When Camp Mystic wanted to build in Central Texas areas considered high risk on Federal Emergency Management Agency flood insurance maps, it did what many property owners do: It hired engineers to take a closer look. FEMA's flood maps are often outdated and can be inexact, particularly in areas where the agency hasn't performed detailed studies. This left an opening for Camp Mystic to generate a more precise analysis of the terrain and to ask FEMA to change how its flood zones were designated. This was a common practice. From 2013 to 2020, the overnight girls camp asked FEMA to redesignate the flood risk for 65 of its buildings at its sprawling facilities in Texas' Guadalupe Valley — more buildings than previously known. The government agency told NBC News that in that period of time it altered the status of 60 buildings, changing the risk designations of both decades-old buildings and its new construction from "high" to "moderate" or "low," on paper. Five buildings remained in high-hazard zones. This summer, late on July 4 into the early hours of July 5,a flood beyond anyone's expectationskilled 28 campers and staffers, inundating many of the Camp Mystic buildings. A review of documents related to Camp Mystic — from county floodplain development records, an engineering study, FEMA flood map determinations and federal flood insurance studies — offers a window into a process that experts say plays out for thousands of properties each year, quietly shrinking the footprint of the nation's flood risk on paper, even as climate change makes flooding a more severe threat on the ground. The trail of documents from Camp Mystic details the ease with which property owners can remold how the federal government assigns flood risk. And they spotlight a national issue relevant beyond Texas: In some areas, FEMA's main tool for assessing flood risk is stuck in time. FEMA's mapping alongside Cypress Creek, where the camp expanded in 2018, is 15 years old and represents a rough estimate of flood risk. It relies on imprecise topography maps, and the rainfall data it uses was last updated in the 1970s. Moreover, the FEMA maps of the area do not account for modern projections for storms intensified by climate change. "One of the problems with FEMA is it appears to be negotiable as opposed to an empirical or science-based understanding of risk," said Jeremy Porter, the chief economist at First Street, a research firm that studies housing and climate risks. "It's based on the ability to create an engineering study and negotiate with FEMA." Properties in flood zones are often required to get costly flood insurance, and being in this area, on paper, can drive down property values. In most areas, construction in flood zones is more tightly regulated. In Kerr County, site of the most devastating Texas floods last month, property owners have toelevate structures at least 12 inches above the base flood leveland take other flood precautions if they're found to be in the floodplain. Better flood maps might not have made a difference for Camp Mystic, whose ownerspoke often about the perils of living near the waterways. The deluge dumped two to three months' of rainfall in about six hours, surprising local officials with its intensity. But the documents reviewed by NBC News show how widely flood maps can vary for a place like Camp Mystic, depending on who draws them. Flood insurance requirements and stricter local development rules in floodplains are designed, in part, to encourage construction away from those areas. It's not clear whether the new maps affected the camp's decision-making or development plans. "The unknown is whether being drawn out of those flood lines affected the choices the camp made of where to develop further cabins and structures," said Sarah Pralle, an associate professor of political science at Syracuse University's Maxwell School who has studied federal flood policy. In a statement, FEMA said its maps undergo an "extensive review and approval process" with community involvement. "Flood hazards change over time. Flood maps are structured to deliver a snapshot in time," of the 100- and 500-year floodplains, the agency's statement said. They're not meant to be "predictions" of where floods will happen, nor records of where they've happened before. "There is no such thing as a 'no-risk zone.' Flooding events do not follow lines on a map," the agency said. An attorney for Camp Mystic did not return requests for comment. About three-quarters of the nation's Flood Insurance Risk Maps (FIRMs) produced by FEMA are outdated,according to First Street. FEMA works with communities and local experts to create these maps but has struggled to keep them up to date. The process usually takes a few years and involves several steps of community review. The Kerr County engineer serves as the region's floodplain administrator and approves local permits to develop in the floodplain. These maps define the 100-year floodplain in the eyes of the federal government, using a combination of the area's topography and statistical models of historic rainfall data and bodies of water on the ground. They focus primarily on one kind of storm: A 100-year flood dependent on rainfall over a 24-hour time period rather than scenarios like what happened at Camp Mystic, which was deluged in a matter of hours. The precision of FIRM mapping varies widely. In Camp Mystic's case, FEMA engineers closely studied what level of flow it would take to cause a 100-year flood near its older camp along the South Guadalupe River. But no one had studied areas along Cypress Creek in detail. "FEMA looked at it and said, 'There's probably a floodplain there," said Chris Steubing, an engineer who is the executive director of the Texas State Floodplain Managers Association. The entire map relied on outdated rainfall data. In 2018, the National Oceanic and Atmospheric Administrationupdated its extreme rainfall predictions for Texas, improving on datasets last updated in the 1970s. The update raised expectations for how much rain could be expected to fall in an extreme event in 24 hours by 2-3 inches near Camp Mystic. But because FEMA's maps for this region were last finalized in 2011, they didn't incorporate this critical information. Also, while newer maps measure topography down to the inch using modern technology like LiDAR, portions of FEMA's map were built with imprecise 20-foot contours, according to the area's flood insurance study. The result was an imprecise representation of the slopes and ridges where water could flow — or build up. Porter, the economist from First Street, said FEMA has historically not received the funding needed from Congress to keep its mapping current, which leaves many communities waiting for updates and better data. "FEMA has some of the best engineers in the world," Porter said. "The problem is they're handcuffed." FEMAs future is murky: As of last month, theagency's staffing was down by a third and many senior officials had left. President Donald Trump has discussedboth overhauling or eliminating FEMA. The agency acknowledges on its website that its mapping is limited. That creates the foundations for a system in which property owners seek to redraw federal flood maps and seek exemptions for their property. The process requires property owners to submit new surveying or analysis from an engineering firm. In 2019, Camp Mystic filed floodplain development permits in Kerr County that show Hewitt Engineering completed a "detailed study" of Camp Mystic's property in 2018. The engineering firm's analysis likely formed the backbone of its most recent appeals to FEMA. In this version of the flood map, the 100-year floodplain was about 44% as big as FEMA's, in areas that had not been studied comprehensively by the agency, according to an analysis by NBC News based on previously unreported documents. A few narrow slivers of Hewitt's 100-year flood zone are outside of FEMA's high-risk areas. John Hewitt, Hewitt Engineering's president, did not return calls. Robert Lamer, a professional engineer based in Michigan who was not involved in Hewitt's work, reviewed the study's map for NBC News. Lamer's review indicated that the firm surveyed the topography on the ground and modeled water flow to make its map, both of which in theory could make it more precise. But, the documents are not clear about the variables Hewitt used for its modeling, including whether it used outdated rainfall data like FEMA. Lamer said closer analysis can find that FEMA's approximated maps can overestimate flood risks — something clients are usually hoping for. "I've had lenders say, 'Prove to me we're out of the floodplain' and we do the work and you're 30 feet above the river," Lamer said of FEMA's original mapping. "That's how inaccurate these maps are." Across the nation, appeals to adjust FEMA maps are common before and after FEMA's maps are finalized. Pralle, the Syracuse professor who has studied flood policy, and Devin Lea, another academic, reviewed five years of data on how FEMA maps are revised. They found that over 20,000 buildings in 255 counties across the U.S. werere-mapped outside of special flood hazard zonesfrom 2013-2017 through several appeals processes. More than 700,000 buildings remained in special hazard flood areas in those counties. The agency approves the vast majority of map amendments, Pralle said, and Lamer, who has worked on hundreds of map amendment applications, said he's only had one rejection. In that sense, Camp Mystic's 92% success rate with exemptions is not an anomaly, but the norm. "You don't submit it if it's not going to get approved," Lamer explained, because there is no financial incentive for clients to continue with the process unless the data shows their flood exposure is less than what FEMA has determined. While FEMA's high-risk flood zones often grow after the agency finalizes new maps, property owners and communities can push to shrink those zones later. Changes to special flood hazard zones are more common "where median home values are higher, buildings are newer, and percentage of white populations are higher," according to a study published in Risk, Hazards & Crisis in Public Policy by Pralle and Lea. The pair's research suggests the appeals system's incentives align to shrink federal flood maps. "FEMA does not have the resources to go out and double-check" in the field, Pralle said. A FEMA spokesperson said the agency reviewed Camp Mystic's cases and the elevation data submitted, in accordance with its policies. The agency noted that the amendment approvals "do not materially change the reality of the risk and dangers of flooding." Storms like those that devastated Camp Mystic are expected to happen more often in a warming world. To shore up existing blind spots, independent organizations are building data-rich tools to better predict the growing risk of intense rainfall. For example, First Street incorporates global climate models to forecast weather extremes and incorporate those into its risk maps. The firm provides data and analytics to individuals, banks, investors and governments, among others, for a fee. Nationally, its analysis found more than twice as many buildings fell within 100-year floodplains compared to FEMA maps. The discrepancy was overwhelmingly due to heavy precipitation risk that FEMA maps didn't capture, Porter said. The firm's mapping of Camp Mystic's 100-year flood zone shows that portions of both the older and new camp sites would be flooded in such an event. In some areas, the flood zone is outside of both Hewitt's and FEMA's unstudied 100-year floodplain; in other areas, it's much narrower and hews closer to the work of Hewitt Engineering. Steubing, of the floodplain association, said early indications suggested that the flooding that occurred on July 4 was an event that could be expected once in 800 years, but more work is needed and several engineering firms continue to evaluate the flood's extent. It's not yet clear how precisely the extent of flooding corresponded to the various risk maps. Although First Street's mapping better incorporates climate risk, it has limitations of its own — namely, that it lacks the kind of detailed survey and streamflow analysis work apparently completed by Hewitt. "We don't have feet on the ground," Porter said. In an ideal system, flood mapping would combine detailed on-the-ground engineering, modern rainfall and streamflow data and predictions about future climate risk. Steubing said floodplain managers need more dynamic tools that depict different flooding scenarios — like fast-falling deluges that blanket small areas, and less rapid but persistent storms that last days. That would help determine risk much more precisely for individual communities. Texas is trying to address a slew of historic data gaps to move in that direction, Steubing said. But much of the state, like portions of the landscape near Camp Mystic, has never been studied in detail or hasn't been mapped at all. To address those gaps, thestate has helped fund a new program with FEMA, calledBase Level Engineering. The effort focuses on using high-resolution LiDAR data and modern modeling to estimate base flood levels in places that haven't been studied closely. The maps are meant to complement, not replace FEMA's FIRM maps. Thenew mapping, which is now available statewide and covers the area near Camp Mystic, was released about six months ago, Steubing said, and represents the kind of next-level model that could help prevent the next disaster.
