Thesis (MA)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: During the period 2006-2012, the Langkloof area, situated in the Eden District Municipality, suffered severely from environmental shocks and social stress including drought, flooding, hail, wildfire, heatwaves and reduced labour demand. These events negatively impacted many farmers and their livelihoods. In response to these external shocks and stressors, large-, medium- and small-scale farmers adopted numerous coping and adaptive strategies. This study performed a comprehensive livelihoods analysis of large-, medium- and small-scale farmers in the Langkloof area, using the widely recognized sustainable livelihoods framework developed by the United Kingdom’s Department for International Development, with a particular focus on coping and adaptive strategies against severe environmental shock and social stress. Variables used in the analysis were the vulnerability context in which farmers pursue a livelihood; livelihood assets (social, human, financial, natural and physical); the policies, institutions and processes in the external environment that influence the degree of ownership and access to assets; livelihood strategies pursued; and the various livelihood outcomes ultimately produced. The analysis of coping and adaptive strategies employed by farmers during these periods formed an integral part of this study. Sixteen livelihood asset indicators were identified to determine the total assets (human, social, physical, financial and natural) of the farmers. After scaling the indicators, a Principal Component Analysis (PCA) was used to assign weights to each indicator and to subsequently calculate the total assets of each household. Regarding the coping and adaptive strategies employed by farmers against environmental shock and social stress, the average number of strategies was calculated for each household. Spearman’s rank correlations were calculated for total assets achieved (capacity) and the number of strategies employed against environmental shock and social stress. / AFRIKAANSE OPSOMMING: Die Langkloofgedied, geleë in die Eden Distriksmunisipalitiet, is gedurende die tydperk 2006-2012 geweldig geteister deur omgewings- en sosiale gebeure, wat droogte, vloede, hael, veldbrande, hittegolwe en ʼn gevolglike verlaagde arbeidsaanvraag sluit. Hierdie rampe het verreikende nadelige gevolge op boere se lewensbestaan gehad. Terwyl hulle lewensbestaan tot die uiterste uitgedaag is, is verskeie hanterings- en aanpassingstrategieë teen omgewingskok en sosiale stres geïmplementeer. Met die gebruik van die Verenigde Koninkryk se Departement van Internasionale Ontwikkeling se volhoubare lewensbestaansraamwerk is ʼn gedetailleerde en alomvattende analise van groot, medium- en kleinskaalboere in die Langkloof se lewensbestaan gedoen, met ʼn sterk skem op hul hanterings- en aanpassingstrategieë teen omgewingskok en sosiale stress. Die analise veranderlikes het die kwesbaarsheidskonteks waarin boere hul lewensbestaan aanpak; hul verskeie bates (menslik, sosiaal, finansieel, fisies en natuurlik); die beleide, instansies en prosesse in die eksterne omgewing wat toegang tot en eienaarskap van bates reguleer; hul lewensbestaanstrategieë; en die lewensbestaansuitkomste wat bereik word, behels. Die analise van die boere se hanterings- en aanpassingstrategieë geïmplementeer teen omgewingskok en sosiale stres gedurende hierdie tydperk was ʼn integrale rol van hierdie studie. Sestien aanwysers is geïdentifiseer om die totale lewensbestaansbates (menslik, sosiaal, fisies, finansieel en natuurlik) van die boere te bepaal. Nadat die aanwysers geskaal is, is ‘n Hoofkomponentanalise (PCA) uitgevoer om gewigte aan elke aanwyser toe te ken vir die berekening van totale bates van huishoudings. Die hanterings- en aanpassingstrategieë ingestel deur boere teen omgewingskok en sosiale stres, is die gemiddelde aantal strategieë per huishouding. Spearman se rangorde korrelasies is bereken vir die totale bates (kapasiteit) en die getal strategieë geïmplementeer teen omgewingskok en sosiale stres.

Submitted by Felipe Garcia (felipe.garcia.rs@gmail.com) on 2013-06-17T19:31:46Z No. of bitstreams: 1 Tese de Doutorado Felipe Garcia - Versão Final.pdf: 2832465 bytes, checksum: 1b420a2d381d534eb81e82c8a092fe26 (MD5) / Approved for entry into archive by Suzinei Teles Garcia Garcia (suzinei.garcia@fgv.br) on 2013-06-17T19:45:58Z (GMT) No. of bitstreams: 1 Tese de Doutorado Felipe Garcia - Versão Final.pdf: 2832465 bytes, checksum: 1b420a2d381d534eb81e82c8a092fe26 (MD5) / Made available in DSpace on 2013-06-17T19:48:12Z (GMT). No. of bitstreams: 1 Tese de Doutorado Felipe Garcia - Versão Final.pdf: 2832465 bytes, checksum: 1b420a2d381d534eb81e82c8a092fe26 (MD5) Previous issue date: 2013-06-07 / This thesis consists of three essays, one being on economic development and the two others on economic growth. The first essay is an investigation into how access to electricity by households affects time allocation of children and teenagers from rural Brazil in what regards the choice to attend school or to engage in the labor market. To establish a causal relationship between the availability of electricity and time allocation, we use the criteria of priority works from Luz Para Todos program as a source of exogenous variation in access to electricity of households located in rural Brazil. We apply Regression Discontinuity Design and Difference in Difference estimator with endogenous variables. The results from the latter methodology indicate that the presence of electricity increases the likelihood that children and adolescents are enrolled in school and are not delayed relative to the grade they should be attending at their age. Also, the probability of a child being literate and not being working increases with the presence of electricity. One of the possible channels that could explain these results is the increased participation of mothers in the labor market. However, one cannot discard the hypothesis that the observed results are justified by the increased access to electricity by the schools, making them better and more attractive to students and by the receipt of Bolsa Familia program which is conditional on school attendence by children and teenagers. The second study aims to contribute to the literature on institutions and economic growth through the investigation of how the 1959 socialist revolution in Cuba have affected the trajectory of per capita income in that country. To do so, we apply the synthetic control method to obtain the counterfactual path of income per capita in the absence of such revolution. The results show that the trajectory of the annual GDP per capita of Cuba between 1959 to 1980 was lower than it would have been if the socialist regime had not been implemented. This result is robust to different placebo tests deployed and also to the use of different series of GDP per capita. Alternative hypotheses that could explain the lower path of GDP per capita are discussed based on the existing literature and data. The discussion suggests that institutional change was indeed the cause of the observed negative effect on the Cuban economy during the period of analysis. Finally, the third article investigates the economic cost of a natural disaster in Brazil. The paper looks at the excessive rains that happened in the state of Santa Catarina in November and December 2008. This study contributes with some new evidence to the iv recent literature on natural disasters. To our knowledge, there are no studies in Brazil that empirically measure the impact that a natural disaster has entailed on an affected region. This study is structured in two parts. The first uses synthetic control to measure the impact of rainfall on the industrial production of Santa Catarina. For that, we propose a small accommodation method for treating the effects of possible flooding rains in other states. In the second part, we use the method of difference in difference to measure the impact of the rains on the GDP per capita of the cities. The results show that for a period of two years after the end of 2008, the impact of the rains caused a lower monthly industrial production of 2.0% in Santa Catarina. For municipalities, the estimated effect of the disaster on GDP per capita was around -7.0% in 2008 and -5.0% in 2009. There was not significant effect in 2010. / Esta tese é composta por três ensaios, um na linha de desenvolvimento econômico e dois na linha de crescimento. O primeiro deles trata de uma investigação sobre o papel do acesso à energia elétrica a nível domiciliar na alocação do tempo das crianças e adolescentes do Brasil rural entre frequentar a escola e participar do mercado de trabalho. Para o estabelecimento da relação causal entre energia elétrica e a alocação de tempo se utilizam os critérios de prioridade de obras do programa Luz Para Todos como fonte de variação exógena no acesso à energia elétrica dos domicílios localizados na zona rural do Brasil. Aplicam-se os métodos de Regressão Descontínua e Diferenças em Diferenças com Variáveis Instrumentais. Os resultados obtidos da segunda metodologia apontam que a presença de energia elétrica aumenta a probabilidade das crianças e adolescentes estarem matriculadas na escola, não estarem atrasadas em relação à série que deveriam estar dada sua idade, serem alfabetizadas e não estarem trabalhando. Entre os possíveis canais capazes de explicar estes resultados está a maior participação das mães no mercado de trabalho. Entretanto, não se pode descartar a hipótese de que os resultados observados sejam justificados pelo aumento do aceso à energia elétrica a nível escolar, o que pode tornar as escolas melhores e mais atrativas, ou consequência do aumento do recebimento do programa Bolsa Família que exige frequência escolar das crianças e adolescentes. Já o segundo ensaio objetiva colaborar à literatura de instituições e crescimento econômico através da investigação do impacto da mudança institucional ocorrida em Cuba após a revolução socialista de 1959 sobre a trajetória da renda per capita do país. Para tanto, aplica-se o método de controle sintético para a obtenção do contrafactual da trajetória da renda per capita na ausência da revolução. Os resultados apontam que a trajetória do PIB per capita anual de Cuba entre 1959 a 1980 foi inferior ao que teria sido caso não tivesse sido implantado o regime socialista. Este resultado é robusto aos diferentes testes de placebo implantados e ao uso de distintas séries do PIB per capita de Cuba. Hipóteses alternativas à mudança institucional, mas também capazes de explicar a trajetória inferior do PIB per capita, são discutidas com base na literatura e na descrição de dados. A discussão sugere que foi de fato a mudança institucional a causa do efeito negativo observado sobre a economia cubana durante o período de análise. Por fim, o terceiro estudo investiga o custo econômico de um desastre natural ocorrido no Brasil em 2008, a saber, o excesso de chuvas em Santa Catarina entre os meses de novembro e dezembro daquele ano. Este artigo colabora com mais uma evidência para a recente literatura de desastres naturais. No Brasil não há nenhum trabalho, sob nosso conhecimento, medindo empiricamente o impacto que um desastre natural tenha acarretado a uma região atingida. Este estudo está estruturado em duas partes. Na primeira se utiliza o controle sintético para medir o impacto das chuvas na produção industrial de Santa Catarina. Propõe-se uma pequena acomodação do método para tratar o possível transbordamento dos efeitos das chuvas nos demais estados. Já na segunda parte, utiliza-se o método de diferenças em diferenças para medir o impacto das chuvas no PIB per capita das cidades mais atingidas. Os resultados apontam que para um período de dois anos após o final de 2008, o desastre causou uma produção industrial mensal 2.0% menor do que seria caso as chuvas não tivessem ocorrido. Por municípios, o efeito estimado do desastre sobre o PIB per capita se situou ao redor de -7,0% em 2008 e -5,0% em 2009. Em 2010 não há evidências de efeito.

MURP / Department of Urban and Regional Planning / Disaster is posing serious threats to both human lives, infrastructure and the environment at large. Greater Tzaneen Municipality (GTM) is one of the many municipalities that suffer from flood related disasters. Lack of integration between Disaster Risk Reduction (DRR) and spatial planning has compounded the disaster risk situation in the municipality. This study sought to investigate the prospects of integrating spatial planning with disaster risk reduction in flood prone areas of GTM. The study is guided by three research objectives. First, the study sought to analyse spatial planning attributes that can be valorised for DRR in flood prone areas; Secondly, it sought to analyse spatial planning factors that define vulnerability attributes of households occupying flood prone areas. Finally, the study sought to perform a cluster analytical creation of a typology of households whose resilience to flooding could be enhanced through spatial planning. Twenty-five flood prone areas were analysed on the basis of four main flood vulnerability attributes. In order to identify such vulnerability attributes, the study borrowed critical insights from literatures on flood vulnerability, spatial planning and DRR. Such a critical review of literature was complemented by the use of pattern matching as a qualitative research instrument. Quantitative that was gathered using a structured observation checklist. Quantitative data generated was first subjected to various statistical tests that included Normality and Reliability Tests. Common measures of Normality test used included measures of skewness, kurtosis and the use of Normal Q-Q plots. To assess flood vulnerability, Hierarchical Cluster Analysis (HCA) was used. HCA was used to identify clusters of flood prone areas which had common characteristics in terms of the four main study constructs proposed by the study which included the physical/engineering, socio-economic, ecological/natural and political or governance conditions characterizing each area. HCA was then used to identify main clusters exhibiting similar characteristics and the associated level of vulnerability of such of communities occupying such clusters. Study results revealed 2 main clusters of flood prone areas whose differences lay in interactions that existed between the physical/engineering, socio-economic, ecological/natural and political or governance conditions characterizing each area. Such clusters depicted 2 levels of vulnerability that is high, and moderate. A number of opportunities and constraints were generated using the SWOT matrix strategy with the main results showing that spatial planning elements characterizing flood prone areas could be transformed into critical urban risk management options for DRR. This is because a spatial planning elements were found to have a direct influences on critical factors of DRR such as location of activities. The study concluded by recommending a number of spatial planning strategies that can be vaporized for DRR. Such strategies are systematically aligned to the unique vulnerability context conditions associated with the two flood vulnerability solution arrived at using HCA. / NRF

Le but de cette recherche est d’évaluer l’importance du paysage culturel dans la résilience des communautés urbaines post-catastrophes. Ce travail se concentre sur le quartier du Lower Ninth Ward dans la ville de La Nouvelle-Orléans (États-Unis) après le passage de l’ouragan Katrina en 2005. Les catastrophes naturelles prennent une envergure et causent des dommages considérables lorsqu’elles touchent des villes. La reconstruction post -désastre est donc très dispendieuse pour les villes et les gouvernements, d’autant que certaines régions sont dévastées au point qu’elles doivent être reconstruites au complet. Cependant, le coût le plus lourd à assumer reste celui en vies humaines et si rebâtir les éléments concrets d’une ville est une tâche difficile à entreprendre, reconstruire une communauté est considérablement plus complexe. Dans le but de comprendre une telle démarche, cette recherche se concentre sur les éléments intangibles, comme l’attachement au lieu et les réseaux sociaux, dont une communauté a besoin pour se reconstituer de façon durable et résiliente. Le concept de résilience est très contesté dans la littérature et plusieurs chercheurs se sont essayés à le mesurer. Cette recherche adopte une perspective critique sur le concept et le revisite d’un point de vue holistique pour mettre en lumière sa complexité. Cette démarche permet de remettre en question l’importance de mesurer un concept finalement en perpétuelle redéfinition dans le temps et selon les échelles géographiques. De plus, en établissant une relation entre résilience et paysage culturel, il a été possible de mieux comprendre la complexité de la résilience. Touchant à plusieurs disciplines (architecture de paysage, urbanisme et sociologie), cette recherche utilise une méthodologie qui reflète son aspect multidisciplinaire : les méthodes mixtes. Ces dernières permettent la collecte de données quantitatives et qualitatives qui produisent une vue globale de la situation post-Katrina à travers le regroupement de recensions statistiques, d’observations de terrain et d’articles de journaux. Parallèlement, des entretiens ont été réalisés avec des résidents du quartier ainsi qu’avec des professionnels pour mieux comprendre les différents points de vue. Cette méthodologie a permis de produire des résultats au niveau du cas d’étude autant qu’au niveau théorique. La recherche valide l’importance de prendre en compte le paysage culturel dans les situations post-catastrophes, (en particulier) dans la mesure où il s’agit d’un élément souvent négligé par les urbanistes et les acteurs locaux. En effet, les éléments constitutifs du paysage culturel tels que l’attachement au lieu et les réseaux sociaux, participent d’un sentiment d'appartenance (« home ») et d’une volonté, pour les résidents, de reconstruire leurs habitations, leur communauté ainsi que leur quartier. Toutefois, il faut reconnaître que ces éléments ne suffisent pas à retrouver ce qu’ils ont perdu. Ainsi, l’étude du paysage culturel permet non seulement de mieux comprendre la complexité de la résilience, mais démontre également que cette dernière est une construction sociale. / The purpose of this research is to determine the importance of using the cultural landscape in evaluating the resilience of an urban community after the occurrence of a natural disaster. The focus is on the neighborhood of the Lower Ninth Ward after Hurricane Katrina in 2005 in the city of New Orleans. Natural disasters are gaining significance and magnitude when they hit cities, which are becoming more and more populated over the years. The damage these disasters cause is colossal. It is very costly for cities to undergo major disasters and sometimes, large sections of cities need to be entirely rebuilt. The costliest price is the human life, and as history marks it, too many lives have perished due to disasters. While rebuilding is a challenging task, yet feasible, rebuilding a community is not as tangible as rebuilding the infrastructure. This research focuses on the many intangible aspects, like place attachment and social networks, a community needs to rebuild itself in a sound and resilient way. The concept of resilience is very contested in the literature and many have attempted to measure it. This research takes a step back and scrutinizes the concept of resilience from a holistic perspective, which highlights its complexity. This leads to questioning the importance of measuring the concept, especially that it changes with time and with the different scales of geography. In addition, a relationship between the cultural landscape and resilience is established, which allows for a better understanding of this complexity. Taking a little from multiple disciplines (Landscape Architecture, Urban Planning, and Sociology), this research resorts to a methodology that reflects its multidisciplinary aspect. The methodology is the mixed methods research design, which allows the collection of quantitative and qualitative data. The focus is to gather census data, newspaper articles, and observations to give a general perspective on the post-Katrina situation. Interviews are collected from residents and from professionals so as to tackle the research from different angles. This allows reaching results at the case study level as well as the theoretical level. This research validates the importance of using the cultural landscape in post-disaster situations as planners and government officials overlook it. Some of the elements that constitute it like place attachment and social networks motivate people to return to their original neighbourhoods and rebuild their homes and community. These elements, however, cannot by themselves give people back what they lost in the disaster. By relating the cultural landscape to the concept of resilience, it implies that resilience is a social construction.

This thesis explores the entire cycle of education, from initial access to schooling, over degree completion, to returns to education. Despite recent gains in increasing access, an tens of millions of children worldwide are still out of school. Abolishing school fees has increased enrollment rates in several countries where enrollments were low and fees were high. However, such policies may be less effective, or even have negative consequences, when supply-side responses are weak. The first part of the thesis evaluates the impacts of a tuition waiver program in Haiti, which provided public financing to nonpublic schools conditional on not charging tuition. The chapter concludes that school's participation in the program results in more students enrolled, more staff, and slightly higher student-teacher ratios. The program also reduces grade repetition and the share of overage students. While the increase in students does not directly equate to a reduction in the number of children out of school, it does demonstrate strong demand from families for the program and a correspondingly strong supply response from the nonpublic sector.Pertaining degree completion, it is well established that natural disasters can have a negative effect on human capital accumulation. However, a comparison of the differential impacts of distinct disaster classes is missing. Using census data and information from DesInventar and EMDAT, two large disaster databases, the second part of the thesis assesses how geological disasters and climatic shocks affect the upper secondary degree attainment of adolescents. The chapter focuses on Mexico, given its diverse disaster landscape and lack of obligatory upper secondary education over the observed time period. While all disaster types are found to impede attainment, climatic disasters that are not infrastructure-destructive (e.g. droughts) have the strongest negative effect, decreasing educational expansion by over 40%. The effects seem largely driven by demand-side changes such as increases in school dropouts and fertility, especially for young women. The results may also be influenced by deteriorated parental labor market outcomes. Supply-side effects appear to be solely driven by infrastructure-destructive climatic shocks (e.g. floods). These findings thus call for differential public measures according to specific disaster types and an enhanced attention to climatic events given their potentially stronger impact on younger generations.It is also widely appreciated that natural disasters can have negative impacts on local labor market outcomes. However, the study of differential types of negative capital shocks, the underlying labor market mechanisms, and the context of the poorest countries have been neglected. Following testable predictions of economic theory, the third part of the thesis exploits the exogenous variation of destruction of human and physical capital caused by the 2010 Haiti earthquake to disentangle the differential impact on local individual monetary returns to education. Employing individual-level survey data from before and after the earthquake the chapter finds that the returns decreased on average by 37%, especially in equipment-capital intensive industry. Higher educated individuals adjust into low-paying self-employment or agriculture. The returns are particularly shock-sensitive for urban residents, migrants, males, and people over age 25. / Doctorat en Sciences économiques et de gestion / info:eu-repo/semantics/nonPublished

Between 45-57% of South Africans are estimated to be engulfed by poverty. In an attempt to identify policy instruments that could help change this status quo, the various strategies that have been implemented in countries (e.g. China, Vietnam and Uganda) that are known to have been relatively successful in reducing poverty are reviewed. In the process, this dissertation discusses the literature regarding poverty, with a particular emphasis on the definition, measurement and determinants thereof. Furthermore, South Africa's anti-poverty strategies are discussed. It turns out that these have met limited success. This is largely due to insufficient pro-poor economic growth, weak implementation/administration at the municipal level, slow asset redistribution, high income/wealth inequality, low job generation rate by SMME's, high HIV/AIDS infection rate, public corruption and inadequate monitoring of poverty. Therefore, if meaningful progress towards poverty reduction is to be achieved, the government needs to deal with the foregoing constraints accordingly. / Economics / M.Comm. (Economics)

Between 45-57% of South Africans are estimated to be engulfed by poverty. In an attempt to identify policy instruments that could help change this status quo, the various strategies that have been implemented in countries (e.g. China, Vietnam and Uganda) that are known to have been relatively successful in reducing poverty are reviewed. In the process, this dissertation discusses the literature regarding poverty, with a particular emphasis on the definition, measurement and determinants thereof. Furthermore, South Africa's anti-poverty strategies are discussed. It turns out that these have met limited success. This is largely due to insufficient pro-poor economic growth, weak implementation/administration at the municipal level, slow asset redistribution, high income/wealth inequality, low job generation rate by SMME's, high HIV/AIDS infection rate, public corruption and inadequate monitoring of poverty. Therefore, if meaningful progress towards poverty reduction is to be achieved, the government needs to deal with the foregoing constraints accordingly. / Economics / M.Comm. (Economics)

Tesis por compendio / [ES] Edge Computing es un modelo de computación emergente basado en acercar el procesamiento a los dispositivos de captura de datos en las infraestructuras Internet of things (IoT). Edge computing mejora, entre otras cosas, los tiempos de respuesta, ahorra anchos de banda, incrementa la seguridad de los servicios y oculta las caídas transitorias de la red. Este paradigma actúa en contraposición a la ejecución de servicios en entornos cloud y es muy útil cuando se desea desarrollar soluciones de inteligencia artificial (AI) que aborden problemas en entornos de desastres naturales, como pueden ser inundaciones, incendios u otros eventos derivados del cambio climático. La cobertura de estos escenarios puede resultar especialmente difícil debido a la escasez de infraestructuras disponibles, lo que a menudo impide un análisis de los datos basado en la nube en tiempo real. Por lo tanto, es fundamental habilitar técnicas de IA que no dependan de sistemas de cómputo externos y que puedan ser embebidas en dispositivos de móviles como vehículos aéreos no tripulados (VANT), para que puedan captar y procesar información que permita inferir posibles situaciones de emergencia y determinar así el curso de acción más adecuado de manera autónoma. Históricamente, se hacía frente a este tipo de problemas utilizando los VANT como dispositivos de recogida de datos con el fin de, posteriormente, enviar esta información a la nube donde se dispone de servidores capacitados para analizar esta ingente cantidad de información. Este nuevo enfoque pretende realizar todo el procesamiento y la obtención de resultados en el VANT o en un dispositivo local complementario. Esta aproximación permite eliminar la dependencia de un centro de cómputo remoto que añade complejidad a la infraestructura y que no es una opción en escenarios específicos, donde las conexiones inalámbricas no cumplen los requisitos de transferencia de datos o son entornos en los que la información tiene que obtenerse en ese preciso momento, por requisitos de seguridad o inmediatez. Esta tesis doctoral está compuesta de tres propuestas principales. En primer lugar se plantea un sistema de despegue de enjambres de VANTs basado en el algoritmo de Kuhn Munkres que resuelve el problema de asignación en tiempo polinómico. Nuestra evaluación estudia la complejidad de despegue de grandes enjambres y analiza el coste computacional y de calidad de nuestra propuesta. La segunda propuesta es la definición de una secuencia de procesamiento de imágenes de catástrofes naturales tomadas desde drones basada en Deep learning (DL). El objetivo es reducir el número de imágenes que deben procesar los servicios de emergencias en la catástrofe natural para poder tomar acciones sobre el terreno de una manera más rápida. Por último, se utiliza un conjunto de datos de imágenes obtenidas con VANTs y relativas a diferentes inundaciones, en concreto, de la DANA de 2019, cedidas por el Ayuntamiento de San Javier, ejecutando un modelo DL de segmentación semántica que determina automáticamente las regiones más afectadas por las lluvias (zonas inundadas). Entre los resultados obtenidos se destacan los siguientes: 1- la mejora drástica del rendimiento del despegue vertical coordinado de una red de VANTs. 2- La propuesta de un modelo no supervisado para la vigilancia de zonas desconocidas representa un avance para la exploración autónoma mediante VANTs. Esto permite una visión global de una zona concreta sin realizar un estudio detallado de la misma. 3- Por último, un modelo de segmentación semántica de las zonas inundadas, desplegado para el procesamiento de imágenes en el VANTs, permite la obtención de datos de inundaciones en tiempo real (respetando la privacidad) para una reconstrucción virtual fidedigna del evento. Esta tesis ofrece una propuesta para mejorar el despegue coordinado de drones y dotar de capacidad de procesamiento de algoritmos de deep learning a dispositivos edge, más concretamente UAVs autónomos. / [CA] Edge Computing és un model de computació emergent basat a acostar el processament als dispositius de captura de dades en les infraestructures Internet of things (IoT). Edge computing millora, entre altres coses, els temps de resposta, estalvia amplades de banda, incrementa la seguretat dels serveis i oculta les caigudes transitòries de la xarxa. Aquest paradigma actua en contraposició a l'execució de serveis en entorns cloud i és molt útil quan es desitja desenvolupar solucions d'intel·ligència artificial (AI) que aborden problemes en entorns de desastres naturals, com poden ser inundacions, incendis o altres esdeveniments derivats del canvi climàtic. La cobertura d'aquests escenaris pot resultar especialment difícil a causa de l'escassetat d'infraestructures disponibles, la qual cosa sovint impedeix una anàlisi de les dades basat en el núvol en temps real. Per tant, és fonamental habilitar tècniques de IA que no depenguen de sistemes de còmput externs i que puguen ser embegudes en dispositius de mòbils com a vehicles aeris no tripulats (VANT), perquè puguen captar i processar informació per a inferir possibles situacions d'emergència i determinar així el curs d'acció més adequat de manera autònoma. Històricament, es feia front a aquesta mena de problemes utilitzant els VANT com a dispositius de recollida de dades amb la finalitat de, posteriorment, enviar aquesta informació al núvol on es disposa de servidors capacitats per a analitzar aquesta ingent quantitat d'informació. Aquest nou enfocament pretén realitzar tot el processament i l'obtenció de resultats en el VANT o en un dispositiu local complementari. Aquesta aproximació permet eliminar la dependència d'un centre de còmput remot que afig complexitat a la infraestructura i que no és una opció en escenaris específics, on les connexions sense fils no compleixen els requisits de transferència de dades o són entorns en els quals la informació ha d'obtindre's en aqueix precís moment, per requisits de seguretat o immediatesa. Aquesta tesi doctoral està composta de tres propostes principals. En primer lloc es planteja un sistema d'enlairament d'eixams de VANTs basat en l'algorisme de Kuhn Munkres que resol el problema d'assignació en temps polinòmic. La nostra avaluació estudia la complexitat d'enlairament de grans eixams i analitza el cost computacional i de qualitat de la nostra proposta. La segona proposta és la definició d'una seqüència de processament d'imatges de catàstrofes naturals preses des de drons basada en Deep learning (DL).L'objectiu és reduir el nombre d'imatges que han de processar els serveis d'emergències en la catàstrofe natural per a poder prendre accions sobre el terreny d'una manera més ràpida. Finalment, s'utilitza un conjunt de dades d'imatges obtingudes amb VANTs i relatives a diferents inundacions, en concret, de la DANA de 2019, cedides per l'Ajuntament de San Javier, executant un model DL de segmentació semàntica que determina automàticament les regions més afectades per les pluges (zones inundades). Entre els resultats obtinguts es destaquen els següents: 1- la millora dràstica del rendiment de l'enlairament vertical coordinat d'una xarxa de VANTs. 2- La proposta d'un model no supervisat per a la vigilància de zones desconegudes representa un avanç per a l'exploració autònoma mitjançant VANTs. Això permet una visió global d'una zona concreta sense realitzar un estudi detallat d'aquesta. 3- Finalment, un model de segmentació semàntica de les zones inundades, desplegat per al processament d'imatges en el VANTs, permet l'obtenció de dades d'inundacions en temps real (respectant la privacitat) per a una reconstrucció virtual fidedigna de l'esdeveniment. / [EN] Edge Computing is an emerging computing model based on bringing data processing and storage closer to the location needed to improve response times and save bandwidth. This new paradigm acts as opposed to running services in cloud environments and is very useful in developing artificial intelligence (AI) solutions that address problems in natural disaster environments, such as floods, fires, or other events of an adverse nature. Coverage of these scenarios can be particularly challenging due to the lack of available infrastructure, which often precludes real-time cloud-based data analysis. Therefore, it is critical to enable AI techniques that do not rely on external computing systems and can be embedded in mobile devices such as unmanned aerial vehicles (UAVs) so that they can capture and process information to understand their context and determine the appropriate course of action independently. Historically, this problem was addressed by using UAVs as data collection devices to send this information to the cloud, where servers can process it. This new approach aims to do all the processing and get the results on the UAV or a complementary local device. This approach eliminates the dependency on a remote computing center that adds complexity to the infrastructure and is not an option in specific scenarios where wireless connections do not meet the data transfer requirements. It is also an option in environments where the information has to be obtained at that precise moment due to security or immediacy requirements. This study consists of three main proposals. First, we propose a UAV swarm takeoff system based on the Kuhn Munkres algorithm that solves the assignment problem in polynomial time. Our evaluation studies the takeoff complexity of large swarms and analyzes our proposal's computational and quality cost. The second proposal is the definition of a Deep learning (DL) based image processing sequence for natural disaster images taken from drones to reduce the number of images processed by the first responders in the natural disaster. Finally, a dataset of images obtained with UAVs and related to different floods is used to run a semantic segmentation DL model that automatically determines the regions most affected by the rains (flooded areas). The results are 1- The drastic improvement of the performance of the coordinated vertical take-off of a network of UAVs. 2- The proposal of an unsupervised model for the surveillance of unknown areas represents a breakthrough for autonomous exploration by UAVs. This allows a global view of a specific area without performing a detailed study. 3- Finally, a semantic segmentation model of flooded areas, deployed for image processing in the UAV, allows obtaining real-time flood data (respecting privacy) for a reliable virtual reconstruction of the event. This thesis offers a proposal to improve the coordinated take-off of drones, to provide edge devices with deep learning algorithms processing capacity, more specifically autonomous UAVs, in order to develop services for the surveillance of areas affected by natural disasters such as fire detection, segmentation of flooded areas or detection of people in danger. Thanks to this research, services can be developed that enable the coordination of large arrays of drones and allow image processing without needing additional devices. This flexibility makes our approach a bet for the future and thus provides a development path for anyone interested in deploying an autonomous drone-based surveillance and actuation system. / I would like to acknowledge the project Development of High-Performance IoT Infrastructures against Climate Change based on Artificial Intelligence (GLOBALoT). Funded by Ministerio de Ciencia e Innovación (RTC2019-007159-5), of which this thesis is part. / Hernández Vicente, D. (2023). Analysis Design and Implementation of Artificial Intelligence Techniques in Edge Computing Environments [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/192605 / Compendio

Recurring drought conditions have always been an endemic feature of climate in South Africa, limiting maize development and production. However, recent projections of the future climate by the Intergovernmental Panel on Climate Change suggest that due to an increase of atmospheric greenhouse gases, the frequency and severity of droughts will increase in drought-prone areas, mostly in subtropical climates. This has raised major concern for the agricultural sector, particularly the vulnerable small-scale farmers who merely rely on rain for crop production. Farmers in the Luvuvhu River catchment are not an exception, as this area is considered economically poor, whereby a significant number of people are dependent on rain-fed farming for subsistence. This study was therefore conducted in order to improve agricultural productivity in the area and thus help in the development of measures to secure livelihoods of those vulnerable small-scale farmers. Two drought indices viz. Standardized Precipitation Evapotranspiration Index (SPEI) and Water Requirement Satisfaction Index (WRSI) were used to quantify drought. A 120-day maturing maize crop was considered and three consecutive planting dates were staggered based on the average start of the rainy season. Frequencies and probabilities during each growing stage of maize were calculated based on the results of the two indices. Temporal variations of drought severity from 1975 to 2015 were evaluated and trends were analyzed using the non-parametric Spearman’s Rank Correlation test at α (0.05) significance level. For assessing climate change impact on droughts, SPEI and WRSI were computed using an output from downscaled projections of CSIRO Mark3.5 under the SRES A2 emission scenario for the period 1980/81 – 2099/100. The frequency of drought was calculated and the difference of SPEI and WRSI means between future climate periods and the base period were assessed using the independent t-test at α (0.10) significance level in STATISTICA software. The study revealed that planting a 120-day maturing maize crop in December would pose a high risk of frequent severe-extreme droughts during the flowering to the grain-filling stage at Levubu, Lwamondo, Thohoyandou, and Tshiombo; while planting in October could place crops at a lower risk of reduced yield and even total crop failure. In contrast, stations located in the low-lying plains of the catchment (Punda Maria, Sigonde, and Pafuri) were exposed to frequent moderate droughts following planting in October, with favorable conditions noted following the December planting date. Further analysis on the performance of the crop under various drought conditions revealed that WRSI values corresponding to more intense drought conditions were detected during the December planting date for all stations. Moreover, at Punda Maria, Sigonde and Pafuri, it was observed that extreme drought (WRSI <50) occurred once in five seasons, regardless of the planting date. Temporal analysis on historical droughts in the area indicated that there had been eight agricultural seasons subjected to extreme widespread droughts resulting in total crop failure i.e. 1983/84, 1988/89, 1991/92, 1993/94, 2001/02, 2002/03, 2004/05 and 2014/15. Results of Spearman’s rank correlation test revealed weak increasing drought trends at Thohoyandou (ρ = of 0.5 for WRSI) and at Levubu and Lwamondo (ρ = of 0.4 for SPEI), with no significant trends at the other stations. The study further revealed that climate change would enhance the severity of drought across the catchment. This was statistically significant (at 10% significance level) for the near-future and intermediate-future climates, relative to the base period. Drought remains a threat to rain-fed maize production in the Luvuvhu River catchment area of South Africa. In order to mitigate the possible effects of droughts under climate change, optimal planting dates were recommended for each region. The use of seasonal forecasts during drought seasons would also be useful for local rain-fed maize growers especially in regions where moisture is available for a short period during the growing season. It was further recommended that the Government ensure proper support such as effective early warning systems and inputs to the farmers. Moreover, essential communication between scientists, decision makers, and the farmers can help in planning and decision making ahead of and during the occurrence of droughts. / Agriculture, Animal Health and Human Ecology / M. Sc. (Agriculture)