demonstrated to have advantages in estimation of
PM2.5
. This
is because
GWR
model allows the relationships of variables to
vary across geographic space.
GTWR
extends the traditional
GWR
by considering temporal dimension, accounting for the
nonstationary variations in both space and time, which is
closer to the real conditions of the
PM2.5
variations.
A fundamental component of
GTWR
is the spatio-temporal
weight matrix which determines the contribution of each
individual observation in the dataset for the parameters
estimation at a specific location. In general, the observations
closer to estimated point over space and time have greater
weights during local relationships construction. The temporal
distance in
GTWR
is linear distance. However,
PM2.5
variations
have seasonal characteristics, which behave nonlinearly in
the temporal dimension. The observations in the same season
have greater weights. The temporal closeness between the
estimated point and observed points in
IGTWR
is measured by
incorporating linear distance and seasonal nonlinear distance.
Therefore, the
IGTWR
model can describe
PM2.5
variations bet-
ter than
OLS
and other
GWR
-based models.
Conclusions
This paper develops an improved
GTWR
model which incor-
porates seasonality into spatiotemporal modeling to retrieve
daily
PM2.5
concentrations in Hong Kong from 2012 to 2014
based on
AOD
data. Since the current
MODIS AOD
products at
10 km and 3 km resolutions cannot satisfy the requirement
in a small area, we employ the simplified aerosol retrieval
method to retrieve the
MODIS AOD
product with 500m spatial
resolution. In order to improve the prediction accuracy of
PM2.5
, the meteorological and land use parameters are added
as auxiliary data in
IGTWR
model. To assess the performance
of
IGTWR
, the
PM2.5
predictions are validated with ground
observations.
IGTWR
model is demonstrated to outperform
OLS
,
GWR
and
GTWR
with highest R
2
and lowest
RMSE
by capturing
the seasonal variations. The averaged
PM2.5
distribution at 500
m resolution from 2012 to 2014 is derived using
IGTWR
model
and
SARA AOD
. The
IGTWR
PM2.5
concentrations show similar
distribution with ground observations and reveal more details
of spatial variations of
PM2.5
. This research is significant for
PM2.5
predictions and can be useful for local government mak-
ing contingency plans for haze.
Acknowledgments
The authors would like to acknowledge the Environment
Protection Department of the government of Hong Kong
Special Administrative Region (
HKSAR
) for providing the
PM2.5
concentration data and Hong Kong Observatory for provid-
ing meteorological data. The authors also thank Mr. Hongyu
Liang for the programming support.
References
Akaike, H., 1974. A new look at the statistical model identification,
IEEE Transactions on Automatic Control
, 19(6):716-723.
Bai, Y., L. Wu, K. Qin, Y. Zhang, Y. Shen, and Y. Zhou, 2016. A
geographically and temporally weighted regression model for
ground-level pm2.5 estimation from satellite-derived 500 m
resolution AOD,
Remote Sensing
8(3):262.
Bilal, M., J.E. Nichol, M.P. Bleiweiss, and D. Dubois, 2013. A
Simplified high resolution MODIS aerosol retrieval algorithm
(SARA) for use over mixed surfaces,
Remote Sensing of
Environment
136:135-145.
Bilal, M., J.E. Nichol, and P.W. Chan, 2014. Validation and accuracy
assessment of a Simplified Aerosol Retrieval Algorithm (SARA)
over Beijing under low and high aerosol loadings and dust
storms,
Remote Sensing of Environment
, 153:50-60.
Bilal, M., and J.E. Nichol, 2015. Evaluation of MODIS aerosol retrieval
algorithms over the Beijing-Tianjin-Hebei region during low to
very high pollution events,
Journal of Geophysical Research:
Atmospheres
120(15):7941-7957.
Bilal, M., M. Nazeer, and J.E. Nichol, 2017. Validation of MODIS and
VIIRS derived aerosol optical depth over complex coastal waters,
Atmospheric Research
, 186:43-50.
Bilal, M., J.E. Nichol, and S.N. Spak, 2017. A new approach for
estimation of fine particulate concentrations using satellite
aerosol optical depth and binning of meteorological variables,
Aerosol and Air Quality Research
, 17(2):356-367.
Brunsdon, C., A.S. Fotheringham, and M.E. Charlton, 1996.
Geographically weighted regression: A method for exploring
spatial nonstationarity,
Geographical Analysis
28(4):281-298.
Chen, C.-C., C.-F.Wu, H.-L.Yu, C.-C. Chan, and T.-J.Cheng, 2012.
Spatiotemporal modeling with temporal-invariant variogram
subgroups to estimate fine particulate matter PM2.5
concentrations,
Atmospheric Environment
, 54:1-8.
Chu, H.-J., B. Huang, and C.-Y. Lin, 2015. Modeling the spatio-
temporal heterogeneity in the PM10-PM2.5 relationship,
Atmospheric Environment
, 102:176-182.
Clean Air Network (CAN), 2013.
2012 Hong Kong Air Quality Review
.
Elminir, H.K. 2005. Dependence of urban air pollutants on
meteorology,
Science of theTotal Environment,
350(1-3):225-237.
Environmental Protection Department, Emission Trends (1997-2015),
URL:
data/emission_inve.html,
last date accessed: 29 September 2018)
Fotheringham, A.S., C. Brunsdon, and M. Charlton, 2002.
Geographically Weighted Regression: The Analysis of Spatially
Varying Relationships
, Wiley, Chichester 284 p.
Ghosh, J.K., M. Wilhelm, J. Su, D. Goldberg, M. Cockburn, M. Jerrett,
M. and B. Ritz, 2012. Assessing the influence of traffic-related
air pollution on risk of term low birth weight on the basis of
land-use-based regression models and measures of air toxics,
American Journal of Epidemiology
175(12):1262-1274.
Guo, H., 2003. Particle-associated polycyclic aromatic hydrocarbons
in urban air of Hong Kong,
Atmospheric Environment
,
37(38):5307-5317.
Guo, Y., Q. Tang, D.-Y. Gong, and Z. Zhang, 2017. Estimating ground-
level PM 2.5 concentrations in Beijing using a satellite-based
geographically and temporally weighted regression model,
Remote Sensing of Environment,
198:140-149.
He, Q., and B. Huang, 2018. Satellite-based high-resolution PM2.5
estimation over the Beijing-Tianjin-Hebei region of China using
an improved geographically and temporally weighted regression
model,
Environmental Pollution,
236:1027-1037.
Huang, B., B. Wu, and M. Barry, 2010. Geographically and temporally
weighted regression for modeling spatio-temporal variation in
house prices,
International Journal of Geographical Information
Science
24(3):383-401.
Levy, R.C., S. Mattoo, L.A. Munchak, L.A. Remer, A.M. Sayer, F.
Patadia, and N.C. Hsu, 2013. The collection of MODIS aerosol
products over land and ocean,
Atmospheric Measurement
Techniques,
6(11):2989-3034.
Li, X., Y.J. Feng, and H.Y. Liang, 2017. The impact of meteorological
factors on PM2.5 variations in Hong Kong,
IOP Conference
Series: Earth and Environmental Science
78:012003.
Liu, Y., C.J. Paciorek, and P. Koutrakis, 2009. Estimating regional
spatial and temporal variability of PM(2.5) concentrations
using satellite data, meteorology, and land use information,
Environmental Health Perspective,
117(6):886-892.
Malakar, N.K., D.J. Lary, A. Moore, D. Gencaga, B. Roscoe, A.
Albayrak, and J. Wei, 2012. Estimation and bias correction of
aerosol abundance using data-driven machine learning and
remote sensing,
Intelligent Data Understanding (CIDU)
).
Mao, L., Y. Qiu, C. Kusano, and X. Xu, 2012. Predicting regional
space-time variation of PM2.5 with land-use regression model
and MODIS data,
Environmental Science and Pollution Research
International,
19(1):128-138.
768
December 2018
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
