9

16 7 Meteorological Disturbances and Precipitation during the Niigata-Fukushima and the Fukui Heavy Rainfall in July 2004 Ryohei MISUMI Advanced Technology Research Group, National Research Institute for Earth Science and Disaster Prevention, Japan misumi@bosai.go.jp Abstract Meteorological characteristics of the Niigata-Fukushima heavy rainfall on 13 July and the Fukui heavy rainfall on 18 July 2004 are studied using the outputs of the Regional Spectral Model (RSM) of the Japan Meteorological Agency (JMA), the upper soundings, the 10-minute data of the Automated Meteorological Data Acquisition System (AMeDAS), the Radar-AMeDAS rainfall, the radar echoes derived by the JMA radars, hourly observations at meteorological observatories, and the data of the vertically looking Micro Rain Radar and the Snow-Rain Intensity Meter at the Nagaoka Institute of Snow and Ice Studies. During the Niigata-Fukushima heavy rainfall, a meso-α-scale depression is found just north of the heavy rainfall area, and the northeasterly associated with the depression and the southwesterly from the East China Sea strongly converge over the sea. Another convergence zone is found along the coast line in the south part of Niigata Prefecture, which is formed at the leading edge of the northwesterly from the meso-α-scale depression. Fall velocity of raindrops measured with the Micro Rain Radar indicates that large raindrops are produced below the 1.5 km level. These results suggest that the convective clouds formed in the convergence zone over the sea move toward the other convergence zone along the coast line, and the raindrops strongly grow in the lower layer of the latter convergence zone. For the Fukui heavy rainfall, a passage of a meso-β-scale depression is suggested by the surface pressure suppression by 1 hpa during the rainfall event, although no mesoscale disturbance is found in the output of the RSM. The air at the middle level of the troposphere is relatively dry in this case and a cold-air pool is produced at the surface of the rainfall area. Westerly from the Sea of Japan converges at the edge of the cold-air pool and radar echoes is broadened and persist longer periods over the land area than over the sea. It is suggested that the convective rainfall associated with the meso-β-scale depression is enhanced in the convergence zone along the edge of the cold-air pool. Because mesoscale depressions are considered as the key factors of the heavy rainfall events in both cases, some monitoring systems for such disturbances over the Sea of Japan are required to predict the disasters. Key words : Rainfall, Disaster, Landslides, Floods, Mesoscale depression 1. 16 7 12 13 13 421mm 2004 4 11 300 16 2004a 5 7 18 9

283mm 1 96mm 2004 3 69 JR 6 4 1 19 2004b 16 7 16 7 2 km km 1,000km 16 7 2004a, b 2 2004 Kato and Aranami 2005 10 2 2 2. 2.1 421mm 10 1 10 Fig. 1 Time variation of 10 minute rainfall at Tochio, Niigata Prefecture. 1 Table 1 List of people killed by the Niigata-Fukushima heavy rainfall in July 2004 (based on the articles of Niigata Nippou and Fukushima Minyu newspapers). 1 7 12 21:00 7 13 6 4mm 24mm/h 13 7 10mm 60mm/h 13 7 15 8 312.5mm 40mm/h 8 17:50 9.5mm 57mm/h 1 75 70 2 14 2 13 14 1 10

16 7 2 7 13 mm Fig. 2 Daily precipitation analyzed with the Radar-AMeDAS on 13 July 2004 (units are mm) and the location of the cities and the towns where people were killed. 13 70m 2004a 12 52 2004b 7 13 18 15 7 13 2 100mm 40km 150km 421mm 476mm 200mm 9 208mm 2.2 3 7 12 21 13 9 21 7 12 21 500km H 13 9 13 21 4 7 13 9 850hPa 500hPa 850hPa L 500hPa 500hPa 850hPa 5g/kg 5-6.5 /km 7 13 3 3 28 21 6 1 2 3 13 9 12 15 11

3 Fig. 3 7 12 21 13 9 21 RSM Distribution of pressure (solid lines), temperature (broken lines) and wind vectors at the surface at 21 JST on 12 July and 09 JST and 21 JST on 13 July 2004. The data are the initial values for the JMA Regional Spectral Model (RSM). 4 Fig. 4 7 13 9 850hPa 500hPa m g/kg RSM Geopotential height (solid lines; m), mixing ratio of water vapor (shadings; g/kg) and wind vectors at the 850 hpa and the 500 hpa levels at 09 JST on 13 July 2004. The data are the initial values for the JMA Regional Spectral Model. 12

16 7 5 Fig. 5-6.5 /km Distribution of surface temperature (contours) and wind (vectors) derived by the AMeDAS. Data at observatories are interpolated with the bilinear method. Elevation corrections for temperature are made with -6.5 /km. Thick broken lines are the convergence lines drawn subjectively. 13

6 7 13 9 e e Fig. 6 Profiles of the potential temperature ( ), equivalent potential temperature ( e) and the equivalent potential temperature of a hypothetically saturated atmosphere ( e ). 18 6 7 13 9 e e 1978 e 6,200m e 5,700m 5,700m -2.1K 2.3 10 7 3 7 13 3 A 6 A B 9 A B 12 B C C 15 B B 18 8 138 7 P1-P2 139 7 Q1-Q2 138 A B C 11 3 139 138 B 6 10 B 139 138 2.4 SR-2A 9 1 0.005mm 10 7 12 13 1 0.005mm 1 9:12 1.84mm =110.4 mm/h 1 11 2004 250 4 1 120 4 10b 3 Matsumoto and Akiyama 1969 METEK MRR 24GHz 12 characteristic falling velocity 1 14

16 7 7 7 13 3 18 Fig. 7 Rainfall intensity estimated with the JMA Radar composites from 03 JST to 18 JST on 13 July 2004. 15

8 Fig. 8 7 13 138 7 P1-P2 139 7 Q1-Q2 Latitude-time sections of rainfall intensity estimated with the JMA radar along 138 E (P1-P2 in Fig. 7) and 139 E (Q1-Q2 in Fig. 7) on 13 July 2004. 9 Fig. 9 Snow-rain intensity meter made by Tamura Snow Measurement Laboratory. (f) f 12a 1.5km 40dBZ 12b 5 2km 6 1.5km 1.5km W 6m/s 12c 6 6 6m/s 1.8mm 6 1.5km 1 1 13 0.80 13 222.1mm 164.3mm 223mm 25 W 14 3 10 CONV 2 x 0, y 0 x 1, y 1 x 2, y 2 u 0, v 0 u 1, v 1 u 2, v 2 u, v 1 2004 3 16

16 7 10 Fig. 10 a 1 b 120 1-minute rainfall at Nagaoka Institute of Snow and Ice Studies (Suyoshi, Nagaoka) measured with the Snow- Rain Intensity Meter and (b) its 120-minute moving average. 11 Fig. 11 7 13 2:57-20:00 1,024 S PER Relationship between the periodogram (S PER) and the period calculated for the 1,024 data of 1-minute rainfall observed from 2:57 to 20:00 on 13 July 2004. 17

12 Fig. 12 a dbz b m/s c 6 m/s Time height sections of (a) radar reflectivity (dbz), (b) characteristic falling velocity (m/s) and (c) the areas where the characteristic falling velocity is greater than 6 m/s. 18

16 7 13 1 Fig. 13 Relationship between rainfall intensity every 1 minute derived by the Micro Rain Radar and the Snow-Rain Intensity Meter. 14 Fig. 14 Locations of Teradomari, Kashiwazaki and Irihirose where are used for the calculation of surface wind convergence. Location of Nagaoka is also shown. 15 Fig. 15 Time variation of the horizontal convergence within the triangle area surrounded by Teradomari, Kashiwazaki and Irihirose. 4 5 6 7 8 2 15 6 6 15 1 10-4 s -1 1987 km 10-4 s -1 40km 6 5 1.5km 12c 19

2.5 300mm 2 2004 9 1 3 RIEGL FG21-HA 2m 16 2 13 25 45 11.9m 112.1m 2m 2 5 10 12 3m 9m 1, 3, 4, 7, 8, 9, 11, 13 9m 6 1 7 8 2 cm 2m 3. 3.1 2004 7 18 18 0:00 5:20 5:40 22.5mm 135mm/h 17 9 7:10-8:00 1 75mm 16 Fig. 16 1 Surface of the failure slope in Photo 1 measured with the laser distance meter and estimation methods for slope length, angle and depth. 2 Table 2 Angle, length, width at the foot, and depth of failure slopes. 20

16 7 3 Table 3 List of people killed or missing by the 2004 Fukui heavy rainfall (based on the articles of Fukui newspaper). 1 2 6 Photo 1 Landslide at Tsuchigaya, Tochio City ( No.6 in Table 2). 2 Photo 2 Shallow landslide (left) and deeper landslide (right) at Tsuchigaya, Tochio City. 17 7 18 10 Fig. 17 10-minute rainfall at AMeDAS Miyama (Asayajima, Miyama town, Asuwa-gun) on 18 July 2004. 5:20-6:10 1 95.5mm 283mm 9 4 1 3 60 9:00 18m 2004 7 18 18 250mm 283mm 365mm 338mm 2004b 3.2 19 7 17 21 18 21 H 21

18 7 18 mm Fig. 18 Daily precipitation analyzed with the Radar-AMeDAS on 18 July 2004 (units are mm) and the location of the cities or the towns where people were killed or missing. 7 13 3 17 21 18 9 850hPa 20 13g/kg 500hPa 500hPa 18 21 19 21-6.5 /km 18 1 W 3 5 C 20 500hPa 7 C 9 4 22 7 18 9 e e 6 e e 1,000m e 1km 1,000m 3,200m e e 5.5 4 850hPa 20 3.3 23 7 18 1 D 20mm/h 1 34mm 3 D D E E 2 5 D 100km F F 5 7 6 F 11 24 136 23 R1- R2 136.5 23 S1-S2 D E F 22

16 7 19 7 17 21 18 9 21 RSM Fig. 19 Distribution of pressure (solid lines), temperature (broken lines) and wind vectors at the surface at 21 JST on 17 July and 09 JST and 21 JST on 18 July 2004. The data are the initial values for the JMA Regional Spectral Model (RSM). 20 7 18 9 850hPa 500hPa m g/kg RSM Fig. 20 Geopotential height (solid lines; m), mixing ratio of water vapor (shadings; g/kg) and wind vectors at the 850 hpa and the 500 hpa levels at 09 JST on 18 July 2004. The data are the initial values for the JMA Regional Spectral Model. 23

21 Fig. 21-6.5 /km Distribution of surface temperature (contours) and wind (vectors) of AMeDAS. Data at observatories are interpolated with the bilinear method. Elevation correction for temperature is made with -6.5 /km. Thick broken lines are the convergent lines drawn subjectively. 24

16 7 22 7 13 9 e e Fig. 22 Profiles of the potential temperature ( ), equivalent potential temperature ( e) and equivalent potential temperature of a hypothetically saturated atmosphere ( e ). D 136 0:00 2:30 136.5 0:30 4:00 E F E 136 136.5 3:00 7:00 4 F 136 4:00 11:00 F 23 136.5 4:30 35.8 N 18:00 3.4 25 26 3 10 2 8 7 17 18:00-23:00 5m/s 18 0:00 6:00 12:00 7:50 7m/s 13:00 7 17 12:30 18 0:00 1 2:00 2 4:00 5:00 3 9:00 17 12:00 18 0:00 0:00 2:00-1.3hPa 4 6:00-7:00 20km 200km 18 1:00 3:00 10-4 s -1 10-4 s -1 1987 3:00-7:00 8:00 21 3.5 300mm 18 16 9 5 2.5 4 1 3 1 31.5m 2 3 27 L 4 2 P1 3 4. 4.1 16 7 13 1 2 km 25

23 7 18 1 11 Fig. 23 Rainfall intensity estimated with the JMA Radar composites from 01 JST to 11 JST on 18 July 2004. 26

16 7 24 7 18 136 136.5 Fig. 24 Latitude-time sections of rainfall intensity estimated with the JMA radar along 136 E and 136.5 E on 18 July 2004. 25 Fig. 25 10 1 10 Time variation of wind, temperature, precipitation (every 10 minute), pressure (every 1 hour) and convergence of surface wind (every 10 minute) at Fukui Meteorological Observatory. 27

4 Table 4 Angle, length, width at the foot and depth of failure slopes caused by the Fukui heavy rainfall. 26 Fig. 26 Locations of Mikuni, Koshino and Ohno where are used for the calculation for convergence of surface wind. Locations of Fukui and Miyama are also shown. 3 4 2 Photo 3 Landslide at Ainoki, Imadate Town (at No.2 in Table 4). 3 6200m 4 5 km 3 6 4 10-4 s -1 6m/s 1.5km 27 Fig. 27 A rough sketch of the landslide at Ainoki, Imadate Town. 28 2 1 1 28

16 7 28 29 Fig. 28 Schematic illustration of the Niigata-Fukushima Fig. 29 Schematic illustration of the Fukui heavy heavy rainfall. rainfall. 2 1.5km 4 1 Matsumoto and Akiyama 1969 3 4.2 16 7 18 1 2 850hPa 3 4 5 500hPa 6 7 1hPa 4 8 10-4 s -1 29 7 29

4.3 16 7 160mm 13 5:00 120mm 2004b 24 120mm 150mm Kato and Aranami 2005 1.5km 200mm 1 Jiang Furen 1 2004 16 7 16 7 30 2 2004 16 7 19 3 2004 16 9 7 4 Kato, T. and Aranami, K. 2005 Formation factors of 2004 Niigata-Fukushima and Fukui heavy rainfalls and problems in the predictions using a cloud-resolving model SOLA, 1, 001-004. 5 Matsumoto, S. and Akiyama, T. 1969 Some characteristic features of the heavy rainfalls observed over the western Japan on July 9, 1967, Part I. Mesoscale structure and short period pulsation. J. Meteorol. Soc. Japan, 47-4, 255-267. 6 1987 192pp 7 2004a 16 7 48 16 9 10 8 2004b 16 7 34 16 8 27 9 2004a 16 7 14 10 2004b 16 7 15 11 2004 16 7 16 7 30

16 7 23 12 2004 2003 51-1, 31-34. 13 1978 249pp 14 2004 16 7 2004 A103 15 2004a 2004 7 12 13 23-2 293-302 16 2004b 2004 7 18 23-3 443-452 17 2004 21-103 2005 10 20 31

10 1 16 7 1 1.5km 1 1hPa 4 32