Page 1 of 2

Overland Survey Loop

PostPosted: Apr 24, 2011 11:11 pm
by Dan Green
The recently connected Heavy Breather System (653m deep) in Canada has a very accurate 1020m of overland survey between the two entrances; when compiled as part of the loop this 1020m is reduced to 1017m (this total that could change with future underground loops).

Q: Should the length of the system be:
1. The total survey subtract the length of the overland survey (1020m)?.....OR
2. The total survey subtract the length of the adjusted overland survey (1017m)?

Dan

Re: Overland Survey Loop

PostPosted: Apr 25, 2011 8:48 am
by Crockett
3?. Keep the overland survey static and allow presumably less accurate underground survey shots to adjust to close the loop?

Re: Overland Survey Loop

PostPosted: Apr 25, 2011 4:57 pm
by Bob Thrun
You're concerned about 3 meters out of a kilometer?!!

I would use the distances originally measured in the cave as the basis of surveyed cave length, not the adjusted survey shots. Fred Wefer proved that a loop closure always reduces the length of the loop. You can read his proof in the 1974 Speleo Digest.

There are many ambiguous situations in deciding what survey shots or parts of survey shots should be counted toward measuring the length of a cave. I am not going to give examples because the list is too long. The "length" of a cave is always an approximation.

Unless the in-cave survey conditions are really bad, the in-cave part of a survey is likely to be more accurate than an outdoor survey because the in-cave shots are shorter. The optimum length of survey shots for the most accuracy is in the 5 to 10 meter range. This is where the errors due to angle measurements are the same as the errors due to length measurment. If there is a significant difference between the in-cave and outdoor measurements, there are some data reduction programs that allow you to set the weights on the measurements that go into an adjustment.

Re: Overland Survey Loop

PostPosted: Apr 25, 2011 6:42 pm
by Dan Green
Thanks, good ideas. All things being relatively unequal, I'm thinking it's best to consider the entire survey with all the loops as a single entity (underground + surface), and just subtract the adjusted surface distances I don't want. The real motivation here is a correct depth number, which swings considerably with or without the surface loop that rises 153m. Either way I guess I'm picking fly sh*t out of pepper here...

Re: Overland Survey Loop

PostPosted: Apr 25, 2011 9:46 pm
by Bob Thrun
I thought your concern was the surveyed length, rather than the depth. For the length you want to sum up the original measurements, not the back-calculated values after network adjustment. I never saw much sense in the survey program giving what the distance and angles would get changed to after the adjustment.

A change of survey program or a change in the shot weighting factors will make a change in the depth. The changes will be well within the range of errors. Changing the order of the survey shots will give different results with a sequential adjustment program.

The cavers keeping the Lechuguilla data settled on a depth number and always quote that. Otherwise every survey that could be part of a new route to the lowest point would change the depth slightly and the value would be jittering up and down. That is for the low point staying the same.

Re: Overland Survey Loop

PostPosted: Apr 26, 2011 1:47 am
by Dan Green
Bob Thrun wrote: For the length you want to sum up the original measurements, not the back-calculated values after network adjustment.


Yes, that makes sense. I have always just blindly taken the length and depth that Walls compiles, but of course these are both adjusted numbers that only really correlate exactly to the line plot. I now see the actual total of 6076m under the "Geometry" tab. Back to my original inquiry- now I'll subtract the total overland of 1020m from 6076m instead of the adjusted 1017m from 6046m. The adjusted depth is good to go at -653m. The other Canadians will be happy to see some of their Rockies caves getting bigger during the long winter pub season. Thanks, I appreciate this assistance.

Re: Overland Survey Loop

PostPosted: Apr 26, 2011 5:47 pm
by George Dasher
Holy Cow Batman!! Nothing like a "shallow" cave!!

Bob is correct--the length and depth should come from the uncorrected data.

As far as closing and using an overland survey in your data--I would do it on a case by case situation.

I know, when I included the overland closed loop for the Sinks of Gandy, it just made the cave turn absolutely strange. I thus elected not to include the overland survey.

People don't understand that closing loops changes the relationship of the survey shots. Thus the best technique is first, careful surveying, and second, solving the problems in the raw data.

Where is the Smilie for laying on your back in amazement? Over your depth!

Re: Overland Survey Loop

PostPosted: Apr 26, 2011 8:06 pm
by Bob Thrun
George Dasher wrote:Bob is correct--the length and depth should come from the uncorrected data.

If there is more than one surveyed route between the highest and lowest points, you have to use the adjusted station coordinates for the depth.

Re: Overland Survey Loop

PostPosted: Apr 27, 2011 1:13 am
by Dan Green
Bob Thrun wrote: If there is more than one surveyed route between the highest and lowest points, you have to use the adjusted station coordinates for the depth.

The notable thing about this particular overland is that it's not just a single traverse, but two that I snaked up different parts of the mountain between the entrances. And then I linked them to each other with two 'rungs' to effectively create five loops within the overland itself- I figured this would minimize errors.

Even calculated independently (un-closed), the two half-km traverses are less that 0.5m apart vertically, so when closed I've always been confident that the two entrances were 152 metres apart vertically and think I need to include the corrected overland for the depth calculation. I consider the overland as just part of the survey- just not the underground part, and so therefore the length is subtracted.

George Dasher wrote:Where is the Smilie for laying on your back in amazement? Over your depth!

Yep, it was quite the exciting connection last season! A lot of hard work by a small team of persistent Rockies cavers. It's been lots of fun to be a part of.

Re: Overland Survey Loop

PostPosted: Apr 27, 2011 8:02 am
by Scott McCrea
Pictures? Trip report? Otherwise, we are forced to think you are making it up. :tonguecheek: :wink:

Re: Overland Survey Loop

PostPosted: Apr 27, 2011 8:10 am
by George Dasher
I've been re-thinking your questions...

And I think, regarding the depth, sometimes you're going to have to use the adjusted data.

That part of the question could go either way, and you're going to have to make a case-by-case decision, based on the quality of the data, the problems in the cave, etc.

Re: Overland Survey Loop

PostPosted: Apr 27, 2011 8:45 am
by Crockett
How did you do the overland survey? Total station? Thats what I assumed when you implied it was more accurate. If you used the same techniques as the underground survey why would the surface survey be more accurate?

Re: Overland Survey Loop

PostPosted: Apr 27, 2011 8:51 am
by Crockett

Re: Overland Survey Loop

PostPosted: Apr 27, 2011 9:19 am
by Dan Green
Crockett wrote:How did you do the overland survey? Total station? Thats what I assumed when you implied it was more accurate. If you used the same techniques as the underground survey why would the surface survey be more accurate?

No not total station. You are right, it should be no more or less accurate. I'd assumed more accurate because the two traverses were fairly accurate relative to each other, but I've never compared the surface closure to loops underground.

Scott,
Some photos here from the first thru-trip:
http://www.flickr.com/photos/47581893@N07/sets/72157624795060618/with/4922668296/

A line plot of the cave and a profile map of the connection area:
Image
Image

Re: Overland Survey Loop

PostPosted: Apr 27, 2011 9:30 am
by Scott McCrea
Alright, I guess I believe you now. Nice work and map and pics!