Understanding the accuracies of satellite-derived sea surface salinity (SSS) measurements in
depicting temporal changes and the dependence of the accuracies on spatiotemporal scales
are important to capability assessment, future mission design, and applications to study
oceanic phenomena of different spatiotemporal scales. This study quantifies the consistency
between Aquarius Version-4 monthly gridded SSS (released in late 2015) with two widely
used Argo monthly gridded near-surface salinity products. The analysis focused on their
consistency in depicting temporal changes (including seasonal and non-seasonal) on various
spatial scales: 1˚ x1˚ , 3˚ x3˚ , and 10˚ x10˚ . Globally averaged standard deviation (STD)
values for Aquarius-Argo salinity differences on these three spatial scales are 0.16,
0.14, 0.09 psu, compared to those between the two Argo products of 0.10, 0.09,
and 0.04 psu. Aquarius SSS compare better with Argo data on non-seasonal (e.g.,
interannual and intraseasonal) than for seasonal time scales. The seasonal Aquarius-Argo
SSS differences are mostly concentrated at high latitudes. The Aquarius team is
making active efforts to further reduce these high-latitude seasonal biases. The
consistency between Aquarius and Argo salinity is similar to that between the two
Argo products in the tropics and subtropics for non-seasonal signals, and in the
tropics for seasonal signals. Therefore, the representativeness errors of the Argo
products for various spatial scales (related to sampling and gridding) need to be
taken into account when estimating the uncertainty of Aquarius SSS. The globally
averaged uncertainty of large-scale (10˚ x10˚ ) non-seasonal Aquarius SSS is
approximately 0.04 psu. These estimates reflect the significant improvements of
Aquarius Version-4 SSS over the previous versions. The estimates can be used as
baseline requirements for future ocean salinity missions from space. The spatial
distribution of the uncertainty estimates is also useful for assimilation of Aquarius SSS. |