Katia Ferraz | Universidade de São Paulo (original) (raw)
Papers by Katia Ferraz
Biodiversidade Brasileira, Dec 28, 2023
The Species Distribution Modeling (SDM) that relate the distribution of species with spatial info... more The Species Distribution Modeling (SDM) that relate the distribution of species with spatial information derived, e.g., remote sensing data can help in the understanding of the spatial distribution of organisms optimizing the understanding of ecological processes such as seed dispersal and interactions animal - plant, at different scales. Whereas in situations of high availability of fruits, the diet of the bat's Sturnira lilium can be 100% composed of Solanum variabile, that when are placed networks to capture of bat near to Solanum variabile areas, always are captured Sturnira lilium and that due to the high mobility of bat not possible say the occurrence of Solanum variabile in your areas, the objective of this work was to predict areas of environmental suitability for the species Solanum variabile and Sturnira lilium and assess whether the presence of Solanum variabile can be predicted from the SDM Sturnira lilium. For this we used data of species occurrence and selected env...
Instituto de Investigación de Recursos Biológicos Alexander von Humboldt, Jun 20, 2017
Brazilian Journal of Mammalogy, 2021
Ecology, 2019
Scientists have long been trying to understand why the Neotropical region holds the highest diver... more Scientists have long been trying to understand why the Neotropical region holds the highest diversity of birds on Earth. Recently, there has been increased interest in morphological variation between and within species, and in how climate, topography, and anthropogenic pressures may explain and affect phenotypic variation. Because morphological data are not always available for many species at the local or regional scale, we are limited in our understanding of intra‐ and interspecies spatial morphological variation. Here, we present the ATLANTIC BIRD TRAITS, a data set that includes measurements of up to 44 morphological traits in 67,197 bird records from 2,790 populations distributed throughout the Atlantic forests of South America. This data set comprises information, compiled over two centuries (1820–2018), for 711 bird species, which represent 80% of all known bird diversity in the Atlantic Forest. Among the most commonly reported traits are sex (n = 65,717), age (n = 63,852), b...
Abstract. This paper deals with the use of CBERS2 satellite images (sensor CCD) in order to ident... more Abstract. This paper deals with the use of CBERS2 satellite images (sensor CCD) in order to identify aquatic habitats for Caiman crocodilus e Paleosuchus palpebrosus in the Luiz Eduardo Magalhães reservoir, State of Tocantins, Brazil. Unsupervised classification was performed using the bands 2, 3 and 4. Field observation data were used to calculate the species density in aquatic habitats. Five habitat classes (patches) were identified, and the highest density was observed on patch 4 (61 animals/km2), which represents flooded ...
Understanding the distribution patterns of threatened species is central to conservation. The Ama... more Understanding the distribution patterns of threatened species is central to conservation. The Amazonian distribution of the northern tiger cat (N-tiger cat, Leopardus tigrinus) and its interspecific relationship with the ocelot, its potential intraguild killer, are intriguing. Here, we combined presence/absence records with species distribution models (SDMs) to determine N-tiger cat occurrence in the Amazon. We also modeled the ocelot density from 46 published estimates. The N-tiger cat’s presence in the Amazon was negatively influenced by the ocelot density and net primary productivity and positively influenced by savannas and precipitation in the driest month. The best-fitting model predicted highly patchy N-tiger cat occurrence over an area of 236,238.67 km2, almost exclusively in savanna enclaves. Additionally, 312,348 camera trap-days at 49 sites in the Amazon revealed no N-tiger cats. The ocelot densities were significantly higher in denser vegetation cover and warmer habitats...
In this slides we share an ongoing work presented at AGU Fall Meeting 2020. We present an approac... more In this slides we share an ongoing work presented at AGU Fall Meeting 2020. We present an approach and methods we have used to date. Reviewing four case studies in this domain (Xi et al., 2016; Jean et al., 2016; Suel et al., 2019; Ayush et al., 2020) we found a common three-stage learning methodology: (1) establish a "preliminary task" (pre-trained model) consisting to train a CNN (convolutional neural network) using a large dataset of images, aiming to intensively learn the relationship between the input and their images annotations (intermediate outputs); (2) extract a feature vector from the CNN output which will be used as the transferability learning, so that each input image would correspond to a feature vector and could be annotated with a socioeconomic indicator; and (3) use a simpler regression model to predict poverty measures from the corresponding CNN feature vector output.<br>
Oecologia Australis, 2019
Biodiversity in Agricultural Landscapes of Southeastern Brazil, 2016
Ecological Indicators, 2016
ABSTRACT O presente trabalho teve como objetivo mapear as áreas aptas ao material de origem do Eu... more ABSTRACT O presente trabalho teve como objetivo mapear as áreas aptas ao material de origem do Eucalyptus grandis Hill ex Maiden no Brasil para o clima atual, e predizer possíveis mudanças nestes locais frente os cenários climáticos futuros. Para tal se utilizou a Modelagem de Distribuição de Espécies (MDE), gerando áreas potenciais na Austrália e projetando-as para o Brasil no tempo presente e futuro, utilizando-se o princípio da máxima entropia (Maxent, 3.3.3k.). Foram utilizados 70 pontos de ocorrência natural da espécie na Austrália e sete variáveis bioclimáticas, sendo: temperatura média anual, variação da temperatura anual, precipitação anual, precipitação do mês mais chuvoso, precipitação do mês mais seco, variação da precipitação e altitude. Para a modelagem do clima atual, o período de tempo utilizado foi de 1950 a 2000. As projeções climáticas foram retratadas por meio do cenário A1B e o modelo HadCM3 para os três períodos de tempo: 2010 - 2039, 2040 - 2069 e 2070 - 2099. Todos os modelos foram significativos (p<0,001), apresentaram elevados valores de AUC (> 0,95) e baixos erros de omissão. A área adequada para o material genético testado no Brasil no tempo presente foi de, aproximadamente, 1.500.000 km², concentrando-se nas regiões sul, sudeste e centro-oeste. Para os cenários futuros a mesma área sofreu redução de 2,8, 4,7 e 3,8% para os cenários 2010 - 2039, 2040 - 2069 e 2070 - 2099, respectivamente. As principais mudanças foram a diminuição da área na região sudeste e aumento na região norte. Para os cenários futuros, a modelagem mostrou uma diminuição da área total da espécie. Embora novas áreas tenham sido consideradas aptas, houve uma diminuição das áreas já conhecidas como adequadas. O uso da modelagem pode ser útil no planejamento do melhoramento genético e na expansão do material genético para novas regiões, além de auxiliar na identificação de áreas em que a cultura se torne mais vulneráveis ao clima, doenças e pragas.
Brazilian Journal of Biology, 2006
Capybara, 2012
When the Iberian colonists arrived in South America in the late fifteenth century, they encounter... more When the Iberian colonists arrived in South America in the late fifteenth century, they encountered a diverse and previously unimagined fauna. The unusual anatomy and behavior of these species intrigued the early explorers. In their reports they named the new-found endemic animals after the most analogous European species. In 1576, for example, Pero de Gândavo (2004) described the capybara (Hydrochoerus hydrochaeris) as “a type of pig.” However, capybaras were sufficiently unlike any known European species for most explorers to simply adopt a phonetic representation of the local name. Therefore, in 1557, the capybara was called catiuare by the German Hans Staden (1557), capiyuara in 1560 by the Spaniard Jose de Anchieta (1997), and capijuara in 1625 by the Portuguese Fernao Cardim (1980). The name capybara actually originates from a word in the indigenous Tupi, which in the sixteenth century was the most widely spread language in South America: kapii’gwara meaning grass eater (ka’pii = “grass” + gwara = “eater”; Houaiss et al. 2004).
Journal of Biogeography, 2007
Biodiversidade Brasileira, Dec 28, 2023
The Species Distribution Modeling (SDM) that relate the distribution of species with spatial info... more The Species Distribution Modeling (SDM) that relate the distribution of species with spatial information derived, e.g., remote sensing data can help in the understanding of the spatial distribution of organisms optimizing the understanding of ecological processes such as seed dispersal and interactions animal - plant, at different scales. Whereas in situations of high availability of fruits, the diet of the bat's Sturnira lilium can be 100% composed of Solanum variabile, that when are placed networks to capture of bat near to Solanum variabile areas, always are captured Sturnira lilium and that due to the high mobility of bat not possible say the occurrence of Solanum variabile in your areas, the objective of this work was to predict areas of environmental suitability for the species Solanum variabile and Sturnira lilium and assess whether the presence of Solanum variabile can be predicted from the SDM Sturnira lilium. For this we used data of species occurrence and selected env...
Instituto de Investigación de Recursos Biológicos Alexander von Humboldt, Jun 20, 2017
Brazilian Journal of Mammalogy, 2021
Ecology, 2019
Scientists have long been trying to understand why the Neotropical region holds the highest diver... more Scientists have long been trying to understand why the Neotropical region holds the highest diversity of birds on Earth. Recently, there has been increased interest in morphological variation between and within species, and in how climate, topography, and anthropogenic pressures may explain and affect phenotypic variation. Because morphological data are not always available for many species at the local or regional scale, we are limited in our understanding of intra‐ and interspecies spatial morphological variation. Here, we present the ATLANTIC BIRD TRAITS, a data set that includes measurements of up to 44 morphological traits in 67,197 bird records from 2,790 populations distributed throughout the Atlantic forests of South America. This data set comprises information, compiled over two centuries (1820–2018), for 711 bird species, which represent 80% of all known bird diversity in the Atlantic Forest. Among the most commonly reported traits are sex (n = 65,717), age (n = 63,852), b...
Abstract. This paper deals with the use of CBERS2 satellite images (sensor CCD) in order to ident... more Abstract. This paper deals with the use of CBERS2 satellite images (sensor CCD) in order to identify aquatic habitats for Caiman crocodilus e Paleosuchus palpebrosus in the Luiz Eduardo Magalhães reservoir, State of Tocantins, Brazil. Unsupervised classification was performed using the bands 2, 3 and 4. Field observation data were used to calculate the species density in aquatic habitats. Five habitat classes (patches) were identified, and the highest density was observed on patch 4 (61 animals/km2), which represents flooded ...
Understanding the distribution patterns of threatened species is central to conservation. The Ama... more Understanding the distribution patterns of threatened species is central to conservation. The Amazonian distribution of the northern tiger cat (N-tiger cat, Leopardus tigrinus) and its interspecific relationship with the ocelot, its potential intraguild killer, are intriguing. Here, we combined presence/absence records with species distribution models (SDMs) to determine N-tiger cat occurrence in the Amazon. We also modeled the ocelot density from 46 published estimates. The N-tiger cat’s presence in the Amazon was negatively influenced by the ocelot density and net primary productivity and positively influenced by savannas and precipitation in the driest month. The best-fitting model predicted highly patchy N-tiger cat occurrence over an area of 236,238.67 km2, almost exclusively in savanna enclaves. Additionally, 312,348 camera trap-days at 49 sites in the Amazon revealed no N-tiger cats. The ocelot densities were significantly higher in denser vegetation cover and warmer habitats...
In this slides we share an ongoing work presented at AGU Fall Meeting 2020. We present an approac... more In this slides we share an ongoing work presented at AGU Fall Meeting 2020. We present an approach and methods we have used to date. Reviewing four case studies in this domain (Xi et al., 2016; Jean et al., 2016; Suel et al., 2019; Ayush et al., 2020) we found a common three-stage learning methodology: (1) establish a "preliminary task" (pre-trained model) consisting to train a CNN (convolutional neural network) using a large dataset of images, aiming to intensively learn the relationship between the input and their images annotations (intermediate outputs); (2) extract a feature vector from the CNN output which will be used as the transferability learning, so that each input image would correspond to a feature vector and could be annotated with a socioeconomic indicator; and (3) use a simpler regression model to predict poverty measures from the corresponding CNN feature vector output.<br>
Oecologia Australis, 2019
Biodiversity in Agricultural Landscapes of Southeastern Brazil, 2016
Ecological Indicators, 2016
ABSTRACT O presente trabalho teve como objetivo mapear as áreas aptas ao material de origem do Eu... more ABSTRACT O presente trabalho teve como objetivo mapear as áreas aptas ao material de origem do Eucalyptus grandis Hill ex Maiden no Brasil para o clima atual, e predizer possíveis mudanças nestes locais frente os cenários climáticos futuros. Para tal se utilizou a Modelagem de Distribuição de Espécies (MDE), gerando áreas potenciais na Austrália e projetando-as para o Brasil no tempo presente e futuro, utilizando-se o princípio da máxima entropia (Maxent, 3.3.3k.). Foram utilizados 70 pontos de ocorrência natural da espécie na Austrália e sete variáveis bioclimáticas, sendo: temperatura média anual, variação da temperatura anual, precipitação anual, precipitação do mês mais chuvoso, precipitação do mês mais seco, variação da precipitação e altitude. Para a modelagem do clima atual, o período de tempo utilizado foi de 1950 a 2000. As projeções climáticas foram retratadas por meio do cenário A1B e o modelo HadCM3 para os três períodos de tempo: 2010 - 2039, 2040 - 2069 e 2070 - 2099. Todos os modelos foram significativos (p<0,001), apresentaram elevados valores de AUC (> 0,95) e baixos erros de omissão. A área adequada para o material genético testado no Brasil no tempo presente foi de, aproximadamente, 1.500.000 km², concentrando-se nas regiões sul, sudeste e centro-oeste. Para os cenários futuros a mesma área sofreu redução de 2,8, 4,7 e 3,8% para os cenários 2010 - 2039, 2040 - 2069 e 2070 - 2099, respectivamente. As principais mudanças foram a diminuição da área na região sudeste e aumento na região norte. Para os cenários futuros, a modelagem mostrou uma diminuição da área total da espécie. Embora novas áreas tenham sido consideradas aptas, houve uma diminuição das áreas já conhecidas como adequadas. O uso da modelagem pode ser útil no planejamento do melhoramento genético e na expansão do material genético para novas regiões, além de auxiliar na identificação de áreas em que a cultura se torne mais vulneráveis ao clima, doenças e pragas.
Brazilian Journal of Biology, 2006
Capybara, 2012
When the Iberian colonists arrived in South America in the late fifteenth century, they encounter... more When the Iberian colonists arrived in South America in the late fifteenth century, they encountered a diverse and previously unimagined fauna. The unusual anatomy and behavior of these species intrigued the early explorers. In their reports they named the new-found endemic animals after the most analogous European species. In 1576, for example, Pero de Gândavo (2004) described the capybara (Hydrochoerus hydrochaeris) as “a type of pig.” However, capybaras were sufficiently unlike any known European species for most explorers to simply adopt a phonetic representation of the local name. Therefore, in 1557, the capybara was called catiuare by the German Hans Staden (1557), capiyuara in 1560 by the Spaniard Jose de Anchieta (1997), and capijuara in 1625 by the Portuguese Fernao Cardim (1980). The name capybara actually originates from a word in the indigenous Tupi, which in the sixteenth century was the most widely spread language in South America: kapii’gwara meaning grass eater (ka’pii = “grass” + gwara = “eater”; Houaiss et al. 2004).
Journal of Biogeography, 2007