13 December 2010

AGU Fall Meeting – Haiti earthquake session

Posted by Dave Petley

The AGU Fall meeting started with a bang this morning with a 2 hour session on the Haiti earthquake. I am going to have a go at blogging live from the session using my iPad, so here goes. I am going to be careful to be general about what the speakers are saying so as not to compromise potential publications from them.

First up is Eric Calais, who is now based in Haiti for a year working for UNDP. He has described the earthquake as the largest event in terms of losses to hit a single country. Approximately 220,000 lives lost, 2 million people affected and $8 billion losses = 100% of Haiti’s annual GDP.

Eric pointed out that the earthquake occurred in an area with known seismicity – especially a large earthquake in 1770 to the west of Port-au-Prince. GPS measurements suggest that the underlying tectonics are a left lateral shear at 8 mm per year and shortening (convergence) at about 5 mm per year. Simple relationships suggested a moment magnitude = 7.1 earthquake should have been anticipated.

The earthquake showed westward movement on the Enriquillo fault, but there was no evidence of surface rupture on that fault. Post earthquake investigations suggest shortening as well as strike slip movement. The data suggest a different fault, not the Enrriquillo fault, was responsible for the earthquake. This new fault, which is orientated sub-parallel to the Enriquillo Fault, is now known as the Leogane Fault, with a combination of strike slip and dip slip movement.

Despite this a Key point is that to the scientific community the earthquake was not a surprise, but to the local people and agencies it was. This suggests a breakdown of communication. Some aspects of the earthquake itself were a surprise from science perspective – especially that movement occurred on an unknown fault. This means that we do not understand the earthquake hazard well enough, and still don’t.

Next up is David Applegate (USGS) talking about the role of science and engineering in building resilient earthquake communities. He has started by pointing out the different impacts of the Haiti and Chile earthquakes. A donor meeting for Haiti in March identified need to undertake a number of keys tasks. The first is to undertake hazard assessment of earthquakes, flooding and landslides. Second was need to address engineering issues with structures and, interestingly, areas susceptible to landslides and liquefaction. Third is a need to build capacity in country, especially around education and training. Finally there is a long term data need around geohazards – e.g. Topo data, remotely sensed data, etc.

USGS identified within 20 mins of the earthquake that 2 million peopled been affected. Early deployment focused on structural engineers, then they deployed the Earthquake Disaster Assistance Team to quickly deliver hazard maps. In particular, geological investigations such as the the fault dynamics, land level changes and the landslides, especially areas subject to future landslide potential. Second was deployment of seismic instruments. Third, improving seismic hazard hazards. He pointed out that GSHAP earthquake hazard map underestimated seismic hazard in Haiti. Finally, training, outreach and capacity building was a key aspect.

The Workshop is online – google “rebuilding for resilience Haiti”.

Next up is a talk on networks in disasters by Brian McAdoo and colleagues, mainly a focus on the social science aspects of the disaster. So how can geoscience reduce impacts of earthquakes? In Haiti there are over 10,000 NGOs working. Team looked at how communication networks operate – of course these are very complex. Pointed out that environmental degradation has been profound in Haiti – especially deforestation. Poverty in the hills has driven urban growth, mostly in very poor structures.

He focused in particular on the ways in which political changes in the country caused land degradation, and sediment production, and this was then reflected in the rate at which the coastline prograded (I.e. The coastline moved towards the sea due to sediment deposition). Very interesting and surprising, and would make a very interesting detailed study elsewhere.

Finished by emphasising the need to communicate our geohazards understanding to wider community, and to use a variety of tools to accomplish this.

Next up is Wayne Pennington, but presented by David Applegate as he has been unable to attend for personal reasons. Starts by pointing out that there are disconnects between needs of scientists and the needs of the community. Points out that the challenge for Haiti is to move from effective response to effective recovery.

Immediate focus after a disaster is humanitarian rather than scientific. After this the science community can play a role in reconnaissance, but remote sensing needs to play a greater role.

Interestingly, a question was asked about the heavy roofs in Haiti, which caused many buildings to collapse. These heavy roofs exist because of the need to protect against rainfall associated with hurricanes. Interesting point about the way that one hazard affects another.

Next up is a talk on the fault by Prentice and colleagues. The Enriquillo fault was thought to be the source as the location appeared to be coincident. However, it was very quickly evident from remote sensing data that there was no surface rupture, but there was evidence of coastal uplift. Measurement of the uplift suggested movement of about 64 cm over a distance of 50 km of coast. Prentice pointed out that the fault model suggested by Calais is one of two – the other is proposed by Hayes et al that suggests movement on three faults. This is published in Nature Geoscience. Both models suggest that most slip was on south dipping blind fault.

Reconnaissance mapping suggests offsets on the Enriquillo Fault – these are probably associated with the 18th century earthquake events. Typically up to 3.3 m offset – suggest an earthquake of 7.1 to 7.6 magnitude event. Thus, the 18th century earthquake was larger and closer to Port-au-Prince. In 2010 no strain was released in the section of fault closest to Port-au-Prince, so the risk to that city remains very high.

Next is Lepinay and colleagues from the Haiti-OBS scientific team talking about aftershocks and marine geophysical data. This is a fine effort given that he broke his arm yesterday! Pointed out that there is only one strong motion record for the event! Therefore aftershock assessment is key. But because the earthquake was small the aftershock sequence was not strong. The team deployed an ocean bottom seismic (OBS) network in February 2010. 21 instruments deployed. The instruments were recovered 3 weeks and 3 months after deployment.

Distribution of measured aftershocks is homogeneous along active fault. Three aftershocks were located by the global network, but the local instruments suggest that all these locations need to be shifted to the east. The aftershocks were nearly all compressive.

Also undertook a bathymetric survey. From this they have interpreted the fault structures. Also identified several large recent landslides in the bathymetry – may or may not be due to the earthquake.

Next in Jean Altidor, a speaker from Haiti Bureau of Mines and Energy. This is a very welcome talk. Worked with USGS and the Canadian natural resource bureau to deploy seismometers in March 2010 to measure aftershocks. Also worked with various other groups to undertake geological and geodetic measurements.

Found that downtown Port-au-Prince underlain by soft sediments, which will have amplified shaking. Also the was substantial topographic amplification. In terms of structures, there were many ground floor soft storey failures.

In a question there was a strong suggestion that the fault was actually offshore. Interesting comment “where the hell is the fault”. Calais emphasised that the fault geometry is probably clear – not offshore.

Final presentation is very controversial. Wdowninski and colleagues from Miami suggesting that the timing of the earthquake has been affected by hurricanes and deforestation. Pointed out that in 2008 therevwas a massive hurricane season that caused lots of floods and landslides. Are the two events linked.

In 2008 Haiti was affected by 4 tropical cyclones. The authors suggest that the hurricanes trigger surface unloading due to rapid erosion, that then triggers the earthquake. Unloading may increase vertical shear and decrease normal stress. In the case of the earthquake, movement on the fault was atypical (dip slip), which suggests that the hurricanes induced rapid erosion in the mountains – this is shown in the satellite imagery. Calculated sediment deposition in the delta of the river draining the mountains. Gives an erosion rate of 6 mm per year – higher than expected, due to deforestation.

Calculated stress change at hypo centre from unloading using a DEM to give distribution of erosion. Then modelled resultant stress change on the fault. This shows that unloading promotes fault movement on a south dipping fault. Showed that magnitude of change is at least hypothetically possible to trigger the earthquake. In particular the very strong 2008 hurricane season induced sufficient change in coulomb stress to bring fault closer to rupture.

So why was there a delay between the hurricane and the earthquake? Maybe increased soil moisture kept normal stress high. Fault movement needed soils to dry out.

Finished by pointing out that a combination of increased sea surface temperature, inducing more hurricanes, and deforestation may be responsible for increasing likelihood of rupture events.

All in all this was a fascinating session. What a great start to the meeting!