Earth system assimilation for the cryosphere: Status and way forward [TD-01]

This report outlines plans for activities to be performed in the first phase of the Data Assimilation and Numerical Testing for Copernicus Expansion Missions (DANTEX) project for the Copernicus Imaging Microwave Radiometer (CIMR) and Copernicus Polar Ice and Snow Topography Altimeter (CRISTAL). In order to put these plans in context, the report first provides a non-exhaustive review of the state of the art for radiative transfer (RT) models, tools and data assimilation (DA) techniques applicable to CIMR and CRISTAL, focusing primarily on the numerical weather prediction (NWP) system at the European Centre for Medium-Range Weather Forecasts (ECMWF). It then gives an overview of wider long-term developments relevant to the exploitation of CIMR and CRISTAL observations, before outlining a roadmap of specific CIMR- and CRISTAL-related activities to be performed during the current phase of the DANTEX project. The aim of the developments planned here is to prepare and advance scientific capabilities for the exploitation of CIMR and CRISTAL-like observations in ECMWF’s coupled DA system, with a particular focus on the cryosphere. As agreed in the DANTEX project contract, the following list of geophysical variables in the cryosphere are considered:

• Sea ice concentration (SIC)
• Sea ice extent (SIE)
• Sea ice thickness (SIT)
• Snow depth
• Surface elevation of polar glaciers

Information on SIC, SIE and SIT can be provided by CIMR, whereas CRISTAL provides information on SIT, snow depth and surface elevation.  It should be noted that SIE is a diagnostic variable derived from SIC based on an SIC threshold, which is usually taken to be 15%[1]. The representation of glaciers and ice sheets in the current ECMWF system is not mature enough to enable the exploitation of satellite data to constrain polar glaciers. Therefore, the work conducted in DANTEX will focus on SIC, SIE, SIT and snow depth.

At ECMWF, the general aim is to assimilate observations with as little processing as possible. For satellite observations this means using instrument data (level 1 (L1) in World Meteorological Organization (WMO) terms) as opposed to retrieved geophysical products (level 2 (L2) in WMO terms). This acknowledges that satellite instruments are sensitive to many different aspects of the earth system and not just the target geophysical variables. The assimilation of instrument data (level 1) means that all the information can potentially be used, not just the narrower list of target geophysical variables (as available at level 2). This approach also allows more flexibility for the DA system to respond (possibly exploiting information from different Earth system components) and make the accounting of observation errors simpler. Following this approach, for CIMR, we will be assimilating radiances. On the other hand, for CRISTAL, the aim for the current phase of the project is to assimilate freeboard measurements, with the understanding that this is a tractable stepping stone towards fuller exploitation of data provided by the instruments on board the satellite. In any case, whatever the assimilated variable is, the target geophysical variables (SIC, SIT etc.) are produced by the DA system analysis. Such geophysical variables will also be featured in forecast products.