GSI/WRF Regional Data Assimilation System and Its Application in theWeather Forecasts over Southwest Asia - Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications - page 623

b

a

c d

Fig. 25.15 Bias (model - observation) of temperature ( a : ı C), geopotential height ( b : gpm), zonal
wind ( c :ms 1 ) and meridional wind ( d :ms 1 ) at 500 hPa for 24-h forecasts averaged over the
1-month period of May, 2006 (Updated from Xu et al. ( 2009 ))

Note that SS in ( 25.6 ) is a function of the DA forecasts ( d ), the CTRL reference

forecasts ( ). The MSE (d, a) and MSE (r, a) are

as defined in ( 25.2 ) indicating the mean-square-error of DA and CTRL forecasts
relative to the analysis, respectively. Therefore, the greater positive SS values reflect
increasing positive skill over the performance of the reference forecasts.
Figure 25.16 depicted the results for the 2-m temperature, 10-m wind speed and
precipitation forecasts over the nine locations defined in Fig. 25.1 . With regards to
the 2-m temperature forecasts, the statistical analysis (Fig. 25.16 a) indicates that
all SS in the different locations are positive for the 24- and 48-h forecasts, but the
SS for 48-h forecasts in most regions is greatly diminished in relation to that of
the 24-h forecasts. The SS in the north Iranian Zagros Mountains (B) and west
Himalaya Mountains (E) is about 10-20 % less than that in the lower mountains or
plain areas. Compared with the results shown in Fig. 25.2 , we find that the forecast
errors in the high mountain areas are mainly from the model systematic errors and
the nonsystematic errors make a relatively smaller contribution to the total forecast

r ), and the analyzed quantity ( a