Bob, I know you were trying to be reasonable in your assessment, but that was not how it reads on paper. What it comes across as is this is the number from the tape drop, and if I throw out these high numbers for a reason I can rationalize - then my results will pretty closely match the tape drop numbers. If this were actually true the numbers generated would be a subjective value and the results from the entire process could be questioned. The scatter present in a data set is one thing that serves to validate the process, that demonstrates that the results are real. You need to have a reason to throw out the high numbers - in this case you stated that your laser reads high by 0.5 yards against a bright sky. This should be applied to all of the measurements against a bright sky. Perhaps I am being over sensitive in light of how emails stolen from a climate research lab is being used by climate deniers. Sentences have been taken out of context, and published as proof of a great climate conspiracy regardless of what they actually mean in the emails or to what they refer. http://www.youtube.com/watch?v=7nnVQ2fROOg So perhaps I am thinking about how something reads, rather than what is meant by the words in context. I know you are dedicated beyond all rational reason to producing accurate numbers for all of the trees you measure and do not fudge the data. My comment was directed at the perception of what was written.
We have discussed some of this before. There is a tendency among some members to arbitrarily throw out the high measurements and go with a more "conservative" lower value. Unless you can demonstrate that the high numbers are in error by random chance, or that these high numbers are systematically wrong, they are as valid as any other measurement. Going with a so called conservative number is introducing greater error into the analysis by overweighting the lower values and throwing out valid higher readings. An average of a cluster is acceptable because it may include angle errors and other miscellaneous errors from shake, etc. From a theoretical standpoint, with many measurements on a top the average of a cluster should be the average of the highest cluster because of how the laser works. Calibration and systematic errors in the laser can lead to erroneously high reading. The high reading with a bright background pattern is a type of systematic error. Otherwise the laser can not read longer than the the target. Therefore if multiple high readings show a high cluster, and the average of that highest cluster is the one closest to the truth - (after you correct for calibration errors, and in this case background errors.) It is the numbers in the lower cluster that should be discarded from the averaging. I have enjoyed reading your recent adventures. Ed Check out my new Blog: http://nature-web-network.blogspot.com/ (and click on some of the ads) ----- Original Message ----- From: [email protected] To: [email protected] Sent: Sunday, December 06, 2009 7:25 PM Subject: Re: [ENTS] Settling the issue on Thoreau Ed, Well, I surely regret that I've given such an undesirable impression. What I was attempting to show were the results that follow from making certain defensible assumptions, at least defensible on the surface. I'm not saying any particular result is absolutely the right one, just showing patterns, asking questions, and stating preferences based on my experience with the equipment. I've tested the Nikon 440 enough to know the patterns it gives. But I don't always factor them in because I don't have input coming from an independent source, such as a tape drop. On any given measurement, I can't be sure of whether to leave a distance alone or reduce it by 0.5 yards. However, examining the patterns can point to a choice. I suppose the number I would have like to have settled on is 158.2 feet - the high. But, I am obviously not doing that. The actual height of Thoreau most likely falls between 156.0 and 157.0 feet - if the laser returns are from the top most twig. The safest course to avoid a number that is too high is to adjust the top distance readings by reducing each by 0.5 yards. I can make a good case for leaving the bottom distances alone. If I follow this strategy for the set of measurements taken and average the results, I get 156.7 feet. From how my Prostaff 440 laser behaves on bright versus not bright targets, and bright versus non-bright backgrounds, this is a defensible course, whatever the result. It happens to be 156.7. I wish it were 158.2, but can't accept that value. So, why go through the process at all? Because, Andrew's tape drop added an additional piece of information that steered me away from the high end measurements. Bob ----- Original Message ----- From: "Edward Frank" <[email protected]> To: [email protected] Sent: Sunday, December 6, 2009 2:15:37 PM GMT -05:00 US/Canada Eastern Subject: Re: [ENTS] Settling the issue on Thoreau Bob, You can't pick and choose among the various measurements you get for a tree to generate a number you want to achieve. There needs to be a consistent criteria for accepting or rejecting certain numbers if you feel they are invalid and if they meet the rejection criteria. Juggling the laser numbers to match the results of a tape drop doesn't count. I realize you are not trying to manipulate the measurement data, but that is what you are effectively doing in this situation. The data is what it is, and should not be arbitrarily prettied up to make it better. That is certainly the impression anyone reading these posts will get. Ed Check out my new Blog: http://nature-web-network.blogspot.com/ (and click on some of the ads) -- Eastern Native Tree Society http://www.nativetreesociety.org Send email to [email protected] Visit this group at http://groups.google.com/group/entstrees?hl=en To unsubscribe send email to [email protected] -- Eastern Native Tree Society http://www.nativetreesociety.org Send email to [email protected] Visit this group at http://groups.google.com/group/entstrees?hl=en To unsubscribe send email to [email protected] -- Eastern Native Tree Society http://www.nativetreesociety.org Send email to [email protected] Visit this group at http://groups.google.com/group/entstrees?hl=en To unsubscribe send email to [email protected]
