Assuming I am remembering correctly, the issue is actually the oscillations of water. It's similar to why a vibration detector can more accurately pick up sound off a vibrating brick wall than through water, and is why water stops bullets better than concrete of equal mass. The water cavitates on impact, creating it's own vibrations and wavelengths, which are MUCH BIGGER and slower than ones through air. They also go further. This is why sound carries better in water.
However, the trick then is distortion. RADAR distorts through water much more severely than water because of it's higher frequency. This is a similar principle to why the bullet stops, actually, because the path distorts quickly, causing loss of momentum through directional changes in water movement and pressure, which has much more drag and weight than air. Basically, RADAR doesn't work in water because what gets reflected back is inconsistent garbage sound.
sourcelinkQUOTE
While radar would eventually be capable of detecting submarines on the surface, the required radio frequencies are quickly attenuated in water, making this technique ineffective for detecting submerged submarines.
And from the document previously linked:
The reason [why radar is not used to find subs] is mainly because radar has a harder time penetrating large volumes of water. Contacts made by submarines are often dozens of miles away, and radar would have to be EXTREMELY powerful to reach that far in water, while sound (a mechanical wave) can make it that far.
GPR is another matter, but keep this in mind: GPR can be made to go further with LOWER frequency (coming closer to sonar in wavelength) but the
detail of the image (which you'd need to find fish smaller than whales) goes to shit. So here is what I think for fish.
Ground Penetrating Radar comes closest to the shadowrun rating 4 radar (based on your comparison for the bunker) and so I'd use it as a comparison. It is known to inaccurately image freshwater lakes. Assuming shadowruns version is futuristic-ally better (GPR still sucks at penetrating water) I would say that it would have a barrier rating of 1 per 2-5 foot depth of standing
fresh water.GPR works less accurately the more electrically conductive a material is, in addition to attenuation problems. Salt water is MORE electrically conductive than fresh (lemme head arguments off at the past here: look it up first if you disagree) and GPR doesn't penetrate seawater with nearly the same depth as fresh. I would suggest a barrier rating for salt water along the lines of the numbers given previously, maybe less: 4 per 2-5 foot depth of seawater. So a building surrounded by a saltwater aquarium, with grounded water tanks, would be MUCH harder to penetrate than your previously mentioned bunker, and this is in keeping with current technical norms.
Since software correction plays a big part, GMs make want to use programming as well, depending of your depth of detail
sourcelinkQUOTE
The most significant performance limitation of GPR is in high-conductivity materials such as clay soils and soils that are salt contaminated. Performance is also limited by signal scattering in heterogeneous conditions (e.g. rocky soils).
Other disadvantages of currently available GPR systems include:
Interpretation of radargrams is generally non-intuitive to the novice.
Considerable expertise is necessary to effectively design, conduct, and interpret GPR surveys.
Relatively high energy consumption can be problematic for extensive field surveys.
Recent advances in GPR hardware and software have done much to ameliorate these disadvantages, and further improvement can be expected with ongoing development.
TL:DR version - 1 per 2-5 feet fresh water. 4 per 2-5 feet salt water, for finding fish smaller than 50 pounds. Submarine sized images would allow for further depth, obviously.