Pamir Meets Tien Shan: The Complex Faulting During the 2008 Nura Earthquake (Kyrgyzstan) Revealed by SAR Data Analysis
Sudhaus, Henriette; Teshebaeva, Kanayim; Echtler, Helmut; Schurr, Bernd
Helmholtz Centre Potsdam GFZ, GERMANY

In the Pamir and Tien Shan mountains the large total crustal shortening is accommodated at many, generally east-west-oriented thrust faults. The analysis of recent GPS velocity data across this region, however, show a strong concentration of strain at the Main Pamir Thrust (MPT). North of the MPT are narrow valleys that extend from the Tajik basin in the West and the Tarim basin in the East and divide the Northern Pamir from Tien Shan. At the eastern side of the MPT, where the triple junction of Kyrgyzstan, China and Tajikistan is located, these valleys have been closed already and a major drainage divide has formed such that the two orogenes Pamir and Tien Shan are touching. On October 5th in 2008 this place was struck by a M6.6 earthquake, which caused a number of fatalities and heavily damaged the settlement Nura. Due to the remote mountainous location of Nura at political borders the tectonics in the area are not studied in detail. One can, however, infer from digital elevation models a drastic change in the general faulting style. The east-west-oriented thrust faults dominating along the western side of the Pamir front are replaced by southwest-northeast-oriented oblique thrust faults in the footwall of the Pamir in the Nura region. These oblique thrust faults prolong to the East towards the Talas-Fergana fault, a major strike slip fault cutting through the Tien Shan mountains. Within the last 50 years already to more M>6 earthquake have occurred in the Nura region. So far, the two earlier ones in the year 1974 were attributed to ruptures at the Main Pamir Thrust fault. The recent Nura earthquake, however, could be precisely located due to a close by installation of a temporal seismic network and it turned out that it ruptured the Pamir Frontal Thrust fault cropping out several kilometers north of the Main Pamir Thrust. The seismic data are further suggesting a very shallow source at only 4 kilometers depth with a clear east-west-striking thrust mechanism.
To study the Nura earthquake in more detail we use SAR data from the ALOS (L-band) and ENVISAT (C-band) satellites for measurements of the co-seismic surface displacement. These measurements proved to be difficult. High mountains and extended snow cover in the source region cause strong phase decorrelation in the interferograms and we have a high loss of information, particularly in the C-band data, due to layover. With radar interferometry we are therefore only able to measure the deformation in the footwall of the Frontal Thrust fault. Pixel offset measurements on the other hand, while showing an acceptable coherence in the ALOS data, are strongly distorted due to ionospheric influences on the radar waves that cause a highly anisotropic imprint on the signal.
Despite these obstacles we could detect clear surface rupture north of the EW-oriented Frontal Thrust fault along the southwest-northeast-oriented Irkeshtam fault in the Pamir footwall, suggesting a co-seismic activation. Such strain partitioning during the earthquake is not anticipated when analyzing seismic data alone and may likely have occurred also during past events. To not merely observe this strain partitioning, but to also be able to quantify it, we need a careful implementation of the incomplete and highly error-affected data. Our way of data error estimation and data implementation for an earthquake source study will be a focus in our presentation.