27l ‎(2)‎

• Board Meeting Date: 11/19/2008
• Description: CF Section - En-Bloc Non-Reimbursable Investment Recommendations - MODFLOW Enhancements for CDSS
• Board Meetings - Doc Type: Memo

Water Project Construction Program Project Data
Non - Reimbursable Investment

Grantee : County:
Colorado Water Conservation Board Statewide
Project Name:
t Systems
Project Type:
Groundwater model code enhancement
Drainage Basin: Water Source:
Statewide N/A
Total Study Cost: Funding Sources:
$ 62 , 5 00 CWCB Const . Fund
Type of Grantee:
Federal Agency
CWCB Non - Reimbursable Inv . :
$ 5 0 ,000

Summary

During the current development of the alluvial groundwater model in the South Platte Decision Support
System, a need was found for enhancements to MODFLOW, a USGS groundwater model, to make the
modeling process more efficient. These enhancements include t he following two items: 1) The capability
of tracking individual model stress input components, from model - input fields to the volumetric budget
output tables generated by MODFLOW; and 2) A user - specified option for continuation of model
execution despite solver nonconvergence.

In explanation of the first item, MODFLOW input packages (e.g., Recharge, Well, etc.) are often used to
represent more than one different type of model input per package. For example, the Recharge package
may be used to input to the model both irrigation recharge from irrigated lands and precipitation recharge
from areas of native vegetation. It is often very helpful in the modeling process to be able to track the
different input components of a package from beginning to end of the m odel run. This is not easily done
with the current version of MODFLOW, and the effort required is time - consuming and potentially prone
to error. The first proposed enhancement would make this process more efficient and accurate.

The second proposed enhanc ement involves a user - specified option to continue model execution despite
solver nonconvergence. During the model calibration process, automated model analyses such as
parameter estimation and sensitivity analysis are often used to make the process more e fficient. Solver
nonconvergence can prevent the completion of lengthy modeling runs utilizing these automated modeling
aids. This enhancement would solve this problem but print a warning to avoid inadvertent application.