Geoscience Reference
In-Depth Information
a
b
c
d
Fig. 25.11 Diurnal cycle variation of 2-m temperature ( ı C) in ( a ) GTS observation, ( b )CTRL
model forecasts, and 10-m wind speed (ms 1 )in( c ) NCAR GFS analysis, ( d ) CTRL model
forecasts. Twenty-four-hours forecasts only (Updated from Xu et al. ( 2009 ))
model forecasts. Note that the analysis data suggests a strong diurnal cycle variation
over northwest Iran and north Afghanistan. Overall, however, the model forecasts
of the diurnal cycle are consistent with the analysis.
25.6
Impact of Satellite Data Assimilation
Results from the previous section suggest that, aside from the 2-m temperature,
errors in forecast variables are dominated by nonsystematic errors, which are caused
by uncertainties in the model initial conditions or unresolvable differences in scales
between the forecasts and observations ( Nutter and Manobianco 1999 ). The model
initial conditions are very important factors affecting model forecasts. For the
purpose of understanding the role of initial conditions in the accuracy of forecasts,
we will now consider satellite observation data assimilation in this section.
 
Search WWH ::




Custom Search