Dolphins determine their surroundings by listening to echoes. Can they mimic those echoes and communicate it to other dolphins, transmitting an "image"? Does the same go for all echolocators?
As I understand your question, dolphins (and other echolators) can more or less map out their "surroundings" (which might include nearby ice floes, for example). It sounds like you're asking if they can then transmit that knowledge to other members of their species.
I don't know the answer, and I suspect no one does. However, it's hard to imagine why one dolphin would need to tell another dolphin about its surroundings when the second dolphin ought to be able to map its surroundings all by itself.
Cetaceans do communicate with each other, presumably alerting each other to danger (e.g. predators) and the presence of prey.
Incidentally, I worked on an acoustic study of bowhead whales in the Arctic Ocean years ago. We recorded bowheads and Arctic seals, which make some unbelievable sounds. We operated under the presumption that they were primarily "sounding out" their environment - searching for open areas in the ice where they could breathe, for example. It's possible that one whale might say "Hey, I found a breathing hole!" without mimicking the echoes that alerted it to that hole.
Sorry, that's not really a direct answer, but it might help put things in perspective.
I have just asked a similar question in scifi.stackexchange and the reason is, I believe your idea can be found in a book written perhaps 60 years ago. I asked who the author was and/or the name of the book.
I am pretty sure this has been substantiated by researchers -- not recalling how the experiment was set up but I think it essentially involved simulating the dolphin echolocation process and then comparing sounds the dolphin emitted when shown the object or perhaps the researchers produced the sounds for a dolphin and asked the dolphin the choose from among objects.
So while a human might draw an object to explain to another human what they wanted, etc., a dolphin could do the same thing with sound. Perhaps their nouns all resemble the signals that come back during echolocation but I would also bet they abbreviate them since their probably is a lot of excess information in the signal from an actual object.
They would have the natural idea of "instance" and "class" (I speculate). And if cetacean language does consist of abbreviated signals from objects, I do not wonder why they have such large brains.
Here is a link that seems to indicate the answer to your question is "yes": https://www.cymascope.com/cetacean.html