Modelos de abundância e distribuição espacial de peixes de riachos
Data
2019-11-08
Autores
Gomes, Cristal Coelho 
Orientadores
Ferreira, Fabio Cop
Tipo
Dissertação de mestrado
Título da Revista
ISSN da Revista
Título de Volume
Resumo
A distribuição e abundância de peixes de riachos é influenciada por características ambientais associadas a múltiplas escalas espaciais e temporais. Esses organismos apresentam histórias de vida complexas e ocupam hábitats distintos ao longo do seu ciclo de vida. Conhecer em que medida as escalas espaciais, e seus respectivos filtros se relacionam à dinâmica e estrutura das populações de riachos é fundamental para o entendimento de processos e padrões ecológicos que se manifestam da associação dos organismos com os ambientes que ocupam. Uma expressão fundamental da associação entre os organismos e o ambiente é a abundância local, mas a geração dessa informação é frequentemente laboriosa e depende de estimativas. Entretanto, estimativas da abundância local são influenciadas por erros de origens diversas, como aqueles oriundos de processos observacionais, do delineamento do estudo, do estimador utilizado, entre outros. Lidar com estes erros é fundamental para a adequada aplicação da estimativa gerada. Nessa dissertação apresentaremos dois artigos, nos quais trataremos desses dois processos complementares para o entendimento das dinâmicas naturais de populações de peixes de riachos. No Capítulo I, em uma primeira seção, estimamos a abundância do bagre R. transfasciatus por meio de um estudo de marcação e recaptura, contrastando diferentes hipóteses sobre como diferentes fontes de erros de observação afetam a estimativa de abundância local. Verificamos que o modelo que considera variabilidade na probabilidade de captura proveniente de heterogeneidade individual apresenta o melhor ajuste entre os modelos testados, tendo sido estimada por esse modelo abundância mais provável de 1.023 indivíduos associados aos 400 m de riacho estudado. Na segunda seção do Capítulo I realizamos uma simulação computacional para avaliar como diferentes delineamentos amostrais afetam a precisão da estimativa. Foi verificado que a intensificação de esforços de amostragem, seja em número de indivíduos capturados por ocasião de amostragem ou pelo número de eventos de amostragem, aumenta a precisão da estimativa, permitindo a aplicação do método para objetivos que demandam menores margens de erro na estimativa gerada. No Capítulo II foram ajustados modelos hierárquicos zero-inflacionados para descrever os padrões de ocorrência e abundância de quatro espécies de peixes de riacho em função de variáveis ambientais consideradas em diferentes escalas espaciais, partindo da premissa de que essas variáveis poderiam atuar, de forma diferenciada para cada espécie, como filtros ambientais para a ocorrência e abundância dessas espécies. Foi observado que as espécies respondem a variáveis ambientais em diferentes escalas de maneira simultânea. Para cada espécie foi verificado um conjunto específico de variáveis preditoras, tendo sido possível observar que o efeito do aninhamento oriundo da configuração dos ambientes de riachos foi adequadamente transmitida pelas diferentes escalas consideradas nos modelos
Stream fish distribution and abundance are influenced by environmental characteristics associated with multiple spatial and temporal scales. These organisms present complex life histories and occupy distinct habitats throughout their life cycles. The knowledge about the extent to which spatial scales, and their respective environmental filters relate to the dynamics and structure of stream populations is fundamental to the understanding ecological processes and patterns. Beyond the ecological process itself, a fundamental expression of the association between the organisms and the environment is local abundance. However, estimates of local abundance are affected by errors of variable origins, such as the ones associated with observational process, to the study design, to the estimator used, among others. Dealing with those errors are thus essential to environmental planning for species conservation. In this dissertation we present two articles in which we will deal with these two complementary processes that influence our comprehension of dynamics of natural populations. In a first section of Chapter I we estimated the abundance of the catfish R. transfasciatus by means of a mark-recapture study contrasting different hypotheses on how the behavior of the observational errors affect the estimate of local abundance. In a subsequent section of this Chapter, we performed a computational simulation to evaluate how different sampling designs affect the local abundance estimate accuracy. We verified in first section that the model which presented the best fit were the one whose considers the individual heterogeneity in capture probabilities. The most probable abundance estimate of R. transfasciatus generated by this model was of 1023 individuals associated to the 400 m long studied stream segment. In second section it was observed that the intensification of sampling, by capturing more individuals in each sampling occasion or by increasing the number of sampling occasions, enhances the estimates performances, raising estimate accuracy. In Chapter II it was fitted zero-inflated hierarchical models to describe patterns of abundance and occurrence of four stream fish species as functions of environmental variables considered at different scales. With these models we evaluated which variables, considered in different spatial scales, represent environmental filters to local abundance and occurrence. It was observed that species respond to environmental variables of different scales simultaneously. To each specie was observed a different set of environmental variables acting as environmental filters, have been successfully reproduced the nested environmental structure of riverine environments through the different scales considered
Stream fish distribution and abundance are influenced by environmental characteristics associated with multiple spatial and temporal scales. These organisms present complex life histories and occupy distinct habitats throughout their life cycles. The knowledge about the extent to which spatial scales, and their respective environmental filters relate to the dynamics and structure of stream populations is fundamental to the understanding ecological processes and patterns. Beyond the ecological process itself, a fundamental expression of the association between the organisms and the environment is local abundance. However, estimates of local abundance are affected by errors of variable origins, such as the ones associated with observational process, to the study design, to the estimator used, among others. Dealing with those errors are thus essential to environmental planning for species conservation. In this dissertation we present two articles in which we will deal with these two complementary processes that influence our comprehension of dynamics of natural populations. In a first section of Chapter I we estimated the abundance of the catfish R. transfasciatus by means of a mark-recapture study contrasting different hypotheses on how the behavior of the observational errors affect the estimate of local abundance. In a subsequent section of this Chapter, we performed a computational simulation to evaluate how different sampling designs affect the local abundance estimate accuracy. We verified in first section that the model which presented the best fit were the one whose considers the individual heterogeneity in capture probabilities. The most probable abundance estimate of R. transfasciatus generated by this model was of 1023 individuals associated to the 400 m long studied stream segment. In second section it was observed that the intensification of sampling, by capturing more individuals in each sampling occasion or by increasing the number of sampling occasions, enhances the estimates performances, raising estimate accuracy. In Chapter II it was fitted zero-inflated hierarchical models to describe patterns of abundance and occurrence of four stream fish species as functions of environmental variables considered at different scales. With these models we evaluated which variables, considered in different spatial scales, represent environmental filters to local abundance and occurrence. It was observed that species respond to environmental variables of different scales simultaneously. To each specie was observed a different set of environmental variables acting as environmental filters, have been successfully reproduced the nested environmental structure of riverine environments through the different scales considered
Descrição
Citação
GOMES, Cristal Coelho. Modelos de abundância e distribuição espacial de peixes de riachos. 2019. 88 f. Dissertação (Mestrado em Biodiversidade e Ecologia Marinha e Costeira) - Instituto do Mar, Universidade Federal de São Paulo, Santos, 2019.