Categories: Maps and Cartography

Using GIS to Track Historical Land Cover Change and Growth Rate at Fort McCoy

Located in the heart of the Upper Midwest, Fort McCoy is the only U.S. Army installation in Wisconsin. The installation has provided support and facilities for the field and classroom training of more than 100,000 military personnel from all services each year since 1984. The Fort McCoy complex is situated on 60,000 acres, 46,000 of which are contiguous live-fire and maneuver areas. Fort McCoy provides reserve- and active-component forces with the networked, integrated, interoperable training resources required to support the Army’s training strategies using a full spectrum of facilities, ranges, and training areas. From 1990 to the present day, new construction projects have served to modernize the post’s infrastructure, facilities, and training areas (The Real McCoy, 2019).

Ad

In order to minimize maintenance costs and ensure the long-term utility of military training lands, it is necessary to inventory and classify the lands relative to their environmental condition and their ability to sustain various kinds and intensities of military training in the future (Warren, Diersing, Thompson, and Goran, 1989).

Spatial information has always been important to military commanders; an understanding of terrain, for example, is an essential military skill. GIS has a key

role to play in creating, editing, analyzing, querying, and displaying geographical data in order to help the commander understand the influence of terrain on the conduct of the battle (Swann, 1999). Geographic Information Systems (GIS) play a pivotal role in military operations. As the operational picture and battlefield develop, everything is essentially spatial in nature.

The Department of Defense (DoD) is responsible for administering more than 25 million acres of federally owned land in the United States making it the fifth largest federal land managing agency. Often the factors affecting land condition are of different scales and their values are of different magnitudes (Mendoza, Anderson, and Gertner, 2002).

Different missions and focuses play an integral part in the asset and resource planning related to any such land usage and mission sets, from training considerations, environmental considerations, and organizational considerations. As stated by Swann (1999), asset and resource management have always been a problem.

Due to the nature and intensity of the activities occurring on many military training areas, management of those lands can be a complex problem. The scale of the problem is enormous. The number of buildings, the length of roads, complexity of infrastructure, and area of land involved are similar to that of a large local government user. With assets and resources for action and usage being often dispersed nationally and internationally, effective management is problematic (Swann, 1999).


Advertisement


In this study, the Fort McCoy military installation was analyzed using imagery spanning 18 years to identify land use change in the following classes: (a) complexes, (b) blacktop, (c) agricultural/green space, and (d) forested lands. The maximum likelihood classification tool and imagery interpretation was used to explore imagery to evaluate land use changes amongst the 2004 and 2018 years to explore changes in land use. Time periods selected were based on availability of finer resolution imagery, time periods of global military needs, and resource constraints.

Methods

Study Area

Fort McCoy is a United States Army installation on 60,000 acres between Sparta and Tomah, Wisconsin, in Monroe County (Figure 1).

Figure 1. Fort McCoy study area located in west-central Wisconsin. Proximity of next major towns: 9 miles east of Sparta, WI, 12 miles west of Tomah, WI, 28 miles south of Black River Falls, WI.

Fort McCoy is located in Monroe County between the cities of Sparta and Tomah and roughly 30 miles east of La Crosse in west-central Wisconsin. The installation is divided by State Highway 21. Since its creation in 1909, the post has been used primarily as a military training center (Figure 2).

Figure 2. Fort McCoy military installation located in Monroe County, WI. The installation, which occupies a land area of approximately 60,000 acres, is divided by State Highway 21.

From 1990 to the present day, new construction projects have served to modernize the post’s infrastructure, facilities, and training areas. The installation has provided support and facilities for the training of more than 100,000 personnel annually since 1984. Today, the post provides full-scale support to its customers at each juncture of its training triad — transient, institutional, and exercise (The Real McCoy, 2019).

Most of Fort McCoy’s 1,000 buildings with 5 million square feet of area are within the triangular shaped Cantonment Area that covers approximately 2,600 acres. The Cantonment Area is surrounded by approximately 114,000 acres of maneuver area, 7,600 acres of impact area, 1,400 acres of ranges, and 640 acres in airfield of land owned by the government.

Work Flow

The analysis entailed a variety of processes from initiation to completion. According to Bangerte (2017), logical data flows focus on what happens in a particular information flow and what general processes occurred.

General work flow processes involved in analysis included data collection, data preparation, classification system establishment, classification tool identification, training sample creation, raster output analysis, raster to GRID conversions, class acreage change analysis, results, and discussion as to explanations for land cover changes. ArcMap was used for classifications, conversions, and overall analysis of land cover changes.

Data Collection

Imagery Acquisition

Imagery for this study came from a variety of sources and was evaluated for image resolution and suitability for image classification. The Fort McCoy Public Affairs and Geographic Information Systems/Integrated Training Area Management office were contacted in attempts to obtain installation GIS and imagery/vector data. All installation procured data is treated as FOUO (For Official Use Only) and thus unavailable for general public distribution without the proper declassification. Consequently, ample public imagery sources were available and subsequently used for input data for the study.

Publicly available imagery included aerial imagery from the National Agriculture Imagery Program (NAIP) within the United States Department of Agriculture (USDA), U.S Geological Survey (USGS) EarthExplorer, and the Monroe County Land Information Office (GIS Office).

Imagery downloaded came in digital ortho quarter tiles (DOQQs) or compressed county mosaics (CCM). Imagery came in either .tiff or .sid files depending on the downloading source.

Data Preparation

Satellite imagery gathered of the Fort McCoy (Figure 3) area included years 2004 and 2018.

Figure 3. Ft. McCoy area of interest/study area. Fort McCoy serves as a Total Force Training Center that supports the year-round training of Reserve, National Guard and active component U.S. military personnel from all branches of the armed services (Real McCoy, 2019).

Preparing, exploring, and later grouping the images into one image (mosaic) was an essential step to start the study in order to successfully run classification tools.

Imagery had to be mosaiced for the study area before classification tools could be run successfully (Figures 4 and 5). A mosaic is a combination or merge of two or more images. Mosaicking the imagery made the process streamlined only using one image as the raster input and made a variety of clustered images into a standardized image set for each year, creating an easier method of visualizing the study area as one image instead of multiple images.

Input imagery data sources were mosaiced together using ESRI’s ArcGIS raster dataset tools. The Mosaic tool merges multiple existing raster datasets into a single existing raster dataset.

Figure 4 (left) 2004 USGS imagery mosaic used for study area. Figure 5 (right) 2018 USGS imagery mosaic used for the study area.

Classification System

Military land often serve a variety of additional uses such as: timber production, agriculture, livestock grazing, off-road vehicle recreation, and hunting. In light of the potential cumulative effects of larger-scale and more intense military training, coupled with other uses, the military community has become increasingly aware of the need to maintain or improve the condition of its lands (Warren and Bagley, 1992).

The Land Condition Trend Analysis (LCTA) program is the Army’s standard for land inventory and monitoring, employing standardized methods of natural resources, data collection, analyses, and reporting designed to meet multiple goals and objectives within owned lands (Anderson, Guertin, and Price, 1996).

Standard LCTA methodology was used in combination with the Anderson classification system (Anderson, Hardy, Roach, and Witmer, 1976) to create the following classes (a) Complexes, (b) Blacktop, (c) Agricultural/Green Space, and (d) Forested Lands (Table 1).

Maximum Likelihood Classification

Maximum likelihood classifier (MLC) is the most widely adopted parametric classification algorithm (Manandhar, 2009). An MLC was chosen due to errors in object classification within unsupervised classification.

Rozenstein and Karnieli (2011) state training maximum likelihood classifiers are more properly applied to an area for which one is more familiar with, and as such, creating classifiers with knowledge of the area provided rationale in utilizing MLC for this study.

ClassDescription
Forested LandsAreas characterized by tree cover. Includes areas with deciduous, evergreen, and/or mixed forest types.
ComplexesA group of similar buildings or facilities i.e barracks, training facilities, administrative buildings, etc.
Agricultural/ Green SpaceAreas characterized by vegetation managed for production of food, feed, or fiber. Includes pasture, hay, row crops, small grains, fallow, open fields and recreational land.
BlacktopAreas characterized by a constructed material including low/high density land surfaces such as roads and parking lots that repel rainwater and do not permit it to soak into the ground (i.e roads, parking lots, highways, gravel lots).
Table 1. Image class descriptions.

Training Samples

Maximum image classifications require training classifiers to assign pixels or objects to a given class using training samples. Representative training samples

for all land cover types (classes) identified in the image had to be attained before conducting land change assessments.

Histogram analysis was used for training sample grouping (Figure 6 and Figure 7). Due to the histograms following a normal distribution, having similar peaks, and overlaps between samples created, they were merged into their respective identified classes and used to create separate signature files for each year of imagery. Training samples that had close peaks and overlap between identified samples were merged and saved as a signature file and used as input for the maximum likelihood tool for classification output.

Figure 6. Agriculture class histogram example merged into one class due to overlap training samples.
Figure 7. Blacktop training sample histogram class analysis.

Analysis

Maximum likelihood classifications were conducted for each respective mosaic image of the study area. Classification area was defined by creating a polygon around the area of interest defined by what is called the Cantonment Area (Figure 8).

A polygon was drawn encompassing the study area of the Cantonment Area for land cover analysis characterized by zones of development within the general triangular base structure. The polygon was used as the processing extent dimensions for the classification tool in raster outputs, but does not encompass the maneuver, impact, ranges, and airfield areas.

Figure 8. Polygon used as the processing extent for classification tool.

A maximum likelihood classification was conducted on the raster bands and produced a classified raster output clipped to the study area polygon encompassing approximately 1,987 acres. (Figure 9).

Figure 9. Raster output for the 2004 study area consisting of approximately 1,987 acres. Maroon = complexes, black= blacktop, light green= agricultural/green space, dark green= forestry.

While the classification process utilized the study area polygon, the output extend for a raster is defined by the rectangular extent of the polygon as shown in Figure 9. The final analyzed area contains land fully operated by the Fort. The size of the maximum likelihood classification results is approximately 4% of all the total land contained within the Fort McCoy property.

Landcover area changes were evaluated by determining changes in area between both images of 2004 and 2018 after classifications were calculated.

Calculating Areas

Maximum likelihood outputs were converted to integer grid formats to be able to determine area changes in square meters for each land class (Figure 10).

Figure 10. Export raster to GRID integer format data export process.

Changes in land area were evaluated for 2004 and 2018. Using the count field in the value attribute table for each raster dataset, area totals in square meters were calculated for each land use category and subsequently converted into acres as land parcels typically use acres for defining parcel size and ownership.

Calculating class conversions between the years was conducted through converting raster data to features. Each raster was converted using the Raster to Polygon tool. Amount of change between each land cover classification was calculated by using the Union tool and Dissolve tool. The Union tool was run to create overlapping features from both polygon feature datasets and the Dissolve tool was run to merge like attributes using the class attribute for both years.

Ad

Results

Land Cover Change Analysis

The acreage for each land cover class was calculated for all four categories along with percentage of change from 2004 to 2018 in an area of approximately 1,987 acres (Table 2).

Class2004 Acreage2018 AcreagePercent Change
Forested Lands572.51421.31-26.41%
Complexes353.91282.75-20.10%
Agricultural/Green Space656.23756.8315.33%
Blacktop403.38524.7730.09%
Table 2. The amount, in acres, of each land cover classification and percent change between 2004 and 2018.

All classes saw change over the 14-year span. From 2004 to 2018, blacktop realized the greatest increase of 30%. Agricultural/green space lands increased by 15%, complexes saw a decrease of land by 20%, and forested lands saw a decrease of land by 26%. Differences in land cover classes are reflected in Appendix B and Appendix C.

In 2018, the majority of land cover from 2004 were forested lands losing about 39% to agricultural/green space; complexes losing 38% to blacktop, agricultural/green space losing 16% to blacktop, and blacktop losing about 17% respectively. Table 3 represents class changes between 2004 and 2018.

Table 3. The comparison of land cover classification changes, in acres lost, between unique classes between 2004-2018.

Accuracy Assessment

The classification process was straightforward; however, the result can be incorrect if training samples were not well chosen. New signature files were created for each year set but errors could have still taken place. Image resolution in earlier time periods and interpretation of the area, can be subjective and also could have added to the potential inaccuracy of the study.

Image resolution in between time periods differed with images used. The tiling format of the 2018 NAIP imagery was based on a 3.75′ x 3.75′ quarter quadrangle with a 300-pixel buffer on all four sides using 4 band colors and 32-bit pixels and a .6 x .6 cell size as compared to the 2004 NAIP imagery based on a 3.75′ x 3.75′ quarter quadrangle with a 360 meter buffer on all four sides using only 3 band colors with 24 bit pixels and 1 x 1 cell size.

The limited number of identified classes in this study did seem to cause some issues but not impact classification categories in repeated attempts to refine classes. Reject fraction for both classifications were set at 0.0 which means that every cell was classified into the assigned four categories.

Conclusions

Land cover changes are continually occurring due to a variety of reasons including land purchases, eminent domain, new zoning, and joint planning efforts the study area and the land surrounding it can be can be expected to follow the general flux of this trend as more land is acquisitioned, expanded upon, and developed as the trends seems to suggest.

Historically, large military installations represent high economic multiplier effects in the local economy. According to the Mississippi River Regional Planning Commission (2013), the estimated economic impact of Fort McCoy on the local economy in FY 2011 was $1.31 billion; in FY 2002, it was $357.8 million, which is 27.0% of the 2011 number (Appendix E). Appendix E from the Mississippi River Regional Planning Commission (2013) shows the economic impact from 2001-2010 that Fort McCoy has on the local communities.

According to Department of Defense Instruction 3030.03, it is DoD policy to work toward achieving compatibility between military installations and neighboring civilian communities by a joint compatible land use planning and control process conducted by the local community in cooperation with the local military installation.

The expansion and significant land cover increases can be attributed to the continued growth of the local community, the symbiotic relationship between Fort McCoy and local communities, and the ever-changing economic and global environment.

References

Anderson, A.B., Guertin, P.J., and Price,D.L. 1996. Land Condition Trend Analysis Data: Power Analysis (No. CERL-TR-97/05). CONSTRUCTION ENGINEERING RESEARCH LAB (ARMY) CHAMPAIGN IL.

Andersen, M.C., Thompson, B., and Boykin, K. 2004. Spatial Risk Assessment Across Large Landscapes with Varied Land Use: Lessons from a Conservation Assessment of Military Lands. Risk Analysis: An International Journal, 24(5), 1231–1242. https://doi-org.xxproxy.smumn.edu/10.1111/j.0272-4332.2004.00521.x.

Army Public Health Center (APHC). 2016. Fort McCoy Installation Compatible Use Zone Study. Retrieved October 2019 from https://phc. amedd.army.mil.

Copley, R., and Wagner, E. 2008. Improved Situational Awareness through GIS and RFID in Military Exercises. 2008-12-12]. http://proceedings.esri.com/library/userconf/proc06/papers/abstracts/a2350.html.

Cully, J.R., Jack, F., and Winter. S. 2000. Evaluation of Land Condition Trend Analysis for Birds on a Kansas Military Training Site. Environmental Management Vol. 25, No. 6, pp. 625–633.

Department of Defense Instruction 3030.30. Joint Land Use Study (JLUS) Program, July 2004 Retrieved December 2019 from https://www.esd.whs.mil/ Portals/54/Documents/DD/issuances/dodi/303003p.pdf?ver=2018-10-05-072003-127.

Fleming, S., Jordan, T., Madden, M., Usery, E.L., and Welch, R. 2009. GIS applications for military operations in coastal zones. ISPRS Journal of Photogrammetry and Remote Sensing, 64(2), 213-222.

Johnson, S., Wang, G., Howard, H., and Anderson, A.B. 2011. Identification of superfluous roads in terms of sustainable military land carrying capacity and environment. Journal of Terramechanics, 48(2), 97-104.

Lozar, R.C., Ehlschlaeger, C.R., and Cox, J. 2003. A Geographic Information Systems (GIS) and Imagery Approach to Historical Urban Growth Trends Around Military Installations (No. ERDC/CERL-TR-03-9). ENGINEER RESEARCH AND DEVELOPMENT CENTER CHAMPAIGN IL CONSTRUCTION ENGINEERING RESEARCH LAB.

Mendoza, G.A., Anderson, A.B., and Gertner, G.Z. 2002. Integrating multi-criteria analysis and GIS for land condition assessment: Part I–Evaluation and restoration of military training areas. Journal of Geographic Information and Decision Analysis, 6(1), 1-16.

Mendoza, G.A., Anderson, A.B., and Gertner, G.Z. 2002. Integrating multi-criteria analysis and GIS for land condition assessment: Part 2—Allocation of military training areas. Journal of Geographic Information and Decision Analysis, 6(1), 17-30.

Mississippi River Regional Planning Commission (MRRPC). 2013. Fort McCoy/Monroe County Joint Land Use Study. Retrieved December 1,2019 from http://www.mrrpc.com/Misc_pdfs/JLUS_final_draft_2.25.pdf.

Price, D.L., Anderson, A.B., Guertin, P.J., McLendon, T., and Childress, W.M. 1997. The US Army’s Land-Based Carrying Capacity (No. CERL-TN-97-142). CONSTRUCTION ENGINEERING RESEARCH LAB (ARMY) CHAMPAIGN IL.

Rozenstein, O., and Karnieli, A. 2011. Comparison of methods for land-use classification incorporating remote sensing and GIS inputs. Applied Geography, 31(2), 533-544.

Swann, D. 1999. Military applications of GIS. International Journal of Geographical Information Systems, 2(2), 889-899.

Warren, S.D., Diersing, V.E., Thompson, P.J., and Goran, W.D. 1989. An erosion-based land classification system for military installations. Environmental Management, 13(2), 251-257.Warren, S.D., and Bagley, C.F. 1992. SPOT imagery and GIS in support of military land management. Geocarto Internatio

About the Author

Christian Rodriguez is currently a graduate student at the Missouri University of Science and Technology. This research article/technical paper is a continuation of a land cover change project he worked on for his graduate program in Geological Engineering. It involved remote sensing, land cover change analysis, imagery classifications, and the implementation of  raster analysis tools.

Ad
Share
Published by

Recent Posts

Mapping Every Tree

The California Forest Observatory, created by Salo Sciences, has applied a new artificial intelligence tool…

September 22, 2020

Self-employment in GIS

Self-employment in the GIS industry may seem daunting, but a recent MapScaping podcast with Kurt Menke, founder…

September 21, 2020

Survey of GIS Professionals

A quick look at GIS professionals, their educational background, and the software tools and programming…

September 20, 2020

Using GIS to Improve Hurricane Evacuation and Preparedness

As hurricane season gets underway in the Atlantic, GIS is being use to better prepare…

September 16, 2020

Using GIS to Monitor the West Coast Wildfires

With fires raging in California, Oregon, and Washington, scientists, activists, geospatial companies, and others are…

September 13, 2020

Earthquake Detection Using Smartphones

Using aggregated accelerometer data across many smartphones, many of us can now better access a…

September 8, 2020