The BOREAS Information System
1.1 Data Set Identification
BOREAS TGB9-IFC1: above canopy NMHCs at SSA-OBS site.
1.2 Data Set Introduction
BOREAS TGB09 non-methane hydrocarbon (NMHC) data from the SSA-OBS site.
1.3 Objective/Purpose
To provide a quantitative inventory of non-methane hydrocarbons, both anthropogenic and biogenic, at the SSA-OBS site. To provide ambient concentration data for biogenic hydrocarbons suitable for calculating the flux of biogenic hydrocarbons at the SSA-OBS site using the gradient method.
1.4 Summary of Parameters
Concentrations in the parts per trillion by volume (pptv) - parts per billion by volume (ppbv) range are reported for a variety of biogenic and anthropogenic non methane hydrocarbons. Local sources and transport from distant sources contribute to the non-methane hydrocarbon inventory in the boreal forest region. A quantitative inventory of ambient hydrocarbon concentrations can aid in identification of sources important to the boreal atmosphere. In particular, gradient measurements allow biogenic emissions from the forest itself to be quantified.
1.5 Discussion
Whole air samples were simultaneously collected from two heights on the flux tower for several 30 minute periods during IFC1. These sampling periods were distributed between 8:00 am and 8:00 PM on five days in late May - early June using gas chromatographic (GC) techniques. Permissible injection volumes for capillary GC are below 1 cc (gas). Typical concentrations for the hydrocarbon species that we measure are in the range of 10 pptv to 10 ppbv, which, for isoprene as example, gives an analyte range of 0.06 to 60 fg/cc. The constraint on injection volume for capillary GC combined with the low concentrations of analyte require that the samples be preconcentrated prior to GC analysis. Sample preconcentration is accomplished by passing a known volume of the air sample through a trap filled with fine glass beads that is cooled to -180C. With this technique, the volatile hydrocarbons of interest are quantitatively retained in the trap, whereas the bulk constituents of air (nitrogen, oxygen, etc.) are not. The nitrogen and oxygen are collected in a vessel of known volume, the reference volume. From the pressure in the reference volume, the total volume of air which passed through the preconcentration system is known. This volume is used to calculate the mixing ratio of each compound in the original air sample after Gas Chromatographic (Flame Ionization Detector, GC-FID) analysis. The sample trapped cryogenically on the glass beads is thermally desorbed into a stream of ultra-pure helium and re-trapped on the surface of a fine stainless steel capillary cooled to -180C. This second cryogenic trapping stage "focuses" the sample into a small linear section of tubing. The cold stainless steel capillary is ballistically heated (by electrical resistance) and the focused sample quickly desorbs into the helium stream and is transferred to the chromatographic column.
The volatile components of the sample are carried through the GC column by the mobile phase, the ultra-pure helium. Capillary GC columns are typically a long (10 - 100m) section of fused silica capillary (0.18-0.53mm inner diameter, I.D.). The inside surface of the column has a thin (0.05-5m) coating of a stationary phase designed to interact with the components of a mixture passing through the column. For any individual hydrocarbon, the amount of time taken to traverse the length of the column (retention time) is determined by the component's affinity for the stationary phase, which, in turn, is a function of the chemical nature of the component. This differential affinity of various species for the stationary phase makes it possible for the GC column to separate a complex mixture as it travels through the column. Controlled variation of column temperature over the course of the analysis (temperature program) also affects the separation of the compounds, the retention time for each compound, and the total analysis time for the sample. Compounds are identified by their characteristic retention times.
Compounds eluting from the end of the column are detected using a flame ionization detector (FID), in which the column effluent is introduced into a hydrogen-air diffusion flame. Electrically charged species formed through radical reactions in the flame are collected by applying a potential across the flame and the current is amplified by a sensitive electrometer. The FID displays a linear response to carbon atoms over a large dynamic range. The FID is very sensitive to hydrocarbons, but is insensitive to water or carbon dioxide.
In addition to the FID detector, some of our samples are analyzed using a mass selective detector (MSD). The MSD produces mass spectra of the species eluted from the column. The MSD is less sensitive than the FID, but provides a more certain identification of the species eluted by relying not only on the retention time, but also on the mass spectrum of the effluent. As compounds exit the column, they are hit by the high energy stream of electrons and a fraction of the molecules are converted to ions. Some molecules are fragmented by the electron impact, and produce ions with a mass less than the original molecule. The ions are electrically accelerated into a quadrupole mass filter, which allows only ions within a narrow range of mass-charge ratio to pass to an electron multiplier. The electron multiplier counts the number of ions which strike it. By scanning a range of mass-charge ratios, the MSD can construct a mass spectrum for each compound exiting the column. The pattern and abundance of ions produced is characteristic of the chemical structure of a compound, and thus may be used to identify species eluting from the column. In general, we use the MSD results to qualitatively identify unknown compounds present in the samples, while the FID results provide quantitative data. Unlike the FID, the MSD is sensitive to water and carbon dioxide.
Water and carbon dioxide are present in ambient air in amounts several orders of magnitude larger than any of the non-methane hydrocarbons. Both water and carbon dioxide are quantitatively trapped in the preconcentration step and although small amounts will not affect FID operation, typical ambient concentrations do pose chromatographic problems, and problems with the MSD operation. Water and carbon dioxide have a detrimental effect on the chromatographic separation of hydrocarbons. Their presence in the preconcentrated sample greatly increases its volume and prevents the hydrocarbons from being tightly focused on the head of the column. Because of its polar nature, water interferes with the interaction between the hydrocarbons and the stationary phase of the column. Water can also clog the glass bead trap, cryo-focuser, or column during the analysis, and may extinguish the FID flame. In the MSD, high water and carbon dioxide signals increase the background signals for other ions such that the detection limit is raised to an unsuitable level. In order to avoid these problems, water and carbon dioxide are removed from the air samples prior to preconcentration. A cold trap (-20 to -60 _C) removes water in sufficient amounts to allow chromatographic analysis to proceed without any clogging or FID quenching. A potassium carbonate trap at 80 _C removes carbon dioxide. We have tested these traps extensively to assure that the concentrations of the hydrocarbons of interest are not affected.
Each sample was analyzed on two separate GC-FID systems. In one, the column and temperature program were chosen to optimize quantitation of C2 - C6 hydrocarbons. In the other, the column and temperature program were chosen to optimize quantitation of C5 - C10 hydrocarbons.
1.6 Related Data Sets
TGB10, TF02
2.1 Investigator(s) Name and Title
Professor Hiromi Niki
2.2 Title of Investigation
Ambient Measurements of Non-Methane Hydrocarbons
2.3 Contact Information
Professor Hiromi Niki or Dr. Valerie Young Centre for Atmospheric Chemistry 006A Steacie Science Library York University North York, ON M3J 1P3 Canada (416) 736-5410 email: valy@bobcat.ent.ohiou.edu FAX: ?????????????
Estimates of the hydrocarbon flux above the canopy by the gradient method require measurement of the hydrocarbon concentration gradient. To accomplish this, simultaneous samples were collected at two heights above the canopy and analyzed for a range of hydrocarbons. The flux of any particular hydrocarbon (HC) can then be determined by the following relationship :
FLUX(HC) = k(z) * ([HC]" - [HC]') / (z" - z')
where ([HC]" - [HC]') is the difference in [HC] between the two heights z" and z'. The eddy diffusivity coefficient k(z) should be obtained from other researchers performing simultaneous measurements at the same sites. Measurements of the hydrocarbon concentrations are made by collecting whole air samples over a 30 minute period in evacuated, electropolished stainless steel canisters which are then transported to our laboratory for analysis. Before being shipped to Saskatchewan for sample collection, the canisters were heated to 80C and evacuated to 10 e-6 Torr for two hours.
4.1 Sensor/Instrument Description
A. The sample collection system filled two sample canisters simultaneously over 30 minutes. The system consists of:
B. Sample analysis system for identification and quantification of hydrocarbons. The system consists of:
4.1.1 Collection Environment
4.1.2 Source/Platform
4.1.3 Source/Platform Mission Objectives
To identify and quantify the non methane hydrocarbons present in the boreal forest, using a sampling strategy that will allow calculation of fluxes.
4.1.4 Key Variables
Concentration of individual non methane hydrocarbons in air.
4.1.5 Principles of Operation
4.1.6 Sensor/Instrument Measurement Geometry
The Teflon sampling tubes are mounted on the main tower at each site. The height of the inlet above the ground, in meters, is given for each sample. Both inlets are above the mean canopy height.
4.1.7 Manufacturer of Sensor/Instrument
A. Sampling system assembled by investigators. All tubing chromatographic grade stainless steel.
B R C / Rasmussen1121 N.W. Donelson StreetHillsboro, OR 97124(503) 690-1077
Tylan CorporationTorrance, CA 90501
KNF Neuberger, Inc.Princeton, NJ(609) 799-4350
B. Analysis system
Hewlett-Packard (Canada)5150 Spectrum Way H80Mississauga, ON L4W 5G1(905) 206-4725
Tekran, Inc.Toronto, ON
4.2 Calibration
A. Sampling system flow rate calibrated using soap bubble flow meter
B. GC-FID systems calibrated with standard mixtures from Environment Canada
(Dr. Daniel Wang, Environment Canada, 351 St. Joseph Blvd., Hull, PQ K1A 0H3) and/or NIST (Dr. Eric Apel, National Center for Atmospheric Research, 1850 Table Mesa Drive, Boulder, CO 80303), and checked with secondary standards mixed in our lab.
Calibration of the GC systems is accomplished using prepared gas mixtures in air of known concentration to determine the characteristic retention time and FID response for each compound in the mixture. Measured FID responses are checked against the expected response based on the number of carbon atoms in each compound, since the FID response is linear over a large range, and is essentially mass responsive to carbon atoms.
4.2.1 Specifications
A. Cans can not be pressurized above 25 psig because a constant flow rate can not be maintained above that pressure.
B. Each calibration run was compared with previous runs. Values obtained did not vary significantly over the course of the project.
4.2.1.1 Tolerance
A. Constant flow rate of 0SYMBOL 45 \f "Symbol"500 sccm air at STP, within SYMBOL 177 \f "Symbol"10%.
B. Detection limits and precision
Stainless steel canisters and Mass flow controllers
Detection limit: 5 ppt V = 0.005 ppbv
Precision:
+/-5pptv for [HC] < 0.1ppbv
+/-5% for 0.1ppbv < [HC] < 1.0ppbv
+/-1% for [HC] > 1.0ppbv
Metal bellows pump
Detection limit: 0.1 ppbv
Precision: not used for quantitative analysis
4.2.2 Frequency of Calibration
A. Weekly
B. Weekly using standard from NIST and/or Environment Canada.
Daily using secondary standard.
4.2.3 Other Calibration Information
A. None.
B. The two samples in each pair were run consecutively, on the same day by the same operator on the same GC. Before the project began, each operator demonstrated that he could achieve the precision quoted in 4.2.1.1 for multiple runs of the same sample. Each day, one sample was repeated by another operator on the same GC to test reproducibility between operators. The concentration of isoprene obtained from the two GCs was compared for each sample.
The two outlets of the pumping system were connected to evacuated stainless steel canisters. The inlets of the pumping system were connected to Teflon lines mounted on a tower so that air could be sampled simultaneously from two heights above the forest canopy. The flow rate for the mass flow controllers was set such that each can would fill to a pressure of about 20 psig during 30 minutes of sampling. The cans were sealed by shutting the valve, then the full cans were shipped to our laboratory for analysis.
Approximately 500 - 700 mL (at standard temperature and pressure) of sample was used for GC-FID analysis. The sample handling procedure is described in more detail in Section 4.
6.1 Data Notes
6.2 Field Notes
SSA-OBS: The mean canopy height was 9 m. Sampling inlets were placed at 12.4 m and 23.3 m above ground level. In order to bring in electrical power, Hydro cut down trees along a path about 50 m wide from the road to the tower. Data taken when the wind was from the direction of the road may be suspect because of this.
The data table contains the date and time sample collection was initiated (GMT), the height above ground at which the sample was collected, and the observed mixing ratio in ppbv of approximately 30 non-methane hydrocarbons. We include all of the biogenic hydrocarbons which we currently quantify, and a number of significant anthropogenic hydrocarbons. Concentrations of other non methane hydrocarbons are available from the principle investigator upon request.
7.1 Spatial Characteristics
7.1.1 Spatial Coverage
Samples were collected only from the tower at each site. Each sample may be considered as an ambient sample whose spatial coverage could be determined by back-trajectory analysis for the 30 minute sampling period.
7.1.2 Spatial Coverage Map
7.1.3 Spatial Resolution
The resolution, for gradient calculations, is no better than the footprint of forest considered in flux measurements. As ambient measurements, the spatial resolution is the same as the spatial coverage.
7.1.4 Projection
7.1.5 Grid Description
7.2 Temporal Characteristics
7.2.1 Temporal Coverage
This data set is for the first intensive field campaign. Data sets exist for the other two intensive field campaigns.
7.2.2 Temporal Coverage Map
7.2.3 Temporal Resolution
On given days, samples were collected at 1-3 hour intervals, usually from late morning to evening, but occasionally for a complete 24 hour period, when there was no precipitation.
7.3 Data Characteristics
7.3.1 Parameter/Variable
DATA_UNIT_ID OBS_DATE CANISTER_NUMBER HEIGHT ISOPRENE QA_CODE FILE_NUM REVISION_DATE CRTFCN_CODE
7.3.2 Variable Description/Definition
DATA_UNIT_ID -
OBS_DATE - DATE THAT THE MEASUREMENT WAS TAKEN
CANISTER_NUMBER - SAMPLING CANISTER NUMBER FOR ISOPRENE DATA
HEIGHT - HEIGHT AT WHICH MEASUREMENTS WERE TAKEN.
ISOPRENE - AMBIENT CONCENTRATION OF ISOPRENE
QA_CODE - QUALITY ASSURANCE CODE.
1=GOOD, 2=MISSING, 3=VALUE OUTSIDE EXPECTED RANGE,
4=UNEXPECTED REVERSE GRADIENT
FILE_NUM - THE FILE NUMBER FOR THE RAW DATA IN TEAMS ARCHIVES
REVISION_DATE -
CRTFCN_CODE -
7.3.3 Unit of Measurement
DATA_UNIT_ID - OBS_DATE - DATE CANISTER_NUMBER - N/A HEIGHT - METERS ISOPRENE - PARTS PER BILLION QA_CODE - N/A FILE_NUM - N/A REVISION_DATE - DATE CRTFCN_CODE - N/A
7.3.4 Data Source
7.3.5 Data Range
7.4 Sample Data Record
8.1 Data Granularity
(BORIS and ORNL DAAC to fill in)
8.2 Data Format(s)
9.1 Formulae
[Sample] = (sample pk area) * ([cal std] / (cal std pk area)) * ((cal std vol) / (sample vol))
Where:
9.1.1 Derivation Techniques and Algorithms
Calculations performed by HP-Chemstation software
9.2 Data Processing Sequence
9.2.1 Processing Steps
9.2.2 Processing Changes
None.
9.3 Calculations
Peak areas converted to concentrations (mixing ratios) by formula in Section 9.1
9.3.1 Special Corrections/Adjustments
Data Below Detection Limit: When the NMHC of interest is not detected by our GC-FID system, it is assigned a concentration of 0.002 ppbv, which is approximately the average between our detection limit and 0.0.
Conversion to GMT: Time at start of sampling was originally noted in local Saskatchewan time. This was converted to GMT by adding 6 hours.
9.3.2 Calculated Variables
9.4 Graphs and Plots
None.
10.1 Sources of Error
The user is advised to check meteorological data carefully before using this data to calculate fluxes by the gradient method, and to know the limitations of this method. Because compound identification is based on residence time, it is possible to misidentify a compound which is present in a real sample but not present in our calibration standard. We have adjusted our temperature programs to achieve good separation for the compounds in our calibration standards, but additional species present in real samples may co-elute. This is unlikely to cause problems for our C2 - C6 data, because the calibration standard for light hydrocarbons is comprehensive, but there may be interferences in the C5 - C10 data.
10.2 Quality Assessment
Laboratory analysis of the air samples was conducted, the samples were examined for outliers. When an outlier was spotted, the original chromatogram was checked for correct integration and quantification.
10.2.1 Data Validation by Source
Time series plots for each hydrocarbon and hydrocarbon distribution plots for eaune. Samples were collected in evacuated stainless steel canisters and shipped to our laboratory in Toronto, ON for analysis by GC-FID. Concentrations in the pptv - ppbv range are reported for a variety of biogenic and anthropogenic non methane hydrocarbons. Our measurements are sufficiently precise for use in calculating fluxes by the gradient method. They are also valuable as a record of the ambient trace gas concentrations present in the region.
10.2.2 Confidence Level/Accuracy Judgment
We are quite confident in our C2 - C6 data. The C10 concentrations may require adjustment later, but are reasonable.
10.2.3 Measurement Error for Parameters
Absolute concentrations are considered accurate within 10%, limited by the uncertainty in the concentrations in our calibration standards. Uncertainty in gradients calculated from these concentrations is determined by the precision of our analysis, which is quoted in section 5.2.1.1.
10.2.4 Additional Quality Assessments
As data from other investigators becomes available, we will compare our results with theirs and update our data and documentation as necessary.
10.2.5 Data Verification by Data Center
(For BORIS and ORNL DAAC Use)
11.1 Limitations of the Data
11.2 Known Problems with the Data
Three samples in this data set contain unusually high levels of anthropogenic hydrocarbons. They are:
12.4m 02-Jun-94 13:25
12.4m 02-Jun-94 14:35
23.3m 04-Jun-94 12:00.
This is not a problem with our GC analysis; values from the two GC-FID systems for C6 hydrocarbons agree well. The high concentrations of light hydrocarbons such as ethane and propane indicate that contamination of the canisters is not a factor; these species are removed very effectively by our cleaning procedure. In fact, 23.3m 04-Jun-94 12:00 was collected in a new canister purchased for this project. We suggest referring to other measurements of trace gas concentrations at these times to determine whether a significant, transient, local anthropogenic source existed.
Canisters with samples:
23.3m 31-May-94 20:00
12.4 04-Jun-94 16:00
arrived at our laboratory empty, the valves having leaked.
C10 data for some samples is missing. Their chromatograms are being reintegrated to improve accuracy. These concentrations will be provided in a future revision.
11.3 Usage Guidance
11.4 Other Relevant Information
14.1 Software Description
HP 5970 GC-MS Workstation v. 3.2 software
HP PC-Chemstation v. 1.01 software
14.2 Software Access
15.1 Contact Information
Ms. Beth McCowan Bldg 22. Room G87 Code 923 NASA GSFC Greenbelt, MD 20771 (301) 286 4005 (301) 286 0239 (fax) beth@ltpmail.gsfc.nasa.gov
15.2 Data Center Identification
See 15.1
15.3 Procedures for Obtaining Data
Users may place requests by letter, telephone, electronic mail, FAX, or
personal visit.
15.4 Data Center Status/Plans
None.
16.1 Tape Products
16.2 Film Products
None.
16.3 Other Products
None.
17.1 Platform/Sensor/Instrument/Data Processing Documentation
Sample analysis is performed by standard chromatographic techniques. No knowledge specific to the GC-FIDs we use is required in order assess or interpret our data. Any analytical instrumentation text should give the user sufficient information about GC-FID.
17.2 Journal Articles and Study Reports
A comprehensive review of sample analysis and data analysis in our laboratory may be found in the Ph.D. dissertation of B.T. Jobson:
B. T. Jobson, Seasonal Trends of Non methane Hydrocarbons at a Remote Boreal and High Arctic Site in Canada, Department of Chemistry, York University, 1994.
17.3 Archive/DBMS Usage Documentation
BOREAS - BOReal Ecosystem-Atmosphere Study BORIS - BOREAS Information System DAAC - Distributed Active Archive Center EOS - Earth Observing System EOSDIS - EOS Data and Information System GSFC - Goddard Space Flight Center NASA - National Aeronautics and Space Administration ORNL - Oak Ridge National Laboratory URL - Uniform Resource Locator
Return to top of document.
20.1 Document Revision Date
Written: 22-Nov-1994
Last updated: 12-Dec-1996
20.2 Document Review Date(s)
BORIS Review:
Science Review:
20.3 Document ID
(For BORIS and ORNL DAAC Use)
20.4 Citation
20.5 Document Curator
(For BORIS and ORNL DAAC Use)
20.6 Document URL
(For BORIS and ORNL DAAC Use)
