Author Contributions
J.-W.L. wrote the first draft, assisted with data curation, and performed formal analysis, investigation, software coding, and visualization of the research article. K.-H.M. conceptualized the paper, supervised and administered the project, performed formal analysis, provided resources and funding, and reviewed and edited the paper. Y.-H.L. and G.L. supported with funding, provided methods, and reviewed and edited the paper. All authors have read and agreed to the published version of the manuscript.
Figure 1.
Locations of the S-band radars (blue dots), X-band radars (red dots), automated weather station (AWS) sites (gray dots), and radiosondes (yellow dots), with the radar coverage areas in circles.
Figure 1.
Locations of the S-band radars (blue dots), X-band radars (red dots), automated weather station (AWS) sites (gray dots), and radiosondes (yellow dots), with the radar coverage areas in circles.
Figure 2.
Flow chart for assimilating radar reflectivity with null-echo observation operators.
Figure 2.
Flow chart for assimilating radar reflectivity with null-echo observation operators.
Figure 3.
(a) Domain configuration and topography (shaded) of D01, D02, and D03; (b) the border lines of city and provinces used in D03 (the borders are shown as a red line for Seoul, an orange line for Gyeonggi, and a blue line for Hwanghae).
Figure 3.
(a) Domain configuration and topography (shaded) of D01, D02, and D03; (b) the border lines of city and provinces used in D03 (the borders are shown as a red line for Seoul, an orange line for Gyeonggi, and a blue line for Hwanghae).
Figure 4.
Synoptic analysis for (a) 0000 UTC 2 July 2017, (b) 1200 UTC 22 July 2017, and (c) 0000 UTC 28 August 2018.
Figure 4.
Synoptic analysis for (a) 0000 UTC 2 July 2017, (b) 1200 UTC 22 July 2017, and (c) 0000 UTC 28 August 2018.
Figure 5.
Comparison of (a) the radar reflectivity (dBZ), reflectivity of (b) CTRL, (c) DA, and (d) DA_NP and the difference in qv between (e) DA and CTRL and between (f) DA_NP and CTRL at 0900 UTC 2 July 2017.
Figure 5.
Comparison of (a) the radar reflectivity (dBZ), reflectivity of (b) CTRL, (c) DA, and (d) DA_NP and the difference in qv between (e) DA and CTRL and between (f) DA_NP and CTRL at 0900 UTC 2 July 2017.
Figure 6.
Same as
Figure 5 except for 2100 UTC 22 July 2017.
Figure 6.
Same as
Figure 5 except for 2100 UTC 22 July 2017.
Figure 7.
Same as
Figure 5 except for 0900 UTC 28 August 2018.
Figure 7.
Same as
Figure 5 except for 0900 UTC 28 August 2018.
Figure 8.
Cumulative precipitation (mm) distribution for Case 1 from the (a) AWSs, (b) CTRL, (c) DA, and (d) DA_NP at D03.
Figure 8.
Cumulative precipitation (mm) distribution for Case 1 from the (a) AWSs, (b) CTRL, (c) DA, and (d) DA_NP at D03.
Figure 9.
Same as
Figure 8 except for Case 2.
Figure 9.
Same as
Figure 8 except for Case 2.
Figure 10.
Same as
Figure 8 except for Case 3.
Figure 10.
Same as
Figure 8 except for Case 3.
Figure 11.
Verification statistics of the (a) accuracy, (b) critical success index (CSI), and (c) equitable threat score (ETS) for CTRL (black lines), DA (red lines), and DA_NP (blue lines).
Figure 11.
Verification statistics of the (a) accuracy, (b) critical success index (CSI), and (c) equitable threat score (ETS) for CTRL (black lines), DA (red lines), and DA_NP (blue lines).
Figure 12.
Vertical profiles biases of (a) water vapor mixing ratio, (b) temperature, (c) U wind component, and (d) V wind component for CTRL (black lines), DA (red lines), and DA_NP (blue lines) at 2017.07.02.1200 UTC (solid lines) and 2018.08.28.1200 UTC (dashed lines).
Figure 12.
Vertical profiles biases of (a) water vapor mixing ratio, (b) temperature, (c) U wind component, and (d) V wind component for CTRL (black lines), DA (red lines), and DA_NP (blue lines) at 2017.07.02.1200 UTC (solid lines) and 2018.08.28.1200 UTC (dashed lines).
Figure 13.
Contoured frequency by altitude diagram (CFAD) percentiles of the (a) Kwanak Mountain Mountain (KWK) observed and (b) CTRL, (c) DA, and (d) DA_NP simulated radar reflectivity for Case 2. The horizontal black dotted line in each panel represents the model’s 0 °C height (b, c, and d). The cumulative reflectivity frequencies of the 25th, 50th, and 75th percentiles are marked with black solid lines, the average reflectivity factor in linear units (mm−6·m−3) is marked by a white solid line, and the average reflectivity is marked by a red solid line (KWK) and blue solid line (models).
Figure 13.
Contoured frequency by altitude diagram (CFAD) percentiles of the (a) Kwanak Mountain Mountain (KWK) observed and (b) CTRL, (c) DA, and (d) DA_NP simulated radar reflectivity for Case 2. The horizontal black dotted line in each panel represents the model’s 0 °C height (b, c, and d). The cumulative reflectivity frequencies of the 25th, 50th, and 75th percentiles are marked with black solid lines, the average reflectivity factor in linear units (mm−6·m−3) is marked by a white solid line, and the average reflectivity is marked by a red solid line (KWK) and blue solid line (models).
Figure 14.
Same as
Figure 13 except for Case 3 forecast period.
Figure 14.
Same as
Figure 13 except for Case 3 forecast period.
Figure 15.
Time–height cross-sections for Case 2 at the KWK radar site for (a) the observations, (b) CTRL, (c) DA, and (d) DA_NP.
Figure 15.
Time–height cross-sections for Case 2 at the KWK radar site for (a) the observations, (b) CTRL, (c) DA, and (d) DA_NP.
Figure 16.
Same as
Figure 15 except for Case 3 forecast period.
Figure 16.
Same as
Figure 15 except for Case 3 forecast period.
Table 1.
Summary of observations used in this study.
Table 1.
Summary of observations used in this study.
Purpose | Observations | Variables | Horizontal Resolution | Vertical Resolution | Time Period | Time Resolution |
---|
Data assimilation | Radar | Reflectivity (dBZ), Doppler radial velocity (m·s−1) | 1, 3 km * | 200 m | 3 h | 30 min |
AWS | Surface pressure (hPa), sea level pressure (hPa), 10-min average wind speed (m·s−1) and direction (°), temperature and dew point temperature (K) | - |
Verification | Precipitation (mm) | 1 km * | - | 9 h | 1 h |
Radiosonde | Water vapor mixing ratio (g·kg−1), temperature (K), wind speed (m·s−1) and direction (°) | Point | 50 hPa | 0000 or 1200 UTC * |
Table 2.
Summary of the Weather Research and Forecasting (WRF) v3.9.1 model configurations.
Table 2.
Summary of the Weather Research and Forecasting (WRF) v3.9.1 model configurations.
| D01 | D02 | D03 |
---|
Resolution | 9 km | 3 km | 1 km |
Number of Grids | 240 × 240 × 60 | 151 × 151 × 60 | 226 × 196 × 60 |
Cumulus | Multiscale Kain–Fritsch scheme |
Microphysics | WRF Double Moment 6 class scheme |
Planetary Boundary Layer | Yonsei University Scheme |
Surface Layer | Revised MM5 Monin–Obukhov scheme |
Land Surface | Unified Noah land surface model |
Long-Wave Radiation | Rapid radiative transfer model scheme |
Short-Wave Radiation | Dudhia scheme |
Initial and Boundary Conditions | National Center for Environmental Prediction Global Forecast System 0.25 Degree Historical Archive |
Table 3.
Experimental design.
Table 3.
Experimental design.
| Assimilated Observation Data |
---|
CTRL | No data assimilation |
DA | AWS + radar radial velocity + radar reflectivity |
DA_NP | AWS + radar radial velocity + radar reflectivity + null echoes |
Table 4.
Selected cases for the numerical experiments and their storm characteristics.
Table 4.
Selected cases for the numerical experiments and their storm characteristics.
| Forecast Period | Total Cumulative Precipitation (mm) | Maximum Rain Rate (mm·h−1) | Maximum CAPE * (J·kg−1) |
---|
CTRL | DA and DA_NP |
---|
Case 1 | 2017.07.02.0600 UTC ~ 1800 UTC | 2017.07.02.0900 UTC ~ 1800 UTC | 55 | 55.0 | 1233 |
Case 2 | 2017.07.22.1800 UTC ~ 2017.07.23.0600 UTC | 2017.07.22.2100 UTC ~ 2017.07.23.0600 UTC | 110 | 67.0 | 1521 |
Case 3 | 2018.08.28.0600 UTC ~ 1800 UTC | 2018.08.28.0900 UTC ~ 1800 UTC | 180 | 73.0 | 1468 |
Table 5.
Contingency table matrix (2 × 2) for skill score calculations.
Table 5.
Contingency table matrix (2 × 2) for skill score calculations.
| Observation | Total |
---|
Yes | No |
---|
Forecast | Yes | Hits = A | False alarms = B | Forecast Yes |
No | Misses = C | Correct negative = D | Forecast No |
Total | Observed Yes | Observed No | Total = N |
Table 6.
Cumulative precipitation error for CTRL, DA, and DA_NP.
Table 6.
Cumulative precipitation error for CTRL, DA, and DA_NP.
| Experiment | Bias (mm) | MAE (mm) | RMSE (mm) |
---|
Case 1 | CTRL | −15.3 | 19.2 | 28.6 |
DA | −9.1 | 20.0 | 30.0 |
DA_NP | −8.8 | 18.2 | 26.7 |
Case 2 | CTRL | −31.4 | 34.0 | 47.3 |
DA | 12.1 | 28.8 | 40.0 |
DA_NP | 2.3 | 21.3 | 31.4 |
Case 3 | CTRL | −48.5 | 48.7 | 62.7 |
DA | 17.5 | 31.2 | 43.6 |
DA_NP | 2.8 | 23.3 | 35.4 |
Average | CTRL | −31.7 | 34.0 | 46.2 |
DA | 6.9 | 26.7 | 37.9 |
DA_NP | −1.2 | 20.9 | 31.2 |