Quantifying Observation Impact for a Limited Area Atmospheric Forecast Model - Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications - page 138

a

Cloud Winds

60N

50N

40N

30N

20N

10N

140w 130w 120w 110w 100w 90w 80w 70w 60w

− 3 − 2.2 − 1.7 − 1.2 − 1 − 0.7 − 0.5 − 0.3 − 0.1 − 0.05 0.05 0.1 0.3 0.5 0.7 1.2

b

Cloud Winds

60N

50N

40N

30N

20N

10N

140w 130w 120w 110w 100w 90w 80w 70w 60w

− 3 − 2.2 − 1.7 − 1.2 − 1 − 0.7 − 0.5 − 0.3 − 0.1 − 0.05 0.05 0.1 0.3 0.5 0.7 1.2

Fig. 6.6 Impacts of feature track satellite winds at observation locations (multiplied by 1,000) for
the same forecasts in Fig. 6.5 .The horizontal grid spacing was ( a )90kmand( b )30km.Values
are in units of J kg 1

for three different metrics. First, the error was calculated in a dry energy norm,
but only for the the lowest ten model layers (roughly the extent of the planetary
boundary layer). Then a moisture term was added to the energy norm calculated over
the same area. Finally, the error was calculated in terms of modified refractive index
(only depends on the temperature and humidity variables). The results for these
different metrics are shown in Figs. 6.7 and 6.8 . Although the units are different,