PosterLetsinger2006_v2 [Converted].ai - Indiana University

3 downloads 72285 Views 2MB Size Report
(georeferenced vector files –. Byron Stone). Digital Elevation Model of each morphostratigraphic unit at the same grid resolution. (georeferenced raster files).
Dressing the Emperor: The Role of GIS in the Development of Three-Dimensional Hydrogeologic Models Sally L. Letsinger ([email protected]), Center for Geospatial Data Analysis, Indiana Geological Survey, Indiana University, Bloomington, Indiana. Greg A. Olyphant ([email protected]), Department of Geological Sciences, Center for Geospatial Data Analysis, Indiana University, Bloomington, Indiana. Cristian R. Medina ([email protected]), Department of Geological Sciences, Center for Geospatial Data Analysis, Indiana University, Bloomington, Indiana.

Abstract No. 114594 Stone and others (USGS, 2001) mapped structure contours for the tops of each of 20 individual units in intersecting and overlapping glacial morphosequences in Berrien County, Michigan (1,350 km2 ), as part of the mapping program of the Central Great Lakes Geologic Mapping Coalition (CGLGMC). We have developed a methodology to translate this detailed morphostratigraphy first into a solid three-dimensional geologic model, and then into a three-dimensional block of data that can be used as input to a finite-difference groundwater-flow model. The technique involves a hybrid approach involving geographic information systems (GIS), threedimensional information visualization software (3DIVS), and customized data-processing code. The methodology begins by converting Stone’s structure contours (they are attributed vector contours) for each individually mapped unit into a raster surface at a defined grid resolution (200 m x 200 m). The top of the geologic model is the surface topography (digital elevation model), which is also used to derive the drainage network that is an important boundary condition in the groundwater-flow model. The bottom of the geologic model is the bedrock topography, which was also mapped and contoured by Stone (USGS, 2001). Stone constructed his structure contour model such that the bottom of each map unit is described by the surface contours of the unit that lies immediately below it. Complex interrelationships dictate that the tops of a number of individually mapped units are sometimes required to describe the bottom surfaces of laterally more extensive units. Once all of the requisite raster grids have been derived, they can be manipulated to provide input that is necessary for development of a detailed solid geologic model using 3DIVS. GIS software and custom code are also used to assign hydrogeologic attributes to the elements of the final three-dimensional finite-difference geologic model.

Model Output Processing and Analysis

Preparation of Geologic Surface Data Sets The surficial geologic mapping completed by the Central Great Lakes Geologic Mapping Coalition is the basis for the geologic model used in this groundwater modeling project. Several steps are required in the conversion of the two-dimensional geologic mapping to a threedimensional geologic model. I. Morphostratigraphic Unit Surface Contours. Byron Stone (USGS) deconstructed the surficial geologic mapping into separate morphostratigraphic units. He then identified the extent of each unit, the shape and configuration of the unit, and its elevation. Each unit was provided to the CGDA groundwater modeling team as a georeferenced vector GIS data layer (shapefile), with attributed contour lines that represent the elevation of the top of the morphostratigraphic unit. The example shown below is for a deltaic unit in the southeastern corner of the county.

Using GIS software to convert model output to visual images in order to assess model results.These examples show the ability of GIS software to visualize the water-table surface, which is a resultant output data set generated by the solution of the steady-state groundwater model of Berrien County, Michigan. I. Because the water table data layer has elevation attributes at the same resolution as the ground-surface topographic layer, the two surfaces can be intersected to show surface ponding, seeps, and areas of

IV. Conceptual depiction of "virtual well field." This is the method by which the morphostratigraphic units are converted from surfaces to three-dimensional volumes. This figure shows the regular X-Y grid imposed on the domain, and the variable Z-direction grid spacing.

ground-water contribution to surface water bodies.

II. This image (top) shows areas where the water table intersects the ground surface. These areas can be compared with known areas of perennial streams, lakes, and wetlands (right image) to evaluate model performance. Areas that are not modeled correctly are calibration targets for future model improvement.

Preparation of Boundary Condition Datasets 0

5

10

km

I. Upper boundary is ground-surface topography, and lower boundary is bedrock-surface topography. The elevation data were converted to digital elevation models (DEMs) using GIS utilities.

THREE-DIMENSIONAL GROUNDWATER MODELING Applied to Berrien County, Michigan

II. Creation of Raster Surfaces for each Morphostratigraphic Unit. Using utilities in GIS software (ESRI ArcGIS), the unit surface contours are converted to raster surfaces at the model resolution (200m x 200m).

PRE-PROCESSING Data preparation

POST-PROCESSING Model output visualization and analysis

MODELING

GEOLOGY

TOPOGRAPHY

INPUTS

MODEL Groundwater flow simulation

0

5

V. Custom computer code is used to identify the elevation of each morphostratigraphic unit surface encountered in each "virtual well." Then, custom programming assigns a code to each grid cell in the three-dimensional array that represents the geologic model. The 3-D array is then imported into 3-D information visualization software (3DIVS, in this case, C Tech's Environmental Visualization System software package).

10

km

II. The perennial stream network provides fixed-head boundary conditions in the model. The flow chart below shows the steps to create the stream network using Hydrology Tools in the Spatial Analyst extension of the ESRI ArcGIS software package. DEM

FILL SINKS

FLOW DIRECTION

FLOW ACCUMULATION

The figures below show the output stream network used as a boundary condition in the current version of the groundwater flow model. Challenges arise to establish a representative perennial stream network using a spatial resolution that will also represent the geologic conditions and allow computational efficiency.

III. Raster unit surfaces stacked in threedimensions from bedrock surface to the ground surface. This view resembles the surficial geologic map for Berrien County by the CGLGMC.

Littoral Sand, Aeolian, and Fluvial

Legend

Contoured tops of morphostratigraphic units and underlying bedrock surface (georeferenced vector files – Byron Stone)

Conversion from ASCII to proprietary data formats

Numerical Finite-Difference Model (computer algorithm) Visualize water table

DECISION: model resolution

Steady state solution Visualize areas of seepage and surface ponding

Data files containing all necessary information on geology in model domain

Digital Elevation Model of each morphostratigraphic unit at the same grid resolution (georeferenced raster files)

Diamicts

Undifferentiated Deltaic Export to ASCII format

Coarse Grained Deltaic

Superimpose a 3-D grid through the solid model at the center of each x,y grid cell; store morphostratigrapic unit code of each x,y,z grid cell in the Solid Model (3-D FiniteDifference Grid and associated IDCELL Array for Groundwater Flow Modeling)

2-D stream network array data input file

Visualize 3-D geologic model of study area

Calibrate

Visualize flow vector field

(1) U.S. Geological Survey. 2001. Surficial geologic map of Berrien County, Michigan; Edited and integrated by B.D. Stone: U.S. Geological Survey Digital Open-file Report 01-256, scale 1:100,000, explanatory text.

Locations and elevations of perennial streams and the Lake Michigan Shoreline at the same x,y grid resolution of the geological model

Stone, B.D., Kincare, K.A., O’Leary, D.W., Lundstrom, S.C., Taylor, E.,M., and Brown, S.E., 2003, Glacial and Postglacial Geology of the Berrien County Region of Michigan: 49th Midwest Friends of the Pleistocene Field Conference, May 16-18, 2003, Benton Harbor, Michigan, 70 pp.

Adjust flow accumulation threshold

no

Bedrock

Data file containing hydraulic properties of porous media associated with each morphostratigraphic unit (2)

References:

Diamicts (basal till, moraine)

Littoral Sand, Aeolian, and Fluvial

Are areas of seeps and surface ponding properly located?

yes

Digital Elevation Model of ground-surface topography

Create “virtual wells” at the center of each grid cell (georeferenced “well field;” each “well” contains elevations of tops of all morphostratigraphic units encountered at that x,y coordinate)

Create the “solid model” by filling the space between successive tops of morphostratigraphic units encountered at each x,y grid cell (3-D Geologic Model)

no

Digital Elevation Model of bedrock-surface topography

Diamicts

Recent Alluvium

yes

Raster stream network (ArcInfo GRID format)

Data Files containing all necessary information on boundary conditions in model domain

Transient solution

Diamicts

Does flowaccumulation threshold produce a stream network that represents perennial streams?

Mathematical groundwater flow model (3)

Geologic and hydrogeologic parameterization

no Raster Calculation: This step uses a flow-accumulation threshold to identify cells in the raster data model that represent streams.

Geologic Mapping (USGS Central Great Lakes Geologic Mapping Coalition Team) (1)

KEY Geologist

Does flow network represent perennial streams?

Hydrogeologist GIS Application

yes

yes

no

Is 3-D geologic model representative?

Conversion from proprietary data models to ASCII format

3DIVS Application Custom programming Input files for custom programming

(2) Schapp, M.G., Leij, F.J., and van Genuchten, M.T., 1995, Estimation of the soil hydraulic properties, in L.G. Wilson, L.G. Everett, S.J. Cullen, W.G. Wilson (eds.): Vadose Zone Characterization and Monitoring, Chapter 3, pp. 501 – 509. Stone, B.D., J.R. Stone, J.P. Masterson, and D.W. O’Leary, 2004. Integrating 3-D Facies Analysis of Glacial Aquifer Systems with Ground-water Flow Models: Examples from New England and the Great Lakes Region, USA, in Berg, Richard C., Russell, Hazen and Thorleifson, L. Harvey, (eds): Three-Dimensional Geological Mapping For Groundwater Applications: 49th annual Meeting Geological Association of Canada, Mineralogical Association of Canada, St. Catherines, Ontario, Canada, Workshop Extended Abstracts. Illinois State Geological Survey, Open File Series 2004-8, 100p. (3) Freeze, A.R., 1971. Three-dimensional, transient, saturated, unsaturated flow in a groundwater basin: Water Resources Research, Vol. 7, No. 2, p. 347-366.

Published work

This research is part of a multi-institutional effort by the Central Great Lakes Geologic Mapping Coalition. This material is based upon work supported by the National Science Foundation under Grant No. EIA-0116050.

Funded By: