Identifying Growth of Structures in the Zagros Fold and Thrust Belt: Initial Time Series Results and Evaluation of Precipitable Water Vapor Effects

In this work, we combine InSAR observations with independent satellite-based measurements of precipitable water vapor in an attempt to identify aseismic deformation in a fold and thrust belt. We use the precipitable water vapor data to assess the magnitude of deformation that may be induced by atmospheric noise that is correlated in time given a particular data set. The Zagros Fold and Thrust Belt (ZFTB) of western and southwestern Iran accommodates active collision between the Arabian and Eurasian plates. Current convergence rates between Arabia and Eurasia are 25-36 mm/yr, with roughly half of the shortening accommodated in the ZFTB. While rates of seismicity in the ZFTB are high, large (>Mw 7.0) surface rupturing earthquakes are rare, and calculations of seismic moment rates suggest that only 10-30% of the total shortening is accommodated seismically. Thus, a significant amount of strain must be accommodated aseismically, presumably though the deformation of the large whale-back anticlines characteristic of the ZFTB. In this work, we present the initial results of an InSAR time series analysis in the Bandar-e-Abbas region of the southern ZFTB. We aim to identify interseismic strain accumulation and growth of the folds in order to better our understanding of the relative roles of seismic and aseismic deformation on the formation of long lived geologic structures. We implement the Small Baseline Subset (SBAS) time series approach to calculate line of sight (LOS) surface velocities from ENVISAT and ERS descending Tracks 435 and 163 between February 1993 and September 2010. A key assumption of SBAS is that correlated atmospheric noise is random in sign over time and should have zero effect on the time series if the number of scenes is large enough. To investigate if the seasonal distribution of available dates induces a biased atmospheric noise signal that may contribute to our computed LOS velocities, we use MODIS and MERIS precipitable water vapor data to analyze the potential effects of our data distribution. We create synthetic “interferograms” by differencing scenes acquired in the same month as the SAR imagery and compute a time series using a generalized SBAS algorithm. We explore a stochastic approach where we generate many combinations of acquisition dates that are within the same month of the SAR acquisitions. When considering velocities of a single pair, we observe a very strong correlation of precipitable water vapor with topography. To the contrary, when we compare the full time series to topography, we see very little, if any, correlation to topography and considerably lower velocities. This implies that our distribution of SAR acquisitions does not impart a seasonal bias. The lack of correlation with topography in time also demonstrates that while this technique cannot “correct” noise errors, it is capable of providing a metric for discerning between real signals of interest and correlated atmospheric noise.