H21B-0817 INVITED
How much water forms in the Southern Ocean? A maximum entropy approach to global water mass analysis.
A new water mass analysis method is developed to quantify the fraction of the interior ocean volume that was last ventilated in different parts of the surface ocean. The new method is formulated in terms of a deconvolution problem in which tracer observations are used to constrain the Green function, 𝒢, for propagating prescribed surface tracer concentrations into interior concentrations. The deconvolution is regularized by maximizing the information entropy of 𝒢 subject to the constraints provided by observed surface and interior tracer concentrations. The new algorithm can be used to analyze combined steady and transient tracer data, and can incorporate in a straightforward way prior information about 𝒢 obtained from a dynamical circulation model. Applied to hydrographic data from the World Ocean Atlas (2005) the analysis suggests that approximately 70% of the total ocean volume below 1625 m is ventilated from the Southern Ocean.
H21B-0818
Idealized Tracer Transport Models with Time-Varying Transport: Applications to Ocean boundary-current
One-dimensional advection-diffusion and advection-diffusion-dilution (or "leaky-pipe") analytical models have been widely used to interpret a variety of geophysical phenomena, such as the propagation of tracers (e.g., CFCs) in the North Atlantic Deep Western Boundary Current (DWBC). In all applications we are aware of, transport is assumed to be in steady state. Here, we relax this assumption, and solve analytically for tracer signals in both models with small-amplitude time-varying transport. For a tracer boundary condition that steadily increases in time, the resulting tracer field exhibits fluctuations due to the transport acting on the gradients induced by boundary forcing. We compare the transport-induced tracer fluctuations to "propagated fluctuations" occurring in steady-state models when the tracer at the boundary is fluctuating in time. Using coefficients fit to DWBC tracer observations we find that in the North-Atlantic propagated tracer fluctuations are larger but by the sub-tropics transport-induced fluctuations dominate. This is in contrast to the common that subtropical and tropical DWBC fluctuations in tracer such as CFCs, temperature and salinity are propagated signals from the northern formation region. However, even the transport-induced fluctuations are smaller than observed sub-tropical DWBC CFC fluctuations.
H21B-0819
An Estimate of the Global Ocean Transit Time Distribution Computed With an Eddy Rich General Circulation Model
Transit time distributions (TTDs), or age spectra, provide a powerful conceptual framework for evaluating and interpreting the advective-diffusive transport properties of geophysical flows. Most calculations and practical applications of TTDs to date have considered only the case of steady circulations with parameterized turbulent mixing. We have computed the first decades of the global ocean TTD for the unsteady, but statistically stationary, flow using an ensemble of impulse boundary propagators in an eddy-rich general ciculation model. In this presentation we will provide an overview of the geographical and temporal structure of the computed TTD, consider the convergence properties of the TTD as a function of ensemble size, space-time coarsening, and lag time, and consider the utility of TTDs as a metric for testing eddy mixing parameterizations in coarse resolution models.
H21B-0820
The Power of Climatological Tracer Observations to Constrain Ocean Transport Pathways: Adapting Hide's Theorem to an Inverse Model
The contributions of surface source regions to the composition of any water-mass in the oceanic interior can be attempted using observed tracer distributions. Previous studies typically proceeded by assuming a steady-state and specifying that only a small number of predetermined source water masses fill the deep ocean, making the solution unique so long as the number of observed tracers equals the number of specified sources minus one. Such an approach artificially limits the range of possible solutions, however, and it is preferable to admit each surface point of the ocean as a potential source. Here we take advantage of the fact that, under steady-state conditions for conservative tracers, the properties of any subsurface location must be a combination of the adjacent locations. This is an adaptation of Hide's Theorem from meteorology, namely that no local minima or maxima of a tracer may exist in the interior of the ocean. Therefore, for an ocean divided into grid boxes, only five tracers (and mass conservation) are needed to unambiguously determine the relationship between each location and its neighbors. Through inverting the relationship between every interior point and its respective neighbors, it is possible to estimate the transport pathways throughout the entirety of the world's ocean, and thereby trace back to the source regions which contributed to any given grid box. We invert for the transport pathways of the modern ocean using climatologies for temperature, salinity, oxygen-18/oxygen-16 isotope ratio, phosphate, nitrate, and oxygen --- a collection which we find to be adequate to obtain a unique solution after amending the previous method to account for nonconservative tracers. Thus, we can diagnose the pathways of all source regions into the interior ocean. In other words, an observationally-constrained steady-state ocean transport operator is found and its uncertainty can be calculated. The method is exquisitely sensitive to departures from steady-state behavior, which is not surprising given that this assumption clearly does not hold at all space and timescales.
H21B-0821
An Ensemble of CFC-11 Tracer Distributions From a Global Eddying Ocean Circulation Model
CFCs, introduced into the atmosphere in the 1930s, have now entered much of the world ocean, where they are advected as a passive tracer. They have been widely used both as a fluid tag to estimate the time elapsed since water last saw contact with the atmosphere, and as a metric for assessing model skill. Results will be shown from a global eddying ocean model simulation of ensemble CFC-11 members and transit-time distributions. Changes in pCFC age expected from internal ocean variability will be assessed, and it will be shown how CFC-derived ages change spatially and temporally. pCFC ages will also be compared with the mean age derived from the transit-time distribution. Finally, implications of eddy-induced variability in the oceanic CFC distribution for comparison with observations will be discussed.
H21B-0822
Tracers vs. trajectories in a coastal region
Two different methods of estimating the water exchange through a Baltic coastal region have been used, consisting of particle trajectories and passive tracers. Water is traced from and to a small discharge region near the coast. The discharge material in this region is treated as zero dimensional particles or tracers with neutral buoyancy. The real discharge material could be a leakage of radio-nuclides through the sea floor from an underground repository of nuclear waste. Water exchange rates between the discharge region and the model domain are estimated using both forward and backward trajectories as well as passive tracers. The Lagrangian trajectories can account for the time evolution of the water exchange while the tracers give one average age per model grid box. Water exchange times such as residence time, age and transient times have been calculated with trajectories but only the average age (AvA) for tracers. The trajectory calculations provide a more detailed time evolution than the tracers. On the other hand the tracers are integrated "on-line" simultaneously in the sea circulation model with the same time step while the Lagrangian trajectories are integrated "off-line" from the stored model velocities with its inherent temporal resolution, presently one hour. The sub-grid turbulence is parameterised as a Laplacian diffusion for the passive tracers and with an extra stochastic velocity for trajectories. The importance of the parameterised sub-grid turbulence for the trajectories is estimated to give an extra diffusion of the same order as the Laplacian diffusion by comparing the Lagrangian dispersions with and without parameterisation. The results of the different methods are similar but depend on the chosen diffusivity coefficient with a slightly higher correlation between trajectories and tracers when integrated with a lower diffusivity coefficient.
H21B-0823
Modelling water exchange between Baltimore Harbor and Chesapeake Bay using artificial tracers
The contaminants discharged from adjacent industries and densely populated areas have a great threat to the water quality and ecosystem in Baltimore Harbor and Chesapeake Bay. Variety of physical processes complicate quantifying the contribution of dominant sources and processes that lead to the environmental problem in Baltimore Harbor. In order to understand the contaminants and water exchange between Baltimore Harbor and Chesapeake Bay, the conservative artificial tracers are used to investigate the transport processes through a three-dimensional hydrodynamic model (CH3D), in which the three-layered circulation pattern in Baltimore Harbor are well reproduced. Several numerical experiments are performed to investigate the variability of the water exchange in response to the external forcing and to quantify the fate and transport timescales of those biogeochemical substances that released from different source regions. The results indicate seasonal variations of the water exchange are mainly controlled by Susquehanna River discharge. It takes less time in the flooding season (about 35 days) than in the drought season (about 115 days) for the dissolved substance to be transported from Susquehanna River into the Baltimore Harbor. In contrast, it takes more time in the flooding season (about 65 days) than in the drought season (about 45 days) for the dissolved substance to be transported from the lower Chesapeake Bay into the Baltimore Harbor (reference to the releasing spot). The residence time of the water in Baltimore Harbor are investigated under different hydrological conditions. It shows high correlation with the flow structure in the Harbor. The typical synoptic events are also examined in order to estimate the direct response timescales of the water exchange processes between the Harbor and the mainstem of the Bay. The results, from the environmental perspective, provide useful information for understanding the transport processes of the dissolved biogeochemical substances between Baltimore Harbor and Chesapeake Bay.
H21B-0824
A Model Study of the Variation of the Water Age and Long-term Transport Timescale of the Changjiang Estuary, China
Abstract Changjiang Estuary is one of the most important waterways in China. To meet the rapidly economic growth in China, a pair of 50 km long dike and 19 groins along the waterway has been constructed in the North Passage, one of the three waterways connecting the estuary to the open sea. With the modification of the waterway, the natural flow divergence between the South Passage and North Passage has been altered. Consequently the circulation patterns have been changed and multiple turbidity maximums have been observed in summer. To understand the influence of change of circulation pattern on the transport processes due to human activities in the estuary under different hydrodynamic conditions, a three- dimensional transport model with realistic forcings was applied to the Changjiang estuary to determine the change of long-term transport timescale. The variations of water age distributions under different hydrological and stratification conditions are investigated. The model results of the residence time and water age distributions in the Changjiang estuary exhibit strong seasonal variation in transport of water mass and circulation pattern. It suggests that the manmade construction increases estuary stratification, thus enhances the gravitational circulation significantly. Consequently, the transport timescale is changed. Contributions of three major physical transport mechanisms, tide, river discharge, and wind, to transport timescale are compared and quantified. The results indicate that freshwater inflow is one of the dominant factors controlling transport timescale in the Changjiang estuary. Keywords: Age of water; Residence time; Stratification; Changjiang Estuary; Numerical modeling.
H21B-0825
Field Evaluation of Travel Times and Flow Mechanisms in the Mississippi Delta Vadose Zone Using Tracers
In the Bogue Phalia basin in the Delta region of Mississippi, as in many farmed areas, intensive application of agricultural chemicals has led to their detection in surface and ground water; however, contributing unsaturated zone processes are not well understood. The fine textured soils often exhibit surface ponding and runoff after irrigation and rainfall as well as extensive surface cracking during extended dry periods. Fields are typically land-formed to promote surface flow into irrigation ditches and streams that feed into larger river ecosystems. Downward flow of water below the root zone is considered minimal; regional ground- water models predict only 5 percent or less of precipitation recharges the heavily-used alluvial aquifer. In this study we assessed transport within and below the root zone of a fallow soybean field with a 2-m ring infiltration test including tracers and subsurface instrumentation for sampling and for measuring water content and matric potential. Seven months after tracer application, we collected 47 continuous cores for tracer extraction to define the extent of water movement. Water movement was rapid below the pond, traveling up to 0.21 cm/s, indicating the importance of vertical preferential flow paths. Lateral flow of water at shallow depths was extensive and spatially non uniform, reaching 10 m from the pond within 3 months. Within 2 months, the wetting front reached a textural boundary between the silty soil and sandy alluvium at 5 m. The aquifer was historically confined by the silty material prior to extensive irrigation pumping that has lowered the water table from about 2 m to 12 m. Preliminary results indicate that after 7 months any water breaking through the soil-alluvium boundary, which now acts as a capillary barrier within the vadose zone, likely does so as unstable finger flow (fine-over-coarse layering has been observed in lab studies to initiate fingering) which is difficult to detect with point measurements. Preferential flow in the vadose zone may contribute to contaminant transport in both surface water, through shallow lateral flow to irrigation ditches, and ground water, through vertical finger flow.
H21B-0826
A Field Investigation of Matrix Diffusion Under Variably-Saturated Conditions: Implications for Contaminant Transport in the Vadose Zone of the Eastern Snake River Plain, Idaho
We present results from a field-scale infiltration tracer test involving two non-reactive solutes with different free-water diffusion coefficients to demonstrate that matrix diffusion is an important transport process in the variably-saturated, fractured basalt of the Eastern Snake River Plain (ESRP) vadose zone. Although diffusive mass transfer between a variety of flowing and stagnant fluid domains (matrix diffusion) has been postulated as part of the conceptual model of flow and transport in the ESRP basalt this is the first study to demonstrate this process experimentally. The shallow subsurface of the ESRP is unsaturated except where contacts between units of contrasting permeability give rise to areas of local saturation (perched water). Tracer breakthrough curves from wells completed at the base of the surficial alluvium did not exhibit a matrix diffusion signature, but breakthrough curves in two wells completed at the base of the basalt layer that underlies the alluvium had clear matrix diffusion signatures. Based on multiple lines of evidence, we believe this matrix diffusion occurred near the base of the basalt layer where water was perched and diverted horizontally through a rubble zone in the basalt directly above a low-permeability sedimentary interbed. Tracer breakthrough curves obtained from a suction lysimeter installed near the bottom of the interbed and almost directly below one of the basalt wells displayed a matrix diffusion signature that was consistent with the signature in the basalt. The matrix diffusion parameters deduced from analyses of the breakthrough curves in the two basalt wells and the lysimeter were in good agreement with each other, suggesting that these parameters may be broadly applicable to fractured basalts within the shallow ESRP vadose zone.
H21B-0827
Perchlorate as a Ground-Water Tracer Along the Lower Colorado River
Anthropogenic perchlorate was first observed in the lower Colorado River (NV and AZ) in 1997. The
perchlorate source was traced upstream from Hoover Dam and Lake Mead to Las Vegas Wash. Perchlorate
migrated through the local surface- and ground-water systems to the Wash from nearby manufacturing
facilities in Henderson, NV, which had been operating since the 1940s. The Nevada Division of Environmental
Protection (NDEP) began monitoring perchlorate in the lower Colorado River at Willow Beach, located about
18 km south of the Dam, in 1997. A 3 μg/L reduction was observed at Willow Beach in 2003-2004,
coincident with remediation at the Henderson site in 1999-2004. This observed decrease indicates that the
effects of remediation rapidly propagated through the surface-water system below the Dam.
In July 2008 water samples were collected and analyzed for perchlorate from eight springs along the lower
Colorado River below Hoover Dam, from a discharge tunnel in the country rock at Hoover Dam, and from
Lake Mead (above and below the thermocline). Lake Mead water collected above the thermocline east of
Sentinel Island contained 3.9 μg/L perchlorate, while water below the thermocline contained 1.8
μg/L. Perchlorate concentrations were lower than the 2 to 4 μg/L quantitation limit for the six
springs located more than 2 km south of the Dam. Samples from Pupfish Springs, about 0.9 km south of the
Dam, contained 6.4-6.8 μg/L perchlorate. Water collected from the discharge tunnel in the Dam
contained 8.2 μg/L perchlorate.
Perchlorate concentrations observed at Pupfish Springs and the discharge tunnel in the Dam in 2008 are
similar to those reported downstream at Willow Beach prior to 2003-2004 by NDEP indicating that the ground
water travel time from the Dam to Pupfish Springs is between 4 and 70 years and the maximum flow velocities
are between about 13-200 m/y. These rapid velocity estimates suggest that faults and fractures in the area
are an important control on discharge points in the vicinity of the Dam. The presence of perchlorate at two
sites and absence or background concentrations at the other sites indicates the presence of multiple flow
paths between the perchlorate source and the springs near the Dam or that the flow paths are more complex
than previously thought.
http://nevada.usgs.gov/water/projects/blackcanyon.htm
H21B-0828
Tracer discharge testing and modeling used to assess the dynamics of surface water - ground water exchange across a range of discharges
We explored surface water - ground water (hyporheic) exchange processes in multiple reaches of two streams in central coastal California, the Pajaro River and Corralitos Creek, using 28 tracer discharge tests at flow rates of 100 to 1200 L/s. Many earlier studies have quantified hyporheic exchange in streams using tracer discharge testing, but most studies involve a smaller number of tests, often conducted at similar discharge rates. In addition, most earlier studies were conducted in discharge-neutral or gaining streams, whereas both the Pajaro River and Corralitos Creek are losing streams. During our tests, two tracers were injected at a constant rate at an upstream location, and breakthrough curves of tracer passage were measured at one or more downstream locations. Observational data are fit to results of numerical calculations that represent transport, storage, and exchange processes, using an optimization scheme to infer quantitative parameters that describe these processes. Data were analyzed to resolve transport, storage, and exchange metrics as a function of location and channel discharge, including the dependence of three length metrics on discharge: dilution length, loss length, and exchange length. Within one section of the Pajaro River where channel losses were small, the dilution length and storage length were found to be insensitive to channel discharge, whereas the dilution length increased with discharge (longer lengths indicate less seepage). In another section of the Pajaro River, where channel losses were greater, both the loss length and the dilution length increased with discharge, whereas the storage length decreased with discharge. In contrast, tracer discharge tests of two sections of Corralitos Creek did not suggest systematic changes in length metrics with discharge. The observed differences in discharge dependence of length metrics in these two streams may result from fundamental differences in channel morphology and sediment load. These results suggest that repeated tracer discharge testing may be helpful for resolving exchange processes in systems where there are functional dependences on channel discharge or other parameters that vary throughout the year.
H21B-0829
Residence Time Distribution in Dynamically Changing Hydrologic Systems
Spatial and temporal variability of weather and climatic forcings induce a dynamic response in hydrologic systems. Regional groundwater systems and stream hyporheic zones are examples of hydrologic systems driven by forcings varying at several time scales, such as daily, seasonal, interannual, decadal and longer. Hydrologic systems are characterized by flow paths and residence time distributions. Residence times vary in space, with positions further along flow paths exhibiting older ages. If the hydrologic flow system is in steady state the flow paths do not change in time and water present at a given point has a stable residence time distribution. But hydrologic flow paths and residence times can change dynamically with weather and climate temporal variability. Traditionally, this dynamic response is ignored and modeled and observed residence times are evaluated as if the flow was in steady state. A finite element scheme is used to model the transient flow and transport of an ideal tracer into a Thothian-like domain, and to illustrate the effect of dynamically changing systems on residence-time estimation.
H21B-0830
Using Multiple Environmental Age Tracers to Investigate Interactions between Hydrothermal and Shallow Local Systems in the Norris Geyser Basin Area, Yellowstone National Park.
Multiple age tracers are used to constrain the evolution of groundwater and interactions between shallow local flow and the hydrothermal system in the area surrounding Norris Geyser Basin in Yellowstone National Park. Springs, ranging in temperature from 4°C – 92°C, were sample for dissolved chloroflourocarbons (CFC-11, CFC-12, and CFC-113) concentrations and tritium concentrations. Sample springs had a wide distribution of elevations, aspects, and volcanic deposits. CFC concentrations indicate a short (< 50 years) mean residence time for the shallow, cool-water system that circulates through the rhyolite and tuff flows adjacent to the geyser basin. Mixing processes and gas phase stripping can be assessed using the three different time dependent input curves for CFC-11, CFC-12, and CFC-113. Comparison of CFC and tritium concentrations provides verification of CFC ages and is an additional measure of mixing processes. All cool samples (0-20°C) contain CFC's and tritium, and indicate little to no mixing of pre-modern waters, with a mean CFC age of ~ 40 years. The relative concentrations of CFC's and tritium reveal mixing processes in the warm and thermal waters. Springs warmer than 20°C show increased mixing of pre-modern water with increasing temperature. Thermal waters (50 – 90°C) are well described by a binary mixing model of modern and pre-modern waters with minimum pre-modern fractions of .75 to .9. The use of multiple age tracers allows for interpretation of different age stratification and mixing models, and provides insight into other processes affecting the groundwater system such as phase distribution and boiling. Age tracers allow for the integrated study of the hydrothermal and local groundwater systems in the Norris Geyser Basin area and help constrain interactions between the two using non-invasive techniques.
H21B-0831
Measuring Pore Water δ2H and δ18O Values as Natural Tracers for the Migration of Under-Saturated, Meteoric Water Across a Mudstone Confining Unit Using the Radial Diffusion Method
The potential migration of under-saturated brines into formations containing otherwise saturated brine pore water is of importance because of the capacity of such brines to dissolve evaporites. The Williston Basin is an intracratonic sedimentary basin comprised of numerous aquifers and aquitards whose dissolved constituents tend to increase with depth in Southern Saskatchewan. At the study location a thin mudstone confining unit separates an overlying porous zone, containing under-saturated brine with δ2H and δ18O values that fall on the meteoric water line, from the underlying Dawson Bay Formation. Here the Dawson Bay Formation is near NaCl saturation and the δ2H and δ18O values indicate mixing between meteoric water and more evolved basin brines. Five core samples, between 10 and 16 cm long and 6 cm in diameter, spanning 20 m of formation, were used to create radial diffusion cells. A 1.9 cm diameter reservoir was drilled along the axis of each core and filled with 2H and 18O enriched deionized water and allowed to come to equilibrium with the pore water. Following equilibrium, the entire reservoir was sampled and analysed for δ2H, δ18O and major ion geochemistry. A semi-analytical model was used to determine the effective diffusion coefficients and effective porosity of each core. Results from pore water analyses are used to create a depth profile of pore water δ2H, δ18O and NaCl saturation across the confining unit. Based on the analyses of the diffusion experiments, numerical simulations of the transit times of under-saturated brine across the confining unit were conducted. These simulations indicate that although some migration would be expected under natural conditions, if a downward gradient is imposed the rate of migration is augmented. This conclusion has significant implications for the interpretation of regional scale flow dynamics, as well as the possible dissolution of evaporites at depth.
H21B-0832
Using Nanotechnology to Identify and Characterize Hydrological Flowpaths in Agricultural Landscapes
Applying the power of nanoscale technology to answer landscape-scale questions constitutes an exciting new frontier in science and engineering. In this project, we propose a possible method of reducing the "nonpoint" problem associated with nonpoint source (NPS) pollution, a problem that has hampered agricultural sustainability and water quality protection for decades. We are developing superparamagnetic polylactic acid (PLA) microspheres incorporating DNA "nanobarcodes" as potential tracers. The eventual goal of this project is to develop technologies for identifying and characterizing different flowpaths at field and watershed scales by using multiple sets of polymer microspheres, each coded with unique DNA sequences, of which there are essentially limitless combinations, i.e., many flowpaths can be uniquely coded. Our ultimate vision is to have the capacity of introducing microsphere-encapsulated DNA at different points in a watershed and collecting these microspheres elsewhere in the watershed; using quantitative, real-time polymerase chain reaction targeted at the specific DNA, we would be able to determine the hydrological linkages and transport times between the collection point(s) and the points of DNA introduction. The potential advantages of this nanotechnology strategy compared to conventional tracers are the elimination of background interferences, the ability to segregate superimposed flowpaths through the design of strictly unique DNA tracers and the biodegradability of the tracers. This presentation highlights recent advances, new challenges, and potential applications for this tracer technology.
H21B-0833 INVITED
Effect of transfer time correlations on the transfer of reactive substances through a series of media with different reactivities
Transfer time distributions of inert substances are often used to characterize the advective transfer through environmental systems. For linear reaction equations and for media with homogeneous reactivities, transfer distributions of reactive substances may be derived in straightforward manner from the inert substance transfer distributions. However in a lot of systems, reactivities vary in different parts of the system. Transfer of substances in such systems is commonly described using a convolution of transfer functions through the different subsystems. The transfer function of reactive substances in a subsystem is derived from the reactivities and the inert tracer transfer functions in the subsystem. The latter may be derived from a deconvolution of the transfer distributions that are observed in different combinations of subsystems. However, this approach often presumes no correlation of transfer times or unconditional transfer distributions in the different subsystems. We used reactive transport of organic substances in soils as an example for which this assumption is generally not valid. In soils, reactivities (degradation rate and sorption constants) of organic substances decrease considerably with depth whereas advection velocities are correlated over a larger spatial scale, which is reflected in a scaling behaviour of the macroscale dispersion coefficient that increases with increasing travel distance. The effect of transfer time correlation in soil layers with different reactivity was accounted for using conditional transfer time distributions. The transfer time correlation was derived for different scaling behaviors of the macroscopic dispersion parameter ranging from perfect linear scaling (or perfect correlation of transfer times) to a constant dispersion (or no correlation of transfer times). For each type of scaling, reactive substance transfer was also predicted using inert tracer transfer functions that were derived assuming no transfer time correlation in the different soil layers. This approach is similar to using an advection dispersion equation with a macrodispersion that increases with travel distance or depth. A comparison between the two approaches demonstrated that neglecting travel time correlation in different soil layers leads to an underestimation of the reactive tracer transfer through the layered soil. As a consequence, a macroscale advection dispersion equation with a depth dependent dispersion, though reproducing transfer times of inert substances, cannot predict transfer distributions of reactive substances. It also demonstrates the importance of considering transfer time correlations between different environmental subsystems for predicting the transfer of reactive substances.