Transient Inverse Calibration of the Site-Wide Groundwater Flow Model (ACM-2): FY03 Progress Report

Abstract

(PNNL) have embarked upon a new initiative designed to strengthen the technical defensibility of the groundwater flow and transport model at the Hanford Site in Southeast Washington State and to develop a more robust capability to incorporate uncertainty into the model. One aspect of the initiative is developing and using a three-dimensional transient inverse modeling approach to estimate the hydraulic conductivities, specific yields, and other site-wide scale parameters that incorporates data on the transient behavior of the unconfined aquifer system resulting from Hanford Site waste management since 1943. Over the historical period of Hanford operations, the large volumes of wastewater discharged to a variety of waste facilities resulted in large water table changes over most of the Hanford Site and created significant groundwater mounds (in excess of 20 m) under waste management facilities in the central part of the site. Since 1988, the mission of the Hanford Site has changed from producing weapons to restoring the environment. Thus wastewater discharges have declined significantly, which has caused significant water table declines. The three-dimensional transient inverse calibration, which an external peer review panel recommended to DOE, is being performed using UCODE, a universal inverse modeling code developed jointly by the U.S. Geological Survey and the International Groundwater Modeling Center of the Colorado School of Mines. The work uses the existing consolidated site-wide groundwater model (SGM) implemented with the Coupled Fluid Energy and Solute Transport (CFEST) code, which is the forward model whose parameters are estimated by UCODE. The transient inverse calibration uses approximately 70,000 water level measurements made at the Hanford Site since the mid-1940s. Compared with the prior model, the initial baseline transient inverse calibration effort (Cole et al. 2001a) significantly improved the ability of the baseline model to simulate historical trends in water table changes over the entire site for the 1943-1996 period of calibration. Most notably improved were the historical trends of water table changes and mound building observed near major discharge facilities in the 200 West Area. The focus of the inverse modeling initiative in the following year (Vermeul et al. 2001) was to use the developed inverse calibration methodology to test an alternative conceptual model (ACM-1) that no longer considered the underlying basalt bedrock system the no-flow base of the unconfined aquifer system as it had in the past. This alternative SGM model was used to examine and evaluate a variety of mechanisms that affect intercommunication between the Hanford Site unconfined aquifer system and the underlying upper basalt-confined aquifer system. The focus of the ACM-2 inverse modeling initiative, as documented in this report, has been to 1) address data and model implementation limitations identified in the ACM-1 model, 2) complete the implementation and evaluation of the facies-based approach for representing...

Three-Dimensional Groundwater Models of the 300 Area at the Hanford Site, Washington State

“…Historical assumptions range from 560 to 1400 m 3 per meter of shoreline, based on values of 36 million to 90 million m 3 /year for the entire Hanford Site aquifer (summarized in ). Recent site-wide groundwater modeling studies produced estimates of 480 and 700 m 3 per meter of shoreline, based on estimates of 31 million and 45 million m 3 /year for the entire Hanford Site (Vermeul et al 2003 andThorne et al 2006, respectively). also estimated the aquifer flux to the river around spring 9 in the 300 Area from results of multilevel piezometer measurements.…”

Identifying the Potential Loss of Monitoring Wells Using an Uncertainty Analysis

“…The final statistic used was the standard deviation ( σ ) calculated from the error measurement estimate ( E m ) and the time and spatial weights ( W t , W s ) and was calculated as (Vermeul et al 2003) …”

“…However, the volume of water discharged is well documented. These specified artificial discharges, whose magnitude is estimated to be much greater than natural recharge, keeps the regression problem well-posed by breaking the typical correlation between hydraulic conductivity and recharge (Vermeul et al 2003).…”

Vadose Zone–Attenuated Artificial Recharge for Input to a Ground Water Model

“…This capability was developed to handle contaminant transport through various successive pathways at Hanford from inventory through hundreds of vadose zone disposal sites and then to ground water and ultimately to the Columbia River (Nichols et al 2005). This capability was readily adapted to efficiently transmit fluid discharges from the numerous vadose zone site models to a ground water model using a technique described in Vermeul et al (2003) based on using a variable saturation finite‐difference model to solve the coupled nonlinear Richards’ and water mass conservation equations for the vadose zone. The technique described in Vermeul et al (2003) adapted the finite‐difference numerical model of the vadose zone to compute the artificial component of recharge as:…”

“…In the first technique, the influence of the vadose zone is ignored and the sources are applied directly to the ground water model (as was done for Hanford ground water modeling prior to 2003). The second technique uses the method discussed in Vermeul et al (2003) that explicitly accounts for the vadose zone using a steady‐state formulation. Finally, the new technique introduced in this article uses multiple simulations of the vadose zone to explicitly account for transient conditions in partitioning artificial recharge from total simulated recharge.…”