Abstract:
A 6-compartment biogeochemical model of nitrogen cycling and plankton productivity has been coupled with a 3D general circulation model in an enclosed environment (the Black Sea) so as to quantify and compare, on a seasonal and annual scale, the typical internal biogeochemical functioning of the shelf and of the deep sea as well as to estimate the nitrogen and water exchanges at the shelf break. Model results indicate that the annual nitrogen net export to the deep sea roughly corresponds to the annual load of nitrogen discharged by the rivers on the shelf. The model estimated vertically integrated gross annual primary production is 130gCm-2yr-1 for the whole basin, 220gCm-2yr-1 for the shelf and 40gCm-2yr-1 for the central basin. In agreement with sediment trap observations, model results indicate a rapid and efficient recycling of particulate organic matter in the sub-oxic portion of the water column (60-80m) of the open sea. More than 95% of the PON produced in the euphotic layer is recycled in the upper 100m of the water column, 87% in the upper 80 m and 67% in the euphotic layer. The model estimates the annual export of POC towards the anoxic layer to 4 1010molyr-1. This POC is definitely lost for the system and represents 2% of the annual primary production of the open sea. Final Revised Paper (PDF, 3116 KB) Discussion Paper (BGD) Special Issue Citation: Grégoire, M. and Beckers, J. M.: Modeling the nitrogen fluxes in the Black Sea using a 3D coupledhydrodynamical-biogeochemical model: transport versus biogeochemicalprocesses, exchanges across the shelf break and comparison of the shelf anddeep sea ecodynamics, Biogeosciences, 1, 33-61, doi:10.5194/bg-1-33-2004, 2004. Bibtex EndNote Reference Manager XML

Abstract:
A 6-compartment biogeochemical model of nitrogen cycling and plankton productivity has been coupled with a 3D general circulation model in an enclosed environment (the Black Sea) so as to quantify and compare, on a seasonal and annual scale, the typical internal biogeochemical functioning of the shelf and of the deep sea as well as to estimate the nitrogen and water exchanges at the shelf break. Model results indicate that the annual nitrogen net export to the deep sea roughly corresponds to the annual load of nitrogen discharged by the rivers on the shelf. The model estimated vertically integrated gross annual primary production is 130 g C m-2yr-1 for the whole basin, 220 g C m-2yr-1 for the shelf and 40 g C m-2yr-1 for the central basin. In agreement with sediment trap observations, model results indicate a rapid and efficient recycling of particulate organic matter in the sub-oxic portion of the water column (60-80m) of the open sea. More than 95% of the PON produced in the euphotic layer is recycled in the upper 100m of the water column, 87% in the upper 80 m and 67% in the euphotic layer. The model estimates the annual export of POC towards the anoxic layer to 4 1010mol yr-1. This POC is definitely lost for the system and represents 2% of the annual primary production of the open sea.

Abstract:
Carl Gustav Jung, Swiss psychiatrist, Freud’s collaborator from whom he parted, is the author of the concept of synchronicity. It is defined as a coincidence of simultaneous facts, occurring without causal connection, but linked together through meaning. Synchronicity is different from simple synchronism because of this meaningful connection, offering those who experience it, new insights about their life.The concept of synchronicity assumes that the universe is governed by a principle of unity that C.G. Jung calls the unus mundus. Because of this unity, we are committed, through the unconscious, to a dialog with the universe that surrounds us. The unconscious provokes coincidences that reveal to us who we are and echoes our most intimate questions. Time, space, and consciousness are accomplices in the synchronicity event, giving us thus the strong impression that a greater Meaning than ourselves is calling us, indicating the path to follow, and enlightening us in our decisions. We have all experienced this sort of coincidences, inexplicable through simple chance. This article summarizes the theory of C.G. Jung about the phenomenon of synchronicity and its role on the psychic development and the awareness of Self.

Abstract:
The Black Sea north-western shelf (NWS) is a~shallow eutrophic area in which seasonal stratification of the water column isolates bottom waters from the atmosphere and prevents ventilation to compensate for the large consumption of oxygen, due to respiration in the bottom waters and in the sediments. A 3-D coupled physical biogeochemical model is used to investigate the dynamics of bottom hypoxia in the Black Sea NWS at different temporal scales from seasonal to interannual (1981–2009) and to differentiate the driving factors (climatic versus eutrophication) of hypoxic conditions in bottom waters. Model skills are evaluated by comparison with 14 500 in-situ oxygen measurements available in the NOAA World Ocean Database and the Black Sea Commission data. The choice of skill metrics and data subselections orientate the validation procedure towards specific aspects of the oxygen dynamics, and prove the model's ability to resolve the seasonal cycle and interannual variability of oxygen concentration as well as the spatial location of the oxygen depleted waters and the specific threshold of hypoxia. During the period 1981–2009, each year exhibits seasonal bottom hypoxia at the end of summer. This phenomenon essentially covers the northern part of the NWS, receiving large inputs of nutrients from the Danube, Dniestr and Dniepr rivers, and extends, during the years of severe hypoxia, towards the Romanian Bay of Constanta. In order to explain the interannual variability of bottom hypoxia and to disentangle its drivers, a statistical model (multiple linear regression) is proposed using the long time series of model results as input variables. This statistical model gives a general relationship that links the intensity of hypoxia to eutrophication and climate related variables. The use of four predictors allows to reproduce 78% of hypoxia interannual variability: the annual nitrate discharge (N), the sea surface temperature in the month preceding stratification (T), the amount of semi-labile organic matter in the sediments (C) and the duration of the stratification (D). Eutrophication (N, C) and climate (T, D) predictors explain a similar amount of variability (~35%) when considered separately. A typical timescale of 9.3 yr is found to describe the inertia of sediments in the recovering process after eutrophication. From this analysis, we find that under standard conditions (i.e. average atmospheric conditions, sediments in equilibrium with river discharges), the intensity of hypoxia can be linked to the level of nitrate discharge through a non-linear equation (p

Abstract:
In this article, we describe the first operational implementation of the GHER hydrodynamic model. This happened onboard the research vessel "Alliance", in the context of the Turkish Straits System 2008 campaign, which aimed at the real-time characterization of the Marmara Sea and (south-western) Black Sea. The model performed badly at first, mainly because of poor initial conditions. Hence, as the model includes a reduced-rank extended Kalman filter assimilation scheme, after a hindcast where sea surface temperature and temperature and salinity profiles were assimilated, the model yielded realistic forecasts. Furthermore, the time required to run a one-day simulation (about 5 min of simulation, or 10 min with pre-processing and data transfers included) is very limited and thus operational use of the model is possible.

Abstract:
We investigate the Landau parameters for the instabilities in spin and charge channels in the nondegenerate extended Hubbard model with intersite Coulomb and exchange interactions. To this aim we use the spin rotationally invariant slave boson approach and we determine the necessary inverse propagator matrix. The analytically derived spin Landau parameter $F_0^a$ for the half filled band uncovers the intrinsic instability of the nondegenerate Hubbard model towards ferromagnetism --- negative intersite exchange interaction triggers a ferromagnetic instability at half filling before the metal-insulator transition, indicated by the divergence of the magnetic susceptibility at $F_0^a=-1$. This result is general and the instability occurs in the strongly correlated metallic regime for any lattice, in three or two dimensions. Next as an illustrative example we present numerical results obtained for the cubic lattice with nearest neighbor exchange $J$ and Coulomb $V$ elements and arbitrary electron density. One finds that the range of small doping near half filling is the most unstable one towards spin polarization, but only in the case of ferromagnetic intersite exchange $J<0$. Charge Landau parameter $F_0^s$ is lowered near half filling by increasing $U$ when the intersite Coulomb interaction $V$ is attractive, but in contrast to $F_0^a$ at $J<0$ it requires an attraction beyond a critical value $V_c$ to generate the divergence of the charge susceptibility at $F_0^s=-1$ in the metallic phase. This instability was found for a broad range of electronic filling away from half filling for moderate attraction.

Abstract:
We construct an example of a G_2-instanton over a twisted connected sum G_2-manifold, by applying a gluing result of the third author and Walpuski to instantons constructed from holomorphic bundles (see arXiv:1310.7933). In this example, the moduli spaces of the instantons that are glued are non-trivial, and their images in the moduli space on the cross-section intersect transversely. The holomorphic bundles are obtained using a twisted version of the Hartshorne-Serre construction, which can be used to produce many more examples.

Abstract:
A recent paper of Edman et al. has taken a combinatorial approach to measuring the anonymity of a threshold mix anonymous communications system. Their paper looks at ways of matching individual messages sent to individual messages received, irrespective of user, and determines a measure of the anonymity provided by the system. Here we extend this approach to include in the calculation information about how many messages were sent or received by a user and we define a new metric that can be computed exactly and efficiently using classical and elegant techniques from combinatorial enumeration.

Abstract:
Deep Boltzmann machines are in principle powerful models for extracting the hierarchical structure of data. Unfortunately, attempts to train layers jointly (without greedy layer-wise pretraining) have been largely unsuccessful. We propose a modification of the learning algorithm that initially recenters the output of the activation functions to zero. This modification leads to a better conditioned Hessian and thus makes learning easier. We test the algorithm on real data and demonstrate that our suggestion, the centered deep Boltzmann machine, learns a hierarchy of increasingly abstract representations and a better generative model of data.

Abstract:
The Boltzmann machine provides a useful framework to learn highly complex, multimodal and multiscale data distributions that occur in the real world. The default method to learn its parameters consists of minimizing the Kullback-Leibler (KL) divergence from training samples to the Boltzmann model. We propose in this work a novel approach for Boltzmann training which assumes that a meaningful metric between observations is given. This metric can be represented by the Wasserstein distance between distributions, for which we derive a gradient with respect to the model parameters. Minimization of this new Wasserstein objective leads to generative models that are better when considering the metric and that have a cluster-like structure. We demonstrate the practical potential of these models for data completion and denoising, for which the metric between observations plays a crucial role.