Hi,
I'm conducting an I-Tree eco evaluation of a smaller neighborhood in Baton Rouge Lousiana.
I'm trying to figure out the number and size of plots that would be ideal.
I'd like to achieve the same standard of error found in 200 plot, 1/10th acre samplings of larger cities.
I have 5 or so strata. Some of these strata are only an acre or so in size.
Is there a formula that can tell me the standard of error for sampling a given #/size of plots from a given size of stratum?
I'd like to be able to alter the # or size of my plots and see the effect it has on my standard of error.
Many thanks,
-John
Sampling Smaller Areas
Moderators: i-Tree Support, i-Tree Team
Hello,
Unfortunately, there is no easy way to predict or calculate the standard error as unknown variables such as the tree species composition and distribution within your study are unknown. Project managers can certainly utilize smaller plots or less plots when working with smaller areas of interest. However, the distribution of vegetation within the study area may still cause higher variance if unevenly distributed.
There is another Forum post which discusses plot options and also has a link to the article where the Eco guidelines were derived from. http://forums.itreetools.org/viewtopic.php?t=49 The journal article has examples of the number of plots used by some cities and the resultant standard error achieved.
We are also seeing more Eco project managers submit their data for analysis and then evaluate if they need to sample more plots within a given strata to increase the accuracy of the estimates. This has become a more viable option especially since processing of Eco data is automates for users in the U.S and results are returned generally in a few hours.
You mentioned that some of the strata in your area of interest are only about an acre which may be small for sampling with plots. You may want to consider options like combining smaller strata if possible. Some people may consider using the Eco complete inventory option in an acre area although the results are more limited than sample based.
Hope this will give you some ideas to consider.
Al
Unfortunately, there is no easy way to predict or calculate the standard error as unknown variables such as the tree species composition and distribution within your study are unknown. Project managers can certainly utilize smaller plots or less plots when working with smaller areas of interest. However, the distribution of vegetation within the study area may still cause higher variance if unevenly distributed.
There is another Forum post which discusses plot options and also has a link to the article where the Eco guidelines were derived from. http://forums.itreetools.org/viewtopic.php?t=49 The journal article has examples of the number of plots used by some cities and the resultant standard error achieved.
We are also seeing more Eco project managers submit their data for analysis and then evaluate if they need to sample more plots within a given strata to increase the accuracy of the estimates. This has become a more viable option especially since processing of Eco data is automates for users in the U.S and results are returned generally in a few hours.
You mentioned that some of the strata in your area of interest are only about an acre which may be small for sampling with plots. You may want to consider options like combining smaller strata if possible. Some people may consider using the Eco complete inventory option in an acre area although the results are more limited than sample based.
Hope this will give you some ideas to consider.
Al
A member of the i-Tree Team
Thanks for the reply.
I want to double check that I understand this correctly.
The Itree model calculates standard error based on the number of plots sampled and variance between those plots. Regardless of how large the study area is.
So whether I'm studying a university campus, or the city of Chicago, if I want a low standard of error, I need to sample 150+ plots and make those plots larger to decrease the variance between them.
I want to double check that I understand this correctly.
The Itree model calculates standard error based on the number of plots sampled and variance between those plots. Regardless of how large the study area is.
So whether I'm studying a university campus, or the city of Chicago, if I want a low standard of error, I need to sample 150+ plots and make those plots larger to decrease the variance between them.
Eco plot information
Generally, increasing plot numbers or increasing the size of plots will produce more accurate estimates regardless of the size of the study area. The recommendations on page 8 & 9 of 200 plots for a city is the same for small cities or large cities. Ultimately, individual managers will have to decide what is suitable for their individual situation.
Users working with smaller parcels like campuses generally will use less plots than a citywide sample or they may consider reducing the plot size from the typical 1/10 acre in conjunction with a high quantity of plots.
Users working with smaller parcels like campuses generally will use less plots than a citywide sample or they may consider reducing the plot size from the typical 1/10 acre in conjunction with a high quantity of plots.
A member of the i-Tree Team