GCOS GOOS WCRP/OOPC IX/3 

page 65 

are only necessary to correct any large scale satellite biases. Simulations were used to define 

a potential satellite bias error as a function of in situ data density. Buoy data were simulated at 

different grid resolutions to show their ability to correct the satellite biases. The goal of this 

study was to define the in situ network which would reduce simulated satellite biases of 2C 

below 0.5C. Results of the simulations showed that a buoy density of two buoys on a 10

o

spatial grid was required. The present in situ SST observing system was evaluated to define 

an equivalent buoy density, which allows ships to contribute along with the buoys. Seasonal 

maps of the equivalent buoy density were computed to determine where additional buoys 

were needed. The results will influence future buoy deployments. In addition, average 

potential satellite bias errors could be computed from the equivalent buoy density. This allows 

the evaluation of the present in situ observing system for climate SST. 

4. Other SST products 

An Integrated SST and Land Surface Temperatures (LST) analysis has been 

developed by NOAA/NCDC. This analysis combines separate analyses of SST and LST. The 

analysis of SST is ERSST.v2. The land analyses are derived from the NCDC Global 

Historical Climate Network (GHCN) temperature data. Global analyses with error estimates 

are produced on a global 5

o

 grid beginning in 1880. The ERSST.v2 paper has been published 

as Smith, T. M., and R. W. Reynolds, 2004: Improved Extended Reconstruction of SST 

(1854 1997). J. Climate, 17, 2466 247. The combined SST and LST paper has been 

submitted to the Journal of Climate as: Smith, T. M., and R. W. Reynolds, 2004: A global 

merged land and sea surface temperature reconstruction based on historical observations 

(1880 1997). 

The UK Met office is producing new analyses using a more flexible gridding system 

with improved bias corrections. These analyses include sampling, measurement and bias 

correction uncertainties at each grid value. A journal paper on these results is being prepared.