on the night of October 28, 2025, Category 5 Hurricane Melissa came ashore near New Hope on Jamaica’s southern coast with sustained winds around 185 miles per hour, the strongest storm ever recorded to make landfall on the island (NASA, 2025). More than three-quarters of the country lost power. Whole neighborhoods in the western parishes were unrecognizable by morning. In the days that followed, teams at Jamaica’s National Spatial Data Management Branch assessed roughly 960,000 buildings and identified about 191,000 as damaged, close to a fifth of every structure in the country, and pushed that picture to international partners within a week (ICEYE, 2025).

That number is the story, and it is worth sitting with. A country of fewer than three million people produced a near-complete national damage assessment while much of it was still in the dark. Prime Minister Andrew Holness, accepting Esri’s President’s Award for the Government of Jamaica at the company’s user conference in San Diego this month, put the reason plainly from the stage. None of it, he said, was improvised.

Jamaica stood up a national emergency response GIS capability in the early 1990s and kept funding it, year after year, through governments and budget cycles and quieter seasons when no storm was on the radar. By the time Melissa formed, the Government of Jamaica had consolidated most of the country’s spatial data into a national repository and placed a single ArcGIS platform at the service of every ministry and agency (MEGID, 2026). The National Emergency Response GIS Team activated before landfall, building a common operating interface so that agencies could move into coordinated work the moment conditions turned. The tools were already in the hands of people trained to use them.

The map was never the deliverable

Most people outside the field hear “GIS” and picture a map. That framing undersells the thing by an order of magnitude, and the gap between the two readings is exactly where the value sits. A map shows you a place. A geographic information system lets you stack dozens or hundreds of independent datasets onto that place and ask questions across all of them at once. Population and building footprints. Power lines and water mains. Shelter capacity, road status, flood extent, satellite damage classification. The location is the join key. What you overlay on it is the product.

During Melissa, that stack meant every agency and every partner worked from the same evidence. Synthetic aperture radar imagery from ICEYE and damage maps from the Copernicus service flowed into Jamaica’s operational dashboards, giving responders near real-time visibility into flood extent and building damage (MEGID, 2026). Teams spent their hours answering questions instead of arguing over whose numbers were right. One shared picture, not a dozen conflicting spreadsheets.

Contrast that with what happens when the layer never reaches the individual.

Thirteen months before Melissa, Hurricane Helene drove historic rainfall into the southern Appalachians. Rain totals across parts of western North Carolina exceeded a one-in-a-thousand-year event (NOAA, 2024). At least 250 people died across six states, 107 of them in North Carolina, 43 in Buncombe County alone. Asheville went 53 days without restored city water (WRAL, 2025). Monica Youngman, chief scientist at the National Weather Service and a western North Carolina resident who lived through the storm, told the same San Diego audience what she heard on her own street while clearing debris with her neighbors. People kept saying they did not know this was coming. What they meant, she realized later, was that they could not picture how it would land on their house, their family, their block. The forecast existed. The translation to personal consequence did not.

Youngman’s team is now rebuilding how the Weather Service communicates, moving from a model built around radar coverage in the 1990s toward one built around the people who act on the forecast, embedding forecasters with emergency managers and pushing water inundation predictions down to the neighborhood scale. Same science, different last mile. The difference between Jamaica and Asheville was never the quality of the data. It was whether the decision layer reached a person in time to change what they did.

The backbone under that neighborhood-scale forecast is itself a mapping story. Ed Clark, director of NOAA’s National Water Center, and Fernando Salas, who leads its service development work, run the National Water Model, a system that forecasts flow across millions of miles of American rivers and turns those numbers into flood inundation maps at the street level. The intellectual roots trace to David Maidment at the University of Texas at Austin, who spent a career arguing for a geographic approach to water, the idea that a river forecast is only useful once it is joined to the specific places and people downstream of it. That is the same principle Jamaica applied to a hurricane and RFI applies to a train station. Water data becomes a decision only when it is tied to a location someone cares about.

The capability is built in ordinary time

Jamaica’s readiness looks like luck only if you skip the thirty years. The clearer lesson for a public sector CIO sits in the peacetime cases, where organizations built the same muscle with no hurricane forcing their hand.

Rete Ferroviaria Italiana manages Italy’s rail infrastructure, more than 2,000 stations carrying over two million passengers a day. A few years ago its planners studying Roma Trastevere, a station near Rome’s center, found that pedestrians from some neighborhoods faced a thirty-minute walk to the main entrance while the station’s northern side bustled (Forbes, 2026). That finding did not come from a track diagram. It came from a platform, first called StationLAND and later MobiLAND, that layers more than 400 categories of data onto every station in the country, from demographics and catchment areas to hospitals, schools, and heritage sites (Esri, 2026). Luigi Contestabile, who heads planning and station services development, describes the origin as noticing that the old approach forgot the people who lived and worked around the stations. Pier Paolo Olla, RFI’s director of station operations, has since built an internal GIS school of more than 200 specialists to keep the capability in-house. RFI went on to map the interiors as well, roughly 75 million square feet of indoor space, and now models station changes in three dimensions before committing capital. When the company’s leaders had to decide whether GIS was core business or a support function, they answered that it was core.

The scale of RFI can make this look like a big-budget move. Allentown, Pennsylvania, argues otherwise. The city runs its entire GIS operation with a two-person team, Elisa Coyle and Scott Rawhouser, building applications on Esri’s prepackaged solution templates rather than from scratch (Esri, 2020). The payoff showed up in a crisis that had nothing to do with weather. When the pandemic hit in 2020, Allentown led a four-city contact-tracing initiative across the Lehigh Valley, and it could move fast for one reason: it already had the ArcGIS platform in place and staff who knew how to run it. The investment was made in a period of resource scarcity, years before anyone needed it. That is the pattern. Two people and a standing platform beat a large team assembled after the emergency starts.

The private sector already prices this in

Commercial operators reached the same conclusion, and they did it with their own money on the line. Travelers, the sixth-largest insurer in the United States, helped form the National Insurance Crime Bureau’s Geospatial Intelligence Center, a consortium effort that flies ultra-high-definition aerial photography across the country on a fixed annual schedule to set a baseline (Esri, 2024). When a hurricane or wildfire hits, the aircraft launch as soon as conditions allow and shoot the same ground again. The comparison between the two flights is the damage assessment, and it lets the insurer engage customers before a single adjuster physically reaches the neighborhood.

Underwriting runs on the same logic in reverse. Carriers layer flood models, wildfire fuel loads, elevation, and historical loss onto individual addresses to price risk before a policy is written, and some have used that analysis to exit entire regions where the math no longer works. The baseline imagery and the address-level risk model are the same investment Jamaica made, aimed at a different loss. Build the geographic picture in calm conditions so the answer is ready when the conditions are not.

A utility makes the case in physical assets. Florida Power & Light, the operating company under NextEra Energy and the largest electric utility in the country, has spent more than a decade hardening its grid on a geographic foundation, mapping every pole, line, and device so it can choose which neighborhoods to underground based on outage history and which lines feed hospitals and 911 centers. During the 2024 hurricanes, its smart grid technology avoided roughly 2.7 million customer outages, and its underground neighborhood lines held up five to fourteen times better than overhead lines through Debby, Helene, and Milton (FPL, 2025). None of that targeting works without knowing where each asset sits, down to the pole. The same map that speeds restoration after a storm prevents the crew from cutting its own line on an ordinary Tuesday.

Five organizations, five sectors, one discipline. A national government, a state railway, a two-person municipal team, a private insurer, and a continental utility all treat location as the layer that everything else hangs from, and all of them paid for it before they needed it.

The loss does not require a hurricane to show up on a balance sheet. Excavation strikes on buried power, water, and telecom lines cost the United States around $30 billion a year, across nearly 200,000 damage incidents in 2024, and roughly a quarter trace to crews that dug without checking what was underground first (Common Ground Alliance, 2025). That is a location-data failure priced in real dollars, and it repeats in quieter forms across the enterprise: a retailer that opens a store where the catchment analysis was guesswork, a utility that cannot find its own assets, a logistics operator routing around infrastructure it never mapped. The decision made without the geographic picture is the expensive one, whether the trigger is a storm or a spreadsheet.