Drought conditions in Amazonia are associated with increased fire incidence, enhancing aerosol emissions with degradation in air quality. Quantifying the synergic influence of climate and human-driven environmental changes on human health is, therefore, critical for identifying climate change adaptation pathways for this vulnerable region. Here we show a significant increase (1.2%–267%) in hospitalisations for respiratory diseases in children under-five in municipalities highly exposed to drought. Aerosol was the primary driver of hospitalisations in drought affected municipalities during 2005, while human development conditions mitigated the impacts in 2010. Our results demonstrated that drought events deteriorated children's respiratory health particularly during 2005 when the drought was more geographically concentrated. This indicates that if governments act on curbing fire usage and effectively plan public health provision, as a climate change adaptation procedure, health quality would improve and public expenditure for treatment would decrease in the region during future drought events.

Between 1998 and 2017, climate-related disasters represented 91% of all occurrences worldwide, causing approximately US$ 2.245 billion of direct economic losses. In the Amazon region, fire is used as a widely spread technique for land clearing, agricultural management, hunting, and religious rituals. However, over the past 20 years, severe droughts caused a major amplification of fire occurrences, leading to several socioeconomic and environmental impacts. Particularly in Acre state, located in the southwestern Brazilian Amazon, the occurrence of extensive fires, associated with extreme climatic events, has been reported since 2005. However, fire dynamics, land tenure relationships, and associated impacts are poorly quantified. In this study, we aim to investigate the following: (1) The spatiotemporal variability of fire dynamics during anomalously dry and regular climate conditions; (2) the attribution of fire occurrence and land tenure relationship, and (3) the environmental, social, and economic impacts caused by fires and its consequences for Acre’s economy. We analyzed information on the spatial patterns of fire, its direct impacts on land use and land cover, carbon stocks, CO₂ emissions, the indirect impact on human illness, and finally the costs of these impacts from 2008 to 2012. During the studied period, burned areas were concentrated around the major cities and roads, forming polygons up to 0.6 km². However, in 2010, an extremely dry year, fires spread to remote areas, impacting protected private areas and sustainable-use conservation areas. In 2010, the total area affected by forest fires was approximately 16 times greater than in meteorologically normal years. The total economic loss estimated in 2010 was around US$ 243.36 ± 85.05 million and for the entire period, US$ 307.46 ± 85.41 million. These values represent 7.03 ± 2.45% and 9.07 ± 2.46% of Acre’s gross domestic product (GDP), respectively.

Fire is one of the main factors directly impacting Amazonian forest biomass and dynamics. Because of Amazonia’s large geographical extent, remote sensing techniques are required for comprehensively assessing forest fire impacts at the landscape level. In this context, Light Detection and Ranging (LiDAR) stands out as a technology capable of retrieving direct measurements of vegetation vertical arrangement, which can be directly associated with aboveground biomass. This work aims, for the first time, to quantify post-fire changes in forest canopy height and biomass using airborne LiDAR in western Amazonia. For this, the present study evaluated four areas located in the state of Acre, called Rio Branco, Humaitá, Bonal and Talismã. Rio Branco and Humaitá burned in 2005 and Bonal and Talismã burned in 2010. In these areas, we inventoried a total of 25 plots (0.25 ha each) in 2014. Humaitá and Talismã are located in an open forest with bamboo and Bonal and Rio Branco are located in a dense forest. Our results showed that even ten years after the fire event, there was no complete recovery of the height and biomass of the burned areas (p < 0.05). The percentage difference in height between control and burned sites was 2.23% for Rio Branco, 9.26% for Humaitá, 10.03% for Talismã and 20.25% for Bonal. All burned sites had significantly lower biomass values than control sites. In Rio Branco (ten years after fire), Humaitá (nine years after fire), Bonal (four years after fire) and Talismã (five years after fire) biomass was 6.71%, 13.66%, 17.89% and 22.69% lower than control sites, respectively. The total amount of biomass lost for the studied sites was 16,706.3 Mg, with an average loss of 4176.6 Mg for sites burned in 2005 and 2890 Mg for sites burned in 2010, with an average loss of 3615 Mg. Fire impact associated with tree mortality was clearly detected using LiDAR data up to ten years after the fire event. This study indicates that fire disturbance in the Amazon region can cause persistent above-ground biomass loss and subsequent reduction of forest carbon stocks. Continuous monitoring of burned forests is required for depicting the long-term recovery trajectory of fire-affected Amazonian forests.

Urban environments are regions in which spectral variability and spatial variability are extremely high, with a huge range of shapes and sizes, and they also demand high resolution images for applications involving their study. Due to the fact that these environments can grow even more over time, applications related to their monitoring tend to turn to autonomous intelligent systems, which together with remote sensing data could help or even predict daily life situations. The task of mapping cities by autonomous operators was usually carried out by aerial optical images due to its scale and resolution; however new scientific questions have arisen, and this has led research into a new era of highly-detailed data extraction. For many years, using artificial neural models to solve complex problems such as automatic image classification was commonplace, owing much of their popularity to their ability to adapt to complex situations without needing human intervention. In spite of that, their popularity declined in the mid-2000s, mostly due to the complex and time-consuming nature of their methods and workflows. However, newer neural network architectures have brought back the interest in their application for autonomous classifiers, especially for image classification purposes. Convolutional Neural Networks (CNN) have been a trend for pixel-wise image segmentation, showing flexibility when detecting and classifying any kind of object, even in situations where humans failed to perceive differences, such as in city scenarios. In this paper, we aim to explore and experiment with state-of-the-art technologies to semantically label 3D urban models over complex scenarios. To achieve these goals, we split the problem into two main processing lines: first, how to correctly label the façade features in the 2D domain, where a supervised CNN is used to segment ground-based façade images into six feature classes, roof, window, wall, door, balcony and shop; second, a Structure-from-Motion (SfM) and Multi-View-Stereo (MVS) workflow is used to extract the geometry of the façade, wherein the segmented images in the previous stage are then used to label the generated mesh by a “reverse” ray-tracing technique. This paper demonstrates that the proposed methodology is robust in complex scenarios. The façade feature inferences have reached up to 93% accuracy over most of the datasets used. Although it still presents some deficiencies in unknown architectural styles and needs some improvements to be made regarding 3D-labeling, we present a consistent and simple methodology to handle the problem.