The policy answer to the communications failures of 11 September 2001 is written down in documents that carry dates and numbers, so it can be checked rather than recalled. A presidential directive in 2003 ordered a national incident management system, the system was issued in 2004 and has been revised twice, and a statute in 2012 settled the spectrum question the 9/11 Commission had raised eight years earlier. This piece, the last of four on the communications record of the response in New York, sets that record out so a reader can establish what their own region adopted and keeps current, and what it adopted once and filed.

The directive, the system, and the editions that followed

President George W. Bush issued Homeland Security Presidential Directive 5, titled Management of Domestic Incidents, on 28 February 2003, and the two things it set in motion are the doctrinal spine of everything that came afterwards. It directed the Secretary of Homeland Security to develop and administer a National Incident Management System, and it required the department to produce a National Response Plan, so a single directive is the origin of both the structure that agencies train to and the plan that governed how the federal government would come to an incident. The directive also named the Secretary of Homeland Security as the principal federal official for domestic incident management, which is the provision most often skipped in summaries even though it is the reason the system was issued by that department rather than by FEMA, which at that point sat inside it.

The system itself was released on 1 March 2004 by Homeland Security Secretary Tom Ridge, after a long period of coordination with federal agencies and with state and local organisations, and it has been revised twice since. FEMA published a revision in December 2008, by which point the agency had been given statutory responsibility for maintaining the system through the Post-Katrina Emergency Management Reform Act of 2006, and it published a third edition in October 2017 which supersedes the 2008 document. The third edition reorganised the material around three components, being resource management, command and coordination, and communications and information management, so if you are working from a course handout or a departmental policy that quotes the 2008 component list you are quoting a superseded edition. A reader citing the system in a plan should take the edition and date from the document on FEMA’s site rather than from a training slide, because the slides in circulation are of several different vintages.

The response plan had a shorter and more complicated life. The National Response Plan came into effect in December 2004, replacing the older Federal Response Plan, and a notice of change was issued in 2006 that incorporated preliminary lessons from the 2005 hurricane season. The plan was then superseded by the first National Response Framework in 2008, with published accounts giving 22 March 2008 as the date the older plan ceased to apply, and the framework’s own introduction describes the change as streamlining the guidance and integrating lessons learned from Hurricane Katrina and other incidents. Further editions of the framework followed, with a second edition in 2013, a third in 2016 and a fourth in 2019, and because a further revision may have issued since I last checked the document, confirm the current edition on FEMA’s site before you cite an edition number.

The difference in function between the two documents is worth being exact about, because they are frequently treated as one thing. The incident management system describes how any incident is organised and managed at any level of government, including command structures, common terminology, resource typing and qualification, and it applies to a single engine company on a dumpster fire as well as to a national catastrophe. The response framework describes how the response is coordinated above the local level, what the federal roles are, and how support arrives, so it is the document that tells a county emergency manager what can be requested and through which channel rather than the document that tells an incident commander how to build a command structure. One further point of chronology matters for the next two sections, which is that the directive of February 2003 and the first issue of the system in March 2004 both preceded the publication of the 9/11 Commission Report in July 2004, so the national structure was not created in response to the Commission’s recommendation about incident command.

What the directive actually required of a local agency

The mechanism that moved this from federal doctrine into local practice was money. The directive stated that beginning in fiscal year 2005, federal departments and agencies were to make adoption of the incident management system a requirement, to the extent permitted by law, for providing federal preparedness assistance through grants, contracts and other activities, and that single sentence is why a volunteer fire department in a county with no federal grant history suddenly had a training requirement. States passed the condition down, adoption was reported upward through annual compliance reporting, and the practical content of compliance for most agencies became a set of completed courses, an adoption instrument signed by a governing body, and the use of the national terminology in written plans.

The training requirement is the part that every reader of this article has personally experienced. The introductory course on the incident command system and the two awareness courses on the national system and the response framework became the baseline for responders, with the intermediate and advanced command courses expected of people who would function as command and general staff, and that curriculum has been stable enough that a certificate earned fifteen years ago still names a course that exists. What arrived alongside it was common terminology, meaning common position titles, a common written form for an incident radio communications plan, and the expectation that a communications function on a large incident is held by a qualified position rather than by whoever owns the radio cache. The site has separate pieces on building that plan and on the communications unit leader role for small departments, and I am not going to repeat either here beyond noting that both exist because the national system named the position and the form.

What the requirement did not produce is the harder half, and I want to state it as a documented property of the compliance scheme rather than as an accusation. Compliance was reported by the jurisdiction about itself, adoption was typically a one-time act evidenced by a resolution or an executive order, and no part of the scheme required a jurisdiction to demonstrate that agencies which do not share a radio system had successfully passed information between them under time pressure. In the agencies I worked with as a communications consultant between 2001 and 2024, the adoption instrument was usually findable within a day if anyone knew which clerk held it, and the date on it was very often within eighteen months of the grant condition taking effect with no revision since, which tells you that the structure was adopted and then left alone.

Two later documents sit on top of this layer and are worth knowing by name so that a reader can look them up. Presidential Policy Directive 8, titled National Preparedness, was issued in 2011 and replaced the earlier preparedness directive from 2003, and it is the origin of the national preparedness goal and the system of core capabilities that state and local plans are now written against. Separately, the Post-Katrina Emergency Management Reform Act of 2006 created an emergency communications office within the Department of Homeland Security and required a national emergency communications plan, the first of which was published in 2008, with later versions now maintained by the Cybersecurity and Infrastructure Security Agency, so check that agency’s site for the edition in force rather than relying on a figure or a target quoted from an older version.

What the Commission recommended about command and about spectrum

The 9/11 Commission Report was published in 2004 by the National Commission on Terrorist Attacks Upon the United States, and its recommendations on emergency response sit in the twelfth chapter, in the section dealing with protecting against and preparing for terrorist attacks, rather than in the ninth chapter which contains the narrative and analysis of the response in New York. The United States Government Accountability Office published a summary of the Commission’s recommendations that cites page 397 for the spectrum recommendation, and that page reference is the most reliable way I know to send somebody to the actual text. The distinction between the chapters matters because the Commission’s findings about what happened on the day and its recommendations about what to do next are written in different registers and are separated by about eighty pages.

On command, the Commission’s text is short and specific. It states that “Emergency response agencies nationwide should adopt the Incident Command System (ICS)”, adds that when multiple agencies or multiple jurisdictions are involved they should adopt a unified command, and characterises both as proven frameworks for emergency response. The Commission’s executive summary carries the funding mechanism as a separate bullet, recommending that Congress “Make homeland security funding contingent on the adoption of an incident command system”, with the same bullet calling for a regional approach, and a neighbouring bullet recommending that federal funding for emergency preparedness be based solely on risks and vulnerabilities.

On spectrum, the recommendation reads that “Congress should support pending legislation which provides for the expedited and increased assignment of radio spectrum for public safety purposes”, and it continues with a second element that almost never survives the paraphrase, namely that high-risk urban areas such as New York City and Washington should establish signal corps units to ensure communications connectivity among civilian authorities, local first responders and the National Guard, with federal funding of such units given high priority by Congress. Read the sentence again and notice what is absent from it. The text names no band, specifies no quantity of spectrum, proposes no network and contemplates no federal authority, because the Commission asked Congress to support legislation that was already pending and to move faster on assignments the Federal Communications Commission would make.

This is the point in the series where I am most often asked to confirm something the text does not say. The Commission did not recommend the network that eventually got built, and anybody who tells a class that the 9/11 Commission called for a nationwide public safety broadband network is describing what happened afterwards rather than what the report says. Treat that as a myth and label it as one in the same breath, because the interesting fact is the opposite of the myth, which is that a recommendation of one paragraph about assigning spectrum turned into a ten year legislative argument with an outcome the recommendation did not anticipate.

Tracking each recommendation to what became of it

The incident command recommendation was already being implemented when it was published, which is the cleanest case in the set. The directive of February 2003 had ordered the system, the system was issued in March 2004, and the grant condition taking effect in fiscal year 2005 supplied exactly the funding contingency the Commission’s executive summary asked for, so the recommendation and the policy converged rather than one causing the other. If you want a defensible way to describe this in a class, say that the Commission endorsed and reinforced a structure that was already in issue, and that the funding contingency it recommended was already written into the directive.

The recommendation on basing preparedness funding solely on risk was implemented in part and remains contested. The Implementing Recommendations of the 9/11 Commission Act of 2007 revised how the principal homeland security grant programmes allocate money among states, and statutory minimum shares for each state have continued to exist in some form since, which is the reason the argument about risk-based allocation recurs in every appropriations cycle. I am not going to print a current percentage or a current allocation formula, because those change, and the place to get them is the statute as amended together with the current notice of funding opportunity from the Department of Homeland Security for the grant programme you actually receive.

The signal corps element of the spectrum recommendation went nowhere that I can document. I have not been able to confirm any jurisdiction that established a signal corps unit in the form the Commission described, and the phrase dropped out of the policy conversation almost immediately, with the spectrum sentence carrying the whole weight of that recommendation in every subsequent summary. If a reader can point me to a unit created under that recommendation I would like to see the documentation, and until then I treat it as a recommendation that was published and not adopted.

The spectrum recommendation itself moved slowly and in stages. The Intelligence Reform and Terrorism Prevention Act of 2004 included communications provisions and, according to a Congressional Research Service report on emergency communications legislation prepared for the 110th Congress, a sense of Congress provision stating that the following Congress should pass legislation supporting the Commission’s recommendation to expedite the release of spectrum. Progress on the substance came in the Deficit Reduction Act of 2005, which a Congressional Research Service report on spectrum auctions describes as setting a definite date of 17 February 2009 for the release of the 700 MHz spectrum then held by broadcasters, and the argument about who would get the contested portion of that spectrum then ran until 2012. The former commissioners, working as the 9/11 Public Discourse Project, issued a final report card on the recommendations in December 2005 that graded the spectrum recommendation a failure, and if you cite the grades, take each one from the report card itself rather than from a news summary of it.

The television transition and the block public safety wanted

The spectrum argument is the longest running consequence of 2001 in this whole record, and it is badly understood because the sequence gets compressed. Congress had already designated a block of the upper 700 MHz band for public safety in the 1990s, when television channels 60 to 69 were reallocated, and a Congressional Research Service report on the first responder network records the problem with that designation plainly, which is that the statute required broadcasters to transition off the spectrum but provided an indefinite timeline for doing so, with the result that the spectrum was not immediately available to public safety. The Commission was writing into that condition in 2004, when public safety held an allocation on paper in a band still occupied by television stations and the assignment could not be expedited by anybody until there was a date certain for the broadcasters to leave.

The Deficit Reduction Act of 2005 supplied that date, setting 17 February 2009 as the definite date for the release of the 700 MHz spectrum held by broadcasters, according to the Congressional Research Service’s published work on spectrum auctions, which converted an indefinite transition into a deadline and made the subsequent auction possible. Congress later moved the consumer-facing transition date to 12 June 2009 in legislation passed early that year, and if you need to cite the statute that made that change by name and public law number, verify it against the congressional record rather than taking it from this article, because I have not been able to re-verify the citation while writing.

The block at the centre of the argument has a precise identity, and using the right frequencies is the difference between a checkable statement and radio folklore. The Federal Communications Commission’s implementation document published in the Federal Register in October 2012 identifies the D Block as 758 to 763 MHz paired with 788 to 793 MHz, which is 10 megahertz of spectrum, and identifies the existing public safety broadband spectrum as 763 to 768 MHz paired with 793 to 798 MHz, another 10 megahertz, with a guard band at 768 to 769 MHz paired with 798 to 799 MHz between them. The public safety broadband portion had been licensed to a single national licensee, the Public Safety Spectrum Trust, and the D Block sat directly adjacent to it, which is the entire reason the D Block and no other block was the one public safety organisations wanted.

The arrangement the Federal Communications Commission attached to the D Block was neither a gift to public safety nor a straightforward auction. The block was offered on the condition that the winning commercial bidder reach a network-sharing agreement with the public safety licensee and assume the cost and responsibility of building a shared network, an arrangement the Congressional Research Service described as a new business model that departed from past auction policy. That model is the thing that failed, and its failure is what produced four more years of legislative argument, so it is worth separating in your own mind from the question of whether the spectrum was ever going to be made available at all.

The auction that did not clear, and the statute that ended the argument

The 700 MHz auction, officially Auction 73, opened in January 2008 and closed on 18 March 2008, and the Congressional Research Service reports that it grossed 19,592,420,000 dollars, which was well above the congressional revenue estimate that trade coverage at the time put at 10.2 billion dollars. Four of the five blocks met their aggregate reserve prices, while the D Block did not, and the rule the Federal Communications Commission had set in advance was that if the results in a block failed to satisfy the relevant aggregate reserve price, none of the licences in that block would be assigned on the basis of the auction, with a contingent subsequent auction available under previously announced terms.

The bidding on the D Block is a short story in the record. Urgent Communications reported on 24 January 2008 that the first round produced a single bid of 472 million dollars for the nationwide D Block licence, which was the minimum acceptable bid set by the Federal Communications Commission against a reserve price of 1.3 billion dollars, and the same report notes that a startup which had promoted the public-private partnership model, Frontline Wireless, had closed about three weeks before the auction began. Under the anonymous bidding rules the identity of the bidder was not disclosed until the auction finished. The block went unassigned, the sharing arrangement with the public safety licensee was never tested in practice, and the Federal Communications Commission was left to evaluate its options for a band it had already reallocated for a purpose that the auction had not delivered.

What followed was a four year argument with two positions in it. One position held that the D Block should be re-auctioned commercially with revised conditions, and the other held that the block should be allocated directly to public safety and combined with the adjacent broadband spectrum to give a single national licensee 20 megahertz to work with. The major national public safety organisations campaigned for direct allocation, the dispute ran through two Congresses, and the resolution came attached to a payroll tax bill rather than to a homeland security bill.

The statute is the Middle Class Tax Relief and Job Creation Act of 2012, Public Law 112-96, 126 Statutes at Large 156, enacted on 22 February 2012, and its sixth title carries the public safety communications and spectrum provisions. That title establishes the First Responder Network Authority, and the Federal Communications Commission’s implementation document of October 2012 records the effect on the spectrum, which is that the Bureau reallocated the D Block at 758 to 763 and 788 to 793 MHz for public safety services, deleted the rules governing the earlier public safety broadband licence, and granted the new authority a licence covering the public safety broadband spectrum, the guard band and the D Block, so that the 758 to 769 MHz and 788 to 799 MHz bands are allocated for the authority’s nationwide network. The same title contained the relocation mandate for the 470 to 512 MHz band that was later repealed, which this site covers separately.

This article stops at that licence grant, and the boundary is deliberate. How the network was built, what priority and pre-emption actually mean on it, what a subscription involves and whether it suits a particular agency are all live operational questions with a separate article on this site devoted to them, and none of them belong in a policy record. The endpoint of the ten year spectrum argument is a statute enacted on 22 February 2012, an authority established under its sixth title, and a licence with defined band edges granted later that year, and anything past that point is a procurement question for an individual agency rather than a consequence of 2001 that a reader can audit.

What did not change, and how you can tell

Radio coverage inside large structures is governed by the same physics it was governed by in 2001, and no statute, directive or system edition has repealed any part of it. Building materials attenuate signals, reinforced concrete and metallised glazing attenuate them severely, a portable radio transmitting from a stairwell or a basement has a fraction of a watt working against several floors of structure, and the failure is worst precisely where crews are working when the incident is worst. Part two of this series dealt with what the investigations established about radios inside the towers, and the relevant point here is that the engineering answer to the problem, which is an in-building system that repeats or distributes signal through the structure, is installed unevenly and maintained worse. Model fire code provisions on emergency responder radio coverage exist and have been adopted state by state, with real variation in whether they apply to existing buildings, and the separate article on this site about in-building coverage deals with what the code may require in a given jurisdiction.

The maintenance problem is the one I would put money on finding in any jurisdiction. An in-building amplifier or distributed antenna system is commissioned when the building is new, it is signed off by somebody who is not in the fire service, and then it is subject to a renovation that removes an antenna, a battery that ages out, a power supply nobody tests, or a system retune when the radio system migrates to a new band or a new mode. Nothing about that failure announces itself, because the building looks the same and the radio works fine in the lobby where anybody would test it, and the way it is found is a walk test with a portable in the stairwell and the lowest level, which is roughly a half day of work that almost nobody schedules and that produces a written record of dead spots nobody previously had.

Command across agencies remains local and, in practice, voluntary. The national system supplies a structure and a vocabulary, and it does not create authority, because who commands a multi-agency incident in your jurisdiction is determined by state law, by county and municipal plans, and by mutual aid agreements that agencies signed with each other. That means the arrangement still depends on whether the people who would be standing at the same table have worked together before, and that dependence is not a defect in the national system so much as a description of what a national system can and cannot do from Washington. A structure can be handed down, and the habit of building one plan with a neighbouring agency has to be built locally at local cost, which is why it varies so much between adjacent counties with identical paperwork.

Interoperability still appears in published after action reports, and I am going to name only the ones I can cite by title and year rather than gesture at a trend. The United States Department of Justice, through its Office of Community Oriented Policing Services, published Critical Incident Review: Active Shooter at Robb Elementary School in January 2024, and communications problems are a substantial part of what that review documents about the response of the many agencies that arrived. The Marjory Stoneman Douglas High School Public Safety Commission delivered its Initial Report in January 2019 to the Governor and Legislature of Florida, and radio performance during the response is one of its subjects, including the behaviour of the county radio system under load and the position of agencies that arrived on systems of their own, which is material a reader should read in the report itself rather than in my summary of it. The After Action Report for the Response to the 2013 Boston Marathon Bombings, published in 2014 and available through the Massachusetts Emergency Management Agency, is older than the other two and I include it because it is a multi-agency example with a published section on communications.

Three reports do not establish a national pattern, and I am not going to claim one from them. What they establish is that the finding survived the doctrinal fix in at least those responses, which is enough to justify the question the action section below asks you to put to your own last three reports. If your own after action reports contain no communications finding at all, that is worth examining before you conclude it is good news, because the most common reason a report contains no communications finding is that nobody was assigned to evaluate communications.

Why one half was fixable and the other half was not

The part of the record that got fixed after 2001 is the part that could be fixed by writing a structure down and funding a procurement, and each of those fixes had an owner with the authority to act alone. A President can sign a directive, a department can issue a system and revise it in 2008 and again in 2017, Congress can set a transition date and pass a statute allocating spectrum, and a regulator can grant a licence with defined band edges. Every item in the first six sections of this article has a date, an issuing body and a document you can download, which is exactly why it is checkable, and that checkability is itself a feature of problems with single owners.

The part that did not get fixed has no owner of that kind. Testing radio coverage inside a privately owned high-rise requires a building owner, a fire marshal, a radio technician and two crews with portables, and no one of those four can compel the other three. Running an exercise designed to fail requires several agencies to give up an operational day, to spend money on a facilitator and evaluators, and to be seen failing in front of neighbours they will have to work with next week, and no federal grant condition has ever been written that can make a group of chiefs want that. The command arrangement for a multi-agency incident requires an agreement that an elected body has to approve, with all the local politics that carries, which is a different kind of work from completing a course.

That asymmetry is a fact about how institutions change rather than a criticism of anybody in this business. Adoption is a single act with a date, it can be certified by the jurisdiction about itself, and it stays adopted without further effort, whereas the capability to pass information between agencies under load decays as people retire, systems migrate and buildings are renovated, and it has to be re-established on a cycle by people who have no incident in front of them to justify the cost. The evidence for the asymmetry is attached to it in both directions, which is that the doctrinal items are dated and verifiable while the practice items keep reappearing in the after action reports named above.

The useful conclusion for a working agency is that both halves are measured with dates, and that the dates are not hard to collect. On the doctrinal side the dates are the one on your adoption instrument, the edition of the national system your written policy cites, and the revision date of your state’s interoperability plan, all of which somebody in your building can produce this week. On the practice side they are the date of the last multi-agency exercise in your region that carried a communications objective capable of being failed, the date of the last walk test with a portable in your largest occupancies, and the date the last communications corrective action out of an after action report was actually closed rather than assigned. A reader who collects those six dates has a more honest picture of their own region’s position since 2001 than any compliance report will provide, and the gap between the two sets of dates is usually where the work is. This is the fourth and last part of the series, and if you use any of it in a class or a plan, cite the earlier parts for the findings about the response itself and cite this part only for the policy record, because the documents named here are the ones that carry dates you can check.

What to do at your agency

• Ask your emergency manager or the municipal clerk for the resolution, ordinance or executive order by which your jurisdiction adopted the national incident management system, read the date on it, and record that date at the top of your communications annex so that every future reader knows how old the adoption is.
• Have whoever maintains your policy manual compare the edition of the national incident management system your policies cite against the current edition on FEMA’s site, and if your documents still reference the 2008 revision, note the discrepancy in writing for the next policy review.
• Have your training officer print the current roster of personnel holding the intermediate and advanced incident command courses, set it beside the list of people who would be command and general staff on a night shift, and write down the gap as a number rather than an impression.
• Ask your statewide interoperability coordinator for the current statewide communication interoperability plan and its revision date, then confirm that your agency’s systems, talkgroups and mutual aid channels appear in it correctly and send the corrections in writing if they do not.
• Find the date of the last multi-agency exercise in your region and read its written objectives, and if no objective asked whether a specific piece of information crossed from one agency’s system to another and how long it took, put that objective on the agenda of the exercise design or chiefs meeting that already meets in your area.
• Pull your last three after action reports, index every communications finding on one page with the name of the person who owned the corrective action and whether it closed, and bring that page to your next staff meeting rather than filing it.
• Pick the two largest occupancies in your response area that have in-building radio systems, obtain the most recent test record from the building owner or the fire marshal, and have two crews walk a portable test this month in the stairwell and the lowest level, writing down each dead spot with the date, the radio model and the talkgroup used.

Takeaways

• Homeland Security Presidential Directive 5, Management of Domestic Incidents, was issued by President George W. Bush on 28 February 2003, and it directed both the national incident management system and the national response plan.
• The national incident management system was released on 1 March 2004 by Homeland Security Secretary Tom Ridge, revised by FEMA in December 2008 and issued in a third edition in October 2017, while the National Response Plan that took effect in December 2004 was superseded by the first National Response Framework in 2008 with later editions in 2013, 2016 and 2019, so confirm the current edition before citing an edition number.
• The directive, not the 9/11 Commission, created the requirement that federal preparedness assistance be conditioned on adoption of the system beginning in fiscal year 2005, and the first issue of the system in March 2004 preceded the Commission’s report of July 2004.
• The Commission’s recommendations on emergency response sit in the twelfth chapter of The 9/11 Commission Report, with the Government Accountability Office’s published summary citing page 397 for the spectrum recommendation, which asked Congress to support pending legislation for expedited and increased assignment of radio spectrum and also called for signal corps units in high-risk urban areas.
• The Commission named no band, no quantity and no network, so the claim that it called for a nationwide public safety broadband network is a myth and should be labelled as one wherever it is repeated.
• Auction 73 closed on 18 March 2008 having grossed 19,592,420,000 dollars according to the Congressional Research Service, and the D Block drew a single first-round bid of 472 million dollars against a 1.3 billion dollar reserve as reported by Urgent Communications in January 2008, so the block went unassigned and the public-private partnership model was never built.
• The Middle Class Tax Relief and Job Creation Act of 2012, Public Law 112-96, enacted 22 February 2012, established the First Responder Network Authority, and the Federal Communications Commission’s October 2012 implementation document records the reallocation of the D Block and the grant of a licence covering 758 to 769 MHz and 788 to 799 MHz to that authority.
• Interoperability findings outlived the doctrinal fix in published reviews including Critical Incident Review: Active Shooter at Robb Elementary School of January 2024 and the Marjory Stoneman Douglas High School Public Safety Commission Initial Report of January 2019, and the honest local test is what your own last three after action reports say.