by Mark MacAllister
June 21, 2002
Page 7 : Determining homerange: Kernel Method
Unlike Jennrich-Turner and MCP, the kernel method for determining homeranges allows us to look at homeranges based not only on locations but also on the number of times an animal uses a specific location. By focusing on the frequency of use, we can more closely approach the true definition of homerange, and in doing so can see more clearly the region that an animal patrols or otherwise depends upon as day-to-day habitat. A kernel homerange map depicts a homerange as a series of concentric, asymmetric, oval-like polygons. These polygons are multi-colored, with each color representing a probability. In the first map in the Media Gallery, the lightest polygon represents a 50% probability region; that is, 50% of the animal's possible locations are notated within the borders of the white area on the map. The region bordered by the aqua polygon signifies 75% probability (note that this would include locations in the white polygon as well). Finally, the green polygons represent 95% probability, and also account for locations in the white and aqua polygons. As we approach the center of the overall homerange map—the kernel of the map—the location points are of course more dense. Put another way, the probability of finding an animal increases as we move toward the center of the kernel homerange on the map.
The first map in the Media Gallery is a kernel homerange map based on Desiré locations with LC values between "3" and "B." What conclusions can you reach using this map that were unavailable to you when using either J-T or MCP methods? How would you contrast the way outliers are handled in J-T, MCP, and kernel homerange maps?
Following that is a kernel homerange map for Desiré data with LC values ranging from "3" to "0." Again, how would you compare this to equivalent maps using MCP and J-T methods? What does this map tell you about Desiré's travels in the southern part of the overall region? If you were to go to Cameroon to look for Desiré, where would you begin your search? Would you attempt to make that judgement using the J-T or MCP maps you saw previously?
While maps like these help researchers understand elephant locations and homeranges, it is also important to put those homeranges into a larger context. For example, does a particular elephant's homerange lie close to populated areas? How far is an elephant's homerange from a river or perennial stream? How frequently does an elephant range across the Cameroon border and into a country where elephant protection is poor? Let's look at one more map to see how these analyses are important.
Next Page : The homerange in context
Pages: 1, 2, 3, 4, 5, 6, 7, 8
View printer-friendly version
