1 Introduction

In late December 2019, viral pneumonia with an unknown agent was reported, in Wuhan, China. The virus, named COVID-19 (SARS-COV-2), quickly crossed the borders and dispersed worldwide on March 25, 2020, affecting 196 countries [1]. In February 2020, the world health organization)WHO( named the new disease COVID-19 [2]. The United Nations have considered COVID-19 as the central public, human, and economic event with many unfavorable effects on various countries [3].

Although, countries have adopted different strategies such as house quarantine, mask use, and social distancing [4], the desease can still have potential devious effects on general health [5].

Notwithstanding considerable advances in all countries of the world for disease control, infectious diseases are important in epidemiology and social health [6]. Thus, all required information should be instantly accessed and utilized to assess the risk of COVID-19 and identify the locations that are currently at high risk. Affecting more than 185 countries indicates that the fast pandemic of COVID-19 requires modern technology to assess the risk of specified locations [7]. Hence, GIS can play a significant role in the spatial analysis of this pandemic virus [8]. One of the major applications of epidemiology is to facilitate the identification of geographical locations and teams of people in danger, providing the most appropriate health and social choices to mitigate risk factors [9]. Health experts have been utilizing maps for spatial analysis of diseases for 150 years [10]. Geographical epidemiology is part of anatomical epidemiology that analyzes the geographical spread of infection and mortality [11].

The first stage in analyzing geographical data is visualization and representing data in the form of geographical maps [12]. geographic information system (GIS) has been widely used in agriculture, business, environment, natural resources, urban design, and various disease modeling for years [13]. Due to the vastness of the geographical distribution and activeness of health care service centers, familiarity with GIS skills is essential for the managers of this sector [14]. Geospatial analysis capability in GIS has a direct application in modeling the spatial pandemic of diseases and their connection with environmental features and the health care system. Currently, GIS technology is considered a significant tool in the health investigation and management of infectious diseases [15, 16]. Iran is one of the countries in the world in Asia, where the first of outbreak COVID-19 was reported on March 2, 2020, in Qom province [17]. According to the Ministry of Health of Iran, the rate of the virus has increased strongly from early July to late June; approximately 23,475 to 64,695 have been confirmed, during this period with an increase in a pandemic in most provinces and terms of the population of Iran with about 5,954,962 million people [18]. It is currently one of the top countries in terms of confirmation compared to the entire population of the country [19].

The most important limitations of this study were the downward trend and the small number of confirmed cases in the first weeks of COVID-19 onset and the start of two doses of Sinofarm and Astrazenka vaccines in Iran. With the pandemic of COVID-19 in January 2020, the investigation has focused heavily on understanding the space–time dynamics of the prime wave of the disease and examining the outcomes of inhibition measures.

Hui et al. [20] examined the pandemic of COVID-19 and its trends as well as why the virus poses a long-term threat to public health. Zhonghua and Xue [21] evaluated the Spatial–temporal profile of COVID-19 in Guangdong Province and if the virus transmission chain-breaking strategies were effective. Liu et al. [22] used a statistical analysis to determine the spatial pandemic of the virus.

Xie et al. [23] examined Wuhan population density and potent economic link as the main reasons for the pandemic of the virus. They also recounted other factors, such as population pandemic, average temperature, access to transportation, and medical facilities. Guliyev [24], in a perusal of the association between COVID-19, mortality, and improved use of spacecraft models, developed a map of the virus pandemic in China. The results showed that spatial influences, directly and accidentally, affect the pandemic of the virus. Rahnama and Bazargan [25] evaluated the analysis of Spatial–temporal patterns of the COVID-19 pandemic in Iran using descriptive-analytical methods and spatial distribution modeling. They discovered that the most significant geographical factor for the pandemic of COVID-19 is the distance and vicinity of the provinces affected by COVID-19 in the Iran, which follows the pattern of the spatial distribution of adaptation. Abolfazl et al. [26] modeled the pandemic of COVID-19 geographically and its variability across the United States. They utilized GWR and multi-geographic weighted regression (MGWR) to study the pandemic of the virus. The results indicated that GWR has reasonable performance in spatial analysis of COVID-19 pandemic as compared to the MGWR model. Pourghasemi et al. [27] analyzed the risk factors for COVID-19 pandemic to identify the locations at extreme danger of infection and assess infection behavior in Fars province in Iran. For this purpose, a GIS-Based machine learning algorithm was employed to measure the risk of COVID-19 pandemic.

Jia et al. [28] presented a dynamic model to diagnose the transmission of COVID-19 considering the relationship between virus pandemic and air quality conditions. The results showed that the air index is the most important climatic factor in virus diffusion.

Aral et al. [29] find the Spatio-temporal pattern of the COVID-19 pandemic in Turkey. They also employ spatial regression to uncover the related factors affecting the COVID-19 cases. In the COVID-19 model, they discovered that population density and the elderly dependency ratio are necessary. Aldana et al. [30] investigated the spatial distribution of COVID-19 cases in Iran. They identified important spatial clusters of cases and the effect of socio-economic features of Iranian provinces on the number of cases. They found a spatial correlation within Iranian provinces in terms of COVID-19.

GIS due to having different types of spatial statistical analysis can be very important in the production of time and space maps of COVID-19 pandemic in Iran. In particular, the development of GIS methods, especially in space spatial and general image of the primary centers of the pandemic of the virus in different areas, is possible [31].

The main purpose of this paper is (1) Evaluation of Spatio-temporal patterns of COVID-19 pandemic using confirmed data, deaths, recoveries and population information from March 2, 2019 to November 31, 2021. (2) Quantification of changes in COVID-19 and (3) Comparison of annual changes of COVID-19 in the provinces of Iran. This activity will stop the pandemic of the virus at different times and regions of its transmission chain, which in itself can help managers and policymakers to be able to adopt new strategies to mitigate this crisis in Iran.

Spatial statistics have emerged as a useful tool for the analysis of spatial pandemic, concerning mapping and statistical analyses of spatial and Spatio-temporal incidences of different pathogens. The value of this paper is to perform spatial analyses which allow us to better understand the COVID-19 pandemic in Iran. Therefore, the result is expected to help the Iran government create more appropriate policies and strategies to reduce the virus pandemic.

2 Methods

The present paper examines Iran as the second-largest country in the Middle East (Fig. 1), the area of this country is 1873959 \({\mathrm{Km}}^{2}\) and it has an arid and semi-arid climate [32]. Iran has 31 provinces due to its political characteristics according to several studies, Iran has suffered the most harm from the virus in the world and ranks fifth in terms of vulnerability [33]. Iran, with a collection of 5,973,457, confirmed cases and 127,053 deaths, is the eighth location in the world in terms of the number of confirmed cases of COVID-19 in this paper.

Fig. 1
figure 1

Location of study area. a The global location of Iran, b Provinces of Iran

This paper is based on secondary data, the data related to confirmed, deaths, recovered, hospital beds and hospitals from the website of the Ministry of Health of Iran, and the data related to population statistics received from the website of the Statistical Center of Iran from March 2, 2019, until the end of November 2021. the data were classified in the software Excel in six indicators for 31 selected provinces of Iran (Table 1).

Table 1 Data was used for spatial analysis in terms of provinces of Iran to model the COVID-19 pandemic [34]

Then, analysis of the Spatio-temporal distribution of four confirmed indices, deaths and recovered and population from the GWR model and Getis-OrdGi * (G-I-star) statistical model ArcGIS10.3 environments in the form of two different models were used. To implement the GWR model, three indicators of confirmed, deaths, and recovered and population, number of hospital beds and hospitals in Excel environment were added to the shapefile of the study area. From the modeling spatial relationships section of ArcGIS10.3 software, the GWR statistical model was called for.

The input of this model was selected as the shapefile of the study area, to which 6 data were added, then three indicators of confirmed, deaths, and recovered as dependent variables and 3 indicators of the number of hospital beds, hospitals, and population as independent variables, were introduced to the model. Hot and Cold spots were determined using the Getis-OrdGi * (G-I-Star) statistical model in the GIS environment to quantify the Z-score and P-values. Finally, the inverse distance weighting (IDW) method was implemented to produce a spatial layer that depicts hot spots and cold regions using the Z-score as the input data. In this paper, Getis-OrdGi * geostatistical model was used for spatial clustering, and Spatial Autocorrelation Morans I analysis for spatial autocorrelation between COVID-19 indices in the provinces of Iran.

2.1 Geographically weighted regression (GWR model)

Regression analysis is generally used to examine the relationship between parameters. When it is more focused on the relationship between an affiliate parameter and one or more independent parameters, it is used for modeling [35]. There are two types of regression models: global multivariate regression and local multivariate regression models. Global multivariate regression model (OLS, spatial error and spatial delay). These models can clarify the importance of statistical relationships between the independent and affiliate variables by an equation. The GWR model reveals the location of the regression parameters. This model examines the parameters that are points one by one by the local weight ordinary least squares (OLS) [36]. The relationship of the model in question is analogous to the global regression models; however, the variables are different in different places [36]:

$${y}_{i}=\beta 0+\upbeta 1{\mathrm{ix}}_{1\mathrm{i}}+\upbeta 2{\mathrm{ix}}_{2\mathrm{i}}+\dots +\mathrm{ \beta knXkn}+\mathrm{ \delta i}$$
(1)

I where β is the estimated parameter vector, x is the independent variables, and y is the observed values' vectors.

2.2 Hot and cold spots model (Getis-OrdGi *)

In this model, the regions with the highest rates of incidence were identified as hot spot areas [37]. The hot spot analysis could be beneficial in studying the evidence of identifiable spatial patterns. This technique, implemented in ArcGIS software, pinpoints statistically meaningful spatial bunches of higher value (hot spots) and lower value (cold spots). In Getis-Ord Gi*, the significance and intensity of clustering are assessed based on an assurance surface and Z-scores. For positive Z-scores, higher Z-scores reveal many intense clusters (hot spots). However, for minus Z-scores, low Z-score values show more severity of bunches of lower values (cold spots) [38]. The Getis-Ord-J. K statistic is calculated by Eqs. 2, 3, and 4.

$${G}_{i}^{*}=\frac{\sum_{j=1}^{n}{w}_{ij}{x}_{j}-\overline{X}{w }_{ij}}{\sqrt[S]{\frac{[n\sum_{j=1}^{n}{w}_{ij}^{2}-\left(\sum_{j=1}^{n}{w}_{ij}\right)]}{n-1}}}$$
(2)

In this formula Xj, the value of the attribute for complication j, Wij represents the spatial weight between the complication and i, j and n the collected number of complications.

$$\overline{X }=\frac{\sum_{j=1}^{n}{x}_{j}}{n}$$
(3)
$$S=\sqrt{\frac{\sum_{j=1}^{n}{x}_{j}^{2}}{n}}-(\overline{X }{)}^{2}$$
(4)

Since Gi itself is a type of Z-score, it does not need to be recalculated.

2.3 Spatial autocorrelation (Global Moran’s I)

Spatial autocorrelation (Global Moran’s I) analysis reveals the correlation between the same values and variables in different locations, strong spatial autocorrelation occurs when the values are randomly distributed in space with no relationship between them [39]. Spatial correlation analysis is calculated as spatial autocorrelation (Global Moran’s I) by Eq. (5).

$$I=\frac{\mathrm{N}}{\mathrm{W}}\frac{{\sum }_{i}{\sum }_{j}wij ({x}_{i}-\overline{x }) ({x}_{j}-\overline{x })}{{{\sum }_{i}({x}_{i}-\overline{x })}^{2}}$$
(5)

where Xi is the coefficient of the distance or relative variable in location unit i, n is the number of location units, and w (i, j) is the connectivity spatial weight between j and I.

3 Results and discussion

3.1 Analysis of the Spatio-temporal of COVID-19 pandemic with the GWR model

Figure 2 shows the COVID-19 pandemic in Iran in the period from March 2, 2019, to the end of November 2021. Using the GWR model, each indicator separately on an annual basis indicates that on March 24, 2019, patients with COVID-19 are observed only in Gilan and Golestan provinces. However, in the provinces of Tehran and Khorasan Razavi, the rate of this index in 2020 had the highest pandemic. Again Gilan province was reported as the most critical province with the highest pandemic of the virus in 2021. The reason for the higher pandemic of this virus in urban and suburban traffic had been the non-compliance with health protocols in neighboring provinces such as Tehran, Qom, and central Ardabil to this province. The number of confirmed cases of this virus in Iran was 28 cases, 86.67% of which were reported in Qom province.

Fig. 2
figure 2

Spatial distribution of COVID-19 pandemic. a confirmed 2019, b confirmed 2020 c confirmed 2021, d death 2019, e death 2020, f death 2021, g recovered 2019, h recovered 2020, i recovered 2021

Since March 1, 2019, of the COVID-19 pandemic in the country has been rising sharply, reaching over 1000 infected people in the country. Studies show that COVID-19 from the provinces of Tehran, Qom, Gilan, Markazi, Mazandaran, and Isfahan in the surrounding areas is expanding quickly. From March 2, 2019, until the end of November 2021, the number of confirmed infected people with COVID-19 in Iran reached 5,117,766 people showing an increase in the number of infected people in the country. traveling to pilgrimage and tourist cities such as Mashhad, Qom, Mazandaran, Golestan, and Gilan, without following the health instructions required by the headquarters in fighting the COVID-19 caused the rapid pandemic of the virus to most parts of the country.

From March 2, 2019, to the end of November 2021, the number of deaths in the country reached 133,880, and the most deaths in 2019 were reported in Lorestan and Markazi provinces. However, in 2020, due to weddings and the lack of hygiene items and crowds, and the lack of virus detection kits caused the transmission of the COVID-19 chain to the provinces of West Azerbaijan and Yazd and a major number of people died. In addition to the high number of patients, Gilan province had many deaths in 2021, the most important reason for the high number of patients and deaths in this province was the excessive travel of people to this province on weekends (Fig. 2). With the observance of health protocols, home quarantine, and closure of pilgrimage and recreation centers by the COVID-19 headquarters, the number of recovered people in the whole country reached 10,317,609 people.

Of these, Markazi province in 2019 with 26.7% has the highest approval in the country compared to other provinces of Iran. Meanwhile, in 2020, West Azerbaijan province with 17.18% and Isfahan province in 2021 with 7.9% have recovered the incidence of COVID-19. The results of the COVID-19 pandemic for the three confirmed case, deaths and recovery are presented in (Tables 2, 3 and 4). Using geographic weighted regression as a visual technique can reveal interesting patterns in geographic data. Spatial distribution of confirmed, deaths and recovered patients in relation to the number of hospitals (dependent variable), the number of beds in each hospital and the population of each province (explanatory variable) shows that there is a strong relationship in some provinces between the variable There is a relationship and explanation, and in some provinces this relationship is negative. One of the important parameters is the coefficient of determination (\({R}^{2}\)) which expresses the goodness and accuracy of the model, and the closer this parameter is to 1, it means that the explanatory variable used has been able to explain the changes in the dependent variable well, in This research on the spatial distribution of COVID-19 is very well estimated and shows that the model with the dependent variable was able to explain the pandemic of COVID-19 spatial distribution in 31 provinces of Iran for confirmed 81%, deaths 58% and recovered 67% (Tables 2, 3 and 4).

Table 2 Results of the confirmed cases of the COVID-19 pandemic relative to population, hospital bed and hospitals
Table 3 Results of the deaths of the COVID-19 pandemic in relative to the population
Table 4 Results of the recovered of the COVID-19 pandemic in relative to the population, hospital bed and hospitals

Past research has shown that a large number of socio-economic and environmental parameters are used to analyze the of COVID-19 and the number of deaths by means of using GIS tools [40]. Also at the provincial level in Italy, Giuliani et al. [41] modeled the Spatio-temporal dynamics of contagion and mortality due to COVID-19. As in Italy, the local pandemic is many heterogeneous. Its major focus is in the north, but it is gradually penetrating the southern provinces.

There is strong proof that hard control measures implemented in some provinces effectively break the cycle of infection and limit the pandemic to adjacent regions. Spatial autocorrelation analysis based on distance shows that at a distance of 383.8 km from the provinces of Qom, Tehran, Razavi and Gilan, the spatial distribution of infected and deceased is positive; However, at a distance of 762.6 km from the provinces of Qom, Tehran, Khorasan Razavi and Gilan, the spatial distribution of recoveries is positive, indicating that from this distance, the number of confirmed and deaths due to COVID-19 has decreased and the number of recoveries has increased.

3.2 Spatial distribution pattern of three indicators of confirmed, the deaths and the recovery in Iran

Table 5 shows the provinces of Ilam, Golestan, Gilan Isfahan, Tehran, and West Azerbaijan in terms of spatial distribution results of Moran correlation from March 2, 2019, to the end of November 2021. The Z-score is 3.357. However, this is 0.478 and 4.938 deaths rate due to the COVID-19 pandemic in Yazd, West Azerbaijan, Tehran, Isfahan, and Qazvin provinces, respectively.

Table 5 Estimating the geographic pattern of the COVID-19 pandemic in Iran using spatial autocorrelation Morans I

While the Moran value for the recovery index was 0.550 and the Z-score was 3.309 in the central provinces of West Azerbaijan and Isfahan. The results of Moran's spatial correlation showed that the Z-score for these provinces was positive in terms of both confirmed and deceased indicators and was statistically significant because in these provinces the number of confirmed cases and deaths of COVID-19 has been too much.

3.3 The spatial pandemic of COVID-19 using the hot and cold spot model

Figure 3 show the COVID-19 pandemic in the country using the hot and cold spot model for March 2, 2019 to the end of November 2021 for the indicators of infection, deaths, recovered relative to hospital, hospital beds and population. As it can be seen in 2019, COVID-19 patients are found only in the provinces of Tehran, the northern parts of Mazandaran and Qom.

Fig. 3
figure 3

Spatial distribution of COVID-19 pandemic with Getis-Ord Gi model a confirmed 2019, b confirmed 2020, c confirmed 2021, d death 2019, e deaths 2020, f deaths 2021, g recovered 2019, h recovered 2020, i recovered 2021

However, in 2020, there is a fast COVID-19 pandemic in the provinces of Tehran, East Azerbaijan and West Azerbaijan. The non-observance of health protocols caused many people in these provinces to become infected with COVID-19.

The announced conditions showed their results with a 99% confidence level in 2021 in the provinces of Tehran and Khorasan Razavi. Studies have shown that the spatial spread of COVID-19 pandemic from the provinces of Tehran, Qom, West Azerbaijan, East Azerbaijan, Isfahan and Mazandaran to the surrounding areas was growing rapidly.

With the rapid of COVID-19 in the provinces from March 2, 2019 to the end of November 2021, a major number of people died due to a lack of medical kits, hospital facilities and disregard for health protocols the provinces of Tehran and Isfahan, Khuzestan and Ilam had hot spots compared to other provinces in terms of deaths, with a 95% confidence level (Fig. 3).

The provinces of East Azerbaijan, Khorasan Razavi, Tehran and Fars had 95% of the recovery areas with internal points. However, the provinces of North Khorasan, Golestan, Kurdistan, Yazd, Isfahan, Qazvin and Zanjan as cold provinces with a 90% confidence level, in terms of cold and hot spots, didn’t show significant results in terms of recovery index (Fig. 3).

3.4 Assessment I anselin local moran

Anselin Local Moran I analysis was used to obtain stronger results in terms of hot spots and cold spots. Due to a series of weight characteristics, this method determines statistically considerable hot spots, cold spots, and spatial outlets. As shown in (Table 6), from March 2, 2019, to November 31, 2021, the confirmed index, deaths and recovered of the COVID-19 were found only in one high cluster in central Iran, including Tehran, Isfahan, Fars, and East Azerbaijan.

Table 6 Estimation of geographical pattern of COVID-19 pandemic in Iran using Getis-Ord Gi *

This very high cluster shows that the risk of COVID-19 infection in these four provinces is much higher compared to other provinces of Iran, and the deaths and recovery rates are high compared to other provinces of Iran. Brasil [42], analyzed the spread and spatial pandemic of COVID-19 and identified the occurrence of clusters of COVID-19 in northeastern Brazilian cities, the results of his research show that COVID-19 is a dangerous crisis for health [43, 44]. Many studies have examined the spatial dynamics of the virus, however they have used a combination of spatial and spatial–temporal time-cycle methods on a small number of cases to analyze the COVID-19 pandemic.

Overall, the result of our research shows the exponential growth of mortality in the central regions of Iran, and the quick dispersion of this virus compared to other regions. Other studies have also been performed about two widespread diseases and COVID-19, and the bulk showed an affirmative relationship between patients with COVID-19 and other diseases for example, blood pressure, diabetes and cardiovascular diseases [45, 46]. Tabarej and Minz [47] spatial–temporal variation pattern in the epicenter footprint: a case study of confirmed, recovered and deceased cases of COVID-19 in India. The study found a sudden change in the hotspot region and a similar shift in the footprint from August. The changing pattern of the hotspot’s footprint will show that October is the riskier month for the first wave of COVID-19. Monte Carlo simulation with 999 simulations is taken to find the statistical significance. So, for the 99% significance level, the p-value is taken as 0.001.

3.5 Effect of population density on the COVID-19 pandemic

The geographical analysis of the COVID-19 pandemic in Iran shows that the spread of COVID-19 in the country is in the surrounding areas of Tehran province, north, west, and northwest of the country, and the highest spatial COVID-19 pandemic in Tehran provinces. Alborz, Ardabil, Bushehr, Ilam, and Khorasan Razavi due to high population density. Spatial and geographical studies show that the most significant factor in the COVID-19 pandemic in Iran is population density and spatial location, which has accelerated the COVID-19. In other words, up to a radius of 382 km from Tehran provinces, the COVID-19 due to provincial travel, non-compliance with health protocols, overcrowding in sales centers and population displacement due to capital accumulation and development of service and industrial sectors caused the dispersion. The virus increased, but from 763 km due to increasing focal length and decreasing population density, the pandemic of the virus gradually in 2021 showed a somewhat decreasing trend (Fig. 4). The results show that most of the confirmed cases are from provinces with high population density such as Tehran as the capital of Iran and Khorasan Razavi as a pilgrimage city though with medical facilities, many hospital beds. At the end of November 2021, the COVID-19 headquarters in these provinces were able to decrease the pandemic of COVID-19 as much as possible (Fig. 5). In the meantime, studies have been done on the origin, transmission and epidemiological characteristics of COVID-19 due to high population density [48, 49].

Fig. 4
figure 4

Population density in the provinces of Iran

Fig. 5
figure 5

Relationship between population density and the COVID-19 pandemic at the level of the provinces of Iran

Isazadeh and Argany [50] showed that the intensity of the desease became more in the monthly COVID-19 pandemic in Qom province in an 8-month perusal era. In particular, from February to April for a two-month reduction, the number of cases increased anew in the next months. It seems that observance of health protocols including social distancing, regular hand washing, use of masks and gloves, lazing at home by people and closing of pilgrimage and tourist centers could reduce the disease in May and June. In China's Hubei province, a study was conducted, the results of this study showed that space clusters follow an upward pattern and undergo a sudden change, and due to high population density, the rapid pandemic of the COVID-19 was observed in this province [51].

4 Conclusion

The pandemic of COVID-19 has become a clinical threat to communities around the world. Today, the study of the geographical distribution of infectious viruses in the world is very important in discovering the cause and conditions of the pandemic of viruses in each region, so the WHO began using geographic information technology in the spatial analysis of viruses in 1993. In this paper, we modeled the spatial of COVID-19 pandemic during the period March 2, 2019, to the end of November 2021 using GWR, Getis Ord Gi * models (hot and cold spot analysis) and Moran's spatial self-solidarity in Iran we also used Moran's spatial correlation to deduce the 6 indicators of confirmed cases, deaths, recoveries, population, hospital beds, and hospital. Using Getis Ord Gi * (hot and cold spot) and GWR, we were able to identify which provinces were most likely to show confirmed, deaths and recovered events. Therefore, the link between COVID-19 hot and cold spots and population may be useful in future studies to investigate spatial insulin Moran correlation. The highest annual space distribution using the GWR model for approvals was seen from March 2, 2019, to the end of November 2020 in Ilam and Golestan provinces reaching over 5,973,457 confirmed in the country. With the continuation of this upward trend in the country, the spatial COVID-19 in its surrounding areas in densely populated provinces such as Tehran, Isfahan, Gilan, West and East, and Central Azerbaijan has been rapidly increasing by over 60%. of the population becoming infected with the virus. Not following the health protocols, pilgrimage trips and tourism to the neighboring provinces caused many citizens to lose their lives. Moran's spatial correlation in these provinces has been positive and statistically significant, and follows the cluster pattern. Following the health protocols of the COVID-19 headquarters and observing social distance and advertised warnings on social media, and tightening traffic caused the number of deaths to decrease by the end of November 2021. The spatial analysis models (hot and cold spot) show results different from the geographically weighted regression model for 31 provinces of Iran, in terms of the confirmed index of deaths and recoveries relative to the population, hospital, and hospital bed. In the central and southern parts of the provinces of Iran, including the provinces of Alborz, Tehran, Isfahan, Tabriz, Yazd and Fars, 99% confidence levels are hot spots. Nevertheless, the central provinces showed to be very hot spots with a 95% confidence level in terms of the index of deaths relative to the population compared to other places. In terms of index of deaths and recoveries as the safest provinces, the provinces of North Khorasan, Golestan, Kurdistan, Qazvin, and Zanjan showed cold spots with a level of 90% confidence, and were scheduled the least dangerous areas among the 31 provinces of Iran on March 2, 2019, to November 31, 2021. The results of Moran's spatial autocorrelation analysis of hot spots of COVID-19 for patients in 2019 showed that the Z-score was negative, indicating that the number of patients across the country has been declining on a random pattern. The Spatio-temporal hazard presented in this paper shows that the Spatio-temporal hazard model, based on the annual release of COVID-19, offers a good grasp of the changes caused by the virus throughout the country.