On the evening of December 11, 1995, an explosion and fire destroyed much of the Malden Mills textile complex on the line between Lawrence and Methuen, Massachusetts, and fire departments from well beyond both cities were called in on mutual aid. The fire is remembered mostly for what the owner did afterward. It also belongs to a category every radio administrator should study, which is the large mutual aid response arriving on a radio plan sized for ordinary days. This piece separates what the record supports from what retellings add, and works through what that kind of load does to a shared channel.

December 11, 1995: what the settled record says about the fire

Malden Mills Industries was a family-owned textile company best known for Polartec fleece. In 1995 its main plant was a complex of mill buildings straddling the line between Lawrence and Methuen in the Merrimack Valley of northeastern Massachusetts. On the evening of December 11, 1995, an explosion in the complex was followed by a fire that spread through several of its buildings. Workers on the evening shift were inside when it started, and a number of them were badly burned. Contemporary coverage and later accounts agree that nobody was killed, and that fact more than any other separates this fire from the others it gets confused with.

The weather that night appears in every account. The fire burned in severe cold with strong winds that drove it between buildings. I have seen specific wind speeds attached to the night in retellings, but I have not confirmed any of them against a weather record, so I leave the number out. Injury counts also vary. Figures in the low-to-mid thirties appear in news coverage and later summaries, and I have not been able to reconcile them against a primary source. Anyone who needs a firm number should start with the Massachusetts Department of Fire Services, which houses the State Fire Marshal’s office, and with contemporaneous regional newspaper coverage.

Most accounts place the initial explosion in the part of the complex that produced flock, a fabric made by applying short fibers to an adhesive-coated backing, and fine fiber dust of that kind is a recognized combustible dust hazard. That is as far as I will go on origin. The cause was the subject of investigation and later litigation, and I am not going to name one or rank the candidates. Readers who need the cause determination should look for the State Fire Marshal’s findings and the court record rather than a secondhand summary, including this one.

The fire is famous outside the fire service because of what the owner, Aaron Feuerstein, did afterward. He announced that he would keep paying the plant’s employees while the company rebuilt, and the story was covered nationally for years as an example of corporate responsibility. That story is accurate in its broad outline and well documented elsewhere. I mention it only because it has crowded the firefighting and communications side of the night almost entirely out of the popular record.

Which mill fire this was, and which ones it was not

Massachusetts had several large fires in old industrial buildings during the 1990s, and the one most often blended with Malden Mills is the Worcester Cold Storage and Warehouse fire of December 3, 1999. The two fires happened four years apart, each in early December and each in a large brick building from the state’s industrial past, in cities about an hour’s drive from each other. The outcomes were completely different. Six Worcester firefighters died after crews became lost inside the warehouse, a vacant building with almost no windows, and the investigations that followed, including a NIOSH firefighter fatality report, examined search, accountability and communications inside that structure. Those findings belong to Worcester. An account of Malden Mills that mentions firefighter deaths or crews lost inside a windowless building has absorbed details from the 1999 fire.

Lawrence also carries an older mill disaster in its history. The Pemberton Mill collapsed in January 1860 while workers were inside, and fire broke out in the wreckage during the rescue. Death toll estimates for Pemberton vary between sources, and the event was a nineteenth-century structural failure with no bearing on radio. It surfaces in retellings because writers reach for local history, and occasionally a summary blurs the two Lawrence-area mill disasters into one narrative about dangerous mills.

A quieter source of confusion is the sheer number of vacant mill fires across New England in that period, many of which drew large mutual aid responses of their own. A fire service anecdote about radio chaos “at the mill fire” may come from any of them. When I see a radio story attached to Malden Mills, I ask whether the teller names the date and the department, because a story that cannot be placed should not be used to teach this incident.

The radio record: what is documented and what is not

The angle I was asked to write is channel congestion during the mutual aid response, and I have to be straightforward about the record. Fire service retellings of Malden Mills sometimes describe the radio picture that night as overwhelmed, with too many companies on too few channels. I have not been able to tie that description to a published after-action report, a fire marshal’s finding or a recording I could review myself. The description may well be accurate, since it fits what large mutual aid responses of that era usually looked like, but fitting a pattern does not make it documented, so I treat it here as an account rather than a finding.

The scale of the response is better established. Coverage at the time described fire departments from many surrounding communities working the fire, and the complex sits close enough to the New Hampshire line that departments from both states were in the natural mutual aid pool. I have seen alarm counts and company totals quoted, but they vary from one retelling to the next and I could not confirm any of them, so I do not repeat a figure. Describing the response as dozens of companies is consistent with the coverage, and I would not go further than that without a dispatch log in hand.

Radio recordings from a 1995 incident at a municipal fire department may or may not still exist. Retention schedules for routine recordings were often short, and many departments of that period recorded the dispatch channel but not the fireground. If a recording or a written communications review survives, it would most likely be held by the Methuen or Lawrence fire departments or by the state. Anyone with access to a primary source on the radio traffic that night can reach me through the contact page, and I will correct this article against it. For the rest of this piece I describe how a shared channel behaves when a large mutual aid response arrives on a 1995-style radio plan, which is general mechanism, and I keep it separate from claims about this particular night.

Mutual aid radio in 1995: home channels and a few shared frequencies

In the mid-1990s a typical municipal fire department in eastern Massachusetts or southern New Hampshire ran conventional analog radio, most often in the VHF high band, on wideband channels. The department might have one channel for dispatch and a second for fireground operations, and a larger city might have a few more. Trunking existed in some places, but conventional systems built around a single repeater or a base station plus simplex were the norm for small and mid-sized departments. Each department’s home channel was licensed and configured for that department’s daily traffic, and the radio plan assumed that one department and perhaps a couple of neighbors would be talking at a time.

For mutual aid the region depended on a small number of shared frequencies. In much of the Northeast the best known of these is 154.280 MHz, which the fire service has long used as a mutual aid and fireground channel and which the National Interoperability Field Operations Guide now lists as VFIRE21 alongside other fire mutual aid channels. Check the current NIFOG for the exact channel list and the conditions on its use, since I am describing the arrangement in general terms. The point that matters for 1995 is that only a handful of these channels existed, and every visiting company that wanted to talk to anyone outside its own department had to use them.

The radios themselves set hard limits. Mobile radios in apparatus usually carried more channels than portables, and some portables still in service at the time held only a few channels. A visiting engine company could arrive with a mobile that had the mutual aid channel and portables that did not, which left its interior crew able to talk only on its home channel, to people miles away who could not see the fire. Even where every radio carried the shared channel, the home department’s own traffic and the visiting departments’ traffic back to their own dispatchers competed with the mutual aid traffic for a fixed number of frequencies over the same patch of ground.

How a shared channel saturates as companies keep arriving

A conventional channel carries one transmission at a time. When two radios key up together, an FM receiver usually locks onto whichever signal is stronger at its location, so different listeners may hear different talkers, or a receiver may produce a garbled mix of both. Neither talker knows the other was on the air. The site’s article on Tenerife covers what one such collision cost in aviation, and on a fireground the usual result is a lost request that nobody realizes was lost until someone asks why it was never answered.

Airtime is the resource that runs out. Every arriving company on a shared channel generates traffic before it does any work, starting with an arrival report and a request for an assignment, then the reply, and often a repeat because the first attempt was covered by someone else. If an exchange like that takes fifteen seconds and a dozen companies arrive within a few minutes, the check-ins alone consume three minutes of channel time, and that figure is my illustration rather than a measurement from any incident. During those minutes the interior crews and anyone with emergency traffic are waiting behind people announcing that they have arrived. The load grows with the number of units while the useful capacity of the channel shrinks, because collisions produce repeats and repeats produce more collisions.

Mutual aid adds an identification problem on top of the capacity problem. Most departments have an Engine 1, and on a shared channel a call for Engine 1 is ambiguous unless every unit prefixes its town. Visiting crews bring the procedures and habits of their home departments, including how long they wait before repeating a call and whether they still use ten-codes. The officer running the fire cannot see most of the companies talking to him and often does not know their staffing or equipment. Each of these adds seconds to every exchange, and those seconds multiply by the number of exchanges on the channel.

The usual human response makes the problem worse. When a talker cannot get through, he keys up again sooner and talks longer, and the officers on the channel begin answering the loudest or most urgent-sounding traffic first. In the recorded mutual aid traffic I reviewed for quality assurance over the years, the companies that went quiet were rarely the ones with nothing to say. More often they had given up trying to get a word in, and on the tape their silence was easy to mistake for an absence of problems.

Staging and divisions: keeping arrivals off the working channel

The Incident Command System handles channel overload mostly through organization rather than hardware. Arriving resources report to a staging area, where a staging area manager tracks them and releases them when operations asks for them. Tactical work is broken into divisions and groups, each with a supervisor who talks to his own crews and passes summarized traffic up to the operations chief or the incident commander. When the communications plan matches that structure, the interior crews of one division share a channel only with the other units in that division, and the incident commander talks to a few supervisors on a separate command channel instead of to every company on the scene. The ICS-205 is the form where that plan gets written down.

By 1995 ICS was well established in the fire service on paper, having grown out of the FIRESCOPE program in California in the 1970s, but its use at large structure fires varied a great deal from region to region, and many departments still ran big fires with the incident commander talking directly to every company. I am not claiming Malden Mills was run that way, because I have not seen a record of how that fireground was organized. The general point stands without it, which is that the radio capacity of a fireground depends mostly on the command structure. A dozen companies reporting to one officer on one channel will saturate it no matter how good the radios are, while the same dozen split among divisions on separate channels can work with ordinary equipment.

Mutual aid in much of New England has long relied on predetermined running assignments, the box or mutual aid card system, in which a dispatcher sends a set list of companies as each alarm level is struck. Those cards settle who comes and in what order. The communications counterpart is a card that settles which channel each alarm level reports to and who on scene answers it, and in my experience that half of the card is where plans are thinnest. A department can know exactly which five engines will arrive on the third alarm and still have no written answer to which channel those engines should call on when they get there.

What changed afterward, and what this fire cannot be credited with

The major changes in mutual aid communications since 1995 are well documented. In February 2003 President George W. Bush issued Homeland Security Presidential Directive 5, which directed the creation of the National Incident Management System and led to federal preparedness funding being tied to its adoption. Federal grant guidance in the mid-2000s pushed urban areas to write tactical interoperable communications plans. The national interoperability channels were given standardized names, and the National Interoperability Field Operations Guide gave field users a single reference for them. The Communications Unit Leader position gained a formal training course and a credentialing path, and the ICS-205 became an expected product on large incidents. The FCC’s narrowbanding deadline of January 1, 2013 then rebuilt many of the radio fleets those plans described, which the site covers in its own article.

I cannot point to a document that ties any of those changes to Malden Mills. The documented drivers are the September 11 attacks, the inquiries that followed them and the federal grant structure that grew out of that period, and the site’s articles on the World Trade Center and on the post-2001 directives cover that history. Malden Mills is part of the institutional memory of fire departments in northeastern Massachusetts and southern New Hampshire, and people who worked it may well have carried its lessons into their own regional planning. I have no record of that happening, though, and I will not write that the fire produced a specific plan, channel or procedure unless someone shows me the paper.

At the state level, Massachusetts has a statewide fire mobilization plan built around regional fire districts, and New Hampshire departments work through their own regional mutual aid compacts. I am not going to describe current channel assignments or the history of either arrangement, because I would be guessing at details that change. Verify current requirements and communications annexes with the Massachusetts Department of Fire Services, the New Hampshire Division of Fire Standards and Training and Emergency Medical Services, or your own state emergency management agency.

Your own large-incident recordings: finding the overload in advance

Any agency that has run a large mutual aid incident in the last few years already holds better evidence about its own channel capacity than a 1995 story can provide, because the recording is sitting on the logger. The useful exercise is to play back the busiest hour of the largest recent incident and measure it rather than listen for impressions. Count the transmissions on each channel, note the overlaps and the near-overlaps where one talker keyed within a second of another unkeying, mark every call that went unanswered on the first attempt, and tally how much of the working fireground channel went to arrivals and requests for assignment rather than tactical traffic.

Then lay the recording against the ICS-205 for that incident, if one was written. In my experience the gap between the plan and the tape usually sits in the first twenty minutes, before anyone has had time to build the structure the plan assumes. Mutual aid companies get dispatched to the original alarm channel because that is what the dispatch message told them, and staging gets its own channel only after the incident commander notices the working channel is jammed. That early sequence is the one worth fixing, because it falls in the period when interior crews are most exposed and the command structure is least developed.

Logger retention matters here as well. If recordings roll over in a short window, the largest incident of the year may already be gone by the time anyone thinks to review it. A large mutual aid incident should trigger a hold on the recordings for every channel involved, including the mutual aid frequencies, which some loggers do not capture at all. The site’s article on reviewing communications after the incident covers the review process itself, and the specific failure to guard against is a logger that never captured the mutual aid channels or overwrote them before the review was scheduled.

What to do at your agency

• The radio system administrator checks whether the logger records every mutual aid and fireground frequency that the agency’s units can transmit on, including the VFIRE channels programmed in apparatus mobiles, and writes a list of any channel that is not captured.
• The communications center supervisor pulls the recording of the largest mutual aid incident in the past year, counts the arrival reports and requests for assignment on the working fireground channel during the first twenty minutes, and presents that count at the next supervisors’ meeting already on the calendar.
• The chief officer responsible for mutual aid takes the existing running cards for the second alarm and above and has a staging channel assignment written onto each card beside the company list.
• The communications manager confirms that the CAD response plan or dispatch script for a second alarm and above announces the staging location and staging channel by name, and corrects every script where it does not.
• The radio shop or the agency’s COML compares the mutual aid zone in its own codeplug against the codeplug of one neighboring department that regularly responds into the jurisdiction, and records any channel that is missing or sits in a different position.
• The training officer reviews existing training material for any reference that attributes firefighter deaths or post-2001 reforms to Malden Mills, and corrects each one against the date and outcome described here.

Takeaways

• The Malden Mills fire began with an explosion on the evening of December 11, 1995, at a complex on the Lawrence and Methuen line, and contemporary coverage agrees that nobody was killed.
• Published injury counts for the fire vary in the low-to-mid thirties, and the cause was the subject of investigation and litigation, so neither should be taught from a secondhand summary.
• The Worcester Cold Storage fire of December 3, 1999, which killed six firefighters, and the Pemberton Mill collapse of 1860 are separate events whose details should never be attached to Malden Mills.
• Descriptions of overwhelmed radio channels at Malden Mills fit the pattern of large mutual aid responses of that era, but I could not tie them to a published after-action report or recording, so they remain an account rather than a finding.
• Mid-1990s mutual aid in the region depended on a handful of shared VHF fire mutual aid frequencies, and visiting companies whose portables lacked those channels could not talk to anyone on scene outside their own department.
• A shared channel fails under mutual aid load mainly because arrival traffic, collisions and ambiguous unit identifiers consume airtime, and a staging channel with its own manager removes most of the avoidable load.
• NIMS, the national interoperability channel names, NIFOG and the COML program trace to post-2001 directives and grant guidance, and no document I know of credits them to Malden Mills.
• An agency’s own logger recordings from its largest recent mutual aid incident are the best available evidence of where its channel plan will saturate, provided the logger captured the mutual aid channels and kept them.