What a traverse actually measures
A traverse is a chain of measured legs — angle, distance, angle, distance — connecting a sequence of survey stations. You measure with a total station at each setup, then carry your coordinates from a known starting point through every leg until you arrive back at a known point. That last sentence is the entire game: the leg observations alone never tell you whether your work is right. Closing back to a known position does.
- Closed-loop traverse. Starts and ends at the same physical station. Best self-check geometry — both the angular sum and the coordinate sum have known expected values, so any error shows up immediately.
- Closed-route traverse. Starts on one known station and ends on a different known station. Most common in field practice — you tie into published control at both ends instead of double-occupying.
- Open traverse. Starts on a known point but never closes back to one. No self-check. Acceptable only for preliminary work; never for legally defensible measurement.
The angles have to sum correctly first
Before you compute coordinates, check the angles. The angular closure check tells you whether the instrument setups were rotated consistently — if the angles don't sum to their expected value, the distances are computed against a corrupted rotation frame and adjusting positions later won't save you.
Typical tolerance for boundary work: ±30″ × √n — thirty arc-seconds times the square root of the number of stations. Tighter for control work, looser for topographic. Outside that, re-occupy the worst angle before computing anything else.
From legs to latitudes and departures
Once the angles balance, convert each leg from polar (azimuth + distance) to rectangular (latitude + departure). Latitude is the north–south component, departure is the east–west component. The two together fully describe the leg as a coordinate delta.
For a closed-route traverse the sums equal the known coordinate difference between the starting and ending stations, not zero. Same math, just a different expected value for the right-hand side.
Linear closure and precision ratio
Whatever's left after summing lats and deps is your closure error vector. Its magnitude — the linear closure — divided by the total perimeter gives the precision ratio, which is how the spec sheets and standards talk about traverse quality.
Running a traverse end-to-end
The math is half the job — the other half is consistent field procedure. Watch how this looks on real ground: setup, backsight, foresight, instrument transport, and recording the angle + distance for every leg without breaking the chain.
Bowditch (Compass) rule, in practice
With the closure error in hand, you distribute it back through the legs. The Bowditch (also called Compass) rule assumes angles and distances are equally trustworthy and spreads the correction in proportion to each leg's length. It's the default for survey-grade traversing with modern total stations.
- Compass rule. Apply when angles and distances are observed with comparable precision. The modern total-station default.
- Transit rule. Distribute corrections in proportion to lat and dep magnitudes instead of leg lengths. Use when angles are observed with higher precision than distances — rare with today's instruments.
- Least squares. Rigorous adjustment that uses redundant observations to weight every measurement statistically. Required for high-order control networks; overkill for routine boundary work.
Edit the legs below to see how a closure error of a few centimetres distributes back through the traverse. The corrections column shows what Bowditch shaves off each leg.
Try this: with the seed values the closure is small but visible. Now bump leg-1 distance from 120 to 122 — watch the red dot drift away from start, the orange closure-error arrow grow, and the precision ratio drop. Bowditch (compass rule) silently absorbs the bump by stretching every leg proportionally to its length. On the exam you'll be graded on whether you can compute closure and apply a Bowditch adjustment by hand — the visual is just here to build intuition for what the math is doing.
How tight does the closure need to be?
The acceptable precision ratio depends on the work the traverse supports. State boards and federal standards publish minimums — here are the numbers you'll see on most US spec sheets.
| Order / class | Minimum precision | Typical use |
|---|---|---|
| 1st-order control | 1 : 100,000 | Geodetic primary control, GNSS-tied control networks |
| 2nd-order Class I | 1 : 50,000 | High-accuracy boundary, major engineering control |
| 2nd-order Class II | 1 : 20,000 | Control densification, photo-control horizontal |
| 3rd-order | 1 : 10,000 – 1 : 5,000 | Topographic, boundary, most engineering layout |
| Construction layout | 1 : 3,000 – 1 : 5,000 | Site grading, building corners on rough work |
Modern total stations and GNSS-tied control routinely hit better than 1 : 50,000 on medium-length loops. The targets above are floors, not aspirations — aim better and you have margin against future re-survey.
Mistakes traverses hate the most
- Mixed angle conventions. Interior vs deflection vs angle-right is the #1 source of angular blunders. Pick one convention for the whole loop and label every page of the field book.
- Reflector constant drift. A single mis-set prism offset poisons every distance. Cross-check the constant at the start of each day and after any equipment swap.
- Tribrach not levelled to the same precision as the angles. A 30-arc-second tilt at the tribrach is the same physical error as a 30-arc-second angle blunder, and it propagates through every backsight + foresight at that setup.
- Forcing closure by editing the field book. If a leg won't close, the right answer is almost always to re-occupy — not to adjust an angle or a distance after the fact. The Bowditch correction is the only edit that gets to touch your raw observations.
How well did it stick?
A quick 5-question check on Traverse Closure and Adjustments. See where you stand and what to review.