No.8

Introduction
Data Policy
Data
Documentation
Contact Information



Cape Verde Observatory

-NOx-

back
Station Contact Person(s) Observation Data and Quick Plot

Observation
Category : Air sampling observation
Situation : ongoing
Time zone : UTC


Sampling
Sampling height : 5
Description : continuous
Sampling and analysis frequency : Measurement of both NO and NO + NO2 occurs on the same instrument channel. The chemiluminesence analyser runs on a 10 minute measurement cycle, 5 minutes of the signal received from NO and 5 minutes of the signal from NO + NO2. Each 5 minute period consists of 150 s of pre-chamber zero readings, followed by 150 s of NO signal. Here we just report NOx (NO + NO2).
Sampling environment : Except from local source emissions from passing ships, the level of NOx in the Cape Verde region (tropical marine boundary layer) is mainly influenced by emissions in the northern hemisphere. There, emissions of NOx are dominated by anthropogenic sources from urban or industrial regions, with little or no emission from oceanic or uninhabited areas. Oxidation of emitted NOx to nitric acid (HNO3), peroxyacetyl nitrate (PAN) and other alkyl nitrate compounds occurs in less than a day in the continental boundary layer, therefore NOx levels in remote regions are usually very low (<100 pptv).

In the presence of NO, VOC oxidation leads to catalytic production of O3, where peroxy radicals are recycled to OH, leading to further VOC oxidation. At higher levels of NOx, such as those found in urban or industrials areas of Europe or North America, OH radicals can be lost via reaction with NO2 to produce HNO3, which can be a reservoir species for NOx. At low levels of NOx, peroxy radicals self-react leading to net destruction of O3 via its photodissociation in the presence of water vapour. It is the competition between the photochemical loss and production of O3 that leads to a non-linear dependency of O3 production on NOx with the O3 production efficiency being greater for lower NOx levels.

A particularly crucial NOx range is 5 – 100 pptv, which depending on the time of year, location in the troposphere and levels of O3 and H2O, contains the so called O3 compensation point, a level of NOx at which there is a net production of O3.
Description for sampling analysis : 10 m 1/4 inch teflon tubing. Filter fitted on inlet.


Instrument and Analysis
Measurement method : Chemiluminescence
Current status and history of instrument : Measurements obtained since October 2006.

Single channel Chemiluminescence instrument manufactured by Air Quality Design (AQD) Inc, Colorado, USA.
Description of instrument : This technique is highly selective and sensitive, and has a linear response over a wide dynamic range (1 ppt – 500 ppb).
The NO2 converter used in the set-up at CVAO utilises a solid-state light source with a narrow band of emission wavelengths to achieve interference-free NO2 to NO conversion.Typically the NO2 to NO conversion efficiency of this set-up is between 35 - 40 %, with a gas resonance time in the irradiated area of around 1 second. The instrument runs on a 10 minute measurement cycle of which 5 minutes is with the LEDs off followed by 5 minutes with the LEDs on. Each 5 minute period consists of 150 s of pre-chamber zero readings, followed by 150 s of NO signal.
The instrument sensitivity is determined every 37 hrs when the instrument is calibrated. The sensitivity to NO varies from 2.2 – 3.1 counts s-1 pptv-1.

The Limit of Detection (LOD) can either be determined from the zero count rate variability, directly related to the photon counting of the detector; or by considering the signal scatter of the processed 10 minute nighttime data (little or no local pollution so assume all NO is converted to NO2 during transport). These methods give values of 1.5 pptv and 1.7 pptv respectively and so we quote the LOD of NO as below 2 pptV.

The accuracy of the measurement considers the Sample and calibration gas mass flow controller accuracies, the concentration of the NO standard, and the error in determining the NO artifact which gives a total uncertainty of 23 %. The NO2 artifact is assessed and combined with the extra uncertainty of the conversion of NO2 to NO, means that the overall uncertainty for NO2 measurements is around 30%.

We report measurements of NOx (NO+NO2)


Calibration
Current scale employed in the measurement : BOC NPL 2006
Measurement calibration : Calibrations are carried out using the standard addition method where a small flow of calibration air is added to the ambient sample flow. This is beneficial because it allows sensitivities to be calculated in air that is the same humidity as ambient. Water vapour is highly effective at quenching the NO2 excited state and so changes in humidity can significantly alter the instrument sensitivity. Calibrations and zeros are performed every 37 hours and consist of nine 10 minute measurement cycles. The first 3 cycles calibrates for NO sensitivity using a flow of 4 sccm of 5 ppmv NO gas in nitrogen in a sample flow of 1600 sccm (giving an NO concentration of ~12 ppbv). The next three cycles use NO with added O3, which titrates approximately 90% of the NO calibration gas to NO2, giving a known NO2 concentration. Following this, NOx free zero air (Ecophysics PAG003) is sampled for three cycles to give an estimation of the instrument artifact. This artifact is caused by a so called “fake NO” signal, first reported by [Drummond et al., 1985], and is not removed by subtraction of the signal observed when O3 is added to the pre-chamber. A better way to ascertain the NO artifact is to use the ambient NO signal at night, when no NO should be present (provided there are no local emitters). We believe this is the case for almost all times during the measurements period at the CVAO and hence the nighttime signal is used for determination of the instrument NO artifact.
Scale and calibration(treasability) : The concentration of the NO standard (5 ppmV): Known to an accuracy of 1%, as stated by BOC (British Oxygen Company). BOC certificates that NO/N2 mix is stable for 5 years;


Data Processing
Measurement unit : ppt
Data processing : The 10 minute averaged NO and NO2 measurements were filtered for outlier points by comparison with the running three day average. Points that were outside 3 standard deviations of the three day running mean were removed from the data series before the hourly averaging took place. It is believed that the outlier points are mainly associated with local shipping and occasionally from cars visiting the area of the site. On analysis of the raw 1 second data, it appears that most of these pollution episodes lasted less than 30 seconds. Due to the 10 minute data acquisition cycle, there were occasions when the spike in the data only appeared during the NO measurement and not when the instrument was measuring NO2. This causes an apparent negative value of NO2 to be reported from the instrument, caused by subtraction of the polluted NO. These data were also removed from the dataset.
Processing for averaging : Processing for Hourly Data:
6 x 10 minute points are averaged to produce the hourly data.
Processing for Daily Data:
24 x hourly points are averaged to produce the daily data.
Processing for Monthly Data:
The hourly data is averaged between 01/xxx 00:00 to 01/yyy 00:00.
Data flag : DATA FLAGS: Error Flag = 0 Good data
Error Flag = 1 Reduced quality data
Error Flag = 2 Below detection limit
Error Flag = 3 Invalid or missing data
Data remarks :


Other Information
Scientific aim : To obtain one of the first sets of longterm measurements of NOx in the tropical marine boundary layer, a region previously data sparse.
NOx levels and therefore tropospheric O3 destruction, are dependent on air mass origin, with higher NOx observed in air from the African continent compared to Atlantic oceanic air masses. In model simulations calculated for average annual conditions, the O3 compensation point would occur at about 25 pptv NO, which is higher than early measurements suggest (7-8 pptv). Data taken at the observatory in subsequent years should allow a more complete investigation into the O3 compensation point and hence how the O3 destroying capability of the region may change with possible future changing NOx.
Reference : Read, K. A., A. S. Mahajan, L. J. Carpenter, M. J. Evans, B. V. E. Faria, D. E. Heard, J. R. Hopkins, J. D. Lee, S. J. Moller, A. C. Lewis, L. Mendes, J. B. McQuaid, H. Oetjen, A. Saiz-Lopez, M. J. Pilling and J. M. C. Plane (2008), Extensive halogen-mediated ozone destruction over the tropical Atlantic Ocean, Nature, 453(7199), 1232-1235.

Lee et al, Year round measurements of nitrogen oxides and ozone in the tropical North Atlantic marine boundary layer, submitted to JGR, 2009.


submitted by University of York

The WDCGG is operated by the Japan Meteorological Agency
in collaboration with the World Meteorological Organization