key: cord-0810689-ia0d2x1g
authors: Etchie, Tunde Ogbemi; Etchie, Ayotunde Titilayo; Jauro, Aliyu; Pinker, Rachel T.; Swaminathan, Nedunchezhian
title: Season, not lockdown, improved air quality during COVID-19 State of Emergency in Nigeria
date: 2021-01-21
journal: Sci Total Environ
DOI: 10.1016/j.scitotenv.2021.145187
sha: 223ffad17d3bdcb5c5a07d54a0085dcf0c4555f5
doc_id: 810689
cord_uid: ia0d2x1g

Globally, ambient air pollution claims ~9 million lives yearly, prompting researchers to investigate changes in air quality. Of special interest is the impact of COVID-19 lockdown. Many studies reported substantial improvements in air quality during lockdowns compared with pre-lockdown or as compared with baseline values. Since the lockdown period coincided with the onset of the rainy season in some tropical countries such as Nigeria, it is unclear if such improvements can be fully attributed to the lockdown. We investigate whether significant changes in air quality in Nigeria occurred primarily due to statewide COVID-19 lockdown. We applied a neural network approach to derive monthly average ground-level fine aerosol optical depth (AODf) across Nigeria from year 2001–2020, using the Multi-angle Implementation of Atmospheric Correction (MAIAC) AODs from Terra and Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) satellites, AERONET aerosol optical properties, meteorological and spatial parameters. During the year 2020. we found a 21% or 26% decline in average AODf level across Nigeria during lockdown (April) as compared to pre-lockdown (March), or during the easing phase-1 (May) as compared to lockdown, respectively. Throughout the 20-year period, AODf levels were highest in January and lowest in May or June, but not April. Comparison of AODf levels between 2020 and 2019 shows a small decline (1%) in pollution level in April of 2020 compare to 2019. Using a linear time-lag model to compare changes in AODf levels for similar months from 2002–2020, we found no significant difference (Levene’s test and ANCOVA; α=0.05) in the pollution levels by year, which indicates that the lockdown did not significantly improve air quality in Nigeria. Impact analysis using multiple linear regression revealed that favorable meteorological conditions due to seasonal change in temperature, relative humidity, planetary boundary layer height, wind speed and rainfall improved air quality during the lockdown.

Following the declaration of World Health Organization on March 11, 2020 that the novel coronavirus (COVID-19) is a global pandemic (Cucinotta and Vanelli, 2020) , many countries took some form of actions to prevent or reduce this virus spread. Among others, these actions involved mandatory shutdown of economic activities and restriction of nonessential travel. The global response to COVID-19 provided researchers a unique opportunity to assess the short-term impact of intervention measures on ambient air quality. Such studies are critically important considering that about nine million people, worldwide, die every year from stroke, heart disease, respiratory infections, chronic obstructive pulmonary disease and lung cancer caused by exposures to ambient fine particulate matter (PM 2.5 ) pollution (Cohen et al. 2005; Burnett et al., 2018; Yin et al. 2020 ). Exposure to PM 2.5 pollution has also been linked to increased deaths from diabetes mellitus (Meo et al., 2015; Weinmayr et al., 2015) and more recently, COVID-19 (Hendryx and Luo, 2020; Setti et al., 2020; Wu et al., 2020) .

Several studies reported substantial improvements in air quality during COVID-19 lockdown compared to pre-lockdown (Abdullah et al., 2020; Chen et al., 2020; Li et al., 2020; Menut et al., 2020; Rodríguez-Urrego and Rodríguez-Urrego, 2020; Tobias et al., 2020; Wang et al., 2020; Zambrano-Monserrate et al., 2020; Zheng et al., 2020) . Other studies comparing PM 2.5 pollution levels during lockdown with baseline average also found substantial declines in PM 2.5 pollution levels during the lockdown period. For example, between 10% and 54% reductions in PM 2.5 pollution levels were reported for different localities in India during lockdown compared to baseline averages (Chauhan and Singh, 2020; Kumar et al., 2020; Mahato et al., 2020; Ranjan et al., 2020; Sharma et al., 2020) .

Similar or lower reductions in PM 2.5 pollution levels were also reported for localities in China (Pei et al., 2020; Zheng et al., 2020) , Brazil (Dantas et al., 2020; Nakada and Urban, 2020) , J o u r n a l P r e -p r o o f Journal Pre-proof United States and Canada (Adams, 2020; Berman and Ebisu, 2020) , and in many parts of the world (Baldasano, 2020; Chauhan and Singh, 2020; Muhammad et al., 2020) .

April. During this period, most countries transitioned from one weather season to the next.

For example, India transitioned from winter to summer, while Nigeria moved from dry to rainy season. Some seasons such as winter are known to aggravate air pollution, while others such as rainy seasons reduce pollution level. Indeed, air pollution levels were found to be strongly influenced by season (Etchie et al. 2018a) . For instance, studies conducted in India observed higher levels of PM 2.5 pollution during winter compared with summer (Pandey et al. 2013; Singla et al., 2012; Tiwari et al., 2012) , or during dry (summer) compared to rainy (monsoon) season (Etchie et al., 2017) . Recently, Chauhan and Singh (2020) attributed about 20% and 30% reduction in PM 2.5 pollution level during COVID-19 lockdown period in New York and Los Angeles, respectively, to rainfall. Likewise, the analysis of long-term trends (2015-2020) in PM 2.5 pollution during the months of January to May in New York did not find a significant change in air pollution resulting from COVID-19 lockdown (Zangari et al., 2020) . Improvement in air quality during the lockdown months in New York was attributed to seasonal change and previous environmental interventions (Zangari et al., 2020) . To the best of our knowledge, no study has examined the long-term seasonal changes in air quality in tropical countries (such as a transition from dry to rainy season) in the months of COVID concentrations due to their high spatial (global) and temporal (~20 years of daily observations) coverages. Several methods have been utilized to derive ambient PM 2.5 concentrations at a high spatiotemporal resolution using satellite AOD products ). The methods include: semi empirical models (Lin et al. 2015; Zhang and Li, 2015) ; chemical transport models (van Donkelaar et al. 2019) ; and statistical models such as linear model (Chudnovsky et al. 2013) , mixed-effect model (Chudnovsky et al. 2014; Kloog et al. 2015; Xiao et al. 2017; Zhang et al. 2019 ), generalized additive model (Ma et al. 2016 ) and

geographically and/or temporally weighted regression models (Hu et al. 2013; He and Huang 2018 Yang et al. 2020; Jiang et al. 2021) and extremely randomized tree (Wei et al. , 2020 (Wei et al. , 2021 (Guo et al. 2011; Cheng et al. 2013 ), South Africa (Kumar et al. 2014) , United States (Tang et al. 2017) and India (Mahato et al., 2020; Ranjan et al., 2020) . Satellite AODs however represent both natural and anthropogenic aerosol pollution level, and may have high measurements error compared to ground-level fine mode AODs (AOD f ), which have radii between 0.1 and 0.25 (Park et al. 2019 ) and are mainly anthropogenic.

In this study, we utilized NN to derive monthly average ground-level AOD f across localities in Nigeria from the year 2001 to 2020 using satellites AOD products, AERONET aerosol optical properties, meteorological parameters and spatial predictors. We statistically assessed the long-term trends (2001 to 2020) in AOD f in order to ascertain whether changes in air quality occurred in Nigeria due to COVID-19 lockdown or not. We believe that ours is a first study to statistically examine the long-term trends in air quality in a tropical country with high baseline air pollution level, high frequency of rainfall and no environmental intervention measures prior to 2020.

Nigeria detected the first COVID case in Ogun State on February 28, 2020 traced to an Italian visitor to Nigeria through the Murtala Muhammed International Airport in Lagos.

On 26 th March 2020, the first COVID fatality was recorded in Nigeria, prompting the Federal Government to declare a nationwide lockdown in effect from 11 pm on 30 th March 2020. The lockdown, spanning from 30 th March to 3 rd May 2020, banned all non-essential international and domestic travels, and shutdown economic activities including schools.

Between 4 th May and June 1 st of 2020, the first phase of easing of the lockdown (easing phase-1) took place. During this period, intrastate movement resumed but commercial drivers were allowed to take only 60% of their normal carriage capacities. Also, overnight J o u r n a l P r e -p r o o f Journal Pre-proof curfew from 8 pm to 6 am was exercised but was later shortened from 10 pm to 4 am in most

States of Nigeria. Senior public offices and private businesses were allowed to work only on Mondays, Wednesdays and Fridays. In the second phase of easing (easing phase-2), which was from 2 nd June to 29 th June 2020, banks resumed normal working hours. Religious gatherings less than 20 persons were allowed to hold just one service per week. Government offices opened from Monday to Friday but working hours were from 9 am to 2 pm. This second phase was later extended to September 03, 2020. During this extension, domestic travels resumed followed by international flights. Vehicles were mandated to travel with 50% of their carriage capacities, with compulsory use of face mask.

We obtained daily Multi-angle Implementation of Atmospheric Correction ( 

We downloaded different monthly average meteorological predictors over Nigeria: rainfall (RF) (mm), relative humidity (RH) (%), air temperature at 2 m above the ground (T) (K), wind speed at 10 m above ground ( 

Journal Pre-proof datasets were processed and downloaded using GEE cloud platform with its built-in country boundary feature for Nigeria.

First, we evaluated the satellite MAIAC AODs (AOD MAIAC ) against the corresponding ground-level AERONET AOD datasets at 550 nm (AOD AERONET ) at Ilorin using simple linear regression, after removing outliers (n = 9). The validation result is shown in Equation 1:

The correlation coefficient (r) We utilized the multilayer perceptron, neural network (MLP-NN) procedure, which is a feedforward architecture from the input layer through the hidden layer to the output layer.

The input parameters were location, year, month, AOD MAIAC , RF, RH, Temp, WS, PBLH, NDVI and PD, while the output parameter was ground-level AOD f . We obtained the number of nodes in the hidden layer from training without external interference. The main function of J o u r n a l P r e -p r o o f 

Where: 0 is intercept for year 2020, 1 is coefficient for time, X. Here, X is covariate for month of the year (January-August). 1 is slope for 2020, while and are the intercept and slope, respectively, for each n th year (2002) (2003) (2004) (2005) (2006) (2007) (2008) (2009) (2010) (2011) (2012) (2013) (2014) (2015) (2016) (2017) (2018) (2019) . is time lag for each year (2002) (2003) (2004) (2005) (2006) (2007) (2008) (2009) (2010) (2011) (2012) (2013) (2014) (2015) (2016) (2017) (2018) (2019) (2020) and is error term (Zangari et al., 2020).

We performed ANCOVA using F-test for type III sums of squares, testing for homogeneity of the regression intercepts (change in air pollution levels) and homogeneity of the regression slopes for time, X (rate of change of air pollution) for each previous year relative to 2020, using a dummy covariate for year (2002-2019 = 0; and 2020 = 1). The analyses were performed using IBM SPSS 23 statistics package.

J o u r n a l P r e -p r o o f

We investigated the impact of local meteorological conditions on the AOD f , using multiple linear regression, of the form:

= 0 + 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 +

Where y is AOD f , 0 is intercept, 1 to 8 are the coefficients for time (X, is month of the year from January to August), season (S: dry season is January-February, while rainy season is March-August.), relative humidity (RH), rainfall (RF), temperature (T), wind speed (WS), planetary boundary layer height (PBLH) and year (Z, is 2001-2020), respectively. is error term.

The monthly average AODs across Nigeria during pre-lockdown, lockdown and easing phases in the year 2020 are shown in Fig. 1 . There were substantial reductions in AOD levels across the States in Nigeria during the period of lockdown or easing phases compared to pre-lockdown in year 2020 (Fig. 1 ). During lockdown, the average air pollution level in

Nigeria fell by about 69% in comparison to the pre-lockdown value, with further reduction of about 23% during easing phase-1. In Lagos, the pollution levels changed by about 81% and 6% respectively for the above two phases, and the reduction was observed to be about 50% and 56% respectively for these phases in FCT area. Generally, air pollution reductions during lockdown appeared to be greater in Southern States than in the Northern States. Also, the lowest pollution levels were observed during easing phase-1, but not during lockdown. The pollution levels increased considerably, particularly in the Central States, during easing phase-2. The percentage increase in pollution levels in Nigeria as a whole, Lagos or FCT during easing phase-2 compared with easing phase-1 was approximately 44%, 11% or 381% respectively.

A time-series analysis of historical changes in AOD levels from 2010 to 2020 (Fig. 2) did not reveal a considerable change during lockdown or phase easing in 2020 compared to the past ten years. Seasonal influence and local meteorological conditions, rather than lockdown, appear to account for the reduction in air pollution level during lockdown or phase to reduced (or delayed) rainfall, we could not suggest the same for the AOD spike for the same period in 2018.

The COVID lockdown measures in Nigeria is expected to impact mainly anthropogenic pollution sources. Therefore, we statistically assessed the long-term trends The impact analysis using multiple linear regression shows that seasonal change and variations in local meteorological conditions are significant (α = 0.05) factors influencing fine mode aerosol levels across Nigeria ( Table 1 ). The prevailing temperature has the greatest effect on fine aerosols pollution in Nigeria, followed in a decreasing order by month of the year, relative humidity, PBLH, season and wind speed. The negative sign in Table 1 indicates negative correlation between the parameter of interest and AOD f . For example, when the surface temperature increases, the convective currents become stronger dispersing ground level aerosols leading to a decrease in AOD f levels. Similarly, increase in surface relative humidity favors the hygroscopic size growth and accumulation of fine mode aerosols into coarse mode aerosols (Hu et al. 2010) resulting in lower AOD f levels. High wind speed reduces AOD f levels by dispersion. Aerosols interact strongly with meteorological parameters within the PBLH. PBLH characterize the convective and turbulent processes, entrainment and dispersion of aerosols and was found to correlate negatively with surface relative humidity (Zhang et al. 2013; Miao et al. 2019) , except over desert region dominated with stable PBL regime (Zhang et al., 2018) . Therefore, a decrease in PBLH (i.e. increase in surface relative humidity) may lead to the growth/accumulation of coarse mode aerosols from fine mode aerosols, leading to a reduction in AOD f levels.

For the coastal state of Lagos, the effect was greatest for temperature followed in a decreasing order by month of the year, rainfall, wind speed, season, year and PBLH. The result for Lagos suggests that rainfall may also have statistically significant effects on aerosol pollution levels in other southern States of Nigeria where the intensity of rainfall is high.

Heavy rainfall can reduce aerosol pollution levels by directly scavenging the pollutants or J o u r n a l P r e -p r o o f Journal Pre-proof indirectly, by limiting outdoor activities that cause the pollution. For instance, both economic and commercial activities are affected by heavy rainfall. Rainfall also reduces household burning of solid fuels and wastes. In periods of heavy downpour, rainfall degrades roads and causes flood, leading to reductions in non-essential movements. Floods also pollute drinking water resources considerably (Adewuyi et al., 2014; Etchie et al., 2013 Etchie et al., , 2014 Etchie et al., , 2020 . Kumar 2020; Kumar et al. 2020; Mahato et al., 2020; Muhammad et al., 2020; Ranjan et al., 2020; Rodríguez-Urrego and Rodríguez-Urrego, 2020) . However, studies that analyzed the long-term changes in air pollution in temperate localities found no significant improvement resulting from the lockdown measures (Adams, 2020; Zangari et al., 2020) . Using groundlevel fine mode aerosol levels across Nigeria from 2001 to 2020, we have, for the very first time, shown that the COVID-19 shutdown measures did not contribute significantly to air quality improvement in Nigeria. Although there was substantial decline in the pollution levels during the COVID lockdown and phasing easing months (April-August, 2020) compared with pre-lockdown months (January-March, 2020), the change in pollution was similar in magnitude to reductions occurring during the same period in past years (2001 to 2019). The impact analysis revealed that seasonal change to favorable meteorological conditions in Nigeria is responsible for the decline in air pollution during the COVID lockdown period.

We note some limitations of this study. First, we derived the levels of fine aerosols across Nigeria using AERONET's ground-level measurements due to the lack of groundlevel monitoring stations for PM 2.5 or gaseous pollutants such as No x and SO 2 in Nigeria.

Thus we could not interpret the magnitude of the changes in fine aerosols in terms of actual concentrations or risk to human health in Nigeria. Secondly, our estimates of fine aerosols were based on datasets from the only AERONET monitoring station in Nigeria, which is at Ilorin. Therefore, the spatio-temporal variability of fine aerosols across Nigeria is dependent J o u r n a l P r e -p r o o f Journal Pre-proof on the variability in MAIAC AODs, spatial predictors and meteorological parameters utilized.

From this study, we conclude that there were significant improvements in air quality across Nigerian localities during the COVID-19 lockdown and easing phases, compared with pre-lockdown period in year 2020. However, the air quality improvements were not due to the lockdown measures, but seasonal change of favorable meteorological conditions. Namely, improvement in air quality resulting from the 2020 COVID lockdown measures in Nigeria was not statistically significant because of the strong effects of seasonal weather changes.

Possibly, this conclusion may also be applicable to other tropical countries that transitioned into a season characterized by low pollution levels at the time of COVID shutdown, and should be investigated in future studies. To our knowledge, ours is the first study that has statistically assessed the effect of COVID-19 lockdown intervention on air quality in a tropical country with high baseline pollution level, strong seasonality and no environmental intervention history before 2020.

This study was made possible by the efforts of numerous persons over many years Liu, B.C., Binaykia, A., Chang, P.C., Tiwari, M.K., Tsao, C.C., 2017. Urban air quality forecasting based on multi-dimensional collaborative Support Vector Regression 

Air quality status during 2020 Malaysia Movement Control Order (MCO) due to 2019 novel coronavirus (2019-nCoV) pandemic

Air pollution in Ontario, Canada during the COVID-19 State of Emergency

Health risk assessment of exposure to metals in a Nigerian water supply

Spatial and temporal variations of the particulate size distribution and chemical composition over Ibadan

Validation of MODIS C6.1 and MERRA-2 AOD using AERONET Observations: A Comparative Study over Turkey

COVID-19 lockdown effects on air quality by NO2 in the cities of

Characteristics and health impacts of particulate matter pollution in China

Fine particulate matter predictions using high resolution Aerosol Optical Depth (AOD) retrievals

A critical assessment of high-resolution aerosol optical depth retrievals for fine particulate matter predictions

The global burden of disease due to outdoor air pollution

WHO Declares COVID-19 a Pandemic

Levels and sources of hourly PM2.5-related elements during the control period of the COVID-19 pandemic at exposure in Nagpur

Spatiotemporal variation trends of satellite-based aerosol optical depth in China during

Correlation between PM concentrations and aerosol optical depth in eastern China

Particulate matter air quality assessment using integrated surface, satellite, and meteorological products: 2. A neural network approach

Satellite remote sensing of particulate matter and air quality assessment over global cities

Satellite-based mapping of daily high-resolution ground PM2.5 in China via space-time regression modeling

COVID-19 prevalence and fatality rates in association with air pollution emission concentrations and emission sources

An emerging ground-based aerosol climatology: Aerosol optical depth from AERONET

Hygroscopicity of inorganic aerosols: Size and relative humidity effects on the growth factor

Estimating ground-level PM2.5 concentrations in the southeastern U.S. using geographically weighted regression

Estimation of hourly fullcoverage PM2.5 concentrations at 1-km resolution in China using a two-stage random forest model

Estimating daily PM2.5 and PM10 across the complex geo-climate region of Israel using MAIAC satellite-based AOD data

Long-term (2003-2013) climatological trends and variations in aerosol optical parameters retrieved from MODIS over three stations in South Africa

Satellite-derived high resolution PM2.5 concentrations in Yangtze River Delta Region of China using improved linear mixed effects model

Effect of lockdown amid COVID-19 pandemic on air quality of the megacity Delhi

Impact of lockdown measures to combat Covid-19 on air quality over western

Effect of environmental air pollution on type 2 diabetes mellitus

COVID-19 pandemic and environmental pollution: a blessing in disguise?

COVID-19 pandemic: impacts on the air quality during

Heavy metals in industrially emitted particulate matter in Ile-Ife

Intercomparison and Validation of Satellite and Ground-Based Aerosol Optical Depth (AOD) Retrievals over Six AERONET Sites in West Africa

Chemical compositions and source identification of particulate matter (PM2.5 and PM2.5-10) from a scrap iron and steel smelting industry along the Ife-Ibadan highway

Characterization and source identification of airborne particulate loadings at receptor siteclasses of Lagos Mega-City

Temporal distribution of fine particulates (PM2.5, PM10), potentially toxic metals, PAHs and metal-bound carcinogenic risk in the population of Lucknow City

Estimation of ground-level particulate matter concentrations through the synergistic use of satellite observations and process-based models over South Korea

Response of major air pollutants to COVID-19 lockdowns in China

Effect of lockdown due to SARS COVID-19 on aerosol optical depth (AOD) over urban and mining regions in India Sci

Spatiotemporal prediction of fine particulate matter during the 2008 northern California wildfires using machine learning

Air quality during the COVID-19: PM2.5 analysis in the 50 most polluted capital cities in the world

SARS-CoV-2 RNA found on particulate matter of

Characterization and mutagenicity assessment of PM2.5 and PM10 PAH at Agra

Trends and spatial patterns of fine-resolution aerosol optical depth-derived PM2.5 emissions in the Northeast United States from 2002 to 2013

Validation and accuracy assessment of MODIS C6.1 Aerosol Products over the heavy aerosol loading area

Statistical evaluation of PM10 and distribution of PM1, PM2.5 and PM10 in ambient air due to extreme fireworks episodes (Deepawali festivals) in megacity Delhi

Changes in air quality during the lockdown in Barcelona (Spain) one month into the SARS-CoV-2 epidemic

Regional estimates of chemical composition of fine particulate matter using a combined geoscience-statistical method with information from satellites, models, and monitors

Deriving hourly PM2.5 concentrations from Himawari-8 AODs over Beijing-Tianjin-Hebei in China

Changes in air quality related to the control of coronavirus in China: Implications for traffic and industrial emissions

Estimating 1-km resolution PM2.5 concentrations across China using the space-time random forest approach

Improved 1 km resolution PM2:5 estimates across China using enhanced space-time extremely randomized trees

Reconstructing 1-km-resolution high-quality PM2.5 data records from

Long-term exposure to fine particulate matter and incidence of type 2 diabetes mellitus in a cohort study: effects of total and traffic-specific air pollution

Exposure to air pollution and COVID-19 mortality in the United States: a nationwide cross-sectional study

Synergy of satellite and ground-based observations in estimation of particulate matter in eastern China

Full coverage high-resolution daily PM2.5 estimation using MAIAC AOD in the Yangtze River Delta of China

Estimating PM2.5 concentrations in Yangtze River Delta region of China using random forest model and the Top-of-Atmosphere reflectance

Higher Risk of Cardiovascular Disease Associated with Smaller Size-Fractioned Particulate Matter

Interaction between planetary boundary layer and pm2.5 pollution in megacities in China: A review

On the summertime planetary boundary layer with different thermodynamic stability in China: A radiosonde perspective

Trends in planetary boundary layer height over Europe

Spatiotemporal prediction of continuous daily PM2.5 concentrations across China using a spatially explicit machine learning algorithm

Remote sensing of atmospheric fine particulate matter (PM2.5) mass concentration near the ground from satellite observation

Evaluation of MAIAC aerosol retrievals over China

Significant changes in the chemical compositions and sources of PM2.5 in Wuhan since the city lockdown as Covid-19

Spatial modeling of PM2.5 concentrations with a multifactoral radial basis function neural network

Tunde Ogbemi Etchie: Conceptualization, Methodology, Validation, Project Administration and Funding Acquisition, Writing -Original Draft

Conceptualization, Data Curation, Validation, Software, Formal Analysis, and Writing -Review and Editing

Aliyu Jauro: Visualization, and Writing -Review and Editing

Pinker: Resources, Visualization, Supervision, Writing -Review and Editing

Nedunchezhian Swaminathan: Conceptualization, Visualization, Supervision, Writing -Review and Editing

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.J o u r n a l P r e -p r o o f