In the summer of 1988, fire burned about 793,880 acres of Yellowstone National Park, roughly 36 percent of the park by the National Park Service’s own figures, and a large area of the national forests around it. The popular explanation is that the park simply let the fires burn. That explanation is a myth in its simple form, because several of the largest fires started outside the park and human-caused fires were fought from the first day. This guide covers the natural fire policy in place that year, the drought and wind that overwhelmed it, what the acreage figures mean, and the federal reviews that rewrote wildland fire policy by 1995.

A century of suppression and the 1972 natural fire plan

Yellowstone was established in 1872. The U.S. Army administered the park from 1886 until the National Park Service, created in 1916, took over the job, and both the Army and the Park Service treated wildfire as something to put out. That matched the practice of federal land agencies generally through most of the twentieth century. The Forest Service’s best-known statement of it was the 10 a.m. policy adopted in 1935, which directed that a fire be controlled by ten o’clock on the morning after it was reported. For decades the measure of success on federal land was how fast a fire went out and how small it stayed.

That thinking began to change in the 1960s. In 1963, a committee chaired by the biologist A. Starker Leopold delivered a report to Secretary of the Interior Stewart Udall on wildlife management in the national parks, and the Leopold Report, as it became known, argued that parks should be managed to preserve natural processes. Fire was one of the processes it discussed. In 1968 the Park Service revised its policy to recognize fire as a natural force in park ecosystems, and over the following years individual parks wrote fire management plans under that policy. Yellowstone adopted its plan in 1972, allowing some lightning-caused fires in designated backcountry zones to burn under defined conditions. The Forest Service was experimenting with similar programs in some wilderness areas during the same period, which matters later in this story.

The track record before 1988 looked reassuring. The Park Service’s published history of the 1988 fires states that from 1972 through 1987 the park allowed 235 lightning-caused fires to burn, and that together they covered about 34,000 acres. Most of them stayed small and went out on their own when weather changed or they ran out of fuel. Fifteen years of that experience gave park managers and the public a reasonable expectation that natural fires in Yellowstone behaved modestly, and the expectation was based on years that turned out not to include anything like 1988.

How a prescribed natural fire was supposed to work

The term in use at the time was prescribed natural fire. A fire started by lightning inside a designated zone was assessed against a prescription, meaning a set of conditions covering weather, fuel moisture, expected fire behavior and the risk to people and developments. If the fire fell within the prescription, managers monitored it rather than attacking it. If it threatened developed areas, moved outside its zone or burned under conditions outside the prescription, it was to be suppressed. The popular name, let-burn, made the program sound passive, though on paper every monitored fire carried a standing decision that could be reversed.

Human-caused fires were outside the program entirely. A fire started by a campfire, a cigarette, a vehicle or a power line was to be suppressed regardless of where it started or how it was behaving. Fires that started on the national forests surrounding the park were governed by the plans of those forests, not by Yellowstone’s plan, so a fire burning toward the park boundary from outside could be under a different set of rules from the fire it eventually joined.

The weakness lay in how the decisions added up. Each fire was judged against conditions at the time it was assessed, and a fire that was reasonably within prescription in late June could be very large by the time conditions moved outside it. Once a fire in heavy timber has grown for several weeks, reversing the decision to monitor it does not put it out, because the suppression resources needed to stop a large fire running in dry fuel under wind may not exist, or may not be able to work safely in front of it. The review that followed the 1988 season concluded, as its findings are commonly summarized, that the plans did not adequately account for extreme drought, which is the mechanism this section describes.

A wet spring, a record dry summer and the end of let-burn in July

The season began with no obvious warning. The Park Service’s history reports that precipitation in April and May of 1988 was above normal, which supported the expectation that the summer would resemble earlier years. Then the rain stopped. The Park Service describes June, July and August of 1988 as the driest summer in the park’s recorded history, and the fuels dried accordingly, from the grasses through the large dead logs on the forest floor that ordinarily hold moisture through most of a summer.

Lightning started fires in June, as it does most years. Several fires inside the park, including the Shoshone and Red fires in the south and the Fan fire in the northwest, were allowed to burn under the plan at first, and the Storm Creek fire, which started on national forest land north of the park, was handled under the Forest Service’s own natural fire plan. Through early July these fires grew more than fires in previous years had, and the acreage began to draw attention from the surrounding communities and from Washington.

By mid-July the park had stopped allowing new natural fires to burn, and around the third week of July it was attempting suppression on all fires in the park. Published accounts give somewhat different dates for these steps, with July 15 and July 21 both appearing, partly because the decision was made in stages, first for new starts and then for fires already burning. Anyone who needs the exact sequence should take it from the Park Service’s own history rather than from a retelling, including this one. By that point, however, the fires already burning were large enough that suppression could not contain them in the conditions that followed.

Weather drove what happened next. Dry cold fronts crossed the region repeatedly through August and September, bringing strong winds without useful rain. Under those winds, fires in lodgepole pine moved through the canopy as crown fires, and accounts describe embers starting spot fires a mile or more ahead of the main front, which made firelines and natural barriers such as rivers and roads unreliable. Crews could protect a developed area, but they could not stop a fire front moving that way.

Where the large fires started: national forests, a cigarette and a power line

The origins of the individual fires are the clearest refutation of the let-burn myth. The North Fork fire started on July 22, 1988, on the Targhee National Forest in Idaho, just west of the park boundary. It is attributed to a cigarette discarded by woodcutters, and as a human-caused fire it was suppressed from its first day. It nonetheless became the largest fire of the season, burning east into the park, through the area around Old Faithful and north toward Mammoth Hot Springs.

Other major fires also began outside the park. The Storm Creek fire was started by lightning on June 14 in the Absaroka-Beartooth Wilderness on the Custer National Forest, north of the park, and was initially managed under the Forest Service’s natural fire program before it was fought. In early September it threatened the gateway communities of Cooke City and Silver Gate. The Hellroaring fire is attributed to an outfitter’s camp on the Gallatin National Forest, so it too was human-caused. The Huck fire started on August 20 on the Bridger-Teton National Forest when a tree fell on a power line, and the lightning-caused Mink Creek fire started on the same forest and burned into the park. Inside the park, lightning-caused fires such as the Clover and Mist fires, which later merged, along with the Shoshone and Red fires, made up the part of the season that the natural fire program does bear on.

The suppression effort, once it was fully committed, was enormous. The Park Service’s history puts the total at about 25,000 people over the season, at a cost of roughly 120 million dollars, and those figures cover the whole greater Yellowstone effort rather than the park alone. Active-duty military units were assigned to fireline work alongside civilian crews. Several incident management teams ran fires that crossed the boundaries between the Park Service and multiple national forests, and accounts describe an area command being established over them so that resources could be allocated among competing fires. For a reader whose job is coordination, the setting deserves attention: two federal departments, several land units, state and local agencies and military forces were all working in the same smoke, with fires that did not respect any of their boundaries.

Readers who want the complete list of fires, their causes and their final sizes should consult the Park Service’s published history of the 1988 fires and the Forest Service’s records for the affected national forests. I have named only the fires whose origins I can state with confidence, and the list above is incomplete.

Black Saturday, Old Faithful and the snow that ended the run

The day that came to be called Black Saturday was August 20, 1988. A cold front brought sustained high winds across the region, and the Park Service’s history reports that more than 150,000 acres burned across the greater Yellowstone area on that single day. That figure exceeds four times the acreage of all the natural fires the park had allowed to burn in the previous fifteen years combined, by the Park Service’s own count, and it came weeks after the park had moved to full suppression.

On September 7, the North Fork fire reached the Old Faithful area. The Old Faithful Inn survived, which accounts credit to structure protection work including sprinklers and wetting of the building, although a number of cabins and other structures in the area were destroyed. The scene was broadcast nationally, and much of the public’s lasting picture of the 1988 fires comes from that day. Over the following days the North Fork fire threatened Mammoth Hot Springs, and the Storm Creek fire threatened Cooke City and Silver Gate, where residents were evacuated.

Weather ended the run. Snow and rain arrived around September 10 and 11 and slowed the fires substantially, though they continued to smolder in places until further snow in November. Most accounts of the season, including the Park Service’s, conclude that the change in weather, rather than suppression, stopped the fires from spreading. Crews did important work protecting communities and developed areas, and that protection is where suppression succeeded. The Park Service’s history reports two deaths connected with the fire effort, both outside the park: a firefighter struck by a falling tree, known as a snag, and a pilot whose aircraft crashed while supporting the fires.

The communication problem around the fires was as serious as the fire behavior. Residents of the gateway towns, businesses dependent on summer tourism, members of Congress and the national press all wanted to know why the fires had not been put out, and the answers they received often described the natural fire policy without explaining that the largest fires were under full suppression and still growing. Superintendent Robert Barbee and other officials faced sustained criticism, and the impression that the park had chosen to let itself burn hardened during those weeks, before the acreage was even mapped.

793,880 acres, 36 percent and the mosaic inside the perimeters

The claim that “one third or more” of Yellowstone burned in 1988 needs a number and a source behind it. The National Park Service gives the figure as about 793,880 acres within the park, which it describes as 36 percent of the park’s area. That is the figure I would cite, attributed to the Park Service. Press reports during and immediately after the fires sometimes gave larger fractions, and some retellings still say that half the park burned. I have not been able to tie those larger fractions to any authoritative mapping, and I would not repeat them.

For the greater Yellowstone area as a whole, including the surrounding national forests, published totals disagree. Accounts based on Park Service material commonly give a figure of about 1.2 million acres, while other published accounts give a figure closer to 1.4 million acres. I cannot reconcile the two from sources I trust. One possible explanation, which I cannot confirm, is that some totals count acreage inside the outer fire perimeters while others count only acreage that actually burned, and that different accounts include different surrounding lands. A planner citing either number should name the source and say which it is.

That distinction matters because the burned area was a mosaic rather than a uniform loss. Mapping after the fires classified burned land into categories that included canopy burn, where crown fire killed the trees outright, mixed burn, where surface fire killed some trees and spared others, and burned meadow and grassland. Inside the fire perimeters there were also substantial islands of forest that did not burn at all. The canopy burn areas were large and visually dramatic, and nobody should overcorrect to the point of calling the fires mild, but much of the land within the perimeters was not a total loss in the sense the television images suggested.

The ecological record since then has been studied closely. Lodgepole pine, the dominant tree across much of the park, includes trees with serotinous cones that are sealed by resin and open in the heat of a fire, and researchers documented dense lodgepole regeneration across much of the burned area in the years that followed. Fire historians, including William Romme and Don Despain, reconstructed Yellowstone’s fire history from tree rings and reported evidence of large fires in earlier centuries, notably in the early 1700s. Their work, as it is commonly summarized, suggests that infrequent, large, stand-replacing fires are part of the natural regime in these high-elevation lodgepole forests, which complicates the separate claim that decades of suppression alone set up the 1988 fires. That question is still discussed among fire ecologists, and readers should take it from their published work.

The 1988 review team and the conditions for bringing natural fire back

While the fires were still burning, the Secretaries of Agriculture and the Interior suspended the natural fire programs on the lands their departments managed, so that all fires would be suppressed, and they established a joint Fire Management Policy Review Team to examine the policy. The team reported its findings to the two secretaries in late 1988, and its report was published in 1989. The fires had also drawn congressional attention, and the political pressure at the time ran toward abandoning natural fire altogether.

The review did not abandon it. As its conclusions are commonly summarized, the team found that the objectives of allowing natural fire to play its role in parks and wilderness were sound, but that the fire management plans in place were not adequate for the conditions of 1988. It recommended that the plans be rewritten before natural fire programs resumed, with stronger prescriptions tied to measures of drought and fire danger, and with explicit consideration of whether suppression resources would be available if a monitored fire had to be fought. It also emphasized coordination between neighboring agencies, because fires in 1988 had crossed between Park Service and Forest Service land under different plans.

I have given the review’s conclusions at the level of a summary because I have not worked from the full text recently, and its recommendations are more detailed than any summary. Readers who need the specific recommendations should obtain the report itself through federal depository libraries or the agencies’ historical collections. The broad direction is well documented, though: the natural fire principle survived, and it came back attached to stricter conditions and a requirement that every plan account for the worst season rather than the average one.

Yellowstone later returned to managing some lightning-caused fires under a revised plan. Park Service and Forest Service units elsewhere did the same as their plans were approved, so that by the early 1990s the practice had resumed in a more constrained form. The national framework that holds it today, however, came from a second review prompted by a different fire.

South Canyon and the 1995 federal wildland fire policy

In July 1994, fourteen firefighters were killed on the South Canyon fire on Storm King Mountain near Glenwood Springs, Colorado. That fire and the severe 1994 season around it prompted the Departments of the Interior and Agriculture to undertake a joint review of federal wildland fire policy as a whole, a broader exercise than the 1988 review, which had focused on natural fire programs. The resulting Federal Wildland Fire Management Policy and Program Review issued its final report in December 1995, and the policy it set out became the common framework for the federal wildland fire agencies.

The 1995 policy placed firefighter and public safety first in every fire management activity, a priority that reads differently in light of South Canyon. It recognized wildland fire as an essential ecological process to be reintroduced into ecosystems where appropriate, which carried forward the principle behind Yellowstone’s 1972 plan. It required fire management plans for areas with burnable vegetation, so that every unit would have decided in advance how it would respond to a fire, and it called for interagency standardization so that neighboring agencies would not be operating under incompatible rules, a gap that 1988 had exposed. It also stated the federal role in the wildland urban interface: federal agencies would conduct wildland firefighting, fuels reduction, prevention and technical assistance, while structural fire suppression would remain the responsibility of tribal, state and local governments.

That last point is the part of the 1995 policy most relevant to a local agency. When a federal incident management team arrives to manage a wildfire burning toward a town, the protection of structures belongs to the local fire department and the jurisdiction behind it, under arrangements that should already be in a cooperative agreement. The policy was reviewed and updated in 2001, and later implementation guidance, including guidance issued in 2009, changed the terminology again so that a single fire can be managed for more than one objective. Readers should take the current terms and the present form of these principles from the National Interagency Fire Center and their state forestry agency rather than from the 1995 text alone.

Taken together, the two reviews explain what changed after Yellowstone. The idea that fire belongs in some landscapes survived the 1988 season, and the reviews surrounded it with requirements for plans that account for extreme conditions, for safety as the first consideration and for agencies that coordinate across the boundaries fires cross.

What to do at your agency

• The county emergency management coordinator in any jurisdiction bordering federal land asks the fire management officer of each adjoining national forest, park or Bureau of Land Management unit for a summary or map of its current fire management plan, and records in the wildfire annex which units share a boundary with the county and where lightning-caused fires may be managed for resource objectives.
• The fire chief or county fire coordinator writes one paragraph into the existing wildfire annex stating who directs structure protection when a federal incident management team is managing a wildfire in or near the county, citing the cooperative fire agreement by name, and confirms the current terms of that agreement with the state forestry agency.
• The communications center manager or radio system administrator confirms with the interagency dispatch center which channels a federal incident management team would assign to local cooperators, checks that the county’s mobile and portable radios are programmed and licensed or authorized to use them, and completes a test with the dispatch center this month.
• The public information officer adds a paragraph to the public information annex requiring that every published fire size state whether it is perimeter acreage or burned area, name the agency that produced it and give the time of the figure.
• The emergency management coordinator places on the agenda of the next scheduled county fire chiefs’ meeting a review of the drought and fire danger indicators the state forestry agency publishes, and the meeting records the indicator level at which the county will begin daily wildfire briefings.
• The training officer schedules one company-level session before the next fire season covering the Yellowstone 1988 and South Canyon 1994 cases, using the Park Service’s published history of the 1988 fires and the official South Canyon investigation report as the source documents.

Takeaways

• Yellowstone adopted a plan in 1972 allowing some lightning-caused fires in backcountry zones to burn under prescription, and the Park Service reports that 235 such fires burned about 34,000 acres from 1972 through 1987.
• Human-caused fires were always to be suppressed, and fires starting on surrounding national forests were governed by those forests’ plans rather than by the park’s plan.
• After a wet spring, the summer of 1988 was the driest in the park’s recorded history according to the Park Service, and the park moved to suppress all fires during July, with sources giving slightly different dates.
• The North Fork fire, the largest of the season, started on the Targhee National Forest from a discarded cigarette and was fought from its first day, so the claim that let-burn alone caused the 1988 fires is a myth.
• The Park Service gives the burned area as about 793,880 acres, or 36 percent of the park, while totals for the greater Yellowstone area range from about 1.2 to about 1.4 million acres depending on the source.
• The burned area was a mosaic of canopy burn, mixed burn, burned meadow and unburned islands rather than a uniform loss, and lodgepole pine regenerated across much of it.
• The 1988 to 1989 federal review kept natural fire but required stronger plans that account for extreme drought, and the December 1995 federal policy review that followed South Canyon made safety the first priority and required fire management plans and interagency standardization.
• The 1995 policy assigns structural fire suppression to tribal, state and local governments, so a local agency should know before the fire season how it will protect structures alongside a federal incident management team.