Key facts
| Quantity | Value | Fishing consequence |
|---|---|---|
| Facies classes | 4 | Rock, gravelly muddy sand, muddy sand, mud |
| Samples checked | 60 | Surface samples used as ground truth |
| Correct predictions | 43 of 60, or 72 percent | The map is a hypothesis, not a survey |
| Misses | 17 | Each one class away |
| Short Bank depth | 40–110 m | Rock mapped in deep water |
| Outcrop relief | 12 m | A rock edge is a vertical face |
| Backscatter resolution | Metre horizontal, centimetre vertical | Outlines patches, not texture |
Four classes, one decision tree
The acoustic class is the first useful split, and it is a hypothesis until something touches the bottom. A USGS benthic-habitat map of central Santa Monica Bay classified the seafloor into four facies: rock, gravelly muddy sand, muddy sand and mud. The map was built as a decision tree from acoustic data, then checked against 60 surface samples. It matched 43 of them, or 72 percent. The four classes are not four equal boxes. Rock is a substrate, gravelly muddy sand is a mixture, muddy sand is a mixture with more mud, and mud is the fine end. The decision tree put them in that order because their acoustic signatures differ in a consistent direction. Hard, rough ground returns more sound and a sharper pulse; soft, smooth ground returns less and a longer, fuzzier echo. The classification is a prediction. The acoustic data says where the bottom changes. The samples say what the change is. A fishing boat with a single-beam sounder is doing the same thing with less data: proposing a class from one number.
What a 72 percent match rate means
A 72 percent match is good for a map built without touching every acre, but it is not a promise that every pixel is correct. The useful detail is the shape of the errors. None of the 17 misses jumped two classes. A pixel mapped as gravelly muddy sand came out as rock or muddy sand, not as mud. Every error was one class away. That pattern says the boundaries are gradational. Sediment does not switch from rock to mud at a line. It passes through gravel, sand and mixed material over a distance that the map compresses into a colour change. The practical reading is to treat the map as telling you which two classes you are between, not which one you are on. Mud and sandy mud were not statistically distinguishable in the survey. That is the pair to treat as a range. A soft return in a muddy area is a soft bottom of uncertain grain, and the rig should be chosen for the soft end rather than the firm one.
Backscatter measures two things at once
Backscatter is the strength of sound returned from the bed. Hard, smooth bottoms return more sound than soft or bumpy ones. Two properties drive the number: acoustic impedance, which is set by density and sound speed, and roughness, which scatters the pulse in directions that miss the transducer. Those two properties do not always move together. A smooth rock face returns a very strong echo. A rough rock face returns less, because the pulse scatters. A smooth mud flat returns a weak but coherent echo. A gravel bed returns a strong echo because it is hard, and a noisy one because it is rough. The single number on the display mixes hardness and texture, which is why a bright return cannot name the sediment. Acoustic impedance is density multiplied by sound speed. Sound travels at about 1,500 metres per second in seawater, roughly 1,650 to 1,800 metres per second in saturated sand, and several thousand metres per second in solid rock. The reflection coefficient at the bed is the difference between the two impedances divided by their sum, so a large contrast returns a large echo and a small one returns almost nothing. Mud and water are close in both density and sound speed, which is why a mud bottom can be nearly transparent to a sounder. Rock and water are far apart, which is why a pinnacle shows as a hard, bright line even at low transmit power. The number on the display is a decibel value derived from that echo, not a measurement of grain size. Multibeam systems resolve metre-scale horizontal features and centimetre-scale vertical ones. That is enough to outline a reef, a channel edge or a gravel patch. It is not enough to separate muddy sand from mud without a sample, and it is not enough to tell a smooth rock slab from packed sand.
Reading each bottom
Rock gives very high backscatter and high relief. At Short Bank, the mapped rock sits in 40 to 110 metres of water, and the outcrops rise as much as 12 metres above the surrounding sediment. That relief is what makes the class useful. A rock edge is a vertical face, and fish relate to the face rather than to the acreage on top. Gravel is hard and rough, and it sits between rock and sand in both return strength and stability. Gravel patches are often thin, so one pass shows a strong return and the next shows almost nothing. The edges are the better mark, where loose stones meet sand and the current sorts the boundary. The classes map onto grain size, and the boundaries explain why a bottom can change character over a short distance. Gravel begins at two millimetres; sand runs from two millimetres down to 0.0625 millimetres; silt and clay, which together read as mud, sit below that. A bed of coarse sand and fine gravel therefore behaves like both classes, holding a jighead well enough to fish but shifting enough to bury it after a storm. Sorting matters as much as size. A well-sorted sand returns a clean, even echo, while a mixed bed of shell, gravel and sand scatters the pulse and produces the broken, noisy trace that anglers read as rough ground. Sand is smooth and returns a moderate, steady echo. It is the most common bottom in the bay map and the one most often misread as gravel when it is coarse or shelly. Sand holds fish that sit low and feed along the surface, and it is the easiest bottom on which to feel what the lure is doing. Mud is soft, smooth and returns little sound. It is the class that best rewards a slow presentation, because a bait that sinks into it disappears. A soft return in a muddy area is a range rather than a fixed type, and the rig should be set for the soft end.
Rig and presentation by bottom
Rock calls for a vertical approach. A heavy jig or a metal lure dropped along the face keeps contact with the structure and can be lifted quickly over the top. The strike usually comes as the lure falls or just after the lift, so the rod should be ready at the end of the drop. A short leader reduces the chance of the hook finding the rock before the fish does. Gravel rewards bottom contact. A jighead bumped along the stones, or a bait on a light Texas rig, keeps the offering in the layer where fish feed. The risk is snagging between stones, so a steady lift-and-drop beats a long drag. Gravel edges are the better mark: the transition to sand concentrates bait and gives fish a reason to sit there. Sand is the most forgiving class. A lighter jighead, a soft plastic or a bait on a simple sinker can all be worked with a slow lift and a controlled fall. The aim is to keep the lure just above the surface, not buried in it. On clean sand, fish often follow a long way before taking, so a steady retrieve with pauses reads better than constant hopping. Mud wants a float or a scent trail. A bait suspended just above the soft layer stays findable, while a lead that sinks in leaves a smell but no target. Use the lightest weight that holds and move it slowly. The bottom tells you when it is too heavy.
Ground-truth with a lead
A lead is the cheapest sampler. A lead that rings and stops is rock or packed gravel. One that thuds and sinks is sand. One that sinks slowly and brings up grey paste is mud. Ten seconds of sampling beats ten minutes of guessing, and it tells you whether the class in your head matches the bottom under the boat. A chart gives you the class at metre scale. The lead gives you the class under the keel. Where the two disagree, believe the lead and note the boundary. The disagreement is information: it marks a transition that the map smoothed away, and transitions are where the fish usually are. Sample more than once. A single drop on a patch of shell can read as rock, and a single drop in a weed bed can read as mud. Three drops across twenty metres will tell you whether you are on a patch or on a boundary.
Where the acoustic class fails
The acoustic class fails at three edges. The first is a thin veneer: a layer of sand over rock, or shell over mud, returns the signature of the top layer. The second is weed: a canopy of kelp or seagrass returns a soft, fuzzy signal that reads like mud. The third is a slope: a steep face mixes returns from several depths, and the class becomes an average of the face. The Santa Monica map handled the first problem by treating the classes as gradational and by checking against samples. A fishing sounder has no such check unless the angler supplies it. That is why the lead, the jig and the first drop matter more than the colour palette. Depth also changes the reading. A soft bottom in deep water returns less than the same bottom in shallow water, because the pulse spreads and attenuates. Compare traces at similar depths, or compare the same line at different states of the tide rather than across a depth range.
Sample, then commit
Read the class as a range. Treat every mapped patch as the boundary between two classes until a sample says otherwise. Drop a lead before rigging. Ring, thud or paste tells you which of the four bottoms is under the boat. Fish the transition. The edge between hard and soft concentrates bait and gives fish a reason to hold. Match the rig to the bottom, not the species. A rock rig on mud fishes too fast; a mud rig on rock loses gear. Check the weed. A fuzzy return under a canopy is not necessarily soft sediment. Re-sample after a storm. Sand moves, gravel patches thin and mud spreads; last season's class is a guess.
Sources
- Predicted Seafloor Facies of Central Santa Monica Bay, California — U.S. Geological Survey · T1
- How does backscatter help us understand the sea floor? — NOAA Ocean Service · T1
- What is multibeam sonar? — NOAA Ocean Service · T1