Fig. 1
Schematic diagram for the ASDA method.
Fig. 2
(a) Geographical areas of domains 1, 2, and 3 and (b) locations of radar observation sites operated by Korea Meteorological Administration (black), Korea Air Force (red), and United States of America Air Force (blue). Locations of automatic weather station (black cross) and radiosonde (green triangle) observation sites over South Korea are also indicated.
Table 1. Brief description of numerical experiments
CONTROL––3DVAR3D-Var1800 UTC 26 July 20064DVAR4D-Var1800 UTC 26 July 2006ASDAASDA (proposed in this study)1800 UTC 26 July 20063DVAR_30min3D-Var1830 UTC 26 July 2006PMVPMV (Huang et al., 1997)1830 UTC 26 July 2006
Fig. 3
Observed 18-hours accumulated rainfall (mm 18h−1) distribution from 1800 UTC 26 to 1200 UTC 27 July 2006. Locations of Seoul, Hongcheon, and Osan are also indicated.
Fig. 4
Synoptic environments at 0000 UTC 27 July 2006. (a) Geopotential height (black solid, contour interval of 30 m), temperature (red dashed, contour interval of 3°C), water vapour mixing ratio (shaded, greater than 0.012 kg kg−1), and wind vector (knot) at 850 hPa, (b) geopotential height (black solid, contour interval of 30 m), wind speed (red solid, contour interval of 5 knots), and divergence (shaded, 10−5 s−1, only negative values are plotted) at 850 hPa, (c) geopotential height (black solid, contour interval of 60 m), temperature (red dashed, contour interval of 5°C), relative vorticity (shaded, 10−5 s−1, only positive values are plotted), and wind vector (knot) at 500 hPa, (d) geopotential height (black solid, contour interval of 120 m), wind speed (green solid, contour interval of 25 knots), and divergence (shaded, 10−5 s−1, only positive values are plotted) at 200 hPa.
Fig. 5
(a) Hodograph and (b) skew T-log p diagram of Osan at 0600 UTC 27 July 2006.
Fig. 6
The analysis increment responses at model level 12 (~600 hPa) to a single radial velocity observation located at 37.00°N, 127.58°E, 4000 m from the Pyeongtaek radar site (36.96°N, 127.02°E, 52 m). The single observation is assumed to be at 1930 UTC 26 July 2006, i.e. the assimilation window is from 1800 UTC to 1930 UTC 26 July 2006. (a) Zonal wind and (b) meridional wind for the 3DVAR experiment, (c) and (d) same as (a) and (b) except for the 4DVAR experiment, (e) and (f) same as (a) and (b) except for the ASDA experiment.
Fig. 7
Eighteen-hours accumulated rainfall (mm 18h−1) distributions from 1800 UTC 26 to 1200 UTC 27 July 2006 for the (a) CONTROL, (b) 3DVAR, (c) 4DVAR, (d) ASDA, (e) 3DVAR_30min, and (f) PMV experiments.
Fig. 8
Time series of hourly rainfall amount (mm h−1) at (a) Seoul and (b) Hongcheon (or the corresponding grid points in case of the model experiment) for the observations (black), CONTROL (blue), 3DVAR (green), 4DVAR (yellow), and ASDA (red) experiments. Maximum hourly rainfall amount is also indicated.
Fig. 9
Vertical distributions of fit to the observations using sounding observations over South Korea at 0000 UTC 27 July 2006 for the CONTROL (blue), 3DVAR (green), 4DVAR (yellow), and ASDA (red) experiments. (a) Zonal wind (m s−1), (b) meridional wind (m s−1), (c) temperature (K), and (d) water vapour mixing ratio (kg kg−1).
Fig. 10
RMSEs of radial velocity (m s−1) as a function of forecast length for the CONTROL (blue), 3DVAR (green), 4DVAR (yellow), and ASDA (red) experiments. Radar radial velocity data from 14 radar observation sites over the Korean Peninsula are used.
Fig. 11
Horizontal distributions of 850-hPa equivalent potential temperature (EPT) (contour interval of 4 K) at 1800 UTC 26 July 2006 for the (a) CONTROL, (b) 4DVAR, and (c) ASDA experiments. In case of data assimilation experiments, the analysis increments of EPT (shaded, K) are also shown.
Fig. 12
Horizontal distributions of 500-hPa absolute vorticity (shaded, 10−5 s−1) and vertical wind shear vector between 500 and 800 hPa (m s−1 hPa−1) at 0600 UTC 27 July 2006 for the (a) CONTROL, (b) 4DVAR, and (c) ASDA experiments.