Yael Etzion : PhD Student

Supervisors: Dr. David Broday, Dr. Zafrir Kulat and Prof. Maxim Shoshany

Real time characterization of ground-level ambient pollutant concentrations and their spatial distributions is important for environmental risk assessment and public health considerations. At present, the answer for that need is mainly based on spatial interpolations of discrete local readings from air quality monitoring networks (AQMN) whose limited surface density casts doubt on the reliability of these interpolations. While laboratory equipment has already evolved to provide molecular scale analysis of air samples, real time environmental monitoring of specific pollutants, especially particulate matter (PM), is somewhat restricted. In order to evaluate real-time health and environmental risks in ambient air there is a need for a more precise information regarding pollutant characteristics. The latter, also known as the “signature”, comprises complex chemical and physical attributes, such as the chemical composition, the ratios among specific components, the thermo-physical state, the particle size distribution and morphology, etc. Currently, monitoring of air pollutants concentrates on greenhouse gases (GHGs), urban smog related and health risking pollutions (O₃, NO_x, SO₂, benzene, toluene, and xylene). Environmental stressors (halogen oxides, CFCs, etc.) are also monitored, but at a much larger sampling interval (seasonal, annual). Separately monitored, either locally or on a global scale, is suspended particulate matter (PM) of different size classes. Exposure to PM has been associated with adverse health effects.

Electro-optical technologies enable real time remote sensing of pollutants based on their interaction with electromagnetic (EM) radiation which is emitted either from the sun, natural blackbodies, or artificial radiation sources. Radiation scattering and absorbance by gaseous and particulate pollutants are imprinted in the EM spectrum as spectral signatures. While chemical composition signatures are usually imprinted by radiation adsorption in the mid-IR range of the EM radiation (2.5-25µm), physical attributes such as size and shape, which are relevant to PM, are mostly imprinted by radiation scattering in the visible-NIR range (0.4-1.2µm). Atmosphere pollutants show high variability in space and time, and often mark only a subtle signature of complex nature that can be difficult to reveal. Hyperspectral imaging offers the possibility to recognize a variety of unique spectral signatures, assembled as spatial data by measuring tens and hundreds of frequency bands simultaneously for every pixel in the image. However, hyperspectral remote sensing (HRS) is restricted by the radiation source used. When using solar radiation, the illumination conditions are not stable and the time period for detection is limited. Yet the use of artificial light sources restricts the spatial range of detection. The key questions to be answered after acquiring hyper-spectra are to what degree these spectra represent their surrounding, how to decipher the physical interpretation out of the signal and at which temporal and spatial scale remote sensing can be used. Nowadays, spectral remote sensing is mostly obtained by satellite-borne and ground based platforms of sunphotometers and spectroradimeters. In all these configurations the signals recorded are not always relevant to air pollution near the ground, since the measurement is done through the entire atmosphere and at limited revisit periods. Data from aviated sensors, though available, is inclined to instability because of aircraft vibrations and shifts, and it is also practically less accessible. Assessing air quality by ground based hyperspectral imaging camera and using a quasi-horizontal open path can offer a new tool with the advantages of spatial and temporal resolution of imaging through the atmospheric layer of most interest. Moreover, since currently remote spectral measurements of ambient PM exploit solar radiation, they are restricted to daytime. By imaging remote artificial illumination by a hyperspectral camera in the visible range we aim to examine the possibility for spatial monitoring of ambient PM near the ground after sundown.