Water-induced soil erosion and shallow landslides interact with each other and need to be studied in an integrated approach to understand hillslope sediment yields. The principal aim of this thesis was to study and model soil erosion and shallow landslides in an integrated way. The thesis presents results from laboratory and catchment-scale studies and modelling. A laboratory flume under a rainfall simulator was used for shallow landslide and soil erosion experiments using sandy and silty loess soils. In the experiments, landslide initiation, retrogressions and slip surface depths were measured and monitored directly or by using video camera recordings. Sediment and runoff were collected from the flume outlet every minute during landslides and every 10 minutes before and after landslides. Changes in the soil slope, after landslides, were recorded. Initially, six experiments including two repetitions were conducted using sandy soils at a 30º and 10º compound slope configuration, but with different soil profile depths. The experimental results showed that total and landslide-driven sediment yields were affected by the original soil profile depth; the greater the depth, the higher the sediment yield. Later, twelve other experiments were conducted on different slopes using silty loess soils. The experimental observations were used to validate an integrated modelling approach which includes WEPP for runoff and soil erosion modelling, a slope stability model for simulating shallow landslides, and a simple soil redistribution model for runout distance prediction. The model predictions were in good alignment with the observations. In all (sandy and silty loess) experiments, peak sediment discharges were related to the landslide events, proximity to the outlet and landslide volume. The post-failure soil erosion rate decreased as a function of changes in the slope profile. The GeoWEPP-SLIP modelling approach was proposed for catchment-scale modelling. The approach simulates soil erosion using the Hillslope and Flowpath methods in WEPP, predicts shallow landslides using a slope stability model coupled with the WEPP’s hillslope hydrology and finally uses a simple rule-based soil redistribution model to predict runout distance and post-failure topography. A case study application of the model to the Bowenvale research catchment (300 ha) showed that the model predictions were in good agreement with the observed values. However, the Hillslope method over-predicted the outlet sediment yield due to the computational weighting involved in the method. The Hillslope method predicted consistent values of sediment yield and soil erosion regardless to the changes in topography and land-cover in the post-failure scenarios. The Flowpath method, on the other hand, predicted higher values of sediment yield in the post-failure vegetation removal scenario. The effects of DEM resolution on the approach were evaluated using four different resolutions. Statistical analyses for all methods and resolutions were performed by comparing the predicted versus measured runoff and sediment yield from the catchment outlet and the spatial distribution of shallow landslides. Results showed that changes in resolution did not significantly alter the sediment yield and runoff between the pre- and post-failure scenarios at the catchment outlet using the Hillslope method. However, the Flowpath method predicted higher hillslope sediment yields at a coarser resolution level. Similarly, larger landslide areas and volumes were predicted for coarser resolutions whereas deposition volume decreased with the increase in grid-cell size due to changes in slope and flowpath distributions. The research conducted in the laboratory and catchment presented in this thesis helped understand the interactions between shallow landslides and soil erosion in an integrated approach.

Cities have adopted evolving urban digitization strategies, and most of those increasingly focus on data, especially in the field of public transportation. Transportation data have intuitively spatial and temporal characteristics, for they are often described with when and where the trips occur. Since a trip is often described with many attributes, the transportation data can be presented with a tensor, a container which can house data in $N$-dimensions. Unlike a traditional data frame, which only has column variables, tensor is intuitively more straightforward to explore spatio-temporal data-sets, which makes those attributes more easily interpreted. However, it requires unique techniques to extract useful and relatively correct information in attributes highly correlated with each other. This work presents a mixed model consisting of tensor decomposition combined with seasonal vector autoregression in time to find latent patterns within historical taxi data classified by types of taxis, pick-up and drop-off times of services in NYC, so that it can help predict the place and time where taxis are demanded. We validated the proposed approach using the experiment evaluation with real NYC tax data. The proposed method shows the best prediction among alternative models without geographical inference, and captures the daily patterns of taxi demands for business and entertainment needs.

The separation of aqueous and organic mixtures with thermal separation processes is an important and challenging task in the chemical industry. Rising prices for energy, stricter environmental regulations and the increasing demand for high purity chemicals are the main driving forces to find alternative solutions to common separation technologies such as distillation and absorption. These are mostly too energy consumptive and can show limited separation performance, especially when applied to close boiling or azeotropic mixtures. Pervaporation can overcome these thermodynamic limitations and requires less energy because only the separated components need to be evaporated. This separation technology is already well established for the production of anhydrous solvents, but not yet widely distributed in the chemical and petrochemical industry due to some crucial challenges, which are still to overcome. Besides the need of high selective membranes, the development of membrane modules adapted to the specific requirements of organoselective pervaporation needs more research effort. Furthermore, only few modelling and simulation tools are available, which hinders the distribution of this process in industrial scale. In this work, these issues are addressed in a combined approach. In close collaboration with our cooperation partners, a novel membrane module for organophilic pervaporation is developed. A novel technology to manufacture high selective polymeric pervaporation membranes is applied to produce a membrane for an industrially relevant organic-organic separation task. A three step modelling approach ranging from a shortcut and a discrete to a rigorous model is developed and implemented in a user interface. A hydrophilic and an organophilic membrane are characterised for the separation of a 2-butanol/water mixture in a wide range of feed temperature and feed concentration in order to establish a generally valid description of the membrane performances. This approach is implemented in the three developed models to simulate the novel membrane module in industrial scale. The simulations are compared to the results of pilot scale experiments conducted with the novel membrane module. Good agreement between simulated and experimental values is reached.

Aux Antilles, la disponibilité limitée des ressources en contexte insulaire et l'activité agricole à aux niveaux d'intrants fragilisent les ressources en eau. Dans les zones bananières à forts niveaux d'intrants, des phénomènes de pollution des eaux sont particulièrement marqués, notamment du fait de l'usage d'un pesticide, la chlordécone. La connaissance du fonctionnement des systèmes hydrologiques à l'échelle du bassin versant représente un enjeu primordial pour pouvoir estimer l'exposition de l'écosystème aux pollutions et pour prévoir l'évolution des contaminations dans le temps. L'objectif de la thèse a été de développer une modélisation mécaniste représentant les processus hydrologiques de surface et souterrains à l'échelle d'un bassin versant sur substrat volcanique, sous climat tropical humide, situé en zone bananière en Guadeloupe. En premier a été développé un modèle parcellaire de bilan hydrique adapté au cas des cultures bananières. Il a pour originalité de prendre en compte l'importante redistribution de pluie opérée par le couvert bananier et d'en simuler les effets en matière d'intensité et d'hétérogénéité intra-parcellaire sur les termes du bilan hydrique. Les résultats d'analyse de sensibilité montrent que la redistribution de la pluie augmente le ruissellement de surface ainsi que la percolation, en cohérence avec les observations de terrain, mais impacte peu ou temporairement l'évapotranspiration et l'humidité du sol. Le calage du modèle sur des données expérimentales indiquent une performance améliorée de la simulation du ruissellement par rapport à un modèle ignorant le mécanisme de redistribution. En second, le bassin versant expérimental de Féfé (17.8 ha) a fait l'objet d'une approche de modélisation intégrant processus hydrologiques de surface et souterrains basée sur un chaînage itératif des modèles MHYDAS et MODFLOW. Confrontée à une année hydrologique de mesures de débits à l'exutoire et de piézométries, l'approche de chaînage de modèles de surface et souterrain apparaît pertinente. Une limite majeure est toutefois la non prise en compte de la zone non saturée dans le processus de recharge des aquifères. L'analyse des simulations et de leurs écarts avec les données observées conforte plusieurs hypothèses issues des analyses expérimentales : un ruissellement de surface fortement hortonien, une contribution majeure des écoulement souterrains au débit à l'exutoire. Elle indique toutefois également une indétermination des processus majeurs lors des périodes de fortes pluies. Différentes hypothèses sont proposées qu'il conviendra d'évaluer dans des travaux futurs. Ce travail constitue une première étape pour évaluer les chemins d'écoulement majeurs et les dynamiques des contaminations par les produits phytosanitaires dans un milieu volcanique tropical sous culture bananière. / In the French West Indies (FWI), limited resources supply on island and farming with extensive uses of pesticides have damaged water resources. In environments under intensive banana production, water pollution can be of particular concern, with regards to the use of chlordecone, an insecticide to control the banana weevil. Understanding the hydrological behaviour of a catchment is a challenge in assessing the exposure of the ecosystem to pollutions and in predicting the long-term contamination dynamics. This thesis aimed at developing a model to simulate de surface and underground hydrological processes at the catchment scale on volcanic deposits in a humid tropical area covered by banana plantations in Guadeloupe. First, we developed an original water budget model at the plot scale, adapted to the banana canopy. It takes into account the high rainfall redistribution by banana cover and simulates the effects of modified rainfall intensities and within-plot heterogeneities on the water balance components. The sensitivity analysis showed that rainfall redistribution promotes surface runoff and percolation, in accordance with the field observations, but influences little or only temporarily the average field evapotranspiration and soil moisture. The model calibration tested on experimental data indicated improved runoff production performances compared to a model without rainfall redistribution. Secondly, the Féfé experimental catchment (17.8 ha) was studied with a linked iterative modelling approach (of MHYDAS and MODFLOW) that includes surface and underground hydrological processes. Tested against a year of outlet discharge and water table depth measurements, the linked modelling approach seems appropriate. However, the main limit of this approach was that it does not consider the transfer through the unsaturated zone when simulating the aquifers' recharge. The analysis of the results and of the differences between measured and simulated variables supported the hypothesis, from experimental analyses, that: the surface runoff is mainly Hortonian, groundwater flow is the main contributor to runoff at the catchment outlet. However, there was still uncertainty concerning the main processes during wet periods. Various hypotheses were suggested and should be investigated in future studies. This work represents a first step towards the evaluation of the major flow paths and contamination dynamics of pesticides on volcanic deposits in a humid tropical area covered by banana plantations.

The diploma thesis deals with the possibilities of using the optimization modelling software system GAMS in risk management. According to the assignment, emphasis is placed on a detailed approach to the program for those, who are interested in its use in the field of risk engineering applications. The first part of the thesis contains the knowledge to understand what the GAMS program is and what it is used for. The next part of the work provides instructions on how to download, install, activate the program and what the user interface of the program looks like. Thanks to mathematical programming, it will be explained on a project on the distribution of lung ventilators, what basic approaches may be used in risk modelling in the GAMS program on a deterministic model. The following are more complex wait-and-see models, which contains the probability parameters and here-and-now models, where we work with demand scenarios and verify whether if they meets the requirements of other scenarios or calculate costs for the highest demands. The two-stage model is also one of the here-and-now models, but it is significantly more complex in its size and range of input data, it includes additional price parameters for added or removed pieces of lung ventilators from the order.

The separation of aqueous and organic mixtures with thermal separation processes is an important and challenging task in the chemical industry. Rising prices for energy, stricter environmental regulations and the increasing demand for high purity chemicals are the main driving forces to find alternative solutions to common separation technologies such as distillation and absorption. These are mostly too energy consumptive and can show limited separation performance, especially when applied to close boiling or azeotropic mixtures. Pervaporation can overcome these thermodynamic limitations and requires less energy because only the separated components need to be evaporated. This separation technology is already well established for the production of anhydrous solvents, but not yet widely distributed in the chemical and petrochemical industry due to some crucial challenges, which are still to overcome. Besides the need of high selective membranes, the development of membrane modules adapted to the specific requirements of organoselective pervaporation needs more research effort. Furthermore, only few modelling and simulation tools are available, which hinders the distribution of this process in industrial scale. In this work, these issues are addressed in a combined approach. In close collaboration with our cooperation partners, a novel membrane module for organophilic pervaporation is developed. A novel technology to manufacture high selective polymeric pervaporation membranes is applied to produce a membrane for an industrially relevant organic-organic separation task. A three step modelling approach ranging from a shortcut and a discrete to a rigorous model is developed and implemented in a user interface. A hydrophilic and an organophilic membrane are characterised for the separation of a 2-butanol/water mixture in a wide range of feed temperature and feed concentration in order to establish a generally valid description of the membrane performances. This approach is implemented in the three developed models to simulate the novel membrane module in industrial scale. The simulations are compared to the results of pilot scale experiments conducted with the novel membrane module. Good agreement between simulated and experimental values is reached.

Foundation Programme (FP) students at the University of Namibia (UNAM) Oshakati Campus display inadequate academic writing abilities. As their aim is to gain admittance to UNAM main campus science-related courses, it is vital to have effective academic writing skills. This action research (AR) study is a comparison of three writing programmes, the process approach, the modeling approach, and the process genre approach which were implemented separately to three different class groups in 2008 and 2009 to improve the writing skills of students and the teaching practice of the researcher. The effects of the interventions were examined using a combination of the quantitative and qualitative research methods. Data were collected using questionnaires, pre- and post-intervention essays and laboratory reports and interviews. The findings indicate that all three approaches improved the academic writing skills of FP students. The process genre approach had a higher rate of effect than the other two approaches. / English Studies / M. A. (Specialisation TESOL)

Foundation Programme (FP) students at the University of Namibia (UNAM) Oshakati Campus display inadequate academic writing abilities. As their aim is to gain admittance to UNAM main campus science-related courses, it is vital to have effective academic writing skills. This action research (AR) study is a comparison of three writing programmes, the process approach, the modeling approach, and the process genre approach which were implemented separately to three different class groups in 2008 and 2009 to improve the writing skills of students and the teaching practice of the researcher. The effects of the interventions were examined using a combination of the quantitative and qualitative research methods. Data were collected using questionnaires, pre- and post-intervention essays and laboratory reports and interviews. The findings indicate that all three approaches improved the academic writing skills of FP students. The process genre approach had a higher rate of effect than the other two approaches. / English Studies / M. A. (Specialisation TESOL)