1. DOWNSCALED SOIL MOISTURE DATA FROM THE METOP ASCAT SCATTEROMETER FOR HYDROLOGICAL USERS IN EUROPE Stefan Hasenauer (1), Barbara Zeiner (2), Alexander Jann (2) and Wolfgang Wagner (1) (1) (1) Vienna University of Technology, Institute of Photogrammetry and Remote Sensing, Vienna, Austria (2) (2) Central Institute for Meteorology and Geodynamics (ZAMG), Vienna, Austria REFERENCESCONTACT Stefan Hasenauer, TU-Wien, Austria sh@ipf.tuwien.ac.at http://www.ipf.tuwien.ac.at/radar http://hsaf.meteoam.it EUMETSAT/ESA SCAT Science Conference Darmstadt, Germany, 11-13 April 2011 This work is presenting the soil moisture product generation chain as well as the product characteristics of the value-added 1km downscaled surface soil moisture product SM-OBS-2. The input is the 25km Advanced Scatterometer Instrument (ASCAT) surface soil moisture product, available in near real-time BUFR format with nearly daily coverage for the whole globe since 2007 from the Metop satellite. The downscaled SM-OBS-2 product is generated by exploiting the linear relationship between the mean surface soil moisture characteristics of a small-scale local area and its regional-scale mean. This task is obtained by analysing multi-annual ENVISAT ASAR (C-band) and Metop ASCAT backscatter time series, stored in a parameter database for each European grid point. Finally, the SM-OBS-2 downscaled soil moisture is generated with the processor WARP_H (Water Retrieval Package for H-SAF) by applying the linear relationship of the European parameter database on the resampled regional-scale measurements resulting in relative soil moisture values at 1km scale in near real-time.INTRODUCTION H-SAF: Satellite Application Facility on Support to Operational Hydrology and Water Management Lifetime: H-SAF Continuous Development and Operations Phase (CDOP): Sep 2010 – Feb 2012 Conceptual architecture for soil moisture products: - Near-real time acquisition of METOP ASCAT (EUMETCast service) - Generation of SM-OBS-1 (25km product) - Timeliness: ~130 min - Format: BUFR - at ZAMG - Generation of SM-OBS-2 (downscaled 1km product) - Concept: Temporal stability of backscatter from ENVISAT ASAR - Format: BUFR - at ZAMG - Generation of SM-DAS-2 (assimilated product) - at ECMWF Fig. 9: The original ASCAT input image at 25 km resolution over region in central Europe. No-data values (e.g. North Sea, Alps) are masked Fig. 7: The coverage of ASAR GM data over Europe for generating downscaling parameters (Mar 2011) 3) MetOp ASCAT Downscaling (1 km) - Applying linear regression model: DSSM(local) = c + d * SSM(regional), where: DSSM…downscaled surface soil moisture, c…intercept, SSM……surface soil moisture resampled, d…slope; - Quality flag characterisation: checks of nominal data ranges and product applicability; - Output file specification: BUFR. 1) Pre-processing: ENVISAT Parameter Database (1 km) - Geocoding: Range Doppler approach using precise orbit files (ENVISAT DORIS); - Resampling: fixed grid, sampling interval 15 arc seconds; - Normalisation: backscatter-normalisation to a reference incidence angle at 30°; - Scaling: between dry and wet references, derived from long time-series of ERS-1/2 scatterometers; - Correlation: moving window of 1 km (local) and aggregated 25 km (regional), calculation of correlation, intercept c and slope d for Europe. Fig. 10: Example of downscaled surface soil moisture (DSSM) at 1 km resolution, available in near-real time. No-data values are masked and given a quality flag information 2) MetOp ASCAT Resampling (25 km  1 km) - Fixed grid, sampling interval 15 arc seconds, area of Europe only; - Delaunay triangulation; - Linear interpolation to 1 km resolution of: Fig. 8: Correlation between local and regional ASAR GM backscatter over central Europe. Envisat METOP Series Sentinel-1 20002005 2020 20102015 H-SAF project PrototypeCDOP Radar satellites Validation approaches: 1) Ground based systems (in-situ) - Single campaign datasets - Continuous monitoring sites (from ISMN network) 2) Modelled data - Hydrological model data - Soil-vegetation-atmosphere-transfer model 3) Inter-comparison with other satellite products - Active vs. passive microwave products Table 1: Overview of validation results for SM-OBS-2 (Development phase, recent) Fig. 5: Hydrological validation test countries Product examples Fig. 1: Lifetime of project and sensors used Fig. 2: The soil moisture product generation chain H-SAF BACKGROUND METHODOLOGY VALIDATION RESULTS New downscaled satellite soil moisture product available for Europe: - Downscaled soil moisture (with ENVISAT ASAR, BUFR, NRT): 22 Sep 2009 – now Positive results with in-situ and hydrological model data comparison - Further work ongoing with hydrological model data/data assimilation Challenges and research needs - File sizes and large amount of data sets - Investigating spatial variability and error characterisation of scaling approaches, e.g. Blöschl 2009 Outlook - Improving geographic coverage (Africa, Australia) - Sentinel-1 investigations Temporal stability concept (Vachaud et al., 1985) - Local soil moisture is often highly correlated with the regional soil moisture - Soil moisture patterns tend to persist in time Hypothesis - Temporally stable soil moisture patterns lead to temporally stable radar backscatter patterns (Wagner et al., 2008) Fig. 4: 20 time domain reflectometry (TDR) stations (5 cm depth), Duero basin (Spain) and their mean (bold black diamonds), 2003- 2005 1 km 25 km Blöschl, G., Komma, J., Hasenauer, S. (2009): “Hydrological downscaling of soil moisture”. Final report of the Visiting Scientist Activity to the Satellite Application Facility on Support to Operational Hydrology and Water Management (H-SAF). 64 p. Vachaud, G., A. Passerat de Silans, P. Balabanis, M. Vauclin (1985): Temporal Stability of Spatially Measured Soil Water Probability Density Function, Soil Sci. Soc. Am. J., 1985, 49, 822-828. Wagner, W., C. Pathe, M. Doubkova, D. Sabel, A. Bartsch, S Hasenauer, G. Blöschl, K. Scipal, J. Martínez-Fernández, A. Löw (2008): Temporal stability of soil moisture and radar backscatter observed by the Advanced Synthetic Aperture Radar (ASAR), Sensors, 8, 1174-1197THEORY SUMMARY AND OUTLOOK Fig. 3: Correlation regional vs. local soil moisture Fig. 6: The processing chain concept. Left: offline part (TU-Wien). Right: near real-time (ZAMG) Fig. 11: First downscaled Image over Australia a)soil moisture; b)error of soil moisture; c)quality flags. The authors would like to thank the visiting scientist Luca Brocca (CNR-IRPI, Italy) for providing exhaustive validation results. ACKNOWLEDGEMENTS