[Logged] Recommended Changes to the Sonar Model

[Ping Jockey]

I’m a former US Navy Sonar Technician who bought CMANO the first week it was released and have been enjoying it and CMO ever since. Great job overall – BZ to the entire development team. [&o]
Early on I noticed a couple of fundamental problems with the modeling/rules associated with sonar detection; specifically layer modeling and convergence zones (CZs). Sorry it’s taken me so long to propose corrections (truth to tell, I was hoping someone else would do it <g> ).

I understand that the current sonar model in CMANO / CMO does not include environmental factors such as weather, sea state (as it relates to mixing in the near-surface layer) or – most importantly - seasonal variations. This is unfortunate, but I realize that implementation of these factors would be very difficult – more so than modeling worldwide terrain, since mountains do not move and change height on a seasonal basis. Consequently, I restrict my recommendations to modifications to the existing static environmental models.

Layer Strength

The difference between the temperature of the near-surface mixed layer and the ocean deeps determines the strength of the acoustic layer. The greater the difference, the stronger the layer. The ocean deeps keep a fairly constant temperature of ~39° F (~4° C), thus a warmer mixed near-surface layer gives a stronger acoustic layer effect.
Command elegantly models this by indexing the layer strength to latitude on the map. The strongest layer (0.7) is at the equator, and every 18° change in latitude reduces the layer strength by 0.1 until it gets to 0.2 at the North Pole (in CMANO, it was 0.01 per 1.8° (1° 48’)).

Note: The model actually uses an 18° band at the equator that is split between the northern & southern hemispheres and thus is 9° on each side of the line. Similarly, the radius of the 0.2 layer strength at the pole is 9° to give it a diameter of 18°. Thus the current layer strength bands are:

81°N – 0.2
63° N – 0.3
45° N – 0.4
27° N – 0.5
9° N – 0.6
0° 00’ – 0.7
9° S – 0.6
27° S – 0.5
45° S – 0.4
63° N – 0.3
… to Antarctica

The Problem
Although near-surface mixed layers exist in polar waters (where there isn’t pack ice) above 60° - 70° latitude, the temperature of these layers is usually close to - or even colder than – the deep water. Without the warm-over-cold thermocline, you don’t get the shadow zones “below the layer” in which a submarine can hide. To put it simply, the acoustic layer doesn’t exist in far northern or southern waters.

Why it’s Important
The Soviet Union put a lot of effort into building large ASW ships – Kresta-II, Kara, Udaloy and even Kiev classes; at first to attack US Polaris missile-launching SSBNs, then later to defend Soviet SSBN bastions when improved SLBM ranges moved enemy SSBN away from – and allowed Soviet SSBNs to patrol closer to – Soviet home waters in the Barents Sea.
Although surface ships are usually at a strong acoustic disadvantage against submarines, the layer-free “terrain” of northern waters helps to mitigate this, and the Soviets/Russians built their ships and tactical doctrine accordingly. This affects ASW scenarios in what would likely be a hotly-contested area in a general war.

How to Fix It
North of the equator, change the latitude increment in which the layer strength changes by 0.1 from 18° to 10°.
South of the equator, change the latitude increment in which the layer strength changes by 0.1 from 18° to 9°.

70° N – 0.0
60° N – 0.1
50° N – 0.2
40° N – 0.3
30° N – 0.4
20° N – 0.5
10° N – 0.6
0° 00’ – 0.7
9° S – 0.6
18° S – 0.5
27° S – 0.4
36° S – 0.3
45° S – 0.2
54° S – 0.1
63° S – 0.0

Convergence Zones

Section 9.2.3 (page 238) of the CMO manual (and 8.7 (page 134) of the CMANO manual) states,

"CZ detections are possible only if the local depth provides at least 600ft/200m clearance under the target."

and

"CZ intervals range from 40nm in the poles to 20nm in the equator, depending also on local temperature."

The bad news is that both statements are almost completely wrong.

The good news is that both should be easily fixable (based on my understanding of Command’s sonar model – I can hear the programmers’ eyes rolling from where I sit). When correctly modeled, the statements should read,

"CZ detections are possible only if the water depth is at least 1700 – 4600m (5600 – 15100ft), depending on local water temperature near the surface. This is modeled as a function of the sensor’s latitude above or below the equator."

and

"CZ intervals range from 20nm near the Arctic/Antarctic circle to 40nm at the equator, depending again on local water temperature. CZs are not available from 70-degrees latitude to the poles."

Correcting the CZ parameters is slightly more complicated ([8|]) than correcting the layer because in addition to correcting the CZ ranges, we are changing the minimum water depth figure for CZ availability from a constant to a variable (note that the actual minimum depth is MUCH deeper than is currently modeled). Also, since the presence of Convergence Zones depends on the presence of a surface layer (in which the negative sound gradient focuses – i.e. “converges” the sound), if there is no Layer (or it is very weak), there is no CZ. As above, there can be no Convergence Zones in Arctic or Antarctic waters.

Currently (as with Layer Depth), CZ ranges are indexed to latitude on the map. The longest range to the first CZ (40nm) is close to the North Pole (above 87° 45’N), and every 4.5° (4° 30’) change in latitude reduces the first CZ range by 1nm until it gets to 20nm near the Equator (2° 15’N to 2° 15’S).

To more accurately model Convergence Zone ranges and availability, set the maximum CZ range (40nm) at the Equator. Set the minimum depth for CZs at the Equator to 5380m. North of the Equator, the latitude interval should be 3.4° (3° 24’) and south of the Equator it should be 2.9° (2° 54’).

For each interval north or south of the Equator, reduce the CZ range by 1nm and the minimum depth allowable by 220m.

Latitude 1stCZ 2ndCZ 3rdCZ 4thCZ Min Depth (m)
69° 42’ No CZ No CZ No CZ No CZ ---- n/a
66° 18’---20-----40-----60-----80------980
62° 54’---21-----42-----63-----84-----1200
59° 30’---22-----44-----66-----88-----1420
56° 06’---23-----46-----69-----92-----1640
*
*
*
8° 30’----37-----74----111----148-----4720
5° 06’----38-----76----114----152-----4940
1° 42’----39-----78----117----156-----5160
0° 00’----40-----80----120----160-----5380
1° 27’----39-----78----117----156-----5160
4° 21’----38-----76----114----152-----4940
7° 15’----37-----74----111----148-----4720
*
*
*
47° 51’---23-----46-----69-----92-----1640
50° 45’---22-----44-----66-----88-----1420
53° 39’---21-----42-----63-----84-----1200
56° 33’---20-----40-----60-----80------980
59° 27’ No CZ No CZ No CZ No CZ ---- n/a

Like the rest of the sonar model, this is a linear simplification of a curved function, but it is 95% accurate to within 100m of typical depth minimums.

Thank you for your hard work continued dedication to making this the best modern war sim ever.

References:

Urick, Robert J. “Principles of Underwater Sound” 3rd Edition Peninsula Publishing 1983


Urick, R. J. “Sound Propagation in the Sea” DARPA U.S. Government Printing Office 1979

Cle´ment de Boyer Monte´gut, Gurvan Madec, Albert S. Fischer, Alban Lazar,
and Daniele Iudicone, “Mixed layer depth over the global ocean: An examination of profile data and a profile-based climatology” Published in the Journal of Geophysical Research, Vol. 109, 2004

[SSN754planker]

Good post. It's like I was time warped back into a classroom at Groton. Some subtle changes in CZ's may up the realism. Got people talking for sure.

[DWReese]

Excellent analysis. Thanks for sharing that.

Doug

[ultradave]

This is what I love about this forum. I learn new stuff constantly! Thanks for posting this.
Dave

[BDukes]

Nice info!

Mike

[Ping Jockey]

ORIGINAL: SSN754planker

Good post. It's like I was time warped back into a classroom at Groton. Some subtle changes in CZ's may up the realism. Got people talking for sure.

Yeah, it took me back a few decades!

For the minimum depth "problem" I made a graph to compare the current CMO model versus interpolated data from graphs in the standard references and the proposed change I recommended above:

[Dimitris]

Hi Rick,

Thanks a lot both for your very kind words and your suggestions. This kind of constructive critique is precisely the feedback we need in order to make this product even better, both as a realistic simulation and as a rewarding game.

Just as a very slight correction, the current sonar model does take the weather & environment into account in some of the calculations, for example on surface ducting (as the wind & sea-state increases, surface ducting becomes increasingly less effective). The bottom depth, and proximity to the polar ice packs, also have a huge effect on the reverberation modifiers.

We are discussing internally how to best adjust the model parameters to adapt to the information you have provided, and also some other recent research findings.

[Dimitris]

This may be a good opportunity to list some other factors related to underwater warfare and under consideration. We have been discussing these with various SMEs and have been trying to prioritize development as well as try to understand the relative after-affects (as you are probably familiar from your Harpoon days, often the hardest part is to fix/adjust things that a change breaks, rather than the change itself).

You will notice that some of these are "soft" factors that have little to do with the physics of sound transmission. As you no doubt know from your own experience, tech and physics are only part of the story.

* A "softer"/fuzzier/ more variable sonar FOM value. Currently the bulk of our sonar logic WRT detection range revolves around taking nominal "perfect day" values from the DB, applying various modifiers to them and arriving at a figure-of-merit value that is used as the "practical detection range". To our knowledge this is also standard practice by sonar operators, done e.g. on a daily basis. The problem is the currently this FOM figure is too "sharp"/absolute, for example if the figure is 12.3 nm then at 12.2 nm there will always be a detection and at 12.4 there will never be one. Obviously in the real sea environment this is not so clear-cut, as regular fluctuations in the environment (salinity, local temperature, currents etc.) can continuously shift this threshold either way. So we are considering to incorporate a "+/-" factor in the calculated FOM that will reflect this. The FOM figure could also be further modifier by operator proficiency and alert level (see below).

* Improve ambient background noise: While currently we do model the effect of passive sonar masking by platforms (e.g. a sub can hide from a passive listener by aligning with a noisy tanker down the same bearing), we do not model the increased ambient noise level borne out from the presence of multiple noisy platforms (e.g. inside the inner ring of a carrier group, even without direct masking the increased ambient noise level makes it much harder to detect a submarine passively). Need to properly represent this.

* Submarine "Tactical"/best-silent speed. Currently the machinery/propeller and flow noise generated by a submarine or ship rises proportionally (though not linearly) to the shaft (desired) and actual speed respectively. Currently we do not explicitly model the fact that noise actually rises much more gradually up to a certain "tactical"/best-silent speed (which is distinct and different on each sub class), and sharply beyond that.

* Platform variable alertness levels. Currently Command models manned platforms (incl. ships/subs) as being continuously on a high state of alert, and all "soft"-factor mechanics (incl. reaction times) assume this posture. In reality most platforms cannot maintain a high alert for a prolonged duration due to accumulated fatigue. Platforms at a lower alert level suffer disadvantages not only in reaction time and damage control (ex: carrier Shinano) but also in man-in-loop sensor performance (e.g. sonar FOM). Need to model the different alert levels and the fatigue buildup & release mechanics that govern them, together with the measures force groups undertake to mitigate them (e.g. rotating alert/watch shifts).

* Noisemakers & decoys as "real" objects. Currently decoys & noisemakers (and tactical mobile decoys like MOSS, Korund etc.) are handled abstractly in the point-defence step just like all soft-kill countermeasures, so they do not instantiate as distinct entities. In RL both decoys and noisemakers can stay present for a significant time, both confusing ASW weapons but also contributing to ambient noise increase and direct masking. It is therefore desirable to implement such countermeasures as distinct entities fired just like weapons.

* Aircraft/helos heard at low altitude. Currently in Command subs cannot hear low-alt aircraft or helos, so they often get surprised by sonobuoys (and subsequently air-dropped torpedoes). AFAIK IRL subs with sensitive sonars can frequently hear such noise sources and maneuver to reduce the chance of detection.

* Combat system capacity. A ship/sub with 50s/60s electronics & combat system is not going to have the same contact handling capacity as a modern platform. Currently Command does not represent this as a platform is able to perform detailed tracking, classification/ID-ing and TMA on an unlimited number of contacts regardless of its tech level. Need to better represent the different capacities in this (this BTW could lead to some serious AI headaches: How should the virtual crew decide which contacts to prioritize for attention?).

* Auto-generated false contacts. One of the problems Command currently has with ASW ops is that each underwater contact is quite likely to be a legit target (even a friendly one), because most scen authors neglect to generously pepper the area of interest with false contacts and/or biologicals. One possible way to remedy this is to auto-generate such false targets in the vicinity of each potential ASW asset, so that the virtual sonar operators (and TAOs) are faced with a similar "which of those 100 blips should I give priority to?" conundrum as their RL counterparts - thus making evasion significantly easier for submarines.

* Bistatic/Multistatic sonar systems. This would allow modelling some advanced systems that are currently operational, including EER sonobuoys.

* Do not use sonobuoys for initial search: Aircraft currently are allowed to pre-emptively (no initial contact) saturate a search area with sonobuoys as part of their search. This is problematic for two reasons: (a) It does not reflect the fact that sonobuoys are a rather scarce resource and are thus usually used only for localization of an existing initial contact, and (b) because of their tight spacing, they make it much more likely to detect a sub in the area being sowed with them. Therefore need to add an (optional?) setting that restricts aircraft from using buoys on general search.

[thewood1]

"* Auto-generated false contacts. One of the problems Command currently has with ASW ops is that each underwater contact is quite likely to be a legit target (even a friendly one), because most scen authors neglect to generously pepper the area of interest with false contacts and/or biologicals. One possible way to remedy this is to auto-generate such false targets in the vicinity of each potential ASW asset, so that the virtual sonar operators (and TAOs) are faced with a similar "which of those 100 blips should I give priority to?" conundrum as their RL counterparts - thus making evasion significantly easier for submarines."

Aren't the false contacts a matter of systems and operator levels? Wouldn't the best tech and operator have fewer false contacts and older/poorer tech and inexperienced operators generate more. As a commander, don't you depend on the operator to filter those out for you? I would think an abstraction would be called for with better equipment and operators showing fewer false contacts.

"* Do not use sonobuoys for initial search: Aircraft currently are allowed to pre-emptively (no initial contact) saturate a search area with sonobuoys as part of their search. This is problematic for two reasons: (a) It does not reflect the fact that sonobuoys are a rather scarce resource and are thus usually used only for localization of an existing initial contact, and (b) because of their tight spacing, they make it much more likely to detect a sub in the area being sowed with them. Therefore need to add an (optional?) setting that restricts aircraft from using buoys on general search."

Why is there a disconnect between RL and CMO approach to sonobouys? Should you punish the player for using them differently then in real life? I'm just curious what the mechanism is that is making CMO players use them differently than a real life commander. Is it just limited assets in real life? Should the player have fewer devices to deploy?

[p1t1o]

ORIGINAL: Dimitris

* Aircraft/helos heard at low altitude. Currently in Command subs cannot hear low-alt aircraft or helos, so they often get surprised by sonobuoys (and subsequently air-dropped torpedoes). AFAIK IRL subs with sensitive sonars can frequently hear such noise sources and maneuver to reduce the chance of detection.

I did just see one of my subs detect a helicopter whilst at "shallow" (-131ft) via "acoustic intercept". I remember checking because it is unusual for a sub to get an air contact. Is this a distinct mechanic?

[Dimitris]

ORIGINAL: p1t1o

ORIGINAL: Dimitris
* Aircraft/helos heard at low altitude. Currently in Command subs cannot hear low-alt aircraft or helos, so they often get surprised by sonobuoys (and subsequently air-dropped torpedoes). AFAIK IRL subs with sensitive sonars can frequently hear such noise sources and maneuver to reduce the chance of detection.

I did just see one of my subs detect a helicopter whilst at "shallow" (-131ft) via "acoustic intercept". I remember checking because it is unusual for a sub to get an air contact. Is this a distinct mechanic?

It must have detected its dipping sonar.

[c3k]

Great information. I realize that I don't even know what I don't know.

[Parel803]

Nice info, thank you.
with regards GJ

[WSBot]

0014439

[p1t1o]

ORIGINAL: Dimitris

ORIGINAL: p1t1o

ORIGINAL: Dimitris
* Aircraft/helos heard at low altitude. Currently in Command subs cannot hear low-alt aircraft or helos, so they often get surprised by sonobuoys (and subsequently air-dropped torpedoes). AFAIK IRL subs with sensitive sonars can frequently hear such noise sources and maneuver to reduce the chance of detection.

I did just see one of my subs detect a helicopter whilst at "shallow" (-131ft) via "acoustic intercept". I remember checking because it is unusual for a sub to get an air contact. Is this a distinct mechanic?

It must have detected its dipping sonar.

Checked and confirmed - it was the dipping sonar, spot on.

[Dimitris]

One question for clarification, if anyone is familiar with this.

Currently, an assumption on our sonar model is that the "deep sound channel" depth band is located right under the thermal layer, and because sound waves generated in this depth range are trapped between the layer above and the higher water pressure below, propagation is excellent in this depth band. As a result, if both the noise emitter and the receiver are within this band, we grant a 200% boost to the effective detection range. (This is why it is unwise to hang in this depth band if facing adversaries with VDS or towed arrays).

However, this assumption of "trapping" holds true only if there is a strong layer present. The weaker the layer, presumably the less sound waves will actually be trapped within this conduit for propagation. With this in mind, perhaps a more accurate modifier for detection range boost within the DSC would be "100% + [percent strength of the layer]" ?

[Ping Jockey]

Dimitris,
Broadly speaking, you are correct. I am away from my references right now, but will be able to give a more quantified answer this weekend.

[Dimitris]

OP changes implemented, currently in testing.

[ultradave]

^^^^ This is what I love about this game. The developers are always responsive to good, new, info. That was fast!

Dave

[cmanouser1]

ORIGINAL: Dimitris

OP changes implemented, currently in testing.

As we're talking about submarine changes, could we get a "Inside Layer" depth setting [&o].
It would be so useful instead of having to micromanage the submarine in fear the layer moves.