David L. Rudolph

Role of the riparian zone in controlling the distribution and fate of agricultural nitrogen near a small stream in southern Ontario

Abstract Uncultivated riparian areas can play an important role in reducing nutrient loading to streams in agricultural watersheds. Groundwater flow and geochemistry were monitored in the riparian zone of a small agricultural watershed in southern Ontario. Hydraulic and geochemical measurements were taken along a transect of monitoring wells extending across the riparian area into an agricultural field. Chloride and nitrate concentrations in groundwater samples collected from the agricultural field were much higher than in samples from the riparian area. A sharp decline in both nitrate and chloride concentrations was observed near the field–riparian zone boundary. It appears that increased recharge within the riparian zone, as compared to the artificially drained field, caused nitrate-rich groundwater from the field to be diverted downward beneath the riparian zone, thus limiting the input of agrochemicals to the riparian area and consequently protecting the stream from potential contamination. Geochemical data also indicated that nitrate was attenuated in the downward moving groundwater. Patterns of dissolved oxygen concentrations and redox potential in the subsurface coincided with the pattern defined by groundwater nitrate. These patterns indicated that conditions within the riparian zone and at depth near the field–riparian zone boundary were conducive to denitrification. A linear relation between the δ 15 N and δ 18 O values of nitrate from the monitored transect also supported denitrification as the primary nitrate removal mechanism. This study provides a new conceptual model of how riparian zones may prevent nitrate contamination of streams, and highlights the need for a complete understanding of both groundwater flow and geochemistry in riparian environments.

Spatial Averaging of Water Content by Time Domain Reflectometry: Implications for Twin Rod Probes with and without Dielectric Coatings

The averaging of apparent relative dielectric permittivities by time domain reflectometry (TDR) is examined for properties varying along TDR waveguides and in the plane perpendicular to a TDR probe. A square root averaging model with uniform weighting factors describes the measured apparent relative dielectric permittivity for axially varying materials; variations in the transverse plane are described by an inverse averaging model with nonuniform spatial weighting. We define the sensitivity of a TDR probe as the change in the measured travel time for a given change in the average soil water content for rods of a fixed length. For rods with dielectric coatings, this sensitivity increases with the dielectric permittivity of the coating and the rod radius and decreases with the soil dielectric permittivity. Due to the dependence of the sensitivity on the soil water content, coated rods will not measure the actual average water content for conditions of axially variable water contents.

The sample areas of conventional and alternative time domain reflectometry probes

We define the sample area in the plane perpendicular to the long axis of conventional and alternative time domain reflectometry (TDR) probes based on the finite element numerical analysis of Knight et al. [1997] and the definition of spatial sensitivity ofKnight [1992]. The sample area of conventional two- and three-rod probes is controlled by the rod separation. Two-rod probes have a much larger sample area than three-rod designs. Low dielectric permittivity coatings on TDR rods greatly decrease the sample area. The sample area of coated rod probes decreases as the relative dielectric permittivity of the surrounding medium increases. Two alternative profiling probes were analyzed. The separation of the metal rods of Hook et al. [1992] probes controls the size of the sample area. Reducing the height or width of the rods improves the distribution of sensitivity within the sample area. The relative dielectric permittivity of the probe body does not affect the sample size. The sample size of the Redman and DeRyck [1994] probe is also controlled by the rod separation. Two alternative surface probe designs [White and Zegelin, 1992; Selker et al., 1993] are shown to have similar sample areas.

Analysis of long-term land subsidence near Mexico City: Field investigations and predictive modeling

The Mexico City region has several flat plains formed on exceptionally porous (60–90%) lacustrine deposits overlying a highly productive regional aquifer. Severe land subsidence due to consolidation of the lacustrine aquitard caused by aquifer exploitation has resulted in restrictions on pumping in the core of Mexico City. This has led to large increases in aquifer pumping in the outlying lacustrine plains where satellite communities are rapidly expanding. The Chalco Basin is one of these lacustrine areas where pumping began in the 1950s and greatly increased in the 1980s. The lacustrine sequence in the Chalco area is significantly thicker than anywhere else in the Basin of Mexico averaging 100 m and reaching a maximum thickness of 300 m. Consequently, this area is susceptible to the highest potential land subsidence effects as a result of groundwater extraction of anywhere in the basin. Land subsidence in the central part of the Chalco Basin has increased to 0.4 m/yr since 1984 and by 1991 total subsidence had reached 8 m. The rapid land subsidence in this area is causing the accumulation of meteoric waters during the rainy season resulting in extensive flooding of farmland. This study first demonstrates a methodology for combining hydraulic data from a network of monitoring wells, geotechnical data from core samples, and a compilation of historical information on land surface elevation to quantify groundwater flow and land subsidence phenomena within the rapidly subsiding Chalco Basin. Then a one-dimensional mathematical model is employed to develop predictions of future land subsidence under a range of pumping conditions. The model permits the hydraulic properties of the aquitard to vary as transient functions of hydraulic head and porosity. Simulations suggest that under current pumping rates, total land subsidence in the area of thickest lacustrine sediment will reach 15 m by the year 2010. If pumping is reduced to the extent that further decline in the potentiometric surface is prevented, total maximum subsidence would be significantly less, ∼10 m, and the rate would nearly cease by 2010.

Contamination in Ontario farmstead domestic wells and its association with agriculture: 2. Results from multilevel monitoring well installations

Abstract Multilevel monitoring wells (MLWs) were installed at 144 farms which were part of a province-wide survey of farm drinking water wells conducted in 1991–1992 in Ontario, Canada. The multilevel sites were selected in areas characterized by coarse-textured soils, on farms typifying local agricultural enterprises. The MLWs were installed in cultivated fields adjacent to the location of the drinking water wells on each farm (within 200 m). On 16 of these farms, MLWs were also installed in a woodlot adjacent to the field site. Water samples were collected on two occasions (winter and summer) and analyses were conducted for nitrate (NO 3 − ), typical bacteria and a selected suite of common pesticides. At 23% of the sites, concentrations in 50% or more of the monitored intervals exceeded the provincial drinking water standard (MAC) for NO 3 − -N during both sampling periods. Significantly higher frequencies of total coliform contamination were encountered in the winter (66%) than in the summer (36%). Very few pesticide detections were recorded. The average concentration of NO 3 − with depth in multilevel wells decreased from approximately 10 mg N l −1 near the water table to 3 mg N l −1 at a depth of about 6.5 m. Bacteria concentrations remained more uniform with depth but decreased significantly in the summer. For most analytes, contaminant frequency was similar for both the drinking water wells and multilevel wells. The occurrence of elevated levels of contamination in the water wells appeared to be associated more with activities on the cultivated fields than with on-farm point sources of contamination. Groundwater quality determined using drinking water wells was consistent with conclusions drawn from multilevel monitoring wells, indicating the utility of water well survey data for assessing groundwater quality within the rural environment. No correlations were observed between the type of cropping practice and the frequency of groundwater contamination. Farms where manure spreading occurred, however, had significantly higher occurrences of contamination from NO 3 − and bacteria than farms where manure was not applied along with commercial N-fertilizer. The uncultivated conditions beneath the woodlot areas appeared to provide an environment conducive to attenuation of nitrate but not bacteria.

A numerical analysis of the effects of coatings and gaps upon relative dielectric permittivity measurement with time domain reflectometry

Fluid-filled gaps or dielectric coatings around parallel-wire transmission lines affect the ability of time domain reflectometry (TDR) to measure the water content of soils and other porous materials. We use a steady state, two-dimensional, finite element numerical solution of Laplaces equation to analyze these effects. We prove that the numerically determined electrostatic potential distribution and boundary fluxes can be used to calculate the equivalent relative dielectric permittivity measured by TDR by comparing the results of the numerical model with those obtained using existing analytical solutions for special cases. We then analyze the effects of fluid-filled concentric gaps that completely or partially surround TDR rods. The results show that an analytical solution due to Annan [1977b] for nonconcentric gaps can be used as a good approximation to predict the effect of concentric gaps or coatings that completely surround the rods. Coatings or gaps filled with low relative dielectric permittivity materials have a greater impact on the measured relative dielectric permittivity than those filled with high dielectric media. An increase in the thickness of the gap or coating for given rod diameters and separations increases the impact of the coating. To a lesser degree, the impact of a coating of a given thickness decreases with an increase in the ratio of the rod diameter to the rod separation. A gap or coating of a given thickness and relative dielectric permittivity will have a greater impact on the response of a three-rod probe than on that of a two-rod probe with the same rod diameter and separation of the outermost rods. Partial air gaps surrounding less than 30° of the rod circumference are not likely to affect the probe response significantly.

Groundwater Flow and Solute Transport in Fractured Lacustrine Clay Near Mexico City

A network of piezometers was installed in a surficial lacustrine clay aquitard overlying a thin saline water aquifer of volcanoclastic origin at a study site near Mexico City in the Basin of Mexico. The aquifer is underlain by additional lacustrine sediments which in turn overlie a thick regional freshwater aquifer. The regional aquifer provides approximately 70% of the water supply for 20 million people in the Basin of Mexico. In the study area, major ions, oxygen 18, and deuterium in the pore water of the surficial aquitard exhibit large variations with depth. The nature of these variations suggests that the saline pore water is being displaced downward by infiltrating meteoric water. The infiltration has been induced by strong downward hydraulic gradients imposed two to three decades ago when heavy aquifer pumping of the thin saline water aquifer began. One-dimensional analytical models representing solute transport in both fractured and unfractured porous media were used to simulate the geochemical profiles in the surficial aquitard. The fractured porous medium model, using a realistic mean hydraulic gradient and fracture spacing (1.5 m) and small but significant fracture aperture (30 μm) provide nearly an exact match to the field data. From this we infer that, because of vertical fractures, there is a much greater potential for downward leakage of water and contaminants through the Mexico City clay into underlying aquifers than has been previously thought.

Origin of methane in the Elk Valley coalfield, southeastern British Columbia, Canada

Coalbead methane is part of the non-conventional gas reservoirs and makes a significant contribution to gas production in some parts of the world. Initially, it was assumed that coalbed methane was of thermogenic origin, but most recent studies based on isotope and chemical data and taking into account the hydrogeology of the basin have demonstrated that secondary biogenic gases are formed in many coal-bearing basins. This study, using a similar approach, evaluated the origin of the gas in the Elk Valley coalfield located in British Columbia, Canada. Isotope data in methane samples collected from testhole wells and piezometers show a range that varies from −51.8‰ to −65.4‰ and −303‰ to −415‰ for δ13C and δ2H, respectively. The δ2H data, which are among the most depleted data reported for coalbed methane, have to be related to the very depleted δ2H values of the groundwater (−148‰ to −163‰). Isotope and chemical data collected from DIC show a trend of increasing δ13C values (−11.9‰ to +34.9‰) associated with an increase in DIC concentration (216 to 1650 mg/l). The most 13C depleted DIC and low DIC waters are found in the shallow groundwater flow system representing conditions close to recharge areas, while the most 13C enriched DIC and high concentration DIC waters are found in the discharge areas associated with a deep groundwater flow system. The DIC pattern, which is typical for methanogenic aquifers, and the isotope data obtained in methane samples clearly indicate that the gas found in the Elk Valley coalfield is mainly biogenic in origin. This study reaffirms that an approach that combines an evaluation of the groundwater flow system, the isotopic characterization and concentration pattern of the main carbon pools (CH4 and DIC), and the isotopic characterization of the groundwater is needed to fully evaluate the origin of gases in coal basins.

Landscape restoration after oil sands mining: conceptual design and hydrological modelling for fen reconstruction

Extraction of oil sands in the relatively dry Western Boreal Plains near Fort McMurray, Alberta, destroys the natural surface cover including fen peatlands that cover upto 65% of the landscape. Industry and environmental monitoring agencies have questioned the ability to reclaim fen peatlands in the post-mine landscape. This study proposes a conceptual model to replace fen systems with fen peat materials supported by groundwater inflow from a constructed watershed. A numerical model is used to determine the optimum system geometry, including the ratio of upland to fen area, thickness and slope of sand materials, and thickness of peat and of the liner that would result in flows that sustain peat wetness to a critical threshold soil water pressure of −100 cm of water at a peat depth of 10 cm. We also test the sensitivity of the system to variations in the value and spatial configuration of the hydraulic conductivity (K) of locally available materials. The optimal conditions were achieved using an upland area at least twice that of the fen, underlain by a sloping (3%) layer of fine-grained material with hydraulic conductivity (K) of 10−10 m/s, that maintains lateral groundwater flow in a sand layer with K of 10−4 to 10−5 m/s. Using daily climate inputs that included 1998, the driest summer on record, the model suggests that adequate wetness can be sustained in the fen for the growing season, and that the extent of water table recession was similar to undisturbed systems during that period.

The non-invasive characterization of pumping-induced dewatering using ground penetrating radar

Abstract Ground penetrating radar (GPR) profiling is a non-invasive geophysical technique that has been used by Endres et al. [Ground Water 38 (2000) 566] to successfully image pumping-induced drainage in an unconfined aquifer. However, the drained water volume calculated from the GPR data was significantly less than the actual pumped volume. To investigate the reasons for this discrepancy, a seven-day pumping test and five-day recovery test was performed at Canadian Forces Base Borden in Ontario, Canada. A dense spatial coverage of GPR profiles was used to better quantify variations in drainage due to small-scale aquifer heterogeneity. In addition, a neutron moisture content probe was used to directly observe drainage and the soil moisture profile at a sealed well near the pumping well. Neutron logging indicated that the transition zone translated downward during pumping without undergoing significant extension. Comparison of the GPR- and neutron-derived transition zone drawdowns show nearly equal responses. Both of these observations support the hypothesis that the behaviour of the GPR reflection is an accurate measure of the transition zone response. In contrast, transition zone drawdown obtained from both GPR and neutron logging are significantly delayed relative to potentiometric drawdown, resulting in an extended capillary fringe. The drained water volume was determined from the GPR-derived transition zone drawdown data using a number of different approaches. Methods that incorporated information about spatial variations in drainage gave larger estimates of drained water volume; however, those estimates were still lower than the actual pumped volume. The unaccountable volume of water could be a result of several factors—aquifer heterogeneity may still not be adequately represented by the increased GPR coverage, and/or leakage from the underlying aquitard may be providing a significant volume of water.