Effects of climatological parameters on the outbreak spread of COVID-19 in highly affected regions of Spain.

The coronavirus (COVID-19) pandemic is infecting the human population, killing people, and destroying livelihoods. This research sought to explore the associations of daily average temperature (AT) and air quality (PM2.5) with the daily new cases of COVID-19 in the top four regions of Spain (Castilla y Leon, Castilla-La Mancha, Catalonia, and Madrid). To this end, the authors employ Pearson correlation, Spearman correlation, and robust panel regressions to quantify the overall co-movement between temperature, air quality, and daily cases of COVID-19 from 29 February to 17 July 2020. Overall empirical results show that temperature may not be a determinant to induce COVID-19 spread in Spain, while the rising temperature may reduce the virus transmission. However, the correlation and regression findings illustrate that air quality may speed up the transmission rate of COVID-19. Our findings are contrary to the earlier studies, which show a significant impact of temperature in raising the COVID-19 spread. The conclusions of this work can serve as an input to mitigate the rapid spread of COVID-19 in Spain and reform policies accordingly.

Introduction

A novel coronavirus disease (COVID-19) epidemic was first reported in Wuhan (China) in December 2019, which spread quickly around the globe (Anderson et al. 2020; Li et al. 2020; Gorbalenya 2020; Wu et al. 2020a). The transmission of viruses can be affected by many factors, including climate conditions (such as temperature, air quality, and humidity), populated areas, and healthcare quality (Wang et al. 2020). Therefore, understanding the relationship between climatic variation, the air quality of a country, and the transmission of COVID-19 is key to playing in the fight against the virus SARS-CoV-2. The two main factors include high temperature and humidity, directly correlated with the spreading of COVID-19 in any region (Wang et al. 2020). However, COVID-19 presents similar patterns in terms of spread rate as compared with the global spread of the Spanish Flu pandemic in 1919 (which affected hundreds of millions of people) and an outbreak of 2002–2003, severe acute respiratory syndrome (SARS) (3.5% for COVID-19 and 11% for SARS).

The climatological parameters could affect the virus spread via various channels. For example, the temperature and air quality level may increase or decrease the risk, survival period of the virus, transmission rate, and virus life in the air or surface. Hence, the sensitivity analysis of climatology parameters on COVID-19 infection rate is essential to ensure the survival of human beings (Jahangiri et al. 2020). To forecast the fast and slow down the pandemic and end time, it is necessary to understand the relationship between the transmission of COVID-19 and weather (extreme hot and cold conditions). For example, a respiratory disease influenza outbreak spreads quickly in winter conditions (Lowen et al. 2007; Tamerius et al. 2013). Few of the recent works reported the evidence that transmission of influenza decreased in high temperature and humidity (Shaman and Kohn 2009; Lipsitch and Viboud 2009; Steel et al. 2011; Park et al. 2019). Two conditions help to spread the infectious disease; first, dry air and respiratory droplets contain the longer airborne duration of the influenza virus and more stable in cold temperature (Lowen and Steel 2014; Tellier 2009). Second, due to the weak immune system of hosts in cold and dry weather, the virus has been susceptible to effect quickly to the hosts (Eccles 2002; Kudo et al. 2020).

The variables that affect the survival and contagion of the virus SARS involve optimal temperature, humidity, and wind velocity (Yuan et al. 2006). In this regard, the clinic symptoms of critically ill patients infected with COVID-19 are probably similar to those of SARS and MERS (Wang et al. 2020). The first reports showed that the contagion of the virus was limited in spreading from people to people (Chen et al. 2020) and can be able to transmit through the air (Mostafa et al. 2020). A study on influenza disease resulted that temperature and humidity are the leading cause of disease outbreak in Korea, and a significant positive association was observed for diurnal temperature range (DTR) (Park et al. 2019). In mid-March, Chinese researchers indicated that there is a correlation between meteorological factors (e.g., temperature, humidity) and COVID-19 outbreak (Oliveiros et al. 2020; Wang et al. 2020). Climate variables, e.g., temperature variation and air quality, can also be a direct cause of biological interactions between the SARS virus and humans (Yuan et al. 2006). Therefore, we assume that the weather conditions might also contribute to the spread of COVID-19.

Spain is located in southwestern Europe, with the capital city of Madrid, situated in the middle of Spain. The first case of COVID-19 in Spain was recorded on 31 January 2020, and now it is in the top ten countries with the highest COVID-19 cases with 260,555 confirmed cases as of 18 July 2020 (Ministry of Health Spain 2020).

Figure 1a illustrates the overall COVID-19 outlook with daily new cases, deaths, cases per million, and deaths per million in Spain. According to the reports of the World Health Organization and Spain authorities, the major affected areas of Spain are Madrid, Catalonia, Castilla-La Mancha, and Castilla y Leon respectively until 17 July 2020 (Fig. 2). According to World Health Orgnization records, there are more than 22 million confirmed pateints and approximately 0.8 million fatiliites globally till 21 August 2020 caused from COVID-19 (WHO 2020). (WHO 2020; report 180).

Fig. 1

a COVID-19 outlook for Spain. Source: WHO (2020). b Scatterplot for COVID-19 new cases in Spain

Fig. 2

COVID-19 variation across the four most affected regions of Spain

This research work joins the strand of few recent works, which also argued that COVID-19 has correlation relationships with temperature, air quality, and other weather conditions (Fareed et al. 2020; Iqbal et al. 2020; Tosepu et al. 2020). The mentioned studies argued that the SARS outbreak in 2003, in Guangdong, ended with the rise in temperature (Wang et al. 2020). Temperature, air quality, and humidity were documented to be practical factors in the spread of SARS disease. Few of the recent works suggested that COVID-19 patients might be similar to those of SARS and MERS (Wang et al. 2020; Shi et al. 2020). More recently, it was suggested that viruses could be transmitted by several factors, including climatic conditions (such as temperature, air quality, and humidity) and population density (Dalziel et al. 2018). The extreme weather conditions that accompany long-term climate variations may also contribute to the spread of the West Nile virus in the USA and Europe (Epstein 2001). The average temperature of Spain is around 57 °F (∼ 14 °C) year-round. Figure 3a shows the temperature difference between different regions of Spain. Recent studies on COVID-19 situation in China and Iran suggested that the low temperature and absolute humidity may also be essential factors of COVID-19 outbreak and mortality rate (Yueling et al. 2018; Mohsen et al. 2020). The discussion, as mentioned earlier, guides us that research on climate change and COVID-19 is still limited, and this research will contribute to an effort to prevent COVID-19 disease. Given the existing literature, the current study chooses temperature and air quality (PM2.5) as indicators of COVID-19 spread in Spain regions.

Fig. 3

a Temperature variation across Spain regions. b Scatterplot of temperature variation with trendline

The current research mainly focuses on the research problem given as Does the COVID-19 outbreak correlates with climatology parameters (temperature and air quality) in Spain? To the best of the authors’ knowledge, this is probably the first work to explore the association between climatology parameters with COVID-19 for the case of Spain. The climatology data are based on information between 29 February and 18 July 2020, in Spain, the duration of fast and peak outbreak of COVID-19 pandemic. However, the parameters were subjected from the top four most populated regions of Spain (Madrid, Catalonia, Castilla-La Mancha, and Castilla y Leon), which can result in infection prevalence and are more in the exposure to COVID-19 identified.

It is important to mention here that the theme and objective of the current study are in line with the scientific reasons and join the strand of recent literature. For instance, more recently, the literature based on meteorological parameters suggested that air quality, population density, temperature, and pollution significantly influence the COVID-19 infections spread in different regions of China and Turkey (Shi et al. 2020; Iqbal et al. 2020; Fareed et al. 2020). On the contrary, Jahangiri et al. (2020) concluded that weather and climate might not affect the COVID-19 spread in Iran. There is still a dearth of literature research regarding COVID-19 and weather parameters. The existing studies have mainly reported mixed findings regarding the determinants and causing factors of COVID-19. However, the effects of different environmental factors on the incidence of COVID-19 remain to be elucidated. The prime objective of this research is to unveil the effects of temperature and air quality (PM2.5) with the daily COVID-19 cases in Spain.

To this end, the authors employ the detailed correlation analyses and panel data regression methods on the panel data of four profoundly affected regions of Spain.

The “Climatology and weather conditions in Spain” section describes the climatology and weather conditions in Spain. The “Materials and methods” section explains the data and methods with graphical analysis. The “Empirical results” section reports the empirical results with discussion and scientific reasons, while the “Conclusion” section concludes the paper.

Climatology and weather conditions in Spain

Currently, the major part of Spain population feels health and safety risks. During recent years, over 15 million people’s health in Spain was adversely affected by just a few of the pollution paradoxes. Currently, the most severely affected areas due to COVID-19 are Madrid and Barcelona, but the regions of Andalusia, Extremadura, Castilla y León, and Castilla-La Mancha and Catalonia are also battling with high-risk levels.

The climate in Spain is extremely varied. There are five types of climate: the climate of the Atlantic coast, cold, humid, and rainy. Therefore, Spain is not always warm and sunny. However, if you exclude the north-facing coast, in much of Spain, the summer is hot, dry, and sunny. The average temperature ranges from 10 to 40 °C.

Three primary pollutants plague Spain: nitrogen dioxide (NO2), PM10/PM2.5 particulate matter, and finally ozone, a pollutant linked to the others. According to Spain Health reports, over 17.5 million Spaniards were hospitalized in 2019. At the same time, the poor air quality and temperature were regarded as key factors to influence the health of the population. Meanwhile, the Ministry for Ecological Transition insists it is impossible to measure the effects on health with any precision. The rising level of health issues was started after economic recovery in 2015, due to burning of fossil fuels and non-renewable sources.

Among air quality indicators, nitrogen dioxide (NO2) is a byproduct of combustion at high temperatures, for instance, in-car motors and electricity plants. The most affected regions with NO2 were Madrid, Catalonia, Barcelona, and Granada. A report indicated that Madrid’s atmosphere contains more than 50% of NO2 emissions, particularly by diesel vehicles (Inventory of Contaminating Emissions). The problem with PM10/PM2.5 particulates is more than other pollutants. The PM2.5 particulate matter consists of dust, ash, soot, and similar substances produced by traffic as well as central heating systems, industry, construction, demolition works, but traffic can also be a source. It affected most parts of Granada, Madrid, Málaga, Castilla-La Mancha, and Asturias, Jaén province. Out of 126 big cities of Spain, 36 were affected due to ozone gas. Most affected areas were Madrid (Sierra Norte region severely hit in Madrid) from where the wind blows the ozone gas from the center to Castilla y León and Castilla-La Mancha. The more cars and pollution in the cities, the bigger the ozone problem in the outlying areas.

Breathing in polluted air affects our respiratory system and aggravates chronic respiratory diseases in several ways, e.g., asthma or COPD (chronic obstructive pulmonary disease) and vascular problems, such as strokes and heart attacks. The polluted particles such as PM10/PM2.5 can penetrate the human circulation, triggering inflammatory mechanisms that damage different organs. The worst affected are the lungs, the heart, and the blood vessels. In 2015, in Spain, there were 38,600 premature deaths, and 27,900 people were severely affected due to particulate pollution of air (European Environment Agency 2018).

Recently, due to COVID-19 crisis, implementation of lockdown restrictions adopted by states and the federal government has resulted in stopping of economic activities such as industrial projects, construction projects, and transportations observing the stagnant phase (Shakoor et al. 2020). Consequently, the levels of NO2 and PM2.5 particles of air contaminant have declined 64% in major Spanish cities due to COVID-19 precaution measures. In Barcelona, the pollution decline to 83%, and Madrid and València dropped the pollution level in the air by about 73% and 64%1.

Materials and methods

Study area and data specification

In this study, we collected daily data of the top four regions of Spain in COVID-19 infections. These regions include Castilla y Leon, Castilla-La Mancha, Catalonia, and Madrid, while the time of data ranges from 29 February to 17 July 2020 (European union 2020).According to World Bank records, Spain has a population of almost 47 million, with a population growth rate of 0.3% per year. The main reason behind selecting Spain is it is currently in the top ten world’s most infected country2 in the world with 260,555 confirmed cases and 28,420 total deaths. The meteorological data, including daily average temperature (AT) and daily PM2.5 (a proxy of air quality) of each region, were retrieved from the World Air Quality organization (https://aqicn.org). However, the data for daily COVID new cases is accessed from the EL PAÍS: the global newspaper3 of Spain. The daily data consists of COVID-19 new confirmed cases, the average temperature in degree celsius (°C), and real-time air quality as PM2.5.

Figure 1b illustrates the scatterplot for COVID-19 daily new cases with a trend line4. The scatterplot clearly shows that the COVID-19 transmission rate was high in March and April, then it suddenly reduced. The COVID-19 daily news cases in Madrid, Catalonia, Castilla-La Mancha, and Castilla y Leon, respectively, are shown in Fig. 2. We observe that Madrid and Catalonia (red and dark green lines) are the most affected regions of Spain, followed by Castilla-La Mancha and Castilla y Leon (green and pink lines).

Similarly, from Fig. 3a, we note that the highest fluctuation of daily temperature is recorded in Castilla-La Mancha and Madrid (pink and blue lines). From the scatterplot of temperature, we observe that Madrid and Castilla-La Manch have high temperature and more fluctuations. In the same line, we observe the PM2.5 from Fig. 4, and Madrid shows the highest level of fluctuation of PM2.5 while Catalonia and Castilla y Leon indicate the lowest level. From a graphical analysis, the authors observe that weather changes in Spain regions are similar to the COVID-19 fluctuations in Spain. Such facts and scientific reasons motivate the researchers to draw a more detailed analysis to attain our primary study objective.

Fig. 4

PM2.5 variation across Spain regions

Due to normality issues, all three variables are transformed into a natural logarithm. Table 1 shows the descriptive statistics of variables with and without log transformation. The summary statistics display that data is average, and there are no outliers in the panel data.

Table 1

Descriptive statistics

Variables	Obs	Mean	S.Dev	Min	Max	p1	p99	Skew.	Kurt.
COVID-19	560	341.948	513.348	0	3419	0	2331	2.517	10.458
PM2.5	560	39.977	18.227	5	109	6	87	.375	3.251
Temperature	560	22.196	6.984	6	39	10	37	.294	2.074
LnCovid	560	4.76	1.714	0	8.137	.693	7.754	− .404	2.74
LnTemp	560	3.048	.327	1.792	3.664	2.303	3.611	− .278	2.568
LnPM2.5	560	3.558	.558	1.609	4.691	1.792	4.466	− .994	3.775

Note: Descriptive statistics are shown with and without log transformation. The descriptive is based on the panel data of top four Spain regions in COVID-19. Ln values show descriptive with log transformation

Estimation strategy

For empirical analysis, we use the step by step approach. For instance, first, we checked the normality by descriptive analysis (Table 1). Descriptive statistics are shown with and without log transformation. The descriptive is based on the panel data of the top four Spain regions in COVID-19. In the second step, we perform a detailed correlation analysis to check the co-movement between temperature, air quality, and COVID-19 daily cases of Spain. In the next step, we normalized the data by transforming the variables into a natural logarithm.

Following the recent study of Tosepu et al. (2020), we further apply the Pearson correlation technique between three variables of COVID-19, temperature, and PM2.5 (Table 2). We further use the pairwise correlation analysis of Spearman correlation methods to examine the robustness between three variables of COVID-19, temperature, and PM2.5 (Table 3).

Table 2

Pairwise correlation analysis (Pearson correlation)

Variables	LnCovid	LnPM2.5	LnTemp
COVID-19	1
Temperature	− 0.3166*	1
PM2.5	0.371**	0.1926*	1

Note: * shows significance at the 5% level

Table 3

Pairwise correlation analysis (Spearman correlation)

Variables	LnCovid	LnPM2.5	LnTemp
COVID-19	1
Temperature	− 0.3917*	1
PM2.5	0.0238	0.1647*	1

Note:* shows significance at the 5% level

After performing the correlation analysis between our primary variables of interest, we utilize the panel regression to check if temperature and air quality exert pressure on COVID incidents of Spain. In doing so, we use panel regression, quantile regression, pooled OLS (robust Driscoll and Kraay standard errors), and fixed effects methods. The panel regression methods are employed on our baseline model given as

In Eq. (2), the variables are used with log transformation, where Temp mentions the impact of daily average temperature, Air illustrates the effect of air quality, and ε shows the error term.

Empirical results

Main findings

Table 1 shows the descriptive statistics for daily COVID-19 cases, PM2.5, and temperature variables. During the study period (29 February 2020 to 17 July 2020), there were 260,555 COVID-19 confirmed cases in Spain. Temperatures ranged from 10 to 40 °C, and PM2.5 ranged from 10 to 100 in the top four regions of Spain used in this research work. The average temperature and PM2.5 during this period were 25 °C and 45, respectively. The mean concentrations of PM2.5, temperature, and COVID-19 were 341.95, 39.98, and 21.2, respectively.

Table 2 presents the coefficients of the Pearson correlation for all possible pairs of our variables of interest. The correlation between all the variables is not the same. The values of coefficients are 1 (PM2.5 and temperature), 0.37 (COVID-19 and PM2.5), and − 0.3166* (temperature and COVID-19), respectively. Our results show that temperature and COVID-19 variables tend to show a negative and significant direction, while COVID-19 and PM2.5 show a positive and significant correlation. Table 3 presents results from Spearman’s rank correlation among all pairs of variables. All the correlation value trends show similar values nearly the same to those obtained from Pearson correlation analysis. The values of coefficients are 1 (PM2.5 and temperature), 0.0238 (COVID-19 and PM2.5), and − 0.3917* (temperature and COVID-19), respectively.

Table 4 presents the results from competing regression models. We have employed three regression models for analysis: quantile regression, pooled OLS, and fixed effects regression model. Notably, coefficients of air quality (PM2.5) are positive with significant values. The results suggest that fixed effects regression and pooled OLS provide similar coefficients for PM2.5, close in value (0.210* and 0.3181), whereas the quantile regression model presents a lower positive and significant value (0.049*) than the other two. Each 1% increase in PM2.5 results in an increase of 0.63%, 0.21%, and 0.41% in COVID-19 cases, according to pooled OLS, quantile regression, and fixed effects regression, respectively.

Table 4

Panel regression empirics

Variables	Pooled OLS	Quantile regression	Fixed effects
Temperature	− 1.7816** [2.480]	− 2.644*** [-9.450]	− 1.795*** [-8.930]
PM2.5	0.3181** [2.2100]	0.049** [0.300]	0.210** [1.600]
Constant	9.0463** [3.570]	12.889*** [13.790]	9.472*** [13.240]
R2	0.110	0.122	0.129
Observations	560	560	560

Note: The symbols *, **, and *** denote the significance level at 10%, 5%, and 1%, respectively. t-statistics of the corresponding coefficients are reflected in brackets

Coefficients of the temperature are negative and significant in the case of all competing models, suggesting that there is no role of temperature in COVID-19 infections. Results showing a pattern of temperature, fixed effects regression, and pooled OLS provide similar coefficients for temperature, close in value (− 1.795*** and − 1.7816**), whereas the quantile regression model presents a bit larger negative value (− 2.644***) than other two. Overall, the panel regression empirics imply that temperature may be negatively associated with the COVID-19 spread, and air quality may induce more infections in Spain.

Discussion of analysis

Our results are consistent with recent studies suggesting that lousy air quality leads to an increase in the number of new COVID-19 infections (Park et al. 2019). Temperature is not associated with the daily new COVID-19 infections and not helpful to slow down the spread of this disease, quite the opposite to the expectations of many experts and some earlier studies (Tosepu et al. 2020; Wang et al. 2020). The rise in temperature may reduce the COVID-19 outbreak, as indicated by the results of our correlation and regression analysis. The findings suggest that temperature may slow down the virus’ effects or reduce its survival period in air and surface. Similar values obtained from multiple models add to the robustness of our results. Recent studies in China and Iran support our results indicating no evidence was found to support the COVID-19 cases could decline when the weather becomes warmer (Xie and Zhu 2020; Jahangiri et al. 2020).

Previously, some studies reported that the temperature is the main factor in the survival and transmission of SARS-CoV and MERS-CoV diseases (Bi et al. 2007; Casanova et al. 2010; Chan et al. 2011; Tan et al. 2005; Van Doremalen et al. 2013). We compared our results with the main findings of these studies. Another study indicated that the optimum environmental temperature related to SARS cases was from 16 to 28 °C based on data from Hong Kong, Guangzhou, Beijing, and Taiyuan (Tan et al. 2005). Moreover, Bi et al. (2007) reported that temperature had a negative relationship with SARS transmission in Hong Kong and Beijing in 2003. Another laboratory study found that coronavirus on smooth surfaces was stable for over 5 days when the temperature was at 22–25 °C (Chan et al. 2011). In brief, most studies showed that there was an optimum temperature (22–25 °C) for coronavirus, and the high temperature was harmful to its viability. However, our findings could not observe an adverse effect of high temperature on COVID-19 infection. A possible reason may be that the study period starts in winter and ends in early summer, with a maximal mean temperature of 30 °C. Further laboratory studies also need to be conducted to determine the underlying mechanism.

Our analysis of PM2.5 is similar to previous reports that suggested that the impact of delicate particulate matter only 1 μg in PM2.5 has a positive and significant association with an increase in the COVID-19 spread (Wu et al. 2020b). Empirically data analyzed by Yao et al. (2020) suggested that the spatial correlation of PM10 and PM2.5 has a positive correlation with COVID-19 cases and rate in China. Furthermore, the researchers suggested that pollution caused by the particulate matter influences the prognosis of the patients. Fattorini and Regoli (2020) analyzed the association of air quality and COVID-19 outbreak in Italy. They concluded that continuous exposure to environmental pollution (PM10 and PM2.5) in Northern Italy is the primary cause of COVID-19 cases. Similarly, Shahzad et al. (2020) and Piazzalunga-Expert (2020) for the case of China, Italy, and India supported similar findings.

Keeping in touch with earlier literature (Amuakwa-Mensah et al. 2017), recent studies are focusing on how population size and population density affect both the current and future spread of COVID-19 disease (Jahangiri et al. 2020; Rocklöv and Sjödin 2020; Zhang et al. 2020).

However, it might be possible that the incidence rate of the infection depends on population density and other factors associated with warmer and drier weather (Byass 2020). It also needs to know why, to date, the epidemic has hit so hard several highly densely populated areas around the world, e.g., New York in the USA and Madrid metropolitan area in Spain and densely populated areas of Brazil, Italy, Germany, India, and so forth. These findings may be limited in application to a European country (Spain) due to its unique climate, social setup, and several other factors related to healthcare facilities and densely populated regions data. Nevertheless, the result provides a valuable perspective on climate-infectious-diseases-control studies. Since the outbreak of COVID-19 is still out of appropriate control in Spain at the moment, the data related to new infections and climate factors are in the process of formation. Additional studies at the end of this outbreak will be more helpful in understanding the role played by the climate factors in COVID-19 spread or control.

Conclusion

This is probably the first article to investigate the correlations between average temperature, air quality, and COVID-19 in Spain population. The authors employed extensive panel data ranging from February to 17 July 2020 to report more recent and conclusive evidence. The present study utilizes correlation and panel regression methods to investigate the association between temperature, air quality, and COVID-19 across four regions of Spain with high infections. The empirics conclude that weather might be a significant factor in determining the incidence rate of COVID-19 in Spain. The correlation findings suggest that air quality may significantly increase the COVID-19 transmission with 0.37% in all regions of Spain.

On the contrary, the correlation empirics conclude that temperature may not induce the COVID-19 spread. Instead, it might act as a contributory factor to reduce its transmission rate with 0.39%. Interestingly, all of the panel regression findings support the correlation results and argue that temperature may reduce the COVID-19 spread. More precisely, 1 °C change in temperature may reduce the COVID-19 spread with 1 to 2 new cases, while the increase in air quality may enhance it from 0.5 to 0.21% influence. The findings of this work are very innovative and support the narrative of a few existing studies, that temperature negatively affects the COVID-19 daily infections. The results of this work can assist the policymakers in suppressing COVID-19 disease in Spain for humanity-saving efforts.

Recently, the Spanish health authorities have warned that young people are being admitted into intensive care as acute cases of COVID-19, which is very alarming and dangerous for Spain. In July 2020, Spain is facing an average daily 1200 to 1500 new cases of COVID-19, even with the young and older population. The authors argue that COVID-19 will be a window for research for many decades, and future research will guide us more about its causes, treatment, and adverse effects.