Orientador: Jansle Vieira Rocha / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Agricola / Made available in DSpace on 2018-08-11T14:29:10Z (GMT). No. of bitstreams: 1 Weber_LianedeSouza_D.pdf: 866084 bytes, checksum: 6f14fe14fef394e2bd299f45b7c98915 (MD5) Previous issue date: 2005 / Resumo: O presente trabalho descreve um estudo sobre estimativa de safras cujo principal objetivo foi criar uma metodologia de integração de dados de produção, dados espectrais e indicadores meteorológicos por meio de redes neurais artificiais, estabelecendo correlações entre índices de vegetação e de produtividade, com o propósito de estimar a produtividade de cana-de-açúcar. O estudo foi dividido em duas etapas: a primeira correspondeu à obtenção e organização dos dados em um banco de dados com padrões de entrada/saída; a segunda, à implementação e ajuste das redes neurais, por meio de ensembles. O estudo foi realizado em unidades amostrais de produção de uma usina sucroalcooleira no município de Araras-SP. A primeira etapa consistiu na obtenção dos coeficientes de produtividade (kp), por meio da inversão do modelo agrometeorológico de Doorenbos e Kassam (1979), a partir da determinação do balanço hídrico. O resultado deste procedimento mostrou a sensibilidade do coeficiente à variabilidade da produtividade nos talhões. Os dados espectrais das imagens Landsat 7 ¿ ETM+ foram obtidos de correlações descritas na literatura estabelecidas entre o Índice de Vegetação Greenness (GVI), a banda do infravermelho próximo (B4) e a produtividade da cana-de-açúcar. A estratégia para treinamento dos ensembles foi baseada no aprendizado supervisionado aplicado a uma arquitetura Multilayer Perceptron (MLP), com uma camada escondida, método de aprendizado de 2ª ordem e feedforward. Na etapa de treinamento e validação, as redes neurais tiveram como variáveis de entrada os valores de kp, GVI e B4, e como variável de saída a produtividade, que definiram os padrões de entrada/saída. A fase de teste consistiu em implementar a metodologia em um grupo de padrões de entrada não utilizados nos treinamentos. Os resultados mostraram valores de EQM entre 0,03 e 0,51 ton/ha, enquanto que a estimativa da usina errou em média 9,93 ton/ha, o que garantiu o correto ajuste da rede neural quanto à topologia, ao número de iterações e aos algoritmos de aprendizagem. Esta etapa mostrou a capacidade de generalização da rede neural, já que os treinamentos foram realizados a partir de unidades amostrais. O estudo ratificou a aplicação desta metodologia na determinação da estimativa de produtividade de cana-deaçúcar, empregando-a como técnica complementar aos atuais métodos de estimativa agrícola, sugerindo a ampliação da escala de aplicação para o ambiente de produção da usina / Abstract: The present thesis describes a study on crop forecast. Its main purpose was to create a methodology for integrating production, spectral and meteorological data indicators through artificial neural networks, establishing correlations between vegetation index and yield coefficients, aiming at the estimate of sugarcane yield. The study was divided in two parts: the first corresponded to obtaining and organizing data in a database with input/output default; the second corresponded to the implementation and adjustment of the neural network. The study was carried out in sample production units (fields) of a sugarmill agricultural area located in the municipality of Araras-SP, Brazil. The first part consisted in obtaining yield coefficients (kp) through the inversion of the Doorenbos-Kassam (1979) agrometeorological model, based on the determination of the water balance. The result of this procedure showed the coefficient¿s sensitivity to the variability of yield within the sample fields. The spectral data of the Landsat 7 ¿ ETM+ images were obtained from correlations, available in scientific literature, between the Greenness Vegetation Index (GVI), near infrared band, and sugarcane yield. The strategy for training the neural network was based on supervised learning applied to a Multilayer Perceptron (MLP) architecture, with a hidden layer, second order learning method and feedforward. For the training and validation stage, the neural network had as input variables kp, GVI and B4 values, and as output variable the yield, both obtained in the input/output database. The test stage consisted of implementing the methodology in a set of input patterns not used for the trainings. The results showed Mean Square Error (MSE) values between 0,03 and 0,51 ton/ha, while the average error of the sugarmill estimates were 9,93 ton/ha, which showed the correct adjustment of the network concerning topology, number of iterations and learning algorithms. This showed the generalization capacity of the neural network once the trainings were carried out based on sample units. The study ratified the application of this methodology for determining sugarcane yield estimate, employing it as a complementary technique to the present methods of agricultural forecast, suggesting the increase of the application scale to a broader area of the sugarmill production environment / Doutorado / Planejamento e Desenvolvimento Rural Sustentável / Doutor em Engenharia Agrícola

Submitted by Neusa Fagundes (neusa.fagundes@unioeste.br) on 2018-06-15T19:26:06Z No. of bitstreams: 2 Carlos_Cattani2018.pdf: 4966391 bytes, checksum: ea80b3cc7502421dface694a3d4967f9 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-06-15T19:26:06Z (GMT). No. of bitstreams: 2 Carlos_Cattani2018.pdf: 4966391 bytes, checksum: ea80b3cc7502421dface694a3d4967f9 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-02-05 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / The temporal dynamics of agricultural crops can be translated by Vegetation Indices (IV) on multiple dates. The use of IVs in certain stages of development, or throughout their entire cycle, are related to their biophysical parameters. There are several IVs, and the most commonly used are those based on the red (RED) and near infrared (NIR) bands, although studies with IVs that used the NIR and medium infrared (SWIR) bands have shown good results in the estimation of biophysical parameters of agricultural crops. In this context, the objective of this work was to characterize the spectral-temporal profiles of different vegetation indexes (RVI, MSR, NDVI, MSAVI, NDRE, WRDVI, and NDMI) and to correlate these profiles with the biophysical parameters (IAF) and photosynthetically active intercepted radiation (RFAI). This research was carried out in commercial farming areas of wheat and beans. For the characterization of the IRs, two terrestrial remote sensors were employed: the FieldSpec4 passive resistive passive sensor (FS) and the GreenSeeker 505 Handheld (GS) active sensor. Both LAI-2200C and LI-191R sensors, respectively, were used to obtain the biophysical parameters, leaf area index (LAI), and photosynthetically active intercepted radiation (RFAI). The biophysical variable IAF for both cultures showed better results when estimated by the NDMI FS index, presenting the highest values of correlation (rs) and coefficient of agreement (dr), and the lowest errors (ME and RMSE). For wheat, the RFAI variable obtained the best fit with the NDMI index, presenting a satisfactory result according to performance index. For beans, RFAI presented a higher correlation with the MSAVI index, but showed a low degree of agreement between the adjusted and the observed data, obtaining low efficiency in the model. / A dinâmica temporal das culturas agrícolas pode ser traduzida por meio de índices de vegetação (IV) em múltiplas datas. A utilização de IVs em determinados estádios de desenvolvimento, ou durante todo o seu ciclo, possuem boas relações com os seus parâmetros biofísicos. Há diversos IVs, sendo que os mais comumente utilizados são os baseados nas bandas do vermelho (RED) e infravermelho próximo (NIR); porém, estudos com IVs que utilizaram as bandas do NIR e a do infravermelho médio (SWIR) têm demostrado bons resultados na estimativa de parâmetros biofísicos de culturas agrícolas. Neste sentido, o objetivo deste trabalho foi caracterizar os perfis espectro-temporais de diferentes índices de vegetação (RVI, MSR, NDVI, MSAVI, NDRE, WRDVI e NDMI) e correlacionar esses perfis com os parâmetros biofísicos, sendo eles o índice de área foliar (IAF) e a radiação fotossinteticamente ativa interceptada (RFAI). Este trabalho foi desenvolvido em áreas agrícolas comerciais, abordando trigo e feijão. Para a caracterização dos IVs foram utilizados dois sensores remotos terrestres: o sensor passivo hiperespectral FieldSpec4 Standart-Res (FS), e o sensor ativo GreenSeeker 505 Handheld (GS). Para a obtenção dos parâmetros biofísicos, índice de área foliar (IAF) e radiação fotossinteticamente ativa interceptada (RFAI), foram utilizados os sensores LAI-2200C e LI-191R, respectivamente. A variável biofísica IAF, para ambas as culturas, mostrou melhores resultados ao serem estimadas pelo índice NDMI FS, apresentando os maiores valores de correlação (rs) e coeficiente de concordância (dr) e os menores erros (ME e RMSE). Para o trigo, a variável RFAI obteve o melhor ajuste com o índice NDMI, apresentando resultado satisfatório, segundo índice de performance. Já para o feijão, o RFAI apresentou maior correlação com o índice MSAVI; entretanto, mostrou baixo grau de concordância entre os dados ajustados e observados, obtendo baixa eficiência no modelo.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-10-31T13:32:58Z No. of bitstreams: 1 isabelafernandamoraesdepaula.pdf: 8651561 bytes, checksum: 51b9a252199579213a423569927ae229 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-11-09T14:05:14Z (GMT) No. of bitstreams: 1 isabelafernandamoraesdepaula.pdf: 8651561 bytes, checksum: 51b9a252199579213a423569927ae229 (MD5) / Made available in DSpace on 2017-11-09T14:05:14Z (GMT). No. of bitstreams: 1 isabelafernandamoraesdepaula.pdf: 8651561 bytes, checksum: 51b9a252199579213a423569927ae229 (MD5) Previous issue date: 2017-07-27 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / A cobertura vegetal, entendida como o conjunto das tipologias arbórea, arbustiva e rasteira, constitui um importante elemento a compor os centros urbanos em função dos diversos benefícios de cunho ecológico, social e estético que proporcionam. Nesse sentido, o conhecimento de sua ocorrência no espaço torna-se fundamental tanto para se identificar as funcionalidades destes no meio quanto para se instituir políticas ambientais que visem melhorias na sua disposição. Sendo assim, considerando a relevância deste recurso natural, esta pesquisa buscou identificar a vegetação existente nas Regiões Administrativas do município de Juiz de Fora. Os resultados obtidos demonstram uma distribuição destituída de homogeneidade, evidenciando maiores fragmentos nas áreas periféricas e reduzidos remanescentes no sentido noroeste-sudeste, revelando um processo de uso e ocupação da terra mais intenso nesses locais. Nota-se, dessa forma, a ocorrência de manchas de vegetação mais pontuais e espaçadas nesses trechos, típicas da categoria Isolated, definida pelo pesquisador Jim como dominante de áreas impermeabilizadas, enquanto nas áreas próximas as bordas dos limites, encontram-se o tipo Connected, caracterizado pela presença de verde urbano mais contíguo e de maiores extensões. Tais aspectos refletem diretamente nos resultados dos índices de cobertura vegetal obtidos, no qual a Região Centro abrange os menores valores em relação as demais. No que se refere aos espaços de integração urbana e espaços livres de construção, observa-se escassez do elemento vegetativo nesses sistemas, visto que o primeiro abrange menos de 2,1% das regiões e o segundo menos de 1,5%, o que demonstra a necessidade de melhorias nesses ambientes, de forma a contribuir mais efetivamente ao bem estar humano. / The vegetal covering, understood as a set of arboreal typology, creeper shrub, consists of an important element that composes the cities and provide a lot of befits such as ecological, social, and esthetical. In this way, the knowledge about vegetal covering throughout the cities is an important issue in order to identify its best features for communities as well as to promote environmental policies for improves the vegetal covering displacement around cities. Thus, regarding the relevance of this natural resource, the aim of this work was identify the vegetal covering that exists at administrative regions of Juiz de Fora city. On the one hand the results shown that the vegetal covering is homogeneity distributed with larges fragments at peripheral regions. On the other hand a small remaining of vegetal covering was found at northwest and southeast regions of city which reveal an intensive process of use and occupation of the ground. Thus, we found sparse and punctual vegetal stains at downtown, typical of isolated category, defined by researcher Jim as impermeable dominants areas. While at cities’ board areas, it can be found connected type vegetal covering which is characterized by continuous green urban presence with larger extension. Such issues reflect directly at vegetal covering index results we have found, in which the downtown indexes are smaller than others regions evaluated. As far as urban spaces of integration as well as building free regions are concerned we can observe scarce vegetative elements. At urban spaces of integration we find only 2,1 % of regions have vegetative elements, while at building open spaces has 1,5 % as well. Therefore, the results have shown that there are demands for improvements in these environments in order to contribute effectively for the human well-being.

The use of Normalized Difference Vegetation Index (NDVI) time series over the last decades has increased our understanding of vegetation change dynamics from global to regional scale through quantitative analysis of inter-annual trends in NDVI and climatological parameters (rainfall and temperature). Change in land cover induced by human activities such as livestock grazing and deforestation for large-scale farming (subsistence and mechanized) has influenced the ecological pattern of the Guinea savannah region (GSR) of Nigeria, thereby resulting in loss of biodiversity and changes in vegetation cover. In the context of the GSR of Nigeria where agriculture still plays a major role in people’s economy, it is important to identify the relationship between climatic variables, vegetation productivity and human activities which can be used to understand the on-going transition processes. This study, therefore, examines the spatial and temporal relationship between NDVI and climate parameters, land use land cover change (LULCC) and the perspective of local people on vegetation change dynamics in the study region. In order to do this, bi-monthly NDVI3g time series datasets from Global Inventory Modeling and Mapping Studies (GIMMS), monthly rainfall datasets from Tropical Applications of Meteorology Satellite (TAMSAT), monthly temperature datasets from Climate Research Unit (CRU), national land use land cover (LULC) data of Nigeria from Forestry Management Evaluation & Coordination Unit (FORMECU), global land cover datasets from European Space Agency, Landsat imagery and socio-economic field data collection were used in order to understand vegetation change dynamics across the Guinea savannah regions of Nigeria. Time series analysis (TSA) was applied to both NDVI and climate data used in order to examine the temporal dynamics of vegetation cover change and to detect NDVI-climate relationship during the period from 1983 through 2011. Both parametric and non-parametric statistical models were employed for the assessment of long-term inter-annual trend on the decomposed time series datasets for the whole region (Guinea savannah region) and selected locations. In addition to the TSA, harmonic regression analysis was performed on NDVI and rainfall datasets in order to examine change in seasonality and phyto-phenological characteristics of vegetation. Detection of change in land use and land cover was done by extracting information from existing land cover datasets (ancillary datasets). CLASlite was used for the assessment of the extent of deforestation, while linkage between remotely sensed data and social science was carried out via field surveys based on questionnaires in order to understand the drivers of vegetation change. The study reveals that about 90 % of the Guinea savannah region show positive NDVI trends which indicate greening over time, while about 10 % of the region shows negative trends. This greening trends are closely related to regions where intensive agriculture is being practiced (also along inland valleys) while regions with negative trends show significant loss in woodlands (forest and shrublands) as well as herbaceous vegetation cover due to over-grazing by agro-pastoralism. The result confirms that there is a good relationship (statistically significant positive correlation) between rainfall and NDVI both on intra-annual and inter annual time scale for some selected locations in the study region (> 65 %), while negative statistical correlation exists between NDVI and temperature in the selected locations. This implies that vegetation growth (productivity) in the region is highly dependent on rainfall. The result of the harmonic regression analysis reveals a shift in the seasonal NDVI pattern, indicating an earlier start and a more prolonged growing season in 2011 than in 1983. This study proves significant change in LULC with evidence of an increase in the spatial extent of agricultural land (+ 30 %) and loss of woodlands (- 55 %) between 2000 and 2009 for Kogi State. The results of the socio-economic analysis (people’s perception) highlight that vegetation change dynamics in the study region are the resultant effects of increased anthropogenic activities rather than climatic variability. This study couples data from remote sensing and ground survey (socio-economics) for a better understanding of greening trend phenomena across the Guinea savannah region of Nigeria, thus filling the gap of inadequate information on environmental condition and human perturbation which is essential for proper land use management and vegetation monitoring.

Příloha k disertační práci: Abstrakt v AJ (Mgr. Josef Laštovička) Abstract This dissertation thesis deals with the study of forest ecosystems in the central Europe with the time series of multispectral optical satellite data. These forest ecosystems have been influenced by biotic and abiotic disturbances for the last decade. The time series of the satellite data with high spatial resolution allow the detection and analysis of forest disturbances. This thesis is mainly focused primally on free available Landsat and Sentinel-2 data, these two data types were compared. From methods, the difference time series analyses / algorithms were used. The whole thesis can be divided into two main parts. The first one analyses usability of classifiers for detection of forest ecosystems with per-pixel and sub-pixel methods. Specifically, the Neural Network, the Support Vector Machine and the Maximum Likelihood per-pixel classifiers were used and compared for different types of data (for data with high spatial resolution - Landsat or Sentinel-2; very high spatial resolution - WorldView-2) and for classification of protected forest areas. The Support Vector Machine were selected as the most suitable method for forest classifications (with most accurate outputs) from the list of selected per-pixel classifiers. Also, Spectral...

Digital soil mapping typically involves inputs of digital elevation models, remotely sensed imagery, and other spatially explicit digital data as environmental covariates to predict soil classes and attributes over a landscape using statistical models. Digital imagery from Landsat 5, a digital elevation model, and a digital geology map were used as environmental covariates in a 67,000-ha study area of the Great Basin west of Fillmore, UT. A “pre-map” was created for selecting sampling locations. Several indices were derived from the Landsat imagery, including a normalized difference vegetation index, normalized difference ratios from bands 5/2, bands 5/7, bands 4/7, and bands 5/4. Slope, topographic curvature, inverse wetness index, and area solar radiation were calculated from the digital elevation model. The greatest variation across the study area was found by calculating the Optimum Index Factor of covariates, choosing band 7, normalized difference ratio bands 5/2, normalized difference vegetation index, slope, profile curvature, and area solar radiation. A 20-class ISODATA unsupervised classification of these six data layers was reduced to 12. Comparing the 12-class map to a geologic map, 166 sites were chosen weighted by areal extent; 158 sites were visited. Twelve points were added using case-based reasoning to total 170 points for model training. A validation set of 50 sites was selected using conditioned Latin Hypercube Sampling. Density plots of sample sets compared to raw data produced comparable results. Geology was used to stratify the study area into areas above and below the Lake Bonneville highstand shoreline. Raster data were subset to these areas, and predictions were made on each area. Spatial modeling was performed with three different models: random forests, support vector machines, and bagged classification trees. A set of covariates selected by random forests variable importance and the set of Optimum Index Factor covariates were used in the models. The Optimum Index Factor covariates produced the best classification using random forests. Classification accuracy was 45.7%. The predictive rasters may not be useful for soil map unit delineation, but using a hybrid method to guide further sampling using the pre-map and standard sampling techniques can produce a reasonable soil map.

No description available.

No description available.

Habitat degradation and hunting pose the most proximate threats to many primate species, while climate change is expected to exacerbate these threats (habitat and climate change combined henceforth as "global change") and present new challenges. Madagascar's lemurs are earth's most endangered primates, placing added urgency to their conservation in the face of global change. My dissertation focused on the critically endangered golden-crowned sifaka (Propithecus tattersalli; hereafter, "sifaka") which is endemic to fragmented forests across a gradient of dry, moderate, and wet forest types in northeastern Madagascar. I surveyed sifakas across their global range and investigated factors affecting their densities. I explored sifaka diets across different forest types and evaluated if nutritional factors influenced sifaka densities. Lastly, I investigated sifaka range-wide genetic diversity and conducted a connectivity analysis to prioritize corridor-restoration and other potential conservation efforts. Sifaka densities varied widely across forest fragments (6.8 (SE = 2.0-22.8) to 78.1 (SE = 53.1-114.8) sifakas/km²) and populations have declined by as much as 30-43% in 10 years, from ~18,000 to 10,222-12,631 individuals (95% CI: 8,230-15,966). Tree cutting, normalized difference vegetation index (NDVI) during the wet season, and Simpson's diversity index (1-D) predicted sifaka densities range-wide. Sifakas consumed over 101 plant species and spent 27.1% of their active time feeding on buds, flowers, fruits, seeds, and young and mature leaves. Feeding effort and plant part consumption varied by season, forest type, and sex. Minerals in sifaka food items (Mg (β = 0.62, SE = 0.19) and K (β = 0.58, SE = 0.20)) and wet season NDVI (β = 0.43, SE = 0.20) predicted sifaka densities. Genetic measures across forest fragments indicated that sifaka populations are becoming more isolated (moderate FIS values: mean = 0.27, range = 0.11-0.60; high M-ratios: mean = 0.59, range = 0.49-0.82; low overall effective population size: Ne = 139.8-144 sifakas). FST comparisons between fragments (mean = 0.12, range = 0.01-0.30) supported previous findings that sifakas still moved across the fragmented landscape. Further validation of these genetic results is needed. I identified critical corridors that conservation managers could protect and/or expand via active reforestation to ensure the continued existence of this critically-endangered lemur. / Doctor of Philosophy / Worldwide, many species of primates are threatened with extinction due to habitat degradation, hunting, and climate change (habitat and climate combined threats, henceforth, "global change"). These threats work at different time scales, with hunting being the most immediate and climate change likely to have its fullest impact experienced from the present to a longer time frame. Lemurs are a type of primate found only on Madagascar, an island experiencing rapid global change, which puts lemurs at a heightened risk of extinction. My dissertation research focused on the critically endangered golden-crowned sifaka (Propithecus tattersalli; hereafter, "sifaka"), a species of lemur found only in a few isolated forests across a dry to wet gradient in northeastern Madagascar. To better understand their extinction risk, I conducted surveys to estimate the number of sifakas remaining and investigated several factors that might determine how many sifakas can live in one place. I then explored how sifaka diets varied depending on the forest type that they inhabit and tested whether nutrients in their food might determine sifaka numbers. Lastly, I calculated sifaka genetic diversity to assess their ability to adapt to new environmental conditions and to determine whether sifakas can move across the landscape to find new mates and to potentially colonize new areas of habitat. Sifaka densities varied widely across their range (6.8-78.1 sifakas/km² ). Only 10,222-12,631 sifakas remain, which is 30-43% less than the range of estimates obtained 10 years ago (~18,000 sifakas). Tree cutting, normalized difference vegetation index (NDVI; a measure of plant health or "greenness" obtained from satellite data), and a tree species diversity index were useful measures to predict sifaka densities. Sifakas ate different plant parts (buds, flowers, fruits, seeds, and leaves) from over 101 plant species. The amount of time they spent eating each day varied by the time of year, forest type, and sex. On average, they spent a quarter of their day eating. Magnesium and potassium concentrations in sifaka food items also were useful nutrition-related measures to predict sifaka densities. Genetic analyses suggested that sifaka populations are becoming more isolated and inbred, meaning sifakas are breeding with other sifakas to which they are closely related. However, it appears that sifakas still can move between forest patches to find new mates and to potentially colonize new areas, if such areas are created. Further validation of these genetic results is needed. I also identified critical areas that will be important to protect and reforest to ensure that movements between populations can continue.

The Appalachian National Scenic Trail is a unit of the National Park System created by the National Trails Act of 1968. Commonly referred to as the Appalachian Trail, or the AT, this National Park has some of the longest boundaries of any park. The AT is routed more than 2000 miles along the mountains of the eastern United States. The land purchased for the protection of the AT creates a separate boundary on each side of the trail. Monitoring these boundaries for intrusions or encroachments is a difficult and time-consuming task when done totally by field methods. This thesis presents a more efficient and consistent monitoring process using remote sensing data and change detection algorithms. Using Landsat TM images, Normalized Difference Vegetation Index (NDVI), and image difference change detection, this research shows that major boundary encroachments can be detected. Detection of sub-pixel vegetation index decreases identifies specific locations for field inspection. Assuming low cost multispectral Landsat imagery is available, simple NDVI difference calculation allows this technique to be applied to the entire AT one or more times per year. This procedure would improve the response time for encroachment mediation. The producer's accuracy for finding possible encroachments was 100 percent and the consumer's accuracy for possible encroachments indicated was 78.3 percent. Due to limited image availability, this study only examines change between one pair of Landsat images. Further refinement of these techniques should investigate other Landsat images at other times. Use of other remote sensing systems and change detection algorithms could be the focus of further research. / Master of Science