The Sinking of the IRIS Dena and the Law of Naval Warfare

On 04 March 2026, within the context of armed conflict between Iran, on one hand, and the US and Israel on the other, a USN submarine torpedoed and sank the Islamic Republic of Iran Ship (IRIS) Dena. It is reported that the attack took place ‘in international waters’ to the south-west of Sri Lanka. Other reports indicate that the warship was somewhere between 20 to 40 nautical miles west-south-west of Galle in Sri Lanka.

IRIS Dena was a relatively modern (commissioned 2021) Iranian-built Moudge-class frigate of 1800 tonnes, with 130-180 (reports differ) ship’s company. Having recently taken part in a multinational naval exercise hosted by India in the Bay of Bengal, the ship was apparently seeking permission to enter port in Sri Lanka. There are also reports that IRIS Dena had been offered port entry by India three days prior to the sinking.

Sri Lanka quickly responded to the incident upon receiving a distress call from the sinking warship. There are also reports that the US quickly reported the incident and the location of the survivors to the Sri Lankan authorities in whose SAR area the sinking occurred. Sri Lankan armed forces and agencies ultimately rescued 32 of the ship’s company and recovered 84 bodies. Around 60 of the ship’s company remain unaccounted for. The rescued sailors are currently in Sri Lankan medical and other facilities, and the recovered bodies of the Iranian sailors have been repatriated to Iran. Another Iranian naval vessel, IRIS Bushehr – citing engineering issues – entered the Sri Lankan port of Trincomalee shortly after the sinking and a third Iranian navy vessel – IRIS Lavan – has been allowed into Kochi port in India. Both Bushehr and Lavan and their crews have been interned by Sri Lanka and India respectively.

This is the fourth known sinking of a warship by submarine-fired torpedo in the context of armed conflict since the end of WWII. The first was the sinking of the INS Khukri on 09 December 1971 by a Pakistan Navy submarine during the 1971 India-Pakistan War. The second was the sinking of the ARA General Belgrano by HMS Conqueror on 02 May 1982, during the Falklands / Malvinas War. The third was the sinking of ROKS Cheonan by a DPRK submarine on 26 March 2010, in the context of either – as the ICC OTP report concluded [para 53] – a reactivation of the 1950-1953 Korean hostilities via an end to the ceasefire established in the (still in force) 1953 Armistice, or a new international armed conflict (IAC) between DPRK and ROK.

To analyse the law relating to the attack on IRIS Dena, and its aftermath, the following questions are relevant:

What rule-set applied at the time? LOAC / LoNW, or something else?

There is zero doubt that the US and Israel are engaged in an IAC with Iran as per common article 2 of the 1949 Geneva Conventions. Because the law of armed conflict (LOAC) relevant to IAC applies as between the belligerents, the relevant rule-set to apply as between US and Iranian warships in relation to hostilities at sea is, consequently, the specialised law of naval warfare (LoNW) as applicable in IAC situations.

Where did the attack occur?

Under LoNW, the belligerents (in this case, US and Iran) can engage in hostilities in all sea areas outside of neutral waters. Neutral waters are the sovereign waters of neutral states – such as, in this case, Sri Lanka. The attack occurred around 20-40 nm off the coast of Sri Lanka, so it was outside the Sri Lankan territorial sea, which extends to a maximum of 12 nm from Sri Lanka’s baselines (LOSC article 3). The fact that the attack occurred in the Sri Lankan EEZ is of no relevance here because EEZ’s are areas of ocean space where the relevant coastal state has sovereign rights – predominantly resource rights – but not sovereignty. As far as LoNW is concerned (see San Remo Manual, rule 34; Newport Manual (2^nd Edn), para 4.1.2), the EEZ is essentially the same as the high seas (see LOSC articles 58, 87) in terms of being a valid venue for engaging in hostilities against the enemy. The only caveat one would add here is that belligerents must still pay ‘due regard’ to the coastal state’s EEZ rights in their operations against each other in that neutral coastal state’s EEZ, but this is a limited obligation relating to the EEZ rights specifically, not to other rights a coastal state may claim (LOSC articles 56, 58-59). There is nothing in this situation which would indicate any due regard obligation was even enlivened as there were no Sri Lankan EEZ rights affected.

Does LoNW allow a USN warship to attack an Iran Navy warship?

Yes – enemy warships are lawful targets during IAC at sea. First, there is no question that IRIS Dena was a warship as it clearly met all of the definitional elements of the LOSC article 29 definition:

For the purposes of this Convention, ‘warship’ means a ship belonging to the armed forces of a State bearing the external marks distinguishing such ships of its nationality, under the command of an officer duly commissioned by the government of the State and whose name appears in the appropriate service list or its equivalent, and manned by a crew which is under regular armed forces discipline.

There is also no dispute that this definition also applies within the context of LoNW – indeed, it is worth recalling that this definition is pretty much a direct lift from a LoNW instrument – 1907 Hague Convention VII on the conversion of merchant ships into warships (articles 1-4). Consequently, the IRIS Dena was clearly targetable. Informed commentary on the attack has almost universally (and correctly) recognised the lawfulness of the attack (eg, here, here, and here).

There is, however, an interesting point as to whether warships are lawful targets by virtue of the 1977 Additional Protocol I (AP I) article 52(2) definition of ‘military objective’, or by operation of pre-AP I customary LoNW, which simply held warships to be lawful targets ab initio. This issue is complicated by the wording in AP I article 49(3) which states that the provisions of Section I of Part VI of AP I (which includes the definition of military objective) apply to attacks from sea to shore, and where ‘sea warfare which may affect the civilian population, individual civilians, or civilian objects on land’, but do ‘not otherwise affect the rules of international law applicable in armed conflict at sea…’.

On one hand, this may imply a difference in test given the military objective test is two-limbed, while the customary (pre-AP I) LoNW test is effectively a singular question – is it an enemy warship. In this regard, the issue really is whether the second limb of the military objective test – ‘…and whose total or partial destruction, capture or neutralization, in the circumstances ruling at the time, offers a definite military advantage’ – applies in any materially restrictive way to clearly military material (such as a warship, except of course where that warship has indicated it is hors de combat, which is generally evidenced by striking its colours of where the crew are observed to be taking to lifeboats). In effect, however, the alternative approaches do not really herald any material difference in outcome. This is because the orthodox approach to the article 52 test is that neutralisation of military combat objects – tanks, warships, military aircraft – always offers the required military advantage.

Second, there is no doubt as to the right of a US warship – in this case a submarine – to conduct attacks on their enemy’s warships (and auxiliaries). This is because warships (although not auxiliaries) have ‘belligerent rights’ (see, for example, Newport Manual, para 3.1),  including the right to attack enemy warships and auxiliaries with lawful means and methods of warfare such as torpedoes.

What counter-arguments have been put in relation to the lawfulness of the attack?

There appear to be two broad arguments in circulation that call the lawfulness of the attack into question. The first is about vessel location and conduct. This argument is evident in some commentary around the fact that the attack occurred some way from the main area of operations, and / or that the IRIS Dena had not otherwise been involved in Iran’s operations against the US and Israel. The implication appears to be that this makes the attack in some way more legally questionable or even ‘treacherous’. Indeed, in one commentary, a US expert is reported to have stated that

…the attack on the Iris Dena was illegal. The ship was transiting home after participating in training exercises hosted by the Indian navy at the time of the strike. India had convened naval assets from 74 countries for a March exercise.

“Was that warship actively posing a threat or participating in hostilities?” asked [the expert]. “You cannot say that this warship was an imminent threat to anyone. By targeting it, is the Trump administration saying that the imminent threat is all of Iran’s government and military? If so, that’s an incredibly dangerous example of military overreach.”

With respect, this is not the case – this is an operational or political argument that has no bearing on the legality of the attack. Enemy warships outside of neutral waters are liable to attack regardless of how close or far they are from the main area of operations, and the expansion of the hostilities (as another expert has noted) is politically significant, but not (in this case) legally significant. There is no ‘conduct of hostilities’ test in the assessment of warship targetability – it is a platform-status assessment, and IRIS Dena was clearly an enemy warship liable to attack. It is notable that similar – and similarly legally irrelevant – claims were made as to the UK sinking of the ARA General Belgrano during the Falklands / Malvinas war in 1982, because at the time of the attack, the Argentinian warship was outside the UK declared 200nm Total Exclusion Zone around the Falklands / Malvinas islands (see chapter 16 in this collection). This is, once again, a political issue not a legal issue – the law is crystal clear that enemy warships may be attacked wherever they are found, so long as they are outside of neutral waters.

The second argument is a more purely legal argument, but it is also wrong because it is a lex ferenda argument rather than a lex lata argument. This is the assertion that because the US and Israel armed attack was a breach of the jus ad bellum, this illegality contaminates or limits the jus in bello available to the aggressor. This critique is a subset of a broader, highly principled, persuasively argued, but in my view nevertheless erroneous, emerging interpretive endeavour to subordinate (or enhance) the availability or legitimacy of jus in bello rights to a conflict’s jus ad bellum assessment. That is, if the war is illegal, the jus in bello available to the aggressor state is more circumscribed (eg geographically, or in terms of requiring that enemy units commit hostile acts before they are targetable); if the war is defensive and existential (in the Nuclear Weapons advisory opinion para 96 sense), the jus in bello correspondingly expands.

With deep respect, I cannot agree and remain in this regard a ‘purist’, holding fast to the orthodox interpretation that illegality in the jus ad bellum does not infect the jus in bello, or create ab initio illegality in the jus in bello, or vice versa. In my view, the purposes of this de jure non-permeability – keeping the different tests separate, mitigating the potential for a downward reciprocity spiral in the conduct of hostilities, and so on – remain as relevant and important today as they have been for the last 80 years.

What law relates to the issue of rescue?

There has been some commentary to the effect that the US did not fulfil the 1949 Geneva Convention II article 18 obligation to rescue the sick, wounded and shipwrecked in the aftermath of the attack. Three issues are of direct relevance here. The first is the addressee of the obligation, in that it is not the Commanding Officer (CO) of the belligerent warship but rather the state that holds the obligation: As article 18 indicates, ‘After each engagement, Parties to the conflict shall, without delay, take all possible measures to search for and collect the shipwrecked, wounded and sick…’. This is not to say that the CO of the attacking warship is not a state agent who must implement this obligation where this is possible. However, the obligation does not rest upon the on-scene CO alone, or even primarily.

The second issue of relevance is the nature of the article 18 obligation as it applies specifically to submarines. There is no doubt that in general submarines have the same obligations as surface warships, as the 1936 London Protocol (which continued in force article 22 of the then-expired 1930 Treaty of London for the Limitation and Reduction of Naval Armaments), and the IMT Nuremberg trial of Admiral Doenitz (on this issue, in relation to the Laconia Order), have reiterated. However, as some commentary has correctly observed (eg, here and here), submarines are perhaps to be treated differently in respect of this obligation because of their vulnerability if detected, and their impracticability as a rescue platform. Indeed, the 2017 ICRC Commentary on Geneva Convention II  article 18 (paras 1642-1644) specifically notes that

1643. Provided doing so does not render the submarine detectable to the enemy (this assessment belongs to the analysis of the security/military considerations discussed in paras 1649–1652), Article 18 may require its commander to alert his or her own authorities and, where possible, other entities, to the location of the attack and to the possibility that there may be survivors, thereby allowing the Party to the conflict to assess which ‘possible measures’ may be taken, for example sending other vessels to the area. While doing so may not be possible seconds after launching its weapon, it may become feasible once the submarine has moved away from the area.

The third issue is that the fact of IAC at sea does not alleviate neutral states of their own obligations – eg as the regional SAR coordinator, or in terms of the SOLAS obligations of masters of merchant vessels of their nationality – from also complying with their (peacetime, or regular) rescue obligations.

The consequence of these three factors is that the GC II article 18 rescue obligation must be assessed holistically and contextually. In this regard, there are reports that the US notified Sri Lankan SAR authorities and also had other plans in place for the rescue of the survivors (on which see the 2017 ICRC Commentary, paras 1637-1641). Additionally, the fact that the Sri Lankan rescue operation was so quick and so effective is also important to this ‘good faith, based on the circumstances and the information reasonably available to both the commanders on the spot or nearby and to the other organs acting on behalf of the Party to the conflict’ assessment (Commentary para 1636).

What is the law relating to the internment of belligerent warships and crews?

The neutral obligation (1907 Hague XIII, article 24) to intern, for the duration of the IAC, any belligerent naval personnel who come into that neutral state’s power, as well as belligerent warships or auxiliaries that seek – in essence – sanctuary in their neutral ports, is uncontentious. Perhaps, until recently, the most famous example was the Argentinian internment of the crew of the German warship Graf Spee on 17 December 1939, for the remainder of WWII. This occurred after Graf Spee had scuttled in the vicinity of Montevideo harbour (see chapter 14 in this collection).  In the present case, Sri Lanka has appropriately retained control of the IRIS Dena survivors and the IRIS Bushehr crew, as well as the IRIS Bushehr itself. India has likewise retained control of the IRIS Lavan crew and the vessel itself. These are clearly actions in compliance with the neutral internment obligations of Sri Lanka and India. However, this neutral obligation is an ongoing one until the general close of military operations and subsequent repatriation arrangements are made, or – while the IAC is ongoing – where the belligerents agree to a repatriation.

Conclusion

As a general proposition of law, it is fair to state three things about the sinking of IRIS Dena. The first is that the attack on IRIS Dena by a US submarine was unambiguously lawful, and it is important to distinguish this legal assessment from assessments that the location, or politics, or the jus ad bellum surrounding the attack somehow erode the attack’s lawfulness as a matter of jus in bello. The second is that there does not appear to be any readily sustainable argument on the basis of the facts as presented to date, that the US breached its GC II article 18 obligation as to rescue of the sick, wounded and shipwrecked after the attack. The third is that Sri Lanka and India appear to have complied closely with their neutral obligations in terms of internment of belligerent warships (or auxiliaries) and their crews.