Topic: Countour Creation. Survey requirements.

[therock003]

On a more technical note, how many points must you survey out in the field in order to create an practical surface model with corresponding contours being accurate? I've been asked to survey a land and create contours. Although i know how to do this on civil, i haven't had any practical experience out in the field.

So basically, I need to know about field point distribution and contour accuracy determination.

[Dinosaur]

Our surveyors would usually shoot a roughly 50' by 50' grid PLUS any features ie: top of bank / toe of slope, flowline of any swales, utilities and structures.  This topography ranged up to 8% slopes or better and the results were generally satisfactory for 2' contour intervals.  If your area is flat and you are looking at smaller intervals, I have seen small areas shot with a 20' - 25' grid.

[therock003]

So you set up an orthogonal grid and you stake out the points that intersect nside that Grid?

Well actually the area in question is not flat at all. There's a creak that divides one parcel into two pieces, and some surveing must be done to determine elevations and create a drawing and overlay the contours.

IS there any way to estimate the contour accuracy? I.e. +/- 2cm precision per point extrapolated from the model?

[Alan Cullen]

Thats a bit hard to do today, considering that most surveying in "bush" areas is done using satellites, which only give an accuracy of, say, +/- 5cm (at min). But actual field survey using standard methods should give you a result to +/- 2 cm. Doesn't really matter...the overall DTM is very representative of the area in question. Then when you take into account constuction tolerances, you are home and hosed. It all works out quite well.

[therock003]

What do you mean by bush areas? You mean areas with lots of flowers and plants? Cause that's true for my case. Question here is, how do you suggest i work with this one?

[Alan Cullen]

What do you mean by bush areas? You mean areas with lots of flowers and plants? Cause that's true for my case. Question here is, how do you suggest i work with this one?

Just do your survey and don't get too concerned with survey tolerances. They will be eaten up during design and construction. Normal site tolerances for survey work is around 1:75,000. So you are quite safe. Flat ground changes nothing. A bulldozer has a tolerance of +/- 300mm. Once construction equipment gets on site, all your worry about contours suddenly means absolutely nothing. Just do your job as you would normally do it, and stop worrying about the fact that it is flat country with a creek going through the guts of it.

What I'm saying is....don't worry about how tight you have to set your grid up to close your suvey up to get super accurate contours...because at the end of the day, it means nothing. If I drive a dozer over your survey, I have just stuffed up all your accuracy, and I haven't even put the bloody blade down yet.

Think about it. We all chased accuracy when we were young.

[Bakerman]

In the "old" days we did set up a grid and collect data at each grid point. That was because the topo drawing was done by hand. Now, with computers, it is much more important to get the information at the critical locations. Just imagine what you would need to have to draw a representative topo and collect those shots in the field.  Sometimes 20 may be enough.........sometimes 2000....just depends. 

[sinc]

Yes, we no longer lay out grids.  The 50x50 grid that Dino mentioned is a rough spacing between points, and it's the spacing we would often use to get a topo with relatively high accuracy.  With some surveys, we don't need that accurate a surface, and we might go more like 100' or more spacing between points.  If we are trying to fix a drainage issue in a relatively flat area, we might go 10' spacing between points.

The critical thing is that we do NOT shoot a grid.  We shoot all the important feature points, such as edges of asphalt and concrete, distinct grade breaks, etc.  Then, in areas where there is no distinct feature to shoot, we make sure we get a shot roughly every 50', to fill in the areas between distinct features.

The big problem with field surveying is that it is mostly a mixture of experience and common-sense.  So there are few hard-and-fast rules.  As Surveyors, one of the most important things we do is know how to deal with error.  There is ALWAYS some amount of error.  As Surveyors, we must be able to make judgment calls about how we do our work.  That means we must know the purpose of our survey, and the choices we make will take into account the demands of the situation.  If we went out and did every topo survey with a 25' grid (again, this doesn't actually mean we lay out a grid, just that we have a max of roughly 25' between points), it would be a waste of time, and we would spend far too much time on each job for the accuracy we need.  That would not make good business sense.

This is why Surveyors typically start out as a second- or third-man (or fourth- or fifth-man, if you work for some government agencies) on a Survey crew - they have a LOT to learn from the more-experienced Surveyors before they can really do the job themselves.

[Alan Cullen]

In the "old" days we did set up a grid and collect data at each grid point. That was because the topo drawing was done by hand. Now, with computers, it is much more important to get the information at the critical locations. Just imagine what you would need to have to draw a representative topo and collect those shots in the field.  Sometimes 20 may be enough.........sometimes 2000....just depends. 

Thank you mate, that is exactly what I'm trying to say here. Forget accuracy for accuracies sake, and just understand......pick up critical points (high and low points)...and all will be sorted out. Once it gets to design phase and then construction phase, all your extra work will count for zero. Because it really doesn't matter at the end of the day.

What Bakerman is trying to say is......pick up high and low points and don't worry about having a super accurate contour over the site......because designers don't really care. As long as they have a DTM that represents the site, they don't care that you have given them contours at 0.1m intervals.

[sinc]

IS there any way to estimate the contour accuracy? I.e. +/- 2cm precision per point extrapolated from the model?

Yes and no.  The important thing to remember is that "all contours are lies".  They illustrate the rough elevation of a point, but they should not be considered to be precise.  If you need a precise measurement, the field surveyor should take a shot at the point in question, and label the exact spot elevation at that point.

The general rule-of-thumb is that at any point in your survey, the actual ground surface is probably ±1/2 the countour interval.  So if you have 1m contours, at any point on the survey, you can expect that the real ground surface is probably within ±0.5m of the contour.

This isn't always the case, though, and in some rare situations, you might have an actual ground that differs from the contour elevation by almost the entire contour interval.  For example, pretend you are doing a topo of an area, and you are creating 1m contours.  The area is a big flat pad, and the top of the pad is the high point of the survey, at Elev=40.95m.  Everything else is lower than that top of pad.  If all you do is put 1m contours on the survey, the highest contour you will see is the 40m contour.  According to the contours, it will look like the top of the pad is at Elev=40.00m.  So the pad would have to be specifically labeled with its correct 40.95m elevation - the contours do not yield this information.

The important point I'm getting at is don't put too much stock in contours.  They are most-important for seeing the relative layout of the land - i.e., where are grade breaks, ditches, ridges, how steep are slopes, etc.  If you need precise elevation measurements, rely on spot elevations, not on contours.

[mjfarrell]

To amplify what SINC has said.

Get the Hard Edges of stuff that has them, Buildings, abutments, headwalls, sidewalks, Roadways.

Get the Fuzzy Edges, Berms, Earth Ditches, and Dikes, Edges of Streams.

Get the natural ground, if the terrain is regular space ones shots accordingly, if irregular, take enough shots to define it.

And the contour data is an interpolation, so the more care, and data one puts into the model the better ones results will be.

[Swift]

First thing to note is depending on your area this may, by law anyway, be a duty of a licensed land surveyor.

As far as accuracy goes it depends
- the terrain being surveyed
- the technology employed (total station, robotics, gps, level rod)
- the most most important determining factor thou is denisity of shots

[Dent Cermak]

At our company the distance between shots is related to the drawing scale. If your final drawing is to be drawn at 1"=30', then the MAXIMUM distance between shots will be 30 feet. ALL breaks and planimetric features falling between grid lines will be shot. You MUST have ALL of the breaks. Shooting on a set grid will NOT get it done, unless you are surveying a football field.

[Dent Cermak]

IS there any way to estimate the contour accuracy? I.e. +/- 2cm precision per point extrapolated from the model?

Yes and no.  The important thing to remember is that "all contours are lies".  They illustrate the rough elevation of a point, but they should not be considered to be precise.  If you need a precise measurement, the field surveyor should take a shot at the point in question, and label the exact spot elevation at that point.

The general rule-of-thumb is that at any point in your survey, the actual ground surface is probably ±1/2 the countour interval.  So if you have 1m contours, at any point on the survey, you can expect that the real ground surface is probably within ±0.5m of the contour.

This isn't always the case, though, and in some rare situations, you might have an actual ground that differs from the contour elevation by almost the entire contour interval.  For example, pretend you are doing a topo of an area, and you are creating 1m contours.  The area is a big flat pad, and the top of the pad is the high point of the survey, at Elev=40.95m.  Everything else is lower than that top of pad.  If all you do is put 1m contours on the survey, the highest contour you will see is the 40m contour.  According to the contours, it will look like the top of the pad is at Elev=40.00m.  So the pad would have to be specifically labeled with its correct 40.95m elevation - the contours do not yield this information.

The important point I'm getting at is don't put too much stock in contours.  They are most-important for seeing the relative layout of the land - i.e., where are grade breaks, ditches, ridges, how steep are slopes, etc.  If you need precise elevation measurements, rely on spot elevations, not on contours.

That accuracy requirement is based on National Map Accuracy Standards for small scale mapping (1:2500; 1:50,000 etc.). There are new Standards for engineering scale drawings, ie. 1"=30' etc. They are a bit tighter. I downloaded a copy of the New Accuracy Standards the other day. When I get back to the office I will see if I can get y'all a link to the site.

[sinc]

Off-hand, I couldn't find the new standards you are talking about.  I only know about the ones that have been in force for some time.  This is the most recent thing I found in a (brief) search:

http://egsc.usgs.gov/isb/pubs/factsheets/fs17199.html

I definitely could not find anything about requiring one shot every 30' for a 1" = 30' drawing.  Maybe that came from these new standards you're talking about?

In any case, this Standard was created for federal agencies that produce maps, and I don't think it has any greater jurisdiction.  We have never worked for the USGS, or for any other government agency that produces maps, and we have never had a client request something that conforms to these Map Accuracy Standards.  It's probably a moot point, anyway, since pretty much any topo survey we do exceeds the accuracy requirements for these maps.

In our line of work, it's far more important to know why the topo survey was requested, so we can make sure to get the level of detail necessary for the engineer to do his job.  That's our goal, and Map Accuracy Standards mean nothing.

(Of course, there are also ALTA surveys, but those have their own accuracy requirements, and again, those are higher than the Map Accuracy Standards.  And ALTA surveys generally do not include a topo.)