Estimation of the vertical distribution of particle matter (PM<sub>2.5</sub>) concentration and its transport flux from lidar measurements based on machine learning algorithms
<p>The vertical distribution of aerosol extinction coefficient (EC) measured by lidar systems has been used to retrieve the profile of particle matter with a diameter <span class="inline-formula"><2.5</span> <span class="inline-formula">µm</span&...
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
Copernicus Publications
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/a0df4129578d4bec8fe1a53740265387 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | <p>The vertical distribution of aerosol extinction
coefficient (EC) measured by lidar systems has been used to retrieve the
profile of particle matter with a diameter <span class="inline-formula"><2.5</span> <span class="inline-formula">µm</span>
(PM<span class="inline-formula"><sub>2.5</sub></span>). However, the traditional linear model (LM) cannot consider the
influence of multiple meteorological variables sufficiently and then
induce the low inversion accuracy. Generally, the machine learning (ML)
algorithms can input multiple features which may provide us with a new way
to solve this constraint. In this study, the surface aerosol EC and
meteorological data from January 2014 to December 2017 were used to explore
the conversion of aerosol EC to PM<span class="inline-formula"><sub>2.5</sub></span> concentrations. Four ML
algorithms were used to train the PM<span class="inline-formula"><sub>2.5</sub></span> prediction models:
random forest (RF), <span class="inline-formula"><i>K</i></span>-nearest neighbor (KNN), support vector machine (SVM) and extreme gradient boosting decision tree (XGB). The mean absolute error
(root mean square error) of LM, RF, KNN, SVM and XGB models were 11.66
(15.68), 5.35 (7.96), 7.95 (11.54), 6.96 (11.18) and 5.62 (8.27) <span class="inline-formula">µg</span>/m<span class="inline-formula"><sup>3</sup></span>, respectively. This result shows that the RF model is the most
suitable model for PM<span class="inline-formula"><sub>2.5</sub></span> inversions from EC and meteorological data.
Moreover, the sensitivity analysis of model input parameters was also
conducted. All these results further indicated that it is necessary to
consider the effect of meteorological variables when using EC to retrieve
PM<span class="inline-formula"><sub>2.5</sub></span> concentrations. Finally, the diurnal and seasonal variations of
transport flux (TF) and PM<span class="inline-formula"><sub>2.5</sub></span> profiles were analyzed based on the lidar
data. The large PM<span class="inline-formula"><sub>2.5</sub></span> concentration occurred at approximately
13:00–17:00 local time (LT) in 0.2–0.8 km. The diurnal variations of
the TF show a clear conveyor belt at approximately 12:00–18:00 LT in
0.5–0.8 km. The results indicated that air pollutant transport over Wuhan
mainly occurs at approximately 12:00–18:00 LT in 0.5–0.8 km. The TF near
the ground usually has the highest value in winter (0.26 mg/m<span class="inline-formula"><sup>2</sup></span> s),
followed by the autumn and summer (0.2 and 0.19 mg/m<span class="inline-formula"><sup>2</sup></span> s, respectively),
and the lowest value in spring (0.14 mg/m<span class="inline-formula"><sup>2</sup></span> s). These findings give us
important information on the atmospheric profile and provide us sufficient
confidence to apply lidar in the study of air quality monitoring.</p> |
|---|