Dakota Home Report Archive FY92 Ann Rep

Kansas Geological Survey, Open-File Rept. 93-1
Annual Report, FY92--Page 5 of 20

Geohydrology, continued

Ground-Water Flow Simulation

Large-Scale Steady-State Modeling of the Regional Flow System in Southeastern Colorado and Western Kansas

The primary result of the FY92 work on the regional model was the development of an interface between the KGS geographic information system (GIS), ARC/INFO, and the ground-water flow model, MODFLOW, that will be used to simulate the regional flow system. All of the research effort was put into modifying an existing interface designed by the U.S. Geological Survey. Unfortunately, no work could be completed on assembly of the regional model because of the time required for putting the geologic and hydrologic databases into a format that could be accepted by the ARC/INFO-MODFLOW interface.

The advantage of linking the GIS to MODFLOW is the pre- and postprocessing capability that allows for easier input of data, evaluation of model results during calibration and simulation of hydrologic processes, and dissemination of model results to users. The linkage of ARC/INFO with the MODFLOW ground-water model is intended to facilitate movement of information in the Dakota Aquifer Program databases from the GIS to the model and subsequently to return the output in a form that allows for easy postprocessing. Output can result in tabular data sets or maps, either of which can be created quickly and distributed to users for evaluation of model runs. With the interface it is possible to automate many aspects of the modeling process, which will make the process of constructing model runs more efficient. This aspect becomes important, especially when modeling a system as large and complex as the Dakota aquifer. It also allows for interactive editing of parameters at the nodes in the model cells, either visually by plotting the grid on the screen or by modifying the relational database of INFO. The post-processing capability of the GIS also makes it convenient to create maps of the input and output. The interface itself does not transform or manipulate the data and is used only for the storage and movement of the data into and out of the GIS.

The U.S. Geological Survey interface was designed for a Prime computer system and was primarily based on the Prime Control Program Language, INFO, and Fortran programs. A considerable effort was invested in the conversion of the interface to the KGS Data General computer system. Four features of the U.S. Geological Survey interface make it attractive: (1) there is no need for changes to the ground-water model code; (2) the data structure is designed within the GIS to match MODFLOW; (3) the interface is flexible enough to allow for all the options available in MODFLOW; and (4) the interface is efficient in terms of computer time and space. The first feature proved to be the most difficult in the conversion process. In the converted version of the interface, changes in the way that the model opens data files passed on from the GIS through the interface were made in the MODFLOW code. As written, files for input and output were opened before entry into the MODFLOW program. However, the Data General operating system requires that the opening of files take place within the program itself. In the Data General version this does not cause any change in the model output. However, this may cause an inconvenience whenever the interface and the MODFLOW program are moved to another platform. Steps are now underway to make the interface query the user for file locations and then to make the necessary modifications based on the response(s) given by the user. The other three features have been retained in the Data General version of the interface. To test the performance of the interface, a simple hypothetical example including a small grid and sample data set was set up. The results indicate that the interface properly converts GIS input into data files that can be used by MODFLOW.

Five sequential operations occur in the interface program to translate the GIS input into a form that can be used by MODFLOW. The interface (1) queries INFO to determine which packages are in use and which parameters are required for the model run, (2) runs the INFO package output program using the control files containing the parameters and outputs system files for the model run, (3) executes MODFLOW, which opens and uses the system files that were output in the previous step, (4) executes a Fortran program that opens selected output files and reformats the records for input back into INFO, and (5) executes one of two programs that allows the user to produce GIS coverages of either heads and drawdowns or flow terms. Between steps 4 and 5 the user can view the output files to determine whether the results are reasonable. If the desired results are not obtained, the user can interactively correct the input controls or values and re-execute the interface and model. If, however, the results appear to be satisfactory, one or both of the final programs can be executed to produce GIS coverages (which can then be used to produce maps). A simplified flow chart of the processes and options of the interface appears in the section on Geographic Information Systems and data-base management activities for FY92. The interface allows input of data in several forms, including the GIS databases, and also the original file format that MODFLOW was initially designed to handle. Output can either consist of ARC/INFO databases in the form of coverages or the original MODFLOW output file in an ASCII format.

Ground-Water Flow Systems and Water-Resources Potential of the Dakota Aquifer in Parts of Republic, Cloud, Clay, and Washington Counties

The results of this research are described in detail in a recently completed report that is under review and is expected to be released for publication early in FY94 (Wade, 1993). Below is a summary of the research findings from this report.

The Dakota aquifer in Republic, Washington, and northern Cloud and Clay counties is approximately 40% sandstone and 60% mudstone and is composed mainly of the Dakota Formation. Test holes reveal that the sandstones are largely isolated from each other by mudstone with the exception of a sheetlike amalgamated fluvial channel sandstone at the base of the Dakota Formation, which is the most important conduit of lateral flow in the aquifer.

Field measurements of water levels, depths to water from driller's logs, ground-water quality, and surface features of ground-water discharge show that flow systems in the aquifer are both regional and local. The regional system is driven by the topographic slope of the Great Plains toward the east. This system brings halite-solution brine, originally derived from Permian strata, into the area from the west. The incised watersheds of Salt Creek in Republic County and Mill Creek in Washington County drain the regional flow system, creating local flow systems that are recharged in the outcrop area. The local systems have potentiometric surfaces that are subdued forms of the local relief.

Ground water in local flow systems is pumped for domestic, irrigation, stock, and municipal use. A steady-state flow model indicates that the recharge to local flow systems in southwestern Washington County is 0.25 in./yr (0.6 cm/yr). Simulations incorporating pumping suggest that water quality and availability have not been significantly affected by pumping in this area and that the aquifer probably could sustain an increase in pumping to double the current level without drawing in water of inferior quality. This is in agreement with limited long-term monitoring data. Mean drawdowns estimated for this pumping rate are between 3 ft and 6 ft (1-1.8 m).

Steady-State Modeling of Ground-Water Flow in Southeastern Colorado and Southwestern Kansas

The U.S. Geological Survey is developing a simulation of the flow system in a subarea of the regional Dakota aquifer system. The project involves modeling on a smaller scale in southeastern Colorado and southwestern Kansas, thereby allowing more detail in the area where the largest amount of water is currently withdrawn from the Dakota aquifer. The study includes development of a conceptual model of the aquifer and its relation to the overlying High Plains aquifer, underlying Permian units, and stream-aquifer interaction in the Arkansas River valley. From the conceptual model of the aquifer system, a three- dimensional finite-difference multilayer digital simulation is being formulated. The geologic and hydrologic information compiled for the subarea, including test-hole drilling and hydrologic testing during FY90-92 are being used in the construction of the model. The computer model is based on the program MODFLOW.

A model grid has been created that is a subdivision of the 6-mi node spacing grid being developed for the large-scale regional model. The grid spacing selected, 1.5 mi, will provide the detail for answering questions concerning well spacing and well interference for ground-water management of the Dakota aquifer. Smaller grid spacing was found to produce such a large number of nodes that computer processing would be too complex and time-consuming to be practical. Updated GIS coverages of the Dakota aquifer, adjacent formation tops and thicknesses, and predevelopment water levels were provided by the KGS so that input data for both the large-scale regional and smaller-scale subregional models would agree. Preliminary boundary conditions and hydraulic parameters have been assigned to the several aquifer and adjacent layers, based on the conceptual model of the aquifer system.

Previous page--Monitoring Sites || Next page--Effect of River Valleys Introduction
Start of this report || Table of Contents
Kansas Geological Survey, Dakota Project
Updated Jan. 1997
Scientific comments to P. Allen Macfarlane
Web comments to webadmin@kgs.ku.edu