Variational Data Assimilation for the Global Ocean - Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications - page 309

a

Temperature Prediction Error (C) 0 M Depth Temperature Prediction Error (C) 0 M Depth

<0 .2 .4 .6 .8 1 1.2> <0 .2 .4 .6 .8 1 1.2>

Temperature Prediction Error (C) 0 M Depth

<0 .2 .4 .6 .8 1 1.2>

Fig. 13.5 Temperature ( ı C) background error standard deviations valid 20 January 2012 in
global HYCOM analysis domains: Atlantic, Indian, and Pacific. ( a )0Mdepth,( b ) 150 M depth,
( c ) 300 M depth

13.3.5 Observation Error Variances

The observation errors and the background errors are assumed to be uncorrelated,
and errors associated with observations made at different locations and at different
times are also assumed to be uncorrelated. As a result of these assumptions, the
observation error covariance matrix R is set equal to 1 C E o along the diagonal

and zero elsewhere. Note that E o represents observation error variances (e o ) nor-
malized by the background error variances interpolated to the observation location
(E o e b ). Observation errors are computed as the sum of a measurement

error and a representation error. Measurement errors reflect the accuracy of the
instruments and the ambient conditions in which the instruments operate. These
errors are fairly well known for many ocean observing systems, although the
errors can change in time due to calibration drift of the instruments and other

e o = D