O objetivo deste trabalho foi propor uma metodologia para investigar os possíveis efeitos das mudanças climáticas na disponibilidade hídrica e as alterações da demanda hídrica no futuro, através de uma abordagem estocástica, que considera projeções climáticas, hidrológicas, populacionais e agropecuárias. O trabalho foi aplicado à bacia hidrográfica do rio Ijuí, noroeste do Rio Grande do Sul, Brasil. A metodologia foi composta por cinco módulos, envolvendo: i) a modelagem hidrológica para transformação da precipitação e evapotranspiração em vazão, utilizando Redes Neurais Artificiais (RNAs), ii) a correção e análise dos cenários climáticos para o futuro, oriundos do modelo Eta CPTEC/HadCM3, iii) a modelagem estocástica das vazões mensais no futuro, iv) a modelagem estocástica das variáveis populacionais e agropecuárias para geração de séries de demanda hídrica no futuro e, v) a simulação do balanço hídrico para geração de curvas de regularização, objetivando uma análise da relação entre a disponibilidade e a demanda hídrica no futuro. Como resultados relacionados ao processo de modelagem hidrológica, destaca-se que o modelo com RNAs mais adequado para a simulação das vazões mensais apresentou apenas três variáveis de entrada, obtendo um coeficiente de Nash-Sutcliffe igual a 0,904. Através da análise de sensibilidade, foi observado que a RNA escolhida relacionou corretamente as variáveis de entrada com a saída da rede, respeitando os princípios físicos envolvidos no sistema hidrológico. Quanto à análise dos cenários climáticos e vazões resultantes do processo de modelagem hidrológica, as diferenças entre os valores simulados com base no modelo Eta e os valores observados, no período de avaliação dos modelos (1976-1990), atingiram erros algumas vezes superiores a 20%. A vazão média de longo período, por exemplo, apresentou uma alteração de 141,6 m³/s (1961-1990) para 200,3 m³/s (2011-2040). Também foi observado um incremento na vazão média e no desvio padrão mensal entre os meses de janeiro e outubro. Entre os meses de fevereiro e junho, o percentual de aumento na vazão média mensal foi mais acentuado, superando o índice de 100%. Considerando-se os intervalos de confiança das estimativas de vazão para o futuro, pode-se concluir que existe uma tendência de aumento na variabilidade hidrológica no período entre 2011 e 2040, o que indica a possibilidade de ocorrência de séries temporais com períodos mais acentuados de estiagem e de cheias. Quanto às alterações na demanda hídrica, foi constatado que a tendência de crescimento das atividades agrícolas irrigadas no período analisado é bem superior à tendência observada em relação à criação animal e ao abastecimento humano. Mantida a tendência e os resíduos modelados entre 2003 e 2010, a média das séries estocásticas geradas para o futuro indica que haverá 1.954 km² de áreas irrigadas em 2040, fazendo com que a demanda aumente de 6,3 m³/s (2011) para 28 m³/s (2040), no mês de maior demanda hídrica (janeiro). Na etapa final, ao calcular a razão entre a demanda para usos consuntivos e a disponibilidade hídrica (demanda/disponibilidade), no período entre 2011 e 2040, foi observada uma tendência de aumento neste percentual ao longo dos anos. Em média, a relação demanda/disponibilidade em 2011 foi de apenas 6,06%, variando entre 0,81% (maio) e 20,15% (dezembro). Já em 2040 esta proporção aumentou para 13,82%, variando entre 1,09% (maio) e 43,3% (dezembro). Quanto às mudanças nas curvas de regularização obtidas através da simulação do balanço hídrico em um reservatório fictício, os resultados atestam que, em caso de confirmação do cenário de mudança climática utilizado, haverá a necessidade de reservatórios com capacidade cada vez maior para atender à demanda para usos consuntivos, em virtude do agravamento das estiagens no início do verão. / The purpose of this study was to propose a methodology to investigate the possible effects of climate change on water availability and changes in water requirement in the future, through a stochastic approach that considers climate, hydrological, agricultural and population projections. The method was applied to Ijuí river basin, northwest of Rio Grande do Sul, Brazil. The methodology consisted of five modules, involving: i) hydrological modeling of monthly flows using Artificial Neural Networks (ANNs), ii) correction and analysis of climate scenarios for the future, derived from the Eta model CPTEC / HadCM3, iii) the stochastic modeling of monthly flows in the future, iv) the stochastic modeling of population and agricultural variables to generate water requirement series in the future and, v) the simulation of the water balance for the generation of curves regularization aiming an analysis of the relationship between water availability and water requirement in the future. Regarding the results of the hydrologic modeling, it is highlighted that ANN model more suitable model for the flow simulation presented only three input variables, obtaining a Nash-Sutcliffe coefficient equal to 0.904. It was observed, through sensitivity analysis, that the ANN related correctly chosen input variables with the output of the network, respecting the physical principles involved in the hydrological system. The analysis of climate scenarios and flows resulting from the hydrologic modeling process showed that the differences between the simulated values based on the Eta model and the observed values for the period of assessment models (1976-1990), errors sometimes reached more than 20 %. Therefore, one must consider that these uncertainties will be replicated in future scenarios, as to analysis of the effects of climate change on water availability. Overall, the results related to stochastic modeling of monthly flows for the future (2011-2040) showed a tendency to increase in flows. The average flow of long period, for example, introduced an amendment to 141.6 m³ / s (1961-1990) to 200.3 m³ / s (2011-2040). We observe an increase in the average flow and monthly standard deviation between January and October. The percentage increase in the monthly average flow was more pronounced between the months of February and June, exceeding the rate of 100%. Considering the confidence intervals of the estimates of flow for the future, it can be concluded that there is an increasing trend in hydrological variability in the period between 2011 and 2040, which indicates the possibility of time series with more severe periods of drought and flood. We found an increasing trend of irrigated agricultural activities above the trend towards livestock and human consumption. If the trend and residues modeled between 2003 and 2010 is maintained, irrigated areas in 2040 should reach 1,954 km², increasing water demand of 6.3 m³ / s (2011) to 28 m³ / s (2040), in the month of higher water demand (in January). The final step is to calculate the ratio between the demand for consumptive uses and water availability (demand / availability), we observe an increasing trend in the percentage in the period between 2011 and 2040. On average, the demand / availability in 2011 was only 6.06%, with values between 0.81% (May) to 20.15% (December). By 2040, this proportion increased to 13.82%, with values between 1.09% (May) to 43.3% (December). Finally, with regard to changes in the curves obtained for regularization by simulating the water balance in a fictitious reservoir, the results show that there is a need for reservoirs with increasing capacity to meet the demand for consumptive uses, upon confirmation of the scenario climate change used, because of worsening drought in early summer.

Kyoto University (京都大学) / 0048 / 新制・課程博士 / 博士(工学) / 甲第17261号 / 工博第3663号 / 新制||工||1557(附属図書館) / 30018 / 京都大学大学院工学研究科都市環境工学専攻 / (主査)教授 椎葉 充晴, 教授 寶 馨, 准教授 立川 康人 / 学位規則第4条第1項該当

The accuracy of numerical models analyzing hydrologic and hydraulic processes depends largely on how well the input terrain data represents the earth's surface. Modelers obtain terrain data for a study area by performing surveys or by gathering historical survey data. If a modeler desires to create a predictive model to simulate the addition of man-made features such as channels, embankments and pits, he must modify the terrain data to include these features. Doing this by hand is tedious and time consuming. In 2001 Christensen implemented a tool in the Surface-water Modeling System (SMS) software package for integrating man-made geometric features into surveyed terrain data. He called this process feature stamping. While Christensen's feature stamping algorithms decrease the time and effort required to integrate geometric features into existing terrain data, they only function on centerline-based features having a constant trapezoidal cross-section. In addition to placing geometric limitations on the features they stamp, Christensen's feature stamping algorithms also possess several instabilities. These instabilities arise when stamping features that leave the bounds of the terrain data, and when modifying and re-stamping features that have already been stamped. This thesis presents changes and enhancements made to Christensen's feature stamping algorithms. These changes and enhancements completely eliminate the instabilities found in Christensen's feature stamping algorithms and make it possible for numerical modelers to stamp more complex geometric features having compound slopes, asymmetric cross-sections and varying cross-sections along their length. Finally, additional feature stamping algorithms make it possible to stamp radial features such as mounds and pits.

No description available.

The hydrograph at the watershed outlet was simulated using the time-area curve concept implemented in a geographic information system (GIS). The goal of this study was to determine if hydrograph prediction accuracy would be improved by accounting for spatial and temporal variation of excess rainfall. Three models with different methods of estimating excess rainfall were developed: the Distributed Curve Number (DCN) model uses a CN for each cell, generating spatially distributed excess rainfall using the Soil Conservation Services curve number method (SCS, 1972); the Uniform Curve Number (UCN) model uses a single "average" CN for the whole watershed, thus generating a uniform excess rainfall; the Phi index model which uses the Phi-index method to generate uniform excess rainfall. With the aid of a GIS, the cumulative flow time to the watershed outlet is estimated for each cell in the watershed and the isochrones of equal travel time are developed. The time-area curve is developed in the form of an S curve. The spatially distributed 1-hr unit hydrograph is derived from the S curve as the difference between the S curve and its value lagged by 1-hr. The models used in this study describe the physical processes and flow mechanisms. They also reflect effects of watershed characteristics (slope, landuse, soil drainage potential) and excess rainfall intensity on the resulting hydrograph at the watershed outlet. Surface flow is divided into channel flow and overland flow based on the upstream drainage area. Flow is routed to the watershed outlet through a channel network derived from the watershed Digital Elevation Model (DEM). The models developed were tested against observed rainfall-runoff data from the 1153-ha Virginia Piedmont watershed (Owl Run). A total of 30 storms were simulated, with statistical comparison of peak flow rate, time to peak flow rate, and the hydrograph shape. The hydrograph shape was compared both visually and statistically. Results indicated that the two models which account for temporal variation in excess rainfall (DCN and UCN) predicted the output hydrograph much more accurately than the Phi model which lacks the ability to capture the temporal variation of excess rainfall. For this watershed, results showed that the spatial variability in excess rainfall which was accounted for by the DCN model did not improve the prediction accuracy over the UCN model which lacks that ability. However, a sensitivity analysis for the effect of the spatial distribution of the excess rainfall indicated that can be a significant effect of spatial distribution on the predicted hydrograph. / Master of Science

La cuenca del Río Guayas es considerada como la zona de mayor producción agrícolaganadera del Ecuador, por tanto su estudio es importante para tener una producción sustentable y sostenible en el tiempo. Su área abarca una superficie de 3 millones de ha, está conformada por zonas altas ubicadas sobre la Cordillera Occidental, zonas medias situadas en las laderas de la cordillera y zonas bajas que corresponde a las grandes llanuras aluviales de la costa donde se asienta la mayor concentración de población. Debido a la estructura morfológica de la cuenca y a los procesos antropogénicos que se desarrollaron sin tomar en cuenta planes conservacionistas adecuados en manejo de cuencas hidrográficas, las zonas bajas son las que mayormente se han visto afectadas durante años, por eventos de inundación frecuentes y prolongados, que generan grandes pérdidas materiales e incluso vidas humanas, afectando severamente su potencial económico y limitando el transporte fluvial. Siendo por este motivo, de fundamental importancia la determinación de la producción de caudales y sedimentos que se aportan al Río Guayas. El objetivo de esta tesis es la modelación hidrológica de la microcuenca del Río Toachi, tributario del Río Guayas, para cuantificar la producción de caudales líquidos y sólidos que se generan a partir de la caracterización morfológica e hidrometeorológica. Se aplicaron modelos hidrológicos y empíricos integrando los Sistemas de Información Geográfica (SIG) en el tratamiento de información espacial, generando una gran base de datos georreferenciada de la cuenca de estudio. Como resultado, se obtuvieron los parámetros morfológicos del relieve, de la forma y relativos a la red hidrográfica a nivel cuenca de estudio y sus unidades hidrológicas (subcuencas), destacando la forma alargada de la cuenca con pendientes fuertes en la cabecera, accidentadas en la zona media y suave en la desembocadura, con una superficie de 50.371,74 ha. Presenta valores de densidad de drenaje bajos con crecientes de menor magnitud y menor probabilidad de ocurrencia de eventos torrenciales, con un bajo poder erosivo. Se generaron los mapas temáticos como: topográfico, hidrográfico, orientación, iluminación y pendientes de las laderas, suelo, uso del suelo y cobertura vegetal, isotermas, isoyetas, zonas climáticas. Se realizó un estudio de los parámetros hidrometeorológicos y su dinámica en la cuenca de estudio a nivel zonas para cuantificar los caudales líquidos por medio del HECGeo-HMS. Se calibró y validó el modelo hidrológico HEC-HMS ® cuantificando el caudal simulado a la salida de la cuenca de estudio para que pueda ser usado en otras cuencas de iguales características morfométricas. Se calcularon los factores de la Ecuación Universal de Pérdida de Suelo (USLE) y la Tasa de Erosión Hídrica Superficial de la cuenca del Río Toachi y de las respectivas unidades hidrológicas. Además se realizó la cartografía particular de los parámetros de la USLE de la cuenca del Río Toachi con SIG para un análisis geoespacial. / Guayas River Basin is considered one of the greatest farming land and crop yield production areas of Ecuador. So this study is important to have a sustainable production through the time. It comprises 3 million hectares; which is compound by highlands located on the West mountain range, mountain passes located on the slopes and valleys that correspond to vast alluvial plains where most of the population is concentrated. Because of the basin morphological structure and the anthropogenic processes that have been developed regardless of preservation planning programs appropriate for hydrographic basin management, valleys have been mainly affected by frequent and long floods over years. These latter events have cost both damages and even losses of lives drastically altering the economic potential and limiting fluvial transportation. The fact of determining the water level and sediment production that are provided to Guayas River is, thus, of utmost importance. The aim of this thesis is the hydrologic modeling of Toachi River micro basin located within Guayas River basin to quantify liquid and solid flow production that stems from the morphological and hydrometeorological characterization. To this purpose, hydrological and empirical models were applied integrating the Geographic Information Systems (GIS) in the analysis of spatial data, thus, generating a great georreferencial database of the studied basin. As a result, morphological parameters of the relief, the shape and relating to the hydrographic network at a basin studied level and its hydrological units (sub-basin) were obtained. The elongated shape of the basin with steep slopes at the headwaters source, rugged in the passes and smooth in the mouth of the River with an extension of 50.371,74 ha. are highlighted. It presents low density drainage values with a growing water level of less scale and less probability of torrential events resulting in a low erosive power of the basin. Topic maps were generated like: topography, hydrography, orientation, illumination and hillside slopes, soil, soil use and vegetation cover, isotherms, isohyets, climate zones. Hydrometeorological parameters study was made and its dynamic in basin studied level to quantify the liquid flows by HECGeo-HMS. The hydrologic model HEC-HMS ® was calibrated and validated quantifying the water level simulated at the end of the studied basin in order to be used by other basins with similar morphometric characteristics. Universal Soil Loss Equation (USLE) and Superficial erosion rate factors of the Toachi River basin and its hydrologic units were calculated. And the particular cartography of the USLE parameters of the Toachi River basin with GIS for a geospatial analysis was made.

The purpose of this study is to determine if Low Impact Development (LID) can effectively mitigate flooding under projected climate scenarios. LID relies on runoff management measures that seek to control rainwater volume at the source by reducing imperviousness and retaining, infiltrating and reusing rainwater. An event-driven hydrologic/hydraulic model was developed to simulate how climate change, land use and LID scenarios may affect runoff response in the Bowker Creek watershed, a 10km2 urbanized catchment located in the area of greater Victoria, British Columbia, Canada. The first part of the study examined flood impacts for the 2050s (2040-2069) following the A2 emissions scenario. For the 24-hour, 25-year local design storm, results show that projected changes in rainfall intensity may increase flood extents by 21% to 50%. When combined with continued urbanization flood extents may increase by 50% to 72%. The second part of the study identified potential locations for three LID treatments (green roofs, rain gardens and top soil amendments) and simulated their effect on peak in-stream flow rates and flood volumes. Results indicate that full implementation of modeled LID treatments can alleviate the additional flooding that is associated with the median climate change projection for the 5-year, 10-year and 25-year rainfall events. For the projected 100-year event, the volume of overland flood flows is expected to increase by 1%. This compares favourably to the estimated 29% increase without LID. In term of individual performance, rain gardens had the greatest hydrologic effect during more frequent rainfall events; green roofs had minimal effect on runoff for all modelled events; and top soil amendments had the greatest effect during the heaviest rainfall events. The cumulative performance of LID practices depends on several variables including design specifications, level of implementation, location and site conditions. Antecedent soil moisture has a considerable influence on LID performance. The dynamic nature of soil moisture means that at times LID could meet the mitigation target and at other times it may only partially satisfy it. Future research should run continuous simulations using an appropriately long rainfall record to establish the probabilities of meeting performance requirements. In general, simulations suggest that if future heavy rainfall events follow the median climate change projection, then LID can be used to maintain or reduce flood hazard for rainfall events up to the 25-year return period. This study demonstrates that in a smaller urban watershed, LID can play an important role in reducing the flood impacts associated with climate change. / Graduate

Made available in DSpace on 2014-12-17T15:03:28Z (GMT). No. of bitstreams: 1 RaquelFT_DISSERT.pdf: 4757642 bytes, checksum: d1473228d1b65e6a6e6b97e612b8485f (MD5) Previous issue date: 2012-02-28 / The management of water resources in the river basin level, as it defines the Law n? 9433/97, requires the effective knowledge of the processes of hydrological basin, resulting from studies based on consistent series of hydrological data that reflect the characteristics of the basin. In this context, the objective of this work was to develop the modeling of catchment basin of the river Jundia? - RN and carry out the study of attenuation of a flood of the dam Tabatinga, by means of a monitoring project of hydrological data and climatology of the basin, with a view to promoting the development of research activities by applying methodologies unified and appropriate for the assessment of hydrological studies in the transition region of the semiarid and the forest zone on the coast of Rio Grande do Norte. For the study of the hydrological characteristics of the basin was conducted the automatic design of the basin of the river Jundia?, with the aid of programs of geoprocessing, was adopted a hydrological model daily, the NRCS, which is a model determined and concentrated. For the use of this model was necessary to determine some parameters that are used in this model, as the Curve Number. Having in mind that this is the first study that is being conducted in the basin with the employment of this model, it was made sensitivity analysis of the results of this model from the adoption of different values of CN, situated within a range appropriate to the conditions of use, occupation and the nature of the soil of this basin. As the objective of this study was also developing a simulation model of the operation of the Tabatinga dam and with this flood control caused in the city of Maca?ba, it was developed a mathematical model of fluid balance, developed to be used in Microsoft Excel. The simulation was conducted in two phases: the first step was promoted the water balance daily that allowed the analysis of the sensitivity of the model in relation to the volume of waiting, as well as the determination of the period of greatest discharges daily averages. From this point, it was assumed for the second stage, which was in the determination of the hydrograph of discharges effluent slots, that was determined by means of the fluid balance time, on the basis of the discharges effluents generated by a mathematical equation whose parameters were adjusted according to the hydrograph daily. Through the analyzes it was realized that the dam Tabatinga only has how to carry out the attenuation of floods through the regularization of the volume of waiting, with this there is a loss of approximately 56.5% on storage capacity of this dam, because for causing the attenuation effect of filled the shell of this dam has to remain more than 5m below the level of the sill, representing at least 50.582.927m3. The results obtained with the modeling represents a first step in the direction of improving the level of hydrological information about the behavior of the basins of the semiarid. In order to monitor quantitatively the hydrographic basin of the river Jundia? will be necessary to install a rain gauge register, next to the Tabatinga dam and a pressure transducer, for regular measurements of flow in the reservoir of the dam. The climatological data will be collected in full automatic weather station installed in Agricultural School Jundia? / A gest?o de recursos h?dricos em n?vel da bacia hidrogr?fica, como define a Lei n? 9433/97, exige o conhecimento efetivo dos processos hidrol?gicos da bacia, decorrentes de estudos baseados em s?ries consistentes de dados hidrol?gicos que espelhem as caracter?sticas da bacia. Neste contexto, este trabalho teve como objetivo desenvolver a modelagem hidrol?gica da bacia hidrogr?fica do rio Jundia? RN e realizar o estudo da atenua??o de cheia da barragem Tabatinga visando a implanta??o de bacia representativa e experimental, atrav?s de um projeto de monitoramento de dados hidrol?gicos e climatol?gicos da bacia, com o intuito de favorecer o desenvolvimento de atividades de pesquisa aplicando-se metodologias unificadas e apropriadas para a avalia??o de estudos hidrol?gicos na regi?o de transi??o do semi?rido e a zona da mata no litoral do Rio Grande do Norte. Para o estudo das caracter?sticas hidrol?gicas da bacia foi realizado o delineamento autom?tico da bacia hidrogr?fica do rio Jundia?, com o aux?lio de programas de geoprocessamento; adotou-se um modelo hidrol?gico di?rio, o NRCS, que ? um modelo determin?stico e concentrado. Para utiliza??o deste modelo foi necess?rio determinar alguns par?metros que s?o utilizados nesse modelo, como o Curva- N?mero. Tendo em vista que este estudo ? o primeiro que est? sendo realizado nesta bacia com o emprego deste modelo, foi feita an?lise de sensibilidade dos resultados deste modelo a partir da ado??o de diferentes valores de CN, situados dentro de uma faixa adequada ?s condi??es do uso, ocupa??o e natureza do solo desta bacia. Como o objetivo deste estudo foi tamb?m desenvolver um modelo de simula??o de opera??o da barragem Tabatinga e com isto controlar as inunda??es ocasionadas na cidade de Maca?ba, foi elaborado um modelo matem?tico de balan?o h?drico, desenvolvido para ser aplicado em planilha Microsoft Excel. A simula??o foi realizada em duas etapas: na primeira etapa promoveu-se o balan?o h?drico di?rio que permitiu analisar a sensibilidade do modelo em rela??o ao volume de espera, assim como a determina??o do per?odo de maiores vaz?es m?dias di?rias. A partir disso, partiu-se para a segunda etapa, que constituiu na determina??o do hidrograma das vaz?es efluentes hor?rias, que foi determinado atrav?s do balan?o h?drico hor?rio, tendo como base as vaz?es afluentes geradas por uma equa??o matem?tica cujos par?metros foram ajustados de acordo com o hidrograma di?rio. Atrav?s das an?lises percebeu-se que a barragem Tabatinga s? tem como realizar a atenua??o de cheias atrav?s da regulariza??o do volume de espera, com isso h? uma perda de aproximadamente 56,5% na capacidade de armazenamento desta barragem, pois para provocar o efeito de atenua??o de cheia o reservat?rio dessa barragem tem que permanecer mais de 5m abaixo do n?vel da soleira, representando no m?nimo 50.582.927 m3. Os resultados obtidos com a modelagem representam um primeiro passo na dire??o de melhorar o n?vel de informa??o hidrol?gica sobre o comportamento das bacias do semi?rido. Nesse intuito, para monitorar quantitativamente a bacia hidrogr?fica do rio Jundia? ser? necess?rio instalar um pluvi?grafo, pr?ximo a barragem Tabatinga e um lin?grafo de press?o, para medi??es regulares de vaz?o no reservat?rio da barragem. Os dados climatol?gicos ser?o coletados na esta??o meteorol?gica autom?tica completa instalada na Escola Agr?cola Jundia?

This dissertation describes a proposed methodology for Cumulative Effects Assessment (CEA) with the purpose of improving the process by making it both more substantive and quantitative. The general principles of the approach include the following: use of effect-based analyses where selected Valued Component (VC) sensitivities are identified first and then effect pathways are determined building bottom-up linkages from VC sensitivities to potential stressors or combinations of stressors to effect drivers and forces behind the drivers. Models were developed based on statistical or historic trend analysis or literature review that predicted the responses of the VCs to changes in effect drivers. Further, scenarios of divergent futures were created that involved different developments of each effect driver or force, and finally the models were applied to each scenario to project the state of the studied VCs. A practical implementation was conducted to demonstrate the use of the proposed methods on future population trends of two anadromous salmon species from the Cowichan River, British Columbia, Chinook and Coho. The assessment was conducted for both early freshwater and marine phases of their life. For the freshwater phase, the assessment focused on two main factors affecting salmon survival, streamflow and stream temperature and established two main drivers affecting these stressors, land use and climate change, and two main forces behind these drivers, Local and Global human development driven change, respectively. Effects of stream temperature and streamflow on salmon freshwater survival were simulated using two models; one was based on Chinook freshwater survival correlations with stream temperature and was developed only for Chinook, and the other was based on literature-derived temperature and streamflow thresholds and was developed for both species. Connections between the stressors (stream temperature and streamflow) and drivers (land use and climate change) were established through a hydrologic model and stream temperature regression model. For the marine environment, models were created using Pearson correlation and stepwise regression analysis examining links between survival of Cowichan River Chinook and Strait of Georgia hatchery-raised and wild Coho and various environmental variables of the nearshore zone of Strait of Georgia and Juan de Fuca Strait. The models were applied to project future salmon survival under four future scenarios for 2050 that were created by combining two opposite scenarios of land use in the watershed, forest conservation and development, and two climate change scenarios, extreme and moderate. Scenario projections showed a decrease in overall (combined early freshwater marine) survival by 2050 for all three studied salmon populations. None of them are likely to survive in scenarios with extreme climate change, while scenarios with moderate climate change showed positive survival rates although lower than present-day baseline levels. Analysis also showed that land use management within the Cowichan River watershed can also affect freshwater survival of both Chinook and Coho and marine survival of Chinook through influence of river discharge on nearshore processes. However, our land-use management scenarios have considerably weaker effect than climate change on salmon survival. Therefore, we conclude that land use management alone is not sufficient to offset effects of climate change on salmon survival. / Graduate

Streamflow forecasting is a fundamental component of various water resources management systems, ranging from flood control and mitigation to long-term planning of irrigation and hydropower systems. In the context of floods, a probabilistic forecasting system is required for proper and effective decision-making. Therefore, the primary goal of this research is the development of an advanced ensemble-based streamflow forecasting framework to better quantify the predictive uncertainty and generate enhanced probabilistic forecasts. This research started by comprehensively evaluating the performances of various lumped conceptual models in data-poor watersheds and comparing various Bayesian Model Averaging (BMA) modifications for probabilistic streamflow simulation. Then, using the concept of BMA, two novel probabilistic post-processing approaches were developed to enhance streamflow forecasting performance. The combination of the entropy theory and the BMA method leads to an entropy-based Bayesian Model Averaging (En-BMA) approach for enhanced probabilistic streamflow and precipitation forecasting. Also, the integration of the Hydrologic Uncertainty Processor (HUP) and the BMA methods is proposed for probabilistic post-processing of multi-model streamflow forecasts. Results indicated that the MACHBV and GR4J models are highly competent in simulating hydrological processes within data-scarce watersheds, however, the presence of the lower skill hydrologic models is still beneficial for ensemble-based streamflow forecasting. The comprehensive verification of the BMA approach in terms of streamflow predictions has identified the merits of implementing some of the previously recommended modifications and showed the importance of possessing a mutually exclusive and collectively exhaustive ensemble. By targeting the remaining limitation of the BMA approach, the proposed En-BMA method can improve probabilistic streamflow forecasting, especially under high flow conditions. Also, the proposed HUP-BMA approach has taken advantage of both HUP and BMA methods to better quantify the hydrologic uncertainty. Moreover, the applicability of the modified En-BMA as a more robust post-processing approach for precipitation forecasting, compared to BMA, has been demonstrated. / Thesis / Doctor of Philosophy (PhD) / Possessing a reliable streamflow forecasting framework is of special importance in various fields of operational water resources management, non-structural flood mitigation in particular. Accurate and reliable streamflow forecasts lead to the best possible in-advanced flood control decisions which can significantly reduce its consequent loss of lives and properties. The main objective of this research is to develop an enhanced ensemble-based probabilistic streamflow forecasting approach through proper quantification of predictive uncertainty using an ensemble of streamflow forecasts. The key contributions are: (1) implementing multiple diverse forecasts with full coverage of future possibilities in the Bayesian ensemble-based forecasting method to produce more accurate and reliable forecasts; and (2) developing an ensemble-based Bayesian post-processing approach to enhance the hydrologic uncertainty quantification by taking the advantages of multiple forecasts and initial flow observation. The findings of this study are expected to benefit streamflow forecasting, flood control and mitigation, and water resources management and planning.