You are here:
Using Genetic Algorithm and MODFLOW to Characterize Aquifer System of Northwest Florida
Citation:
HUANG, J., C. J. Richards, AND M. E. GONSOULIN. Using Genetic Algorithm and MODFLOW to Characterize Aquifer System of Northwest Florida. Presented at MODFLOW and More 2011: Integrated Hydrologic Modeling Conference, Golden, CO, June 05 - 08, 2011.
Impact/Purpose:
Presentation for the MODFLOW and More 2011: Integrated Hydrologic Modeling Conference in Golden, CO, June 5-8, 2011
Description:
By integrating Genetic Algorithm and MODFLOW2005, an optimizing tool is developed to characterize the aquifer system of Region II, Northwest Florida. The history and the newest available observation data of the aquifer system is fitted automatically by using the numerical model combined with Genetic Algorithm. A large scale numerical model is created and employed. The aquifer system is divided into 26 zones; a group of hydrogeological parameters, including horizontal and vertical conductivity, river bed conductance and general head boundary conditions, is defined in each of these zones. The study analyzes the sensitivity of parameters on the observation data and investigates the uncertainty of parameter estimation. Specifically, the Monte Carlo technique is employed when bounding the uncertainty of parameter estimation. Statistically analyzing distribution of best parameters for each Monte Carlo realization, the behavior of uncertainty of parameter estimation can be recognized. The study also predicts the uncertainty of model simulation based on the uncertainty of parameter estimation. It is the intent of this work to lay a solid technical foundation and create a more reliable numerical model from which the local community can effectively implement the aquifer protection and optimize water supply project. The work also provides a practical example of using large scale numerical and global optimization technology to manage the ground water resources in a region scale.