The technique of open-path monitoring with a Fourier transform spectrometer is being vigorously developed.Many issues for the routine acquisition of data have been addressed and are fairly well understood. The primary questions that remain unanswered at the present time are how to measure the quantities accuracy and precision. When the FTIR system makes measurements, it uses a long, open path, and for that configuration, even the concepts of the accuracy and precision are difficult to define. The classical definition of accuracy is in reality a measure of how well the instrument responds to a known stimulus. Precision is classically defined as the standard deviation of a set of responses of an instrument to the same stimulus. While these definitions maybe applicable to FTIR measurements, actually challenging the instruments with a known quantity of gas presents many problems. One suggestion has been to place a short cell filled with a high concentration of gas in the beam. Then, if the concentration in the cell is known, both the accuracy and the precision can be determined. The use of a short cell poses several experimental problems that need clarification. For example, during routine operation the FTIR measures the absorbance due to gases along the path. If a short cell containing a high concentration of gas is to be used to measure either precision or accuracy, then should it be in the beam but empty while the routine measurements are being made. What time is to be used or the accuracy and precision measurements. Since the gas in the cell is added to the gas along the path, how can the variability of the atmospheric component be accounted for unless the acquisition time is short compared to the variability of the data. The answer to the acquisition time question obviously depends on the end use of the monitoring program data. It has been suggested that a naturally occurring gas such as N₂O may be used for the precision and accuracy measurements. That idea suggests the following question: If a measurement of N₂O is used for the determination of accuracy, can the accuracy of a measurement of toluene be determined simultaneously. The answer to that question cannot be given as a simple yes or no. Another possibility may be to use the absorbance due to water vapor. Water vapor exhibits absorbance throughout the mid-IR region and is easily measured by independent techniques.

A method for the evaluation of cloud temperature and transmittance from ground-based measurements of the downwelling spectral radiance is proposed. The method uses the strong emission band of ozone at 9.6 micrometers as a natural source of IR radiation for probing clouds. Temperature and transmittance are derived from the cloudy sky radiance measured in two narrow spectral channels. Cloud parameters are found by solving a system of equations requiring an estimate of the cloud height and thickness obtained from lidar returns and of atmospheric temperature and humidity profiles. Tests performed on experimental spectra verify the method with typical temperature errors smaller than 2 K for the spectral measurements coincident in time with the lidar measurements. The main sources of limitation are identified and analyzed.

Measurements of stratospheric OH at 308 nm with the sun as light source have been carried out using a FTS. The spectra have been recorded in the high Arctic at Ny-Alesund during summer 1996. During clear sky conditions the spectra recorded with the FTS were found to be of sufficient quality with regard to the SNR of the analysis of OH. However, already slightly cloudy conditions increased the scintillation noise to an extent, preventing the OH analysis. The retrieval of the total OH columns was done in two ways; by studying single solar absorption spectra and by utilizing the Doppler shift of two spectra, recorded on the east and west side of the solar disk. The results of both methods agreed and gave total zenith OH columns of about 6 X 10¹³ molec. cm^-2 for SZA of 60 degrees.

New legislation requiring the monitoring and reduction of polluting gases in both Europe and the US has increased the demand for reliable and affordable Open Path Monitoring Systems, suitable for hazardous area operation. We report on the research and development of a UV Fourier-transform spectrometer based system for open path monitoring of both hazardous and environmental gases. The device has no moving parts and is designed for unattended operation. Laboratory tests have successfully detected the present of and differentiated between SO₂ and H₂S present in a 1 meter test cell down to 1.0 ppm.m levels. the real time response of the system allows for the use of maximum entropy modeling to predict the size and location of a gas leak. Initial field tests with open air gas releases of SO₂ and H₂S have verified this. The multiple gas feature of the instrument allows for additional possible applications with regards to environmental monitoring.

A nonlinear spectral curve fitting computer algorithm for the reduction of Fourier transform spectrometer data is being developed at the Air Force's Arnold Engineering Development Center. The algorithm is an adaptation of an existing algorithm utilized for plume signature data and optical gas diagnostics data reduction. The nonlinear algorithm is not constrained by the linear absorbance assumptions required for classical least squares (CLS). Therefore, the nonlinear algorithm allows for the determination of gas concentrations from high absorption level spectra as well as determination of other nonlinear parameters such as spectra shift and spectral resolution. In addition, the currently implemented algorithm has the ability to simultaneously determine the background spectra or I degrees. These capabilities could make the nonlinear algorithm a welcome addition as a quality assurance and data reduction tool. Open path Fourier transform IR spectra from various sources have been analyzed utilizing this algorithm in an attempt to determine its capabilities and limitations. As various data sets are analyzed the algorithm is further refined to make its application easier and faster. A brief description of the algorithm and the results of a comparative study between the nonlinear algorithm and CLS is presented.

Fourier transform IR (FTIR) spectroscopy has become a powerful analytical tool for the detection and measurement of atmospheric pollutant gases. This work describes the application of concentration analysis techniques to data recorded with a versatile FTIR spectroscopy system, developed at the University of Reading PHysics Department. Spectra were recorded at three separate sites, each possessing a distinct source of atmospheric pollution gases. The two sites monitored in the active mode were a traffic congested town center at rush hour and a dairy farm cow shed. The site monitored passively contained three 5 m high methane burners. The analysis techniques have been designed to provide rapid and accurate analysis of the spectrometer data, without the need for high computing power, thus making analysis possible in the field using a laptop PC. In an attempt to enhance the resolution of the spectral data, and therefore resolve overlapping spectral lines, a super- resolution algorithm has been tested on part of the recorded data. The results of applying the algorithm has been tested on part of the recorded data. The results of applying the algorithm, predominantly an image processing technique, are shown and improvements to the algorithm are discussed. Results from the urban and agricultural sites show that CO, CH₄, and NH₃ can be measured to a ppm level with a maximum uncertainly of 8 percent.

Increasing regulatory standards for the limitation of greenhouse gas and other pollutants emissions from industrial facilities require new environmental control strategies and innovative and cost-effective measurements techniques. Elevated flare stacks present a difficult measurement challenge, because it is extremely difficult to determine concentrations in the post-combustion gas of operational flares by sampling techniques. A remote sensing measurement system has to be used, therefore. Fourier transform IR spectroscopy measurements were realized using a mobile environmental laboratory which is equipped with a high resolution K300 spectrometer. A multi-component air pollution software MAPS was used to determine CO₂, CH₄, CO and NO concentrations. It is based on radiative transfer line-by-line calculations and least-square fit procedures. The paper focuses on the results of FTIS flare emission measurements at natural gas production facilities. Different types of flares have been investigated under varying operational conditions. Emission rates and combustion efficiencies are calculated which indicate that the environmental impact of methane emissions from flares is rather small.

Fourier transform IR spectroscopy (FTIR) is well suited for establishing emission rates from sources of for example CH₄, N₂O, CO₂ as important greenhouse gases and for NH₃ as a pollutant causing severe damages to natural ecosystems. Since the gases mentioned above have spectral liens broadened to 0.2 cm^-1, the resolution used for these measurements is 0.2 cm^-1 too on a mobile instrument allowing resolutions up to 0.06 cm^-1. In a first attempt, calibration spectra were taken from HITRAN92 based FASCODE calculations. Since test measurements revealed problems in determining accurate concentrations, own spectra were taken at a dynamic gas-mixing station. The absorption coefficients differ strongly from the simulations, for example for CH₄ deviations are 19 percent max., for N₂O 16 percent for NH₃ 15 percent. Calibration showed good linearity of absorbance vs. concentration up to absorbencies of 0.4 at the resolution of 0.2 cm^-1.

Ground based emission spectroscopy measurements are performed in the high arctic. A commercial uncooled FTIR interferometer has been sued. The emission spectra of the polar atmosphere are collected to examine the variation of trace gases in the dark wintertime. The emission of these trace gases is calculated by the FASCOD3 algorithm. Additionally tropospheric aerosols have to be considered in the analysis. The concentration variations of the gases and aerosols are due to the absence of photolytical reactions in the wintertime and to transport of polluted airmasses from the midlatitudes to the Arctic.

Commercial Fourier transform based FTIR spectroradiometers can be used in an easy way as ground-based remote sensing systems. Information of the ozone column can be obtained from the spectral IR radiance in the 900-1200 cm^-1 range. Tropospheric and stratospheric ozone are sounded by an appropriate selection of the experimental conditions. A theoretical study of the best choice for the wavenumbers of zenith angles for the direction of view is presented. The problem of the superimposed water emission is analyzed for the selected wavenumbers. Finally, some experimental results are presented, with a fast method to obtain the water and ozone columns from a spectral radiance spectrum.

The problems of the stratospheric ozone decline during the last decades is the subject of scientific efforts to discover the main factors influencing the processes leading to its disappearance. The exact understanding of chemical- physical processes in the Earth's atmosphere needs high- quality representative measurements for time-series and modeling study. The present paper deals with the problems of the polarization characteristics of DOAS instruments and the corresponding requirements for instrumental orientation relative to the scattered plane. To avoid the polarization effects upon the measurement's accuracy the instrument has to be 'unpolarized'. A fiber optic and an appropriate diffraction grating are used to do this. Experimental measurements have been carried out to check the possibility to use a fiber optic bundle for DOAS applications. The results obtained from this experiments how that the tested fiber bundles do not completely destroy the polarization state of the incoming radiation. At the fiber exit the degree of the polarization reaches up to 10-15 percent in the measured spectral intervals. On the other hand, accordingly to some of the existing methods, to remove Ring effect from the measured zenith spectra, the DOAS instrument used has to be strongly 'polarized'. This achieved by means of a polarizer inserted into the instrument. The assumption is that, applying such kind of measurements makes it possible to separate the Rayleigh from the Raman scattering. The latter e is assumed as unpolarized, which is in conflict with the theory. A new method base on the deconvolution is offered by use as a possible way of overcoming the problems with the Ring effect.

The scanning imaging absorption spectrometer for atmospheric cartography (SCIAMACHY) is a passive optical instrument applied for the remote sensing of the earth atmosphere. SCIAMACHY is the German/Dutch contribution to the first ENVISAT mission defined by ESA and will be launched on the Polar Platform in 1999. Dornier Satellitensysteme GmbH has the prime-contract by the German agency DARA and Fokker Space B.V the prime contract by the Dutch agency NIVR with participation of the Belgian institute BIRA. The primary scientific objective of SCIAMACHY instrument is to observe the atmosphere by measuring sun- and moonlight which is transmitted, reflected and scattered by the earth's atmosphere. Based on the atmospheric absorption of light in the 240 nm to 2385 nm bandwidth concentrations of ozone, greenhouse and trace gases will be determined. The instrument was conceived to improve our knowledge and understanding of a variety of issues of importance to chemistry and physics of the earth's atmosphere such as ozone hole chemistry, troposphere-stratosphere exchange and tropospheric pollution.

In the mid-IR spectral region between 3 micrometers and 25 micrometers lead salt diode lasers are the most common light sources for high-resolution spectroscopy. Other light sources as CO-, CO₂-lasers or sideband-lasers are expensive, non- portable and poorly tunable. To overcome the cryogenic cooling of the lead salt diode lasers and the describe disadvantages of the other light sources the generation of tunable laser radiation by difference frequency generation in nonlinear crystals as an alternative method is well known. Because of the low efficiency of the nonlinear process dye, solid state or gas lasers are the preferred pump sources for difference frequency generation up to now. To use the flexibility of the nonlinear process with the advantages of the visible- or near-IR-diode lasers we developed a portable tunable cw difference frequency spectrometer applying the nonlinear crystal AgGaS₂ and two diode lasers as pump sources. High resolution scans over CO, N₂O and OCS absorption lines near 2107 nm^-1 demonstrate the capabilities of the spectrometer. Moreover, we present high resolution line broadening measurements of the R(2) line of ¹³C¹⁶O in the fundamental band as well as experiments for trace gas detection of ¹²C¹⁶O in the fundamental band.

The use of catalytic converters in cars with gasoline engine results in a tremendous reduction of the emission of pollutant gases. The optimal operation of the exhaust treatment systems is being checked and maintained periodically, but there is always a significant percentage of cars with a malfunction of the catalytic converter causing a substantial percentage of the total emission. Roadside emission monitoring of individual cars in the running traffic could be used to indicate these gross polluters, arrange maintenance of their vehicles and thus reduce total emission. Present monitoring systems use non- dispersive IR spectroscopy. Other systems are based on mid- IR diode laser spectroscopy offering a higher signal to noise ratio, higher selectivity for detection of specific compounds and better optical quality for long open path measurements, but these systems depend on liquid nitrogen cooling. In this work a compact mid-IR (MIR) laser diode system for roadside measurements will be presented, that is cooled thermoelectrically using a Peltier element. Sensitivity and time resolution of the system have been determined and found to be suitable for detection of single gross polluters in the running traffic. The presented system demonstrates the feasibility of high sensitive, selective and fast field MIR laser diode spectroscopy together with ruggedness and low maintenance expense.

This paper reports on investigations regarding the visualization of environmental relevant gases normally invisible to the human eye. The principle of this system is based on the backscatter absorption gas imaging technique developed by Lawrence Livermore National Laboratory and Laser Imaging Systems. Main advantages of this approach in comparison to conventional laser-based remote gas-sensing methods are reachable speed of localization, high sensitivity as well as the detection of multileakages. The applied method is based on infrared (IR)-technology consisting of a thermal imaging device and a tunable CO₂-laser. Main aim of a project carried out at the Laser Zentrum Hannover e.V. is the computer based processing of spectroscopic and IR-image information provided by this technique. For this, several important parameters like the wavelength tuning precision, the dependence of the wavelength on the average output power of the laser source as well as the optical resolution have been investigated. The results show, that the BAGI-technology offers a large potential for future developments related to the observation of transparent gases. In addition to this the authors presents a new concept for an enhanced system that would not only be able to visualize gases but also capable of determining their direction of moving, computing their volume and forecasting their possible further spreading.

A photodiode-array spectrometer for the detection of atmospheric trace gases has been developed to take diffuse solar zenith and, using a n artificial light source, horizontal long path measurements. Of the numerous factors involved in determining the minimum amount of a detectable gas, including its spectral characteristics, atmospheric phenomena and the algorithm used, the present study examines only the features of the spectrometer and sensor. The spectrometer's spectral dispersion is accurately calculated and its internal stray light kept to a minimum. The linear image sensor's dynamic range and aperture-response function are analyzed along with those phenomena that can alter the actual signal such as veining glare, blooming, etaloning and dark current; the latter two are treated in depth. Etaloning is linked to optical interference in the sensor's passivating layer. A simple model of it, as well as a check of the relative stability over time of the transmission peaks, are included. The analysis of the dark current indicates the existence of a photo-induced component that is inversely proportional to the incident radiation flux. It yields a simple analytical equation describing the phenomenon for the sensor's various elements. This makes it possible to derive the correct dark current value in relation to both the incident radiant energy and the spectral range investigated.

A monitoring campaign of atmospheric pollutants was conducted in February 1993 by several of Italy's CNR institute in heavily polluted greater Milan. This metropolitan area, the largest one in northern Italy, is situated in the northernmost part of the Po Valley and, because of its topography and orography is frequently marked by low ventilation and inversion phenomena, a fast that promotes the accumulation and vertical layering over the city of pollutants. The need for more detailed information on air circulation and changes occurring in the lower atmospheric layers, as well as to understand why air-mass exchange does not take place, thereby impeding the dispersion of pollutants, was the project's goal- orientation. Measurement of NO₂, SO₂, O₃, HNO₂ were carried out over a 1.7 Km path in the city center by means of a DOAS system called GASCOD developed by remote sensing group of FISBAT-CNR at Bologna. The light source has been equipped with a remote-controlled occulting devices in order to separate the sky light scattered into the field of bye of the receiving system, which can interfere with the lamp spectra during daytime. The light from the source is collected by a Cassegrain telescope and focused on the spectrograph's entrance slit receiving system; the detector is a linear image sensor featuring an array of 512 MOS photodiodes. Data recorded in the same and boundary areas by a conventional analyzer from city's air-pollution monitoring network are reported for comparison. The statistical correlation of concentration values to the main weather and atmospheric stability parameters are stressed.

Remote sensing techniques have been developed to measure concentrations of atmospheric pollutants in the vicinity of diffuse pollution sources. In order to derive emission rates from these measurements numerical dispersion models have to be used. Three main types of dispersion models are currently available: Gaussian dispersion models, Eulerian models, and Lagrangian models. Gaussian models base on an analytical solution of the diffusion equation which describes the horizontal and vertical mean and turbulent transport of airborne matter from a source. Eulerian models require the definition of a grid in the volume of interest. At these grid points the budget equations for mass, momentum, heat, moisture, and pollutants will be solved numerically. Lagrangian models do not need a predescribed grid.Here the budge equations will be solved for particles or very small atmospheric volumes moving with the mean wind. For the application of Lagrangian models the wind field must be known a priori from measurements or from model simulations. Advantages, disadvantages, and application of these models is discussed in this paper.

A simple layout for a DOAS system based on a commercially available spectrometer with CCD-detector is presented. The open-path light beam over several kilometers is controlled automatically to maintain an optimal beam alignment. The DOAS system is operated in parallel with a scintillation anemometer (scidar), which optically measures the path averaged crosswind. This arrangement provides fluxes of atmospheric constituents under the constrain of homogeneous concentrations along the light paths. In the framework of the EU-project 'VOTALP' 5 combined systems were used to evaluate the budget of NO₂, O₃, formaldehyde, toluene and SO₂ fluxes in the Mesolcina valley in Switzerland. Two scidar/DOAS systems were set up in different heights across the northern end of the valley, two across the southern end and one parallel to the valley sidewall. Beam lengths ranged from 0.6 to 2.7 km. Results of the flux measurements are presented.

Inverse modeling with a Gaussian dispersion model is used to determine emission rates of diffuse sources from experimental data. The path-integrated gaseous concentrations of an exhaust are measured downwind of the source by open-path FTIR absorption spectroscopy. Additionally, a Doppler mini sodar was used to determine the vertical wind profile and an aerosol lidar to measure the spatial distribution of particulate emissions. The particle- size distribution was contributed by a high-volume aerosol impactor. Results of emission investigations form a piggery are discussed.

Agriculture is looked at contributing over 90 percent of the known NH₃ emission in Germany. A large share is due to spreading liquid manure; at sunshine, the loss off nitrogen ia NH₃ emissions may be up to 90 percent compared to the NH₄-bound nitrogen in the slurry. In a practice-oriented experiment, four different spreading techniques were compared. Four plots 36 m wide were laid out along the two main wind directions in a valley. In between two plots each an open path Fourier-transform IR (FTIR) spectrometer was placed on a rotating bearing to allow 180 degree swings on the altogether four IR sources. In combination with meteorological data emission rates could be estimated and compared. Even at rather bad weather conditions the application technics could be classified clearly by their ammonia emission.

The Global Ozone Monitoring Experiment (GOME) is a new atmospheric chemistry instrument on-board the ERS-2 satellite which was launched in April 1995. The GOME is designed to measure a range of atmospheric trace constituents, with particular emphasis on global ozone distributions. The ground segment for the GOME sensor is with the German Remote Sensing Data Center (DFD). Major components of the GDP are the complete GOME data archive, the Earth-shine spectra calibration step, the total ozone column retrieval process, and the integration into the D-PAF data management system (DMS). Raw GOME data re converted into 'calibrated radiances' during the Level 0 to 1 processing by applying a series of calibration algorithms using in-flight observations and pre-flight instrument calibration parameters. Total column abundances of ozone and other trace gases can be derived from the Level 1 Product, comprising the Earth-shine radiance and the extra- terrestrial solar irradiance, by applying three designated algorithms in the Level 1 to 2 processing step. The Initial Cloud Fitting Algorithm (ICFA) uses the spectral features close to and within the O₂ A-band around 760 nm to determine the fractional cloud cover of the pixel scene. The differential optical absorption spectroscopy technique is used for the operational retrieval of ozone and nitrogen dioxide form data in the UV and visible regions of the spectrum. The slant column densities are converted to vertical columns by division with an appropriate Air Mass Factor (AMF), derived from radiative transfer simulations. If clouds are detected by ICFA, an averaged AMF is calculated from the intensity-weighted AMFs to ground and to cloud top. Since the end of July 1996 the GOME data processing is performed operationally at the DFD.

One major problem in open path Fourier transform IR spectroscopy is the generation of a suitable background single beam spectrum. An overview over widely used methods with advantages and disadvantages will be given. In addition the shifting method as a new derivative technique will be introduced. A set of guidelines as a result of an extensive study on the shifting method will be given.