I am not a botanist but need to estimate the average percent vegetation cover of my study area. I have researched this and found articles that use the "releve" method for determining plot size and sample size. It seems very subjective but reasonable if you are very familiar with your study site. Is this still standard protocol among botanists?
Thanks
Yes, it's just a random quadrat method, so it will work :)
I have developed my own method for my fauna studies. There are a few apps (Canopy Cover, Canopeo, CanopyApp, HabitAp) out there that can be used to determine canopy cover using your phone/tablet.
I carry out point descriptions of habitats and the process is relatively simple. The point description is based on the assessment of structural characteristics within 20m-30m of the sample point. Canopy cover is determined by averaging the cover, as determined by using the apps, from 6 points within the 20m-30m radius.
I standardise the height at which the cover is determined by using a hiking pole to rest the phone while taking the photo (I use HabitApp). I take one measure at the sample point and then walk in random directions (approx 10m) before taking another measurement. The six measures are averaged and entered into the data entry sheet/app.
I hope that makes sense. The method adds some tedium to the assessment but it standardises a measure that is important in assessing fauna habitat characteristics for the groups I study.
I don't know about botanists, but in Remote Sensing one has different methods to measure fractional cover in the field. Several methods are described in the following document:
Conference Paper Practical Seminar II: Hemispherical Digital Photography (HDP...
However, 'Hemispherical Digital Photography' or in short HDP, is a technique which is applied very often and which is quite accurate. You need a camera with 'fish eye' objective as well software to process your data. You will have to buy this camera of which different types (mostly Nikon's) are commercially available. Just Google them. The sampling strategy is explained on the conference paper. The software to process your photographs is freely available from Marie Weiss in France. She developed very good soft, to extract different bio-geophysical variables from your hemispherical photographs, whereunder LAI and fractional vegetation cover. Marie Weiss currently works at the "Environnement Méditerranéen et Modélisation des Agro-Hydrosystèmes (EMMAH)", French National Institute for Agricultural Research. Marie does research in Environmental Science and Remote Sensing.
Download the software which is named:'Can Eye' from:
https://www6.paca.inrae.fr/can-eye
Have a look at fish eye objectives and camera's (Nikon) at:
https://www.google.be/search?client=firefox-b&ei=mphxX5u9EubjkgWVvo2ABg&q=Nikon+Fish+Eye+camera&oq=Nikon+Fish+Eye+camera&gs_lcp=CgZwc3ktYWIQAzIGCAAQDRAeMggIABAWEAoQHjIICAAQCBANEB4yCAgAEAgQDRAeOgQIABBHOgUIABCRAjoECAAQQzoECC4QQzoGCAAQBxAeOggIABAHEAoQHjoGCAAQChBDOgQIABANOgYIABANEAo6BAgAEAo6BggAEBYQHlDeE1izkAFguJIBaAJwAXgAgAGOAYgB1BCSAQQxNS43mAEAoAEBqgEHZ3dzLXdpesgBA8ABAQ&sclient=psy-ab&ved=0ahUKEwib8qG2sYvsAhXmsaQKHRVfA2AQ4dUDCAw&uact=5
Success and best regards,
Frank
Thank you all for the comments. Maybe I can be more specific with my question, since I can't seem to decide the best strategy for my project. I have 3 large areas (14 km^2, 21 km^2, and 91 km^2). I'd like to estimate % cover of vascular plants in each. There is a paper that I am modeling my design after, because I'd like to be able to compare my results with theirs (they are on one Hawaiian island, I'm on another). Their study area is an entire national park on Maui (134 km^2), and study design is the following: 4 transects each at a different elevation range. Each transect has 4 ,10m x 10m plots that are spaced equally on the linear transect. They fail to mention the length, orientation, or placement criteria for the transects. I would not consider my sampling areas homogenous, so I hesitate to only use 1 linear transect with 4 relatively small plots for each area. It seems like randomly placing the plots would provide better representation of the area. And it seems like many small plots (1m^2 plots) would represent the areas better than few, larger plots (100m^2 plots). Person-hours are also a limiting factor, so that must be considered. Any ideas appreciated!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size