Abstract

New automated observation systems for use in passenger aircraft to measure atmospheric carbon dioxide (CO2) and other trace species have been developed and are described in this paper. The Continuous CO2 Measuring Equipment (CME) is composed mainly of a nondispersive infrared analyzer, a datalogger, and two calibration cylinders for in situ CO2 measurements. The Automatic Air Sampling Equipment (ASE), on the other hand, is designed for flask sampling; the instrument, connected to a metal bellows pump, is made up of a specially designed control board and can accommodate 12 flasks. The CME platform can be used to conduct high-frequency measurements of CO2 for obtaining a detailed spatial observation over a wide area, while ASE, despite the limited flight frequency, can provide useful distributions not only of CO2 but also various trace gas species, as well as their isotopic ratios. ASE and CME are installed on the racks in the forward cargo compartment of the aircraft and the air bypass intake is mounted on the air-conditioning duct upstream of the recirculation fan. Both sets of sampling equipment are automatically controlled through input of relevant flight parameters from the aircraft data system. Their deployment in a Boeing 747-400 aircraft was approved by the aviation regulatory agencies in the United States and Japan through issuance of the supplemental type certificate (STC), while the approval for installation of CME in a Boeing 777-200ER was also obtained via STC. First measurement results of CO2 variations obtained by CME and ASE deployed on Japan Airlines (JAL) aircraft are reported herein.

1. Introduction

Atmospheric greenhouse gases have been increasing exponentially because of anthropogenic activities, such as large deforestation and fossil fuel combustion. The growth rate of atmospheric carbon dioxide (CO2) for the 1995–2005 period was 1.9 ppm yr−1, the largest in any decade over the last 200 yr (Forster et al. 2007). To predict the future CO2 level with a sufficient degree of reliability, a quantitative understanding of the global carbon cycle is necessary.

Top-down inverse methods to estimate regional carbon sources and sinks employ atmospheric transport models constrained by observed atmospheric CO2. For example, Gurney et al. (2004) estimated the seasonal variations of CO2 fluxes in 11 land and 11 ocean regions, averaged for the period from 1992 to 1996, while Baker et al. (2006) obtained interannual variability of CO2 sources and sinks over the 1988–2003 period for regions grouped into land and ocean components in three broad latitude bands. In these and other cases involving inverse estimates, it has been clearly identified that more CO2 measurements in the atmosphere are needed to reduce the uncertainties in CO2 flux estimates.

Atmospheric CO2 has been measured by a large number of ground-based stations and cruise ships as part of the existing global monitoring networks [e.g., WMO 2006; additional information available from the National Oceanic and Atmospheric Administration (NOAA)/Earth System Research Laboratory (ESRL)/Global Monitoring Division (GMD), online at http://www.esrl.noaa.gov/gmd/]. Comparatively, CO2 measurements in the free troposphere above the planetary boundary layer (PBL) by research aircraft using onboard measurement systems are sparse, both in time and space (e.g., Gerbig et al. 2003; Machida et al. 2003; Vay et al. 2003; Sawa et al. 2004). The high cost and the “snapshot” nature of atmospheric CO2 observations associated with aircraft measurements have prevented one from obtaining a relatively good characterization of spatial and temporal CO2 variability above the PBL. However, a partial remedy to this situation has been obtained by making CO2 measurements using commercial airliners that make regular and frequent flights. Some examples include flask air sampling using commercial aircraft to examine vertical and horizontal distributions of CO2 over Japan (Nakazawa et al. 1993), Australia (Pearman and Beardsmore 1984), the western Pacific (Nakazawa et al. 1991; Matsueda et al. 2002), from Europe to the tropics (Brenninkmeijer et al. 1999), and over northern high-latitude regions around North America, Scandinavia, and the Arctic (Bolin and Bischof 1970). These airline observations using flask sampling systems have provided a database for studies into atmospheric transport processes such as convection, interhemispheric transport through the upper troposphere, and exchange between the upper troposphere and lowermost stratosphere (UT/LS; e.g., Taguchi et al. 2002; Shia et al. 2006). Regular and long-term CO2 flask measurements obtained on Japan Airlines (JAL) aircraft have been demonstrated to be quite useful in reducing inverse flux estimate uncertainties (Maksyutov et al. 2003), as well as validating satellite-based CO2 observations in the free troposphere (Chédin et al. 2003; Engelen and McNally 2005; Peylin et al. 2007). Recently, Stephens et al. (2007) reestimated CO2 fluxes in the northern and tropical land regions using vertical CO2 data obtained by aircraft measurements at 12 sites in the world. However, more systematic and long-term aircraft measurements of CO2 with wide spatial coverage are needed, not only to constrain further regional inverse flux estimates, but also to increase our understanding of the global carbon processes, particularly in regards to the atmospheric transport.

The commercial aircraft–based observations using in situ measurement systems have made significant contributions to various atmospheric chemistry research and monitoring projects, such as Civil Aircraft for the Regular Investigation of the Atmosphere Based on an Instrument Container (CARIBIC; Brenninkmeijer et al. 2007), Measurement of Ozone and Water Vapor by Airbus In-Service Aircraft (MOZAIC; Marenco et al. 1998), and Measurements of Nitrogen Oxides and Ozone along Air Routes (NOXAR; Brunner et al. 2001). In particular, the CARIBIC program focused on ozone, many reactive gases, and aerosols, which cannot be observed by grab air-sampling methods. Airborne CO2 measurement systems were originally developed for campaign-style research (e.g., Machida et al. 2003; Sawa et al. 2004); therefore, installation of similar systems on commercial aircraft required significant modifications of the systems, in order to avoid interference with aircraft operations and to adhere to the strict safety regulations of commercial airlines.

During the first phase (1993–2005) of the aircraft observation program in collaboration with JAL, Automatic Air Sampling Equipment (ASE) was installed on the Boeing 747-200 to collect air samples over the western Pacific between Australia and Japan at an altitude of about 10 km (Matsueda et al. 2002). To continue the observations after 2005 (second phase), the ASE instrumentation had to be removed from the Boeing 747-200 and reinstalled on a newer plane, the Boeing 747-400. With the continued collaboration of JAL, we took this opportunity to install Continuous CO2 Measuring Equipment (CME), a new instrument for in situ CO2 measurements, on Boeing 747-400 and 777-200ER aircraft. Because CME can be placed on board an aircraft for 1–2 months, we have been able to extend the observation frequency from twice a month (by ASE) to almost every day, and to extend the spatial coverage to include routes to Europe, Asia, Hawaii, and North America. With CME we have been able to obtain vertical profiles of CO2, as well as enough data spatially to delineate certain atmospheric dynamics that influence CO2 transport, such as the air exchange between UT/LS and between the Northern and Southern Hemispheres. The development of the new instrumentation had taken approximately 3 yr, from 2003 to 2005, and involved collaborations of scientific institutes in Japan, the JAL Foundation, JAL, various aircraft engineering companies, and the aviation regulatory agencies. In this paper we provide a detailed description of the technical and practical aspects of the ASE and CME instruments, as well as some preliminary measurement results demonstrating the capability of the instruments operated on board commercial passenger aircraft.

2. Installation of observing system

a. System for Boeing 747-400

Figure 1 shows the configuration of the observing system for both ASE and CME on the Boeing 747-400 aircraft. Sample air is collected from the air-conditioning system in the aircraft, as described in Matsueda and Inoue (1996). Fresh air outside the aircraft is compressed by an engine and fed into the air-conditioning ducts by passing it through a pneumatic system and air cycle packs for removing water vapor and destructing ozone; it is then mixed with the cabin air at the recirculation fan. A new bypass intake port for the air sampling (one for ASE and another one for CME) is mounted on the air-conditioning duct upstream of the recirculation fan so that the sample air is not contaminated with the cabin air. Because the air-intake ports are located under the floor of the main cabin, stainless-steel tubes—one 13 m long with a diameter of 3/8 in. for ASE and another one 9 m long with a diameter of 1/8 in. for CME—are laid under the cabin floor and extended toward the forward cargo compartment. A filter (230-μm mesh size) is connected to both air sample tubes to remove coarse dust from the air-conditioning duct. After installing the air sample tubes, a careful leak test around the tubing connectors was performed using a helium leak detector to ensure there was no contamination from the surrounding air outside the tubes.

Fig. 1.

3D schematic of equipments and airflow tubes below the cabin floor in a Boeing 747-400. Two air intakes for the ASE and CME are mounted on the air-conditioning duct. ASE and CME are installed on racks in the forward cargo compartment. Details of the system are described in the text.

Fig. 1.

3D schematic of equipments and airflow tubes below the cabin floor in a Boeing 747-400. Two air intakes for the ASE and CME are mounted on the air-conditioning duct. ASE and CME are installed on racks in the forward cargo compartment. Details of the system are described in the text.

The entire system (ASE, CME, and a pump) is placed in the forward cargo compartment, where it is easily accessible by aircraft engineers for maintenance service and wiring of the electric power supply and aircraft data buses (Fig. 1). The ASE and CME packages are installed on two aviation-approved racks positioned in a vacant space next to the water tanks at the back part of the forward cargo compartment. The instrument packages can be easily placed or removed from the racks by the aircraft system engineering company for such purposes as regular maintenance and data collections. Several stainless steel flexible tubes attached to the racks are used to connect the sample inlets to the vents on ASE and CME using quick connectors. A metal bellows pump (Metal Bellows, MB-302, P/N 45067, 7.3 kg) used for ASE is located on a stand at the sidewall of the forward cargo compartment to avoid the vibration. The flow rate of the sampled air using the MB-302 pump is more than 8 L min−1 STP, easily flushing the long tubing from the air-intake port to the actual ASE instrument. The ASE pump is operated using the AC power supply (115 V, 400 Hz, 3 phase) from the aircraft, although the operation of the ASE and CME instruments is run by a 28-V direct current (dc) electric power supply. The ASE–CME system has been installed on two Boeing 747-400 aircraft (JA8917 and JA8921) operated by JAL.

b. System for Boeing 777-200ER

For the Boeing 777-200ER aircraft, only the CME instrument package is installed in the vacant space on the sidewall of the forward cargo compartment. The setup is similar to the one described above for the Boeing 747-400. The instrument has been installed on three Boeing 777-200ER aircraft (JA703J, JA705J, and JA707J) operated by JAL.

3. Automatic ASE

a. Instrumentation description

The basic design of the new ASE for flask air sampling is similar to the previous ASE model developed more than 12 yr ago (Matsueda and Inoue 1996), but all of the components have been replaced with new ones. One of the most significant advancements in the new ASE instrument is the sampling operation system, which can be automatically controlled using real-time monitoring of flight navigation data from the ARINC, 429 data bus of the aircraft. Although the sampling intervals in the previous ASE model were determined by a preset time sequence, sampling locations for the new ASE instrument are determined by a predesignated set of numbers that refer to latitude, longitude, or altitude.

Figure 2 shows a schematic diagram of the new ASE instrument, which is separated into two packages (ASE-1 and ASE-2) for easy handling; each package has a dimension of 430 mm length × 215 mm width × 531 mm height, with ASE-1 and ASE-2 weighing 17.9 and 15.6 kg, respectively. Three-dimensional views of ASE are shown in Fig. 3. The main components of each ASE package are six sample flasks, solenoid valves, and two pressure sensors (Fig. 2), so that 12 air samples in total can be collected during one flight. Before introduction of air into the flasks, a filter (60 μm) is used to completely remove dust and other small particles. The cylindrical flasks, each with an internal volume of about 1.7 L (260 mm long, 99 mm I.D.), are made of titanium with a thickness of 0.9 mm to reduce the total weight of the instrument. At both ends of each flask, solenoid valves (28-V dc operation) with orifices of 1.2 mm (CKD, HVB112-X0453) are attached.

Fig. 2.

Schematic of airflow for the ASE separated into two packages of ASE-1 and ASE-2. The 12 titanium air sample flasks are connected via solenoid valves to the airflow. Details of the sampling procedure are described in the text.

Fig. 2.

Schematic of airflow for the ASE separated into two packages of ASE-1 and ASE-2. The 12 titanium air sample flasks are connected via solenoid valves to the airflow. Details of the sampling procedure are described in the text.

Fig. 3.

3D views of (left) ASE-1 and (right) ASE-2.

Fig. 3.

3D views of (left) ASE-1 and (right) ASE-2.

Pressure sensor 1 with a pressure range of 0–100 psia (Druck, A3203-10A), placed in ASE-1, is connected to the sample flow tubing to monitor the sampling pressure in the flasks in both ASE-1 and ASE-2. On the other hand, pressure sensor 2 with a pressure range of 700–1100 hPa (Druck, A3203-05B) is separated from the sample flow tubing to monitor the cabin pressure. After takeoff, the cabin pressure of the aircraft decreases, eventually settling around 800 hPa at the cruising altitude of about 9–12 km. This information is used by the pressure sensor 2 to count the number of flight legs during a round-trip of the aircraft. Thus, the sampling starting time as dictated by the cabin pressure sensor is independent of any delay in departure time of the aircraft.

The automated sampling operation is performed by specially designed control boards attached to ASE-1 and ASE-2. These boards control the pump and solenoid valves for air sample collection using information from the ARINC 429 data bus, as well as from pressure sensors 1 and 2. An ARINC 429 Serial Port Adapter (RTX Systems) interface in ASE is connected to three channels of the ARINC bus of the aircraft data system. The information received by the ASE control boards are date, time, pressure altitude, radio altitude, latitude, longitude, ground speed, and static air temperature. Before each flight, the sampling operation parameters (such as sampling locations, altitudes, and sample air pressure) in the control boards can be set by a laboratory computer. During the flight, the control board records the values of the ARINC parameters at each of the 12 air-sampling locations, as well as the information related to valve and pump operations so that the sampling procedure can be checked after the flight.

To prevent any compromise of the aircraft safety by ASE, the instrument is designed to shut down immediately when anomalous conditions are detected. For example, an abnormal air pressure increase inside the tubing can be detected by pressure sensor 1 to shut the pump off to stop the airflow. In addition, a relief valve connected to the tubing in ASE-1 will mechanically relieve any pressure greater than 30 psig. Temperature sensors on the main control board monitor the heating condition around the ASE instrument to stop the electric power supply when the temperature exceeds 70°C. The cutting off of the electric power under the same condition is also accomplished mechanically by a thermal fuse. Thus, redundancy is an important safety component in an automated sampling system installed in a commercial aircraft.

b. Flask sample analysis

To reduce the flask storage time between sampling and analyses, air samples are collected during return flight to Narita, Japan. Immediately upon arrival at the airport, the ASE-1 and ASE-2 packages containing the flasks are unloaded from the plane and transported to the National Institute for Environmental Studies (NIES) in Tsukuba, Japan, where trace gas analyses are performed within 2 days. The analysis port on each of the ASE packages is connected to an automated analytical system that introduces samples from the flasks to various analyzers for trace gas analyses. Because the solenoid valves on both ends of each flask are controlled by a PC of the analytical system, the risk of contamination and leakage during the coupling and decoupling of the sample flasks is minimized.

An air sample from each flask is introduced to four analyzers of the analytical system for simultaneous measurements of CO2, methane (CH4), carbon monoxide (CO), molecular hydrogen (H2), nitrous oxide (N2O), and sulfur hexafluoride (SF6). Prior to the analyses, the sample air is dried by passing it through a glass trap situated in a bath cooled to −80°C. The CO2 mixing ratio is analyzed by a nondispersive infrared gas analyzer (NDIR; LI-COR, either LI-6252 or LI-6262), while other trace gases are analyzed by gas chromatographs with different detectors. A gas chromatograph equipped with a flame ionization detector (GC-FID; Agilent Technologies, HP-5890) is used for CH4; the same gas chromatograph with a reduction gas detector (Trace Analytical, RGD-2) is used for analyzing CO and H2. A gas chromatograph with an electron capture detector (GC-ECD; Agilent Technologies, HP-6890) is used for N2O and SF6. Signals from the NDIR and the gas chromatographs are collected on the same PC. Analytical precision for repetitive measurements is less than 0.03 ppm for CO2, 1.7 ppb for CH4, 0.3 ppb for CO, 3.1 ppb for H2, 0.3 ppb for N2O, and 0.3 ppt for SF6.

All measurements are referenced to four multicomponent working standard gases in synthetic air. The working standards are regularly calibrated by the primary standard gases at NIES to ensure there is no significant drift. All mixing ratios in this paper are reported in parts per million (ppm), parts per billion (ppb), or parts per trillion (ppt) by mole fraction in dry air traceable to the NIES primary standards, which were prepared by a gravimetric method. After the analyses of the trace gas concentrations, isotopic analyses are made for carbon (δ13C) and oxygen (δ18O) of CO2 at NIES, and carbon (δ13C) and hydrogen (δD) of CH4 at Tohoku University.

c. Storage test for the ASE flasks

The internal surface of each flask is treated by electrochemical buffing to minimize the drift in trace gas mixing ratios. In addition, flasks are conditioned by repeatedly evacuation and flushing with highly purified air before use. To evaluate the magnitude of the concentration drift of each trace gas in the flask, storage experiments using dry air from a high pressure cylinder were performed. Figure 4 shows changes in the CO2, H2, and CO concentrations in three flasks over a 2-day period (because our analyses have been performed within 2 days of sampling air). No significant changes in the CO2 and H2 mixing ratios were found in all three flasks, although CO showed a small increase of +1.3 ± 0.2 ppb over 2 days. It is interesting to note that a similar drift in the CO concentration in the previous ASE instrument was also reported by Matsueda et al. (1998). Thus, all CO measurements have been corrected for the storage effect. We found no significant drifts for CH4 and N2O in the ASE flasks within our analytical precision.

Fig. 4.

Changes in (a) CO2, (b) H2, and (c) CO concentrations in three air sample flasks during a 2-day storage experiment. The experiment for each flask is made 3 times, and the error bars represent analytical errors. Details of the experimental procedure are described in the text.

Fig. 4.

Changes in (a) CO2, (b) H2, and (c) CO concentrations in three air sample flasks during a 2-day storage experiment. The experiment for each flask is made 3 times, and the error bars represent analytical errors. Details of the experimental procedure are described in the text.

4. Continuous CME

a. Instrumentation description

The basic design of the CME instrument for continuous CO2 measurement is based on the measuring system that has been used on a research aircraft and unmanned aerial vehicle (Machida et al. 2003; Watai et al. 2006), but improved with new operating features. One of the new features is an automated operation system that can be controlled based on real-time monitoring of the flight navigation data from the aircraft. Another new feature is an automated regular calibration on board the aircraft using two standard gases pressurized into high pressure cylinders. With these automated new features, CME can be left on its own to operate for 1–2 months until the pressure in the cylinders decreases to 2 MPa.

Figure 5 shows a schematic diagram of the CME package (264 mm length × 330 mm width × 570 mm height) weighing about 25 kg. Figure 6 shows 3D views of the instrument package. The main components of the CME instrument consist of a dryer, pump, solenoid valves, flow controller, pressure controller, NDIR, two standard gas cylinders, and a programmable controller/datalogger device. Air is drawn by a diaphragm pump (KNF, NMP 830 KNDC) from the intake port mounted on the air-conditioning duct and then introduced into the analyzer. In front of the pump, a dryer tube packed with 50 g of CO2-saturated magnesium perchlorate is used to remove water vapor before the air sample is analyzed for CO2 by the NDIR (LI-COR, LI-840). The loss of CO2 by passing the air sample through the magnesium perchlorate is examined to be negligible by laboratory experiments. The flow rate of air into the NDIR cell is kept constant at 150 standard cc per minute (sccm) by a mass flow controller (HORIBA STEC, SEC-E40). In addition, the absolute pressure of the sample air in the NDIR cell is maintained at 0.110 MPa by using an auto pressure controller (HORIBA STEC, UR-7340) to avoid a signal drift of the NDIR associated with changes in the cabin pressure. Because instability of the flow and pressure in the NDIR cell has large effects on the precision of the CO2 measurements, they are monitored carefully and recorded on the datalogger, to be used later for quality assurance of the data.

Fig. 5.

Schematic of airflow for the CME. The main components are a dryer, pump, solenoid valves, two standard gas cylinders, and NDIR. Details of the measuring procedure are described in the text.

Fig. 5.

Schematic of airflow for the CME. The main components are a dryer, pump, solenoid valves, two standard gas cylinders, and NDIR. Details of the measuring procedure are described in the text.

Fig. 6.

3D views from the side containing the (left) electrical parts and (right) containing the bottles for CME.

Fig. 6.

3D views from the side containing the (left) electrical parts and (right) containing the bottles for CME.

The switch from air sample to standard gases at an interval of 10 min during ascent and descent, and 20 min during level flight, is accomplished by solenoid valves. Two standard gases (CO2-in-air mixture) of about 340 and 390 ppm are pressurized to about 12 MPa in 2.3-L high pressure aluminum cylinders (Luxfer, NO11). The cylinders and their valves comply with the specifications of the Compressed Gas Association, Inc., of the U.S. Department of Transport (CGA/DOT). They also meet the standard set by the High Pressure Gas Safety Institute of Japan [Koatsu Gasu Hoan Kyokai (KHK)] for pressurization of CO2-in-air cylinders in compliance with the High Pressure Gas Safety Law of Ministry of Economy, Trade and Industry (METI) in Japan. The cylinders are tested to ensure that there is no long-term significant CO2 drift. A small high pressure regulator (NISSAN TANAKA, S1-00G) connected to each cylinder is specially developed to fit a U.S. connector type. After 1–2 months, the cylinders are recalibrated at NIES for any drift.

The automated operation of CME is performed by a programmable controller/datalogger device (Campbell Scientific, Inc., CR10X-XT). This device controls the pump, the solenoid valves for the flow selection, and the NDIR using information from the flight data system on the aircraft. To obtain the required flight parameters, an ARINC 429 Serial Port Adapter interface on CME is connected to three channels of the ARINC buses of the aircraft data system. The parameters received by the controller/datalogger device are date, time, pressure altitude, radio altitude, latitude, longitude, ground speed, static air temperature, wind direction, and velocity. Whenever electric power is supplied from the aircraft, the controller/datalogger device records every 10 min the cabin pressure (measured by the pressure sensor 2), as well as the temperature of the cabin air, the NDIR cell, the pump and the inside air of the CME frame. It also switches on the electric power to warm up the NDIR during taxiing before takeoff. Air sample analysis is not performed within 1200 ft from the ground surface measured by radio altimeter to avoid analysis of polluted air. The CME operation is terminated when the ground speed decreases to less than 50 kt after touchdown. Thus, ground speed and radio altitude from the aircraft are key parameters for the automated CME operation. Information regarding flow rate and cell pressure, as well as data from the NDIR, pressure sensor 1, and the aircraft data system, is recorded every 10 s (corresponding to about 100 m in altitude) during ascent and descent, and every 1 min (corresponding to about 10–20 km in horizontal distance) at cruising altitude, by the programmable controller/datalogger device with a memory capacity to last more than 2 months.

As is the case with ASE, redundant safety features are built into the CME operation, to ensure noninterference with the safe operation of the aircraft. The CME instrument is programmed to shut down when air pressure increases beyond 25 psig and/or when temperature exceeds 75°C. In particular, the temperature sensors attached to the pump, the NDIR analyzer, and the controller/datalogger device will initiate the shutdown of the electrical power supply if overheating of the instrument is detected. Even if the sensors do not work, the CME operation is terminated by the relief valves and the thermal fuses if pressure greater than 90 psig or temperature greater than 100°C, respectively, is detected.

b. CME performance test

To evaluate the analytical precision of CME, a standard gas of CO2 in air was used as an air sample that was pumped through the dryer, and its mixing ratio was determined by using two standard gas cylinders in CME. Before the experiment, the CO2 contents in these three standards were precisely calibrated by the NIES scale. Figure 7 shows an example of CO2 variation from CME for a period of about an hour. More than 90% of the 10-s-averaged data are within ±0.2 ppm from the overall average. This result indicates a good stability of the sample flow and pressure in the NDIR cell, because the measurement variation primarily is due to the electric noise in the NDIR. The 1-h-averaged CO2 concentration was determined to be 368.87 ppm, which agreed well with the assigned value for the standard gas of 368.86 ppm. This agreement indicates no significant change in the CO2 concentration by passing the gas through the dryer, pump, and solenoid valves in CME. There is also a good linearity in the NDIR response. The linearity of the NDIR response is confirmed by further experiments using five standard gases in a range between 349 and 393 ppm, with the deviations less than 0.12 ppm. This has allowed us to use only two standard gas cylinders in CME.

Fig. 7.

Time variations of CO2 of a standard gas measured by CME for about 10 h. Detailed experimental procedures are described in the text.

Fig. 7.

Time variations of CO2 of a standard gas measured by CME for about 10 h. Detailed experimental procedures are described in the text.

To check the reliability of the CME instrument in measuring real air samples, natural air was taken from outside the NIES building and its CO2 mixing ratio was measured by CME and a high-performance measuring system using a different NDIR (LI-COR, LI-6262) with a higher precision of ±0.2 ppm in 1-s data that were used in previous aircraft observation campaigns (Machida et al. 2003). Figure 8 shows the comparison of atmospheric CO2 variations measured by CME and the LI-6262 system. The agreement between the two systems is relatively good, giving confidence in the CME capability to measure atmospheric CO2 relatively accurately. Because of the low flow rate, the response time of CME is slower than the LI-6262 system; in the figure, data from CME have been shifted forward by 50 s. The smaller error bars, representing one standard deviation of 1-s data, associated with CME are due to the slower response of the instrument, as compared to the LI-6262 system.

Fig. 8.

A comparison of atmospheric CO2 variations measured by CME (open squares and thick line) and LI-6262 system (open circles and thin line). Error bars represent standard deviations of 1-s data. Details of the experimental procedures are described in the text.

Fig. 8.

A comparison of atmospheric CO2 variations measured by CME (open squares and thick line) and LI-6262 system (open circles and thin line). Error bars represent standard deviations of 1-s data. Details of the experimental procedures are described in the text.

5. Approval to install ASE and CME in passenger airliners

After various tests following the procedures for airborne equipment specified in the Radio Technical Commission for Aeronautics (RTCA) document (RTCA/DO-160D), a supplemental-type certificate (STC) to install ASE and CME on the Boeing 747-400 aircraft series was issued from the Federal Aviation Administration (FAA) on 26 October 2005. A similar STC document was issued from the Japanese Civil Aviation Bureau (JCAB) on 2 November 2005. Thus, the ASE–CME instrument package can be installed on any Boeing 747-400 aircraft operated by other airlines with only minor adjustments/changes. A certificate to install CME on the Boeing 777-200 series was obtained from the FAA on 22 March 2006, followed by JCAB approval on 31 March 2006.

6. Evaluation of preliminary observational results

a. Comparison between ASE and CME

The CME instrument is installed on all five planes (two Boeing 747-400 and three Boeing 747-200ER) operated by JAL, but ASE is installed only on the Boeing 747-400 planes that fly twice a month between Sydney, Australia, and Narita to perform flask measurements of CO2 and other trace gases, adding to the already existing CO2 record from the first phase of the JAL project (Matsueda et al. 2002). This route is of special interest because of its flight over the tropical region where measurements are scarce. Figure 9 shows an example of ASE CO2 measurements obtained during the JAL flight (Boeing 747-400, JA8921) from Sydney to Narita on 22 April 2006. Data obtained by CME for the same flight are also plotted in Fig. 9. There is a good agreement between ASE- and CME-measured CO2 values; this level of agreement was also observed for data obtained from another JAL flight (Boeing 747-400, JA8917). Data from all the flights show agreement to within ±0.2 ppm, which is close to the analytical uncertainty of the CME system. The data agreement indicates a high reliability in CO2 measurements by CME, in spite of the fact that we use a small NDIR, a small pump with a lower flow rate, and magnesium perchlorate in CME. As can be seen in Fig. 9, the flask measurements by ASE can provide a gross latitudinal distribution of CO2 but miss fine spatial structural variations identified by CME. Some of these significant structural variations detected by CME are discussed below. More detailed discussions about the comparison between ASE and CME were described in another paper (Matsueda et al. 2008).

Fig. 9.

Latitudinal distributions of CO2 measured by ASE (open squares) and CME (black dots) during the JAL flight from Sydney to Narita on 22 Apr 2006. Thin line represents the flight altitude.

Fig. 9.

Latitudinal distributions of CO2 measured by ASE (open squares) and CME (black dots) during the JAL flight from Sydney to Narita on 22 Apr 2006. Thin line represents the flight altitude.

b. Frequent CME observations

Figure 10 shows CME CO2 latitudinal distributions at about 10-km altitude obtained from eight flights between Jakarta, Indonesia, and Narita during 6–10 November 2005. All flights show similar CO2 distributions to within ±0.5 ppm. All of the latitudinal profiles show a clear drop of 1–2 ppm in the CO2 concentration from the equator to around 2°N. The location over the western Pacific coincides with the position of the intertropical convergence zone (ITCZ) as indicated by the outgoing longwave radiation (OLR) data provided by NOAA/Office of Oceanic and Atmospheric Research (OAR)/ESRL (online at http://www.cdc.noaa.gov/cdc/data.interp_OLR.html) for this time period. This ITCZ-related CO2 distribution is similar to the one obtained during the Biomass Burning and Lightning Experiment Phase A (BIBLE A) in the same season in 1998 (Machida et al. 2003), but with an overall increase of about 13 ppm in the CO2 level from 1998 to 2005. This increase is consistent with the global CO2 increase observed during the first phase of the JAL project (Matsueda et al. 2002).

Fig. 10.

Latitudinal distributions of CO2 taken by CME at an altitude of about 10 km on eight different flights between Jakarta and Narita during 6–10 Nov 2005.

Fig. 10.

Latitudinal distributions of CO2 taken by CME at an altitude of about 10 km on eight different flights between Jakarta and Narita during 6–10 Nov 2005.

Another interesting feature that is quite evident in Fig. 10 is a sharp increase in CO2 observed around 12°N during a round-trip between Narita and Jakarta on 7 November 2005. This feature has been simulated by a 3D transport model, indicating a rapid intrusion of enriched CO2 air masses from the boundary layer to the upper troposphere near the Philippines (T. Maki 2007, personal communication). A similar synoptic-scale process resulted in CO2 increases around 15° and 32°N on 6 November 2005. These results attest to the value of in situ continuous measurements by CME. Finally, we would like to note that large gradients in CO2 near the tropopause have been detected by CME during flights to Europe and North America. Detailed analyses of these and other measurements by CME will be published separately.

7. Conclusions and outlook

With cooperation and encouragement from Japan Airlines, we have been actively carrying out systematic measurements of CO2 and other trace gases in the atmosphere using commercial aircraft since 1993. During the first phase of the JAL project (1993–2005), an automated flask sampling system was used to obtain measurements of CO2 and other trace gases over the western Pacific between Japan and Australia. In the second phase we have significantly expanded the program to cover additional regions from Japan to Europe, Asia, Hawaii, and North America, using five JAL aircraft. We have made substantial improvements in the flask sampling system (ASE) and developed a new instrument (CME) to obtain continuous in situ measurements of CO2. The CME and ASE instrument packages have been certified by the FAA and JCAB to operate on the Boeing 747-400 series, while CME has been certified to operate on the Boeing 777-200 series as well. Thus, these instruments can be used by any airline that employs any of these two Boeing series aircraft, providing a very powerful observational platform for making precise and accurate measurements of trace gases deep in the atmosphere on a global scale.

Acknowledgments

We would like to acknowledge Tamiki Suenaga of JAL Foundation for coordinating the developing project of ASE and CME. We are grateful to many engineers of the Japan Airlines, JAMCO Tokyo, and JAMCO America for developing, testing, and installing the ASE and CME instruments. We also thank Akihiro Kudo of JANS, Co., Ltd., for his skillful improvement in the design and manufacturing of ASE. We would like to acknowledge Kaz Higuchi for his comments on the manuscript. Three anonymous reviewers made valuable comments that helped to improve the manuscript. This work is financially supported in Japan by the Special Coordination Funds for Promoting Science and Technology of the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) and Global Environment Research Coordination System of the Ministry of Environment (MOE).

REFERENCES

REFERENCES
Baker
,
D. F.
, and
Coauthors
,
2006
:
TransCom 3 inversion intercomparison: Impact of transport model errors on the interannual variability of regional CO2 fluxes, 1988–2003.
Global Biogeochem. Cycles
,
20
.
GB1002, doi:10.1029/2004GB002439
.
Bolin
,
B.
, and
W.
Bischof
,
1970
:
Variations of the carbon dioxide content of the atmosphere in the Northern Hemisphere.
Tellus
,
22
,
431
442
.
Brenninkmeijer
,
C. A. M.
, and
Coauthors
,
1999
:
CARIBIC—Civil aircraft for global measurement of trace gases and aerosols in the tropopause region.
J. Atmos. Oceanic Technol.
,
16
,
1373
1383
.
Brenninkmeijer
,
C. A. M.
, and
Coauthors
,
2007
:
Civil aircraft for the regular investigation of the atmosphere based on an instrumented container: The new CARIBIC system.
Atmos. Chem. Phys.
,
7
,
4953
4976
.
Brunner
,
D.
,
J.
Staehelin
,
D.
Jeker
,
H.
Wernli
, and
U.
Schumann
,
2001
:
Nitrogen oxides and ozone in the tropopause region of the Northern Hemisphere: Measurements from commercial aircraft in 1995/1996 and 1997.
J. Geophys. Res.
,
106
,
27673
27699
.
Chédin
,
A.
,
S.
Serrar
,
N. A.
Scott
,
C.
Crevoisier
, and
R.
Armante
,
2003
:
First global measurement of midtropospheric CO2 from NOAA polar satellites: Tropical zone.
J. Geophys. Res.
,
108
.
4581, doi:10.1029/2003JD003439
.
Engelen
,
R. J.
, and
A. P.
McNally
,
2005
:
Estimating atmospheric CO2 from advanced infrared satellite radiances within an operational four-dimensional variational (4D-Var) data assimilation system: Results and validation.
J. Geophys. Res.
,
110
.
D18305, doi:10.1029/2005JD005982
.
Forster
,
P.
, and
Coauthors
,
2007
:
Changes in atmospheric constituents and in radiative forcing.
Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 129–234
.
Gerbig
,
C.
,
J. C.
Lin
,
S. C.
Wofsy
,
B. C.
Daube
,
A. E.
Andrews
,
B. B.
Stephens
,
P. S.
Bakwin
, and
C. A.
Grainger
,
2003
:
Toward constraining regional-scale fluxes of CO2 with atmospheric observations over a continent: 1. Observed spatial variability from airborne platforms.
J. Geophys. Res.
,
108
.
4756, doi:10.1029/2002JD003018
.
Gurney
,
K. R.
, and
Coauthors
,
2004
:
TransCom 3 inversion intercomparison: Model mean results for the estimation of seasonal carbon sources and sinks.
Global Biogeochem. Cycles
,
18
.
GB1010, doi:10.1029/2003GB002111
.
Machida
,
T.
,
K.
Kita
,
Y.
Kondo
,
D.
Blake
,
S.
Kawakami
,
G.
Inoue
, and
T.
Ogawa
,
2003
:
Vertical and meridional distributions of the atmospheric CO2 mixing ratio between northern midlatitudes and southern subtropics.
J. Geophys. Res.
,
108
.
8401, doi:10.1029/2001JD000910
.
Maksyutov
,
S.
,
T.
Machida
,
H.
Mukai
,
P. K.
Patra
,
T.
Nakazawa
, and
G.
Inoue
, and
TransCom-3 Modelers
,
2003
:
Effect of recent observations on Asian CO2 flux estimates by transport model inversions.
Tellus
,
55B
,
522
529
.
Marenco
,
A.
, and
Coauthors
,
1998
:
Measurement of ozone and water vapor by Airbus in-service aircraft: The MOZAIC airborne program, an overview.
J. Geophys. Res.
,
103
,
25631
25642
.
Matsueda
,
H.
, and
H. Y.
Inoue
,
1996
:
Measurements of atmospheric CO2 and CH4 using a commercial airliner from 1993 to 1994.
Atmos. Environ.
,
30
,
1647
1655
.
Matsueda
,
H.
,
H. Y.
Inoue
,
Y.
Sawa
,
Y.
Tsutsumi
, and
M.
Ishii
,
1998
:
Carbon monoxide in the upper troposphere over the western Pacific between 1993 and 1996.
J. Geophys. Res.
,
103
,
19093
19110
.
Matsueda
,
H.
,
H. Y.
Inoue
, and
M.
Ishii
,
2002
:
Aircraft observation of carbon dioxide at 8-13 km altitude over the western Pacific from 1993 to 1999.
Tellus
,
54B
,
1
21
.
Matsueda
,
H.
,
T.
Machida
,
Y.
Sawa
,
Y.
Nakagawa
,
K.
Hirotani
,
H.
Ikeda
,
N.
Kondo
, and
K.
Goto
,
2008
:
Evaluation of atmospheric CO2 measurements from new flask air sampling of JAL airliner observations.
Pap. Meteor. Geophys.
,
59
,
1
17
.
Nakazawa
,
T.
,
K.
Miyashita
,
S.
Aoki
, and
M.
Tanaka
,
1991
:
Temporal and spatial variations of upper tropospheric and lower stratospheric carbon dioxide.
Tellus
,
43B
,
106
117
.
Nakazawa
,
T.
,
S.
Morimoto
,
S.
Aoki
, and
M.
Tanaka
,
1993
:
Time and space variations of the carbon isotopic ratio of tropospheric carbon dioxide over Japan.
Tellus
,
45B
,
258
274
.
Pearman
,
G. I.
, and
D. J.
Beardsmore
,
1984
:
Atmospheric carbon dioxide measurements in the Australian region: Ten years of aircraft data.
Tellus
,
36B
,
1
24
.
Peylin
,
P.
, and
Coauthors
,
2007
:
Evaluation of Television Infrared Observation Satellite (TIROS-N) Operational Vertical Sounder (TOVS) spaceborne CO2 estimates using model simulations and aircraft data.
J. Geophys. Res.
,
112
.
D09313, doi:10.1029/2005JD007018
.
Sawa
,
Y.
, and
Coauthors
,
2004
:
Aircraft observation of CO2, CO, O3 and H2 over the North Pacific during the PACE-7 campaign.
Tellus
,
56B
,
2
20
.
Shia
,
R-L.
,
M-C.
Liang
,
C. E.
Miller
, and
Y. L.
Yung
,
2006
:
CO2 in the upper troposphere: Influence of stratosphere-troposphere exchange.
Geophys. Res. Lett.
,
33
.
L14814, doi:10.1029/2006GL026141
.
Stephens
,
B. B.
, and
Coauthors
,
2007
:
Weak northern and strong tropical land carbon uptake from vertical profiles of atmospheric CO2.
Science
,
316
,
1732
1735
.
doi:10.1126/science.1137004
.
Taguchi
,
S.
,
H.
Matsueda
,
H. Y.
Inoue
, and
Y.
Sawa
,
2002
:
Long-range transport of carbon monoxide from tropical ground to upper troposphere: A case study for Southeast Asia in October 1997.
Tellus
,
54B
,
22
40
.
Vay
,
S. A.
, and
Coauthors
,
2003
:
Influence of regional-scale anthropogenic emissions on CO2 distributions over the western North Pacific.
J. Geophys. Res.
,
108
.
8801, doi:10.1029/2002JD003094
.
Watai
,
T.
,
T.
Machida
,
N.
Ishizaki
, and
G.
Inoue
,
2006
:
A lightweight observation system for atmospheric carbon dioxide concentration using a small unmanned aerial vehicle.
J. Atmos. Oceanic Technol.
,
23
,
700
710
.
WMO
,
2006
:
The state of greenhouse gases in the atmosphere using global observations through 2005.
WMO Greenhouse Gas Bull.
,
2
,
1
4
.

Footnotes

Corresponding author address: Toshinobu Machida, National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba 305-8506, Japan. Email: tmachida@nies.go.jp