Geophysical Research Letters Supporting Information for Gravity increase before the 2015 Mw7.8 Nepal earthquake Shi Chen1, Mian Liu2, Lelin Xing3, Weimin Xu1, Wuxing Wang4, Yiqing Zhu5, and Hui Li3 1. Institute of Geophysics, China Earthquake Administration, Beijing, 100081, China 2. University of Missouri, Columbia, MO, 65211, USA 3. Hubei Earthquake Administration, Wuhan, 430071, China 4. Institute of Earthquake Science, China Earthquake Administration, Beijing, 100036, China 5. Second Crust Monitoring and Application Center, China Earthquake Administration, Xi’An, 710054, China. Contents of this file Text S1 Figures S1 to S2 Introduction This supporting information provides the details of the forward modeling using gravity change and figures of the vertical displacement of the gravimetric stations from the time series of continuous GPS measurements and GRACE data of gravity changes at the four stations in southern Tibet. 1 Text S1. Forward modeling of gravity anomaly of a disk body with a uniform density change. (1) Equation of gravity anomaly βπ(π, π) We used the analytical solution to forward gravity change with a disk geometry [Singh, 1977]. βπ(π₯, π¦) = 2πΊπ | π 2 −π₯ 2 √(π₯+π )2 +π¦ π π«(π) + √(π₯ + π )2 + π¦ 2 π¦(π) + 0.5ππ¦π¬(π, π) − ππ¦| 2 π+π» where the βπ is the gravity change with a uniform density π, πΊ is gravitation constant. The x and y are the horizontal and vertical position of observed station, respectively. The R, Z and H are the radius, top depth and thickness of disc model, respectively. As shown in Figure 2, π = πβ + ππ‘ and πβ is the known quantity and equal to the elevation, list in Table 1. The K(k) is the first kind complete elliptic integral: π/2 Κ(k) = ∫ 0 ππ‘ √1 − π 2 π ππ2 π‘ and E(k) is the second kind complete elliptic integral: π/2 Ε(k) = ∫ √1 − π 2 π ππ2 π‘ππ‘ 0 and the Heuman Lambda function is: π¬(π, π) = π¦(π)πΉ(π, π) + πΎ(π)π¦(π) − πΎ(π)πΉ(π, π) where k and φ as follows: 4π π§ (π₯ + π )2 + π§ 2 π x−R φ = + tan−1 2 z k2 = (2) Bouguer layer equation for correcting the effects of elevation change: βπ = −(3.086 − 0.419π)ββ ππΊππ/ππ Assuming crustal density to be π = 2.7 − 2.9π/ππ3 for the Tibetan crust [Chen, et al., 2004], the gravity vertical gradient is from -1.87~1.95 ππΊππ/ππ. We used the −1.9ππΊππ/ππ in this paper. (3) Strain rate estimations According to the definition of dilatation Θ: Θ = Δπ/π where π is a volume element, Δπ is the incremental change of the volume element. After the deformation, volume element is π1 = π + Δπ = (1 + Θ)π If the total mass is conserved, the density can be expressed as 1/π1 = (1 + Θ)/π0 where π0 is the density of element before deformation, and π1 is the density after deformation. Therefore, the density change Δπ is: Δ π = π1 − π0 = − Θ π 1+Θ 0 The equation can be transformed as: 2 Θ=− Δπ Δπ + π0 Using the plane stress equation [Turcotte et al., 2002], there is: Θ= 2(1 − 2υ) π1 1−υ where υ is the Poisson ration, and π1 is the horizontal strain. Figure S1. Time series of daily solution of vertical displacements at the Naqu, Lhasa, Shigatse, and Zhongba stations from continuous GPS recording. Red lines are the linear fits. The GPS stations are collocated with the gravimetric stations. We fit the average vertical motion rate at these stations and use the results to correct for the effects of elevation changes on gravity at these stations. 3 Figure S2. Gravity change observed by the Gravity Recovery and Climate Experiment (GRACE) satellite. We computed the series of average monthly gravity change at the four gravimetric stations using a 300 km Gauss filter. The GRACE data show clear seasonal variations and provide the long-wavelength gravity change. The average rate is -0.048μGal/yr for the Lhasa station, -0.046μGal/yr for the Shigatse station, -0.042μGal/yr for the Zhongba station, and -0.018μGal/yr for the Naqu stations. These trends of decreasing gravity are consisted with the continuing uplift of the Tibetan Plateau. The long-wavelength gravity decrease from the GRACE data are is opposite to the gravity increase observed by absolute gravity measurements at these stations. Hence gravity increases at these stations are not artifacts of background gravity variations; we suggest that they reflect mass change in a broad source region related to the tectonic processes leading to the 2015 Nepal earthquake. References Chen,W. P., and Z. Yang (2004), Earthquakes beneath the Himalayas and Tibet: Evidence for strong lithospheric mantle, Science, 304, 1949-1952. Singh, S.K. (1977), Gravitational attraction of a vertical right circular cylinder, Geophysical Journal of the Royal Astronomical Society, 50, 243–246. Turcotte, D. L., and G. Suchubert, G (2002), Geodynamics, 2nd ed. Cambridge, U. K.:Cambridge Univ. Press. 4