Aquarius Salinity Retrieval Algorithm Abstracts


Aquarius Salinity Retrieval Algorithm

The Aquarius salinity retrieval algorithm beyond Version 3.0
Contact author: Frank Wentz, <>
All Authors:
Frank Wentz, Remote Sensing Systems
Thomas Meissner, Remote Sensing Systems
Joel Scott, Remote Sensing Systems
Kyle Hilburn, Remote Sensing Systems

Our presentation will first give an overview of updates and improvements that we plan to make in the Aquarius salinity retrieval algorithm for the next major release. Then we will describe in more detail the following two improvements:
1. Repeat of On-Orbit Simulations using antenna pattern with proper spillover value: Analysis of Aquarius observations have shown that the true spillover of the Aquarius antenna pattern is between 1 and 1.5% less than the GRASP 2012 simulations, which were used for the previous On-Orbit Simulations. These simulations are used to specify a number of tables used by the Aquarius Operational Processor. We will repeat the simulations using revised antenna patterns having the correct spillover.
2. Improvement of land correction: The sidelobe land correction that is used in V3.0 estimates the increase in the brightness temperature of Aquarius footprints close to the coast due to emission from land that enters the antenna sidelobes. The estimate is based on land emission radiative transfer model that uses auxiliary input for soil type, vegetation type, soil moisture and land surface temperature. We have compared this land emission RTM to measured Aquarius TB over land and found discrepancies that exceed +/- 20 K over large areas. We therefore expect that the use of this land emission model will result in significant errors in the land correction algorithm and that it can be improved by using a land emission model that is based on measured Aquarius rather than RTM TB. We have derived such a model. For deriving the land emission correction it is necessary to know the land emission far off boresight and therefore it is necessary to extrapolate the measured Aquarius land TB to the full range of Earth incidence angles between 0 and 90 deg. We have also adjusted the antenna gain pattern for calculating the TB contribution from the sidelobe. We will recomputed the land-correction tables using the new land TB model in conjunction with the new antenna patterns discussed in Item 1.

Mitigation of large scale biases in the Aquarius salinity retrievals
Contact author: Thomas Meissner, <>
All Authors:
Thomas Meissner, Remote Sensing Systems
Frank Wentz, Remote Sensing Systems
Joel Scott, Remote Sensing Systems

The Aquarius Version 3.0 salinity retrievals have salty biases at mid-high latitudes and fresh biases in the tropics and subtropics when compared to ARGO or HYCOM. In addition we observe biases between the three Aquarius beams. These biases are largely due to imperfection in the geophysical model function that us used in the salinity retrieval algorithm, mainly the dielectric constant, the oxygen absorption and the surface roughness model. Version 3.0 provides an empirical post-hoc adjusted salinity that mitigates these biases by stratifying them as function of SST and removing them from the retrieved value. For upcoming releases it is necessary to obtain a better understanding of the physical cause for these biases and also to perform the correction at the TB level rather the salinity level. We have analyzed measured minus expected TB as 2-dimensional function of SST and wind speed. The result suggests that the deficiencies on the geophysical model function are a combination of several effects:
1. Small errors in the dielectric constant model.
2. Small errors in the oxygen absorption model.
3. Small uncertainties in the auxiliary SST fields.
4. The assumed SST dependence in the wind induced emissivity.
We have derived and analyzed the magnitude and channel signature of each of these error sources. We find that the largest part of the observed biases is due to errors in the SST dependence in the wind induced emissivity. Based on our analysis, we derive an adjustment to the geophysical model function and discuss its impact on the performance of the salinity retrievals, in particular the large-scale biases and the residual interbeam biases.

Effect of sea water dielectric constant model and sea surface temperature ancillary data on remote sensing of sea surface salinity
Contact author: Emmanuel Dinnat, <>
All Authors:
Emmanuel Dinnat, Chapman University/NASA-GSFC
Jacqueline Boutin, CNRS/IRD/UPMC/MNHN
David Le Vine, NASA-GSFC

ESA’s Soil Moisture and Ocean Salinity (SMOS) and NASA’s Aquarius use L-band (1.4 GHz) radiometers to measure emission from the sea surface and retrieve SSS. Significant differences in SSS retrieved by both sensors are observed, with SMOS SSS being generally lower than Aquarius SSS, except for very cold waters where SMOS SSS is the highest overall. Differences are mostly between -1 psu and +1 psu (psu, practical salinity unit), with a significant regional and latitudinal dependence. We investigate the impact of the vicarious calibration and retrieval algorithm used by both mission on these differences.
One notable difference between the two missions is the sea water dielectric constant model. SMOS uses the model by Klein and Swift (1977) [1] and Aquarius uses the model by Meissner and Wentz (2012) [2]. The dielectric constant model is used: 1/ to calibrate the instruments by comparing radiometric measurements to forward model simulations, and 2/ to invert SSS from surface brightness temperature (Tb). In order to assess the impact of the dielectric constant model on the SSS difference, we reprocess the Aquarius data using the model used for SMOS. Specifically, we use the Klein and Swift model for the reference ocean used in the calibration of Aquarius; then we used it again, keeping all other factors the same, to perform the inversion to obtain SSS.
Another important difference between both missions algorithms is the ancillary product used for sea surface temperature (SST). Aquarius uses the daily optimally interpolated (OI) SST from NOAA, that relies on in situ (ship and buoys) measurements and satellite data from the Advanced Very High Resolution Radiometer (AVHRR) infrared (IR) sensor [3]. We compare this SST product to the Operational Sea Surface Temperature and Sea Ice Analysis (OSTIA) produced by the Met Office [4] that uses microwave sensors in addition to IR sensor. Similarly to what was done for the dielectric constant study, we reprocess the Aquarius data, including the calibration and retrieval steps, using the OSTIA SST instead of the NOAA OI SST.
We will present the impact of the dielectric constant model and the SST product on the differences between SMOS and Aquarius and show comparisons with in situ data from the Argo global network of free-drifting floats.
[1] L. A. Klein and C. T. Swift, “An improved model for the dielectric constant of sea water at microwave frequencies,” IEEE Transactions on Antennas and Propagation, vol. AP-25, no. 1, pp. 104–111, 1977.
[2] T. Meissner and F. J. Wentz 2012, “The Emissivity of the Ocean Surface Between 6 and 90 GHz Over a Large Range of Wind Speeds and Earth Incidence Angles”, IEEE Trans. Geosci. Remote Sens., vol. 50, no. 8, pp. 3004–3026, Aug. 2012.
[3] R. Reynolds et al., “Daily High-Resolution-Blended Analyses for Sea Surface Temperature”, J. Climate, vol. 20, pp. 5473–5496, 2007.
[4] C. J. Donlon et al., “The Operational Sea Surface Temperature and Sea Ice Analysis (OSTIA) system”, Remote Sensing of Environment, vol. 116, pp. 140-158, Jan. 2012.