Hydrodynamic Modeling to Enhance a Conceptual Model of a Tidally-driven Creek and Superfund Site
Jeremy Grush, LimnoTech, Inc., Ann Arbor, MI
Dan Herrema, LimnoTech, Inc., Ann Arbor, MI
Andrew Hopton, CH2M HILL, Philadelphia, PA
Rich Galloway, Honeywell International, Morristown, NJ

Solute Flux Rate Uncertainty Evaluation at a Monitored Boundary
David E. Langseth, WRE Gradient Corporation, Cambridge, MA

Using Genetic Algorithms on Groundwater Modeling Problems in a Consulting Setting
Karen M. Madsen, AECOM, Westford, MA
A. Elizabeth Perry, AECOM, Westford, MA

Hydrodynamic Modeling to Enhance a Conceptual Model of a Tidally-driven Creek and Superfund Site

Jeremy Grush, LimnoTech, Inc. 501 Avis Drive, Ann Arbor, MI, 48108, Tel: 734-332-1200, Fax: 734-332-1212, Email: jgrush@limno.com
Dan Herrema, LimnoTech, Inc. 501 Avis Drive, Ann Arbor, MI, 48108, Tel: 734-332-1200, Fax: 734-332-1212, Email: dherrema@limno.com
Andrew Hopton, CH2M HILL, 1717 Arch Street, Suite 4400, Philadelphia, PA, 19103, Tel: 215-563-4220 ext 49031, Fax: 215-563-3828, E-mail:  Andrew.Hopton@ch2m.com

Rich Galloway, Honeywell International, 101 Columbia Road, Morristown, NJ, 07962, Tel: 973-455-4640, Email: rich.galloway@honeywell.com 

Ackermans Creek is a tidally-driven system composed of narrow man-made channels that drain the adjacent phragmites wetlands to Berrys Creek.  A two-dimensional model was developed to represent the hydrodynamics of Ackermans and Berrys Creek in order to understand present-day tidal flows and local sediment stability.  The model is being used to support a remedial investigation and to assist in remedy selection for this Superfund site. The main goal of model application was to develop and refine a conceptual site model of the Ackermans Creek system.  Hydrodynamic model predictions were calibrated to local water level and velocity measurements, and were placed in context by comparison with other available data including sediment bed particle size distributions, contaminant data, and field visual observations of sediment bed and bank characteristics.  Model applications were conducted to assess spatially- and temporally-variable tidal velocities and shear stresses, providing insight into erosional and depositional patterns within the system.  Model applications illustrate the range of water depths, velocities, and shear stresses that occur over a lunar cycle, highlighting areas of maximum velocity and shear stress.  Building on the developed understanding of hydrodynamics, particle-tracking simulations were also used to characterize suspended sediment movement within the water system.  Model predictions were generally consistent with other data and provided a well-constrained and useful conceptual model of the system. This presentation will describe model development and application, and how results are relevant to the confirmation of an accurate site understanding.

Solute Flux Rate Uncertainty Evaluation at a Monitored Boundary

David E. Langseth, Sc.D., P.E., D. WRE Gradient Corporation, 20 University Road, Cambridge, MA, 02138, Tel: 617-395-5000, Fax: 617-395-5001, Email: dlangseth@gradientcorp.com

Limitations on data collection density, aggregated spatial and temporal modeling scales, and variations in time can lead to substantial uncertainty in estimated ground water and solute flux rates.  Difficulties associated with quantitative characterization of uncertainty often lead to use of conservative assumptions in management decision-making processes, which can lead to non-optimal resource management.  This poster describes a computational method for estimating solute flux rate uncertainty at a monitored boundary that can be programmed on an ordinary spreadsheet application and discusses parameter estimation approaches.  This method produces results comparable to those produced by Monte Carlo simulation under a reasonable range of assumptions.  Two alternative conceptual models at the monitored boundary can be used, one in which the entire area of interest is treated in an integrated manner and the other in which the area of interest is segmented into zones.  This poster addresses only the segmented approach.  Case study results not shown here indicate that the conceptual model choice has an impact on the analysis results.  Use of uncertainty analysis to evaluate likelihood of exceeding target impacts on receiving waters is also illustrated.

Using Genetic Algorithms on Groundwater Modeling Problems in a Consulting Setting

Karen M. Madsen, AECOM, 2 Technology Park Drive, Westford, MA 01886, USA, Tel: 978-589-3427, Email: karmadsen@yahoo.com
A. Elizabeth Perry, P.G, AECOM, 2 Technology Park Drive, Westford, MA 01886, USA, Tel: 978-589-3167, Email: elizabeth.perry@aecom.com

This paper presents a practical application for writing and applying simple genetic algorithms (GAs) for MODFLOW.  The method employed by GAs is derived from the driving forces of evolution in the natural world.  They employ functions that mimic natural evolutionary processes including selection, mutation, and genetic crossover.  A GA solves mathematical problems where a desired outcome to the problem is defined (for example, calibration targets or remediation goals), but the inputs needed to arrive at this outcome are unknown, such as optimization and/or parameter estimation.  Our paper includes an introduction to genetic algorithms, the pseudocode of our genetic algorithm for MODFLOW, and the results of an experiential application.  Due to the lack of commercially available GAs for MODFLOW, we coded a simple algorithm in Visual Basic Script and applied it to an example model.  In the example model, the GA was used to conduct parameter estimation on a MODFLOW model of a river basin in New England that we had previously developed and calibrated in our practice.  The calibration target used was net groundwater flow into the river.  Four model input parameters were selected as chromosomes for the GA to act on: recharge, river conductance, and two general head boundaries.   An initial population of 100 models was developed by varying the value of the gene parameters.  The GA ran a MODFLOW simulation for each member of the population, extracted each output file, and established the error of each model from the calibration target.  It then evolved the entire population of models towards the calibration target.  The GA converged on a single input set that established best-fit values for all of the chromosome parameters.  Genetic algorithms provide a practical alternative to trial-and-error and automated statistical calibration procedures, and can also be used for optimization.

Top