Repositório RCAAP

This thesis covers three major topics related to wildfires, remote sensing and meteorology: (i) quantifying and forecasting fire danger combining numerical weather forecasts and satellite observations of fire intensity; (ii) mapping burned areas from satellite observations with multiple spatial and spectral resolution; and (iii) modelling fire progression taking into account weather conditions and fuel (vegetation) availability. Regarding the first topic, an enhanced Fire Weather Index (FWI) is proposed by using statistical methods to combine the classical FWI with an atmospheric instability index with the aim of better forecasting the fire danger conditions favourable to the development of convective fires. Furthermore, the daily definition of the classical FWI was extended to an hourly timescale, allowing for assessment of the variability of the fire danger conditions throughout the day. For the second topic, a method is proposed to map and date burned areas using sequences of daily satellite data. This method, tested over several regions around the globe, provide burned area maps that outperform other existing methods for the task, particularly regarding the consistency and accuracy of the date of burning. Furthermore, a method is proposed for fast assessment of burned areas using 10-meter resolution satellite data and making use of Google Earth Engine (GEE) as a tool for preprocessing and downloading of data that is then used as input to a deep learning model that combines a coarse burned area map with the medium resolution data to provide a refined burned area map with 10-meter resolution at event level and with low computational requirements. Finally, for the third topic, a method is proposed to estimate the fire progression over a 12-hour period with resource to an ensemble of models trained based on the reconstruction of past events. Overall, I am confident that the results obtained and presented in this thesis provide a significant contribution to the remote sensing and wildfires scientific community while opening interesting paths for future research on the topics described.

No summary/description provided

No summary/description provided

A escassez de estudos sistemáticos sobre a utilização pedagógica de software educativo e a quase ausência de padrões de qualidade pedagógica, foram algumas das constatações abordadas na reflexão que tivemos oportunidade de fazer no primeiro seminário SACAUSEF1 e que, em nossa opinião, continuam a justificar a necessidade de trabalho cuidado e em profundidade no domínio da avaliação da qualidade deste tipo de materiais de apoio ao ensino e à aprendizagem (Costa, 2005). Como nessa altura também defendemos, é necessário, por outro lado, que os critérios de avaliação usados acompanhem a evolução tecnológica que se tem verificado nos últimos tempos e nos permitam, igualmente, uma avaliação adequada da qualidade dos diferentes tipos de conteúdos educativos cada vez em maior número disponíveis na Internet. No entanto, como estes conteúdos podem assumir propósitos e configurações muito diversas, será prudente que qualquer sistema de avaliação comece por definir o que se entende por “conteúdos educativos” e, só a partir daí, estabeleça os critérios de qualidade mais pertinentes e relevantes.

Malaria is a vector-borne disease that affected up to 241 million people just last year, causing more than 602 thousand deaths in Africa alone. Global strategies for malaria control include artemisinin-based combination therapies (ACTs) as the recommended first line treatment for uncomplicated Plasmodium falciparum malaria. Unfortunately, drug resistance remains a recurring obstacle. An area still under-studied is the host-pathogen interactions in terms the emergence of drug resistance. Host pharmacogene variation can influence the dynamics of drug resistance selection. Therefore, the dual aim of this thesis is to identify new genetic causality for Plasmodium falciparum resistance against present and in the pipeline combination therapies, while considering the scarcely studied contribution of the patient pharmacogenetic profile. In Chapter I, by targeted sequencing we profiled main antimalarial drug-related pharmacogenes of a Gabonese population. Out of the 347 identified genetic variants, 18 were novel with notable deleterious predicted consequences. Importantly, 16.7% were population-specific, highlighting the diversity of the understudied African genetic landscape. In Chapter II we accessed the influence of CY2C8 polymorphisms on the efficacy and tolerability of artesunate-amodiaquine (ASAQ) treatment in Zanzibar. The notable prevalente CYP2C8*2 and CYP2C8*3 reduced-activity alleles, 17.5% and 2.7% respectively, were associated with increased occurrence adverse events, indicating the influence patient genotype status on ASAQ tolerability. In Chapter III we monitored the spread of piperaquine-related molecular markers in African settings. The unique long active follow-up of the trial allowed to capture the increase prevalence of pfcrt 76T and 356T, significantly associated with shorter re-infection time. In addition, the background increased of parasites harbouring copy number variation (CNV) of pfpm3, spotlight the need of extended follow-ups when treatment strategies include long half-live partner drugs. In Chapter IV we determined baseline frequencies of plasmepsins 2/3 CNV in Tanzania and Kenya and evaluated current distribution across Africa based on a literature review.

Human activities are causing major changes in the composition of species communities worldwide. Oceanic islands are at the epicentre of these changes, being among the most threatened ecosystems. Drastic changes in island bird communities resulting from both species extinctions and introductions are likely to have important repercussions on the composition and function of island avifaunas, potentially causing severe knock-on effects on island ecosystems. The main objective of this thesis was to explore the taxonomic and functional implications of bird extinctions and introductions for the avifauna of oceanic islands worldwide. It focuses on large oceanic islands (> 100 km2; n = 74), distributed across three oceans. Using island lists of extinct and extant, native and introduced species, plus species functional traits and island characteristics, the thesis first addresses the potential competition between native and introduced island bird species. Then, it evaluates the impacts of bird extinctions and introductions on island functional diversity, investigating island-level consequences, taxonomic and functional homogenization among archipelagos, and patterns and drivers of spatial and temporal taxonomic and functional changes. Overall, the results suggest that marked niche differentiations between native and introduced birds explain the apparent rarity of competition-driven extinctions. Similarly, introduced birds differ from extinct birds because they mostly prefer human-modified landscapes, which might explain why they do not compensate for the lost functional diversity due to extinctions. Island bird assemblages are becoming functionally poorer because they are losing unique species while gaining widespread functionally more redundant species, and therefore are becoming taxonomically and functionally homogenized. Anthropogenic pressure and human connectivity are driving the taxonomic and functional changes across oceanic islands. Assessing temporal changes in taxonomic and functional diversity is key to understand anthropogenic impacts, predict future changes and propose effective conservation tools to safeguard the resilience of ecosystem functioning.

The critical status of the European eel (Anguilla anguilla) population is an ongoing concern for fisheries and environmental managers. The European Eel Regulation approved in 2007 established the framework for the eel recovery, although incomplete knowledge about the stock hampers management that ensure the species’ sustainability. The present research investigated the dynamics of the European eel in a coastal lagoon, contributing to increase the knowledge about the species and to the assessment of the status of the stock in Portuguese inland waters. The information gathered on eel recruitment, and on the species exploitation and life history traits in a Portuguese brackish system, the Santo André lagoon, provided relevant data to develop stock assessment tools, and to analyse and understand the impact of management decisions on the local eel stock. Data on glass eel from the rivers Minho and Lis provide insights on recruitment trends in Portugal in recent decades and on the drivers of glass eel ingress into Portuguese coastal systems and were used as a proxy to the recruitment dynamics in Santo André lagoon. Results suggest that eel recruitment shows no evidence of a dramatic decline in Portugal, compared to trends reported for other European rivers. The dynamics of the eel population in continental waters was studied from a socio-economic and biological perspective in Santo André lagoon, where the eel yield is significant. The fishery statistics since the 1980’s and the field work conducted in 2011/12 and in 2015/16 have made possible to gather relevant information on the exploitation and biology and have provided the necessary data to characterize the local eel population. Data collected were used to calibrate a Bayesian state-space Integral Projection Model (SSIPM) that described the eel dynamics in the lagoon and estimated relevant eel biomass indicators. The eel population in the lagoon is male-skewed and characterized by a relatively young population (average age is 2.3 years), with only 13.7 % of individuals above 350 mm (13.7%). Eel growth in the lagoon is among the fastest reported for the species and explains the early onset of maturity in males, which metamorphose into silver eels, on average, at three years of age. The male dominance in the lagoon is most likely related to the high eel density estimates obtained with the SSIPM, ranging between 6 and 139 kg ha-1 in the years studied (2008-2017). The high productivity of the system supports one of the highest eel fishing yields in the species’ range, which represents a substantial share (50%) of local fishermen’s income. The increase in the catch per unit effort between 2006 and 2017 obtained from fisheries statistical data, was hypothesized to be a consequence of fishing effort reduction under the eel regulation, combined with an increase in natural recruitment. These results point out to the positive signs of the adaptive approach adopted in local fisheries management, although the eel regulation targets (silver eel biomass escapement) may be compromised. The closed nature of the lagoon translates into a late escapement of silver eels, which occurs only in the spring when the connection to the sea is artificially re-established, making them very vulnerable to capture by the fishery. Despite being an illegal activity, silver eel fishing occurs, showing that the management system must be improved. There are conditions to implement fisheries co-management in this socio-ecological system, and the estimated fishing yield and silver eel escapement under different management scenarios suggest that there are opportunities to improve the sustainability of the local European eel population. The SSIPM developed proved to be a useful tool to describe the European eel population in Santo André lagoon and can be used to support the evaluation of the Portuguese eel management plan, particularly in coastal brackish systems. The framework can be adapted to other eel habitats where eel abundance time-series are available, ideally incorporating information on sex ratio, growth, and length at silvering of the local population, since those life history traits are environmentally dependent.

In the marine environment, sound can be an efficient source of information. Indeed, several marine species, including fish, use sound to navigate, select habitats, detect predators and prey, and to attract mates. Therefore, all the abiotic, biotic and manmade sounds that comprise the soundscape, have the potential to be used to assess and monitor species and marine environments. Passive acoustic monitoring (PAM) involves the use of acoustic sensors to record sound in the environment, from which relevant ecological information can be inferred. This thesis studied marine soundscapes, with special attention on fish communities, anthropogenic noise, and applied several methods to analyse acoustic recordings. Most of the focus was on the Tagus estuary, where the presence of two highly vocal species is known: the Lusitanian toadfish (Halobatrachus didactylus) and the meagre (Argyrosomus regius). Azorean and Mozambique soundscapes were also analysed. Several methods were applied to extract information and to visualize soundscape characteristics, including sound recognition systems based on hidden Markov models to recognize fish sounds and boat passages. Analysis of several types of marine environments and time scales showed several advantages and disadvantages of different methods. The use of sound pressure level on different frequency bands allowed the quantification of daily and seasonal patterns. Ecoacoustic indices appear to be cost-effective tools to monitor biodiversity in some marine environments. Using automatic recognition, vocal rhythms (diel and seasonal patterns) and vocal interactions among individuals were also characterized. Furthermore, boat noise effects on fish were studied: we encountered impacts on the audition, vocal behaviour and reproduction. Overall, we used PAM as a tool to remotely assess and monitor soundscapes, biodiversity, fish communities’ seasonal patterns, fish behaviour, species presence, and the effect of anthropogenic noise aiming to contribute for the management and conservation of marine ecosystems.

Seabirds are marine top predators regarded as indicators of the environmental changes occurring in their supporting ecosystems. The analytical lens of this thesis focusses on seabird belonging to the order Procellariiformes, which have similar life-histories characterised by high life expectancy and delayed sexual maturity. Furthermore, despite acting as central place foragers during breeding, most procellariiform seabirds can perform foraging trips covering thousands of kilometres by extracting energy from the wind through a flight behaviour known as "dynamic soaring". The overarching aim of my thesis is to understand the pathways through which wind and oceanographic processes affect the demography, population dynamics, foraging ecology and spatial distribution of seabirds. Focussing on the black-browed albatross (Thalassarche melanophris) as a model organism, we developed integrated population models to investigate the effects of wind and oceanographic fluctuations on the population breeding and survival processes. By analysing a demographic database spanning nearly two decades, we found that the population breeding parameters were negatively impacted by higher sea surface temperatures and positively affected by stronger winds, presumably through bottom-up environmental processes modulating food availability and accessibility. Survival was relatively constant and was only influenced by deeper ecosystem changes acting at larger spatio-temporal scales. Furthermore, our results revealed the high sensitivity of the population to the survival rate of the poorly understood sub-adult life history stages, which comprised approximately half of the total population size. We then studied the occurrence of albatross chick mortality events not caused by predation. Our results showed that, while albatross chicks weighed less in years with warmer sea temperatures, chick malnutrition and environmentally-driven food regulation did not explain the observed patterns of mortality. Rather, nestlings mortality events unrelated to predation were clustered at small scales in time and space, suggesting that part of the pronounced inter-annual variability in albatross breeding success was modulated by the prevalence of an unidentified infectious disease. By developing state-space models, we quantified a previously hypothesised, but never empirically documented "habitat-mediated" pathway linking environmental conditions to the breeding processes of a social monogamous population. Specifically, we found a higher prevalence of divorce in challenging years characterised by warmer sea surface temperatures, documenting the direct disruptive effects of ocean warming on the social monogamous bonds of albatrosses. Our work then focussed on the hypermobile Desertas petrel (Pterodroma deserta) and Bulwer's petrel (Bulweria bulwerii) as model organisms to investigate role of winds in shaping the flight behaviour and the foraging ecology of dynamic soaring seabirds during the breeding season. Desertas petrels used favourable winds to maximise their ground speed and distance covered throughout their round-trip foraging movements, among the longest recorded in any animal. Bulwer's petrels, on the other hand, exploited the stable North Atlantic trade winds, exhibiting a striking selectivity for crosswinds and engaging in crosswind zig-zag flight throughout large sections of their tracks. Under stable winds, this strategy enabled them to maximise the distance travelled and the probability of detecting odour plumes along the round trip. Crucially, the movement patterns of these two species suggest that seabirds have a priori knowledge of the regional winds and can plan their round-trip with an expectation of predicted wind conditions and costs of flight to return back to their colony. Collectively, the findings of my thesis highlight the sensitivity of seabirds to changes in oceanographic conditions and their reliance on winds to sustain their extreme life-history. Given the accelerating pace of global change and its dramatic effects on marine ecosystems, monitoring the diagnostic responses of these "sentinels" of the global ocean and, crucially, predicting their future performance is a conservation goal of upmost importance.

In this thesis we obtain an abstract continuity theorem for the drift associated with a product of isometries in both Gromov hyperbolic spaces and symmetric spaces as well as the Lyapunov exponents for a product of linear operators over some Hilbert space. We obtain these results by following a recipe of having large deviations estimates and an avalanche principle; a result which allows us to take conclusion of global nature from local hypothesis. As a main example, we apply the results to cocycles over Markov systems, where we prove the aforementioned large deviations estimates hold, thus providing a large class of examples. Upon presenting the linear setting we also mention the case of quasi-periodic linear cocycles. Whilst exploring Markov systems we also obtain a Fürstenberg type formula. From the perspective of Gromov hyperbolic spaces, we prove their group of isometries is a topological group and how random products of isometries follow a multiplicative ergodic theorem for the drift, thus describing the behaviour of typical orbits.

Breast cancer represents the main cause of cancer-related deaths in women. Nonetheless, the mortality rate of this disease has been decreasing over the last three decades, largely due to the screening programs for early detection. For many years, both screening and clinical diagnosis were mostly done through Digital Mammography (DM). Approved in 2011, Digital Breast Tomosynthesis (DBT) is similar to DM but it allows a 3D reconstruction of the breast tissue, which helps the diagnosis by reducing the tissue overlap. Currently, DBT is firmly established and is approved as a stand-alone modality to replace DM. The main objective of this thesis is to develop computational tools to improve the visualization and interpretation of DBT data. Several methods for an enhanced visualization of DBT data through volume rendering were studied and developed. Firstly, important rendering parameters were considered. A new approach for automatic generation of transfer functions was implemented and two other parameters that highly affect the quality of volume rendered images were explored: voxel size in Z direction and sampling distance. Next, new image processing methods that improve the rendering quality by considering the noise regularization and the reduction of out-of-plane artifacts were developed. The interpretation of DBT data with automatic detection of lesions was approached through artificial intelligence methods. Several deep learning Convolutional Neural Networks (CNNs) were implemented and trained to classify a complete DBT image for the presence or absence of microcalcification clusters (MCs). Then, a faster R-CNN (region-based CNN) was trained to detect and accurately locate the MCs in the DBT images. The detected MCs were rendered with the developed 3D rendering software, which provided an enhanced visualization of the volume of interest. The combination of volume visualization with lesion detection may, in the future, improve both diagnostic accuracy and also reduce analysis time. This thesis promotes the development of new computational imaging methods to increase the diagnostic value of DBT, with the aim of assisting radiologists in their task of analyzing DBT volumes and diagnosing breast cancer.

Not many documents qualify as material evidence of 17th century holographs in Portugal. However, personal miscellanies can provide valuable insight into textual composition and transmission in the literary scene of the time. One of the examples worth exploring is a miscellaneous volume of papers collected by D. Francisco Manuel de Melo (1608-1666), which includes several manuscripts either written by the owner himself or by the hand of fellow contemporary poets. These are clean copies with few or no layers of revision, but still relevant to a certain kind of genetic criticism without drafts (GRÉSILLON, 1993), focused on the interstices of scribal activity and distributed authorship in the early modern period. The article examines little-explored documentation from the national archive Torre do Tombo, thus broadening the scope of genetic studies into the realm of Portuguese baroque literature.

Apesar de permitir o acesso a quantidades inimagináveis de recursos, a Internet traz, na maior parte das vezes, grandes problemas em termos de selecção da informação, independentemente do tipo de utilizadores e dos objectivos que justifiquem a sua utilização. Na Escola, para fins educativos, uma das estratégias mais interessantes para ajudar os alunos a tirarem partido da riqueza de informação disponível são as WebQuests, também conhecidas, em língua portuguesa, por Aventuras na Rede. Muito embora sejam uma proposta de trabalho originalmente pensada para orientar os alunos na pesquisa, são bastantes os desafios que a sua concepção e desenvolvimento coloca também aos professores. Partindo das principais vantagens para os alunos, normalmente associadas à criação e utilização das WebQuests, pretende-se perspectivar também as suas potencialidades em termos dos próprios professores e da sua preparação para uma utilização mais esclarecida e crítica do uso do computador e da própria Internet para fins educativos.

In this work, a fully combined tsunami model was developed, by coupling a sequence of 3 distinct numerical models, with different characteristics, for particular phases of the tsunami lifecycle. The computational codes that compose the fully combined tsunami model are the GeoClaw code, the FUNWAVE-TVD code and the OpenFOAM code, via the olaFlow solver. The coupling of GeoClaw with FUNWAVE-TVD was designated as the combined model 1 (CM1) and the combination of FUNWAVE-TVD/CM1 with the CFD code was designated as the combined model 2 (CM2). The full combination of both CM1 and CM2 resulted in the fully combined tsunami model CM. To achieve the coupling between numerical models, individual coupling methodologies were approached, tested and analysed. For the CM1, we choose a refined covered gauge domain coupling methodology and for the CM2 a timeSeries condition coupling methodology was used, which applied waveType wavemaker and the waveTheory tveta, from the olaFlow module. The validation of the individual numerical codes and of the combined model patches was performed with both numerical and physical test cases. Several physical experiments were carried out to generate both solitary and N-waves and a novel first-order theoretical formulation, necessary to generate N-waves experimentally, by means of a piston wave generating system, was developed and detailed in this work. The large-scale physical experiments were performed in the wave basin and in a beach composed by a 1:15 plane slope and a 1:30 plane slope. The generated solitary and N-waves were classified according to their Stokes number. Experimental free surface elevation, run-in, run-up and pressure measurements were retrieved from the physical experiments. Run-in, run-up and pressure laws were proposed for solitary waves and N-waves respectively. The experimental measurements were compared with numerical simulation results. The objectives of the development of the fully combined tsunami model were (1) to join the advantages of the individual models in a single one, attempting to increase the accuracy, efficiency and regime of validity, and (2) to bring a contribution in the tackling of some of the existing problems and challenges of tsunami science, such as the frequency dispersion in long distance tsunami propagation, the complex tsunami on land propagation and fluid flow interactions with river courses and with the coastal and urban areas. The fully combined tsunami model CM simulation results for a Mω 8.5 Earthquake and Tsunami hitting the Portuguese coast showed the ability of the combined model to cover all the tsunami stages. We show that with a 2DV simulation of the CFD code for the Marina of Cascais bathymetric and topographic profile it was possible to observe the vortices behind the breakwater. The analysis of the free surface elevation, velocities and pressure of the tsunami waves was performed. This allowed us to understand the consequence of three diferent tsunami waves scenarios after the breakwater zone. It was possible to draw some brief conclusions considering the tsunami impact. The fully combined tsunami model achieved in this work is a novelty, since it is composed by a sequence of distinct numerical models, including the three-dimensional component granted by the CFD code. With this combined model, it is possible to perform the simulation of real case tsunami events and hypothetical scenarios, applying real or synthetic tsunami-type wave profiles, studying and researching the impact and the tsunami interaction with the coastal areas.

Extreme wildfires are increasingly rising to intense and uncontrolled fires, with dimension and destructive potentials that are greater than what has been seen and dealt with. The hazards posed by these fires increase significantly when they approach the wild–urban interface, with relevant environmental and socio-economic consequences. The 2009 Victorian bushfires and the 2017 Portugal wildfires are powerful reminders, and they have demonstrated the need to better understand why mitigation plans have failed to protect the community in these events and to improve community resilience. The year 2017 is a milestone in the history of wildfires in Portugal, not only because of the vast burned area but also due to the high number of fatalities. The two occurrences were at different times (June and October) but were geographically close (region of centre of Portugal). A total of 117 deaths occurred in both events and 92% of the victims were in wild–urban interface areas. This paper analyses and discusses the characteristics and causes of death of the victims of these two events: age, place of death, distance from place of death to place of residence and last-minute choices to aim to understand the actions that people took in the face of the approaching fire, which led to their death. In both cases, most people died fleeing the fire without any information from the competent authorities. In the end, it is possible to identify risk factors that lead to the death of civilians due to wildfires, such as the increase in demand for rurality by young people from big cities with no previous contact with wildfires; on the other hand, there is the ageing of the population residing in forest areas, who were previously physically and structurally prepared to deal with fires and are currently no longer able to.

No decurso da investigação em torno da luminária da extinta Casa Real Portuguesa, constatamos a existência de uma quantidade significativa de luminária francesa neorococó do último quartel do século XIX nos palácios nacionais. A mesma tendência foi seguida na Alemanha; esta propensão em comum intensificou-se após a guerra franco-prussiana (1870-1871). Nos dois países desenvolveram-se determinadas características que importam dar aqui a conhecer. Este estudo é sustentado pelo cotejamento de bibliografia e documentação coeva, foca-se nas diferentes abordagens e contextualizações com novas reflexões, colmatando assim a escassez de estudos sobre luminária oitocentista.

Os programas de jovens ocupam hoje um lugar central na oferta educativa de muitos museus internacionais, em particular nos museus de arte contemporânea na América do Norte e no Reino Unido. Alinhados com as características específicas deste grupo etário, enquanto público independente, estes programas podem ser agrupados em três grandes tipologias – pontual, curta duração e longa duração. À dimensão temporal aliam-se objetivos e estratégias específicos que pretendem oferecer aos jovens diferentes possibilidades de acesso e participação nos museus, definindo o que se designa como programação por etapas. No contexto museológico português, o investimento na relação com os jovens, fora do âmbito escolar, é um fenómeno recente, mas em crescimento. Tendo por base a análise dos programas para jovens de cinco instituições da grande Lisboa: o Museu de Arte, Arquitetura e Tecnologia (MAAT), a BoCA – Biennial of Contemporary Art, a Casa da Cerca – Centro de Arte Contemporânea, as Galerias Municipais e o LU.CA – Teatro Luís de Camões, é discutido o potencial de uma programação por etapas interinstitucional.

For most hospitals, the operating room (OR) is a significant source of expenses and income. A critical point of effective OR scheduling is the prediction of OR time for a patient procedure. An inefficient schedule results in two scenarios: underestimated or overestimated OR times. A solution reported in the literature is the implementation of machine learning (ML) models that include additional variables to improve the accuracy of these predictions. This project goal is to improve the OR schedule efficiency in a hospital center by achieving precise OR time predictions. This goal was accomplished by developing two ML models (Multiple Linear Regression (MLR) and Random Forest (RF)), through two different approaches. Firstly, for all the specialties on the dataset (All Specialties Model). Second, a specialty-specific model for each (Urology, General Surgery, and Orthopedics Models). This leads to eight models where the predictive features were identified based on the literature along with consultations with the professionals. The All Specialties Model presented a surgery median time of 115.0 minutes, with an R-squared surrounding 0.7. Urology had a median time of 70.0 minutes, with an R-squared of 0.822 and 0.831 and a MAE of 21.7 and 20.9 minutes for MLR and RF models, respectively. General Surgery had a median time of 110.0 minutes with an R-squared of 0.826 and 0.825 and a MAE of 26.2 and 26.1 minutes for MLR and RF, respectively. For Orthopedics, the RF was the only one able to model all the data with an R-squared of 0.683 and a MAE of 27.1 minutes. When compared with the current methods, considering a 10% threshold, the models achieved reductions in underestimation surgeries (41%), and an increase of within predictions (19%). However, with a 22% increase in overestimation predictions. We conclude that using ML approaches improve the accuracy of OR time predictions.

En 2010, Erika Masanet y Beatriz Padilla publicaban un análisis comparativo de los flujos mi gratorios, factores motivacionales y perfiles socioeconómicos de los brasileños residentes en Portugal y España, aludiendo a la existencia de un “sistema ibérico de migración”. Esta expre sión se justificaba con arreglo a las similitudes existentes en el espacio ibérico como marco geográfico que, más allá del nivel nacional, constituye un vector de atracción migratoria para los ciudadanos brasileños. Doce años después, varias de las tendencias identificadas en aquel artículo se han consolidado, al tiempo que han emergido nuevos perfiles, ajustados a las nue vas dinámicas políticas, económicas y sociales que han marcado desde entonces la historia eu ropea. Es el caso de la crisis económico-financiera de 2008, la pandemia sanitaria provocada por el COVID-19 y, más recientemente, la Guerra de Ucrania y sus efectos sobre el mercado energético y la inflación en la zona euro. Ante este panorama general, se hace necesaria una reflexión, sobre las características y moti vaciones de la migración brasileña para y en el espacio ibérico, reflexión que justifica la elabo ración de este monográfico. Importa, pues, formular algunas cuestiones relevantes, a saber: ¿qué es lo que sabemos de la inmigración brasileña en España y en Portugal?, ¿qué temas y problemáticas han centrado las investigaciones sobre esta comunidad de inmigrantes en am bos países?, ¿qué nos queda por saber sobre estas migraciones en el contexto ibérico?, ¿cuáles son los retos que la movilidad ibero-brasileña representa en el futuro? La mayoría de las inves tigaciones se ha ocupado de los factores contextuales (condiciones políticas, sociales y econó micas de los países de origen y de destino), relacionales (redes de contacto, sobre todo, en el espacio transnacional) e individuales (capital cultural, intelectual y lingüístico) para explicar la elección y las trayectorias migratorias de los ciudadanos brasileños en Portugal (Malheiros, 2007; Pinho, 2014) y España (Solé et al., 2011).

A fragmentação sócio-espacial aumentou globalmente, juntamente com as crescentes desigualdades, a diversidade de estilos de vida e padrões de consumo. Entende-se como o resultado de processos que levam a formas de ocupação fragmentadas, novas realidades na esfera social e na produção e apropriação do espaço, com enclaves sem continuidade com as estruturas sócio-espaciais circundantes. A fragmentação é frequentemente estudada na perspectiva da habitação com a emergência de condomínios fechados, mas a minha intenção é integrar o retalho de luxo com o imobiliário devido à forte componente externa tanto na cadeia de fornecimento (desenvolvimento e processo de investimento) como na nova procura (turistas e residentes não permanentes) que traz para um novo nível os conflitos e contradições em torno do direito à Cidade, tendo Lisboa como estudo de caso. O foco principal é a mudança no mercado imobiliário e a oferta de novos tipos de produtos, complexos de utilização mista e condomínios fechados, tendências de consumo de luxo, a formação de um enclave de luxo dentro da cidade central de Lisboa, e o apoio fornecido pela política pública. A análise identifica tendências e desafios que requerem uma investigação renovada e respostas políticas até agora inexistentes ou simplesmente experimentadas sem avaliação