New Page 1

In 2002, the Arc Hydro data model was published by ESRI Press. Arc Hydro is a geospatial and temporal data model for water resources that operates within ArcGIS. The Arc Hydro data model has been adopted by a wide range of organizations as a standardized format for storing water resource data. The original Arc Hydro data model was primarily focused on surface water data, the wide use of spatial information in groundwater studies and attempts to integrate groundwater and surface water models necessitate the extension of the current model to describe the groundwater environment. Integration of the groundwater data model into Arc Hydro will enable the representation of surface water and groundwater features simultaneously within ArcGIS and will support the connectivity between models, thus providing a better understanding of the hydrological cycle. Representing the 3-dimensional characteristics of groundwater systems introduces an additional level of complexity to Arc Hydro. Features such as geological formations, wells and the representation of flow in aquifers require 3-dimensional representation and visualization capabilities.

Currently, the data model is being designed by collaboration of ESRI, The Environmental Modeling Research Laboratory at Brigham Young University, and the Center for Research in Water Resources at the the University of Texas, Austin. The first draft of the geodatabase design has been completed and distributed within the groundwater community for review. This document presents the geodatabase design, implementation examples, and ArcGIS tools and code..

Arc Hydro groundwater geodatabase (Download in word format)

The following is a proposed geodatabase design for representing groundwater systems that will be integrated with the Arc Hydro surface water data model to provide a better representation of hydrologic systems within a geodatabase.

In designing the data model, we concluded that it would be impossible to anticipate all of the different types of groundwater data that could potentially be stored in the geodatabase. We concluded that it was best to keep the data model simple enough to be workable and efficient, yet flexible enough that it would be adapted for a wide variety of uses. We have attempted to focus on ground water data in terms of raw field data and a conceptual representation of the primary feature of a hydrogeologic system, rather than formatting the data in a fashion that favors a specific model type (MODFLOW, for example). This allows the data model to be used as a tool for archiving and sharing groundwater data for a variety of applications.

The Hydrogeology feature dataset is a set of vector objects (points, lines, polygons, and multipatches) that represent hydrogeologic features used in groundwater studies. The Modeling Feature Dataset is a set of vector feature classes that can represent common modeling objects such as cells and elements. This feature dataset is primarily used to post the results of a modeling study within a GIS. The Surfaces group includes both rasters and TINs and is used to define elevations or spatially variable aquifer parameters.

In addition to the above components, a Time Series component will be added to represent time dependent information. A design for representation of temporal information within a geodatabase is being developed in a parallel study. Once the design is complete, it will be implemented in the groundwater data model. The temporal representation will include four basic concepts:

Temporal components will be integrated within the other components of the data model. For example, there will be both static and transient rasters in the Surfaces group.

Hydrogeology Feature Dataset

The hydrogeology feature dataset is a set of vector feature classes that represent two- and three- dimensional objects which are common in groundwater studies.

There are ten feature classes in the hydrogeology feature dataset. These include two- and three-dimensional points, lines, and polygons, and multipatches which are used to represent three-dimensional solid objects. The following shows examples of the hydrogeology feature classes.

Aquifer: Aquifer is a polygon feature class that is used to represent the boundaries of aquifers and zones within the aquifer. Examples of zones within the aquifer are outcrop and downdip areas within the aquifer boundary (Figure 3)

Well: Well is a point feature class that represents well locations (x, y) in map view and associated attributes. Various types of wells (e.g. pumping, monitoring) can exist within the feature class and will be differentiated by a feature type (FType) field. Most of the information related to a well can be stored in relation to the well point. Properties such as the type of the well, day it was drilled, well depth, owner, pumping rate etc. can all be stored as attributes of the well feature. A large variety of well attributes exist and these vary between studies, thus attributes will not be attached to the well in advance and users will add their attributes of interest. Exceptions are the attributes that describe the well’s screen. These include:

Each well will have attributes to describe one screen and if necessary users will be able to add additional screen attributes to represent multiple screens at the same well.

BoreLine: BoreLine is a three dimensional polyline feature class (PolylineZ) that represents the hydrostratigraphy observed along a borehole. Figure 4 shows an example of borelines within the Beaufort Aquifer (North Carolina), each vertical colored segment represents a hydrogeologic unit observed along the borehole.

GeoArea, GeoLine and GeoPoint: These are general feature classes that can represent a variety of groundwater related information. GeoArea features are two-dimensional polygons that can represent areal features such as geologic formations from geologic maps, areas with a common hydraulic conductivity, recharge or discharge zones, and evapotranspiration zones. GeoPoints are three-dimensional points (PointZ) that can represent features such as geologic contacts, or measurements from geophysical instruments, and GeoLines are three-dimensional lines (PolylineZ) that represent features such as faults and folds from geologic maps. Figure 6 shows examples of GeoArea, GeoLine and GeoPoint features.

GeoSection: This is a three-dimensional polygon feature class (PolygonZ) that represents vertical slices of the subsurface, and can represent fence diagrams and cross-sections. These can be created by interpolation from GeoPoints or BoreLines. Figure 6 shows GeoSections created from BoreLines in the Beaufort Aquifer.

GeoVolume: GeoVolume is a multipatch feature class that can be used to represent volumes. The multipatch is a set of three-dimensional triangles that can define the surface of a volume element. GeoVolumes can be constructed by connecting a set of GeoPoints, BoreLines, or GeoSections. Currently ArcGIS does not support the automatic creation of volume objects and the calculation of attributes such as the volume and surface area of the element. However, multipatches can be created by linking ArcGIS and external computational geometry programs and attributes computed in the external programs can be written as an attribute of the multipatch. Figure 7 shows an example of GeoVolumes created for the Beaufort aquifer and its top confining layer generated from BoreLines.

Surface water features: Surface water features include a polygon feature class, SurfaceWaterAreas, and a line feature class, SurfaceWaterLines. The surface water features represent sources and sinks to the groundwater system such as rivers, lakes, wetlands etc. Each feature within these feature classes will have the following attributes:

Hydrogeologic Unit Table

The hydrogeologic unit table provides for a conceptual description of hydrogeologic units within the geodatabase. The table is a generic description of units that can be linked to spatial representations in the form of BoreLines, GeoPoints, GeoLines, GeoAreas, GeoSections, and GeoVolumes. The concept is to use the table as an integrative method to link together varying spatial representations of the same hydrogeologic unit. These spatial representations will not hold the strata properties as attributes of the features; rather they will point to a hydrogeologic unit in the table. Additional properties could be added to the table by users. Figure 8 shows an example of a hydrogeologic unit table related to spatial features.

Figure 8. Example of a hydrogeologic unit table linked to spatial representations

Modeling Feature Dataset

The Modeling Feature Dataset is a set of vector feature classes that can represent common modeling objects. The modeling feature dataset includes four feature classes:

These feature classes can be used to represent commonly used computational grids such as finite difference grids and finite element meshes, and enable storage and presentation of model inputs and outputs related to the computational grids. Figure 9 shows the Modeling Feature dataset.

Boundary is a polygon feature class that represents the two dimensional extent of a model. Within the boundary two-dimensional (Cell2D) and three-dimensional (Cell3D) objects can be generated to represent the cells in a finite difference grid or elements in a finite element mesh, and the nodes of the mesh will be stored in the Nodes feature class. Figure 10 shows examples of two-dimensional cells and nodes used to represent finite element meshes and finite difference grids, and Figure 11 shows an example of a three-dimensional representation of a finite difference grid.

GeoRasters and Raster Series

GeoRasters and Raster series will be stored as raster catalogs. This is a new feature within ArcGIS 9.0 which allows cataloged rasters to be stored within a Geodatabase. Attributes of the rasters can be stored in the catalog, and the catalog can be used to index rasters with time values. A set of static rasters can be used to represent hydrogeologic units. GeoRasters can define boundaries of units by describing the unit’s top and bottom elevations and can also describe distributions of properties, such as hydraulic conductivity, transmissivity, and porosity, within the unit. Raster series usually represent parameters related to the water within a hydrogeologic unit. Examples of Raster series may include the distribution of contaminant concentrations over time or the change of a potentiometric surface within an aquifer. Figure 12 shows examples of GeoRasters and Figure 13 shows an example of Raster series indexed by time.

Figure 12. GeoRasters defining top and bottom boundaries of the Woodbine Aquifer (Texas)

GeoTin

The GeoTin is a standalone TIN which can be used to represent surfaces. TINs are widely used to represent discontinuous surfaces, which are common in subsurface studies that deal with geologic objects such as faults and fractures. The GeoTin provides a mechanism to store and represent such surfaces which are difficult to represent using rasters. In ArcGIS, TINs cannot be stored within a personal geodatabase, thus it will reside in a separate folder.

Example for the North Carolina coastal aquifer system

(Download the example geodatabase and an ArcScene document ~120 MB)

The geodatabase presented above is being implemented with information from the North Carolina coastal aquifer system. This example shows how spatial information related to groundwater studies can be stored using the proposed geodatabase. The information in the example is maintained at the Center for the Analysis and Prediction of River Basin Environmental Systems at Duke University (www.env.duke.edu/cares/neuse/data.html), as part of a CHUASI (Consortium of Universities for the Advancement of Hydrologic Science Inc) prototype study of the Neuse River Basin, which overlies the coastal aquifer system.

The geology of the North Carolina Central Coastal Plain can be characterized as a gently eastward dipping and eastward thickening wedge of sediments and sedimentary rock. The basement surface elevation ranges between 100 feet above sea level to 4500 ft below see level. The hydrogeologic system consists of eight major aquifers and the confining units that separate them: the Surficial, Yorktown, Castle Hayne, Beaufort, Peedee, Black Creek, Upper and Lower Cape Fear aquifers (* Lautier, 2001). The layout of the units is shown in the following figure (from Lautier, 2001).

The example geodatabase is based on the data model described above. Some relationships were created between the feature classes of the geodatabase to enable easy querying of feature. For example selection of all wells screened in a certain aquifer or querying for the BoreLines of a selected well. A formal design for the relationships and attributes is not yet included in the data model design and the relationships and attributes in the example database are not currently part of the data model, although relationships and attributes will definitely be established as we progress in the design process.

A feature dataset named Base Data is included in the example geodatabase; this feature dataset holds the original datasets as they were retrieved, before putting them into the data model framework. This feature data set is not part of the data model; it is included for demonstration purposes and to enable tracking back to the pre-processed information.

Two feature subtypes exist within the aquifers feature class: Boundary, and Zone. The boundary subtype defines the extent of the aquifer in plan view and the zones classify specific zones within the aquifer boundary. The following figure shows an example of the water quality zones within the boundary of the Black Creek aquifer.

Two types of wells are represented: Hydrostratigraphy wells which have information regarding the stratigraphy of the subsurface and USGS wells that are related to water measurements (Water elevation, pumping, etc.). In the geodatabase all wells are stored as point features in the Well feature class, and well types are distinguished through the FType (feature type) field. The USGS wells are wells within the Neuse River Basin, the datasets are prepared as part of a CHUASI (Consortium of Universities for the Advancement of Hydrologic Science Inc) project in the Neuse River Basin.

The BoreLines feature class was created from the hydrostratigraphy well data, to represent the hydrostratigraphy observed at the wells. The data from the hydrostratigraphy wells was first converted from their original form into a structured table (see hydrostratigraphy table in the geodatabase) from which the BoreLines were generated. Each BoreLine feature in the BoreLine feature class represents a segment along a drilled well with a specific lithology or hydrologic unit associated to the segment. BoreLines in the geodatabase are related to wells from the Well feature class, for example all the BoreLines of a specific well can be queried by selecting the well and querying the related features. The following figure shows the creation of the hydrostratigraphy table from the well information.

In the example dataset there are three types of GeoLines: Geologic Structures, Dikes, and Contour lines. The types of information are distinguished by using different feature types (FType field in the geodatabase). The following figure shows the types of GeoLines.

The geologic formations include attributes describing features such as faults, folds, and scarps, while the contour lines describe the tops of geologic formations. The attributes of each data type is stored in a separate table, and only the spatial description of the feature appears within the GeoLine feature class. The specific attributes for each data type can be retrieved using the relationships established in the geodatabase. The following figure shows an example of such a relationship, where GeoLine 52 is a geologic structure and is related to the geologic structures table from which we can learn the structure’s name and ID.

The example datasets include two types of GeoAreas, which are distinct by the FType field: Geologic formations and recharge/discharge zones. The geologic formations details are stored in a separate table in the geodatabase (Geologic formations) and a relationship is established between the GeoArea feature class and the Geologic formations table. The following image shows the two types of GeoAreas in the geodatabase.

The water areas and water lines in the example dataset are the river network and water bodies in the Neuse River basin.

The example dataset includes a set of GeoVolumes that represent the hydrostratigraphy of the studied area. The solid model of the subsurface was created by interpolation of the BoreLines using the Horizons Method in GMS (Groundwater Modeling System). The hydrostratigraphy from the BoreLines was exported to GMS, then solid interpolation was done in GMS and the results were read back into ArcGIS as a multipatch feature (for more details on constructing GeoVolumes see the sources and samples section). The following image shows the solid model of the subsurface in ArcScene.

The example geodatabase includes a set of GeoRasters that represent the tops and bottoms of the hydrogeologic formations. Each raster in the GeoRasters catalog represents the top surface of a specific unit. The surfaces were generated from the contour lines in the GeoLine feature class by converting the 3D lines to TINs and from TINs to rasters. The following image shows the raster catalog containing the GeoRasters.