The Lewis and Clark Historic Landscape Project

Abstract

This project seeks to:

1. develop a simply-formatted, easily-accessible computer database of the GLO surveyor field notes (1815-1819) of the original rectangular land survey that cover Missouri’s portion of Lewis and Clark’s route (1804-1806) on the Missouri River and,
2. reconstruct the landscape and course of the old Missouri River and produce graphic products that will depict the land much the same as was experienced by the Lewis and Clark Expedition members.

This project is being done as part of Missouri’s contribution to the national celebration of the Lewis and Clark Expedition Centennial (2004-2006).

A database of this historic corridor’s original land survey notes will provide us with a better understanding of the structure, composition and ecological processes (i.e. human and natural disturbance regimes) of the native ecosystems before they were altered by modern human activities. Such a database will help us understand the patterns and distribution of native ecosystems and early human settlement along the river. Additionally, this database should greatly increase our capabilities to more thoroughly understand the morphology of the Missouri landscape where nature is inclusive of human interaction with the land.

The database will importantly capture vegetative and non-vegetative aspects of the GLO (General Land Office) notes. Our goal is to make one pass through the GLO notes and move all information contained within the notes to a searchable computer database. Examples of types of information captured or recorded in the notes include: topography; suitability of soil for agriculture; general nature of the forest and undergrowth; prairies, grasslands, and other non-point vegetative associations; perennial streams (including their course, relative velocity, and measured width); evidence of coal, salt, lead, and other minerals; evidence of fire and wind-damaged stands of trees, Indian trails, roads, cabins, cultivated fields and fences, and various other human activities. These data are all recorded and entered into the project’s database structure. These records permit the use of these data within historical, hydrological, cultural, and archaeological studies, thus opening this information to a much wider group than conservation, natural resources, or biological research. Because all information is being captured in the database, and because this is historic information, there should very seldom be a need to amend, update, or add to this database once it has been created.

The deliverables that will occur as a result of this task are:

1. A stand-alone CD-ROM survey-notes product with graphic capabilities.

2. An increased opportunity for researchers to be exposed to and educated in the utility of the GLO survey and the Lewis and Clark notes within their discipline.

3. An increased and enhanced use of these data for public, educational, and research purposes.

Proposal Summary

With the exception of the Platte Purchase in northwest Missouri, almost all of the rapidly-settled area around the Missouri River was surveyed by the U.S. General Land Office (GLO) before 1820 and much had been completed within a decade of the return of the Lewis and Clark Expedition (1804-1806). This offers us the opportunity to create a database of this historic corridor’s original land survey notes and provide a much better understanding of the structure, composition and ecological processes (i.e. human and natural disturbance regimes) of the native ecosystems before they were altered by modern human activities. Such a database will help us understand the patterns and distribution of native ecosystems and early human settlement along the river. Additionally, this database should greatly increase our capabilities to more thoroughly understand the Missouri landscape where nature is inclusive of human interaction with the land.

The database will importantly capture both vegetative and non-vegetative aspects of the GLO notes. Our goal is to make one pass through the GLO notes and move all information contained within the notes to a searchable computer database. Examples of information captured or recorded in the GLO notes include: topography, suitability of soil for agriculture; general nature of the forest and undergrowth; prairies, grasslands, and other non-point vegetative associations; perennial streams (including their measured width); evidence of coal, salt, lead, and other minerals; evidence of fire and wind-damaged stands of trees, and evidence of human activity including: Indian trails, roads, cabins, cultivated fields and fences. These data are all recorded and entered into the project’s spatial and tabular database structure. These records permit the use of these data within conservation, natural resources, or biological research but also enable expanded use within historical, hydrological, cultural, and archaeological studies, thus opening this information to a much wider group. Because all information is being captured in the database, and the information is historic in nature, there should be no need to amend, update, or add to this database once it has been created.

The field notes of the rectangular land survey of the U.S. General Land Office (GLO) form the most comprehensive, detailed record of the nature of the land surface in the nineteenth century. With a grid cell covering virtually every square mile of the public domain, they form a systematic collection of reasonably objective information obtainable from no other source. Although the survey was conducted to facilitate the alienation of the public domain into private ownership, the field notes of the survey now stand as a historical document of major proportion.

Usually done for any tract of land just before it was officially opened to settlement by the U.S. government, the survey captures land conditions before wholesale transformation by farmers, stockmen, lumbermen, and others. Though the land surface is undergoing constant change by both natural and human agencies, the rapidity and thoroughness of change brought about by white settlement suggests that the land survey was conducted at a critical time in environmental transformation and social organization of the land and is sometimes referred to as "presettlement" conditions. Thus, most of the uses to date of the GLO field notes for research have been for reconstructing the natural environment, assuming that the "presettlement" survey reflected natural ecosystems relatively undisturbed by Euro-American settlement. Foresters and ecologists, especially, use the field notes to help restore ecosystems to a more natural condition. In the last score of years, however, other researchers have found the rich detail of the field notes a veritable mine for interpreting human activity within the environment. The field notes provide a wealth of detail unavailable elsewhere for this purpose.

The most extensive use of the Missouri GLO field notes for a single piece of research was achieved by Walter Schroeder in his mapping of the pre-settlement prairies of Missouri (1981). Anthropologist Michael O’Brien used the field notes to reconstruct vegetation patterns of the Mark Twain Reservoir area and developed models for prehistoric and early historic settlement (1984). Geographer Harbert Clendenen analyzed the field notes to uncover the presence of squatters and early settlers in the Current River region of the inner Ozarks (1973). Several theses relied on the field notes for reconstruction of historic vegetation in localized regions (Howell and Kucera 1956; Batek 1994; Porter 1998), and several agencies similarly relied on the field notes for small tracts; such as state parks, national forest purchase areas, and historic sites.

Why Missouri?

During the time of the survey in Missouri, the land was bought and occupied by settlers of various backgrounds: French Creoles, American hunter-woodsmen, American planters from the South with slaves, American farmers from north of the Ohio River, thousands of German immigrants, Mormons, and Utopians. All had to fit their settlements into the grid of the survey. For example, German immigrants could not reconstruct the compact villages they had known in Germany and were forced to separate themselves onto farms of the square and fat rectangles of the survey. Mormons and Utopians, in contrast, bought up large blocks of sections to lay out their planned towns and villages. Hunter-woodsmen defied the cadastral structure of the survey and continued to use the land as if it were public domain. This proposed database would provide information on how different cultural groups acquired and used land along the Missouri River.

The physical geography of Missouri included several thousand square miles of timbered swamps in the southeast, closed forests and open savannas in the Ozark hills, and rolling prairies in the west and north. Missouri’s varying terrain challenged the ingenuity of the surveyors. In the rough Ozarks the surveyors initially added an arbitrary extra mile to every six-mile township to compensate for running their chains up and down the steep slopes. Bodies of iron ore hidden deep below the surface created local magnetic anomalies that the surveyors were hard pressed to comprehend. The wide Missouri River, which bisects the state, posed a problem, never solved, of how to survey across it (the Ohio and Mississippi lay on state boundaries and no survey lines were required). Surveyors abandoned their initial attempts to survey in swamps as physically impossible and the survey was delayed in those areas for nearly thirty years. Missouri posed as a very difficult testing ground for surveying in these different environments.

Finally, in a larger geographical context, Missouri occupies a transitional location in plant and animal distribution of the continental U.S. That is, subtropical species range northward into southern Missouri, Great Lakes species range southward into northern Missouri, Great Plains species range into western Missouri, and eastern forest species range into eastern Missouri. The state’s flora and fauna contain isolated remnants from past climates (relic plants in Ozark refugia where they had spread during Wisconsin glaciation). Additionally, plant ecologists recognize the upland Ozarks as one of the continental centers of plant endemism, where new species have emerged or have persisted. All of these areas would be sampled during our reconstruction of the Missouri River corridor and the meticulous detail of the land survey in regard to plant species, numbers, sizes, and precise location would help the understanding of Missouri’s key ecological role in native plant distribution.

The Rectangular Land Survey’s Role in American History and Landscape

The rectangular land survey began with the creation and consolidation of the public domain. It was a necessary prelude to the opening of the public lands for sale to individuals. Developed by Thomas Jefferson and others, the survey divided the surface into a straight-line geometry of squares and rectangles in order to facilitate Jefferson’s notion/dream of the American Yeoman Farmer spreading out and settling the land westward. It brought cadastral order to the country. All land ownership in the United States that was covered by the survey is tied into the system, including city lots and subdivisions (Thrower 1966).

The survey’s great benefit lies in the simplicity and brevity of property description, down to the surveyed section in size. It is unambiguous. It covers the land surface entirely and can be extended over space indefinitely. Monumented and marked lines and corners can easily be relocated, if lost. Any person, by using monuments and markers, can readily find any tract on the land (Donaldson 1884).

Historians have long recognized how well the rectangular land survey enabled Americans to spread across the country in an equitable and organized manner. The system allowed any settler to cheaply obtain the land of his choice with secure title. The system worked against land speculation and monopoly, emphasizing a yeomanry of family farmers. It promoted widespread dispersal of settlers and discouraged clustered ethnic villages of the European style. Once the elements of the survey had been worked out in the Northwest Territory, only minor modifications had to be made for changing environments, such as using sod mounds in prairies due to the lack of trees to blaze. The disposal of the public domain was predicated upon this well-conceived , well-conducted land survey. Historians have long admired the process of alienating the public domain into private ownership and conclude that it promoted a nation of independent, self-reliant, productive citizens. Congress used the public domain to promote the settlement and economic development of the country, such as grants for railroad and canal construction, grants for timber culture on the plains, relinquishment of swamps to states for drainage and development, and grants for higher education. Also, by reserving one thirty-sixth of the land for support of public schools (section 16 of every township), Congress greatly promoted local education efforts (Gates 1996, Thrower 1966).

Methodology

The field notes of the GLO comprise, by far, the most useful collection of land information for pre-European settlement conditions in Missouri. They include both quantitative and qualitative information on the natural environment, but especially trees, since trees were both a major resource for settlers and a major indicator of land quality. Information in the field notes is systematic; instructions to surveyors were reasonably standardized as to what was to be recorded and how. Information was gathered geographically in a systematic fashion according to the lines and points of the land survey. Lastly, information was gathered, in general, just before permanent settlement of the land was expected and therefore reflects conditions of a "natural" or least disturbed (at least by white settlers) environment.

Other pieces of information are also recorded such as human habitation, stream widths, landscape features, and other bits of information. The level of detail recorded varies by surveyor and date of survey as the rules for documentation within the survey changed through time. The field notes reflect conditions at only one specific point in time within a constantly changing natural environment, but no other alternatives exist for such a complete and systematic reconstruction of natural environmental conditions.

Other sources, such as travelers' accounts and soil surveys, are subject to their own, more serious problems for analysis, although they do serve as checks for the land survey field notes. In sum, the GLO field notes are recognized as the single best historic landscape/ecological database against which change can be measured.

The GLO conducted its land survey of the area of Missouri beginning with the running of the Fifth Principal Meridian in 1815 and extended into the 1850s. Different tracts were surveyed by different surveyors under separate contracts. Some tracts were re-surveyed. The field notes exist in two forms, the original notes and notes rewritten under WPA auspices. Both forms exist in sets of bound volumes and on microfilm. The rewritten notes have been shown to be virtually perfect copies and are preferred by users because of the relative ease of reading them.

When the survey was being conducted in Missouri, the instructions given in each surveyor's contract sometimes varied. However, for the purposes of this project, it is believed that the information was sufficiently standardized in the instructions and as they were carried out in the field. One congressional township of 36 square miles yields data on 133 points (of which 25 would be duplicated when aggregating townships) and 84

linear miles (of which 12 would be duplicated). All facets of the surveyor notes are captured for query and retrieval purposes within the computer.

This capture has been facilitated by the development of AMLs (Arc Macro Language Programs) and user-interface menus to guide and standardize the recording of the information contained in these notes within the computer system. The data are entered directly into the computer system and a system of checks has been created to ensure quality of data entry. These programs reside at the Geographic Resources Center and were developed during a contract with The Nature Conservancy and further modified through following efforts. We currently have on-site, 5 microfilm readers (used during the pilot project), the statewide land-net (24k and 100k PLSS), as well as the ties to the GLO notes and their availability.

Some biases are recognized in the recording of this information. For example, lacking specific guidelines as to which trees they should select or even in the presence of such guidelines, surveyors differed in their selection of trees for field marking and recording. This surveyor bias can be checked by comparing one surveyor's tract against another. Also, it is thought that many surveyors chose not to blaze cedars because of the difficulty of accessing and blazing their trunks; cedars would then be underestimated. Species of smaller size, like dogwood, redbud, and wahoo, may have gone undocumented because they were not chosen for blazing. Also, the largest-diameter or oldest trees may be underestimated because it is thought that surveyors selected trees that would have a longer life after being blazed. These are problems inherent in the survey that an experienced, knowledgeable, and careful interpreter needs to acknowledge for greatest confidence in the results.

Surveyors recorded quantitative information on tree species, tree diameter, and distance and direction of up to four trees from each section corner (one tree in each quadrant) and two trees at each quarter section. Other trees along lines may also have been noted. This information is spatially available for a regular square grid of points one-mile apart, at section corners, with supplementary information at half-mile intervals along section lines. Information is lacking for the interior of sections. However the quadrant distribution of data lend themselves to statistical and spatial analysis and inference. For example, density and basal area of trees may be calculated as well as relative importance of various species. These procedures have been used and accepted by researchers for some time. By accepted spatial statistical techniques, the very random nature of the surveyed grid lines over the uneven landscape are conducive to statistical analysis.

In addition to the quantitative information on trees, qualitative information is available through verbal description of the topography; suitability of soil for agriculture; general nature of the forest and undergrowth, prairies, grasslands, and other vegetative associations, along every line surveyed. Information is also available on all perennial streams, including their measured width; evidence of coal, salt, and other minerals; evidence of fire and wind damage of trees, trails, roads, cabins, cultivated fields and fences, tree deadening, and other human intrusions. These data are recorded and entered into the database structure that was developed and adopted by an interagency oversight committee.

The methods that are used to accomplish the compilation, organization, and

representation of these historic data are a culmination of the thesis research of two successful GGRC graduate students, Mike Batek MA ‘94 and Shannon Porter MA ‘98, and the efforts that are currently taking place in the GGRC on the project. These procedures have been refined over time to be effective in all of the data from the GLO (General Land Office) notes and digitizing these into a GIS (Geographic Information System) environment. These procedures can be logically separated into the categories of 1) data entry, 2) data organization, and 3) data visualization and representation.

Data Entry

The GLO notes are received from the Missouri State Archive offices in the form of microfilm rolls. These rolls are catalogued by survey date and are cross-referenced to survey township and range on a grid map of the State of Missouri. The following are the general steps taken in the data entry process:

1. The data entry technicians (student interns) check this map for the appropriate roll number, retrieve the roll from the secured cabinet in the GGRC lab, and install the microfilm in a microfiche reader. The microfilm is then advanced to the appropriate page by referring to the written headers which contain survey date, contract date, surveyor name, survey township, and survey range.

2. The entry of the data into the GIS computer database is facilitated through the use of a menu-driven user-interface program that prompts the technician for information on a step-by-step basis. Technicians first enter the numbers of the applicable township and range. The program then retrieves the necessary geo-referenced line data from the Public Land Survey boundary files that are stored within the GGRC’s archived GIS data.

3. A pop-up window depicting the land survey grid appears and prompts the technician to enter the start point on one of the four corners of the township, the direction of travel along the arc or line (N, S, E, W), and the end point or distance traveled along this line measured in survey chains and links. This establishes the line or arc upon which the information as written by the original surveyor will be attributed.

4. Another pop-up window appears and prompts the technician to enter the technician name, the current entry date, the microfilm number (film in use), the microfilm volume (volume number of archived data), the page number on which the data begins, the surveyor name, the survey contract date (if given), and the date of the survey itself (Fig. 1).

5. A pop-up then appears and prompts the technician to identify whether the following data will be entered as points or lines. The character of the GLO survey data exists as these two geographic features since the surveyors measure and record the landscape at a certain place (point) or generally describe the landscape when traveling along a certain distance (line). The technician first selects the point category and is prompted to enter data concerning either trees or water features.

6. Tree data are entered as either a bearing (witness) tree (a tree identified off of the section corner points or quarter-section corner points at a certain distance and direction) or as a line tree (a tree identified as being located on the survey line). The technician is prompted to enter in the given information as to the species name (i.e. white oak, hickory, etc.), the distance along the line (chains and links), the diameter of the tree at breast height, and the bearing and distance to the tree (if a bearing tree and given as i.e. S 30 degrees W at # chains, # links). The tree points are additionally coded as to whether the trees are to be considered section-corner trees (usually four identified with one in each quadrant), quarter-section-corner trees (usually two identified), or as a line tree encountered along the survey line (Fig. 2).

7. Water features are similarly entered as they were either identified at a certain direction and distance or were encountered along the survey lines. The technician selects the water features category and is prompted to enter the distance along the survey line (chains and links), the direction of travel (N,S,E,W), the type of water feature (lake, pond, river, creek, brook, slough, etc.), the bearing or direction to the feature, the width of the feature, the velocity of the feature if appropriate (swift, slow-moving, still, etc.), and the cultural description of the feature usually given as a name (Missouri River, Chariton River, Jacob’s Slough, Muddy Creek, etc.) (Fig. 4).

8. After finishing the entry of the point data, the technician returns to the main menu and selects ‘arcs’ (lines) as the next feature entry. Here they are prompted to select the ‘vegetation’, ‘land’ or ‘human’ categories. First the technician enters in the general descriptions of the vegetation along the survey line (Fig. 3). The data entered consists of distance along the line (chains and links including start and stop points), the general description and condition of the vegetation (i.e. timber, scattered timber, prairie, barrens, brushy, burned, fired, etc.), and the more specific description of the vegetation if written in the notes (i.e. oak, hickory, maple, hazel undergrowth, blackberry bramble, elm grove, etc.). This procedure importantly stores information as to where forests, prairies, swamps and such were entered and exited by the surveyors thus giving approximations to both size and location of these features. In addition, information about the burning of the vegetation yields possible connections to human interaction with the land at the time.

9. Following the entry of the vegetation information, the technician selects the ‘land’ category. The data entered consists of the original surveyor’s own words in describing the land’s quality in relation to settlement, (i.e. good soil, fertile soil, fit for cultivation, unfit for settlement, rocky, too steep, too wet, prone to flooding, etc.), land elevation, (high ridge, low bottoms, level prairie, etc.), and general land description (rolling, hilly, flat, etc.) (Fig. 6).

10. The final arcs or line entries are concerned with cultural or human features encountered by the surveyors (Fig. 5). These data concern information on American Indian inhabitants, the early American migrant, and the European settlers of Missouri in the early nineteenth century (Missouri, Osage, French, Spanish, Irish, German, etc.). The entries include structures (i.e. barns, houses, fences, corrals, Indian mound, etc.), human

vegetation (corn field, pasture, garden, farmed, etc.), roads and trails; descriptions as to

location and type (i.e. "Osage Indian trail bears N and S" or "Mormons heading north on road"), and names of the features if applicable (i.e. Mr. Jones’ farm, Mr. Jones’ field, or Mr. Jones’ mule). This information most importantly yields where and how humans were interacting with the land at the time of the surveys.

11. After the entry of the human features, the database is then checked by another technician to maintain and ensure quality control. This consists of the QC technician operating another program that examines the structure of the newly created file in search of any anomalous data and also prompts the technician to visually check the data on screen in search of any problem areas. Data found to be erroneous is then corrected by the QC technician or referred to supervisory personnel for further investigation.

Figure 1: User-interface menu for entry of general information.

Figure 2: User-interface menu for entry of tree and survey point descriptions.

Figure 3: User-interface menu for entry of vegetative line or arc descriptions.

Figure 4: User-interface menu for entry of water feature descriptions.

Figure 5: User-interface menu for entry of human feature descriptions.

Data Organization

The entire dataset consists of files of either points or lines. Each of these features has its own inherent items that are created to maintain the topological integrity or feature position and identity in relation to other like features within the same file. In other words, every point and line has its own solely occupied place and its own unique identification number.

Items that are generically inherent within the arc or line files are:

FNODE-point or node from which the line begins,

TNODE-point or node at which the line ends,

LPOLY#-not applicable in line-only files,

RPOLY#-not applicable in line-only files,

LENGTH-length of the line in meters,

FILE#-unique internal feature number,

FILE-ID-another unique internal feature number,

LINEDIR-the direction of travel along the section line.

Items that are generically inherent in the point files are:

AREA-not applicable (points have no area),

PERIMETER-not applicable (points have no perimeter),

POINT#-unique internal feature number,

POINT-ID- another unique internal feature number.

CHAINPT-the distance along the line measured in survey chains,

LINKPT-further distance measured in survey links.

Other items that are generic to all files are:

TECHNICIAN-the name of the person entering the information into the computer,

ENTDATE-the date that the information is entered into the computer,

FILMNO-the microfilm number from which the information is read,

VOLNO-the volume number of the original survey notes,

PAGENO-the volume page number of the original survey notes,

SURVEYOR-the name of the original surveyor who wrote the notes,

CONTDATE-the contract date of the survey operation (often not given),

SURVDATE-the date that the survey was conducted along the section line.

Items that are attributable to the vegetation description files (arcs) are:

LINEDIR-the direction of travel along the survey line,

TVEG1-the first species of timber listed along the survey line,

TVEG2 through TVEG10-other timber species,

TVEGCT-the total number of timber species listed along the survey line,

UVEG1-the species of understory vegetation listed along the survey line,

UVEG2 through UVEG10-other understory species,

DVEG1-the general description of the vegetation (i.e. timber, prairie),

DVEG2 through DVEG10-other descriptions of vegetation.

Items that are attributable to land description files (arcs) are:

LINEDIR-the direction of travel along the line again,

QLAND1-the surveyor’s opinion of the quality of the land (i.e. fit for cultivation),

QLAND2 through QLAND3-further opinions if given,

DLAND1-the description of the landscape (i.e. rolling, hilly, flat),

DLAND2 through DLAND3-further descriptions if given,

ELAND1-the relative elevation of the land (i.e. high, low),

ELAND2 through ELAND3-further elevation data if given.

Items that are attributable to the human description files (arcs) are:

SHUMAN1-a human structure listed (i.e. barn, cabin),

SHUMAN2-another human structure listed,

VHUMAN1-a type of human vegetation or cultivation listed (i.e. garden, corn field),

VHUMAN2-other human vegetation listed,

RHUMAN1-a road or trail listed,

RHUMAN2-another road listed,

DHUMAN1-a general description of human activity (i.e. trappers traveling north),

DHUMAN2-another general description of human activity,

NHUMAN1-a person’s name listed as attributable to the location (i.e. Mr. Jones’ farm),

NHUMAN2-another name listed.

Items that are attributable to the water description files (points) are:

TYPE-the type of water feature (i.e. creek, lake),

BWAT-the bearing or direction of the water from the point,

WWAT-the width of the water feature in survey links,

VWAT-the velocity of flowing water (i.e. swift, sluggish),

DWAT-the description or name of the water feature (i.e. Chariton River),

Items that are attributable to the survey point and tree files are:

PTLN-a numeric item that is internally calculated to distinguish the type of point (The numbers are 6-9 with 6 = a survey point (no tree), 7 = a tree encountered along the survey line, 8 = a tree used as a witness to a quarter-section corner point, and 9 = a tree used as a witness to a section corner point.),

SPECIES-C-the species of the tree listed,

DIAM-the diameter of the tree in inches,

DIR1-the first cardinal direction,

BEARING-degrees of the direction,

DIR2-the second cardinal direction,

DIST-the distance to the tree in survey links,

FEATURE-a marker placed by the surveyors at the point (i.e. stake, stone, mound).

As stated earlier, all of the above-listed data are stored and inherently linked to geo-referenced (x and y geo-coordinated) lines and points within the GIS environment. This presents the opportunity to examine any or all of the attributes within this database for statistical spatial analyses and/or the production of powerful graphics, including maps for visual analysis and publication.

Data Representation and Visualization

This database is being evaluated using spatial modeling techniques within the Geographic Information System ARC/INFO. All data entries, whether specific point data or line data, will be ascribed to a specific point or series of points to be used in further analysis. Each attribute within the database can provide a separate and distinct view of any given area. For example, we might be interested in extracting only those points referring to the occurrence of white oak and from these points construct a presettlement distribution map of that species. We might also want to resample these white oak points to break out species size distributions for viewing the distribution of size classes in the region. These maps could then be combined with maps of other species to form a

composite map that would show the community structure of the presettlement vegetation. In this manner we have reconstructed the presettlement vegetation as it occurred during the survey period for the pilot region. These data can then be related to other databases being gathered on the site for further analysis of possible interrelationships of presettlement vegetation and soils, geology, etc. This data layer is also useful in identifying possible 'old growth' areas within the area. The field notes also record occurrence of fire and windthrow within the ecosystem which can be used as a measure of natural disturbance cycles and distributions.

The specific analytical methodologies and techniques have been developed in concert with the review and evaluation by an interagency oversight committee. From these findings and exploratory analyses, a standard analysis was developed which creates a single series of maps and spatial databases depicting the historic vegetation of Missouri. This includes several components such as overstory, density, size classes, understory, and cultural features if applicable.

The analyses that have been performed by others using GLO notes, have generally been concerned with establishing the areal extent and type of vegetation or land cover that was present at the time of the early surveys. Many across the U.S. have used the bearing trees at the section corners as a systematic sample in analyzing what the vegetation might have been. While this procedure has been substantiated through its implementation in a variety of research projects, the procedure generally ignores the information that was collected while the surveyors were traveling along the section lines and only yields data points whose closest neighbors are one mile away in distance.

The challenge for this project, in relation to historic land cover, has been in combining the bearing tree data, the line tree data, and the descriptive information recorded along the survey line, as well as interpolating into the interior of the mile-wide sections. As discussed earlier, the line descriptions usually include a list of timber species, a list of understory or brush species, and a general description of the vegetation along the way. The following are the general procedural steps that are currently being used to consider both the line and point data and establish historic land cover/land use for Missouri:

1. The line data are analyzed by key words and phrases in the surveyors’ descriptions to establish whether the features can be considered grassland (i.e. "prairie", "barrens", "glade", "meadow") or forest/woodland (i.e. "timber", "stands", "woods", "groves"). These data are then selected out and separated into their respective files as grass or woods.
2. These files or coverages are then converted into 50-meter resolution grid or raster files, which are much more amenable to spatial analysis techniques.
3. The point data are analyzed and those points that are representative of trees are selected, separated, and also converted into a 50-meter resolution grid file.
4. A type of euclidean distance analysis is performed on these grids that assigns probability values as a function of distance to every grid cell in the analysis area. This results in grid files of grass and trees containing values ranging from 0 to 100 across the region.
5. These grid files are combined for boolean analyses which establish a general land cover grid based upon the probability values for being either forest/woodland or grassland.
6. This general land cover grid is then visually checked and edited against a DEM (digital elevation model) and the hydrography, where logical modifications are made according to the topography. This generally involves the smoothing of forest areas along slope contours, the extension of fingers of prairie on ridgelines, and the extension of wooded riparian corridors that cut through and between the 1-mile sections.
7. A frequency analysis is then performed on the tree point data to establish the most often reported species of trees in the analysis region. This usually yields a report of 15-20 high frequency species.
8. These high frequency species are selected out of the point file and converted into their own respective 50-meter grid files.
9. These tree species grids are combined into a ‘stack’ (grid analysis package available in GIS) and a ‘maximum likelihood cluster’ analysis is performed to aggregate the grid cells into classes or clusters based upon their spatial covariance or tendencies to exist together or apart in space and value. This yields a signature file that describes the species, probability values, and total cell numbers for any given number of classes. The signature file is then used to establish the different types of forest/woodland by species and number of species across the analysis region. These forest/woodland types or associations are then integrated into the earlier general land cover thus yielding a much more specific land cover.
10. Similar analyses continue to establish even further specifics according to whether the land could be considered high prairie, wet bottomland prairie, upland forest, bottomland hardwoods, brush/understory components, etc. The tree data are also analyzed at the corner points for tree diameter and density across the analysis region to establish an idea, based upon the calculation of canopy cover, of the extent of actual forest, woodland, and open woodland or savanna.

After these analyses have been performed more informative, intuitive databases can be created which enables the GRC to produce highly informative maps to accompany any raw data that is disseminated to the user-public.

Products and Approaches

Phase 1 - Internet Based Browse and Search Extensions for Missouri State Archive GLO Notes:

The GLO database will include a browsable spatial metadata descriptor in verbose format as HTML documents that are accessed from the data set listings page. These files are maintained on-line. By creating maps and graphics within ArcInfo of these spatial data sets and through the use of a graphic compression program called MrSID these images will be linked to the data set listing area, FTP area, and metadata area of the GLO database. With this graphic browse capability a user will be able to preview the GLO databases and their mapped holdings within structured pieces of Missouri's geography. Through this funding we will also compile and/or install the tools for both keyword (indian, trail, prairie, date of survey, etc.) and spatial (township-range, county, etc.) search requests from the net browser.

These tools will enable clients of Archive GLO information to query specifically identified fields of information and define what operators will work with those fields. They will also permit the standardization of the results, and their format, that are to be returned to the client based on their request. We may expand the keyword attribute set to recognize other fields that are components of the Missouri site if they are requested at high frequencies by users.

The deliverables that will occur as a result of this task are:

1. Historic Vegetation Map (hard copy and digital versions with tables of point

information by vegetation category)

2. Point, line, and composite databases, coverages, and grids for the top 20 species

within each unit.

1. Select databases, coverages, and grids of understory components as determined as having significance within the unit.

4. Surveyor map and database for interpreting potential bias or other phenomena.

5. Density map for the unit to assess spatial structural components of vegetation

categories.

6. Size class map for the unit to assess size structural components of vegetation

categories.

7. Data bases for:

a. Corner trees d. Hydrologic elements

b. All points e. Landscape elements

c. Human elements f. Line descriptions

9. Increased and enhanced use of Missouri State Archive's GLO data in support of

this event as well as other uses of this data.

Phase 2 - Missouri State Archive Lewis and Clark Expedition CD-ROM:

The deliverables that will occur as a result of this task are:

1. Stand-alone Missouri State Archive CD-ROM GLO product with mapping and

query capabilities showing the data and landscapes of the early 1800's. Possible

links to scanned images and other data concerning the Lewis and Clark Expedition

are highly possible.

2. Missouri teachers, librarians, and other researchers exposed to and educated in the utility of GLO within their area as demonstrated through this project.
3. Increased and enhanced use of Missouri State Archive's GLO data for the Missouri River corridor.

References