RESUMO
During the 3 years of the ObseRvations of Aerosols above CLouds and their intEractionS (ORACLES) campaign, the NASA Orion P-3 was equipped with a 2D stereo (2D-S) probe that imaged particles with maximum dimension (D) ranging from 10 < D < 1280 µm. The 2D-S recorded supermicron-sized aerosol particles (SAPs) outside of clouds within biomass burning plumes during flights over the southeastern Atlantic off Africa's coast. Numerous SAPs with 10 < D < 1520 µm were observed in 2017 and 2018 at altitudes between 1230 and 4000 m, 1000 km from the coastline, mostly between 7-11° S. No SAPs were observed in 2016 as flights were conducted further south and further from the coastline. Number concentrations of refractory black carbon (rBC) measured by a single particle soot photometer ranged from 200 to 1200 cm-3 when SAPs were observed. Transmission electron microscopy images of submicron particulates, collected on Holey carbon grid filters, revealed particles with potassium salts, black carbon (BC), and organics. Energy-dispersive X-ray spectroscopy spectra also detected potassium, a tracer for biomass burning. These measurements provided evidence that the submicron particles originated from biomass burning. NOAA Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) 3 d back trajectories show a source in northern Angola for times when large SAPs were observed. Fire Information for Resource Management System (FIRMS) Moderate Resolution Imaging Spectroradiometer (MODIS) 6 active fire maps showed extensive biomass burning at these locations. Given the back trajectories, the high number concentrations of rBC, and the presence of elemental tracers indicative of biomass burning, it is hypothesized that the SAPs imaged by the 2D-S are examples of BC aerosol, ash, or unburned plant material.
RESUMO
Recently, we have witnessed Deep Learning methodologies gaining significant attention for severity-based classification of dysarthric speech. Detecting dysarthria, quantifying its severity, are of paramount importance in various real-life applications, such as the assessment of patients' progression in treatments, which includes an adequate planning of their therapy and the improvement of speech-based interactive systems in order to handle pathologically-affected voices automatically. Notably, current speech-powered tools often deal with short-duration speech segments and, consequently, are less efficient in dealing with impaired speech, even by using Convolutional Neural Networks (CNNs). Thus, detecting dysarthria severity-level based on short speech segments might help in improving the performance and applicability of those systems. To achieve this goal, we propose a novel Residual Network (ResNet)-based technique which receives short-duration speech segments as input. Statistically meaningful objective analysis of our experiments, reported over standard Universal Access corpus, exhibits average values of 21.35% and 22.48% improvement, compared to the baseline CNN, in terms of classification accuracy and F1-score, respectively. For additional comparisons, tests with Gaussian Mixture Models and Light CNNs were also performed. Overall, the values of 98.90% and 98.00% for classification accuracy and F1-score, respectively, were obtained with the proposed ResNet approach, confirming its efficacy and reassuring its practical applicability.