Project in Fall 2010 by Jennifer Smith
GIS Analysis of Grey Wolf Habitats in Michigan
Current and potential wolf habitats were examined using five criteria: land coverage, road density, population density, farm density, and prey (deer) density. The series of maps shows areas based on these criteria, which can be considered favorable grey wolf habitats.
Starting with European settlement of North America, wolf populations faced endangerment and extinction in many parts of the United States. Legal hunting of these animals ensued and their numbers dwindled. During the later part of the 20th Century wolves became protected under the Endangered Species Protection Act in 1967 (Roell, Brian et al. 2009). Since their classification as endangered, wolves have been increasing their numbers, but are only found in Alaska, Idaho, Michigan, Minnesota, Montana, Wyoming , Wisconsin. Grey wolf populations in Michigan and the Great Lake area have been increasing, which is the area of study for this GIS analysis. At this time, wolves have reached population numbers high enough to be taken unlisted as threatened; however, due to bureaucratic reasons, wolves are still listed on the Federal Endangered Species List (Roell et al. 2009).
According to Wydeven et al. (1999) wolf territories range from 20 to 120 sq. miles and Average pack size is 4.1 wolves. Research of favorable wolf territories in the Great Lakes area focuses on either habitat, including road density, population density and land use, or prey density, deer populations; so either a “Habitat Area Model” or a “Prey Based Model” can be used (Wydeven et al. 1999). Expanding wolf populations and ranges calls for analysis of suitable habitats in order to predict trends and educate the public about coexisting with these predators. For this reason, Wolf Management Plans have been developed and employed in Michigan and the Great Lake area (see Roell et al. 2009). Mladenoff et al. (1999) state that growing predator populations, especially of predators that are “not highly habitat specific,” necessitates analyzing population growth potential and habitat range.
The purpose of examining wolf populations in the Great Lake area, specifically Michigan and Wisconsin, is to determine which areas wolf populations will expand into in future years. By using specific criteria, such as human density, land use, road density, and prey population density (deer), we can attempt to predict where wolf packs may go in the future. The current range of wolf packs is known to a degree based on telemetry data; in order to determine how growing populations will interfere with human populations we must evaluate the conditions of existing wolf ranges. The two most important factors for determining what makes an area favorable are land use and deer density.
Some may wonder why we should bother examining such data and why we should care about wolf territories. Analyzing possible habitats is important because wolves are a vital part of the ecosystem; specifically, wolf populations keep the deer population in check. Without the balance of predator and prey species, deer populations increase to unmanageable numbers. Thus, it is beneficial for human populations coexisting with wolf populations to allow their numbers to increase. Estimating which parts of the Great Lake area wolves may occupy in the future can help us inform the public now of possible disturbance by wolf packs. Educating the public is the best way to prevent unnecessary killing of wolves and also wolf predation on domestic animals. Additionally, habitat analysis can help biologists determine population recovery potentials; and this in turn can help manage “animals to minimize human conflict” (Mladenoff et al. 1999, 37)
A series of analysis was completed using ArcGIS; each step focused on a specific criterion.
First, overall land coverage of the Upper Peninsula of Michigan was examined. The land coverage map was created using a shapefile from the Michigan Geographic Data Library. The shapefile classifies land in the Upper Peninsula into numerous categories; the area is separated into categories based on type of forest, bodies of water, and various urban categories. In order to make the map easier to interpret, I changed some categories that could be grouped together to the same symbol color. Various categories of coniferous forest were changed to the same symbol, and the same was done for all urban areas, for example. This map shows that the majority of the Upper Peninsula remains forest land but a significant portion is farm or agricultural land (symbolized as red). Also, all black areas refer to urban centers, airports, roads and human recreational space. The results are figure 1 in the results section. Clearly, the majority of the map is displayed in green, thus, most of the area is suitable wolf habitat. The next step was to determine suitable habitat based on more specific criterion.
Road density is one of the most important factors; analysis was done by finding the mileage of highway per square mile of land. According to previous research the best density is less than 0.7 miles per square mile. Land with more than 1 mile of road per square mile of land is least suitable and unlikely to become wolf territory (Wydeven et al. 1999). The numbers were calculated using Topographic Map Quads accessed through ESRI; the area of each Topo Quad averages at 53 square miles. A major highway layer (ESRI) was added and clipped to Michigan state boundaries; then the distance of each highway segment was calculated in miles. The sum highway mileage for each Topo Quad was calculated and joined to the Topo Quad attribute table. Next, the road density for each Topo Quad was found using the field calculator; the calculation was: sum distance in miles divided by total square miles for each Topo Quad. Numerous Topo Quads had zero mileage of highway, thus they are the most suitable for wolf territory. The lightest green on the map (figure 2) indicates the most suitable territory. Topo Quads with less than 0.1 miles of road per square mile are also favorable habitat and road densities up to 0.7 miles of road per square mile should be considered suitable habitat, as well. Road density values above this should not be considered likely wolf pack territory. Since wolf territories range from 20 to 120 sq. miles, ideally areas with that much space between major highways could be inhabited by wolves. The percent area as favorable habitat was calculated using the area of selected TopoQuads (less than 0.1 miles of road per square mile area) divided by the total area of Michigan; the value for the area of Michigan came from the Topo Quad data. The results showed that, solely based on road density, 29.8% of Michigan can be considered favorable wolf habitat.
Wydeven et al. (1999) suggest favorable habitats have less than 4.0 people per square mile. General population density was examined using Census Block Groups from ESRI. The results are as expected; denser populations are found in the Lower Peninsula, clustered around urban centers. The Upper Peninsula has low population density for the most part, however, only block groups with less than 5 people per square mile were chosen as favorable habitat. Block groups with higher population densities do contain wolf populations, thus these are the areas likely to face the most human-predator conflict. The percent favorable habitat was also calculated using population density; the total area of selected block groups was divided by the total area of Michigan based on block group data. The results were 14.6% as favorable habitat, which is substantially lower than the estimate based on road density. The Population Density map is shown with Road Density in Figure 2.
Analysis of farm density shows that the entire Upper Peninsula is suitable territory since all counties have less than one farm per square mile; some of the Upper Peninsula counties have zero farms per square mile. Northern Wisconsin shows the same, but the southern portions of both states have higher farm density; the counties with over one farm per square mile would likely prove to be problematic for wolf packs. It is not that wolves could not live in close proximity to agricultural area; it is just that higher the farm density makes an area less suitable. The higher the farm density, more likely wolves would prey on domestic animals and killing of wolves would occur. Therefore, knowing where farm density increases will help biologists inform the public about ways to coexist with wolf populations.
The criterion of farms per square mile helps determine the best habitats; however, it is a very general category considering each county is hundreds or thousands of square miles. I found it better to examine the total number of farms in each county and select counties with less than 100 farms as favorable territory. Also, these counties were selected with the additional criterion of population density equal to less than 50 people per square mile; the population density for this selection is different from above because in this case county boundaries are used, not census block groups. Selecting counties with both of these attributes resulted in 16 counties in Michigan and Wisconsin. Clearly, wolves will not colonize according to county boundaries but it is the easiest way to break apart the entire area.
Figure 3 in the results section shows a series of three maps: the first shows farm density based on square mileage for each county, the second shows the number of farms per county, and the last shows selected counties based on my criteria. The criteria used for the last map is: less than 100 farms and less than 50 people per square mile for each county. Each selected county is over 120 sq. miles; therefore, they fit within the average size for wolf ranges.
The next image in the results section (figure 4) shows that my selected counties generally match up with the known extent of wolf packs. I matched the selected counties from my attribute analysis with a shape file of known wolf pack territories. Each selected county has a wolf population, thus farm density and population density are very important factors. The wolf ranges shown do encounter counties with higher farm and population densities; these counties may find larger human-predator conflict and are not ideal territories. However, predators such as wolves cannot be contained to a specific area. Therefore, instead of attempting to control the public should be educating about wolves and how to successfully coexist while minimizing conflict. The last resort in these situations would be to kill problematic wolves, meaning those venturing into urban or highly populated areas; this solution should be avoided if at all possible.
The last piece of analysis is determining deer density; the results are shown in Figure 5. Wolves will follow prey, which in this case is mostly deer. Knowing where the highest and lowest prey density is can also help determine where wolves may resort to other food sources; specifically the areas in which wolves may prey on domestic animals. Data for deer density per county was found at Quality Deer Management Association; I used the deer densities given and attributed a value (1 through 4) to each county. Class one is less than 15 deer per sq. mile, class two is 15 to 30 deer per sq. mile, class 3 is 30 to 45 deer per sq. mile and class 4 is over 45 deer per sq. mile. Each class was entered individually into the new field “Deer Density” in the attribute table for counties; the county layer was accessed through ESRI. The map shows that the Upper Peninsula does not have the highest deer density although, it is still favorable habitat; Wydeven et al. (1999) assessed that the average wolf habitat has 22 deer per sq. mile. This means class one and class two are suitable habitats. The center of the Lower Peninsula has the greatest deer density but this area also has higher population and farm densities; therefore, it is not the most favorable habitat despite its abundance of deer.
Overall, it may seem increasing wolf populations would become problematic for farmers and urban populations; overabundance of deer populations generally causes more harm. Without natural predators, specifically wolves and mountain lions, deer populations can only be controlled by human efforts. Thus, allowing wolves to return to their natural habitats and increase numbers will benefit the ecosystem. The habitat analysis I have done and analysis done by others shows that certain characteristics lend themselves to land being favorable habitat, however, wolves will not restrict themselves to these areas. Knowing that wolf packs may often be located in highly populated areas will help biologists manage the populations.
Clearly, the wolf populations in Michigan, and the Great Lake area, are expanding and it seems inhabit the majority of favorable land in the Upper Peninsula. The Lower Peninsula on the other hand does not show much potential for wolf habitats. I recommend using this type of analysis to examine other northern states; the criteria discussed for favorable wolf habitats in Michigan can be used to examine areas east and west of the Great Lakes area. Other states and regions may present possibilities for Grey wolf population growth and range expansion.
Mladenoff, David J., Sickley, Theodore A., and Wydeven, Adrian P. 1999. “Predicting Grey Wolf Landscape Recolonization: Logistic Regression Models vs. New Field Data. Ecological Operations 9(1): 37-44.
Roell, Brian, Beyer, Dean E. Jr., Lederle, Patrick E., Lonsway, Donald H., and Sitar, Kristie L. 2009. Michigan Wolf Management 2009 Report.
Wydeven, Adrian P., Mladenoff, David J., Sickley, Theodore A., and Haight, Robert G. 1999. Appendix C: GIS Evaluation of Wolf Habitat and Potential Populations in The Great Lakes States. Online Resource, http://www.timberwolfinformation.org/info/wolfmanplan/final/appendix/appendix_c.htm Accessed December 5, 2010.
Wolf Pack image from http://fangedwolf.tripod.com/Wolf.htm. Accessed December 13, 2010.