Helene tested what continuity plans take for granted: staff, signal and signatures

The week

Helene formed as a tropical storm in the northwestern Caribbean on September 24, 2024, and became a hurricane the next day near Cancun. Over very warm Gulf water it intensified rapidly on September 26. Its wind field was unusually large, bigger than about 90 percent of storms of similar strength in that area, according to the National Hurricane Center.

Late on September 26, Helene came ashore about ten nautical miles southwest of Perry, Florida, in the Big Bend region, the curve of coastline where the peninsula meets the Panhandle. It arrived as a Category 4 storm with winds of about 140 miles per hour. No stronger hurricane had made landfall in the Big Bend since reliable records began around 1900.

It did not stop at the coast. Helene was moving at about 30 miles per hour, which carried hurricane-force gusts deep inland across Georgia and the Carolinas. The worst of what followed happened hundreds of miles from the sea.

The mountains of western North Carolina had already been soaked. On September 25 and 26, a stalled front and the terrain itself produced heavy rain well ahead of the storm, what forecasters call a predecessor rain event. Soils were saturated and rivers were rising before Helene’s core arrived.

On the morning of September 27, rain fell at two to three inches an hour on top of that. The water had nowhere to go. The highest total the Hurricane Center recorded was 30.78 inches, at Busick in Yancey County. The U.S. Geological Survey documented more than 2,000 landslides, most of them in western North Carolina.

Rivers broke records that had stood since the great flood of 1916. The French Broad River at Asheville crested at 24.8 feet, against a previous record of 23.1 feet. The Swannanoa River at Biltmore reached 27.33 feet, far above its old record of 20.7. Catastrophic flooding also hit parts of Tennessee, Virginia and South Carolina, in towns that had never thought of themselves as hurricane country.

Around Asheville, roads and bridges were destroyed. A stretch of Interstate 40 along the Pigeon River, near the Tennessee line, was washed out. Across the border, Interstate 26 in Unicoi County was closed for weeks. In nearby Erwin, the county hospital flooded almost to its roof, and staff moved 70 people, 11 of them patients, onto the roof to be rescued by helicopter.

Rescue crews worked across the whole region. The Hurricane Center counted at least 2,700 people rescued from high water in the Southeast, more than 1,000 of them from freshwater flooding in western North Carolina. The Coast Guard alone rescued at least 865 people there.

Power failed across large areas, and so did cellular service, the mobile phone network that most people, and most modern offices, now depend on for almost everything. The city’s water system was severely damaged. For many residents, the ordinary infrastructure of daily life simply stopped: no lights, no phone, no running water, and no easy way out.

The scale was enormous. The National Hurricane Center, the federal forecasting office that tracks and later documents every Atlantic storm, counted at least 250 deaths in the United States, 107 of them in North Carolina. That made Helene the deadliest hurricane in the contiguous U.S. since Katrina in 2005. The center estimated that 7.4 million customers lost power, about 1.18 million of them in North Carolina.

Federal, state and local agencies, utilities and volunteers mounted a long response, and federal disaster declarations followed within days. Inside the local governments at the center of the disaster, a less visible test was under way: of the plans those agencies had written for how they would keep working when things went wrong.

Those plans, it turned out, rested on assumptions that Helene removed all at once. People would be available to do the work. Phones and internet would let them communicate and reach their systems. And the officials who sign off on spending and emergency decisions would be reachable.

Every public agency of any size has, or should have, a continuity of operations plan. Emergency managers call it a COOP. It is the agency’s answer to a simple question: if something bad happens, how do we keep doing the things the public depends on? Issue benefits, answer 911 calls, pay staff, process permits, keep the water running.

Most continuity of operations plans for public agencies are built around the loss of a building or a system. The headquarters floods, so staff move to an alternate site. The server room catches fire, so the agency switches to a backup copy of its data. These are real risks, and planning for them is worthwhile.

Helene removed something more basic across a whole region. Offices stood but could not be reached. Systems in distant data centers kept running, but nobody could connect to them. Staff were unavailable because they were survivors themselves, searching for family, without water or fuel.

An agency keeps serving the public in a disaster only if three inputs hold: people to do the work, a way to communicate and reach systems, and someone with authority to decide. Call them staff, signal and signatures. Plans usually assume all three. Helene removed them together. They have different owners and different lead times, which is why they need separate preparation and why they fail together when nobody owns the whole.

The argument holds across levels of government, with differences of degree. A town with a few dozen employees has little depth in any of the three. A large county has depth but more dependencies. State agencies and federal field offices can shift work to staff in other regions, if they have planned for it. The sections below take the plan’s hidden assumptions first, then each input in turn, then one function under stress, and finally what is at stake for the people accountable.

The assumptions buried in a continuity plan

A continuity plan typically lists the agency’s essential functions, the staff who perform them, where they will work if the main office is lost, which records and systems they need, and who takes charge if the leader is unavailable. Good plans are exercised: people walk through a scenario around a table and check whether the plan holds up.

What the federal guidance already covers

The ingredients are familiar. FEMA’s continuity guidance has long organized a plan around a standard set of elements. They are essential functions, orders of succession, delegations of authority, continuity facilities, continuity communications, essential records, human capital, testing and exercises, devolution and reconstitution.

Federal executive branch agencies must follow Federal Continuity Directives. State, local, tribal and territorial governments are pointed instead to FEMA’s Continuity Guidance Circular, which is guidance rather than a mandate. FEMA had issued an updated circular in August 2024, a month before Helene.

Two of those elements matter most here. Devolution is the transfer of essential functions to staff at another location when the primary organization cannot perform them. Reconstitution is the return to normal operations afterward. Many local plans treat devolution as a single sentence naming a partner, with no accounts, training or written agreement behind it. The gap stays invisible until the region itself is the problem.

The weakness usually lies less in what the plans say than in what they assume without saying. A plan that says “staff relocate to the alternate site” assumes staff can get there and are free to come. A plan that says “systems are hosted in the cloud and remain available” assumes someone can connect to them. The cloud means computers in someone else’s data center, reached over the internet; if the internet connection is gone, they are out of reach however well they run. A plan that says “the county manager authorizes emergency spending” assumes the county manager can be found.

The three inputs work like the legs of a stool. Most planning strengthens one leg at a time, with a better backup site or a better data backup, but a regional disaster kicks out all three at once, and the stool falls no matter how sturdy any single leg is.

Why regional failure is different

Single-site planning assumes that the surrounding region still works. Staff from an unaffected office drive over. The phone company fixes one line. A deputy at home takes the call. A regional disaster removes that background, because the same roads, power grid, carriers and fuel suppliers serve everyone in the area. Failures that would be independent in a building fire become correlated in a flood.

Vendors share the exposure too. A local managed service provider, the regional office of a payroll processor, the bank branch that takes deposits and the print shop that produces checks may all be in the same flooded valley. A dependency list that records only the vendor’s name, and not where its people and equipment are, overstates resilience. The fix is a location column in the vendor register and a question at each renewal about the vendor’s own continuity arrangements.

Take a county in a mountain region with a well-written plan. Its data sits safely in a vendor’s data center two states away. Its alternate office is across town. Its finance director has authority to approve emergency purchases. Then a regional flood hits. The alternate office is fine, but the road to it is gone. The data is safe, but cell towers are down and nobody can log in. The finance director’s house is cut off and she cannot be reached for four days. Every element of the plan worked. The plan as a whole did not.

Staff: essential personnel are also survivors

Plans list essential personnel and expect them to report. In a regional disaster many cannot, and should not be pressed to until their households are safe. Homes are lost. Schools and childcare close. Fuel is scarce for days. A dispatcher whose own street is flooded is not going to be at the console, and a good employer would not want her there until her family is safe.

Agencies that coped well treated staff welfare as an operational task rather than a human resources nicety. They checked on employees, supplied water, fuel or lodging where they could, and planned for partial availability lasting weeks rather than days.

The planning consequences are specific. Each essential function needs more than one trained person, and at least one should be based outside the likely impact zone, whether in another office of the agency, a partner jurisdiction or a contractor. Remote-work arrangements should name which out-of-area staff can pick up which work, with access already provisioned. “Provisioned” means their accounts, permissions and equipment are set up in advance, rather than requested in the middle of the crisis when the people who grant access may themselves be unavailable.

Devolution partners and what they need

A partner jurisdiction can only pick up work if it can legally and technically do it. Some functions travel easily: answering a call line, processing invoices, updating a website. Others carry rules with them. A backup dispatcher who can see criminal justice information must meet the FBI’s CJIS Security Policy, which requires fingerprint-based background checks for people with unescorted access to that information.

Benefits work has its own constraints. Eligibility workers need roles in state benefits systems, and those roles are granted by the state, not by the county. Permitting staff may need local code knowledge that a distant partner lacks. The functions that look most clerical are often the hardest to hand over.

Each pairing therefore needs a short written agreement: which functions, under what trigger, using whose systems, with what access, and who pays. Reciprocity helps, because two jurisdictions in different regions can each serve as the other’s backup. The agreement should be tested at least once a year with a real transaction, such as a partner clerk processing a batch of invoices in the live system.

The security trade-off

Pre-provisioned access for outsiders widens the attack surface. Accounts that exist all year for an emergency that may never come are exactly what attackers look for. The usual compromise is to create the accounts in advance but keep their privileges disabled or narrow, with a documented way to enable them quickly.

That enabling step must itself work without the local network. In effect, more than one administrator, in more than one place, must be able to do it. Quarterly reviews then confirm that the named people still work there and still need the access.

Mutual aid also matters. The Emergency Management Assistance Compact is an agreement among states to lend each other personnel and resources in emergencies, and many states have similar arrangements among their own local governments. These agreements are usually seen as a way to bring in firefighters and police. They should also cover administrative and technical roles, such as permitting clerks, payroll staff and network engineers. The back office keeps the front line running.

The mechanics differ by level. EMAC operates state to state. Assistance flows when a governor has declared an emergency, requests run through each state’s emergency management agency, and the compact settles questions of cost and liability. Its members are all 50 states, the District of Columbia, Puerto Rico, Guam, the U.S. Virgin Islands and the Northern Mariana Islands.

EMAC also supports virtual missions, so an administrator in another state can do the work remotely without traveling into a damaged area. County-to-county help usually runs through a separate in-state agreement, and that agreement should name back-office roles explicitly rather than leave them to interpretation.

Lead time and pay

Training lead time is the constraint. A backup for a benefits eligibility worker or a 911 dispatcher takes months to prepare. That work must happen in calm years, when nobody feels urgency, which is exactly why it is so often deferred.

Pay continuity belongs here too. Staff who cannot reach a timekeeping system still need to be paid on schedule. Missing a paycheck in the middle of a disaster compounds the hardship of the people the agency most needs. So payroll needs a documented fallback, such as repeating the prior period’s pay and reconciling later, approved before it is needed.

Planning for weeks, not days

Most continuity plans assume a disruption measured in days. Helene showed that staff availability can stay partial for weeks. Asheville went without drinkable tap water for more than seven weeks. Employees living with that cannot work normal shifts, and the plan should say so. It should rotate people, cap hours and bring in relief staff before exhaustion sets in.

Leave and pay policies need the same foresight. Some employees will need paid time to deal with lost homes, insurance claims and children out of school. Others will work long overtime. Public employers should settle in advance how emergency leave, overtime and compensatory time will work, so that supervisors are not inventing policy unit by unit. Consistent rules also reduce later grievances and audit findings.

Small organizations face the hardest version of this. A town with a single finance clerk cannot train an in-house backup. Its realistic options are a partner town, the county, a contracted bookkeeper outside the region, or a state association that keeps a roster of volunteers. Any of them is better than discovering the gap on the first payday after the storm.

Signal: the network beneath every procedure

Modern agency work depends on cell networks, internet access and power, usually all at once. When towers lose power or backhaul, staff cannot call each other, reach cloud systems, receive alerts or approve a login on a phone app. “Backhaul” is the connection that carries traffic from a cell tower back into the wider network. A tower can be standing and powered and still useless if the fiber cable or link behind it is cut. Helene’s flooding damaged infrastructure that took days to weeks to repair.

Resilience here is about diversity of paths: more than one way to talk and more than one way to connect. Key sites need satellite terminals and the skills to set them up. Radio systems need to interoperate with neighboring jurisdictions, so that a deputy from one county can talk to a road crew from the next. Generators need fuel contracts that assume every other customer will be calling the same supplier on the same morning.

Priority services and their limits

Federal programs help at the margins. CISA runs three priority telecommunications services for people with national security and emergency preparedness roles, and state, local, tribal and territorial governments are eligible. GETS gives priority to calls on landline networks. WPS gives priority to calls on cellular networks. TSP tells carriers to prioritize the installation and restoration of an organization’s enrolled voice and data circuits.

Each has limits worth knowing before a storm. WPS helps when a network is congested, not when the tower is dark. TSP helps only for circuits enrolled in advance, and it shapes the order of repairs; it cannot conjure a crew or a bridge. Enrollment is an administrative chore: naming eligible staff, setting up their phones and recording which circuits serve which essential functions. Few agencies do it outside a crisis, and fewer keep it current.

Where the data lives and how to reach it

Helene did not show that cloud hosting was a mistake. Servers in a flooded basement would have been lost outright, while data in a distant data center survived. The weakness was the path between staff and the data. An agency can design for that by giving staff more than one way in, from more than one place, and by keeping local copies of the few records needed in the first days.

Those records make a short, specific list: staff and contractor contact details, the vendor master file, the last payroll register, open purchase orders, the delegations of authority and the continuity plan itself. Exported on a schedule, encrypted and kept on devices held by named people, they let work begin before connectivity returns. Each copy is also a data-protection risk, so the list should stay short and the devices should be tracked.

The login problem

Identity is easy to miss. Most organizations now use multi-factor authentication: to sign in, you need your password plus a second proof, very often a notification sent to an app on your phone that you tap to approve. It is a sound security practice. But if every login requires a push notification to a phone, no one can sign in while cell service is down.

Hardware security keys, small physical devices that plug into a computer or tap against it, or pre-issued backup codes held securely, keep privileged access working without a signal. “Privileged access” means the accounts with power to change systems, the ones administrators need most in a crisis.

The detail matters. A one-time code from an authenticator app is generated on the phone itself and needs no signal, unlike a push notification or a text message. But the laptop still needs some connection to reach a cloud identity service, so offline codes solve only half the problem. Hardware keys have the same limit. They remove the dependence on the cell network, not on the internet.

Major identity providers, Microsoft among them, recommend keeping at least two emergency access accounts, often called break-glass accounts. These are excluded from the normal phone-based checks, protected by hardware keys or certificates, and held by different people in different places. Any use should trigger an immediate alert, because an attacker who finds one holds the keys to everything. The answer to that tension is close monitoring rather than weaker protection.

Paper still matters

Offline copies matter too: contact lists, procedures, maps, delegations and key records, printed or on encrypted devices. The emergency plan stored only on a shared drive is no help if the shared drive cannot be reached.

Radio fills part of the gap. North Carolina runs a statewide public safety radio network known as VIPER, and many states operate similar systems. These networks were built for responders, though, and few administrative staff carry radios or know how to use them. Deciding in advance which back-office roles get a radio, and training those people, costs little and extends the network to the staff who handle money and records.

Security policy has to anticipate emergency connectivity as well. A satellite terminal that the agency’s network rules forbid connecting is unusable, so the rules should define in advance how emergency links are authorized and monitored. Otherwise staff face a bad choice in the moment: break the rules to serve the public, or follow them and leave the public unserved.

Signatures: who may decide when the chain is broken

Spending, emergency procurement, fee waivers, deadline extensions, emergency declarations and staff redeployment all require authority. In normal times it sits with named officials. In a regional disaster those officials may be unreachable for days.

Orders of succession and delegations of authority are standard continuity elements. An order of succession says who takes over if the leader is unavailable; a delegation of authority says who may make specific decisions, such as approving purchases, and within what limits. They are often stale or unknown to the people who would use them. A delegation that names a deputy who retired two years ago is worse than useless, because it creates the illusion of a plan.

They should be current, specific about dollar limits and duration, and stored where they can be found offline. Updating them may need a council or board vote, which gives this input the longest lead time of the three. Training a backup takes months; getting a governing body to approve new delegations can take a full budget cycle.

What a usable delegation contains

A delegation that works in a disaster answers six things on one page. It names the trigger, such as a declared state of emergency or a set number of hours without contact. It names positions and people, at least three deep. It lists the decisions covered and the dollar limit for each. It sets a time limit after which authority lapses unless renewed. It requires a written record of every use. And it sets a date for the governing body to ratify what was done.

State law frames all of this. State emergency management statutes typically let a mayor or board chair declare a local state of emergency and exercise defined powers, and local ordinances and charters fill in the rest. Those powers vary widely between states, and between cities and counties in the same state. Counsel should confirm which officials may delegate which authority, and whether the delegation itself needs a vote.

The case against broad delegations

The strongest objection is accountability. Wider emergency authority invites mistakes, favoritism and fraud, and disaster spending has a long history of all three. A governing body is right to be wary of signing away control.

The answer is design rather than refusal: tight time limits, caps, a log, prompt public reporting and ratification at the next meeting. Narrow, bounded authority that can be used is safer than broad authority that nobody can find. Without it, the realistic alternative is decisions made with no authority at all, as the example later in this article shows.

The paperwork with a long tail

Emergency purchasing deserves its own attention because the paperwork has a long tail. When a disaster is federally declared, FEMA, the Federal Emergency Management Agency, can reimburse state and local governments for a large share of eligible costs through its public assistance program. But FEMA public assistance reimbursement depends on documentation of what was bought, why, at what price and under which emergency procurement rule.

The federal rules are specific. Under the Stafford Act, FEMA public assistance covers at least 75 percent of eligible costs, and the federal procurement standards in 2 CFR Part 200 apply to purchases paid for with that money. Those standards allow noncompetitive purchasing when a public emergency will not permit the delay of a competitive process. That justification lasts only as long as the emergency does, and the agency must document it.

Time-and-materials contracts deserve particular care. Federal rules allow them only when no other contract type is suitable, and they require a ceiling price and close oversight. Standing contracts help most when they were competitively procured to federal standards and include the clauses federal rules require. A contract that skipped those steps can still be used in an emergency, but part of its cost may not be recovered.

Agencies should keep pre-approved emergency procurement procedures, standing contracts for debris removal and supplies, and a simple logging method that works on paper. Records kept in the first chaotic weeks determine how much is recovered months later. A crew chief who buys chainsaws with a personal card and keeps no receipt has solved a problem today and created one for the finance office next year.

A workable logging method is simple. One form per purchase records the date, item, quantity, price, vendor, who authorized it, under which authority, and which essential function it served. A photo of the form and the receipt, uploaded once a connection returns, preserves the record twice. Someone in finance should own the log from the first day, rather than reconstructing it when the auditors arrive.

One week of paying suppliers after a flood

Return to the mountain county, and follow one ordinary function through the first week after a regional flood: paying emergency suppliers.

On day one, the county needs bottled water, generator fuel and a contractor to clear a blocked road to the hospital. The purchasing officer who normally handles emergency orders lives in a valley that is now cut off. Her backup works at the next desk and lives three miles away, also without power. Staff has failed.

The deputy county manager, who is reachable, tries to log into the procurement system to raise an order. The system is fine, running in a data center in another state. But his login requires a tap on a phone app, and his phone has no service. The IT administrator who could reset his access has the same problem. Signal has failed.

The deputy decides to proceed anyway, placing orders by radio and on paper. But the county’s delegation of authority caps his spending at a level far below what the road contract will cost, and the two officials who could approve more cannot be reached. He goes ahead because the hospital road matters more than the rule. Signatures have failed too, and that creates a second problem months later, when auditors and FEMA reviewers ask under what authority the contract was signed and where the paperwork is.

Now replay the same week with preparation. A purchasing clerk in a partner county two hundred miles away, trained last year and given access in advance, raises the orders remotely. The deputy signs in with a hardware key that needs no phone signal. A current delegation, approved by the board the previous spring, lets him authorize emergency contracts up to a set limit for thirty days. A standing debris contract means the road crew is already under agreement. Every decision is logged on a printed form designed to match FEMA’s documentation needs.

The second version requires nothing expensive or technically difficult. All of it had to be done before the flood.

What the second version cost

The prepared county spent its effort in four places. Finance and human resources negotiated a reciprocal agreement with a partner county and trained two of its clerks. IT bought hardware keys for a handful of administrators and set up and tested break-glass accounts. The county attorney drafted a delegation and put it on a board agenda. Purchasing competed and signed standing contracts for debris removal, fuel and water.

Each item was modest on its own. Together they needed time from several departments over a year, and one board vote. The harder obstacle was coordination. Each task belonged to a different department, and no single department felt the risk keenly enough to push the others, which is why ownership comes up again below.

Service, money and authority on the line

Elected officials and agency heads can easily treat continuity as a technical matter for emergency managers. Each of the three inputs, though, touches something they are accountable for.

• Public service and trust. Residents judge government by whether it shows up when things are worst. Benefits that stop, 911 calls that go unanswered or permits that freeze during recovery leave lasting damage.
• Money. Poor documentation can mean unrecovered disaster costs that land on local budgets for years.
• Staff. How an agency treats employees who are also survivors shapes retention and morale long after the water recedes.
• Legal exposure. Decisions taken without clear authority can be challenged later, and so can spending outside procurement rules.

Well-prepared agencies have at least two trained people for every essential function, one of them out of the region. Their privileged accounts can be reached without a phone signal. Delegations are current, specific and printed. Emergency purchasing procedures and contracts exist before the storm. And one person or group is responsible for seeing all three inputs together, rather than leaving each to its separate department.

The usual failures follow the same pattern: a backup who lives next door to the primary; an emergency plan that lives only in the cloud; a delegation nobody has read since it was signed; an exercise that assumes everyone is in the room.

How the exposure differs by type of government

Small towns have the least depth and the most to gain from partnerships. One clerk, one IT contractor and a part-time manager can each be a single point of failure, so their continuity depends on the county, a neighboring town or a contracted service outside the region.

Counties carry the widest range of essential functions: 911 dispatch, the sheriff, elections, social services, public health and often water. Several come with state or federal rules attached, such as CJIS for dispatch and state systems for benefits, which complicates handing work to a partner.

Special districts and public utilities face a revenue problem as well as an operational one, because their income depends on billing that may have to stop. State agencies are better placed, since they have staff in unaffected regions. Federal executive branch agencies are bound by the Federal Continuity Directives to plan for devolution. Their weakness tends to be the local field office that serves the disaster area and is itself inside it.

School districts deserve a mention because their buildings often serve as shelters and distribution points. Their continuity plan has to cover two roles at once: education, which stops, and community support, which suddenly expands.

The case for doing less

The strongest argument against this work is cost against probability. A mountain county may see a storm like Helene once in several generations. Staff time spent on reciprocal agreements and board delegations has an opportunity cost in a year of tight budgets. Federal aid, state support and mutual aid will arrive eventually.

That argument is weaker than it looks. Most of these preparations cover any regional disruption, from ice storms and wildfire to a carrier outage or a cyberattack on a shared provider. Their cost is mainly staff time, not capital. And the period before outside help arrives is exactly when they matter. Federal aid reimburses money later; it does not supply authority, logins or trained clerks in the first week.

Testing all three together

The three inputs fail together, so they should be tested together. Tabletop exercises usually assume participants can meet and talk. A more realistic drill removes some participants without notice, takes away phones and shared drives, and asks the remaining staff to find the procedures, the connection and the authority they need.

Human resources owns staff, IT and facilities own signal, and legal and executive leadership own signatures. An exercise is often the only place all three owners see the same gap at once.

A useful exercise follows three rules. Participants learn on the day which of them are unavailable, and those people leave the room. Phones and the shared drive are declared out of service. And the exercise ends with a real transaction in a live or test system, such as a payment approved by the out-of-area backup. Findings go to a named owner with a date, and the next exercise checks whether they were fixed.

Where staff, signal and signatures are thinnest

1. For each essential function, is at least one trained person outside our region, with working accounts and equipment already in place?
2. If cell service is down across the county, how do our system administrators sign in?
3. Are our delegations of authority current, with named people, dollar limits and time limits, and does anything need a board vote to fix?
4. Do we have standing emergency contracts and a paper logging method that would satisfy FEMA documentation requirements?
5. Who, by name, owns the whole picture across staff, signal and signatures?

With hindsight

Recovery in western North Carolina took many months and continued through 2025. The National Hurricane Center’s final report, first issued in March 2025 and revised as late as February 2026, called Helene the most devastating natural disaster in the region’s history. NOAA put total U.S. damage at about $78.7 billion, the seventh-costliest U.S. hurricane after adjusting for inflation.

Asheville restored drinkable water to its whole system on November 18, 2024, 53 days after the storm. That is a long time for a city to go without safe tap water, and it is a useful benchmark for planners who assume disruption will last days. Main roads took longer still. Interstate 40 through the Pigeon River gorge reopened with reduced lanes in March 2025, and its permanent rebuilding was expected to take years.

Revenue is part of continuity

The city also paused utility billing until the whole system was back to drinkable water. Its published approach is instructive. Water consumption was charged at zero for the period when the water was not drinkable. Sewer service, which had kept running, was billed at the regular rate. Flat fees for the missed two-month billing cycle were doubled on the next bill, and the city said it would not assess delinquent fees until after March 1, 2025.

Each of those choices was both humane and practical, and each needed a working billing system, meter readers and someone with authority to change the rules. Continuity covers revenue and records as well as service. A utility that cannot bill cannot easily fund its own repairs, and the systems and staff that handle billing need their own recovery plan, just as the pipes do.

The federal debate

Helene also fed a broader argument about federal disaster support. On January 24, 2025, the new administration created a council to review FEMA, citing the federal response to Helene. The order asked for a report comparing FEMA’s performance with state, local and private responses, including communications and electricity. It also asked whether FEMA should support state control of disaster relief rather than supplant it.

The council was continued by further executive orders in January and March 2026. The second extension ran until ten days after the council delivered its report, or May 29, 2026, whichever came first. The debate over FEMA’s future role continued through that period.

Whatever the outcome of that debate, it gave state and local agencies a further reason to plan for the first weeks of a disaster on their own resources. In practice that means the same three things Helene exposed: out-of-area staff who can pick up the work, independent communications that do not depend on the local cell network, and authority that does not depend on a single official being reachable.

What held up

The guidance was never the problem. FEMA had updated its Continuity Guidance Circular for nonfederal organizations in August 2024, a month before landfall, and its elements already covered succession, delegations, communications and devolution. What Helene tested was whether those elements had been turned into accounts, agreements and votes. The long recovery that followed confirmed that disruption is measured in weeks, and that revenue, records and authority need planning alongside facilities and data. The federal review’s own focus on communications and electricity pointed at the same shared dependencies that every local plan in the region had relied on.

These preparations are cheap in calm years and nearly impossible in the middle of a crisis. Training a backup takes months. Changing a delegation may take a board vote. Signing an emergency contract takes a procurement cycle. Nothing on that list can be done after the roads wash out. The agencies that will cope best with the next regional disaster are the ones doing that unglamorous work now, while it feels optional.