rounding

General discussions of measurement techniques and the results of testing of techniques and equipment.

Moderators: edfrank, dbhguru

User avatar
sradivoy
Posts: 422
Joined: Fri Jun 03, 2011 8:23 pm

rounding

Post by sradivoy » Wed May 27, 2015 11:42 am

Anyone else think it's a good idea to round to the nearest foot (or perhaps to the lowest foot) for all handheld measurements? Personally I think using decimal places to the nearest tenth should be reserved exclusively for tripod measurements, regardless of how inherently accurate your equipment happens to be. I mean a shaky hand can make all the difference, especially when you're dealing with tenths of a foot.

User avatar
Don
Posts: 1560
Joined: Tue Mar 09, 2010 12:42 am

Re: rounding

Post by Don » Wed May 27, 2015 12:55 pm

sradivoy-
I share your opinion...large broadleaf trees can "shrink and expand" in response to diffusion pressure deficits that occur during the day, sufficient to vary more than a tenth of a foot...it is fine to record data in the field as you find it (including the time and weather), but in terms of tree registries, I favor rounding down to the nearest foot.
Don Bertolette - President/Moderator, WNTS BBS
Restoration Forester (Retired)
Science Center
Grand Canyon National Park

BJCP Apprentice Beer Judge

View my Alaska Big Tree List Webpage at:
http://www.akbigtreelist.org

User avatar
edfrank
Posts: 4217
Joined: Sun Mar 07, 2010 5:46 pm

Re: rounding

Post by edfrank » Wed May 27, 2015 11:06 pm

I think a tenth of a foot is fine. People who understand the methodology should understand the basics of significant digits. As for rounding down - NO. If you are going to round a number it should be to the closest number up or down. There are standard rounding criteria. Where the number ends in a 5, 6, 7, 8, or 9 you round up, if it ends in 1, 2, 3, or 4 your round down. Changing standard math rounding protocols doesn't make your numbers any better. Being conservative in numbers doesn't make your data set better, it just introduces a bias in one direction into your set. It is better to round numbers as is generally done rather than adding a personal bias to the rounding procedure because you want to be conservative in your measurements. The numbers should be what they are, not adjusted to suit. In the case of single trunks versus double trunks, the questionable ones should be included in the best interpretation rather than all being discounted and considered a double when the best guess is that it is a single just because of some degree of doubt. It adds bias to the data set. It adds a systematic one-directional bias rather than random non-directional error which tends to cancel out when crunching large sets of data.
"I love science and it pains me to think that so many are terrified of the subject or feel that choosing science means you cannot also choose compassion, or the arts, or be awe by nature. Science is not meant to cure us of mystery, but to reinvent and revigorate it." by Robert M. Sapolsky

User avatar
Don
Posts: 1560
Joined: Tue Mar 09, 2010 12:42 am

Re: rounding

Post by Don » Wed May 27, 2015 11:49 pm

Ed-
In general, I agree with you. We all understand rounding rules from our arithmetic and mathematics classes.

But a tree measured at 135.7 isn't 136 feet tall. Period. It just doesn't measure up (to 136').

Is it 135''? Yes. And more. But not another foot more.

These are not statistics, they're trees. Would we round it up to 140' because we like a tidy one's column?
Don Bertolette - President/Moderator, WNTS BBS
Restoration Forester (Retired)
Science Center
Grand Canyon National Park

BJCP Apprentice Beer Judge

View my Alaska Big Tree List Webpage at:
http://www.akbigtreelist.org

User avatar
bbeduhn
Posts: 1090
Joined: Thu Feb 10, 2011 4:23 pm

Re: rounding

Post by bbeduhn » Thu May 28, 2015 7:43 am

Don,
If all things are done correctly and the tools are working properly, that 135.7' is generally a 136+'. When extreme measures are taken on a champion tree, the reading can be within a tenth of a foot but in most cases, we are looking up at fairly high angles and tend to under measure the top twig. I don't see a need to round up or down to the nearest foot. I always round to the nearest tenth, however.
Brian

User avatar
sradivoy
Posts: 422
Joined: Fri Jun 03, 2011 8:23 pm

Re: rounding

Post by sradivoy » Thu May 28, 2015 8:40 am

Another aspect to consider is that trees are moving targets that generally get taller over time. A tree measured five years ago will be less accurate than a tree measured last week. Although 135.7" trees isn't a 136' tree yet it more than likely will be within a year or two. Rounding up buys a little time for the need to measure again. In other words, assuming that the top doesn't die back or break off, a year of two later after first measured the 136"' rounding will be more accurate than the 135' rounding. I'm good with any of these as long as we're on the same page. I just think that handheld measurements are way too shaky, especially with the clinometer, to have an accuracy within a tenth of a foot. Tripod measurements are different story. No need for rounding there.

User avatar
mdvaden
Posts: 882
Joined: Tue Sep 07, 2010 6:30 pm

Re: rounding

Post by mdvaden » Thu May 28, 2015 9:06 am

It almost doesn't seem to matter if it's handheld.

But I don't see why anyone would want to round up. But it would seem almost rewarding to round down by a foot, or even two feet. That way if its remeasured, the greater height feels even better if the number grows.

Sometimes I provide a "zone".

Like with Tsunami the tall hemlock I found in Prairie Creek redwoods. Hand held, I got around 272 ft.. Late the same day, when I told Steve Sillett about it, I said it should fall somewhere between 270 to 274 ft. That was completely accurate. Further measurement the same week with a tripod put it close to 272 ft.
M. D. Vaden of Oregon = http://www.mdvaden.com

200 Pages - Coast Redwoods - http://www.mdvaden.com/grove_of_titans.shtml

Portraits & Weddings - http://www.vadenphotography.com

User avatar
Will Blozan
Posts: 1153
Joined: Fri Mar 12, 2010 7:13 pm

Re: rounding

Post by Will Blozan » Thu May 28, 2015 4:46 pm

For what it is worth, tape drops as a reference for laser measurement workshops have often been within a tenth of a foot- never a full foot. In the Tsuga Search Project ground-based heights were on the average 1.4 feet (so +/- .7'- if I recall correctly) different from the tape drop BUT this range was exacerbated by not actually hitting the real top until the climb.

I am comfortable with nearest 1/10th with my equipment as verified by tape drop. Can't speak for others though...

Will

User avatar
Will Blozan
Posts: 1153
Joined: Fri Mar 12, 2010 7:13 pm

Re: rounding

Post by Will Blozan » Thu May 28, 2015 4:48 pm

Also, I measured a 306' sequoia with a Nikon 440 and clinometer side by side with Bob Van Pelt and we were 1/10' different. He was using Trupulse on tripod. Lucky? I think not.

User avatar
dbhguru
Posts: 4470
Joined: Mon Mar 08, 2010 9:34 pm

Re: rounding

Post by dbhguru » Sat May 30, 2015 12:18 pm

NTS,

This is a useful discussion to have because there is no single answer on when to apply rounding rules, and to what degree. The answer is case dependent on:

1. Accuracy of equipment being used and the state of its calibration
2. Expertise of measurer
3. Distinctness of target
4. Tripod or no tripod
5. Amount of effort put into verifying measurements (very, very important)
6. Method of measurement (sine vs tangent)
7. Situation involving measurement (championship status, scientific paper, general article, our database, etc.)
8. To whom a measurement is being reported

In the case of champion tree reporting, rounding to the nearest foot, probably rounding down, makes sense today. In the future, as more candidates are certified by National Cadre members, rounding down to the nearest half-foot may be justifiable, and even to a tenth in certain circumstances. Regardless, It would be useful to include a range. For example, if the target is distinct, and I take my time, I can confidently state an accuracy to +/- 0.5 feet with the TruPulse 360. I have repeatedly tested this instrument and obtained an accuracy range of between +/- 0.25 and 0.3 feet when recording at the point of changeover on the display and using a tripod. The TruPulse 200X is accurate to +/- 1.5 inches as tested against industrial red beam lasers accurate to +/- 1.5 millimeters. Interestingly, my Nikon Forestry Pro is pretty darned accurate as well. Here is a just completed comparison between the three instruments for distance.
Comparison3Lasers.png
Note that comparison of straight averages doesn't tell the story. The average of the absolute differences between two instruments is the operative comparison, Note how close the two TruPulses are. Their maximum difference is 0.5 feet and the average is 0.24. To my surprise, the overall performance of the Nikon is quite good, but it hiccups every now and then, and the greater distance tend to be most problematic, although not always.

Note that these numbers do not mean that in height calculations, the three instruments will yield averages that are as close as the above. For height, as opposed to simple distance, we must take into account the returns of the tilt sensors. In the next calibration tests, I'll show some results on the angles, and then put distance and angle together. At that point, we can get into the finer points of when we should call for rounding to a particular level and for what purpose. Lots to discuss. Hopefully, others will contribute, e.g. Michael Taylor.

As a final point, measurers like Michel Taylor and Will Blozan seem to have accuracy meters built into their heads. If they report a measurement to a tenth of foot, I would be reluctant to expect them to round in any direction. I would accept reporting to the tenth of foot. However, for less experienced measurers, rounding down to the whole foot makes sense, especially at the level of reporting for champion tree competitions where rules are needed.

Bob
Robert T. Leverett
Co-founder, Native Native Tree Society
Co-founder and President
Friends of Mohawk Trail State Forest
Co-founder, National Cadre

Post Reply

Return to “Measurement and Dendromorphometry”