Insight into vaccination and meteorological factors on daily COVID-19 cases and mortality in Bangladesh

Abstract

The ongoing COVID-19 contagious disease caused by SARS-CoV-2 has disrupted global public health, businesses, and economies due to widespread infection, with 676.41 million confirmed cases and 6.77 million deaths in 231 countries as of February 07, 2023. To control the rapid spread of SARS-CoV-2, it is crucial to determine the potential determinants such as meteorological factors and their roles. This study examines how COVID-19 cases and deaths changed over time while assessing meteorological characteristics that could impact these disparities from the onset of the pandemic. We used data spanning two years across all eight administrative divisions, this is the first of its kind––showing a connection between meteorological conditions, vaccination, and COVID-19 incidences in Bangladesh. We further employed several techniques including Simple Exponential Smoothing (SES), Auto-Regressive Integrated Moving Average (ARIMA), Auto-Regressive Integrated Moving Average with explanatory variables (ARIMAX), and Automatic forecasting time-series model (Prophet). We further analyzed the effects of COVID-19 vaccination on daily cases and deaths. Data on COVID-19 cases collected include eight administrative divisions of Bangladesh spanning March 8, 2020, to January 31, 2023, from available online servers. The meteorological data include rainfall (mm), relative humidity (%), average temperature (°C), surface pressure (kPa), dew point (°C), and maximum wind speed (m/s). The observed wind speed and surface pressure show a significant negative impact on COVID-19 cases (−0.89, 95% confidence interval (CI): 1.62 to −0.21) and (−1.31, 95%CI: 2.32 to −0.29), respectively. Similarly, the observed wind speed and surface pressure show a significant negative impact on COVID-19 deaths (−0.87, 95% CI: 1.54 to −0.21) and (−3.11, 95%CI: 4.44 to −1.25), respectively. The impact of meteorological factors is almost similar when vaccination information is included in the model. However, the impact of vaccination in both cases and deaths model is significantly negative (for cases: 1.19, 95%CI: 2.35 to −0.38 and for deaths: 1.55, 95%CI: 2.88 to −0.43). Accordingly, vaccination effectively reduces the number of new COVID-19 cases and fatalities in Bangladesh. Thus, these results could assist future researchers and policymakers in the assessment of pandemics, by making thorough efforts that account for COVID-19 vaccinations and meteorological conditions.

Graphical abstract

1. Introduction

In late December 2019, a disease having flu-like symptoms was first discovered in some patients in the Wuhan city of China (Dhama et al., 2022; A. Islam et al., 2022a, Islam et al., 2022b, Islam et al., 2022c, Islam et al., 2022d). The disease-causing agent was detected as a novel straincoronavirusSARS-CoV-2 (Islam, 2021a, Islam et al., 2021b; Sakib et al., 2021). Despite the vaccination process, the COVID-19 pandemic is still lingering for three years––making it impossible to estimate its dynamics (M. A. Islam et al., 2023; Rakib et al., 2021). To manage the ongoing pandemic conditions, it is important to assess the factors linked with the rapid spread of COVID-19 (A. Islam et al., 2022a, Islam et al., 2022b, Islam et al., 2022c, Islam et al., 2022d). Regardless of the vaccination status, a number of factors, such as the virus' infectivity, the host's behavior and defensive mechanisms, environmental conditions, population density, etc., can affect the spread of COVID-19. Environmental factors, including various metrological situations or parameters, are among them and are in charge of the spread, diffusion, and potential contagiousness of COVID-19 variants. Previous studies have shown that metrological factors affect the growth and activity of respiratory viral diseases including SARS-CoV (Bi et al., 2007; Park et al., 2020; Tan et al., 2005). Experimental studies have shown that SARS-CoV-2 is highly active in conditions with low relative humidity and temperature, and genetic materials decay rapidly in an environment with high relative humidity and temperature (Ahmed et al., 2021; Altamimi and Ahmed, 2020; Jakariya et al., 2021). When the virus is exposed to increased relative humidity and temperatures, as well as simulated solar light, the virus becomes even less stable (half-life, 3 min). Besides, the factor of temperature, it appears that rainfall and wind speed, among other weather/meteorological factors, influence the spread of COVID-19 (Ahmadi et al., 2020; Nasirpour et al., 2021). Immunization is recognized as one of the most active weapons to control the ongoing pandemic. Several COVID-19 vaccines (e.g., Pfizer-BioNTech, Moderna, Oxford-AstraZeneca Sinopharm, and Sputnik Janssen) have already been used against the coronavirus disease but there is no study on the correlation of vaccines with daily cases and death of ongoing pandemic in Bangladesh (M. S. Islam, 2021a, Islam et al., 2021b). The first mRNA COVID-19 vaccine Pfizer-BioNTech was approved by the SAGE-WHO (Strategic Advisory Group of Experts on Immunization) which proved comparatively safe and reliable and showed 95% efficiency against SARS-CoV-2 symptomatic patients (Marks et al., 2023). The AZD1222 (Oxford/AstraZeneca) 63.09%, Janssen Ad26.CoV2·S 85.4%, Moderna (mRNA-1273) 92% active against various variants of SARS-CoV-2 (M. A. Islam et al., 2023).

Bangladesh has been fighting against COVID-19 since reporting the first COVID-19 case on March 08, 2020 (DGHS, 2022). As of 07 February 2023, Bangladesh Government has reported 2,037,543 confirmed cases, and 29,442 death cases; these data are available at http://dashboard.dghs.gov.bd/webportal/pages/covid19.php (DGHS, 2022). Although few studies published from Bangladesh to sort out the association with weather factors and COVID-19 cases using one o two cities (Hridoy et al., 2021; Karmokar et al., 2022). However, there are no studies that examine the effect of meteorological conditions with vaccination data on COVID-19 incidence in Bangladesh (Prata et al., 2020). In this study, we examine the changes in COVID-19 cases and deaths over time in Bangladesh using time series models while documenting the characteristics of meteorological factors and vaccination campaigns that could impact these changes from the beginning of the pandemic.

2. Methodology

2.1. Study area

Bangladesh is a South Asian developing country situated at 20° 34′ North latitude and 92° 41′ East longitude (Fig. 1 ). It is a densely populated country bounded by the Bay of Bengal in the south region. The observed average temperature in Bangladesh was around 26 °C, with a range between 15 °C and 34 °C throughout the year and a yearly average rainfall equal to 2200 mm (mm). The relative humidity is observed higher from June to October in Bangladesh (Supplementary Figure-SF1).

Fig. 1.

Bangladesh map with meteorological parameters (yearly Average of different meteorological variables map) (a) Temperature (^◦C), (b) Rainfall (mm), (c) Wind (m/s) & (d) Relative Humidity (%).

Map (a) shows that higher temperature exists along central to southern parts of Bangladesh, while lower temperature was found along northern parts adjoining the north borderline of Bangladesh and south-eastern parts bordering Myanmar. The highest and lowest temperature observed in Bhola (26.1995 °C) and Lalmonirhat (21.3966 °C) districts respectively. Possible causes can be that Bhola is located adjoining the Bay of Bengal whereas Lalmonirhat is located in the northern part near the foothills of the Himalayan hill ranges (Rimi et al., 2019).

Rainfall in Map (b) depicts an inverse scenario with respect to temperature. Due to lower rainfall along central and most of the southern parts, the air temperature observed relatively higher. On the contrary, higher rainfall in northern Bangladesh and the majority parts of the Chittagong hill tracts lead to relatively lower air temperature. Rainfall in Cox's Bazar (410.0638 mm) and Narail (138.9112 mm) districts noticed higher and lower respectively. Monsoon weather from the south-western of the Bay of Bengal causes higher rainfall in Cox's Bazar district) (Rahman et al., 2015).

Map (c) shows higher wind speeds found across central, south-western, and north-western Bangladesh. Lower speed found across the northeastern and south-eastern parts. The highest wind speed identified in the Narail district (9.044 m/s), thus, due to the highest wind movement, Narail showed the lowest rainfall. Sylhet district showed the lowest wind speed (4.7948 m/s). As the greater Sylhet division has Tripura hills on its south part, these hills work as a barrier to higher wind speed.

Map (d) shows higher relative humidity along the central-southern, central and north-eastern parts of Bangladesh whereas lower humidity is noticed majorly in north-western and south-eastern parts (Mortuza et al., 2019)The highest and lowest relative humidity are found in Lakshmipur (81.8055%) and Rajshahi (77.7943%) districts respectively.

2.2. Daily COVID-19 confirmed and death cases

For confirmed daily COVID-19 cases and death reports from all over Bangladesh including Barisal, Chittagong, Dhaka, Khulna, Mymensingh, Rajshahi, Rangpur, and Sylhet eight divisions (Supplementary File SF1), we used WHO COVID-19 reports (WHO, Data Collection) (https://covid19.who.int/) (Supplementary File SF2). The data set includes daily COVID-19 new confirmed cases, death cases, total counted death number, total death number per million, and total vaccination. The data were collected for the period March 8, 2020, to January 31, 2023, including all of the eight administrative divisions of Bangladesh (Haque et al., 2022; F. E. F. E. Hossain et al., 2021). SF1 contains the division's name, capital, established year, subdivisions (Upazilla, Unions), area, and population density.

2.3. Meteorological factors

We used NASA's data from Worldwide Energy Resources webpage (NASA, 2022)on a daily scale to collate metrological variables including rainfall (mm), relative humidity (RH) (%), temperatures (°C), surface pressure (kPa), dew point (°C), and wind velocity (m/s) at 10 m height (maximum wind speed) (Supplementary File SF3) (M. A. Islam et al., 2022a, Islam et al., 2022b, Islam et al., 2022c, Islam et al., 2022d; M. A. Islam et al., 2022a, Islam et al., 2022b, Islam et al., 2022c, Islam et al., 2022d).

2.4. Time series models

We used four-time series models (SES, ARIMA, ARIMAX, and Prophet) to analyze the trend of COVID-19 cases and deaths. The time series models were selected based on the variables. Using the time series models with the reported COVID-19 data, we predicted the COVID-19 new cases for the next 30-days. We also used the SES model as a standard to compare the accuracy of other models. Previously these models were used among confirmed Covid-19 affected patients in over 192 countries (Chyon et al., 2022).

2.4.1. SES model (simple exponential smoothing)

The simple exponential smoothing model most commonly used model for analyzing time series data (de Livera et al., 2011). SES is a simple tool that uses data as fluctuating around a steady mean (Tseng and Shih, 2020). The R package ‘fpp2’ is used for running the SES model in this study (Chaurasia and Pal, 2020).

2.4.2. ARIMA model (Auto-Regressive Integrated Moving Average)

We used the ARIMA, which is a statistical, data-oriented analysis that uses the structure of the data itself to forecast the trend of daily COVID-19 cases and deaths (Adhikari and Agrawal, 2013). For running the ARIMA model, the R package ‘forecast’ is used in this study (de Livera et al., 2011).

The equation of this model is shownas:

$${\phi }_p\left(B\right){Ф}_p\left(B^5\right){\nabla }_s^d{\nabla }^D{Z}_{t=}{\theta }_q\left(B\right){\Theta }_Q\left(B^5\right)a{t}_{\dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots }\text{.}$$ equation (1)

where p = non-seasonal autoregression, d = Regular differencing, q = Order of non-seasonal MA, s = Length of season, φ = AR operator, Ф = AR variable, ∇d = Differencing operator, ${\nabla }_{\mathrm{s}}^{\mathrm{d}}$ = Seasonal differencing operator, Zt = Observed value at t, θ = MA operator of q, Θ = Seasonal MA parameter of Q and at = Noise component (0, σ2).

2.4.3. Prophet model (automatic forecasting time-series model)

The Prophet model is also used to analyze time-series data using the R package “prophet” to observe the COVID-19cases and deaths. The Prophet model is used for irregular observations in the data set and the model fits very quickly. Advantageously, it controls for missing data and outliers in the data set (44, 45). Equation (2) is used for Prophet analysis (Hasan et al., 2021) as:

$$Y\left(\mathrm{t}\right)=g\left(t\right)+s\left(t\right)+h\left(t\right)+{\in }_{t\dots \dots \dots \dots \dots \dots \dots \dots \dots \dots \dots .}$$ equation (2)

Where g (t), s (t), and h (t) are model factors and ∈t is used for non-periodic changes.

2.4.4. Autoregressive integrated moving average with explanatory variables (ARIMAX)

The ARIMA models accept a direct relationship between the time-series values and attempt to exploit these straight conditions in perceptions, and arrange to extricate nearby designs, whereas removing high-frequency commotion. In this model, the information explanatory variable (X) is added, which is called ARIMAX (p, d, q) for accurate interpretation (Adhikari and Agrawal, 2013).

2.4.5. Empirical evaluation

All the four-time series models were compared using the benchmarks in predicting the case fatality rate in this study. The benchmark is allowed to measure the presence made by its counterparts (Kourentzes and Petropoulos, 2016). The SES model is the best suitable non-seasonal model for time series, permitting any error or trend element . In this study, we investigate and compare the execution of the considered time series models to ensure the prediction, coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE).

2.4.6. Statistical analysis

We look at how COVID-19 cases and deaths have evolved through the time series model. Through the ARIMAX model, we attempt to determine whether there is a link between COVID-19 cases and deaths with meteorological variables with and without including vaccination variable. To perform this analysis, R software was used.

3. Results

According to the government dashboard, the total number of confirmed cases of COVID-19 reported in Bangladesh was 2,037,543 (7.5% of total COVID-19 test) with corresponding 29,442 (1.44% of total positive cases) confirmed death cases, and 1,992,694 (9.77 %of total positive cases) recoveries as of February 07, 2023 (Fig. 2 A–D), (Worldometers.info (2023)) (IEDCR, 2023). The five highly affected districts areDhaka (n = 728,787), Chattogram (n = 129,129), Cumilla (n = 47,333), Sylhet (n = 38,292), and Narayanganj (n = 36,835), selected as COVID-19 red zone from fifty districts (Supplementary file SF-4) (WHO, 2022). The highest number of daily death cases (n = 261) was observed on August 5, 2021, while the highest number of positive cases diagnosed was 16,230 on July 28, 2021 (WHO, 2022). The government of Bangladesh continues to observe COVID-19 protocols including mask-wearing and hygiene whereas 72% of the total population has received one dose of the COVID-19 vaccine and 49% are fully vaccinated during the study time.

Fig. 2.

Graphs represent the COVID-19 cases, death, lab teat, vaccination up to date scenario of Bangladesh. (A) Total COVID-19 lab test and confirmed cases (B) Total vaccination and positive cases, (C) Total lab test and death (D) Total vaccination and Death case. (X-axis: Year; Y-axis: cases/lab test/vaccination/deaths).

Bangladesh is a tropical-moist climate-based area characterized by seasonal diversity in precipitation, moderate ambient temperature, and high relative humidity (Kamruzzaman et al., 2021). Bangladesh has four climatic seasons in a year namely lower temperature in the winter (from December to February), higher temperature in the summer (from March to May), the rainy season from June to September, and the post-monsoon autumn from October to November (Shahid and Khairulmaini, 2009). From the observation of daily temperature data in the last 61 years (1960–2021), the average temperature of 25.2 °C occurs in July, minimum temperature of 12.9 °C in January while the maximum temperature of 33.5 °C is observed in April (Supplementary Fig-SF1). We also observe maximum precipitation of 496 mm in July, and a minimum of 4 mm in January (Kamruzzaman et al., 2021).

Table 1 describes the descriptive statistics of temperature, precipitation, dew point, relative humidity, wind speed, and surface pressure for the study period of COVID-19 cases and deaths. In the observation period of COVID-19 cases, the minimum temperature is identified as 14.15 °C, while the highest and average temperatures are 33.51 °C and 25.72 °C. We identify the average rainfall of 6.20 mm from the beginning of the pandemic with a range detected from 0.00 to 123.12 mm. The average wind speed is noticed between a minimum of 0.55 to a maximum of 7.26 m/s (53). Our study results show the average dew point, relative humidity, and surface pressures are 21.10, 78.14, and 100.73 KPa respectively. In the observation period of COVID-19 deaths, the minimum temperature is identified as 14.15 °C, while the highest and average temperatures are 33.51 °C and 25.82 °C. We observe an average rainfall of 6.26 mm from the beginning of the pandemic with a range detected from 0.00 to 123.12 mm. The average wind speed is observed between a minimum of 0.55 to a maximum of 7.26 m/s. Our study results show the average dew point, relative humidity, and surface pressures are 21.20, 78.21, and 100.72 KPa respectively.

Table 1.

Descriptive statistics of metrological parameters for confirmed cases and deaths.

	Wind Speed (m/s)	Temperature (°C)	Dew Point (°C)	Rainfall (mm)	Relative Humidity (%)	Surface Pressure (kPa)
Mean	2.05	25.72	21.10	6.20	78.14	100.73
Median	1.77	27.85	22.78	1.69	81.75	100.71
Maximum	7.26	33.51	27.88	123.12	95.31	101.88
Minimum	0.55	14.15	6.42	0.00	33.50	99.08
Std. Dev.	1.01	4.54	5.37	10.74	12.35	0.51

The highest standard deviation for COVID-19 cases in the meteorological parameters is recorded in relative humidity, with a variation of 12.35, followed by rainfall with a variation of 10.74. In contrast, the lowest variation is detected in surface pressure at 0.51 (Table 2 ). The highest variance for COVID-19 deaths in the meteorological parameters is recorded in relative humidity, with a variation of 12.37, followed by rainfall with a variation of 10.77. In contrast, the lowest variation is evident in surface pressure at 0.51 (Table 2).

Table 2.

Descriptive statistics of meteorological parameters for deaths using different dataset.

	Wind Speed (m/s)	Temperature (°C)	Dew Point (°C)	Rainfall (mm)	Relative Humidity (%)	Surface Pressure (kPa)
Mean	2.06	25.82	21.20	6.26	78.21	100.72
Median	1.77	27.88	22.91	1.72	81.84	100.70
Maximum	7.26	33.51	27.88	123.12	95.31	101.88
Minimum	0.55	14.15	6.42	0.00	33.50	99.08
Std. Dev.	1.01	4.46	5.29	10.77	12.37	0.51

We found a trend between observed and predicted COVID-19 cases using the SES model, with R², RMSE, and MAE values of 95.49%, 622.17, and 268.32 respectively (Table 3 and Fig. 3 ). We identified the trend between observed and predicted COVID-19 deaths using ARIMAX and ARIMAX + Vaccination models, with R², RMSE, and MAE values of 96.71% and 96.66%, 531.42 and 639.51, and 245.49 and 329.18, respectively. We also observed a substantial growing trend between observed and predicted COVID-19 deaths in the ARIMA and Prophet models, with R², RMSE, and MAE values of 96.69% and 38.86%, 532.94 and 2290.63, and 240.23 and 1588.01, respectively (Table 3).

Table 3.

The summary of the SES model (Simple Exponential Smoothing), ARIMA model (Auto-Regressive Integrated Moving Average), and Prophet Model (Automatic Forecasting Time-series Model).

	Confirmed Cases			Confirmed Deaths
Method & Period	R2	RMSE	MAE	R2	RMSE	MAE
Simple Exponential Smoothing	95.49%	622.17	268.32	97.55%	7.35	4.47
Auto-Regressive Integrated Moving Average	96.69%	532.94	240.23	97.66%	7.18	4.32
Auto-Regressive Integrated Moving Average with explanatory variables	96.71%	531.42	245.49	97.67%	7.16	4.36
Auto-Regressive Integrated Moving Average with explanatory variables + Vaccination variable	96.66%	639.51	329.18	97.65%	7.29	4.45
Automatic Forecasting time-series model	38.86%	2290.63	1588.01	70.45%	25.17	18.46

RMSE: Root Mean Square Error; MAE: Mean Absolute Error.

Fig. 3.

The predicted results of daily COVID-19 deaths. Graphs show daily cases using the automatic forecasting time-series model (Prophet), ARIMA, SES model, ARIMAX without vaccination, and ARIMAX with vaccination.

The ARIMAX model with vaccination variable shows all other models in terms of accuracy (with better R², RMSE, and MAE values). The model showed a higher coefficient of determination and smaller errors than the ARIMA, ARIMAX, Prophet, and benchmark SES models. The COVID-19 deaths are predicted to rise significantly in the next 15 days according to both forecasted models. The forecasting of the regional daily deaths of COVID-19 for each model is shown in Fig. 3.

We also found a consistent trend between observed and predicted COVID-19 confirmed deaths in the SES model, with R², RMSE, and MAE values of 97.55%, 7.35, and 4.47 respectively (Table 3 and Fig. 4 ). We also noticed a substantial growing trend between observed and predicted COVID-19 cases using the ARIMAX and ARIMAX + Vaccination models, with R², RMSE, and MAE values of 97.67% and 97.65%, 7.16 and 7.29, and 4.36 and 4.45, respectively. Moreover, we detected a substantial growing trend between observed and predicted COVID-19 cases and deaths in the ARIMA and Prophet models, with R², RMSE, and MAE values of 97.66% and 70.45%, 7.18 and 25.17, and 4.32 and 18.46, respectively (Table 3).

Fig. 4.

The predicted results of daily cases of COVID-19. Graphs show daily cases using the automatic forecasting time-series model (Prophet), ARIMA, SES model, ARIMAX without vaccination, and ARIMAX with vaccination.

The ARIMAX model with vaccination variable outperformed all other models in terms of accuracy (with better R², RMSE, and MAE values). The model showed a higher coefficient of determination and smaller errors than the ARIMA, ARIMAX, Prophet, and benchmark SES models. The COVID-19 cases are predicted to rise significantly in the next 15 days according to both forecasted models. The forecasting of the regional daily cases of COVID-19 for each model is shown in Fig. 4.

Table 4 shows that the average wind speed and surface pressure have a significantly negative relation withCOVID-19 daily confirmed cases (−0.89, 95% CI: 1.62 to −0.21) and (−1.31, 95%CI: 2.32 to −0.29), respectively. A 100-dose improvement in vaccination coverage is associated with −1.19 (95% CI: 2.35, 0.38) times reduction in COVID-19 incidence.

Table 4.

Factors associated with COVID-19 cases using the ARIMAX model.

	Confirmed Cases
Variables	Without Vaccination			With Vaccination
	Coef.	95%CI	P-value	Coef.	95%CI	P-value
Wind speed	−0.89	−1.62, −0.21	0.015	−2.39	−4.10,-0.44	0.023
Average temperature	−2.24	−9.77,5.51	0.492	5.74	−15.39,27.04	0.576
Dew point	1.64	−7.59,11.12	0.845	−10.53	−36.77,15.87	0.512
Rainfall	−0.03	−0.61,0.61	0.870	0.36	−1.10,1.70	0.811
Relative humidity	0.03	−5.77,5.19	0.881	7.99	−8.86,23.82	0.440
Surface pressure	−1.31	−2.32,-0.29	0.032	−3.66	−7.69,-0.51	0.032
Vaccination	-	-	–	−1.19	−2.35, 0.38	0.021

Table 5 shows that the wind speed (−0.87, 95%CI: 1.54 to −0.21) and surface pressure (-1.51, 95%CI: 2.88, −0.01) have a significantly negative relation withCOVID-19 daily confirmed deaths. These findings are similar when we include the effect of vaccination in the model with meteorological variables.

Table 5.

Factors associated with COVID-19 deaths using the ARIMAX model.

	Confirmed Cases
Variables	Without Vaccination			With Vaccination
	Coef.	95%CI	P-value	Coef.	95%CI	P-value
Wind speed	−0.87	−1.54, −0.21	0.012	−3.11	−4.44, −1.25	<0.001
Average temperature	−2.69	−10.55, 4.78	0.471	9.23	−17.31, 35.70	0.540
Dew point	2.39	−6.87, 11.81	0.606	−14.85	−46.93, 17.22	0.546
Rainfall	−0.03	−0.49, 0.69	0.916	0.31	−1.38, 1.51	0.812
Relative humidity	−0.65	−6.47, 5.38	0.845	11.49	−8.10, 30.47	0.366
Surface pressure	−1.51	−2.88,-0.01	0.049	−4.91	−8.76, −0.11	0.037
Vaccination	-	-	–	−1.55	−2.88, 0.43	0.040

4. Discussion

This study investigated the correlation between meteorological parameters and pandemic outcomes in Bangladesh by considering many potential predictors. Our findings showed a substantial linkage between climatic factors and daily COVID-19 confirmed cases and deaths. The outcomes of the ARIMAX model and trend study imply that daily COVID-19 positive cases and deaths are linked to wind speed and surface pressure. Rainfall shows a positive correlation with both confirmed cases and deaths of COVID-19. Several research conducted to ascertain whether meteorological variables correlate with the expansion of SARS-CoV-2 confirm the existence of the relationship (Kumar et al., 2022; Sarkodie and Owusu, 2021a; Vadiati et al., 2022). Research on the relationship between meteorological characteristics and infectious diseases (including avian influenza A/H5N1, SARS-CoV, and MERS-CoV) reported the significance of metrological factors on the transmission of previous epidemics/pandemics (Sarkodie and Owusu, 2021b). For example (Sarkodie and Owusu, 2021b), found a negative correlation between metrological parameters (e.g., wind speed, solar radiation, and humidity) on the COVID-19 pandemic in Iran (Sarkodie and Owusu, 2021b). Shi et al. (2020) reported meteorological parameters are correlated with the transmission of SARS-CoV-2 infection in China. Few studies in different countries also supported the related results such as, in Malaysia (Suhaimi et al., 2020), Germany (Sarkodie and Owusu, 2021b), Norway (Menebo, 2020), and Indonesia (Sarkodie and Owusu, 2021b).

Our findings showed that temperature is linked with COVID-19 mortality, which is consistent with previous research studies and confirmed the conclusion of other investigations. Christophiet al (2021) found the chance of increasing daily numbers of SARS-CoV (2003 pandemic) was 18 times higher on days with a lower environmental temperature than on days with a higher temperature (temperature greater than 24.6 °C used as the reference standard) (Christophi et al., 2021). SARS-CoV in different climatic conditions and reported high temperature and RH showed a synergistic role in SARS-CoV virus inactivation, whereas low temperature and RH increased viral lifetime (Sabarathinam et al., 2022). As a result, the environmental parameters of tropical countries (e.g., Chile, Malaysia, Egypt, and Thailand) were found not suitable for the long-term existence of the virus. Based on the SARS-CoV-2 data from 166 country sites, it can be concluded that temperature adversely correlated with daily cases and fatalities of COVID-19 (Y. F. Wu et al., 2020). The incidence of COVID-19 in China reduces as the temperature rises (F. Y. Wu et al., 2020). In another study, Shi et al. (2020) reported an increase in temperature and RH are connected with a decrease in COVID-19 transmission in 100 Chinese cities. However, lots of findings showed higher temperature, humidity, and UV radiation did not affect COVID-19 incidence (Luo et al., 2020; Yao et al., 2020; Wang et al., 2020). Jinjarak et al. (2020) analyzed the COVID-19 death rates associated with meteorological variables.

In these current findings, we found a negative relation between rainfall and COVID-19 daily cases and deaths. However, several studies found a positive connection between rainfall and the transmission of influenza (Gomez-Barroso et al., 2017; Mahamat A, Dussart P, Bouix A, Carvalho L, Eltges F, Matheus S, Miller M.A, Quenel P, 2013). The data suggested that the influenza virus or short-range transmission was prevalent in tropical and subtropical regions. Droplets or aerosols formed during coughing, snoring, speaking, singing, or breathing can transfer viruses into the air (Bi et al., 2007). Aerosol virus survival and infectivity are affected by ambient stress temperature (Jayaweera et al., 2020). Hence,SARS-CoV-2can be active for 3 h in spray form (<5 μm) but shows higher viability on plastic surfaces and stainless metals, copper body, cartons, and glassware for up to 72 h in droplets (>5 μm) (van Doremalen et al., 2020). The results indicated the viability of SARS-CoV-2 reduced substantially due to the evaporation of the droplets at high temperatures. Similar to our findings, another study reported rainfall has a negative association with the spread of COVID-19 in India and Pakistan (M. S. M. S. Hossain et al., 2021). The precipitation rate responsible for the aggregation and washout process of aerosols in the environment and SARS-CoV-2 RNA correlated with the longevity of the virus in the atmosphere, hence, inactive to disperse further. Another mechanism might be that people often stay in residence on drizzling days, and this could decrease the proliferation of COVID-19. However, due to COVID-19, variants of the coronavirus are ongoing not a clear association between vaccination exposure with an increase or decrease in the number of cases. Therefore, it is still necessary to study and follow up on this issue to learn how meteorological factors and vaccination is related to reducing the spread of COVID-19.

The growing of COVID-19 incidence and fatality in Bangladesh could be attributable to various factors including the enhanced number of asymptomatic patients (silent spreader of COVID-19) confirmed by lab testing, the introduction of dexamethasone and additional medical treatment improvements for serious patients, the experience of public health-related occupations, improved public consciousness, and protection against COVID-19 infections. To control the spread of COVID-19, vaccination plays a significant role, particularly in declining new cases and death rates (Mortuza et al., 2019; Rimi et al., 2019). Our findings support that an increase in COVID-19 vaccination reduces daily cases and deaths. COVID-19 vaccines are negatively correlated with the number of ICU patients. Another public health study found every 10% upgrade of vaccines is related to8% reduction in mortality rates (Suthar et al., 2022).

5. Limitations of the study

The publicly available data may contain underreported statistics ofCOVID-19 positive patients and deaths, which may affect our investigations. The original scenario of meteorological effects and vaccination may differ due to variability in air pollution, human immunity, individual migrations and mobility, behavior and habits, economic and lifestyle, and cultural conditions––which may affect COVID-19 mortality and incidence by acting as confounders. In addition, our study was based on metrological outdoor data, however, SARS-CoV-2 transmission can be affected quite differently by indoor air conditions. These criteria should be included while evaluating the combined meteorological indicators and COVID-19 in future studies.

6. Recommendation

The manufacturers of vaccines could take precautions to avoid meteorological influences while maintaining or improving the effectiveness (Lin et al., 2020). Studies evaluating the effectiveness of vaccinations when given in sequential doses by different manufacturers using the same technology should be conducted by researchers working with meteorological data. Additionally, tests involving the mixing or subsequent administration of vaccinations in various meteorological variations ought to be conducted to see if they may provide a wide range of cross-protection against recently developing varieties (Haga et al., 2022). To quickly resolve this issue, enough financing, rigorous research, and extensive assessments are needed. The conflicting results on the impact of weather on the COVID-19 outbreak, however, highlight the need for additional study on multiple vaccination application settings and longer time series datasets. Additional studies could be undertaken to examine the association between meteorological variables and COVID-19 verified cases, as well as the daily testing rate and other social and environmental factors as control parameters. Datasets for human serum antibody levels and SARS-CoV-2 vaccination are still lacking. Therefore, additional research could be done to explore the factors affecting vaccination (Donzelli et al., 2022).

7. Conclusion

This research is the first to analyze the linkage between meteorological factors and COVID-19 daily cases and deaths in Bangladesh using two years of data across all eight administrative divisions. COVID-19 positive reports and deaths are substantially correlatedwith climatic parameters such as rainfall, relative humidity, temperature, wind speed, surface pressure, and dew point either positively or negatively. Further investigations are required to understand the COVID-19 daily patterns and pathogenicity of the virus at the host level. The analysis of our results demonstrates that COVID-19 vaccination successfully reduces the daily cases and deaths in Bangladesh. Our findings are useful to better understand, monitor and control the transmission of SARS-CoV-2.

Ethics statement

The work did not involve any human subject and animal experiments.

Declaration of competing interest

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.

Appendix A. Supplementary data

The following are the Supplementary data to this article.

Data availability

Data will be made available on request.