Hello,
I'm interested in using UFORE for a county park that I am currently interning at. I know the size of the park (181 acres), but unfortunately I don't have the time or resources to collect data on 200 plots throughout the park.
I'm wondering if anyone has any suggestions as to how to scale down my study area and the plots within it and still have UFORE be a useful tool? I think that having the type of information that UFORE can produce for even a subset of the park will be useful and will hopefully spark an interest from park officials to do a broader scale study.
Any tips?
Thanks,
Amanda
UFORE in a County Park
Moderators: i-Tree Support, i-Tree Team
Nice question! I have wondered about this a lot myself. I think much depends on what you want/need to get out of the project.
In general, as I understand it, there is no particular reason to do 200 plots, especially if 1) the variation from plot to plot of what you are interested in is relatively low, and/or 2) you don't have a need for highly accurate results. I once saw a graph created by Dave Nowak that showed that the true mean of tree count, for example. is contained by the results from relatively small sample sizes, although the error is large.
In particular, though, plots in parks can have extremely high variation when--as I often see--there are some areas with large open spaces, and others with a completely closed canopy. If a small number plots are laid on such a configuration, you might not actually have very many trees in your sample since a particular random plot selection could fall mostly in ballfields, entertainment venues, etc. One way around this that occurred to me--and I will be interested to see how the i-Tree Team reacts to this--might be to define multiple study subdomains within your park.
Depending on your specific configuration, you could have something like:
I hope this might give you some useful ideas.
In general, as I understand it, there is no particular reason to do 200 plots, especially if 1) the variation from plot to plot of what you are interested in is relatively low, and/or 2) you don't have a need for highly accurate results. I once saw a graph created by Dave Nowak that showed that the true mean of tree count, for example. is contained by the results from relatively small sample sizes, although the error is large.
In particular, though, plots in parks can have extremely high variation when--as I often see--there are some areas with large open spaces, and others with a completely closed canopy. If a small number plots are laid on such a configuration, you might not actually have very many trees in your sample since a particular random plot selection could fall mostly in ballfields, entertainment venues, etc. One way around this that occurred to me--and I will be interested to see how the i-Tree Team reacts to this--might be to define multiple study subdomains within your park.
Depending on your specific configuration, you could have something like:
- a closed-canopy domain
- an open-canopy domain
- a domain without canopy
I hope this might give you some useful ideas.
Last edited by Jerry on Fri Feb 12, 2010 11:39 am, edited 1 time in total.
Eco/UFORE in a County Park
Yes, I agree with Jerry that this is a good question and there seems to be more users interested in adapting the model to smaller scale projects such as parks and school campuses. As Jerry pointed out, there can be some variation of vegetation in some of these instances depending on the different elements and decreasing the amount of plots may decrease the amount of trees in your sample. Therefore, we are also suggesting that people consider using a higher quantity of smaller plots instead of the usual 1/10 acre.
I have attached a map of a 35 acre school property with 20 random plots selected from a grid overlay. I believe that (4) plots have trees in them which may be representative of the overall property which has much open space, fields and buildings. The user was satisfied with this plot design for their specific goals as every situation is somewhat unique. I also attached an extracted graph showing the relationship between plot qty and variance from a UFORE FAQ document. Hopefully this will give you some more ideas as you consider your options.
Al
I have attached a map of a 35 acre school property with 20 random plots selected from a grid overlay. I believe that (4) plots have trees in them which may be representative of the overall property which has much open space, fields and buildings. The user was satisfied with this plot design for their specific goals as every situation is somewhat unique. I also attached an extracted graph showing the relationship between plot qty and variance from a UFORE FAQ document. Hopefully this will give you some more ideas as you consider your options.
Al
You do not have the required permissions to view the files attached to this post.
A member of the i-Tree Team
Thanks!
The park I'm working with is relatively forested with few open areas, so I don't think there would be much variation.
I've been trying to figure out why the i-tree manual suggests 200 plots in general even though the size of the survey area can vary greatly depending on the project. My best guess is that it is because itree takes into account so many different parameters?
I would be interested in how the i-tree team came up with this number so that I could possibly tailor it more towards my project. If anyone has any explanations, I would be happy to hear them!
~Amanda
The park I'm working with is relatively forested with few open areas, so I don't think there would be much variation.
I've been trying to figure out why the i-tree manual suggests 200 plots in general even though the size of the survey area can vary greatly depending on the project. My best guess is that it is because itree takes into account so many different parameters?
I would be interested in how the i-tree team came up with this number so that I could possibly tailor it more towards my project. If anyone has any explanations, I would be happy to hear them!
~Amanda
It has to do with statistics and the Central Limit Theorum which basically states that the more samples you have, the lower your statistical error, and the more precise your results become. I think that 200 plots in a 181 acre park is a bit much (1 plot represents less than an acre). You may want to ask yourself how many plots you can get done in a certain timeframe and then go with that. After receiving your results, you can take a look at the statistical error. If that error is too high for your liking, then put in more plots the next season until you get the error below your comfort level. The iTreetools.org website briefly explains this under Resources in the faq_UFORE document. Here is a link to a publication done in 2008 by the UFORE developers that explains in more detail.
http://www.urbanforestrysouth.org/resou ... re%20plots
Hope this helps.
http://www.urbanforestrysouth.org/resou ... re%20plots
Hope this helps.