Measuring basal area in mixed vs variable plot

elduderino260 · July 30, 2014, 01:49:40 AM

Sorry to bug y'all again, but my new question (based on my previous post) has boiled down to this:

I am performing an ecological restoration experiment in the tropics, assessing 4 different restoration strategies. I have five 20x30m replicate plots for each treatment. In each plot, I will sample various soil parameters at 4 randomly selected locations. I plan to measure percent ground and canopy cover at each of the 4 locations to stay consistent. I also wanted to measure stem density, basal area, and vegetation species richness. There are several issues that have arisen as I am learning about the logistics of sampling in the plots.

Because of the density of vegetation and size of the plots, measuring and recording each stem is clearly impossible. I initially considered recording stem DBH (for individuals >=5cm DBH) in four 2m radius plots centered on the aforementioned subsample locations (and extrapolate out to the whole plot; essentially a fixed plot survey). However, with this method, either the subplots were too small and I missed some huge trees which would have completely changed the basal area of the plot or they were too big to be manageable both physically (getting to each tree) and time-wise (this is only one of many metrics we are recording); especially with dense understory.

I could, however, see the boles of the trees, so I thought that I could use a wedge prism from the plot center to get a more accurate BA measurement for the plot.

However, I am wondering whether that is valid (essentially using a variable plot method on a fixed experimental plot)...I suppose the main problem with this is that I might include individuals that are outside of the experimental plots in the sample. I suppose I could limit this by demarcating the boundaries of the plot and not count anything outside that (it may not be necessary due to low visibility in the dense veg), but my doubts remain.

Thanks for your patience and help.

clearcut · July 30, 2014, 12:06:09 PM

I do not see a problem with using both fixed plots and variable radius plots together. We often use different plot sizes and techniques from the same plot center to estimate specific variables. You are already using subplots for soils and and cover estimates. Variable radius plots simply represent a subplot for each tree diameter.

Nested plot designs using a larger fixed radius or variable radius plot for trees, a smaller fixed radius plot for shrubs, and an even smaller fixed radius plot for ground cover are common in research. Transects from the plot center to estimate canopy cover and fuel loading are also frequently used. You are simply matching the sampling technique to the variable being measured.

If you estimate basal area using a prism or angle gauge, you will need to count every "in" tree even if it is outside of the 20X30m replicate plot. Each diameter tree, has its own plot size - that is the variable radius. The technique assumes that you are counting all of the "in" trees in a circular sweep. You do not have to measure all of those trees. The circular sweep gives the estimate of Basal Area - (Count*BAF = BA in m^2/ha) of the stand at that point. You will have to measure the diameter and distance of borderline trees and trees that are difficult to see from the plot center.

elduderino260 · July 30, 2014, 03:17:47 PM

Quote from: clearcut on July 30, 2014, 12:06:09 PM
I do not see a problem with using both fixed plots and variable radius plots together. We often use different plot sizes and techniques from the same plot center to estimate specific variables.

Good to know.

Quote from: clearcut on July 30, 2014, 12:06:09 PM
If you estimate basal area using a prism or angle gauge, you will need to count every "in" tree even if it is outside of the 20X30m replicate plot. Each diameter tree, has its own plot size - that is the variable radius. The technique assumes that you are counting all of the "in" trees in a circular sweep. You do not have to measure all of those trees. The circular sweep gives the estimate of Basal Area - (Count*BAF = BA in m^2/ha) of the stand at that point. You will have to measure the diameter and distance of borderline trees and trees that are difficult to see from the plot center.

Perhaps I am being stupid here, but what do you mean that each diameter tree has its own plot size? I think you mean that the larger the DBH, the further away it can be and still considered "in", but how will that affect the calculation of BA? I thought all I needed to do is count the trees that are in according to the prism, measure the DBH for borderline trees, multiply that by the plot radius factor, and count that individual if the product is greater than the distance to the observer; and then multiply the number of in trees by the BAF to get the m^2/ha BA, which I can than use to calculate the avg BA for my plot...

Clark · July 30, 2014, 04:05:35 PM

elduderino260 - You've got the idea of how to use a BAF properly, don't worry about the "different size plot for each diameter tree". You know and understand enough that one day it will come to you but it isn't important to understand that particular statement now.

Using fixed and variable radius plots is quite common. Often a small fixed plot will measure the regeneration present, typically trees under 5" DBH down to 1' tall. The variable radius plot measures the trees larger than 5" DBH.

If the 20x30 plots are just a sample of the replication then counting trees that are outside the plot but inside the replication shouldn't be an issue. However, if the replications are 20x30 in size then counting trees outside that area would be "incorrect". Now, there is a way to mitigate the "problem" of counting trees that are outside the designated area. It's called a "phantom plot" and it involves more work in the field.

Clark

elduderino260 · July 30, 2014, 04:12:28 PM

Quote from: Clark on July 30, 2014, 04:05:35 PM
elduderino260 - You've got the idea of how to use a BAF properly, don't worry about the "different size plot for each diameter tree". You know and understand enough that one day it will come to you but it isn't important to understand that particular statement now.

Good! Thanks.

Quote from: Clark on July 30, 2014, 04:05:35 PM
If the 20x30 plots are just a sample of the replication then counting trees that are outside the plot but inside the replication shouldn't be an issue. However, if the replications are 20x30 in size then counting trees outside that area would be "incorrect". Now, there is a way to mitigate the "problem" of counting trees that are outside the designated area. It's called a "phantom plot" and it involves more work in the field.

Unfortunately, I believe my case is the latter. We have five 20x30 m plots for each of our four treatments (so 20 20x30 m plots). The samples themselves are all taken within that area. Could you provide an explanation or link to such, as a cursory google search revealed nothing...

All this being said, the visibility is pretty short, so we might not even have that problem as we can clearly see which trees are within the confines of the plot...