Fog Simulation with the WRFPAFOG Coupled Model Utilizing

Fog Simulation with the (WRF+PAFOG) Coupled Model Utilizing Meteorological Tower Data Seong Soo Yum 1, Wonheung Kim 1, Jae-In Song 1 and Chang Ki Kim 2 1 Department of Atmospheric Sciences, Yonsei University, Korea 2 New and Renewable Energy Resource Center, Korea Institute of Energy Research, Korea

Introduction ü Prediction of fog using numerical models is difficult because fog formation is usually determined by local meteorological conditions that are hard to be measured and modeled with sufficient resolution. ü In this study we uses the coupled system of the 1 D PAFOG model and the 3 D WRF model to simulate fogs formed at a southern coastal region of Korea, where the National Center for Intensive Observation of Severe Weather (NCIO) is located. ü Here we demonstrate how the coupled model simulation of fog can be improved with the utilization of meteorological tower data.

The Models 3 D MODEL WRF V 3. 7 Horizontal Resolution 18, 6, 2 km Vertical Layers 65 (top~20 km) Initial field NCEP FNL data Radiation Process RRTMg scheme (both SW & LW) PBL Process MYNN scheme Surface physics Unified Noah land-surface model Microphysics Morrison 1 D MODEL (PAFOG, PArameterized FOG, Bott and Trautmann, 2002) Vertical Layers PAFOG L 400 (up to the model top, 2. 5 km) Dynamics Bott et al. (1990) turbulence Radiation Microphysics 2. 5 level model of Mellor and Yamada(1974) Δ-two straem approximation of Zdunkowski et al. (1982) Nickerson et al. (1986) and Chaumerloac et al. (1987) Vegetation Siebert et al. (1992)

Model Domain Coupling Method WRF_D 2 WRF_D 3 WRF_D 1 Selected Cases (2014. 05. ~2015. 06. ) Number of Fog Cases 22

Study Region Ø Location: Boseong, Korea (34. 76 o N, 127. 16 o E). Ø Instrumentation : 300 m Met. Tower (T, Td every 30 m) • Ka-band cloud radar • Microwave Radiometer • Ceilometer • Parsivel 2 300 met. tower

Average vertical profiles of Tz-T 0 and RH (Met Tower) Non-fog days Fog days

Nudging Method Time Weighting Tower (t 0 min) Tower (t 30 min) Height Weighting Tower (t 60 min)

Skill Scores 2014. 05. ~ 2015. 06. (22 cases) Observation counts Forecast Yes No Yes a b No c d HR CSI WRF 0. 05 ± 0. 12 0. 11 ± 0. 23 WRF+PAFOG 0. 27 ± 0. 30 0. 14 ± 0. 18 WRF+PAFOG (Initial _TOWER) 0. 60 ± 0. 48 0. 23 ± 0. 18 WRF+PAFOG (nudging_TOWER) 0. 81 ± 0. 21 0. 45 ± 0. 22

Skill Scores 2014. 05. ~ 2015. 06. (22 cases) count Foreca sts Observations Yes No Yes a b No c d WRF+PAFOG(nudging_TOWER) for radiation fog cases (13) HR CSI 0. 93 ± 0. 24 0. 65 ± 0. 16 HR WRF+PAFOG(nudging_TOWER) for cases WRF with prior precip. (9)0. 05 0. 68 ± 0. 24 CSI 0. 26 ± 0. 13 ± 0. 12 0. 11 ± 0. 23 WRF+PAFOG 0. 27 ± 0. 30 0. 14 ± 0. 18 WRF+PAFOG (Initial _TOWER) 0. 60 ± 0. 48 0. 23 ± 0. 18 WRF+PAFOG (nudging_TOWER) 0. 81 ± 0. 21 0. 45 ± 0. 22

Time Variation of Met. Variables A case with prior precipitation (2015. 1. 22) Precip. Soil moisture variation due to precipitation considered Precip.

Skill Scores 2014. 05. ~ 2015. 06. (22 cases) Observations count Yes No a b HR CSI No c d WRF+PAFOG(nudging_TOWER) for radiation fog cases (13) 0. 93 ± 0. 24 0. 65 ± 0. 16 WRF+PAFOG(nudging_TOWER) for cases with prior precip. (9) 0. 68 ± 0. 24 0. 26 ± 0. 13 Foreca sts HR WRF+PAFOG(nudging_TOWER) WRF for cases with prior precip. (9) (soil 0. 05 moisture variation considered) WRF+PAFOG 0. 27 WRF+PAFOG(Initial _TOWER) WRF+PAFOG(nudging_ TOWER) 0. 82 ± 0. 07 CSI 0. 50 ± 0. 08 ± 0. 12 0. 11 ± 0. 23 ± 0. 30 0. 60 ± 0. 48 0. 14 ± 0. 18 0. 23 ± 0. 18 0. 81 ± 0. 21 0. 45 ± 0. 22

Skill Scores 2014. 05. ~ 2015. 06. (22 cases) count Foreca sts Observations Yes No a b Yes HR CSI No c d WRF+PAFOG(nudging_TOWER) for radiation fog cases (13) 0. 93 ± 0. 24 0. 65 ± 0. 16 WRF+PAFOG(nudging_TOWER) for cases with prior precip. (9) 0. 68 ± 0. 24 0. 26 ± 0. 13 HR WRF+PAFOG(nudging_TOWER) WRF for cases with prior precip. (9) (soil 0. 05 moisture variation considered) WRF+PAFOG 0. 27 0. 82 ± 0. 07 CSI 0. 50 ± 0. 08 ± 0. 12 0. 11 ± 0. 23 ± 0. 30 0. 14 ± 0. 18 WRF+PAFOG (Initial _TOWER) 0. 60 ± 0. 48 0. 23 ± 0. 18 WRF+PAFOG (nudging_TOWER) 0. 81 ± 0. 21 0. 45 ± 0. 22 WRF+PAFOG (nudging_TOWER) 0. 88 ± 0. 18 0. 59 ± 0. 13

Skill Scores 2014. 05. ~ 2015. 06. (22 cases) Observations count Foreca sts Yes No a b Yes HR CSI 0. 93 ± 0. 24 0. 65 ± 0. 16 WRF+PAFOG(nudging_TOWER) 0. 68 ± 0. 24 for precipitation effect case HR (not considered) WRF+PAFOG(nudging_TOWER) 0. 05 ± 0. 82 0. 12± 0. 07 0. 26 ± 0. 13 No c d WRF+PAFOG(nudging_TOWER) for radiation fog case CSI But initial soil moisture values were ± 0. 08 0. 110. 50 ± 0. 23 for precipitation effect case fixed at minimum value for all cases! WRF+PAFOG (considered) 0. 27 ± 0. 30 0. 14 ± 0. 18 WRF+PAFOG (Initial _TOWER) 0. 60 ± 0. 48 0. 23 ± 0. 18 WRF+PAFOG (nudging_TOWER) 0. 81 ± 0. 21 0. 45 ± 0. 22 WRF+PAFOG (nudging_TOWER) 0. 88 ± 0. 18 0. 59 ± 0. 13

Time Variation of Met. Variables Observed Soil Moisture: 0. 394 m 3/m 3 0. 266 m 3/m 3 0. 394 m 3/m 3 Observed Soil Moisture: 0. 266 m 3/m 3 0. 394 m 3/m 3

Skill Scores 2014. 05. ~ 2015. 06. (22 cases) Observations count Foreca sts Yes No Yes a b No c d WRF+PAFOG (nudging_TOWER) Precipitation effect considered WRF+PAFOG (nudging_TOWER) Initial soil moisture varied HR CSI 0. 88 ± 0. 18 0. 59 ± 0. 13 HR 0. 89 ± 0. 07 CSI 0. 64 ± 0. 13

Skill Scores 2014. 05. ~ 2015. 06. (22 cases) Observations count Foreca sts Yes No Yes a b No c d WRF+PAFOG HR CSI 0. 27 ± 0. 30 0. 14 ± 0. 18 Nudging effect WRF+PAFOG (nudging_TOWER) HR 0. 81 ± 0. 21 CSI 0. 45 ± 0. 22 VS Soil moisture effect WRF+PAFOG (Soil moisture variation) HR 0. 53 ± 0. 35 CSI 0. 27 ± 0. 19

Comparison of WRF and UM Coupled Simulations 2012. 06. ~ 2013. 07. (22 sea fog cases over the sea near IIA) UM_RE+PAFOG UM_LO+PAFOG WRF_CO+PAFOG WRF_NE+PAFOG MBE RMSE T -0. 32 ± 0. 91 1. 70 ± 1. 12 -0. 75 ± 0. 77 1. 85 ± 1. 16 3. 20 ± 4. 73 1. 98 ± 1. 14 3. 38 ± 4. 75 1. 93 ± 1. 38 Td 0. 51 ± 0. 61 1. 59 ± 1. 12 0. 04 ± 0. 74 1. 44 ± 0. 69 1. 73 ± 2. 93 1. 96 ± 1. 12 2. 28 ± 3. 06 2. 03 ± 1. 35 rw CSI CSI 0. 37 ± 0. 34 0. 22 ± 0. 31 0. 23 ± 0. 32 0. 17 ± 0. 27

Summary ü WRF+PAFOG coupled system was set up to use for fog simulation over the southern coastal region of Korea. ü With the utilization of meteorological tower data, performance of the coupled model improved significantly. ü For these coastal fog cases, soil moisture information was found to be critically important for better simulation. ü Next step is to elucidate fog formation mechanism based on the model simulation results. ü With proper adjustment, the coupled model system can be used as a fog prediction model during the Pyeong. Chang Olympics.