Auxiliary Material for Improving InSAR geodesy using global atmospheric models. Romain Jolivet, (Seismological Laboratory, Department of Geological and Planetary Sciences, California Institute of Technology, California, USA.) Piyush Shanker Agram, (Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109, United States of America.) Nina Y. Lin, Mark Simons, (Seismological Laboratory, Department of Geological and Planetary Sciences, California Institute of Technology, California, USA.) Marie-Pierre Doin, (Institut des Sciences de la Terre, UniversitŽ Joseph Fourier, CNRS, UMR 5275, BP 53, 38041 Grenoble, France) Gilles Peltzer, (Department of Earth and Space Science, University of California Los Angeles, USA.) Zhenhong L. (COMET+, School of Geographical and Earth Sciences, University of Glasgow, UK.) Journal of Geophysical Research, Solid Earth, 2013 Introduction The supplementary materials consist in three sections. We first present figures for comparison between atmospheric delay corrections derived from Global Atmospheric Models (GAMs) and from an empirical approach. We then present examples of comparison between GAMs-derived corrections and MERIS-derived corrections, corresponding to cases in figure 4 in the main manuscript. We finally list the interferograms used to produce figure 8 of the main manuscript. Comparing predictions from Global Atmospheric Models and from an empirical approach. In this section, we show a comparison between the prediction derived from atmospheric models and an empirical approach for the case of figures 3, 6, 7, 9, 10 and 11. This empirical method, described in Lin et al, 2010, allows to estimate the correlation coefficient between the digital elevation model and the interferogram by decomposing the phase and the DEM into different spatial frequency bands. Discarding the highest (i.e. short wavelength noise) and the lowest (i.e. long wavelength related to orbital inaccuracy) frequencies leads to a robust estimate of the linear relationship describing the stratification of the troposphere. Figures S01, S02, S03, S04, S05, S06 and S07 show the original interferogram, to the left, the prediction from a global atmospheric model, center column, and the empirical estimation of the stratified delay, right column. We show the residuals between the data and the GAM prediction and between the data and the empirical estimation in the center of each corresponding column. Finally, we show these residuals after removing a best fitting first-order polynomial function in both range and azimuth, that describes residual orbital errors, in the bottom of each column. FigureS01.pdf: Comparison with an empirical correction method (see Fig. 3 in Main Manuscript) - Left: Interferogram between Envisat acquisitions on 08/27/2004 and 03/05/2004 over the Pakistani Makran. On color cycle is 10 cm along the Line-Of-Sight. Each contour line is 5 cm. The background shading is from SRTM DEM. Center Column: Prediction from ERA-Interim (top), the corresponding residuals (center) and flattened residuals (bottom). Right Column: Empirical estimation (top), the corresponding residuals (center) and flattened residuals(bottom). FigureS02.pdf: Comparison with an empirical correction method (see Fig. 6 in Main Manuscript) - Left: Interferogram between ERS SAR acquisitions on 10/26/1993 and 11/30/1993 over central California. On color cycle is 15 mm along the Line-Of-Sight. Each contour line is 5 cm. The background shading is from SRTM DEM. Center Column: Prediction from ERA-I (top), the corresponding residuals (center) and flattened residuals (bottom). Right Column: Empirical estimation (top), the corresponding residuals (center) and flattened residuals(bottom). FigureS03.pdf: Comparison with an empirical correction method (see Fig. 7 in Main Manuscript) - Left: Interferogram between ALOS PALSAR acquisitions on 04/12/2009 and 01/19/2010 over Hawaii. On color cycle is 75 mm along the Line-Of-Sight. Each contour line is 50 mm. The background shading is from SRTM DEM. Center Column: Prediction from NARR (top), the corresponding residuals (center) and flattened residuals (bottom). Right Column: Empirical estimation (top), the corresponding residuals (center) and flattened residuals(bottom). FigureS04.pdf: Comparison with an empirical correction method (see Fig. 8 in Main Manuscript) - Left: Interferogram between Envisat ASAR acquisitions on 01/13/2007 and 08/11/2007 over the northern coast of Chile. On color cycle is 100 mm along the Line-Of-Sight. Each contour line is 50 mm. The background shading is from SRTM DEM. Top Row: Prediction from ERA-Interim (left), the corresponding residuals (center) and flattened residuals (right). Bottom row: Empirical estimation (left), the corresponding residuals (center) and flattened residuals(right). FigureS05.pdf: Comparison with an empirical correction method (see Fig. 9 in Main Manuscript) - Left: Interferogram between Envisat acquisitions on 11/25/2005 and 09/16/2005 over the Pakistani Makran. On color cycle is 10 cm along the Line-Of-Sight. Each contour line is 5 cm. The background shading is from SRTM DEM. Center Column: Prediction from ERA-Interim (top), the corresponding residuals (center) and flattened residuals (bottom). Right Column: Empirical estimation (top), the corresponding residuals (center) and flattened residuals(bottom). FigureS06.pdf: Comparison with an empirical correction method (see Fig. 10 in Main Manuscript) - Left: Interferogram between Envisat acquisitions on 02/09/2004 and 09/26/2005 over Northern Chile on track 096. On color cycle is 20 cm along the Line-Of-Sight. Each contour line is 5 cm. The background shading is from SRTM DEM. Center Column: Prediction from ERA-Interim (top), the corresponding residuals (center) and flattened residuals (bottom). Right Column: Empirical estimation (top), the corresponding residuals (center) and flattened residuals(bottom). FigureS07.pdf: Comparison with an empirical correction method (see Fig. 10 in Main Manuscript) - Left: Interferogram between Envisat acquisitions on 05/08/2004 and 08/06/2005 over Northern Chile on track 368. On color cycle is 20 cm along the Line-Of-Sight. Each contour line is 5 cm. The background shading is from SRTM DEM. Center Column: Prediction from ERA-Interim (top), the corresponding residuals (center) and flattened residuals (bottom). Right Column: Empirical estimation (top), the corresponding residuals (center) and flattened residuals(bottom). Comparing predictions from Global Atmospheric Models and MERIS: Additional examples Figures S08 and S09 show two additional examples of good prediction of the tropospheric delay by Global Atmospheric Models, corresponding to interferograms 16 and 4 in figure 4 of the main text. Figures S10 and S11 show two additional examples of bad prediction of the tropospheric delay by Global Atmospheric Models, corresponding to interferograms 14 and 3 in figure 4 of the main text. Figures S12 and S13 show two additional examples of average prediction of the tropospheric delay by Global Atmospheric Models, corresponding to interferograms 17 and 8 in figure 4 of the main text. FigureS08.pdf: Validation with MERIS - Left: An interferogram over the Makran area (South Pakistan) from Envisat SAR acquisitions on 11/30/2007 and 08/17/2007 on track 449. Six panels to the right show the corresponding wet, total and total de-ramped tropospheric delay predictions from ERA-I (Top) and MERIS (Bottom). One color cycle corresponds to 5 cm along the Line-Of-Sight direction and 10 cm contour lines are indicated. Background shading is from SRTM DEM. This case corresponds to the interferogram #16 in figure 4 of the main text FigureS09.pdf: Same as figure S08 for Envisat SAR acquisitions 03/25/2005 and 02/18/2005. This case corresponds to the interferogram #4 in figure 4 of the main text. FigureS10.pdf: Same as figure S08 for Envisat SAR acquisitions 08/12/2005 and 03/25/2005. This case corresponds to the interferogram #14 in figure 4 of the main text. FigureS11.pdf: Same as figure S08 for Envisat SAR acquisitions 03/19/2010 and 01/08/2010. This case corresponds to the interferogram #3 in figure 4 of the main text. FigureS12.pdf: Same as figure S08 for Envisat SAR acquisitions 07/23/3004 and 01/30/2004. This case corresponds to the interferogram #17 in figure 4 of the main text. FigureS13.pdf: Same as figure S08 for Envisat SAR acquisitions 08/12/2005 and 06/03/2005. This case corresponds to the interferogram #8 in figure 4 of the main text. Interferograms used to produce figure 8 of the main manuscript. The table presenting the interferograms used to produce figure 8 of the main manuscript is in the file IfgForFigure8.txt. IfgForFigure8.txt: Text file containing the following informations Column ÒNameÓ: Name of the interferogram, referring to figure 8 of the main tex. Column ÒSatelliteÓ: Name of the satellite. Column ÒAcq. Date 1Ó: First acquisition date (yyyymmdd). Column ÒAcq. Date 2Ó: Second Acquisition date (yyyymmdd). Column ÒRadar BandÓ: Radar Band and wavelength, cm.