Proper site characterization is a critical component in making risk-based decisions and in selecting an appropriate action for a site, whether it is active remediation, containment or natural attenuation. The overall purpose of this work is to investigate innovative techniques for characterizing the factors controlling the transport and fate of organic chemicals at contaminated sites. It is expected that results from this work will lead to improved and more cost-effective methods for characterizing contamination at hazardous waste sites. The information gained from using these methods may lead to a better understanding of factors controlling contaminant transport at sites and therefore more informed risk-based decision making and selection of remediation strategies. The results indicate that (1) the presence of porous media heterogeneity and distinct zones of dense nonaqueous liquid (DNAPL) saturation lead to reduced performance (reduced accuracy) of the partitioning tracer test for measuring DNAPL saturation in saturated subsurface systems, (2) gas-phase tracer tests have the potential to accurately measure water contents for a system with uniform water content and homogeneous porous media, (3) the diffusivity-tracer test method can be used to determine whether diffusion-mediated processes are significant at a particular site, and (4) for a 2-dimensional flow cell flushing experiment the magnitude of trichloroethene concentration and the shape of the trichloroethene elution curves varied as a function of location and sampling type and that the less than solubility concentrations observed at almost all ports were caused by the nonuniform NAPL distribution and porous media heterogeneity, rather than by rate-limited interphase mass transfer at the pore-scale.

A Sediment-Resistant flume has been tested and analyzed in the laboratory and field to measure flow rate with high sediment concentration for natural and irrigation streams. This flume is basically an improvement of the well studied long-throat flume which has an additional chute proposed by J. A. Replogle. The chute, where a new gauging station is located at the half-way point of its length equal to two times the throat length, worked satisfactory for a chute slope of 3%. For clear-water flow the chute showed a supercritical and curvilinear flow having a hydrostatic pressure consistently located at the halfway of the chute. For sediment-laden flow with a concentration of 3% by weight of clear mortar sand, behavior was similar to the clear-water. Sediment deposits filled the approach channel of the flume and plugged the intake pipe of the supercritical gauging station causing it to fail. This failure was compensated for by the use of the new supercritical additional gauging station located in the chute, which remains clean and continues to measure the flow-rate. Hydraulic behavior of the stage-discharge curves for the sediment-laden flow flume was similar to the clear-water flow after the approach channel was filled. Prior to filling an inconsistent condition existed caused by the sediment movements in the approach channel. It was shown that this flume might be used in the field with errors of 5% or less. In addition, a computer program was developed to compute the stage-discharge curve using the dynamic equation of gradually varied flow, which may be used to compute the stage-discharge curve in new flumes. Further laboratory and field research is to be expected to compare these results with future data to improve reliably.

A Snow-Atmosphere-Soil Transfer (SAST) model has been developed to extend the point snowmelt model to vegetated areas using the parameterization concepts of the Biosphere-Atmosphere Transfer Scheme (Dickinson et al. 1993). The model applications for short-grass and forest fields show that the simulated surface temperature, albedo, and snow depth have close agreement with observations. In addition, because of biases in simulated runoff in the high-latitudes, a Shuffled Complex Evolution (Sorooshian et al. 1993) scheme for automatic calibration has been connected with the SAST model to determine the realistic distribution of runoff components from different soil layers and search the optimized parameter set. The calibrated runoff closely matches observations. Because the Community Climate Model version 3 (CCM3) coupled with the SAST model overestimates snow depth and precipitation and underestimates surface temperature over the Rocky Mountains, remotely sensed snow depth data have been assimilated in the model to alleviate model discrepancies based on energy and mass balances. The improved surface temperature simulations result from the decreased snowmelt and albedo in winter and spring and from the weakened evaporation in summer due to drier soil. Meanwhile, modeled summer precipitation over the Rocky Mountains has a minor improvement. The relationship between the variations of tropical Pacific SST and snowpack anomalies in the western United States (U.S.) has been studied by comparing observations and CCM3 output. The results indicate that in the northwestern U.S., the warm tropical Pacific phase of the El Nino-Southern Oscillation (ENSO) is associated with diminished snowpack while its cool phase is related to enhanced snowpack. This relationship is largely determined by winter precipitation variability for the observations; however, it relies heavily on the variations of temperature due to the biases in atmospheric patterns for the model output. In the southwestern U.S., positive snowpack anomalies for both observations and simulations result from the strong warm phase of the ENSO and negative ones are connected with exaggerated local precipitation in fall.

The North American Monsoon System (NAMS) is an important climatological feature of much of southwestern North America because it is responsible for large portions of the annual rainfall in many otherwise arid and semi-arid environments. This dissertation explores issues related to numerical simulation of the North American Monsoon climate. Simulation studies using both an atmospheric general circulation model (AGCM) and a regional climate model (RCM), forced by model analyzed boundary conditions, are presented. The RCM was run for a single season with three different convective parameterization schemes for a single season to assess the sensitivity to convective representation. The main conclusion from these simulations was that substantial differences in both the time-integrated thermodynamic and circulation structures of the simulated July 1999 NAM atmosphere evolve in the simulations when different convective parameterization schemes (CPSs) are used. All simulations reproduced the maximum of precipitation along the western slope of the Sierra Madre Occidental. However, root mean squared errors and model biases in precipitation and surface climate variables were substantial, and showed strong regional dependencies between each of the simulations. There are large differences in the modeled monthly-total surface runoff between simulations. These differences appear to be more closely related to differences in local, precipitation intensity than to time-average or basin-average intensity. It was found that many features of the North American Monsoon were poorly simulated by the AGCM used in its current configuration when using a yearly repeating cycle of sea-surface temperatures. In particular, the model is unable to simulate the regional patterns of monsoon circulation and rainfall. Modeled rainfall over the southwest U.S. and Mexico is much too low, while tropical precipitation is overestimated. Anomalous sea-surface temperature forcing in the Pacific Ocean also induced model responses that resemble observed responses suggesting that sea-surface temperatures may play a modest role in establishing the monsoon circulation and hence in the generation of monsoon rainfall.

Soil-Vegetation-Atmosphere-Transfer Schemes (SVATS) are used in global climate studies to simulate and help understand the complex interactions between the climate and the biosphere. There currently exists a multitude of SVATS of varying complexity differing in terms of the modeled physics and the manner and sophistication with which the processes are represented. This analysis uses systems-based multi-criteria techniques to investigate the performance and sensitivity of various SVATS and their parameters. Results indicate that, once complexity reaches a certain level, incorporating more physics does not necessarily result in improved simulations or reduced errors and that several parameters in the models are insensitive regardless of the input data (i.e., vegetation type). To better understand SVATS performance in semi-arid regions, and to evaluate the various impacts of data on the parameter estimation problem, an intensive calibration and validation study is undertaken. Findings show that calibrated parameters result in improved performance over default, proxy site parameters result in similar performance for many time periods, and there is a need to include wet periods with elevated latent heat to capture the variability of climatic conditions such as the monsoon and El Nino winters. Last, a preliminarily investigation of the performance of the BATS2 model is undertaken to evaluate the capabilities to simulate carbon (along with energy and water) fluxes in semi-arid regions. Results show poor performance for carbon flux simulations and that improvements are needed to better represent C4 vegetation for semi-acid regions. Future research will be directed toward integrated modeling (carbon, energy, and water) in semi-arid regions.

Multiobjective decision support systems (DSS) are gaining acceptance as tools to evaluate resource management systems. Before applying a DSS, a matrix of decision criteria and alternative management systems is populated using information from measured data, expert opinion or simulation models. As each information source exhibits differences in data availability and accuracy, the extent to which outcomes from the DSS are influenced by the source of information remains an important issue. A conceptual framework links the Prototype Decision Support System (P-DSS) developed by the USDA-ARS Southwest Watershed Research Center in Tucson, Arizona, to a conservation practice physical effects matrix. Four rangeland practices of yearlong (YL) and rotation (ROT) grazing, with mesquite trees retained (+M) and removed (-M), are evaluated against eight decision variables that consider soil, water, plants and wildlife habitat. Each decision variable is quantified using data from four experimental watersheds on the Santa Rita Experimental Range, expert opinions, and two simulation modeling approaches. The simple approach uses the Curve Number method, RUSLE and MUSLE, while the complex approach uses the CREAMS hydrology and erosion models. Outcomes from the P-DSS are sensitive to the source of information. When measured data and complex models quantify the decision variables, the YL-M and ROT-M management systems dominate the current system of YL+M. The simple modeling approach identifies ROT+M in addition to YL-M and ROT-M. However, when a frequency of rank methodology is used, the simple and complex modeling approaches identify ROT-M as the preferred system, while the measured data and expert opinion identify YL-M. Ranking the four management systems quantified by simple models matches the ranking obtained from the expert survey. Rank ordering using the complex models agrees with the opinion of the most knowledgeable expert. Simple and complex modeling estimates of sediment yield are significantly different, as are estimates of peak runoff rate. The results suggest that model complexity improved information accuracy but had limited effect on the outcomes from the P-DSS. The effect of information sources on the outcomes from the P-DSS may become more pronounced if the evaluation changes from a relative assessment to one involving quality standards.

Laboratory experiments were conducted on vesicular basalt cores to estimate hydraulic properties. Properties included dry bulk density, effective porosity, skeletal density, saturated hydraulic conductivity and determination of moisture characteristic curves. Unsaturated hydraulic properties estimated included hydraulic conductivity and diffusivity as a function of matrix suction. Infiltration tests were run on a larger block of the same basalt. Infiltration curves were developed and saturated hydraulic conductivity estimated.

The purpose of this study is to examine the effects of land use and hydrology on surface water quality in a semi-arid watershed. Six synoptic sampling events were performed along the upper San Pedro River, AZ, USA before, during, and after the 2002 monsoon season. Water samples were analyzed for conservative solutes, nutrients, and organic matter. During non-monsoon baseflow periods, conservative solutes indicated limited hydrologic connection between regions. Protected reaches had significantly higher DOC concentrations and agricultural reaches had significantly higher DON and NO₃-N levels. In contrast, solute concentrations during the monsoon season indicated all regions were hydrologically linked. DOM and NO₃-N concentrations increased as terrestrially derived solutes were flushed into the stream. Nutrient loads were variable suggesting that changes in nutrient concentrations were related to individual reaches. This research demonstrates that hydrologic flowpaths and land cover are important controls on surface water quality at the reach and river scales.

Groundwater models are often developed as tools for environmental decision-making. However, sparse data availability can limit a model's utility by confounding attempts to select a single structural representation of a system or to find a unique and optimal set of model parameters. As a result, estimates of prediction uncertainty and the value of further data collection may be important results of a modeling effort. The Discrimination/Inference to Reduce Expected Cost Technique (DIRECT) is a new method for developing an ensemble of models that collectively define prediction uncertainty in a manner that supports risk-based decision making and monitoring network design optimization. We apply aspects of DIRECT to a modeling investigation of an aquifer system in Central Utah where a major Coalbed Methane gas field is located and a new approach for stimulating gas production is being explored. In the first stage of this study we develop an ensemble of regional MODFLOW models and calculate their relative likelihood using a set of observation data. These regional results and likelihoods are then transferred to a regional MT3D residence time model and to a local advective transport model to provide further information for the well design. A cost function is applied to the transport results to assess the relative expected costs of several proposed well field designs. The set of hydrologic results and associated likelihoods from the ensemble are combined into cost curves that allow for the selection of designs that minimize expected costs. These curves were found to be a useful tool for visualizing the ways that design decisions and hydrologic results interact to generate costs. Furthermore, these curves reveal ways in which uncertainty can add to the cost of implementing a design. A final analysis explored the cost of having uncertain model results by applying and manipulating synthetic likelihood distributions to the transport results. These results suggest the value that may be added by reducing uncertainty through data collection. Overall, the application of DIRECT was found to provide a rich set of information that is not available when ensemble methods and cost consideration are omitted from a modeling study.

In the United States, the flood frequency analysis guidelines described in Bulletin 17B are followed to provide reliable flood discharge magnitude estimates for urban floodplain planning and flood insurance studies. The statistical analysis in Bulletin 17B has various assumptions, including that floods are generated by the same type of atmospheric mechanism (flood homogeneity). However, these assumptions should be carefully assessed before proceeding since they might not always be valid and could increase the potential for flood risk. This study focuses on flood frequency analysis from the perspective of flood heterogeneity, the hydrometeorological genesis of each flood event, in Arizona. This was done by analyzing the occurrence and magnitude of individual flood events, which were classified by their flood-producing atmospheric mechanism. Flood frequency curves were derived for each mechanism and combined using a new approach involving the Partial Duration Series peaks. The combined frequency curves were then compared to curves derived from the standard Bulletin 17B method. Results showed that in southern Arizona, the dominant flooding mechanism is characterized by brief, intense, and localized convective precipitation in the summer. However, the dominant flood-producing mechanism in the central Arizona topographic transition zone and at higher elevations is characterized by prolonged and widespread precipitation from synoptic activity in the winter. Tropical cyclone-enhanced precipitation is also an important, but infrequent, flood-producing mechanism throughout the state. Overall, the dominant mechanism does not necessarily produce the largest floods. In such cases flood heterogeneity can have a strong influence on the discharge estimates for the most extreme upper tail probabilities calculated from the flood frequency analysis. Thus, the most frequent floods may impose very little risk of flooding while uncommon floods can impose a much larger one. These results suggest that the flood homogeneity assumption is not valid in many Arizona watersheds. To produce the most accurate discharge estimates possible, it is critical that both analysts and flood managers become aware of the potential repercussions if these details are overlooked.