Find the length of a crest or sag vertical curve from the two grades and the stopping sight distance it must provide. Enter the approach and departure grades as percentages and the required sight distance, and this calculator identifies whether the curve is a crest or a sag, applies the appropriate sight distance formula including the case where sight distance exceeds curve length, and returns the required curve length, the K value and the rate of grade change.
Grade difference
Curve length, sight distance shorter than the curve (S < L)
Curve length, sight distance longer than the curve (S > L)
Design parameters
G₁, G₂ — approach and departure grades, percent, signed in the direction of travel
A — algebraic difference of grades, always taken positive
S — stopping sight distance the curve must provide
L — required length of the parabolic vertical curve
K — curve length per unit grade change, the parameter design tables are indexed on
RC — rate of grade change per metre, the reciprocal of K
h₁, h₂ — driver eye height and object height assumed for the crest case
A road rising at 3.5 percent meets a falling grade of 2.5 percent. The design speed calls for a stopping sight distance of 120 m.
A = |3.5 − (−2.5)| = 6.000 %
G₂ is below G₁, so the curve is a crest and the divisor is 658.
L = A·S² / 658 = 6 × 120² / 658 = 86 400 / 658 = 131.3 m
Is L greater than S? 131.3 > 120, so yes — the sight line does fall entirely within the curve and the expression used was the right one. The result is reported as Crest, S < L.
K = L / A = 131.3 / 6 = 21.88 m per %
RC = A / L = 6 / 131.3 = 0.04569 %/m
Grade difference A 6.000 %
Required curve length 131.3 m — Crest, S < L
K value 21.88 m per % · Rate of change 0.04569 %/m
A valley: the road falls at 3.0 percent and then rises at 3.0 percent. The same 120 m stopping sight distance applies.
A = |−3.0 − 3.0| = 6.000 %, identical to the crest example.
But G₂ is above G₁, so this is a sag and the divisor becomes 120 + 3.5 × 120 = 540.
L = 6 × 120² / 540 = 86 400 / 540 = 160.0 m
160.0 > 120, so the S < L branch is correct: Sag, S < L.
K = 160.0 / 6 = 26.67 m per %, against 21.88 for the crest.
The sag needs 22 percent more curve for the same grade change and the same sight distance.
Crest 131.3 m, K = 21.88
Sag 160.0 m, K = 26.67
Same A, same S, 28.7 m more curve required in the valley
In daylight nothing obstructs the view across a sag, so sight distance is not the issue. The controlling condition is night driving: a headlight sits about 0.6 m above the road and its beam spreads only around one degree upward, so as the road curves away beneath it the illuminated length grows slowly. That is a more demanding geometry than a driver's eye at 1.08 m looking over a crest at a 0.6 m object, and it is why the sag divisor is smaller and the curve longer.
A gentler crest: 3.0 percent rising to 2.0 percent falling, so A = 5.000 percent, with the same 120 m sight distance.
L = 5 × 120² / 658 = 109.4 m
But 109.4 is less than the 120 m sight distance, so the assumption behind that expression — that the whole sight line lies on the curve — is false. The answer cannot be used.
L = 2S − 658/A = 240 − 658/5 = 240 − 131.6 = 108.4 m
The calculator reports this as Crest, S > L.
S < L form 109.4 m — invalid here, since 109.4 < 120
S > L form 108.4 m — the governing answer
K value 21.68 m per %
The two forms cross over when A equals the constant divided by the sight distance — at S = 120 m that is A = 658/120 = 5.48 percent for a crest. Below that grade difference the S > L branch governs, and the gap between the two widens as A falls. Push A low enough and 2S − 658/A goes negative, meaning sight distance alone requires no vertical curve; the comfort and appearance criteria then take over, and this calculator does not check either.
| Input | SI unit | Imperial unit | Accepted | Notes |
|---|---|---|---|---|
| Approach grade G₁ | % | % | any sign, typically −10 to +10 | Positive uphill in the direction of travel |
| Departure grade G₂ | % | % | any sign, typically −10 to +10 | Below G₁ gives a crest, above gives a sag |
| Stopping sight distance S | m | ft | > 0 | Taken from the design speed — see the table below |
Stopping sight distance by design speed — IRC values, level road
| Design speed, km/h | SSD, m | Crest K at this S | Sag K at this S |
|---|---|---|---|
| 30 | 30 | 1.37 | 4.00 |
| 40 | 45 | 3.08 | 7.30 |
| 50 | 60 | 5.47 | 10.91 |
| 60 | 80 | 9.73 | 16.00 |
| 65 | 90 | 12.31 | 18.62 |
| 80 | 120 | 21.88 | 26.67 |
| 100 | 180 | 49.24 | 43.20 |
Crest and sag length for a range of grade differences — at S = 120 m
| A, % | Crest L, m | Branch | Sag L, m | Branch |
|---|---|---|---|---|
| 2 | 0.0 | none required | 0.0 | none required |
| 3 | 20.7 | S > L | 60.0 | S > L |
| 4 | 75.5 | S > L | 105.0 | S > L |
| 5 | 108.4 | S > L | 133.3 | S < L |
| 6 | 131.3 | S < L | 160.0 | S < L |
| 8 | 175.1 | S < L | 213.3 | S < L |
| 10 | 218.8 | S < L | 266.7 | S < L |
Which criterion governs a vertical curve
| Criterion | Applies to | Governs when | Checked here? |
|---|---|---|---|
| Stopping sight distance | Crest and sag | Most of the time | Yes |
| Headlight sight distance | Sag | Built into the sag formula | Yes |
| Passenger comfort | Sag | High design speed, large A | No |
| Appearance / minimum length | Crest and sag | Very small A | No |
| Drainage near the apex | Crest and sag | Long flat curves, kerbed roads | No |
| Overtaking sight distance | Crest | Two-lane rural roads | No |
| Headroom under a structure | Sag | Bridge or gantry over a sag | No |
Sign convention for the two grades
| G₁ | G₂ | A = |G₁ − G₂| | Curve type |
|---|---|---|---|
| +3.5 | −2.5 | 6.0 | Crest — over a hill |
| +4.0 | +1.0 | 3.0 | Crest — rise easing off |
| −3.0 | +3.0 | 6.0 | Sag — through a valley |
| −4.0 | −1.0 | 3.0 | Sag — descent easing off |
| +2.0 | +2.0 | 0.0 | No curve — grades equal |
Vertical Curve Calculator — multicalci.com. Sight distance criterion only. No check on passenger comfort, appearance minimum length, drainage gradient near the apex, overtaking sight distance or headroom under a structure. No chainage table, high or low point location, or unsymmetrical curves. Results are indicative and must be verified against IRC, AASHTO or the governing highway standard.