Run a closed traverse right around a field and back to where you started, and the calculation should bring you exactly home. It never quite does. The small gap between where the computation says you finished and where you actually began is the error of closure.
That gap is the single most useful number in traverse work, because it tells you how good the whole survey is.
What the Error of Closure Is
If the traverse has an error of closure, it will not close when plotted on paper. The end point A′ of the closed traverse ABCDE will not coincide with the starting point A. The distance AA′ by which A′ falls short of coinciding with A is the error of closure.
It is also called the closing error.
Why It Exists at All
Geometrically, a closed traverse must close — you physically walked back to your starting peg. So the gap is not a property of the ground; it is entirely the accumulated effect of measurement error in the lengths and angles.
That is what makes it so valuable. The closing error is a direct, honest measure of how much error crept into the whole survey, obtained without any external reference.
Components of the Closing Error
For a closed traverse, the x and y components are simply the sums that ought to have been zero:
ex = ΣD ey = ΣL
where ΣD is the algebraic sum of all departures and ΣL the algebraic sum of all latitudes.
Why These Sums Should Be Zero
Walk around a closed loop and return to your starting point. Every metre you travelled north must be cancelled by a metre travelled south; every metre east by a metre west.
So in a perfect traverse, ΣL = 0 and ΣD = 0. Whatever they actually come to is precisely the amount by which the traverse failed to close — which is why the sums are the error components.
The Magnitude of the Closing Error
Since ex and ey are perpendicular components, the total closing error follows from Pythagoras:
e = √(ex2 + ey2)
Direction of the Closing Error
tan θ = ex / ey
The signs of ex and ey define the quadrant in which the closing error lies.
Note the Order of the Ratio
The formula is ex over ey — departure over latitude, or x over y. This looks upside down compared with the usual tan θ = y/x from coordinate geometry, and students often invert it by habit.
The reason is that surveying measures bearings clockwise from north, not anticlockwise from east. North is the y axis, so the angle from north to the error direction has the x component opposite it and the y component adjacent — giving tan θ = x/y.
This is the same convention that makes a whole circle bearing start at north and increase clockwise.
Worked Illustration
If both components are positive:
tan θ = (+) / (+) = positive → closing error lies in the 1st quadrant
A positive ex means the computed end point finished too far east; a positive ey means too far north. North-east is the first quadrant, so the result is consistent.
The Correction and Its Quadrant
The sign of the corrections Cx and Cy will be opposite to ex and ey:
Cx = −ex Cy = −ey
Note: the closing error and the correction will always lie in diagonally opposite quadrants.
Why Diagonally Opposite
This follows from the master rule met in the chaining corrections topic — correction equals minus error.
Reversing the sign of both components turns a direction through exactly 180 degrees. And a rotation of 180 degrees moves you into the diagonally opposite quadrant every time:
| Error in Quadrant | Correction in Quadrant |
|---|---|
| 1st (NE) | 3rd (SW) |
| 2nd (SE) | 4th (NW) |
| 3rd (SW) | 1st (NE) |
| 4th (NW) | 2nd (SE) |
It makes physical sense too. If your traverse drifted north-east, the correction must pull it back south-west — the exact opposite direction.
Note that this is the same 180-degree reversal that relates a fore bearing to a back bearing. The closing error and its correction are the same line travelled in opposite directions.
Closing Error for a Link Traverse
A link traverse does not return to its start. It runs between two known control points, so the check is different:
ex = X′ − X ey = Y′ − Y
| Symbol | Meaning |
|---|---|
| X′, Y′ | Computed coordinates of the final control point |
| X, Y | Known coordinates of that same point |
The Same Idea, a Different Reference
Both cases ask the identical question — where did the computation say I ended up, versus where do I know I actually ended up?
- In a closed traverse, the known finishing point is the starting point itself, whose coordinates are already established. So the check reduces to ΣL and ΣD being zero.
- In a link traverse, the known finishing point is a different control point, so you compare computed against known coordinates directly.
An open traverse has neither — it ends nowhere known, so no closing error can be computed at all. This is exactly why open traverses are inherently unchecked and must be run with extra care.
Relative Error of Closure
Relative error of closure = Error of closure ÷ Perimeter of traverse = e / p
Generally it is expressed as a fraction with numerator one, that is, in the form 1 / (p/e).
The relative error of closure is also called the relative accuracy or the degree of accuracy.
Why Relative Rather Than Absolute
This is the crucial idea, and it is what makes the quantity meaningful.
Suppose two traverses each close with an error of 0.1 m. Are they equally good? Not remotely:
| Traverse A | Traverse B | |
|---|---|---|
| Closing error | 0.1 m | 0.1 m |
| Perimeter | 100 m | 10,000 m |
| Relative error | 1 in 1,000 | 1 in 100,000 |
Traverse B is a hundred times more accurate, despite an identical absolute error, because it achieved that error over a hundred times the distance.
Dividing by the perimeter puts every traverse on a common footing regardless of size, which is why specifications state permissible accuracy as a ratio like 1 in 10,000 rather than as a distance.
The Connection to Precision Matching
That figure of 1 in 10,000 should look familiar. It is exactly the linear precision required to match a 20-second theodolite, as derived from tan δθ = δl / l.
So the relative error of closure is the quantity that lets you verify in practice what the precision-matching formula predicted in theory. Compute it at the end of the traverse and you know whether the survey lived up to its instruments.
Formula Summary
| Quantity | Expression |
|---|---|
| Closing error components, closed traverse | ex = ΣD, ey = ΣL |
| Closing error components, link traverse | ex = X′ − X, ey = Y′ − Y |
| Magnitude of closing error | e = √(ex2 + ey2) |
| Direction of closing error | tan θ = ex / ey |
| Corrections | Cx = −ex, Cy = −ey |
| Relative error of closure | e / p, expressed as 1/(p/e) |
Quick Revision Notes
- The error of closure is the distance AA′ by which the computed end point fails to coincide with the starting point. Also called the closing error.
- For a closed traverse, ex = ΣD and ey = ΣL, both of which should be zero.
- tan θ = ex / ey — note x over y, because bearings run clockwise from north.
- The signs of ex and ey define the quadrant of the closing error.
- Cx = −ex and Cy = −ey.
- The closing error and the correction always lie in diagonally opposite quadrants, because reversing both signs is a 180° rotation.
- For a link traverse, ex = X′ − X and ey = Y′ − Y, comparing computed with known coordinates.
- Relative error of closure = error of closure ÷ perimeter, expressed as 1/(p/e).
- It is also called the relative accuracy or degree of accuracy.
- An open traverse has no closing error, since it ends at no known point.
Mistakes Students Commonly Make
- Inverting the direction formula. It is tan θ = ex/ey, not ey/ex.
- Swapping the components. ex comes from departures; ey from latitudes.
- Giving the correction the same sign as the error. It is opposite.
- Placing the correction in an adjacent quadrant. It is always diagonally opposite.
- Judging accuracy by the absolute closing error. Only the relative error is meaningful, since it accounts for the size of the traverse.
- Dividing the perimeter by the error and calling that the relative error. The relative error is e/p; it is merely expressed as 1/(p/e).
- Attempting to compute a closing error for an open traverse, which is impossible.
- Forgetting that ΣL and ΣD must be algebraic sums with signs respected.
Conclusion
A closed traverse must physically close, so any gap in the computation is pure measurement error — which makes the closing error a direct measure of survey quality obtained with no outside reference. Its components are the latitude and departure sums that ought to have been zero, its direction comes from tan θ = eₓ/e_y, and its correction is simply its own reverse, landing in the diagonally opposite quadrant. Divide it by the perimeter and you get the relative error of closure, the figure that lets you compare any two traverses fairly and check whether the survey matched the precision its instruments promised.
Frequently Asked Questions
What is the error of closure?
The distance by which the computed end point of a closed traverse fails to coincide with its starting point. It is also called the closing error.
How are the components of the closing error found in a closed traverse?
The x component equals the algebraic sum of all departures and the y component equals the algebraic sum of all latitudes, both of which should be zero in a perfect traverse.
How is the direction of the closing error found?
From tan θ = eₓ/e_y, with the signs of the two components determining the quadrant in which the closing error lies.
Why is the ratio x over y rather than y over x?
Because surveying measures bearings clockwise from north, so north is the reference direction. The x component is opposite the angle and the y component adjacent to it.
What is the relationship between the error and the correction?
The corrections are equal and opposite to the error components, so Cₓ = −eₓ and C_y = −e_y.
Why do the error and correction lie in diagonally opposite quadrants?
Because reversing the sign of both components rotates the direction through 180 degrees, which always lands in the diagonally opposite quadrant.
How is closing error computed for a link traverse?
By comparing the computed coordinates of the final control point with its known coordinates, so eₓ = X′ − X and e_y = Y′ − Y.
What is the relative error of closure?
The error of closure divided by the perimeter of the traverse, generally expressed as a fraction with numerator one. It is also called the relative accuracy or degree of accuracy.
Why is relative error used rather than absolute error?
Because the same absolute error means something very different over a short traverse than over a long one. Dividing by the perimeter puts traverses of any size on a comparable footing.
Can an open traverse have a closing error?
No. An open traverse neither returns to its starting point nor ends on a known control point, so there is nothing to compare the computed finishing position against.