SMOS/ELDAS workshop

Last update: 6 December 2002

Introduction

In order to better link the SMOS and the ELDAS projects and to improve the collaboration between the working groups, we defined an 'SMOS Simulator Assimilation Experiment', which would ideally be carried out within the framework of ELDAS at the end of the project (end 2004). This experiment covers a limited period of 2 weeks over Europe, and is designed to evaluate the possible impact of the future SMOS sensor on the products generated by the ELDAS data assimilation system.

We kindly ask the Invited Speakers to review this experiment plan with respect to a) the SMOS specific features (antenna sampling, overpass times and frequency) b) the expected accuracy of the brightness temperatures c) potential inversion schemes for the data assimilation d) preferable data assimilation techniques. Recommendations can either be based on synthetic experiments or on previous field experiments with L-band radiometers. Below a sketch of the proposed experimental design is outlined.

Agenda

The agenda for the workshop is as follows:

Chair: Matthias Drusch

• 09:30 start & welcome
• 09:45 Bart vd Hurk: overview of ELDAS data base and outline of the SMOS experiment
• 10:15 Yann Kerr: General description of SMOS and potential relevance to ELDAS-like applications
• 10:45 break
• 11:00 A. Camps: THE SMOS simulator
• 11:45 J.C. Calvet: brightness temperature calculation and retrievals
• 12:25 Ryan Teuling: presentation of PhD research plan "Surface heterogeneity and data assimilation"
• 12:30 lunch break
• 13:30 Henning Wilker: the radiative transfer model coded in TESSEL
• 14:00 Gisela Seuffert: first results of microwave data assimilation in TESSEL
• 14:30 break
• 14:45 Discussion
• 16:00 Matthias Drusch: Conclusions
• 16:15 Closure

Draft Plan: SMOS Simulator Assimilation Experiment

We propose to use the ELDAS data base and the SVAT models (e.g. TESSEL) to produce a 14 day data set for Europe, which comprises the relevant input parameters for the SMOS simulator (soil moisture content, soil temperature, vegetation and roughness characteristics, atmospheric profiles). Based on this 'true' data set, the SMOS simulator will be used to derive top of the atmosphere brightness temperatures with the instrument's characteristic resolution and scan geometry.

The set of brightness temperatures will then be assimilated into the ELDAS model suite using the assimilation scheme developed within the project's framework. Currently, this scheme assumes that data can be offered as a gridded database, and data assimilation is carried out in model space. In order to represent 'reality', noise will be added to this brightness temperature data set, derived from perturbed input data to the simulator (e.g. incomplete and / or erroneous land surface characteristics). In addition, noise will emanate from the difference in radiative transfer models in use in the SMOS simulator and in the ELDAS-infrastructure.

The impact of the (synthetic) 1.4 GHz brightness temperatures on the retrieved soil moisture and atmospheric parameters will be analysed with respect to:

• the imposed errors: where and / or when are errors in the brightness temperature too high to be of use for soil moisture assimilation?
• temporal resolution of SMOS data: is the currently proposed resolution adequate for retrieving deeper soil moisture from brightness temperature data?
• horizontal resolution of SMOS data: does the complexity of the SMOS antenna gain function allow to represent the brightness temperature as a gridded database, or is it necessary to do the data assimilation in observation space?
• importance of ancillary information: is simultaneous assimilation of other data necessary to avoid potential bias problems in the retrievals (e.g. surface temperature from Meteosat, or higher frequency microwave brightness temperatures from AMSR)