Developing a dynamic correction mechanism for aethalometer results of actual urban aerosols
Yuzhe Zhang,Guorui Zhi,Wenjing Jin,Shijie Liu,Lei Wang,Zhengying Li,Rong Shi,Peng Zhang,Yun Shu,Jingnan Hu
DOI: https://doi.org/10.1016/j.atmosres.2021.105529
IF: 5.965
2021-06-01
Atmospheric Research
Abstract:<p>The aethalometer (AE) is usually used to monitor light-absorbing aerosols. However, since the AE must first collect particles on a filter tape and then detect the optical changes caused by the filter/particle system, the primary data reported by the AE are not in-situ absorption coefficients (<em>b</em><sub>insitu</sub>), but light attenuation coefficients (<em>b</em><sub>ATN</sub>). The shadow effect, multiple scattering effect, and particle-scattering effect combine to generate artifacts. The intent of the present study was to develop a new correction methodology that could dynamically translate <em>b</em><sub>ATN</sub> into <em>b</em><sub>insitu</sub>. A photoacoustic extinctiometer (PAX), which is usually accepted as a reference instrument for <em>b</em><sub>insitu</sub> measurement, was concomitantly operated with an AE situated in wintertime Beijing, China. (i) During the observation period, the AE collected 303 consecutive quartz spots (<em>n</em> = 1, 2, 3..0.303) laden with particles. Each spot had a series of consecutive attenuation readings (<em>ATN</em><sub>n,i</sub>(λ), <em>i</em> = 1, 2, 3....) dependent on the ambient black carbon (BC) concentrations. Based on the 11-spot moving average of the compensation factors, <em>f</em><sub>moving,n</sub>(λ), a dynamic shadow effect correction coefficient was devised (Eq. <a class="workspace-trigger" href="#fo0050">(10)</a>). (ii) The <em>b</em><sub>ATNn,i</sub> for each of seven wavelengths could be calculated from <em>ATN</em><sub>n,i</sub>(λ), and the absorption Ångström exponent <em>α</em><sub>n,i</sub> was derived by considering the relationship between <em>b</em><sub>ATN,n,i</sub>(λ) and λ. (iii) Meanwhile, the PAX continuously recorded <em>b</em><sub>insitu,n,i</sub>. The composite scattering correction coefficient, <em>CC</em><sub>n,i</sub>(λ), could be calculated as Eq. <a class="workspace-trigger" href="#fo0055">(11)</a>. (iv) A liner relationship was discovered between the <em>CC</em>(λ) and <em>α</em>. At 880 nm, for example, the relationship was <em>CC</em>(λ) = 1.79<em>α</em> + 2.63 (<em>r</em><sup>2</sup> = 0.90). (v) Apparently, the two key parameters, <em>R</em><sub>n,i</sub>[<em>f</em><sub>moving,n</sub>(λ)] and <em>CC</em><sub>n,i</sub>(λ), are both derivable from the AE measurement data, which allows the translation from <em>b</em><sub>ATN,n,i</sub> to <em>b</em><sub>insitu,n,i</sub> (Eq. <a class="workspace-trigger" href="#fo0025">(5)</a>) from the AE alone, independent of other instruments. Applying the dynamic correction algorithm to the data from a different time period revealed that <em>b</em><sub>cor,n,i</sub> agreed well with <em>b</em><sub>insitu,n,i</sub>, directly supporting our correction algorithm. Future studies are proposed to address the ways in which the methodology is temporally or regionally dependent, and whether the aerosols from sources or affected by dust advection also fall within the scope of this correction algorithm.</p>
meteorology & atmospheric sciences