Raster Data – Site Selection
Using fuzzy logic to model Bald Eagle habitats
There are several endangered and threatened species, including Bald Eagle near Big Bear Lake that the national forest service wants to protect. The national forest service wants to model the natural habitat of Bald Eagle and has provided certain criteria where bald eagles preferred to live. These include living away from human disturbances, such as roads, power lines, cities; living close to water; and in areas with 40 percent and 70 percent tree cover. To address this problem, I will be using fuzzy logic as this method accounts for uncertainties.
The MXD file contains the Fuzzy membership inputs group layer that contains distance to human disturbance, distance to water, and reclassified land cover raster layers. These layers were created using Euclidean distance and reclassification tools.
First, I ran the Fuzzy Membership tool on the Water layer. I chose membership type Small so that a larger distance to water (undesirable) will be assigned a value closer to 0 and smaller distances to water (desirable) will be assigned value closer to 1 (desirable). The closer the distance to water, the higher the fuzzy membership. The other values were 1500 for midpoint and 5 for spread. The tool created a new raster layer based on a 0 to 1 scale. I imported the symbology and assigned the layer transparency of 60%.
I repeated the procedure on the Distance to Human Disturbance layer. However, instead of assigning the membership type Small, I assigned membership type Linear. The bald eagles nest away from human disturbances. Therefore, the ideal location would be far away from humans with the membership value closer to 1 (full membership).
Lastly, I used the Fuzzy Membership tool to create a fuzzy membership layer for the land cover variable. I used membership type Near with Midpoint value of 5 and Spread of 0.1. This membership type allowed me to select areas with tree cover between 40% and 70%.
Once I created all three layers, I used Fuzzy Overlay tool to combine these layers.
Fuzzy membership type for the Human Disturbance layer (click on the image to enlarge)
The green areas are farther away from humans, values closer to 1 (full membership), whereas the red areas are closer to humans, value closer to 0.
Application and reflection.
Fuzzy logic analysis has applications in studying disease outbreaks. Most infectious disease outbreaks do not follow a predictive pattern. Early identification is key to prevention of the spread of the outbreak Dengue is an acute viral illness and is transmitted between humans by mosquitoes. It is the most common neuroinvasive arboviral disease of humans in the world. Dengue mostly affects older children and adults and may result in dengue hemorrhagic fever, a severe form of dengue infection. In North Carolina, clinicians by law are required to report all suspected or confirmed cases of dengue to the local health departments. There is no vaccine available to prevent dengue fever.
Problem: The NC public health department wants to develop a system for accurately predicting future dengue outbreaks in the state using spatial data.
Data needed: The NC public health department has a listing of all cases occurred in the state and includes the demographic and year, week, and County information. Data on other risk factors such as rainfall, temperature, altitude, vegetation, social, and demographic profile will be obtained from several sources such as NASA, USGS, NOAA, and Census Bureau.
Analysis: Once the data are imported into ArcGIS 10.6.1, ran the Fuzzy Membership tool on the layers and assign fuzzy membership values between 0 and 1. I will then I used Fuzzy Overlay tool to combine these layers.