Harnessing satellite data for pandemic response: Lessons from COVID-19

GRANADA, Spain — Humanitarians have long used satellite data as a tool to help coordinate responses to myriad natural disasters — from flooding and hurricanes to disease outbreaks such as Ebola. As countries began going into lockdown earlier this year, it made sense that satellite data might help governments in a response to the COVID-19 pandemic.

By the end of March, private data companies such as Maxar and open-source data cooperatives like MapAction reported a proliferation of satellite image requests from all over the world.

“Satellite images give you access to data without having to break social distancing restrictions,” said Rhiannan Price, the director of sustainable development practices at Maxar. “I think a lot of people saw potential in them.”

One such project aimed to assess the health care needs — such as routine vaccinations and follow up care — of mothers and children under the age of 5 in rural Peru during quarantine. Another project in South Africa aimed to use satellite images as a tool to help to identify potential relocation sites for people living in high-density urban slums who were considered particularly at-risk for contracting the coronavirus.

The successes and shortcomings of these projects can provide useful lessons for stakeholders hoping to harness satellite data for development goals.

The challenge of adapting data to address an unprecedented crisis

While satellite imagery is widely used in humanitarian settings for disaster response, it was unclear exactly how the data might best be leveraged to address the many fallouts from the pandemic, according to Pete Ward, director at the GAL Center, an education group based in Cuzco, Peru.

So organizations experimented. In the beginning, many of the pilot projects followed a certain strategic logic, he said.

“We started off by trying to identify which areas in Cuzco were the most at-risk for infection by overlaying census data over base maps, and also combined data sets that included indicators like population density, access to clean water and age of residents,” he said.

However, by the time it completed the project two weeks later and presented the results to government partners, the Cuzco COVID-19 response had found an alternative to support at-risk populations.

“The pandemic response moved so quickly, it was difficult to keep up,” he said.

Trial and error

The organization decided to try something else. For the next modified version of the project, government partners requested help identifying children under 5 living in rural areas who might be at risk of missing out on non-coronavirus-related health care during quarantine — for example vaccinations or treatment for anaemia.

To comply with the request, GAL worked with open-source mapping collaborative Humanitarian OpenStreetMap Team to use satellite images from Maxar and Microsoft Bing of the city and surrounding regions. Next, they sent out a request to Humanitarian OpenStreetMap Team volunteers to map buildings and roads. They then combined these base maps with data from private government health databases with the goal of creating an interactive dashboard and map that would allow health care workers to track children’s health needs at the house level.

This project also came up short, as the databases they planned to use in combination with the base maps were out of date and incomplete, and it was unrealistic to update the database in the midst of a pandemic, he said.

While neither of these projects led to actionable data, Ward said that lessons can be drawn from them.

“Your base maps are only as good as the data you are trying to overlay,” he said. “If it isn’t complete, it probably won’t end up working.”

It also became clear to Ward and other partners that making satellite data useful during an unprecedented crisis would take time, trial, and error, he said.

A supporting role

Ultimately, GAL did contribute to a successful COVID-response project: supporting the distribution of oxygen tanks to patients seriously affected by the virus. An oxygen factory nearby told the local government that it could provide oxygen to COVID-19 patients at home for free if the government could manage the logistics.

As a part of a multisector partnership, a local organization helped move, fill, and distribute tanks, while GAL used satellite images and open-source tools to create a database that mapped the locations of the homes of COVID-19 patients who required oxygen.

This geodatabase also helped paint a broader picture of coronavirus hotspots — including the infection rate and spread — that stakeholders hope can help government partners better plan for future responses.

“By collecting data on where the people who are most affected by COVID live — and understanding their support network — we are creating a valuable database that shows us where oxygen is needed, and where the virus is spreading most,” Ward said.

Satellite images for relocation planning under quarantine

In South Africa, quarantine and social distancing provisions have provided a catalyst leading to the transformation of disaster management and resettlement planning — with a geographic information system playing a pivotal role, according to Louis Roodt, managing director at South African urban and regional planning organization Promoveo Consult.

The Buffalo City government was concerned that dense informal settlements would become COVID-19 hotspots. Prior to the pandemic, it had already planned to build new public housing options. However, the outbreak of coronavirus made the construction of these new communities even more urgent, according to Roodt.

“With people living in such proximity, social distancing is impossible,” he said. “We realized that we needed to de-densify these settlements in the shortest possible period of time.”

This presented several challenges. Typically, the preplanning stage would take six months or more, requiring numerous field surveys, first to the existing settlements to count and understand the needs of the families that needed to be relocated, and then to buildings sites to assess their suitability and connectivity to existing infrastructure and other services.

“Usually, we’d do all of this in person,” Roodt said. “But we were under a hard lockdown. I couldn’t even leave my house — so out of necessity, we had to rely on satellite imagery.”

To his surprise, he found that he was able to do much of the field work remotely. By zooming in on vegetation growing in a specific area he was able to estimate if the soil was suitable for building. He could also reasonably assess a potential sites’ accessibility to infrastructure by tracking road connectivity to hospitals and markets, and could even identify electrical cables and heavy water pipes to ensure the potential building site had access to electricity and hot water.

But what surprised him the most was the efficiency of this new system.

“What would have taken six months in the past took us six weeks from start to finish,” he said. “It’s opening up new avenues and methods of doing work which we wouldn’t have thought to do in the past.”

A need for greater accessibility

During the early days of the pandemic, both Roodt and Ward emphasized the need for trial and error before they were successful at arriving at the most useful applications of satellite technology to meet their specific needs.

However, Roodt said that this sort of experimentation is best done before the disaster takes place.

Right now, first responders are able to access satellite images — often for free — when a disaster hits, thanks to programs such as Maxar’s open-data wing and the International Charter Space and Major Disasters.

Roodt would like to see such affordable, up-to-date satellite images made more easily available to practitioners working in different aspects of development — from urban planners, like himself, to health ministries — particularly in low-resource countries that are often unable to afford them.

“Satellite images have tremendous potential for all sorts of applications, but right now they are underutilized,” Roodt said. “And that’s because the cost is prohibitive. If this technology was more accessible, who knows what sorts of solutions we’d be able to find.”

Visit the Data for Development series for more coverage on practical ways that satellite data can be harnessed to support the work of development professionals and aid workers. You can join the conversation using the hashtag #DataForDev.