No description available.

Natural tropical deciduous forest (TDF) is considered with a medium to small height (< 15 m). Particularly, in Mexico TDF shows a remnant of 36.2% of primary forest driving changes in the structure and species composition. This vegetation in Mexico is mainly transformed into grassland for cattle raising, and agriculture, primarily for self-consumption. More information about the ecology and the social pressures on this vegetation can be seen in Chapter I. The general methods, including sampling allocation and collection, characteristics of the study site, as well the procedure of the research proposal is presented in Chapter II. The main aim of this thesis is to improve the accuracy of predictions of net carbon emissions and the spatial distribution of AGB in the Tropical Deciduous Forest of Mexico. To address this aim, it is important to take into consideration the forest structure, spatial patterns and processes in a natural forest in a multi-scale analysis; also, it is necessary to characterize the spatial socio-economic drivers that influence current AGB losses. With the understanding of such elements, it is possible to reconstruct the potential carbon stocks and estimate the allocation of net carbon emissions due to deforestation and forest degradation. This study shows that it is possible to count net carbon emissions caused by deforestation and forest degradation at a landscape scale. To come to such estimates, it was necessary to reduce the different sources of uncertainty. Chapter III explores different elements that drive the AGB allocation in a mature forest. The AGB in the mature forest was considered as the potential AGB that the forest could get assuming that it has reached its steady state. Different field sampling strategies and allometric equations were evaluated to account for uncertainty in the AGB estimations. The results showed that small sampling design (300-400 m2) and large-sized plots (4 ha) produce the same tree distribution for trees: ≥30 cm in DBH as well as in AGB. These results contradict what has been reported for others (Chave et al., 2004 and 2005) when they refer to the general definition of tropical forest. However, those other studies referred to forests with a much higher precipitation and which can be classified as tropical rain (perennial) forest (Chave et al., 2004). In the tropical deciduous forest, the kind considered in this study, AGB tends to be allocated in small-sized trees. Diverse biophysical characteristics that may drive AGB allocation were considered over different spatial scales. Water stress was the main driver for AGB density at different spatial scales. Nutrients showed little significance to explain AGB as other studies have suggested in secondary forests and/or chronosequences. With this understanding, Chapter IV shows the use of different multi-variable models. Parsimonious models were the result of the variables selection and sensitivity test. Most of the methodologies showed a better performance to explain AGB allocation than a null-model. However, when they were contrasted with independent observations over different spatial resolutions, it was possible to conclude that only GLM was capable of reproducing the spatial patterns, and its estimations were close to observations. Nevertheless, some observations with very large AGB densities were underestimated by the model. This underestimation was related to the presence of few very large-sized trees. These two chapters depict the possibility of accounting for the potential AGB, and the uncertainty, namely whether the landscape could reach it with the absence of human disturbance. Once the potential AGB map was built and validated, it was transformed to carbon stock, using a local carbon concentration estimate. This potential carbon stock map was contrasted to the different available maps of current carbon stocks. Consequently, it was possible to estimate net carbon emissions due to deforestation and forest degradation (Chapter V), suggesting that the general models tend to agree in the total carbon loss. However, there are some spatial discrepancies in the magnitudes of change. Main differences between maps can be reduced by diverse socio-ecological constraints that dominate the landscape. This is important because it may be possible to make future adjustments that would reduce variability, enabling more accurate AGB estimations. However, to individually account for deforestation and forest degradation, more detailed sources of local information are necessary, such as socio-economic variables. Therefore models with a bottom-up perspective would lead to a better understanding and representation of the landscape. Finally, the growing rural population will have larger demands for wood and food, so while remote or protected areas may have the potential for storing high AGB, forest close to settlements and access routes are likely to continue being disturbed, unless affordable alternatives are available for the sustainable use of the forest. In conclusion, the estimation of spatial heterogeneity of AGB in the landscape is of great importance when measuring carbon stocks and ecological dynamics. Various elements influence the AGB allocation in the mature forest. Among all of them, water availability played the most decisive part of various spatial scales. My models support the hypothesis that water availability plays the major role in explaining AGB in Mexico on a local, sub-regional and landscape scale. Model selection produced contrasting AGB estimates and patterns. Moreover, the results of this study tell us that there is not a clear consensus among various current AGB maps. However, they also show that with a multi-model comparison it is possible to identify carbon emissions drivers and calculate total carbon emissions due to forest disturbances. Socio-economic variables played the major role in explaining AGB losses. Therefore, future studies should look into a bottom-up approach for a better understanding and representation of current AGB.

Wind erosion is a global issue and affecting millions of people in drylands by causing environmental issues (acceleration of snow melting), public health concerns (respiratory diseases), and socioeconomic problems (costs of damages and cleaning public properties after dust storms). Disturbances in drylands can be irreversible, thus leading to natural disasters such as the 1930s Dust Bowl. With increasing attention on aeolian studies, many studies have been conducted using ground-based measurements or wind tunnel studies. Ground-based measurements are important for validating model predictions and testing the effect and interactions of different factors known to affect wind erosion. Here, a machine-learning model (random forest) was used to describe sediment flux as a function of wind speed, soil moisture, precipitation, soil roughness, soil crusts, and soil texture. Model performance was compared to previous results before analyzing four new years of sediment flux data and including estimates of soil moisture to the model. The random forest model provided a better result than a regression tree with a higher variance explained (7.5% improvement). With additional soil moisture data, the model performance increased by 13.13%. With full dataset, the model provided an increase of 30.50% in total performance compared to the previous study. This research was one of the rare studies which represented a large-scale network of BSNEs and a long time series of data to quantify seasonal sediment flux under different soil covers in southern Utah. The results will also be helpful to the managers for controlling the effects on wind erosion, scientists to choose variables for further modeling or local people to increase the public awareness about the effects of wind erosion.

Les lacs jouent un rôle particulier dans le cycle de l’eau dans les bassins versants urbains. La stratification thermique et le temps de séjour de l’eau élevé favorisent le développement phytoplanctonique. La plupart des métaux sont naturellement présents dans l’environnement à l’état de traces. Ils sont essentiels pour les organismes vivants. Néanmoins, certains métaux sont connus pour leurs effets toxiques sur les animaux et les humains. La concentration totale des métaux ne reflète pas leur toxicité. Elle dépend de leurs propriétés et de leur spéciation (fractions particulaires, dissoutes: labiles ou biodisponibles et inertes). Dans les systèmes aquatiques, les métaux peuvent être absorbés par des ligands organiques ou minéraux. Leur capacité à se complexer avec la matière organique dissoute (MOD), particulièrement les substances humiques, a été largement étudiée. Dans les lacs, le développement phytoplanctonique peut produire de la MOD non-humique, connue pour sa capacité complexante des métaux. Pourtant, peu de recherche sur la spéciation des métaux dans la colonne d’eau des lacs urbains a été réalisée jusqu’à présent.Les objectifs principaux de cette thèse sont (1) d’obtenir une base de données fiables des concentrations en métaux traces dans la colonne d’eau de lacs urbains représentatifs; (2) d’évaluer leur biodisponibilité via une technique de spéciation adéquate ; (3) d’analyser leur évolution saisonnière et spatiale et leur spéciation; (4) d’étudier l’impact des variables environnementales, en particulier de la MOD autochtone sur leur biodisponibilité; (5) de lier la concentration des métaux au mode d’occupation du sol du bassin versant.Notre méthodologie est basée sur un suivi in-situ des lacs en complément d’analyses spécifiques en laboratoire. L’étude a été conduite sur trois sites: le lac de Créteil (France), le lac de Pampulha (Brésil) et 49 lacs péri-urbains (Ile de France). Sur le lac de Créteil, plusieurs dispositifs de mesure en continu nous ont fourni une partie de la base de données limnologiques. Dans le bassin versant du lac de Pampulha, la pression anthropique est très importante. Le climat et le régime hydrologique des 2 lacs sont très différents. Les 49 lacs de la région d’Ile de France ont été échantillonnés une fois pendant trois étés successifs (2011-2013). Ces lacs nous ont fourni une base de données synoptique, représentative de la contamination métallique à l’échelle d’une région fortement anthropisée.Afin d’expliquer le rôle des variables environnementales sur la concentration métallique, le modèle Random Forest a été appliqué sur les bases de données du lac de Pampulha et des 49 lacs urbains avec 2 objectifs spécifiques: (1) dans le lac de Pampulha, comprendre le rôle des variables environnementales sur la fraction labile des métaux traces, potentiellement biodisponible et (2) dans les 49 lacs, comprendre la relation des variables environnementales, particulièrement au niveau du bassin versant, sur la concentration dissoute des métaux. L’analyse des relations entre métaux et variables environnementales constitue l’un des principaux résultats de cette thèse. Dans le lac de Pampulha, environ 80% de la variance du cobalt labile est expliqué par des variables limnologiques: Chla, O2, pH et P total. Pour les autres métaux, le modèle n’a pas réussi à expliquer plus de 50 % de la relation entre fraction labile et variables limnologiques. Dans les 49 lacs, le modèle Random Forest a donné un bon résultat pour le cobalt (60% de la variance expliquée) et un très bon résultat pour le nickel (86% de la variance expliquée). Pour Ni les variables explicatives sont liées au mode d’occupation du sol : « Activités » (Equipements pour l’eau et l’assainissement, entrepôts logistiques, bureaux…) et « Décharge ». Ce résultat est en accord avec le cas du lac de Créteil où la concentration en Ni dissous est très élevée et où les catégories d’occupation du sol « Activités » et « Décharges » sont dominantes / Lakes have a particular influence on the water cycle in urban catchments. Thermal stratification and a longer water residence time in the lake boost the phytoplankton production. Most metals are naturally found in the environment in trace amounts. Trace metals are essential to growth and reproduction of organisms. However, some are also well known for their toxic effects on animals and humans. Total metal concentrations do not reflect their ecotoxicity that depends on their properties and speciation (particulate, dissolved: labile or bioavailable and inert fractions). Trace metals can be adsorbed to various components in aquatic systems including inorganic and organic ligands. The ability of metal binding to dissolved organic matter (DOM), in particular humic substances, has been largely studied. In urban lakes, the phytoplankton development can produce autochthonous DOM, non humic substances that can have the ability of metal binding.. But there are few studies about trace metal speciation in lake water column.The main objectives of this thesis are (1) to obtain a consistent database of trace metal concentrations in the water column of representative urban lakes; (2) to access their bioavailability through an adapted speciation technique; (3) to analyze the seasonal and spatial evolution of the metals and their speciation; (4) to study the potential impact of environmental variables, particularly of dissolved organic matter related to phytoplankton production on metal bioavailability and (5) to link the metal concentrations to the land use in the lake watershed.Our methodology is based on a dense field survey of the water bodies in addition to specific laboratory analysis. The research has been conducted on three study sites: Lake Créteil (France), Lake Pampulha (Brazil) and a panel of 49 peri-urban lakes (Ile de France). Lake Créteil is an urban lake impacted by anthropogenic pollution. It benefits of a large number of monitoring equipment, which allowed us to collect a part of the data set. In Lake Pampulha catchment, the anthropogenic pressure is high. Lake Pampulha has to face with many pollution point and non-point sources. The climate and limnological characteristics of the lakes are also very different. The panel of 49 lakes of Ile de France was sampled once during three successive summers (2011-2013); they provided us with a synoptic, representative data set of the regional metal contamination in a densely anthropized region.In order to explain the role of the environmental variables on the metal concentrations, we applied the Random Forest model on the Lake Pampulha dataset and on the 49 urban lake dataset with 2 specific objectives: (1) in Lake Pampulha, understanding the role of environmental variables on the trace metal labile concentration, considered as potentially bioavailable and (2) in the 49 lakes, understanding the relationship of the environmental variables, more particularly the watershed variables, on the dissolved metal concentrations. The analysis of the relationships between the trace metal speciation and the environmental variables provided the following key results of this thesis.In Lake Pampulha, around 80% of the variance of the labile cobalt is explained by some limnological variables: Chl a, O2, pH, and total phosphorus. For the other metals, the RF model did not succeed in explaining more than 50% of the relationships between the metals and the limnological variables.In the 49 urban lakes in Ile de France, the RF model gave a good result for Co (66% of explained variance) and very satisfying for Ni (86% of explained variance). For Ni, the best explanatory variables are landuse variables such as “activities” (facilities for water, sanitation and energy, logistical warehouses, shops, office…) and “landfill”. This result fits with Lake Creteil where dissolved Ni concentration is particularly high and where the “activities” and “landfill” landuse categories are the highest

Thesis (PhD (Process Engineering))--University of Stellenbosch, 2010. / Dissertation presented for the Degree of DOCTOR OF PHILOSOPHY (Extractive Metallurgical Engineering) in the Department of Process Engineering at the University of Stellenbosch / ENGLISH ABSTRACT: Fault diagnosis is an important component of process monitoring, relevant in the greater context of developing safer, cleaner and more cost efficient processes. Data-driven unsupervised (or feature extractive) approaches to fault diagnosis exploit the many measurements available on modern plants. Certain current unsupervised approaches are hampered by their linearity assumptions, motivating the investigation of nonlinear methods. The diversity of data structures also motivates the investigation of novel feature extraction methodologies in process monitoring. Random forests are recently proposed statistical inference tools, deriving their predictive accuracy from the nonlinear nature of their constituent decision tree members and the power of ensembles. Random forest committees provide more than just predictions; model information on data proximities can be exploited to provide random forest features. Variable importance measures show which variables are closely associated with a chosen response variable, while partial dependencies indicate the relation of important variables to said response variable. The purpose of this study was therefore to investigate the feasibility of a new unsupervised method based on random forests as a potentially viable contender in the process monitoring statistical tool family. The hypothesis investigated was that unsupervised process monitoring and fault diagnosis can be improved by using features extracted from data with random forests, with further interpretation of fault conditions aided by random forest tools. The experimental results presented in this work support this hypothesis. An initial study was performed to assess the quality of random forest features. Random forest features were shown to be generally difficult to interpret in terms of geometry present in the original variable space. Random forest mapping and demapping models were shown to be very accurate on training data, and to extrapolate weakly to unseen data that do not fall within regions populated by training data. Random forest feature extraction was applied to unsupervised fault diagnosis for process data, and compared to linear and nonlinear methods. Random forest results were comparable to existing techniques, with the majority of random forest detections due to variable reconstruction errors. Further investigation revealed that the residual detection success of random forests originates from the constrained responses and poor generalization artifacts of decision trees. Random forest variable importance measures and partial dependencies were incorporated in a visualization tool to allow for the interpretation of fault conditions. A dynamic change point detection application with random forests proved more successful than an existing principal component analysis-based approach, with the success of the random forest method again residing in reconstruction errors. The addition of random forest fault diagnosis and change point detection algorithms to a suite of abnormal event detection techniques is recommended. The distance-to-model diagnostic based on random forest mapping and demapping proved successful in this work, and the theoretical understanding gained supports the application of this method to further data sets. / AFRIKAANSE OPSOMMING: Foutdiagnose is ’n belangrike komponent van prosesmonitering, en is relevant binne die groter konteks van die ontwikkeling van veiliger, skoner en meer koste-effektiewe prosesse. Data-gedrewe toesigvrye of kenmerkekstraksie-benaderings tot foutdiagnose benut die vele metings wat op moderne prosesaanlegte beskikbaar is. Party van die huidige toesigvrye benaderings word deur aannames rakende liniariteit belemmer, wat as motivering dien om nie-liniêre metodes te ondersoek. Die diversiteit van datastrukture is ook verdere motivering vir ondersoek na nuwe kenmerkekstraksiemetodes in prosesmonitering. Lukrake-woude is ’n nuwe statistiese inferensie-tegniek, waarvan die akkuraatheid toegeskryf kan word aan die nie-liniêre aard van besluitnemingsboomlede en die bekwaamheid van ensembles. Lukrake-woudkomitees verskaf meer as net voorspellings; modelinligting oor datapuntnabyheid kan benut word om lukrakewoudkenmerke te verskaf. Metingbelangrikheidsaanduiers wys watter metings in ’n noue verhouding met ’n gekose uitsetveranderlike verkeer, terwyl parsiële afhanklikhede aandui wat die verhouding van ’n belangrike meting tot die gekose uitsetveranderlike is. Die doel van hierdie studie was dus om die uitvoerbaarheid van ’n nuwe toesigvrye metode vir prosesmonitering gebaseer op lukrake-woude te ondersoek. Die ondersoekte hipotese lui: toesigvrye prosesmonitering en foutdiagnose kan verbeter word deur kenmerke te gebruik wat met lukrake-woude geëkstraheer is, waar die verdere interpretasie van foutkondisies deur addisionele lukrake-woude-tegnieke bygestaan word. Eksperimentele resultate wat in hierdie werkstuk voorgelê is, ondersteun hierdie hipotese. ’n Intreestudie is gedoen om die gehalte van lukrake-woudkenmerke te assesseer. Daar is bevind dat dit moeilik is om lukrake-woudkenmerke in terme van die geometrie van die oorspronklike metingspasie te interpreteer. Verder is daar bevind dat lukrake-woudkartering en -dekartering baie akkuraat is vir opleidingsdata, maar dat dit swak ekstrapolasie-eienskappe toon vir ongesiene data wat in gebiede buite dié van die opleidingsdata val. Lukrake-woudkenmerkekstraksie is in toesigvrye-foutdiagnose vir gestadigde-toestandprosesse toegepas, en is met liniêre en nie-liniêre metodes vergelyk. Resultate met lukrake-woude is vergelykbaar met dié van bestaande metodes, en die meerderheid lukrake-woudopsporings is aan metingrekonstruksiefoute toe te skryf. Verdere ondersoek het getoon dat die sukses van res-opsporing op die beperkte uitsetwaardes en swak veralgemenende eienskappe van besluitnemingsbome berus. Lukrake-woude-metingbelangrikheidsaanduiers en parsiële afhanklikhede is ingelyf in ’n visualiseringstegniek wat vir die interpretasie van foutkondisies voorsiening maak. ’n Dinamiese aanwending van veranderingspuntopsporing met lukrake-woude is as meer suksesvol bewys as ’n bestaande metode gebaseer op hoofkomponentanalise. Die sukses van die lukrake-woudmetode is weereens aan rekonstruksie-reswaardes toe te skryf. ’n Voorstel wat na aanleiding van hierde studie gemaak is, is dat die lukrake-woudveranderingspunt- en foutopsporingsmetodes by ’n soortgelyke stel metodes gevoeg kan word. Daar is in hierdie werk bevind dat die afstand-vanaf-modeldiagnostiek gebaseer op lukrake-woudkartering en -dekartering suksesvol is vir foutopsporing. Die teoretiese begrippe wat ontsluier is, ondersteun die toepassing van hierdie metodes op verdere datastelle.

The Caledonian pinewoods are a habitat of crucial environmental and cultural importance, and the sole home of many rare species. However, they have seen steady decline in recent centuries, through the establishment of hunting estates and forestry plantations. A recent trend in management is the attempted transformation of existing plantations (dense communities with a regular spatial structure and low variance in size and age) towards a state mimicking the perceived natural condition, which has a lower density, irregular spatial pattern, high variance in size and age. This presents a problem for traditional forestry practices, which were conceived primarily with “even-aged” plantation populations in mind. The shift towards management of an uneven-aged structure requires a more in-depth consideration of individual trees’ lifecycles and their effect upon long-term population dynamics. In recent years, great advances in computational and mathematical models for spatially interacting populations have been made. However, certain complications have prevented them from being utilised to their full potential for the purposes of forest management. Forest communities are not only spatially structured; the size of each tree plays a role in its ability to acquire resources for growth and survival. Existing models of population dynamics are discussed, and their extension to incorporate both size- and spatially- structured interactions is presented. The key aspects of populations’ structural development are studied. Data from both plantation and semi-natural Scots Pine stands in Scotland allow parameterisation of a stochastic individual-based model, which in turn provides insights into the behaviour of real populations, and the importance of spatial effects and heterogeneity in individuals. A partial differential equation (moment) approximation to the stochastic model is presented. While this is analytically intractable, numerical integration and heuristic analysis of the equations enable clearer identification of the drivers of population structure. Many results are concordant with existing models of both qualitative forest stand development and theoretical dynamics of spatially-structured populations, while others are specific to joint size-space structure. This deeper understanding of the population dynamics allows robust recommendations for diverse uneven-aged stand management objectives to be made. Approaches to accelerating the transformation of plantation stands towards a “natural” state (using two key operations: thinning – removal of trees, and planting) are investigated. Finally, approaches to so-called “continuous cover forestry” – the practice of maintaining a quasi-natural state while also obtaining economic value from a forest – are also considered. In both cases, the model’s simplicity enables clearer conclusions than would be possible using other approaches.

Forests are an important component in the earth system. They cover nearly one third of the land surface, store about as much carbon as the entire atmosphere and host more than half of the planet’s biodiversity. Forests provide ecosystem services such as climate regulation and water cycling and they supply resources. However, forests are increasingly at risk worldwide, due to anthropogenic deforestation, degradation and climate change. Concepts for counteracting this development require abilities to monitor forests and predict possible future developments. Given the vast size of forest cover along with the variety of forest types, field measurements and experiments alone cannot provide the solution for this task. Remote sensing and forest modeling enable a broader and deeper understanding of the processes that shape our planet’s forests. Remote sensing from airborne and spaceborne platforms can provide detailed measurements of forest attributes ranging from landscape to global scale. The challenge is to interpret the measurements in an appropriate way and derive biophysical properties. This requires a good understanding of the interaction between radiation and the vegetation. Forest models are tools that synthesize our knowledge about processes, such as tree growth, competition, disturbances and mortality. They allow simulation experiments which go beyond the spatial and temporal scales of field experiments. In this thesis, several major challenges in forest ecology and remote sensing were addressed. The main variable of interest was forest biomass, as it is the most important variable for forest carbon mapping and for understanding the role of vegetation in the global carbon cycle. For the purpose of biomass estimation, remote sensing derived canopy height and structure measurements were combined with field data, forest simulations and remote sensing simulations. The goals were: 1) to integrate remote sensing measurements into a forest model; 2) to understand the effects of spatial scale and disturbances on biomass estimation using a variety of remote sensing metrics; 3) to develop approaches for quantifying biomass changes over time with remote sensing and 4) to overcome differences among forest types by considering several structural aspects in the biomass estimation function. In the first study, a light detection and ranging (lidar) simulator was developed and integrated in the forest model FORMIND. The model was parameterized for the tropical rainforest on Barro Colorado Island (BCI, Panama). The output of the lidar simulator was validated against real airborne lidar data from BCI. Undisturbed and disturbed forests were simulated with FORMIND to identify the most well suited lidar metric for biomass estimation. The objective hereby was to achieve a low normalized root mean squared error (nRMSE) over the entire range of forest structures caused by disturbances and succession. Results identified the mean top-of-canopy height (TCH) as the best lidar-derived predictor. The accuracy strongly depended on spatial scale and relative errors < 10% could be achieved if the spatial resolution of the produced biomass map was ≥ 100 m and the spatial resolution of the remote sensing input was ≤ 10 m. These results could provide guidance for biomass mapping efforts. In the second study, forest simulations were used to explore approaches for estimating changes in forest biomass over time based on observed changes in canopy height. In an ideal situation, remote sensing provides measurements of canopy height above ground which allows the estimation of biomass stocks and changes. However, this requires sensors which are able to detect canopy surface and terrain elevation, and some sensors can only detect the surface (e.g., X-band radar). In such cases, biomass change has to be estimated from height change using a direct relationship. Unfortunately, such a relationship is not constant for forests in different successional stages, which can lead to considerable biases in the estimates of biomass change. A solution to this problem was found, where missing information of canopy height was compensated by integrating metrics of canopy texture. Applying this improved approach enables estimations of biomass losses and gains after disturbances at 1-ha resolution. In mature forests with very small changes in height and biomass all tested approaches have limited capabilities, as was revealed by an application using TanDEM-X derived canopy height from BCI. In the third study, a general biomass estimation function, which links remote sensing-derived structure metrics to forest biomass, was developed. General in this context means that it can be applied in different forest types and different biomes. For this purpose a set of predictor metrics was explored, with each predictor representing one of the following structural aspects: mean canopy height, maximal possible canopy height, maximal possible stand density, vertical canopy structure and wood density. The derived general equation resulted in almost equally accurate biomass estimates across the five considered sites (nRMSE = 12.4%, R² = 0.74) as site-specific equations (nRMSE = 11.7%, R²= 0.77). The contributions of the predictors provide a better understanding of the variability in the height-to-biomass relationship observed across forest types. The thesis has laid foundations for a close link between remote sensing, forest modeling and forest inventories. Several ongoing projects carry this further, by 1) disentangling and quantifying the uncertainty in biomass remote sensing, 2) trying to predict forest productivity based on structure and 3) detecting single trees from lidar to be used as forest model input. These methods can in the future lead to an integrated forest monitoring and information system, which assimilates remote sensing measurements and produces predictions about forest development. Such tools are urgently needed to reduce the risks forests are facing worldwide.

We analyzed recent total suspended solids (TSS) data from 45 gages on 36 rivers throughout the state of Minnesota. Watersheds range from 32 to 14,600 km2 and represent a variety of distinct settings in terms of topography, land cover, and geologic history. Our study rivers exhibited three distinct patterns in the relationship between discharge and TSS: simple power functions, threshold power functions, and peaked or negative power functions. Differentiating rising and falling limb samples, we generated sediment rating curves (SRC) of form TSS = aQb, Q being normalized discharge. Rating parameters a and b describe the vertical offset and steepness of the relationships. We also used the fitted SRCs to estimate TSS values at low flows and to quantify event-scale hysteresis. In addition to quantifying the watershed-average topographic, climatic/hydrologic, geologic, soil and land cover conditions, we used high-resolution lidar topography data to characterize the near-channel environment upstream of gages. We used Random Forest statistical models to analyze the relationship between basin and channel features and the rating parameters. The models enabled us to identify morphometric variables that provided the greatest explanatory power and examine the direction, form, and strength of the partial dependence of the response variables on individual predictor variables. The models explained between 43% and 60% of the variance in the rating curve parameters and determined that Q-TSS relation steepness (exponent) was most related to near-channel morphological characteristics including near-channel local relief, channel gradient, and proportion of lakes along the channel network. Land use within the watershed explained most variation in the vertical offset (coefficient) of the SRCs and in TSS concentrations at low flows.

Much of the planned installation of wind turbines in Sweden will be located in the northern region, characterized by a lower population density so that problems related to sound pollution and visual acceptance are of lower concern. This area is generally distinguished by complex topography and the presence of forest, that significantly affects the wind characteristics, complicating their modelling and simulation. There are concerns about how good an industrial code can simulate a forest, a question of paramount importance in the planning of new onshore farms. As a first step, a sensitivity analysis was initially carried out to investigate the impact on the ow of different boundary conditions and cell discretization inside the forest for a 2D domain with a homogeneous forest. Subsequently, a comparative analysis between the industrial code WindSim and Large Eddy Simulation (LES) data from Segalini. et al. (2016) was performed with the same domain. Lastly, simulations for a real Swedish forest, Ryningsnäs, was conducted to compare a roughness map approach versus modelling the forest as a momentum sink and a turbulence source. All simulations were conducted for neutral stability conditions with the same domain size and refinement. The main conclusions from each part can be summarized as follows. (i) The results from the sensitivity analysis showed that discretization of cells in the vertical direction inside the forest displayed a correlation between an increasing number of cells and a decreased streamwise wind speed above the canopy. (ii) The validation with the LES data displayed good agreement in terms of both horizontal mean wind speed and turbulence intensity. (iii) In terms of horizontal wind speed for Ryningsnäs, forest modelling was prevailing for all wind directions, where the most accurate simulation was found by employing a constant forest force resistive constant (C2) equal to 0.05. All forest models overestimated the turbulence intensity, whereas the roughness map approaches underestimated it. Based solely on the simulations for Ryningsnäas, a correlation between lower streamwise wind speed and higher turbulence intensity can be deduced.

This thesis is concerned with developing methodologies that enable existing models to be effectively reused. Results of this thesis are presented in the framework of Quantitative Structural-Activity Relationship (QSAR) models, but their application is much more general. QSAR models relate chemical structures with their biological, chemical or environmental activity. There are many applications that offer an environment to build and store predictive models. Unfortunately, they do not provide advanced functionalities that allow for efficient model selection and for interpretation of model predictions for new data. This thesis aims to address these issues and proposes methodologies for dealing with three research problems: model governance (management), model identification (selection), and interpretation of model predictions. The combination of these methodologies can be employed to build more efficient systems for model reuse in QSAR modelling and other areas. The first part of this study investigates toxicity data and model formats and reviews some of the existing toxicity systems in the context of model development and reuse. Based on the findings of this review and the principles of data governance, a novel concept of model governance is defined. Model governance comprises model representation and model governance processes. These processes are designed and presented in the context of model management. As an application, minimum information requirements and an XML representation for QSAR models are proposed. Once a collection of validated, accepted and well annotated models is available within a model governance framework, they can be applied for new data. It may happen that there is more than one model available for the same endpoint. Which one to chose? The second part of this thesis proposes a theoretical framework and algorithms that enable automated identification of the most reliable model for new data from the collection of existing models. The main idea is based on partitioning of the search space into groups and assigning a single model to each group. The construction of this partitioning is difficult because it is a bi-criteria problem. The main contribution in this part is the application of Pareto points for the search space partition. The proposed methodology is applied to three endpoints in chemoinformatics and predictive toxicology. After having identified a model for the new data, we would like to know how the model obtained its prediction and how trustworthy it is. An interpretation of model predictions is straightforward for linear models thanks to the availability of model parameters and their statistical significance. For non linear models this information can be hidden inside the model structure. This thesis proposes an approach for interpretation of a random forest classification model. This approach allows for the determination of the influence (called feature contribution) of each variable on the model prediction for an individual data. In this part, there are three methods proposed that allow analysis of feature contributions. Such analysis might lead to the discovery of new patterns that represent a standard behaviour of the model and allow additional assessment of the model reliability for new data. The application of these methods to two standard benchmark datasets from the UCI machine learning repository shows a great potential of this methodology. The algorithm for calculating feature contributions has been implemented and is available as an R package called rfFC.