Evaluating effects of spectral training data distribution on continuous field mapping performance

Continuous field mapping has to address two conflicting remote sensing requirements when collecting training data. On one hand, continuous field mapping trains fractional land cover and thus favours mixed training pixels. On the other hand, the spectral signature has to be preferably distinct and thus favours pure training pixels. The aim of this study was to evaluate the sensitivity of training data distribution along fractional and spectral gradients on the resulting mapping performance. We derived four continuous fields (tree, shrubherb, bare, water) from aerial photographs as response variables and processed corresponding spectral signatures from multitemporal Landsat 5 TM data as explanatory variables. Subsequent controlled experiments along fractional cover gradients were then based on generalised linear models. Resulting fractional and spectral distribution differed between single continuous fields, but could be satisfactorily trained and mapped. Pixels with fractional or without respective cover were much more critical than pure full cover pixels. Error distribution of continuous field models was non-uniform with respect to horizontal and vertical spatial distribution of target fields. We conclude that a sampling for continuous field training data should be based on extent and densities in the fractional and spectral, rather than the real spatial space. Consequently, adequate training plots are most probably not systematically distributed in the real spatial space, but cover the gradient and covariate structure of the fractional and spectral space well. (C) 2009 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS). Published by Elsevier B.V. All rights reserved.

How much protein and energy are needed to equilibrate nitrogen and energy balances in ventilated critically ill children?

BACKGROUND & AIMS: Protein and energy requirements in critically ill children are currently based on insufficient data. Moreover, longitudinal measurements of both total urinary nitrogen (TUN) and resting energy expenditure (REE) are lacking. The aim of this study was to investigate how much protein and energy are needed to equilibrate nitrogen and energy balances in ventilated critically ill children on the basis of daily measurements of TUN, REE and protein and energy intakes. Comparisons were made with the guidelines of the American Society for Parenteral and Enteral Nutrition and the Dietary Reference Intakes. METHODS: Children with an expected duration of mechanical ventilation ≥72 h were prospectively recruited. TUN was measured by chemiluminescence, and REE was measured by indirect calorimetry. Generalised linear models for longitudinal data were used to study the relation between protein intake and nitrogen balance and to calculate the minimum intake of protein needed to achieve nitrogen equilibrium. A similar approach was used for energy. Results were compared to the recommended values. RESULTS: Based on 402 measurements performed in 74 children (median age: 21 months), the mean TUN was high at 0.20 (95% CI: 0.20, 0.22) g/kg/d and the REE was 55 (95% CI: 54, 57) kcal/kg/d. Nitrogen and energy balances were achieved with 1.5 (95% CI: 1.4, 1.6) g/kg/d of protein and 58 (95% CI: 53, 63) kcal/kg/d for the entire group, but there were differences among children of different ages. Children required more protein and less energy than the Dietary Reference Intakes. CONCLUSIONS: In critically ill children, TUN was elevated and REE was reduced during the entire period of mechanical ventilation. Minimum intakes of 1.5 g/kg/d of protein and 58 kcal/kg/d can equilibrate nitrogen and energy balances in children up to 4 years old. Older children require more protein.

Vigor and field emergence in "high oleic" sunflower seeds, in southeast of Buenos Aires province

Sunflower crop was based, as yet, on high linoleic cultivars, but in the last years request for oil with higher content of oleic acid has increased, due to their dietary characteristics. At the beginning, high oleic cultivars were used to be sown in warm regions, but then the concern about growing it in temperate areas, as the south-east of Buenos Aires Province, was posed. In this region, early sowings are recommended, so that grain filling matches with a period of appropriate hydric and light conditions, as to result in greater yields. However, early sowings are limited by low soil temperature, that delays seedling emergence, resulting in heterogeneous stand establishment. The aim of this work was to evaluate seed performance of four high oleic cultivars in the southern area of Buenos Aires Province, by means of vigor tests and field trials. Germination, cold, tetrazolium viability, tetrazolium viability with cold, accelerated ageing tests and three field sowings at different soil temperatures were performed. Data were analyzed by Anova using generalised linear models, and tests and cultivars were contrasted among themselves. Similar seedling emergence under optimal and suboptimal temperatures for high oleic and high linoleic cultivars was recorded. The success of seedling establishment does not appear to be related to the acidic composition of seeds.

There is no difference in nitric oxide metabolites and neonatal outcome between premature infants born to pre-eclamptic and those born to normotensive women

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Ground beetle (Carabidae) abundance data from north-eastern Australian rainforest, between June 2008 - January 2010, using pitfall traps

Understanding how the environment influences patterns of diversity is vital for effective conservation management, especially in a changing global climate. While assemblage structure and species richness patterns are often correlated with current environmental factors, historical influences may also be considerable, especially for taxa with poor dispersal abilities. Mountain-top regions throughout tropical rainforests can act as important refugia for taxa characterised by low dispersal capacities such as flightless ground beetles (Carabidae), an ecologically significant predatory group. We surveyed flightless ground beetles along elevational gradients in five different subregions within the Australian Wet Tropics World Heritage Area to investigate (1) whether the diversity and composition of flightless ground beetles are elevationally stratified, and, if so, (2) what environmental factors (other than elevation per se) are associated with these patterns. Generalised linear models and model averaging techniques were used to relate patterns of diversity to environmental factors. Unlike most taxonomic groups, flightless ground beetles increased in species richness and abundance with elevation. Additionally, each subregion consisted of distinct assemblages containing a high level of regional endemic species. Species richness was most strongly positively associated with the historical climatic conditions and negatively associated with severity of recent disturbance (treefalls) and current climatic conditions. Assemblage composition was associated with latitude and current and historical climatic conditions. Our results suggest that distributional patterns of flightless ground beetles are not only likely to be associated with factors that change with elevation (current climatic conditions), but also factors that are independent of elevation (recent disturbance and historical climatic conditions). Variation in historical vegetation stability explained both species richness and assemblage composition patterns, probably reflecting the significance of upland refugia at a geographic time scale. These findings are important for conservation management as upland habitats are under threat from climate change.

Efficient training of RBF networks for classification

Radial Basis Function networks with linear outputs are often used in regression problems because they can be substantially faster to train than Multi-layer Perceptrons. For classification problems, the use of linear outputs is less appropriate as the outputs are not guaranteed to represent probabilities. In this paper we show how RBFs with logistic and softmax outputs can be trained efficiently using algorithms derived from Generalised Linear Models. This approach is compared with standard non-linear optimisation algorithms on a number of datasets.

DVMS 1.5: A user manual (the data visualization and modeling system)

The data available during the drug discovery process is vast in amount and diverse in nature. To gain useful information from such data, an effective visualisation tool is required. To provide better visualisation facilities to the domain experts (screening scientist, biologist, chemist, etc.),we developed a software which is based on recently developed principled visualisation algorithms such as Generative Topographic Mapping (GTM) and Hierarchical Generative Topographic Mapping (HGTM). The software also supports conventional visualisation techniques such as Principal Component Analysis, NeuroScale, PhiVis, and Locally Linear Embedding (LLE). The software also provides global and local regression facilities . It supports regression algorithms such as Multilayer Perceptron (MLP), Radial Basis Functions network (RBF), Generalised Linear Models (GLM), Mixture of Experts (MoE), and newly developed Guided Mixture of Experts (GME). This user manual gives an overview of the purpose of the software tool, highlights some of the issues to be taken care while creating a new model, and provides information about how to install & use the tool. The user manual does not require the readers to have familiarity with the algorithms it implements. Basic computing skills are enough to operate the software.

Prevalence and associations of Trypanosoma spp. and Sodalis glossinidius with intrinsic factors of tsetse flies

Trypanosomiasis has been identified as a neglected tropical disease in both humans and animals in many regions of sub-Saharan Africa. Whilst assessments of the biology of trypanosomes, vectors, vertebrate hosts and the environment have provided useful information about life cycles, transmission, and pathogenesis of the parasites that could be used for treatment and control, less information is available about the effects of interactions among multiple intrinsic factors on trypanosome presence in tsetse flies from different sites. It is known that multiple species of tsetse flies can transmit trypanosomes but differences in their vector competence has normally been studied in relation to individual factors in isolation, such as: intrinsic factors of the flies (e.g. age, sex); habitat characteristics; presence of endosymbionts (e.g. Wigglesworthia glossinidia, Sodalis glossinidius); feeding pattern; host communities that the flies feed on; and which species of trypanosomes are transmitted. The purpose of this study was to take a more integrated approach to investigate trypanosome prevalence in tsetse flies. In chapter 2, techniques were optimised for using the Polymerase Chain Reaction (PCR) to identify species of trypanosomes (Trypanosoma vivax, T. congolense, T. brucei, T. simiae, and T. godfreyi) present in four species of tsetse flies (Glossina austeni, G. brevipalpis, G. longipennis and G. pallidipes) from two regions of eastern Kenya (the Shimba Hills and Nguruman). Based on universal primers targeting the internal transcribed spacer 1 region (ITS-1), T. vivax was the predominant pathogenic species detected in flies, both singly and in combination with other species of trypanosomes. Using Generalised Linear Models (GLMs) and likelihood ratio tests to choose the best-fitting models, presence of T. vivax was significantly associated with an interaction between subpopulation (a combination between collection sites and species of Glossina) and sex of the flies (X2 = 7.52, df = 21, P-value = 0.0061); prevalence in females overall was higher than in males but this was not consistent across subpopulations. Similarly, T. congolense was significantly associated only with subpopulation (X2 = 18.77, df = 1, P-value = 0.0046); prevalence was higher overall in the Shimba Hills than in Nguruman but this pattern varied by species of tsetse fly. When associations were analysed in individual species of tsetse flies, there were no consistent associations between trypanosome prevalence and any single factor (site, sex, age) and different combinations of interactions were found to be significant for each. The results thus demonstrated complex interactions between vectors and trypanosome prevalence related to both the distribution and intrinsic factors of tsetse flies. The potential influence of the presence of S. glossinidius on trypanosome presence in tsetse flies was studied in chapter 3. A high number of Sodalis positive flies was found in the Shimba Hills, while there were only two positive flies from Nguruman. Presence or absence of Sodalis was significantly associated with subpopulation while trypanosome presence showed a significant association with age (X2 = 4.65, df = 14, P-value = 0.0310) and an interaction between subpopulation and sex (X2 = 18.94, df = 10, P-value = 0.0043). However, the specific associations that were significant varied across species of trypanosomes, with T. congolense and T. brucei but not T. vivax showing significant interactions involving Sodalis. Although it has previously been concluded that presence of Sodalis increases susceptibility to trypanosomes, the results presented here suggest a more complicated relationship, which may be biased by differences in the distribution and intrinsic factors of tsetse flies, as well as which trypanosome species are considered. In chapter 4 trypanosome status was studied in relation to blood meal sources, feeding status and feeding patterns of G. pallidipes (which was the predominant fly species collected for this study) as determined by sequencing the mitochondrial cytochrome B gene using DNA extracted from abdomen samples. African buffalo and African elephants were the main sources of blood meals but antelopes, warthogs, humans, giraffes and hyenas were also identified. Feeding on multiple hosts was common in flies sampled from the Shimba Hills but most flies from Nguruman had fed on single host species. Based on Multiple Correspondence Analysis (MCA), host-feeding patterns showed a correlation with site of sample collection and Sodalis status, while trypanosome status was correlated with sex and age of the flies, suggesting that recent host-feeding patterns from blood meal analysis cannot predict trypanosome status. In conclusion, the complexity of interactions found suggests that strategies of tsetse fly control should be specific to particular epidemic areas. Future studies should include laboratory experiments that use local colonies of tsetse flies, local strains of trypanosomes and local S. glossinidius under controlled environmental conditions to tease out the factors that affect vector competence and the relative influence of external environmental factors on the dynamics of these interactions.

As rodovias e as aves de rapina nocturnas relação entre a densidade, o habitat e a mortalidade

As estradas e tráfego inerente surgem como a criação antrópica mais conspícua e penetrante na paisagem natural, sendo considerados os principais agentes causadores de fragmentação e destruição de habitats, assim como representam um obstáculo físico sem precedentes, limitando as relações directas entre os indivíduos, por diminuição da frequência de dispersão e aumento da mortalidade por atropelamento, impedindo o fluxo natural de genes e suscitando o aumento de fenómenos de inbreeding e perda de heterozigotia. Todos os impactes deletérios associados às rodovias são claramente perceptíveis em vertebrados, onde as aves de rapinas nocturnas não são excepção. Uma vez que estas rapaces beneficiam das suas bermas e orlas, como locais de poiso, nidificação ou como corredores de dispersão através da paisagem, são frequentemente vítimas de mortalidade por atropelamento em estradas, sendo esta problemática considerada actualmente uma das mais recentes e importantes formas de mortalidade não natural em rapinas nocturnas e vinculada como um dos maiores problemas de conservação que afecta este grupo. Não obstante, esse mútuo efeito de atracção/repulsa das estradas a estas rapaces, as rodovias criam uma barreira específica que limita a dinâmica, comportamento e densidade populacional das espécies residentes, reconhecendo-se que o isolamento daí resultante, pode comprometer a viabilidade populacional a longo prazo, podendo mesmo conduzir a altos riscos de extinção das populações locais devido a efeitos estocásticos. Mediante esta problemática, este trabalho debruçou-se sobre um único objectivo principal: a avaliação do impacte das rodovias e do tráfego, na densidade das aves de rapina nocturnas. Este estudo foi efectuado na região Alentejana, abrangendo uma área de cariz tipicamente mediterrânico, delimitada pelas localidades de Montemor-o-Novo, Arraiolos e Évora, sendo seccionada por 143 quilómetros de estradas, divididas em autoestrada, rodovias com elevada e reduzida densidade de tráfego. A monitorização das rapinas nocturnas foi conduzida em dois anos amostrais (2005 e 2007), tendo sido focalizada sobretudo em duas espécies de Strigiformes, a Coruja­ do-mato Strix aluco e o Mocho-galego Athene noctua, recorrendo ao uso de playbacks com reprodução de vocalizações de indivíduos conspecíficos. Foram usadas 32 variáveis explicativas integradas em três grandes grupos: variáveis de estrada, métricas da paisagem, uso do solo, tendo sido analiticamente testadas, recorrendo à aplicação de Modelos Lineares Generalizados. Os principais resultados obtidos demonstram que as variáveis de estrada, aliadas à densidade de tráfego e ruído inerente à sua circulação, são provavelmente, responsáveis por um comportamento de repulsa das espécies de aves de rapina nocturnas em estudo, apresentando estas densidades mais elevadas longe de áreas antropicamente perturbadas e, portanto, de menor qualidade que se encontram adjacentes às rodovias. Todavia a presença de habitat favorável a estas rapaces é provavelmente o descritor com maior poder estatístico no que concerne à sua distribuição e densidade, sendo os montados densos e a presença de zonas agrícolas de sequeiro, positivamente correlacionadas, respectivamente com a densidade de casais reprodutores de Coruja-do-mato e Mocho-galego. Mediante a observação dos resultados será veemente a aplicação de medidas de mitigação específicas, que fundamentalmente considerem o afastamento dos efectivos populacionais longe das estradas e tráfego, conservando e assegurando as características estruturais, requisitos e qualidade dos habitats, de modo a incrementar e garantir a viabilidade e densidade das populações, fidelizando a territorialidade e permanência destas aves nestas áreas. Adicionalmente deverá investir-se na conectividade das manchas de paisagem fragmentada pelas rodovias, criando opções de conservação estratégicas, em zonas ecologicamente mais sensíveis, que não somente minimizem o efeito de repulsa reconhecido nestas aves, mas também os níveis de mortalidade por atropelamento, tornando os ecossistemas mais funcionais para a sobrevivência destes rapaces. ABSTRACT; Roads and traffic are the most conspicuous and pervasive human creation, being the great responsible for fragmentation and habitats destruction, reducing animal movement through landscape, which implies decrease of gene flow and loss of variability that can fragmented populations, thus reducing their sizes and densities. All deleterious impacts associated with roads are clearly visible in vertebrates, where owls aren't exception, being frequent victims of road mortality, since they can use roadside habitats and edges for hunting, nesting or dispersal corridors through the landscape, being nowadays one of the most recent and important causes of non­natural mortality in owls and has been recognized as one of the largest conservation problems affecting this group. However, the attractive and avoidance effect of roads and his edges on owls creates a barrier effect that limits dynamics, behaviour and breeding density of resident species, recognizing that possible isolation, could compromise populations survival, make them more vulnerable to high risks of local extinction due to stochastic effects. Despite that, several authors suggest that owls use roads to hunt, as marginal habitats, or for navigation corridors through the landscape ln this context, the major aims of this study was to verify if there are negative effects of roads on the density of owls, considering traffic as an influencing factor. This study was conducted in Alentejo, covering a typical Mediterranean area, including three main localities Montemor-o-Novo, Évora and Arraiolos, being sectioned by 143 km of roads, including highway, roads with high traffic density, and the remaining with low traffic density. The owl census was conducted in two sampling years (2005 and 2007) and was focused mainly on Little owl Athene noctua and Tawny Owl Strix aluco species, using the playback technique, with conspecific calls. We used 32 explanatory variables, mainly included in three groups: road variables, landscape metrics and land use, having been analytically tested, with application of Generalised Linear Models. The main results show that noisy roads with high traffic density are probably the most responsible for the avoidance behaviour of owls, under the study area, showing density depression near high anthropogenic disturbed areas adjacent to roads. However, the presence of habitat quality to these birds is probably the descriptor with greater statistical power, considering its distribution and density, with the dense oak woodland and croplands and arable lands, positively correlated, respectively with Tawny owl and Little owl density. ln consequence, the great conservation effort should be done in order to keep breeding populations away from roads and traffic, ensuring the structural features, requirements and quality of its habitats in order to enhance and ensure the viability and density of owl's populations in these areas. ln addition, it is important to invest in connectivity between roadside fragmented patches, creating strategic conservation options, in sensitive areas, which minimize the avoidance effect, recognized in owls, but also road-kill levels, making ecosystems more functional to survival of these top predators.

Analysis of the claims data of a life insurance portfolio

Mestrado em Ciências Actuariais

Geographical and environmental correlates of big and small game in Andalusia (southern Spain)

In Andalusia, southern Spain, each game estate applies its own rules and presents its results in annual hunting reports, which have been mandatory for Spanish game estates since 1989. We used the information about hunting yields, included in 32134 annual hunting reports produced during the period 1993/94 to 2001/02 by 6049 game estates, to determine the current distribution of hunting yields of big and small game species in Andalusia. Using generalised linear models and a geographic information system, we determined the most favourable municipalities to big and small game, respectively, and delimited potential areas to attain good hunting yields for big and small game at a 1-km2 resolution. Municipalities and areas favourable to big game are mainly located in the Sierra Morena and the westernmost fringe of the Betic Range, while those favourable to small game occupy the upper Guadalquivir River valley. There is a clear segregation between big and small game species according to the physiography and land uses of the territory. Big game species are typical of Mediterranean woodland areas, while the most emblematic small game species prefer agricultural areas. Our results provide a territorial ordination of hunting yields in southern Spain and have several potential applications in strategic planning for hunting activities and biodiversity conservation in Andalusia that can be extrapolated to other regions.

Non-linear time series models

The last three decades have seen quite dramatic changes the way we modeled time dependent data. Linear processes have been in the center stage in modeling time series. As far as the second order properties are concerned, the theory and the methodology are very adequate.However, there are more and more evidences that linear models are not sufficiently flexible and rich enough for modeling purposes and that failure to account for non-linearities can be very misleading and have undesired consequences.

Binary classification of dyslipidemia from the waist-to-hip ratio and body mass index: a comparison of linear, logistic, and CART models.

BACKGROUND: We sought to improve upon previously published statistical modeling strategies for binary classification of dyslipidemia for general population screening purposes based on the waist-to-hip circumference ratio and body mass index anthropometric measurements. METHODS: Study subjects were participants in WHO-MONICA population-based surveys conducted in two Swiss regions. Outcome variables were based on the total serum cholesterol to high density lipoprotein cholesterol ratio. The other potential predictor variables were gender, age, current cigarette smoking, and hypertension. The models investigated were: (i) linear regression; (ii) logistic classification; (iii) regression trees; (iv) classification trees (iii and iv are collectively known as "CART"). Binary classification performance of the region-specific models was externally validated by classifying the subjects from the other region. RESULTS: Waist-to-hip circumference ratio and body mass index remained modest predictors of dyslipidemia. Correct classification rates for all models were 60-80%, with marked gender differences. Gender-specific models provided only small gains in classification. The external validations provided assurance about the stability of the models. CONCLUSIONS: There were no striking differences between either the algebraic (i, ii) vs. non-algebraic (iii, iv), or the regression (i, iii) vs. classification (ii, iv) modeling approaches. Anticipated advantages of the CART vs. simple additive linear and logistic models were less than expected in this particular application with a relatively small set of predictor variables. CART models may be more useful when considering main effects and interactions between larger sets of predictor variables.

Generalized linear and generalized additive models in studies of species distributions: setting the scene

An important statistical development of the last 30 years has been the advance in regression analysis provided by generalized linear models (GLMs) and generalized additive models (GAMs). Here we introduce a series of papers prepared within the framework of an international workshop entitled: Advances in GLMs/GAMs modeling: from species distribution to environmental management, held in Riederalp, Switzerland, 6-11 August 2001.We first discuss some general uses of statistical models in ecology, as well as provide a short review of several key examples of the use of GLMs and GAMs in ecological modeling efforts. We next present an overview of GLMs and GAMs, and discuss some of their related statistics used for predictor selection, model diagnostics, and evaluation. Included is a discussion of several new approaches applicable to GLMs and GAMs, such as ridge regression, an alternative to stepwise selection of predictors, and methods for the identification of interactions by a combined use of regression trees and several other approaches. We close with an overview of the papers and how we feel they advance our understanding of their application to ecological modeling.

Effects of alternative sets of climatic predictors on species distribution models and associated estimates of extinction risk: a test with plants in an arid environment

Aim To evaluate the effects of using distinct alternative sets of climatic predictor variables on the performance, spatial predictions and future projections of species distribution models (SDMs) for rare plants in an arid environment. . Location Atacama and Peruvian Deserts, South America (18º30'S - 31º30'S, 0 - 3 000 m) Methods We modelled the present and future potential distributions of 13 species of Heliotropium sect. Cochranea, a plant group with a centre of diversity in the Atacama Desert. We developed and applied a sequential procedure, starting from climate monthly variables, to derive six alternative sets of climatic predictor variables. We used them to fit models with eight modelling techniques within an ensemble forecasting framework, and derived climate change projections for each of them. We evaluated the effects of using these alternative sets of predictor variables on performance, spatial predictions and projections of SDMs using Generalised Linear Mixed Models (GLMM). Results The use of distinct sets of climatic predictor variables did not have a significant effect on overall metrics of model performance, but had significant effects on present and future spatial predictions. Main conclusion Using different sets of climatic predictors can yield the same model fits but different spatial predictions of current and future species distributions. This represents a new form of uncertainty in model-based estimates of extinction risk that may need to be better acknowledged and quantified in future SDM studies.