Main Narrative


The week commencing 12 April 2015 saw what is believed to be the largest loss of life at sea in the recent history of the Mediterranean. On 12 April, 400 people died when an overcrowded boat capsized due to its passengers’ excitement at the sight of tugboats approaching to rescue them. Less than a week later, on 18 April, a similar incident took an even greater toll in human lives, leading the deadliest single shipwreck recorded by the United Nations High Commissioner for Refugees (UNHCR) in the Mediterranean.   Over 800 people are believed to have died when a migrants’ vessel sank after colliding with a cargo ship that had approached to carry out a rescue operation due to a mis-manoeuvre. More than 1200 lives were thus lost in a single week. As Médecins sans frontieres (MSF) commented at the time, these figures eerily resemble those of a war zone.  

Beyond their huge death toll, what is most striking about these events is that they were not the result of the reluctance to carry out rescue operations, which has been identified as a structural cause of migrants’ deaths in the Mediterranean Sea and which we have documented in previous reports.   In these two cases, the actual loss of life occurred during and partly through the rescue operation itself. The detailed reconstruction of these two successive tragedies provided in this report in the “Black Week” section shows, however, that in all likelihood the merchant vessels involved complied with their legal obligations and did everything they possibly could to rescue the passengers in distress. While it could appear that only the ruthless smugglers who overcrowded unseaworthy boats to the point of collapse are to blame, the report focuses on the deeper responsibilities of EU agencies and policy makers. It demonstrates that the latter’s policy of retreat from state-led Search and Rescue (SAR) operations shifted the burden of extremely dangerous search and rescue operations onto large merchant ships, which are ill-fitted to conduct them. In this way, EU agencies and policy makers knowingly created the conditions that led to massive loss of life in the April shipwrecks.

About the report

The following report, produced by Forensic Oceanography – a research team based within the Forensic Architecture agency at Goldsmiths (University of London) that specialises in the use forensic techniques and cartography to reconstruct cases of deaths at sea – in collaboration with WatchTheMed and in the framework of the ESRC-supported “Precarious Trajectories” research project, seeks to understand the conditions that made these events possible.

It does so by mobilising a vast array of methodologies and techniques. First, the report offers, through a series of visualizations, diagrams and figures, a detailed spatio-temporal reconstruction of various cases of shipwrecks. This work was an exercise in the culling of disparate data that was eventually recombined in an effort to assemble a coherent spatial narrative of the chain of events. The reconstructions provided by the report are in fact based on numerous sources, in particular survivors’ testimonies, distress signals, Search and Rescue (SAR) reports provided by Frontex, Automatic Information System (AIS) vessel tracking data, judicial documents obtained from public prosecutors’ offices in Sicily investigating these cases, and photographs taken during the events by rescue teams. At times, elements of information were also extracted from secondary sources such as news reports and human rights reports by international organizations such as Amnesty International.

The policy decisions that led to these shipwrecks, however, sought precisely to keep state-operated assets at a distance from the area in which these were occurring. Focusing exclusively on the reconstruction of the events, then, would not have allowed for an accurate description of the mechanisms of this form of killing by omission. Therefore, in addition to case reconstruction, the report undertook an analysis that could be characterized as a “policy forensics”. This consisted of a comprehensive textual analysis of various technical assessments produced by Frontex, official statements by policy makers and EU officials, minutes of operational meetings between Frontex and other member states agencies, and transcripts of debates in the European Parliament and in its Civil Liberties, Justice and Home Affairs (LIBE) committee. This endeavour was necessary, first of all, in order to gain a fine-grained understanding of the successive institutional steps that led to the retreat of state-led SAR operations. Secondly, it has allowed us to assess with precision the degree of knowledge concerning the risks to migrants’ lives the actors taking these decisions possessed. On this basis, the report points to the responsibility of the various agencies and individuals that took those decisions.

Finally, the report seeks to attend to the materialisation of these policies at sea, in terms of: the operational zones, operational logics and practices of state actors; how these policy shifts affected the practices of other actors operating at sea, such as smugglers and merchant ships; and the conditions and danger of migrants’ crossings. Here key sources were: spatial analysis of operational zones; interviews with state officials (the Italian Coast Guard, Customs Police and Frontex) concerning their operations at sea; and statistical data referring to migrant arrivals, deaths and SAR operations.

The diversity of sources and types of data required the report to draw upon the methodologies and expertise of a variety of disciplines. The material has thus been analysed in collaboration with experts in the relevant fields of geographic information science, vessel tracking technologies, image forensics, oceanography, statistical analysis, EU policy, international law and migration studies.

Report outline

After this introduction, in “Mare Nostrum’s Demise” we introduce the characteristics of the namesake Italian Navy’s rescue operation launched in October 2013 and refute the criticism that was mounted against it to justify its termination.

In “Institutionalised Neglect” we reconstruct the institutional processes and the operational decision-making that led to the launch, on 1 November 2014, of the Triton operation led by Frontex, the European border agency. Compared to Mare Nostrum, this operation deployed more limited assets in an area further away from the Libyan coast and did not have rescue as operational priority. Through operational documents and meeting minutes, the report demonstrates that EU member states, political bodies and agencies – in particular Frontex – knew that the ending of MN would cause increased risk for migrants and ultimately lead to more deaths at sea, but discarded this information to prioritize the aim of deterrence.

In the Rescue Gap” analyses the effects these policy decisions had on the practices of actors at sea and the danger of crossing. Through forensic case reconstruction, together with spatial and statistical analysis, the report demonstrates that Triton operation left an increasingly dangerous gap in SAR capabilities.

April’s Black Week” reconstructs in detail the 12 and 18 April 2015 shipwrecks in which more than 1200 people died. While it could appear then that only the ruthless smugglers who overcrowded unseaworthy boats to the point of collapse are to blame, in light of the sequence of events and policy decisions we have reconstructed, we can see another level of political responsibility emerging. The EU’s policy of retreat of state-operated rescue at sea left in fact ill-adapted commercial vessels to bear most of the responsibility for rescuing and this, in turn, led to assistance becoming deadly. Death by rescue, we conclude, was the outcome of the EU’s policies of non-assistance.

In section 6, “After the Shipwrecks”, the report describes the main policy shifts brought about by the April shipwrecks, most notably, the extension of Triton operational area as well as the launch of the military operation EUNAVFOR MED, both of which continued to prioritize security concerns over saving lives.

Finally, in the “Conclusions” section, the report summarises its findings concerning the responsibilities involved in the twin April shipwrecks. We further argue that while the current exclusionary migration regime stands, migrants will continue to be forced to resort to smugglers in order to reach EU territory, and will encounter situations of distress. Proactive state-led SAR operations are thus necessary. However, because even in the presence of the record means deployed by the Mare Nostrum operation, the danger of crossing remained high, rescue is insufficient to bring the deaths of migrants to an end. Only a fundamental reorientation from a policy that seeks to select and block migrants’ movements to one that would grant legal and safe passage, would make both smugglers and the very need to rescue migrants at sea obsolete. This is the condition to stop the list of more than 20,000 recorded cases of deaths at sea since the beginning of the 1990 from growing ever longer.


The Mare Nostrum (MN) operation was launched in response to the tragic shipwreck that occurred on 3 October 2013. On the morning of that day, a boat carrying more than 500 migrants sank less than one kilometre from the coast of Lampedusa, causing the death of at least 366 people.   The public outcry that followed this event forced EU politicians to respond. Between 7 and 8 October 2013, ministers gathered at the Justice and Home Affairs meeting of the European Council in Luxemburg “held an exchange of views on the actions that are needed to avoid such tragedies”.   As proposed by Italy, the “Mediterranean Task Force” was established “to identify the tools which the EU has at its disposal and which could be used in a more effective way”.  

In the following days and weeks, however, it became clear that the “actions” and “tools” identified by the EU to avoid massive loss of lives at sea would be geared only in the direction of strengthening border control, not rescue at sea. In the immediate aftermath of the 3 October shipwreck, the EU President, José Manuel Barroso, announced the final implementation of the EU-wide Eurosur surveillance system and the reinforcement of Frontex, the EU border control agency.  

This direction continued to be reaffirmed over the next six months of successive reports and meetings between the different EU bodies. During these meetings, no common measures granting migrants legal access to EU territory and safe transport, would even be considered. Such measures, if initiated, would end precarious means of crossing, and importantly, the necessity to resort to smugglers. Moreover, EU member states and institutions made no substantial efforts to support the rescue of migrants at sea. Assessing this institutional process that unfolded after October 2013, Amnesty International’s Secretary General, Salil Shetty, stated on 21 January 2014:

“Two European Summits and one task force later ... all we have is the re-affirmation of the exact same Policy principles that are at the root of the problem: surveillance and prevention, not rescue and protection.”  
Salil Shetty, Amnesty International Secretary General

The Italian government decided to adopt a different strategy and, in light of migrants’ increasing crossings and deaths, Italy prioritized rescue at sea. On 18 October 2013, the Italian government single-handedly launched what has been by far the largest “humanitarian and security” operation in the Mediterranean: Mare Nostrum.   This operation deployed one 135-metre-long amphibious vessel, two frigates, two corvettes, four helicopters, three planes and unmanned aerial vehicles. These patrolled the sea for over one year with the specific tasks of intercepting and rescuing all migrants’ vessels departing from the Libyan coasts.   The scale of MN was unprecedented; so was the inscription of the humanitarian “duty” of saving lives at the core of MN’s mission, which was coupled with the aim of bringing to justice those deemed responsible for putting their lives at risk, i.e. the smugglers.   This shift of mission produced several breaks in the way rescue and border enforcement at sea had been practiced until then. With MN, rescue activities reached an unprecedented expansion in their spatial deployment. Until then, search and rescue (SAR) operations in the (undeclared) Libyan SAR zone were a rare event and the majority of migrants’ boats reached Italian and Maltese coasts on their own or were simply “escorted” for the last few nautical miles.   During MN, military vessels were continuously positioned in close proximity of the Libyan coast, and intercepted and rescued every migrants’ boat that they encountered. Moreover, while disembarkation had constituted a thorny problem for many years   , with MN, migrants rescued in the central Mediterranean were taken to Italy by default. MN thus constituted a clear shift away from the reluctance of states to carry out rescue operations and the criminalisation of those who engaged in this, which had led to cases of non-assistance in the past.

Synthetic map and figures of the sitation in the central Mediterranean during 2014. Frontex and Coast Guard data show that migrants were mainly being rescued very close to Libyan shores, south of the Mare Nostrum operational area. During this period, the Italian Navy was by far the main SAR actor, accounting for 50 percent of all rescues. Both the number of deaths and mortality rate increased over the summer. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi.

MN, however, had ambivalent results from a humanitarian point of view and became increasingly contested, both within Italy and at the EU level.   Critics of MN argued in particular that the operation did not constitute a viable response to the drama of migrants dying at sea because (1) it constituted a “pull factor” and (2) it inadvertently led to more deaths instead of saving more lives, since it led to more crossings and smugglers, as a consequence of its presence, started sending migrants off in increasingly precarious conditions. While it is a fact that during Mare Nostrum’s period of operation a record number of migrants arrived on EU shores and a record number of deaths were reported, an analysis of migration patterns before, during and after the operation shows that MN was not the root cause of the trends that unfolded during this period.


One criticism of MN was that it led to an increase in the number of arrivals on Italian shores. The availability of a great number of assets dedicated to rescue operations close to the Libyan coast, it was argued, was actually encouraging more migrants to start the crossing on unseaworthy vessels and to put their lives in the hands of unscrupulous smugglers. As a result, the argument continued, to curb arrivals it was necessary to end MN. Politicians across the EU such as the UK Foreign Office Minister Lady Anelay echoed the Italian xenophobic Lega Nord party in denouncing search and rescue operations as a “pull factor”, thereby justifying her government’s refusal to fund the continuation of the operation.  

At first glance, the data on arrivals during 2014 seems to confirm this argument. Arrivals in Italy went from slightly over 40,000 in 2013 to over 170,000 in 2014. Looking at 2013 trends more closely, however, we can see that the increase in the number of crossing had started well before MN. This tendency was indicated by Frontex itself, which in its risk analysis regarding the July to September 2013 period (i.e., before the beginning of MN), observed the most significant increase in crossings since 2011.

“Irregular migration in the Central Mediterranean increased staggeringly between the second and third quarters of 2013. Compared to detections during every other quarter in 2012 and 2013 the increase was both sudden and dramatic to a total of over 22 000 detected migrants.”  
Frontex Risk Analysis

With this increase, 75 percent of the yearly arrivals were reported in only four months.   The agency further noted that, despite it being winter, which leads to less crossings, 2,476 crossings were recorded in December 2013, “the largest for a month of December since 2008”.  

Frontex figures for July–September 2013. The graphs show a strong increase during this period, and Syrians becoming the top nationality crossing the central Mediterranean.

Two main factors can be mentioned to explain the increase recorded in 2013, which would continue to mark 2014 as well. The first one is the Syrian exodus, which in the summer of 2013 went, as a UNHCR report tellingly put it, “from slow boil to breaking point”.   Until the summer of 2013, Syrians had mostly sought refuge in other areas within their own country or in neighbouring countries (Lebanon, Jordan, Turkey, Iraq and Egypt).   However, the increasing numbers of refugees quickly exceeded both the capacity of humanitarian organisations and of neighbouring countries, which began reintroducing visa obligations for Syrians.   Over 2013, as the civil war became more and more entrenched, the Syrian exodus increased markedly – with the number of Syrian refugees doubling from one million in March 2013 to two million in September 2013.   As a result, Syrians became increasingly likely to seek refugee ever further, including by crossing the sea from Libya. This trend would continue apace throughout 2014.

The second key factor was the increasing level of violence that resulted from the progressive collapse of the transition process in Libya. The process of political fragmentation led to mounting violence that affected Libyan society at large but in had a particularly grave influence on migrants, who were subjected to multiple forms of violence at different moments of their trajectory.   This led a growing number of foreign nationals who had been residing in the country (including Syrians, 14,056 of whom were registered with the UNHCR in Libya by 30 September 2013)   to attempt the crossing of the central Mediterranean.

The intensification of crossings can thus not be attributed solely to the presence of MN, since they resulted from deeper regional political factors that were leading to this trend before MN. As we will discuss in more detail in the “After Mare Nostrum” section, the fact that MN was not the major cause is further confirmed by the comparable scale of crossings after MN.


The other main criticism put forth to justify ending support for MN was that the operation increased the danger of the crossings, as it prompted a shift in smugglers’ practices. Smugglers, it was argued, were relying on the presence of MN to operate rescue at an early stage and were thus providing subpar crossing conditions, thus provoking more deaths. Frontex’s executive director, Gil Arias, epitomized this position in his answer to the Committee on Civil Liberties, Justice and Home Affairs (LIBE) of the European Parliament on 4 September 2014:

“I’m afraid of saying that it has been a pull factor but obviously the smugglers have abused of the proximity of the operation to the Libyan coast to, on the one hand to put more people on the sea, with the assumption that they will be rescued very soon, and this also made it cheaper for them, as I said, because they put and they are putting less fuel, less food, less water on the vessel, which at the same time also increases the risk for the migrants.”  
Gil Arias, Frontex’s executive director

Let’s analyse this claim. First, it is a fact that while MN allowed for a record number of people to be rescued, a record number of 3,186 deaths were also reported in the central Mediterranean in 2014 alone.   During the period of Mare Nostrum’s operation, then, the crossing did not become less dangerous and the mortality rate actually slightly increased from 15/1,000 in 2013 to 18.4/1,000 in 2014, with a peak in the summer.   So while MN did not make the crossing substantially more dangerous, as its critics held, it is also true that it did not make it less dangerous. It is also a documented fact that smuggling practices evolved during the time of the MN operation, and partly adapting to its presence. It is certainly likely that this shift contributed to making the crossing more dangerous, leading to “a new strategy of deliberately meeting, instead of bypassing controls, but also using even more unseaworthy boats, loading them ever more, and providing less food and fuel”.  

Average number of people loaded per boat, 2013-2014.

The problem with the argument outlined by Frontex’s executive director, Gil Arias, however, is that such a claim assumes a causal chain between MN, smuggling practices and increased risk for migrants. Such an understanding however is insufficient to explain the rise of deaths and slight rise in mortality. First of all, shifts in the practices of smugglers attributed to MN emerged before MN. Secondly, there were factors other than smugglers’ practices that contributed to the rising death toll.

Smugglers in Libya had been operating a well-established business since the beginning of the century thanks to their stable relations with the Qaddafi regime.   The fall of the Libyan regime, however, led to changes in smuggling practices that have made the conditions of the crossing more dangerous. In particular, as Nancy Porsia details in her report included in the annex, the practices of smugglers had already started to considerably change during 2013 in parallel to the dwindling hopes for a stable and prosperous post-revolutionary future. The Libyan political fragmentation led in fact to more volatile relations between smugglers and the factions in control of particular areas but also allowed new actors – who offered lower prices but did not always possess the willingness or the know-how to organise safe crossings – to enter the smuggling market.   This in turn meant that to guarantee a profitable margin, smugglers had to resort to subpar navigation equipment or to loading more migrants on board their unsafe boats. In its 19 November 2013 report for the “Task Force Mediterranean”, the European External Action Service (EEAS) noted the trend towards “increasing numbers of migrants on the boats, which are overloaded and/or unseaworthy, leading to an increasing risk of fatalities especially along the central Mediterranean route”.   As this document was produced only days after the start of MN, it cannot describe a shift brought about by MN, but rather a pre-existing situation. This tendency only increased in the following months and, as Porsia notes, “as of late 2013, the standards of the smugglers’ service drastically dropped: very often the migrants were put at sea with no life jacket, no satellite phone or GPS”.   The shipwrecks of October 2013 and the resulting peak in mortality at the time, as indicated in the data put together by Philippe Faruges and his colleagues (figure below), may be seen as the result of these tendencies in smugglers practices.

Migrant mortality rate in the central Mediterranean, 2011-2015.

The changes in the practices of smugglers and the increase in the danger of crossing were thus well underway before MN was launched and cannot be attributed to the latter. Rather, it seems that MN could not make up for and may have exacerbated a shift in the practices of smugglers that was already underway. This shift would probably have occurred anyway without MN as a result of the combination of the increasing number of people willing to cross the sea and the volatility of the Libyan context mentioned above. The peak in the mortality rate recorded in October 2013 provides an indication of the danger migrants might have continued to face without MN – a peak that, we should note, was reached during the period of Mare Nostrum’s operation. What is clear is that MN could only address this increasing danger of the crossings partially, as MN assets deployed close to the Libyan coast came to operate as a “half-way bridge to Europe”,   still forcing migrants to resort to the service of smugglers for the first stretch of their journey.

Furthermore, we believe that beyond the practices of smugglers, other factors contributing to migrants’ deaths at sea were at work. Firstly, while providing detailed reconstructions of the cases of deaths that occurred during 2014 is beyond the scope of this report, a cursory glance reveals that several of the largest cases of deaths at sea of summer 2014 amounted to simple murder, and that this murderous behaviour cannot be attributed to the effects of MN.   Secondly, the spike in deaths at sea and mortality during the summer of 2014 (June–September) coincided with the worsening political context in Libya (see Porsia in annex), which certainly contributed to increased crossings   and may have impacted the practices of smugglers described above.   It seems to have been beyond the capacity of MN to respond to this high number of crossings occurring in highly precarious conditions. MN assets themselves were overwhelmed by the novel scale of crossings and the shift in their modus operandi, which is indicated by the increasing recourse to commercial vessels to commercial vessels over this same period.

Frontex figure showing that the number of people rescued by the shipping industry suddenly reached more than 20 percent in June 2014 and remained consistently high after that.

What emerges from the above is that MN began operating in the midst of a growing storm, in which the combined exodus of Syrians (and of other nationalities), together with growing instability in transit countries such as Egypt and Libya, and the resulting changing practices of smugglers were leading to record numbers of crossings in increasingly dangerous conditions.   Despite the vast resources put in place, MN did not manage to curb the increasing mortality rate at sea. However, it is also clear that MN was not the main cause of this increase.

Considering the trends of late 2013, it is probable that without MN in place, a similar number of arrivals would have been reached, and both the number of deaths and mortality would have continued to soar, as the twin October 2013 shipwrecks indicate. The peak in the mortality rate observed in summer 2014 further suggests, it was not less of MN that was needed to bring the number of deaths and the danger of crossings down, but more of it.

The ending of MN was thus justified on the argument that MN was the cause of more arrivals and more deaths. However, if, as we have demonstrated, increased crossings and deaths were related to deeper regional factors, discontinuing Mare Nostrum would not lead to less crossings but to more deaths at sea. This forecast was available to EU policy makers and agencies, and yet, as the following sections will show, they decided to end MN and (not)replace it with a more limited Frontex operation in all knowledge of the deadly consequences this policy would have.


While the criticism that MN led to more crossings and more deaths was thus unfounded, these arguments were put forward by EU member states and agencies to respond negatively to Italy’s calls to hand over MN to the EU by transforming Frontex’s central Mediterranean operation into a Europeanised Mare Nostrum. Instead, a much more limited operation called “Triton” would be implemented by Frontex.

The deadly effects of the rapid removal of MN were predicted by several actors, including EU agencies, but across EU institutional bodies and agencies the warnings were ignored. In what follows, we offer a reconstruction of this process of institutional wilful neglect that spans from April 2014 to January 2015.

This reconstruction is based on operational documents, previously unpublished minutes of four meetings (20.08.2014; 24.09.2014; 9.10.2014; 16.10.2014) that Frontex officers had with their Italian and Maltese counterparts, an extensive review of parliamentary discussions and transcripts of official meetings, press releases and official statements by NGOs.


Deeming that the costs of the operation (9.5 million euros per month) should be shared by all EU member states,   already on 16 April 2014 the Italian Ministry of Interior Angelino Alfano declared to the Italian parliament that “Frontex will have to take on a leading role in directing and coordinating patrolling in the Mediterranean.”   On 11 June, the Italian parliament voted on a motion proposing an exit strategy from Mare Nostrum and its replacement by a “conjoint action by the international community”. At the same time, Alfano revealed to the press that he had a long conversation with Home Affairs Commissioner Cecilia Malmström, in which he invoked an exit strategy from Mare Nostrum that involved the EU.   On 24 June, presenting his priorities for the then upcoming Italian presidency of the Council of the EU, Italian Prime Minister Renzi stated that Italy would request at an EU Council summit a few days later that “MN becomes an operation part of Frontex”.  

Over the summer of 2014, a long and sinuous institutional process unfolded during which the contours of the Frontex mission were debated and defined, ultimately leading to a much more limited mission that was a far cry from Italy’s request that Frontex “take on the duties of Mare Nostrum”.   At the 26–27 June European Council meeting, the Council concluded vaguely that “Frontex should reinforce its operational assistance, in particular to support Member States facing strong pressure at the external borders.” This was enough for Prime Minister Matteo Renzi to state on 29 June 2014 that “today, Italy is less alone in the Mediterranean”, asserting that the agreement regarding migration policies reached at the European summit is “a truly significant step forward” because it established “the basis for being able to finally breathe life into a Frontex ‘-plus’, broadening its operations&rdquo   However, the limits of the MN substitute operation proposed by the EU became clear when, following the EU Minister of Interiors and Minister of Justice Informal Meeting on July 9 in Milan, Home Affairs Commissioner Malmström said that she was discussing with Alfano what was needed for a “scaled-down” version of Mare Nostrum:

“Frontex is a small agency and cannot take over Mare Nostrum tomorrow. […] Frontex can do a lot, but we do not have the means to totally substitute (Mare Nostrum) unless all other member countries also contribute with vessels or helicopters or staff or money.”  
Home Affairs Commissioner Cecilia Malmström 

The reticence of member states was however already voiced as they exited the meeting. While German Interior Minister Thomas de Maiziere told Deutsche Welle at the interior ministers’ meeting in Milan that “it cannot be the permanent task of the Italian navy to take in refugees,” he argued that “the idea of passing the task on to Frontex is unrealistic in my eyes.”  

Discussions between Italy, Frontex and the Commission continued throughout the summer. On 26 August, Italian officials met with Frontex representatives and a European Commission delegation in Rome. The results of their talks were discussed during a meeting between Malmström and Alfano in Brussels on 27 August.   Following her meeting with Alfano, in an official press release Commissioner Malmström stated that:

“We have decided that the two ongoing Frontex operations Hermes and Aeneas will be merged and extended into a new upgraded operation. The aim is to put in place an enlarged ‘Frontex plus’ to complement what Italy has been doing.”  
Home Affairs Commissioner Cecilia Malmström 

While the Commission has not granted us the minutes of the meeting between Alfano and Malmström, we have accessed a number of documents that allow us to get an impression of the negotiations between different actors at the operational level following this policy decision, and to assess the knowledge they had of the consequences of ending MN in terms of the risk of crossing for migrants.


On 28 August, Frontex issued a “concept document” on the future of Frontex’s Hermes and Aeneas operations in the central Mediterranean (see (see annex)). Written at a time when the ending of MN had not yet been formally announced, the document proposed that operation Hermes 2014 either be extended in the case of the continuation of MN, or be upgraded to new operation Triton if MN was terminated. The document contains two sections: first, an assessment aiming to “provide analytical support to decision making at Frontex”; and second, the proposal that emerged from this decision making. Crucially, the assessment notes that:

“the withdrawal of naval assets from the area, if not properly planned and announced well in advance, would likely result in a higher number of fatalities.”  
Frontex concept document for operation Triton

As a result, the following recommendation is formulated concerning the operational area:

“if operation Mare Nostrum discontinues in the future, then Frontex should consider extending the operational area of the JO Hermes 2014 and consider to which extent the objectives of effective border control and contribution to saving lives can be achieved.”  
Frontex concept document for operation Triton

In the actual proposal however, considering the scenario in which MN is terminated, the operational zones of Hermes and Aneas are effectively merged into a new Triton operation, but are not extended further south, thus remaining far from the extent of MN (see map comparing zones).

Map comparing the operational zones of Italian Navy’s Mare Nostrum and Frontex’s Triton. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi.

Frontex was thus aware that ending MN abruptly could lead to an increased risk of migrants’ deaths, and that as such specific counter-measures should have been taken, including extending the operational zone. It decided, however, to ignore this risk and not to follow its own internal recommendation in the actual proposal for the new operation.

The minutes of the meetings held between August and October 2014 by Frontex, Italian and Maltese agencies, and the EU Commission (see annex)provide further examples of how these concerns were not considered at an operational level. In these meetings, Italy repeatedly requested that more important means be dedicated to the Frontex operation being defined, including Navy assets, but Frontex responded that this was beyond the means allocated to the operation and defended its operation’s spatial retreat.   On one particular occasion, an official of the Italian Navy did state explicitly that:

“in the case Mare Nostrum finishes, the new operation in the area will have surely extra work as concerns SAR.”  
Italian Navy Official

During none of the meetings between Italy and Frontex, including after the ending of MN was formally announced in October, were concrete measures planned by Frontex or Italy in response to this warning, nor to the increased risks to the lives of migrants it would entail. While these risks were being publically evoked on numerous occasions (see below), they always remained outside the purview of these operational meetings.


While the risk to the lives of migrants was simply absent from Frontex’s operational meetings with member states, there were other opportunities within EU institutional bodies to give heed to this concern. This was the case for instance of the hearing at the European Parliament’s committee on Civil Liberties, Justice and Home Affairs (LIBE) held on 3 and 4 September. On that occasion, Commissioner Malmström and Frontex Interim Executive Director Gil Arias provided further clarifications to Members of the European Parliament, who in turn responded with questions and comments.

It is worth quoting the minutes of this meeting (see annex) at length. The intervention of Arias is particularly important. In it, Arias reiterated the criticism of Mare Nostrum being a pull-factor and representing an increased danger for migrants at sea that we have tackled above. As a consequence, he specified that:

“joint operation Triton will not replace Mare Nostrum. Neither the mandate, nor the available resources, allow for that replacement.”  
Gil Arias, Frontex Interim Executive Director

He then clearly presented Triton as a retreat from MN’s spatial extent and its proactive mission to rescue:

“This operation will be closer to EU shores than Mare Nostrum. […] In any case, the difference between Mare Nostrum and Triton is fundamentally the nature of the two operations. While Mare Nostrum is clearly a search and rescue operation, Triton will be with a main focus on border control, border management, although as it is obvious saving lives is an absolute priority, and in fact very frequently the control operations, the border control operations coordinated by the agency turn into search and rescue operations, and this is how it works in practice.”  
Gil Arias, Frontex Interim Executive Director

At no point in his presentation, however, did Arias evoke the additional risks for the lives of migrants that such a retreat might entail – and of which he must have known, since this risk had been assessed by his own agency. This risk was clearly recalled to him in the response by Italian MEP Barbara Spinelli:

“Now we know that Mare Nostrum is not going to be replaced by Frontex or Frontex Plus, nor by Triton or Hermes, so there is not actually going to be any search and rescue operation. I can draw two conclusions here. Basically Italy is being left on its own, and secondly there is going to be no more Mare Nostrum programme and people are going to be dying in the Mediterranean again. Mr Arias Fernandez, are you aware of those two conclusions?”  
Barbara Spinelli, Member of the European Parliament

Arias did not fundamentally address these comments, but limited himself to reiterating the mission of Frontex:

“It might be worth to repeat once again: while saving lives is an absolute priority, the operations coordinated by Frontex have as main focus border management, and contribution to search and rescue activities which are under the responsibility of the national competent authorities.”  
Gil Arias, Frontex Interim Executive Director

From what precedes, it appears that the additional risk for the lives of migrants that the ending of MN and its replacement by Triton would entail was mentioned at several moments within EU institutional bodies and agencies, but that these warnings were deliberately ignored and at times, as Arias’s statement exemplifies, not properly communicated to EU politicians by the EU border agency.

This is all the more striking considering the many warnings and criticisms formulated publically from without EU institutional bodies once the limitations of the Triton operation became clear.

On 15 September, days after the LIBE hearing, Amnesty International vocally criticised this turn in policy in its report, Lives Adrift:

“People will attempt the crossing, with or without Operation Mare Nostrum in place. But without it, many, many more will die.”  
Amnesty International

On 17 October 2014 the UNHCR stated:

“UNHCR is concerned over the announcement of the ending this month of the Italian operation Mare Nostrum without a similar European search and rescue operation to replace it. This will undoubtedly increase the risk for those trying to find safety in Europe, and could lead to more refugees and migrants perishing at sea.”  

On 5 December 2014, François Crépeau, United Nations Rapporteur on the rights of migrants warned:

“The fear is that, next summer, without an operation like Mare Nostrum, thousands of people will die. Turning a blind eye isn’t a solution: people will continue to cross and, because of Europe’s inaction, to die”  
François Crépeau, United Nations Rapporteur on the rights of migrants

These warnings stemming from within EU agencies, EU institutional bodies, UN organisations and human rights NGOs were not heeded to. On 9–10 October, the Justice and Home Affairs Council welcomed the launch of Triton, due on 1 November, and Italy’s Interior Minister Angelino Alfano formally announced the that Operation Mare Nostrum would end by the end of the year.   After a period of “phasing out” between November and December, from 1 January 2015, the central Mediterranean was left with Triton, which had a much smaller budget – initially 2.9 million euros per month – and fewer available assets patrolling a smaller area, extending up to thirty nautical miles from Lampedusa.  

Why were the increased risk to the lives of migrants either disregarded or not mentioned at all by Frontex and EU member states? François Crépeau, United Nations Rapporteur on the rights of migrants, provided his own interpretation on 30 October 2014. On this occasion he was responding to the statement by UK Foreign Office Minister Lady Anelay, who justified her decision to axe UK support to SAR operations in the Mediterranean by accusing them of being a pull-factor.   Crépeau replied:

“It’s like saying, let them die because this is a good deterrence. […] To bank on the rise in the number of dead migrants to act as deterrence for future migrants and asylum seekers is appalling.”  
François Crépeau, United Nations Rapporteur on the rights of migrants

While Crépeau’s statement was only a well-grounded deduction, the newly released Frontex’s Tactical focused assessment for Triton from 14 January 2015 confirms that the rationale for this retreat of state-operated rescue was indeed to act as a deterrent for migrants and smugglers in the aim of stemming crossings:

“The end of Operation Mare Nostrum on 31 December 2014 will have a direct impact on the JO Triton 2014. The fact that most interceptions and rescue missions will only take place inside the operational area could become a deterrence for facilitation networks and migrants that can only depart from, the Libyan or Egyptian coast with favourable weather conditions and taking into account that the boat must now navigate for several days before being rescued or intercepted.”  
Frontex Tactical Focused Assessment for operation Triton

From all of the above, what emerges is that in the ending of MN and the planning of Triton, the aim of Frontex and EU member states of seeking to deter the arrivals of migrants took precedent over taking into account the information concerning the additional risks that the cutting back of interception/rescue operations would entail for the lives of migrants. More than that, the increased risks to the lives of migrants that were being publically evoked on numerous occasions were never considered in any operational decision and consistently remained outside the purview of EU decision makers.


While in the previous section we addressed the institutional process of ending Mare Nostrum (MN) and not replacing it by the Frontex operation Triton, we now turn to the effects these policy decisions produced in the operational situation that emerged at sea in early 2015, and how the increased risks to the lives of migrants that was forecast but disregarded materialised. As already mentioned, compared to MN, Triton had fewer available assets that were patrolling a smaller area extending up to thirty nautical miles from Lampedusa, instead of proactively patrolling the waters immediately off the Libyan coast as MN had done.   Moreover, the aim of Triton was border control and not rescue at sea, and it therefore involved a very different spatial and operational logic. Search and Rescue (SAR) activities were now only to be operated as a secondary outcome of its border patrols, and Frontex assets would only be deployed towards SAR operations if called upon to do so by the Italian Coast Guard.


This is precisely what happened in early November, just days after the beginning of Triton. Faced with the gap in rescue capabilities determined by the phasing out of MN, the Maritime Rescue and Coordination Center (MRCC) in Rome started to call upon Frontex assets to operate SAR operations that, as predicted by the human rights advocates mentioned in the previous section, continued to occur outside Triton operational area and close to the Libyan coast. In response to this, on 25 November 2014, Klaus Rösler, Frontex Director of Operations Division, wrote a letter to Giovanni Pinto, Italian General Director of Immigration and Border Police, to voice his “concerns about engagement of Frontex deployed assets in activities outside the operational area”, as the letter’s subject line reads. Quotes from the letter were leaked to the press already in early December 2014, but the full document was released to us by Frontex only recently (see annex).  

In the letter, highlighting that “the operational aim of Frontex joint maritime operations is […] to control irregular migration flows towards the territory of the European Union and to tackle cross border crime”, Rösler refers to two main SAR events that had occurred in November 2014, the second of which is the most interesting in the frame of our current discussion. In this event, MRCC Rome requested that a Frontex asset not support a specific SAR event but that it sail to “a given area according to the fact of numerous satellite phone calls received”. Here Rösler argued that Frontex assets should only be dispatched after “indication of a state of emergency (Uncertainty phase; Alert phase; Distress phase)” had been defined. While formally reaffirming Frontex’s readiness “to support well-grounded national efforts of MS [Member States] as regards international Search and Rescue (SAR) obligations”, the letter ends with the suggestion “to ensure management of Frontex deployed assets in SAR events […] in line with Operational Plan”, which, as stated earlier, defines border control as its priority.

The letter shows the extent to which Frontex was using technical arguments to voice its reticence about carrying out rescue operations and to seek to limit its assets being called upon to intervene outside Triton’s operational area. This was at least the interpretation of a group of Members of the European Parliament, including the above-quoted Barbara Spinelli who, a few days later, at the 11 December 2014 LIBE meeting, asked Angelino Alfano, in his role as representative of the Italian Presidency of the EU and Italian Minister of Interior, “to oblige Frontex to abide to its SAR duties enshrined in EU regulations.”  


The first effects of the retreat of state-led SAR capabilities in the central Mediterranean began to make themselves felt in a series of incidents that occurred between early December and early March 2015. In this period, contrary to the assumptions that led to the demise of MN, migrants’ crossings continued unabated and the number of arrivals recorded in the first four months of 2015 remained almost identical to the first four months of 201 (more than 26,000). Furthermore, the smugglers’ modus operandi did not change, and the use of subpar navigation equipment continued to produce the necessity for SAR interventions very close to the Libyan coast. However, there was now no longer a fleet of state-operated vessels positioned close to the coast to detect and rescue them.

Synthetic map and figures of the situation in the central Mediterranean, January – May 2015. Frontex and Coast Guard data show that migrants continued to be mainly rescued very close to Libyan shores, as was the case during 2014 when the Mare Nostrum operational was in place. However, the Triton operational area was now much further north in relation to the location of situations of distress. In this phase, the share of rescues operated by the Italian navy dropped from 50 (2014) to 26 percent, while the share of commercial vessels soared to 30 percent. While arrivals stayed stable in relation to 2014, the number of deaths peaked, leading to the migrant mortality to increase 30-fold, from 2 deaths in 1,000 crossings to 60 in 1,000, with peaks in February and April in function of the cases of deaths investigated by this report. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi.

Two main mechanisms led to increased deaths as a result of this vacuum in SAR capabilities. First, in some instances, boats were left to drift for days before being detected and, once detected, it took many more hours for rescue vessels to reach the location of distress, due to their positioning much further away from the location of SAR events. This was made apparent by the cases of the 22 January and 8 February that cost in total an estimated toll of 365 deaths, producing a peak in the mortality rate for the period.   Secondly, as began to be apparent with the case of 3 March, in order to cope with this SAR vacuum, MRCC Rome was increasingly forced to mobilise commercial vessels operating in the area for SAR activities. However, filling this gap with inadequately equipped commercial vessels would lead to cases of death by rescue.

These cases thus represented clear warnings of the effects of the EU’s policy of retreat from state-led search and rescue operations. Almost every single one of these incidents was reported in the press and critically commented upon by state actors and the human rights community. These warnings of the predictably larger tragedies to come remained unheeded to by EU policy makers and agencies.


22.01.2015 CASE

As they approached the dock in Malta, the passengers of the 22 January case gaze back on the rubber boat on which they had drifted for around eight days. Photograph credit: Chris Mangion/ Malta Today

According to testimonies collected by Amnesty International (AI) in the Safi migrant detention centre in Malta, on 15 January 2015, a small rubber boat carrying some 122 people left the shores of Garabouli in Libya at about 16:00 GMT.   hey were told by armed smugglers to continue straight ahead and they would reach Italy. They had no telephones, water or food, and no life jackets. After a few days, fuel ran out and they started to drift. The dinghy started taking in water and they had no buckets to empty it out. The passengers became extremely cold, thirsty and hungry. A survivor described people “losing their minds” and jumping into the water. Other drank seawater. In its summary of the case, Amnesty describes the detection of the migrants’ boat as follows:

“their boat had been drifting for around eight days before a fishing boat spotted them some 2.5 nm east of Maltese shores at 7:00am. Within 30 minutes, two Armed Forces of Malta patrol boats, one of which operating under Triton, reached the boat in distress.”
Amnesty International

Neither the Maltese authorities nor Triton assets had been able to detect the boat despite it entering Malta’s territorial waters and almost reaching the country’s shores. Up to 34 people died. 88 young men from Sub-Saharan Africa were saved, but one died in hospital shortly after the rescue.  

This incident offered a first signal of two trends: first, the fact that the smugglers sent out the boat in the middle of the winter without even providing a satellite phone indicated the continuation of the increasingly reckless practices of smugglers, whom, with or without Mare Nostrum, were offering substandard security levels, even in relation to the practice established several years prior by smuggling networks for whom satellite phones had been part of the standard “kit”. On the other hand, the fact that the vessel was allowed to drift for around eight days and to reach the island of Malta, 400 km further north, was certainly indicative of the lack of patrols operating in this space after the end of Mare Nostrum, and the consequent gap in detection and rescue capabilities.

08-09.02.2015 CASES

The position of three out of the four migrant boats that left the same time from Garabulli and AIS data showing vessels in the area. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi. Data: MarineTraffic

On 8 February 2015, MRCC Rome directed its assets towards several boat in distress located at about 120 nm from Lampedusa and 40nm from Tripoli. However, because of the SAR vacuum, the deployment of rescue took several hours, which cost more lives as the passengers faced severe meteorological conditions. Furthermore, some of the passengers even perished after rescue had taken place, during the 18 hours that it took for the patrol boats to sail back to Lampedusa. The survivors interviewed by Amnesty estimated that about 105 people were on board each of the four dinghies. With only 86 passengers rescued, it is probable that some 330 of their fellow travellers perished at sea.

Amnesty International (AI) interviewed the 86 survivors of this series of shipwrecks soon after their arrival at the beginning of February 2015. Additionally, we have accessed the detailed SAR report from Frontex (hereafter abbreviated as FX), and AIS data (for which, probably because of the bad weather conditions, coverage was poor). The following chain of events emerges from these sources. Based on its interviews, Amnesty summarised the first phases of the events as follows:

“Some 420 migrants left together from the Libyan port town of Garabouli, 40km west of Tripoli, in four inflatable dinghies. Most were young men from West Africa. People smugglers had kept them near Tripoli to await the journey after charging them the equivalent of around 650 euros. On the evening of 7 February, the smugglers, armed, made them board the dinghies, which were numbered 1 to 4. The boats were powered by small outboard motors, and the smugglers had not provided enough petrol for the trip. […] Early on 8 February, the boats drifted in the Mediterranean Sea north of Libya, in serious danger. High waves were washing people off the dinghies and into the sea.”  
Amnesty International

According to the detailed SAR report provided to us by Frontex, at 11:50 GMT, MRCC Rome received a call from an operator of the reception centre, Corcolle. The caller reported that she in turn had been contacted by one of the passengers on board via satellite phone, who said there were about 300 migrants and that they had been drifting for about 13 hours. At 12:03 GMT, the satellite telephone in question was located in position 33°49'N - 013°41 'E (FX), that is, approximately 120 nm from Lampedusa and 40 nm from Tripoli. This was the first of several contacts with the migrants’ boats as indicated in the map above.

At 13:30 GMT, MRCC Rome managed to reach the passengers via the satellite telephone number. This time the migrants reported to be 105 persons on board, adrift and without water or food. MRCC Rome declared this a SAR event (FX). At 13:50 GMT, MRCC Rome sent out the following distress signal to all vessels transiting through the area with the position it had established for the vessel at 12.03 GMT:

IN 33-50N 013-41E AT 081203Z FEB.
PHONE: 39 065 908 4527, 39 065 908 4409,
FAX: 390 6592 2737, 3906 5908 4793,
E-MAIL: [email protected]
( 081350Z FEB 2015 )

The poor coverage of the AIS vessel tracking data at the time (probably as a result of the severe meteorological conditions) does not allow us to ascertain with precision the location of the different vessels at the time. However, because of the time it took them to navigate towards the location of the SAR event, it appears that the coast guard vessels were located near Lampedusa (this is confirmed by Frontex for at least two assets). Neither were any commercial vessel accounted for by the AIS data located in proximity at the time of the distress signal. The Italian Coast Guard’s vessels CP 302, CP 305, CP 312 and CP 319 (two of which were coordinated by the Frontex operation Triton) were dispatched to the location of the vessels in distress.   According to Amnesty, “despite prohibitive weather conditions, with exceptionally strong winds and several metres-high waves, Italian coast guard responders managed to reach the boat in distress after approximately 6.5 hours of navigation at 9 PM [20:00 GMT]”.   In addition, two merchant vessels operating off the Libyan coast – the Bourbon Argos and the Saint Rock – and were requested by MRCC Rome to direct themselves towards the migrants’ vessels.   While AIS data shows the Saint Rock changing its course towards the location of the boat in distress at 12:17 GMT, the Bourbon Argos, for which only the position at 12:17 GMT is available, was the first to reach one of the rubber boats. It was unable to conduct a rescue operation but waited for the coast guard and sought, in the meantime, to shield the boat from high waves.  

When the coast guard vessels arrived on location, they were able to rescue 105 people from the first rubber boat, but 29 people died of hypothermia (FX) within the 18 hours that it took for the patrol boat to navigate back to Lampedusa in extreme weather conditions. Throughout this time, the passengers had to remain on deck for lack of space inside and could enter the cabin only for short periods, on rotation. The CP 302 and 319 arrived at the island at 14:40 GMT on 9 February, carrying 58 people including the 29 people who had died; the CP 305 and 312 returned at 18:25 GMT, carrying 46 people (FX).

An Italian Coast Guard vessel arrives in Lampedusa in the afternoon of 9 February 2015. Video credit: Italian Coast Guard

One of the coast guards on board, Salvatore Caputo (whom the authors of this report also met in Catania in October 2015), recalled the return journey as follows to Amnesty:

“They were exhausted, thirsty, very hungry… As we proceeded to transfer the men onto our vessels, with the merchant vessel trying to shelter us, the sea became even rougher and we could not see much. We gave them foil blankets and heat packs, but they were not much use… It was very cold, perhaps zero degrees. Some were so drenched they took off all their clothes… To keep them warm we made them rotate inside the cabin, but it was all very difficult. We were all feeling sick and scared. We feared for our lives… I felt so enraged: saving them and then seeing them dire like that…”
Salvatore Caputo, Italian Coast Guard

In an interview, the doctor on board of one of the patrol vessels, Gabriella Lattuca from the Order of Malta, concurred concerning with Caputo regarding the dramatic situation on board.   Following her account, the rescuers run out of survival blankets on board since these kept flying away due to strong winds. The “heating pads” that were distributed were not enough to keep the survivors warm and due to the dire meteorological conditions it was impossible to request aerial rescue forces. Among the eight members of the crew (six coast guard personnel, one doctor and one paramedic), all of whom were also feeling sick due to extreme conditions, only the scuba diver was equipped to go on deck to check the conditions of the passengers. The crew found out about the death of some of the passengers only when it was too late.

While from our reconstruction it appears clear that the Italian Coast Guard crew did all they possibly could to rescue the migrants in extremely difficult conditions, this was insufficient to avoid the death of 29 people, given the lack of larger SAR vessels and the long distance they had to travel.

Commenting on the incident, doctor Pietro Bartolo, chief health care official on Lampedusa, suggested:

“It is likely that with MN in place we would have not had these deaths: it is not possible to go and rescue migrants 100-120 nm from Lampedusa and to then carry them to Lampedusa in prohibitive weather conditions. MN allowed Navy ships to reach those wretched people, bring them on board, shelter and revive them. Now this is much more difficult.”
Pietro Bartolo, chief health care official on Lampedusa

As for the second rubber boat, according to Amnesty it was never found and left no survivors.

The third vessel was sighted at 13:40 GMT in pos. 33'56'N – 014’4'E by the Bremer Martha cargo ship with approximately 15 persons on board (FX). At 13:50 GMT, the offshore supply ship Bourbon Argos was redirected to the area. At 14:09 GMT, the Atlantic aircraft M.M. reported that there were approximately 30 migrants on board the aforementioned black rubber dinghy. However, when the Bourbon Argos and the Bremer Martha conducted the rescue, they realised that only seven passengers were still alive (AI and FX). At 16:47 GMT, the rescue operation by the two merchant vessels was complete (FX).

The fourth and last boat was localised at 34-12N 013-32E at 21.28 GMT, as stated in a hydrolant distress signal sent out by MRCC Rome to vessels transiting through the area at 05:05 GMT of the following day:

HYDROLANT 329/2015 (53,56)
INMARSAT-C: 424744220,
PHONE: 39 065 908 4527, 39 065 908 4409,
FAX: 390 6592 2737, 3906 5908 4793,
E-MAIL: [email protected]
( 090505Z FEB 2015 )

According to the Frontex report, which only provides scant information regarding this event, the Bourbon Argos tugboat was re-directed to this area, which it reached at 12.48 GMT, and recovered the two remaining passengers in 34'01'N - 013'52'E.

In total, the survivors interviewed by Amnesty believe that some 330 of their fellow travellers perished; they estimated that about 105 people were on board each of the four dinghies and only 86 people survived.

As in the 22 January case described above, the deaths of migrants in these four interrelated cases were directly linked to the gap in SAR capability left by the ending of MN and its (non-)replacement by Triton. While the smugglers behaved murderously by sending the migrants out in extreme meteorological conditions, the gap left in rescue capacity meant that there were no assets located near the Libyan coast, which might have allowed for the rescue of the migrants before the weather worsened. Even once they could be rescued, the unavailability of larger SAR vessels that could shelter the migrants below deck led to further deaths, as the case of the first boat shows clearly. With MN in place, it would have taken less time to reach the passengers; furthermore, the large military ships of the MN operation – some of which had medical operating rooms and doctors on board and could shelter large number of migrants below deck – would surely have provided better conditions to care for the passengers during the long hours of navigation back to Lampedusa. Finally, in addition to the vacuum left by the ending of MN, it appears that the two largest vessels part of Triton – the Icelandic Tyr and an Italian navy vessel, presumably the Libra – were in maintenance in Malta and at a port in Sicily and therefore unable to intervene, further diminishing the already reduced SAR capability in the central Mediterranean.   When contacted concerning this issue, Frontex provided the following explanation:

“We can confirm that two of Frontex co-financed assets were in the port at the time of the incident. The reasons were maintenance, crew change, bunkering etc. At the same two other open patrol vessels co-financed by Frontex were patrolling as scheduled, which allowed the rescue of the migrants in this incident and therefore didn’t jeopardize Frontex’s capacity to assist Italian authorities in SAR events. […] Every time a Frontex-deployed asset remains in the port for refuelling, crew change, refurbishing of food, medical supplies or technical maintenance, the national authorities are informed well in advance about this in order for them to take this into account when coordinating border control and search and rescue activities on their waters.”
Frontex Press Office

Both the 22 January and the 8–9 February cases, as well as the data on deaths at sea at the end of February 2015, were thus already proving the human rights community’s predictions correct: ending MN was not leading to less crossings or deaths at sea, but to continuing crossings in more dangerous conditions. The peak in mortality rate observed in February 2015 (see statistical annex) clearly indicated that what had been sound predictions were now an unfolding reality, of which the first signs announced more tragedies to come.

In the aftermath of these shipwrecks, several high level politicians and international organisations criticised Operation Triton, Frontex and the EU at large for failing to address the humanitarian catastrophe that was, once more, unfolding in the Mediterranean Sea. Laura Boldrini, then president of the Chambre of Deputies in Italy, tweeted: “Horror off Lampedusa. People didn’t die because of a shipwreck, but of cold. These are the consequences of the post-Mare Nostrum era.”   The UNHCR urged the EU to prioritise the saving of lives, with UN High Commissioner for Refugees, António Guterres, stating: “There can be no doubt left after this week’s events that Europe’s Operation Triton is a woefully inadequate replacement for Italy’s Mare Nostrum.”   Amnesty International called upon the EU to increase search and rescue capacities, suggesting that “the humanitarian crisis that sparked the need for Mare Nostrum has not gone away. […] EU member states must stop burying their heads in the sand whilst hundreds keep dying at sea.”   These calls, however, were ignored by EU policy makers and agencies, who took no action to prevent further tragedies.


With both the Italian government and EU institutions remaining firm in their decision not to (re-)launch a large-scale SAR operation that could fill the deadly gap left by the end of MN, MRCC Rome sought to patch it by relying on the inadequate means of commercial vessels. While commercial vessels were already involved in some of the cases above, an incident in early March offered a clear indication of the deadly effects this recourse could have.

03.03.2015 CASE

Position of three boats in distress and AIS tracks of vessels in the area. AIS data shows the OOC Cougar diverting its course towards the boat in distress after 11:08 GMT. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi.

On 3 March 2015, a shipwreck occurred just as a migrants’ vessel was reaching a tugboat that had come to rescue them, in a eerie anticipation of the events of the Black Week that we reconstruct in the next section. In order to reconstruct this case, we have relied on a detailed SAR report released to us by Frontex (FX), distress signals, AIS data, interviews with a few of the survivors conducted by Amnesty International (AI)   and photographs released by the company managing the vessel in question. The following chain of events emerges from these sources.

As the Frontex report indicates, the boat had left in the night of 2/3 March 2015 carrying 153 people, including some 4 women and 2 children. While the information in our possession does not allow us to clarify at what time a distress call was placed, at 11:49 GMT, MRCC Rome sent out the following distress signal, indicating the position of three boats in distress in the area (see map above):

HYDROLANT 524/2015 (52,53,56)
(Cancelled by HYDROLANT 606/2015)
33-23N 012-19E, 250 PERSONS ON BOARD.
33-17N 013-19E, 100 PERSONS ON BOARD.
33-16N 013-11E, 100 PERSONS ON BOARD.
INMARSAT-C: 424744220,
PHONE: 3906 5908 4527, 3906 5908 4409,
FAX: 390 6592 2737, 3906 5908 4793,
E-MAIL: [email protected]
( 031149Z MAR 2015 )

AIS data shows that after 11:08 GMT, the OOC Cougar changed direction and headed towards the location of the boat in distress “A” referred to in the distress signal.

At 13:20 GMT, the Frontex report states that a “wooden boat” was sighted by the tugboat OOC Cougar at position 33 27’N 012 24’E”, approximately 6 nm north-east of the position of boat A indicated on the Hydrolant message and showed on the map. We can thus assume that these positions refer to the same migrants’ boat. However, as the migrants’ boat approached the large tug boat (73.60 M X 16.0 M) and the passengers tried to catch the ladder that the crew of the tug boat had thrown to them, the movement of the passengers on board caused the boat to capsize – according to the Frontex report at 14:44 GMT.

A 25 year old Palestinian man from Lebanon interviewed by Amnesty International recalled the incident as follows:

“At 5pm [15:00 GMT], an American ship [editor’s note: the Liberian flag flown by the OOC Cougar is similar to the American one] was coming, we saw it. It came close to our boat. [...] They threw a rope ladder […]. Many tried to get on it and the boat capsized […]. I fell into the water, I was the first one. I couldn’t breathe. When we were in the water it was like a war scene. There were helicopters and boats around us […]. Immirdan, a Syrian woman, about 35, died with her one-year old son. They couldn’t swim. She had asked me for some bread, chocolate, cheese, I gave it to her. 20 minutes later, the boat capsized. I saw her.”
Survivor of the 3 March case

The Opielok company that manages the OOC Cougar has released several photographs that allow us to see how this tragic event unfolded:

Photographs taken by the crew of the OOC Cougar showing the sequence of events that led the boat to capsize just as it was approaching the supply ship to be rescued. Photo credit: OOC Opielok Offshore Carrier.

The photographs and their metadata show the migrants’ boat approaching at 14:36 GMT, and the crew on board the OOC Cougar swiftly lowering a rope to the migrants. We also see a crewmember with a megaphone communicating with the passengers. However, the next photograph taken at 14:48 GMT shows the passengers in the water, a life raft having been lowered. The last photograph that was released was taken after the rescue at 16:57 GMT and shows the passengers on board the OOC Cougar draped in survival blankets and the rafts being lowered. We also see another tugboat of the Asso family in the distance, probably the Asso Ventiquattro, which, according to AIS data, was present in vicinity.

According to the press release of the Italian Coast Guard, shortly after the migrants’ boat capsized, the Italian coast guard vessel, Dattilo (CP 940) joined the rescue efforts with 318 people rescued in a previous operation already on board. The Italian Navy vessels Fugolsi (financed by Frontex) and Bersagliere also took part. According to Frontex, At 21:49 GMT, 98 people had been rescued by the OOC Cougar, 12 by the Asso vessel, 11 people and 10 dead bodies recovered by the Bersagliere. Of the 153 people who were initially on board, more than 30 died.

According to the press release by the Italian Coast Guard, that day, 941 people were rescued in seven SAR events, all in an area close to 50 nm from the coast of Libya. Three different merchant ships, two coast guard vessels and one ship of the Italian Navy financed by Frontex participated in the rescue efforts of a total of seven migrants’ boats. This information does not only give a scale of the vast amount of SAR events that were happening at the time, but also of the huge involvement of the shipping industry.

The 3 March shipwreck was thus indicative not only of the growing trend towards the privatization of recue as the result of the retreat of state assets dedicated to SAR close to Libyan coast, but also of its potentially lethal effects.

The commercial shipping community had already emerged as a crucial actor during MN, when it started to be involved in rescue operations on an unprecedented level compared to previous years: over the whole year of 2014, according to the data compiled by the Italian Coast Guard, some 600 commercial vessels were diverted to rescue 42,061 people in 2014, representing 25 percent of the total number, second to the Italian Navy. However, as the data compiled by Frontex indicates, the share of commercial vessels (indicated as “civilian vessels”) grew during the phasing out of MN: in December 2014, it represented close to 40 percent of the total number diverted to conduct rescue operations. Thus, it is unsurprising that with the ending of Mare Nostrum, this trend grew and consolidated: of the 39,250 people rescued between 1 January and 20 May 2015, 11,954 were rescued by cargo ships. This represented 30 percent of the total of the rescued people, thus making the shipping industry the primary actor conducting rescue operations in the central Mediterranean.

Frontex figure showing that the number of people rescued by the shipping industry suddenly reached more than 20 percent in June 2014 and remained consistently high after that. Source: Frontex, Annual Risk Analysis 2015, April 2015, p. 20.  

Previously long opposed and criminalized, the mobilisation of merchant ships for rescue activities had since 2014 not only been encouraged but actively called upon by the Italian Coast Guard, and increasingly so since the end of MN. While commercial vessels have contributed to saving thousands of people, their involvement has also posed serious challenges in terms of safety. Commercial ships are not designed to safely approach boats that are much smaller, overcrowded and unstable. Furthermore, they often have a very limited crew, who are not specifically trained or equipped to carry out the extremely perilous operations necessary to rescue an overcrowded boat on the open seas.   As such, and without diminishing the importance of the efforts of the shipping community, it is not surprising that their massive involvement in rescue operations led to cases of shipwrecks. Despite their intentions, the commercial vessels became not merely involved in the rescue efforts, but in the sequence of events that led to the situation of distress and ultimately death of the migrants. Stuck between the shifting practices of smugglers whose operational mode had been adapted to the presence of MN on the one hand, and the EU policy makers’ reluctance to provide assistance at sea on the other, the excessive mobilisation of private vessels for rescue operations led to assistance becoming deadly.

The 3 March incident was not unprecedented in demonstrating the risks of privatised rescue. Already on 7 June 2014, a raft carrying 107 migrants overturned just as it was approaching the an oil tanker flying the Maltese flag, the Norient Star. The accident caused the death of 5 people.   While this incident was of too small a scale to attract public attention, the shipping community did warn early on that the burden it was being made to carry was too heavy.

The International Chamber of Shipping warned at the end of October 2014 that:

“The shipping industry is therefore very concerned by reports that the new EU Frontex operation ‘Triton’ will have a third of the budget of the current Italian ‘Mare Nostrum’ operation which it replaces, that its primary focus will be border control, and that search and rescue operations may be reduced in international waters. It will clearly be much more difficult for merchant ships to save lives at sea without the adequate provision of search and rescue services by EU Member States.”, (last accessed 12 April 2016)." data-html="true">
International Chamber of Shipping

Frontex was well aware that by pulling state-operated vessels back from the Libyan coast, a larger burden would be left to the shipping community. In its “tactical assessment” for Triton dated 14 January 2015 (see annex), it noted:

“In 2014, it became apparent that facilitation networks were exploiting the presence of civilian vessels in the central Mediterranean area because according to international maritime law, when a distress call is transmitted, the ship, being the nearest, is obliged to render assistance to those in distress at sea and to deliver them to a place of safety. In this regard, it is assumed that facilitation networks will continue to exploit the presence of civilian merchant ships in the central Mediterranean during 2015 to reach Italy.”
Frontex Tactical Focused Assessment

Following this event, and the materialisation of the increasing recourse to the shipping industry, a coalition of shipping industry organisations, led by European Community Shipowners’ Associations (ECSA) and the International Chamber of Shipping (ICS), wrote an open letter to the EU member states and institutions on 31 March 2015.   Given that this call was premonitory but was not headed to, it is worth quoting at length:

“The humanitarian crisis in the Mediterranean Sea is spiralling out of control. […] There is a terrible risk of further catastrophic loss of life as ever-more desperate people attempt this deadly sea crossing. […] We believe it is unacceptable that the international community is increasingly relying on merchant ships and seafarers to undertake more and more large-scale rescues […]. Commercial ships are not equipped to undertake such large-scale rescues […]. In the short term, we therefore feel that the immediate priority must be for EU and EEA Member States to increase resources and support for Search and Rescue operations in the Mediterranean, in view of the very large number of potentially dangerous rescues now being conducted by merchant ships […]. In addition to increasing SAR resources, there is also an urgent need for EU and EEA Member States to develop a political solution. […] The shipping industry believes that the EU and the international community need to provide refugees and migrants with alternative means of finding safety without risking their lives by crossing the Mediterranean in unseaworthy boats.”  
European Community Shipowners’ Associations and the International Chamber of Shipping

This letter, written less than two weeks before the large scale shipwrecks of April, by the actors who, despite themselves, were being put on the frontline of rescue in the central Mediterranean as the result of the retreat of EU member states and agencies, could not be more clear in terms of the increasing risk entailed by the excessive mobilisation of merchant ships. But just as with the warnings that followed the 8 February shipwrecks that had resulted from the vacuum in SAR capability, these signals and calls denouncing the dangerous privatization of rescue used to fill it were left attended to and unanswered. In this context, the April shipwrecks were only waiting to happen.


The week starting on 12 April 2015 saw two successive shipwrecks, which together cost the lives of more than 1,200 people. The tragedy that had been announced by the shipwrecks of early 2015 and predicted by the human rights community, the shipping industry and MEPs, had taken place.

The month of April usually sees a strong increase in the number of crossings, with migrants taking advantage of improving meteorological conditions after the winter months. 2015 was no exception: the arrivals in Italy went from 2,283 in March to 16,063 by the end of April according to UNHCR data. More than 8,000 people were rescued in the four days from 10 to 13 April 2015 alone. The 12 April represented a peak, with 3,791 people rescued that day.   Given the scale of the events, it is not surprising that the limited SAR capacities left after the start of Triton proved inadequate, leading to the 12 April shipwreck.

12.04.2015 case

Animation based on Automatic Identification System (AIS) vessel tracking data of the 12 April shipwreck. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi. Data: MarineTraffic

The reconstruction of this case is based on the following sources: interviews conducted by Forensic Oceanography with two groups of survivors in a reception centre in Italy – the first composed of fourteen English speakers (abbreviated as ENG), and the second of seven French speakers (FR);   AIS vessel tracking data analysis; photographs taken during the events by one of the intervening ships; and SAR reports.

The migrants’ vessel departed from the port of Zuwara in the night of 11/12 April 2015. Before the departure, some had been waiting in a safe house in Tripoli for a week (FR 18). As of midnight, the migrants were taken in small rubber boats towards a bigger wooden fishing boat with three decks (FR 8, 43). There were between 600 and 700 people on board (FR 56). The boat left Zuwara around 03:00 GMT on 12 April.

However, shortly after having departed, the boat started encountering difficulties. Engine problems appeared to have taken one to two hours to repair (ENG 43), and the bilge pump stopped functioning and vessel began to take on water. One survivor recalled: “Inside the boat, they say there is a pump that takes the water out, but when this problem occurred the pump stopped so the boat was taking a lot of water, and people were just taking the water out manually, using clothes and bottles” (ENG 43–46).

According to survivors who were located on the upper deck of the boat near the captain, at around 12:00 GMT a distress call was placed. One of the survivors recalled the interaction as follows:

“He called the Red Cross, he was told that he hadn’t arrived yet, he should continue. We continued until 13:00 and we called them again, he was told that he should continue, that he only had 110 km left before being taken. Then he called again, and he was told that there were only 20 km left. Then, as we were moving, we saw boats” (FR 64–68).
Survivor of the 12 April shipwreck

As the Italian Coast Guard has not provided us with the full SAR report concerning this event, we do not know if they took any specific action at this stage when the distress call was made. It appears that since the migrants’ vessel was located close to the Libyan coast and was still able to advance with a functioning engine, the Coast Guard indicated to the driver to continue navigating in the direction of Lampedusa so as to come closer to SAR vessels – of which very few are accounted for by the AIS data (see for example vessel CP 324 departing Lampedusa after the boat has capsized).

Automatic Identification System (AIS) vessel tracks in the Mediterranean at the time of the 12 April shipwreck. The vessels that first arrived on scene are underlined. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi.

However, as migrant vessel continued northward, some time after 13:00 GMT, they saw other vessels in the distance and began to discern an oil platform surrounded by ships (FR 70–77; ENG 188). This is collaborated by the AIS data, which shows the platform ENSCO 5004 in operation at the Bouri oil fields less than 1 km from the location of the migrant vessel capsizing. As the boat approached the nearby ships, doubts emerged between the migrants on board as to whether these would be Libyan or of other nationalities. The boat advanced slowly, stopping and starting again, to the rhythm of the arguments of the passengers. As they approached the unknown vessels near the oil rig many migrants moved to see the unknown vessels and the now unstable migrant vessel began to roll to the port side (left). Some migrants attempted to return their vessel to a point of equilibrium, but the unstable, overloaded migrant vessel continued to roll over and capsized (FR 80–90). The position later provided in a distress signal by the Maritime Rescue Coordination Centre (MRCC) in Rome as the location of the event was 34-05.0N 012-24.8E. AIS data shows the tracks of the vessels present at the time and location of the event.

Automatic Identification System (AIS) vessel tracks of the three vessels participating in the SAR operations after the the 12 April shipwreck. The red squares mark the four locations from which the OOC Jaguar took the pictures analysed below. The white square marks the point of capsizing of the migrants’ boat. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi

Comparing surviving migrant testimony and the AIS data allows us to corroborate the identity of the vessels involved. The survivors recalled the colour of the first ship they encountered as “blue” (FR 95; ENG 134), and of a shape characteristic of oil platform supply vessels. The presence of the writing, “St. John’s” and “Rescue zone” on the side of the first ship was confirmed by all testimonies. The OOC Jaguar tug, accounted for by AIS data, is indeed a blue supply vessel and bears the writing “St. John’s” on its hull, below its name. This first vessel would then be followed by two other vessels that the survivors remembered as “red” (FR 137; ENG 188); the Asso Ventuno and Asso Ventiquattro, the presence of both of which has been documented by AIS data, have a red hull. Moreover, the Asso Ventiquattro bears the writing “Rescue Zone” and, when shown the pictures of the Asso Ventuno, all migrants recognized this as the ship that assisted in the rescue operations.

The OOC Jaguar (72mx16m) and the Asso Ventiquattro (70m × 17m). Photo credits: and

Photographs released to the press in the immediate aftermath of the shipwreck by Opielok Offshore Carriers GmbH, the company managing the OOC Jaguar, provide further evidence of the unfolding of events. The first photograph shows the migrants’ boat clearly tipping to the left side in the distance, and this despite a calm sea (see the oceanographic analysis in annex). By extracting the metadata from the digital photograph’s file, we can determine that this photograph was taken with a small compact camera at 17:11:44. While the metadata does not specify the time zone, we assess this to be Central European Time, which at the time was GMT+2. Thus, the photograph was taken at 15:11:44 GMT. The focal length used is 15.40 mm or 87 mm in 35 mm equivalent focal length, what is normally considered a long-focus lens, and the angle of vision in relation to the surface of the water is low as the horizon remains visible. This indicates that the photograph was taken when the OOC Jaguar was still relatively far from the migrants’ boat. This is corroborated by the vessel’s location according to the AIS track at the time the photograph was taken.

Photograph of the unstable migrant vessel taken by the crew of the OOC Jaguar at 15:11:44 GMT. Photo credit: OOC Opielok Offshore Carrier.

The next photograph provided by Opielok Offshore Carriers GmbH shows the boat already having capsized and with the passengers in the water, some of them still clinging to the overturned boat, others adrift. The photograph’s metadata indicates this image was taken at 15:54:49 GMT. This time the focal length is 10.70 mm or 60 mm in 35 mm equivalent focal length, close to what is considered a normal lens, and the angle of vision is higher in relation to the surface of the water (no horizon is visible), indicating that the photograph was taken when the boat was very close to the migrants. Once again, this is corroborated by the position at which the vessel was located according to the AIS track at the time the photograph was taken.

Photograph taken by the crew of the OOC Jaguar at 15:54:49 GMT. Photo credit: OOC Opielok Offshore Carrier.

The sequence of events that occurred in the time between the two released photographs – the 43 minutes that ultimately saw the boat capsize – is as of yet not entirely clarified. Important elements that would be necessary for a complete reconstruction are missing. In particular, MRCC Rome has refused us access to the detailed SAR report for this event, while, despite a productive dialogue, Opielok Offshore Carriers GmbH has not yet shared with us the elements of evidence they possess (such as the ship’s log book, VHF voice recordings and the testimony of the captain). What we offer here is thus a tentative reconstruction based mainly on survivor testimonies that may be amended should we receive further information from the other actors directly involved.

One of the survivors located near the migrant vessel captain recalls how their boat approached the OOC Jaguar as follows:

“They were not coming towards us, so now we had to start our engine again, to move towards them […], but at that time, the boat was tilting to one side and then tilting to the other side again, and we moved until we were very closed to them now. Then everyone, those from the lower deck, started to come outside, and those from the top were starting to come down, because everyone was hassling to be rescued first, so it was like a push and pull. People were shouting at each other, ‘go down go to your place!’, but no one was listening to one another, and then when the boat tilted to this side, and then everyone shouted, ‘return to your place’, and when they came back to the other side, then the boat sank” (ENG 95-105).
Survivor of the 12 April shipwreck

The francophone survivors described a similar sequence of events, but did not specify their timing (FR 81–91). It appears that as the migrants moved closer to port side of their vessel to be rescued, they became agitated and the agitation grew out of control. It is precisely to avoid such situations that professional rescuers often keep their vessel at a distance from a boat in distress and deploy one or two Rigid Hull Inflatable Boats (RHIB) towards it. This allows personnel on board the RHIB to assess the situation and maintain order on board the boat.   The trained personnel necessary to implement this particular knowhow so as to manage the extremely difficult and dangerous rescue of an overcrowded and unseaworthy boat, however, were not available to the OOC Jaguar.

While the exact moment of capsizing in not known to us, the behaviour of the OOC Jaguar would seem to indicate that it occurred shortly after 15:30 GMT. The AIS track of the ship, which until that moment had been advancing at minimal speed, shows in fact at 15:36 GMT a manoeuver to approach the position of the boat when it capsized, with the OOC Jaguar accelerating considerably. A few minutes later, at 15:48 GMT, the vessel slows down again and takes a sharp turn (or moves back in reverse). This might be a manoeuvre to back off from the location of boat just after it capsized. Finally, at 15:51 GMT, it moves forward again the capsized location. It is at this time that the second and third photographs were taken showing the capsized boat.

Several of the survivors interviewed by Forensic Oceanography have claimed that a relatively long time (15 to 30 minutes, ENG 263) passed before the crew of the OOC Jaguar began to carry out a rescue operation by deploying its RHIB. During this time, the survivors claim, the crew of the OOC had been taking pictures but not providing them with assistance.

Since evidence from MRCC Rome and Opielok Offshore Carriers GmbH has not been released to us, it is not possible to corroborate or disprove this account, or to explain the actions the crew was undertaking during this time to deploy its rescue boats (such as communicating with MRCC Rome, preparing the crew to respond and preparing the life rafts). However, a third photograph was taken at 15:54:53 GMT, seconds after the previous one showing migrants clinging onto the capsized boat, but showing a different perspective (the horizon is visible again). This time, the photograph shows a small rescue boat approaching the drifting migrants.

Photograph taken by the crew of the OOC Jaguar at 15:54:53 GMT. Photo credit: OOC Opielok Offshore Carrier.

A fourth photograph taken at 16:45:51 GMT shows the small rescue boat aside the OOC Jaguar with survivors on board. Without knowing with certainty the exact time at which the boat capsized, we cannot estimate the time that was necessary for the crew of the OOC Jaguar to deploy its rescue boat.

Photograph taken by the crew of the OOC Jaguar at 16:45:51 GMT. Photo credit: OOC Opielok Offshore Carrier.

Between 16:35 and 16:43 GMT, two other tugboats, the Asso Ventuno and Asso Ventiquattro, arrived on scene and started to carry out rescue operations, as their AIS tracks indicate. Immediately before that, at 16:37 GMT, MRCC Rome had sent out a distress signal to all vessels in the vicinity requesting assistance for the capsized passengers and indicating the following position: 34-05.0N 012-24.8E.

HYDROLANT 865/2015 (52,53,56)
DNC 08,
DNC 09. F/V,
CAPSIZED AND ADRIFT IN 34-05.0N 012-24.8E.
INMARSAT-C: 424744220,
PHONE: 39 065 908 4527, 39 065 908 4409,
FAX: 39 06 592 2737, 39 065 908 4793,
E-MAIL: [email protected]
( 121637Z APR 2015 )

According to the Frontex SAR report, further assets were deployed to assist with the rescue operations: the Italian Coast Guard fixed wing aircraft “Manta 10-03”, the CP 324 vessel (financed by Frontex), the CP 287 and the Italian Navy ship Bersagliere, which is also depicted as present on the scene in another photo taken from the OOC Jaguar.

Photograph taken by the crew of the OOC Jaguar during the rescue operations. As no metadata is available for this photo, it is not possible to determine with precision the exact time at which it was taken. Photo credit: OOC Opielok Offshore Carrier.

What is clear, however, is that the only available rescue vessels were not sufficient to respond to the numerous people adrift at sea and desperately trying to stay afloat without life jackets. As already recalled, at the same time many other vessels were providing assistance to various migrants’ boats in distress at other locations.

Video by the Italian Coast Guard showing that at exactly the same of the events, the Coast Guard vessel CP 920 and the Amazigh cargo ship were undertaking another rescue operation not far from the location of the events we have described. Video credit: Italian Coast Guard

All 145 rescued migrants and 1 dead body were transhipped onto the Navy vessel, while 8 other bodies were transhipped onto the CP287. Considering that there were at least 600 people on board the boat (FR 56), the death toll surely surpasses the 400 mark.

While questions remain open concerning the precise unfolding of the events that ultimately led to this capsizing accident, it appears that the crew of the OOC Jaguar did everything it could to rescue the passengers in distress both before and after the boat capsized. While preventing the terribly overcrowded boat from overturning and sinking would have been challenging even with the highest level of training and most adapted rescue means, the means and knowhow available to the crew of the OOC Jaguar imposed further limitations on the rescue effort that ultimately contributed to the tragic outcome.

The worst predictions of the human rights community were thus taking shape. However, these events were still not enough to spur EU member states and agencies into taking urgent action. One week later, an even greater toll in human lives would be taken in another case of “death by rescue”.

18 April 2015 case

Animation based on Automatic Identification System (AIS) vessel tracking data of the 18 April shipwreck. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi. Data: MarineTraffic

On 18 April 2015, a boat carrying more than 800 people capsized after ramming into the King Jacob, a large commercial vessel that was approaching it to carry out a rescue operation. Only 28 people could be rescued and 24 bodies were retrieved. It is thus believed that close to 800 people died in this incident.

The reconstruction of this case is based on interviews we conducted in Catania in October 2015 with two French-speaking survivors, S. from Ivory Coast and O. from Senegal (abbreviated as SO), and with an English-speaking survivor, S. from Sierra Leone (abbreviated as S).   We then corroborated their testimonies with further elements of evidence. We first relied on materials produced in the frame of the legal proceedings launched by the Tribunal of Catania against those accused of being the smugglers driving the boat (“Ordinanza Mohammed Ali Malek”, abbreviated as OMAM, and documents provided by the defense lawyer, Avv. Ferrante, abbreviated as AF). We also used Automated Identification System (AIS) vessel tracking data and a Hydrolant distress signal. We should note here that neither the Italian Coast Guard nor OSM Shipping GmbH & Co. KG, the company that managed the King Jacob, have shared any evidence directly with us. This lack of transparency, however, has not prevented us from achieving a coherent understanding of the unfolding of events. The following chain of events emerges from the combination of these different sources.

Interview with the survivors of the 18 April shipwreck. Credit: Forensic Oceanography. Editing: Laure Vermeersch. Research assistant: Sabine Llewellyn

The two survivors from Senegal and Cote d’Ivoire recounted to us that, prior to embarking, they had been held captive in a detention centre in Tripoli, from where, at around 19:00 GMT on 17 April 2015, they and more than one hundred fellow captives were brought by the guards directly to the beach in trucks (SO 16–28). They remembered the beach as that of “Garapoli”, probably corresponding to the town of Garabulli (SO 67) in the East of Tripoli. When they arrived on the beach, they saw many other people already waiting in lines, under the orders of 11 to12 armed smugglers. Around midnight, they began bringing the migrants in small zodiac boats containing around 100 people at a time to the larger fishing vessel anchored off the coast, a process that lasted until 03:00–04:00 GMT on the morning of 18 April 2015 (SO 124–34). Based on the number of journeys needed to fill the boat, the survivors estimated that there were more than 800 people in total on board (SO 99–109). As they boarded the large fishing boat, they were dispatched to different sections, with two of the interviewees being seated in the top cabin, close to the driver (S 195; SO 198–203). As they boarded the wooden fishing vessel, O. remembered thinking: “It’s over. Deaths awaits us. […] I did not believe that we would survive. Because so many people like this, on a boat, we were not well organised. […] It seemed like they wanted to kill us even, because they threw us to the sea” (SO 256–60).

No photograph of the fishing vessel has been accessed to date, and it is probable that none exists just prior to the capsizing since the incident occurred during the night. However, an underwater expedition directed by the Italian Navy was able to locate the wreck, provide a sonar estimation of its dimensions (length: 21 m; width: 8 m; height of the hull above the seabed: 8 m), identify signs of collision compatible with the reconstruction provided by the captain of the King Jacob (an impact on the boat’s bow, damage on its left side and the balustrade collapsed) and ascertain the presence of several bodies inside the boat. While the complete documentation was not released to us, the images released publicly by the Italian Navy show the hull of the boat in resting on the sea floor.

Underwater video of the 18 April 2015 shipwreck. Credit: Italian Navy.

Around 04:00 GMT on 18 April, the migrants’ boat started navigating. It advanced slowly throughout the day, with the boat appearing unstable from the start (SO 262). Around 17:00 GMT on the same day, SO recall the driver placing a call with a phone to request assistance (SO 293). Based on reports to the prosecutor of Catania, as well as further evidence, this event and its timing was confirmed by Maritime Rescue Coordination Centre (MRCC) in Rome in the context of the already mentioned legal proceedings: at 17:35 GMT a distress call was received by MRCC Rome from a boat located within the (non-declared) Libyan SAR zone. At 18:43 GMT, MRCC Rome was able to locate the exact position at the moment of the distress call at coordinates: Lat. 33°51’50’’ N – Long. 014°26’13’’ E (AF).

Automatic Identification System (AIS) vessel tracks in the Mediterranean at the time of the 18 April shipwreck. The vessels that first arrived on scene are underlined. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi.

The following events unfolded as a response to the distress call. According to the testimony of the captain of the King Jacob, a large cargo ship 146.42 m long and 22.7 m   wide flying the Portuguese flag operated at the time by OSM Shipping GmbH & Co. KG, at 19:00 GMT, he received a call from the Italian Coast Guard (OMAM 17) and diverted the the King Jacob’s course to the location of the vessel in distress. AIS data shows the vessel already heading towards the location of distress at 18:52 GMT. MRCC Rome also dispatched the Italian Coast Guard vessel Gregoretti (CP 920) to the location of the call (OMAM 2), which is accounted for in the AIS data as navigating towards the location of distress as of at least 19:30 GMT (no data is available for before this). Moreover, AIS data also shows the vehicle carrier City of Lutece also navigating to the distress location, although from the information currently in our possession it is unclear when it was actually contacted by the MRCC in Rome.

According to the survivors we met, after contacting MRCC Rome, the migrants’ boat remained adrift with the engines shut down for some time, until around 19:00 GMT, when the survivors recall a helicopter that hovered over their boat and left (SO 304). After this, they resumed navigation.

The King Jacob, a 146.42 m-long and 22.7 m-wide cargo ship. Photo credit: Maritime Connector.

Around 21:00 GMT, according to the captain of the King Jacob, while the migrants’ boat was not observed directly in the dark, the cargo ship’s radar indicated the presence of a small vessel 6 nm away, towards which the Portuguese ship continued to navigate (OMAM 18). The migrants’ boat was proceeding without lights (OMAM 14–15). This timing corresponds to that remembered by the survivors who describe seeing a large vessel in the distance at around 21:00 GMT. The driver lit a small torch to attract the vessel’s attention (SO 323–28). This probably corresponds to the moment at which, according to the testimony of the captain of the King Jacob, a small light coming from the migrants’ boat was first sighted at a distance of three nm. At that point, the captain of the King Jacob gave orders to switch on the right spotlight of the ship, which was not able, however, to shed light on the boat. The left spotlight was then also switched on, while the ship continued approaching the small light coming from the boat (OMAM 18). At 21:11 GMT, at the moment of this first sighting, the King Jacob provided to the Italian Coast Guard vessel Gregoretti the following position for the migrants’ boat: Lat. 33°56’ N – Long. 014°28’ E (AF).

Automatic Identification System (AIS) vessel tracks of the first three vessels in the 18 April shipwreck. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi.

Between 21:11 and 21:20 GMT, at about 1 nm from the migrants’ boat, the captain of the King Jacob realized that the boat was terribly overloaded and therefore changed course four times in order to avoid collision (for a total duration of about 8 minutes) since the migrants’ boat continued to follow the King Jacob with a course that would inevitably lead to collision. After this, the King Jacob switched off its engines at the following coordinates: Lat. 33°51'.9N - Long. 14°26'.2E. (OMAM 18). The AIS data does not allow to show each one of these maneuvers, but the King Jacob vessel’s track is compatible with the overall description of events. At 21:20 GMT, the captain of the King Jacob ordered all the crew to go onto the main deck to assist in the embarkation of the migrants. According to his testimony, the ships’ spotlights continued to illuminate the migrants’ boat. While the latter initially moved slowly towards the cargo ship, all of a sudden, at a distance of about 100 m, the migrants’ boat accelerated and changed course, moving towards the cargo ship’s left side and eventually crashing its bow against it. After the collision, came up alongside the King Jacob, rubbing against its side in a backwards motion. According to the captain of the King Jacob, at the time of the collision the engines were turned off and the ship was not moving, so the backward movement of the migrants’ boat was due to the latter boat reversing. Lawyer Ferrante disagreed with this reconstruction, claiming that his client (the driver of the migrants’ boat) did not activate the engines and that it was indeed the King Jacob that was moving headway without propulsion. While the exact dynamic of the impact is still being ascertained, what is certain is that the migrants’ boat then capsized on its right side and sank within 5 minutes (OMAM 18).

The survivors interviewed for this report recalled the incident in a similar way; however, they recalled that the spotlight was turned on just as the boat began advancing, spurring the reaction of the captain (SO 378–93, see also S 246–49). One of the survivors drew a sketch of the collision, describing it in these terms:

“The other boat stopped. It stopped 5 minutes something like that. We continued advancing slowly. The other boat put its spotlight on us. At that movement, our diver accelerated, he accelerated until here, and hit the boat in the middle and slid all the way to here, and sunk” (SO 426–33).
O., survivor of the 18 April shipwreck
Interview with the survivors of the 18 April shipwreck, here drawing the trajectory of the vessels involved in the collision. Credit: Forensic Oceanography. Editing: Laure Vermeersch. Research assistant: Sabine Llewellyn

While some of the survivors’ testimonies collected by the Public Prosecutor's office in Catania do not mention a variation of speed in the moments immediately preceding the impact (OMAM 14–15), it was recalled by survivor K., who attributed the wrong manoeuvres that led to the collision to the fact that the captain had been trying to “escape” once he realized that the cargo ship’s crew was not Italian. Furthermore, another survivor claimed that as the migrants’ boat was approaching, all of a sudden it accelerated as if it wanted to escape (source of both testimonies: OMAM 14–15). In addition, all survivors heard by the Procura agree that there were three collisions between the two ships (OMAM 14–16). Traces of the collision were reported on the side of the King Jacob upon its inspection in port.

S., the survivor from Sierra Leone interviewed by Forensic Oceanography, was thrown from the migrants’ boat by the collision, and landed on the deck of the cargo ship. From there, he could observe the tragic scene unfolding before his eyes:

“From there I sat down, watching people going. People are shouting ‘Allahu Akbar’!, ‘Help me, help me, help me!’. [] I started seeing two-thee friends now are swimming, and the boat started to throw ropes for people. People came climbed up. [….] the boat started to go down slowly, slowly, slowly in the sea, while people are coming out the boat, until they stood at the tip of the boat, before the boat went down. I saw everything.” (S 263-275)
S., survivor of the 18 April shipwreck

Shortly after the collision, the King Jacob carried out a so-called Williamson turn to go back to the place of the shipwreck and search for people in the water (AF), which we can see indicated in the loop formed by the AIS data track within the 20 minutes following the moment of collision. At 21:28 GMT, the King Jacob communicated to Italian Coast Guard vessel Gregoretti that there had been a collision with the migrants’ boat during the approach phase (AF). The position of the collision was determined by the Gregoretti with its onboard instruments as Lat. 33°57.5’ N – Long. 014°28.3’ E (AF). At 21:32 GMT, the intervention of a helicopter was requested to search for the shipwrecked (OMAM 2). At 21:42 GMT, the Gregoretti, which in the meantime was appointed “On Scene Commander”, communicated that it had rescued 10 people (OMAM 2). By 22:25 GMT, 28 people had been rescued (one of them was evacuated by helicopter to the Cannizzaro hospital in Catania) and 24 bodies retrieved. The bodies were taken by the Gregoretti ship to Malta, while the survivors were taken to Catania (OMAM 2).

Despite the rescue efforts, the death toll in this shipwreck was simply cataclysmic. As with the 12 April shipwreck, it appears that no direct responsibility for these tragic events can be attributed to the King Jacob and its crew. This was also the conclusion reached by the public prosecutor in Catania. The King Jacob is a 147 m long container ship of, i.e. approximately seven times longer than the migrants’ boat it was tasked to rescue. Its sides are very high and utterly unfit for boarding a boat of that size and transferring its passengers. The crew of such a ship normally counts no more than 20 people, who usually lack specific training for rescue operations of this kind. Moreover, despite the calm sea (see oceanographic analysis in annex), the darkness certainly did not help to facilitate the rescue operation. What appears certain is that, considering the horrendous conditions in which the migrants were being sent out to the sea, only the highest level of professionalism and the most adapted means could hope to lead to safe rescue operations.

In the week between 12 and 18 April 2015, more than 1200 people had been swallowed by the currents as a result of their boats capsizing just as they were about to be rescued. Beyond the huge death toll, what is most striking about these events is that they were not the result of the reluctance to carry out rescue operations, which has been identified as a structural cause of migrants’ deaths in the Mediterranean Sea.   In these two cases, the actual loss of life has occurred during and partly through the rescue operation itself.

While it could appear that only the ruthless smugglers who overcrowded unseaworthy boats to the point of collapse are to blame, in light of the sequence of events and policy decisions we have reconstructed in our report, we can see another level of political responsibility that created the conditions in which these tragedies were bound to occur. The EU’s policy of retreat of state-operated rescue at sea left ill-adapted commercial vessels to bear most of the responsibility for rescuing and this, in turn, led to assistance becoming deadly. Death by rescue was thus the outcome of the EU’s policies of non-assistance.

The disastrous effects of this phenomenon are tragically confirmed by the data on arrivals, crossing and deaths at sea for the first four months of 2015: while in the first four months of 2014, more than 26,000 had crossed and 60 deaths had been recorded, in the same period of 2015 an almost identical number of crossings had occurred, but the number deaths had increased to 1,687. The probability of dying at sea was had thus increased 30 fold, jumping from 2 deaths in 1000 crossings to 60 in 1000 (see statistical annex).

Migrant mortality rate for 2014 and 2015 compared, on the basis of UNHCR and IOM data.

The 18 April 2015 tragedy led to a public outcry and numerous scandalised statements. The UNHCR titled its 19 April 2015 press release “New Mediterranean boat tragedy may be biggest ever”, and called for urgent action by the EU.  

Despite the statements expressing their deep concern for the dead, little consideration was given to the survivors. Despite their trauma, nineteen out of the 28 survivors were brought to the infamous centre for asylum seekers (CARA) in Mineo, one of the biggest in Europe. Among them were the survivors we interviewed. The Territorial Commission, responsible for examining asylum requests in Italy, recommended the Police headquarter to grant them with a resident permit on humanitarian grounds, which would have allowed them to leave Mineo. Furthermore, on the 22 April, Domenico Lucano, Mayor of Riace, offered to host them in his town, which has become in recent years an example of hospitality for refugees. However, nobody informed the survivors of these possibilities until six months later, they met a representative of the Catania Antiracist League, on the occasion of our interview. The survivors of the 18 April tragedy were confronted with the same indifference and precarity faced by thousands of others migrants who have arrived in Italy in the recent years.

After the Shipwrecks

Synthetic map and figures of the situation in the central Mediterranean, June – September 2015. Frontex and Coast Guard data show that migrants continued to be mainly rescued very close to Libyan shores, as was the case during 2014 when the Mare Nostrum operational was in place. The Triton operational area was extended further south up to the limit of Malta SAR zone, and further joined by the EUNAVFOR MED and nongovernmental rescue operations operating very close to the Libyan coast. In this phase, the share of rescues operated by all the above-mentioned actors increased while the share of commercial vessels dropped from 30 (Jan-May) to 4 percent. Crossings, deaths and mortality rates resumed to levels comparable to those recorded during Mare Nostrum in 2014. Credit: Forensic Oceanography. GIS analysis: Rossana Padeletti. Design: Samaneh Moafi.

Like the prior twin shipwrecks of October 2013, the April shipwrecks caused a public outcry and a policy upheaval. At the debate in the European Parliament on 29 April 2015, the President of the European Commission, Jean-Claude Juncker, gave a speech in which he admitted that:

“It was a serious mistake to bring the Mare Nostrum operation to an end. It cost human lives.”  
Jean-Claude Juncker, President of the European Commission

However, the ending of Mare Nostrum cannot adequately be described as a “mistake”. It was a clear decision taken by the Italian government, to which EU policymakers and agencies responded in a tragically inadequate way by (non-)replacing it with Frontex’s more limited Triton operation. As we have demonstrated in the successive sections of our report, EU policy makers and agencies were not now “discovering” the terrible reality they had brought fourth inadvertently. Even more so after the tragic events of the Black Week, this policy and its effects must be characterized as an act of killing by omission.

We could expect that the conduct of the agency that played a leading role in shaping policy-makers decisions towards what was described as a “mistake” by the Commission’s own president, i.e. Frontex, would fall under severe scrutiny and be deeply overhauled. Instead, in the very same speech Juncker announced that Frontex’s budget would be trebled.

At the same time, however, Juncker also announced that Triton’s operational zone would be expanded as a way of “restoring something that we had lost along the way.” This was the first major shift produced by the April shipwrecks. On 13 May 2015, the European Commission declared that “search and rescue efforts will be stepped up to restore the level of intervention provided under the former Italian ‘Mare Nostrum’ operation.”   On 26 May, Frontex adopted a new operational plan for Operation Triton, with an increased budget, additional assets and an expanded operational area from 30 up to 138 nautical miles south of Lampedusa, almost reaching the extent that had been covered earlier by MN.   However, contrary to Juncker’s and subsequent statements’ description of Frontex’s expansion as “restoring” the situation prior to the ending of MN, we should note that Triton did not become a European Mare Nostrum. While Frontex’s operational zone was expanded further south, it still did not reach the extent Mare Nostrum had. Furthermore, its operational priority continued to be border control as opposed to saving lives, and its assets would thus continue not be deployed proactively towards SAR.

A second impressive shift bought about by the Black Week has been the dramatic decrease in the rate of mobilisation of commercial ships for the purpose of rescue operations: the number of people rescued by commercial ships went down from 11,954 in the first five months of 2015 to only 3,689 from June to September, thus dropping from a contribution of 30 percent of the total of all rescues to 4 percent (Italian Coast Guard data, see annex).

Thirdly, the institutional process that saw the extension of Triton also led to the decision to launch a novel military operation at sea: the European Union military operation in the Southern Central Mediterranean (EUNAVFOR MED). The operation was launched following a statement on 23 April by the EU Council that mentions the EU’s commitment to “undertake systematic efforts to identify, capture and destroy vessels before they are used by traffickers” and to “disrupt trafficking networks, bring the perpetrators to justice and seize their assets.”   To this effect, Federica Mogherini, the EU’s chief foreign and security policy coordinator, was “invited to immediately begin preparations for a possible CSDP (Common Security and Defence Policy) operation.”   In the following weeks, the operation was further defined according to three operational phases, ranging from surveillance activities, interception and destruction of vessels used for smuggling on the high seas, to direct military action against smugglers inside Libyan territorial waters.   From the formal beginning of EUNAVFOR MED on 22nd June 2015, the operation’s command started coordinating the several military vessels that had been deployed by different states in a more or less chaotic manner in the immediate aftermath of the April shipwrecks.   With at least five planes and four ships, deployed by twenty-two different countries on a rotational basis close to the Libyan cost, EUNAVFOR MED came to reconstitute the naval force that the ending of MN had left vacant.   However, just like the Triton operation, at the core of its mission have been security concerns, not humanitarian ones. As such, rather than a “humanitarian and military” operation similar to MN, at work has been “a police operation with military means”, as Rear Admiral Hervé Bléjean, the Deputy Operation Commander in the Mediterranean, describes it. “The adversaries”, he contended, “are not combatants but criminals, and the aim is not to eliminate them but to bring them to justice.”   As a consequence, saving the lives of migrants has been far from the mission’s operational priority. This was clearly illustrated when it was revealed that the UK’s HMS Entreprise had not rescued a single migrant after almost eight weeks of deployment on intelligence-gathering missions near the Libyan coast.   While it appears that after the summer more rescue operations were conducted, and by the end of 2015 (after six months of activity) 8,500 people had been rescued by assets operating within EUNAVFOR MED;   this number pales in relation to the rescue operated by Médecins Sans Frontières (MSF) for example, which in eight months of activity rescued 20,129 people.  

Cloud of smoke left by the destruction of a boat that had been used by migrants after they had been rescued. Photograph credit: MSF.

The fourth and last major shift that followed the April shipwrecks is connected precisely to the intervention of independent NGOs in the Mediterranean. While the operational priority of providing assistance to migrants at sea was stripped of the mission of state-led operations Triton and EUNAVFOR MED, non-governmental humanitarian actors took the initiative to launch a series of rescue operations, constituting a veritable civilian rescue flotilla.   In early April 2015, MSF (Holland) had already announced that it would join the Migrant Offshore Aid Station (MOAS, in operation since 2014) to provide medical assistance on board the Phoenix.   In the aftermath of April 18, MSF launched two further rescue missions of its own on board the boats Bourbon Argos (MSF Belgium) and Dignity I (MSF Spain).   In May 2015, Seawatch, an independent first-aid and rescue operation initiated by a group of German citizens, sailed to the central Mediterranean. Finally, an additional initiative of this kind called SOS Mediterranée has recently been launched.   The main patrolling and rescuing zone of the vessels constituting this civilian flotilla has been immediately outside the Libyan territorial waters, between Tripoli and Zuwara – an area that had been covered by MN. While the civilian rescue activities have remained trapped in the “half-way bridge” conundrum that had already proven its limits in the frame of the MN operation – as their intervention could not prevent migrants from resorting to smugglers in order to reach them – their impact have been impressive. By the end of October 2015, after which point these different operations either stopped for the winter or moved their activities to the Eastern Mediterranean, non-governmental vessels had rescued over 18,000 people, accounting for 7.6 percent of all rescued people (see annex).  

As we can see from these successive shifts, in the aftermath of April 2015, state-led presence at sea has risen again to a level similar to that of MN, with the addition of nongovernmental actors. While this did somewhat diminish the danger of crossing in the second half of 2015, 2,892 deaths have been recorded by the International Organisation for Migration (IOM) in the central Mediterranean in 2015. This figure is almost identical to that of 2014 (3,186), and the mortality rate for both years is also comparable (see annex).

The only factor that succeeded in significantly curbing the danger of crossing in 2015 has not been a state or non-state operation at sea, whether aimed at policing the border or at rescuing people. Rather, this has been the decision of migrants themselves to change their route as of May 2015 from the central to the eastern Mediterranean; that is, from a longer and much more dangerous route to a much shorter and relatively safer stretch of sea.   While 806 deaths have been recorded this year in the eastern Mediterranean, this number is proportionally much lower in relation to the 856,723 arrivals in Greece than it would have been in the central Mediterranean, leading to a dramatic decrease in the overall mortality rate for the Mediterranean crossing as a whole from 15 percent in 2014 to 3.7 percent in 2015 (see annex). This is, in relation to the figures that have been calculated to date, the lowest mortality rate in the last 15 years.   This trend, however, risks being reversed in the aftermath of the EU’s agreement with Turkey at the end of 2015. Turkey has, at the end of 2015, demonstrated efforts to enforce tougher controls in the Aegean – which has unsurprisingly coincided with increasing cases of deaths in Aegean Sea, despite the increasing number of NGOs, activists and humanitarian actors carrying out rescue operations at sea in that area.   The recent deployment of a NATO operation in the Aegean Sea to “stem illegal trafficking and illegal migration” risks only reinforcing this trend.   The data analysed by the Unravelling the Mediterranean Migration Crisis (MEDMIG) research project for the first three months of 2016 shows that the danger of crossing the Aegean is rising.   While migrants’ crossings in the central and eastern Mediterranean show no sign of abating, we can expect more tragedies to come over the coming year.

As a result of the redeployment of EU member states and agencies, joined by nongovernmental rescue missions, in the aftermath of the April shipwrecks, the reliance on merchant ships for SAR operations has all but stopped. Few cases of death by rescue have been recorded, despite the notable exception of the 5 August case, in which close to 300 people died when the Irish military ship LÉ Niamh approached in order to rescue them. Despite the crew of the Niamh deploying two Rigid Hull Inflatable Boats (RHIB) according to standard procedure,   the survivors we interviewed told us that their boat capsized because it was taking in water and there were big waves that destabilised it. This case only demonstrates that such tragedies can occur despite the highest level of means dedicated to SAR and that even increasing state-led search and rescue is not in and of itself a sufficient solution.

As the report demonstrates, even in the presence of the record means deployed by the Mare Nostrum operation, the danger of crossing remained high, because without avenues for legal and safe migration available, migrants continued to need to resort to smugglers and perilous means of crossing. Only a fundamental reorientation from a policy that seeks to select and block migrants’ movements to one that would grant legal and safe passage, thereby making both smugglers and the very need to rescue migrants at sea obsolete, may stop the list of more than 20,000 recorded cases of deaths at sea since the beginning of the 1990 from growing ever longer.