No abstract

We studied the temporal and spatial characteristics of extreme typhoon rainfall in Taiwan using Central Weather Bureau hourly precipitation data from 21 surface stations during the past 51 years . Extreme rainfall is defined as 95th percentile intensity of total rain events, or equivalently, rain events greater than 9 mm hr -1 which contribute 40% to the total rain amount in Taiwan. It was found that approximately 70% (20%) of extreme rain is in the typhoon season (Mei-Yu) from July to October (from May to June). There are significant variations of typhoon extreme rainfall over the annual and decadal time scales, with larger extreme rainfall values and events in the periods of 1960-1976 and 1994-2010, and less in the 1977-1993 period. The recent 1994-2009 period has the most extreme rainfall and events, as well as, inter-annual variability. In contrast, there are strong inter-annual variations of Mei-Yu extreme rainfall, but no significant decadal variations. The averaged typhoon rain intensity, however, is about the same, being 19 mm hr -1 in all these three periods. Our analysis indicates that the typhoon extreme rainfall spatial pattern is phased locked with the Central Mountain Range, Taiwan. In general, the amount of extreme rainfall was related to the typhoon translation speed and duration time, but not typhoon intensity. Slower speeds and longer duration time lead to larger extreme rainfall values. Our analysis also indicate that the mean duration time of Taiwan landfall typhoons with northern tracks (tracks north of 23 degrees latitude) is about 3 hours longer than that of southern track typhoons in the last 51 years, and is more likely to produce three times as much extreme rainfall. The interactions of summer or winter monsoons with typhoons are also important factors that may contribute to the extreme rainfall in Taiwan. Examples of extreme rainfall due to typhoon circulation interaction with summer and winter monsoon flows are presented. Monsoon water vapor supply, typhoon slow translation speed, and mesoscale convection due to typhoon-monsoon flow interactions are the key factors in extreme precipitation events.

No abstract

Typhoons with “deflection tracks” (DTs) within a 200-km distance of the mountainous island of Taiwan are examined. We analyze 84 landfalling typhoons that compose 49 DT cases turning to the left-hand side, including 18 with very large deflection angles (DA > 20°) and another 7 having looped tracks (LTs). Most of the large DA and LT cases are “northern landfall” type, reaching Taiwan’s east coast poleward of 24°N and originally possessing relatively slow translation speeds (~4 m s−1). Their average translation speeds, however, increase by 50% in the 3 h prior to landfall. The WRF Model is used to simulate DT cases, and potential vorticity (PV) tendency diagnosis is used to interpret the contributions of the horizontal advection (HA), vertical advection (VA), and diabatic heating (DH) terms. The northern landfall tropical cyclones (TCs) possess significant cross-mountain flow to the south of the storm near the coast, resulting in vorticity stretching (the VA effect) and subsidence warming. The subsidence suppresses storm convection and produces heating asymmetries (the DH effect) that can induce significant southwestward deflections. The cross-mountain VA and DH effects are weaker for the “southern landfall” storms. The results explain well the observed increase of translation speed prior to landfall in DT cases and show that the HA effect, in general, does not contribute to the track deflection. Our results highlight the impact of topography on TC track by the vorticity stretching effect and by asymmetric diabatic heating.

Recent observational and numerical modeling studies of the mechanisms which transport moisture to the stratosphere by deep convective storms at mid-latitudes are reviewed. Observational evidence of the cross-tropopause transport of moisture by thunderstorms includes satellite, aircraft and ground-based data. The primary satellite evidence is taken from both conventional satellite of thunderstorm images and CloudSat vertical cloud cross-section images. The conventional satellite images show cirrus plumes above the anvil tops of some of the convective storms where the anvils are already at the tropopause level. The CloudSat image shows an indication of penetration of cirrus plume into the stratosphere. The aircraft observations consist of earlier observations of the "jumping cirrus" phenomenon reported by Fujita and recent detection of ice particles in the stratospheric air associated with deep convective storms. The ground-based observations are video camera records of the jumping cirrus phenomenon occurring at the top of thunderstorm cells. Numerical model studies of the penetrative deep convective storms were performed utilizing a three-dimensional cloud dynamical model to simulate a typical severe storm which occurred in the US Midwest region on 2 August 1981. Model results indicate two physical mechanisms that cause water to be injected into the stratosphere from the storm: (1) the jumping cirrus mechanism which is caused by the gravity wave breaking at the cloud top, and (2) an instability caused by turbulent mixing in the outer shell of the overshooting dome. Implications of the penetrative convection on global processes and a brief future outlook are discussed.