Jump to content

HCDB ship RCS review / calibrating surface detection ranges

Recommended Posts

This discussion branched off from the EW hands on Red flag topic into a full blown DB discussion, so let me copy it over here, the proper place. The bonus is that Enrique's post from 2012 is just few topics away from here, that one is probably the Philosopher's Stone of RCS so far!!

(If forum admin could replace this post with posts #17-#31 from the "EW hands on: Red Flag - Russians invited" topic in General that would be even better.)

(( Damn, forum doesn't allow me to just quote the other posts, too many quotes ..., attempt to copy them unquoted.))

Here it is:


Grumble, on 31 May 2014 - 02:13 AM, said:

I just played my first scenarios with the Westpac battleset and the new radar model got me by surprise, enough to have me return to this EW topic. Though I've read Brad's AEW&C / AWACS performance testing some time ago, I must have been holding onto some Sauronistic faith of the "all seeing eye". :blink::)


So I went back to Red Flag for testing.

Westpac E2C AEW ranges vs. surface combatants:



Shocking. A Dergach can get a cheap shot with it's Sunburn missiles and even the Hawkeye will not detect it (before firing), shipborne radars are even worse off.

:o 38nm really? Any views? Understandably there is not much data around but for example "The Naval Institute Guide to World Naval Weapons Systems, 1997-1998" has that the "APS-138 can reportedly track a cruise missile at a range of 150nm". The Dergach is small ship, but it's still 200 feet of metal. And the Group II has the APS-145, digital and all. 38? Need to rethink carrier defense.

What do you do to protect your ships from missile boats?


The new AESA radars are also cut to size by RCS:

Westpac APG-81 AESA ranges vs. surface ships


No bonus for the AESA. The Dergach might get a visual sooner than radar detection.



TonyE, on 31 May 2014 - 05:16 AM, said:

Dergach doesn't look particularly stealthy to me. Methinks you have the beginnings of a case to modify a lot of ship RCS values in the database. You've shown a few examples of perhaps too short detection ranges. At what range do you expect detection of those platforms? Is there a general pattern across all ships (or within one or more types of ships) that would be handy for Brad (i.e. all Frigates need their RCS values increased by 4)?



CV32, on 31 May 2014 - 3:15 PM, said:

The RCS values are mostly guesswork based on comparisons with the size, length, shape and stealth features (if any) of other ships. This is probably just one that needs adjustment.



Grumble, on 01 Jun 2014 - 12:10 AM, said:


Tough question, as my expectations are based on the legacy battlesets :) and not much more. My gut feeling was "more than 100nm", but I'd search for data first. I found just one useful looking source (referenced many places), an RCS Ship Table at: www.mar-it.de itself based on "Williams/Cramp/Curtis: Experimental study of the radar cross section of maritime targets, 1978".

(Are you aware of any other source for experimental RCS values?)


This table lists a good range of commercial vessels and one "warship". Apparently "warship" has higher relative RCS (lot of masts, turrets, etc., reasonable) and the other difference that warships are rated by "displacement" while commercial ships use "gross tonnage" (good reading this one), no conversion, great :).

Still length can be used to correlate these values to ours I guess.

Fit a function to it and take a closer look at the DB where there is extreme deviation vs DB RCS.


After some crunching and guessing and cleaning the base table is this:






CV32, on 01 Jun 2014 - 3:30 PM, said:

These are all factors already in play for determining the RCS value to be used.


At the rate which platforms were added, however, there has no doubt been some degradation of the accuracy of the relative values. Again, its very much relative. Because two "warships" are no more equal than two "container ships".


As for AESA and other peculiarities, keep in mind that we have a single radar model. There is no accounting for monopulse, pulse Doppler, synthetic aperture, etc, etc. Their range values are drawn from H4 values where possible, and if not, then extrapolated or drawn from other sources.



Grumble, on 01 Jun 2014 - 11:51 PM, said:

True I'm aware, and without actual facts I can only data mine the DB for relative mismatches, but even the out of place values might just mean that that ship is special.

Ok, before I waste any more time would you check these if this adds any value:

I looked at ships grouped by subtype (CV,CG,DD,FF,PT,...) hoping that there is more RCS correlation among similar purpose vessels. Started with Frigates, sorted them by length and looked for out of series RCS values. Then checked the DB description for any mention of stealth.

The first nuggets:


"RCS" is the RCS from the DB

"RCS by Len" is the avarage RCS of similar length ships.


Is any of these a valid correction?


broncepulido, on 02 Jun 2014 - 07:58 AM, said:

I think ship's RCS must at least be considered basically as function of lenght, wide and height (as HCE models not the different angles of radar incidence against the radar target, i.e., frontal or lateral at least). I use lenght x wide halved, plus height.


Edited: and later I use a table, of course. And as ship's height is a value usually not reflected on books, I use the generic height values of 19, 29 and 40 meters employed in the DB.


broncepulido, on 02 Jun 2014 - 8:57 PM, said:

About modelling AESA and PESA radars an idea can come from this:

- PESA radars increased his range some 50% in relation with a conventional radar of the same power.

- AESA radars increased his range some 100% in relation with a conventional radar of the same power.



But those increased radars ranges are implemented in the radar ranges reflected on the DB, the "increased PESA/AESA range" would be a plus on those previous ranges.



broncepulido, on 02 Jun 2014 - 9:02 PM, said:

And yes, me too for years was thinking a "warship" has a greater RCS than a tonnage equivalent civilian warship, with a lot of radar antennas, turrets, weapons mounts and other things (and what about an aircraft carrier with the deck full of planes ?!).

Also, in the DBs aircrafts I add some penalty when the plane is equipped with search radars (+1, example, fighter search radar) or big search radars (+2, example, AWACS radar).


Edited: also, a ship with high "sides" or boards (example: oiler, merchant, civilian passenger cruise ship) must have a higher radar return.




broncepulido, on 02 Jun 2014 - 9:24 PM, said:

I like the idea of a Iranian PG Toragh/Boghammar with the correspondent "inshore fishing vessel" 142 RCS as reflected on the tables ("inshore fishing vessel" is almost the same size than PG Toragh), as opposite to the current 165 RCS, perhaps it's a solution to how to sneak with a small size boat near a conventional warship (but a think it requires a lot of testing).


Edited: also, a small boat gets more "cover" from sea's radar return clutter, more if tthe radar is tuned to bigger targets to avoid false returns because sea state.




TonyE, on 04 Jun 2014 - 04:41 AM, said:

Keep going guys, looking good so far. Brad is away for a week or so and may not be on the forum but I'm sure he'll jump back in when he returns.


I don't agree with the 142 RCS for a Boghammer (a max detect range of 9nm when using a 360nm ranged radar) but I appreciate the arguments.


In the Boghammer case I would look at vessels or sensors likely to be facing off against the Boghammer, then come up with a reasonable detection range, then work the RCS backwards from that. A starting place perhaps anyway.




donaldseadog, on 04 Jun 2014 - 09:52 AM, said:

Just to add some humour, and illustrate some problems in getting it perfect, I remember one time picking up three contacts very close to each other but some distance away, turned out to be a super tanker :)




broncepulido, on 04 Jun 2014 - 5:49 PM, said:

And don't forget:

- Máximum surface-to-surface radar range is 25-27 nm, because the radar horizon (remember, Earth is sphere-shaped :P ).

- If you active surface-to-surface radar, you will be picked by enemy ESM and fired upon! (as very usual case, the USSR/Russia Komar and Osa patrol missile boats are equipped with ESM, but not the export versions).

Link to comment
Share on other sites

So far we had no cornerstone to work with, just theory, i.e. what do we "expect" the detection range for a Dergach or other should be, but you just need to look the right place, the HG Database forum! ^_^ 2012 Enrique posted a link to a study of FLAR systems by the US Coast Guard, that has some great references.


The study is from 1984, evaluates the (then) new APS-134 FLAR vs the in use AN/APS-127, AN/APS-133, and AN/APN-215 systems.


Key information for the RCS calibration:

  • The APS-134 FLAR detected a target of 1m2 (meter square) radar cross section at ranges of 20.0 to 35.9 nm in sea states 1 through 3.
  • The APS-134 proved capable of initial target detection of 100 m2 targets at ranges of 74.0 to 80.5 nm in sea states I through 3.
  • During both experiments, the primary search objects were anchored, 13- to 18-foot open fiberglass boats without engines or other substantial metal equipment, similar fiberglass boats with a 5-foot wooden post and radar reflector, and 4- to 6-man canopied rubber/fabric life rafts with and without radar reflectors. On any given search day, four to six targets were set on two search legs as depicted in Figure 1-4.
  • For the SAR mission, targets 16 to 25 feet long were considered to be representative search objects. Targets of this type equate to an approximate radar cross section of 1m2.
  • For ELT, drug interdiction represents the most significant problem, with the primary target being "mother ships" 60 to 300 feet long. To represent these targets, an intermediate value for radar cross section of 100 m2 was selected.
And there is also a table (2-4) referencing RCS to length:
  • Target Cross Section
  • 1m2 Small targets (<20ft)
  • 50m2 40 to 60 foot targets
  • 100m2 ~100 foot targets
Also very interesting notes about the huge effect of sea state on surface detections!


My view is that this anchors the small Frigate detection well beyond the 100nm range, perhaps closer to 200 for a +300nm SS AEW.

Armed patrol boats are also in the +50nm range.

  • Like 1
Link to comment
Share on other sites

Understandably there is not much data around but for example "The Naval Institute Guide to World Naval Weapons Systems, 1997-1998" has that the "APS-138 can reportedly track a cruise missile at a range of 150nm".

Not all "cruise missiles" are created equal, obviously, and if anyone believes this applies across the board to the performance of the APS-138 (or just about any radar), I have some lovely beachfront property in Arizona to sell to you. :P

Link to comment
Share on other sites

I suppose that APS-138 data are relative to AS-4 or AS-6 flying high, hehehe

Still, the APS-138 data against the other targets in calm seas is giving you some data points. How about using those to make an educated statement about some of the existing RCS values in the DB?


Don't forget the Radar RCS workbook, http://harpgamer.com/harpforum/index.php?/files/file/28-hc-rcs-worksheet/

Link to comment
Share on other sites

I hinted at this earlier in the thread, said it again on IRC last night, and I guess I will say it again here.


We are working with a very simple model.


Don't get too hung up on trying to achieve 'perfect' or even operationally realistic values for things like radar performance.


I see mention of how warships might (in real life) be more radar reflective than a merchant of comparable size because of their antennae, guns, missile canisters, etc, etc. A fair point.


At the same time, however, we must also recognize that our model does not even take into account such things as target aspect, i.e. whether the target is facing head-on or sitting broad side to the sensor.


Given those realities, does the radar reflectivity of a gun turret now matter? I think not.


So there is considerable tail chasing going on here.


The DB has thousands of platforms. It is in a constant state of adjustment and will NEVER be 100% accurate, even if that were possible with the model we have. (Which most certainly is NOT).


Long story short ...


The thread is interesting, at least theoretically so, but has limited value to DB work for our purposes.

Link to comment
Share on other sites

Speaking of calibrating, also don't forget H4.1 Smarter Radar



Vow, great. We got two corners for the puzzle.


Do you know if the "Naval SITREP #17" mentioned in there is available on the web?

(Has it been published as a book or as a web page?)

Link to comment
Share on other sites


The thread is interesting, at least theoretically so, but has limited value to DB work for our purposes.


Brad, I take this as a challenge. (Just need to find the time for it.)



Heh, if you can find the improvements to the radar (and sensor) model generally, I'm all for it. ^_^


(And 'Smarter Radar' is already in play).

Link to comment
Share on other sites

Naval Sitrep 17 shouldn't be available for free. It also isn't on the wargame vault as yet:

http://www.wargamevault.com/browse.php?&keywords=naval sitrep&manufacturers_id=5404

There do appear to be a few physical copies for sale out on the Internet (e-Bay and some game resellers).

You could also ask Bond & Carlson directly for that particular article and they would likely send it to you for free.

Link to comment
Share on other sites


Ok, thanks for the look, but I since realized that the radar model will not help much with checking the RCS values which is what started this topic. There is actually an abundance of radar model and equation published out there, but practically no experimental RCS examples and just one empirical formula thanks to Mr. Skolnik.

Also for radar equation I can just stick to your RCS workbook (thanks!), this is anyway the best match for the GE model.

Link to comment
Share on other sites

Ok, so this is it. I start with expectation setting (I’m preparing for storm ^_^) So, keep in mind, this thread started when a Dergach FFL seemed to be too stealthy, it turned out that as per HCDB she has RCS 170, an E-2C grp2 Hawkeye, would detect it only inside 38nm. This does not “feel” right, but can we verify it. Well, no, not surprisingly there is not much factual data on the net about naval ship RCS, all I got is the HCDB. As second best I settle to check HCDB consistency across ship sizes and compare it with the few available data from outside.

References,what can I work with:

  • (1) An 1984 Coast Guard study testing the APS-134 FLAR against 1m2 to 100m2 RCS surface targets, also referencing these to 1m to 30-90m length.
  • (2) The Harpoon 4.1 Smarter Radar document listing expected radar detection ranges of most DB sensors against
    • Large 3,000,000 m2
    • Medium 300,000 m2
    • Small 30,000 m2
    • Very Small 3,000 m2
    • Stealthy 300 m2
  • Note: H4.1 seems to use ship m2 RCS values on a different scale from the Coast Guard study and also from Skolnik. (And both are on the m2 scale, so the difference is not due to dbsqm representation.) Will need to convert.
  • (3) An empirical formula from M. Skolnik that references radar frequency & displacement to naval ship RCS m2:
    rcs = 52f1/2D3/2
    where f is radar frequency in MHz and D is ship displacement in kilotons.
  • (4) Tony’s RCS worksheet (RCSW now on) which is practically the GE routines to calculate between RCS m2 <> RCS dbm (as in the HCDB) <> detection range (radar).
  • (5) HCDB-130924

Data scoop, the factual (like) data from the references:

  • (5) HCDB: APS-134 is 150nm SS radar
  • (5) HCDB: APS-145 is 360nm AS/SS radar
  • (1) APS-134 vs 1m2 target: 20.0 to 35.9 nm in sea states 1 through 3.
  • (1) APS-134 vs 100m2 target: initial target detection 74.0 to 80.5 nm in sea states I through 3.
  • (1) 1m2 RCS :: ships < non metallic 8m (25 feet) long (Targets 16 to 25 feet long were considered to be representative search objects. Targets of this type equate to an approximate radar cross section of 1m2)
  • (1) 100m2 RCS :: Ships 60 to 300 feet long. (To represent these targets, an intermediate value for radar cross section of 100 m2 was selected.)
  • (2) APS-134 performance vs Lrg/Med/Small/Vsmall/Stealthy 150/150/96/54/30
  • (2) APS-145 SS performance vs Lrg/Med/Small/Vsmall/Stealthy 250/195/111/63/35
  • (3) I'm going to use S band, 3000 MHz freq. for the Skolnik formula.

Note: HCDB stores RCS values of dbm2 like logarithmic scale (dubbed “HCRCS” in the RCSW), ranges between 160-250, for our “feels like” assessment I convert these to square meter RCS values, comparing ship length to m2 gives at least some “feeling” unlike comparing it to a logarithmic value. I use the RCSW formulas for this.

0) Initial checks, correlating (1) the Coast Guard study to HCDB and HC GE, per the RCSW:

  • for an APS-134 (150nm SS radar) to detect a ship target at 20nm the target has to have >500m2 RCS or >177 HCRCS.
  • for an APS-134 (150nm SS radar) to detect a ship target at 80nm the target has to have >100000m2 RCS or >219 HCRCS.

Now I create a function that calculates RCS for the ships based on other data available from the DB, I’m going to use displacement and length as input here.
My hope is that it will be possible to match this function to the DB values on average, discrepancies are either purpose made (like stealthy ships) or values that need to be corrected (my goal to find).

Excel number crunching:

  • Export HCDB ship annex into a excel
  • Convert HCRCS values to m2 values (HCRCSm2 values) for human comparison.
  • Then combine ship length and displacement values into a single normalized-displacement value (longer ships get “bonus” to displacement and vice versa).
  • Calculate an RCSm2 with the Skolnik formula based on the adjusted displacement.
  • My expectation is that the Skolnik function will have similar shape to the DB values on average.
  • Then adjust the Skolnik function values (add a constant to it) to match the real life experiments, i.e. (1).
  • By now we should have an RCS like curve that matches the few examples we have. This can then be used to pinpoint out-of-ordinary values in HCDB (candidates for further investigation) and also to generate baseline values for new ships when no other source is available.
  • Of course most of the outstanding values will be intentional deviation, due to differences in ship design, like advancement in ship stealth technology. For this most common case I added a Stealth column to the table, where a “stealth level” can be entered, each level diminishes the Skolnik value with power of two (or increases it for negative values.)
  • Finally there is also an override column. Sometimes you just don’t want to argue with a function.

Ok, enough talk, do it.
I’m splitting the ship annex into smaller groups (like carriers, cruisers, destroyers, etc.), there is less scatter and exception in smaller groups, easier to follow a trend.

I demo with destroyers: steps 1) and 2), HCRCSm2 of destroyers presented on the length scale:

Steps 3) 4) 5) HCRCSm2 and Skolnik curve:

Now the shape looks ok (if there is any trend in the red curve, the blue could well be one :) ) but the low point is way too low (for the HC way of radar model)
The leftmost, smallest destroyer is Egypt’s Tariq, 91m, 1475t, HCRCS=186 == 1685m2,
this would give a mere 27.19nm detection range for the APS-134 (per HCSW), which it can even achieve (according to (1)) against 1m2 liferaft.
So I will boldly shift the Skolnik curve up with (say) 80000, Step 6):

This yields a 74.81nm detection range for the APS-134 against the Tariq, still modest but in line with (1) measurements against 60 to 300 feet long targets.
Also according to (2) the APS-134 achieves 96/54nm against small/vsmall targets, no problem here either.

Lets also check the other end of the DD line how does the +80k impact that,
ignore the Zumwalt and Kongo classes just for now,
the Improved Spruance had 78nm APS-134 range according with the original DB value, the adjusted Skolnik gives 85nm, within bounds I say.
Ok, so I convinced myself that Skolnik with the +80k shift gives a reasonable approximation.

Now comes the more tedious task of checking ships with larger deviation vs Skolnik
Starting from the largest

  • Zumwalt, obviously intentionally small RCS, I set it lvl 10 stealth (Skolnik RCS decreased to it’s 1/1024 part)
  • Kongo, ok again, set lvl 5 stealth
  • Several Spruance classes, ok no change needed
  • Atago classes, give it lvl 5 stealth
  • KDX-3, lvl 2
  • Then Kolkata Type 15A, Udaloys, Delhi Type 15, Shirane, nothing special, keep as is

Step 8) Skolnik and DB curve after few of the largest destroyers were adjusted for stealth:

Ok, definitely some more work in here but on track, to be continued.

I leave it for now with the shifted Patrol Boat curve, without the manual adjustment yet, better match than with DDs.






Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Create New...