Results of chemical analyses revealed elevated levels of organic surrogate parameters which included dissolved organic carbon, trihalomethanes, total organic halides, and UV absorbance in Tucson groundwater impacted by artificial recharge activity. Inorganic parameters associated with recharge water were also elevated in Tucson groundwater impacted by artificial recharge. The mobility of organochlorine during deep well injection of chlorinated Colorado River Water into the Las Vegas Valley aquifer was also investigated. The Las Vegas Valley aquifer was found to have little sorptive capacity for organo-chlorine.

In 1959, the Bureau of Reclamation proposed a channelization project through the Cibola Valley along the Lower Colorado River. The project entailed rerouting the river through a dry cut in order to lower groundwater levels in the Palo Verde Irrigation District upstream, thereby improving irrigation drainage. In conjunction with this, Cibola National Wildlife Refuge was created in 1964 to mitigate the effects of habitat loss from the channelization project. Aerial photographs of the Cibola Valley were analyzed for 1959, and vegetation community types were determined. A vegetation type map was developed for 1959 and compared with vegetation type maps for 1976 and 1986. Between 1959 and 1986, a lowering of the water-surface level in some parts of the refuge resulted in the draining of some lakes and the creation of slow-moving backwaters. Cottonwood-willow and marsh communities saw a significant reduction in area while the salt-cedar community saw rapid growth.

A sealed double ring infiltrometer (SDRI) was used at Page Ranch, Arizona to measure saturated hydraulic conductivity (Ksat in a test clay pad, as well as to characterize preferential flow and transport mechanisms. Ksat varied from 3.5 x 10⁻⁹ to 2.2 x 10⁻¹⁰ cm/sec depending on treatment of matric potential at the wetting front. These in-situ Ksat values were about one order of magnitude less than the laboratory values that ranged from 10⁻⁷ to 10⁻⁹ cm/sec. Although the pad was not instrumented to detect a shallow wetting front, the dye front and water content data indicated that flow penetrated 4 to 6 cm by the end of the 75 day test. Tracer data suggested that bromide moved to 18 or 20 cm by diffusion and that the effective diffusion coefficient was 15 to 21 x 10⁻¹⁰ m²sec. Preferential flow, as judged by the uniformity of the dye front and bromide tracer movement, was insignificant.

This thesis has analyzed some aspects of stochastic flow and transport in geologic media with a random stationary and statistically isotropic hydraulic conductivity field. Explicit expressions for cross-covariances between velocity and head, and velocity and log conductivity as well as covariances of velocity under steady state uniform mean three-dimensional flow with an exponential log conductivity covariance are derived to first order and their structure is examined. An exact early time solution due to Batchelor for the mean concentration is compared with other existing stochastic solutions and its range of validity is determined for the case of an instantaneous point source. This early time solution is simpler and more general than any other stochastic transport solution at early time. A Monte Carlo simulation scheme is developed to study the ensemble behavior of solute particles traveling in such a field. The thesis concludes with a concentration estimation scheme conditioning on site measurements.

An important wetland type in the southwestern United States requiring study is the cienega. Cienega is a Spanish term used for wetlands found in semi-arid grasslands, associated with perennial springs and headwater streams. A field study was conducted at Canelo Hills Cienega in Southeastern Arizona for one year in which changes in water levels, soil water content, and stream flow were monitored. Plant composition, soil classification, and basic geological characteristics were determined as well. Water level fluctuations and flow gradients indicate that this cienega is mostly groundwater dependent and is an effluent system maintaining the perennial nature of the adjacent stream. Mean water levels ranged from.9 cm above the surface to 0.85 m below. Water level fluctuation within the cienega ranged from 1.12 m/yr to 0.18 m/yr. Vegetation occurred in distinct patterns which varied across the cienega in relation to water levels and availability. Dominant genera included Eleocharis, Carex, Juncus, Poa, and Bidens.

Replicate column experiments were done to quantify the effects of temperature and bacterial motility on advective transport through repacked, but otherwise unaltered, natural aquifer sediment. Greater microsphere removal observed at the higher temperature agreed with the physical-chemical model, but bacteria removal at 18°C was only half that at 4°C. The sticking efficiency for non-motile A0500 (4°C) was over three times that of the motile A0500 (18°C), 0.073 versus 0.022 respectively. Motile A0500 bacteria traveled twice as far as non-motile A0500 bacteria before becoming attached. Once attached, non-motile colloids detached on the time scale of 9 to 17 days. The time scale for detachment of motile A0500 bacteria was shorter, 4 to 5 days. Results indicate that bacterial attachment was reversible and detachment was enhanced by bacterial motility. The kinetic energy of bacterial motility changed the attachment-detachment kinetics in favor of the detached state.

Various humic substances were reacted with a coarse loamy soil in both batch studies and column experiments. Solution pH was not buffered, rather monitored to more closely approximate a "real life" scenario. Humic substances were also reacted with soil that had been pre-saturated with various concentrations of cadmium nitrate solution. The batch experiments were characterized by interference from antecedent natural organic matter in the soil, causing high nonlinearity in the data. Humic substance sorption in the column experiments was characterized by early breakthrough of a small fraction, followed by long extended tailing. Complete breakthrough was never observed in any of the column studies. Sorption phenomena is believed to be primarily controlled by the structural configuration of the humic substances, with solution pH and the valence of cations in the electrolyte solution having the greatest effect on sorption. Implications to humic substance transport in natural systems will be discussed.

New methods were tested to evaluate the ground cover parameters used in the WEPP model. Rainfall simulator plot data, collected from fifteen rangeland ecosystems across the western United States, were used to evaluate modifications to the ground cover components for rill and interrill areas in the model. Observed and predicted values of average sediment yield were compared for six different configurations of the model. The modifications made to WEPP to incorporate measured ground cover distributions resulted in a lower r2 for predicted sediment yield than any of the configurations in which ground cover distribution was not modified. This result was due to a reduction in the calculated value of total interrill ground cover. The model equations used to adjust interrill erodibility for the effect of interrill ground cover were adversely affected. The result was that interrill erodibility values were increased for many of the plots tested.

The spatial and temporal variability of water content, soil water tension, and derived hydraulic conductivity parameters are analyzed using geostatistical methods. The measured data sets were obtained from a 1985 experiment near Las Cruces, New Mexico. Post-irrigation water content and tension measurements had been recorded over 44 days at 455 sampling locations along a 91 x 1.5-meter transect. Unsaturated hydraulic conductivity values are derived using the instantaneous profile formula, and an exponential model is used to obtain values of saturated hydraulic conductivity and pore-size distribution parameters. The exponential model is found to inadequately describe the conductivity data for tensions near saturation, because excessively large saturated hydraulic conductivity values are derived. Semivariogram analysis shows ranges of dependence of three to 32 meters for water content and six to 34 meters for tension. As water content decreases, the coefficient of variation and variance are found to increase.

The relationship between runoff peak and volume was examined for ten sub-watersheds in the Beaver Creek, Arizona watershed. Least square analyses, linear regression, and coefficient analysis were utilized to evaluate the effects of treatments. Different treatments, such as clear cutting, uprooting, herbicide, and thinning, were applied on watersheds. For most of the watersheds, treatments were shown to have no significant effect on the relationship. However, herbicide treatment and clear cutting treatment in watersheds 3 and 12 respectively showed highly significant increases in peak flows. Significant increase in peak for the smaller events resulted from scattered patch cutting in watershed #10.