Enhancing the Applicability of Satellite Remote Sensing for <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" id="M1"> <a:msub> <a:mrow> <a:mtext>PM</a:mtext> </a:mrow> <a:mrow> <a:mn>2.5</a:mn> </a:mrow> </a:msub> </a:math> Estimation Using Machine Learning Models in China
Jun Chai, Jun Song, Yawen Xu, Le Zhang, Bing Guo
Abstract
Numerous studies and monitoring data indicate that fine particle ( <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" id="M2"> <a:msub> <a:mrow> <a:mtext>PM</a:mtext> </a:mrow> <a:mrow> <a:mn>2.5</a:mn> </a:mrow> </a:msub> </a:math> ) pollution in China is still comparatively severe. Given the sparse and uneven distribution of air quality monitoring base stations established in China and the limitation of geographical conditions, inversion of aerosol optical depth by satellite remote sensing can achieve low-cost air quality monitoring in global areas. In this study, we use the machine learning algorithm XGBoost to build a prediction model to achieve nationwide average <c:math xmlns:c="http://www.w3.org/1998/Math/MathML" id="M3"> <c:msub> <c:mrow> <c:mtext>PM</c:mtext> </c:mrow> <c:mrow> <c:mn>2.5</c:mn> </c:mrow> </c:msub> </c:math> concentration prediction. Meanwhile, we used aerosol data from Moderate Resolution Imaging Spectroradiometer (MODIS) in a specific band, combined with a land use regression (LUR) model as predictors of surface <e:math xmlns:e="http://www.w3.org/1998/Math/MathML" id="M4"> <e:msub> <e:mrow> <e:mtext>PM</e:mtext> </e:mrow> <e:mrow> <e:mn>2.5</e:mn> </e:mrow> </e:msub> </e:math> concentrations in China, for the period Dec. 2019-Nov. 2021. In order to provide more accurate <g:math xmlns:g="http://www.w3.org/1998/Math/MathML" id="M5"> <g:msub> <g:mrow> <g:mtext>PM</g:mtext> </g:mrow> <g:mrow> <g:mn>2.5</g:mn> </g:mrow> </g:msub> </g:math> concentration prediction, the correspondence between <i:math xmlns:i="http://www.w3.org/1998/Math/MathML" id="M6"> <i:msub> <i:mrow> <i:mtext>PM</i:mtext> </i:mrow> <i:mrow> <i:mn>2.5</i:mn> </i:mrow> </i:msub> </i:math> and aerosol optical depth (AOD) under different seasons was studied. The coefficients of determination (R2) for different seasons are 0.86 (spring), 0.80 (summer), 0.90 (autumn), and 0.88 (winter), indicating that the fit is best for autumn and worse for summer. The study shows the potential usefulness of using the LUR model with the XGBoost algorithm for predictive assessment of <k:math xmlns:k="http://www.w3.org/1998/Math/MathML" id="M7"> <k:msub> <k:mrow> <k:mtext>PM</k:mtext> </k:mrow> <k:mrow> <k:mn>2.5</k:mn> </k:mrow> </k:msub> </k:math> spatial distribution.