Every survey job starts with control. You pull up the project, locate your benchmarks, and start work. The assumption — sometimes stated, often just implied — is that those control points are correct relative to each other. But assumption is not verification. And the difference between the two can mean the difference between a deliverable that closes and one that has to be redone from scratch.
Leveling between control points before you begin work is the step that confirms your vertical framework is sound. It is not a formality. It is the foundation everything else is built on.
What “Leveling Between Control Points” Actually Means
When a project has two or more benchmarks — published NGS monuments, project control set by a previous surveyor, construction benchmarks established by the engineer — those points each carry an elevation value. The question is whether those elevations are consistent with each other in the field.
A differential level run between control points answers that question directly. You run a loop or a traverse from one benchmark to another, check your closure, and either confirm the control is consistent or identify that something is wrong before you use it.
This matters because benchmarks are not immune to disturbance. Monuments get hit by equipment, frost-heaved, undermined by erosion, or simply disturbed over time. A benchmark that was set correctly five years ago may not be at the same elevation today. Published elevations don't update automatically when the monument moves.
What Goes Wrong When You Skip It
You build on a false foundation
If you accept control elevations at face value and begin setting grade, running pipe inverts, or establishing finished floor elevations — and one of your benchmarks is wrong — every observation you make from that point is contaminated. The error doesn't stay local. It propagates through the entire job.
This is particularly dangerous on construction surveys where your work informs physical decisions. A benchmark that is 0.15 feet high doesn't just produce a bad survey — it produces bad concrete, wrong grades, and drainage problems that get discovered after the contractor has already poured.
You lose the ability to detect error in the field
Running a level loop between control points gives you a closure check. That check is your early warning system. If the loop closes within tolerance, you can proceed with confidence. If it doesn't, you know something is wrong before you've staked anything.
Without that check, you have no mechanism to detect error until the deliverable is complete — at which point correcting it means going back to the field. The cost of a return visit almost always exceeds the time it would have taken to run the level loop in the first place.
You have no documentation if something is disputed
If a client or contractor later questions your elevations, your field notes are your defense. A documented level run between control points — with backsights, foresights, instrument heights, and a calculated closure — demonstrates that you verified your control before you used it. Without that documentation, the burden of proof shifts to you in the worst possible direction.
Acceptable Closure and When to Investigate
The standard tolerance for differential leveling is typically expressed as a function of distance: third-order leveling allows ±0.05 feet times the square root of the distance in miles. For most construction and boundary surveys, you want to be well inside that.
A practical rule: if your loop closure exceeds 0.02 feet on a short run, investigate before you proceed. Common causes include:
- A disturbed monument — physically check the benchmark for signs of movement
- A rod reading error — re-run the section where error is suspected
- Instrument out of adjustment — run a peg test to check for line-of-sight error
- Heat shimmer or atmospheric refraction — relevel during cooler morning hours
- An incorrect published elevation — compare against a third benchmark if available
Do not average two conflicting benchmarks and move on. That is not a resolution — it is a decision to carry half the error into your work. If your control doesn't close, find out why.
Running the Loop Efficiently
The objection to leveling between control points is usually time. It adds setup, it adds turning points, and on a tight schedule it can feel like overhead on a job that should be straightforward. That objection is understandable. It is also shortsighted.
A well-run level loop between two benchmarks a few hundred feet apart takes fifteen to twenty minutes. A return visit to re-run a job because the control was wrong takes half a day, minimum, plus the cost of rescheduling and the credibility hit with the client.
The efficiency gains come from running the loop systematically from the start:
- Set up on the way in. Run your level loop as you walk the site before you do anything else. You're already there — use the time.
- Use a direct route with stable turning points. Avoid soft ground, heavy traffic areas, and anything that might settle between backsight and foresight.
- Record everything. BS, HI, FS, and calculated elevation at each turning point. If something is questioned later, your notes need to tell the full story.
- Calculate closure before you leave the loop. Don't wait until you're back at the office to find out whether the run closed. Do the math in the field so you can investigate immediately if something is off.
The Documentation Standard
Your level loop notes should be complete enough that someone else could check your work. That means:
- Benchmark identifications and stamped elevations for each point used
- A complete record of every backsight, instrument height, and foresight
- Calculated elevations at each turning point
- The calculated closure and whether it meets tolerance
- Any anomalies observed and how they were resolved
This is the same standard that holds up in a dispute, satisfies a reviewing surveyor, and protects your license if a job is ever questioned. Field notes taken in the moment are the only contemporaneous record of what actually happened on a job site. Treat them accordingly.
The Bottom Line
Leveling between control points is not optional on any job that uses more than one benchmark. It is the verification step that separates a survey built on confirmed vertical control from one built on assumptions. The time it takes is measured in minutes. The cost of skipping it — when something is actually wrong — is measured in days and dollars.
Run the loop. Record the work. Close the notes before you leave the site.