Overview of the KMA/NIMS Atmospheric Research Aircraft (NARA) and its data archive: Annual airborne observations over the Korean peninsula

This study describes the Korea Meteorological Administration/National Institute of Meteorological Sciences (KMA/NIMS) Atmospheric Research Aircraft (NARA) and its observational data archive. NARA has been performing annual observation flights, which are aimed at reducing the uncertainty of atmospheric observations in the observation data gap area around the Korean peninsula, since January 2018. An online system has also been constructed to provide data management, transmission, quality control and simple visualization. The mission strategy of NARA is subdivided into four individual units, namely observations of severe weather (SW), climate monitoring (CM), environmental monitoring (EM) and cloud physics and weather modification experiments (CP). As of December 2020, NARA has been launched operationally for 325 flights (corresponding to 9.02 flights per month), typically over the West Sea, mid‐inland areas (34.4% and 25.4%, respectively), and below an altitude of 3 km (51.9%). Results of intercomparison tests confirmed that NARA measurements have reasonable offsets (<0.9%) to each other in terms of pressure and temperature. Moreover, the monthly average temperature profile in the East Sea area showed a seasonal variation was detected in monthly variation. From these results, it is evident that NARA data will contribute significantly to enhancing the level of scientific understanding of atmospheric observations and the applications (i.e. long‐term study) thereof as the amount of data accumulated increases.


| INTRODUCTION
Atmospheric observations have been performed regularly over the past few decades and are based on numerous different platforms, including fixed ground stations, aircrafts, automobiles, ships, satellites and mobile platforms such as balloons, cell phones and unmanned aerial vehicles (Stith et al., 2018).Generally, airborne observations have the advantages of higher spatial and temporal resolution compared with other observation platforms (e.g.satellites).In addition, direct airborne observations of the maritime atmosphere or upper troposphere are possible, allowing scientists to fill the gap in observational blank areas.Owing to these advantages, they have focused on continuing to improve airborne observations for research purposes (Geerts et al., 2018).
Applications of airborne observations are being continuously expanded to areas where this was formerly impossible.Examples include monitoring storm and typhoon dynamics, weather modification experiments, exploration of cloud microphysics, solar and terrestrial radiation, monitoring of air quality and satellite validation (Bushnell et al., 1973;Goyer et al., 1966;Herman, 1977;Knollenberg, 1970;McCormick et al., 1979;Reynolds et al., 1975).Moreover, the expansion of airborne observations meant that these have become a crucial component of modern atmospheric observations (McBeath, 2014).Several airborne observation campaigns have occurred in and around the Korean peninsula.These campaigns have been conducted for numerous applications, including characterization of the physical-optical-chemical properties of aerosols and determining the distribution and trans-boundary transport of solid/aqueous/gaseous phase air pollutants (Herman et al., 2018;Huebert et al., 2003;Kim et al., 2014;Ramana et al., 2010;Stith et al., 2009).
However, except for several functional studies, most of these airborne observations were conducted within a short period campaign that targeted a specific space and time, although annual operational observations are crucial for investigations of sub-annual variations such as monthly and seasonal variation (Andrews et al., 2004;Sawa et al., 2012;Sheridan et al., 2012).Similarly, airborne observation campaigns conducted in Korea have typically been conducted over short periods.Moreover, the campaigns were generally conducted using either leased commercial aircrafts or with the cooperation of external organizations.
The National Institute of Meteorological Sciences (NIMS), a division of the Korea Meteorological Administration (KMA), initiated the operation of the KMA/NIMS Atmospheric Research Aircraft (NARA) at the beginning of 2018 with the aim of contributing to reducing the uncertainty of atmospheric observations around the Korean peninsula (Lee et al., 2011).NARA was launched for annual observational flights around the Korean peninsula more than 3 years ago.In addition, with the marked increase in the number of earth observations based on various platforms, data archives have garnered increased attention as valuable and essential tools for researchers (Kiemle et al., 2016).Numerous research organizations have also produced data archives (Adler et al., 2003;Noone et al., 2020;Schneider et al., 2013).In addition to launching NARA, the NIMS has built an online data archive to encourage scientists to utilize NARA data.
In this study, NARA and its data archive of annual airborne observation datasets for the Korean peninsula.Were introduced.Section 2 describes the NARA, including the atmospheric instruments used for observations; Section 3 discusses its science and mission strategies; Section 4 presents an online data management system; Section 5 explains the data archive; and this study is summarized in Section 6.

KMA/NIMS ATMOSPHERIC RESEARCH AIRCRAFT (NARA)
NARA is a modified KingAir 350HW (Beechcraft Inc.) aircraft.Its maximum altitude and flight hours are up to 10.7 km and 5.5 h, respectively.Its crew capacity is two pilots, two operators, and one scientist with 25 atmospheric instruments (Table 1).Different combinations of atmospheric instruments are utilized depending on the purpose of the scientific mission.Observations preceding severe weather (SW) are performed by measuring the atmospheric profile, surface wind speed, liquid water contents and gamma rays using dropsondes, SFMR, GVR and RSX-3 (Franklin et al., 1993; Klotz that NARA data will contribute significantly to enhancing the level of scientific understanding of atmospheric observations and the applications (i.e.long-term study) thereof as the amount of data accumulated increases.
Figure 1 illustrates the atmospheric instruments aboard the NARA, which are located on both its interior and exterior, considering aerodynamic characteristics.In addition, inlets are placed in front of the engine to minimize the effects of exhaust gases and flares.The dropsonde launcher is located underneath the tail to prevent collisions.Iso-kinetic/window inlets are used for aerosols/gas analyzers.Additionally, pressure stabilizers are employed in the gas analyzers, CCNC and SP2 for sampling and precision control of flow/pressure.Water traps and valves are installed at the inlets of the CRDS and gas analyzers to avoid sample contamination and protect the instruments.
To prevent the hygroscopic growth of aerosol particles, a hotwire heater was installed at the sampling tube of the nephelometer and Sky-OPC to maintain sample air dryness (<40% RH).
Temperature and RH measurements were duplicated for backups in terms of data from the AIMMS-20, total temperature sensor and dew point hygrometer.M300 is a centralized data system used for operation, display and acquisition (Delene, 2011).M300 shares GPS time information to ensure that all observation data have an identical time.NARA also contains a spare margin (space, power and payload) for potential use for creating synergy with existing nephelometer.it is also being considered that installation of an aerosol absorption instrument, such as Tricolour Absorption Photometer (TAP, BMI Inc.), which is a commercialized version of the Continuous Light Absorption Photometer (CLAP, NOAA; Kim et al., 2019;Ogren et al., 2017).

| SCIENCE AND MISSION STRATEGY
The details of NARA missions were explained in Lee et al. (2011).Table 2 describes NARA missions and mission codes from its inception in early 2018; the mission code and sub-mission codes represent the scientific goals and details of the actual target, respectively, expressed as Arabic numerals.For example, the combination of the  mission and sub-mission codes, EM-01, indicates that the scientific target of the flight is environmental monitoring, predominately air quality monitoring over the West Sea area.The combined code system employed by NARA is effective for managing data archives.SW, EM and CP have the same sub-mission, which is described as 'Validation of satellite data retrieval algorithms', but they differ in terms of meteorological parameters to be validated.For example, SW-04 denotes a mission to validate fundamental weather variables (e.g.temperature and RH) and EM-02 aims to validate atmospheric constituents related to air quality (e.g.gas concentration and aerosol size distribution).
Figure 2 presents the main target areas (boxes) and the representative flight routes of each mission (coloured lines).Although the target area of each mission depends on the location of the weather phenomena of interest, many mission flights are conducted in the boxed areas.Figure 2a shows target areas for SW, including some in the West, East and South seas.For example, the SW flight on September 16, 2020, was targeted to the West Sea area close to the border between China and Korea.At an altitude of approximately 4.5 km, NARA launched 4-8 dropsondes along the West Sea border from the north to the south (the number of sonde launched was generally in this range during not only this example flight but also, the most of the flights).The West Sea area is essential for observing the airmass entering the Korean peninsula with the westerly wind.Figure 2b describes the key areas for the CM, including Anmyeon-do, an island home to one of the regional global atmosphere watch stations (AMY) in the western part of the Korean peninsula.The CM flight on 27 April 2019 was a spiral observation covering the altitude range approximately 0.6-9 km around the AMY located on the west coast of the Korean peninsula.By vertical profiling above the AMY, NARA measured the concentrations of greenhouse gases (GHGs) at this location.Vertical profiling has two advantages: it can be used to validate both ground-based in-situ measurements and retrievals of ground-and space-based remote sensing; furthermore, it improves our understanding of the vertical distribution of GHGs. Figure 2c indicates EM target areas, including the West Sea and the southern inland area.The EM flight on April 19, 2019, targeted an area in the West Sea, situated close to the border between China and Korea.Air quality measurements were performed on several layers from 0.6 to 1 km within the boundary layer.The West Sea area is essential for monitoring pollutants, including short-lived climate forcers transported to the Korean peninsula due to Asian continental outflow (He et al., 2003).Figure 2d illustrates the target areas of the CP, including the northeastern and southern parts of the inland areas.For example, the CP flight on 27 March 2020 targeted the northeastern area around the Taebaek Mountains, close to the east coast of the Korean peninsula.By maintaining an altitude of 2 km, an AgI flare was burned as part of cloud seeding to enhance precipitation.This area is favourable for investigating the nature of the nucleation of cloud droplets due to the topographical effect of the moisture flowing with the easterly wind over the mountain region.
NARA observations have, so far, been limited to the Korean peninsula.However, international collaboration is planned in the future; NIMS is currently collaborating with the Taiwan-area Atmospheric and Hydrological

| NARA OPERATION AND MANAGEMENT SYSTEM (NOMS)
After the observation flight, measurement data were transferred to the M300 and then transmitted to the NARA data archive.The web-based NARA Operation and Management System (NOMS) 1 provides multiple functions, including data archive management, distribution, quality control, simple visualization and auxiliary information such as in-flight log-notes and instrument history (Figure 3).Anonymous users can access most of their functions, and users who log in can access the full functions.Users can obtain data through FTP protocols located in the ASCII files in the ICARTT 2 format.After the data are obtained from the archive, automated quality control (AQC) and manual quality control (MQC) are performed.AQC algorithms have been developed by adding empirical optimizations to the manufacturer's QC algorithms.Figure 4 shows an example of the dropsonde AQC algorithm.Once raw data are obtained, the AQC algorithm determines whether the status flags accurately represent the actual launch status and then feeds the raw data into Aspen, the manufacturer's algorithm. 3Aspen conducts an extensive series of QC protocols that remove contaminated data points to ensure that data quality is high (Martin & Suhr, 2021).The data are subsequently forwarded for algorithm checks, including validity and latency of the GPS connection and early termination.The dropsonde data can then be used to prove that the data are in AQC level.Likewise, similar AQC procedures (a combination of manufacturer and NIMS empirical optimization) are also applied to the other instruments aboard NARA.MQC is triggered at user demand by flagging certain data points.Details of the flags are provided by the regulations of EMEP (European Monitoring and Evaluation Programme; Tørseth et al., 2012).Additionally, NOMS also provide simple data visualization and metadata (i.e.maintenance log and photo).Figure 6 indicates the total annual flights and purpose of each flight for the 2018-2020 period.SW accounted for 36.7% of all the missions due to the ICE-POP.Since 2019, the number of CM flights has increased as a result of the initiation of CM-02.EM accounted for approximately 11% of all the missions and CP increased from 21.1% to 34.1% in 2019 and 2020, respectively, due to the contribution of IJCO-WCE.The number of PO (Preliminary Operation) missions conducted in 2018 was the highest thus far; this number decreased steadily as the system increased stabilization. Figure 7 illustrates the flight frequency according to  7a indicate the most frequent operations among the missions and were selected based on the flight altitude of the target area.More than 60% of the flights were over highly populated areas and had an altitude of <4 km.Notably, 51.9% of the flights were concentrated below 3 km.Flights at 4-5 km accounted for 16.0%, which corresponds to the SW cruising altitude over the West Sea area.For the SW flights, a dropsonde was launched at either 4-5 km or around 10 km.This altitude is regulated by an avionic permit for launching a dropsonde.For the remainder of the SW flights, 22.7% were above 5 km including 6.2% at 8-9 km. it is because the spiral flight of CM-01 reaches a maximum altitude at 9 km.EM was conducted mainly to observe the air quality near the boundary layer.CM flights play a key role in vertically measuring GHGs throughout the entire altitude range (0-10 km).CP flights were primarily performed at an altitude of 1-3 km with the purpose of enhancing precipitation using seeding methods for cloud microphysics research.

| Simple statistics on the flights
Figure 7b shows the number of NARA flights in and around target areas for the 2018-2020.Flights took place in and around the Korean peninsula.The specific target area of each flight was determined by considering factors such as the scientific goal, air space condition, and the fairway routes of commercial airlines.Of the flights, 34.4% were performed in the West Sea, which constitutes the largest portion of the analysis area.This area is located upwind of the Korean peninsula, considering prevailing winds.In addition, 15.6% of flights were performed over the East Sea.Satellite validation and atmospheric profiling during heavy snowfall events can also be conducted in this area.In terms of the South Sea, 8.2% of the flights were conducted in this area, which was the smallest portion of the analysis area.Preceding observation flights in this area can aim specifically at the accurate forecasting of typhoon trajectories.Of the flights, 25.4% were performed in the mid-inland region of the Korean peninsula.Cloud seeding and GHG monitoring can be conducted in mid-inland areas.Of these kinds of flights, 16.4% were conducted in the southern-inland region of the Korean peninsula.Additionally, particular flights can be conducted in the southern-inland area to monitor radioisotopes 2-3 times per year.
Figure 9 shows the intercomparison results of the vertical profiles measured by the dropsonde and AIMMS-20.Each data profile was averaged using a 10 m vertical resolution, and the pressure profiles showed significant agreement as shown in Figure 9a.The two pressure profiles were decreased with increasing altitude in a logarithmic relation, consistent with the equations of state and hydrostaticity.Temperature also showed a linear decreasing trend with increasing altitude in Figure 9b, and its lapse rate was approximately −6.4 K km −1 , which is close to the ideal lapse rate (Moore, 1956).RH profiles were highly variable, particularly below 3 km, as shown in Figure 9c.To compare the differences among the profiles, the coefficient of difference (CD) was defined by Equations (1 and 2) based on 'coefficient of variance' and 'root mean square difference' (RMSD): where the number in the subscript represents each profile, subscript i is the ith element of the profile, and the overbar (¯) is the average.The CDs of pressure, temperature and RH were 0.1%, 0.9% and 15.8%, respectively (arbitrary reference: dropsonde; averaged across altitudes).The results show that the pressure and temperature measurements offsets are in reasonable range.Although the CD of RH appeared to be greater than those of pressure and temperature, the three RH profiles showed similar structures.A high RH of approximately 70-90% was (1)  of the dropsonde are described in Section 5.2.Atmospheric profiles were by a 4.0 km altitude because of release altitude restrictions for the West Sea area.This study did not consider spatial differences and weather conditions.A seasonal variability in temperature was observed, with temperature increases recorded from April to August and decreases from August to November.In January, a surface temperature of 280 K was recorded; however, in August, this temperature was recorded at an altitude of 4 km.Although significant uncertainty can arise due to the inhomogeneity of the number of dropsondes released per month (Figure 10), the temperature variation tendency observed was consistent with the results of previous studies (Ahn et al., 1984;Peel et al., 2007).With an increase in datasets accumulated in the data archive, for example it can be possible to conduct long-term studies of atmospheric characteristics for investigating on the trend analysis (Jung et al., 2019;Ku et al., 2020;Li et al., 2020).

| SUMMARY
The has increased gradually, and most are conducted the West Sea area (34.4%) and below altitude of 3 km (51.9%).The results of an intercomparison of atmospheric instruments aboard NARA showed significant agreement and the potential capability for long-term studies.The significance of the NARA data archive can be summarized as follows.
• Diversity: multiple missions and numerous meteorological parameters • Continuity: annual operational observations over the Korean peninsula • Openness: online data management system Although airborne observations are not recent, they have numerous potential applications, particularly in regions where access is limited, such as the Korean peninsula.Moreover, airborne observations provide the research required for the long-term analysis of meteorological phenomena, which could contribute to reducing the level of uncertainty and enhancing the level of scientific understanding of tropospheric mechanisms.

F
I G U R E 1 Atmospheric instruments aboard the KMA/NIMS Atmospheric Research Aircraft (NARA) (solid line: top, broken line: bottom) T A B L E 2 Descriptions of NARA missions and sub-missions and their corresponding codes Observation and Prediction Experiment (TAHOPE, Taiwan) of the Korean Precipitation Observation Program (KPOP), which will contribute to the Tropical cyclones-Pacific Asian Research Campaign for Improvement of Intensity estimations/forecast (T-PARC-II) and the Prediction of Rainfall Extreme Campaign In the Pacific (PRECIP).

Figure 5
Figure 5 shows the number of monthly NARA flights conducted annually since 2018-325 in total, corresponding to a total of 9.02 flights per month.In 2018, 2019 and 2020, respectively, 106, 109 and 110 flights were operated.The months with fewer flights was attributed to maintenance repair and overhaul (August 2018, March and September 2019, July and December 2020) and auxiliary repairs (May 2018 and February 2020).The high number of flights conducted from February to March 2018 resulted from collaboration

F 2
Primary mission areas (dashed box) and sample of actual flight pattern (solid lines) for (a) SW (September 16, 2020), (b) CM (April 27, 2019), (c) EM (April 19, 2019), and (d) CP (March 27, 2020) missions.The cross marks the location of Gimpo airport altitude and mission area from 2018 to 2020.The mission codes in Figure

F
I G U R E 3 Site map for the 'NARA Operation and Management System (NOMS)' (black: before login, red: after login) F I G U R E 4 Automated quality control (AQC) algorithms for dropsondes

E 5
Distribution of monthly flights for the 2018-2020 period F I G U R E 6 Distribution of flights for the 2018-2020 period according to mission purposedetermined at an altitude lower than 2 km; again, value rapidly declined to 30% at 2.5 km and showed a trend which was gradually decreased with increasing altitude.The results show that water vapour presents more spatial/temporal variability than pressure and temperature(Trenberth et al., 2005).

5. 3 |
Figure10shows the monthly variation of vertical temperature profiles from 254 dropsonde measurements in the West Sea area for the 2018-2020 period.The specifications Korea Meteorological Administration (KMA)/ National Institute of Meteorological Sciences (NIMS) Atmospheric Research Aircraft (NARA) has conducted 325 flights around the Korean peninsula since 2018.The NARA missions focus on severe weather, environmental and climate monitoring and cloud physics and weather modification experiments.The NARA Operation and Management System (NOMS) was designed to manage data obtained during NARA missions, including quality control and distribution.The number of annual flights F I G U R E 9 Comparison of the vertical profiles recorded by AIMMS-20 and a dropsonde on August 30, 2019: (a) pressure, (b) temperature and (c) relative humidity; green lines: AIMMS-20 descending, blue lines: AIMMS-20 ascending and red: dropsonde F I G U R E 1 0 Monthly variation of the temperature profile for the West Sea area from 2018 to 2020 (the numbers on top are the number of dropsondes per month)

T A B L E 1 Atmospheric Instruments aboard NARA Missions Atmospheric instruments (model, manufacturer) Parameters
CCN counter (CCN-200, DMT) Cloud condensation nuclei size distribution Cloud Probe (CCP, DMT) Cloud droplet size distribution and image Precipitation Probe (PIP, DMT) Precipitation droplet size distribution and image WCM-2000 (SEA) Liquid Water Content (LWC), Total Water Content (TWC)