Obtaining and Importing GIS Data for Modeling
GEOS 5313: Applied Surface Water Modeling
Spring 2007

As GIS data products become more numerous, it becomes more difficult to decide which provider to use. For the purposes of this class, we'll rely on the USGS Seamless Data Server, and the public-domain products it provides for most of our data. Hydrologic soil group maps will be obtained from EPA's BASINS data. Here is a brief tutorial:

1. Downloading From USGS-Seamless
   1. open the Seamless site (Fig. 1)
   2. select the ``Zoom In'' tool, and progressively narrow the display region until it is close to the final extent desired
   3. Layer:Display Tab on lower left is used to toggle on/off items displayed in the central window
      • Topography: before or during the zoom-in process, under the ``Layers:Display'' tab on the right side of the tool enable NED 1'' (or 1/3'' for higher resolution) shaded relief
      • Roads: at large scale (close-up view) you'll probably want the ``BTS Roads'' (Bureau of Transportation Safety) on, at small scale ``National Atlas Roads'' (main highways) will be best
      • Hydrography: Often ``National Atlas Streams'' will help you to be certain you download data for the entire watershed of interest
   4. switch to ``Layer: Download'' tab on the left
   5. select desired data, usually one elevation set (NED), roads (BTS), land-use (see categories below)
   6. select ``Define Download Area'' from toolbox on lower left, then drag a rectangular area (Fig. 1)
   7. a ``Products to Download'' dialog will open (Fig. 2)
   8. gridded datasets should be downloaded in ``BIL'' format for use in WMS . For use in ArcGIS, the ArcGrid format is probably best. To change the format from that listed, select ``Modify Data Request'' and make appropriate choices. Select ``Save and Return'' from that tool to return to download
   9. select desired products and download at will

   Figure 1: USGS Seamless server, ready to download products for UTD. Green line shows download area defined by user.

   

   Figure 2: USGS Seamless server, download products dialog. User selects ``Download'' button to retrieve each item, or ``Modify Data Request'' to change formats (e.g. gridded info should be downloaded in ``BIL'' format for use in WMS ).

   

2. Re-project Data. In Spring 2007 there was no re-projection facility in the Seamless server, and the native data format is Geographic Projection, NAD 83 geoid. For WMS a rectilinear projection is required (effectively UTM), and changing geoid reference is not supported. To reproject in WMS :
   1. begin with a blank project (start WMS and/or choose File/New)
   2. load a shapefile or grid to be reprojected. In WMS -7.0 reprojecting multiple objects appears to fail (i.e. you may have to do them one at a time)
   3. loading a geographic grid should automatically bring up the coordinate transformation dialog, otherwise choose Edit/Change Coordinates
   4. make appropriate dialog entries (be sure to set the local coordinates first)
   5. if reprojecting one object at a time, save as individual file (saving land-use grids in WMS -7.0 requires saving the entire project, of which only the *.lus file is needed)

3. Loading land-use grids in WMS
   1. unzip the land use grid downloaded from Seamless
   2. with the Map or DEM module active choose File/Open
   3. set the file type to Land Use Grid (*.*)
   4. from the landuse grid subdirectory created in step 3a select the *.hdr file and choose Open

4. Land Use Data Types

   At this time, three categories of land-use are available:

   LCUC

   Land Use/Land Cover, based on 1977-83 aerial photographs and satellite images. Available from BASINS, WebGIS, etc. as polygon shapefiles. To use download the appropriate shapefiles and import directly into WMS (usually into a ``LandUse'' coverage). Categories based on Anderson et al. (1972). See Fig. 3.

   NLCD

   National Land Cover Data, based on 1992 Landsat5 TM images, available in raster format, BIL-compressed is best for use in WMS . 21 basic categories (Fig. 3), corresponding roughly to the Classification Level I categories of Anderson et al. (1972)

   NLCD-2001

   A new raster dataset circa 2001 developed from Landsat-7's enhanced TM (ETM) is underway and parts of it can be obtained directly from MRLC Viewer, or via the Seamless Server. Caategory definitions are somewhat expanded over NLCD-1992, see online documentation

   Local

   Many municipalities will have assembled their own database. For example NCTCOG Land Use files are available free online (e.g. last image, Fig. ).

   Figure 3: Top: Level I & II LULC land use codes (appropriate for assignments made from satellite data), as defined by Anderson et al. (1972). Bottom: NLCD land use codes.

   

   In general, SCS Curve Number ``theory'' was developed with LULC-type data (airphoto interpretation) in mind. While NLCD data is more detailed (30-m pixels, top, Fig. 4) and up-to-date, it is less specific (fewer categories) and should be field-checked or adjusted using airphotos. Ultimately assignment of curve number is somewhat unconstrained, and the most accurate approach is to determine for each basin using measured velocities.

   Figure 4: LULC (first image, circa 1975, airphoto-interpreted), NLCD (second image, circa 1992, satellite-image interpreted), and NCTCOG (last image, 2001, airphoto interpreted) land use for UTD watersheds, as plotted in WMS. UTM zone 14 coordinates shown, landuse data source labeled in upper right corner.

   

5. Soils Data

   Soil type data is not currently available via the USGS Seamless server. The following data types and sources are accessible via Internet:

   STATSGO

   ``State Soil Geographic Database'', consolidated Soil Conservation Service soil maps at 1:250,000 scale. Not useful for small urban watersheds, but most appropriate for class projects. See STATSGO webpage for links to documentation and downloads

   SSURGO

   Detailed soil maps (``Soil Survey Geographic Database''). Usually electronic versions of county soil maps generated by SCS, 1160 datasets avaliable as of Spring 2002 out of 3066 counties in U.S. See SSURGO webpage

   BASINS

   EPA has a watershed water quality research program, with accompanying software BASINS. Models and data are meant to operate at the HUC level (i.e. perfect for class projects!).

6. Using BASINS soils data: BASINS provides data by HUC, and so may be the simplest source for soils data. To download soils data from the BASINS website, and import into WMS. Note importing of the shapefiles can also be done via the Map Module, but for large files this can be time consuming. Using the GIS Module allows you to discard unneeded polygons.
   1. Obtain and unpack data:
      1. find your HUC at EPA's Surf Your Watershed
      2. navigate to the BASINS data parent directory and enter the desired HUC subdirectory
      3. download the file [HUC#]_core31.exe, which is a self-extracting WinZip file
      4. execute the [HUC#]_core31.exe file, and select Run WinZip (if possible)
      5. extract the STATSGO shapefiles statsgo* . Note these shapefiles are projected in geographic coordinates, NAD83 datum (same as the Seamless data, must be reprojected to rectilinear coordinates for WMS)
   2. Import soils data into WMS (see also WMS Tutorial 10.2)
      1. run WMS (assuming version 7.1 here), switch to GIS Module (globe icon)
      2. right click on GIS Layers in the Data Tree, load the statsgo.shp shapefile. Select Attribute mapping... button if you wish, note the hydrologic soil group ``HYDGRP'' is not in the attribute list
   3. Add ``HYDGRP'' attribute to shapefile
      1. right click on the statsgo.shp entry in the data tree, select Join Table to Layer
      2. open the database table statsgoc.dbf (contains ``HYDGRP'' attribute, indexed to polygons in statsgo.shp)
      3. retain ``MUID'' as the shapefile and table join fields (i.e. the polygon ID), select HYDRGRP from near the bottom of the ``Table Data Field'' list (Fig. 5)
      4. reproject into UTM coordinates

      Figure 5: WMS 7.1 DBF file import dialog. This is the final step in adding the HYDGRP attribute to the statsgo.shp soils shapefile.

      

   4. Make local soil type coverage (this step not needed if you use a landuse GIS Layer instead of polygon shapefile)
      1. right click on Coverages in the data tree, create a new coverage, make its type Soil Type
      2. return to GIS Module, zoom in to your project area if necessary
      3. Select soil-type polygons needed for your model: switch to the Select shapes tool and select all soil-type polygons in your zoomed-in view
      4. Map these to the Soil Type coverage: choose Mapping/Shapes->Feature Objects pulldown menu (Fig. 6). Note that ``HYDGRP'' is already mapped to ``SCS Soil Type''.
      5. select this coverage when computing ``GIS Attributes'' (composite ) in the Hydrologic Modeling module/Calculators

      Figure 6: WMS 7.1 mapping soil polygons to local coverage. A subset of the STATSGO soil polygons is mapped from the GIS module into the SoilType coverage. See also the WMS help on ``Mapping to Feature Objects''.

      

7. other good summary of WMS data accession is available from BYU-EMRL

Bibliography

J. R. Anderson, E. F. Hardy, J. T. Roach, and R. E. Witmer.

A land use and land cover classification system for use with remote sensor data.
Prof. paper 964, U. S. Geol. Survey, Washington, D.C., 1972.
URL http://landcover.usgs.gov/pdf/anderson.pdf.
41.