H51C-0636
Midwinter Snowmelt Generated by Ground Heat Conduction: Implications for Catchment Hydrology
Ground heat conduction is commonly ignored in modelling of snowpack energy exchanges and snowmelt runoff due to its perceived insignificance relative to other energy sources. Snowmelt at the base of a snowpack was continuously measured during the winter of 2006-2007 with 4 m2 lysimeters at six different sites within a 3.5 km2 continental, mountainous catchment in southeast British Columbia. Soil wetness, soil temperature, and air temperature were also continuously measured at each site. Snowmelt during a three month midwinter period with sub-zero air temperatures ranged from 11 to 113 mm, comprising 3 to 38 % as much as the annual peak snow water accumulation. Given the lack of surface melt, this midwinter snowmelt was driven by ground heat conduction. Spatially, total melt was positively associated with shallow soil moisture content, likely because midwinter snowmelt maintained soil wetness at or near field capacity at three of six sites. There also was likely a positive feedback between soil moisture and melt rate, due to the association between soil thermal conductivity and soil wetness. It is hypothesized that midwinter melt caused by ground heat conduction may be important for enhancing catchment response by maintaining hydrologic connectivity between upslope areas and the channel network.
H51C-0637
Surface Temperatures, Groundwater Recharge Temperatures, and Noble Gases: An Investigation in the Wasatch Mountains, Utah
To understand better the systematics of noble gases in tracing groundwater recharge, we monitored dissolved noble gases in surface water and groundwater as well as temperature of the air, ground, and surface water in a high alpine basin in the Wasatch Mountains of Utah. Temperature probes were installed in local recharge and discharge areas to continuously monitor ground temperatures at 5 depths down to 1 m. Water temperatures are continuously being monitored in a small stream and a shallow well (depth ~1.5 m) located in the discharge area. Air temperatures are recorded at a meteorological station located within the basin. Noble gas and tritium samples were collected approximately every other week from February 2007 to present from both the stream and the well. Preliminary noble gas results show that groundwater recharge temperatures for samples taken from the well warmed by as much as 1.5 °C over a one month period from February to March, indicating that noble gases dissolved in groundwater exhibit temporal effects. When snow cover is present, ground and surface water temperatures are not influenced by fluctuations in air temperature. While ground temperatures at the recharge site did not record the spring snow melt event as expected, ground temperatures at the discharge site cooled by ~1 °C throughout the month of May, indicating possible movement of snow melt through the shallow aquifer. Following the snow melt event, ground and surface water temperatures mimicked the same warming and cooling trends found in the air temperatures, but were on average 5 °C to 10°C cooler than the air. We anticipate that continuous monitoring of temperatures and noble gases will offer insight into the dynamic interplay between air/soil temperatures, movement of water through the vadose zone, and recharge events in high alpine environments.
H51C-0638
A Field Test of Noble Gas Temperature Systematics from an Instrumented Monitoring Well: Implications for Gas Transport in the Capillary Fringe
Recent studies have found a significant bias of noble gas temperatures (NGTs) to values well below the average ground temperature at the water table (Hall et al., 2005; Castro et al., 2007). In order to test models that explain this bias, a new monitoring well has been drilled within ~30m of the well studied in Hall et al. (2005). The well is 5cm in diameter, has a total depth of 24.4m, is screened over the bottom 12.2m and has a typical depth to the water table of 13.1m. Preliminary measurements of CO2 concentration in the screened region within 0.5 m above the water table plus dissolved O2 in groundwater near the water table are consistent with total CO2 and O2 partial pressures of about 0.1 atm, in good agreement with the estimate made by Hall et al. (2005) with their oxygen depletion (OD) model for NGTs. An important feature of the OD model is that the extra noble gas partial pressures reduce the size and importance of the "excess air" component that forms the basis of all standard NGT models. Noble gases in water samples from eight different levels of the well were measured and the He concentrations were uniformly above those expected if groundwater were in equilibrium with the atmosphere. Indeed, the He concentration gradient is about 5 orders of magnitude lower than what would be expected if there was rapid gas diffusion through the unsaturated zone from air above a water table that is in equilibrium with groundwater. Standard models of gas transport suggest that gas diffusion can have extremely high tortuosity at the base of the unsaturated zone and this appears to be necessary to explain the apparently low upward helium flux. Given the potential for restricted transport within the gas phase and the relatively low apparent excess air needed with the OD model, it was decided to re-examine the Partial Re-equilibration (PR) model (Ballentine & Hall, 1999) in light of the possibility that the rate limiting step is diffusion within the gas phase rather than diffusion in water. A boundary layer model with diffusive re-equilibration of excess air at a rate proportional to D2/3 gives an excellent fit to the data of Hall et al. (2005), yielding correct NGTs and an improvement of the fit over the OD model with unfractionated excess air. Because gas diffusion is less mass dependent than diffusion in water and because this revised NGT model needs comparatively low values of excess air, predicted isotope ratio anomalies are low and within measurement error of actual Ar isotope ratios. Therefore, it may not be strictly necessary to use NGT models that are continuously within equilibrium when measured isotopic ratios are close to their atmospheric values. Ballentine & Hall, 1999, Geochim. Cosmochim. Acta, 63, 2315-2336. Castro et al., 2007, Earth Planet. Sci. Lett., v257, 170-187. Hall et al., 2005, Geophys. Res. Lett., 32, L18404, doi:10.1029/2005GL023582
H51C-0639
Water Transport Through the Unsaturated Zone and Shallow Groundwater Response Following Flood Irrigation in Three Different Soils of Northern New Mexico
Deep percolation from surface irrigation may be important for groundwater recharge in irrigated corridors. This is particularly important in irrigated valleys of northern New Mexico, where no or minimal chemicals are used in crop production. The study site is located at the New Mexico State University-Alcalde Sustainable Agriculture Science Center, in the agricultural corridor between the Rio Grande and a main irrigation ditch. The objective of this study was to characterize the continuum of water transport through the unsaturated zone and into the aquifer in different soil types. Six 12 m by 12 m infiltration plots were instrumented to measure, both the amount of water being transported through the unsaturated zone and the water level rise following flood irrigation. The infiltration plots were installed in three soil types (Fruitland sandy loam, Werlog clay loam, and Abiquiu-Peralta complex) that overlie a shallow aquifer that ranges from 1.5 m to 4 m depth from soil surface, depending on proximity to the river. These three soil types account for over eighty five percent of the total acreage in the agriculture corridor. Soil moisture, soil temperature, water level, and climatic data were collected to determine plot water budget, velocity of water transport through the soil profile, and water level response for the different soil profiles been evaluated. The model Hydrus was used for simulating the water movement through the soil and the water balance. Preliminary results showed a good agreement between field-measured and simulated results. Results of this study are being used to calculate the crop seepage contributions to the recharge of the shallow aquifer in a Rio Grande-main ditch irrigated corridor. Key words: Deep percolation, water level response, flood irrigation, water balance, Hydrus.
H51C-0640
Hydrological Observation on the Water Budget at the Cornfield in Northern Kyushu, Japan
If global warming will proceed constantly, we may face the crisis of decreasing available water. To use limited water resource for agriculture efficiently, it is important to understand the hydrological cycle and to analyze the water budget in agricultural land. Although there are various factors related to hydrological cycle in agricultural land, it is not well known how these factors contribute to the process of hydrological cycle. In order to reply to this problem, we observed the components on the water budget at the feed-crop cornfield in northern Kyushu, Japan. The study cornfield is on the pyroclastic flow deposits upland at the foot of Mt. Aso caldera. The observation period is from May to August, 2007 that corresponds to the period between seeding and harvesting of the corn. When corns were harvested, they were about 2.5m tall. During three months growing periods, it is thought that the content of the water budget was different at each growing stage. As corns grew taller and taller, corn leaves intercepted rainfall more and more to prevent the direct free-throughfall to the soil surface because of shapes of corn leaves. For this study, we defined three water budget equations. They were based on 1) distribution of gross rainfall through the corn canopy, 2) components of evapotranspiration, and 3) water budget around the soil surface, respectively. Because these equations must be correlated each other, we could evaluate the interception and the fluctuation of soil water content which are related above three water budget equations. Also we observed the fluctuation of soil water content during rainfall-infiltration process by using the electrical resistivity tomography (ERT). We also evaluate the water budget during the whole period between seeding and harvesting by using the all monitored hydrological elements; gross rainfall, throughfall, stem flow, interception loss, surface evaporation, transpiration, soil water content, surface flow, and infiltration.
H51C-0641
Modeling Vegetative Controls on the Water Balance in Shallow Desert Soils
Vegetative controls on the water balance in shallow desert soils near the Amargosa Desert Research Site, Nevada were investigated with transient, one-dimensional models. Movement of liquid water, water vapor, and heat were simulated in sparsely vegetated and non-vegetated soils. Dominant transport mechanisms simulated within the root zone (upper 100 cm) were liquid water flow subsequent to precipitation and thermal-water-vapor flow during dry periods. Liquid water fluxes were greater beneath non-vegetated soils. Thermal-water-vapor fluxes beneath vegetated soils were as much as 10 times greater than non-vegetated soils and up to 1000 times greater than liquid water fluxes during warm dry periods. Total head gradients beneath non-vegetated soils were predominantly downward while those below vegetated soils were seasonally trending with upward gradients during warm dry periods and downward gradients during cool wet periods. Moisture content beneath non- vegetated soils greatly exceeded residual moisture content observed and simulated beneath vegetated soils. Sustained root-zone moisture accumulation, which does not exist beneath vegetated soils, increases the potential for downward migration of water and contaminants.
H51C-0642
A Method for Computing Infiltration and Redistribution in a Discretized Moisture Content Domain
In an effort to improve the computational efficiency and robustness of one-dimensional vadose zone flow calculations in the development of large-scale coupled surface and subsurface flow interaction models, alternatives to the Richards' Equation (RE) are sought. This presentation will describe an infiltration and redistribution method that utilizes a discretized moisture content domain. The range of moisture content within a given soil is discretized into vertically continuous, interactive bins. The entry and vertical movement of water in each bin are simulated by wetting and drying fronts that move according to explicit infiltration approximations based on capillary and gravitational driving forces. Wetting front advances within bins create pore-water deficits that are satisfied by capillary-driven inter-bin flow. The method inherently provides numerical stability and robustness because it precludes the need to include the non-linear constitutive models estimated by other vadose zone flow approaches. Comparison of this method results with RE solutions and experimental data will be presented.
H51C-0643
Parameter Estimation for Infiltration using Double Ring Infiltrometers
Rainwater infiltration rates are critical to hydrologists' understanding of the behavior of a soil during rainfall events, as well as for predictive estimates of runoff potential. However, current infiltration studies fail to account for inherent procedural error associated with this technique. In August and September of 2007, infiltration rates of soils determined to be statistically similar on the basis of antecedent moisture, slope, soil texture, and vegetation were measured and analyzed. The rates were used to develop error parameters for standard use of double- ringed infiltrometers. Another set of statistically similar soils was infiltrated to establish the viability of these parameters. The objective of this study is to establish error parameters for use with double-ringed infiltrometers. Our hypothesis is that infiltration rates of soils of statistically similar antecedent moistures, vegetation types, and soil textures will fall within one standard deviation of the mean, accounting for a more uniform distribution rather than normal distribution.
H51C-0644
Estimation of Soil Hydraulic Parameters From Ring Infiltration Measurements by a Method Based on the Scaling of Dimensionless Numerical Solutions to the 3-D Axisymmetric Richards' Equation
A method is developed to estimate the soil hydraulic parameters using simple ring infiltration tests, assuming that the 3-D axisymmetric Richards equation governs the water movement within a soil. In this method, measured volumes of infiltration vs time, VI-t, are matched by scaled dimensionless volumes, VI*-t*, calculated numerically. To do so, the method utilizes a similar approach to that of the "Beerkan method" (Haverkamp et al., 1996) which assumes that the shape parameter(s) of the hydraulic conductivity and water retention curves, can be obtained from textural information and other easy-to-collect data. Then, we show that the two scaling factors making possible to pass from the dimensionless solution to the measured infiltration signal are only dependent on: i) the hydraulic conductivity at saturation (Ks), and ii) the normalisation parameter of the water retention curve (hae). We test the method numerically and for experimental data of two contrasting soils: a well graded silt and a structured coarse sand, both having null pressure heads at the surface and uniform initial water contents. The results indicate that good estimations of the parameters Ks and hae can be obtained, although they are dependent upon accurate estimations of the shape parameter(s).
H51C-0645
Characterizing Hydrological Processes in Vadose Zone by Direct Infiltration Water Sampling.
These days, planted forest mountainside was roughly maintained due to the population descent and small birth rate. Because thinning operation would delayed, forest was always dark and floor weed was rare. Management induced non point source pollution like surface soil erosion was suspected, however, we could not approach to the source with the stream water analysis. Therefore, direct soil water sampling device using glass fiber capillary force was developed to examine hydrological processes in watershed. In our design, water was collected just by the capillary force and let the excess water down through so that infiltration water was truly sampled and solute concentration kept the same quality as in soil water. The experiment was conducted at two neighboring Japanese cedar planted forest under different management, i.e., south slope was roughly maintained and west slope was well maintained by thinning operation. Load discharges were higher in south slope and lower in west slope. Infiltration water analysis revealed that ion concentration was gradually decreased at west slope, however in south slope, it dropped to lower level in soil water and increased again in stream water. The trend showed that soil buffering function was poor in south slope. Actually, disk permeameter survey revealed that hydraulic conductivity was small in south slope; TOC and biological activity were lower. This entire soil environment explained the water environmental differences in stream water. Because changes in soil environment affects water environment in the future, monitoring or examination of soil environment was considered as preventive measure for environmentally sound water and solute circulation in watershed.
H51C-0646
Site Specific Evaluation of Multisensor Capacitance Probes
Multisensor capacitance probes (MCPs) are widely used for measuring soil water content (SWC) at the field scale. Although manufacturers supply a generic MCP calibration, many researchers recognize that MCPs should be calibrated for specific field conditions. MCPs measurements are typically associated with small soil volumes, and are subsequently scaled up to the plot or field scale. Research is needed to understand how representative are these measurements for water monitoring studies that operate with the elementary area from one to tens square meters. The objectives of this study were: (a) to test the accuracy of SWC field measurements using generic and laboratory MCP calibrations; (b) to test applicability of a single MCP calibration for SWC measurements at different depths; and (c) to compare the accuracy of two and three-parameter equations using scaled frequency (SF). Four 1x1 m plots were equipped with MCPs to measure SWC at 9 depths at the OPE3 USDA-ARS research site at Beltsville, MD. Within each plot, three undisturbed soil cores were taken with a 100 cm3 soil auger. SWC sampling was made on three different dates when soil water contents were distinctly different. To compare MCP measurements with observed SWC, the SF was converted into SWC using: (a) the manufacturer generic calibration; and (b) calibration obtained in laboratory for a mesic Aquic Hapludult soil. Parameters of three different calibration equations were also obtained by fitting the equations to the water contents measurements at the plots. This fit was done: (a) for all observations regardless the depth, (b) for observations at each genetic horizon, and (c) for each depth separately. Results show that the manufacturer and the laboratory calibrations provided a satisfactory fit to the field-measured SWC at depths of 30, 40 and 50 cm. The fit was about two times less accurate at depths of 10, 20, 60, 70 80 and 90 cm. A minor improvement was obtained at depths of 10 and 20 cm after calibration equations were parameterized with observations from all depths pooled together. A significant improvement (P<0.02) in accuracy of MCP measurements for depths of 10, 20, 60 and 70 cm was obtained after SWC data were grouped into subsets either according to soil genetic horizon or according to depth. The three-parameter calibration equations did not improve accuracy compared to two-parameter calibration equations. Overall, more accurate measurements of SWC in a layered soil using MCPs could be obtained by employing site and depth-specific relationships between SF and plot-averaged water contents.
H51C-0647
2-D water flux using a penta-probe heat-pulse sensor: laboratory experiment and numerical evaluation
The fate of precipitation and snowmelt are not only important sources for water supply, but also crucial inputs to hydrological modeling and therefore measurements of subsurface flux in soils are of great interest. Recent work using heat-pulse measurements to assess water flux in soil have shown promise for determination of infiltration rates. Within the research realm, heat-pulse based sensors typically have one heater probe and one pair of upstream and downstream temperature sensing needles for flux determination. One-dimensional sensors may not be suitable for inferring hillslope percolation rates due to the liklihood of multidimensional water flux. Advanced heat-pulse based sensors with one heater probe and two pairs of orthogonally arranged temperature sensors have been proposed to measure horizontal and vertical fluxes simultaneously. In order to evaluate the penta-probe heat-pulse (PPHP) method a laboratory experiment was performed in conjunction with numerical modeling. A cylindrical flow cell comprised of an outer lexan wall and inner porous stainless steel wall was divided into four sections (i.e., two inlets and two outlets) leaving the center of the cylinder packed with porous media (glass beads). By adjusting pressure head at each inlet and outlet, a 2-D flow field was formed at the center where the sensor was located. We used the analytical model by Wang et al. (2002) to estimate 2-D water fluxes (e.g., vertical and horizontal) from the temperature increases measured by each pair of temperature sensing needles. The calculated 2-D flux was compared to the results of a CORE2D V4 (Yang et al., 2006) numerical model which simulated flow and heat transport in the flow cell. Water velocities derived from numerical simulations and temperature responses show good agreements.
H51C-0648
Rapid Streamflow Generation from Subsurface Flow
Traditional streamflow forecasting from snowmelt-dominated basins has been based on snowpack dynamics. A weakness of this approach is a failure to accommodate the increasingly common mid-winter rainfall events, which are often responsible for major flooding. We recently combined a snowmelt and soil water balance model in order to predict the spatial and temporal distribution of water flow through the root zone (through flow) in a snowmelt-dominated watershed (Reynolds Mountain in Idaho, USA). No attempt was made to calculate subsurface flow through the underlying fractured bedrock. We found: 1) a measurable soil water storage component and 2) extremely rapid streamflow response to through flow input even when the water source (snow drift) is hundreds of meters from the stream channel. The implication is that subsurface water throughout the watershed is interconnected and that streamflow response is insensitive to the spatial location of inputs. At present, it is difficult to know how far the second observation can be generalized, except to say that it is consistent with widespread observations that flooding events are often dominated by old, subsurface water that must be similarly interconnected.
H51C-0649
Rivulets Link the Hydrology of Soils to Hill-Slopes
Hydrologists dealing with run-off formation in small catchments often complain about much faster responses to precipitation at the catchment scale than what was predicted with soil hydrological approaches to infiltration and drainage. Constant velocity of the wetting front was observed, for instance, over a depth of 2 m and lasting more than 14 hours during an infiltration-drainage experiment in a tank that was homogeneously filled with sand. Constant velocity means neither acceleration nor deceleration of the moving water. This is only possible if the flow-driving and the flow-impeding forces continuously balance behind the wetting front. Rivulets are the conceptual units that dynamically balance forces during flow. They are considered thin streaks of flow that can be superimposed to rivulet ensembles and to water-content waves. Rivulets keep moving under the impact of gravity unless one of the following processes decelerates them: (i) increase of momentum dissipation due to reduced widths of the flow paths, (ii) capillarity-induced water sorption from the rivulets, and (iii) reduction of the rivulets' momentum due to the cessation of input. The velocities of the rivulets' wetting fronts collected so far in various systems like soils, unconsolidated sediments, fissured granites, and formations of chalk and karst, that extend over distances between 0.1 to 2000 m vary within about a factor of 50, thus offering a way of up-scaling soil hydrological processes to hill-slopes and small catchments that is based on rivulets.
H51C-0650
Soil Moisture Time Series Analysis for a Hillslope located Gwangneung National Arboretum
Understanding the hydrological processes at a hillslope scale can be achieved through intensive in situ monitoring of an intermediate hydrologic variable, soil moisture, during rainfall events. A soil monitoring system was installed to efficiently represent the spatial and temporal features of soil moisture for a hillslope located Gwangneung national Arboretum in South Korea. The soil moisture responses to sequential rainfall events were obtained as multiple time series. This paper attempts to explore an issue about how digital terrain analysis of many topologically based hydrology models can be addressed in terms of measured soil moisture histories. Time series analysis provides a systematic method of evaluating the stochastic characteristics of hydrologic variable. A derivation of the soil moisture transfer mechanism can be used as the physical basis of soil moisture time series analysis. After recording the soil moisture response patterns for a few consecutive rainfall events, a time series modeling procedure was applied to configure the characteristics of soil moisture. Characterizations of the variation in soil moisture variation were discussed through the interpretation of the time series models that were selected based on their terrain attributes.
H51C-0651
A Probabilistic Analysis of the Impact of the Topography on Slope Stability
This work investigates the influence of topography on hillslope stability in partially saturated soils, by using a distributed hydrological model, GEOtop [Rigon et al., 2006] coupled to a simple geo-mechanic model, GEOtop-FS [Simoni et al., 2007], which includes the theory of Lu and Likos, [2004]. GEOtop is a dynamic hydrological model, and describes the transient infiltration by solving Richards equation; GEOtop-FS uses the pore pressures computed by GEOtop to return, for every time step, probabilistic stability classes for any point in the catchment. Three variables are defined for each cell of the analyzed catchments: PFS = P[ FS < 1] [Duan and Grant, 2000] represents the stability range expressed in terms of failure probability; it can assume one of the following ranges of values 0-20% (very low probability of failure), 20-40% (low), 40-60% (moderate), 60-80% (high), 80- 100% (very high); α is the slope angle and mrph represents one of the three basic morphological attributes, convergent, divergent and planar, as derived from the planar curvature of the sites. We investigate the probability that a given range of stability/instability, in a given morphology, mrph, occurs at a slope angle α, which is given by p(α | PFS, mrph). In order to assess whether the slope angle is really a discriminant factor for slope stability, it is necessary to correlate the information provided by p(α | PFS, mrph) with the one provided by p(PFS | α , mrph), which describes the probability that a cell characterized by an α slope angle and by a morphology mrph, falls within a certain range of PFS. The analysis is performed using different, representative, soil types and three assumptions on soil depth: a uniform soil depth distribution, a soil depth calculated based on the topography derived from field geophysical measurements and a soil depth derived from the application of simple theories on soil-sediment evolution. The analysis shows that in partially saturated soils, where rainfall infiltration occurs mainly in the vertical direction, topography does not influence hillslope hydrology at local scale. However, when lateral redistribution occurs as a consequence of water table rise, some effects are evident at the catchment scale. The analysis suggests that planar and divergent sites tend to be stable, whereas for convergent sites instability seems to be evenly distributed among different stability ranges for slope angles close to soil internal friction angle.
H51C-0652
The Role of Vegetation on Elevated Soil Nitrate Concentrations in Northern Nevada
Soil profiles in the Spanish Springs area of Reno, Nevada have indicated high concentration of nitrate 2-3 meters below the soil surface. The main vegetation of the area is sagebrush (Artemesia tridentata) and rabbitbrush (Chrysothamnus nauseosus) with ground water depth varying from very deep (~ 28-30 m) to very shallow (~3m). We are investigating the interaction of vegetation with the elevated soil nitrate at these locations. Root samples were collected from 4 different locations (three sites with deep ground water and one with shallow ground water, with rabbitbrush population) 15 cm depth up to a depth of 45-60 cm using a root auger. Physio-chemical properties of soil were analyzed and leaf elemental status was determined. Our initial analyses indicate that the higher C:N in conjunction with shallow groundwater supports a higher root density with greater denitrification potential. Also indicated from this initial analysis is the absence of significant distinction between sites in relation to the sagebrush (Artemesia tridentata) and rabbitbrush (Chrysothamnus nauseosus) plant leaf nitrate concentrations. Preliminary conclusions indicate that nitrate levels in soil are directly influenced by groundwater levels which also appear to directly influence the C:N in the soil and the potential denitrification. Future work will include lysimiter experimentation within the laboratory which will focus on understanding the nitrate uptake by the local plants. Seasonal change in plant nitrate uptake and denitrification of the soil is also planned for the selected sites.
H51C-0653
Nutrient Retention by Spodosols in an Impacted Isolated Wetland
Agricultural soils within the Lake Okeechobee drainage basin can pose a nutrient management challenge because of their low nutrient retention capacity and hydrological setting. While hydrological restoration of isolated wetlands within the hydroscape may mitigate some phosphorus (P) loss from pastures, P retention mechanisms within these wetlands are difficult to control, and depend primarily on vegetation, antecedent soil conditions, and climate. The objective of this research was to: (i) Characterize soil properties responsible for nutrient retention such as, texture, bulk density, particle density, porosity, soil organic matter (OM) and total P (TP) concentrations, and distinct hydric soil indicators; and (ii) Use steady-state diagenetic equations to describe the effect of physical and biological processes using a low authigenic phosphate (PO4) precipitation rate on measured soil pore water profiles. Results indicate that the spodosols exhibit a modified A-horizon within the upper few centimeters, a sandy illuviated E-horizon, a loamy Bt-horizon interspersed with Fe-Al redoximorphic features, a Bh-horizon with slightly higher soil TP concentrations, contained over a high density-low conductivity argillic clay horizon, and at least three distinct hydric soil indicators observed within the top 15 cm. An inverse relationship between porosity (17-50 %) and bulk density (0.8-2.2 g.cm-3) was attributed to physical stresses such as tillage and compaction. Based on the strong correlation between soil OM and TP we conclude that a large fraction of the soil P is organic, and not derived from inorganic agricultural amendments. However, the low sorption capacity of P by spodosols can pose a threat to downstream aquatic systems through ground water pathways, restricted only by the presence of semi-confining clayey horizons (below 120 cm depth) that may behave as preferential flow paths, and possibly retaining some of the P from solution onto its reaction sites. Lastly, pore water PO4 concentrations could not be explained using steady-state diagenetic equations in the upper 30 cm, however, below 30 cm pore water PO4 concentrations could be a result of antecedent soil conditions, and a slow precipitation rate of phosphatic minerals.
H51C-0654
Three-dimensional modelling of water flow through a heterogeneous vadose zone
Stormwater infiltration basins have become a common alternative practice to traditional stormwater pipe networks in urban areas. They are often built in permeable subsurface soils, such as alluvial deposits. These sedimentary deposits are highly heterogeneous which generate preferential flow paths that may cause non- uniform transport of contaminants at great depths. Thus, the hypothesis of using homogeneous deposit cannot be considered valid at the scale of infiltration basin. Therefore, it is required a fine scale three-dimensional numerical simulations of fluid flow and solute transport to understand how subsurface heterogeneities affect fluid flow and contaminant transfer. The aim of our study is to improve the understanding of the water flow mechanisms occurring at the lithofacies scale in the heterogeneous vadose zone of infiltration basin. The studied basin, located in the eastern suburbs of Lyon (France), is built in quaternary glaciofluvial deposits. Results from previous studies show that the glaciofluvial deposit is composed of 4 main lithofacies: sands, bimodal sandy gravels, heterometric sandy gravels, and openframework gravels. These lithofacies are organized in braided river deposits architecture. Ground-Penetrating Radar was assessed to characterize the structural units of the studied deposits. Hydraulic properties of each lithofacies were characterized on analogous modern glaciofluvial lithofacies, leading to the definition of hydrofacies. Geophysical measurements (Ground-Penetrating Radar and Electrical Resistivity) performed on a limited zone of the basin was interpreted to define a lithofacies distribution. Hydrofacies properties were used to build a three- dimensional hydrostratigraphic model of the glaciofluvial deposit. This model was implemented in the finite element program Hydrus3D. The results show that openframework gravels cause capillary barrier effects leading to funneled flows, i.e. non- vertical water flows along lithofacies interfaces. These results may explain the heterogeneous repartition of heavy metals concentration measured in other areas of the basin.
H51C-0655
A Linking Test That Explores The Non Uniqueness Of Soil And Vegetation Parameters Of A Unsaturated Flow Model
Groundwater recharge can be accurately computed by using models that solves the Richards' equation. A drawback of such models is that they require a considerable number of parameters that need to be determined: Mualem-van Genuchten, interception and crop factor. It is preferable to estimate these parameters indirectly from in situ measurements since values retrieved from sample analysis poorly characterize field conditions. Groundwater recharge can be modelled once the hydraulic and the vegetation parameters have been optimised simultaneously. The optimisation can be achieved by matching solely observed and simulated time series of soil moisture profiles. A good match between simulated and measured soil moisture can be obtained but the parameters are non-unique, leading to the problem of equifinality. We found that a unique groundwater recharge could be obtained only if the vegetation parameters were known. We investigated why the optimum hydraulic parameters were not unique by using the Linking Test. The Linking Test is a simple inverse method developed by the authors that takes into account the non uniqueness of the parameters. The Linking Test shows that very different soils could give similar groundwater recharge. This test also provides all the different combinations of the hydraulic parameters that give similar groundwater recharge. The Linking Test showed that the hydraulic parameters were linked due to "range" and "compensation" equifinalities. We showed that only 2 functional parameters are required to get a satisfactory simulation. It should notably simplify the determination of the hydraulic parameters in a region where various soil types are encountered and thus help in designing upscaling methods for these parameters. Other foreseen applications of knowing how the hydraulic parameters are linked are an improvement (1) of the sensitivity analysis to the hydraulic parameters, (2) the uncertainty analysis of the groundwater recharge by using soil water flow models, (3) the interpretation of the physical meaning of the hydraulic parameters determined by inverse modelling. The Linking Test could be applied to any inverse problem prior to modelling to determine its feasibility and to establish if the non-uniqueness of the parameters is of concern for the problem to be solved.
H51C-0656
Simulation of Soil Moisture Development in Flood Protecting Earth Dams
Extreme floods represent an increased risk for urban areas and agriculture. Time to time the protective earth dams are destroyed by a suddenly increased amount of water with destroing or even cathastrophic consequences. A numerical study of the soil moisture development within the earth body during the flood is simulated under a selection of boundary conditions. Several soil materials are considered. Simulations are performed firstly for homogeneous materials using the 2D single domain approach, in the second step the dual permeability simulations are done assuming inhomogeneities in the construction which may lead to the preferential flow. Results for saturated as well as for unsaturated part of the dam are analyzed. Using the appropriate simulation model may help to design safer flood dams and evaluate the reason of possible failures to prevent future disasters. The research has been performed in the frame of research project VZ 04 CEZ MSM 6840770005.
H51C-0657
Modeling Plant-Scale Root Zone Water Dynamics in an Oak Savanna
Study of water exchange between soil, plants, and the atmosphere in response to seasonal or periodic droughts is critical to modeling the hydrologic cycle and biogeochemical processes in water-controlled ecosystems. The difficulties in such studies arise from insufficient understanding of the complex interactions between the various processes and their scale-dependence. The purpose of our study is to establish and calibrate a plant biophysical model that couples plant-soil and plant-atmospheric interactions to calculate the water exchange through the soil-plant-atmosphere continuum at a plant scale (~10 m2), with the regulation of root water uptake and evaporative fluxes by water deficits and climatic conditions explicitly considered. The complexity required for modeling water dynamics at the plant scale is investigated in this study. We start with coupling a big-leaf biophysical model with a bucket soil water balance model, with soil water loss regulated by soil water availability in a linear fashion. The alternative biophysical models with increasing complexities include the dual-source model that divide the canopy into shaded and sunlit parts and a multi-layer 1-D model with sophisticated radiation transfer and energy balance modules. The level of detail in subsurface water dynamics is adjusted by changing the dimensionality of the Richard's equation. The impact of soil water availability on water loss is modified to a nonlinear pattern as desired. The models are calibrated and compared using a cluster of measurements collected on single trees, which includes multiple soil moisture probes that monitor soil moisture profile vertically and laterally and sap flow sensors at different tree heights for measuring tree transpiration. This study forms the basis for scaling up the water dynamics to a stand scale (~100 to ~10000 m2) or other larger scales.
H51C-0658
Modeling The Water Table In The Middle Rio Grande River Riparian Corridor
The Middle Rio Grande River (MRGR) is the main source of fresh water for the state of New Mexico. An arid area with low water resources created a situation where water is extensively diverted or stored to supply the high demand for municipalities and agricultural activities. The extensive water diversions over the last few decades has affected the composition of the native riparian vegetation such as cottonwood and coyote willow and enhanced the spread of invasive species harmful to the river system such as Tamarisk and Russian Olives. The river aquatic system has also been badly affected. The need to study the river hydrological processes and their relation with its health is important to preserve the river ecosystem. The water table within the riparian zone is intrinsically connected to the flows in the river. Large withdrawals of water by Tamarisk affect the surface flows, which coupled with the large diversions for irrigation result in a complicated river management problem. In this paper we describe the methodology used to spatially model the water table depth between the river and the adjacent drains parallel to the river. Water table readings are used to check the model. Evapotranspiration by the riparian vegetation is estimated and included in the soil moisture balance. The model runs as an application in ArcGIS. Spatial layers include soils and riparian vegetation maps obtained from the classification of airborne high resolution multispectral imagery.