Estimating the Effects of Lock-Down for the First and Second Waves of COVID-19 on Air Pollution indices in the Iranian Provinces with Google Earth Engine Technology

doi:10.21203/rs.3.rs-2251698/v1

Download PDF

Research Article

Estimating the Effects of Lock-Down for the First and Second Waves of COVID-19 on Air Pollution indices in the Iranian Provinces with Google Earth Engine Technology

https://doi.org/10.21203/rs.3.rs-2251698/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Introduction: The COVID-19 restrictions provided a unique situation for environmental policymakers to examine their hypothesis and present alternative policies for controlling air pollution indices globally. The aim of this study is to estimate the COVID-19 lockdown effects on the air pollution with satellite dataset indices in all 31 provinces of Iran in the first wave from 2020-03-14 to 2020-04-20 vs from 2019-03-14 to 2019-04-20 and second wave from 2021-04-13 to 2021-04-25 vs from 2020-04-13 to 2020-04-25.

Methods: The Carbon Monoxide (CO), Water vapor (H2O), Nitrogen Dioxide (NO2), Ozone (O3), Sulfur Dioxide (SO2), Absorbing Aerosol Index (AER), and Atmospheric Formaldehyde (HCHO) as air pollution indices and Pressure, Total precipitation rate, Air temperature and Wind speed as weather indices are downloaded from Google Earth Engine (GEE). Three statistical analysis methods are used: 1) T-test and Wilcoxon rank-sum test for estimating the Unadjusted Effect, 2) Generalized Additive Mixed Model (GAMM) for adjusted Effects and 3) Functional clustering for clustering all provinces.

Result: The first wave of the COVID-19 lock-down improves the air quality against the same period in the previous year but the same conclusion is not inferred from data for the second wave. The reduction of CO (-0.00077), NO2 (-0.0000110) and AEI (-0.275) and increment of H2O (106.476), O3 (0.0042) and SO2 (0.000104) are statistically significant in the first wave. But the reduction of CO (-0.00082), H2O (-354.52), O3(-0.00258) and increment of NO2 (0.0000089) and HCHO (0.0000135) are statistically significant in the second wave with GAMM models. The functional clustering with functional principal component analysis (FPCA) shows three major clusters of provinces.

Conclusions: We conclude that although the air quality in the first wave is better than the second wave, the Containment and Health Index in the second wave is greater than the first wave.

Air Pollution

COVID-19

Iran

Generalized Additive Mixed Model

Functional Data Analysis

Google Earth Engine

Long and short exposure to air pollution like PM 2.5 increased the chance of COVID-19 infection and mortality rates in the US (Wu, Nethery et al. 2020, Wu, Nethery et al. 2020), China (Ma, Zhao et al. 2020, Xie and Zhu 2020, Zhu, Xie et al. 2020), Italy (Conticini, Frediani et al. 2020) with Spain, France & Germany (Ogen 2020), Asia (Baniasad, Mofrad et al. 2021) and other countries (Pansini and Fornacca 2021). It also acts as a confounder variable in studying the relationship between racism and COVID-19 diagnosis and deaths in the US(Millett, Jones et al. 2020). Not only air pollution but also weather conditions such as high temperatures have significant effects on COVID-19 infectious in the world (Scafetta 2020).

The COVID-19 restrictions on transportation and industrial activities provided a unique situation for environmental policymakers to examine their hypothesis and present alternative policies for controlling air pollution indices globally; The fast decrease in air pollution indices in India is observed (Shehzad, Sarfraz et al. 2020) and they suggest using a two or four days lock-down policy for controlling air pollution with considering other factors like economics and seasonal variations (Mahato, Pal et al. 2020). The economic effects of air pollution during COVID-19 are studied and some suggestions for stringent environmental policies are discussed (Bashir, Ma et al. 2020). On the other hand, these restrictions did not decrease the severe air pollution in some regions of China due to the metrological and chemical conditions (Wang, Chen et al. 2020). But lock-down policies have a positive impact on the perception of people around the world about air quality improvement (Barbieri, Lou et al. 2020, Lou, Barbieri et al. 2022).

New technologies measure different atmospheric conditions; for example, monitoring air pollution with the AURA satellite from NASA EOS (Earth Observation System) in China and they recognize the reduction of CO and NO2 during COVID-19 because of transportation restrictions and decrease the economic activities (Filonchyk, Hurynovich et al. 2020). Decreasing of PM2.5 with Moderate-resolution imaging spectroradiometer (MODIS) data in many countries (Bherwani, Kumar et al. 2021). Studies show that global air pollution declined using Google Earth Engine (GEE) and Apple mobility data(Venter, Aunan et al. 2020). It is statistically significant in different air pollution indices in Iran and its neighboring countries. (Fayaz 2022).

The geographic region of this study is Iran and all 31 provinces during the first and second waves of national COVID-19 lockdown days. We compare the air pollution and weather indices during the lockdown days and the previous year. The statistical analysis has three steps, including 1) Comparing the average values of indices with the T-test and nonparametric Wilcoxon rank-sum test, 2) Adjusting the air pollution indices with generalized additive mixed models (GAMM) (Wood 2006, Harezlak, Ruppert et al. 2018) and weather conditions and estimating the lock-down effects and 3) hierarchal clustering of the provinces with Functional Principal Components Analysis (FPCA) with principal components analysis through the conditional expectation (PACE) method (Yao, Müller et al. 2005) and Functional Clustering (Chiou and Li 2007). The result and conclusion follow after them.

The methods describe briefly in 4 subsections.

Data Gathering And Management

Iran has 31 provinces (Wikipedia.org) and their shape files are available from ArcGIS Online (www.arcgis.com). The Carbon Monoxide (CO), Water vapor (H₂O), Nitrogen Dioxide (NO₂), Ozone (O₃), Sulfur Dioxide (SO₂), Absorbing Aerosol Index (AER), and Atmospheric Formaldehyde (HCHO) from Sentinel-5P: Near Real-Time (NRTI) Satellite are air pollution indices and Pressure, Total precipitation rate, Air temperature and Wind speed from NASA Global Land Data Assimilation System Version 2 (GLDAS-2) are weather indices of this study. They are downloaded with Google Earth Engine Data Catalog (https://developers.google.com/earth-engine/datasets). (Table A1_1 in Appendix A)

The analysis time period is: 1) comparing the first wave (from 2020-03-14 to 2020-04-20) with the previous year (from 2019-03-14 to 2019-04-20) and 2) comparing the second wave (from 2021-04-13 to 2021-04-25) with the previous year (from 2020-04-13 to 2020-04-25) of lock-down days in Iran. (Wikipedia.org) (Table A1_2 in Appendix A)

Unadjusted Effect: T-test And Wilcoxon Rank-sum Test

The summary statistics and statistical tests are independent sample t-test in two conditions for equal/unequal variance, F test for variance comparisons, Shapiro-Wilk Normality Test and nonparametric Wilcoxon Rank Sum test with R version 4.1.3 (Team 2022). The hypothesis is:

$$\left\{\begin{array}{c}{H}_{0}: {\mu }_{COVID19}={ \mu }_{Previous Year}\\ {H}_{1}:{\mu }_{COVID19}\ne { \mu }_{Previous Year}\end{array}\right. \left(1\right)$$

for each air pollution indices of each province in first and second lock-down waves. They are unadjusted sample averages because they don’t adjust with any covariates.

Adjusted Effects: Generalized Additive Mixed Model

The weather conditions are not the same in the two time period of analysis, therefore a recent study adjust the air pollution indices with meteorological variables and simple linear regression (Venter, Aunan et al. 2020). The limitation of this adjustment is 1) considering only the linear effects and 2) country-specific estimates. The Generalized Additive Mixed models (GAMM) can capture the nonlinear effects for example with smoothing spline, $f(.)$, and it estimates the overall random effects for provinces. In this regard, the confounding variables are pressure, total precipitation rate, Air temperature, and Wind speed of the time-period of analysis. The model is:

The COVID Status covariate have two values (0- Previous Year and 1- COVID-19) and the provinces, ${U}_{i}$, enter the model as random effect. The model comparisons are with adjusted ${R}^{2}$, Akaike information criterion (AIC) and Bayesian information criterion (BIC). All calculations are with gamm() and s() functions in mgcv package with R (Wood 2006, Wood and Wood 2015, Harezlak, Ruppert et al. 2018)

Clustering: Functional Data Analysis

The functional clustering (FC) methods from Functional Data Analysis (FDA) toolbox (Ramsay, Silverman et al. 2005) are new and promising methods for air quality and meteorological time series studies such FC based on the Partitioning Around Medoids (PAM) algorithm for air quality monitoring network of NO₂, PM₁₀ and O₃ and functional hidden dynamic geostatistical models in Italy (Ignaccolo, Ghigo et al. 2008, Maranzano, Otto et al. 2022), for climate zone identifications (Di Giuseppe, Jona Lasinio et al. 2013) and Functional Zoning (Ignaccolo, Ghigo et al. 2013). The FC methods are also widely used in environmental statistics such as: Yield Pattern Analysis (Pascucci, Carfora et al. 2018), river network data (Haggarty, Miller et al. 2015, Jamaludin 2017) and water quality functional spatial clustering with calibrating Ocean Colour data from MERIS, SeaWiFS and MODIS satellite data (Maritorena, d'Andon et al. 2010, Carlo, Paolo et al. 2017). The curves and time series observation have dependency to each other, and FDA methodology consider this structure with advanced statistical methods.

The air pollution indices ,${X}_{ijwC}\left(t\right)$, are functional data with ${\{X}_{ijwC}\left(t\right), i=1,\dots ,7 , j=1,\dots ,31, w=\text{1,2} , C=\text{1,2} \},$ and $i$ is air pollution indices, $j$ is the province name, $w$ is the wave (first and second waves) and $C$ is the COVID-19 status (Previous year and COVID-19) (Ramsay, Silverman et al. 2005). Air pollution indices have different units and in order to enter them in the model, the preprocessing steps for ${X}_{ijwC}\left(t\right)$are scaling and centering. Then, the functional average of all air pollution indices, ${\stackrel{-}{X}}_{jwC}\left(t\right)$, for each province, waves and covid status are:

Therefore, each province has four new functional variables including ${\stackrel{-}{X}}_{j11}\left(t\right)$ for the first wave and previous year, ${\stackrel{-}{X}}_{j12}$ for the first wave and COVID-19 year, ${\stackrel{-}{X}}_{j21}\left(t\right)$ for the second wave and previous year and ${\stackrel{-}{X}}_{j22}\left(t\right)$ for the second wave and COVID-19 of each province $j=1,\dots ,31$. If $w=1$, then $t \in \left[\text{1,37}\right]$ and if $w=2$, then $t\in \left[\text{1,12}\right]$.

In the next step, the functional principal component analysis (FPCA) for each of ${\{\stackrel{-}{X}}_{j11}\left(t\right),{\stackrel{-}{X}}_{j12},{\stackrel{-}{X}}_{j21}\left(t\right),{\stackrel{-}{X}}_{j22}\left(t\right), for j=1,\dots ,31\}$ among all provinces decompose them and reduce their dimensions (Yao, Müller et al. 2005). The functional variables have some missing values at some observation time points and they are spare, and the principal components analysis through conditional expectation (PACE) approach (Yao, Müller et al. 2005) is FDA methodology to deal with them (Ramsay, Silverman et al. 2005). The summary statistics of ${\stackrel{-}{X}}_{jwC}\left(t\right)$are:

$${E\stackrel{-}{X}}_{wC}\left(t\right)={\mu }_{wC} t \in T \left(4\right)$$

$${COV(\stackrel{-}{X}}_{wC}\left(s\right),{\stackrel{-}{X}}_{wC}\left(t\right))={G}_{wC}\left(s,t\right) t,s \in T (5)$$

The decomposition of the covariance is:

$$\left\{\begin{array}{c}{G}_{wC}\left(s,t\right)=\sum _{k}{\lambda }_{kwC}{\varphi }_{kwC}\left(s\right){\varphi }_{kwC}\left(t\right) \\ {\varphi }_{kwC}\left(.\right):eigenfunctions \\ {\lambda }_{kwC} :eigenvalues \\ t,s \in T\end{array} \left(6\right)\right.$$

Now, the $j$th random curve is approximately equivalent to the ${\stackrel{-}{X}}_{jwC}\left(t\right)={\mu }_{wC}\left(t\right)+ \sum _{k}{\xi }_{jkwC}{\varphi }_{kwC}\left(t\right)$where ${\xi }_{jkwC}$ are uncorrelated random variables(Yao, Müller et al. 2005). The $k$ is estimated with the 95% fraction of variance explained (FVE). Finally, the clustering procedure is clustering the ${\widehat{\xi }}_{jkwC}$ with hierarchal clustering algorithm and EM Clustering with hclust() from stats package and FClust() from fdapace R package (Carroll, Gajardo et al. 2020). The result plot is ${\widehat{\xi }}_{jk11} vs {\widehat{\xi }}_{jk12}$ and ${\widehat{\xi }}_{jk21} vs {\widehat{\xi }}_{jk22}$.

Unadjusted Effects

The Table 1 and 2 show the estimated mean differences and growth rate (%) of all air pollution indices with parametric and nonparametric p-value tests between the COVID-19 lock-down days and the same period previous year for the first and second waves, separately.

Table 1

The estimated mean differences, growth rate (%) and the P-values group by provinces for the first wave of COVID-19.
	CO	H₂O	NO²	O³	SO²	AEI	HCHO
01_AL	-0.0007; -2.50% (0.22;0.06)	54.3; 9.55% (0.42;0.1)	-0.000075; -37.44% (0.14;0.61)	0.0022; 1.46% (0.41;0.22)	-0.00003; -17.54% (0.72;0.62)	-0.31; 42.33% (0;0)	0.0000041;6.83% (0.68;0.53)
02_AR	-0.0018; -5.86% (0;0.01)	-44.4; -6.36% (0.43;0.52)	-0.000011;-14.9% (0;0)	0.003; 1.92% (0.21;0.24)	0.000004; 5.26% (0.94;0.67)	-0.22; 24.29% (0.03;0.01)	-0.0000042;-7.65% (0.74;0.87)
03_AE	-0.0012; -4.14% (0;0)	52.8; 9.72% (0.19;0.31)	-0.000009;-12.53% (0;0)	0.0021; 1.34% (0.38;0.43)	0.000213;191.89% (0.35;0.98)	-0.27; 32.54% (0;0)	0.0000067;19.2% (0.35;0.71)
04_AW	-0.0008; -2.84% (0.08;0.12)	74.4; 15.25% (0.05;0.06)	-0.000008;-11.79% (0;0)	0.0015; 0.96% (0.54;0.66)	0.000135;195.65% (0.34;0.93)	-0.27; 33.53% (0;0)	0.0000126;46.67% (0.15;0.52)
05_BU	-0.0002; -0.60% (0.59;0.85)	-118.7;-8.28% (0.34;0.77)	-0.000001;-1.67% (0.64;0.39)	0.008; 6.28% (0;0)	0.000099;48.06% (0.17;0.28)	-0.15; 48.57% (0.21;0.1)	0.0000021;2.56% (0.81;0.97)
06_CB	-0.0005; -1.96% (0.31;0.5)	49.8; 9.89% (0.46;0.15)	-0.000004;-6.75% (0.02;0.01)	0.0054; 3.96% (0.01;0.02)	0.000043; 74.14% (0.46;0.33)	-0.32; 68.65% (0;0)	-0.000003;-7.81% (0.67;0.6)
07_FA	-0.0004; -1.41% (0.29;0.35)	-76.3; -7.99% (0.38;0.76)	-0.000001;-1.63% (0.68;0.09)	0.0086; 6.81% (0;0)	0.000002; 1.8% (0.96;0.35)	-0.31; 64.86% (0;0)	-0.0000063; -9.98% (0.29;0.2)
08_GI	-0.0012; -3.69% (0.05;0.03)	0.8; 0.09% (0.99;0.77)	-0.000014; -14.27% (0.05;0.01)	0.0028; 1.8% (0.28;0.33)	0.00009; 128.57% (0.27;0.26)	-0.12; 14.34% (0.28;0.26)	-0.0000026; -4.19% (0.87;0.77)
09_GO	-0.0002; -0.61% (0.77;0.3)	19.9; 2.02% (0.83;0.63)	-0.000008; -10.99% (0;0)	0.0049; 3.26% (0.06;0)	-0.00003; -26.55% (0.63;0.43)	-0.21; 27.3% (0.03;0.02)	-0.0000174; -23.93% (0.08;0.07)
10_HA	-0.0007; -2.57% (0.12;0.17)	60.1; 11.27% (0.26;0.17)	-0.000012; -15.99% (0.01;0)	0.0018; 1.21% (0.48;0.47)	0.000054; 79.41% (0.53;0.82)	-0.39; 54.11% (0;0)	0.0000025;7.02% (0.76;0.62)
11_HO	-0.0001; -0.31% (0.77;0.98)	10.4; 0.7% (0.94;0.87)	-0.000004; -6.8% (0.01;0.01)	0.0096; 7.85% (0;0)	0.000031; 39.24% (0.4;0.38)	-0.22; 97.56% (0.05;0.05)	-0.0000021; -2.74% (0.71;0.68)
12_IL	-0.0006; -1.86% (0.16;0.23)	27; 2.62% (0.8;0.68)	-0.000007; -10.73% (0;0)	0.0019; 1.31% (0.44;0.46)	0.00024; 135.59% (0.23;0.31)	-0.28; 40.89% (0.02;0.01)	-0.0000051; -8.79% (0.58;0.93)
13_IS	-0.0005; -1.75% (0.35;0.4)	110.6; 16.04% (0.15;0.05)	-0.000005; -6.38% (0.01;0.02)	0.0064; 4.62% (0;0)	-0.000009; -9.68% (0.83;0.8)	-0.37; 68.71% (0;0)	0.0000009;1.52% (0.88;0.56)
14_KE	-0.0001; -0.35% (0.75;0.49)	51.4; 5.43% (0.58;0.56)	-0.000004; -5.65% (0.04;0.02)	0.0099; 7.84% (0;0)	0.00003; 34.09% (0.36;0.06)	-0.31; 64.07% (0;0)	-0.0000087; -12.61% (0.14;0.16)
15_KM	-0.0009; -3.04% (0.03;0.05)	55; 7.41% (0.45;0.51)	-0.000007; -9.75% (0.01;0.01)	0.0008; 0.54% (0.75;0.96)	0.000122; 137.08% (0.31;0.68)	-0.3; 35.71% (0;0)	0.0000093;24.35% (0.3;0.2)
16_KN	-0.0008; -2.67% (0.05;0.05)	-50.5; -6.07% (0.51;0.91)	-0.000006; -9.28% (0.02;0)	0.0055; 3.74% (0.04;0)	-0.000068; -144.68% (0.13;0.13)	-0.34; 48.57% (0;0)	-0.0000032; -5.96% (0.72;0.31)
17_KR	-0.0006; -2.00% (0.08;0.17)	-32.3; -3.29% (0.72;0.95)	-0.000004; -5.71% (0.04;0.02)	0.0071; 5.02% (0;0)	0.000021; 39.62% (0.58;0.49)	-0.4; 59.73% (0;0)	-0.0000043; -7.58% (0.49;0.57)
18_KS	-0.0004; -1.38% (0.28;0.32)	48.8; 5.06% (0.56;0.47)	-0.000003; -4.21% (0.07;0.09)	0.01; 7.57% (0;0)	0.000004; 10% (0.9;0.04)	-0.41; 84.27% (0;0)	-0.0000073; -13.06% (0.32;0.73)
19_KH	-0.0003; -0.89% (0.47;0.57)	-82.2; -6.08% (0.59;0.99)	-0.000002; -3.12% (0.33;0.14)	0.0047; 3.41% (0.02;0.02)	0.00023; 138.55% (0.11;0.06)	-0.17; 29.29% (0.09;0.02)	-0.0000012; -1.28% (0.88;0.92)
20_KB	-0.0008; -2.86% (0.2;0.39)	71.5; 8.92% (0.51;0.43)	-0.000005; -7.79% (0;0)	0.0068; 5.17% (0;0)	0.000099; 76.15% (0.22;0.34)	-0.22; 40.79% (0.02;0.01)	-0.0000152; -27.89% (0.15;0.51)
21_KU	-0.001; -3.62% (0.01;0.01)	38.2; 7.46% (0.4;0.33)	-0.000007; -10.85% (0.01;0)	0.0018; 1.18% (0.48;0.7)	0.000134; 291.3% (0.23;0.13)	-0.36; 46.36% (0;0)	0.0000143;71.5% (0.07;0.14)
22_LO	-0.0007; -2.50% (0.18;0.15)	30.5; 4.42% (0.71;0.45)	-0.000006;-9.02% (0.01;0)	0.0021; 1.45% (0.4;0.47)	0.000106; 100% (0.41;0.87)	-0.43; 64.83% (0;0)	0.000003;7.65% (0.71;0.46)
23_MA	-0.0004; -1.47% (0.32;0.33)	63.9; 10.3% (0.38;0.11)	-0.000022;-23.4% (0.01;0.01)	0.0022; 1.5% (0.38;0.41)	0.000018; 12.08% (0.8;0.72)	-0.43; 58.47% (0;0)	0.0000072;14.66% (0.32;0.32)
24_MD	-0.0014; -4.68% (0.01;0.02)	0.5; 0.07% (1;0.63)	-0.000014; -13.82% (0.03;0.01)	0.004; 2.65% (0.13;0.05)	-0.000044; -23.16% (0.59;0.16)	-0.22; 28.13% (0.01;0.01)	-0.0000096; -10.12% (0.52;0.93)
25_QA	-0.0009; -3.16% (0.09;0.11)	109.1; 18.65% (0.08;0.09)	-0.000018; -17.41% (0.03;0.14)	0.0021; 1.39% (0.41;0.3)	-0.000002; -1.55% (0.99;0.74)	-0.28; 31.37% (0.01;0.01)	-0.0000099; -22.65% (0.38;0.79)
26_QO	-0.0001; -0.33% (0.82;0.78)	75.7; 8.55% (0.45;0.05)	-0.000029; -25.42% (0.02;0)	0.003; 2.05% (0.23;0.18)	-0.000027; -13.37% (0.72;0.75)	-0.3; 42.53% (0.01;0.01)	-0.0000061; -7.26% (0.38;0.31)
27_SE	-0.0007; -2.24% (0.05;0.05)	69.1; 8.35% (0.34;0.09)	-0.000007; -9.9% (0;0)	0.0059; 4.09% (0.02;0)	-0.00002; -23.53% (0.62;0.73)	-0.34; 51.93% (0;0)	-0.0000062; -9.41% (0.28;0.28)
28_SB	0; 0.00% (0.92;0.99)	-26.7; -2.17% (0.8;0.92)	-0.000004; -6.56% (0;0)	0.0106; 8.62% (0;0)	0.000003; 8.82% (0.9;0.47)	-0.3; 108.02% (0.01;0.01)	-0.0000056; -7.97% (0.22;0.25)
29_TE	-0.0009; -3.09% (0.1;0.14)	62; 10.27% (0.34;0.07)	-0.000056; -29.48% (0.01;0)	0.003; 2.02% (0.25;0.11)	0.000094; 40.34% (0.6;0.63)	-0.27; 37.61% (0;0)	-0.0000023; -2.96% (0.81;0.43)
30_YA	-0.0003; -1.02% (0.36;0.32)	105; 12.81% (0.16;0.12)	-0.000003; -5.12% (0.05;0.01)	0.0085; 6.35% (0;0)	-0.000001; -1.33% (0.98;0.35)	-0.41; 91.05% (0;0)	-0.0000033; -5.39% (0.46;0.5)
31_ZA	-0.0009; -3.23% (0.03;0.04)	74.2; 13.9% (0.16;0.11)	-0.000012; -16.26% (0.01;0)	0.0022; 1.44% (0.4;0.54)	0.00009; 225% (0.21;0.79)	-0.28; 32.84% (0;0)	0.0000033;8.11% (0.66;0.89)
Estimated Mean Differences [${\mu }_{COVID}-{\mu }_{Previous}]$, growth rate (%) (P-Value for T-test unequal Variance; P-Value for Wilcoxon )

Table 2

The estimated mean differences, growth rate (%) and the P-values group by provinces for the second wave of COVID-19.
	CO	H₂O	NO²	O³	SO²	AEI	HCHO
01_AL	-0.0013; -4.41% (0.17;0.09)	-242.3; -27.32% (0.23;0.22)	0.000024; 17.44% (0.2;0.15)	-0.0142; -9.24% (0;0)	-0.0001; -102.15% (0.23;0.38)	0.18; -15.77% (0.24;0.14)	0.0000312; 39.39% (0.05;0.08)
02_AR	-0.0003; -0.99% (0.71;0.98)	185.1; 24.32% (0.19;0.35)	0.000005; 7.07% (0.03;0.03)	-0.0155; -9.81% (0;0)	-0.00003; -41.43% (0.46;0.57)	-0.23; 19.91% (0.13;0.2)	0.000017; 30.8% (0.08;0.05)
03_AE	-0.0008; -2.72% (0.18;0.22)	29; 3.92% (0.84;1)	0.000006; 8.96% (0.01;0.02)	-0.0152; -9.7% (0;0)	-0.00004; -42% (0.52;0.51)	-0.03; 2.35% (0.81;0.76)	0.0000145; 30.08% (0.04;0.06)
04_AW	-0.0009; -3.12% (0.11;0.27)	52.9; 7.35% (0.75;0.8)	0.000007; 9.79% (0;0)	-0.0144; -9.22% (0;0)	0.00003; 54.24% (0.59;0.48)	-0.09; 7.75% (0.41;0.44)	0.0000089; 21.71% (0.24;0.32)
05_BU	0.0003; 0.86% (0.59;0.4)	-439.1; -30.03% (0.01;0.02)	0.000013; 20.16% (0;0)	0.0001; 0.08% (0.93;0.63)	0.00028; 106.51% (0.01;0.01)	-0.06; 7.34% (0.79;0.84)	0.0000438; 54.48% (0;0)
06_CB	-0.0005; -1.86% (0.15;0.15)	27.7; 4.62% (0.77;0.93)	0.000007; 11.97% (0;0)	-0.0043; -3.07% (0.15;0.07)	0; -6.52% (0.95;0.76)	-0.32; 37.13% (0.11;0.22)	0.0000241; 57.79% (0.01;0.01)
07_FA	0; 0.00% (0.9;0.93)	-292.3; -26.89% (0.03;0.03)	0.00001; 15.87% (0;0)	0.0005; 0.38% (0.8;0.89)	0.00015; 133.03% (0;0)	-0.07; 7.13% (0.75;0.67)	0.000031; 50.65% (0;0)
08_GI	-0.0007; -2.08% (0.43;0.32)	181.7; 15.53% (0.34;0.41)	0.000012; 14.34% (0.01;0.01)	-0.0161; -10.18% (0;0)	0.00002; 13.68% (0.86;0.88)	-0.22; 19.2% (0.33;0.29)	0.0000248; 35.08% (0.2;0.42)
09_GO	-0.0009; -2.63% (0.21;0.2)	335.7; 33.82% (0.07;0.09)	0.000008; 11.62% (0;0)	-0.0126; -8.11% (0;0)	0.00002; 40.43% (0.67;0.74)	-0.07; 6.77% (0.71;0.84)	0.0000247; 44.67% (0.21;0.29)
10_HA	-0.0012; -4.24% (0.02;0.02)	-104; -13.54% (0.4;0.51)	0.000009;14.04% (0;0.01)	-0.013; -8.57% (0;0)	-0.00004; -61.54% (0.41;0.48)	0.14; -11.08% (0.44;0.2)	0.000025; 48.26% (0.01;0.02)
11_HO	-0.0004; -1.20% (0.46;0.67)	-933.4; -45.46% (0;0)	0.000007;11.79% (0;0)	0.0023; 1.8% (0.16;0.27)	0.00031; 364.29% (0;0)	0.22; -32.35% (0.24;0.38)	0.0000296; 44.05% (0;0)
12_IL	-0.0012; -3.58% (0.18;0.13)	-233.3; -21.66% (0.16;0.24)	0.000007;11.65% (0;0)	-0.0091; -6.17% (0;0.01)	0; -1.61% (0.97;1)	0.1; -10.46% (0.41;0.22)	0.0000049; 5.88% (0.64;0.74)
13_IS	-0.0012; -3.96% (0.02;0.02)	-291.7; -28.5% (0.05;0.07)	0.000009;13.49% (0;0)	-0.0062; -4.31% (0.04;0.03)	-0.00001; -15.19% (0.79;0.71)	0.08; -8.6% (0.58;0.44)	0.0000245; 46.93% (0;0.01)
14_KE	-0.0013; -4.28% (0.04;0.02)	-474.7; -35.51% (0;0)	0.000004;6.23% (0.08;0.3)	0.0032; 2.44% (0.09;0.14)	0.00004; 35.29% (0.34;0.51)	0.05; -5.44% (0.75;0.98)	0.0000208; 39.02% (0.04;0.03)
15_KM	-0.0017; -5.57% (0;0)	-163.4; -19.04% (0.2;0.29)	0.000006; 9.2% (0.01;0.02)	-0.0126; -8.33% (0;0)	-0.00003;-29.91% (0.6;0.84)	0.12; -9.02% (0.42;0.24)	0.0000081; 13.04% (0.38;0.33)
16_KN	-0.0009; -2.92% (0.05;0.02)	147.5; 17.35% (0.25;0.24)	0.000007; 11.2% (0;0)	-0.0102; -6.71% (0;0.01)	-0.00001; -130% (0.52;0.76)	0.01; -1.02% (0.95;0.93)	0.0000197; 45.6% (0.05;0.09)
17_KR	-0.0011; -3.55% (0.01;0.02)	-153.7; -15.8% (0.12;0.22)	0.000007;11.56% (0;0)	-0.0038;-2.61% (0.16;0.14)	0.00003;57.78% (0.43;0.63)	0.18;-16.88% (0.11;0.14)	0.0000174; 36.1% (0.05;0.09)
18_KS	-0.0015; -4.95% (0.01;0.01)	-434.3; -35.56% (0;0)	0.000002;3.25% (0.15;0.15)	0.0031;2.27% (0.17;0.16)	-0.00003;-50.91% (0.17;0.13)	0.18;-17.92% (0.2;0.32)	0.0000126; 25.3% (0.13;0.14)
19_KH	-0.0005; -1.44% (0.38;0.14)	-436.3; -32.86% (0.04;0.13)	0.000018; 25.67% (0;0)	-0.004; -2.83% (0.16;0.09)	0.00008; 30.83% (0.27;0.17)	-0.14; 18.19% (0.33;0.55)	0.0000199; 18.43% (0.15;0.1)
20_KB	-0.0001; -0.34% (0.9;0.79)	-143.4; -14.62% (0.48;0.79)	0.000012; 20.21% (0;0)	-0.0014; -1.03% (0.61;0.41)	0.00002; 10.43% (0.84;0.35)	-0.3; 31.64% (0.16;0.24)	0.0000462; 83.54% (0.02;0)
21_KU	-0.0008; -2.85% (0.08;0.07)	-90.6; -13.22% (0.38;0.44)	0.000007; 11.51% (0;0)	-0.015; -9.72% (0;0)	0.00006;75% (0.55;0.98)	0.09;-7.52% (0.53;0.32)	0.0000222; 61.84% (0.02;0.04)
22_LO	-0.0017; -5.78% (0;0)	-187.3; -22.49% (0.09;0.09)	0.000011; 18.09% (0;0)	-0.0094; -6.4% (0;0.01)	0; -4.29% (0.96;0.93)	0.16;-12.78% (0.29;0.11)	0.0000139; 25% (0.16;0.22)
23_MA	-0.0013; -4.55% (0;0.01)	-177.3; -21.36% (0.13;0.2)	0.000013; 18.23% (0;0)	-0.0115; -7.67% (0;0)	0; -2.25% (0.97;0.71)	0.29; -22.05% (0.06;0.08)	0.0000224; 41.79% (0;0)
24_MD	-0.0011; -3.56% (0.17;0.22)	217.7; 25.2% (0.16;0.16)	0.000016; 18.86% (0;0)	-0.0139; -8.95% (0;0)	0.00002; 23.53% (0.78;0.38)	-0.09; 7.86% (0.62;0.59)	0.0000163; 18.36% (0.36;0.45)
25_QA	-0.0013; -4.38% (0.05;0.07)	-140.7; -15% (0.36;0.48)	0.000016; 20.15% (0.01;0)	-0.0141; -9.17% (0;0)	-0.00004; -51.35% (0.5;0.48)	0.16; -12.03% (0.31;0.2)	0.0000276; 52.47% (0.05;0.03)
26_QO	-0.0016; -4.94% (0;0)	-435.9; -36.59% (0.01;0)	0.000011; 12.88% (0.01;0.01)	-0.0109; -7.24% (0;0)	-0.00004; -44.33% (0.41;0.44)	0.53; -45.98% (0;0)	0.0000134; 16.63% (0.27;0.15)
27_SE	-0.0013; -4.06% (0.01;0.02)	-175.5; -16.56% (0.1;0.24)	0.000005; 7.38% (0.02;0.03)	-0.0094; -6.25% (0.01;0.01)	-0.00002; -52.17% (0.35;0.35)	0.44; -39.22% (0;0.01)	0.00001; 17.36% (0.32;0.24)
28_SB	-0.0009; -2.88% (0.2;0.07)	-732.3; -45.65% (0;0)	0.000003; 5.03% (0.03;0.04)	0.004; 3.1% (0.03;0.04)	0.00016; -2285.71% (0.09;0.22)	0.29; -39.17% (0.15;0.16)	0.0000084; 14.74% (0.16;0.13)
29_TE	-0.0018; -5.79% (0.01;0.05)	-212.2; -23.83% (0.11;0.07)	0.000018; 11.66% (0.23;0.24)	-0.0128; -8.38% (0;0)	0.00001; 15.48% (0.83;0.93)	0.25; -23.14% (0.11;0.09)	0.0000181; 23.51% (0.2;0.2)
30_YA	-0.0012; -3.90% (0.06;0.02)	-273.5; -25.02% (0.01;0)	0.000004; 6.85% (0.02;0.04)	-0.0009; -0.65% (0.76;0.63)	-0.00001; -16.44% (0.67;0.6)	0.1; -10.53% (0.57;0.67)	0.000012; 22.26% (0.15;0.2)
31_ZA	-0.0004; -1.41% (0.42;0.44)	-107.6; -13.81% (0.43;0.59)	0.000006; 9.6% (0;0)	-0.0151; -9.74% (0;0)	-0.00003; -55.77% (0.48;0.37)	0.08; -6.89% (0.59;0.24)	0.0000292; 63.89% (0.05;0.03)
Estimated Mean Differences [${\mu }_{COVID}-{\mu }_{Previous}]$, growth rate (%) (P-Value for T-test unequal Variance; P-Value for Wilcoxon )

In the first wave, the reduction of CO, NO², and AEI is statistically significant for most provinces. And O³ increase significantly. The differences for the others including H₂O, SO² and HCHO do not have any particular significant pattern. It shows the first wave of the COVID-19 lock-down improve the air quality against the same period in the previous year.

But in the second wave, the NO² and HCHO increase significantly for a limited number of provinces. And the reductions of the O³ is significant. The changes for the CO, H₂O, SO² and AEI do not have any significant particular pattern. It shows that the effect of second waves on the reduction of the air pollution indices are not very clear and there is no same conclusion about the improvement of air quality in the second wave.

The further analysis is available in the appendices: The spatial distribution of the air pollution indices for each province in the first and second COVID-19 waves and the previous years (Appendix A). The GEE codes for reproducing this result (Appendix B). The summary statistics including count, mean, standard deviation and median of each air pollution with statistical comparisons results such as T-test and Wilcoxon test group by provinces and COVID-19 waves (Appendix C).

Adjusted Effects

The Table 3 and 4 shows the GAMM results of the adjusted COVID-19 lock down effects with the weather conditions such as cubic splines of the pressure, precipitation, temperature and average wind speed group by air pollution indices for the first and second waves, respectively.

Table 3

The GAMM coefficients and the goodness of fit indices of air pollution indices in the first wave.
Dependent	Parametric Coefficients*		Smoothing Terms**				Model
Dependent	Intercept	COVID_1	PRES	PREC	TEMP	WIND	${{R}}^{2}$ (adj)	AIC	BIC
CO	0.02969	-0.00077	2.38	2.91	3.18	3.22	0.577	-21749.99	-21704.35
CO	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	0.577	-21749.99	-21704.35
H₂O	783.072	106.476	1.67	6.14	3.54	3.87	0.546	31595.73	31641.36
H₂O	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	0.546	31595.73	31641.36
NO²	0.0000808	-0.0000110	0.6	1.5	0	1.1	0.0267	-38513.16	-38467.42
NO²	< 0.01	< 0.01	0.13	< 0.01	0.78	0.01	0.0267	-38513.16	-38467.42
O³	0.1432	0.0042	1.66	2	4.12	2.79	0.483	-14889.09	-14843.28
O³	(< 0.01)	(< 0.01)	(0.0059)	(0.003)	(< 0.01)	(< 0.01)	0.483	-14889.09	-14843.28
SO²	0.000104	0.000059	0	0	2.369	0	0.00739	-26392.6	-26347.55
SO²	(< 0.01)	(< 0.01)	(0.474)	(0.421)	(0.019)	(0.445)	0.00739	-26392.6	-26347.55
AEI	-0.661	-0.275	1.02	1.98	4.44	0.9	0.203	2274.114	2319.92
AEI	(< 0.01)	(< 0.01)	(0.035)	(< 0.01)	(< 0.01)	(0.045)	0.203	2274.114	2319.92
HCHO	0.0000579	-0.0000014	0.00	2.31	2.82	0.82	0.124	-37439.77	-37394.51
HCHO	(< 0.01)	(0.362)	(0.346)	(< 0.01)	(< 0.01)	(0.059)	0.124	-37439.77	-37394.51
Estimated Coefficients (P-Value), * Effective Degrees of Freedom, (P-Value)

Table 4

The GAMM coefficients and the goodness of fit indices of air pollution indices in the first wave.
Dependent	Parametric Coefficients*		Smoothing Terms**				Model
Dependent	Intercept	COVID_1	PRES	PREC	TEMP	WIND	${{R}}^{2}$ (adj)	AIC	BIC
CO	0.03067	-0.00082	2.92	1.24	3.78	2.59	0.692	-7533.961	-7497.14
CO	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	0.692	-7533.961	-7497.14
H₂O	1106.52	-354.52	1.00	6.53	4.17	1.00	0.48	10521.4	10576.5
H₂O	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	(< 0.01)	(0.087)	0.48	10521.4	10576.5
NO²	0.0000718	0.0000089	0.00	1.57	2.19	2.84	0.0357	-14600.22	-14563.34
NO²	(< 0.01)	(< 0.01)	(0.33)	(0.08)	(< 0.01)	(< 0.01)	0.0357	-14600.22	-14563.34
O³	0.14381	-0.00258	1.11	1.68	3.94	2.45	0.496	-5263.24	-5226.37
O³	(< 0.01)	(< 0.01)	(0.02)	(< 0.01)	(< 0.01)	(0.006)	0.496	-5263.24	-5226.37
SO²	0.0000976	0.0000124	2.186	0.000	2.306	0.271	0.147	-10720.6	-10683.85
SO²	(< 0.01)	(0.355)	(< 0.01)	(0.45)	(< 0.01)	(0.24)	0.147	-10720.6	-10683.85
AEI	-1.033	0.013	0.00	6.42	3.47	1.17	0.249	727.2835	764.1583
AEI	(< 0.01)	(0.69)	(0.83)	(< 0.01)	(< 0.01)	(< 0.01)	0.249	727.2835	764.1583
HCHO	0.0000638	0.0000135	0.49	0.00	1.76	1.75	0.221	-13304.98	-13268.19
HCHO	(< 0.01)	(< 0.01)	(0.17)	(0.39)	(< 0.01)	(< 0.01)	0.221	-13304.98	-13268.19
Estimated Coefficients (P-Value), * Effective Degrees of Freedom, (P-Value)

Table 3 shows that the adjusted R-square for CO, H₂O and O³ models are greater or around 0.50 and the other air pollution indices models have lower values than these values. However, the reduction of CO (-0.00077), NO² (-0.0000110) and AEI (-0.275) and increment of H₂O (106.476), O³ (0.0042) and SO² (0.000104) are statistically significant in the first wave. The air quality improvement is statistically significant except the increment of the SO².

In the Table 4, the adjusted R-square is high for CO (~ 0.70) and around 0.50 for H₂O and O³ models and are low for other air pollution indices. However, the reduction of CO (-0.00082), H₂O (-354.52), O³(-0.00258) and increment of NO² (0.0000089) and HCHO (0.0000135) are statistically significant in the second wave. With the reduction of H₂O and O³ and increment in the NO² and HCHO, there is not clear improvement in the air quality in the second lock-down wave in these models. The details of estimated parameters are available (appendix D).

Functional Clustering Result

The FVE for the first and second component of the ${\stackrel{-}{X}}_{j11}\left(t\right)$ is 84% and 15%, and the FVE for the first component of ${\stackrel{-}{X}}_{j12}\left(t\right)$, ${\stackrel{-}{X}}_{j21}\left(t\right)$ and ${\stackrel{-}{X}}_{j22}\left(t\right)$ are greater than 95%.

The hierarchical clustering of FPC scores is in Fig. 1. At least three clusters between heights of 4 and 5 are observed in these figures. Figure 2 shows the first FPC scores of covid-19 (y-axis) and previous year (x-axis) in the first and second waves. Both plots have almost the same pattern with three clusters. The 06_CB (Chahar Mahaal and Bakhtiari) and 21_KU (Kurdistan) are near each other, the 12_IL (Ilam), 24_MD(Mazandaran), 09_GO(Golestan), 26_QO (Qom), 01_AL(Alborz), 11_HO (Hormozgān), 08_GI(Gilan), 29_TE(Tehran), 05_BU(Bushehr) and 19_KH(Khuzestan) are near each other and the other provinces are between of them.

Different sporadic studies show a decrease in air pollution indices during the COVID-19 lock down in Iran's provinces, for example: Decreasing PM_2.5 from 19/02/2020 to 11/03/2020 for all provinces (Asna-ashary, Farzanegan et al. 2020), Studying PM_2.5, PM₁₀ and NO₂ in 12 cities of Iran (Norouzi and Asadi 2022), decreasing NO₂ and CO in Iran and 4 cities with GEE (Shami, Ranjgar et al. 2022), decreasing CO, NO₂, PM_2.5 and PM₁₀ and increasing the SO₂ and O₃ in the first wave and decreasing Air Quality Index (AQI) in four provinces with 150 air pollution monitoring stations (Rad, Shariati et al. 2020), CO, NO₂ and O₃ in Isfahan (Moazeni, Maracy et al. 2022, Roshan, Sarli et al. 2022), Arak province (Karimi, Moradzadeh et al. 2022), Tehran (Khorsandi, Farzad et al. 2021, Hashemi, Hoepner et al. 2022, Keshtkar, Heidari et al. 2022, Namdar-Khojasteh, Yeghaneh et al. 2022), the aerosol optical depths (AODs) in Khuzestan province (Broomandi, Crape et al. 2022), PM_2.5, NO₂ and O₃ in Khorramabad capital city in Lorestan province (Anbari, Sicard et al. 2022, Rashidi, Khaniabadi et al. 2022) and Bojnurd city in North Khorasan (Aboubakri, Ballester et al. 2022).

The Containment and Health Index (C&H) is a combination of the Stringency Index with other indicators from the Oxford Coronavirus Government Response Tracker (OxCGRT) project. The value is between 0 and 100 and it is for reporting the strictness of government policies against COVID-19 for example traffic limitations, mask wearing and …. The mean and standard deviation of the C&H index for Iran in the first COVID-19 lock-down from 2020-03-14 to 2020-04-20 is 37.09 and 3.128, respectively. But the mean and standard deviation of this index in the second wave from 2021-04-13 to 2021-04-25 is 71.916 and 0.572, respectively. We conclude that although the air quality in the first wave is better than the second wave, the C&H index in the second wave is greater than the first wave. (Ritchie, Mathieu et al. 2020)

There are few studies about air pollution in the two waves of COVID-19. For example, the reduction of NO₂ in the first wave is greater than the reduction of it in the second wave globally (Wang, Tan et al. 2021). The NO₂ is the most important factor among other metrological indices that can relate to the COVID-19 incidence and severity in the first and second waves in Spain (Linares, Culqui et al. 2021, Bañuelos Gimeno, Blanco et al. 2022). The NO₂ is higher in the second wave than the first wave of COVID-19 lock-down in India (Dutta and Dutta 2021). On the other hand, the reduction of NO₂ in the second wave is higher than the first wave due to the more strict restrictions in the Australia (Boroujeni, Saberian et al. 2021).

Recent studies about climatic indices in Iran with Fuzzy clustering (Aghelpour, Singh et al. 2021) don’t use any FC methods and from this point of view using and an extension of FC methods is one of the new future directions of this research. Another recent study (Fayaz 2022) and this study are among the first researches that use GEE dataset with FDA methods and it counts as another future direction of this study. Finally, the main limitation of this research is not using the datasets of the air quality monitoring sites from Department of Environment, Iran (www.doe.ir), it accurate the estimates.

Funding

“The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.”

Competing Interests

“The authors have no relevant financial or non-financial interests to disclose.”

Author Contributions

“All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by [***] The first draft of the manuscript was written by [***] and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.”

Aboubakri, O., J. Ballester, H. R. Shoraka, A. Karamoozian and E. Golchini (2022). "Ambient temperature and Covid-19 transmission: An evidence from a region of Iran based on weather station and satellite data." Environmental Research 209: 112887.
Aghelpour, P., V. P. Singh and V. Varshavian (2021). "Time series prediction of seasonal precipitation in Iran, using data-driven models: a comparison under different climatic conditions." Arabian Journal of Geosciences 14(7): 1-14.
Anbari, K., P. Sicard, Y. Omidi Khaniabadi, H. Raja Naqvi and R. Rashidi (2022). "Assessing the effect of COVID-19 pandemic on air quality change and human health outcomes in a capital city, southwestern Iran." International Journal of Environmental Health Research: 1-12.
Asna-ashary, M., M. R. Farzanegan, M. Feizi and S. M. Sadati (2020). COVID-19 outbreak and air pollution in Iran: a panel VAR analysis, MAGKS Joint Discussion Paper Series in Economics.
Baniasad, M., M. G. Mofrad, B. Bahmanabadi and S. Jamshidi (2021). "COVID-19 in Asia: Transmission factors, re-opening policies, and vaccination simulation." Environmental research 202: 111657.
Bañuelos Gimeno, J., A. Blanco, J. Díaz, C. Linares, J. López, M. Navas, G. Sánchez-Martínez, Y. Luna, B. Hervella and F. Belda (2022). "Air pollution and meteorological variables’ effects on COVID-19 first and second waves in Spain." International Journal of Environmental Science and Technology: 1-14.
Barbieri, D. M., B. Lou, M. Passavanti, C. Hui, D. A. Lessa, B. Maharaj, A. Banerjee, F. Wang, K. Chang and B. Naik (2020). "Survey data regarding perceived air quality in Australia, Brazil, China, Ghana, India, Iran, Italy, Norway, South Africa, United States before and during Covid-19 restrictions." Data in brief 32: 106169.
Bashir, M. F., B. Ma and L. Shahzad (2020). "A brief review of socio-economic and environmental impact of Covid-19." Air Quality, Atmosphere & Health 13(12): 1403-1409.
Bherwani, H., S. Kumar, K. Musugu, M. Nair, S. Gautam, A. Gupta, C.-H. Ho, A. Anshul and R. Kumar (2021). "Assessment and valuation of health impacts of fine particulate matter during COVID-19 lockdown: a comprehensive study of tropical and sub tropical countries." Environmental Science and Pollution Research 28(32): 44522-44537.
Boroujeni, M., M. Saberian and J. Li (2021). "Environmental impacts of COVID-19 on Victoria, Australia, witnessed two waves of Coronavirus." Environmental Science and Pollution Research 28(11): 14182-14191.
Broomandi, P., B. Crape, A. Jahanbakhshi, N. Janatian, A. Nikfal, M. Tamjidi, J. R. Kim, N. Middleton and F. Karaca (2022). "Assessment of the association between dust storms and COVID-19 infection rate in southwest Iran." Environmental Science and Pollution Research 29(24): 36392-36411.
Carlo, G., G. Paolo and P. Roberto (2017). "Spatial clustering of curves with an application of satellite data." Spatial statistics 20: 110-124.
Carroll, C., A. Gajardo, Y. Chen, X. Dai, J. Fan, P. Hadjipantelis, K. Han, H. Ji, H. Mueller and J. Wang (2020). "fdapace: Functional data analysis and empirical dynamics." R package version 0.5 5.
Chiou, J. M. and P. L. Li (2007). "Functional clustering and identifying substructures of longitudinal data." Journal of the Royal Statistical Society: Series B (Statistical Methodology) 69(4): 679-699.
Conticini, E., B. Frediani and D. Caro (2020). "Can atmospheric pollution be considered a co-factor in extremely high level of SARS-CoV-2 lethality in Northern Italy?" Environmental pollution 261: 114465.
Di Giuseppe, E., G. Jona Lasinio, S. Esposito and M. Pasqui (2013). "Functional clustering for Italian climate zones identification." Theoretical and applied climatology 114(1): 39-54.
Dutta, A. and G. Dutta (2021). "Association of air pollution and meteorological variables with the two waves of COVID-19 pandemic in Delhi: A critical analysis." Heliyon 7(11): e08468.
Fayaz, M. (2022). "The lock-down effects of COVID-19 on the air pollution indices in Iran and its neighbors." Modeling earth systems and environment: 1-7.
Filonchyk, M., V. Hurynovich, H. Yan, A. Gusev and N. Shpilevskaya (2020). "Impact assessment of COVID-19 on variations of SO2, NO2, CO and AOD over East China." Aerosol and air quality research 20(7): 1530-1540.
Haggarty, R., C. Miller and E. Scott (2015). "Spatially weighted functional clustering of river network data." Journal of the Royal Statistical Society: Series C (Applied Statistics) 64(3): 491-506.
Harezlak, J., D. Ruppert and M. P. Wand (2018). Semiparametric regression with R, Springer.
Hashemi, F., L. Hoepner, F. S. Hamidinejad, A. Abbasi, S. Afrashteh and M. Hoseini (2022). "A survey on the correlation between PM2. 5 concentration and the incidence of suspected and positive cases of COVID-19 referred to medical centers: A case study of Tehran." Chemosphere 301: 134650.
Ignaccolo, R., S. Ghigo and S. Bande (2013). "Functional zoning for air quality." Environmental and ecological statistics 20(1): 109-127.
Ignaccolo, R., S. Ghigo and E. Giovenali (2008). "Analysis of air quality monitoring networks by functional clustering." Environmetrics 19(7): 672-686.
Jamaludin, S. (2017). Streamflow profile classification using functional data analysis: A case study on the Kelantan river basin. AIP Conference Proceedings, AIP Publishing LLC.
Karimi, B., R. Moradzadeh and S. Samadi (2022). "Air pollution and COVID-19 mortality and hospitalization: An ecological study in Iran." Atmospheric Pollution Research: 101463.
Keshtkar, M., H. Heidari, N. Moazzeni and H. Azadi (2022). "Analysis of changes in air pollution quality and impact of COVID-19 on environmental health in Iran: application of interpolation models and spatial autocorrelation." Environmental Science and Pollution Research 29(25): 38505-38526.
Khorsandi, B., K. Farzad, H. Tahriri and R. Maknoon (2021). "Association between short-term exposure to air pollution and COVID-19 hospital admission/mortality during warm seasons." Environmental Monitoring and Assessment 193(7): 1-6.
Linares, C., D. Culqui, F. Belda, J. A. López-Bueno, Y. Luna, G. Sánchez-Martínez, B. Hervella and J. Díaz (2021). "Impact of environmental factors and Sahara dust intrusions on incidence and severity of COVID-19 disease in Spain. Effect in the first and second pandemic waves." Environmental Science and Pollution Research 28(37): 51948-51960.
Lou, B., D. M. Barbieri, M. Passavanti, C. Hui, A. Gupta, I. Hoff, D. A. Lessa, G. Sikka, K. Chang and K. Fang (2022). "Air pollution perception in ten countries during the COVID-19 pandemic." Ambio 51(3): 531-545.
Ma, Y., Y. Zhao, J. Liu, X. He, B. Wang, S. Fu, J. Yan, J. Niu, J. Zhou and B. Luo (2020). "Effects of temperature variation and humidity on the death of COVID-19 in Wuhan, China." Science of the total environment 724: 138226.
Mahato, S., S. Pal and K. G. Ghosh (2020). "Effect of lockdown amid COVID-19 pandemic on air quality of the megacity Delhi, India." Science of the total environment 730: 139086.
Maranzano, P., P. Otto and A. Fassò (2022). "Adaptive LASSO estimation for functional hidden dynamic geostatistical model." arXiv preprint arXiv:2208.05528.
Maritorena, S., O. H. F. d'Andon, A. Mangin and D. A. Siegel (2010). "Merged satellite ocean color data products using a bio-optical model: Characteristics, benefits and issues." Remote Sensing of Environment 114(8): 1791-1804.
Millett, G. A., A. T. Jones, D. Benkeser, S. Baral, L. Mercer, C. Beyrer, B. Honermann, E. Lankiewicz, L. Mena and J. S. Crowley (2020). "Assessing differential impacts of COVID-19 on black communities." Annals of epidemiology 47: 37-44.
Moazeni, M., M. R. Maracy, B. Dehdashti and A. Ebrahimi (2022). "Spatiotemporal analysis of COVID-19, air pollution, climate, and meteorological conditions in a metropolitan region of Iran." Environmental Science and Pollution Research 29(17): 24911-24924.
Namdar-Khojasteh, D., B. Yeghaneh, A. Maher, F. Namdar-Khojasteh and J. Tu (2022). "Assessment of the relationship between exposure to air pollutants and COVID-19 pandemic in Tehran city, Iran." Atmospheric Pollution Research 13(7): 101474.
Norouzi, N. and Z. Asadi (2022). "Air pollution impact on the Covid-19 mortality in Iran considering the comorbidity (obesity, diabetes, and hypertension) correlations." Environmental Research 204: 112020.
Ogen, Y. (2020). "Assessing nitrogen dioxide (NO2) levels as a contributing factor to coronavirus (COVID-19) fatality." Science of the Total Environment 726: 138605.
Pansini, R. and D. Fornacca (2021). "Early spread of COVID-19 in the air-polluted regions of eight severely affected countries." Atmosphere 12(6): 795.
Pascucci, S., M. F. Carfora, A. Palombo, S. Pignatti, R. Casa, M. Pepe and F. Castaldi (2018). "A comparison between standard and functional clustering methodologies: Application to agricultural fields for yield pattern assessment." Remote Sensing 10(4): 585.
Rad, A. K., M. Shariati and M. Zarei (2020). "The impact of COVID-19 on air pollution in Iran in the first and second waves with emphasis on the city of Tehran." Journal of Air Pollution and Health 5(3): 181-192.
Ramsay, J. O., B. Silverman and (2005). Functional Data Analysis, Springer New York, NY.
Rashidi, R., Y. O. Khaniabadi, P. Sicard, A. De Marco and K. Anbari (2022). "Ambient PM2. 5 and O3 pollution and health impacts in Iranian megacity." Stochastic Environmental Research and Risk Assessment: 1-10.
Ritchie, H., E. Mathieu, L. Rodés-Guirao, C. Appel, C. Giattino, E. Ortiz-Ospina, J. Hasell, B. Macdonald, D. Beltekian and M. Roser (2020). "Coronavirus pandemic (COVID-19)." Our world in data.
Roshan, G., R. Sarli and J. M. Fitchett (2022). "Urban heat island and thermal comfort of Esfahan City (Iran) during COVID-19 lockdown." Journal of Cleaner Production 352: 131498.
Scafetta, N. (2020). "Distribution of the SARS-CoV-2 pandemic and its monthly forecast based on seasonal climate patterns." International journal of environmental research and public health 17(10): 3493.
Shami, S., B. Ranjgar, J. Bian, M. Khoshlahjeh Azar, A. Moghimi, M. Amani and A. Naboureh (2022). "Trends of CO and NO2 Pollutants in Iran during COVID-19 Pandemic Using Timeseries Sentinel-5 Images in Google Earth Engine." Pollutants 2(2): 156-171.
Shehzad, K., M. Sarfraz and S. G. M. Shah (2020). "The impact of COVID-19 as a necessary evil on air pollution in India during the lockdown." Environmental Pollution 266: 115080.
Team, R. C. (2022). R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing.
Venter, Z. S., K. Aunan, S. Chowdhury and J. Lelieveld (2020). "COVID-19 lockdowns cause global air pollution declines." Proceedings of the National Academy of Sciences 117(32): 18984-18990.
Wang, H., J. Tan and X. Li (2021). "Global NO 2 dynamics During the COVID-19 pandemic: a comparison between two waves of the coronavirus." IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 14: 4310-4320.
Wang, P., K. Chen, S. Zhu, P. Wang and H. Zhang (2020). "Severe air pollution events not avoided by reduced anthropogenic activities during COVID-19 outbreak." Resources, Conservation and Recycling 158: 104814.
Wikipedia.org. "COVID-19 lockdowns." from https://en.wikipedia.org/wiki/COVID-19_lockdowns.
Wikipedia.org. "Provinces of Iran." from https://en.wikipedia.org/wiki/Provinces_of_Iran.
Wood, S. and M. S. Wood (2015). "Package ‘mgcv’." R package version 1(29): 729.
Wood, S. N. (2006). Generalized additive models: an introduction with R, chapman and hall/CRC.
Wu, X., R. C. Nethery, M. B. Sabath, D. Braun and F. Dominici (2020). "Air pollution and COVID-19 mortality in the United States: Strengths and limitations of an ecological regression analysis." Science advances 6(45): eabd4049.
Wu, X., R. C. Nethery, M. B. Sabath, D. Braun and F. Dominici (2020). "Exposure to air pollution and COVID-19 mortality in the United States: A nationwide cross-sectional study." MedRxiv.
Xie, J. and Y. Zhu (2020). "Association between ambient temperature and COVID-19 infection in 122 cities from China." Science of the Total Environment 724: 138201.
Yao, F., H.-G. Müller and J.-L. Wang (2005). "Functional data analysis for sparse longitudinal data." Journal of the American statistical association 100(470): 577-590.
Zhu, Y., J. Xie, F. Huang and L. Cao (2020). "Association between short-term exposure to air pollution and COVID-19 infection: Evidence from China." Science of the total environment 727: 138704.

No competing interests reported.

AppendixADOC.pdf
Appendix A - Unadjusted Air Pollution and Weather Indices Estimation Map of Iran Provinces in the first and second wave lock-down of COVID-19
AppendixBDOC.docx
Appendix B – The Google Earth Engine Codes.
AppendixBDOC.pdf
AppendixCDOC.docx
Appendix C - Summary Statistics and Comparisons of Unadjusted Estimates
AppendixCDOC.pdf
AppendixDDOC.docx
Appendix D - The full results of Generalized Additive Mixed Models (GAMM)
AppendixDDOC.pdf

Download PDF

Version 1

posted

You are reading this latest preprint version

Estimating the Effects of Lock-Down for the First and Second Waves of COVID-19 on Air Pollution indices in the Iranian Provinces with Google Earth Engine Technology

Status:

Version 1

Abstract

Figures

Introduction

Methods

Data Gathering And Management

Unadjusted Effect: T-test And Wilcoxon Rank-sum Test

Adjusted Effects: Generalized Additive Mixed Model

Clustering: Functional Data Analysis

Results

Unadjusted Effects

Adjusted Effects

Functional Clustering Result

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1