Map Monday: Tracking Relocation and Crowdsourced Mapping during Hurricane Maria

By Matthew Leger • February 11, 2019

  • February 11, 2019
  • GIS

Hurricanes are wreaking havoc across the globe, hitting coastal areas and islands with unprecedented power. As global temperatures have risen, the intensity of hurricanes has risen as well, particularly in the North Atlantic basin. Over the last several decades in the US, very powerful hurricanes have slammed coastal states and territories, disrupting lives, business, and government at enormous costs. Of the ten costliest disasters in US history, seven have been hurricanes with six of those occurring in the last 14 years. This trend is projected to get worse in the coming years.

While studies have shown that hurricanes have had minimal impact on the broader US economy, they have come at a high cost to individual American families. Americans have lost their homes, their jobs and their lives or the lives of their loved ones. Hundreds of thousands, even millions, have been forced to relocate, some temporarily and some permanently, as their hometowns lay in ruin. Hurricane Katrina alone forced 400,000 residents from their homes with many unable to ever return as clean up and rebuilding continues more than a decade later. To this day, no one really knows for sure the full extent of the displacement from Katrina as the US Census Bureau and others struggle to adequately track such a large pattern of migration.

According to the Global Report on International Displacement 2018, just under 1.7 million Americans were displaced due to natural disasters in 2017, ranking fourth in the world behind China (4.5 million), the Phillipines (2.5 million), and Cuba (1.74 million). In total, 18.8 million were displaced globally due to natural disasters.

Graph showing displacement by country
Image from the Global Report on International Displacement 2018

As demonstrated post-Katrina, accurately tracking the large migration patterns before, during and after a disaster is nearly impossible, especially as more hurricanes and other natural disasters pile on. Migration trackers can only estimate where and how many people move.

As people move away from areas struck by disaster, the local and state governments of their new homes can become overwhelmed by the influx of new residents. School systems are forced to pack their classrooms with more children, public transit systems take on more riders, and poverty rates increase as those who lost everything are forced to start over with limited jobs available to support everyone. Many cities also struggle to keep up with the demand for housing as tens of thousands of new residents move in, seemingly overnight in some cases. There is no easy way to prepare for the movement and displacement of people during major disasters. However, more data could help governmental entities better prepare for and serve those seeking shelter.

Mapping Puerto Rico’s Real-Time Migration Patterns Before, During, and After Hurricane Maria

On September 20, 2017 at around 6:15 am, Hurricane Maria touched down in Yabucoa, Puerto Rico as a Category 4 hurricane with 155 mph sustained winds. Maria devastated the island, dropping as much as 38 inches of rain in some areas, knocking out nearly 80% of the island’s power lines and costing around $95 billion in damage. It was the third costliest disaster in US history. Those who stayed on the island through the disaster were left with limited access to basic necessities like food, water, and medicine and much of the island was left without power for months. As a result of the hurricane and its longer-lasting impacts, upwards of 3,000 Puerto Ricans lost their lives, making it one of the deadliest natural disasters in US history.

While the American Community Survey (ACS) and the Puerto Rico Community Survey (PRCS) of the US Census Bureau each track state-to-state migration on an annual basis, they use limited sample sizes and are conducted only once annually. This means those who left and returned to Puerto Rico prior to the survey being conducted would never be accounted for, leaving room for a very large margin of error. While FEMA can track and map applications for disaster relief related to Hurricane Maria, this is dependent entirely on those fleeing disaster to reach out to the US government for help. One new solution can help governments at all levels better prepare for disaster-related displacement by tracking cell-phone movements and mapping migration patterns in real time.

Using aggregated cellphone data from an unidentified carrier following very strict data privacy standards, data analytics company Teralytics mapped the migration patterns in and out of Puerto Rico from August 2017 to February 2018 from a sample of 500,000 Puerto Ricans. They found that 407,465 Puerto Ricans in their sample left the island and 359,813 returned in that timeframe, leaving nearly 50,000 Puerto Ricans on the US mainland. They could show exactly when they left, where they went, and when (or if) they returned to the island. The map below shows travel patterns out of Puerto Rico (in red) and return travel patterns (in blue) exactly as they occurred.

Using this data, Teralytics mapped the US states and counties where Puerto Ricans in their sample chose to relocate. The following map shows that Florida (43%) was the destination of choice for many Puerto Ricans, followed by New York (9%), Texas (7%) and Pennsylvania (6%). Some chose to relocate to Massachusetts, California and Illinois. Relocation data was broken down to the county level, highlighting those counties with the largest influx of Puerto Rican migrants. Seven of the top 10 counties that received the most migrants were in Florida.


Map of migration  

Using this data, all levels of government as well as nonprofit organizations could more effectively and efficiently deliver services to those who need it, exactly where they need it, for as long as they need it. Understanding these patterns of migration could also give emergency managers very valuable information to inform future evacuation plans. Government officials at all levels could also use this data to better prepare resources for deployment to areas of the country that would expect a higher influx of migrants in the event of a future disaster.

Understanding disaster-related migration patterns is very important for recovery efforts. Equally important is the effectiveness of the response efforts on the ground immediately following a major disaster. Once again, location intelligence can play an important role for government as response, clean up and recovery begins.

Volunteer Cartographers Come Together to Help Puerto Rico Recover

During or following a disaster, navigating a city or island is easier for first responders when buildings, landmarks and street signs are still standing. When those structures are gone, response and recovery efforts become far more difficult. Highly detailed maps can alleviate some of that complexity by giving emergency responders a visual aid that helps them navigate the landscape and more effectively deploy resources where they are most needed. For disasters on the US mainland, emergency responders typically have access to very detailed maps that display data about buildings, homes, streets and other public assets; however, that was not the case for Puerto Rico following Hurricane Maria.

Aware of the lack of geospatial data available to emergency responders, volunteer cartographers came together for a Mapathon at Columbia University (as well as other academic institutions) to provide the critical information emergency managers needed for a more effective response. Volunteers analyzed GIS data and satellite images prior to and after the destruction on the island. They added buildings, streets, bridges and other pieces of infrastructure that were previously never present on any maps of Puerto Rico and made note of their current state. They provided a more complete picture of what the island once was and what it currently looked like on an open data platform known as OpenStreetMap, which is used for mapathons all over the world and is accessible and free to use for anyone with internet access.

On the ground in Puerto Rico, responders from the American Red Cross and FEMA relied on OpenStreetMap to deploy resources and navigate the island after Maria. The maps were accessible via mobile devices, allowing teams to remain updated throughout their response missions. They could even update information from the field where they could warn other teams about barriers along the way like blocked roads or damaged bridges, enabling other response teams to get where they needed to be most efficiently.

Collectively, tens of thousands of volunteers from all over the world have contributed to crisis mapping projects over the last decade. Thanks to organizations like the Humanitarian OpenStreetMap Team (HOT), academic institutions like Columbia University, and concerned global citizens from all corners of the planet, areas devastated by natural disasters can rebuild using geospatial data that they never had access to before. While rebuilding will take years, ongoing crowdsourced mapping efforts may make the difference that helps Puerto Rico come back stronger than ever before, giving those forced to relocate a great reason to return home.

About the Author

Matthew Leger

Matt Leger is a Research Assistant for the Innovations in Government Program at the Ash Center. He has a diversity of experiences in research across the public and private sectors, as well as in academia with a primary focus on understanding how technology can be used to help address some of society’s greatest challenges. Matt has worked with the Smart Cities Strategies team at the International Data Corporation (IDC); the NYCx team in the NYC Mayor’s Office of the Chief Technology Officer; and at the research institute CTG-UAlbany. He holds a Bachelor of Arts in Public Policy and a Master’s Degree in Public Administration both from the Nelson A. Rockefeller College of Public Affairs and Policy at the University at Albany in Albany, NY.