Sustainable water resource management requires understanding how hydrologic processes are impacted by environmental management and land-use decisions across multiple spatial and temporal scales. A key concept in this respect is hydrologic ecosystem services (HES), which are the water related ‘goods’ produced by the environment which are valuable to humans. This dissertation assesses a range of topics concerning HES in the Vouga basin (north-central Portugal), and their connection with land-cover and land-use practices. Specifically, the relationship between changes in forest and agricultural land-cover and management practices, and associated changes in HES were examined using a range of statistical and modeling approaches. To quantify the effects of different agricultural scenarios on both HES and potential stakeholders, the ‘Soil and Water Assessment Tool’ (SWAT) was utilized, in conjunction with economic assessment methods. The first research section (Section 6) of the dissertation assesses the trends in streamflow quantity and yield in the Águeda watershed (a sub-basin of the Vouga) over a 75-yr period which coincided with large-scale afforestation of Pinus pinaster and (later) Eucalyptus globulus. Counter to the findings from meta-analysis studies of the effect of forest change on water availability, this study did not detect statistically significant trends in streamflow. By contrast, these findings support the view that there are prerequisite climatic, pedological, and eco-physiological watershed conditions that are necessary to observe hydrologic impacts at the watershed scale (which are not present in the Águeda watershed). By contrast, the significant changes which were detected are related to baseflow, which correspond with different periods of afforestation, and may be attributable to the promotion of soil water repellency under the mature pine and eucalypt stands. In the second research section (Section 7), an assessment is carried out on the hydrologic and nitrate dynamics at the whole basin scale, using the SWAT model. This assessment indicated that there is a high degree of variability in nitrate export from the different parts of the basin, with the highest rates coming from the lower (agriculturally dominated portion) of the basin. The main flow pathways for nitrate export were found to be leaching from agricultural land-cover types, which consistently had the highest export for all land-use and pathways. These findings indicate that the water bodies at the highest risk of nitrate pollution in the Vouga basin are the groundwater aquifers. The final research section (Section 8) utilizes the SWAT model to examine how reduced rates of fertilizer inputs would affect nitrate leaching, crop yields, and agricultural profitability in the lower Vouga basin. This research found that reduced rates of fertilization would reduce the amount of leached nitrate substantially, but that this would also lead to a large decrease in crop yield and profitability. A large difference in the inefficiency (i.e. crop production vs. nitrate export) between different HRUs was found, which could provide a focus for potential management action. This research strongly indicates that such actions may be needed to reduce the negative impacts of this pollution on the value of the groundwater aquifers, and to avoid associated costs which are otherwise passed on to local water users (e.g. through higher water treatment costs). The overall findings of the dissertation highlight the importance of the upper (forested) basin as a drinking water supply area, given the prevalence of nitrate pollution in the lower basin. However, the historic afforestation in the Vouga basin has resulted in a reduction in baseflow, which is negative from a drinking water supply perspective. Therefore, while the forested uplands are beneficial from water quality standpoint (compared to intensive agriculture), they also have altered flow patterns in a manner which will reduce available supply. The findings from the upper basin contrast sharply with the lower basin, where there are potentially large negative HES impacts due to current agricultural practices. These practices will primarily impact groundwater aquifers, and therefore the water quality within the lower basin receive little benefit from the relatively high-quality water from the upper basin. This highlights the importance of considering the interconnectivity of HES across spatial scales, which will depend on the specific site characteristics of the river basin.

Areas cultivated with eucalyptus forests have increased in Southern Brazil, and the effects on hydrosedimentological processes are not well known particularly at the watershed scale and in watersheds where land use has been changing. The Soil and Water Assessment Tool (SWAT) model has been applied in a few forested watersheds to evaluate and predict effects of commercial forest cultivation or land use change in hydrological processes, whilst erosion processes has not been the main goal. The present study aimed (i) to evaluate the effects of eucalyptus plantation on hydrosedimentological processes, and to evaluate the efficiency and limitations of the Soil and Water Assessment Tool (SWAT) model to simulate streamflow and sediment yield (daily and monthly) in nested eucalyptus watersheds and in two paired watersheds with eucalyptus and grassland, and (ii) to identify the contribution of the bank channel, unpaved roads and eucalyptus stands sediment-sources for two different size fractions of sediment (< 0.063 mm and 0.063-2 mm) deposited along the drainage network, and suspended sediment collected in the outlet of small watersheds. SWAT was used to evaluate hydrological and erosion processes for two nested forest watersheds (watershed-0.98 km2 and sub-watershed-0.39 km2) and two paired watersheds in Pampa Biome, one with grassland (1.10 km2) and other with eucalyptus (0.83 km2). Measured streamflow and sediment yield at the watersheds outlets was used to evaluate model sensitivity to selected model parameters, and for calibration from 2009 to 2013 for nested watersheds, and from 2009 to 2013 for paired watersheds. Time series plots and standard statistical measures were used to verify model predictions. For nested watersheds, predicted monthly streamflow was very good during calibration for the watershed and for the sub-watershed, respectively. Predicted monthly sediment yield was very good and satisfactory for the watershed, and unsatisfactory for the sub-watershed. Simulations for daily time-scale were satisfactory to predict streamflow and sediment yield for the watershed. For the sub-watershed, predicted streamflow was satisfactory , but sediment yield was unsatisfactory . For paired watersheds, predicted monthly streamflow was good during calibration for the grassland watershed and for the eucalyptus watershed. Predicted monthly sediment yield was satisfactory for the grassland watershed, and very good and satisfactory for eucalyptus watershed. Daily simulations were satisfactory to predict streamflow in both grassland and eucalyptus watersheds. The results suggest that the SWAT model is a promising tool to evaluate hydrological processes in Brazilian watersheds. However, more field work with continuous monitoring is required and empirical equations of SWAT must change to better represent the processes that predominate in each watershed and to represent the hydrological and erosion processes in sub-daily time-scale and, especially, for small watersheds where the time of concentration is less than one day. Sediment-sources was evaluated using geochemical properties, determined by inductively coupled plasma mass spectrometry and X-ray fluorescence analyses, and the data used to calculate proportional contributions of sediment. Source contributions were determined at points spatially-distributed along the main channel of the watershed. Source determination for in-stream sites was done using samples collected in one spot to evaluate source-contribution of area upstream of this site of interest to indicate how different sources dominate at different downstream locations. To examine whether different size-fractions shared similar origins, two size fractions of both source and suspended samples including fine (<0.063 mm) and coarse (0.063 2 mm) particles were analyzed. Fingerprinting approach indicated particle-size and location of sources within a watershed are major factors affecting the measured contribution of sources for coarse and fine sediments. The closer a sediment sampling site is to a potential source, the more likely this sediment-source will dominate the sampled material, especially for coarse particles. Based on the classification analysis, the relative contribution of each source to eroded sediments was bank channel, i.e. from the stream network. Modeling results and sediment-source identification can provide an improved understanding effect of land use or change in land use on sediment yield in watershed scale, which are useful and may be used as a management tool of natural resources. / Áreas cultivadas com florestas de eucalipto têm aumentado no Sul do Brasil e os efeitos sobre os processos hidrossedimentológicos são pouco conhecidos, especialmente em escala de bacia hidrográfica e para áreas em que ocorreu mudança no uso do solo. O modelo Soil and Water Assesment Tool (SWAT) tem sido utilizado em bacias hidrográficas florestais para avaliar e simular o efeito do cultivo de florestas comerciais ou mudança de uso do solo sobre os processos hidrológicos, mas os processos erosivos não têm sido avaliados. O presente estudo objetivou (i) avaliar os efeitos do plantio de eucalipto sobre os processos hidrossedimentológicos e a eficiência e as limitações do SWAT em simular o escoamento superficial e produção de sedimentos (diários e mensais) em bacias hidrográficas embutidas, ocupadas com eucalipto, e em duas bacias hidrográficas pareadas ocupadas principalmente com eucalipto e com campo natural, e (ii) identificar a contribuição das fontes margem da rede de drenagem, as estradas e os povoamentos de eucalipto na geração de duas diferentes frações de sedimentos (< 0,063 mm e 0,063-2 mm) depositados ao longo da rede de drenagem, e de sedimentos em suspensão coletados no exutório de bacias embutidas. O SWAT foi utilizado para avaliar os processos hidrológicos e de erosão em duas bacias florestais embutidas (bacia-0,94 km2 e sub-bacia-0,39 km2) e em duas bacias hidrográficas pareadas no Bioma Pampa, uma com campo natural (1,10 km2) e outra com eucalipto (0,83 km2). A vazão e a produção de sedimentos medidas nas seções de monitoramento das bacias hidrográficas foram utilizadas para avaliar a sensibilidade dos parâmetros selecionados do modelo e para calibração para as bacias embutidas (2009-2013), e para as bacias pareadas (2009-2013). As séries temporais e os parâmetros estatísticos foram utilizados para avaliar o potencial preditivo do modelo. Para as bacias embutidas, a representação da vazão mensal foi muito boa durante a calibração para a bacia e para a sub-bacia, respectivamente. A representação da produção de sedimentos mensal foi muito boa e satisfatória para a bacia, e insatisfatória para a sub-bacia. As simulações na escala diária foram satisfatórias para a representação da vazão e da produção de sedimentos para a bacia. Para a sub-bacia, a representação da vazão foi satisfatória, mas da produção de sedimentos foi insatisfatória. Para as bacias hidrográficas pareadas, a representação da vazão mensal foi boa durante a calibração para a bacia com pastagem e para a bacia com eucalipto. A representação da produção de sedimentos mensal foi satisfatória para a bacia com pastagem, e muito boa e satisfatória para a bacia com eucalipto. As simulações diárias foram satisfatórias para a vazão em ambas as bacias pareadas. Os resultados sugerem que o modelo SWAT é uma ferramenta promissora para avaliar processos hidrológicos em bacias hidrográficas brasileiras. No entanto, fazem-se necessários mais tempo de monitoramento contínuo das variáveis hidrossedimentológicas e alteração de equações empíricas do SWAT para melhor representar processos erosivos que predominam em cada bacia e para representar os processos hidrológicos e erosivos em escala sub-diária, especialmente para pequenas bacias em que o tempo de concentração é menor do que um dia. As fontes de sedimentos foram avaliadas por meio do uso de propriedades geoquímicas, determinadas por análises de espectrometria de massas com plasma e de fluorescência de raios-X. A contribuição das fontes de sedimentos foi determinada em pontos espacialmente distribuídos ao longo do canal principal da bacia. Amostras de sedimento de um ponto de coleta foram utilizadas para avaliar a contribuição das fontes da área de contribuição à montante desse local de interesse e para indicar o a diferença da contribuição dessas fontes em diferentes localizações à jusante. Para avaliar se as diferentes frações de sedimentos tem origem semelhante, duas frações de tamanho de partículas das fontes e amostras de sedimento fino (< 0,063 mm) e grosso (0,063-2 mm) foram analisadas. A técnica fingerprinting indicou que o tamanho das partículas e localização das fontes dentro da bacia hidrográfica são os principais fatores que afetam a contribuição das fontes para a produção de sedimentos finos e grossos. As fontes mais próximas do local de amostragem são as potenciais fontes de sedimentos, sendo provável que essa fonte de sedimentos será dominante na amostra de sedimentos, principalmente para as frações grossas. Com base na análise de classificação, a contribuição relativa de cada fonte indicou que a principal fonte de sedimentos foi a margem do canal da rede de drenagem. Os resultados da modelagem e identificação das fontes de sedimentos contribuem para uma melhor compreensão do efeito do uso do solo ou da alteração no uso sobre a produção de sedimentos na escala de bacia hidrográfica, que são úteis e podem ser utilizados como ferramenta de gestão dos recursos naturais.

This research was aimed to evaluate the possibility of using satellite precipitation products (SPPs) and Numerical Weather Prediction (NWP) of precipitation for better hydrologic simulations and flood forecasting in the trans-boundary Chenab River Basin (CRB) in Pakistan. This research was divided into three parts. In the first part, two renowned SPPs, i.e., global precipitation mission (GPM) IMERG-F v6 and tropical rainfall measuring mission (TRMM) 3B42 v7, were incorporated in a semidistributed hydrological model, i.e., the soil and water assessment tool (SWAT), to assess the daily and monthly runoff pattern in Chenab River at the Marala Barrage gauging site in Pakistan. The results exhibit higher correlation between observed and simulated discharges at monthly timescale simulations rather than daily timescale simulations. Moreover, results show that IMERG-F is superior to 3B42 by indicating higher R2, higher Nash–Sutcliffe efficiency (NSE), and lower percent bias (PBIAS) at both monthly and daily timescale. In the second part, three latest half-hourly (HH) and daily (D) SPPs, i.e., 'IMERG-E', 'IMERGL', and 'IMERG-F', were evaluated for daily and monthly flow simulations in the SWAT model. The study revealed that monthly flow simulation performance is better than daily flow simulation in all sub-daily and daily SPPs-based models. Results depict that IMERGHHF and IMERG-DF yield the best performance among the other latency levels of SPPs. However, the IMERG-HHF based model has a reasonably higher daily correlation coefficient (R) and lower daily root mean square error (RMSE) than IMERG-DF. IMERG-HHF displays the lowest PBIAS for daily and monthly flow validations and it also represents relatively higher values of R2 and NSE than any other model for daily and monthly model validation. Moreover, the sub-daily IMERG based model outperformed the daily IMERG based model for all calibration and validation scenarios. IMERG-DL based model demonstrates poor performance among all of the SPPs, in daily and monthly flow validation, with low R2, low NSE, and high PBIAS. Additionally, the IMERG-HHE model outperformed IMERG-HHL. In the third and last part of this research, coupled hydro-meteorological precipitation information was used to forecast the 2016 flood event in the Chenab River Basin. The gaugecalibrated SPP, i.e., Global Satellite Mapping of Precipitation (GSMaP_Gauge), was selected to calibrate the Integrated Flood Analysis System (IFAS) model for the 2016 flood event. Precipitation from the Global Forecast System (GFS) NWP, with nine different lead times up to 4 days, was used in the calibrated IFAS model. This study revealed that the hydrologic simulations in IFAS, with global GFS forecasts, were unable to predict the flood peak for all lead times. Later, the Weather Research and Forecasting (WRF) model was used to downscale the precipitation forecasts with one-way and two-way nesting approaches. It was found in this study that the simulated hydrographs in the IFAS model, at different lead times, from the precipitation of two-way WRF nesting exhibited superior performance with the highest R2, NSE and the lowest PBIAS compared with one-way nesting. Moreover, it was concluded that the combination of GFS forecast and two-way WRF nesting can provide high-quality precipitation prediction to simulate flood hydrographs with a remarkable lead time of 96 h when applying coupled hydrometeorological flow simulation.

The economic and social losses of environmental disasters are increasingly higher. Floods are a main concern in many locations around the world. Preventive actions are urgent and necessary. This doctoral thesis addresses topics related to hydro-meteorological risks and water resources management. Its aim is to cope with hydrological risks and water resources management through a flooding risk index, complex watershed modeling, different calibration techniques, and ensemble streamflow forecasting. Specific assumptions and research questions are defined in each chapter of the thesis, and are mostly related to the 12,600 km2 Piracicaba watershed, in Southeast, Brazil. Chapter one has general introductions and explains how the thesis is organized. Chapter two brings an assessment and mapping of flooding risks. Chapter three reviews the watershed modelling topic, through applications of a selected watershed model (the Soil and Water Assessment Tool - SWAT) in Brazil. Chapter four proposes a good practice methodology for calibration of watershed models for different time-steps with available data, having hydrology as main focus. Chapter five explores different methodologies for calibrating hydrological models, using two optimization algorithms and with a multi-site and single site approaches to evaluate related changes in performance. Chapter six has complex watershed modeling for sub-daily time-step, with an automatic hourly calibration module that was included in SWAT-CUP and the application of these models to forecast ensemble streamflow and with a data assimilation approach with optimization to improve the quality of the forecasts. Chapter seven has overall conclusions and chapter eight has a summarized list of other activities developed during the doctoral process. Overall we believe the methodologies and results for the Piracicaba watershed are very good. And that they can be replicated in other watersheds in Brazil and around the world. The proposed mapping assessments of flooding vulnerability and risks can be applied for the entire Brazil, and could be used as a tool in water resources management and planning. The watershed model (SWAT) used on this doctoral thesis also proved to be a versatile and robust model, with several good example applications in Brazil, and in particular for the Piracicaba case study. The step by step calibration methodology, as well as the different calibrations performed can help other modelers on choosing where and how to calibrate their own models. For hourly application, this work is pioneer, in area scale and model used. The results for ensemble flow forecasting and data assimilation show a little of what can be performed with this kind of application, and that it can be a potential tool for real time applications in streamflow forecasting and early warning systems. We believe the lessons learnt in this thesis can improve and aid modeler and water resources managers worldwide. / Os prejuízos econômicos e sociais de desastres ambientais têm sido maiores. Inundações são uma das principais preocupações ao redor do mundo. Ações preventivas são urgentes e necessárias. Esta tese de doutorado aborda temas relacionados à gestão dos recursos hídricos e de riscos hidro-meteorológicos. Possui o objetivo de lidar com riscos hidrológicos através de índices de risco a inundações, modelagem complexa de bacias hidrográficas, diferentes técnicas de calibração, e previsão de vazões por conjunto. Pressupostos e objetivos específicos são definidos em cada capítulo da tese, e são na sua maioria relacionados à bacia hidrográfica do Rio Piracicaba (12.600 km2), Sudeste do Brasil. O Capítulo um traz as introduções gerais e explica a organização da tese. O capítulo dois desenvolve mapeamento de riscos a inundações. O capítulo três revisa o tópico de modelagem de bacias hidrográficas, através de aplicações de um modelo selecionado (Soil and Water Assessment Tool - SWAT) no Brasil. O quarto capítulo propõe uma metodologia de boas práticas para a calibração de modelos de bacias hidrográficas utilizando dados disponíveis, com foco principal na hidrologia. O capítulo cinco explora diferentes metodologias de calibração, utilizando dois algoritmos de otimização e abordagens de calibração em um local e demais locais para avaliar alterações relacionadas ao desempenho da modelagem. O capítulo seis trabalha com modelagem sub-diária, com um módulo de calibração horária automática, que foi incluído no SWAT-CUP, e aplicação destes modelos para previsão de vazões por conjunto e assimilação de dados com otimização, para melhorar a qualidade das previsões. O sétimo capítulo traz as conclusões gerais da tese e oitavo capítulo apresenta uma lista resumida de outras atividades desenvolvidas durante o doutorado. Acreditamos que as metodologias e resultados para a bacia hidrográfica Piracicaba são muito bons. E que podem ser replicados em outras bacias hidrográficas no Brasil e ao redor do mundo. O mapeamento de vulnerabilidade e riscos de inundação propostos pode ser aplicados para todo o Brasil, além de possuir potencial como uma ferramenta de planejamento. O modelo utilizado (SWAT) também provou ser versátil e robusto, com vários bons exemplos de aplicações no Brasil, e em especial para a Bacia do Piracicaba. A metodologia sistemática para calibração, bem como as diferentes calibrações executadas podem auxiliar outros modeladores a escolherem como calibrar seus próprios modelos. Este trabalho é pioneiro no tipo de aplicação horária apresentada. Os resultados de previsão por conjunto de vazões e de assimilação de dados mostram o potencial da metodologia para sistemas de previsão de vazões em tempo real e em sistemas de alerta antecipado. Nós acreditamos que as lições aprendidas nesta tese podem auxiliar modeladores e gestores de recursos hídricos ao redor do mundo.

Over the last decades, fast growing population along with urban and agricultural sprawl has drastically increased the pressure on water resources of the Federal District (DF), Brazil. Various socio-environmental problems, such as soil erosion, non-point source pollution, reservoir silting, and conflicts among water users evoked the need for more efficient and sustainable ways to use land and water. Due to the complexity of processes relevant at the scale of river basins, a prior analysis of impacts of certain land use and/or land management changes is only feasible by means of modeling. The Soil and Water Assessment Tool (SWAT) has been proven to be useful in this context, across the globe and for different environmental conditions. In this thesis, the SWAT model is utilized to evaluate the impact of Best Management Practices (BMPs) on catchment hydrology and sediment transport. However, model applications in tropical regions, such as the DF, are hampered by severe challenges, (i) the lack of input and control data in an adequate temporal and spatial resolution and (ii) model structural failures in representing processes under tropical conditions. The present (cumulative) thesis addresses these challenges in model simulations for two contrasting watersheds, which both are important sources of the DF’s drinking water supply, i.e. (i) the agriculture-dominated Pipiripau river basin where conflicting demands put immense pressure on the available water resources and (ii) the Santa Maria / Torto river basin, which is to large parts protected as national park and, thus, covered by native vegetation of the Cerrado biome. Perhaps one of the most challenging issues facing watershed modelers in tropical regions is the fact that rain gauge networks can usually not reflect the high spatio-temporal variability of mostly convective precipitation patterns. Therefore, an ensemble of different reasonable input precipitation data-sets was used to examine the uncertainty in parameterization and model output. Acceptable streamflow and sediment load predictions could be achieved for each input data-set. However, the best-fit parameter values varied widely across the ensemble. Due to its enhanced consideration of parameter uncertainty, this ensemble approach provides more robust predictions and hence is reasonable to be used also for scenario simulations. BMP scenarios for the Pipiripau River Basin revealed that erosion control constructions, such as terraces and small retention basins along roads (Barraginhas) are promising measures to reduce sediment loads (up to 40%) while maintaining streamflow. Tests for a multi-diverse crop rotation system, in contrast, showed a high vulnerability of the hydrologic system against any increase in irrigation. Considering the BMP implementation costs, it was possible to estimate cost-abatement curves, which can provide useful information for watershed managers, especially when BMPs are supported by Payments for Environmental Services as it is the case in the study area due to the program Produtor de Água. While for agricultural areas the model has proven to generate plausible results, the plant growth module of SWAT was found to be not suitable for simulating perennial tropical vegetation, such as Cerrado (savanna) or forest, which can also play a crucial role in river basin management. For temperate regions SWAT uses dormancy to terminate growing seasons of trees and perennials. However, there is no mechanism considered to reflect seasonality in the tropics, i.e. the phenological change between wet and dry season. Therefore, a soil moisture based approach was implemented into the plant growth module to trigger new growing cycles in the transition period from dry to wet season. The adapted model was successfully tested against LAI and ET time series derived from remote sensing products (MODIS). Since the proposed changes are process-based but also allow flexible model settings, the modified plant growth module can be seen as a fundamental improvement useful for future model application in the tropics. The present thesis shows insights into the workflow of a watershed model application in the semi-humid tropics – from input data processing and model setup over source code adaptation, model calibration and uncertainty analysis to its use for running scenarios. It depicts region-specific challenges but also provides practical solutions. Hence, this work might be seen as one further step toward robust and process-based model predictions to assist land and water resources management. / Starkes Bevölkerungswachstum, ungeplante Suburbanisierung und Landnutzungsänderungen (z.B. Intensivierung in der Landwirtschaft) verstärkten innerhalb der letzten Jahrzehnte zunehmend den Druck auf die Wasserressourcen des Bundesdistrikts Brasilien (zentralbrasilianisches Hochland), in dessen Mitte die junge Hauptstadt Brasília liegt. Damit verbundene negative Umweltauswirkungen, wie Bodenerosion, Stoff- und Sedimenteinträge in Fließgewässer und Talsperren sowie Konflikte zwischen den Wassernutzern erfordern daher dringend effektive und nachhaltige Lösungen im Land- und Wasserressourcen-management. Der Einfluss von möglichen zukünftigen Landnutzungs- und Bewirtschaftungsänderungen auf Wasserverfügbarkeit und -qualität hängt vom jeweiligen, oftmals sehr komplexen, landschaftsökologischen Prozessgefüge ab und kann nur mithilfe von prozessbasierten Simulationsmodellen quantitativ auf der Ebene von Einzugsgebieten abgeschätzt werden. Das “Soil and Water Assessment Tool” (SWAT) ist ein solches Modell. Es findet weltweite Anwendung für verschiedene Umweltbedingungen in Einzugsgebieten der Meso- bis Makroskala, um Landnutzungseffekte auf den Wasserhaushalt und den Transport von Nährstoffen, Pestiziden und Sedimenten zu prognostizieren. Seine Anwendung in tropischen Regionen, wie etwa in Zentralbrasilien, ist jedoch mit erheblichen Herausforderungen verbunden. Das betrifft sowohl die Verfügbarkeit von Eingangs- und Referenzdaten in ausreichender raum-zeitlicher Auflösung, als auch modellstrukturelle Unzulänglichkeiten bei der Prozessabbildung. Die vorliegende kumulative Dissertation zeigt dies anhand von Modellanwendungen für zwei unterschiedliche wasserwirtschaftlich relevante Einzugsgebiete (EZG): Das landwirtschaftlich intensiv genutzte EZG des Rio Pipiripau mit aktuell besonders konfliktträchtiger Wassernutzung, und das Santa Maria/Torto-EZG, welches - geschützt als Nationalpark - durch größtenteils natürliche Vegetationsformationen der brasilianischen Savanne (Cerrado) gekennzeichnet ist. Eine der größten Herausforderungen für die Einzugsgebietsmodellierung in tropischen Regionen liegt in der Abschätzung des Gebietsniederschlages, da vorhandene Messstationsdichten oft nicht ausreichen, um die hohe räumliche und zeitliche Variabilität der meist konvektiven Niederschläge zu erfassen. Mithilfe eines Ensembles verschiedener, plausibel generierter Niederschlagsreihen ist der Einfluss von Niederschlagsdaten-Unsicherheit auf die Modellparametrisierung und -vorhersage explizit berücksichtigt und untersucht worden. Zufriedenstellende Abfluss- und Sedimentfrachtsimulationen waren mit jeder der als Modelinput verwendeten Niederschlagsreihen möglich, jedoch nur bei entsprechender, z.T. stark voneinander abweichender Einstellung der Kalibrierungsparameter. Da diese umfassendere Betrachtung von Parameterunsicherheit zu robusteren Modellvorhersagen führt, wurde der Ensemble-Ansatz auch in der Simulation von Bewirtschaftungsszenarien, dem eigentlichen Modellzweck, verwendet. Die Szenariosimulationen zeigten, dass Maßnahmen zur Erosionsvermeidung (Terrassierung) und zum Sedimentrückhalt (kleine Sedimentrückhaltebecken entlang von Straßen - Barraginhas) die Sedimentfracht des Rio Pipiripau durchschnittlich um bis zu 40% reduzieren können, ohne dabei die Wasserverfügbarkeit zu beeinträchtigen. Modellszenarien mit einer vielgliedrigen Fruchtfolge auf großer Fläche verdeutlichten dagegen die hohe Vulnerabilität des Niedrigwasserabflusses in der Trockenzeit gegenüber jedweder Erhöhung der Bewässerungsmenge. Auf Grundlage von Kostenschätzungen für einzelne Maßnahmen konnten Kostenkurven zur Verringerung der Sedimentfracht und damit nützliche Informationen für das Wasserressourcen-Management abgeleitet werden, insbesondere weil eine Auswahl solcher Agrar-Umweltmaßnahmen im Pipiripau-EZG durch das Programm Produtor de Água finanziell gefördert werden sollen. Während das Modell in landwirtschaftlich genutzten Gebieten plausible Ergebnisse produzierte, wurden erhebliche Schwachstellen in der Simulation ausdauernder Vegetation (z.B. Cerrado) identifiziert. Zur Unterbrechung jährlicher Vegetationszyklen verwendet SWAT eine tageslängenabhängige Dormanzperiode. Diese ist zwar zweckmäßig zur Abbildung der Vegetationsdynamik in den gemäßigten Breiten, steuert aber nicht tropische Vegetationszyklen. Um den Wechsel zwischen Trocken- und Regenzeit in der pflanzenphänologischen Simulation in SWAT abzubilden, wurde daher im Rahmen dieser Arbeit das Pflanzenwachstumsmodul modifiziert, und zwar unter anderem durch Einbeziehung der simulierten Bodenfeuchte zur Unterbrechung der Wachstumszyklen. Das angepasste Modul wurde erfolgreich anhand von Fernerkundungsdaten (MODIS) zum zeitlichen Verlauf von Blattflächenindex und Evapotranspiration getestet. Es ist prozessbasiert und erlaubt flexible Einstellungen, so dass es als grundlegende Modellverbesserung auch für andere SWAT-Anwender von großem Nutzen sein kann. Die vorliegende Dissertation bringt neue Einsichten in verschiedene wichtige Aspekte der integrierten Modellierung tropischer Einzugsgebiete, von der Eingangsdatenaufbereitung über Quellcode-Anpassung, Modellkalibrierung und Unsicherheitsanalyse bis hin zu Szenariosimulationen. Sie veranschaulicht regionsspezifische Herausforderungen, liefert gleichzeitig aber auch praktikable Lösungen und damit einen wichtigen Beitrag für robustere prozessbasierte Modellanwendungen als Entscheidungsunterstützung im Bereich Land- und Wasserressourcenmanagement.

The economic and social losses of environmental disasters are increasingly higher. Floods are a main concern in many locations around the world. Preventive actions are urgent and necessary. This doctoral thesis addresses topics related to hydro-meteorological risks and water resources management. Its aim is to cope with hydrological risks and water resources management through a flooding risk index, complex watershed modeling, different calibration techniques, and ensemble streamflow forecasting. Specific assumptions and research questions are defined in each chapter of the thesis, and are mostly related to the 12,600 km2 Piracicaba watershed, in Southeast, Brazil. Chapter one has general introductions and explains how the thesis is organized. Chapter two brings an assessment and mapping of flooding risks. Chapter three reviews the watershed modelling topic, through applications of a selected watershed model (the Soil and Water Assessment Tool - SWAT) in Brazil. Chapter four proposes a good practice methodology for calibration of watershed models for different time-steps with available data, having hydrology as main focus. Chapter five explores different methodologies for calibrating hydrological models, using two optimization algorithms and with a multi-site and single site approaches to evaluate related changes in performance. Chapter six has complex watershed modeling for sub-daily time-step, with an automatic hourly calibration module that was included in SWAT-CUP and the application of these models to forecast ensemble streamflow and with a data assimilation approach with optimization to improve the quality of the forecasts. Chapter seven has overall conclusions and chapter eight has a summarized list of other activities developed during the doctoral process. Overall we believe the methodologies and results for the Piracicaba watershed are very good. And that they can be replicated in other watersheds in Brazil and around the world. The proposed mapping assessments of flooding vulnerability and risks can be applied for the entire Brazil, and could be used as a tool in water resources management and planning. The watershed model (SWAT) used on this doctoral thesis also proved to be a versatile and robust model, with several good example applications in Brazil, and in particular for the Piracicaba case study. The step by step calibration methodology, as well as the different calibrations performed can help other modelers on choosing where and how to calibrate their own models. For hourly application, this work is pioneer, in area scale and model used. The results for ensemble flow forecasting and data assimilation show a little of what can be performed with this kind of application, and that it can be a potential tool for real time applications in streamflow forecasting and early warning systems. We believe the lessons learnt in this thesis can improve and aid modeler and water resources managers worldwide. / Os prejuízos econômicos e sociais de desastres ambientais têm sido maiores. Inundações são uma das principais preocupações ao redor do mundo. Ações preventivas são urgentes e necessárias. Esta tese de doutorado aborda temas relacionados à gestão dos recursos hídricos e de riscos hidro-meteorológicos. Possui o objetivo de lidar com riscos hidrológicos através de índices de risco a inundações, modelagem complexa de bacias hidrográficas, diferentes técnicas de calibração, e previsão de vazões por conjunto. Pressupostos e objetivos específicos são definidos em cada capítulo da tese, e são na sua maioria relacionados à bacia hidrográfica do Rio Piracicaba (12.600 km2), Sudeste do Brasil. O Capítulo um traz as introduções gerais e explica a organização da tese. O capítulo dois desenvolve mapeamento de riscos a inundações. O capítulo três revisa o tópico de modelagem de bacias hidrográficas, através de aplicações de um modelo selecionado (Soil and Water Assessment Tool - SWAT) no Brasil. O quarto capítulo propõe uma metodologia de boas práticas para a calibração de modelos de bacias hidrográficas utilizando dados disponíveis, com foco principal na hidrologia. O capítulo cinco explora diferentes metodologias de calibração, utilizando dois algoritmos de otimização e abordagens de calibração em um local e demais locais para avaliar alterações relacionadas ao desempenho da modelagem. O capítulo seis trabalha com modelagem sub-diária, com um módulo de calibração horária automática, que foi incluído no SWAT-CUP, e aplicação destes modelos para previsão de vazões por conjunto e assimilação de dados com otimização, para melhorar a qualidade das previsões. O sétimo capítulo traz as conclusões gerais da tese e oitavo capítulo apresenta uma lista resumida de outras atividades desenvolvidas durante o doutorado. Acreditamos que as metodologias e resultados para a bacia hidrográfica Piracicaba são muito bons. E que podem ser replicados em outras bacias hidrográficas no Brasil e ao redor do mundo. O mapeamento de vulnerabilidade e riscos de inundação propostos pode ser aplicados para todo o Brasil, além de possuir potencial como uma ferramenta de planejamento. O modelo utilizado (SWAT) também provou ser versátil e robusto, com vários bons exemplos de aplicações no Brasil, e em especial para a Bacia do Piracicaba. A metodologia sistemática para calibração, bem como as diferentes calibrações executadas podem auxiliar outros modeladores a escolherem como calibrar seus próprios modelos. Este trabalho é pioneiro no tipo de aplicação horária apresentada. Os resultados de previsão por conjunto de vazões e de assimilação de dados mostram o potencial da metodologia para sistemas de previsão de vazões em tempo real e em sistemas de alerta antecipado. Nós acreditamos que as lições aprendidas nesta tese podem auxiliar modeladores e gestores de recursos hídricos ao redor do mundo.

Over the last decades, fast growing population along with urban and agricultural sprawl has drastically increased the pressure on water resources of the Federal District (DF), Brazil. Various socio-environmental problems, such as soil erosion, non-point source pollution, reservoir silting, and conflicts among water users evoked the need for more efficient and sustainable ways to use land and water. Due to the complexity of processes relevant at the scale of river basins, a prior analysis of impacts of certain land use and/or land management changes is only feasible by means of modeling. The Soil and Water Assessment Tool (SWAT) has been proven to be useful in this context, across the globe and for different environmental conditions. In this thesis, the SWAT model is utilized to evaluate the impact of Best Management Practices (BMPs) on catchment hydrology and sediment transport. However, model applications in tropical regions, such as the DF, are hampered by severe challenges, (i) the lack of input and control data in an adequate temporal and spatial resolution and (ii) model structural failures in representing processes under tropical conditions. The present (cumulative) thesis addresses these challenges in model simulations for two contrasting watersheds, which both are important sources of the DF’s drinking water supply, i.e. (i) the agriculture-dominated Pipiripau river basin where conflicting demands put immense pressure on the available water resources and (ii) the Santa Maria / Torto river basin, which is to large parts protected as national park and, thus, covered by native vegetation of the Cerrado biome. Perhaps one of the most challenging issues facing watershed modelers in tropical regions is the fact that rain gauge networks can usually not reflect the high spatio-temporal variability of mostly convective precipitation patterns. Therefore, an ensemble of different reasonable input precipitation data-sets was used to examine the uncertainty in parameterization and model output. Acceptable streamflow and sediment load predictions could be achieved for each input data-set. However, the best-fit parameter values varied widely across the ensemble. Due to its enhanced consideration of parameter uncertainty, this ensemble approach provides more robust predictions and hence is reasonable to be used also for scenario simulations. BMP scenarios for the Pipiripau River Basin revealed that erosion control constructions, such as terraces and small retention basins along roads (Barraginhas) are promising measures to reduce sediment loads (up to 40%) while maintaining streamflow. Tests for a multi-diverse crop rotation system, in contrast, showed a high vulnerability of the hydrologic system against any increase in irrigation. Considering the BMP implementation costs, it was possible to estimate cost-abatement curves, which can provide useful information for watershed managers, especially when BMPs are supported by Payments for Environmental Services as it is the case in the study area due to the program Produtor de Água. While for agricultural areas the model has proven to generate plausible results, the plant growth module of SWAT was found to be not suitable for simulating perennial tropical vegetation, such as Cerrado (savanna) or forest, which can also play a crucial role in river basin management. For temperate regions SWAT uses dormancy to terminate growing seasons of trees and perennials. However, there is no mechanism considered to reflect seasonality in the tropics, i.e. the phenological change between wet and dry season. Therefore, a soil moisture based approach was implemented into the plant growth module to trigger new growing cycles in the transition period from dry to wet season. The adapted model was successfully tested against LAI and ET time series derived from remote sensing products (MODIS). Since the proposed changes are process-based but also allow flexible model settings, the modified plant growth module can be seen as a fundamental improvement useful for future model application in the tropics. The present thesis shows insights into the workflow of a watershed model application in the semi-humid tropics – from input data processing and model setup over source code adaptation, model calibration and uncertainty analysis to its use for running scenarios. It depicts region-specific challenges but also provides practical solutions. Hence, this work might be seen as one further step toward robust and process-based model predictions to assist land and water resources management. / Starkes Bevölkerungswachstum, ungeplante Suburbanisierung und Landnutzungsänderungen (z.B. Intensivierung in der Landwirtschaft) verstärkten innerhalb der letzten Jahrzehnte zunehmend den Druck auf die Wasserressourcen des Bundesdistrikts Brasilien (zentralbrasilianisches Hochland), in dessen Mitte die junge Hauptstadt Brasília liegt. Damit verbundene negative Umweltauswirkungen, wie Bodenerosion, Stoff- und Sedimenteinträge in Fließgewässer und Talsperren sowie Konflikte zwischen den Wassernutzern erfordern daher dringend effektive und nachhaltige Lösungen im Land- und Wasserressourcen-management. Der Einfluss von möglichen zukünftigen Landnutzungs- und Bewirtschaftungsänderungen auf Wasserverfügbarkeit und -qualität hängt vom jeweiligen, oftmals sehr komplexen, landschaftsökologischen Prozessgefüge ab und kann nur mithilfe von prozessbasierten Simulationsmodellen quantitativ auf der Ebene von Einzugsgebieten abgeschätzt werden. Das “Soil and Water Assessment Tool” (SWAT) ist ein solches Modell. Es findet weltweite Anwendung für verschiedene Umweltbedingungen in Einzugsgebieten der Meso- bis Makroskala, um Landnutzungseffekte auf den Wasserhaushalt und den Transport von Nährstoffen, Pestiziden und Sedimenten zu prognostizieren. Seine Anwendung in tropischen Regionen, wie etwa in Zentralbrasilien, ist jedoch mit erheblichen Herausforderungen verbunden. Das betrifft sowohl die Verfügbarkeit von Eingangs- und Referenzdaten in ausreichender raum-zeitlicher Auflösung, als auch modellstrukturelle Unzulänglichkeiten bei der Prozessabbildung. Die vorliegende kumulative Dissertation zeigt dies anhand von Modellanwendungen für zwei unterschiedliche wasserwirtschaftlich relevante Einzugsgebiete (EZG): Das landwirtschaftlich intensiv genutzte EZG des Rio Pipiripau mit aktuell besonders konfliktträchtiger Wassernutzung, und das Santa Maria/Torto-EZG, welches - geschützt als Nationalpark - durch größtenteils natürliche Vegetationsformationen der brasilianischen Savanne (Cerrado) gekennzeichnet ist. Eine der größten Herausforderungen für die Einzugsgebietsmodellierung in tropischen Regionen liegt in der Abschätzung des Gebietsniederschlages, da vorhandene Messstationsdichten oft nicht ausreichen, um die hohe räumliche und zeitliche Variabilität der meist konvektiven Niederschläge zu erfassen. Mithilfe eines Ensembles verschiedener, plausibel generierter Niederschlagsreihen ist der Einfluss von Niederschlagsdaten-Unsicherheit auf die Modellparametrisierung und -vorhersage explizit berücksichtigt und untersucht worden. Zufriedenstellende Abfluss- und Sedimentfrachtsimulationen waren mit jeder der als Modelinput verwendeten Niederschlagsreihen möglich, jedoch nur bei entsprechender, z.T. stark voneinander abweichender Einstellung der Kalibrierungsparameter. Da diese umfassendere Betrachtung von Parameterunsicherheit zu robusteren Modellvorhersagen führt, wurde der Ensemble-Ansatz auch in der Simulation von Bewirtschaftungsszenarien, dem eigentlichen Modellzweck, verwendet. Die Szenariosimulationen zeigten, dass Maßnahmen zur Erosionsvermeidung (Terrassierung) und zum Sedimentrückhalt (kleine Sedimentrückhaltebecken entlang von Straßen - Barraginhas) die Sedimentfracht des Rio Pipiripau durchschnittlich um bis zu 40% reduzieren können, ohne dabei die Wasserverfügbarkeit zu beeinträchtigen. Modellszenarien mit einer vielgliedrigen Fruchtfolge auf großer Fläche verdeutlichten dagegen die hohe Vulnerabilität des Niedrigwasserabflusses in der Trockenzeit gegenüber jedweder Erhöhung der Bewässerungsmenge. Auf Grundlage von Kostenschätzungen für einzelne Maßnahmen konnten Kostenkurven zur Verringerung der Sedimentfracht und damit nützliche Informationen für das Wasserressourcen-Management abgeleitet werden, insbesondere weil eine Auswahl solcher Agrar-Umweltmaßnahmen im Pipiripau-EZG durch das Programm Produtor de Água finanziell gefördert werden sollen. Während das Modell in landwirtschaftlich genutzten Gebieten plausible Ergebnisse produzierte, wurden erhebliche Schwachstellen in der Simulation ausdauernder Vegetation (z.B. Cerrado) identifiziert. Zur Unterbrechung jährlicher Vegetationszyklen verwendet SWAT eine tageslängenabhängige Dormanzperiode. Diese ist zwar zweckmäßig zur Abbildung der Vegetationsdynamik in den gemäßigten Breiten, steuert aber nicht tropische Vegetationszyklen. Um den Wechsel zwischen Trocken- und Regenzeit in der pflanzenphänologischen Simulation in SWAT abzubilden, wurde daher im Rahmen dieser Arbeit das Pflanzenwachstumsmodul modifiziert, und zwar unter anderem durch Einbeziehung der simulierten Bodenfeuchte zur Unterbrechung der Wachstumszyklen. Das angepasste Modul wurde erfolgreich anhand von Fernerkundungsdaten (MODIS) zum zeitlichen Verlauf von Blattflächenindex und Evapotranspiration getestet. Es ist prozessbasiert und erlaubt flexible Einstellungen, so dass es als grundlegende Modellverbesserung auch für andere SWAT-Anwender von großem Nutzen sein kann. Die vorliegende Dissertation bringt neue Einsichten in verschiedene wichtige Aspekte der integrierten Modellierung tropischer Einzugsgebiete, von der Eingangsdatenaufbereitung über Quellcode-Anpassung, Modellkalibrierung und Unsicherheitsanalyse bis hin zu Szenariosimulationen. Sie veranschaulicht regionsspezifische Herausforderungen, liefert gleichzeitig aber auch praktikable Lösungen und damit einen wichtigen Beitrag für robustere prozessbasierte Modellanwendungen als Entscheidungsunterstützung im Bereich Land- und Wasserressourcenmanagement.

<p>Current watershed-scale, nonpoint source (NPS) pollution models do not represent the processes and impacts of agricultural best management practices (BMP) on water quality with sufficient detail. To begin addressing this gap, a novel process-based, watershed-scale, water quality model (WEPP-WQ) was developed based on the Water Erosion Prediction Project (WEPP) and the Soil and Water Assessment Tool (SWAT) models. The proposed model was validated at both hillslope and watershed scales for runoff, sediment, and both soluble and particulate forms of nitrogen and phosphorus. WEPP-WQ is now one of only two models which simulates BMP impacts on water quality in ‘high’ detail, and it is the only one not based on USLE sediment predictions. Model validations indicated that particulate nutrient predictions were better than soluble nutrient predictions for both nitrogen and phosphorus. Predictions of uniform conditions outperformed nonuniform conditions, and calibrated model simulations performed better than uncalibrated model simulations. Applications of these kinds of models in real-world, historical simulations are often limited by a lack of field-scale agricultural management inputs. Therefore, a prototype tool was developed to derive management inputs for hydrologic models from remotely sensed imagery at field-scale resolution. At present, only predictions of crop, cover crop, and tillage practice inference are supported and were validated at annual and average annual time intervals based on data availability for the various management endpoints. Extraction model training and validation were substantially limited by relatively small field areas in the observed management dataset. Both of these efforts contribute to computational modeling research and applications pertaining to agricultural systems and their impacts on the environment.</p>