A decade (2011–2020) of tropical cyclone reconnaissance flights over the South China Sea

Abstract

A decade has passed since the Hong Kong Observatory (HKO) and the Government Flying Service (GFS) jointly commenced operational tropical cyclone (TC) reconnaissance over the South China Sea in 2011, currently the only such programme in the region and a first since the USA terminated aircraft reconnaissance in the western North Pacific in 1987. This paper reviews the HKO–GFS flights between 2011 and 2020, including 24 storm-penetrating flights down to 2000ft and 28 dropsonde missions resulting in 190 profiling measurements, which have supported operational forecasting and warning of TCs as well as contributed to scientific research and regional collaboration. Tropical cyclones (TCs) are a major weather hazard faced by many subtropical coastal cities (Gray, 1968). For Hong Kong, which is situated along the southern coast of China and hence susceptible to weather systems originating from both the South China Sea and the western North Pacific (Mcgregor, 1995; Goh and Chan, 2010), the monitoring and warning of TCs has remained a key focus of the Hong Kong Observatory (HKO) since its establishment in 1883 (Ho, 2003). To support operational TC forecasting, observational data – particularly near-surface distributions of winds and pressure – are vital for the analyses of storms, impact assessment and formulation of warnings (Chavas et al., 2017; Klotzbach et al., 2020; Knaff et al., 2021). In recent decades, meteorological satellites have rapidly gained in spatial and/or temporal resolution of observations, offering an ever-widening spectrum of remotely-sensed and derived parameters, hence becoming an indispensable part of the TC analysis, forecast and numerical prediction process (Wang et al., 2008; Schmetz and Menzel, 2015; Emanuel, 2018). Nonetheless, there are certain inherent limitations in the information provided by satellite observations, particularly for dynamic variables under TC conditions. For example, atmospheric motion vectors derived from infrared images of geostationary satellites, while spatio-temporally dense, tend to cover the upper atmospheric layers when nearby deep convective clouds are present (Hidehiko et al., 2015; Kim and Kim, 2018). On the other hand, ocean surface winds derived from scatterometers on board polar-orbiting satellites are constrained by the number of available overpasses as well as their relatively low horizontal resolution (Figa-Saldana et al., 2002; Wang et al., 2007a; Ribal et al., 2021). For the South China Sea basin, the lack of in situ TC observations, particularly near-surface wind distributions, has remained a constant challenge to operational forecasting. It is against such a background that HKO took forward its TC reconnaissance programme in collaboration with the Government Flying Service (GFS). Taking a global perspective, TC reconnaissance flights have a long history since the mid-twentieth century (Simpson and Starrett, 1955). For example, the Hurricane Hunters of the US National Oceanic and Atmospheric Administration (NOAA) have produced routine reconnaissance data over the Atlantic and the eastern North Pacific basins since the 1960s, with sufficient temporal resolution for spectral calculations, over the Atlantic and the eastern North Pacific basins since the 1960s (Vonich and Hakim, 2018). Furthermore, since its routine introduction of the global positioning system (GPS) dropsonde in 1996, NOAA has accumulated more than 10 000 dropsonde observations inside or near hurricane eyewalls covering over 100 systems (Wang et al., 2015). However, after the termination of aircraft reconnaissance in the western North Pacific by the US in 1987 (Gray et al., 1991), there had been no routine aircraft TC observations over the South China Sea region until 2011, when the HKO–GFS reconnaissance programme officially began (Chan and Tse, 2012). At present, it remains the only such programme in the region. At first, data collection took the form of low-level straight-line flights into stormy areas using an instrumented aircraft, with the aim of sampling the distribution of hazardous winds associated with both outer rainbands and the core region of a TC. Such a demanding undertaking is made possible by the skill and professionalism of the GFS, whose responsibilities include search and rescue (SAR) operations over the South China Sea, often in adverse weather conditions. These low-level flights have resulted in a unique set of observational data within the TC boundary layer, including several eyewall penetrations, which are the first-of-its-kind in the region. Starting from 2017, the mode of reconnaissance saw a major change following the introduction of the dropsonde launching system installed on the new fixed-wing aircraft of the GFS. Dropsonde missions, where typically up to 10 dispensable units are launched from the upper troposphere within the Hong Kong Flight Information Region (HKFIR) each time, have replaced the low-level storm-penetrating flights, offering the new technical possibility of 3-dimensional in-situ profiling of TCs. Operationally, for TCs within the domain 105–125°E, 10–30°N, HKO issues forecast tracks (updated every 3h) containing information on storm position, intensity and hazardous wind radii. Reconnaissance observations of both ‘flavours’ provide an important reference to the forecasting bench in validating storm position and wind distribution (radii of gale, storm and hurricane force winds), which are otherwise based on a combined assessment of numerical weather prediction (NWP) and satellite data together with the small set of available surface observations over the data-sparse South China Sea. In most global NWP models, Hong Kong only subtends a few grid-points in either horizontal direction; local winds can exhibit considerable sensitivity to positional or structural changes of only a couple of tens of kilometres. As such, detailed observation on wind radii is particularly valuable in the formulation of local TC warning strategy. This paper reviews the progress of the HKO–GFS reconnaissance programme over the past decade (2011–2020). The article is organised as follows. First, we describe and summarise the low-level storm-penetrating flights operated between 2011 and 2016. Flight missions after the introduction of the dropsonde launching system in 2017 are documented and discussed next, after which contributions to the scientific literature by the reconnaissance programme are reviewed. Finally, we close with a brief summary and outlook. This section gives an overview of the low-level reconnaissance flights conducted using instrumented aircraft between 2011 and 2016. The Jetstream 41 is a fixed-wing aircraft used by the GFS since 1999. It is a twin turboprop aircraft with wingspan of 18.4m and maximum speed of 250kn. It has a radius of action of about 500 nautical miles or around 900km. Its main roles include performing SAR missions as well as providing support to various government departments. Figure 1 shows the data probe installed on the left wingtip of the aircraft, which forms part of the dedicated meteorological measurement system operated by HKO. The measurement system is the Aircraft Integrated Meteorological Measuring System 20Hz (AIMMS-20) manufactured by Aventech Research Inc. A detailed technical description of the AIMMS-20 can be found in Beswick et al. (2008). Here, we summarise the key features of its implementation in Hong Kong. The data probe measures the true airspeed, flow angle, air temperature and relative humidity. A GPS module determines the position, velocity and true heading of the aircraft. Flight parameters, including rates of rotation and accelerations about three body axes, are measured by an inertial measurement unit. Finally, a central processing module handles real-time data collection and computations as well as storage of output data, which are available at up to 20Hz frequency. The Jetstream 41 was in use until March 2016 until it was decommissioned by the GFS. The replacement fixed-wing aircraft, the Bombardier Challenger 605, is described in the next section. Data collection missions are typically initiated whenever a TC enters the northern South China Sea and may impact the weather in Hong Kong. Normally, the process begins with the design of a flight profile up to a couple of days before the planned mission. A number of factors are normally considered, including the actual and predicted storm structure, and projected movement of the storm. These are based on the latest available observation data and HKO's assessment of the storm's evolution, with reference to NWP model output. There are also practical and operational constraints. For example, air traffic control might prefer timeslots with less traffic at Hong Kong International Airport (HKIA) such as in the morning or early afternoon; occasionally, a mission could not be arranged because of tightly packed air traffic arising from anticipated TC passage. The total length and duration of the mission will be subject to fuel consumption and the boundaries of HKFIR. For GFS, SAR missions always take top priority, which may at very short notice divert the fixed-wing aircraft from a planned reconnaissance flight should such an emergency occur. As such, it is not always possible for every storm with the potential to impact Hong Kong to be measured by flight. In terms of the flight path, in order to sample the spatial distribution of high winds surrounding the TC concerned, straight lines cutting across the key quadrants of the storm are preferred, which also minimises potential data quality control issues when the aircraft is turning. The altitude, or flight level, is chosen as a balance between its representativeness of near-surface wind speed and safety of the flight crew. FL020 (‘Flight Level 20’, around 2000ft or 610m above sea level) is commonly used, which is well within the TC boundary layer. An example is shown in Figure 2. This is taken from the mission from Severe Typhoon Utor on 13 August 2013 (Chan et al., 2014). The left panel shows the flight path overlaid on a concurrent satellite image together with a schematic diagram of the flight profile. The right panel shows, from top to bottom, the time series of flight-level winds, their 10m reduction values and the reduced mean sea-level pressure. The list of low-level TC reconnaissance missions conducted between 2011 and 2016 is given in Table 1. A total of 24 missions have been conducted, covering 19 different systems with intensities ranging from low pressure area (LPA) to severe typhoon (STY) during the time of the respective missions. (The LPA on 16 August 2016 would subsequently develop into TC Dianmu; hence, they count as one system.) In particular, there are seven missions flying into a typhoon (TY) and one into a STY, four into a severe tropical storm (STS), eight into a tropical storm (TS) and two into a tropical depression (TD) 1 . The Jetstream 41 performed 20 missions between 2011 and 2015, while the four missions in 2016 were performed by the Challenger 605. The flight paths covered by these 24 missions are shown in Figure 3. Each mission is represented by a solid, coloured line. The grey dotted line marks the approximate boundaries of the HKFIR. It can be seen from Figure 3(a) that the reconnaissance flight essentially spanned all corners of the HKFIR, with the highest density towards the eastern or southeastern sides. This is because, with the exception of systems forming over the South China Sea, which are relatively rare and may approach from the south or southwest (Wang et al., 2007b), TCs affecting Hong Kong tend to follow a northwest or west-northwest track originating from the western North Pacific. Figure 3(b) offers an alternative view by using a storm-relative coordinate system. This is achieved by subtracting from each flight path its corresponding TC position. Here, we use the HKO analysis positions which for storms in the region concerned are available at three-hourly intervals. No temporal interpolation is applied, and the closest analysis time to a particular flight path is used. It can be seen that the reconnaissance flights predominantly sampled the northeast and northwest quadrants, again reflecting local TC climatology. In particular, 12 flights entered within 1° latitude-longitude from the analysed centre position of a TC. Among these, we counted five eyewall penetrations for TCs at typhoon intensity or above: Tembin, Utor, Kalmaegi, Linfa and Nida. To the best knowledge of the authors, these are the only boundary layer eyewall penetration flights performed over the South China Sea. A dropsonde launching system is installed on the new Challenger 605 fixed-wing aircraft of the GFS (Figure 4). This system was declared operational in March 2017. The Challenger is a twin-turbofan multi-purpose aircraft. It has a cruise speed of 470kn and maximum range of 3500 nautical miles. Its radius of action is around 700 nautical miles with 8h of endurance. In addition to various SAR equipment, it can be equipped with the Airborne Vertical Atmospheric Profiling System (AVAPS-II), which is a dropsonde launching system using GPS-enabled dropsondes. The dropsonde is a parachuted, expendable electronic measurement device that contains sensors for pressure, temperature and humidity as well as a GPS sensor from which three-dimensional wind components are derived. The dropsonde is launched from the aircraft and it continuously measures the state of the atmosphere during its descent, which is slowed and stabilised by the parachute. The measurement data, typically available at 2Hz frequency, are transmitted back to the processing unit on the aircraft in real-time using radio communications. HKO–GFS dropsonde missions are operated within the HKFIR. In each mission, up to 10 dropsondes may be launched, which are usually released from around FL300 (an altitude of 30 000ft, or 11 800m). The operational constraints arise from the need to avoid air traffic within the HKFIR, which is one of the busiest airspaces in the world. Similar to the low-level data collection flights, design of the flight profile is based on forecast movement of a TC, as well as actual and anticipated evolution of its structure. Drop locations, typically at separation of 0.5 to 1° latitude–longitude, are selected to sample the extent of high winds around the TC, particular over those quadrants that are expected to affect Hong Kong. Where possible, vertical profiles are also sampled near the core TC region to help estimate its maximum intensity. The dropsonde data, in WMO BUFR format, are also internationally exchanged on the Global Telecommunication System (GTS) in real-time starting from late 2017 (WMO, 2015). A summary of the HKO–GFS dropsonde missions is given in Table 2. Up to the end of the 2020 TC season, a total of 28 missions have been completed, covering 19 different systems. Around the time of their respective missions, the intensities of the systems spanned the whole range of the TC spectrum, including four missions into an LPA, nine for four for one for two for and one for these flights, a total of 190 dropsonde profiles have been The spatial distribution of these dropsonde observations is in Figure In both of Figure each mission is represented by a different set of coloured It can be seen from Figure by the end of 2020, most of the areas within the have been sampled by dropsonde observations, including the eastern and western corners and the southern The region over the northern of the the on Hong Kong within which no could be local we to a storm-relative coordinate system in a to Figure it can be seen that the missions have accumulated observation data over all four quadrants of a sampled TC. In particular, there are seven missions during which measurement profiles are within 1° latitude-longitude from the analysed centre position of the concerned TC, including one mission into a four into a and one into a and one into a The of is the first dropsonde eyewall measurement over the South China Sea in recent in Figure In Figure we the of TC in was the most in the to various and in its path et al., across the entered the South China Sea with intensity on To sample its a dropsonde mission was performed that The HKO–GFS were made near the southeastern boundary of the HKFIR, the northwest of at a of around to from its centre position. The corresponding wind are the near-surface wind observations at an altitude of around 10m above sea At around the time, the programme et al., also performed a flight mission over the sampling the of This may be counted as a unique where observations are made over different quadrants of a TC over the South China Sea by two different reconnaissance within the short of a few the HKO–GFS reconnaissance programme is at operational TC analysis and forecasting, the observation data it has also contributed to a range of scientific which we will summarise in the following the SAR missions of GFS and the low-level storm-penetrating flights of HKO may the fixed-wing aircraft and its meteorological measurement system into often conditions. A practical need to and quality of such valuable observation In a by HKO and the and on a series of of the data processing of AIMMS-20 using and other system with in wind and temperature up to to offering potential to many of the measurement system including operational and research that spectral of winds and inside the TC boundary layer, particularly over the are et al. the first aircraft measurement of boundary layer winds near the centre of a TC over the South China Sea, the of the spectrum and using the 20Hz AIMMS-20 output data under the by the HKO–GFS missions. This has the to research after more data have been For example, et al. analysed a of including and using data from flights covering five TCs between 2013 and 2016. It the of up to wind of before while the vertical length near the layer of the eyewall region is found to be relatively to wind by et al. analysed four eyewall penetration flights between and and found a between vertical and wind speed at the TC core in to more in the North These have the potential to NWP by the of boundary layer under wind In terms of TC structure, et al. analysed the distribution of winds and during the low-level flight into in with observation data for and 10 other TCs with intensity change along the coastal areas of It was found that the of wind and as well as their spatial an important in the of the TC. et al. performed analysis of the low-level flight data for and and found observational of horizontal in the TC boundary layer at a to from the centre of a which is also a first for TCs in the region. On the the three-dimensional of low-level flight data have been by et al. using the NWP of and by et al. and et al. using the and System technical in the model data processing and used, on various of the analyses and forecast could be The of dropsonde observations was by et al. for the of and by et al. for the of both using the model In both on track prediction are the also in the distribution of surface winds and pressure. The dropsonde data also supported the and of using satellite et al. on a new for of surface pressure based on temperature from which against HKO's dropsonde observations when close to the TC core the real-time analyses of TC structure. Finally, the HKO–GFS reconnaissance flights also in a number of research such as the China et al., and the Tropical Research et al., to scientific and reduction on a regional At the time of the HKO–GFS reconnaissance programme remains the only operational TC reconnaissance over the South China Sea region. It various practical and operational accumulated over a a unique observational which has both operational and TC 2011 and a total of 24 low-level flight missions into the TC boundary layer have been performed at down to about 2000ft above sea using instrumented aircraft equipped with a dedicated 20Hz meteorological measurement including five eyewall penetrations into TCs of typhoon intensity or above – these are the only such flights to have been performed over the South China Sea. the introduction of the dropsonde launching system on the new GFS fixed-wing aircraft, a total of 28 dropsonde reconnaissance missions have been performed between 2016 and 2020, resulting in the sampling of 190 vertical profiles that have spanned all four quadrants of a TC, including near the eyewall region of a In addition to providing valuable real-time support to operational TC analysis and forecasting at the observation data from the reconnaissance programme has also contributed to more than a research covering various scientific including processing of measurement data, boundary layer TC and intensity change as well as The HKO–GFS reconnaissance missions also in research on a regional we are that the present of dropsonde measurement profiles over the South China Sea is at an early up to of than has been in particular, over the Atlantic In the the reconnaissance programme will the with a view to for against other ocean on TC and intensity change will together with on more data HKO will to the latest in the prediction and of which is the weather hazard over the South China coastal The Government Flying Service (GFS) and various of the Hong Kong Observatory (HKO) are for their dedicated contributions to the reconnaissance programme over the images of and are by of the Meteorological

Topics & Concepts