key: cord-264427-frrq4h39
authors: Huang, Ling; Liu, Ziyi; Li, Hongli; Wang, Yangjun; Li, Yumin; Zhu, Yonghui; Ooi, Maggie Chel Gee; An, Jing; Shang, Yu; Zhang, Dongping; Chan, Andy; Li, Li
title: The silver lining of COVID‐19: estimation of short‐term health impacts due to lockdown in the Yangtze River Delta region, China
date: 2020-07-07
journal: Geohealth
DOI: 10.1029/2020gh000272
sha: 
doc_id: 264427
cord_uid: frrq4h39

The outbreak of COVID‐19 in China has led to massive lockdowns in order to reduce the spread of the epidemic and control human‐to‐human transmission. Subsequent reductions in various anthropogenic activities have led to improved air quality during the lockdown. In this study, we apply a widely used exposure‐response function to estimate the short‐term health impacts associated with PM(2.5) changes over the Yangtze River Delta (YRD) region due to COVID‐19 lockdown. Concentrations of PM(2.5) during lockdown period reduced by 22.9% to 54.0% compared to pre‐lockdown level. Estimated PM(2.5)‐related daily premature mortality during lockdown period is 895 (95% confidential interval: 637‐1081), which is 43.3% lower than pre‐lockdown period and 46.5% lower compared with averages of 2017‐2019. According to our calculation, total number of avoided premature death associated PM(2.5) reduction during the lockdown is estimated to be 42.4 thousand over the YRD region, with Shanghai, Wenzhou, Suzhou (Jiangsu province), Nanjing, and Nantong being the top five cities with largest health benefits. Avoided premature mortality is mostly contributed by reduced death associated with stroke (16.9 thousand, accounting for 40.0%), ischemic heart disease (14.0 thousand, 33.2%) and chronic obstructive pulmonary disease (7.6 thousand, 18.0%). Our calculations do not support or advocate any idea that pandemics produce a positive note to community health. We simply present health benefits from air pollution improvement due to large emission reductions from lowered human and industrial activities. Our results show that continuous efforts to improve air quality are essential to protect public health, especially over city‐clusters with dense population.

The outbreak of the tragic Coronavirus disease 2019 by the end of 2019 has caused tremendous impacts on people's life around the world. At the time of this writing (May 6 th , 2020), COVID-19 has made more than 3.6 million people sick and led to more than 257,301 deaths worldwide (https://www.statista.com/statistics/, last access on May 6 th , 2020).

During its peak, the pandemic at one point caused over 15,000 new confirmed cases in China in just one single day back in February, and presently in May very few new local infections are reported in China (http://www.nhc.gov.cn/, last access on May 6 th , 2020). The effective containment of COVID-19 within China is mostly attributed to a series of prevention and control measures implemented rapidly by the Chinese government. Starting from late January 2020, national emergency response policies were launched in China in order to reduce the intensity of the spread of the epidemic to slow down the increase of number of new cases, including but not limited to: schools shut down, traffic strictly restricted, industries and construction activities suspended, mass gatherings and events cancelled or suspended, and social distancing become the new norm. As a result of the massive lockdown, emissions of primary air pollutants from various human and industrial activities decreased substantially and PM2.5 concentrations during COVID-19 lockdown in China has been shown to be much better than previous years during the same period (NASA, 2020; Wang et al., 2020a) . It is well known that poor air quality, with PM2.5 (particulate matters with aerodynamic diameters less than 2.5 μm) being a key criteria pollutant, could have adverse health impacts (Boldo et al., 2011; Cao et al., 2012; Song et al., 2017) and lead to premature mortality (Fang et al., 2016; Liu et al., 2016; Lu et al., 2015b ). An integrated exposure risk (IER) model for PM2.5 exposure-response function is widely used to estimate the premature mortality attributed to PM2.5 exposure. For example, Maji et al. (2018) estimates PM2.5-related long-term premature mortality for 161 cities in China for year 2015 as well as the potential health benefits of air pollution control policies for year 2020. Wang et al. (2020b) calculates the number of premature death due to acute and chronic exposure of ambient PM2.5 in China during 2013-2017. With substantial reductions in PM2.5 concentrations due to COVID-19 lockdown, a follow-up question is what are the health impacts of the short-term changes in air quality.

The Yangtze River Delta (YRD) region is one of the most economic developed and populated regions in China. In the past, the YRD region has frequently experienced heavy haze pollution (Cheng et al., 2013; Wang et al., 2015) . With various control strategies continuously being carried out, the overall air quality over the YRD region has greatly improved for the past few years (Ministry of Ecology and Environment of China, 2019). According to the latest report released by the Ministry of Ecology and Environment of China (Ministry of Ecology and Environment of China, 2020, 40 out of 41 cities in the YRD region has successfully met the goals of reducing PM2.5 concentrations during the 2019-2020 fall and winter season.

In our most recent study , we investigate the air quality changes over the YRD region due to lowered human activities during COVID-19 lockdown using multipollutant observations and photochemical model simulations. In this follow-up study, we attempt to quantify the short-term health impacts associated with PM2.5 changes over the YRD region due to COVID-19 lockdown. We estimate the premature mortality associated with PM2.5 exposure before lockdown and during lockdown periods. Utilizing simulated results based on an integrated meteorology and air quality modeling system, we estimate the number of avoided premature death due to lowered PM2.5 concentrations during COVID-19 lockdown over the YRD region. Methods and results from our previous study are partially adopted in this study to support health related estimation.

2.1 Quantitative analysis PM2.5 changes due to COVID-19 lockdown

The Yangtze River Delta (YRD) region, consisting of Shanghai, Jiangsu, Zhejiang, and Anhui province (Fig. 1) , is one of the most economic developed and populated regions in China. On January 23 th 2020, being one of three earliest provinces (the other two being Hunan and Guangdong provinces), Zhejiang province (located in south of the YRD region) announced provincial lockdown as "Level I" (particularly serious) response, followed by Shanghai and Anhui province on the next day and Jiangsu province two days later. Coincided with the Chinese Spring Festival (January 24 th -February 1 st , 2020), all kinds of human activities were greatly reduced during Level I response period. With the epidemic gradually controlled, emergency response in Anhui and Jiangsu province was downgraded to Level II (serious) on February 25 th , followed by Zhejiang province on March 2 nd . Shanghai announced Level II response on March 24 th due to high numbers of imported infectious cases. Same as previous study, we define pre-lockdown period as January 1 st -January 23 rd , Level I response period as January 24 th -February 25 th , and Level II response period as February 26 th -March 31 st . Fig. 1 Location of the Yangtze River Delta (YRD) region with city-level population To quantify the changes of air quality caused by reduced human activities during COVID-19 lockdown, the integrated Weather Research Forecasting model (WRF) -Comprehensive Air Quality Model with Extensions (CAMx) modeling system is used ). Details of model configurations and input data can be found in and are briefly summarized here. The integrated WRF/CAMx model is applied to simulate air quality over the YRD region ( Fig. 1 ) during pre-lockdown, Level I response, and Level II response periods. Two parallel simulations are conducted with two sets of anthropogenic emissions while keeping all other inputs and model configurations identical. For the base case simulation, the baseline emissions (i.e. emissions from normal activities assuming no lockdown) are used. For the COVID-19 scenario, emissions estimated based on reduced human activities due to lockdown are applied. For emission reductions outside the YRD region during lockdown, we applied the reduction ratio used by Wang et al. (2020a) . The relative improvement factor (RF) is defined as the ratio of simulated concentrations between the two scenarios and is applied to the observed concentrations to obtain the theoretical concentrations of air pollutants that would be if there is no lockdown. Results of model performance evaluation of the COVID-19 scenario show acceptable agreement between simulated and observed results ).

2.2 Premature mortality due to short-term PM2.5 exposure We estimate the premature mortality due to ambient PM2.5 exposure based on a widelyused log-linear exposure-response function below (Fang et al., 2016; Gao et al., 2016) :

where Y is the number of premature deaths caused by ambient PM2.5 exposure due to five leading causes (k=5): cerebrovascular disease (stroke), ischemic heart disease (IHD), chronic obstructive pulmonary disease (COPD), lung cancer (LC) for adults (≥ 25 years), and acute lower respiratory infection (ALRI) for infants (< 5 years). β is the cause-specific exposureresponse coefficients and values reported from a meta-analysis study (Lu et al., 2015a) are utilized in this study. For an increase of 10 μg/m 3 PM2.5, β is 0.63% [95% confidential interval (CI): 0.35% -0.9%] for cardiovascular disease (i.e. stoke、IHD) and 0.75% (95% CI: 0.39% -1.11%) for respiratory disease (i.e. COPD, ALRI, LC). The baseline incidence rate (R) at provincial level is obtained from the Sixth National Population Census (http://www.stats.gov.cn/tjsj/pcsj/rkpc/6rp/indexch.htm, last access on 20 th April, 2020) and the contribution of individual disease to total mortality are based on the national estimates from the Global Burden of Diseases (GBD) project of Institute for Health Metrics and Evaluation (IHME) and Health Effects Institute (HEI) for year 2017 (https://vizhub.healthdata.org/gbd-compare/, last access on 23 rd April, 2020). According to GBD study, stroke, IHD, COPD, LC, and ALRI contribute 20.2%, 16.7%, 9.2%, 6.4% and 1.7% of total deaths in China for year 2017. P is the exposed population for each city in the YRD region and is obtained from statistical yearbooks for year 2018. The exposed PM2.5 concentration in Eq. (4) 3.2 Premature mortality attributable to short-term PM2.5 exposure Ambient PM2.5 exposure leads to higher mortality in infants (<5 years) from ALRI and in adults (≥25 years) due to stroke, IHD, COPD and LC. We calculate the premature mortality due to above-mentioned causes based on the health impact function (Eq. 1) over the YRD region during pre-lockdown, Level I, and Level II period of 2017-2020 (Fig. 4) . During the pre-lockdown period, the total premature mortality attributed to PM2.5 exposure are relatively consistent during 2017-2020 and the number in 2020 is 36.4 thousand (95% CI: 30.4-38.8 thousand) for the whole YRD region. Stroke and IHD contribute to 14.4 thousand (95% CI:

12.0-15.4 thousand) and 11.9 thousand (95% CI: 10.0-12.7 thousand) premature death, together accounting for 72.2% of total PM2.5-related premature death. COPD, LC, and ALRI contribute the rest 27.8% of PM2.5-related death, each causing 6.8 thousand (95% CI: 5.7-7.2 thousand), 3.2 thousand (95% CI: 2.7-3.4 thousand), and 0.05 thousand (95% CI: 0.04-0.06 thousand) premature death during the pre-lockdown period. During Level I and Level II response periods, while the premature morality due to PM2.5 exposure during 2017-2019 fluctuated a little bit, a sharp decrease is observed for year 2020. The total PM2.5-related premature mortality during 2020 Level I and Level II period is 33.2 thousand (95% CI: 23.9-38.5 thousand) and 27.7 thousand (95% CI: 18.6-33.8 thousand), which dropped by 32.3% and 47.7% compared to the 2017-2019 average values. The relative contributions from different diseases remain unchanged since the only changing variable here is the PM2.5 concentration.

In order to directly compare the premature mortality before and during lockdown (Level I plus Level II period), we present the daily premature mortality in Figure S1 . During 2017-2019, the daily premature morality dropped by 3.2% to 12.7% before and during lockdown. For 2020, the daily premature mortality across the YRD region during pre-lockdown is 1.6 thousand (95% CI: 1.3-1.6 thousand) and 0.9 thousand (95% CI: 0.6-1.1 thousand) during lockdown (Level I + Level II), representing a sharp decrease by 43.3%. The significant reduction in premature mortality during lockdown periods, whether it is compared to the prelockdown periods of the same year, or the same periods of historical years, indicate substantial health benefits associated with lowered PM2.5 concentrations due to COVID-19 lockdown. Fig. 4 Premature mortality due to LC, stroke, IHD, COPD during pre-lockdown, Level I and Level II periods of 2017-2020 (data for ALRI is not shown due to low numbers). Estimated premature mortality with the assumption of no-lockdown is also shown for Level I and Level II.

Table S1-S4 shows the city-level premature mortality by period and year. With the same base incidence rate (R) being applied for all cities, city-level premature mortality depends on the exposed population and the PM2.5 concentration for that city. With the largest population (2418.3 thousand in 2018) of all cities in the YRD region (Fig. 5) , Shanghai has the highest PM2.5-related premature mortality during 2017-2019, even though its averaged PM2.5 concentration is ranked the 6 th , 9 th , and 9 th from the bottom during January to March of 2017, 2018 and 2019 (Fig. 5) . On the other hand, high premature mortality for cities like Bozhou and Suqian in Anhui province (for example, each is ranked 3 rd and 7 th during Level Ⅱ period in 2020 in terms of premature mortality) is more associated with the high PM2.5 concentrations, which is ranked 3 rd and 7 th in terms of average PM2.5 concentration and only 13 rd and 19 th in terms of population. Cities with small population and low PM2.5 concentrations, for example, Huangshan (in Anhui province), has the lowest premature mortality. During 2020 pre-lockdown, daily premature mortality due to PM2. An integrated WRF/CAMx model is used to estimate PM2.5 concentrations that would be during Level I and Level II period if no lockdown occurred. The potential health benefits due to lockdown are estimated as the differences of premature mortality calculated based on the observed PM2.5 concentrations and the simulated 'no lockdown' concentrations during the same periods. If no lockdown occurred, the total premature death due to PM2.5 exposure over the YRD region would be 50.2 thousand (95% CI: 42.0-53.3 thousand) during Level I period and 52.9 thousand (95% CI: 40.6-58.5 thousand) during Level II period. These two numbers are lower than the corresponding values of 2017-2019 (except for Level I in 2017 and 2019). The total avoided premature mortality over the YRD region is 17.1 thousand during Level I and 25.1 thousand during Level II (Table S5) , each representing 51.5% and 90.6% of current premature mortality rate due to PM2.5 exposure. In terms of diseases, avoided premature morality due to IHD and stroke are 14.0 thousand and 16.9 thousand, each accounting for 33.2% and 40.1% of current base incident mortality rate due to PM2.5. Avoided premature mortality due to COPD, LC, and ALRI are 7.6 thousand, 3.6 thousand, and 0.06 thousand. Figure 6 shows the city-level total avoided premature mortality during Level I and Level II (avoided premature mortality associated with different diseases are shown in Fig. S2 ). The number of avoided premature death for different cities depends on the base population and the changes in PM2.5 concentrations due to lockdown ( , and Suzhou (Anhui province, 1449, 95% CI: 1027-1672).

A few of uncertainties exist with our estimation of health impacts, which are also recognized in a couple of similar studies Maji et al., 2018; Wang et al., 2020) . First and foremost, there exists uncertainties with the parameters (e.g. concentrationresponse coefficient, threshold concentration) used in the integrated exposure-response function. Although limitation exists, we use values reported by Lu et al.(2015a) , which is developed based on meta-analysis of 59 studies covering 22 cities in Mainland China (3 cities in the YRD region), Hong Kong and Taiwan to reduce uncertainties with the concentrationresponse coefficient. Other simple assumptions when using the exposure-response function include that the exposure-response coefficient does not vary by age and a provincial-level base incidence rate is used for all cities within the province. Secondly, we only calculate the premature mortality associated with PM2.5 exposure while realizing that synergistic effects exist when exposed to other or multiple pollutants (Apte et al., 2015; Billionnet et al., 2012) and our estimation of premature mortality perhaps represents an underestimation. On top of that, the influences of PM2.5 chemical composition, size distribution and sources on health impacts (Ostro et al., 2015) are ignored in this study. When we calculate city-level population exposure, we ignore the heterogeneities of ambient PM2.5 concentrations and population within the city. We simply use the observed PM2.5 concentrations averaged over all monitoring sites within the city as the exposed concentrations. This introduces underestimation when more monitoring sites are located in the rural and less populated areas and overestimation when monitoring sites are more likely to be located in urban and populous areas for a city. Ways to improve the spatial accuracy of health impact estimation include utilizing population distribution with high spatial resolution and interpolated PM2.5 concentrations based on networks of low-cost sensors (Cavaliere et al., 2018; Holstius et al., 2014) or satellite based data (e.g. the aerosol optical depth, Chen et al., 2019; . Finally, estimations of avoided premature death assuming no-lockdown are associated with uncertainties with the meteorology and air quality model. This may all be explored in the near-future when more data are available.

In this study, we attempt to quantify the health impacts associated PM2.5 improvement due to reduced human and industrial activities during COVID-19 lockdown in the Yangtze River Delta region, a region with heavy air pollution during the past years. As a result of reduced human activities, concentrations of PM2.5 during lockdown periods reduced by 22.9% to 54.0% over the YRD region compared to pre-lockdown level. Avoided premature death due to lowered PM2.5 concentrations are estimated to be 42.2 thousand over the YRD region, representing 69.3% of the total present premature mortality due to PM2.5 exposure.

The avoided premature mortality is mostly contributed by reduced death associated with stroke (16.9 thousand, accounting for 40.0%), ischemic heart disease (14.0 thousand, 33.2%) and chronic obstructive pulmonary disease (7.6 thousand, 18.0%). Top five cities with highest avoided premature mortality are Shanghai (5433 persons), Wenzhou (2751), Suzhou (in Jiangsu province, 2660), Nanjing (2000) and Nantong (1969) .

The outbreak of COVID-19 is disastrous. This study is by no means to suggest that pandemics are bringing a positive effect for health. The passive emission reductions during COVID-19 lockdown provide a good opportunity to show the health impacts related to reduction in PM2.5. It merely reinforces our knowledge that PM2.5 has detrimental health effects. Results from our study suggested that substantial health benefits could be achieved with reduced PM2.5 concentrations by emission reductions, which confirms the importance of recent efforts to mitigate the haze pollution nationwide. However, although PM2.5 concentration decreased substantially during lockdown period, the residual PM2.5 concentrations are still much higher than the recommended 24-hr standard by WHO.

Continuous efforts are needed to reduce emissions in the long term and through the most cost-effective ways in order to protect public health.

Addressing Global Mortality from Ambient

Estimating the Health Effects of Exposure to Multi-Pollutant Mixture

Health impact assessment of a reduction in ambient PM2.5 levels in Spain

Fine Particulate Matter Constituents and Cardiopulmonary Mortality in a Heavily Polluted Chinese City

Development of Low-Cost Air Quality Stations for Next Generation Monitoring Networks: Calibration and Validation of PM2.5 and PM10 Sensors, Sensors (Basel)

Extreme gradient boosting model to estimate PM2.5 concentrations with missing-filled satellite data in China

Long-term trend of haze pollution and impact of particulate matter in the Yangtze River Delta

Mortality effects assessment of ambient PM2.5 pollution in the 74 leading cities of China

Improving air pollution control policy in China-A perspective based on cost-benefit analysis

Field calibrations of a low-cost aerosol sensor at a regulatory monitoring site in California

Air quality changes during the COVID-19 lockdown over the Yangtze River Delta Region: an insight into the impact of human activity pattern changes on air pollution variations

Integrating low-cost air quality sensor networks with fixed and satellite monitoring systems to study ground-level PM2.5. Atmospheric Environment

Estimating adult mortality attributable to PM2.5 exposure in China with assimilated PM2.5 concentrations based on a ground monitoring network

Systematic review and metaanalysis of the adverse health effects of ambient PM2.5 and PM10 pollution in the Chinese population

Short-term effects of air pollution on daily mortality and years of life lost in Nanjing

Estimating premature mortality attributable to PM2.5 exposure and benefit of air pollution control policies in China for 2020

China Air Quality Improvement Report

2020. Reports on the Completion of Environmental Air Quality Targets for the Autumn and Winter of 2019-2020 in Key Areas

Associations of Mortality with Long-Term Exposures to Fine and Ultrafine Particles, Species and Sources: Results from the California Teachers Study Cohort

Health burden attributable to ambient PM2.5 in China

Analysis of a severe prolonged regional haze episode in the Yangtze River Delta

Severe air pollution events not avoided by reduced anthropogenic activities during COVID-19 outbreak. Resources, Conservation and Recycling

Acute and chronic health impacts of PM2.5 in China and the influence of interannual meteorological variability

This study is financially sponsored by the Shanghai Science and Technology Innovation Plan Monitoring Service). The baseline incidence rate (R) at provincial level is obtained from ©2020 American Geophysical Union. All rights reserved.http://www.stats.gov.cn/tjsj/pcsj/rkpc/6rp/indexch.htm (The Sixth National Population Census). The contribution of individual disease to total mortality are based on the national estimates from https://vizhub.healthdata.org/gbd-compare/ (The Global Burden of Diseases (GBD) project of Institute for Health Metrics and Evaluation (IHME) and Health Effects Institute (HEI) for year 2017).