CSES Community Deformation Models in Southwest China

Abstract

Geodetic measurements provide important constraints on crustal deformation and earthquake processes. In this study, we describe of such data can be used to produce community deformation models which quantify interseismic strain rates and fault slip rates, which can then be used to inform probabilistic earthquake forecast. We apply the framework to data from the China Seismic Experimental Site (CSES) in the Sichuan-Yunnan area. We followed a workflow, consisting of four steps. (1) We collected and compiled the most complete and up-to-date dataset of GNSS observations, then estimated the velocity at 618 sites after correction for seasonal and co- and post-seismic effects. (2) We determined a continuous 3D velocity field model using VELMAP from the joint inversion of these GNSS data with the addition of InSAR and precise leveling data. We estimated the uncertainties from cross-validation to 1.1 (1.5) and 0.9 (0.9) mm/yr w.r.t. the east and north components of the campaign (continuous) GNSS and 0.5 mm/yr for the leveling data (at the 68% confidence level). (3) We used this velocity field model to determine a continuous strain rate field, which shows the highest strain rates localized along the Xianshuihe-Xiaojiang fault system, where the second invariant of the strain rate tensor reaches 57 nanostrain/year. (4) We developed a block model for the CSES region which fits the east and north GNSS velocities to within 0.6 and 0.7 mm/year respectively (at the 68% confidence level). Our model implies uniform slip rates of 10–14 mm/year along the entire Xianshuihe-Xiaojiang fault system. In the central segment of this fault system, slip is partitioned between the Anninghe-Zemuhe and Daliangshan faults, with slip rates of ~ 6–7 and 4–5 mm/year respectively. The block model might underestimate faulting in the interior of blocks where minor faults exist and conversely overestimate the slip-rates on block boundaries, in particular for Xiaojiang where multiple NS-trending branches exist. Since the accuracy of the deformation models depends primarily on the geodetic accuracies and spatial resolution, future work will focus on increasing the geodetic observations using in particular InSAR for which large archive of acquisitions are available. While the approach presented here could certainly be improved, it provides an effective means to incorporate geodetic and remote sensing data in seismic hazard studies.KeywordsCommunity deformation modelsGeodesyVelocity fieldStrain ratesFault slip ratesCSES