O alisamento dos resultados numa empresa: o caso do Grupo SONANGOL

Dissertação de Mestrado apresentada ao Instituto de Contabilidade e Administração do Porto para a obtenção do grau de Mestre em Contabilidade e Finanças sob orientação de Professor Doutor Adalmiro Alvaro Malheiro de Castro Andrade Pereira

Unmixing hyperspectral data: independent and dependent component analysis

The development of high spatial resolution airborne and spaceborne sensors has improved the capability of ground-based data collection in the fields of agriculture, geography, geology, mineral identification, detection [2, 3], and classification [4–8]. The signal read by the sensor from a given spatial element of resolution and at a given spectral band is a mixing of components originated by the constituent substances, termed endmembers, located at that element of resolution. This chapter addresses hyperspectral unmixing, which is the decomposition of the pixel spectra into a collection of constituent spectra, or spectral signatures, and their corresponding fractional abundances indicating the proportion of each endmember present in the pixel [9, 10]. Depending on the mixing scales at each pixel, the observed mixture is either linear or nonlinear [11, 12]. The linear mixing model holds when the mixing scale is macroscopic [13]. The nonlinear model holds when the mixing scale is microscopic (i.e., intimate mixtures) [14, 15]. The linear model assumes negligible interaction among distinct endmembers [16, 17]. The nonlinear model assumes that incident solar radiation is scattered by the scene through multiple bounces involving several endmembers [18]. Under the linear mixing model and assuming that the number of endmembers and their spectral signatures are known, hyperspectral unmixing is a linear problem, which can be addressed, for example, under the maximum likelihood setup [19], the constrained least-squares approach [20], the spectral signature matching [21], the spectral angle mapper [22], and the subspace projection methods [20, 23, 24]. Orthogonal subspace projection [23] reduces the data dimensionality, suppresses undesired spectral signatures, and detects the presence of a spectral signature of interest. The basic concept is to project each pixel onto a subspace that is orthogonal to the undesired signatures. As shown in Settle [19], the orthogonal subspace projection technique is equivalent to the maximum likelihood estimator. This projection technique was extended by three unconstrained least-squares approaches [24] (signature space orthogonal projection, oblique subspace projection, target signature space orthogonal projection). Other works using maximum a posteriori probability (MAP) framework [25] and projection pursuit [26, 27] have also been applied to hyperspectral data. In most cases the number of endmembers and their signatures are not known. Independent component analysis (ICA) is an unsupervised source separation process that has been applied with success to blind source separation, to feature extraction, and to unsupervised recognition [28, 29]. ICA consists in finding a linear decomposition of observed data yielding statistically independent components. Given that hyperspectral data are, in given circumstances, linear mixtures, ICA comes to mind as a possible tool to unmix this class of data. In fact, the application of ICA to hyperspectral data has been proposed in reference 30, where endmember signatures are treated as sources and the mixing matrix is composed by the abundance fractions, and in references 9, 25, and 31–38, where sources are the abundance fractions of each endmember. In the first approach, we face two problems: (1) The number of samples are limited to the number of channels and (2) the process of pixel selection, playing the role of mixed sources, is not straightforward. In the second approach, ICA is based on the assumption of mutually independent sources, which is not the case of hyperspectral data, since the sum of the abundance fractions is constant, implying dependence among abundances. This dependence compromises ICA applicability to hyperspectral images. In addition, hyperspectral data are immersed in noise, which degrades the ICA performance. IFA [39] was introduced as a method for recovering independent hidden sources from their observed noisy mixtures. IFA implements two steps. First, source densities and noise covariance are estimated from the observed data by maximum likelihood. Second, sources are reconstructed by an optimal nonlinear estimator. Although IFA is a well-suited technique to unmix independent sources under noisy observations, the dependence among abundance fractions in hyperspectral imagery compromises, as in the ICA case, the IFA performance. Considering the linear mixing model, hyperspectral observations are in a simplex whose vertices correspond to the endmembers. Several approaches [40–43] have exploited this geometric feature of hyperspectral mixtures [42]. Minimum volume transform (MVT) algorithm [43] determines the simplex of minimum volume containing the data. The MVT-type approaches are complex from the computational point of view. Usually, these algorithms first find the convex hull defined by the observed data and then fit a minimum volume simplex to it. Aiming at a lower computational complexity, some algorithms such as the vertex component analysis (VCA) [44], the pixel purity index (PPI) [42], and the N-FINDR [45] still find the minimum volume simplex containing the data cloud, but they assume the presence in the data of at least one pure pixel of each endmember. This is a strong requisite that may not hold in some data sets. In any case, these algorithms find the set of most pure pixels in the data. Hyperspectral sensors collects spatial images over many narrow contiguous bands, yielding large amounts of data. For this reason, very often, the processing of hyperspectral data, included unmixing, is preceded by a dimensionality reduction step to reduce computational complexity and to improve the signal-to-noise ratio (SNR). Principal component analysis (PCA) [46], maximum noise fraction (MNF) [47], and singular value decomposition (SVD) [48] are three well-known projection techniques widely used in remote sensing in general and in unmixing in particular. The newly introduced method [49] exploits the structure of hyperspectral mixtures, namely the fact that spectral vectors are nonnegative. The computational complexity associated with these techniques is an obstacle to real-time implementations. To overcome this problem, band selection [50] and non-statistical [51] algorithms have been introduced. This chapter addresses hyperspectral data source dependence and its impact on ICA and IFA performances. The study consider simulated and real data and is based on mutual information minimization. Hyperspectral observations are described by a generative model. This model takes into account the degradation mechanisms normally found in hyperspectral applications—namely, signature variability [52–54], abundance constraints, topography modulation, and system noise. The computation of mutual information is based on fitting mixtures of Gaussians (MOG) to data. The MOG parameters (number of components, means, covariances, and weights) are inferred using the minimum description length (MDL) based algorithm [55]. We study the behavior of the mutual information as a function of the unmixing matrix. The conclusion is that the unmixing matrix minimizing the mutual information might be very far from the true one. Nevertheless, some abundance fractions might be well separated, mainly in the presence of strong signature variability, a large number of endmembers, and high SNR. We end this chapter by sketching a new methodology to blindly unmix hyperspectral data, where abundance fractions are modeled as a mixture of Dirichlet sources. This model enforces positivity and constant sum sources (full additivity) constraints. The mixing matrix is inferred by an expectation-maximization (EM)-type algorithm. This approach is in the vein of references 39 and 56, replacing independent sources represented by MOG with mixture of Dirichlet sources. Compared with the geometric-based approaches, the advantage of this model is that there is no need to have pure pixels in the observations. The chapter is organized as follows. Section 6.2 presents a spectral radiance model and formulates the spectral unmixing as a linear problem accounting for abundance constraints, signature variability, topography modulation, and system noise. Section 6.3 presents a brief resume of ICA and IFA algorithms. Section 6.4 illustrates the performance of IFA and of some well-known ICA algorithms with experimental data. Section 6.5 studies the ICA and IFA limitations in unmixing hyperspectral data. Section 6.6 presents results of ICA based on real data. Section 6.7 describes the new blind unmixing scheme and some illustrative examples. Section 6.8 concludes with some remarks.

Blind hyperspectral unmixing

Hyperspectral unmixing methods aim at the decomposition of a hyperspectral image into a collection endmember signatures, i.e., the radiance or reflectance of the materials present in the scene, and the correspondent abundance fractions at each pixel in the image. This paper introduces a new unmixing method termed dependent component analysis (DECA). This method is blind and fully automatic and it overcomes the limitations of unmixing methods based on Independent Component Analysis (ICA) and on geometrical based approaches. DECA is based on the linear mixture model, i.e., each pixel is a linear mixture of the endmembers signatures weighted by the correspondent abundance fractions. These abundances are modeled as mixtures of Dirichlet densities, thus enforcing the non-negativity and constant sum constraints, imposed by the acquisition process. The endmembers signatures are inferred by a generalized expectation-maximization (GEM) type algorithm. The paper illustrates the effectiveness of DECA on synthetic and real hyperspectral images.

Comparison of small mammal prevalence of Leishmania (Leishmania) mexicana in five foci of cutaneous leishmaniasis in the State of Campeche, Mexico

In the Yucatan Peninsula of Mexico, 95% of the human cases of Cutaneous Leishmaniasis are caused by Leishmania (Leishmania) mexicana with an incidence rate of 5.08 per 100,000 inhabitants. Transmission is limited to the winter months (November to March). One study on wild rodents has incriminated Ototylomys phyllotis and Peromyscus yucatanicus as primary reservoirs of L. (L.) mexicana in the focus of La Libertad, Campeche. In the present study, the prevalence of both infection and disease caused by L. (L.) mexicana in small terrestrial mammals were documented during five transmission seasons (1994-2004) in five foci of Leishmaniasis in the state of Campeche. Foci separated by only 100 km, with similar relative abundances of small mammals, were found to differ significantly in their prevalence of both symptoms and infection. Transmission rates and reservoir species seemed to change in space as well as in time which limited the implementation of effective control measures of the disease even in a small endemic area such as the south of the Yucatan Peninsula.

Study of nuclear reactions relevant for astrophysics by Micro-AMS

Dissertação para obtenção do Grau de Doutor em Física

PRELIMINARY RESULTS OF A LARGE-SCALE TREE INVENTORY OF UPLAND RAIN FOREST IN THE CENTRAL AMAZON

A large-scale inventory of trees > 10cm DBH was conducted in the upland "terra firme" rain forest of the Distrito Agropecuário da SUFRAMA (Manaus Free Zone Authority Agricultural District) approximately 65Km north of the city of Manaus (AM), Srasil. Thegeneral appearance and structure of the forest is described together with local topography and soil texture. Thepreliminary results of the Inventory provide a minimum estimate of 698 tree species in 53 families in the 40Km radius sampled, including 17 undescribed species. Themost numerically abundant families, Lecythidaceae, Leguminosae, 5apotaceae and Burseraceae as also among the most species rich families. One aspect of this diverse assemblage is the proliferation of species within certain genera, Including 26 genera In 17 families with 6 or more species or morphospecies. Most species have very low abundances of less than 1 tree per hectare. While more abundant species do exist at densities ranging up to a mean of 12 trees per ha, many have clumped distributions leading to great variation in local species abundance. The degree of similarity between hectare samples based int the Coefficient of Community similarity Index varies widely over different sample hectares for five ecologically different families. Soil texture apparently plays a significant role In determining species composition in the different one hectare plots examined while results for other variable were less consistent. Greater differences in similarity indices are found for comparisons with a one hectare sample within the same formation approximately 40Km to the south. It is concluded that homogeneity of tree community composition within this single large and diverse yet continuous upland forest formation can not be assumed.

Ecology and natural history of Akodon lindberghi (Rodentia, Sigmodontinae) in southeastern Brazil

We studied the ecology and natural history of the globally threatened and poorly known Akodon lindberghi Hershkovitz, 1990 in Parque Nacional da Serra da Canastra (PNSC) and Juiz de Fora (JF), southeastern Brazil. From November 1998 to September 2001 a total of 131 individuals were captured in wire-cage live-traps and 52 by pitfalls traps. They were all marked and released at the site. The largest abundances were registered during the dry season, and most of the captures occurred in open habitats. The mean body mass of the two populations was significantly different (18.1 g at PNSC versus 13.1 g at JF; H = 46.2678, g.l.=2, p<0.001). In PNSC, individuals were reproductively active from August to February, and juveniles were present from May to August. The results suggest that the changes in vegetation structure caused by deforestation and intensive agricultural activities could increase the predation rate, affecting the mean body mass of the population.

Assemblages of drosophilids (Diptera, Drosophilidae) in mangrove forests: community ecology and species diversity

Brazilian fauna of drosophilids has been researched in various ecosystems, but those in mangrove forests remain overlooked in Brazil and elsewhere. The present study attempts to characterise the assemblages of drosophilids of this environment, based on 28 collections taken in three mangrove areas in Santa Catarina Island, southern Brazil. The three mangroves surveyed were different in their surroundings, which varied from highly urbanised areas to conservation areas with natural vegetation. Overall, 69 species were collected, and no remarkable difference was detected in species composition and abundances or in the richness, evenness and heterogeneity between sites. The species abundance distribution observed fitted to a theoretical lognormal distribution in the three mangroves. The species richness scored and the performance of the species richness estimators showed an unexpectedly high diversity, considering the very low floristic diversity and the harsh conditions of the environment. Regarding species composition and abundances, the drosophilid mangrove assemblages were shown to be more similar to those found in open environments, with a marked dominance of exotic species. Finally, considering the apparent lack of feeding and breeding sites, we suggest that mangrove forests are acting as sink habitats for the drosophilids populations.

Landscape context on generalist and specialist aphid natural enemies

Report for the scientific sojourn at the University of Lund, Sweden, between May and September 2007. A landscape-scale research approach has been highlighted by a growing body of literature as essential for understanding important ecosystem services as biological control. Aphids are victims of a diversity of enemies making the aphid-enemy interaction a nice example for the role of enemy diversity for the functioning of biological control. Here it is examined the effects of landscape complexity on cereal aphids and associated natural enemies that varied in the degree of specialization. Parasitoids wasps abundance did not differ between landscape types but was strongly negatively related to the percentage of arable land. In contrast, abundances of generalist predators like Coccinellidae were significantly higher in simple landscapes since can benefit from the high availability of a variety of alternative resources within cropping systems. Consequently coccinellidae-to-aphid ratio was significantly higher in fields in homogenous landscapes as compared to fields included in an heterogeneous landscape, suggesting that enemy pressure on cereal aphids increases with landscape simplification. The landscape effect will depend mainly on the degree of specialization of functionally dominant natural enemies, so that the results imply that conservation actions aiming to optimise abundance for one taxonomic group in the agricultural landscape will not automatically increase abundance of other groups. Given that the strength of natural enemy impact on biocontrol depends on landscape features and the role of functionally dominant natural enemies. So, therefore it is essential to focus the future empirical work in examining the schedule of agricultural landscapes that maintain a diversity of generalist and specialist natural enemies.

Migration of sharks into freshwater systems during the Miocene and implications for Alpine paleoelevation

Trace-element and isotopic compositions of fossilized shark teeth sampled from Miocene marine sediments of the north Alpine Molasse Basin, the Vienna Basin, and the Pannonian Basin generally show evidence of formation in a marine environment under conditions geochemically equivalent to the open ocean. In contrast, two of eight shark teeth from the Swiss Upper Marine Molasse locality of La Moliere have extremely low delta O-18 values (10.3% and 11.3%) and low Sr-87/Sr-86 ratios (0.707840 and 0.707812) compared to other teeth from this locality (21.1%,22.4%o and 0.708421-0.708630). The rare earth element (REE) abundances and patterns from La Moliere not only differ between dentine and enameloid of the same tooth, but also between different teeth, supporting variable conditions of diagenesis at this site. However, the REE patterns of enameloid from the ``exotic'' teeth analyzed for O and Sr isotopic compositions are similar to those of teeth that have O and Sr isotopic compositions typical of a marine setting at this site. Collectively, this suggests that the two ``exotic'' teeth were formed while the sharks frequented a freshwater environment with very low O-18-content and Sr isotopic composition controlled by Mesozoic calcareous rocks. This is consistent with a paleogeography of high-elevation (similar to 2300 m) Miocene Alps adjacent to a marginal sea.

Toxic phytoplankton response to warming in two Mediterranean bays of the Ebro Delta

The analysis of the phytoplankton and environmental parameters of the time series in Alfacs and Fangar bays (north western Mediterranean) from 1990 to 2009 shows some trends. There is an increase in the average water column temperature, 0.11, 0.01, 0.80 and 0.23 ºC for spring, summer, fall and winter respectively in Alfacs Bay and 1.76, 0.71, 1.33, 0.89 ºC for spring, summer, fall and winter in Fangar Bay. The trends in phytoplankton populations show a shift in the timing of occurrence of Karlodinium spp. blooms and an increase of the Pseudo-nitzschia spp. abundances. There is a lack of correlation between the average seasonal temperatures and the toxic phytoplankton abundances.

Community ecology of the metazoan parasites of white croaker, Micropogonias furnieri (Osteichthyes: Sciaenidae), from the coastal zone of the State of Rio de Janeiro, Brazil

One hundred specimens of white croakers, Micropogonias furnieri (Desmarest 1823) (Osteichthyes: Sciaenidae) collected from Pedra de Guaratiba (23°01'S, 43°38'W), State of Rio de Janeiro, Brazil, from September 1997 to August 1999, were necropsied to study their parasites. The majority of the fish (95%) were parasitized by metazoan. Twenty-eight species of parasites were collected. The nematodes were the 40.5% of the total number of parasites specimens collected. Dichelyne elongatus was the most dominant species. Lobatostoma ringens, Pterinotrematoides mexicanum, Corynosoma australe, D. elongatus, and Caligus haemulonis showed a positive correlation between the host's total length and parasite prevalence and abundance. The monogenean P. mexicanum had differences in the prevalence and abundance in relation to sex of the host. The mean diversity in the infracommunities of M. furnieri was H=0.499±0.411, with correlation with the host's total length and without differences in relation to sex of the host. One pair of ectoparasites showed positive covariation, and two pairs of endoparasites showed positive association and covariation between their prevalences and abundances, respectively. Negative association or covariations were not found. The dominance of endoparasites in the croakers parasite infracommunities reinforced the differences found in sciaenids from the South American Pacific Ocean, in which the ectoparasites are dominant.

Community ecology of the metazoan parasites of pink cusk-eel, Genypterus brasiliensis (Osteichthyes: Ophidiidae), from the coastal zone of the State of Rio de Janeiro, Brazil

Fifty-five specimens of pink cusk-eel, Genypterus brasiliensis Regan, 1903 (Osteichthyes: Ophidiidae) collected from the coastal zone of the State of Rio de Janeiro, Brazil (approx. 21-23°S, 41-45°W), from September 2000 to January 2001, were necropsied to study their parasites. All fish were parasitized by one or more metazoan. Fourteen species of parasites were collected. G. brasiliensis is a new host record for nine parasite species. The larval stages of cestodes and the nematodes were the majority of the parasite specimens collected, with 38.4% and 36.5%, respectively. Cucullanus genypteri was the dominant species with highest prevalence and/or abundance. The parasites of G. brasiliensis showed the typical overdispersed pattern of distribution. Six parasite species showed correlation between the host's total body length and prevalence and abundance. Host sex did not influence prevalence and parasite abundance of any parasite species. The mean diversity in the infracommunities of G. brasiliensis was H= 0.364 ± 0.103, with correlation with the host's total length and without differences in relation to sex of the host. One pair of adult endoparasites (C. genypteri and A. brasiliensis) showed positive covariations between their abundances. Negative association or covariation was not found. Differences between the qualitative and quantitative aspects of the parasite community of G. brasiliensis from Rio de Janeiro and Argentina suggest the existence of two population stocks of pink cusk-eel in the South America Atlantic Ocean.

Population dynamics of intermediate snail hosts of Fasciola hepatica and some environmental factors in San Juan y Martinez municipality, Cuba

The variation of abundances of intermediate snail hosts of Fasciola hepatica in Cuba (Fossaria cubensis and Pseudosuccinea columella) was studied during one year under natural conditions at five sampling sites in San Juan y Martinez municipality, Pinar del Rio province, Cuba. The effect of some environmental variables on the lymnaeid abundances was also studied. A canonical correspondence analysis showed that both species do not generally occur together in the same habitat and that most factors affect them in an opposite fashion, although both of them correlate positively through time to the diversity of the habitats. F. cubensis prefers the sites that are in or closer to the city whereas P. columella is more abundant in rural sites. Lymnaeid abundances are mainly affected by nitrite and nitrate concentrations as well as by the abundance of the thiarid Tarebia granifera. F. cubensis is more abundant in polluted habitats with low densities (or absence) of T. granifera whereas P. columella prefers cleaner habitats and can coexist with the thiarid, even at its higher densities. The implications of divergent preferences of the two lymnaeids for the control of fasciolosis are discussed.

Enriched basaltic andesites from mid-crustal fractional crystallization, recharge, and assimilation (Pilavo Volcano, Western Cordillera of Ecuador)

The origin of andesite is an important issue in petrology because andesite is the main eruptive product at convergent margins, corresponds to the average crustal composition and is often associated with major Cu-Au mineralization. In this study we present petrographic, mineralogical, geochemical and isotopic data for basaltic andesites of the latest Pleistocene Pilavo volcano, one of the most frontal volcanoes of the Ecuadorian Quaternary arc, situated upon thick (30-50 km) mafic crust composed of accreted Cretaceous oceanic plateau rocks and overlying mafic to intermediate Late Cretaceous-Late Tertiary magmatic arcs. The Pilavo rocks are basaltic andesites (54-57 center dot 5 wt % SiO(2)) with a tholeiitic affinity as opposed to the typical calc-alkaline high-silica andesites and dacites (SiO(2) 59-66 wt %) of other frontal arc volcanoes of Ecuador (e.g. Pichincha, Pululahua). They have much higher incompatible element contents (e.g. Sr 650-1350 ppm, Ba 650-1800 ppm, Zr 100-225 ppm, Th 5-25 ppm, La 15-65 ppm) and Th/La ratios (0 center dot 28-0 center dot 36) than Pichincha and Pululahua, and more primitive Sr ((87)Sr/(86)Sr similar to 0 center dot 7038-0 center dot 7039) and Nd (epsilon(Nd) similar to +5 center dot 5 to +6 center dot 1) isotopic signatures. Pilavo andesites have geochemical affinities with modern and recent high-MgO andesites (e.g. low-silica adakites, Setouchi sanukites) and, especially, with Archean sanukitoids, for both of which incompatible element enrichments are believed to result from interactions of slab melts with peridotitic mantle. Petrographic, mineral chemistry, bulk-rock geochemical and isotopic data indicate that the Pilavo magmatic rocks have evolved through three main stages: (1) generation of a basaltic magma in the mantle wedge region by flux melting induced by slab-derived fluids (aqueous, supercritical or melts); (2) high-pressure differentiation of the basaltic melt (at the mantle-crust boundary or at lower crustal levels) through sustained fractionation of olivine and clinopyroxene, leading to hydrous, high-alumina basaltic andesite melts with a tholeiitic affinity, enriched in incompatible elements and strongly impoverished in Ni and Cr; (3) establishment of one or more mid-crustal magma storage reservoirs in which the magmas evolved through dominant amphibole and clinopyroxene (but no plagioclase) fractionation accompanied by assimilation of the modified plutonic roots of the arc and recharge by incoming batches of more primitive magma from depth. The latter process has resulted in strongly increasing incompatible element concentrations in the Pilavo basaltic andesites, coupled with slightly increasing crustal isotopic signatures and a shift towards a more calc-alkaline affinity. Our data show that, although ultimately originating from the slab, incompatible element abundances in arc andesites with primitive isotopic signatures can be significantly enhanced by intra-crustal processes within a thick juvenile mafic crust, thus providing an additional process for the generation of enriched andesites.