Rift to post-rift evolution of a passive continental margin: the Ponta Grossa Arch, SE Brazil

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

Neoproterozoic tectonic and magmatic episodes in the NW sector of Borborema Province, NE Brazil, during assembly of Western Gondwana

The Medio Coreau domain of NE Brazil is located along the northwest margin of Borborema Province, the western branch of a Brasiliano/Pan-African collisional belt that formed during the assembly of Western Gondwana. The early Paleoproterozoic basement of the Medio Coreau domain is composed of migmatitic gneisses and juvenile granulites, overlain by late Paleoproterozoic and Neoproterozoic rocks intruded by syn- to post-tectonic Brasiliano granitoids. According to integrated structural and geochronological data (U-Pb zircon and monazite ages), the Neoproterozoic tectonic evolution of the Medio Coreau is characterized by low-angle thrusting and transcurrent deformation. U-Pb geochronological data from plutons intruded during this compressional regime indicate the collisional evolution began at approximately 622 Ma and continued until about 591 Ma. The continuation of convergence until approximately 560 Ma resulted in the formation of NE-SW and E-W shear zones within the Borborema Province and adjoining West African provinces. The final stage of the ductile tectonism was characterized by uplift and high-angle fault generation between approximately 560 and 545 Ma. The last tectonic event was an extensional phase, resulting in the formation of the Jaibaras graben and intrusion of post-orogenic granites at around 532 Ma. (c) 2007 Elsevier Ltd. All rights reserved.

Evidence for Neoproterozoic continental arc magmatism in the Santa Quitéria Batholith of Ceará State, NW Borborema Province, NE Brazil: Implications for the assembly of West Gondwana

Recent field investigations and geochronological studies of Neoproterozoic rocks in the northwestern part of the Borborema Province, Ceará State, NE Brazil provide important clues pertaining to the nature of convergence between the Borborema Province and the West African-São Luis craton during the assembly of West Gondwana. U-Pb zircon data indicate that the earliest evidence of convergent magmatism along the northwest margin of the Borborema Province occurred around 777 Ma, and was followed by the development of a large continental arc batholith (Santa Quitéria batholith) between ca. 665 and 591 Ma within the central part of Ceará State. These findings, along with supporting geophysical data, suggest that convergence between the Borborema Province and the West African-São Luis craton involved closure of an oceanic realm with subduction polarity to the southeast beneath the northwestern part of the province. Consequently, it seems likely that the Pharusian Ocean was continuous from the Hoggar Province in West Africa into South America during the late Neoproterozoic and additional data suggests that it may have even been connected with the Goianides Ocean of the Brasília Belt farther to the southwest.

U-Pb geochronology of the southern Brasília belt (SE-Brazil): Sedimentary provenance, Neoproterozoic orogeny and assembly of West Gondwana

The Brasília belt borders the western margin of the São Francisco Craton and records the history of ocean opening and closing related to the formation of West Gondwana. This study reports new U-Pb data from the southern sector of the belt in order to provide temporal limits for the deposition and ages of provenance of sediments accumulated in passive margin successions around the south and southwestern margins of the São Francisco Craton, and date the orogenic events leading to the amalgamation of West Gondwana. Ages of detrital zircons (by ID-TIMS and LA-MC-ICPMS) were obtained from metasedimentary units of the passive margin of the São Francisco Craton from the main tectonic domains of the belt: the internal allochthons (Araxá Group in the Áraxá and Passos Nappes), the external allochthons (Canastra Group, Serra da Boa Esperança Metasedimentary Sequence and Andrelândia Group) and the autochthonous or Cratonic Domain (Andrelândia Group). The patterns of provenance ages for these units are uniform and are characterised as follows: Archean- Paleoproterozoic ages (3.4-3.3, 3.1-2.7, and 2.5-2.4Ga); Paleoproterozoic ages attributed to the Transamazonian event (2.3-1.9Ga, with a peak at ca. 2.15Ga) and to the ca. 1.75Ga Espinhaço rifting of the São Francisco Craton; ages between 1.6 and 1.2Ga, with a peak at 1.3Ga, revealing an unexpected variety of Mesoproterozoic sources, still undetected in the São Francisco Craton; and ages between 0.9 and 1.0Ga related to the rifting event that led to the individualisation of the São Francisco paleo-continent and formation of its passive margins. An amphibolite intercalation in the Araxá Group yields a rutile age of ca. 0.9Ga and documents the occurrence of mafic magmatism coeval with sedimentation in the marginal basin. Detrital zircons from the autochthonous and parautochthonous Andrelândia Group, deposited on the southern margin of the São Francisco Craton, yielded a provenance pattern similar to that of the allochthonous units. This result implies that 1.6-1.2Ga source rocks must be present in the São Francisco Craton. They could be located either in the cratonic area, which is mostly covered by the Neoproterozoic epicontinental deposits of the Bambuí Group, or in the outer paleo-continental margin, buried under the allochthonous units of the Brasília belt. Crustal melting and generation of syntectonic crustal granites and migmatisation at ca. 630Ma mark the orogenic event that started with westward subduction of the São Francisco plate and ended with continental collision against the Paraná block (and Goiás terrane). Continuing collision led to the exhumation and cooling of the Araxá and Passos metamorphic nappes, as indicated by monazite ages of ca. 605Ma and mark the final stages of tectonometamorphic activity in the southern Brasília belt. Whilst continent-continent collision was proceeding on the western margin of the São Francisco Craton along the southern Brasília belt, eastward subduction in the East was generating the 634-599Ma Rio Negro magmatic arc which collided with the eastern São Francisco margin at 595-560Ma, much later than in the Brasília belt. Thus, the tectonic effects of the Ribeira belt reached the southernmost sector of the Brasília belt creating a zone of superposition. The thermal front of this event affected the proximal Andrelândia Group at ca. 588Ma, as indicated by monazite age. The participation of the Amazonian craton in the assembly of western Gondwana occurred at 545-500Ma in the Paraguay belt and ca. 500Ma in the Araguaia belt. This, together with the results presented in this work lead to the conclusion that the collision between the Paraná block and Goiás terrane with the São Francisco Craton along the Brasília belt preceded the accretion of the Amazonian craton by 50-100 million years. © 2003 Elsevier B.V. All rights reserved.

Long-term landscape evolution and post-rift reactivation in the southeastern Brazilian passive continental margin: Taubat, basin

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

Lower Cretaceous radiolarians from the northwestern Australian margin

During Leg 123, abundant and well-preserved Neocomian radiolarians were recovered at Site 765 (Argo Abyssal Plain) and Site 766 (lower Exmouth Plateau). The assemblages are characterized by a scarcity or absence of Tethyan taxa. The Berriasian-early Aptian radiolarian record recovered at Site 765 is unique in its density of well-preserved samples and in its faunal contents. Remarkable contrasts exist between radiolarian assemblages extracted from claystones of Site 765 and reexamined DSDP Site 261, and faunas recovered from radiolarian sand layers of Site 765. Clay faunas are unusual in their low diversity of apparently ecologically tolerant species, whereas sand faunas are dominated by non-Tethyan species that have never been reported before. Comparisons with Sites 766 and 261, as well as sedimentological observations, lead to the conclusion that this faunal contrast results from a difference in provenance, rather than from hydraulic sorting. Biostratigraphic dating proved difficult principally because of the paucity or even absence of (Tethyan) species used in published zonations. In addition, published zonations are contradictory and do not reflect total ranges of species. Radiolarian assemblages recovered from claystones at Sites 765 and 261 in the Argo Basin reflect restricted oceanic conditions for the latest Jurassic to Barremian time period. Neither the sedimentary facies nor the faunal associations bear any resemblance to sediment and radiolarian facies observed in typical Tethyan sequences. I conclude that the Argo Basin was paleoceanographically separated from Tethys during the Late Jurassic and part of the Early Cretaceous by its position at a higher paleolatitude and by enclosing landmasses, i.e., northeastern India and the Shillong Block, which were adjacent to the northwestern Australian margin before the opening. Assemblages recovered from radiolarian sand layers are dominated by non-Tethyan species that are interpreted as circumantarctic. Their sudden appearance in the late Berriasian/early Valanginian pre-dates the oceanization of the Indo-Australian break-up (Ml 1, late Valanginian) by about 5 m.y., but coincides with a sharp increase in margin-derived pelagic turbidites. The Indo-Australian rift zone and its adjacent margins probably were submerged deeply enough to allow an intermittent "spillover" of circumantarctic cold water into the Argo Basin, creating increased bottom current activity. Circumantarctic cold-water radiolarians transported into the Argo Basin upwelled along the margin and died en masse. Concomitant winnowing by bottom currents led to their accumulation in distinct radiolarite layers. High rates of faunal change and the sharp increase of bottom current activity are thought to be synchronous with the two pronounced late Berriasian-early Valanginian lowstands in sea level. Hypothetically, both phenomena might have been caused by a glaciation on the Antarctic-Australian continent, which was for the first time isolated from the rest of Gondwana by oceanic seaways as a result of Jurassic and Early Cretaceous seafloor spreading. The absence of typical Tethyan radiolarian species during the late Valanginian to late Hauterivian period is interpreted as reflecting a time of strong influx of circumantarctic cold water following oceanization (Mil) and rapid spreading between southeast India and western Australia. The reappearance and gradual increase in abundance and diversity of Tethyan forms along with the still dominant circumantarctic species are thought to result from overall more equitable climatic conditions during the Barremian and early Aptian and may have resulted from the establishment of an oceanic connection with the Tethys Ocean during the early Aptian.

An insight into the breakup of Gondwana: Identifying events through low-temperature thermochronology from the basement rocks of Madagascar

Fission track analysis was applied to the Precambrian suites of Madagascar in order to identify the lower-temperature cooling histories and their relationships to the Phanerozoic events that affected the island. Apatite ages range from 431 to 68 Ma, and zircon ages range from 452 to 238 Ma. Thermochronologically, the island can be divided into a southern, central, and northern region each with a subdivision on an east-west basis. The southern region is sharply separated from the central region by strongly contrasting apparent apatite ages over the northwest-southeast striking Ranotsara Shear Zone (RSZ). The change in apparent ages over the RSZ is indicative of later reactivation along younger brittle faults. The central region has the oldest ages of the island and has a diffuse contact to the third region northward. Along the entire western margin of the Precambrian basement initial Paleozoic exhumation was followed by heating (burial by sediments) during Jurassic and Cretaceous times. A decrease in ages along the eastern margin from 119 to 68 Ma coincides with the predicted positions of the Marion hot spot after effects of erosion are considered. On the other hand, these ages may represent progressive opening of the margin in a southward direction together with associated denudation of the rift shoulder. The eastern part of the central region has remained very stable since at least Devonian times, undergoing only long-term very slow exhumation at rates of 1–5 m/Myr.

Air Gondwana : Using ICT to create an authentic learning environment to teach basic negotiation skills

In recent years greater emphasis has been placed by many Law Schools on teaching not only the substantive content of the law but also the skills needed for the practice of the law. Negotiation is one such skill. However, effective teaching of negotiation may be problematic in the context of large numbers of students studying in a variety of modes and often juggling other time commitments. This paper examines the Air Gondwana program, a blended learning environment designed to address these challenges. The program demonstrates that ICT can be used to create an authentic learning experience which engages and stimulates students.

Large margin vector quantization

In this paper we describe the Large Margin Vector Quantization algorithm (LMVQ), which uses gradient ascent to maximise the margin of a radial basis function classifier. We present a derivation of the algorithm, which proceeds from an estimate of the class-conditional probability densities. We show that the key behaviour of Kohonen's well-known LVQ2 and LVQ3 algorithms emerge as natural consequences of our formulation. We compare the performance of LMVQ with that of Kohonen's LVQ algorithms on an artificial classification problem and several well known benchmark classification tasks. We find that the classifiers produced by LMVQ attain a level of accuracy that compares well with those obtained via LVQ1, LVQ2 and LVQ3, with reduced storage complexity. We indicate future directions of enquiry based on the large margin approach to Learning Vector Quantization.

Boosting the margin: A new explanation for the effectiveness of voting methods

One of the surprising recurring phenomena observed in experiments with boosting is that the test error of the generated classifier usually does not increase as its size becomes very large, and often is observed to decrease even after the training error reaches zero. In this paper, we show that this phenomenon is related to the distribution of margins of the training examples with respect to the generated voting classification rule, where the margin of an example is simply the difference between the number of correct votes and the maximum number of votes received by any incorrect label. We show that techniques used in the analysis of Vapnik's support vector classifiers and of neural networks with small weights can be applied to voting methods to relate the margin distribution to the test error. We also show theoretically and experimentally that boosting is especially effective at increasing the margins of the training examples. Finally, we compare our explanation to those based on the bias-variance decomposition.

Exponentiated gradient algorithms for conditional random fields and max-margin Markov Networks

Log-linear and maximum-margin models are two commonly-used methods in supervised machine learning, and are frequently used in structured prediction problems. Efficient learning of parameters in these models is therefore an important problem, and becomes a key factor when learning from very large data sets. This paper describes exponentiated gradient (EG) algorithms for training such models, where EG updates are applied to the convex dual of either the log-linear or max-margin objective function; the dual in both the log-linear and max-margin cases corresponds to minimizing a convex function with simplex constraints. We study both batch and online variants of the algorithm, and provide rates of convergence for both cases. In the max-margin case, O(1/ε) EG updates are required to reach a given accuracy ε in the dual; in contrast, for log-linear models only O(log(1/ε)) updates are required. For both the max-margin and log-linear cases, our bounds suggest that the online EG algorithm requires a factor of n less computation to reach a desired accuracy than the batch EG algorithm, where n is the number of training examples. Our experiments confirm that the online algorithms are much faster than the batch algorithms in practice. We describe how the EG updates factor in a convenient way for structured prediction problems, allowing the algorithms to be efficiently applied to problems such as sequence learning or natural language parsing. We perform extensive evaluation of the algorithms, comparing them to L-BFGS and stochastic gradient descent for log-linear models, and to SVM-Struct for max-margin models. The algorithms are applied to a multi-class problem as well as to a more complex large-scale parsing task. In all these settings, the EG algorithms presented here outperform the other methods.

Exponentiated gradient algorithms for large-margin structured classification

We consider the problem of structured classification, where the task is to predict a label y from an input x, and y has meaningful internal structure. Our framework includes supervised training of Markov random fields and weighted context-free grammars as special cases. We describe an algorithm that solves the large-margin optimization problem defined in [12], using an exponential-family (Gibbs distribution) representation of structured objects. The algorithm is efficient—even in cases where the number of labels y is exponential in size—provided that certain expectations under Gibbs distributions can be calculated efficiently. The method for structured labels relies on a more general result, specifically the application of exponentiated gradient updates [7, 8] to quadratic programs.

Margin-adaptive model selection in statistical learning

A classical condition for fast learning rates is the margin condition, first introduced by Mammen and Tsybakov. We tackle in this paper the problem of adaptivity to this condition in the context of model selection, in a general learning framework. Actually, we consider a weaker version of this condition that allows one to take into account that learning within a small model can be much easier than within a large one. Requiring this “strong margin adaptivity” makes the model selection problem more challenging. We first prove, in a general framework, that some penalization procedures (including local Rademacher complexities) exhibit this adaptivity when the models are nested. Contrary to previous results, this holds with penalties that only depend on the data. Our second main result is that strong margin adaptivity is not always possible when the models are not nested: for every model selection procedure (even a randomized one), there is a problem for which it does not demonstrate strong margin adaptivity.