Investigators:
Robert R Arnts
Chris Geron
Fred L Mowry
John T. Walker
Grants:
"Emissions of Biogenic Oxidant and PM Precursors- Very High Reactivity VOCs and Surface Layer Chemistry above Forests", US Environmental Protection Agency, , 2003 - 2005
Description:
Inhalation of fine particulate matter (PM2.5) and ozone have been identified as health risks to humans. In response the EPA regulates atmospheric emissions to reduce these risks. In the case of particulate matter and ozone, these reductions may require that chemical precursors to these hazards be regulated. With ozone, reactive volatile organics (VOCs) and nitrogen oxides govern the generation of ozone in the atmosphere. In the case of fine particulate matter a variety of sources contribute to the usually complex mixture of inorganic and organic substances found therein. Radiocarbon measurements of the organic fraction of particulates from air near Nashville, TN indicate up to 70% has been derived from biogenic sources. Volatile organics emitted by vegetation such as monoterpenes and more recently isoprene and sesquiterpenes have been shown in the laboratory to form aerosols as a result of atmospheric oxidation processes. Some of the oxidation products of monoterpenes and isoprene have been identified in the organic fraction of ambient particulate matter thus probably contributing to the biogenic component observed in the radiocarbon measurements. Furthermore, particle flux measurements in the size range of 10 to 100 nanometeres (range where gas to particle conversion is observed) over forests show positive fluxes - indicating formation of particles. Evidence has been presented that these particle formation events may be more attributable to sesquiterpene than monoterpene emissions.Air quality models such as EPA's Community Multiscale Air Quality (CMAQ) model are incorporating the newly emerging chemistry and physics of gas to particle conversion. In order to accurately simulate this process and thus account for any uncontrollable biogenic contribution to ambient particulate matter, the model needs emission data which specifies which compounds are emitted and at what rate. While the isoprene portion of the inventory has been greatly improved during the 1990s due to the contributions of many research efforts, the non-isoprene portion of the inventory (currently about half of the total emissions) is highly uncertain. Recent National Research Council and NARSTO reviews have called for improvements in this area. The current inventory does not include sesquiterpene emissions at all due to the lack of data. Monoterpene emissions are currently calculated by simple temperature correction to a base emission rate.
This effort will focus on improving our understanding of the what variables control release of volatile organics from a loblolly pine (Pinus taeda) forest canopy. Initial experiments have shown that in addition to monoterpenes we observe several sesquiterpenes emitted. The major sesquiterpene, B-caryophyllene, in leaf enclosure and small seedling experiments (without ozone present) is emitted at rates comparable to the major monoterpene emission, a-pinene. However, in the open atmosphere above the canopy, our flux measurements show only a small fraction of of the a-pinene flux. Rough analysis of the estimated rate of loss of the B-caryophyllene with ozone suggests that the measured flux is consistent with very fast oxidation occurring in the short (30 seconds) transport time from emission at the leaf surface to collection in our sampler. Futhermore, we observe oxygenated species such as n-nonanal, n-decanal, 6-methyl-5-hepten-2-one and methyl salicylate in experiements when pine seedlings are fumigated with ozone at ambient levels. We also see some of these compounds in our flux measurements. Ozone initiated release of volatile organics from vegetative surfaces is not a process that is currently represented in the emission inventories.
The impact of this work will be that new algorithms improving currently poorly defined emissions (monoterpenes) and currently non-represented emissions (sesquiterpenes and ozone instigated products) will be incorporated in biogenic emission inventories for use in air quality models. Because these compounds can be ozone and particulate matter precursors, the accuracy with which the air quality model can simulate ozone and secondary organic aerosol formation will be improved. Furthermore, as a result of studying gas phase ozone-VOC losses and ozone-leaf surface reactions, we will define what mechanisms are operative in contributing to ozone dry deposition. All these improvements to the models will make them more accurate thus improving their efficacy as a tool for regulatory decision for achieving cleaner air.
Presently, equipment installed on the Ring 6 central tower consist of components of a relaxed eddy accumulator (REA) flux measurement system and other standard micrometeorological devices necessary to check for energy balance closure in the mixing layer above the canopy. The REA system samples air continuously from near a sonic anemometer which is positioned about 5 meters above the canopy edge. Mounted on a boom so as to avoid the wake effects of the tower, the boom can be reoriented within a range of about 120 degrees to point into the prevailing winds expected for the measurement day. Two air intakes consisting of 55 mm open face Teflon particle filters draw air continuously from the horizontal plane of the sonic anemometer. One filter draws air (about 2 lpm) for eddy covariance measurements of carbon dioxide and water vapor (photosynthetic and latent heat fluxes). This sample is drawn into a closed path infrared gas analyzer (LICOR 6262) which is housed in weatherproof metal box a few meters away mounted on the tower. A second smaller metal box houses a 120VAC diaphragm type vacuum pump and a mass flow controller to maintain flow. The second Teflon filter draws air at about 15 lpm for possible sampling by the relaxed eddy accumulator system. Just after the air is drawn through the particle filter, the air is continuously blended with a 90 cc/min stream of internal standard gas (2 ppm each of methylcyclohexane and 2-methylheptane). This is supplied by a compressed gas cylinder at the base of the tower, a 1/8" OD stainless steel tube and controlled by a mass flow controller within the Mass Flow Controller module (see below). The length of ΒΌ" OD PFA Teflon tubing quickly transports the sampled air to the REA module where a set of two sample segregator fast acting Teflon valves can be actuated to shunt a portion of the flow to an updraft or downdraft accumulator (sets of metal tubes packed with Tenax-GR adsorbent). The system is controlled through a computer situated in the Ring 6 instrument shed (cabled communications). Another module attached to the tower just below the REA unit houses six mass flow controllers is cooled via a solid state air conditioner. The sonic anemometer, IRGA, mass flow controller, micromet sensors and REA module communicate to the computer through specialized I/O devices (Applied Technologies Data Packers) also mounted on the tower. The REA module also contains a set of thermoelectric coolers to regulate temperatures of the adsorbent collection tubes. A 12 VDC power supply at the base of the tower feeds this unit via a heavy duty (yellow) cable ascending the tower. The micromet sensors on the tower consist of net radiometers, pyranometers, PAR sensors and a set of five precision thermocouples for measuring vertical air temperatures. A heavily insulated module adjacent to the data packer houses reference junctions for the precision thermocouples.
In addition to the signal and power cables, two rigid plastic tubes connect the vacuum pumps (2 blue boxes at the base of the tower with the Mass Flow Controller and IRGA modules. Two stainless steel lines (1/8") conduct zero air and nitrogen from compressed gas cylinders at the base of the tower to the IRGA and Mass Flow Controller modules.
Instrumentation housed in the shed include 1) the REA Controller for on the fly vertical wind monitoring and sampling control, 2) the supervisory computer (laptop) that allows the operator to interface with the controller, display incoming data, perform calibrations and setup experiments. Two sets of four channel mass flow controller readout devices regulate flow operations and supply output data stream to the REA supervisor.
Lastly soil and biomass temperatures data is gathered from a small plot just outside the perimeter of Ring 6. Again these sensors are accompanied by a thermocouple reference box and data packer which are hard wired to the computer in the instrument shed .
Operational protocols
Each set of adsorbent tubes is used to integrate up and downdrafts for up to 55 minutes with up to six sets of tubes loaded in the REA magazine. The sampling schedule is weighted toward high ambient temperature weather in order to capture maximum emission days. In addition sampling days are scheduled every two to three weeks to monitor for seasonal changes in emissions. Sampling days may be added to detect changes in emission pattern associated with phenology or environmental changes (bud break, drought etc.) Selection of specific days for sampling is also governed by anticipated prevailing wind direction (west to southwest winds are optimum) and low probability of storm activity. Measurements are anticipated to commence in April and run through September.
Because equipment is not used for periods of two to three weeks during the measurement season, cables running to the instrument shed are disconnected from the electronics on the tower. This provides some protection against static/lightning overloads of the system. On measurement days cabling is re-connected, gases turn on and equipment powered up shortly after sunrise (alternatively this is done the day before if weather permits). After warm up, flow calibrations are performed by one person on the tower in radio communication with the operator in the shed. The REA is loaded with adsorbent tubes and a series of quality assurance (pressure balancing and contamination checks are run). The REA is reloaded and the system set to run a series of flux measurements. During this period all micromet data is logged into the system. Measurements usually run for the full course of the daylight hours or until the atmosphere becomes stable and unsuitable for flux measurements. Tubes are returned to the laboratory at EPA-RTP for detailed analysis of volatile organic compounds.
Proposed additions to the system for the 2005 growing season.
1. Some measurements from summer 2003 (last measurements) indicate some drift in the internal standard measured in the Balance runs. We believe this may be due to adsorption and desorption phenomenon occurring in the stainless steel line due to temperature fluctuations during the day. To remedy this we propose to add a low power heater cable to this line to stabilize temperatures. We will use self regulating heat cable (UL listed, CSA certified) and cover both heater cable and stainless steel line with Silicone coated fiberglass sleeve. The maximum temperature under full power is only 50 C. The tower base terminus of this cable will be passed to a NEMA 4X (outdoor approved) enclosure which will in turn be provided with a grounded 120 VAC plug.
2. Some of the important volatile organic compounds we are measuring (i.e. the sesquiterpene B-caryophyllene) react very fast with ambient ozone. To strip out the ozone and prevent loss of the sesquiterpene we currently pass the sampled air though a stack of copper wire mesh screens that are coated with manganese dioxide. This removes that ozone but also causes some loss of the sesquiterpene (probably due to adsorption). We would like to try an alternative approach which should remove the ozone but not cause loss of the sesquiterpene. This will be accomplished by adding a stream of nitric oxide in nitrogen to the sampled air,. The nitric oxide reacts quickly with the ozone. The compressed gas cylinder contains 1.5% NO in nitrogen. About 1 cc/min per minute will be continuously added to the 15 lpm total air flow yielding 1 ppm in the sample stream. Most of this will be vented by the vacuum pumps at the base of the tower. A chemical scrubber could be added to this remove NO/NO2 venting to the atmosphere. However, this loading is too small to appreciably elevate NO/NO2 levels in the ring. Pending approval of this component we will add a small (10"X8"X4") enclosure near the REA box to regulate flow of the nitric oxide. The regulated output will be Teed in to the internal standard line for blending with the sampled air.
3. We also plan to add three aerosol measurements to examine the possible fast conversion of biogenic emissions to particles. These include: 1) A condensation nuclei counter for performing flux measurements of small particles. This assembly is about a 20" cube and mounts on the outside of the tower on a cross member. It operates in concert with an RM Young sonic anemometer which will be mounted adjacent to the REA sampling boom. 2) A particle size classification instrument- an electrostatic size classification instrument will be co-located with the nuclei counter. 3) An integrating nephelometer for measurement of light scattering aerosol. This instrument is currently on the tower (used during the 2003 CELTIC study) and resting on one of the upper walkways. This can be mounted on the outside of the tower to improve egress.
Update 5/16/2005:
Addition of ambient ozone, nitrogen oxides and sulfur oxides monitoring to ring 6. In order to understand canopy level ozone/VOC/nitrogen oxides photochemistry and the formation of secondary organic aerosols, we wish to add said O3/NOx/SOx measurements. These will be used to model the oxidant and aerosol formation processes.A 1/2" OD Teflon line with an open faced Teflon particle filter inlet will be placed on one of the CO2 release towers nearest to the instrument shed in ring 6. It will be placed so that air will be drawn at the same level above the canopy as the relaxed eddy accumulator inlet (on the central walkup tower). A pump will continuously draw air through the line to the instrument shed where it will be distributed to each of the three analyzers. Data will be continuously logged into a data logger and in turn transferred to a spreadsheet in a laptop computer.