Скачать презентацию An Autonomous Inexpensive Robust CO 2 Analyzer AIRCOA Скачать презентацию An Autonomous Inexpensive Robust CO 2 Analyzer AIRCOA

a8bbc59aad5f860d255df8aea6d973a3.ppt

  • Количество слайдов: 33

An Autonomous Inexpensive Robust CO 2 Analyzer (AIRCOA) Britton Stephens, Andrew Watt, and Gordon An Autonomous Inexpensive Robust CO 2 Analyzer (AIRCOA) Britton Stephens, Andrew Watt, and Gordon Maclean National Center for Atmospheric Research, Boulder, Colorado, USA

Using high frequency data makes signals bigger, but the annual-mean signals are still very Using high frequency data makes signals bigger, but the annual-mean signals are still very small: To measure 0. 2 Gt. Cyr-1 source/sink to +/- 25% need to measure regional annual mean gradients to 0. 1 -0. 2 ppm Flux footprint, in ppm(Gt. Cyr-1)-1, for a 106 km 2 chaparral region in the U. S. Southwest (Gloor et al. , 1999).

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

NCAR CO 2 and O 2/N 2 Calibration Facility NCAR CO 2 and O 2/N 2 Calibration Facility

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

CO 2 signal averaged over 2. 5 min. measurement cycle CO 2 signal averaged over 2. 5 min. measurement cycle

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Calibration sequence Calibration sequence

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Empirical pressure correction Empirical pressure correction

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Empirical temperature correction SPL 9/4 -9/18 NWR 9/18 Empirical temperature correction SPL 9/4 -9/18 NWR 9/18

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Drying system monitoring A change of 0. 5% RH is approximately 300 ppm H Drying system monitoring A change of 0. 5% RH is approximately 300 ppm H 2 O, which would cause a dilution error of 0. 1 ppm in CO 2

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Nafion absorption effect Flow pulled through Nafion went from 300 to 50 sccm at Nafion absorption effect Flow pulled through Nafion went from 300 to 50 sccm at t = 30 sec

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Empirical flushing correction Empirical flushing correction

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Automated (4 - or 8 -hourly) leak checks A positive trend of 0. 3 Automated (4 - or 8 -hourly) leak checks A positive trend of 0. 3 k. Pa/min would be a leak rate of 0. 1 sccm which if 100 ppm different would cause a 0. 1 ppm bias

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Regulator oven tests Three cylinders were in the oven and one (green dots) was Regulator oven tests Three cylinders were in the oven and one (green dots) was not

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Regulator flushing tests Regulator flushing tests

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Laboratory intercomparisons Field surveillance tanks Laboratory offsets less than 0. 05 ppm (1 -sigma Laboratory intercomparisons Field surveillance tanks Laboratory offsets less than 0. 05 ppm (1 -sigma = 0. 13 ppm) 2. 5 -month average field differences from assigned values 0. 01 to 0. 10 ppm (1 -sigma = 0. 10 to 0. 13 ppm)

Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO Potential source of bias AIRCOA solution Relating to WMO CO 2 Scale Dedicated CO 2 and O 2 calibration transfer facility Short-term IRGA noise Average for 2 minutes to get better than 0. 1 ppm precision Drift in IRGA sensitivity 4 -hourly 4 -point calibrations and 30 -minute 1 -point calibrations IRGA pressure sensitivity Automated 4 -hourly pressure sensitivity measurements IRGA temperature sensitivity 30 -minute 1 -point calibrations, temperature control at some sites Incomplete drying of air Slow enough flow (100 sccm), two 96” Nafion driers, downstream humidity sensor to verify performance Drying system altering CO 2 Continuous flows and pressures through Nafions and run surveillance gas through entire system Incomplete flushing of cell and dead volumes Fast enough flow (100 sccm), alternate calibration sequence low‑tohigh / high-to-low to look for effects Leaks through fittings, solenoid valves, and pumps Automated 8 -hourly positive pressure leak-down and 4 -hourly ambient pressure leak-up checks Pressure broadening without Ar Use calibration gases made with real air Fossil CO 2 in calibration gases and different field and lab 13 C sensitivities Laboratory tests limit current effect to 0. 05 ppm, long-term plans to use cylinders with natural CO 2 Regulator temperature effects Tests suggest effect is negligible, but could be regulator dependent Regulator flushing effects Repeat calibration tests suggest the effect is negligible Whole-system diagnostics and comparability verification Long-term surveillance tank analyzed every 8 hours, co-location with other programs, rotating cylinders, and laboratory comparisons Delay in diagnosis of errors Near real-time data acquisition, processing, and dissemination

Automated web-based output http: //www. eol. ucar. edu/~stephens/RACCOON Automated web-based output http: //www. eol. ucar. edu/~stephens/RACCOON

Power Budget Component Measured Amps @ 12 VDC Watts Purge pump 0. 61 7. Power Budget Component Measured Amps @ 12 VDC Watts Purge pump 0. 61 7. 3 Sample pump 0. 09 1. 1 Computer (incl. PS/DAQ/Relays) 0. 42 5. 0 RH/T sensor: 0. 0014 <0. 02 7 mass-flowmeters < 0. 001 <0. 01 2 solenoid valves 0. 19 2. 3 Li 820 0. 35 4. 2 Total 1. 67 20. 0 Measured total 1. 7 20. 4 Measured AC Power 0. 24 Amps at 120 V / 60 Hz ~28 W