高速公路系統建設可提升區域與區域之間的可及性與易行性，同時達到節省旅行時間和成本、高速公路交流道沿線地區土地價值上升等效果。此外，高速公路系統建設對於周圍地區人口、產業可產生相當程度之影響，例如大量的人口遷移、區域交通影響衝擊、社會和經濟層面產生重大變化等，進而改變國家及區域發展方向。於此情況下，合宜地評估及分析高速公路系統建設開發對國家及區域發展之影響，將有助於其推展並降低對地區之不利影響。 本研究希望以我國從1976至2010年來，共九條高速公路系統 建設為研究主軸，利用台灣城際運輸系統需求模式(TMD2008)之旅行時間資料、內政部戶口普查及工商普查資料為基礎，並運用事前事後比較(before and after analysis)與空間自相關分析(Spatial Autocorrelation analysis)等分析方法對高速公路系統建設所造成區域人口、產業之空間變動進行評估，分析高速公路系統建設對於區域人口、產業及交通可及性之空間分布變化影響。最後則建構高速公路可及性對區域人口及產業空間影響模型，並針對歷年來高速公路系統建設對國家及區域發展的影響進行實證分析。 / The benefits of highway infrastructure investments include promoted inter-regional accessibility and mobility, saved travel time and cost and increased land value in freeway interchange areas. In sum, highway infrastructure investments play an important role in national and regional development. It contributes to the population re-distribution, the growth of economic activities, and more importantly, the land use change. Therefore, how to effectively evaluate the impact of highway investments on national and regional development is an important research topic. The analytical results in this study can be applied to increase the accuracy of population and economic projection, and to manage transportation policy decisions. Between 1976 and 2010, there were nine highway systems had been built in Taiwan. This study emphasizes on the impact of highway investments on the spatial distribution of population and industry, and regional accessibility change. In this paper, we utilize before and after analysis and Spatial Autocorrelation analysis with Geographic Information System to analyze demographic data, economic data and TDM2008 database in order to perform the spatial analysis of population and industry re-distribution along highway constructions. Finally, a spatial gravity model is built to verify and describe the related spatial impacts so that several influential factors can be identified by this empirical study.

Rivers and their catchments are under mounting pressure from direct channel modification, intensification of land use, and from a legacy of decades of channelisation. Recent legislation, in the form of the EU Water Framework Directive, places a greater emphasis on the management of water bodies as holistic systems, and includes the explicit consideration of hydromorphological quality, which describes the hydrologic and geomorphic elements of river habitats. These are defined specifically as hydrological regime, river continuity and river morphology. This appreciates that sediment and flow regimes, along with the channel structure, provides the 'template' on which stream ecological structure and function is built. Invertebrate fauna contribute significantly to the biodiversity of rivers, and often form the basis of monitoring river health. However much of the fundamental ecological knowledge base on the response of invertebrates to hydromorphological change needed to make informed decisions and accurate predictions, is either lacking, inadequate or contradictory. This thesis addresses some of the key potential shortcomings in recent bio-assessment that others have alluded to, but which have rarely been explored in the context of direct channel manipulations. By using two case studies of, realignment in a natural upland catchment, and flood protection engineering in an urban stream, this study investigates the sensitivity of hydromorphological impact assessment methods that rely on biodiversity patterns of benthic riverfly (Ephemeroptera, Plecoptera and Trichoptera) larva. This work employed widely used biomonitoring indices of benthic riverfly larva abundance, species richness, alpha and beta diversity, and community composition, applied over a range of spatial scales, in combination with spatially contemporaneous physical habitat data, to describe and explain community changes in response to disturbance, and patterns of natural variation. The effects of restoration were investigated using a high degree of sample replication within channels and across the wider catchment, as well as contrasting spring and autumn seasons. To assess change in a small urban channel, approaches that explicitly consider spatial elements of community data, using spatial eigenvectors analysis, were applied to spatially detrend community data and directly investigate spatial patterns. Restoration of the Rottal Burn was found to be successful in restoring habitat diversity and geomorphic processes, and in turn increasing reach scale species richness and beta diversity through the gradual arrival of rare and specialist taxa into novel habitats. Catchment scale replication revealed high variation in diversity indices of modified and undisturbed streams, and a strong temporal pattern related to antecedent flow conditions. Channels with greater habitat heterogeneity were able to maintain high gamma diversity during times of high flow stress by providing a number of low flow refuges along their length. The urban Brox Burn had surprisingly high riverfly richness and diversity driven by small scale hydraulic heterogeneity, created by bed roughness resulting in a range of microhabitats. Riverfly community responses to direct channel dredging could not be detected by measurements of average richness and diversity, however distinct changes were seen in gamma diversity, the identity of community members and their arrangement among sample patches. Impacts of sediment pollution release due to engineering were short lived and apparently had little detrimental impact on biodiversity. Strong spatial patterns of community assembly on the stream bed were uncovered, relating to longitudinal, edge and patchy patterns. Significant habitat drivers of community composition were confounded by high amounts of spatial autocorrelation, especially hydraulic variables. Due to the strongly physical and spatial nature of hydromorphological disturbance, turnover of species between sample locations at a range of scales, and the spatial arrangement of habitats and communities is of more use for detecting these types of subtle changes compared to mean richness or diversity. These findings have implications for the targeting of resources for monitoring of restoration, or engineering disturbances, in order to be sensitive to hydromorphological change. Efforts should target the main area of natural variability within the system, either replicating sampling in time or space to distinguish effects of impact. Spatial patterns, measures of beta diversity and species identity can be better exploited to identify systems with functioning geomorphological processes. Channel typologies proved misleading, and quantification of habitat and selection of control sites using multiple pre-defined criteria should be carried out. Studies of restoration operations and engineering impacts provide considerable opportunities for advancing our knowledge of the mechanisms that drive community response under a range of conditions to improve impact detection.

Made available in DSpace on 2016-06-02T19:31:49Z (GMT). No. of bitstreams: 1 2941.pdf: 567635 bytes, checksum: 18b3f5d53030b198ba538da511d6b2e3 (MD5) Previous issue date: 2010-03-02 / Universidade Federal de Minas Gerais / Together with fire and climate changes, soil is considered a major determinant in the Braziliancerrado, the richest savanna in the world. Soil can influence plants by filtering species capableof acquiring resources and compete for them, but can also influence plant patterns ofallocation in defense against herbivory. Although many studies focused on plant soilrelationship in cerrado, few focused on the influence of soil at fine scale. We expectedcommunity descriptors, such as floristic composition, richness, evenness, diversity, and totalabundance to be related to soil features at fine scale within a physiognomy. We also expectedplants on nutrient-poor soils to present higher anti-herbivory defenses. In a cerrado site, weplaced 100 contiguous 25 m2 plots, in which we identified all woody individuals, measuredsoil variables and the following leaf traits: specific leaf area, C:N ratio, water content,toughness, trichomes, latex, and presence of tannins, alkaloids, and terpenoids. We did apartial redundancy analysis to test for relationship between soil features and floristiccomposition, controlled for spatial dependence. We also did multiple regression or spatialautoregressive models to test for relationships between soil features and: (1) the abundance ofthe five commonest species, (2) total abundance, (3) richness, (4) evenness, and (5) diversityand to predict defense traits based on soil features. We found no relationship between soil andfloristic composition, probably due to functional redundancy or limited dispersal. Organicmatter was positively related to Myrsine umbellata, the most abundant species, and totalabundance, and negatively related to evenness, what suggests positive feedbacks to cause thedominance by Myrsine umbellata. We also found a positive relationship between sum of basisand species richness, probably reflecting a fertility gradient. Contrary to our expectations, wefound no relationship between total defenses and total soil fertility or soil variables, whatcould result from low variability in soil fertility at fine scale or of high phenotypic variability.Presence of tannins was positively related to organic matter, possibly reflecting a strategytowards lower tolerance due to low reserve allocation or interactions with other resources.However, since tannins decrease leaf decomposition rates, organic matter could beaccumulating in soil. Overall, we suggest that soil is an important factor structuring cerradocommunity even at fine scales and that the dominance of cerrado species could be related topositive plant-soil feedbacks. / O solo, juntamente com o fogo e as variações climáticas, é considerado um dos principaisdeterminantes do cerrado brasileiro, a savana mais rica do mundo. O solo pode influenciar asespécies de planta selecionando aquelas adaptadas a explorar e competir por recursos, mastambém pode influenciar os padrões de alocações em defesas contra herbivoria. Emboramuitos estudos tenham se voltado a entender as relações entre solo e vegetação no cerrado,poucos se focaram efeitos em escala local. De forma geral, nossa expectativa é de quedescritores da comunidade, como composição florística, riqueza, equabilidade, diversidade eabundância total estejam relacionados com o solo, mesmo em escala local, dentro de umadeterminada fisionomia, e que plantas em solos pobres em nutrientes invistam mais emdefesas contra herbivoria devido ao alto custo em repor as folhas perdidas. Em uma área decerrado, lançamos 100 parcelas contíguas de 25 m2 cada, identificamos todos os indivíduosem nível de espécies, coletamos amostras de solo e medimos os seguintes traços foliares dedefesa contra herbivoria: área foliar específica, razão C:N, quantidade de água, dureza,densidade de tricomas, quantidade de látex, e presença de alcaloides, terpenoides e taninos.Para testar a relação entre a composição florística e as variáveis do solo, usamos uma análisede redundância parcial, controlando a autocorrelação espacial. Para testar a relação entre asvariáveis do solo e (1) a abundância das cinco espécies mais abundantes, (2) a abundânciatotal, (3) a riqueza de espécies, (4) a equabilidade e (5) a diversidade de espécies; para prevera distribuição dos traços de defesa contra herbivoria por meio das variáveis do solo,utilizamos regressões múltiplas ou modelos autorregressivos, na presença de autocorrelaçãoespacial. Encontramos uma baixa relação entre o solo e a composição florística,provavelmente devido à presença de espécies funcionalmente redundantes e espécies comdispersão limitada. O conteúdo de matéria orgânica esteve positivamente relacionado àabundância de Myrsine umbellata, a espécie mais abundante na área, e à abundância total, enegativamente relacionada à equabilidade, o que sugere que um mecanismo deretroalimentação positiva pode ser a causa da dominância de Myrsine umbellata. Tambémencontramos uma relação positiva entre soma de bases e a riqueza de espécies provavelmenterefletindo um gradiente de fertilidade. Contrariamente às nossas expectativas, nãoencontramos relação entre o investimento total em defesas e a fertilidade do solo,provavelmente refletindo uma baixa variação nas variáveis do solo em escala local ouvariações fenotípicas entre indivíduos da mesma espécie ou ambas. Entretanto, a presença detaninos esteve relacionada positivamente com o conteúdo de matéria orgânica, o que poderefletir menor tolerância à herbivoria em solos mais pobres ou uma alta acumulação dematéria orgânica no solo devido à lenta taxa de decomposição de folhas com tanino. De formageral, sugerimos que o solo é um importante fator estruturando a comunidade, mesmo emescala local, e que a dominância de espécies de cerrado pode estar relacionada a mecanismosde retroalimentação positiva.

Made available in DSpace on 2017-07-10T18:33:59Z (GMT). No. of bitstreams: 1 Alessandra J Caumo1.pdf: 3054237 bytes, checksum: 8e5eaef96cdf96ff916b40c6dc2ab4d9 (MD5) Previous issue date: 2012-08-30 / Fundação Araucária / It was observed the spatial occurrence distribution of occupied women on rural families in Southern Brazil. This study showed another point of view about occupied women from Space Data Exploratory Analysis (AEDE). In the familiar agriculture the gender relations are most often, due to the work division imposed by the activity performed by the family members, and, in special the woman, her work is considered merely a help . This kind of analysis helps in a certain way, important changes in Brazil to be known and, mainly, in Southern Brazil in the distribution of occupied women. In this task was used the variable of occupied woman as the main one, but compared in relation to occupied men, where the approach is of a gender perspective. So, it was used the Space Data Exploratory Analysis (AEDE) technique, by the Moran s I statistics and by the clusters identification analysis. Moran s I statistics allowed to evaluate the existent correlation between a region and their neighbors considering a variable. However the clusters identification analysis allowed a simple visualization of next regions conglomerate in a map that pursue similar values, in other words, allow the clusters visualization. The analyzed period was the year 2006, by the data of 2006 Agriculture Census, performed by the Statistics and Geography Brazilian Institute IBGE, these data were studied by an absolute form and rates. The AEDE revealed the occupied women distribution among the states of Paraná, Santa Catarina and Rio Grande do Sul. The results, to the scholarity in the absolute data or by rates, presented relation with the number of occupied women, in this case, where it presented a number of occupied men with a greater scholarity. In the activities out of the establishments, was also verified that in cities with the lesser occupation per woman, the greater was the number of cities the women followed any activity outside the property. With the results were found positive spatial autocorrelation of the absolutely occupied women in the Southern Brazil rural area, showing off that spatial localization interferes in women s values. Having a contribution to gender perspective analysis in Southern Brazil and allowing new researches or concepts for the study, where the data revealed a proportion of about 40% of the occupied women from the total in the Southern Brazil agriculture. / Foi observada a ocorrência da distribuição espacial da mulher ocupada na agropecuária do Sul do Brasil. Este estudo apresentou um quadro das mulheres ocupadas sob outro ponto de vista, pela Análise Exploratória de Dados Espaciais (AEDE). Na agricultura familiar, as relações de gênero estão com maior frequência dispostas de modo a considerar o trabalho das mulheres apenas uma ajuda . Isso é devido à divisão de trabalho imposta pelas atividades exercidas pelos membros da família. Esse tipo de análise, de certa forma, ajuda a conhecer as mudanças importantes no Brasil e, principalmente, no Sul do Brasil, na distribuição espacial da mulher profissionalmente ocupada. Neste trabalho se utilizou a variável de mulher ocupada como principal, mas comparado em relação aos homens ocupados, em que a abordagem é de uma perspectiva de gênero. Desta forma, foi utilizada a técnica de Análise Exploratória de Dados Espaciais (AEDE), por meio da estatística I de Moran e da análise de identificação de clusters. A estatística I de Moran permitiu avaliar a correlação existente entre uma região e seus vizinhos considerando uma variável. O período analisado foi o ano de 2006, por meio dos dados do Censo Agropecuário 2006, feito pelo Instituto Brasileiro de Geografia e Estatística IBGE. Esses dados foram estudados de forma absoluta e por taxa. A AEDE revelou a distribuição das mulheres ocupadas entre os estados do Paraná, de Santa Catarina e do Rio Grande do Sul. Os resultados, para a escolaridade nos dados absolutos ou por taxas, apresentaram relação com o número de mulheres ocupadas. Nesse caso, onde se apresentou maior número de mulheres ocupadas também se apresentou um número de homens ocupados com maior escolaridade. Nas atividades fora dos estabelecimentos também se verificou que, para os municípios com menor número de mulheres ocupadas, maior era o número de municípios em que as mulheres exerciam alguma atividade fora da propriedade. Com os resultados, constatou-se autocorrelação espacial positiva das mulheres ocupadas em dados absolutos no meio rural do Sul do Brasil, demonstrando que a localização espacial interfere em seus valores. Tem-se então uma contribuição para a análise da perspectiva de gênero no Sul do Brasil e a possibilidade de novas pesquisas ou de novos conceitos para o estudo, em que os dados revelaram uma proporção de aproximadamente 40%de mulheres ocupadas do total de ocupados na agropecuária do Sul do Brasil.

Long-term evolution of nature-society interactions can be studied by using data of land use change. In the thesis is used a detailed statistical database of Czechia (developed at Charles University in Prague, Faculty of Science). The database allows monitoring changes of eight categories of land use in the period 1845-2000. Attention is devoted to the growing differentiation of land use at different scale levels, the thesis is concentrated on functional specialization of czech regions. Deepening division of labour between regions is reflected by increasing territorial concentration of land use. This process is monitored by using three methods. Ternary plots present an overview of the differentiation trends of land use. Spatial autocorrelation methods present the general development and specific areas of functional specialization in Czechia. A review by territorial heterogeneity rate provides a detailed look at the change of territorial concentration of land use. The thesis operates with three territorial classifications (hydrological, geomorphological and administrative) at two regional levels and the level of Czechia. Results are interpreted using the driving forces (such as technological development, economics, historical events, society). The final part of the thesis compares the development of...

Seit den 80er Jahren des vergangenen Jahrhunderts erleben schienengebundene Nahverkehrsmittel vielerorts eine Renaissance. Hohe Kapazitäten bei der Personenbeförderung und abgasemissionsfreie Fahrzeuge sind häufig genannte Argumente für den Aus- oder Neubau von Straßen- und Stadtbahnnetzen. Unfälle zwischen Straßen- bzw. Stadtbahnen und anderen, vor allem nichtmotorisierten Verkehrsteilnehmern, sind oft besonders schwer. Hinzu kommt, dass der in Deutschland innerorts seit Jahren zu beobachtende Trend eines Rückgangs von Unfallzahlen und Unfallschwere für Unfälle mit Straßen- und Stadtbahnen nur sehr bedingt zu beobachten ist. In Deutschland existiert grundsätzlich eine Vielzahl aufeinander abgestimmter Verfahren, welche die Verbesserung der Verkehrssicherheit in Straßennetzen sowohl präventiv als auch nachträglich unterstützen. Diese sind leicht verständlich aufgebaut und einfach handhabbar. Sie berücksichtigen jedoch das Unfallgeschehen innerorts ganzheitlich und enthalten keine spezifischen Aspekte zu schienengebundenen innerörtlichen Nahverkehrsmitteln. Ziel der vorliegenden Arbeit ist es, Werkzeuge zur Verfügung zu stellen, die zur gezielten Analyse und Verbesserung der Verkehrssicherheit im Zusammenhang mit dem Unfallgeschehen unter Beteiligung von Straßen- bzw. Stadtbahnen herangezogen werden können. Zur Generierung der erforderlichen Datenbasis wurden zunächst die Unfalldaten der Verkehrsbetriebe dreier Städte systematisch vereinheitlicht, georeferenziert und für die weitere Verwendung aggregiert. In einem ersten Untersuchungsschritt wurde anschließend – in Anlehnung an vorhandene Instrumente - ein Verfahren zur Lokalisierung unfallauffälliger Bereiche für Straßen- und Stadtbahnunfälle entwickelt. Zu diesem Zweck wurden Kenngrößen der räumlichen Autokorrelation herangezogen, um über ein einheitliches, objektiv quantifizierbares Instrument für Vergleiche zwischen Netzen mit unterschiedlichen Sicherheitsniveaus bzw. Variationen von Betrachtungszeitraum und einzubeziehenden Teilkollektiven der Unfälle zu verfügen. Hierbei wurden Kriterien zur Lokalisierung unfallauffälliger Bereiche unter Voraussetzung des Einhaltens räumlicher und zeitlicher Stabilität abgeleitet. Zur Identifizierung und Quantifizierung infrastruktureller und verkehrlicher Einflüsse auf das Unfallgeschehen mit Straßen- und Stadtbahnen wurden in einem weiteren Schritt Unfallprognosemodelle erstellt, die das Unfallgeschehen mit entsprechenden Parametern verknüpfen und auf diese Weise Aussagen zu den Einflüssen einzelner Merkmale erlauben. Dabei wurden unterschiedliche Formen der Modellbildung vorgenommen: Neben Grundmodellen zur gemeinsamen Abbildung aller Unfälle wurden Modelle mit Differenzierungen nach Unfallschwere, infrastrukturellen Ausbauformen und Unfalltypen geschätzt und miteinander verglichen. Dies beinhaltete in einer ersten Stufe auch einen Vergleich zweier Städte untereinander, indem die Modelle zunächst stadtfein geschätzt und anschließend einander gegenübergestellt wurden. Dabei haben sich die typfeinen Modelle als methodisch überlegen in Zuverlässigkeit und Genauigkeit erwiesen, sie wurden daher in einer zweiten Stufe mittels Anwendung auf ein unabhängiges Kollektiv in Form der Daten einer dritten Stadt validiert. Im Ergebnis konnten Kriterien zur Identifizierung auffälliger Bereiche von Unfällen mit Straßen- und Stadtbahnen zur Verfügung gestellt werden. Ferner wurde abgeleitet, welche Ausprägungen entwurfstechnischer und betrieblicher Parameter das Unfallgeschehen in welchem Maße beeinflussen. Damit wird ermöglicht, im Rahmen der Verkehrssicherheitsarbeit Maßnahmen zielgerichtet und effizient in Bereichen, die entsprechende Effekte hinsichtlich der Verbesserung der Verkehrssicherheit erwarten lassen, einzusetzen. Ferner können Maßnahmenalternativen szenarienbasiert analysiert und (z. B. bei Neu- bzw. Umbaumaßnahmen) verglichen werden, indem die Maßnahmenwirkung auf das Unfallgeschehen mit Straßen- und Stadtbahnen quantitativ prognostiziert wird. / Since the 1980s, urban rail has been experiencing a renaissance. High capacities and emission-free vehicles are frequently cited arguments for the expansion or new construction of tram and light rail networks. Accidents between trams and other road users are often serious. Further, the trend of declining accident numbers and accident severity that has been observed in Germany for many years can only be observed to a very limited extent regarding accidents involving trams. In Germany, there are a large number of coordinated procedures that support the improvement of traffic safety in road networks. However, they consider the accident situation in urban areas in a holistic way and do not contain any specific aspects concerning trams. The aim of the present work is to provide tools that can be used for the analysis and improvement of traffic safety regarding trams. In order to generate the required database, the accident data of the transport authorities of three cities were systematically standardized, georeferenced and aggregated. In a first step, a procedure for the localization of black spots was developed. For this purpose, methods of spatial autocorrelation were used in order to have an objectively quantifiable instrument for comparisons between networks with different safety levels or variations of the period of time. As a result, criteria for the localization of black spots were derived under the condition of keeping spatial and temporal stability. In order to identify and quantify infrastructural influences on the occurrence of tram accidents, generalized linear accident prediction models were developed in a further step. Different levels of modeling were used: In addition to basic models for the common representation of all accidents, models with differentiations according to accident severity, infrastructural parameters and accident types were estimated and compared. In a first stage, this also included a comparison of two cities with each other by first estimating the models on a city-by-city basis and then comparing them with each other. The type-specific models proved to be best in reliability and accuracy, and were therefore validated in a second stage by applying them to an independent data set from a third city. As a result, criteria for identifying black spots could be provided. Furthermore, it was derived which characteristics of infrastructural design and operational parameters influence the occurrence of accidents and to what extent. This enables an efficient application of measures in the context of road safety work. Furthermore, alternative measures can be analyzed and compared, e. g. on the basis of scenarios.

[pt] Uma das principais preocupações mundiais atualmente está relacionada às questões ambientais. Essa preocupação é considerada na seleção de projetos de energia e, como resultado, a geração de energia elétrica a partir de fontes renováveis tem experimentado um forte crescimento em todo o mundo, incluindo o Brasil. Em relação às fontes de energia hidrelétrica, as Pequenas Centrais Hidrelétricas (PCHs) são uma alternativa para reduzir o impacto ambiental. Esses projetos produzem entre 5 e 30 megawatts (MW) e sua instalação tem um baixo custo e respeito ao meio ambiente, principalmente por não existir necessidade de reservatórios de regulação, o que não é o caso de grandes usinas hidrelétricas. Nos últimos anos, o número de PCHs tem aumentado bastante, como consequência dos incentivos para geração de eletricidade a partir de fontes renováveis. Como a geração de energia hidrelétrica é fortemente influenciada por regimes hidrológicos, especialmente no caso de usinas a fio d água como as PCHs, melhorar a assertividade das previsões de geração de energia elétrica de maneira estocástica torna-se altamente importante para as distribuidoras. Esta dissertação tem como principal objetivo apresentar o desempenho de um grupo de modelos de previsão aplicados para PCHs de uma distribuidora real de energia elétrica. Para isso foram utilizadas diferentes abordagens, incluindo dados de vazão de usinas hidrelétricas vizinhas como variável explicativa em modelos causais, assim como também modelos univariados. / [en] One of the main world concerns nowadays is related to the environment issues. Such concern is considered in the selection of energy projects and, as a result of that, the generation of electricity from renewable sources has experienced a sharp growth all over the world, Brazil included. Concerning hydropower sources, Small Hydropower Plants (SHPs) are an alternative to reduce environmental impact. These projects produce between 5 and 30 megawatts (MW) and its installation has a low cost and respect to the environment, mainly because there is no need of regulation reservoirs, which is not the case in bigger hydroelectric plants. In recent years the number of SHPs is increasing in a great deal, as a consequence of the incentives to generate electricity from renewable sources. Since hydro power generation is heavily influenced by hydrological regimes, especially in the case of run-of-river plants, as SHPs, improving the assertiveness of electric power generation forecasts in a stochastic way becomes highly important for distributing utilities. This master dissertation has as main objective to present the performance of an arrange of forecasting models applied to SHPs of a real distributing utility. It was used different approaches, including inflow data from neighboring hydro plants as exogenous variable, in causal models and also univariate models.

近年來由於地理資訊系統（GIS）的快速發展發，空間資料分析開始受到重視並在社會科學領域中逐漸扮演重要的角色。雖然一般的統計方法已在傳統資料分析上發展已久，然而它們卻不能有效地說明空間性資料，並且無法充分處理空間相依或空間異質性問題。一般而言，空間資料分析主要有兩個分派:模型導向學派與資料導向學派。本文研究目的在於應用空間統計方法合理且充分地評估房地產價值，研究方法包含地理統計(克利金和共克利金)、地理加權迴歸與空間特徵價格模型等，並且以台中市不動產資料進行實證探究。這項新的研究技術在不動產評價領域中將可提供更好的解析能力，使其在評價過程中或是不動產投資決策時，成為一個更強而有力的分析工具。 / In recent years, spatial data analysis has received significant awareness and played an important role in social science because of the rapid development of Geographic Information System (GIS). Although classic statistical methods are attractive in traditional data analysis, they cannot be executed seriously for spatial data. Standard statistical techniques didn’t sufficiently deal with spatial dependence or spatial heterogeneity issues. Generally, the model-driven method and the data-driven method are mainly the two branches of the spatial data analysis. The purpose of this paper is to apply spatial statistics methods including geostatistical methods (kriging and cokiging), geographically weighted regression, and spatial hedonic price models to real estate analysis. It seems to be completely reasonable and sufficient. The real estate data in Taichung city (Taiwan) is used to carry out our exploration. These techniques give better insight in the field of real estate assessment. They can apply a good instrument in mass appraisal and decision concerning real estate investment.

Made available in DSpace on 2017-07-10T18:33:40Z (GMT). No. of bitstreams: 1 Sandra C Kleinschmitt.pdf: 1716732 bytes, checksum: e8f8b01c3b62129ddcc186836f7688b6 (MD5) Previous issue date: 2009-03-06 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The central objective of this research was to analyze the spatial distribution of homicides in the state of Paraná, over the years 2001 and 2005 and the relationship between this phenomenon and socioeconomic, demographic and urban infrastructure variables as well. The method of Exploratory Spatial Data Analysis (ESDA) was used to measure the degree of spatial autocorrelation among the studied variables, based on the characteristics of the 399 municipalities in the State of Paraná, presently. Initially, the ESDA was used to check the existence of spatial autocorrelation among the state municipalities considering the homicide rates. The results demonstrated the existence of spatial autocorrelation of homicide rates in the municipalities of Paraná State. When considering the mapping of clusters, there was the indicative of four clusters of High High type- (HH) and four large clusters of low-low type (LL). The highest homicide rates were found in clusters of municipalities located in the Metropolitan Mesoregion of Curitiba, in the North Central Mesoregion, West Mesoregion, and in a number of municipalities from Center South, Center West, West and Southwest Mesoregions. The clusters of type LL were in the Northwest and in the Central North Mesoregion, and Pioneiro North and Southwest Mesoregion. The common manifestations of variables in the four clusters of HH type were: the high average of residents per home (ARH), the high number of head of family without income, low number of kindergartens, low number of cultural facilities, low number of people over 60 years and the low incidence of infant mortality. The other tested variables autocorrelated in the cities with the rates of murder, but according to their particular regions. The exception was the variables: water supply and sewerage, per capita Gross Domestic Product (GDP) and head of family over 15 years literate, not significant, at a 5% significance. / O objetivo central desta pesquisa é analisar a distribuição espacial dos homicídios no Estado do Paraná, entre os anos 2001 e 2005, e a relação entre este fenômeno e as variáveis socioeconômicas, demográficas e de infraestrutura urbana. O método de Análise Exploratória de Dados Espaciais (AEDE) foi utilizado para medir o grau de autocorrelação espacial entre as variáveis estudadas, a partir das características dos 399 municípios existentes no Estado do Paraná, na atualidade. Num primeiro momento utilizou-se a AEDE para verificar a presença de autocorrelação espacial entre os municípios do Estado, considerando as taxas de homicídio. Os resultados demonstraram a existência da autocorrelação espacial das taxas de homicídio entre os municípios paranaenses. Ao considerar os mapas de clusters, verificou-se a presença de quatro grandes clusters do tipo Alto-Alto (AA) e quatro grandes clusters do tipo Baixo-Baixo (BB). As maiores taxas de homicídio foram encontradas nos clusters constituídos por municípios situados na Mesorregião Metropolitana de Curitiba, na Mesorregião Norte Central, na Mesorregião Oeste e num conjunto de municípios das Mesorregiões Centro-Sul, Centro-Ocidental, Oeste e Sudoeste. Os clusters do tipo BB constituíram-se nas Mesorregiões Noroeste e Norte Central, na Mesorregião Norte Pioneiro e na Mesorregião Sudoeste. As manifestações comuns das variáveis nos quatro clusters do tipo AA, foram: a alta média de moradores por domicílio (MMD), a alta presença de chefes de família sem renda, a baixa presença de creches, a baixa presença de equipamentos culturais, a baixa presença de população com mais de 60 anos e a baixa incidência de mortalidade infantil. As outras variáveis testadas se autocorrelacionaram nos municípios com as taxas de homicídio, mas de acordo com suas particularidades regionais.

No description available.