Mathematical analysis of COVID-19 pandemic by using the concept of SIR model.

The health organizations around the world are currently facing one of the greatest challenges, to overcome the current global pandemic, COVID-19. It erupted in December 2019, in Wuhan City, China. It spreads rapidly throughout the world within couple of months. In this paper, the data of the COVID-19 have been collected, organized, analyzed and interpreted using the discrete-time model of SIR epidemic model. Moreover, results for several countries from different regions of the world have been obtained. Furthermore, comparative study has been carried out for the countries under consideration. The comparison was performed for the data of different countries on same dates of each month. However, the calculations are carried out for thirteen consecutive weeks, to investigate the rate of spread and the control of the disease in these countries. This guides us to some important concepts like factors favoring the spread of virus and those resisting the spread. Different regions are studied and their data have been evaluated to know which regions are the most effected. This study helps to know the important factors about the behavior of the coronavirus in different environments, such as lockdowns, temperatures, humidity and other restrictions. The proposed concepts and equations can be used to project the upcoming behavior of the pandemic.

Introduction

In December 2019, the first cases of COVID-19 disease were officially reported in Wuhan City, China (Huang et al. 2020; Lu et al. 2020). In no time, it spreads worldwide. COVID-19 disease is caused by the novel coronavirus known as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (Lai et al. 2020). It belongs to the family of coronaviruses, most of which infect the animals (Lu et al. 2020). Some coronaviruses are human coronaviruses (HCoV) that infect human, for instance 229E, NL63, OC43 and HKU1 (Hoek 2007). Sometimes the animal’s coronaviruses evolve and infect people such as severe acute respiratory syndrome coronavirus (SARS-CoV) in 2002–2004 (Lu et al. 2020; He et al. 2003) and Middle East respiratory syndrome coronavirus (MERS-CoV) in 2012 (He et al. 2003). Both of these coronaviruses evolved from the animals (Wit et al. 2016). COVID-19 is also one of these evolved coronaviruses. All three of these coronaviruses infect the respiratory system of the host. In severe cases, when the upper respiratory system cannot cope with the virus, the viruses approach the lungs resulting in fatal pneumonias (Sims et al. 2005.) COVID-19 disease is highly contagious and deadly virus that spreads rapidly. The sickness can be asymptomatic in some patients, while others can have severe symptoms that can take them to ventilators (Furukawa et al. 2020). The virus can be transmitted from both types of patients, i.e., person having symptoms and no symptoms (Furukawa et al. 2020).

COVID-19 is an infectious disease and spreads by direct or indirect contact of people (Chan et al. 2020). Since the virus dwells in the respiratory organs, it spreads through oral secretions carrying viruses such as coughs, sneezes, saliva and talking of the sick person can transfer the coronavirus (Liu et al. 2020). The infection is spread in the healthy community in two ways.

Direct

Inhalation of droplets of infective in the air or by close contact with infected person during handshaking etc.

Indirect

The virus can survive and multiply in the air. Therefore, it can be easily spread through orofecal and other environmental roots such as food, fomite, fingers, feces and flies. Whenever an infected individual coughs, sneezes or talks, he releases tiny droplets containing the viruses. These germs stay in the air or may land on any surface or ground.

A healthy person becomes infected if he/she inhales these air droplets, or through contact with the viral surface because the germs stick to his hands. When he touches his face, mouth, nose or eyes, the virus enters his body and eventually the contact becomes COVID patient (Mackenzie and Smith 2020). Furthermore, studies have shown that SARS-CoV and MERS-CoV also infected the digestive-tract, so there are also chances that COVID-19 may also spread through fecal excretions (Mackenzie and Smith 2020). These viruses can survive up to several hours on different surfaces depending on the nature of the surface (Doremalen et al. 2020).

The most effective tactic to prevent this pandemic from spreading is to reduce the contacts of people. Therefore, the world is following the safety measures like isolating the infected, quarantine the suspected, social distancing and locking down the areas that are most affected. It has been controlled in some countries, and others are still facing the problem.

Since epidemics and pandemics are of great concern to everyone around; researchers from all over the world are putting their efforts to combat the disease. Mathematicians also contributed in the medical sciences and provided numerous mathematical models for epidemic diseases. These models provide information about the forthcoming behavior of the disease. In addition, these models are used to predict the time when a certain disease reaches its peak, and guess when the disease will die off. Likewise, an approximate number of infected people can be found at a given time (Hethcote 2000). Most of these models divide the total population in a number of compartments. The flow of population through these compartments depends on the disease being modeled. Different diseases involve different parameters, such as types of infection, symptoms, vaccination, medication, isolation, and quarantine. Some epidemic models that follow the compartmental division of the total population are SI, SIS, SIR, SIRS, SIRD, SEI, SEIS, SEIR, SEIRS, MSIR, MSEIR, MSEIRS and carrier state (Roberts and Heesterbeek 2003; Anderson 2013; Anderson 1992; Bailey 1975,1982; Becker 1982; Castillo-Chavez 2013; Daley and Gani 2000; Becker 1979; Al-shami 2021; Dietz and Schenzle 1985; Hethcote 1994; Isham and Medley 1996; Lauwerier 1981; Mollison and Denis 1995; Wickwire 1977). Each of these models comprises of different compartments that are labeled by alphabets. The names are given by the arrangement of these alphabets in which the population flows. The choice of model and the compartments depends on the disease. Also, different assumptions are made on each model accordingly. For example, in SI model, S stands for susceptible which contains healthy people that are vulnerable to specific disease and I stands for infected class, the people who can transfer the disease to S community. The name also indicates the population flow that initially everyone lies in S compartment and then moves to I. Similarly, SEI means initially people lie in susceptible (S) then move to exposed (E) compartment which contains people who have been exposed to disease. Finally, the person moves from E to infected (I). Moreover, the letters M and D used in some of the models named above stand for “maternal derived immunity” and “diseased,” respectively. Table 1 describes the flow of population for abovementioned compartmental disease models. Some of these compartmental models can be formulated using difference equations for a discrete-time interval and also by differential equations if the time interval is sufficiently small and a continuous behavior is required (Allen 1994). The discrete-time models are solved numerically whereas the other type of models are solved analytically. Discrete-time models use iterative schemes to be solved. It is a step-wise procedure in which each step depends on the results from the previous step. Moreover, it requires a data for each step to proceed further. On the other hand, the continuous time models do not have such limitations. They are solved by using formulae and different techniques to get a stable solution. The choice of the models depends on the nature of study and the types of required results. As an illustration, the basic SIR model is presented in Fig. 1.

Table 1

Description of different compartmental disease models explained with the flow of population

Model	Process and description
SI	Susceptible → Infected
SIS	Susceptible → Infected → Susceptible
SIR	Susceptible → Infected → Recovered
SIRD	Susceptible → Infected → Recovered or Diseased
SIRS	Susceptible → Infected → Recovered → Susceptible
SEI	Susceptible → Exposed → Infected
SEIS	Susceptible → Exposed → Infected → Susceptible
SEIR	Susceptible → Exposed → Infected → Recovered
SEIRS	Susceptible → Exposed → Infected → Recovered → Susceptible
MSIR	Maternal derived immunity → Susceptible → Infected → Recovered
MSEIR	Maternal derived immunity → Susceptible → Exposed → Infected → Recovered
MSEIRS	Maternal derived immunity → Susceptible → Exposed → Infected → Recovered → Susceptible
Carrier State	Susceptible → Infected → Recovered or Carrier

Fig. 1

This flowchart explains the flow of population in an SIR epidemic model, where S is the group of susceptible, I is infected group and R is recovered group

In this article, an SIR model developed by Kermack and McKendrick (1991) has been presented with some modifications. The modifications made makes the model meet current situations and the data available. For instance, the usage of removed class instead of recovered class, which covers the recovered as well as the fatalities. The SIR epidemic model has been preferred over other compartmental epidemic models because the available data are limited. So, in order to conclude some valuable results using limited data, the SIR model is the most suitable and ideal. The fundamental difference between a discrete-time model and the continuous model is the obviously the time intervals. The continuous model gives a continuous analysis whereas the discrete-time model uses discrete intervals with step-wise analysis. Another difference is the form of equations involved in these models. In a discrete-time model, the equations are difference equations with basic arithmetic operators, which can easily be solved using numerical methods. While in continuous model, the equations contain the differential terms which are solved analytically for a stable solution. Researchers have discussed the continuous form of the models. Some models are used to analyze COVID-19 pandemic (Alkahtani and Alzaid 2020; Pandit 2020; Chen et al. 2020; Cooper et al. 2020; Calafiore et al. 2003). For the contribution of mathematicians in the medical sciences refer to Rebouças Filho et al. (2014); Rebouças Filho et al. 2017; Rodrigues et al. 2018; Parah et al. 2020; Ohata et al. 2020; Dourado et al. 2020). Here, a novel approach has been adopted by using the discrete-time SIR epidemic model. This is so simple, yet provides valuable conclusions. Since, the equations in a discrete-time SIR model are solved by numerical methods, therefore, the results are approximate. Moreover, real data of COVID-19 have been computed in the model. Hence, the results are more practical. There are two main reasons for choosing a discrete-time SIR model for studying the COVID-19; (I) It provides results for each time frame (depending on the data availability and the selection of time interval), (II) The available data were limited. The SIR model was the ultimate match for the available data. All of the other models might not give the satisfactory results due to lack of data. Also, they are difficult to model the situation using a discrete-time interval (Fig. 2).

Fig. 2

Flowchart of the tweaked SIR model which is being adopted in current study. The countries under study are highlighted in red

SIR model divides the entire population in three groups, that is susceptible, infected and removed (Allen 1994; Kermack and McKendrick 1991). Table 2 describes the details of the symbols being used. In Sect. 2, the SIR model is described in detail and the difference equations are derived, modified for the purpose of required outcomes. Additionally, this article aims to compare the spread of disease and removal rates in different countries from different regions of the world at a given time frame using real data of the cases. However, the main focus of this study is the rate at which the infection is spreading. Hence, the cumulative class (removal class) has been taken instead of recovered and diseased. This accelerates and simplifies the calculations, and required results are obtained quickly and easily. On daily basis, the data were collected from the website of Worldometer www.worldometers.info. Then, these data were organized on weekly basis from first week of April to first week of July, those are thirteen weeks. Since this article focuses on comparison of different regions, therefore minimum countries are taken with maximum differences like different global locations, different atmospheres, weather conditions, and epidemic stages. In the given time frame, some countries were in the initial stages (South Africa), while others were in the middle (Australia, USA), peak (Italy) and final stages (China). The countries under study include USA, Italy, China and India from northern hemisphere, while South Africa and Australia from southern hemisphere. Each of these countries is discussed alongside with their data and effective factors in separate subsections. Then, all of these countries are compared by discussing all the differences and similarities between them and the impact of all factors on the spread and control of the pandemic.

Table 2

Chart of units and description of the variables used

Variable	Description	Units
N	Total population	Persons
S	Number of Susceptible individuals	Persons
I	Number of Infected individuals	Persons
R	Number of Removed individuals	Persons
μ	Infection rate	1/Persons×days
α	Removal rate	1/Persons
Δt	Time interval	Weeks

The rest of the paper is arranged as: Sect. 2 contains the mathematical model for the disease and the derivation of equations for infection and recovery rates. In Sect. 3, numerical calculations are carried out. The data of different countries have been analyzed separately. The numerical results are written in the tables and pictorially represented by graphs. In Sect. 4, different regions are compared and the final conclusions are given in the last section.

Model and formulation

In this section, the model for COVID-19 pandemic is explained and the equations for the model are solved for the required variables. The general equation for each of the required variables is obtained.

There are several models for epidemic diseases. Here, an SIR model is being followed (Allen 1994; Kermack and McKendrick 1991). This model divides the total population () into the following three compartments.

Susceptible (): This compartment covers the individuals that are sensitive to virus, they might catch the disease. In other words, the healthy people that are vulnerable to the virus, fall under the title of susceptible.

Infected (): Those people who have been infected by the virus and they are sick fall in this group of the model.

Removed (): Obviously, every sick person will be eventually removed from the infected category; either by recovering with immunity or dying due to infection. Hence, the people recovering of dying move to the removed compartment.

Hence, the sum of susceptible, infected and removed people yields the total population.

The initial conditions for the model arewhere represents the initial susceptible, represents the initial infected and represents the initial removed. If time interval is taken to be one week then , and represent the susceptible, infected and removed, respectively, after weeks. It is assumed that the total population is constant. Therefore, (1) can be rewritten as

After each time interval, people get infected and others are removed from the infection. So, the number of people getting disease move from susceptible compartment to the infected compartment . Therefore, is monotonically decreasing function of time such that as . If is considered as a rate of contact sufficient to transfer disease, between susceptible and an infected person, then the new infectives due to one infected person after time interval is given by . If there is a whole class of infectives then the number of new infectives after time interval is . Hence, people are removed from susceptible compartment and move to infected compartment after each time interval .

On the other hand, people are also getting recovered and losing their lives to the disease, they are removed from the infected compartment and moved to the removed compartment . Hence, is a monotonically increasing function of time . If is the rate of removal then the number of individuals removed from the infectiousness after time interval is given by . Therefore, people leave the infected compartment and join the removed group .

The flow of population between the compartments after unit time intervals () is expressed by the following equations

Equation (5) is solved for and (7) for giveand

Note that the removal rate is the sum of death rate π and recovery rate λ.

Equation (9) is for the rate of infection in the unit time interval and (10) for the removal rate.

Numerical analysis

In this section, the model is applied to the data of COVID-19. Data for several countries were gathered every day. Only six countries were considered for study to cover different environments and stages of pandemic. The calculations are carried out to compare the outcomes and achieve some valuable results. These results have been found in separate subsection for each of the countries under research. The data and outcomes of calculations are expressed in tabular form. In addition, the graphs and bar charts are constructed for visual study and better understanding. The following flowchart in Fig. 3 summarizes the process of calculation process.

Fig. 3

Flowchart for the process of numerical calculations

USA

The estimated total population () of the USA was recorded to be 330,758,784 at the beginning of trial. The first case of COVID-19 in the USA was recorded on 21st January 2020. Soon, a public health emergency was declared on 31st of January by the government https://www.npr.org/sections/health-shots/2020/01/31/801686524/trump-declay-and-restricts-travel-from-c. The data are organized in a tabular form. Table 3 shows the numbers of susceptible, infected , removed , total cases, deaths and recovered individuals for thirteen consecutive weeks from first week of April to first week of July. That are the 11th to 23rd week of pandemic in USA. , and represent the initial numbers of people in susceptible, infected and removed compartments, respectively. The in the subscripts of, and indicates that the data values are of 1st week, i.e., week . Similarly,, , and represent the data of week 1, and , , and denote the number of individuals lying in susceptible, infected and removed compartment, respectively, of week n (n = 1, 2, 3, …, 13).

Table 3

Data of USA. Total Population () = 330,758,784. Complete data of 13 weeks. The subscript in , , and denotes the week number from 0 to 12

Weeks	Total Cases	Deaths	Recovered	Removed (R_n)	Infected (I_n)	Susceptible (S_n)
0	233,246	5738	10,284	16,022	217,224	330,525,538
1	435,169	14,797	22,897	37,694	397,473	330,323,615
2	619,331	27,176	47,069	74,245	545,086	330,139,453
3	903,298	50,954	90,261	141,215	762,083	329,855,486
4	1,078,476	62,535	149,686	212,221	866,255	329,680,308
5	1,279,391	76,206	214,276	290,482	988,909	329,479,393
6	1,430,348	85,197	310,259	395,456	1,034,892	329,328,436
7	1,602,466	95,467	371,795	467,262	1,135,204	329,156,318
8	1,757,130	102,794	494,722	597,516	1,159,614	329,001,654
9	1,901,723	109,140	688,669	797,809	1,103,914	328,857,061
10	2,077,090	115,572	810,350	925,922	1,151,168	328,681,694
11	2,245,568	120,160	919,405	1,039,565	1,206,003	328,513,216
12	2,447,449	123,934	1,025,350	1,149,284	1,298,165	328,311,335

The data available include the total cases, active cases (infected), deaths, recoveries and the total population. The susceptible and removed were calculated from these data. Finding the number of removed individuals is so simple as it is the number of people leaving the infected compartment that are either recovered with immunity or dying of infection. Hence, the sum of dead and recovered yields the removed. However, the assumption of constant population is used for the calculation of susceptible in each week. Hence, (2) is rearranged in the following order

So, can be found using the above equation by simply putting the values of , and for any week. Since the number of . Therefore,

Now suppose that the rate of infection in first week is , then by (9)

From Table 3, the values of in week 0 and 1 are and , respectively, while the value of in week 0 is . Computing these values in the above formula yields as follows:

Similarly, the rate of infection in second week is calculated as

From Table 3, the value of in week 2 is , while the value of in week 1 is . Computing these values in the above formula yields as follows:

Continuing this process for all the weeks, the rates are calculated and written in Table 4.

Table 4

Rates in USA. The calculated rates of infection and removal rates for each week are arranged in an order

Rates of Infection		Rates of Removal
μ_1	2.812×10^-9	α_1	0.0998
μ_2	1.404×10^-9	α_2	0.0920
μ_3	1.578×10^-9	α_3	0.1229
μ_4	6.969×10^-10	α_4	0.0931
μ_5	7.035×10^-10	α_5	0.0903
μ_6	4.633×10^-10	α_6	0.1062
μ_7	5.050×10^-10	α_7	0.0694
μ_8	4.139×10^-10	α_8	0.1147
μ_9	3.790×10^-10	α_9	0.1727
μ_10	4.831×10^-10	α_10	0.1161
μ_11	4.453×10^-10	α_11	0.0987
μ_12	5.096×10^-10	α_12	0.0910

Also, removal rate for each week has been calculated as follows:

From Table 3, and . Computing these values in the above formula yields as follows:and

From Table 3, . Substituting this value in the above formula gives as follows:

In the same way, all the removal rates for each are calculated and written in Table 4.

It is observed from the calculations that the rate of infection was so high in 1st week then it dropped rapidly after four weeks. This indicates that the infection was spreading at high rates in the first 10 to 15 weeks of the epidemic in the region. And it dropped down to almost a constant rate after that. There are many factors that are responsible for this change. Initially, USA reacted slowly to control the spread of disease, so it spreads at higher rates. They started to lockdown the affected areas one after another from 19th of March, that is after 9 weeks of the pandemic https://www.wsj.com/articles/china-reports-no-new-domestic-coronavirus-infections-for-the-first-time-since-outbreak-started-11584611233. Moreover, the weather changed during March and April as the spring starts in USA (https:seasonsyear.com, USAS.). Obviously by locking down, the rate of contacts between people dropped. The virus could not find as many new hosts to infect as before. Therefore, a remarkable drop down of the curve of rate of infection can be seen in Fig. 4. The weather also affects the survival of the virus in an environment on different surfaces. Temperature and humidity have been known to affect the age of the coronavirus; the higher temperature goes the quicker it deactivates (Casanova et al. 2010). From the calculations, the removal rate has been obtained to be around 0.1 infected persons per week that means on average, each week 10% of the infected people were removed from the infectious state, either by defeating the disease and recovering with the immunity, or being defeated by the disease and losing their lives.

Fig. 4

Rate of infection in USA

The associated graphs for the data in the above tables are given below in Figs. 4 and 5.

Fig. 5

Rate of removal in USA

Italy

The total population of Italy was 60,440,954 according to Worldometer (Dourado et al. 2020). Table 5 contains the data of COVID-19 cases for the same time frame that is from the first week of April to first week of July. The disease travelled from China to Italy on 31st of January and spread rapidly with great death rate https://www.corriere.it/cronache/20_gennaio_30/coronavirus-italia-corona-9d6dc436-4343-11ea-bdc8-faf1f56f19b7.shtml. According to these dates, the data in below table are of 10th to 22nd weeks of the epidemic in Italy.

Table 5

Weekly data and calculated results for Italy. Total Population () = 60,440,954

Weeks	Total Cases	Deaths	Recovered	Removed (R_n)	Infected (I_n)	Susceptible (S_n)	Infection Rates(μ_1)	Removal Rates(α_n)
0	115,242	13,915	18,278	32,193	83,049	60,325,712	Nil	Nil
1	139,422	17,699	26,491	44,190	95,262	60,301,532	4.826×10^-9	0.1445
2	165,155	21,645	38,092	59,737	105,418	60,275,799	4.486×10^-10	0.1632
3	192,944	25,969	60,498	86,467	106,527	60,248,010	4.373×10^-9	0.2536
4	205,463	27,967	75,945	103,912	101,551	60,235,491	1.951×10^-9	0.1638
5	215,858	29,958	96,276	126,234	89,624	60,225,096	1.699×10^-9	0.2198
6	222,104	31,106	112,541	143,647	78,457	60,218,850	1.157×10^-9	0.1943
7	228,006	32,486	134,560	167,046	60,960	60,212,948	1.249×10^-9	0.2982
8	231,732	33,142	150,604	183,746	47,986	60,209,222	1.015×10^-9	0.2740
9	233,836	33,601	160,938	194,539	39,297	60,207,118	7.282×10^-10	0.2249
10	236,142	34,167	171,338	205,505	30,637	60,204,812	9.747×10^-10	0.2791
11	238,159	34,514	180,544	215,058	23,101	60,202,795	1.094×10^-9	0.3118
12	238,833	34,675	184,585	219,260	19,573	60,202,121	4.846×10^-10	0.1819

Now, the rates of infection ,, …, and the removal rates are calculated in the same manner as in previous subsection using (9) and (10). The final results are mentioned in Table 5. The corresponding graphs for each of the rates are in Fig. 6 for infection and Fig. 7 for removal.

Fig. 6

Rate of infection in Italy

Fig. 7

Rate of removal in Italy

From the calculated results and their graphs, it can be seen that the infection rate has dropped after 4 weeks. That is 14th week of epidemic in the nation. The lockdown started on 9th of March (Sylvers and Legorano 2020) that is the sixth week of the epidemic in Italy. In this region, March to May is the spring season and the average temperature in these days remains between 5 and 25 °C. (https:seasonsyear.com,Italy.). The removal rate of Italy improved with the passage of time. In earlier stages, there were too many deaths and less recoveries but later on more people started to recover from the infection. The reason for higher death rate in Italy is due to low immune system of its people as there are many old age people. According to Statista, a German data base company, 23.1% of the Italian population are aged 65 and older https://www.statista.com/statistics/569201/population-distribution-by-age-group-in-italy/#:~:text=In%202020%2C%2023.1%20percent%20of,14%20years%20old%20and%20younger.. The removal rate was about 15% of the total infectious people at the beginning but it is a bit higher toward the end, consider Fig. 7. In the eleventh week, the removal rate was the highest, where almost 31% of infected people were removed from the infected class.

India

It was 30th of January when the first case of COVID-19 disease was reported in India https://www.hindustantimes.com/india-news/india-most-infected-by-covid-19-among-asian-countries-leaves-turkey-behind/story-Jjd0AqIsuL3yjMWg29uJ3I.html. It spreads very quickly in the region. It is one of the most populated countries in the world. According to Worldometer, the total population of India was 1,378,270,651, which has been assumed to be constant throughout the calculations of 13 weeks. The available data are from 10th to 22nd week of a public health outbreak in the country. Since the record of total cases was available, therefore the susceptible class has been calculated by the assumption of constant population. The numbers can be seen in the table below (Table 6).

Table 6

Weekly data and calculated results for India. Total Population () = 1,378,270,651

Weeks	Total Cases	Deaths	Recovered	Removed (R_n)	Infected (I_n)	Susceptible (S_n)	Infection Rates(μ_1)	Removal Rates(α_n)
0	2341	68	177	245	2096	1,378,268,310	Nil	Nil
1	6237	186	569	755	5482	1,378,264,414	1.349×10^-9	0.2433
2	12,320	405	1432	1837	10,483	1,378,258,331	8.051×10^-10	0.1974
3	24,434	780	5,457	6,237	18,197	1,378,246,217	8.384×10^-10	0.4197
4	34,780	1,151	9,068	10,219	24,561	1,378,235,871	4.125×10^-10	0.2188
5	56,325	1,889	16,776	18,665	37,660	1,378,214,326	6.365×10^-10	0.3439
6	78,194	2,551	26,400	28,951	49,243	1,378,192,457	4.213×10^-10	0.2731
7	115,800	3,519	47,053	50,572	65,228	1,378,154,851	5.541×10^-10	0.4391
8	165,362	4,711	70,788	75,499	89,863	1,378,105,289	5.513×10^-10	0.3822
9	216,824	6,088	104,071	110,159	106,665	1,378,053,827	4.156×10^-10	0.3857
10	298,021	8,500	146,851	155,351	142,670	1,377,972,630	5.524×10^-10	0.4237
11	372,767	12,389	197,894	210,283	162,484	1,377,897,884	3.802×10^-10	0.3850
12	472,972	14,907	271,688	286,595	186,377	1,377,797,679	4.476×10^-10	0.4697

Table 6 also contains the rates of infection and removal, which have been calculated using the same equations as in preceding subsections. Similarly, Figs. 8 and 9 represent the rates of infection and removal, respectively. It is learned from the graph that the rate of infection was higher in the earlier weeks then it dropped after 4 weeks. There are some ups and downs but the curve is oscillating with a little amplitude around some low rate number. A national level lockdown and restrictions were imposed on 25th of March https://www.indiatoday.in/india/story/novel-coronavirus-covid19-latest-news-update-india-lockdown-confirmed-positive-cases-deaths-uk-usa-italy-iran-china-1658922-2020-03-24. The temperatures range is so vast that depends on the location. In June, the weather is scorching hot in many regions of the country and it is the summer. Here too the virus seems to spread slowly with low rate after fourteen weeks of its arrival in the country.

Fig. 8

Rate of infection in India

Fig. 9

Rate of removal in India

The removal rate is lower in the first weeks. Although the rate is not increasing every week, but the end results show that the removal rate improved as can be seen in Fig. 9. It reaches the highest value in the last week, that is about 0.5 sick people being removed from the infectiousness per week. This average removal rate is so high, that almost half of the infected population got removed every week from the infection.

China

In December 2019, there was a cluster of suspected pneumonia patients in Wuhan, Hubei. Chinese authorities informed the World Health Organization about the situation https://www.who.int/csr/don/05-january-2020-pneumonia-of-unkown-cause-china/en/. Soon the problem got the attentions of everyone. The patients of pneumonia were increasing rapidly, so it was declared an outbreak. It occurred at the occasion of the Chinese New Year. Most of the people were out of homes, touring and travelling in vacations. This helped the virus to spread faster and made it unable to be controlled in its early stages. China started to put travelling restrictions and locked down the cities. Moreover, they extended the lockdown to provincial levels. Following this strategy, they succeeded in containing the epidemic. Things were under control in China.

During the time frame of this study, China had almost completely defeated the disease. People were recovering from the sickness, and there were few hundred infected people left. The lockdowns have been lifted, and the life had become normal.

The population of China was 1,439,323,776. Table 7 contains all the numerical information needed for the calculations of 13th to 26th weeks of epidemic in the so-called mainland.

Table 7

Weekly data and calculated results for China. Total Population () = 1,439,323,776

Weeks	Total Cases	Deaths	Recovered	Removed (R_n)	Infected (I_n)	Susceptible (S_n)	Infection Rates(μ_1)	Removal Rates(α_n)
0	81,589	3318	76,408	79,726	1,863	1,439,242,187	Nil	Nil
1	81,865	3335	77,370	80,705	1,160	1,439,241,911	1.029×10^-10	0.5255
2	82,295	3342	77,816	81,158	1,137	1,439,241,481	2.576×10^-10	0.3905
3	82,804	4632	77,257	81,889	915	1,439,240,972	3.111×10^-10	0.6429
4	82,862	4633	77,610	82,243	619	1,439,240,914	4.404×10^-11	0.3869
5	82,885	4633	77,957	82,590	295	1,439,240,891	2.582×10^-11	0.5606
6	82,929	4633	78,195	82,828	101	1,439,240,847	1.036×10^-10	0.8068
7	82,967	4634	78,249	82,883	84	1,439,240,809	2.614×10^-10	0.5446
8	82,995	4634	78,288	82,922	73	1,439,240,781	2.316×10^-10	0.4643
9	83,021	4634	78,314	82,948	73	1,439,240,755	2.475×10^-10	0.3562
10	83,057	4634	78,361	82,995	62	1,439,240,719	3.427×10^-10	0.6438
11	83,293	4634	78,394	83,028	265	1,439,240,483	2.645×10^-9	0.5323
12	83,430	4634	78,428	83,062	268	1,439,240,346	3.592×10^-10	0.1283

Even though the spread of disease has been contained, the rates of infection and removal rates have been calculated and organized in Table 7. A visualization of each of the rates is given in Figs. 10 and 11. The infection rate was extremely low throughout the experiment, but a great spike appears in the 11th week. As mentioned earlier that China was in its final stages and has almost coped with the virus through its strict and large lockdowns. But as the things got normal, the severity of lockdowns dropped; therefore after two months, a first case of COVID-19 was reported in Beijing. Later, there were dozens of infectives. The testing process accelerated and strict safety measures were ordered. As a result, the matter resolved and the rate of infection got controlled after a week. It was traced that the virus spread through the meat and vegetable market in the city https://www.dw.com/en/coronavirus-fresh-cluster-prompts-partial-beijing-lockdown/a-53793212.

Fig. 10

Rate of infection in China

Fig. 11

Rate of removal in China

On the other hand, the average removal rate was about 0.498 that is much better than other countries. China overcame the sickness by restricting the contacts of people through quarantines and strict lockdowns. The weather was also getting warmer and humid, which helped to quickly deactivate the virus in the environment (Casanova et al. 2010).

3.5 Australia

The total population of Australia was 25,461,791 in April to June 2020. The disease arrived to Australia on 25th of January, when a man who returned from Wuhan, China to Victoria, Australia https://www.health.gov.au/ministers/the-hon-greg-hunt-mp/media/first-confirmed-case-of-novel-coronavirus-in-australia. This was the first wave of epidemic in Australia, which did not do much damage as compared to the second wave which approached during May and June. The data under study are from 10th to 22nd week of pandemic in the region which can be seen in Table 8. The rates of infection and rates of removal for each of the weeks have been calculated and can be seen in Table 8.

Table 8

Weekly data and calculated results for Australia. Total Population () = 25,461,791

Weeks	Total Cases	Deaths	Recovered	Removed (R_n)	Infected (I_n)	Susceptible (S_n)	Infection Rates(μ_1)	Removal Rates(α_n)
0	5137	25	345	370	4767	25,456,654	Nil	Nil
1	6104	51	2987	3038	3066	25,455,687	7.969×10^-9	0.5597
2	6447	63	3686	3749	2698	25,455,344	4.395×10^-9	0.2319
3	6675	79	5136	5215	1460	25,455,116	3.320×10^-9	0.5434
4	6753	91	5715	5806	947	25,455,038	2.099×10^-9	0.4048
5	6896	97	6040	6137	759	25,454,895	5.932×10^-9	0.3495
6	6989	98	6297	6395	594	25,454,802	4.814×10^-9	0.3399
7	7081	100	6472	6572	509	25,454,710	6.085×10^-9	0.2980
8	7150	103	6580	6683	467	25,454,641	5.326×10^-9	0.2181
9	7240	102	6664	6766	474	25,454,551	7.571×10^-9	0.1777
10	7285	102	6761	6863	422	25,454,506	3.730×10^-9	0.2046
11	7391	102	6877	6979	412	25,454,400	9.868×10^-9	0.2749
12	7558	104	6942	7046	512	25,454,233	1.592×10^-9	0.1626

The curve of rate of infection in graph behaves differently than that the curve of other countries. It could be due to the difference in seasons of Australia and the countries in northern hemisphere of the globe. In Australia, it was the beginning of winter, when the temperatures were dropping with the passage of time. Although the rate of infection is also relatively higher than other mentioned countries, but the number of sick people are not so high. Actually, the infection rate indicates that how many new people in susceptible class are being infected by the infective class. Since there were less infectives, so new cases are also less. The higher rate does not mean greater numbers, it depends on the population and current infective class. A national level lockdown started on 23rd of March and ended on 15th of May https://www.straitstimes.com/asia/australianz/australia-starts-lockdown-measures-as-coronavirus-cases-jump. According to these dates, the lockdown was lifted in sixth week of our data. The effects of lifting the lockdown can be observed by the graph in Fig. 12.

Fig. 12

Rate of infection in Australia

The rate of removal is higher as well; Fig. 13. In Australia, the rate of infection was lower in the first wave of pandemic but it crept up to significantly large number in the last week as can be seen in Fig. 12.

Fig. 13

Rate of removal in Australia

South Africa

On 5th of March, the first case of COVID-19 appeared in the country of 59,212,004 people, it was a man who came back from Italy https://www.gov.za/speeches/health-reports-first-case-covid-19-coronavirus-5-mar-2020-0000. Table 9 contains the data of South Africa for thirteen weeks. That are 5th to 17th weeks of the pandemic in the area. Unlike northern hemisphere, in South Africa, the winter season starts in June. The corresponding rates for each week have been calculated and given in Table 9. The graphs for both the rates are given in Figs. 14 and 15.

Table 9

Weekly data and calculated results for South Africa. Total Population () = 59,212,004

Weeks	Total Cases	Deaths	Recovered	Removed (R_n)	Infected (I_n)	Susceptible (S_n)	Infection Rates(μ_1)	Removal Rates(α_n)
0	1380	5	50	55	1325	59,210,624	Nil	Nil
1	1845	18	95	113	1732	59,210,159	5.927×10^-9	0.0438
2	2506	34	410	444	2062	59,209,498	6.446×10^-9	0.1911
3	3953	75	1473	1548	2405	59,208,051	1.185×10^-8	0.5354
4	5647	103	2073	2176	3471	59,206,357	1.190×10^-8	0.2611
5	8232	161	3153	3314	4918	59,203,772	1.258×10^-8	0.3279
6	12,074	219	4745	4964	7,110	59,199,930	1.320×10^-8	0.3355
7	19,137	369	8950	9319	9,818	59,192,867	1.678×10^-8	0.6125
8	25,937	552	13,451	14,003	11,934	59,186,067	1.170×10^-8	0.4771
9	37,525	792	19,682	20,474	17,051	59,174,479	1.641×10^-8	0.5422
10	58,568	1210	33,252	34,462	24,106	59,153,436	2.086×10^-8	0.8204
11	80,412	1674	44,331	46,005	34,407	59,131,592	1.532×10^-8	0.4788
12	106,108	2102	55,045	57,147	48,961	59,105,896	1.263×10^-8	0.3238

Fig. 14

Rate of infection in South Africa

Fig. 15

Rate of removal in South Africa

Here, the graph also behaves differently. Infection rate was lower in first weeks and then it rose up and begin to drop down after 11 weeks. A countrywide lockdown was put on 26th of March https://www.theguardian.com/world/2020/mar/23/south-africa-to-go-into-21-day-lockdown-on-thursday-night. which was lifted on 30th of April https://businesstech.co.za/news/government/388743/south-africas-coronavirus-lockdown-extended-by-2-weeks/. We do not see any drop in the curve of infection rate during third to sixth week, but it did not rise up either, and kept constant for almost a month. Virus began to spread quickly after the lockdown was lifted in 5th week of the data under study. The weather between April to June remains between 0 and 20 °C. Its autumn in March to May and then the winter arrives in June. The temperature is quite unpredictable in the region as it depends on the location. But its range is same as that of autumn.

From Fig. 15, it is observed that the removal rate kept on increasing with the passage of time. The highest removal was in the 10th week, when nearly 14,000 people out of 17,000 infected people got removed. The average removal rate in South Africa was 0.412 people per week that means 41% of infected people were being removed from the sickness.

Comparison

In this section, all the six countries under experiment are compared using graph for rate of infections and bar chart for the removal rates. Figure 16 displays the rates of infection of these countries in same time interval, i.e., from the first week of April 2020 to the first week of July 2020.

Fig. 16

Graph containing rates of infection for all countries with six curves representing the rates of six countries

By studying Fig. 16, it is learned that the rate of infection was higher in Australia and South Africa. Both of these countries are located in the Southern hemisphere. So their seasons are opposite to those in other four countries from northern hemisphere. In southern hemisphere, winter generally starts in June, while it is the start of summer in northern hemisphere. The rates of infection in other countries are much lower as compared to Australia and South Africa. These rates kept on falling as the temperature rose. Basically, the virus deactivates quickly in the higher temperatures (Casanova et al. 2010). As this virus can spread through indirect sources, so the deactivation time plays an important role in the spread. Therefore, it is concluded that the climate has impacts on the rate of infection. Since China was in the final stages of the pandemic in those days, so it had the least infection rate. On the other hand, South Africa is the leading competitor among the six nations.

The rates of removal were least in the USA and the highest in China as seen in Fig. 17. It is learnt from both of these figures that the least affected country was China in the given time frame, while USA is the most affected followed by Italy. The people were getting infections but removal was low. This leaves more people in the infected compartment which means more and more newly infected people, because the new cases are the multiples of rates of infection and the currently infected persons. Also, if the number of infected people increases in South Africa and Australia, then their future looks worrying.

Fig. 17

Bar chart displaying the removal rates in all countries. Each country’s data of twelve weeks have been put in a single figure. The bars represent the rates in a particular week

Since the rate of infection is directly proportional to the infective class, i.e., the higher number of infective people will infect much more people. Thus, in order to control the disease, the contacts between the people must be restricted. Besides this, the infected people must be given proper treatment so they recover as soon as possible to reduce the population in infected compartment.

Conclusions

This study aims to exhibit the technique of modeling an epidemic disease mathematically, and the way mathematical equations and formulae can be applied to the numerical data of the disease. Henceforth, the data for COVID-19 were collected to carry out the analysis and an SIR epidemic model was formulated accordingly. Unlike, many other research papers, the current study uses the discrete-time SIR epidemic model and applies it to the real numeric data of COVID-19 pandemic. The proposed SIR model is formulated for the rates of infection (at which the infection is spreading) and rates of removal (at which the infective population is being removed). Furthermore, an iterative scheme is developed to predict the forthcoming behaviors of the disease. This study solved the modeled equations using the iterative schemes that further beautifies the proposed model. Because iterations help in predicting an approximate number of infected, recovered and dead individuals at any certain time using the available data at present times. In addition, the model is applied to six different countries around the globe with different environments and different stages of the pandemic in the countries. The results were compared and reasoned by some key factors like travel restrictions, lockdown, safety measures, temperature and environment that could affect the spread and containment of the disease. This study also concludes that the temperature and infection rates are inversely proportional. Without vaccines and medications, the lockdowns, travel restrictions and safety measures have been mathematically proven to be the key factors that reduce the spread of disease. In the mean while, the importance of treatment and isolation of the infected people has been verified.