Blog Post 3: Analyzing Data

This is an exciting time! Finally, after a long summer field season, we are able to start analyzing our hard won data! I hope to have some nice clean charts in my final blog post, but first I will simply describe the methods I have been using to analyze our various data sets.

To analyze temperature, humidity, and soil moisture as it relates to movement, I will be running a Pearson Correlation. This test will allow me to determine if any of these three variables have an impact on average daily movement. I plan to look at trends for each individual species instead of each genus, as I looked at species with widely differing life history characteristics and so likely respond differently to each variable. Within each species, I will compile the daily movements for all the frogs we tracked, and compare them with the temperature/humidity/soil moisture readings of that day. Because anurans have very permeable skin, dehydration and dessication are serious threats to their survival. Because of this, I predict that they larger movements will be associated with more humid days and more rainy (moist soil) conditions. Anurans are ectotherms, so I do not believe that high temperature will limit their movement unless it is also a very dry day. However, unusually cool days may serve to limit their movement, as it will slow their metabolism.

I have also been keeping track of the “substrate” that we find each individual anuran on every day. I have defined substrate as the type of cover a frog is using, such as logs, leaf litter, burrows, or bare ground. For each species, I plan to calculate percentages each substrate, to determine what type of cover they are using most often and look for differences between species.

In the long term, I also plan to look at vegetation, canopy, and understory preferences for each species, but analyzing this data will be slightly more complicated and will take a few months. It will take so long as a result of the methods we used to assess these variables – we took pictures of vegetation, canopy, and understory cover at each location where we found a frog, and I will need to go through every picture and classify them individually before we can begin calculating percent coverage for any of these factors. For vegetation, we took a picture of a 1m x 1m quadrat at the location where we located a frog. To determine the percentage of vegetation cover, I will compare the area covered by plants to the area of the whole quad. Similarly, for canopy, we took a picture from 1m above the ground at each location where we found a frog, with the camera facing straight up. After loading the picture into GIS, I can separate out individual pixel cells by color, and thus separate the blue “sky” from the green/brown “canopy.” By comparing cell counts for each color, I can then determine the percentage of canopy cover. For understory, we took four pictures of a “cover pole” from each of the four cardinal directions at each frog location. As with canopy, I will import these pictures into GIS and compare the number of cells that make up the entire pole to the number of cells on the pole that are masked by undergrowth or leaves. After calculating the percentages for these three variables, I can then look for differences in habitat preferences across species and eventually across seasons to see if habitat type changes in the breeding and nonbreeding periods.

I’m very excited to see how much of our data is significant – tentative results are already looking good! I can’t wait to share them with you in my final post!