English   español  
Please use this identifier to cite or link to this item: http://hdl.handle.net/10261/88934
logo share SHARE   Add this article to your Mendeley library MendeleyBASE
Visualizar otros formatos: MARC | Dublin Core | RDF | ORE | MODS | METS | DIDL
Exportar a otros formatos:


Towards an Improved Characterization of Instrumental Biases and Forward Model Errors in SMOS Observations over the Ocean

AuthorsGourrion, Jérôme ; Tenerelli, Joseph; Guimbard, Sébastien
Issue Date16-Apr-2013
CitationSMOS & Aquarius Science Workshop. Abstract book: 13-14 (2013)
AbstractThe Soil Moisture and Ocean Salinity (SMOS) satellite was launched on November 2, 2009 in the framework of the European Space Agency's (ESA's) Earth Explorer opportunity missions. Over the oceans, Sea Surface Salinity (SSS) is retrieved on a global basis with a spatio-temporal sampling appropriate for Ocean dynamics and Earth water cycle studies (Font 2010). The single payload onboard SMOS is the Microwave Imaging Radiometer using Aperture Synthesis (MIRAS), a novel fully-polarimetric L-band radiometer which estimates the brightness temperature by means of two-dimensional aperture synthesis interferometry. It consists of a Y-shaped set of 72 receivers (McMullan, 2008). More than 3 years after launch, the salinity product accuracy has still not reached the mission objective, even in the RFI-free open ocean domain. Main reasons are: 1) the challenging but intrinsically low sensitivity of L-band brightness temperature to sea surface, 2) the imperfection of the forward model used in the inversion procedure, 3) the spatio-temporal biases still present in the reconstructed brightness temperature. The present work is a contribution to adressing above-mentioned points 2 and 3. Several forward model deficiencies have been identified which propagate down to the retrieved salinity. If different studies have recently pointed out roughness dependent SSS errors and proposed updated formulations for the increment of emissivity due to surface roughness (Guimbard et al., 2012; Yin et al., 2012), the agreement in their results suggests that a robust improvement has been achieved. Nevertheless, another critical component of the forward model is the celestial signal scattered by the rough sea surface (Tenerelli 2008). The complexity of the biscattering problem and the large number of parameters involved makes highly difficult the procedure to improve its description empirically from real data. In spite of this, a recent work by J.Tenerelli has produced very promising results. The amplitude of the modeled signal near the specular direction is improved and better mimics the changes due to surface roughness variations. Nevertheless, there is still some discrepancy between parameters obtained when using different datasets, especially when using ascending or descending passes, and between different geometrical observation conditions i.e. incidence angle. Such inconsistency in the model parameters suggest an imperfection of the model physics. As mentioned in the introduction, latitudinal and seasonal biases are also affecting SMOS reconstructed TB ocean images (Tenerelli et al. 2010, Oliva et al. 2012) and retrieved salinity fields (Reul et al. 2012). Results suggest a correlation of the error with the sun illumination of the instrument through thermal effects, but attempts to cancel the corresponding biases at the calibration level are still not conclusive. In this work, it is assumed that such biases are essentially uniform across the field of view.
A key point in this discussion is that celestial reflection model errors and thermal instrumental biases both vary at latitudinal and seasonal scales. In the current approach, forward model updates are contaminated by the imperfect instrumental biases estimates and vice versa. The present work is an attempt to uncouple these two important steps. First, for a specific data subset where the celestial reflection signal is expected to be time-invariant, the temporal biases are estimated, an empirical correction applicable to the brightness temperatures is derived and a corrected data subset obtained. Second, the corrected dataset is used to obtain celestial reflection residuals. Their inconsistency with the current galactic model, primarily in terms of incidence angle dependence is analyzed to derive a modification of the model. Finally, after evaluating its performance, the updated model is evaluated for a much larger dataset and the instrumental biases are now evaluated both at the temporal and orbital scales. For a given latitudinal band, i.e. orbital position, and a limited set of locations in the FOV, a specific geometrical configuration is identified for which the celestial contamination does not significantly vary along the year. A data selection strategy developed for the antenna frame systematic errors study (Gourrion et al., 2012) is refined to characterize the instrumental temporal biases in that particular latitudinal range. Assuming that the thermally-induced instrumental biases are homogenous across the FOV, celestial reflection residuals are derived from a wide range of FOV locations but the same orbital location. Their analysis points out an imperfection in the shape of the bistatic scattering coefficients used in the computation of the celestial signal as scattered by the rough sea surface. Both theoretically-based and empirical ad-hoc modifications are tested to propose a modification of the bistatic scattering function
DescriptionSMOS & Aquarius Science Workshop, 15-17 April 2013, Brest, France
Publisher version (URL)http://www.congrexprojects.com/docs/default-source/13c07_docs3/smos-aquarius-abstract-book.pdf?sfvrsn=2
Appears in Collections:(ICM) Comunicaciones congresos
Files in This Item:
There are no files associated with this item.
Show full item record
Review this work

WARNING: Items in Digital.CSIC are protected by copyright, with all rights reserved, unless otherwise indicated.