Does Environment has any role in recovery from COVID-19 pandemic? A case study from India.

The present study has investigated the role of meteorology and air quality for recovering from the COVID-19 pandemic in India. Using Pearson’s correlation method, we look into if there is any signicant association occurs between the district level recovery case counts and different remote sensing based environmental variables. Among weather parameters, only precipitation and air temperature found to be signicantly correlated with recovery situation. However, all the pollutants’ concentration was negatively correlated with count of recovery cases. It depicts that air quality might has greater importance in recovery from this disease. During late monsoon onwards, recovery rate was getting more than the infections which indicate that lesser temperature and good rainfall could help the air to be freshen. Through air pollution was greater during winter and post monsoon than the summer season in India the higher recovery rate was counted during post-monsoon and winter which suggest that patients may require lesser temperate ambient for better recovery. Spatial patterns also suggest that north-eastern hilly region followed by districts located in the northern mountain had better recovery where the pollutants’ concentration was also quite lower during the study period. Therefore, improving air quality with proper preventive precaution could help to combat the pandemic situation in India.


Introduction
In the year 2020, world has experienced one of the deadliest pandemics in human history ever with a death count of 2501229 and con rm cases count of 112649371 globally as of 03:39 pm CET, February 26, 2021(https://covid19.who.int/). It started since December 2019 when a pneumonia infection broke out in Wuhan, Hubei province, China, which was then named as Novel Coronavirus Pneumonia (NCP), started spreading all over the country . Gradually it has spread all over the world . WHO started with epidemiological and operational updates about COVID-19 on their website since January 21 of 2020 and by March 11, 2020 the director general of the organization briefed the media by declaring COVID-19 as pandemic. In the situation report-10 that was published on January 30, 2020, WHO had already announced the rst con rmed case of India and also stated it to have a travel history to Wuhan.
A few researchers have already exhorted about signi cant association of environmental parameters to COVID-19 transmission (Azuma et al., 2020;Bherwani et al., 2020;Eslami and Jalili, 2020;Gupta and Pradhan, 2020a;Kifer et al., 2021;Rume and Islam, 2020;Saadat et al., 2020;Shakil et al., 2020). There are also a few global as well as regional studies that had been carried out in the context of China, USA, England, Germany, Spain, Italy, Indonesia, Iran, Pakistan, Bangladesh, India, and found signi cant correlation between weather parameters and the COVID-19 cases (Ahmadi et al., 2020;Fadli et al., 2020;Ficetola and Rubolini, 2021;Gupta and Pradhan, 2020b;Islam et al., 2020;Liu et al., 2020b;Ma et al., 2020;Mehmood et al., 2021;Nottmeyer and Sera, 2021;Qi et al., 2020;Tosepu et al., 2020). Studies carried out by Bochenek et al., 2021;Borah et al., 2020;Briz-Redón and Serrano-Aroca, 2020;Emediegwu, 2021;Guo et al., 2021;Mecenas et al., 2020;Runkle et al., 2020;Şahin, 2020;Sil and Kumar, 2020 suggested that warm and humid condition days incubation period, lower the number of cases (Şahin, 2020). Another study that had taken national and region-speci c data from Germany, Italy, Spain and United Kingdom and put them into a non-linear model has concluded that cold and dry environments are likely to facilitate the COVID-19 transmission (Fu et al., 2021). With low temperature, studies have explored mild diurnal temperature range and low humidity to play signi cant role in rise in transmission . Studies have even shown that in the context of China increase in the temperature will signi cantly decrease the doubling rate of cases . These results started piling up hope amongst people from all over the world that the situation will get back to normal in the upcoming summer season (Karapiperis et al., 2020;Shi et al., 2020;Kaplin et al., 2021). However, several studies, have shown completely opposite results (Kumar, 2020;Saha zadeh and Sartoli, 2020;Ceylan, 2021;Hridoy et al., 2021;. A study in the equatorial country Brazil, with the help of principal component analysis and linear regression analysis, has shown that higher mean temperature and average relative humidity favor the COVID-19 transmission (Auler et al., 2020). A very recent study in China added rainfall in the same category i.e., positively correlated with case counts . In India it has been suggested that hot and dry regions of lower altitude shall be most affected by the COVID-19 transmission . A study in Turkey has found positive correlation of number of cases with almost all the parameters (temperature, wind speed, population density etc.) except humidity (Saraç and Koyuncu, 2020). Amidst this turmoil situation, the major question in front of the world now is when shall the situation get back to normal and how?
A Delhi based study, during the initial period of pandemic, estimated that it will take almost 10months for complete recovery from COVID-19, having said that with an increase in 1℃ in the temperature about 30 cases will recover every day (Awasthi et al., 2021). A study in Jakarta, Indonesia shows that exposure to sunlight is showing a signi cantly better recovery among patients (Asyary and Veruswati, 2020). Bhattacharjee et al., 2021 predicts that in India the number of patients will start declining when the case load rate (Con rmed cases -Recovered cases -Deaths) goes lesser than the recovery rate. As of now, there is no signi cant study available to explain how under existing climatic condition of India the recovery of COVID-19 patients shall be in uenced. Hence, this paper aims at taking a close look into the relationship between the meteorological factors and recovery rate in the country.
Scanty studies have been carried out on the importance of freshen air for the melioration of human health and to elude any convalescence (Cragg et al., 2016;Kampa and Castanas, 2008;Manisalidis et al., 2020;White et al., 2019). Likewise, scarcely any study has been fanatical about the signi cance of weather parameters on COVID-19 recovery, even though 93.88% of the sufferers have been healed in India, slightly lesser than the global recovery rate (97.42%) till March 31, 2021. Since a handful of studies had reckoned that the inferior air quality had e caciously swayed to aggravate the COVID-19 emanation, hence it becomes also requisite to investigate whether the convivial air quality has been benevolent to convalesce the COVID-19 patients in India. Till date, seldom study might have inspected the yearlong period of pandemic situation while minimal have rendered any estimation on this. Therefore, the present study has been rmed to analyse the association of COVID-19 incidences (i.e. transmission) and recovery cases with weather and air quality parameters across the country, as well as proffered a comparative assessment for estimating con rm and recovery cases availing the non-linear complex relationship.

Data And Methodology
Counts of Con rmed and Recovered cases for all the available districts in India have been acquired from https://www.covid19india.org/. Spatial data for total six different meteorological parameters -2-m Air Temperature (AT), Bias-corrected Total Precipitation (PRC), Speci c Humidity (Hm), Cloud Cover (CLD), Incoming Short-wave Radiation (ISWR), Wind Speed (WS), and six different air quality parameters -Aerosol Optical Depth (AOD), Tropospheric Column Nitrogen di-Oxide (NO 2 ), Total Column Ozone (O 3 ), Surface concentration of Carbon Monoxide (CO), Sulphur di-Oxide (SO 2 ) and Black Carbon (BC) have been incorporated in the present study. Remote Sensing based datasets are highly useful for various environmental studies (Bhatt et al., 2021;Das et al., 2017;Das and Gupta, 2021;Gupta et al., 2019;Moniruzzam et al., 2018;Nanda et al., 2018;Rousta et al., 2020), hence we have used such dataset from different sources. Monthly mean of these twelve environmental variables were further processed in ArcGIS software, adjoined and related with the monthly cumulative counts of con rmed and recovery cases for each months for each of those selected districts.

Infections and recoveries during 1st COVID-19 waves in India
The district wise spatial distribution of COVID-19 con rmed cases and recoveries during the 1st COVID-19 wave (up to March, 2021) are shown in Figure 1. It shows that most of the districts have registered counts of recoveries within the range of 2500-5000 following the range of 10000-25000. Comprehensively, 20 states had recoveries of that same count; among them only Maharashtra and Kerala had registered more than 10 lakh. All states had more than 90% recovery cases except Maharashtra, Chandigarh and Punjab; whereas Mizoram and Arunachal Pradesh incidentally had more than 99% of recovery. Approximately 15% of the total districts in India mostly from the North-Eastern hilly areas and very few from the interior of Peninsular Plateau region had reported more than 99% of recovery. During summer, the average recovery rate of all provinces was unanticipatedly lower (56.96%), whereas a progressive improvement was perceptible during monsoon (78.87%), post-monsoon (119.62%) and winter (130.02%). Noteworthy, recovery rate of higher than 100% in any province during a month or season indicates that a greater number of people had been recovered than being infected in that duration, i.e. recovery counts had surpassed the count of con rmed cases. The daily count of recovery cases was lower than infective cases throughout the summer and since late September the recovery cases were far ahead in count than infections.

Pearson's correlation test
Pearson's correlation technique was applied to explore the degree of association of environmental parameters with infected and recovered cases at 95% con dence interval using equation 1. A correlogram is also being prepared to better represent the interrelation of the variables of the input dataset.
Here, r is the correlation coe cient, En represents the environmental variables, C denotes the count of con rm and recovery cases individually.

Statistical Measures
The descriptive statistics (mean, median, mode, standard deviation, standard error, range) of the input variables used for the analysis have been put in Table 2 (descriptive statistics). The entire dataset of 4020 observations was alienated into a 70:30 ratio, where 70% of the data were used to train the models and the rest 30% data were used for its evaluation. Furthermore, in order to investigate the spatial in uence of environmental parameters, monthly average values of all 12 variables were appended with monthly cumulative counts of con rmed and recovery cases noted over 335 major affected districts across the country during the yearlong period.

Results And Discussion
Through the Pearson's correlation test we found non-signi cant association between the recovery counts  Pan et al., 2017;Zhang et al., 2004),. The negative correlation between the AT and recovery cases suggests that patients might have recovered comparatively better in lesser temperature conditions. Recently, Y. Liu et al., 2020d;Shahri et al., 2021;Yang et al., 2020b have found that thrombocytopenia could increase the mortality risk among COVID-19 patients. It was also noticed that the daily count of recovery cases exceeded the con rmed cases after mid-September, 2020 and this parallel pattern continued up to the end of January, 2021 when the AT was continuously decreasing. Simultaneously, we have also noticed that the districts located at higher altitudes in north and north-east India, naturally experience relatively lesser AT, had registered more than 97% of recovery which was considerably better than the other districts located in the plain or low-lying region. Since AT started to rise February onwards the daily infections also noted to rise and outpaced the daily count of recoveries. Hence, the surrounding AT can be adjudged as a vital factor both in case of transmission and recuperation of COVID-19 in India. The non-signi cant correlation of recoveries with CLD, WS, Hm still remains subject of further investigations in this context.
Precipitation is a key factor in the reduction of air pollution (Roldán-Henao et al., 2020;Zheng et al., 2019). Greater CLD also slows down the rising trend in AT by radiative cooling effect (Gettelman and Sherwood, 2016). According to Akimoto et al., 2015;Desideri et al., 2007;Steffens, 2020;Wang and Su, 2020, the O 3 decrease during the presence of higher humidity levels, hence the O 3 formation also got reduced during the rainy season and depleted by depositing on water droplets. The negative correlation between ne PM and precipitation is often registered in umpteen studies (Gupta et al., 2020e;Pandey et al., 2017;Wang and Ogawa, 2015;Wu et al., 2018b). Thus the concentration of ne PM including the dust which was earlier conjectured for probable expeditious recrudescence of COVID-19 also got reduced during the monsoon precipitation. Earlier Jayamurugan et al., 2013;Shukla et al., 2008 had discussed the deduction of gaseous pollutants due to higher humidity and raindrop formation processes during the rainy season. Therefore it is well understood that the change in weather pattern reduced the air pollution during monsoon season, and during the post-monsoon and winter seasons the pollution levels had regained their peak due to the shortage of precipitation and humidity. We have also noted that recovery counts had signi cantly negative correlation (p < 0.05) with all the pollutants during the study period. O 3 had better inverse association (-0.384) followed by AOD (-0.355), CO (-0.353), NO 2 (-0.328), BC (-0.264), SO 2 (-0.216). It depicts that convalescents had a facile recuperation where the concentration of air pollutants were considerably lower. It also explains why ner air quality over the hilly regions in the North-East India might have helped to heal better from this disease. Though the overall country had exhibited notable rise in pollution level during post-monsoon and winter, spatial scenarios depict that the pollutant gases namely CO, NO 2 and SO 2 were mostly concentrated over Northern IGP and Eastern Indian region only. BC and AOD values (hence PM levels) were also lesser in Western parts of the country. With a signi cant congruence, the states located in North-East India had registered more than 135% and 160% of recovery during Post-Monsoon and Winter; whereas during the same seasons, the states from West and South India heeded around 124% of average recovery, fairly better than the provinces in IGP (around 98%). Therefore we can reasonably descry the escalating betterment of health in less adulterated areas while a gradual revival over tainted regions in the country, which eventually indicates that the recuperation followed by rehabilitation from this pandemic could be attainable in a less polluted environment. Hence ameliorating the air quality by strict regulations and proper preventive measures could help not only to deduce the dissemination of COVID-19 but also preferably assist to convalesce from this disease sooner.

Conclusion
In short, the correlation of the number of recovered cases with the environmental parameters through the Pearson's correlation test reveals that, increasing PRC and RH have slight impact in increasing the recovery. GRN is strongly positively correlated with the number of recovered cases. It means that, humid areas with higher vegetation cover and precipitation are supposedly going to induce high counts of recovery while the dry areas with lesser precipitation are going to experience surge in the number of con rmed cases. Areas of higher temperature are going to experience a rise in number of con rmed cases along with a slow recovery. Association cloud and wind with both the cases in weak or very weak. Among the other parameters, presence of O3 and AOD shows strong correlation with both the cases, negative with recovery cases. All the other parameters like NO2, BC, CO, OC, SO2 and ISWR are showing weak or moderate correlation with the cases, negative for recovery cases.
In accordance with the outcomes from the studies carried out by Huang et al., 2020b;Sarkodie and Owusu, 2020, we found precipitation to be bene cial for the better recovery. People residing in the area with lesser Temp recovered in greater count while Hm, WS, CLD didn't pose any signi cance as observed during the study period. During or in a later stage of treatment, several patients are also condemned to various post-syndrome infections mostly related to cardiovascular and respiratory disorders which might be the rami cations of virulent vulnerability to the gaseous pollution. Hence it altogether asserts that the environment needs to be ameliorated immensely by its standards to alleviate such deadly pandemic. The injurious impact of tra c pollution has also been discussed through various global and regional studies (Gauderman et al., 2007;Laumbach and Kipen, 2012;Urman et al., 2014). Considering those cognizance along with the observations throughout the study, it may be opined that the complete lockdown with strict regulations could substantially decrease the pollution levels, and so the dissemination rate of COVID-19 too. The pre-eminence of weather parameters on air quality is indubitable and evinced globally (Bhaskar and Mehta, 2010;Fisher, 2002;Yansui Liu et al., 2020e;Panofsky and Prasad, 1967;Qin et al., 2020;Seo et al., 2018;Zhang et al., 2015), hence its signi cant association with COVID-19 dissemination during a year-long period over India has been certainly revealed through the present study. We also postulate that amelioration from this pandemic could be viable, preferably in a freshened (less polluted) environment with non-scorching weather conditions. The outcomes conjectured that patients might recover better breathing freshened moist air and if surrounding air quality can be eloquently moderated, the current pandemic can be assuaged as well. This disquisition will irrefutably succour the government administrations including the policy makers across the country to instigate proper preparedness and effectuate requisite manoeuvres impeding the parlous rami cations of COVID-19 in forthcoming days.

Declarations
Authors declare no con ict of interest.