Bachelor-Arbeit Univ. St. Gallen, 2007.

Techn. Hochsch., 1999--Aachen.

Orientador: Miguel Petrere Junior / Banca: Luzia Ap. Trinca / Banca: João Luis Ferreira Batista / Resumo: É bastante comum no estudo de ecologia de comunidades o uso de índices de diversidade. Historicamente, os principais índices são os derivados por Shannon e por Simpson. Atualmente, existe uma família de índices, derivada da entropia de Tsallis, que inclui os índices supracitados. Igualmente importantes para descrever os padrões observados na natureza, são os modelos de distribuição de espécie/abundância. Sabe-se também que a amostragem em comunidades naturais geralmente é realizada por quadratins e que as populações das espécies na natureza exibem um padrão agregado de distribuição espacial, ao invés de uma distribuição aleatória. Esta dissertação tem como objetivo central avaliar os possíveis efeitos que a estrutura da comunidade imposta por meio de modelos clássicos, como as séries logarítmica e geométrica e a distribuição lognormal, que a agregação populacional e que o método de amostragem (por quadratins ou aleatória de indivíduos) podem ter sobre a estimação dos índices da família de Tsallis. Foram realizadas simulações baseadas no método modificado de Thomas (seguindo outros estudos na área), nas quais foram gerados diferentes cenários, representando comunidades naturais, com os três tratamentos: (i) distribuição de espécie/abundância, (ii) nível de agregação populacional e (iii) método de amostragem. Então foi verificado o desempenho dos métodos de estimação de máxima verossimilhança (EMV), jackknife e bootstrap . O desempenho desses métodos foi medido pelo viés percentual calculado em relação ao parâmetro populacional, pela amplitude do intervalo de confiança e também pela porcentagem de casos em que o intervalo de confiança cobriu o verdadeiro valor do parâmetro populacional. Foi observado que o EMV, em todos os cenários, produziu estimativas mais enviesadas do que os métodos intensivos de estimação... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Community ecology is often studied by the use of diversity indices. Historically, the most important indices are derived from Shannon and from Simpson's formulas. Presently, there is a family of diversity indices, derived from Tsallis entropy, which include the above ones. The species/abundance models are also important to describe the observed patterns in nature. It is well known that, when sampling in natural communities, quadrats are used and that species populations usually exhibit an aggregated pattern of spatial distribution, instead of a random one. The main objective of this dissertation is to evaluate the possible effects that the community structure imposed by the classic species/abundance models, such as the logarithmic and geometric series and the lognormal distribution, the level of aggregation and the sampling method (by quadrats or individual sampling) would interfere with the estimation of the Tsallis diversity indices family. Simulations based on the modified Thomas method were performed (following other studies on the field), in which different scenarios were computationally created, representing the biological communities, with three treatments: (i) the species/abundance model, (ii) the aggregation level and, (iii) the sampling method. Then the performances of the maximum likelihood (MLE), jackknife and boostrap estimators were evaluated. The performances are measured by the relative bias around the population parameter, the length of the confidence intervals and also by the relative number of cases in which the confidence intervals covered the population parameter. It is observed that the MLE, in every scenario, generated more biased estimations than the two evaluated resampling methods. Also it is observed that the geometric series is the species/abundance distribution which caused less bias among the tested models, especially for species richness estimation... (Complete abstract click electronic access below) / Mestre

Made available in DSpace on 2014-06-11T19:23:03Z (GMT). No. of bitstreams: 0 Previous issue date: 2011-02-25Bitstream added on 2014-06-13T20:29:43Z : No. of bitstreams: 1 butturigomes_d_me_botib.pdf: 420214 bytes, checksum: aac06b3aca61cdb7205ba4e36b9d3b5c (MD5) / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / É bastante comum no estudo de ecologia de comunidades o uso de índices de diversidade. Historicamente, os principais índices são os derivados por Shannon e por Simpson. Atualmente, existe uma família de índices, derivada da entropia de Tsallis, que inclui os índices supracitados. Igualmente importantes para descrever os padrões observados na natureza, são os modelos de distribuição de espécie/abundância. Sabe-se também que a amostragem em comunidades naturais geralmente é realizada por quadratins e que as populações das espécies na natureza exibem um padrão agregado de distribuição espacial, ao invés de uma distribuição aleatória. Esta dissertação tem como objetivo central avaliar os possíveis efeitos que a estrutura da comunidade imposta por meio de modelos clássicos, como as séries logarítmica e geométrica e a distribuição lognormal, que a agregação populacional e que o método de amostragem (por quadratins ou aleatória de indivíduos) podem ter sobre a estimação dos índices da família de Tsallis. Foram realizadas simulações baseadas no método modificado de Thomas (seguindo outros estudos na área), nas quais foram gerados diferentes cenários, representando comunidades naturais, com os três tratamentos: (i) distribuição de espécie/abundância, (ii) nível de agregação populacional e (iii) método de amostragem. Então foi verificado o desempenho dos métodos de estimação de máxima verossimilhança (EMV), jackknife e bootstrap . O desempenho desses métodos foi medido pelo viés percentual calculado em relação ao parâmetro populacional, pela amplitude do intervalo de confiança e também pela porcentagem de casos em que o intervalo de confiança cobriu o verdadeiro valor do parâmetro populacional. Foi observado que o EMV, em todos os cenários, produziu estimativas mais enviesadas do que os métodos intensivos de estimação... / Community ecology is often studied by the use of diversity indices. Historically, the most important indices are derived from Shannon and from Simpson’s formulas. Presently, there is a family of diversity indices, derived from Tsallis entropy, which include the above ones. The species/abundance models are also important to describe the observed patterns in nature. It is well known that, when sampling in natural communities, quadrats are used and that species populations usually exhibit an aggregated pattern of spatial distribution, instead of a random one. The main objective of this dissertation is to evaluate the possible effects that the community structure imposed by the classic species/abundance models, such as the logarithmic and geometric series and the lognormal distribution, the level of aggregation and the sampling method (by quadrats or individual sampling) would interfere with the estimation of the Tsallis diversity indices family. Simulations based on the modified Thomas method were performed (following other studies on the field), in which different scenarios were computationally created, representing the biological communities, with three treatments: (i) the species/abundance model, (ii) the aggregation level and, (iii) the sampling method. Then the performances of the maximum likelihood (MLE), jackknife and boostrap estimators were evaluated. The performances are measured by the relative bias around the population parameter, the length of the confidence intervals and also by the relative number of cases in which the confidence intervals covered the population parameter. It is observed that the MLE, in every scenario, generated more biased estimations than the two evaluated resampling methods. Also it is observed that the geometric series is the species/abundance distribution which caused less bias among the tested models, especially for species richness estimation... (Complete abstract click electronic access below)

農林水産研究情報センターで作成したPDFファイルを使用している。

Given a square with integer side length n, we ask for the number of different ways to divide it into sub-squares, considering only the list of parts. We enumerate all possible lists and check whether a placement with those squares is possible. In order to do this, we propose a new algorithm for creating perfect square packings.

Given a square with integer side length n, we ask for the number of different ways to divide it into sub-squares, considering only the list of parts. We enumerate all possible lists and check whether a placement with those squares is possible. In order to do this, we propose a new algorithm for creating perfect square packings.

Termites are important ecosystem engineers in tropical and sub-tropical terrestrial regions where they influence ecosystem processes by altering the physical and chemical structure of the habitat. Termites affect nutrient availability by decomposition and comminution (shredding) of organic matter and act as agents of bioturbation as they re-work substrates during the construction of nests, tunnels and runways. At present we have a relatively good understanding of termite diversity patterns in the tropics through the extensive use of the standardised transect sampling protocol by Eggleton et al. (1995). These diversity data suggest that there is a functional difference in termite assemblage structure, and potentially in termite abundance and biomass, among comparable habitats across continents. However due to the lack of comparable abundance and biomass data from South America this has not previously been confirmed. In this thesis I, therefore, collected extensive data on termite taxonomic and functional assemblage structure in a South American site in Peru. The data were used to compare termite abundance and biomass from two comparable sites in Africa (Cameroon) and south east Asia (Malaysia) in order to gain better understanding of the role termites play in ecosystem processes. I found that there was an intercontinental difference in the abundance and biomass of termite feeding-groups mainly due to the dominance of soil-feeding termites in Cameroon and the absence of fungus-growing termites from Peru. The impact of certain lineages on the intercontinental differences suggests that the differences may be due to biogeographical evolution. Moreover, Eggleton et al. (1998) show that larger-bodied soil-feeding termites in Cameroon process more energy per unit area than predicted by their body size. Due to the need for an examination of the allometric relationships in termite assemblages outside Africa and the development of a more sophisticated feeding-group classification I explore the findings in Eggleton et al. (1998) further using population density - body mass relationships in three termite feeding-groups among the three continental sites in Cameroon, Peru and Malaysia. I found that large-bodied soil-feeding termites in Cameroon and large-bodied wood-feeding termites in Peru had higher population densities than expected by their body masses. As the population density - body mass relationship is inverse to that of the energy - body mass relationship the results suggest that the two feeding-groups also use more energy than expected by their body masses. Further, we have a relatively good understanding of the role termites play as ecosystem engineers e.g. in nutrient cycling and distribution, however, compared with our understanding of wood and litter decomposition in tropical forests quantitative data on the impact of termites in soil processes is poorly understood. In this thesis I conducted, to our knowledge, the first in situ soil macrofauna exclusion experiment using translocated soil in Peru to examine the impact of termites on soil C and N loss. I found that termites promote soil C and N loss which may be linked to the increase in microbial activity due to the passage of soil through the termite gut as well as the affect termites have on bioturbation and nutrient distribution. To conclude, in this thesis I present the first intercontinental comparison of abundance and biomass as well as the first in situ soil macrofauna exclusion experiment to date. The link between termite ecology, biogeography and evolution is discussed as well as the contribution of this thesis to the field of termite ecology.

Esta tesis parte de la hipótesis inicial de que existe una serie de influencias en la creación artística realizada por gemelos y se intentará demostrar a lo largo de la misma en qué aspectos se reflejan y su importancia, que servirá como punto de partida para valorar su singularidad y repercusión. Del mismo modo creemos que es una buena oportunidad para analizar si la influencia genética o ambiental es importante a la hora de seleccionar ser artista y más aún que ambos gemelos seleccionen ser artistas. Con ello seguimos en paralelo a otros estudios sobre gemelos de similares característicos pero ubicados en otra área. Por otro lado, se pretende explorar la propia identidad como gemela, ahondando en propuestas artísticas que reflejen esta inquietud y cómo esto se entrelaza con la teoría del arte contemporáneo. / Rey Jordà, MD. (2013). Anàlisi del procés creatiu i d'identitat d'artistes bessons en l'art contemporani amb propostes d'art al quadrat (2002-2012) [Tesis doctoral]. Editorial Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/33749 / Premios Extraordinarios de tesis doctorales

The thesis analyses the differences that result from using of specific data collection method, computer assisted telephone interviewing (CATI) and computer assisted personal interviewing (CAPI). The main findings are based on election model built by research agency TNS Aisa for elections to the Chamber of Deputies of the Parliament of the Czech Republic in 2013. The aim of the thesis is to determine whether the chosen method of data collection influences the results of the estimated electoral preferences, to determine in which sociodemographic categories it is happening and whether the differences in the obtained data are related to the ideological orientation of the political parties. The dependence of the results on the data collection method is assessed by the chi-quadrate independence test. Further, through personal interviews with researchers, it is determined at what stage of the research process the data may be distorted and what its possible causes are. The benefit of the thesis is the finding that the chosen method of collection influences the established preferences among voters aged 30-44, university graduates and voters living in Prague. In these groups, the left-hand side ČSSD is preferred more by personal interviewing and right-handed TOP 09 through telephone interviewing. The collection of data itself was evaluated by the most risky phase of the research process, in which possible distortion could occur. While the accuracy of personal interviewing depends largely on the interviewer's personality, the sources of distortion of the telephone inquiry result more from the nature of the use of the phone itself.