Catchment modelling and sensitivity analysis to spatial interpolation methods of daily rainfall

Doctoral thesis (2012)

Spatial interpolation of precipitation data is of great importance for hydrological modelling. Geostatistical methods (kriging) are widely applied in spatial interpolation from point measurement to ... [more ▼]

Spatial interpolation of precipitation data is of great importance for hydrological modelling. Geostatistical methods (kriging) are widely applied in spatial interpolation from point measurement to continuous surfaces. The objective of this dissertation is to confront the performances of the several spatial interpolation methods, in particular the geostatistics for daily rainfall data in the nested catchments to realize a sensitivity analysis on discharges modelled at the outlets to the daily rainfall input in the modelling. The previous studies in the literature emphasized a requirement of novel investigation on the essential means to recover the rainfall data and eventually, the quality of the hydrological modelling. This study leaned on 30-year daily rainfall data of 70 raingages in the hilly landscape of the Ourthe and Ambleve catchments in Belgium (2908 km²). Two common deterministic methods are employed here. The Thiessen Polygon (THI) assigns the value from the nearest observation to a certain grid cell. The Inverse Distance Weighting (IDW) is an advanced nearest neighbour approach that allows including more observations than only the nearest points. The value at a certain grid cell is obtained from a linear combination of the surrounding locations. In the geostatistical algorithms, the spatial variation may be better described by a stochastic function. Four versions of kriging are used. The Ordinary Kriging (ORK) is the basic form of Kriging that the prediction is also a linear combination of the measured values. But the spatial correlation between the data, as described by the variogram, determines the weights, assuming that the mean is constant but unknown. The Universal Kriging (UNK) is based on the hypothesis that this mean is a polynomial function of spatial coordinates. So, this type of kriging is not stationary with regard to the mean. The Kriging with External Drift (KED) supposes that the mean of the interest variable depends on auxiliary variables; the theory of this kriging is in fact the same as the theory of UNK, which also contains a non-constant mean. The drift is defined externally through some auxiliary variables. The Ordinary Cokriging (OCK) suggests estimating the variable of interest by weighted linear combination of its observations and the observations of the auxiliary variables. This technique requires the study of the spatial dependence between variables besides the study of the simple spatial dependences. All types of kriging use a variogram model to characterise spatial correlation. A variogram describes in terms of variances how spatial variability changes as a function of distance. Seven semi-variogram models (logarithmic, power, exponential, Gaussian, rational quadratic, spherical and penta-spherical) were fitted to daily sample semi-variogram on a daily basis. These seven variogram models were also adopted to avoid negative interpolated rainfall. The elevation, extracted from a digital elevation model, was incorporated into multivariate geostatistics. Seven validation raingages and cross validation were used to compare the interpolation performance of these algorithms applied to different densities of raingages. The areal rainfalls are calculated for the catchment area and used for analysis of extreme rainfall. The effects of the interpolation methods on the extreme rainfall are analysed. The interpolated rainfalls are also used as rainfall input of the physically-based and distributed EPIC-GRID model. The long series of model results are analysed by comparison with the observed discharges at the different outlets of the catchments. Then, the extreme discharges at the outlets are computed. All of these investigations always take into account of the raingage density and the raingage position for very sparse raingage cases. The main results show that among the methods based on the only rainfall data, the geostatistics and IDW are the best ones. The performances change according to the density of the raingages. For the extreme rainfall and long term modelling results, no big difference is found for a high density of raingage but large difference are found for the case of the scattered raingage. For the latter case, UNK and KED are very sensitive to the position of the raingages. IDW, ORK and OCK are found to be the best performance. However for the extreme flow, KED and OCK is the best whereas IDW is not better for the high density of the raingages. For the case of the scattered raingages, the difference in extreme discharge between the interpolation methods is very large and larger than shown in the extreme rainfall and the long term modelling results. IDW, ORK and OCK always perform better. UNK and KED still are sensitive to the positions of raingages. The index of position is used to describe the form of polygon defined by the four raingages. This index is defined as the relation between the polygon perimeter and a circle perimeter having an area equivalent to that of the polygon. The best raingage position for all interpolation methods used in this research should be arround the catchment area, its index should be close to one. [less ▲]

A multiscale model of sprouting angiogenesis during fracture healing.

Conference (2012, September 18)

Sensitivity of a Greenland ice sheet model to atmospheric forcing fields

in Cryosphere (The) (2012), 6

Predicting the climate for the future and how it will impact ice sheet evolution requires coupling ice sheet models with climate models. However, before we attempt to develop a realistic coupled setup, we ... [more ▼]

Predicting the climate for the future and how it will impact ice sheet evolution requires coupling ice sheet models with climate models. However, before we attempt to develop a realistic coupled setup, we propose, in this study, to first analyse the impact of a model simulated climate on an ice sheet. We undertake this exercise for a set of regional and global climate models. Modelled near surface air temperature and precipitation are provided as upper boundary conditions to the GRISLI (GRenoble Ice Shelf and Land Ice model) hybrid ice sheet model (ISM) in its Greenland configuration. After 20 kyrs of simulation, the resulting ice sheets highlight the differences between the climate models. While modelled ice sheet sizes are generally comparable to the observed one, there are considerable deviations among the ice sheets on regional scales. These deviations can be explained by biases in temperature and precipitation near the coast. This is especially true in the case of global models. But the deviations between the climate models are also due to the differences in the atmospheric general circulation. To account for these differences in the context of coupling ice sheet models with climate models, we conclude that appropriate downscaling methods will be needed. In some cases, systematic corrections of the climatic variables at the interface may be required to obtain realistic results for the Greenland ice sheet (GIS). [less ▲]

Apprentissages automatiques supervisés pour le monitoring environnemental et énergétique d'une chaudière de régénération

Doctoral thesis (2012)

The broad objective of this thesis is to apply and compare supervised learning techniques for prediction of nitrogen oxide pollutant emission from the recovery boiler of a Kraft pulp mill. In this task ... [more ▼]

The broad objective of this thesis is to apply and compare supervised learning techniques for prediction of nitrogen oxide pollutant emission from the recovery boiler of a Kraft pulp mill. In this task, we want to highlight a technique that is most suited and self-adapted to boiler transient operating conditions. The Kraft process is an alkaline process to produce chemical pulp; cellulose fibers are dissociated from lignin by cooking the chips in a solution of sodium hydroxide (NaOH) and sodium sulfide (Na2S), called white liquor. The residual black liquor is concentrated and burned in a recovery furnace to yield an inorganic smelt of sodium carbonate (Na2CO3) and Na2S. The recovery boiler both regenerates the cooking chemicals and produces high pressure steam to the pulp mill, but the boiler is a major source of atmospheric pollutants in the mill. In particular nitrogen oxide formation is very complex because of several chemicals and dynamic mechanisms: thermal NOx, prompt NOx and fuel NOx. Nowadays, there is an increasing demand in such industries for efficient data analysis tools, especially for pollutant monitoring and/or energy management. Literature reviews refer mainly on numerical solutions where a complete description of the process is needed and where stationary condition is often a working hypothesis. This is the case with the advanced data validation and reconciliation techniques that we evaluate. This technique is based on thermodynamic models, chemical and physical relationships within process parameters and equipment. This is helpful to highlight some lack of information about the process, but this approach failed to model accurately steam and fumes utilities operating points. Indeed, in a Kraft recovery boiler, the total nitrogen oxide emission is dependent on several operating factors and heterogeneous conditions, e.g. operating fuels (black liquor or heavy fuel), furnace load, droplet size, air system operation, retention time, biomass characteristics,... For such a complex problem, machine learning techniques may be used as alternative methods in engineering analysis and predictions. They involve algorithms that improve automatically through experience collected in historical databases. Among supervised learning techniques, we focus mainly on neural networks methods (static and dynamic architectures) and additionally on tree-based (regression tree and random forests) and linear ones. For each method, we evaluate its ability to predict NOx pollutant emission in varying conditions. A random forest is a collection of uncorrelated regression trees, induced from bootstrap samples of the training data. Its internal estimates are also used to measure variable importance and allow us to classify relevant variables for a model inputs selection task. Note that we need some additional a priori knowledge to select the final inputs set. Among static neural network structures, the multilayer perceptron is the most widely used, particularly the two-layer structure in which the input units and the output layer are interconnected with an intermediate hidden layer. The model of each neuron in the network includes a nonlinear activation function that is differentiable; this network can perform static mapping between an input space and an output space. Within dynamic architectures, we distinguish those that have only feed-forward connections and those that have feedback (recurrent) connections. In this work, we focus mainly on NARX network (Nonlinear AutoRegressive model with eXogenous inputs) and additionally on Elman recurrent neural network. This last one incorporates an additional layer, called context layer, the nodes of which are the one-step delay elements embedded into the local feedback paths. Nevertheless, Elman's approach has some drawbacks associated with learning parameters scheme and temporal gradient approximation. Particularly, the NARX network is used for input-output modeling of nonlinear dynamical systems. It is a recurrent model: model inputs are applied to a tapped-delay-line memory of n units and outputs are fed back to the input layer through another line of m units. The total model order s=n+m is therefore a key parameter and the method of Lipschitz numbers is a tool for estimating it. An advantage of NARX is that we can use standard backpropagation algorithm for neural network learning scheme. Furthermore, to increase model robustness, we average neural predictions over a set of individual neural predictors, this is helpful for reducing variance prediction across trials. Despite the fact that generalization is done on the worst case configuration possible, we see that ensemble of NARX networks perform well on predicting NOx emissions during transient operations and Lipschitz numbers are very helpful for system orders estimations. We illustrate the potential of a dynamic neural approach compared to the others in the nitrogen oxide prediction task. It is more suited to practical modeling needs and offers a modeling of time and memory. It allows us to monitor NOx pollution and possibly adjusting control variables and performing diagnostics. The thesis is divided into seven chapters covering several publications. Chapter 1 is about the Kraft process and its recovery boiler. We start with a short description of the Kraft pulp mill. Then we describe the Kraft recovery boiler, some chemical reactions in the furnace, the steam production equipments and the atmospheric pollutants. Finally we discuss about nitrogen oxide formation in the furnace, the effects of several operating conditions on its production. Chapter 2 is about data mining, on what it is, on what it is used for and which are the main modeling cultures. This chapter deals with system identification, modeling approaches (white box, grey box, black box), some definitions about learning and modeling, and finally some links between modeling and optimization techniques. Chapter 3 starts with a state-of-the-art about numerical simulation of a Kraft recovery boiler, then we apply and evaluate a data validation scheme for steam and fumes utilities modeling. Finally we discuss the application of artificial intelligence techniques within the framework of a recovery boiler. Chapter 4 aims at selecting model inputs, starting with a supervised selection approach based on random forests. We introduce some methodological insights about tree-based methods, from a simple regression tree to random forests. Random forests internal estimates are used to measure the relative importance of each input variable in predicting a response, i.e. nitrogen oxide emission or high pressure steam production. Finally we discuss about some useful extra knowledge to take in account for the selection of final inputs. Chapter 5 is about neural networks modeling, we introduce the perceptron, the multilayer perceptron, and the associated backpropagation algorithm. We discuss about static and dynamic architectures, especially the Elman recurrent neural network. Finally, we apply a multilayer perceptron and an Elman recurrent neural network for predicting the high pressure steam flow rate from the Kraft recovery boiler. Chapter 6 presents some insights about input-output modeling of nonlinear dynamical systems, especially with NARX network. At the end, we explain the Lipschitz method that is applied for system orders estimation. Chapter 7 summarizes some comparison results about supervised learning techniques applied to predict nitrogen oxide pollutant emission from the recovery boiler. This comparison involves neural network techniques, tree-based methods and multiple linear regression. Finally, some research perspectives are presented and some conclusions are drawn. [less ▲]

Lorsque les limites des sciences et les sentiers de chasse s’entrecroisent: Une ethnographie des laboratoires scientifiques amazoniens en partenariat avec les communautés autochtones locales

Doctoral thesis (2012)

Currently, protected areas and protected natural species are in a state of proliferation all over the world. Landscapes, flagship species, peoples in voluntary isolation, ecosystems, and “biological ... [more ▼]

Currently, protected areas and protected natural species are in a state of proliferation all over the world. Landscapes, flagship species, peoples in voluntary isolation, ecosystems, and “biological corridors” are all protected. In order to achieve this, categories of protection at different geographical levels have been mobilized. The processes that allow any one of these categories to become operational within a precise territory are more or less well standardized and have in common, amongst others things, two components that have caught my attention: Firstly, the central role of previous scientific and technical studies, and secondly, the desire to ensure the participation of the populations concerned. How are relations between scientists and the non-scientific public in participatory environmental management? This thesis seeks to answer this question, from the starting point of an analysis of the cognitive dimension of participatory environmental management, i.e. how the production of knowledge affects the modes of existence within protected areas (and consequently their human and nonhuman inhabitants). Building upon the study of daily activities that bring scientists and indigenous peoples into contact in and around the scientific field stations “Yasuní” and “Juri Juri Kawsay” (in the Ecuadorian Amazon), this research describes the relationship between ways of knowing a conflicting object (for example a natural species, a territorial boundary, an isolated people) and the consequences in terms of configuration of that object (in terms of identity, value, and fragility) and its management (administration, use, protection, recovery or abandonment). By mobilizing the approach of Actor-Network theory, this dissertation attempts to outline how mediation between humans and nonhumans, non-scientific and scientific, occurs. It also offers a reflection on how the Agency of an entity (the competence to make make distributed in the humans and nonhumans actors) comes to influence the joint actions. This research can capture the ambiguity, instability, resistance and complexity that still exist in the Amazon, with some unexpected findings along the way, by showing the changes among things and people throughout development projects, land management initiatives, and conservation and science programs. After describing the trajectory of four entities (laboratories, species conservation, territories and subjects), which meet at the Amazonian interfaces where science, conservation and development are in competition, the findings of this dissertation have met two ambitions: to consider the importance of taking into account the complexity and historicity of scientific data and to answer a broader question: What are the relevant factors to take into account when considering the possibilities for joint action between scientists and indigenous populations? [less ▲]

NORS: Demonstration Network Of ground-based Remote Sensing observations in support of the GMES atmospheric service

Poster (2012, September 17)

NORS (Demonstration Network Of ground-based Remote Sensing Observations in support of the GMES Atmospheric Service) aims at demonstrating the value of ground – based remote sensing data from the Network ... [more ▼]

NORS (Demonstration Network Of ground-based Remote Sensing Observations in support of the GMES Atmospheric Service) aims at demonstrating the value of ground – based remote sensing data from the Network for the Detection of Atmospheric Composition Change for quality assessment and improvement of the GMES Atmospheric Service products. [less ▲]

Performance Measurements of the microPET FOCUS 120 for Iodine-124 Imaging

in IEEE Transactions on Nuclear Science (2012), PP

This study aimed to evaluate the performance of the microPET FOCUS 120 for 124I in terms of counting rate capability and image quality using the NEMA NU 4-2008 methodology. Scanner sensitivity was ... [more ▼]

This study aimed to evaluate the performance of the microPET FOCUS 120 for 124I in terms of counting rate capability and image quality using the NEMA NU 4-2008 methodology. Scanner sensitivity was measured for 124I for comparison and reached 75 cps/kBq, respectively, with the usual 350-650 keV energy window (EW) and 6 ns time window (TW). The noise equivalent count rate (NECR) index was defined as: NECR = RT2/(RP+RGP) (T = true, P = prompt, GP = γ-prompt). A rat phantom maximum NECR of 48 kcps was obtained for the 250-590 keV EW with 6 ns TW. An almost identical maximum NECR of 43 kcps was recorded for 350-590 and 350-650 keV EW and 6 ns TW. The 2 ns TW reduced the sensitivity and NECR by 40-50% for all EW. The mouse phantom NECR study was limited because of the maximum available activity concentration of 124I. The 250-590 keV EW showed the largest scatter and γ-prompt plus scatter fractions with 25.7% and 43%, respectively, for the rat phantom and 12.2% and 27% for the mouse phantom. With the 350-590 keV EW, these fractions decreased to 20% and 33.5% for the rat phantom and to 10% and 21% for the mouse phantom. The image quality was investigated with the NEMA NU 4-2008 dedicated phantom for four (two analytic and two iterative) 2D or 3D reconstruction methods. The lowest spillover ratios (SOR) for the phantom non-emitting regions were obtained for the 350-590 and 350-650 keV EWs. Recovery coefficients (RC) of the hot rods were the highest for the 350-590 keV EW except for the 1 mm rod. Scatter correction led to a large decrease in RC. The combination of the 350-590 keV EW with 6 ns TW appeared to be a good compromise between counting rate capability and image quality for the FOCUS 120, especially when maximum a posteriori reconstruction was used without scatter correction. Moreover this combination enabled the best quantification with an error as low as 0.36%. [less ▲]

Intraocular Lens Adsorbome: a Proteomic Study of Adsorbed Proteins onto Acrylic Materials and Its Implication in Secondary Cataract

Poster (2012, September 17)

The intraocular lens (IOL) is a polymer implant designed to replace the natural lens after cataract surgery. When the implant is introduced into the lens capsule, the polymer starts to interact with the ... [more ▼]

The intraocular lens (IOL) is a polymer implant designed to replace the natural lens after cataract surgery. When the implant is introduced into the lens capsule, the polymer starts to interact with the aqueous humour and the exchange of molecules between the solid and the liquid begins. The nature of exchange in water, ions, and biomolecules may result in several postoperative complications including glistening, calcification, and posterior capsular opacification. The posterior capsular opacification (PCO, also called “Secondary Cataract”) is raised from the over-growth of residual lens epithelial cells. The first step of the over-growth process of the cells is their adhesion to the deposited biomolecules, such as proteins involved in extra-cellular matrices. The purpose of this study is to identify the principal proteins adsorbed onto the acrylic polymers by mass spectrometry. The concept of adsorbome is to generate a list of adsorbed proteins to the hydrophilic and hydrophobic polymers, and then compare the difference to the original component of aqueous humour in order to see the affinity of individual protein to each material. Two kinds of hydrophilic and two kinds of hydrophobic acrylic polymers were tested for their adsorbomes by treating them with an aqueous humour analogue and the major adsorbed proteins were identified by mass spectrometry. Interestingly, the hydrophilic acrylic polymer shows a relative lower protein adsorption rate but shows a higher incidence of secondary cataract. This phenomenon implies the adsorbed proteins play a crucial role in progress of secondary cataract. [less ▲]

An analysis of learning interactions in a cross-border network for sustainable urban neighbourhood development.

in Journal of Cleaner Production (2012), Online access

Learning networks have the potential to support joint learning and collective innovation processes needed for sustainable urban development. However, systematic analyses of joint learning processes in ... [more ▼]

Learning networks have the potential to support joint learning and collective innovation processes needed for sustainable urban development. However, systematic analyses of joint learning processes in such networks are often lacking. In this paper, the Interreg project SUN (Sustainable Urban Neighbourhoods) is taken as a case study to analyse learning interactions in a regional network for sustainable urban neighbourhood development in the Euregio Meuse-Rhine. [less ▲]

Introduction au droit public - Plan général du cours (année 2012-2013)

Learning material (2012)

Le document constitue le Plan général de l'enseignement d'Introduction au Droit public dispensé à l'Université de Liège en 1ère année de bachelier en droit et en 1ère année de bachelier en science ... [more ▼]

Le document constitue le Plan général de l'enseignement d'Introduction au Droit public dispensé à l'Université de Liège en 1ère année de bachelier en droit et en 1ère année de bachelier en science politique. [less ▲]