Supplement to "Arctic Ocean liquid freshwater storage trend 1992 - 2012" B. Rabe, M. Karcher, F. Kauker, U. Schauer (Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany) J. M. Toole, R. A. Krishfield (Woods Hole Oceanographic Institution, Woods Hole, USA) S. Pisarev (Shirshov Institute of Oceanology, Moscow, Russia) T. Kikuchi (Japan Agency for Marine Science and Technology, Yokosuka, Japan) J. Su (Key Laboratory of Physical Oceanography, Ocean University of China, Qingdao, China) Geophysical Research Letters, 2013 Introduction This supplement describes the data sources, processing, and further technical details of the objective mapping and error estimates. It includes a table with observational data sources, a figure showing the data distribution in time, one figure showing the horizontal grid used for the objective mapping, and two figures of the monthly timeseries and statistical errors from the objective mapping. All are contained in one PDF named "supplement01.pdf". The structure of the text is as follows: 1. Observations 2. Data processing procedures for salinity profiles 3. Liquid freshwater calculations 4. Objective mapping 5. Contributors to liquid freshwater trends 6. Uncertainty Table 1. Data sources for salinity observations used in this study. 1.1 Expedition, project or institute 1.2 Year(s) 1.3 Platform 1.4 Source URL (last accessed in June 2013) or contact Table 2. Observed annual mean liquid freshwater content of the upper Arctic Ocean basin 1.1 Year 1.2 Liquid freshwater content (10^4 km^3) Figure 1. Locations of salinity profile observations used to calculate liquid freshwater inventories: (a) 1992 - 2005 and (b) 2006 - 2012. The time of each observation is shown in colour. Figure 2. Maps of the Arctic Ocean with grid points used in the objective mapping of the data in the domain under study. The bathymetry contours are based on the IBCAO database (IBCAO; Jacobson, 2008) with contour lines representing depths of 100, 200, 500, 750, 1000, 2000, 3000 and 4000 m. The domain under study is enclosed by the 500 m isobath and a cut-off at 82degN north of the Fram Strait (see manuscript text, Figure 1). Figure 3. Timeseries of monthly values of LFWC (dots) and the corresponding statistical mapping error (error bars), based on the values at each grid point. Figure 4. Timeseries of monthly values of spatially averaged depth of the 34 isohaline, h (dots) and the corresponding statistical mapping error (error bars), based on the values at each grid point.