6533b7ddfe1ef96bd1273dae
RESEARCH PRODUCT
A spatially consistent downscaling approach for SMOS using an adaptive window
Adriano CampsDavid ChaparroMiriam PablosGerard PortalM. Vall-llosseraLuciana RossatoMaria Pilessubject
Atmospheric ScienceBrightnessTeledeteccióMean squared error010504 meteorology & atmospheric sciencesREMEDHUS0211 other engineering and technologiesHigh resolution02 engineering and technology01 natural sciencesNormalized Difference Vegetation IndexBECComputers in Earth SciencesImage resolution021101 geological & geomatics engineering0105 earth and related environmental sciencesRemote sensingNative resolutionAdaptive moving windowLow resolutionMoving windowRemote sensing:Enginyeria de la telecomunicació::Radiocomunicació i exploració electromagnètica::Teledetecció [Àrees temàtiques de la UPC]Orbit (dynamics)RadiometryEnvironmental scienceSpatial variabilitySoil moistureSòls -- HumitatDownscalingSMOSdescription
The European Space Agency (ESA)'s Soil Moisture and Ocean Salinity (SMOS) is the first spaceborne mission using L-band radiometry to monitor the Earth's global surface soil moisture (SM). After more than 7 years in orbit, many studies have contributed to improve the quality and applicability of SMOS-derived SM maps. In this research, a novel downscaling algorithm for SMOS is proposed to obtain high-resolution (HR) SM maps at 1 km (L4), from the ∼40 km native resolution of the instrument. This algorithm introduces the concept of a shape adaptive moving window as an improvement of the current semi-empirical downscaling approach at SMOS Barcelona Expert Center, based on the “universal triangle”. Its inputs are as follows: the SMOS SM (L3 at ∼40 km spatial resolution), the vertical and the horizontal SMOS brightness temperatures (L1C at ∼40 km), and the HR normalized difference vegetation index and land surface temperature from optical-based sensors. The proposed method has the following advantages: HR SM maps are obtained while maintaining the dynamic range from the original L3 product; energy is conserved, because differences between aggregated L4 and L3 SM maps are negligible; and it can be applied to continental areas, even when they integrate different climates. A comparison of SMOS L3 and L4 products with in situ data for networks allocated in Spain and Australia shows good agreement in terms of correlation and root mean square error. The proposed method is shown to capture 1-km SM spatial variability while preserving the quality of SMOS at its native resolution.
| year | journal | country | edition | language |
|---|---|---|---|---|
| 2017-01-01 |