An Assessment of Relation of Environmental Parameters and COVID-19 Transmission at the Early Stage During March-may 2020 in India

Background The Corona virus disease 2019 (COVID-19) mainly caused by the novel severe acute respiratory syndrome corona virus 2 (SARS-CoV-2) became a global pandemic by March 2020. Actually, there is no strong evidence of weather and COVD-19 spread relation as it is a new virus. This study mainly focuses on the tropical weather impact on the spatio-temporal spread of COVID-19 during the early stages i.e. March-May 2020 in India, which is a large country where the disease has shown an exponential growth. This study is an attempt to assess the relationship of major environmental parameters like solar radiation, air temperature and relative humidity with the positive cases of COVID-19 for the period March-May 2020 which is the summer season or pre-monsoon season over India. The time series and signicant correlation analysis at daily, weekly scale and the spatial analysis of weather and COVID-19 cases are presented. The results show a signicant correlation of solar radiation and atmospheric temperature with COVID-19 cases, both at daily and weekly scale in India whereas relative humidity has low correlation in the study period. But the temperature humidity index (THI), a measure of the thermal stress, shows positive correlation with the disease spread. These results could be a good input

May 2020 which is the summer season or pre-monsoon season over India. The time series and signi cant correlation analysis at daily, weekly scale and the spatial analysis of weather and COVID-19 cases are presented.

Results
The results show a signi cant correlation of solar radiation and atmospheric temperature with  cases, both at daily and weekly scale in India whereas relative humidity has low correlation in the study period. But the temperature humidity index (THI), a measure of the thermal stress, shows positive correlation with the disease spread.

Conclusions
These results could be a good input for developing the integrated modelling framework for the COVID-19 forecasting using state of art numerical weather prediction model and disease process modelling.

Background
Whole world affected with the spread of an epidemic the Corona virus disease 2019 (COVID- 19), initially started in Wuhan, China in late December 2019 [17,16,20]. COVID-19 mainly caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [9]. In rst quarter of 2020, COVID-19 spread across the globe exponentially and the World Health Organisation (WHO) declared the pandemic on 11 March 2020 [32]. The rst positive case of COVID-19 in India was reported in Kerala on 30 Jan 2020 in a student who had a travel history of ying from Wuhan, China. The positive cases crossed 500 by 23 March 2020 and then India faced a nation wide full lockdown for 21 days. The number of positive crossed 100,000 on 18 May, 2020 i.e. in 111 days from date of incidence. The present trend of the new infected cases shows exponential growing rate and it varies differently over different regions in world. As COVID-19 is a new disease, there is no evidence of the relation of weather parameters on the disease transmission or spread. There is a need of better understanding of the disease-climate parameter interaction by quantifying their relation so that weather parameter enabled modelling effort can be useful in predicting the COVID-19 incidence and the disease spread well in advance.
COVID-19 mainly causes the respiratory illnesses like u which include in uenza A and in uenza B but it spreads more faster and easily than u and also the death rate is higher in this disease. There are several studies related to transmission of seasonal u ( in uenza) virus which largely controlled by the seasonal climatic factors like temperature and relative humidity, population density, city size and its structure, etc. [10,8] . In the temperate regions, the seasonal variation of in uenza transmission is controlled by the conditions of absolute humidity, levels of susceptibility and changes in population-mixing and contact rates [25] and past studies have shown seasonal patterns of human coronavirus species and annual variation in species [14]. Study by Chan et al. [7] showed in subtropical area during spring season the SARS coronavirus have better stability at low temperature and humidity. The speci c climatic condition like the optimal temperature, humidity and wind velocity will help for the survival and transmission of SARS virus, whereas low humidity, high barometric pressure and daily temperature uctuations will reduce the transmission [33].
Low absolute humidity during winter increases in uenza virus survival and thus increases in uenza virus transmission e ciency [27] . Relative humidity and temperature is sensitive to the aerosol route and insensitive to the contact route in the transmission of the in uenza [19]. In a temperate region, in uenza incidence increase with low temperature and low/high humidity and also with high diurnal temperature range [22]. The e ciency of transmission of in uenza virus is affected by the humidity and temperature conditions and the transmission of in uenza virus is less dependent on humidity and temperature in the absence of immunity [30].
After the widespread establishment of SARACoV-2 infection, the intensity, and timing of pandemic and post pandemic outbreaks will depend on the time of year. The transmission and the level of crossimmunity that present between beta corona viruses is less dependent on the magnitude of seasonal variation [15]. The seasonality of the SARS-CoV-2 has not been established so far but the climate might play a role in the spread of the disease [28,29]. Some studies have reported that the COVID-19 spread can be affected by a large number of factors including the climate conditions [24,31]. Several recent studies is being carried out by researchers to nd the relation of weather parameters and COVID-19 in different regions, countries and at global scale [1,2,3,4,5,6,21,23,33] but there is lot of divergence in the results as some studies indicated the expansion of COVID-19 is not driven by climate at global scale [6]. A recent study also nds the impact of environmental temperature on COVID-19 exponential growth for US and Italy at regional scale [18]. In few other studies the results shows insigni cant correlation of temperature and UV radiation with COVID-19 transmission [23,33]. Similarly another time series study over mainland China explains the relation of temperature and humidity with the transmission of COVID-19 [24].
Our study is mainly focussed on nding the evidence of in uence of the weather conditions particularly air temperature, relative humidity and solar radiation ux in the early stage spreading of COVID-19 epidemic in India using the data for the period March to May 2020. Pearson correlation coe cients (CC) between the climate variables and COVID-19 cases are computed and presented at daily and weekly scale. The Temperature-Humidity Index (THI) is also an important parameter which generally used for determining the heat stress and in this study, THI computed using the temperature and humidity at all India scale and the correlation of THI and COVID-19 cases is explored.

COVID-19 Spread in India
The daily number of cumulative COVID-19 positive cases reported in India for the period 01 March to 20 May are presented in logarithmic scale in Fig. 1 which shows there is a steady exponential growth in the epidemic numbers for the analysis period. It is clearly indicated that March 4 cases was at 29 and quite linear till April 11 with 8446 cases and there after the growth phase increased very fast and the number became 10453 on April 13 and close to 50000 cases by May 06 and the number crossed 100000 on May 19, 2020. The curve of the COVID-19 spread in India initiated with a pre-exponential phase, which is characterized by a mild logarithmic growth, followed by the exponential growth resulting the disease outbreak. The spatial distribution of the COVID-19 is presented in Fig. 2 represents the number of cases in each state on 14 th March and 16 th May 2020, which clearly shows Maharastra, Tamilnadu and Delhi were the most affected states in India so far. There is a challenge in understanding the regional variation in the growth rates as it is different for different countries and states and one can consider the weather parameters like temperature and relative humidity etc. as factors affecting the COVID-19 outbreak.

Weather Parameter distribution in India
The spatial distribution of weekly averaged solar radiation ux at surface is presented in Fig. 3, clearly indicates the start of strong radiation in the 4 th week onwards all over India. Due to lockdown, a reduction in concentration of the pollutants and aerosol contents in atmosphere is seen in the year 2020, which resulted in more surface radiation compared to other years in summer season. During 29 March-18 April almost all the regions experienced high solar radiation ux. Same spatial distribution analysis for 2m air temperature shows a increasing trend from March to April and in several parts of the country, weekly average temperature crossed 36 o C in the month of May 2020 (Fig. 4). The distribution also shows a gradual increase in temperature, starting in western India, followed by central India and nally eastern India recording high temperature during summer 2020. The relative humidity variability is also presented week wise in Fig. 5 and it clearly indicates the spatio-temporal variability in the distribution of water vapour and shows a increasing trend in rst 2 weeks of March then decreasing trend from March 20 to 2nd week of April and increased in the last week of April and again decreased in May over most of regions in India. In general, the western and central part showed a comparative low humidity as against the north, east, south and north-east India regions during the study period.

Relation of Weather parameter and COVID-19 cases
The daily time series of all India averaged relative humidity (upper panel), air temperature (middle panel) and downward solar ux are presented along with the daily reported number of COVID-19 cases in India  Fig. 7).
As generally the summer season prevails from March to May, THI a measure of the thermal stress, is also compared with COVID-19 cases at all India scale and it is observed that both in daily and weekly averaged analysis, there is strong correlation between the THI and COVID-19 cases (Fig. 8) and the correlation found to be 0.6 at 95% signi cant level.

Conclusions
The COVID-19 spread showing exponential growth rate by May 2020 in India, is going to be more in coming months as the summer season strengthening this year due to low emission and aerosol in atmosphere, both solar radiation and temperature is increasing. The present study con ned over India shows signi cant correlation of solar radiation and atmospheric temperature with COVID-19 cases whereas humidity has low correlation in the study period. Also, the THI a measure of the thermal stress shows positive correlation with the COVID-19 spread. In the coming days, there is a need of nding these relations as the COVID-19 shows rapid transmission in India for robust estimation and evidence. Along with the weather parameter, there is also a requirement of other prevention activities for the control of the ongoing pandemic. Authors' contributions : KCG contributed to the conceptualization of this paper and wrote sections of the paper. All authors contributed to study design, data collection, data analyses, interpretation of data, and drafting of the manuscript. All authors have read and approved the manuscript.