Co-seismic surface structures
In the first discussion with the local authorities and our guide, we also talked about the effects and damages of the El Galpón earthquake. The damages in the village were obvious and I wrote about them in the last post. But for us seismologists it is also interesting to see if there are damages in the landscape. Typical effects would be: cracks in the ground, landslides, water and sand fountains, a new step in the topography or changes in the ground water table. We call this stuff co-seismic effects, because they occur during the shaking of an earthquake. Some of these (i.e. changes in the ground water table) can occur before an event as pre-seismic or after an event as post-seismic events (e.g. landslides).
Luckily, the people understood what we were pointing at and they showed us pictures of big cracks in the ground. They found them directly after the earthquake, so they could be our co-seismic effects. On our last day of fieldwork we had some time left to have a look on these cracks in the ground.
The pictures describe what we saw the best. At several spots we could observe a clear opening in the ground of a few centimeters horizontally and sporadically a vertical offset of up to 8cm. This crack was only one clear line in a nearly N-S orientation (striking 5 degrees). And only the eastern part of the ground was lifted up. Right or left of it there was no comparable feature. At one point we found a step-over and the crack jumped half a meter to the west. We only had access to the crack at 4 spots and we could follow it only for like 10m, because the vegetation was too dense. But our local guide told us they found this feature over a length of about 10 kilometers at several locations and some parts of it are already refilled by the strong monsoon this summer. The northernmost spot was a highway where the crack went perpendicular through the asphalt. Along this highway it was the only place of damage. Parallel to the street we saw additional cracks, sometimes two in parallel. The highway is build on an artificial dam, so these cracks are consolidation cracks. That means due to the shaking of the earthquake the material of the dam got weakened and parts of it moved downward building a crack at the break-off.
2. Additional information
In the area of the earthquake epicenter there is nearly no topography. The estimated epicenter is in the middle of a deep sediment filled basin and extremely flat. So the highway dam is the only steep feature in the area where people could have been affected by landslides. Except for the Colorado range southeast of El Galpón; we found some small rockfalls in the no-mans-land. But much more interesting are the hot water springs in the middle of the basin, quite close to the cracks we saw. Around 100 km to the north and to the south there are more hot springs. Here, the water comes out of the ground with 43° Celsius in an artesian manner. That is very interesting, because it is in a complete flat deep basin (up to 5km of sediment [Iaffa et al. 2011]) and no volcanism nor other sources are obvious.
For the source mechanism of the hot spring water a fault could be a reasonable explanation. In this area a reverse fault was found in seismic lines by Iaffa et al. 2011. Faults as contact zones between two adjacent blocks giving the space for fluids to move along them. Due to the geothermal gradient water from a deeper layer where the hot water originated has to be warmer than water in upper layers. The correlation of the hot water springs and the fault which produced the earthquake in October is quite close, because the ground crack ran through some of the small buildings which were used as hot water bathrooms. But that would mean the crack we found is the surface rupture of the earthquake rupture plane. A different interpretation of the cracks would be liquefaction. Liquefaction is a co-seismic effect which occur in water saturated basins. This effect everybody observed already as a child at the beach when building a drip castle (in German we say Kleckerburg). The beach sand with a lot of water in between the grains get the ability to flow. When you have very wet sand, which cannot flow by itself and shake this sand it also starts to flow. On the beach you have experienced that while standing in the sand and dig your feet deeper in the wet sand by moving your feet and toes. Instead of shaking the few square centimeters under your feet an earthquake shakes some square kilometers. Extreme events (in Japan or New Zealand) dug complete houses and cars into the ground. A very nice video of a dry or stiff upper layer moving on top of the wet sand can be seen here. That could have been a similar effect in here during the earthquake. A colleague of mine who is originally from close to the area told me that in this basin you have big mud lenses in the ground close to the surface. During the dry winter they dry out and get smaller. Due to the shrinking the surface respond to that compressional stress by forming small cracks.
The latter explanation with mud lenses is not very suitable, because we went into the area directly after the monsoon, which was very wet this year and all the cracks should have been gone in the meanwhile. To discuss liquefaction based on the moisture conditions I don’t have any data. But I would assume it is less likely because the earthquake occurred before the rain season and probably the ground was too dry to cause liquefaction. In addition we didn’t find any side effects of liquefaction like water fountains (as you could see in the video) or objects drown into the sand. But the most important point is the lack of additional cracks. We only found one clear cracking with quite consistent strike orientation. With liquefaction you would have much more cracks and which would be oriented much more irregularly. Only from the field observations a surface rupture of the earthquake seems to be the most reasonable explanation. But the estimated depth of the earthquake hypocenter is too deep. Different institutes estimated the depth between 15 and 23 kilometers. The length of the surface rupture of a Mw 5.7 you would expect to be between 10 and 15 kilometers. That is quite consistent, as our maximum distance of observed crack structures is about 13 kilometers. With such a ‘smaller’ magnitude it is not reasonable that the earthquake broke through the whole distance up to the surface. Either we cannot really explain this structure or the depth estimate is not very well. You have to know that the depth estimate is the most erroneous in the whole process of earthquake localization . To keep the error small you need a seismograph very close or on the epicenter. Or when you locate it with teleseismic data (i.e. you can analyze this earthquake with data from seismographs in Europe) you need a very exact model of the subsurface underground. Due to the lack of stations in this area, the second way is the proper one for localization. My supervisor suggested then maybe the velocity model of the underground is too different from the reality. Normally, you will use a global average model for the main parameters you need. But in a deep basin like this the velocities are much slower. That means the speed at which a seismic wave in that area will go through the ground is much lower than the average. Thus, the depth of an earthquake in a slow velocity environment will appear much deeper because the real wave needs much more time to travel through space. That is a hypothesis I will check later on when I have my own data from the area. To calculate a local velocity model of that basin was our plan anyway. So I will keep you in touch.
In the end I hope I could explain a little bit to you what we were struggling with during the field school. Sometimes the answers and reasons for something we observe are not really clear and easy.