O modelo Soil and Water Assessment Tool (SWAT) tem sido utilizado para avaliar os impactos do uso e manejo da terra nos recursos hídricos, sedimentos e agroquímicos em diversas escalas e condições ambientais em todo o mundo. Entretanto, pelo fato de ter sido desenvolvido em centros de pesquisa norte-americanos, alguns parâmetros e rotinas de simulação não refletem adequadamente determinados processos de bacias localizadas em regiões tropicais. Nesse sentido, o presente trabalho visou aprimorar a modelagem hidrológica do SWAT através da revisão e modificação de processos relacionados à simulação da evapotranspiração. Os procedimentos propostos incluíram a alteração das rotinas de dormência vegetal no código fonte do modelo e a modificação dos cronogramas de operações de manejo e parâmetros do banco de dados de crescimento das plantas, visando reproduzir com maior precisão o ciclo das culturas em bacias tropicais. As modificações foram testadas em cinco bacias localizadas no Estado de São Paulo, com áreas entre 42 e 5.959 km², sendo comparados os resultados obtidos antes e depois da implementação das mesmas. Com as alterações, a análise do balanço hídrico anual evidenciou um aumento nos valores de evapotranspiração de cerca de 61% nas bacias, aproximando-se dos totais anuais de evapotranspiração calculados através de métodos empíricos, bem como houve redução significativa do escoamento superficial. Verificou-se uma melhoria da simulação de vazão em todas as bacias, sendo obtidos valores superiores para o Coeficiente de Eficiência de Nash-Sutcliffe (NSE) quando comparados à simulação sem as alterações. A calibração e validação foram realizadas com base na simulação modificada, sendo obtidos valores de NSE mensais entre 0,71 e 0,93 na calibração e 0,53 e 0,88 na validação, enquanto os valores diários de NSE situaram-se entre 0,51 e 0,82 na calibração e 0,38 e 0,83 na validação. A calibração a partir de uma simulação na qual as distorções dos processos hidrológicos da bacia estivessem previamente minimizadas resultou em bons resultados sem alteração excessiva dos parâmetros, indicando uma simulação hidrológica de melhor consistência. / The Soil and Water Assessment Tool (SWAT) has been used to predict the impact of land management practices on water, sediment, and agricultural chemical yields in a wide range of scales and environmental conditions across the globe. However, originally developed in the United States, some parameters and routines are unrealistic for simulating in tropical watersheds. In this sense, this work aims to improve the hydrologic modeling of SWAT model by reviewing and modifying parameters and routines related to evapotranspiration process. In order to adequately represent crop growth in tropical basins, the proposed procedures included changes in dormancy routines of SWAT source code and modifications of scheduled management operations and plant growth database parameters. These modifications were tested in five different basins at São Paulo State, Brazil, with areas ranging from 42 to 5959 km², by comparing the results before and after their implementation. Annual water balance analysis showed an increase in evapotranspiration about 61% for basins, approaching the total annual evapotranspiration estimated by empirical methods. Hence, it was observed that surface runoff and base flow components showed a decrease. The modifications resulted in improved flow simulation for all basins, showing better Nash-Sutcliffe Efficiency Coefficient (NSE) values compared to the unchanged simulation. Calibration and validation processes used the modified simulation database, being achieved monthly NSE between 0.71 – 0.73 at calibration and 0.53 – 0.88 at validation, while daily NSE were 0.51 – 0.82 at calibration and 0.38 – 0.83 at validation. Overall, minimizing distortions in hydrological processes at pre-calibration step resulted in good estimations without excessive modification of parameters at calibration, attesting a consistent hydrological modeling for the basins analyzed.

Seepage is an important component of the water budget in closed basins that do not have surface water drainage features. In the shallow water table environment of Florida, internal drainage of soil controls flooding. With the recent rapid population growth and urban development in the state, a simple, field-based method is needed to estimate seepage rates. In this study five locations in Hillsborough County, Florida, were instrumented with wells with pressure transducers measuring water level fluctuations at 1 minute resolution. For closed basins with lakes, evaporation (E) rates were determined using data from a weather station and Penman-Monteith FAO56 method, and then seepage rates were calculated from a water budget. Seepage rates varied greatly depending on conditions specific to the site. The seepage rates found for the three surface water sites in this study were 1.1 cm/d for a retention pond surrounded with dense vegetation, 0.5-0.8 cm/d for a natural lake located close to a groundwater pumping site, and 0.4 cm/d for another natural lake with no groundwater pumping in the proximity. Two methods to estimate seepage rates into semi-confined aquifers were compared: (a) mass balance approach and (b) Darcy's equation. At one of the sites the rate was 0.1 cm/d, and at the other site (sinkhole) it ranged between 0.8 cm/d during the wet season and 0.2 cm/d during the dry season when the head difference between the surficial and Floridan aquifers became smaller. The results of the study indicate that simple and relatively inexpensive field methods can estimate seepage within a narrow range and give reasonable seepage predictions that can be used in flood modeling. The obtained values indicate that seepage does not provide adequate drainage relief in closed basins. Another important finding is the magnitude of the local recharge to the Floridan aquifer. Further sensitivity studies on hydrological models that use seepage as one of the inputs may indicate that lower data collection resolution or simpler ET estimation methods are acceptable.

Spatially-distributed water consumption was modeled over a segment of the Lower Colorado River, which contains irrigated agricultural and Tamarisk-dominated riparian ecosystems. For the irrigation scheme, distributed evapotranspiration data were analyzed in conjunction with point measurements of precipitation and surface flow in order to close daily and annual water balance. The annual closure error was less than 1% of the total water diversion to the area. In addition, it was found that the soil water storage component of the water balance cannot be neglected if the analysis is performed over time frames shorter than annual (e.g. growing season). Water consumption was highly uniform within agricultural fields, and all the full-cover fields were transpiring close to their potential rates. Mapping several new and existing drainage performance indicators showed that neither soil salinization nor water-logging would be of concern in this irrigation scheme. However, the quality of high-volume return flow must be studied, especially since the degraded water quality of the western US rivers is believed to act in favor of the invasive riparian species in outcompeting native species. Over the Tamarisk forest, the remotely-sensed evapotranspiration estimates were higher than the results of an independent groundwater-based method during spring and winter months. This was chiefly due to the fixed satellite overpass time, which happened at low sun elevation angles in spring and winter and resulted in a significant presence of shadows in the satellite scene and consequently a lower surface temperature estimate, which resulted in a higher evapotranspiration estimate using the SEBAL model. A modification based on the same satellite imagery was proposed and found to be successful in correcting for this error. Both water use and crop coefficients of Tamarisk estimated by the two independent methods implemented in this study were significantly lower than the current approximations that are used by the US Bureau of Reclamation in managing the Lower Colorado River. Studying the poorlyunderstood stream-aquifer-phreatophyte relationship revealed that diurnal and seasonal groundwater fluctuations were strongly coupled with the changes in river stage at close distances to the river and with the Tamarisk water extraction at further distances from the river. The direction of the groundwater flow was always from the river toward the riparian forest. Thus the improved Tamarisk ET estimates along with a better understanding of the coupling between the river and the riparian aquifer will allow the Bureau of Reclamation to re-asses their reservoir release methodology and improve efficiency and water savings.

The principal objective of this dissertation is to develop Bayesian machine learning models for multiple reservoir, irrigation canal, and river system operation and management. These types of models are derived from the emerging area of machine learning theory; they are characterized by their ability to capture the underlying physics of the system simply by examination of the measured system inputs and outputs. They can be used to provide probabilistic predictions of system behavior using only historical data. The models were developed in the form of a multivariate relevance vector machine (MVRVM) that is based on a sparse Bayesian learning machine approach for regression. Using this Bayesian approach, a predictive confidence interval is obtained from the model that captures the uncertainty of both the model and the data. The models were applied to the multiple reservoir, canal and river system located in the regulated Lower Sevier River Basin in Utah. The models were developed to perform predictions of multi-time-ahead releases of multiple reservoirs, diversions of multiple canals, and streamflow and water loss/gain in a river system. This research represents the first attempt to use a multivariate Bayesian learning regression approach to develop simultaneous multi-step-ahead predictions with predictive confidence intervals for multiple outputs in a regulated river basin system. These predictions will be of potential value to reservoir and canal operators in identifying the best decisions for operation and management of irrigation water supply systems.

A reliable estimate of evapotranspiration (ET) in river basins is important for the purpose of water resources planning and management. ET represents a significant portion of rainfall in the water budget; therefore, the uncertainty in estimating ET can lead to the inaccurate prediction of water resources. While remote sensing techniques are available to estimate ET, such methods are expensive and necessary data may not be readily available. Classical methods of estimating ET require detailed land use/cover information that are not readily available in rural river basins. Complementary methods provide simple and reliable approaches to estimate ET using meteorological data only. However, these methods have not been investigated in detail to assess the overall applicability and the needs for revisions if any. In this work, an improved approach to use the complementary methods using readily available meteorological data is presented. The methodology is validated using 34 global FLUXNET sites with heterogeneous land use/cover, climatic, and physical conditions. The method was compared with classical methods using Ghana as a study area where original pioneering studies of ET have been performed. The work was extended to develop global maps of ET and water surplus (precipitation - ET) for the 20th century followed by climate change-induced 21st century estimates for 2040-2069 and 2070-2099 periods. The emission scenario used was the moderate A1B with the global climate models CGCM3.1 and HADGEM1. The results were assessed at different scales from global to regional such as for potential outcomes of climate change on ET and water surplus.

The current study has been conducted in response to the growing problem of water scarcity and the need for more effective methods of irrigation water management. Remote sensing techniques have been used to match spatially and temporally distributed crop water demand to water application rates. Remote sensing approaches using Landsat imagery have been applied to estimate the components of a soil water balance model for an agricultural field by determining daily values of surface/root-zone soil moisture, evapotranspiration rates, and losses and by developing a forecasting model to generate optimal irrigation application information on a daily basis. Incompatibility of coarse resolution Landsat imagery (30m by 30m) with heterogeneities within the agricultural field and potential underestimation of field variations led the study to its main objective, which was to develop models capable of representing spatial and temporal variations within the agricultural field at a compatible resolution with farming management activities. These models support establishing real-time management of irrigation water scheduling and application. The AggieAirTM Minion autonomous aircraft is a remote sensing platform developed by the Utah Water Research Laboratory at Utah State University. It is a completely autonomous airborne platform that captures high-resolution multi-spectral images in the visual, near infrared, and thermal infrared bands at 15cm resolution. AggieAir flew over the study area on four dates in 2013 that were coincident with Landsat overflights and provided similar remotely sensed data at much finer resolution. These data, in concert with state-of-the-art supervised learning machine techniques and field measurements, have been used to model surface and root zone soil volumetric water content at 15cm resolution. The information provided by this study has the potential to give farmers greater precision in irrigation water allocation and scheduling.

Both empirical and process-simulation models are useful for evaluating the effects of management practices on environmental quality and crop yield. The use of these models is limited, however, because they need many soil property values as input. The first step towards modelling is the collection of input data. Soil properties can be highly variable spatially and temporally, and measuring them is time-consuming and expensive. Efficient methods, which consider the uncertainty and cost of measurements, for estimating soil hydraulic properties form the main thrust of this study. Hydraulic properties are affected by other soil physical, and chemical properties, therefore it is possible to develop empirical relations to predict them. This idea quantified is called a pedotransfer function. Such functions may be global or restricted to a country or region. The different classification of particle-size fractions used in Australia compared with other countries presents a problem for the immediate adoption of exotic pedotransfer functions. A database of Australian soil hydraulic properties has been compiled. Pedotransfer functions for estimating water-retention and saturated hydraulic conductivity from particle size and bulk density for Australian soil are presented. Different approaches for deriving hydraulic transfer functions have been presented and compared. Published pedotransfer functions were also evaluated, generally they provide a satisfactory estimation of water retention and saturated hydraulic conductivity depending on the spatial scale and accuracy of prediction. Several pedotransfer functions were developed in this study to predict water retention and hydraulic conductivity. The pedotransfer functions developed here may predict adequately in large areas but for site-specific applications local calibration is needed. There is much uncertainty in the input data, and consequently the transfer functions can produce varied outputs. Uncertainty analysis is therefore needed. A general approach to quantifying uncertainty is to use Monte Carlo methods. By sampling repeatedly from the assumed probability distributions of the input variables and evaluating the response of the model the statistical distribution of the outputs can be estimated. A modified Latin hypercube method is presented for sampling joint multivariate probability distributions. This method is applied to quantify the uncertainties in pedotransfer functions of soil hydraulic properties. Hydraulic properties predicted using pedotransfer functions developed in this study are also used in a field soil-water model to analyze the uncertainties in the prediction of dynamic soil-water regimes. The use of the disc permeameter in the field conventionally requires the placement of a layer of sand in order to provide good contact between the soil surface and disc supply membrane. The effect of sand on water infiltration into the soil and on the estimate of sorptivity was investigated. A numerical study and a field experiment on heavy clay were conducted. Placement of sand significantly increased the cumulative infiltration but showed small differences in the infiltration rate. Estimation of sorptivity based on the Philip's two term algebraic model using different methods was also examined. The field experiment revealed that the error in infiltration measurement was proportional to the cumulative infiltration curve. Infiltration without placement of sand was considerably smaller because of the poor contact between the disc and soil surface. An inverse method for predicting soil hydraulic parameters from disc permeameter data has been developed. A numerical study showed that the inverse method is quite robust in identifying the hydraulic parameters. However application to field data showed that the estimated water retention curve is generally smaller than the one obtained in laboratory measurements. Nevertheless the estimated near-saturated hydraulic conductivity matched the analytical solution quite well. Th author believes that the inverse method can give a reasonable estimate of soil hydraulic parameters. Some experimental and theoretical problems were identified and discussed. A formal analysis was carried out to evaluate the efficiency of the different methods in predicting water retention and hydraulic conductivity. The analysis identified the contribution of individual source of measurement errors to the overall uncertainty. For single measurements, the inverse disc-permeameter analysis is economically more efficient than using pedotransfer functions or measuring hydraulic properties in the laboratory. However, given the large amount of spatial variation of soil hydraulic properties it is perhaps not surprising that lots of cheap and imprecise measurements, e.g. by hand texturing, are more efficient than a few expensive precise ones.

Based on the determination of the zero-flux plane, a water balance procedure for large-scale experimental databases was automated to estimate the soil water balance based on soil water content distribution with depth through time. The automated procedure was verified using data from the BOREAS project obtained in three Intensive Field Campaigns during the spring and summer of 1994. The data used correspond to four tower sites measuring atmospheric fluxes above the forest canopy from the Northern and Southern Study Areas and are designated according to the predominant vegetation in the area as Old Jack Pine and Young Jack Pine. The total hydraulic head through time at these sites is determined to identify the position of the zero-flux plane, which separates that part of the soil profile in which water flow is upward from the region in which the water flow is downward. In conjunction with precipitation and soil water content data, the procedure allows estimation of the actual soil water balance, the water used from the region above the zero-flux plane being evapotranspiration, and the change in soil water content below the mean zero-flux plane being drainage. Prior to this study, no published attempt had been made to automate a water balance procedure for large-scale experimental databases based on the position of the zero-flux plane and soil water content distribution through time. / Graduation date: 1997

.

How much groundwater flows to boreal streams depends on the relative contributions from each landscape unit (forested uplands, lakes, and peatlands) within a catchment along with its hydrogeologic setting. Although there is an understanding of the hydrologic processes that regulate groundwater outputs from individual landscape units to their underlying aquifers (both coarse- and fine-textured) in the boreal forest, less understood is how the topography, typology, and topology (i.e. hydrologic connectivity) of the landscape units regulates groundwater flow to streams. Improved understanding of groundwater-stream interactions in the Boreal Plain of Alberta and Saskatchewan is critical as this region is undergoing substantial environmental change from land cover disturbances for energy and forestry industries and climate change. This thesis determines groundwater-stream interactions during the autumn low-flow period in a 97 km2 glacial outwash sub-catchment of White Gull Creek Research Basin, Boreal Ecosystem Research and Modelling Site, Saskatchewan. The catchment (Pine Fen Creek) is comprised of a large (30 km2) valley-bottom peatland, two lakes, and jack pine (Pinus banksiana) uplands. The pine uplands are important areas of annual groundwater recharge for the catchment. Vertical hydraulic gradients (VHGs) show frequent flow reversals between the lakes and sand aquifer, and spatially diverse VHGs between the peatland and sand aquifer. Groundwater flow nets and lateral hydraulic gradients indicate the stream receives groundwater along its length. Isotopic samples of end members corroborate the hydrometric data. Catchment streamflow response during the 2011 low flow period was not simply the addition of net groundwater inputs from each landscape unit. Instead, the large size, valley-bottom position, and short water ‘memory’ of the peatland were the critical factors in regulation of catchment streamflow during low flow periods. Peatland hydrologic function alternated between a source and sink of runoff (surface and subsurface) to the stream, dependent on the position of the water table; a value of 0.15 m below peat surface was the critical functional tipping point. Given the high percentage of peatlands (21%) within the Boreal Plain, incorporating their runoff threshold is required in parameterizing runoff generation in hydrological models, and thus predicting impacts of peatland degradation and forest clearing on streamflow.