Multibeam and Bathymetric Survey: Reading the Deliverable
By Cheryl Ansel, DOT bridge program engineer, NHI-qualified team leader. Reviewed by Dana Whitcombe, technical reviewer.
Underwater Inspection | Commercial Diving Contractors Alabama A July morning on the Sacramento River, and my survey tech is staring at the screen. We'd run a Reson SeaBat 7125 over the old Interstate 5 bridge pier, and the point cloud was lying to us. Or so I thought. Turns out the data was fine. The pier's upstream face had a scour hole that dove to 14 feet below the riverbed—far deeper than any previous single-beam sounding ever suggested. We'd done the bathtymetric survey properly, but reading the deliverable took another hour of messing with vertical exaggeration and sun angle. That's the real skill. Multibeam bathymetry gives you a million points where you used to have a few dozen. But a million points don't mean a thing if you can't tell whether the riverbed is actually a smooth plane or a series of sand waves messing with your filtering. So here's how to read the thing—not just look at it.
Who's Actually Doing the Sonar
First thing I do when I open a multibeam deliverable is check who ran the boat. Call me old-school, but the surveyors' experience shows up in the data. A good hydrographer will tell you about the tide, the boat wake, the suspended sediment. A bad one just sends a zip file. You're looking for the basics: the frequency (we use 400 kHz for shallow water, but 200 kHz sees through mud better), the ping rate, and the swath width. If they didn't log sound velocity profiles every few hours, walk away. That's the difference between a depth error of 0.2 feet and 2 feet. I've seen both. And ask about the inertial nav unit. An Applanix POS MV or a SBET with post-processing is worth every penny. If they used a $2,000 AHRS, treat the outer half of each swath as decoration.
Reading the Point Cloud
Open the point cloud in something like CARIS or HYPACK, and turn off the pretty shade. You're not looking for colors—you're looking for noise. A clean point cloud over a flat mud bottom should have a vertical scatter of maybe 0.3 feet. If you see a 2-foot jump along one side of a line, that's roll bias, not a hole. If you see a perfectly circular dimple in the middle of nowhere, that's a fish. Here's the trick: display the points as bare dots, then exaggerate the vertical scale to 10x. Scour holes pop. Sand waves show their true shape. And you'll spot the stepping artifact where two lines didn't match in elevation. That's your cue to go back to sound velocity. Then switch to the sun-shaded relief, and let the light come from the north. The geomorphology jumps out—old river channels, riprap, debris. I once found a 600-pound stormwater pipe that had fallen off a truck, just by flipping the sun azimuth. The black-and-white image does more than the fancy color ramp.
Sonar Imagery vs. Bathymetry
Don't confuse the bathymetric point cloud with the backscatter mosaic. The first tells you elevation; the second tells you texture. A multibeam sonar can do both if the rendering is right. But the way they make the mosaic is completely different. For the imagery, look at the backscatter values. A hard surface—rock, concrete, a car tire—returns a strong echo, so it's dark on the mosaic. Soft mud returns weak echoes, so it's light. But here's the catch: the same surface looks different if you're on the edge of the swath. If the mosaic has bright edges and dark centers, the gain is misbalanced. That's a surveyor issue, not a riverbed issue. I usually ask for both layers, because the bathymetry might show a clean slope while the imagery reveals a whole row of 55-gallon drums you'd otherwise step on in a dive.
Ground Truth and Anomalies
Every anomalous patch in the point cloud needs a second opinion. If I see a hump that looks like a large rock but the area is known as a disposal site, that hump might be a leftover chunk of concrete from the last pier project. A quick side-scan pass or a diver's probe settles it. This is where the 400 inspections get me—never trust a single sensor. Multibeam is brilliant, but it misses voids under the mud and can't squeeze under a king pile. A 200-kHz single-beam in the right spot gives you a different perspective. Or a magnetometer, if you're hunting steel. And always check the overlap between survey lines. If the riverbed looks like a washboard in the overlapping area of two lines, you've got a positioning mismatch. The way the deliverable handles that—whether they blend or just chop—tells you how careful they are.
Frequently Asked Questions
What is a bathymetric survey? A bathymetric survey measures water depth and maps the underwater floor, often using multibeam sonar. For bridge inspections, it shows scour holes, debris, and settling around piers. How is multibeam different from side-scan sonar? Multibeam creates a detailed 3D map of the riverbed, giving you elevations. Side-scan gives you an image of texture and objects but no direct depth. Both have their place under a bridge. What accuracy should I expect from a multibeam bathymetric survey? With proper calibration, ±0.3 feet vertically is common. But you must know the tide, sound velocity, and the unit's heave corrections. If they don't report residuals, ask for them. How do I interpret the colored maps in a bathymetry deliverable? Red typically means shallow, blue means deep, but don't rely on the color scale alone. Check the numerical values in the points and the vertical exaggeration. A subtle 2-foot depression might look huge on the screen—or invisible.