GIS Research to Digitize Maps of Iowa 1832-1859 Vegetation
GIS Research to Digitize Maps of Iowa 1832-1859 Vegetation
by Paul F. Anderson, Department of Landscape Architecture and Department
of Agronomy, Iowa State University
The General Land Office (GLO) conducted the original public land survey
of Iowa during the period 1832 to 1859. Approximately 187 Deputy Surveyors
in Iowa produced both township maps and field notes which briefly described
the land and its natural resources (vegetation, water, soil, landform, and
so on) at the time of the survey. These maps and survey notes are among
the few data sources about vegetation distribution before much of Iowa changed
to a landscape of intensive agriculture. Click
here to view an animated GIF file (about 120K bytes) of Iowa maps showing
the year each township was surveyed.
These maps were digitized to prepare a vegetation data layer for GIS data analysis. In addition to digitizing maps, the research team developed a Microfilm Directory Database, researched historical GLO survey methods, and researched GIS modeling procedures to analyze vegetation character, quantity, and distribution from the vegetation map GIS files. Results of GIS spatial and statistical analysis are being used by land managers and researchers in understanding historic vegetation patterns, vegetation changes, and implications for vegetation preservation and restoration.
For more information, see our new GLO Research Web site.
Original paper copies of the surveyors' field notes and maps were not
available for our use. Instead, we used microfilm copies of the notes and
maps. To date, three different sets of microfilm have been used in our research:
Microfilm (WPA series) produced by the Secretary of State's office in Des
Moines with field maps (also called "topographies" or "topographic
sketches") and typescript versions of the field notes. Microfilm from
the National Archives in Washington, DC with handwritten manuscript field
notes and plat maps from one of the three sets drawn at the Surveyor General's
office in Dubuque which were drawn in triplicate at a scale of 2 inches
= 1 mile, using colored ink. Microfilm from the State Historical Society
of Iowa without notes, but with maps from the first set of original official
plats (originally at the Surveyor General's office in Dubuque, then transferred
to the Secretary of State's office, now at the SHSI library).
Township maps were digitized in sets by county. Maps for all 99 counties
were digitized in approximately the same order as the maps appear on the
microfilm (from southeast to northwest): 14 counties in Phase 1 (1992-93),
25 in Phase 2 (1993-94), 31 in Phase 3 (1994-95), and 29 in Phase 4 (1995-96).
Maps were annotated and digitized using the following eight-step procedure:
1. Print hard copy township maps from microfilm
2. Annotate maps with vegetation lines and labels by reading the surveyors'
field notes on microfilm
3. Place each township map on 12"x12" digitizing tablet and rubbersheet
to township lines from NRGIS
4. Digitize lines and label polygons using custom AutoLISP commands in AutoCAD
Rel. 12
5. Verify digitized maps by visually comparing to source maps
6. Import each file into ArcCAD to check for label errors, open polygons,
and slivers
7. Edit changes (if necessary) and verify digitized maps again by visually
comparing to source maps
8. Save files in PC Arc/Info format
Names for 38 vegetation types were taken directly from the township maps
and field notes. We followed the principle of keeping vegetation types separate
and not aggregating vegetation types into fewer categories. In this way,
others who use the data can aggregate according to their own needs.
Features on the township maps that were line data and point data were not digitized.
ALE Swale MEA Meadow SMR Swamp/marsh BAR Barrens OAK Oak barrens SPR Spring BAY Bayou OPE Openings SWA Swamp BOG Bog PON Pond TBR Timber/barrens BRU Brush POO Pool THI Thicket CIT City PPT Pt prairie/pt timber TIM Timber DRA Drain PRA Prairie TSB Timber/scattering/barrens FIE Field RAV Ravine TSO Timber/scattering/openings GRO Grove RIV River (border) VIL Village IFD Indian field ROU Rough WET Wetland ISL Island SAN Sandbar WIL Willows LAK Lake SCA Scattering trees WIN Windfall MAR Marsh SLU Slough
Mike Miller (principal graduate research assistant) completed his masters thesis in May 1995 using digitized GLO data for Fayette and Jackson Counties. In his MLA thesis, Analysis of Historic Vegetation Patterns in Iowa using Government Land Office Surveys and a Geographic Information System, Miller completed a literature review of over 90 sources on GLO survey methods used in Iowa and studies of so-called "presettlement" vegetation in other midwest states. Miller also prepared GIS descriptive models which help define vegetation categories used by deputy surveyors in their maps and notes. Using statistical and spatial modeling procedures, Miller described each of the ten GLO vegetation categories in Fayette County in terms of soils, slope steepness, slope aspect, and existing land use. For Fayette and Jackson Counties, Miller also described each vegetation type in terms of witness tree composition, size, and relative density. Miller used Coefficient of Areal Correspondence (CAC--Minnick) to compare similarities and differences between versions of GLO maps (94.6% for prairie and 88.0% for timber) and maps from the 1875 Andreas Atlas (81.6% for prairie and 56.7% for timber). Miller's research provides a form of metadata to help researchers and land managers evaluate the quality, reliability, and limitations of the digitized vegetation data. In this way, the data can be used in an appropriate way with the knowledge of its limitations due to survey methods of the 1832-1859 period.
Paul Anderson (principal investigator) is
using digitized GLO data in research for the Dallas County Conservation
Board and Office of the State Archaeologist in describing known archaeological
sites and in finding additional sites. Anderson began the research by preparing
GIS descriptive models of 85 known sites and 850 randomly selected non-sites
in terms of their proximity to stream confluences, distance to river valley,
landscape position, land cover, GLO vegetation, and native vegetation based
on soils. Chi square measures reveal that the largest differences between
known sites and non-sites are in their proximity to confluences (717), landscape
position (528), GLO vegetation (452), and proximity to valley (351). In
a comparison of GLO vegetation and native vegetation from soils, the CAC
(Minnick) measure is 31.7%. GIS predictive modeling uses two separate logit
models which have cutpoint scores of 0.44 and 0.56 respectively and improvement
in classification over chance of 55.7% and 51.4% respectively. Current field
work is being used to validate the two predictive models and provide data
on more known sites to aid in future refinements of the GIS predictive models.
Research results include one Arc/Info coverage for each county and two microfilm
directory database files (one database file for each microfilm set purchased).
In addition, researchers documented observations and concerns about each
vegetation type and county. Recommendations include taking caution in using
the data for site-specific research, making coverages and documentation
available on-line, additional historic research about GLO surveyor instructions
and methods, GIS descriptive modeling of statewide vegetation patterns,
and document scanning the surveyors' field notes as an additional hyperlinked
digital research tool.
The research is being done with the cooperation and support of the State
Preserves Board and the Iowa Department of Natural Resources (specifically,
Daryl Howell, John Pearson, Kevin Kane, Jim Giglierano, and Robin McNeely)
and Iowa State University Department of Landscape Architecture and College
of Design. DNR has placed these 99 Arc/Info county coverages on their anonymous
FTP (ftp.igsb.uiowa.edu) and NRGIS
file server. A preliminary version of the statewide coverage is currently
available from Kevin Kane's FTP server (k2.gis.iastate.edu) and the ISU GIS Facility PC database
.
Click here to view a larger, somewhat
detailed map of Iowa GLO vegetation (42K bytes). Click
here to view an even larger, very detailed map of Iowa GLO vegetation
(138K bytes).
The final report for the research is available here on-line. The on-line
report includes 142 pages of text, illustrations, tables, and maps. It's
in Acrobat PDF format. To read the file, you must have Acrobat Reader and
WebLink plug-in on your computer (they (except the computer) are available
free from www.adobe.com). Local information about Acrobat and PDF files
is available from ISU
Extension. You must also configure your Web browser to activate Acrobat
Reader as the viewer for PDF files. If you've done all that and you want
the report in PDF format, you can download the main
report (approximately 750 Kbytes), the appendices
(approximately 1,225 Kbytes), or an example of a county
map (Polk--approximately 72 Kbytes) (Poweshiek--approximately 93 Kbytes) (Jasper--approximately
123 Kbytes) from the report. GLO vegetation
maps of other counties
are also available. 
New research is developing optical character recognition (OCR) scanning
methods to convert the GLO surveyors' field notes to digital form. The new
research also involves designing a database structure and queries to aid
researchers, historians, and resource managers in interpreting and analyzing
the content of the GLO surveyors' field notes. We've already begun to experiment
with document digitizing of the surveyors' field notes for townships in
Dallas County. We're also experimenting with different on-line hyperlink
and searchable techniques (
HTML and Adobe Acrobat) to make the field notes
an additional research tool for studies involving historic vegetation. Jane Chen's MLA thesis
is
part of this research. Support for the project comes from Dallas County
Conservation Board and Iowa State University. After completing the pilot
study in Dallas County, we hope to obtain funding to complete digitizing
of field notes for the remaining counties in Iowa. Recent research in
Polk County and Hamilton County
by Nicole Horst and Said Musli has resulted additional digital field
notes along with information
on GLO surveying practices and vegetation descriptions. Luke Sling's MS thesis
in the Water Resources program is also part of this research. Support for his
project comes from the Department of Landscape Architecture. As part of
his research, Luke developed an ArcIMS
site as a model for digital access to the surveyors' field notes.
Luke Slings is helping create a Web site with more GLO maps and digital field notes. See our new GLO Research Web site still being developed.
More information is available from
Paul F. Anderson, Department of Landscape
Architecture, 146 College of Design, Iowa State University, Ames, IA 50011
(515-294-8943 voice, 515-294-2348 fax, fridolph@iastate.edu).
This project was described in a Des Moines Register newspaper article by Larry Stone.
Last revision: 16 April 02
