An algorithm for estimating Absolute Salinity in the global ocean

Abstract. To date, density and other thermodynamic properties of seawater have been calculated from Practical Salinity, S P. It is more accurate however to use Absolute Salinity, S A (the mass fraction of dissolved material in seawater). Absolute Salinity S A can be expressed in terms of Practical Salinity S P as S A=(35.165 04 g kg-1/35)S P+δ S A(φ, λ, p) where δ S A is the Absolute Salinity Anomaly as a function of longitude φ, latitude λ and pressure. When a seawater sample has standard composition (i.e. the ratios of the constituents of sea salt are the same as those of surface water of the North Atlantic), the Absolute Salinity Anomaly is zero. When seawater is not of standard composition, the Absolute Salinity Anomaly needs to be estimated; this anomaly is as large as 0.025 g kg−1 in the northernmost North Pacific. Here we provide an algorithm for estimating Absolute Salinity Anomaly for any location (φ, λ, p) in the world ocean. To develop this algorithm we use the Absolute Salinity Anomaly that is found by comparing the density calculated from Practical Salinity to the density measured in the laboratory. These estimates of Absolute Salinity Anomaly however are limited to the number of available observations (namely 811). To expand our data set we take advantage of approximate relationships between Absolute Salinity Anomaly and silicate concentrations (which are available globally). We approximate the laboratory-determined values of δ S A of the 811 seawater samples as a series of simple functions of the silicate concentration of the seawater sample and latitude; one function for each ocean basin. We use these basin-specific correlations and a digital atlas of silicate in the world ocean to deduce the Absolute Salinity Anomaly globally and this is stored as an atlas, δ S A (φ, λ, p). This atlas can be interpolated to the latitude, longitude and pressure of a seawater sample to estimate its Absolute Salinity Anomaly. For the 811 samples studied, ignoring the Absolute Salinity Anomaly results in a standard error in S A of 0.0107 g kg-1. Using our algorithm for δ S A reduces the error to 0.0048 g kg−1, reducing the mean square error by a factor of five. The number of sea water samples used to develop the correlation relationship is limited, and we hope that the algorithm and error can be improved as further data becomes available.