant but small average increase in deforestation rates with a more intense dry season: with every mm of water deficit, deforestation tends to increase by 0.13% per year. Deforestation, in turn, has caused an estimated 4% of the recent observed drying, with the south-western part of the Amazon being most strongly affected. Combining both effects, we quantify a reinforcing drought-deforestation feedback that is currently small, but becomes gradually stronger with cumulative deforestation. Our results suggest that global climate change, not deforestation, is the main driver of recent drying in the Amazon. However, a feedback between drought and deforestation implies that increases in either of them will impede efforts to curb both.

The Atlantic rainforest of Brazil is one of the global terrestrial hotspots of biodiversity. Despite having undergone large scale deforestation, forest cover has shown signs of increases in the last decades. Here, to understand the degradation and regeneration history of Atlantic rainforest remnants near São Paulo, we combine a unique dataset of very high resolution images from Worldview-2 and Worldview-3 (0.5 and 0.3m spatial resolution, respectively), georeferenced aerial photographs from 1962 and use a deep learning method called U-net to map (i) the forest cover and changes and (ii) two pioneer tree species, Cecropia hololeuca and Tibouchina pulchra. For Tibouchina pulchra, all the individuals were mapped in February, when the trees undergo mass-flowering with purple and pink blossoms. Additionally, elevation data at 30m spatial resolution from NASA Shuttle Radar Topography Mission (SRTM) and annual mean climate variables (Terraclimate datasets at ∼ 4km of spatial resolution) were used to analyse the forest and species distributions. We found that natural forests are currently more frequently found on south-facing slopes, likely because of geomorphology and past land use, and that Tibouchina is restricted to the wetter part of the region (southern part), which annually receives at least 1600 mm of precipitation. Tibouchina pulchra was found to clearly indicate forest regeneration as almost all individuals were found within or adjacent to forests regrown after 1962. By contrast, Cecropia hololeuca was found to indicate older disturbed forests, with all individuals almost exclusively found in forest fragments already present in 1962. At the regional scale, using the dominance maps of both species, we show that at least 4.3% of the current region’s natural forests have regrown after 1962 (Tibouchina dominated, ∼ 4757 ha) and that ∼ 9% of the old natural forests have experienced significant disturbance (Cecropia dominated).

The South American Monsoon System (SAMS) is a major climatic feature of South America, and its domain extends from Amazon to La Plata basin. The SAMS region is vulnerable to variations of climate and precipitation patterns, which could impact economic activities and lead to potential societal consequences. In the face of a warming future scenario, the importance of the study of the past climate with numerical simulations is to evaluate the climate models and to assure the reliability of future projections. Here we investigate the Mid-Holocene SAMS, evaluating changes in strength, life cycle and associated dynamical mechanisms in ten Earth System Models simulations. Our results show that the SAMS was weaker in the Mid-Holocene than in the pre-industrial climate in December-January-February (DJF), but stronger in September-October-November (SON). This is probably a consequence of insolation variations in the Mid-Holocene, which contributed to changes in the moisture flux from the Atlantic Ocean to the continent, the strength of the upper-level atmospheric circulation, and the amount of precipitation over the SAMS region. Moreover, we suggest that the life cycle of the SAMS was altered during the Mid-Holocene, with an earlier onset and demise. Our results also indicate that Mid-Holocene SAMS changes are connected to precipitation variations near Northeast Brazil, in a dipole configuration of precipitation between western Amazon and Northeast Brazil, due to the influence of the Walker cell. Finally, this study highlights a need for improvement of the numerical models to better simulate the amount of precipitation over South America and the upper-level circulation over western Amazon in SON, which are crucial factors for a more realistic representation of the SAMS.

Environmental disturbances affect ecosystem functioning through changes in organisms’ metabolism (direct effect) and biodiversity loss (indirect or biodiversity-mediated effect). It is still a challenge to separate direct and biodiversity-mediated effects of environmental changes on ecosystem functioning due to the difficulties in isolating ‘true’ biodiversity loss effects. Furthermore, it is still unclear whether biodiversity-mediated effects are as important as direct effects. In this study, we performed an experiment in artificial microcosms to disentangle biodiversity-mediated and direct effects of two major environmental disturbances on the functioning of aquatic ecosystems: increases in temperature and salinity. The ecosystem function analyzed was the microalgae predation by the zooplankton community (zooplankton grazing rates). Temperature and salinity increases affected the zooplankton grazing rates due to changes in community composition and abundance, as well as organism performance. The impact of salinity changes on community structure was higher than that of temperature; however, the importance of biodiversity-mediated and direct effects was similar to regulating the ecosystem functioning, albeit they have presented different directions and magnitude across the treatments. At a moderate level of temperature increase, we observed that the biodiversity-mediated effect was more relevant than the direct effect, with negative effects on the overall grazing rates. Our results suggest that disturbances can affect the functioning of aquatic environments through a set of complex biological mechanisms that balance direct and biodiversity-mediated effects. We concluded that the relative importance of biodiversity-mediated effects depends on the type and intensity of the disturbance.

Concerns regarding water scarcity, food security, and land/soil degradation are increasing, and these issues are far from being independent of bioenergy production. The competing need for land and water resources in food and bioenergy production have been at the forefront of policy debates. In this study, we examined policy debates related to land, food and water use in sugarcane ethanol production, and the challenges to integrated landwaterenergy-food policies. We analyzed large amounts of data obtained from governmental and business documents, Brazilian newspapers, and the bulletins of non-governmental organizations over the last ten years by combining an unsupervised probabilistic latent Dirichlet allocation (LDA) model with sentiment analysis. The results show that land, energy and water are managed in isolation by separate and disconnected institutional entities. Although the discourses of actors have evolved over time in response to certain events (e.g., expansion of sugarcane and water scarcity), the discussions and resultant public policies have failed to consider the interdependence among various sectors. To achieve sustainable resource use, sugarcane ethanol expansion should be considered by decision-makers and companies within a wider governance framework based on nexus dynamics; specifically, trade-offs in land and water use with sectors beyond bioenergy must be acknowledged.

Roughly 60% of all deforested lands in the Brazilian Amazon are covered with pastures, putting cattle raising in evidence as a major driver of deforestation and also of forests’ regrowth. Still, the role of cattle raising diversity in the landscape dynamics of this region remains poorly understood. To contribute to this discussion, we combined data from semi-structured interviews and quantitative spatially explicit methods to characterize and spatialize cattle raising systems and explore the effects of this diversity over secondary vegetation between 2004 and 2014 in Para, a hotspot of deforestation and core region of cattle production. We quantified the use of different pasture management strategies to classify small- and large-scale operations into systems with high or low impact against pastures’ degradation. High-impact systems were mapped in regions with consolidated infrastructure and high accumulated deforestation, where they expanded. On the contrary, low-impact systems were more widespread and found near forest frontiers, shrinking over time. High-impact systems had less secondary vegetation, while under low-impact systems, as a result of strategies with little or no effect against degradation, the historical pattern of concentration of this cover prevailed. Better infrastructure and access to markets as well as higher accumulated deforestation are underlying conditions related to the emergence of intensification and, as it is still unclear whether intensification is indeed capable of sparing land, the expansion of intensive cattle raising systems has the potential to configure landscapes with reduced forested areas, either primary or secondary.

In this study, the erosion of the nocturnal boundary layer (NBL) was analyzed in the central Amazon during the dry season of 2014, using data from the GoAmazon 2014/5 Project and high-resolution model outputs (PArallelized Les Model PALM). The dataset consisted of in situ (radiosonde) and remote sensing instruments measurements (Ceilometer, Lidar, Wind Profiler, microwave radiometer, and SODAR). The results showed that the NBL erosion occurred, on average, two hours after sunrise (06:00 local time), and the sensible heat flux provided more than 50% of the sensible heating necessary for the erosion process to occur. After the erosion, the convective phase developed quickly (175.2 m h-1). The measurements of the remote sensors showed that the Ceilometer, in general, presented satisfactory results in relation to the radiosondes for measuring the height of the planetary boundary layer. The PALM simulations represented well the NBL erosion, with a small underestimation (≈ 20 m) at the beginning of this phase. In the final phase of NBL erosion and in the initial stage of the development of the convective boundary layer (CBL), the model presented satisfactory results, with heights of CBL ranging from 800 m to 1,650 m, respectively.

Roughly 60% of all deforested lands in the Brazilian Amazon are covered with pastures, putting cattle raising in evidence as a major driver of deforestation and also of forests’ regrowth. Still, the role of cattle raising diversity in the landscape dynamics of this region remains poorly understood. To contribute to this discussion, we combined data from semi-structured interviews and quantitative spatially explicit methods to characterize and spatialize cattle raising systems and explore the effects of this diversity over secondary vegetation between 2004 and 2014 in Para, a hotspot of deforestation and core region of cattle production. We quantified the use of different pasture management strategies to classify small- and large-scale operations into systems with high or low impact against pastures’ degradation. High-impact systems were mapped in regions with consolidated infrastructure and high accumulated deforestation, where they expanded. On the contrary, low-impact systems were more widespread and found near forest frontiers, shrinking over time. High-impact systems had less secondary vegetation, while under low-impact systems, as a result of strategies with little or no effect against degradation, the historical pattern of concentration of this cover prevailed. Better infrastructure and access to markets as well as higher accumulated deforestation are underlying conditions related to the emergence of intensification and, as it is still unclear whether intensification is indeed capable of sparing land, the expansion of intensive cattle raising systems has the potential to configure landscapes with reduced forested areas, either primary or secondary.

This study applies three different methods to assess the flood risk and damage from the strongest high-pressure cold front (locally known as ‘Norte’) event in terms of the residual tide from 30 years (1979-2008) of data for Progreso, Yucatan. The most important difference between the three methods is the estimation of flood vulnerability for Progreso. The first method, proposed by Mexico’s National Center for the Prevention of Disasters (CENAPRED) and used by the Mexican government is based mostly on economic asset (household goods) values and flood impacts. The second (CENAPREDv2) and third (FRI) methods are proposals for assessing risk that include 17 socioeconomic indicators. The former includes economic asset values, as is the case for CENAPRED, while the latter does not. The main results of this study show that the modeled ‘Norte’ event flooded 25% of Progreso’s city blocks, with an estimated economic flood risk of $USD 16,266 (CENAPRED) and $USD 223,779 (CENAPREDv2), and flood damage of $USD 48,848 and $USD 671,918, respectively. When calculating flood risk (FRI) and flood damage (FRI_FD) without monetary terms, the risk categories along the back-barrier behind Progreso varied spatially from ‘very low’ to ‘high’, while areas along the coastal side presented a ‘low’ and ‘very low’ risk. These categories increased for the flood damage because the exceedance probability of the flood was not considered as it was for flood risk in the three methodologies. Therefore, flood damage provides the losses caused by a given flood event without considering how probable that loss may be. In conclusion, this study proposes that the selection of the applied method depends on the main objectives and specific interests when assessing flood risk. For instance, if economic damage is the main concern, then the CENAPRED method should be used as it identifies where the larger economic impacts could occur; when a socioeconomic approach is needed then the FRI should be applied, but if both economic damage and socioeconomic aspects are needed, the CENAPREDv2 is recommended. Besides considering economic aspects, the FRI method also includes social variables that can help to map the most vulnerable population in terms of mobility, education, communication access and others. Therefore, the proposed FRI method is very relevant for disaster risk managers and other stakeholders interested in disaster risk reduction.