==============================================================================
TOPIC: Use of Tsuga Search data
http://groups.google.com/group/entstrees/browse_thread/thread/db9804774e510f01?hl=en
==============================================================================

== 1 of 1 ==
Date: Tues, Jan 1 2008 9:54 pm
From: dbhguru@comcast.net


ENTS,

Thanks to Will and Jess, we're awash in high quality data on old growth hemlocks. Attachment 1 is an Excel workbook with 8 spreadsheets that examines a number of OG hemlocks from a trunk modeling standpoint. Spreadsheet DataAndModels compares the actual circumference of the hemlocks at as close to 100 feet up the trunk as Will's data permits to that predicted by several trunk models, more specifically, the paraboloid, cone, and three mixes of paraboloid and cone (50-50, 67-33, 33-67).

Spreadsheet SummaryOf Models provides a convenient summary of the DataAndModels spreasheet. I think I have presented these results before. Spreadsheet 100FtAnalysis is an excerpt from DataAndModels for the cone and paraboloid only. Spreadsheet PctConeVSPara is an excerpt for the cone and parabolid mixes. The last two columns of this spreadsheet show total measured volume and big tree points as applied to only circumference and height for each hemlock. I didn't have crown spread data, but do believe that circumference can be treated as a surrogate for crown spread.

Spreadsheet PtsVSVol regresses volume against points as the independent variable to produce a bivariate linear regression equation with a regression coefficient of 0.82. That's not shabby. Obviously, the idea here is to see how well weighted points can be used to predict trunk volume.

Spreadsheet HgtAndCirVSVol regresses volume against height and circumference as independent variables with a regression coefficient of an impressive 0.89. The last two spreadsheets show the results of the two regressions.

The regression tests show that using height and circumference separately to predict volume is measurably better than using the weighted points approach. Since the weighted formula is less successful at predicting trunk volume than a direct multivariate linear regression of height and circumference onto volume, we are left to ponder the validity of the formula extended to the three dimensions of height, circumference, and spread as a predictor of tree bigness - to the extent that bigness is defined primarily around trunk volume.

Bob

==============================================================================
TOPIC: Excel Workbook
http://groups.google.com/group/entstrees/browse_thread/thread/743348eb8e4d5167?hl=en
==============================================================================

== 1 of 1 ==
Date: Wed, Jan 2 2008 10:08 am
From: dbhguru@comcast.net


ENTS,

If anybody has opened the Excel workbook I sent earlier as an attachment to an e-mail, and are confused by it, I apologize. Please feel free to ask questions about it and I will happily clarify. Communications such as the Excel workbook are often presented by me without any strong expectation of responses from fellow and lady Ents , although I'm happy when I get responses. My main purpose with the technical e-mails is to provide a continuous stream of infromation and data that will become available to researchers when Ed posts the communications threads to the ENTS website. For me, it is easier to present material on these technical subjects in bits and pieces via th e-mail process. However, I hope that some of the material will stimulate interest in others to join the measuring mission of ENTS.

In presenting technically oriented e-mails, there is usually a set of priorities that I am attempting to implement. The first priority is to produce a more accurate statistical description of a forest site and/or individual tree. Rucker analysis and its spin-offs does that. The description priority is usually suffiicient to justify the effort. The second priority is to perfect our measuring craft, push the envelope as it were, but always in a direction of our own choosing. Development of formulas meets the second priority. We don't have the market cornered on tree measuring, but at this point we're by far the best show in town within the geographical East. I don't want us to lose that edge. The third priority is to build databases of scientifically useful research data. I have faith that the Tsuga Search database and my reports on MTSF will eventually be courted by serious scientific researchers. The fourth priority is to provide a means of distinguishing exemplary forest si
tes and individual trees through measurment methods that highlight differences. In the past Tom Diggins has spoken eloquently to this point. The exemplary status of Zoar Valley, NY, was virtually invisible to the local DEC people. Their techniques were (and still are) too crude to reveal the exemplary features of Zoar. About all the DEC representatives have done has been to add up board feet and to misrepresent the true nature of the resource at Zoar Valley (Ooh, my bad).

The above having been said, I still plan to produce a set of examples to help interested Ents explore the value of formulas, methods, and protocals.

Bob