Karl Pohlmann

A Model-Averaging Method for Assessing Groundwater Conceptual Model Uncertainty

This study evaluates alternative groundwater models with different recharge and geologic components at the northern Yucca Flat area of the Death Valley Regional Flow System (DVRFS), USA. Recharge over the DVRFS has been estimated using five methods, and five geological interpretations are available at the northern Yucca Flat area. Combining the recharge and geological components together with additional modeling components that represent other hydrogeological conditions yields a total of 25 groundwater flow models. As all the models are plausible given available data and information, evaluating model uncertainty becomes inevitable. On the other hand, hydraulic parameters (e.g., hydraulic conductivity) are uncertain in each model, giving rise to parametric uncertainty. Propagation of the uncertainty in the models and model parameters through groundwater modeling causes predictive uncertainty in model predictions (e.g., hydraulic head and flow). Parametric uncertainty within each model is assessed using Monte Carlo simulation, and model uncertainty is evaluated using the model averaging method. Two model-averaging techniques (on the basis of information criteria and GLUE) are discussed. This study shows that contribution of model uncertainty to predictive uncertainty is significantly larger than that of parametric uncertainty. For the recharge and geological components, uncertainty in the geological interpretations has more significant effect on model predictions than uncertainty in the recharge estimates. In addition, weighted residuals vary more for the different geological models than for different recharge models. Most of the calibrated observations are not important for discriminating between the alternative models, because their weighted residuals vary only slightly from one model to another.

Description of hydrogeologic heterogeneity and evaluation of radionuclide transport at an underground nuclear test

Realistic models of contaminant transport in groundwater demand detailed characterization of the spatial distribution of subsurface hydraulic properties, while at the same time programmatic constraints may limit collection of pertinent hydraulic data. Fortunately, alternate forms of data can be used to improve characterization of spatial variability. We utilize a methodology that augments sparse hydraulic information (hard data) with more widely available hydrogeologic information to generate equiprobable maps of hydrogeologic properties that incorporate patterns of connected permeable zones. Geophysical and lithologic logs are used to identify hydrogeologic categories and to condition stochastic simulations using Sequential Indicator Simulation (SIS). The resulting maps are populated with hydraulic conductivity values using field data and Sequential Gaussian Simulation (SGS). Maps of subsurface hydrogeologic heterogeneity are generated for the purpose of examining groundwater flow and transport processes at the Faultless underground nuclear test, Central Nevada Test Area (CNTA), through large-scale, three-dimensional numerical modeling. The maps provide the basis for simulation of groundwater flow, while transport of radionuclides from the nuclear cavity is modeled using particle tracking methods. Sensitivity analyses focus on model parameters that are most likely to reduce the long travel times observed in the base case. The methods employed in this study have improved our understanding of the spatial distribution of preferential flowpaths at this site and provided the critical foundation on which to build models of groundwater flow and transport. The results emphasize that the impacts of uncertainty in hydraulic and chemical parameters are dependent on the radioactive decay of specific species, with rapid decay magnifying the effects of parameters that change travel time.

Modeling Groundwater Flow and Transport of Radionuclides at Amchitka Island's Underground Nuclear Tests: Milrow, Long Shot, and Cannikin

Since 1963, all United States nuclear tests have been conducted underground. A consequence of this testing has been the deposition of large amounts of radioactive material in the subsurface, sometimes in direct contact with groundwater. The majority of this testing occurred on the Nevada Test Site (NTS), but a limited number of experiments were conducted in other locations. One of these locations, Amchitka Island, Alaska is the subject of this report. Three underground nuclear tests were conducted on Amchitka Island. Long Shot was an 80-kiloton-yield test conducted at a depth of 700 meters (m) on October 29, 1965 (DOE, 2000). Milrow had an announced yield of about 1,000 kilotons, and was detonated at a depth of 1,220 m on October 2, 1969. Cannikin had an announced yield less than 5,000 kilotons, and was conducted at a depth of 1,790 m on November 6, 1971. The purpose of this work is to provide a portion of the information needed to conduct a human-health risk assessment of the potential hazard posed by the three underground nuclear tests on Amchitka Island. Specifically, the focus of this work is the subsurface transport portion, including the release of radionuclides from the underground cavities and their movement through the groundwater system to the point where they seep out of the ocean floor and into the marine environment. This requires a conceptual model of groundwater flow on the island using geologic, hydrologic, and chemical information, a numerical model for groundwater flow, a conceptual model of contaminant release and transport properties from the nuclear test cavities, and a numerical model for contaminant transport. Needed for the risk assessment are estimates of the quantity of radionuclides (in terms of mass flux) from the underground tests on Amchitka that could discharge to the ocean, the time of possible discharge, and the location in terms of distance from shoreline. The radionuclide data presented here are all reported in terms of normalized masses to avoid presenting classified information. As only linear processes are modeled, the results can be readily scaled by the true classified masses for use in the risk assessment. The modeling timeframe for the risk assessment was set at 1,000 years, though some calculations are extended to 2,000 years. This first section of the report endeavors to orient the reader with the environment of Amchitka and the specifics of the underground nuclear tests. Of prime importance are the geologic and hydrologic conditions of the subsurface. A conceptual model for groundwater flow beneath the island is then developed and paired with an appropriate numerical modeling approach in section 2. The parameters needed for the model, supporting data for them, and data uncertainties are discussed at length. The calibration of the three flow models (one for each test) is then presented. At this point the conceptual radionuclide transport model is introduced and its numerical approach described in section 3. Again, the transport parameters and their supporting data and uncertainties are the focus. With all of the processes and parameters in place, the first major modeling phase can be discussed in section 4. In this phase, a parametric uncertainty analysis is performed to determine the sensitivity of the transport modeling results to the uncertainties present in the parameters. This analysis is motivated by the recognition of substantial uncertainty in the subsurface conditions on the island and the need to incorporate that uncertainty into the modeling. The conclusion of the first phase determines the parameters to hold as uncertain through the main flow and transport modeling. This second, main phase of modeling is presented in section 5, with the contaminant breakthrough behavior of each test site addressed. This is followed by a sensitivity analysis in section 6, regarding the importance of additional processes that could not be supported in the main modeling effort due to lack of data. Finally, the results for the individual sites are compared, the sensitivities discussed, and final conclusions presented in section 7.

Uncertainty Analysis of Radionuclide Transport in a Fractured Coastal Aquifer with Geothermal Effects

Groundwater flow and radionuclide transport at the Milrow underground nuclear test site on Amchitka Island are modeled using two-dimensional numerical simulations. A multi-parameter uncertainty analysis is adapted and used to address the effects of uncertainties associated with the definition of the modeled processes and the values of the parameters governing these processes. In particular, we focus on the effects on radioactive transport of uncertainties associated with conduction and convection of heat relative to the uncertainties associated with other flow and transport parameters. These include recharge, hydraulic conductivity, fracture porosity, dispersivity and strength of matrix diffusion. The flow model is conceptualized to address the problem of density-driven flow under conditions of variable salinity and geothermal gradient. The conceptual transport model simulates the advection–dispersion process, the diffusion process from the high-velocity fractures into the porous matrix blocks, and radioactive decay.For this case study, the uncertainty of the recharge-conductivity ratio contributes the most to the output uncertainty (standard deviation of mass flux across the seafloor). The location of the freshwater–saltwater transition zone changes dramatically as this ratio changes with the thickness of the freshwater lens and the location of the seepage face changing as well. In the context of radionuclide transport from the nuclear test cavity that is located in the area where the transition zone is uncertain, travel times of radionuclide mass from the cavity to the seepage face along the seafloor are significantly impacted. The variation in transition zone location changes the velocity magnitude at the cavity location by a large factor (probably an order of magnitude). When this effect is combined with porosity and matrix diffusion uncertainty, the uncertainty of transport results becomes large. Although thermal parameters have an effect on the solution of the flow problem and also on travel times of radionuclides, the effect is relatively small compared to other flow and transport parameters.

Sensitivity of Solute Advective Travel Time to Porosities of Hydrogeologic Units

An integral approach is proposed to quantify uncertainty and sensitivity of advective travel time to the effective porosities of hydrogeologic units (HGUs) along groundwater flow paths. The approach is applicable in situations where a groundwater flow model exists, but a full solute transport model is not available. The approach can be used to: (1) determine HGUs whose porosities are influential to the solute advective travel time; and (2) apportion uncertainties of solute advective travel times to the uncertainty contributions from individual HGU porosities. A simple one-dimensional steady-state flow example is used to illustrate the approach. Advective travel times of solutes are obtained based on the one-dimensional steady-state flow results in conjunction with the HGU porosities. The approach can be easily applicable to more complex multi-dimensional cases where advective solute travel time can be calculated based on simulated flow results from groundwater flow models. This approach is particularly valuable for optimizing limited resources when designing field characterization programs for uncertainty reduction by identifying HGUs that contribute most to the estimation uncertainty of advective travel times of solutes.

Evaluating Groundwater Interbasin Flow Using Multiple Models and Multiple Types of Data

The groundwater interbasin flow, Qy , from the north of Yucca Flat into Yucca Flat simulated using the Death Valley Regional Flow System (DVRFS) model greatly exceeds assessments obtained using other approaches. This study aimed to understand the reasons for the overestimation and to examine whether the Qy estimate can be reduced. The two problems were tackled from the angle of model uncertainty by considering six models revised from the DVRFS model with different recharge components and hydrogeological frameworks. The two problems were also tackled from the angle of parametric uncertainty for each model by first conducting Morris sensitivity analysis to identify important parameters and then conducting Monte Carlo simulations for the important parameters. The uncertainty analysis is general and suitable for tackling similar problems; the Morris sensitivity analysis has been utilized to date in only a limited number of regional groundwater modeling. The simulated Qy values were evaluated by using three kinds of calibration data (i.e., hydraulic head observations, discharge estimates, and constant-head boundary flow estimates). The evaluation results indicate that, within the current DVRFS modeling framework, the Qy estimate can only be reduced to about half of the original estimate without severely deteriorating the goodness-of-fit to the calibration data. The evaluation results also indicate that it is necessary to develop a new hydrogeological framework to produce new flow patterns in the DVRFS model. The issues of hydrogeology and boundary flow are being addressed in a new version of the DVRFS model planned for release by the U.S. Geological Survey.

Fractional dynamics of tracer transport in fractured media from local to regional scales

Tracer transport through fractured media exhibits concurrent direction-dependent super-diffusive spreading along high-permeability fractures and sub-diffusion caused by mass transfer between fractures and the rock matrix. The resultant complex dynamics challenge the applicability of conventional physical models based on Fick’s law. This study proposes a multi-scaling tempered fractional-derivative (TFD) model to explore fractional dynamics for tracer transport in fractured media. Applications show that the TFD model can capture anomalous transport observed in small-scale single fractures, intermediate-scale fractured aquifers, and two-dimensional large-scale discrete fracture networks. Tracer transport in fractured media from local (0.255-meter long) to regional (400-meter long) scales therefore can be quantified by a general fractional-derivative model. Fractional dynamics in fractured media can be scale dependent, owning to 1) the finite length of fractures that constrains the large displacement of tracers, and 2) the increasing mass exchange capacity along the travel path that enhances sub-diffusion.

Contaminant Boundary at the Faultless Underground Nuclear Test

The U.S. Department of Energy (DOE) and the Nevada Division of Environmental Protection (NDEP) have reached agreement on a corrective action strategy applicable to address the extent and potential impact of radionuclide contamination of groundwater at underground nuclear test locations. This strategy is described in detail in the Federal Facility Agreement and Consent Order (FFACO, 2000). As part of the corrective action strategy, the nuclear detonations that occurred underground were identified as geographically distinct corrective action units (CAUs). The strategic objective for each CAU is to estimate over a 1,000-yr time period, with uncertainty quantified, the three-dimensional extent of groundwater contamination that would be considered unsafe for domestic and municipal use. Two types of boundaries (contaminant and compliance) are discussed in the FFACO that will map the three-dimensional extent of radionuclide contamination. The contaminant boundary will identify the region wi th 95 percent certainty that contaminants do not exist above a threshold value. It will be prepared by the DOE and presented to NDEP. The compliance boundary will be produced as a result of negotiation between the DOE and NDEP, and can be coincident with, or differ from, the contaminant boundary. Two different thresholds are considered for the contaminant boundary. One is based on the enforceable National Primary Drinking Water Regulations for radionuclides, which were developed as a requirement of the Safe Drinking Water Act. The other is a risk-based threshold considering applicable lifetime excess cancer-risk-based criteria The contaminant boundary for the Faultless underground nuclear test at the Central Nevada Test Area (CNTA) is calculated using a newly developed groundwater flow and radionuclide transport model that incorporates aspects of both the original three-dimensional model (Pohlmann et al., 1999) and the two-dimensional model developed for the Faultless data decision analysis (DDA) (Pohll and Mihevc, 2000). This new model includes the uncertainty in the three-dimensional spatial distribution of lithology and hydraulic conductivity from the 1999 model as well as the uncertainty in the other flow and transport parameters from the 2000 DDA model. Additionally, the new model focuses on a much smaller region than was included in the earlier models, that is, the subsurface within the UC-1 land withdrawal area where the 1999 model predicted radionuclide transport will occur over the next 1,000 years. The purpose of this unclassified document is to present the modifications to the CNTA groundwater flow and transport model, to present the methodology used to calculate contaminant boundaries, and to present the Safe Drinking Water Act and risk-derived contaminant boundaries for the Faultless underground nuclear test CAU.

Use of Numerical Groundwater Modeling to Evaluate Uncertainty in Conceptual Models of Recharge and Hydrostratigraphy

The uncertainty associated with multiple conceptual models of groundwater recharge and subsurface hydrostratigraphy and the impacts of this uncertainty on predictions made by a regional groundwater flow model is quantitatively evaluated. The composite prediction formally incorporates the uncertainty in the alternative input models using maximum likelihood Bayesian model averaging. The alternative models are weighted by model probability, which is the degree of belief that a model is more plausible given available prior information and site measurements. Flow predictions are found to be more sensitive to hydrostratigraphy than to recharge. Furthermore, posterior model uncertainty is dominated by inter-model variance as opposed to intra-model variance, indicating that conceptual model uncertainty has greater impact on the results than parametric uncertainty. Without consideration of conceptual model uncertainty, uncertainty in the flow predictions would be significantly underestimated.

Validation Analysis of the Groundwater Flow and Transport Model of the Central Nevada Test Area

The Central Nevada Test Area (CNTA) is a U.S. Department of Energy (DOE) site undergoing environmental restoration. The CNTA is located about 95 km northeast of Tonopah, Nevada, and 175 km southwest of Ely, Nevada (Figure 1.1). It was the site of the Faultless underground nuclear test conducted by the U.S. Atomic Energy Commission (DOEs predecessor agency) in January 1968. The purposes of this test were to gauge the seismic effects of a relatively large, high-yield detonation completed in Hot Creek Valley (outside the Nevada Test Site [NTS]) and to determine the suitability of the site for future large detonations. The yield of the Faultless underground nuclear test was between 200 kilotons and 1 megaton (DOE, 2000). A three-dimensional flow and transport model was created for the CNTA site (Pohlmann et al., 1999) and determined acceptable by DOE and the Nevada Division of Environmental Protection (NDEP) for predicting contaminant boundaries for the site.