Vegetation, Transects, and PVC

The vegetation analysis is in full swing! We have collected all the data needed for the project and are now in the process of analyzing it using sample point. Which brings me to my point today:

One of these quadrat transect methods is not like the other!

One of the projects of the vegetation analysis is to compare two different quadrat transect methods. The first one involved a one meter square PVC quadrat being placed along a vegetation transect every 4 meters for 16 meters. In each quadrat we would count the number of individuals in each plant species and determine the average height of each plant species. This was all done by sight and it was very labor intensive. That aside, the data for the quadrats seemed to vary widely because of human error. Each time sight was the method used to determine the abundance and height. Jesse and I both estimated what we thought was the measurements for the quadrat and then averaged the numbers.

On the other side of the project, Jason from GTM and Sharylin quickly zipped through transects using the image analysis transect method. This procedure used a camera mounted above a stand to take a picture of a 1 meter square quadrat below. The method took two days to complete, mostly because of weather restraints since our camera was not water resistant. The measurements for the quadrats were made a program called sample point (www.samplepoint.org). Once the images were cropped to only show the quadrat sections sample point was used to select 64 random points/pixels. Each point determined to be a type of plant or not and then the data for abundance was put in a excel file to be analyzed with a statistical process. By the time of this writing, all of the transects are completely cropped and only one transect is left to be run through sample point.

All 17 transects were measured using the Image analysis method while only the first 3 transects were measured using the visual method. After doing three transects with the visual PVC method it was obvious that it was not nearly as accurate as the Image method. The transects have not been compared as of yet but we believe that the camera and sample point method will be much more accurate.