An Unprecedented Response

The GEOINT Community's role in the Ebola relief effort

By Kristin Quinn • 2015 issue 1

When the American Red Cross first arrived in West Africa to assist with the 2014 Ebola epidemic, the organization was faced with an empty slate.

“The original map we were looking at for West Africa was almost blank,” said Dale Kunce, a senior geospatial engineer and GIS team lead for the Red Cross. “You didn’t just not know where the buildings were, you didn’t know where the towns were, where the roads were, what roads connected which towns. If a patient at a treatment center said they were from whatever hamlet, you didn’t know where that was.”

Today, thanks in large part to OpenStreetMap (OSM) data, when a patient names their village or hamlet, health officials and aid workers have the resources to quickly determine not only where the village is located, but the number of homes and general population of the village, as well as which route the patient most likely traveled to reach treatment. This information helps pinpoint which locations are most at risk and therefore in need of education campaigns and contact tracing—following up each day for 21 days (the Ebola incubation period) with people who have been in contact with an Ebola patient.

In OSM, the state of California has more mapping content than the entire African continent, according to Kevin Bullock, a product manager at DigitalGlobe overseeing base map products, and also an OSM contributor and advocate.

“Mapping data outside of the U.S. and Europe is inadequate,” Bullock said. “Mapmaking companies have invested in building out map data in developed countries. The degradation of that data is quite steady as you move away from developed areas. Maps of areas in Africa and Southeast Asia are inadequate and not actionable, especially for humanitarian missions … It’s like giving a UPS driver 1,000 packages to deliver in Manhattan without any addresses or locations and saying ‘good luck.’”

In the wake of the West African Ebola epidemic that emerged in 2014, the geospatial community has addressed these extreme obstacles by deploying satellite imagery, crowdsourced crisis mapping programs, and GIS software and training to get updated maps and data into the hands of local officials, as well as the range of military, health, and humanitarian personnel that have descended upon the region.

Geospatial tools have been applied in the Ebola response effort for many mission sets:

  • For triage and determining where to stand up Ebola treatment units (ETUs) and community care centers (CCCs);
  • To monitor and share the locations of various non-governmental organizations (NGOs) deployed to West Africa;
  • To help government and aid organizations transport supplies and patients efficiently;
  • To map dwellings for door-to-door wellness checks, quarantines, and contact tracing;
  • To understand the disease’s epidemiology—the study of the distribution and determinants of health states or events;
  • To plot the locations of corpses;
  • To understand the human geography and sociocultural issues affecting the spread of Ebola through the region, and;
  • To anticipate where the disease might spread.

 

Not only has the Ebola outbreak been uniquely challenging as a result of the size of the affected area, the disease’s potential to spread globally, and the duration of the crisis, but the geospatial community’s response has also been unprecedented in many ways. The number of volunteers involved, the amount of publicly available data, and the level of collaboration being achieved across federal agencies, national governments, academia, and NGOs have all soared beyond that of more static crises.

LAYING THE FOUNDATION

When the Ebola epidemic first gained global attention in spring 2014, DigitalGlobe combed its 15-year commercial imagery archive to provide images of the affected area and began training its satellites on the region more often.

Satellite imagery and map layers serve as the foundation necessary to enable exquisite analysis in underdeveloped areas according to Todd M. Bacastow, director of insight product management for DigitalGlobe. The imagery provides a quick-start for volunteers to trace roads and buildings then later incorporate human geography data to begin recognizing patterns in and making predictions from the data.

“Unless you have foundational map data you can’t perform the analysis required to answer complex questions,” Bacastow said.

The use of geospatial data in epidemiology is nothing new; the technology has just become more sophisticated. The London Cholera epidemic of 1854 was halted after health officials plotted patient residences on a map and traced the disease to a lone water pump.

“DigitalGlobe has been collecting imagery in West Africa for longer than three years not knowing there was going to be a health crisis,” Bullock said. “We were able to go into our archive and publish actionable, relevant imagery. It’s very powerful when put into the hands of the crowd because entire villages, cities, provinces, and countries can be mapped within hours.”

Although government and military personnel can gain access to DigitalGlobe’s imagery and human geography data of West Africa through the National Geospatial-Intelligence Agency’s (NGA) EnhancedView program, the company’s partnership with software provider Mapbox has helped it provide that same imagery to the volunteer Humanitarian OpenStreetMap Team (HOT), which uses volunteer crisis mappers from around the world to develop crowdsourced maps. A scalable Mapbox platform linked to OSM creates a pipeline to process DigitalGlobe imagery and publish it on the laptops, tablets, and smartphones of the HOT volunteers who trace the images to create updated maps.

“When OpenStreetMap contributors log in they expect access to satellite imagery, almost like we expect electricity to work,” Bullock said. “But it actually takes a lot of time and investment to unlock that for them.”

The Ebola relief effort has been the largest undertaking yet for HOT by all measures (Editor’s note: HOT declined to be interviewed for this article). Since March 2014, more than 2,500 volunteers from around the world have mapped more than 12 million objects in Sierra Leone, Liberia, and Guinea, including around 33,000 place names, more than 100,000 kilometers of roads, and 750,000 buildings. HOT products are helping Doctors Without Borders (Medecins Sans Frontieres), the Red Cross, and other NGOs carry out their Ebola relief missions in West Africa.

Kunce said the Red Cross and other humanitarian organizations wouldn’t be as effective in their missions without the contributions of companies such as DigitalGlobe and Mapbox.

“The Red Cross cannot afford a satellite nor would people want us to buy a satellite because the U.S. citizens have already paid for them,” Kunce said. “[They’ve] already paid for the imagery, so being able to use that for humanitarian purposes is amazing.”

But the reach of OSM has grown beyond NGOs since the crowdsourcing platform gained credibility in the aftermath of the 2010 Haiti earthquake. Michael Wellman, a geospatial analyst with the U.S. Centers for Disease Control and Prevention (CDC), said the agency has made significant use of OSM throughout the Ebola crisis.

However, Wellman said CDC is cautious about the veracity of crowdsourced data and has encountered many discrepancies in the range of data sets built during the Ebola response. For example, his team might use satellite imagery to clear up conflicting data on the location of ETUs or villages with the same name (which he said is quite common in the region). The CDC Situational Awareness Team also relies on personnel in the field to confirm information.

“They’re verifying and validating what OSM contributors are doing so we can provide up-to-date maps for our deployers—people running samples to labs—so they know what’s the route I need to take to get there? Is it passable? What’s the status and quality of the road?” Wellman said.

All CDC personnel in the field take GPS devices with them to send back coordinates of new places that need to be mapped as well as to map the conditions of roads, many of which are actually dirt motorcycle paths. Wellman said one team, encountering some terribly muddy roads, plotted GPS points for every place they got stuck in order to prevent others from taking the same routes. In another example, Wellman received an email on Thanksgiving Day from a team in Liberia desperately seeking help to map the area around a village where residents had fled an Ebola outbreak. After the team sent GPS coordinates, geospatial analysts were able to promptly email them a map.

Kunce described the adoption of OSM by government agencies and the U.S. military, as “overwhelming.” OSM is even the base layer for NGA’s public-facing Ebola response website.

CHANGING THE GAME

The popularity of NGA’s public Ebola data has far exceeded that of its restricted information. As a result, Tim Peplaw, director of NGA’s Readiness Response and Recovery Office, said the Ebola relief effort has presented many lessons learned for the agency.

“Generally when we handle a disaster it’s a relatively small area on the ground, such as in the event of an earthquake or flood,” Peplaw said.

His office is responsible for providing unclassified GEOINT products for humanitarian crises. In response to the Ebola crisis, NGA stood up a public-facing website using Esri’s ArcGIS online, which Peplaw said can handle unpredictable surges in web traffic and constantly refresh data as situations on the ground evolve.

“[ArcGIS Online] caters to multiple audiences,” Peplaw said. “You don’t have to be a well-versed GIS user in order to understand how to turn on and off the different layers … users have the ability to control and build their own common operating picture.”

NGA is also providing an unclassified, but password protected website as well as a classified website for Ebola data. But in a speech NGA Director Robert Cardillo gave at the Geography 2050 conference in New York City Nov. 19, he lauded the success of the public website.

“It’s been an unprecedented experience for NGA, with important lessons that will inform our efforts to provide public-facing geospatial content in the future,” Cardillo said.

As of Feb. 10, NGA’s public Ebola website included 495 data layers, 202 products, 68 applications, and 100 percent of its unrestricted elevation data. Since Oct. 23, the site had been viewed more than one million times.

“It shows there’s a high demand for this type of information among nontraditional customers, including volunteers, health workers, and NGOs,” Cardillo said. … “It’s a frank reminder we need to do all we can to provide our data, knowledge, and services at the lowest classification possible, in order to provide context for our customers.”

As a result, Peplaw’s office is working with NGA general counsel to move as much data down to the public website as possible and to make the data not just accessible, but downloadable.

“This is different than anything we’ve ever done before in the unclassified realm,” Peplaw said.

HUMAN GEOGRAPHY

In addition to satellite imagery, DigitalGlobe has also provided detailed human geography data to government agencies and humanitarian relief organizations, offering a better understanding on potential points of disease transmission and to help identify which populations are most at risk. NGA tapped DigitalGlobe to provide human geography layers of affected West African countries alongside the commercial imagery available on the agency’s public Ebola website. Co-registered imagery and human geography layers enable monitoring of demographics, critical infrastructure, economies, ethnicities, education levels, environment, medical facilities, and local events.

“These equip the global relief community with robust and exploitable foundation GEOINT that clearly takes us beyond the pixel into the next realm of responsiveness,” said Christopher Incardona, DigitalGlobe’s senior director of government programs–NGA.

Much like London’s Cholera outbreak was tied to patterns of behavior among patients who used the infected water pump, sociocultural and demographic information remains essential to enable responsiveness in modern health crises.

NGA works to get this data into the hands of those who need it most through its partnership with the U.S. State Department on the Worldwide Human Geography Data (WWHGD) Work Group. The working group, launched in 2011, builds voluntary partnerships around human geography data and mapping to promote human security.

“One of the main contributions we got from the outside world, even before the public-facing website was established, was due to a terrific response from a data cell in the [WWHGD],” Cardillo said.

Ashley Moran, associate director of the University of Texas at Austin’s Robert Strauss Center for International Security and Law and a member of the WWHGD, said the group has been a valuable network for collaboration. With the help of the working group, the Strauss Center leverages geospatial technology to integrate various data sets related to security in West Africa.

Kevin Hughes, a GIS enthusiast and WWHGD member, expressed frustration that many people still don’t understand how ubiquitous GIS is in developed countries.

“It’s kind of like saying the Internet is [only] for email,” Hughes said. “When you look at GIS now, it’s embedded in every element of modern society.”

Hughes hopes resources being poured into West Africa will emerge as a silver lining to the Ebola epidemic, and said GIS has the potential to help these nations advance in terms of civil order and capacity for critical infrastructure, property boundaries, first responder networks, and more.

Moran said in order to help developing countries in the long term, it’s essential for organizations such as those involved in the WWHGD to focus on building local capacity through training and the augmentation of platforms already in place. The Strauss Center has partnered with iLab Liberia, a nonprofit computer lab providing access to cutting-edge technology and IT assistance to leverage technology for the good of the country. iLab frequently hosts free training and tech events.

“There are local organizations with incredible geospatial tools in place that might just need to be supercharged to respond to this crisis,” Moran said.

This is a message the U.S. Army and NGA have paid close attention to as well.

CAPACITY BUILDING

Barry Miller is an NGA analyst deployed to Liberia with the U.S. Army 101st Airborne Division, which was sent to the country to help build ETUs and CCCs. Miller works with analysts from the Liberian Institute of Statistics-Geo-Information Services (LISGIS), the primary provider of geospatial products to not only the Liberian government, but to the United Nations and all NGOs supporting the Ebola relief effort. Miller helps Liberian analysts couple their local knowledge with unclassified imagery to map areas hardest hit by the epidemic, visualize medical and transportation infrastructure, and determine the best locations for helicopter landing zones and standing up ETUs and CCCs.

Miller and the 101st GEOINT team created a Google Earth Common Operational Picture of the response mission, which is replicated on the All Partners Access Network and is available across the U.S. Department of Defense (DoD) and among organizations without access to traditional DoD networks. They also work with the Liberian Ministry of Health to visualize its data and create overall count, infection rate, and density maps on the number of probably, suspected, or confirmed cases of Ebola throughout Liberia. Miller emphasized that the Liberian public is also reliant upon LISGIS data.

“It’s critical to show status updates of ETUs so the Liberian people know where to go to get help,” Miller said, adding that up-to-date road conditions also help patients determine the best route for traveling to a facility.

Miller is confident GEOINT has helped mitigate the spread of Ebola in West Africa.

“We’ve actually seen a significant number drop over here and with GEOINT we’ve been able to publish and disseminate that information to other organizations so they’re able to see how the numbers have dropped and where,” he said.

Perhaps most importantly, the 101st has provided more than 200 hours of GEOINT mentoring and training to LISGIS that will benefit the country long after the epidemic subsides. This includes embedding an analyst with LISGIS daily, studying LISGIS analysts to provide process improvement suggestions, and weekly, five-hour Saturday classes at LISGIS headquarters.

“The partnership between the 101st Airborne Division and NGA with LISGIS has been very productive and beneficial for both parties,” said Cpl. Christopher Byers, a GEOINT specialist with the 101st. “As LISGIS becomes better at their analysis and production, they will be able to shoulder more of the geospatial requirements related to the Ebola fight.”

The Urban and Regional Information Systems Association’s (URISA) GISCorps, which matches volunteer GIS experts with NGO needs, focuses on capacity building by offering assistance and training to nonprofits that cannot afford to hire GIS professionals.

In December, the Corps was interviewing volunteers for its third Ebola response project with the United Nations’ World Health Organization (WHO), which had requested six volunteers in total: five to be deployed to Ghana, Sierra Leone, Guinea, Liberia, and Geneva, Switzerland; and one to work remotely.

“This project is about getting our technology on the ground for 30 to 45 days because [WHO] doesn’t have anyone who knows GIS in their offices,” said Shoreh Elhami, GISCorps co-founder.

Like NGA, WHO is also working in Esri’s ArcGIS Online platform for its ability to cater to even the most novice GIS users. Elhami said all five deploying volunteers must be well versed in the platform and that Esri offered to provide additional training to volunteers immediately before deployment. The aim is for volunteers to pass this knowledge on to WHO personnel in country.

The World Bank’s Global Facility for Disaster Reduction and Recovery (GFDRR) has stood up an Ebola GeoNode with capacity building in mind, according to Benson Wilder, a geographer and analyst with the U.S. State Department’s Humanitarian Information Unit (HIU), which partnered with GFDRR on the project. The goal of the Ebola GeoNode is to provide open, well-curated spatial data and metadata that can be accessed in an environment enabling quick visualization and analysis.

“There’s a mind toward longer term management of the geospatial data and the ability to transfer the content and the governance to entities who will have an interest in using it and keeping it up to date,” Wilder said. “The best map is going to be made by the people who live there.”

However, Wilder added, in the midst of Ebola it’s been challenging to build up a strong community of local contributors.

But as the epidemic subsides, the data will only become more powerful if it is available in the long term and locals continue to map, according to Kunce.

“They will be able to advocate on their own behalf because they will have data about themselves,” Kunce said.

WHAT’S NEXT?

Although the potential uses of the new geospatial data born out of this crisis are promising, the ongoing epidemic and threat of transmission is still acute. The Ebola epidemic has begun to slowly taper, and the CDC is working vigilantly through the application of modeling to maintain and speed up this course.

“We are taking data of the distance between districts and the cases among districts to look at the likelihood of the virus moving to adjoining districts beyond where is already affected,” said Jacqueline Burkholder, a medical epidemiologist with the CDC’s Situational Awareness Team.

Meanwhile, more ETUs and other medical facilities are being built in the field and will need mapped, DigitalGlobe continues to snap images of the region, NGA continues to update its public-facing website, and volunteer and professional mappers alike update the ever-changing data to reflect the real-time situation on the ground.

Stakeholders in the geospatial community should consider how to apply the significant lessons already learned. How will HOT maintain engagement with its new influx of volunteer mappers and utilize them in future crises? How will the advent of OSM and similar open-source platforms change the way the federal government responds to crises?

And how will NGA and other federal agencies learn from the success of NGA’s unparalleled public Ebola website? This is a question NGA Director Cardillo has already begun to ponder.

“In the future, [West Africans] will be able to use their new content and techniques well beyond the current crisis to build a modern geospatial database about their country for a future census and future elections,” Cardillo said during his Geography 2050 address. “While we’ve posted our context on the World Wide Web, we have much to learn, because this will be more and more common, and more and more necessary.”

 

ADDITIONAL RESOURCES

NGA Ebola Response Maps

Ebola Outbreaks 1976 – 2015

21 Maps and Charts that Explain Ebola

Mapping the Ebola Outbreak Region

Before and After Maps of the Ebola Response

WHO Ebola Response Report

TED Talk: Your Health Depends on Where You Live