Every major multi-agency incident review in the United States since the 1970s has said some version of the same thing about radio interoperability, and the recommendations have been adopted, funded, and then written up again as an issue at the next incident. Even though the technology improved across those decades and gateway hardware got cheaper, interoperability stayed mostly a legal and organizational problem wearing an engineering costume, and the part that fails is almost never the radio. This piece walks the record, separates the three layers of the problem, and explains why the memorandum of understanding is the hardest document in the file.

The after-action record, from FIRESCOPE forward

The modern American incident management system exists because of a communications and coordination failure. After the 1970 wildfire siege in Southern California, Congress funded the U.S. Forest Service to develop what became FIRESCOPE, an interagency effort among federal, state and local fire agencies in the region, and the Incident Command System came out of that work in the 1970s. The problems FIRESCOPE was chartered to fix concerned not fire behavior but an inability to place resources from different agencies under one command structure, to speak a common terminology, and to talk to each other on the radio.

The 1991 East Bay Hills fire in Oakland and Berkeley produced the same finding in more expensive form. Twenty-five people died, and published counts of destroyed dwellings vary depending on whether apartment units are tallied separately from single-family homes, with figures generally running from roughly 2,800 to over 3,300. The U.S. Fire Administration’s technical report on that fire and the state review that followed both documented mutual aid companies arriving from across Northern California on radio systems that could not reach the incident’s command structure, alongside the well known hydrant thread incompatibility. Runners and face-to-face contact carried a large share of the tactical coordination.

The 9/11 Commission Report, published in 2004, examined the same category of failure at the World Trade Center. The Commission found that the evacuation order issued by FDNY chiefs in the North Tower was not received by many of the firefighters inside, and it described a situation in which information held by NYPD aviation about the condition of the South Tower did not reach the fire chiefs who needed it, because the two departments were operating separate command posts on separate radio systems. The Commission’s account also notes that command and coordination problems between those departments had been identified after the 1993 bombing of the same complex. The 2006 House select committee report on Hurricane Katrina, titled A Failure of Initiative, and the White House lessons-learned report issued the same month both put communications collapse near the top of their findings.

The record continues past the events people remember. The Marjory Stoneman Douglas High School Public Safety Commission, in its initial report published in January 2019, found that responding agencies at the February 2018 Parkland shooting were on separate radio systems and could not talk directly, and that the county system experienced capacity problems during the response. That is roughly forty-eight years after FIRESCOPE was funded to fix an identical class of problem, and the intervening decades include a national standard, a federal grant program, and a plan for every state.

Technical, procedural and governance interoperability

SAFECOM, the DHS-managed program that has published public safety communications guidance since the early 2000s, describes interoperability along five lanes in a tool called the Interoperability Continuum: governance, standard operating procedures, technology, training and exercises, and usage. The reason that model has survived twenty years of revisions is that it makes a distinction most agencies collapse, which is the difference between having a path and having a practice.

Technical interoperability means a radio signal can physically get from a member of Agency A to a member of Agency B with acceptable audio quality. This is the layer with vendors, budgets and demonstrations. The SAFECOM continuum lays out a maturity progression in this lane that starts with swapping radios between agencies, moves through gateway devices and shared conventional channels, and ends with a standards-based shared system that both agencies use every day. The progression is honest about the fact that a swapped radio is a legitimate interoperability method, and in my experience it is the one that works most reliably at three in the morning.

Procedural interoperability means both agencies know what to call the resource, who assigns it, what belongs on it, and what happens when it is done. Two agencies can be perfectly capable of talking on 8TAC91 and still fail if one of them uses ten-codes on it, if neither has decided who owns the talkgroup during a joint incident, or if the channel is labeled differently in the two codeplugs so that a dispatcher and a company officer are naming different things out loud. This is the layer that an ICS-205 documents and the layer that joint SOPs govern, and it is cheap to build and expensive to maintain because it decays with personnel turnover.

Governance interoperability means somebody has the legal authority to commit both agencies to the first two layers and to keep committing over budget cycles and elections. That includes who owns the shared infrastructure, who pays for its maintenance, who authorizes activation, whose records retention law applies to the recorded traffic, and what happens when the two agencies want the same resource at the same time. Nothing in the technical or procedural layer survives without this one, because a shared channel plan with no signed agreement behind it is a document that expires quietly when the person who negotiated it retires.

Shared channels: the capability nobody is monitoring

The United States has designated nationwide interoperability channels in every public safety band, with standardized names published in the National Interoperability Field Operations Guide, the NIFOG, which the DHS Office of Emergency Communications and its successor organizations have maintained and reissued for years. VHF has VCALL10 and the VTAC series along with VLAW31 and VMED28, UHF has UCALL40 and UTAC channels, 700 MHz and 800 MHz have their own calling and tactical sets. The FCC’s rules require these to be programmed and used according to the national naming and usage conventions, and the licensing path for most agencies runs through a regional planning committee or a statewide license.

The channels themselves are not the problem. Three separate failure modes take them out of service in practice. The first is that they were never programmed into a portion of the fleet, usually because a codeplug revision happened in one agency and not the other, or because a batch of radios came from a different vendor with a different template. The second is that they are programmed but placed in a zone nobody has ever navigated to, which under stress is functionally the same as not being programmed. The third, and the one that defeats otherwise well prepared agencies, is that a calling channel only works if someone on the other end is monitoring it, and almost nobody staffs a dispatch position to monitor VCALL10 continuously.

Simplex operation on these channels also carries a coverage reality that planners routinely gloss over. A portable-to-portable simplex contact on a VHF tactical channel in rolling terrain or inside a structure covers a fraction of the ground that a trunked system with simulcast infrastructure covers, so a channel plan that assigns two agencies to a shared simplex tactical for a wide-area incident has quietly assumed a coverage footprint that does not exist. Repeater pairs exist for several of the interoperability channels, which helps, but the repeater has to be installed, licensed, and known to both agencies before the incident.

Gateways, patches and ISSI: what a stopgap actually costs

An audio gateway takes the receive audio from one radio system and keys a transmitter on another, usually through a device with a bank of radio interface modules or through a console patch built by a dispatcher. It is the single most common interoperability method in the country because it requires no agreement about frequencies, no shared infrastructure, and no changes to anybody’s subscriber fleet. It is also, in SAFECOM’s own maturity framing, an early-stage method rather than a destination, and the reasons are worth listing plainly.

A patch consumes a talkpath on every system it touches, for as long as it exists, so patching two trunked systems together during a large incident is taking channel capacity out of both systems at the moment both are busiest. Unit identification does not survive the patch, so the receiving dispatcher sees the gateway rather than the calling unit, and the emergency button does not cross either, which means a mayday declared by a mutual aid company on a patched talkgroup may not alarm at the console that would act on it. Audio quality degrades because a vocoded signal gets converted to analog audio and re-encoded, and hang times and squelch tails on the two sides create a window where the first syllable of every transmission gets clipped. Doubling increases, because the users on each side cannot hear the other side’s busy indication.

The P25 Inter-RF Subsystem Interface, the ISSI, and its console counterpart the CSSI are the standards-based version of the same idea, and they do preserve unit ID and emergency signaling in ways an audio patch does not. They also require IP connectivity between the two systems, licensing that is usually a paid feature rather than a checkbox, agreement on talkgroup and unit ID numbering so that two systems do not collide, key management if either side is encrypted, and a written agreement about who may roam onto whose system and who pays for the airtime. The technical work on an ISSI link is measured in weeks, while the agreement behind it is measured in fiscal years.

Every one of these methods needs an owner during the incident. A patch that nobody tears down keeps consuming capacity and keeps cross-connecting traffic to units who no longer need it, which is how a fireground channel ends up carrying a law enforcement perimeter conversation. Building the patch and terminating the patch are both jobs, and on any incident large enough to need one, they belong to a communications unit position rather than to whichever dispatcher has a free hand.

The memorandum of understanding is the hard part

Agencies that have solved technical interoperability and still cannot use it are almost always stuck on an unsigned document. The terminology matters here, because a memorandum of understanding is generally a statement of intent that does not create enforceable obligations, while a memorandum of agreement or an intergovernmental agreement can commit money, assign liability and bind successors. Your county attorney will have a strong opinion about which instrument your state’s law supports, and that opinion should be obtained before the drafting rather than after, because a two-year negotiation that produces an unenforceable document is a common outcome.

The provisions that actually stall these agreements are predictable. Money comes first: who pays for the interconnection circuit, who pays for subscriber fees when Agency B’s radios are affiliated on Agency A’s system, who funds the maintenance contract on a gateway that was purchased with a grant, and what the cost recovery looks like when a shared resource is used for a routine incident rather than a declared emergency. Authority comes second: who may request activation, whether a shift supervisor can approve it or whether it takes a chief officer, and whether the request goes through a dispatch center or through a duty officer. Control comes third, meaning whose standard operating procedures govern conduct on a shared talkgroup, who has the authority to remove a user from it, and whose incident commander directs its use in a unified command.

Then come the provisions that lawyers add and operators forget. Recorded radio traffic on a shared resource is a record, and your state’s retention schedule and open records law will apply to it, so the agreement has to say which agency is the custodian. Encryption keys are property with security consequences, so the agreement has to say who generates them, who distributes them, how often they rotate, and what happens when a radio is lost. Liability and indemnification have to be addressed within whatever your state’s sovereign immunity and mutual aid statutes allow. Access to another agency’s system may implicate CJIS security policy obligations if data as well as voice is involved. Term, renewal, and termination provisions matter because an agreement that any party can exit on thirty days notice is not a foundation for a capital investment.

The reason all of this is harder than the hardware is that it requires two or more elected or appointed bodies to give up a small piece of autonomy at the same time, in public, with no incident in front of them. A radio system is a political asset as much as a technical one, and a county that owns the system generally prefers the leverage that ownership carries. The interoperability agreements I have watched get signed were the ones where a specific person with the authority to sign made it a personal priority and a deadline existed, and the ones that died were the ones circulated as a draft and left to find their own momentum.

Grant money bought boxes and skipped the governance

Federal attention to interoperability arrived in force after 2001, though the diagnosis predated it. The Public Safety Wireless Advisory Committee, chartered jointly by the FCC and NTIA, delivered a final report in 1996 that named interoperability as a fundamental deficiency and projected substantial additional spectrum needs for public safety, and the joint Justice and Defense Public Safety Wireless Network program published studies through the late 1990s reaching similar conclusions. Both are worth reading in the original rather than in summary, and both are findable through the FCC and NTIA document archives.

The money followed the attacks. NTIA and DHS jointly ran the Public Safety Interoperable Communications grant program, funded from digital television spectrum auction proceeds, which awarded just under a billion dollars to states and territories in a single round in 2007. Published summaries differ slightly on that total, depending on whether they count the full appropriation or the amount actually obligated, so check the figure with NTIA before citing it. That program required each state to produce a Statewide Communication Interoperability Plan, a SCIP, and the SCIP requirement is arguably the more durable legacy, because it forced every state to name a statewide interoperability coordinator and to convene a governing body. DHS also published the first National Emergency Communications Plan in 2008, with revisions in 2014 and 2019.

What the grant era did not do was fund the recurring costs. A gateway purchased with grant money in 2008 needed a maintenance agreement in 2011, a firmware path in 2014, and replacement around 2018, and none of those appeared in the grant. The dedicated interoperability grant programs of that era, including FEMA’s Interoperable Emergency Communications Grant Program, ran for a few years and ended, after which interoperability projects competed for funding inside broader homeland security grant categories against equipment with more visible constituencies. I have opened plenty of cabinets containing a working audio gateway that nobody had powered on since the acceptance test, sitting next to a binder with a draft MOU in it.

Why the same finding comes back every decade

Interoperability decays for structural reasons rather than because anyone stops caring. Personnel turnover is the largest of them. The knowledge of how to build the patch, which zone the mutual aid channels live in, and who to call at the neighboring county tends to concentrate in one or two people per agency, and when those people promote or retire the capability leaves with them unless it was written into a document that new members are actually trained on. A capability that exists only as institutional memory has a half-life of roughly one personnel cycle.

Codeplug drift is the second mechanism. Fleet programming changes for a hundred good reasons across a few years, and unless one person is responsible for verifying that the interoperability zone survives every revision in every model of radio the agency owns, some portion of the fleet will lose it. The problem compounds when the two agencies revise on different schedules, so that the channel is present on both sides but with different names, different squelch codes, or one side pointing at a repeater pair and the other at simplex.

The third mechanism is that interoperability is rarely used routinely, and capabilities that are not used routinely are not available under stress. The usage lane of the SAFECOM continuum makes this point explicitly, ranking daily use across a region above use only at planned events or major incidents. Agencies that share a talkgroup for automatic aid runs every week will have it working during a mass casualty incident, while agencies that reserve a shared resource for catastrophes will discover on the day of the catastrophe that a firmware update changed the channel order.

Finally, the after-action recommendations themselves tend to be written at a level of generality that no one can be held to. A finding that reads “agencies should improve interoperable communications” produces a purchase and a press release, while a finding that reads “the county and the city shall execute a signed agreement covering activation authority and cost sharing for the shared talkgroup by the end of the fiscal year, with the fire chief and the sheriff as responsible parties” produces an argument, then possibly a signed agreement. The second kind is rarer because it requires the review to name people.

Testing interoperability so the result means something

A monthly radio check on a mutual aid channel is worth doing and proves very little, because it is typically performed by the two people in each agency most likely to know how, from a fixed console, with unlimited time and no other traffic. That test verifies the RF path, and it says nothing about whether a company officer can find the channel, whether a dispatcher knows who to notify, or whether the agreement covering its use is still current.

A test that means something has four properties. It is performed by a randomly selected line member rather than the communications officer, because the question is whether the average user can execute. It starts from the radio’s normal operating position, so that the navigation to the interoperability zone is part of the test. It involves the other agency’s dispatch center answering, which is the step that most often reveals nobody is monitoring. And it is timed, because a capability that takes four minutes to establish is not available for a mayday.

Document what fails and route it to a body that meets anyway. Most counties have a public safety committee, a 911 board, a fire chiefs association meeting, or an EMA-hosted coordination meeting on a standing schedule, and a one-page test result presented at a meeting that already exists gets more traction than a new working group. If your region has a statewide interoperability coordinator, that office generally wants this data and can sometimes bring technical assistance or state contract resources that an individual agency cannot access on its own.

The other test worth running is a document test. Pull your mutual aid agreement with each neighboring agency, check the signature date and the termination clause, and confirm the named positions still exist. I have seen agreements signed by a chief who left in 2009, referencing a radio system that was decommissioned in 2016, still sitting in the binder as the governing document. The agreement is the part that fails silently, because nothing about it stops working until the day you need to rely on it.

What to do at your agency

• Have your training officer pull three portables at random from three different apparatus this month, hand them to line members who did not know the test was coming, and time how long it takes each to reach the national calling channel for your band and make contact with a neighboring agency’s dispatch center.
• Ask your radio system administrator or fleet programmer for the current codeplug template and verify in writing that the interoperability zone is present, correctly named per the NIFOG convention, and identical across every make and model of subscriber your agency owns.
• Pull every mutual aid and shared communications agreement in your files, check the signature date, the named positions, and the termination clause, and give your chief a one-page list of which ones reference people or systems that no longer exist.
• Put one item on the agenda of the county public safety or 911 committee that already meets: who is authorized to request a patch or gateway activation between the two largest systems in your county, and who is responsible for tearing it down.
• Ask the neighboring agency’s communications supervisor directly whether any position in their center monitors the national calling channel continuously, and write the actual answer into your own dispatch procedure rather than assuming someone is listening.
• Write one paragraph into your existing communications SOP covering the capacity cost of a console patch on your system, so that the dispatcher building it at two in the morning knows what it takes out of service.
• Identify the person at your state emergency management agency or statewide interoperability coordinator’s office who handles interoperability technical assistance, get the current SCIP, and confirm which regional governance body your agency is supposed to have a seat on.

Takeaways

• The Incident Command System itself came out of an interagency coordination and communications failure, since FIRESCOPE was funded after the 1970 Southern California wildfire siege, and the same class of finding has recurred in major incident reviews for the five decades since.
• The 9/11 Commission Report documented that the North Tower evacuation order did not reach many firefighters and that NYPD aviation observations did not reach FDNY chiefs operating from a separate command post, and it noted that coordination problems between those departments had been identified after the 1993 bombing.
• SAFECOM’s Interoperability Continuum treats governance, standard operating procedures, technology, training and exercises, and usage as five separate lanes, and agencies routinely advance far down the technology lane while remaining at the start of the other four.
• Nationwide interoperability channels exist in every public safety band with standardized names published in the NIFOG, and they fail in practice because they were never programmed into part of the fleet, because they are buried in a zone nobody navigates to, or because no dispatch position is staffed to monitor the calling channel.
• An audio gateway or console patch consumes a talkpath on both systems, does not carry unit identification or emergency button signaling across, degrades audio through re-encoding, and needs a named person to build it, monitor it, and tear it down.
• The P25 ISSI and CSSI preserve unit ID and emergency signaling that an audio patch drops, but they require IP connectivity, paid licensing, agreement on ID numbering, key management if either side is encrypted, and a signed agreement on roaming and cost.
• The provisions that stall interoperability agreements are money, activation authority, whose SOP governs a shared resource, records custody for recorded traffic, encryption key ownership, liability within your state’s immunity statutes, and the termination clause.
• NTIA and DHS awarded just under a billion dollars in the 2007 Public Safety Interoperable Communications round and required every state to produce a SCIP, but the grant era funded acquisition rather than the maintenance, licensing and replacement cycles that keep the capability alive.
• Interoperability decays through personnel turnover, codeplug drift, and lack of routine use, so a shared resource used weekly for automatic aid will work during a disaster while one reserved for disasters will not.