Title: Temporal Variability of Thermosteric
 1Temporal Variability of Thermosteric  Halosteric 
Components of Sea Level Change, 1955-2005
S. Levitus, J. Antonov, T. Boyer, R. Locarnini, 
H. Garcia, A. Mishonov
NOAA/NESDIS/NODC/Ocean Climate Laboratory, Silver 
Spring, MD
Representativeness of the data in the ocean 
profile database Comparison of NCDC Yearly SST 
anomaly  Levitus et al. (2005) yearly SST anomaly
We have used data from the World Ocean Database 
2005 (WOD05) plus additional data to compute new 
estimates of the temporal variability of ocean 
heat content and the thermosteric component of 
sea level change for 1955-2005. WOD05 contains 
approximately 0.9 million temperature profiles 
not previously available as part of the WOD 
series. These include historical data acquired as 
part of the IOC Global Oceanographic Data 
Archaeology and Rescue project, real-time and 
delayed-mode data sent via the GTS as part of the 
IOC/GTSPP project, Argo profiling float data from 
the GODAE server, and modern data not sent via 
the GTS acquired through international exchange 
as part of the IOC sponsored World Ocean 
Database project. Results presented here are 
still preliminary. Additional quality control is 
being performed on the data.
Hansen et al. (2005) have suggested that lack of 
data may partly be responsible for the strong 
interdecadal variability observed in ocean heat 
content (and hence the thermosteric component of 
sea level change). Gregory et al. (2004) and 
manuscripts by other authors have examined time 
series of the data by 1 squares in the World 
Ocean Database (WOD) and reach similar 
conclusions. Here we present evidence that this 
variability is real. The figure below compares 
the yearly, global average SST anomaly based on 
the relatively sparse data set of WOD with the 
same quantity based on ICOADS and satellite SST 
measurements (constrained by SST observations 
from drifting buoys) from Smith and Reynolds 
(2004) (curve labeled NCDC). The correlation 
between the two series is 0.96. The reason for 
such good agreement between the two series is 
that the anomalies contributing to the global 
average are of large spatial scales (gyre  
basin) and relatively long time scales (Levitus 
et al., 2005). Compositing and analyzing data by 
semester (6-month) and seasonal (3-month) still 
results in the same signal characterized by large 
interdecadal variability. Even relatively sparse 
sampling captures this signal. The objective 
analysis procedure used to generate the WODS 
based SST and heat content fields is basically a 
running spatial average of spatial scale 500 km, 
not data restricted to individual 1 squares. 
Despite this agreement, there is an additional 
need for more data. 
History of the Number of Temperature and Salinity 
Profiles Available Electronically from NODC
Approximately 3 million of the 5.5 million 
increase in temperature profiles since the GODAR 
project began in 1993 are for the pre-1991 
period. It is the rescue of these data that 
have allowed the first estimates of ocean heat 
content and steric sea level components to be 
made.
Basin integrals of the thermosteric component of 
sea level change, 0-700m, 1955-2005
Pacific Ocean
Atlantic Ocean
Linear trends (mm/yr) (1955-1998) of zonally 
averaged steric sea level and its components
The thermosteric  halosteric components of sea 
level change may compensate each other, or act in 
concert to affect sea level. Here, (Levitus et 
al., 2005) we document this for the zonally 
averaged linear trend of these components. Green 
line is the halosteric component of sea level 
 Dashed red line is the thermosteric component of 
sea level change based on all T data Solid red 
line is the thermosteric component of sea level 
based on T obs. that have simultaneous 
measurement of salinity Solid black line is the 
steric sea level sum of halosteric and 
thermosteric (T-S simultaneous).
Indian Ocean
World Ocean 
Indian Ocean