HOW WAS THE TICK THREAT MAP PRODUCED?
As a result of entering our collection and observation data into a GIS, we able to produce tick threat maps for areas that we did not have the time or resources to survey. The map below gives someone using an area, who probably has little knowledge of tick biology or vegetation types, an indication of the likelihood of contacting ticks in their area of operation. We believe that visuals such as this greatly enhance threat awareness, resulting in individuals exercising greater personal protection measures.
A three step process was undertaken to produce the threat map based on aerial photographs.
Step 1. An expert assigned a tick habitat suitability index at various sites throughout the study site.
Step 2. We validated the expert's rating by comparing it spatially to actual tick trap collections.
Step 3. We than looked for correlations between the expert's rating of tick habitat and infrared aerial photographs.
STEP 1. ASSIGNING THE TICK HABITAT SUITABILITY INDEX
An expert, who had been conducting tick surveys for decades, surveyed the study site and at each of the locations shown in red he assigned a Tick Habitat Suitability Index (THIS) of 0 (no expectation) to 10 (highest expectation). He based his assessment on vegetation, ground cover and leaf litter.
STEP 2. VALIDATING THE EXPERT'S TICK HABITAT SUITABILITY INDEX
The legend has three columns. The first is the probability that the colored area contributes to, in this case, 88% of the tick population. The second column, in parentheses, gives the percent of the total area in this probability category. The third column lists the square footage of each probability category.
First we used geostatistical analysis to produce a population probability contour (PPC) map. The PPC is produced by sorting trap counts in descending order and then constructing a cumulative frequency distribution. A frequency is chosen for analysis and those trap sites contributing to the selected population frequency threshold (e.g. 88%) get an indicator value of 1 and those that don’t contribute get an indicator value of 0. Krigging is then conducted on the indicator value to produce the PPC of the sample population, which is assumed to reflect the distribution of the entire population.
The results of the analysis are shown as the colored contoured areas on the map below. The exact explanation of this map is that there is a 0.5 or greater probability that contoured area contains all the habitat that had a THSI value of 6 or greater. The interpretation is that the colored areas are where our expert expected the majority of the ticks to be.
We then conducted the same probability analysis on the results of all ticks captured using dry ice at 103 georeferenced sites. The exact sites were surveyed each of the three years. The map below is the 88% probability contour, i.e., there is 0.5 or greater probability that the colored contoured areas account for 88% of the tick population.
We then used GIS software to subtract the tick data layer from the habitat suitability index layer. The resultant layer is shown below at the right. The grey areas (61.3% of the study site) are where there THSI index and the actual trap catches were in agreement. The light and dark blue areas (13.4% of the study site) are where the THSI Index predicted more ticks than there actually were. Rarely does an organism occupy all of it’s available habitat so we do not feel that this overestimation detracts from the validity of the THSI Index. The yellow and red areas (21.2% of the study site) are where the THSI Index underestimated the tick population. All in all, we feel that the tick data and this analysis has validated the THSI Index. We are currently conducting further analysis on soil type and slope which we feel will further improve the accuracy of the THSI Index.
STEP 3. CORRELATING THE EXPERT'S TICK HABITAT SUITABILITY INDEX WITH AERIAL PHOTOGRAPHY
Shown below is an IKONOS Satellite image, using the near infrared band, of Ft AP Hill. The advantage to using the near infrared band is that it is capable of distinguishing vegetation types (dark red is pine forest; brighter red is deciduous); greenish color is grass; turquoise color is asphalt and gravel; dark blue is water. Using this to look for tick habitat.
We conducted a backward stepwise regression to investigate the correlation between the THSI and the following variables: Slope, aspect, red band, green band, blue band, and the near infrared band (NIR). The following formula was derived as a best fit for the data: THSI AA = 7.727+(red)(-0.0249)+(nir)(-0.00378)
This formula was then used to classify the image into three categories, low (green), medium (yellow), and high (red) risk, as shown below.
Imagine Software was used to despeckle the image and produce the more usable/readable image shown below.
We then returned to AP Hill with the originator of the THSI and another tick expert who ground truthed the map and confirmed that it provided a good representation of high risk tick areas. It was this map that was used to produce the threat map for the jamboree. This map can also be loaded onto a PDA with GIS tracking software so that an individual can see when they are entering a high risk area. This map was also use to produce a guide for tick control.
Related Public Health Command Programs