Risk assessment to curb COVID-19 contagion: A preliminary study using remote sensing and GIS

Shruti Kanga Suresh Gyan Vihar University, Jaipur https://orcid.org/0000-0003-0275-5493 Gowhar Meraj Department of Ecology, Environment and Remote Sensing, Government of Jammu and Kashmir https://orcid.org/0000-0003-2913-9199 Sudhanshu Suresh Gyan Vihar University, Jaipur Majid Farooq Department of Ecology, Environment and Remote Sensing, Government of Jammu and Kashmir https://orcid.org/0000-0002-7813-3474 M. S. Nathawat Indira Gandhi National Open University, New Delhi https://orcid.org/0000-0003-1516-3112 Suraj Kumar Singh (  suraj.kumar@mygyanvihar.com ) Suresh Gyan Vihar University, Jaipur https://orcid.org/0000-0002-9420-2804


Introduction
or Coronavirus disease 2019 has been linked with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and is currently responsible for creating havoc to the whole human civilization 1 . COVID-19 risk mitigation requires considerable planning, if the pandemic has to be eliminated from every area, regardless of developed or developing countries 2 . Highly populated countries of Asia, such as India, Bangladesh, and Pakistan, are particularly more vulnerable to this disease because of the high population densities, weak health care system, and poverty 3 . To control this pandemic, the only established solution till now, has been the isolation and quarantine of the infected persons, through the government imposed national level lockdowns 4 . On the other side, the extended lockdowns for restricting its spread in these countries has posed a signi cant challenge for the governments at different levels, as it has created a state of economic crisis 5 . In India, state governments are planning for riskinformed lockdowns by identifying the areas at varying degrees of risk, using the parameters that have been to date, known to govern the contagiousness of the COVID-19 6 . Furthermore, it is imperative to understand the risk associated with this pandemic spatially, to help the agencies dealing with curbing its spread, for risk-informed decision making 7 .
There are various studies on risk assessments based on vulnerability and exposure of communities to various natural disasters. However, the case of pandemic COVID-19 is different, as the exposure parameter is an infected person, itself, or a thing that has been in his contact 8 . Moreover, there are speci c hypothesized parameters such as BCG vaccinated people and women being immune to the lethality of this virus, which de nes the resilience of the community towards the contagion of this disease. However, such parameters are still under trials 9 . Hence, if the COVID-19 risk assessment has to be performed, hazard and vulnerability parameters need to be de ned 10 . Providing sustainable threat protection from the COVID-19 requires approaches that are proven in resisting the spread of this pandemic particularly when it is fact that it has no drug or vaccine till date 11 . An issue that shall increase the likelihood of the recurrence of this pandemic, unless it is available, as such, the need for a comprehensive risk assessment becomes obligatory.
The framework for the risk assessment of COVID-19 must take all the readily available information and use it for the spatial identi cation of the areas that are currently at high risk. Addressing the issue of the implementation of restricting the virus spread, requires de ning the risk zones according to the nest scale of administrative coverage. COVID-19 risk assessment and mapping (labeled hereafter as CRAM) framework, as proposed here, involves the mapping of the risk, using its de ned hazard and vulnerability components, for informed decision making to declare red, orange, blue, and green zones of containment [12][13] . We used proximity to hotspots and settlements as the hazard components because the pandemic spread is directly proportional to the distance from the hotspot and density of settlements. In contrast, different socio-economic parameters were used as vulnerability components in the assessment In collaboration with the Rajasthan state authorities, the area under Jaipur municipal corporation (JMC) was chosen as the study area for COVID-19 risk assessment, mainly because the city has been witnessing an increase in the number of COVID-19 cases, ever since, it had hit India (Fig. 1) 14 . Moreover, due to the high population density of this area, it has become imperative for the authorities to manage lockdowns without affecting the economy of the state and shall only be achieved through a spatial risk assessment of the COVID-19 threat. As an example of the whole country, the extended lockdown of 40 days, i.e., from 24 March till 03 May in India, has caused an economic loss of approximately $26 billion per week 15 .
Till now, COVID-19 risk mapping has not been reported anywhere in the world. Probably, similar strategies are implemented across the globe, but the formal reporting in any scienti c journal is scarce. Hence, to the best of our knowledge, we believe this study is rst of its kind that has used spatial sciences, remote sensing, and GIS for risk assessment of the COVID-19. To provide scienti c support, we have already submitted the results of our work in Jaipur, to the Government of Rajasthan in facilitating the COVID-19 risk mitigation plans for the whole state, which is signi cantly contributing towards COVID-19 riskinformed planning and management of the area.

Methods
The methodology of COVID-19 risk assessment and mapping (CRAM), for simplicity, is divided into three steps. The rst step involved the generation of GIS layers of various administrative data (ward), hazard data, socio-economic data, and biophysical data. The second step involved the integration of hazard and vulnerability to generate risk assessment. The third and the nal step in CRAM, involved risk mapping for informed decision making and the prioritization of COVID-19 risk areas using ward-level administrative boundaries of the Jaipur city, for prompt action. Individual steps are further brie y discussed below. Fig. 2 demonstrates the complete CRAM framework.
As per the current understanding of the COVID-19, several parameters have been identi ed to affect its lethality and infection. These include various hazard, biophysical and socio-economic parameters, determining the actual risk of an area to COVID-19 disease 16 . Hence, we calculated the risk through integration of hazard and vulnerability to this pandemic, of each zone of concern (i) -wards in our situation. We de ned COVID-19 risk (written-off as, C19R index) as Accordingly, an increase in the hazard and vulnerability will increase the COVID-19 risk of the zone of interest. Table 1 de nes the hazard and vulnerability components of C19R and is discussed in more detail in the following sections.

Data sets
We used census 2011 data of Jaipur city and related speci c socio-economic parameters to COVID-19 vulnerability, such as population, population density, percentage of main workers, and percentages of literates. Groundwater well data used in this study was provided by the central groundwater Commission, Government of India. Further, hotspot locations provided by Jaipur municipal and health authorities, were used as input data for generating various hazard zones, using GIS-based proximity analysis. Preprocessed World-view satellite imagery of September, 21 2019, was used for land-use/ landcover (LULC) mapping to be used as a hazard component.

Hazard
We hypothesized hazard in our study as the potential of COVID-19 posing danger of exposure, to the population of Jaipur. We used proximity to hotspots, i.e., locations with a high density of con rmed positive COVID-19 cases, as the hazard and de ned four levels of hazard zones in consultation with the municipal and health authorities of Jaipur city; Red zone (0-350 m radius), orange zone (350m-700m), blue zone (700m-1050m) and green zone (1050-1400m) (Fig. 3a). Besides, LULC of the wards was used as a hazard parameter as certain LULC types associate with the high probability of COVID-19 infections, such as settlements and agriculture, where people gather and get exposed to the virus. Hence using, ArcMap 10.1, we conducted level-II classi cation of the study area using visual image interpretation technique on the Worldview satellite imagery [17][18][19][20][21][22] (Fig. 3b).

Vulnerability
Vulnerability, in our case, refers to the susceptibility of the area to COVID-19 infection due to its demographic, economic, and availability of clean water for sanitation conditions. As for as the current understanding of this pandemic is concerned, densely populated areas, people that do not have access to clean water for frequent sanitation and people that need to leave their houses for livelihood, are particularly more vulnerable. That is the reason we chose population, population density, well-density (as water supply data was not available), and percent workers, respectively, as the parameters that make an area more vulnerable to the COVID-19 infection. Further, we hypothesized that educated people being more aware of this disease would take requisite precautions such as hygiene and isolation to protect them from getting infected. Hence, we took percent literates also as a parameter in the vulnerability component ( Fig. 4 a-d).
Moreover, we digitized all roads of the Jaipur city, to provide routing information to the agencies dealing with this pandemic in case of emergencies (Fig. 1c). Also, using kriging spatial interpolation technique, we generated well density layer, using well location data in ArcMap 10.1 (Fig. 4e). Finally, the integration of all the layers is carried out using a GIS-based weighted overlay analysis 12 . The weights for different classes are shown in Table 2 and were developed using the current knowledge about these parameters in governing the infection and spread of COVID-19 pandemic 16 .  Fig. 5 a, b and c, respectively. We have categorized them into ve classes each, red (very high) followed by orange (high), blue (moderate), green (low), and pink (very low). The nal risk assessment being the integration of hazard and vulnerability components, is shown in Fig. 5 (c) and is being discussed hereunder. The results indicated that out of the total area of JMC (379 km 2 ), 6.13 km 2 (6.13%) fall in red risk zone followed by 60.38 km 2 (15.91%) in orange, 139.63 km 2 (36.79%) in blue, 164.51 km 2 (43.34%) in green and 8.9 km 2 (2.34%) in pink risk zone. The risk assessment results indicate that majority of areas under high-risk zones (red and orange) concentrate along the north-eastern and south-western zones of the study area with some scattered zones of red and orange in eastern and southeastern zones along the borders of the JMC. As a result, the risk of all the north-eastern and southwestern zone wards of the study area is higher for all the risk categories. As a result, the population in these wards is particularly under a more signi cant threat of the COVID-19 infection.

Results And Discussion
Promoting risk-informed COVID-19 management of JMC study area The results from the CRAM framework depict signi cant spatial variation, indicating a higher risk for the wards of the north-eastern zone as compared to the wards of other zones, as shown in Fig. 5 (c) referring to the higher number of the hotspots of COVID-19 in this zone of the JMC. We propose that globally, the risks of COVID-19 (C19R) infection and spread can only be managed using risk-informed planning until a cure or vaccine is available, so that the economy and livelihood of the people and countries as a whole, do not suffer. By analyzing the spatial distribution of the C19R to prioritize the lowest levels of administrative boundaries (wards in our case) for planning risk mitigation approaches and resource distribution, it can be achieved. The CRAM framework aims to support this purpose for the area falling under Jaipur municipal corporation (JMC), because the framework had a spatial component in the under high risk as similarly, more than 80% of their areas fall under orange risk zone. The results also indicate the wards of east, west, and north zones are comparatively at lower risk due to the lower hazard probability as well as lower population density and greater availability of water for sanitation, as depicted in Fig. 5 (a-c). Conversely, the results also indicate that north-east and south-west zone wards are at very high risk for COVID-19 due to higher total population, population density, and comparatively lesser availability of the water for sanitation. This information is vital for risk-informed planning for the eradication of COVID-19 threat in these areas.
We propose that the wards under very high risk and high risk may be considered as containment zones through extended lockdowns until the daily number of con rmed positive cases come to near zero. Moreover, such areas shall need an increase in the rate of testing for COVID-19 infection to identify infected persons for isolation and quarantine. Overall, we propose a detailed management strategy shown in Table 3 for all the areas under different risk zones shown in Fig. 5 (c) 23 .
Currently, the COVID-19 threat is a global pandemic, and countries can combat it in the long run only through a risk-informed planning, as it appears that the disease is going to remain in the category of noncurable diseases for some time 24 . Moreover, in the developing world, the COVID-19 risk informed decision making has to be taken to the panchayat and local mohalla levels to contain virus without affecting the economy. This work shall serve as a baseline methodology in other regions of the country and the world with a similar setup.

Conclusions And Limitations
Human populations have been threatened due to the ongoing COVID-19 pandemic, predominantly those that are living with the settings that make them more prone to higher levels of its risk. To address this, we present an integrated COVID-19 risk assessment and mapping (CRAM) framework, wherein we used hazard and vulnerability parameters linked with the COVID-19 contagion, to identify the areas that are under different levels of risk. COVID-19 risk (C19Ri) indicators have been proposed based on the current knowledge of the disease infection, lethality, and spread 25 . The results of the assessment identi ed areas that are under very high-risk category, e.g., NE and SE zone wards of the JMC, Jaipur India. Prioritization of regions for decisions regarding containment and isolation is viable using this approach. This approach thus provides opportunities for long term COVID-19 risk management so that the economy and livelihood suffer least in the study area.
The authors acknowledge that there are scores of limitations in the proposed CRAM framework at present, mainly due to the incomplete knowledge regarding the operational mechanism of COVID-19 infection. Moreover, the number of risk indices used are less in number, but since the science related to this pandemic is still evolving, once more indices are available, the same can be used in further studies. At the moment, whatever knowledge is possible, must be utilized to eradicate this threat. Due to data unavailability, we could not use water supply data, instead we used well density data, which we know may sometimes not give a clear picture of the sanitation conditions of the area. We recommend to use water supply data if available. Further, if parameters that govern the resilience of the community to COVID-19 lethality are available, we then advise using hazard-vulnerability-resilience based risk assessment framework, which shall make the CRAM more focused and broader in scope.

CONFLICTS OF INTEREST
The authors declare that there is no con ict of interest regarding the publication of this paper.     For BLUE AND GREEN ZONES, all the relaxation is permitted only with the permission to take care of social distancing. If social distancing is being broken at any place, the relaxations will be stopped immediately.