Sri Lanka: Early Warning of War Crimes?

Speaking to Le Monde journalist Philippe Bolopion on condition of anonymity, a UN official stated the following:

“On savait qu’on se préparait à un carnage. On a tiré la sonnette d’alarme pendant des mois, mais ils n’ont jamais frappé en public sur le gouvernement. Tout le monde a peur que son agence soit jetée dehors.”

“We knew carnage was brewing. We rang the alarm bells for some months but no one ever took the Sri Lankan government to task publicly. Everyone is scared of having their agency removed from the country.”

Bolopion, Le Monde’s special correspondent in Colombo also obtained evidence of text messages (SMS) sent by local UN and NGO staff from the scene of the war crime.

March 9: “Please, ask the Sri Lankan army to stop.”

March 14: “Where is the no-fire zone”?

And as the LTTE starts recruiting by force:

March 12: “Both sides are torturing us.”

March 12: “We’re dying! Two shells landed 10m away.”

March 19: “Youths are being taken to fight, very sad. How is the international community reacting?”

March 21: “Hundreds of people trying to escape were stopped by local dictators. They were beaten with sticks, without distinction by age or gender. I hear they crying very loudly.”

March 21: “Why is the international community staying silent?”

So much for the responsibility to protect.

Applying Earthquake Physics to Conflict Analysis

I’ve long found the analogies between earthquakes and conflicts intriguing. We often talk of geopolitical fault lines, mounting tensions and social stress. “If this sounds at all like the processes at work in the Earth’s crust, where stresses build up slowly to be released in sudden earthquakes … it may be no coincidence” (Buchanan 2001).

To be sure, violent conflict is “often like an earthquake: it’s caused by the slow accumulation of deep and largely unseen pressures beneath the surface of our day-to-day affairs. At some point these pressures release their accumulated energy with catastrophic effect, creating shock waves that pulverize our habitual and often rigid ways of doing things…” (Homer-Dixon 2006).

But are fore shocks and aftershocks in social systems really as discernible as well? Like earthquakes, both inter-state and internal wars actually occur with the same statistical pattern (see my previous blog post on this). Since earthquakes and conflicts are complex systems, they also exhibit emergent features associated with critical states. In sum, “the science of earthquakes […] can help us understand sharp and sudden changes in types of complex systems that aren’t geological–including societies…” (Homer-Dixon 2006).

The Model

To this end, I collaborated with Professor Didier Sornette and Dr. Ryan Woodard from the Swiss Federal Institute of Technology (ETHZ) to assess whether a mathematical technique developed for earthquake prediction might shed light on conflict dynamics. I presented this study along with our findings at the American Political Science Association (APSA) convention last year (PDF).

This geophysics technique, “superposed epoch analysis,” is used to identify statistical signatures before and after earthquakes. In other words, this technique allows us to discern whether any patterns are discernible in the data during foreshocks and aftershocks.

Earthquake physicists work from global spatial time series data of seismic events to develop models for earthquake prediction. We used a global time series dataset of conflict events generated from newswires over a 15-year period. The graph below explains the “superposed epoch analysis” technique as applied to conflict data.

The curve above represents a time series of conflict events (frequency) over a particular period of time. We select arbitrary threshold, such as “threshold A” denoted by the dotted line. Every peak that crosses this threshold is then “copied” and “pasted” into a new graph. That is, the peak, together with the data points 25 days prior to and following the peak is selected.

The peaks in the new graph are then superimposed and aligned such that the peaks overlap precisely. With “threshold A”, two events cross the threshold, five for “threshold B”. We then vary the thresholds to look for consistent behavior and examine the statistical behavior of the 25 days before and after the “extreme” conflict event.

Results

For this study, we performed the computational technique described above on the conflict data for the US, UK, Afghanistan, Columbia and Iraq.

The foreshock and aftershock behaviors in Iraq and Afghanistan appear to be similar. Is this because the conflicts in both countries were the result of external intervention, i.e., invasion by US forces (exogenous shock)?

In the case of Colombia, an internal low intensity and protracted conflict, the statistical behavior of foreshocks and aftershocks are visibly different from those of Iraq and Afghanistan. Do the different statistical behaviors point to specific signature associated with exogenous and endogenous causes of extreme events? Does one set of behavior contrast with another one in the same way that old wars and new wars differ?

Future Research

Are certain extreme events endogenous or exogenous in nature? Can endogenous or exogenous signatures be identified? In other words, are extreme events just part of the fat tail of a power law due to self-organized criticality (endogeneity)? Or is catastrophism in action, extreme events require extreme causes outside the system (exogeneity)?

Another possibility still is that extreme events are the product of both endogenous and exogenous effects. How would this dynamic unfold? To answer these questions, we need to go beyond political science.

The distinction between responses to endogenous and exogenous processes is a fundamental property of physics and is quantified as the fluctuation-dissipation theorem in statistical mechanics. This theory has been successfully applied to social systems (such as books sales) as a way to help understand different classes of causes and effects.

Our goal is to use the same techniques to investigate the questions: Do conflict among actors in social systems display measurable endogenous and exogenous behavior?  If so, can a quantitative signature of precursory (endogenous) behavior be used to help recognize and then reduce growing conflict? The next phase of this research will be to apply the above techniques to the conflict dataset already used to examine the statistical behavior of foreshocks and aftershocks.

World Disaster Report 2009: Early Warning/Action

I just finished reading the International Federation of the Red Cross’s (IFRC) new World Disaster Report 2009 which focused specifically on early warning and early action. I highly recommend it to those of us working on conflict early warning and rapid response.

Since the report is over 200 pages long, below are what I consider to be the most important excerpts (along with some of my own commentary).

People-Centered Approach

The report notes that “the development of a more people-centered approach is clearly essential to ensure that the warnings captured by satellites, computer modeling and other technologies reach at-risk communities and are then acted upon.”

The people-centered approach to early warning focuses on how individuals and communities can understand the threats to their own survival and well-being, share that awareness with others and take actions to avoid or reduce disaster. The risk of disaster is partly caused by external hazards (such as an earthquake, cyclone, surprise disease epidemic, war or economic crisis) that are difficult or impossible to stop. However, communities know that disasters are also about people being vulnerable, not being in the right place at the right time with adequate forms of protection.

I find the reference to war as an external hazard fascinating (and vindicating). I’ve been emphasizing the need for a people-centered approach to conflict early warning since 2003 based on my research in disaster management. But when I did so, I would often get push-back from “experts” who would be highly critical of my drawing on lessons learned from operational responses to “natural” disasters.

And yet, as I would typically reply, the international community almost always fails to correctly predict and/or prevent armed conflict. Hence my insisting that we should treat conflict as a hazard and focus more on preparedness—at least as a plan B if we fail to predict/prevent.

People-centred early warning suggests that rather than being vulnerable, people can be capable, resilient and able to protect themselves. Three basic requirements are that individuals and institutions have the knowledge about what is a threat, that people are able to communicate a change in threat, and that they are in a position to respond. People-centered approaches to early warning, therefore, require the right support from scientific and political institutions to provide the context within which they can become strong.

A key point on which many scholars and practitioners of disaster reduction agree is that ’strategies must extend beyond information provision to engage community members in ways that facilitate their adoption of protective actions.’

There’s no reason why we can’t take a similar approach to conflict early warning and response. In fact, many local communities in conflict have, and continue to do so. The question is, why aren’t we building on this existing capacity to render communities more resilient to conflict?

The elements of early warning systems most likely to fail are the the dissemination and communications of warning, and response capability and preparedness to act. A people-centered approach is especially essential for these two elements, one that focuses not only on the science and technology behind the warnings, but also on the social and psychological aspects of early warning and early action and on activities to build a culture of prevention, rather than a culture of short-term response.

Local-level involvement must start with the first element –building risk knowledge. Hazard assessment and risk mapping can help bridge the gap between scientific and local knowledge.

Communities suffering from food insecurity or disease outbreaks normally know about the impending disaster before the authorities. Early warning systems for slow onset disasters must be locally based and controlled, or at least there must be close coordination between national and local systems, to ensure early detection and early response.

The people-centered approach reinforces the important truth that it is people, not institutions, who should have rights. Institutions should be established in the interests of people. Early warning systems are, therefore, systems or institutions that must serve people’s needs.

Challenges of Technology

The three global early warning conferences (1998, 2003 and 2006) have fostered a new consensus: “early warning is a system, not a technology.”

Early warning is not only the production of technically accurate warnings but also a system that requires an understanding of risk and a link between producers and consumers of warning information, with the ultimate goal of triggering action to prevent or mitigate a disaster.

To be sure, “excessive focus on technological solutions without balancing the other components is not only expensive but it can create a false sense of security.”

Given the rise in information and communication technologies (ICT) and the difficulty in ensuring that only one, authoritative voice issues warnings, the risk of false alarms is increasing. Ideally, warnings would be transmitted by a single authoritative voice, but realistically, this does not always happen and cannot be controlled.

In my opinion, this will be increasingly difficult to control.

Many technical monitoring systems, whether global or national, continue to have a top-down, scientific bias. Warnings based on remotely sensed data or national modeling can miss important dynamics existing at the local level. A major challenge for all technical warning systems is how to build community-level early warning indicators and indigenous knowledge into the system. Early warning systems for slow-onset disasters such as droughts will not be accurate if they ignore community-level indicators. Even for rapid-onset disasters, local indicators can be important elements of the system if properly understood and integrated.

Even with well-coordinated structures, dissemination to remote areas is still difficult in many places and requires a combination of technological and non-technological solutions. There is no ‘one size fits all’ solution.

Studies have been carried out in Sri Lanka to evaluate the effectiveness of various ICT approaches, including satellite radio and SMS sent via cell phones. Even proponents of SMS-based dissemination view it as complementary to other warning channels due to limited cell phone penetration among the most vulnerable, language limitations with SMS, potential damage to cell phone networks during disasters, network congestion and suchlike. The traditional broadcast media remains the most widely used channel globally to disseminate warnings, but the effectiveness of this channel can be compromised if the most vulner able populations lack access to TVs or radios.

Planning and Preparedness

People are more likely to pay attention to warnings if they have been educated about the risks in advance and know what actions to take.

Contingency plans can map out roles and responsibilities in advance and speed the response time, although in many cases, they become routine annual documents rather than living, operational processes. Simulation exercises can be very effective in building response capabilities and bringing preparedness planning to life. They can test the response systems, coordination structures and the knowledge of at-risk populations.

A separation between the producers or operators of the early warning system and those making funding decisions for preparedness and response activities is common. When funding decisions fall in a different ministry or organization, those controlling the resources may require additional verification or analysis before taking action. Some of the most effective early response systems occur when information and analysis units are directly tied into funding units.

Response-oriented contingency planning has been advocated to increase the effectiveness of such time-critical livelihood interventions. Contingency planning should focus on the most probable scenarios and a limited number of feasible, tested response options. In these situations the goal is to reduce the associated timelines for implementation.

Early Action

In slow-onset disasters such as drought, it is difficult to define when an ‘emergency’ begins. Early warning systems monitor a variety of indicators, but without clear indicators to trigger response, it is often delayed until the effects are visible and populations have suffered economic or physical losses.

In order to achieve the last mile, early warning systems need to engage all people at community level, to be locally owned and shown to be cost-effective. Incidentally, systems that are inclusive lead to improved well-being and development of communities at many levels. Lay people with an interest in early warnings need not wait for experts to provide information.

Setting minimum standards in early warning might be a reasonable target for the people-centered approach and those responsible for assisting the well-being of communities. Responsibilities, on the other hand, lie in part with individuals, so that taking effective early action relates also to personal decision-making, the quality of community cohesion, values and ethics.

In the face of rising risks and rising uncertainties, effective early action is more important than ever, [which] works best when it spans a range of timescales, not just providing a more rapid response to a disaster, but also anticipating it days, hours, months, years and even decades in advance, and over time reducing the risk of a range of hazards.

That bridging of timescales is the key to early warning and early action: ‘routinely taking action before a disaster or health emergency happens, making full use of scientific information on all timescales.’ Practical early action, based on early warnings at all timescales, does pay off, reducing risks and saving lives.

I find this reference to timescales particularly fascinating as I’ve made similar arguments with more of an emphasis on bridging timescales with spatial scales for conflict early response. See my post on Scale and Complex Systems here.

Accurate risk maps, showing people and assets at risk, can be a key tool to inform plans and activities. In practice, many [...] early warning products are almost impossible to understand by non-experts. They are often overly technical, sometimes including large uncertainties. As a result, the raw products do not naturally lead humanitarian actors to a decision.

The one, rather major, criticism that I do have of the report is the repeated a-political references to Cyclone Nargis in Burma. At no point does the report acknowledge the political causes of the “natural” disaster and the political constraints placed on the disaster response community by Burma’s junta.

Global Impact Vulnerability Alert System (GIVAS): A New Early Warning Initiative?

Cross-posted on iRevolution.

UN Secretary-General Ban Ki-Moon is calling for better real-time data on the impact of the financial crisis on the poor. To this end, he is committing the UN to the development of a Global Impact and Vulnerability Alert System (or GIVAS) in the coming months.  While I commend the initiative’s focus on innovative data collection, I’m concerned that this is yet another “early warning system” that will fail to bridge alert and operational response.

The platform is being developed in collaboration with the World Bank and will use real time data to assess the vulnerability of particular countries or populations. “This will provide the evidence needed to determine specific and appropriate responses,” according to UNDP. UN-Habitat opines that the GIVA will be a “vital tool to know what is happening and to hold ourselves accountable to those who most need our help.”

According to sources, the objective for the GIVAis to “ensure that in times of global crisis, the fate of the poorest and most vulnerable populations is not marginalized in the international community’s response. By closely monitoring emerging and dramatically worsening vulnerabilities on the ground, the Alert would fill the information gap that currently exists between the point when a global crisis hits vulnerable populations and when information reaches decision makers through official statistical channels.”

GIVA will draw on both high frequency and low frequency indicators:

“The lower frequency contextual indicators would allow the Alert system to add layers of analysis to the real time “evidence” generated by the high frequency indicators. Contextual indicators would provide information, for example, on a country’s capacity to respond to a crisis (resilience) or its exposure to a crisis (transmission channels). Contextual indicators could be relatively easily drawn from existing data bases. Given their lesser crisis sensitivity, they are generally collected less frequently without losing significantly in relevance.”

“The high frequency indicators would allow the system to pick up significant and immediately felt changes in vulnerability at sentinel sites in specific countries. This data would constitute the heart of the Alert system, and would provide the real-time evidence – both qualitative and quantitative – of the effects of external shocks on the most vulnerable populations. Data would be collected by participating partners and would be uploaded into the Alert’s technical platform.”

“The pulse indicators would have to be highly crisis sensitive (i.e. provide early signals that there is a significant impact), should be available in high periodicity and should be able to be collected with relative ease and at a reasonable cost. Data would be collected using a variety of methodologies, including mobile communication tools (i.e. text messaging), quick impact assessment surveys, satellite imagery and sophisticated media tracking systems.”

The GIVA is also expected to use natural language processing (NLP) to extract data from the web. In addition, GIVA will also emphasize the importance of data presentation and possibly draw on Gapminder’s Trendalyzer software.

There’s a lot more to say on GIVA and I will definitely blog more about this new initiative as more information becomes public. My main question at this point is simple: How will GIVA seek to bridge the alert-response gap? Oh, and a related question: has the GIVA team reviewed past successes and failures of early warning/response systems?