Study(2013)

Vortex structure and development process of parent storm of a tornado occurred in Inabe City, Japan

A tornado struck Inabe City, central Japan, at about 15 JST on September 18, 2012. This event is referred to as "Inabe tornado" in this study. The parent storm of the Inabe tornado developed in the humid environment in which tornadoes occur occasionally in Japan. The purpose of this study is to reveal characteristics of a tornado and its parent storm in the humid environment, by a case study of the Inabe tornado. In particular, this study focuses on the vortex structure and development process of the parent storm using three X-band polarimetric radars that covers Inabe city.

Vertical vortices are detected by a method proposed by Donaldson (1970) and Suzuki et al. (2000) to apply the plan position indicator (PPI) data of each elevation angle of all radars. Vertical velocity is estimated by dual-Doppler analysis on a Doppler velocity field obtained by all radars.

Since a mesocyclone whose diameter of approximately 4 km is identified within the parent storm, the Inabe tornado is recognized to be associated with a supercell type storm. A hook echo, weak-echo region under an echo overhang, which are typically observed in the supercell storm over the Great Plains are detected in the parent storm. The updraft velocity in a core of the parent storm is approximately 14 m s-1 at a height of 2 km (Fig. 1), and is weaker than typical supercells over the Great Plains. Cold outflow produced by a downdraft in the core with wind speed of 10 m s-1 at a height of 1 km is balanced with the warm and moist southeasterly inflow. A smaller scale vortex referred to as “misocyclone” is detected within the mesocyclone. A misocyclone is observed adjacent to the mesocyclone a few minutes before the occurrence of the Inabe tornado. It is noteworthy that a pair of positive and negative vortices with the diameter about 1-3 km is detected in a rear-flank of the parent storm before the appearance of the misocyclone.

Figure 2 shows schematic images of the structure of the parent storm of the Inabe tornado. A vortex pair is formed by the tilting of the horizontally elongated vortex tube that has horizontal vorticity by an updraft associated with a rear-flank downdraft and cold outflow. The cyclonic vortex of the vortex pair moves into the mesocyclone just before the deep development of the misocyclone. As the location of the misocyclone in the parent storm corresponds to the damage track of the Inabe tornado, we propose that the formation of the Inabe tornado is attributed to the regional intensification of an updraft by the misocyclone.

fig01

Fig. 1 Vertical cross section of reflectivity (shading), vertical velocity (contour), and storm-relative wind velocity (vectors) on the plane along an inflow through the core of the updraft at 1505 JST. The thick and dashed contours indicate updraft and the downdraft every 2 m s-1 starting from 1 m s-1, respectively.

fig02

Fig. 2 Schematic images of the structure of the parent storm of the Inabe tornado at the stage of (a) the formation of the vortex pair and (b) the development of the misocyclone. The thick arrows indicate updraft and downdraft. The thin arrows indicate the storm-relative wind. The circular arrows indicate the vortex detected by radars. The thick line in (b) indicates the damage track of the Inabe tornado.

Distribution of solid particles in precipitation systems during the Baiu season in Okinawa

To clarify precipitation mechanisms in mesoscale precipitation systems during the Baiu season, distributions of solid particles, such as ice crystals, snowflakes, and graupels, above melting level should be investigated. In this study, we reveal vertical distribution of hydrometeors using data obtained by an X-band polarimetric radar installed at Aguni Island, Okinawa Prefecture. We use total 1392 data of range height indicator (RHI) obtained in 36 days from May 23, 2011 and in 41 days from May 7, 2012.

First we exclude the case that echo-top height is lower than the melting level (approximately at a height of 5 km), and 548 cases still remain. Precipitation systems are divided into 3 types. "A stratiform precipitation system" is defined by the criteria: Echo-top height of 30 dBZ does not exceed above the melting level and a significant bright band is detected. "An isolated convection" is defined by that: Echo-top height of 30 dBZ exceeds above the melting level and no significant bright band is detected. "Convective cells embedded in a stratiform precipitation system" is defined by that: Echo-top height of 30 dBZ exceeds above the melting level and a significant bright band is detected. Figure 3 shows a sample of the three types. According to the definition, 465, 23, and 60 cases are classified into a stratiform precipitation system, an isolated convection, and convective cells embedded in a stratiform precipitation system, respectively. Thus precipitation during the Baiu season in Okinawa is expected to be mainly brought about by a staratiform precipitation system.

A hydrometeor classification analysis is applied to all of the RHI data and vertical distribution of frequency of each solid particle; ice crystal, dry snow, and dry graupel, is analyzed for each type of the precipitation system (Fig. 4). Frequency of ice crystal and dry snow is dominant through all types of the precipitation system above the melting level. Dry graupel is detected very few in a stratiform precipitation system. On the other hand, dry graupel is detected below 11 km and its frequency exceeds 35% at a height of 5.5 km (around the melting level) in an isolated convection. Since frequency of an isolated convection and convective cells embedded in a stratiform precipitation system is quite few in comparison with that of a stratiform precipitation system, the contribution of graupel to produce a heavy rainfall event is expected to be small during the Baiu season in Okinawa. As a result, the coalescence process of raindrops should be mainly attributed to produce a heavy rainfall in the precipitation mechanism.

fig03

Fig. 3 A sample of RHI images of reflectivity for 3 types of the precipitation system: (a) a stratiform precipitation system, (b) an isolated convection, and (c) convective cells embedded in a stratiform precipitation system.

fig04

Fig. 4 Vertical distributions of frequency of solid particles; ice crystal (gray broken line), dry snow (solid line), and dry graupel (dotted line) above a height of 5.5 km (around the melting level) for 3 types of the precipitation system: (a) a stratiform precipitation system, (b) an isolated convection, and (c) convective cells embedded in a stratiform precipitation system.

Structure of a mesoscale convective system observed in PALAU2013: A preliminary result

To clarify structure of mecoscale convective systems (MCSs) over the tropical ocean, we conducted an intensive field experiment in Republic of Palau in June 2013 (PALAU2013). Direct observations of ice particles using hydrometeor videosondes (HYVISs) were conducted within the observation ranges of a polarimetric radar and a Doppler radar. This study shows an overview of the field experiment and preliminary result in PALAU2013.

An X-band polarimetric Doppler radar whose observation range is 60 km was installed at the northern tip of Palau, the Ngarchelong site (Fig. 5). Another X-band Doppler radar whose observation range is 150 km has already been installed at the Aimeliik site since 2004. Both radars made a continuous volume scanning observation at every 7.5 min. HYVIS observations were also conducted at the Aimeliik site. Totally 19 HYVISs were launched in PALAU2013. HYVIS observations identify vertical profiles of ice particle's properties (type, size, and number concentration) as well as the height, air temperature, and humidity.

Passage of MCSs is detected by a large echo area (Fig. 6). Three MCSs that develop to typhoon T1303, T1304, and T1306 after the passage of the observation region are observed on June 06 (DOY=157), 15 (DOY=166), and 26 (DOY=177), respectively. A MCS accompanied leading convective and trailing stratiform regions with meridional length of 200 km and zonal width of 150 km propagated westward over the observation range of the radars on June 15. Four HYVISs were intermittently launched into the MCS. Figure 7 shows particle images at a height of 6.1 km of the first HYVIS launched just behind the convective region. Ice particles of column and plate types and a frozen drop as well as small particles of supercooled liquid water are found in this image.

We will make a dual-Doppler analysis to detect a vortex structure in a convection and a matching analysis on hydrometeor classification obtained by a polarimetric radar and HYVIS observations in near future.

fig05

Fig. 5 The location of radar sites and their observation ranges in Republic of Palau during PALAU2013.

fig06

Fig. 6 Time series of echo-top height (gray bar graph) and echo area at a height of 2 km (solid line) from May 28 (DOY = 148) to July 1 (DOY = 182), 2013. Both echo-top height and echo area are defined by 15 dBZ.

fig07

Fig. 7 A particle image obtained by a HYVIS at a height of 6.1 km (temperature -5.2 Celsius degree) launched at 1445 LST from the Aimeliik Site, Republic of Palau. Horizontal and vertical scales of the image are 1.2 mm and 0.9 mm, respectively.

Evaluation of environmental modulators influencing the intensity change of a tropical cyclone using a coupled atmosphere-ocean non-hydrostatic model

Environmental modulators such as ocean heat content (OHC) and vertical wind shear (VWS) influence on the intensity change of tropical cyclones (TCs). Park et al. (2012; Tropical Cyclone Research and Review) showed that OHC and VWS influence the intensity change of TCs in terms of a two-dimensional phase diagram (hereafter PEH diagram) using reanalysis data of the atmosphere and ocean. However, they were not able to evaluate the relationship between them in comparison with many previous studies. In the present study, we try to evaluate the relationship between OHC and VWS for T1013 (Megi) using a coupled atmosphere-ocean non-hydrostatic model, Cloud Resolving Storm Simulator (CReSS) and Non-Hydrostatic Ocean model for Earth Simulator (NHOES), CReSS-NHOES with horizontal grid resolution of 0.02 degree both for latitude and longitude (approximately 2 km). The simulation is conducted for 7 days from 00 UTC on October 14, 2010, after one day of the formation of T1013.

The simulation well reproduces its track (not shown) and the tendency of central pressure (Fig. 8). The simulated minimum central pressure (889 hPa) corresponds to the observed one (885 hPa).

We define the inner core region of the TC as the maximum wind speed in the vertical column exceeds 25 m s-1, thus the size of the inner core region is flexible. The atmospheric environment in which VWS between 850 and 200 hPa is estimated, is defined as the outer region in the range of 2 degrees of the inner core region. The oceanic environment in which OHC is calculated by heat content above the depth of 26 degree Celsius, is defined as the ocean below the inner core region. Figure 9 shows the PEH diagram obtained by the CReSS-NHOES simulation. In this diagram, the lower right (upper left) area shows high (low) OHC and low (high) VWS; that is favorite (not favorite) on the development of a TC. When T1013 intensifies (initial development period between Oct. 14 and 16, and rapid intensification period between 12 Z on Oct. 17 and 00 Z on Oct. 18), VWS is lesser than 6 m s-1 and OHC is larger than 90 kJ cm-2. These environmental modulators are suitable for the intensification of TCs shown in previous studies. On the other hand, when T1013 maintains its intensity between 00 Z on Oct. 16 and 12 Z on Oct. 17, VWS increases above 7.5 m s-1 and OHC decreases below 90 kJ cm-2. Using the PEH diagram, we reproduce well the relationship between these modulators and the intensification of T1013. In comparison with Park et al. (2012), the impact of the dynamical process around the center of TCs on the environment (VHS) should be needed to evaluate the intensity change of TCs. In addition, the interaction of TCs and ocean should be included to evaluate it.

fig08

Fig. 8 Time series of observed (solid line) and simulated central pressure of T1013. Gray shaded shows that T1013 landfall on Luzon Island, the Philippines.

fig09

Fig. 9 The PEH diagram using the CReSS-NHOES simulation on T1013. Horizontal and vertical axes show OHC and VWS, respectively. Intensification rate of the TC is calculated by 3 hours before and after the time. Dark and light gray marks show the intensification and decaying at the time, respectively. Crosses show the period when T1013 landfall on Luzon Island.

A formation mechanism of gust wind around a meso-β-scale vortex developed along the Japan-Sea Polar-airmass Convergence Zone during the cold-air outbreak

When cold-air outbreak occurs over the Sea of Japan during the winter season, meso-β-scale vortical disturbances (meso-β vortices), with horizontal scale of several tens km, are frequently observed along the Japan-Sea Polar-airmass Convergence Zone (JPCZ) by satellite images. The meso-β vortices are sometimes accompanied with the spiral and comma shaped radar echoes and bring about gust wind after their landing on the coastal region facing the Sea of Japan. However, a mechanism of its gust wind has been unclear, because it is difficult to observe the vortices over the sea; that is, their horizontal scale and lifetime are too small and short to observe. The purpose of the present study is to clarify the formation mechanism of gust wind related to the meso-β vortices using a cloud resolving model: Cloud Resolving Storm Simulator (CReSS). We performed a simulation with horizontal resolutions of 500 m on meso-β vortices observed on January 29, 2011.

Figure 10 shows simulated vertical integrated hydrometeor amount corresponding to cloud coverage of meso-β vortices over the Sea of Japan as well as the visible satellite image at 1430 JST on January 29, 2011. The model simulated well the cloud pattern over the Sea of Japan, and the position and scale of vortices along the JPCZ in comparison with the satellite image. The diameter and thickness of the meso-β vortices shown in "A" in Fig. 10 are approximately 100 km and 1.5 km, respectively. Time-height section of area averaged kinetic energy including the meso-β vortex "A" shows that the layer of maximum kinetic energy is found in the low-level and that of minimum appears around a height of 2 km above the top of the meso-β vortex (not shown). Another maximum layer is found above a height of 3.5 km, thus the low-level maxima of kinetic energy is clearly separated with the upper-level one. As a result, the origin of the gust wind around the surface in the meso-β vortex is not expected to be the advection of the high wind region from the upper-level.

A back trajectory analysis is examined to investigate the origin of the gust wind. Figure 11 shows time series of wind speed, kinetic energy, and pressure gradient force of air parcel intruding the gust wind region of the meso-β vortex along the back trajectory. The air parcel has its origin at a height of 1 km of the north and wind speed is small at that time. The wind speed accelerates as the air parcel approaches the meso-β vortex corresponding to the increase of pressure gradient force. Thus we concluded that the forming mechanism of gust wind around the meso-β vortex is the acceleration by pressure gradient force, not by the vertical advection of the high wind in the upper level.

fig10

Fig. 10 (a) Visible image over the Sea of Japan at 1430 JST on January 29, 2011. (b) Vertical integrated hydrometeor amount (kg m-2) simulated by CReSS at the same time. White color shows the existence of high hydrometeor (dense cloud region).

fig11

Fig. 11 Time series of the back trajectory analysis starting at 2100 JST on January 29, 2011 calculated using the simulation result. The upper panel shows time series of wind speed (gray solid line), height (solid line), and distance from the starting position of the back trajectory (dotted line) of the air parcel. The lower panel shows time series of kinetic energy (gray solid line) and pressure gradient force (solid line) of the air parcel.

Impacts of GPS and Doppler radar data assimilation on precipitation prediction - A heavy rainfall event in the southern Gifu Prefecture on July 15, 2010 -

To Improve reproducibility of a numerical weather prediction on local heavy rainfall events, improving an initial condition should be needed. In particular, a moisture field is one of the less described variables in initial conditions because of its high spatial and temporal variability. The aim of the present study is to clarify the sensitivity of data assimilation of the moisture field on the location and amount of precipitation. We conduct a numerical simulation on a heavy rainfall event in the southern Gifu Prefecture on July 15, 2010 using Cloud Resolving Storm Simulator (CReSS) with horizontal grid resolution of 1 km. We cannot reproduce the rainfall distribution and maximum rainfall amount around the target region without assimilating Global Positioning System derived precipitable water vapor (GPS-PWV) and horizontal wind (HW) derived from Doppler radar. Three-dimensional variational data assimilation (3DVAR DA) system is used to investigate the impact of GPS-PWV. HW is also assimilated by nudging method. This study investigates the impact of three different GPS-PWV data such as "Near Real Time analysis (NRT)," "reanalysis," and "F3" data.

Figure 12 shows the impact of GPS-PWV and HW on the horizontal distribution of total precipitation amount for 9 hours starting from 18 JST on July 15, 2010. The location and amount of precipitation in the southern Gifu Prefecture is improved by assimilating both GPS-PWV and HW in comparison with no assimilation experiment (Fig. 12b). However, total precipitation amount is less than half of the observation. Application of sequential assimilation of GPS-PWV with nudging technique improve the reproducibility of total precipitation amount (Fig. 12c, 12d). It suggests that the sequential assimilation of GPS-PWV is quite useful to quantitative precipitation forecasting (QPF).

Figure 13 shows time series of hourly precipitation of each GPS-PWV assimilation as well as the observed one at Tajimi. Maximum precipitation and its peak time assimilating the "reanalysis" GPS data are reproduced better than those assimilating "NRT" and "F3" data. The "reanalysis" GPS data estimates more accurate PWV using a correct GPS satellite position relative to the Earth surface. However, it cannot release in real time since GPS orbit data before and after 30 days are needed to calculate. Thus the "reanalysis" GPS data cannot apply for the forecasting on heavy rainfall events. However, this result suggests that providing more accurate GPS-PWV data is expected to improve the forecasting on heavy rainfall events.

fig12

Fig. 12 Sensitivity of GPS-PWV and HW on horizontal distributions of total precipitation amount for 9 hours starting from 18 JST on July 15, 2010. White contour shows that observed by JMA at every 100 mm. Result of (a) assimilating only GPS-PWV (NRT) data, (b) assimilating GPS-PWV (NRT) and HW data, (c) conducting sequential assimilation of GPS-PWV with nudging technique for all GPS observation points, and (d) same as (c) except for the GPS observation points that observed PWV is greater than the simulated one are drawn.

fig13

Fig. 13 Time series of hourly precipitation of sequential assimilation of different method on GPS-PWV at a grid including Tajimi AMeDAS station. Observed series (solid line), assimilating "NRT" (dotted line), "reanalysis" (chain line), and "F3" (broken line) GPS data are drawn.


[back]