Zone” newsletter, July, 2004
Lessons Learned —
author, date unknown
One-Year Anniversary Letter
by Kelly Close, FBAN
Declaration on Cramer
Redactions, by James Furnish, April, 2005
FSEEE v. USFS, FOIA
Civil Lawsuit Order,
Request to USFS, December, 2005
FOIA Appeal to USFS,
Management Evaluation Report
Investigation Team Information
Synopsis of the
Cramer Fire Accident Investigation
(facts 1 - 57)
(facts 58 - 201)
(facts 203 - 237)
Resources on the Fire
Cramer Fire Timeline
Fire Behavior and Weather
Equipment Found at H-2 and the Fatalities Site
Fire Policy, Directives, and Guides
OIG FOIA Response,
2nd FOIA Request to OIG,
2nd OIG FOIA Response,
August, 2006, (1.4 mb, Adobe .pdf file)
OSHA Cramer Fire Briefing Paper
• Summary and ToC
• Sections I-IV
• Sections V-VII
• Section VIII
OSHA South Canyon Fire
Letter to District
Ranger, June 19, 2003
OSHA Investigation Guidelines
OSHA News Release
• OSHA Citation 1
• OSHA Citation
OSHA FOIA Letter
Adobe PDF and Microsoft Word versions of documents related to
the Cramer Fire can be downloaded from the U.S.
Forest Service website.
Citations for U.S. Forest Service
Salmon-Challis National Forest
Intermountain Region 4
Six Month Issuance Date — January 22, 2004
V. Accident Description
On Saturday, July 19, 2003, a lighting strike ignited a wildfire in steep,
rugged terrain undergoing extremely dry drought-like conditions. The fire
was located near Cache Bar in the Cramer Creek drainage on the north side
of the Salmon River about 50 miles northwest of Salmon, 10. Late afternoon,
Sunday, July 20, Long Tom Lookout first reported smoke from the fire.
Monday, July 21, 2003, the fire officially transitioned from a type 4
to a type 3 command structure which was operated out of the Cove Creek
Helibase, approximately 7 miles up the Salmon River from the fire and
without view of the fire area. By approximately 4:30 pm, the fire had
been steadily increasing in size to approximately 60 acres since reported
24 hours prior. Over the next several hours the fire increased in size
to approximately 200 acres and remained active until about 3:00 a.m. the
By Tuesday, July 22, 2003, approximately 70 to 80 personnel were working
to contain the fire as a Type 3 incident. Most of the crews were working
below and to the east side of the fire to construct fire-line in the draw
near Cramer Creek. The previous night the Incident Commander had requested
two additional Type-2 medium helicopters to use on the fire that could
not be provided due to limited fire resources.
At around 9:40 a.m., two helitack crew personnel had rappelled from a
helicopter onto a ridge line above and on the north-west side of the main
fire to clear a two- way helispot (H-2) in order for a helicopter to shuttle
fireline ground crews to this area. This plan had been decided and agreed
upon by the Incident Commander and the acting helicopter foreman (rappel
spotter, who later became the helibase manager) during a morning helicopter
recon flight. The provision of a lookout
had been discussed, however no lookout was provided for the rappellers.
Due to vegetation and changing slopes, the rappellers could not see the
active fire from H-2, nor could they see more than approximately 150 to
200 feet down slope in either direction from the ridgeline upon which
they were working (See photos from H-2).
For some reason the job took longer than expected, and the rappellers
repeatedly stated to the helibase radio operator that more time was needed.
It was originally estimated that clearing H-2 would take approximately
1 hour. The rappellers remained working at H-2 for more than 5 hours.
Upon inspection of the helispot, it appears that 15 to 20 larger diameter
(15 to 30 inches across) ponderosa pine had been felled, requiring an
undercut. In addition, a very large number of small diameter trees (less
than 4 inches diameter) had also been felled.
As often occurs, fire activity increased into the afternoon due to increasing
temperatures, decreasing relative humidity, and afternoon winds. Between
1300 and 1400 many personnel on the fire reported increased fire activity.
At around 1400 the fire apparently made a "push" and overran
a lower helispot (H-1) near where most of the fire-line crews were working
near Cramer Creek. The temperature was nearing 100 °F and humidity
was in the low teens.
Several times during the day different individuals on the fire observed
low intensity fire backing down into the Cache Bar drainage below where
the rappellers were working. Sometime between 2:30 and 2:40 p.m. this
fire turned into an active flaming front.
At around 1505, the rappellers requested the Cove Creek helibase for
a pick-up from the helispot site due to significant smoke. A rappeller
stated "Send them in a hurry." Both helicopters H-193 and H-166
were on the ground at the helibase at the time. At 1509 a rappeller again
called the helibase requesting a pickup. When told the helicopter was
on the ground but would leave shortly, the rappeller responded, "We
need them right now."
At 1510 the helibase manager instructed the radio operator to ask the
rappellers if they needed to go to their safety zone. The rappellers replied
they were fine, they were just taking a lot of smoke. At 1512, a rappeller
again called the helibase. When told H-166 needed fuel but would be leaving
right away, the rappeller responded, "Oh God. We just got fire down
below us. So the smoke's coming right at us...uh...just make them hurry
up." Between 1513 and 1520, the lead plane reported the fire was
"blowing-up" on the west side. By the time H-166 reached the
helispot around 1520, it notified the rappellers it was too smoky to land.
At 1524 a rappeller called again in an excited voice, "Could I get
a helicopter up right now?" Shortly thereafter the rappellers stated
to the lead plane L41 they were going to "run for it." Apparently
no other communication was successful with the rappellers (Radio
log and witness statements).
The two rappellers were later found about 420 feet up hill and into the
drainage from the helispot. Neither victim was found inside a fire shelter.
It appears that at least one of the victims may have attempted to deploy
a fire shelter, and the other shelter appeared to be ready for deployment.
Preliminary reports from the Forest Service technical equipment experts
stated that the temperatures were too extreme for the victims to have
survived, even inside properly deployed fire shelters. Temperature estimates
were between 1,300 and 2,000 degrees Fahrenheit. The fire apparently quickly
spread through a stand of timber below the helispot and fatality site,
burning intensely through the crowns of the trees near the victims.
VI. Findings and Causal Factors
A. Extreme Fire Conditions and Local Factors
The managers, incident commander, and key personnel involved with
this fire were employees of this District or Forest and were very familiar
with the current conditions and local factors contributing to the hazardous
conditions involving the fire fighting efforts. The effects of steep terrain,
drought, winds, high temperatures and low humidity were all recognized
by key personnel, yet tactics were not altered to account for the extremely
The fire season on the SCNF normally runs from July 1 to September 15.
The Forest is a high fire-load Forest, one of four Forests in Region 4
with a high fire occurrence. Based on the last 10 years, the SCNF averages
120 fires and 52,000 burned acres per year. Recent fire events have increased
dramatically in size and severity. Two fires, the Clear Creek and the
Salmon-Challis Wilderness, for example, burned 417,000 acres in 2000.
The Salmon River Breaks is "rugged and dangerous country with limited
visibility." The terrain transitions rapidly among distinct features-benches,
steep slopes (=70%), breaks, and saddles. Vegetation covering the rocky
slopes also limits visibility ([————],
Because the topography is steep, it has a significant effect on fire
behavior, fuels; and weather. Fire behavior in the hotter part of the
day can be extreme, with rapid spread through the canyons, draws, and
chutes. Steep slopes not only predispose areas to rapid uphill fire growth
but also cause fire brands to roll down hill. Historically, a fire that
burns in the Salmon River Breaks will burn to the top of the ridges and
then downhill to the Salmon River because of the sheer ruggedness and
steepness of the terrain (Goheen, p 14; Hafenfeld, p. 26;
[————], pp. 9, 13; Forest FMP, Sect. III p. 5,
The topography also has a marked effect on fuels, because it contributes
significantly to differences in moisture and surface heating from place
to place across the landscape. These differences influence variations
in the amount of available moisture as a function of elevation and variations
in surface heating due to aspect (Forest FMP, Sect. III, pp. 58-62).
Vegetation and Fuels:
Vegetation varies considerably by elevation and aspect. In 1985 a stand-replacement
fire burned through most of Fountain Creek drainage and into the adjacent
drainage to the south east above the Cache Bar boat ramp
(Witness statements about "Old burn"). This occurred
in the same Cache Bar drainage above which these rappellers were working
when they were fatally burned over.
Southern aspects consist of grass and scattered shrubs, while the northern
aspects are a mosaic of rocky patches and nearly continuous stands of
2-4 foot shrubs, primarily shiny-leaf Ceanothus. Large, downed, wood fuels
from fallen snags are scattered through these brush fields. Shiny-leaf
Ceanothus burns intensely due to volatile oils in the foliage. The potential
rate of spread through Ceanothus increases as the growing season progresses
from June 20 to September 10. It is possible that the observation of this
green foliage during a helicopter recon flight provided a false sense
of security as to its potential to burn. http://www.fs.fed.us/database/feis/plants/shrub/ceavel/fire_ecology
At 75%-100% live fuel moisture, fires exhibit extreme fire behavior and
burn actively through the night. Live fuel moisture on the SCNF was measured
five times during June and July. Moisture levels in conifers, sampled
on July 21, 2003, ranged from 80-110%. Shiny-leaf Ceanothus, sampled on
July 30, 2003, was 104%. Generally, live conifer fuel moisture levels
were comparable to or lower than samples taken in early August, 2000.
During the 2000 fire season, more than 417,000 acres burned on the Forest,
including the Clear Creek Fire which made a 23,000 acre run in one day
(North Zone Fire Danger Pocket Card; Records L. [————]
Great Basin Live Fuel Moisture Project).
The climate in the Salmon River Breaks varies widely with elevation. Wind
patterns in the area are dominated by westerly winds. Topography influences
local weather conditions, especially winds. Upslope/up-canyon diurnal
winds are common during the summer months. It is normal to have strong
up the Salmon River canyon on hot summer afternoons. These winds are primarily
from the west. Since these winds are channeled by topography, the local
wind conditions are far different from those predicted in the general-area
fire weather forecasts. Nighttime thermal zones are common, especially
in the deeper canyons, and contribute 10 active fire behavior at night
(Forest FMP pp. 58-62).
The Pocatello, ID, office of the National Weather Service (NWS) provides
fire weather information and predictive services for this area of the
Salmon-Challis National Forest, including daily fire weather forecasts
and spot forecasts (forecasts specific for the fire area). The Forest
is required to use forecasts from the NWS servicing center in Pocatello,
even though those from the NWS servicing office in Missoula, Montana are
considered by some on the Forest to be more accurate and useful. In particular,
wind predictions from the Pocatello spot weather forecasts are perceived
to be unreliable, and the forecasts frequently do not reflect actual conditions
on the ground. Weather from previous day is often seen as a more reliable
predictor than forecasts from Pocatello (Hershey, pp. 25-27).
Note: conditions on the North Fork RD, on which the Cramer Fire occurred,
had been the same for several successive days with high temperatures,
low humidity, and afternoon winds ([————],
22-26; [————] NWS Pocatello).
The SCNF fire program relies heavily on data provided by seven Remote
Automated Weather Stations (RAWS) maintained across the Forest. Maintenance
of the RAWS sites is the responsibility of the Forest, however the maintenance
service could also be contracted to staff at the National Interagency
Fire Center (NIFC) at an increased expense. The Skull Gulch RAWS site
is the most representative of the Cramer Fire area, but weather data from
this site was inaccurate and unusable due to a temperature and relative
humidity sensor that was not functioning properly as well as other weather
data being reported erroneously. This problem was not discovered and corrected
until after the Cramer Fire. The RAWS on the Forest had not been maintained
or calibrated since September 2002 (Statement: [————];
Record Skull Gulch RAWS). The person responsible for RAWS maintenance
for the Forest had apparently ordered replacement sensor in June 2003,
but was unable to install them and concurrently perform calibration and
maintenance prior to the busy fire season ([————]
phone interview). Five days after the fatality a NIFC technician
inspected and performed maintenance on the Skull Gulch RAWS and detected
the sensor problems causing the reporting of incorrect data.
Long Tom lookout generally provides weather information to fire suppression
personnel in the area, including the first three days of the Cramer Fire.
The difference in elevation between Long Tom and the fire was too great
for the lookout's weather observations to be valid for the fire area ([————];
Fogel; Forest Net Radio Log; Long Tom Lookout log).
The combinations of inaccurate RAWS data, inappropriate lookout weather
due to elevation differences, and lack of faith in Pocatello National
Weather Service reports may be a contributing factor to the limited attention
paid to forecasts and why a spot weather was not obtained by the IC the
day of the fatalities. Apparently, the electronic spot weather request
form used by dispatch has a section where a requestor can input discrepancies
from the previous day's forecast, in order for the Weather Service to
make adjustments and to make future forecasts more accurate. Apparently
this feature is seldom used. Little additional evidence was provided indicating
the Forest had taken any steps to try to mitigate weather forecast related
Fire Hazard Indicators:
Two indices from the National Fire Danger Rating System (NFDRS) are commonly
used to express fire danger and track seasonal trends -- the Burning Index
(BI) and the Energy Release Component (ERC). The BI reflects the difficulty
of fire control. The ERC reflects the contribution that all live and dead
fuels have to potential fire intensity and is a good indicator of the
overall fire danger resulting from local fuel moisture conditions. These
indices communicate expected fire danger to fire personnel and allow them
to associate the numeric values with real-life experiences in the area.
Forest fire suppression personnel use the ERC to estimate fire growth
potential and in complexity analyses (Record FMP; NWCG
http://famweb.nwcg.gov/pocketcards). The burning index is also
presented on the area's Fire Danger Pocket Cards required under the Thirtymile
Hazard Abatement Plan. The Forest also uses the Haines Index as an effective
indicator of potential fire behavior. The index is included in the daily
zone fire weather forecast and is read by dispatch when the daily fire
weather forecast is transmitted to field units (Forest
FMP, Sect. IV, p. 4-9).
The Forest was experiencing fire conditions comparable to the year 2000
when two fires burned 417,000 acres on the Forest. During July, 2003,
the Forest and eastern Great Basin were at Preparedness Level IV. The
BIs and ERCs were near the high-end — the Forest BI was well above
the 90th percentile and the ERC was in the 96th percentile-indicating
dangerous conditions. The Forest had just experienced a 15,000-acre fire
on the Leadore Ranger District, a district with historically low fire
occurrence and smaller fires. The Eastern Great Basin Coordination Center
issued an updated safety alert that was distributed to all personnel,
emphasizing both the dangerous conditions and the need for aggressive
initial attack. The Fire Danger Pocket Card is a method of communicating
information on fire danger to firefighters. A crew on the Cramer Fire
had reviewed their Pocket Cards a few days prior to the Cramer Fire and
recognized that they were at conditions comparable to those on the Clear
Creek Fire in 2000 that burned 217,000 acres (Hand, pp
18-19, 63; Mills, p. 10; Sever, p. 13; Shaddle, p. 21; Eastern Great Basin
Coordination Center Safety Alert; NWCG Pocket Card http://famweb.nwcg.gov/pocketcards).
The Forest's Fire Management Plan provides much information about the
fire regime and hazards specific to the area such as the trend towards
fires with increasing severity. From 1994 to 2001 the data shows that
50% to 75% of the burnt acres were severe, versus historic data indicting
only 20% were severe (FMP Sect III, p. 28).
B. Indifference to Core Safety Principles
The lack of consideration to safe firefighting principles extended beyond
indifference to just the Ten Standard Orders and Eighteen Watch Out Situations;
The Incident Response Pocket Guide (3 inch by 5 inch booklet) is carried
by every firefighter and includes a wide variety of safety practices in
the first pages of the guide under the Operational section. It includes
Operational Leadership Guide; Risk Management Process; Look-Up, Down and
Around; Common Denominators of Tragedy Fires; Tactical Watch Outs; LCES;
Extended Attack Transition Analysis and more. The Standard Orders and
Watch Outs are prominently located on the back cover. (IRPG Included in
binder and on CD)
LCES (Lookouts, Communications, Escape Routes and Safety Zones)
LCES is simplified system to aid firefighters in remembering the four
main categories into which the Ten Standard Firefighting Orders are divided
(Wildland Fire Fatalities in The United States, 1990-1998).
Lookouts and Communications
Lookouts were not provided for the rappellers at H-2. A competent lookout
was not provided with good vantage points of the fire and H-2 (Fogel,
pp. 49 and 71-75; Fuller, p. 1; [————], pp. 7,
23, 24, 39-42; Hackett, pp.40, 74-77; Hackett OSHA Recorded statement,
Tape 1, 26m50s; Hand 49-54; [————], p. 14; [————],
p. 8; Raddatz, p. 20; [————], p. 39). Lookouts
are to have knowledge of the escape routes and safety zones as well as
knowledge of disengagement trigger points ("Incident
Response Pocket Guide" p. 6). It was the candid opinion of
many experienced firefighters on the investigation team that the provision
of an appropriate lookout would have been the number one factor which
could have prevented these deaths.
Communications were not maintained to provide updates on situation changes.
The rappellers did not have a direct supervisor to whom they were to report
and who would routinely check on their progress (Fogel,
p. 11, 12, 70; [————], p. 8). The Incident
Commander admitted that he would probably be their supervisor, but that
it was common to put experienced firefighters into an area to cut a helispot
without direct supervision (Hackett, p. 93).
The rappellers occasionally communicated with the radio operator at the
Cove Creek helibase, as well as some communications with aerial support
over the fire ([————], p. 6; Fogel,
pp. 11-12 and 70). The incident commander did not have direct communication
with the rappellers. The rappellers were not informed of the decision
to abate the plan to fly crews into H-2, nor were they provided information
on the progression of fire in the drainage below them.
Safety Zones and Escape Routes
It appears that the primary safety zone designated for the victims was
either an open area west of the helispot across and into the Cache Bar
drainage, or down the ridge into the blackened area that had burned the
first three days of this fire. The interview statements were not consistent
as to which area was to be considered the primary safety zone ([————]
, p. 22; Fogel, pp. 6, 8, 19, 66, 67; Fuller, p. 1; Hackett, pp. 24, 27,
41-44, 99-100; Hackett written statement).
The "open area" was considered "old burn," presumably
from the 1985 Fountain Creek fire. It would have been necessary to have
"burned out" the area prior to use, and it was not large enough
to be considered a safety zone. By definition, a safety zone is to be
large enough to be survivable without the deployment of a fire shelter.
As a guideline for a safety zone, the "distance between the firefighter
and the flames must be at least 4 times the maximum flame height"
(Interagency Standards for Fire and Fire Aviation Operations,
pp. 4-10). This area experienced intense fire during the blow-up
which claimed the lives of the rappellers (USFS Accident
Investigation Report, Fire Behavior
Analysis Appendix A, Entrapment Phase).
The other possible safety zone into the black (low intensity surface
fire from first day) was downhill from the helispot. The escape route
along the ridgeline was rocky with unburned fuels along the route between
H-2 and the black. In addition, with fire burning in the Cache Bar drainage,
the escape route is likely to have been compromised from the heat of the
fire and fuels on the slope below the ridge on the Cache Bar side. Furthermore,
with smoke and fire advancing from the same general direction as the down
slope safety zone, without a lookout it would have been difficult for
the rappellers to know where the fire was, and if the escape route and
safety zone had been compromised.
Once conditions had changed, such as an active flaming front below Helispot-2,
no attempt had been made to reevaluate the safety zones and escape routes,
and if necessary identify an adequate alternate escape route and safety
zone. No other measures had been taken to mitigate this hazard. The helicopter
had become the rappellers' primary route to safety ([————],
p. 8; Fogel, p. 13).
C. Management and Organization
Decentralized — Over the past few years, since
the arrival of the new Operations Staff Officer, the Forest has shifted
to a very decentralized structure of fire management. For fires from small
Type 5 incidents to significant Type 2 incidents, primary responsibility
for decisions has been delegated to the Districts. Only for Type 1 Commands
does the Forest Supervisor become involved (Bates, 35-36;
Dudley, 4-7,19-22; Goheen, p. 14; Hafenfeld, pp. 20-21, 83-84; [————],
43-45; Matejko, 10-12, 14-16; Mills, p. 14-15; Forest FMP Sect. IV, p.
21; Memo From Forest Supervisor to District
Ranger on delegation of authority).
The Operations Staff Officer at the Forest level is tasked under the
Forest Supervisor with addressing and ensuring fire operations are handled
appropriately. The Staff Officer has oversight of the Forest FMO and Deputy
FMO, who in turn provide assistance when needed to the District FMOs.
The Operations Staff Officer is also responsible for addressing effectiveness
with the District Rangers (Hafenfeld, p. 84). There
is a concern among many in the Forest staff that the Operations Staff
Officer could not provide an appropriate level of supervision over the
District Ranger for the North Fork and the Middle Fork Districts, as that
Ranger is also the Staff Officer's spouse (Dudley, p. 7;
Goheen, pp. 7-8, 11-12; [————], 32-33; Matejko,
Accountability — It could not be demonstrated
at any time during the past year that any firefighting personnel or others
up the chain of command had ever been disciplined, removed from their
position, or formally addressed concerning unsafe behavior during firefighting.
All SCNF personnel involved with supervising or managing firefighting,
from incident commanders, through the district level to the forest level
were not evaluated based upon their ability to fight fires safely and
in accordance with core firefighting safety principles (Performance
Plan and Appraisal form FS-6100-37 9/97).
Fire staff at the Forest Supervisor's Office did not play a role in ensuring
appropriate safety measures at the Cramer Fire. The Forest Fire Management
Officer, Forest Operations Staff Officer, and Central Idaho Dispatch Manager
were all made sufficiently aware that the fire had grown substantially
each day, suppression strategies and tactics of the ICs had not been successful,
and the two medium helicopters requested by the IC were not available.
On the evening of the 21st, the Incident Commander requested assistance
with overhead and handling logistics (Hackett statement,
pp. 18-19, 88-89; Sever, pp. 9-11). The forest fire managers responded
that resources were spread too thinly and such support was not available.
The Operations Staff Officer and Forest FMO were both involved with a
regional fire preparedness review on July 22, 2003 (Hafenfeld,
pp. 9, 34-35; Mills, pp. 3, 5, 8). None of the forest-level or
district-level fire staff appeared to have taken any further steps to
review the circumstances around the fire and to ensure the appropriate
safe level of incident response was being applied.
On the evening before the fatality, concerns about disorganization and
the competency involved with the Cramer Fire had been brought to the attention
of forest-level fire management by the Forest Aviation Officer (Bates,
p. 31; Hafenfeld, pp. 65-68, Sever, pp. 29-32). These concerns were forwarded
to and later discussed with the North Fork District Ranger on the morning
of July 22. None of the concerns about competency and perceptions of disorganization
were addressed or looked into further ([————],
pp. 7-10, 13-14, 25-26, 29-37).
Despite recognition by the Forest Service of the hazardous nature of
smaller fires (see "Common Denominators of Tragedy Fires" in
Willful Justification below), the Forest Service continued to fail to
adequately monitor safety on this Type 3 fire. No one monitored safety,
which could have included any of the following: the incident commander
(Hackett statements); a safety officer at the fire
(none designated); line officers such as the District Ranger (Bates
statements, p. 6-7 plus); the district FMO (vacant); the zone duty
officer — not involved ([————],
pp. 2-8, 32); forest-level fire staff, such as the Forest FMO/Forest
Duty Officer (Hafenfeld, Mills, Sever).
Resources Spread Thin - Delayed Hiring of District FMOs —
For approximately 12 to 18 months, including the previous 2002
fire season, the North Fork District FMO position was vacant (Bates
pp. 33-35; Matejko, pp. 18, 23). In absence of the DFMO, acting
DFMOs were assigned on a couple of occasions. During the period leading
up to the Cramer Fire there was no acting DFMO for this District and the
Assistant DFMO assumed the responsibilities for the North Fork. This Assistant
DFMO was also the Cramer Fire Type 3 IC, who was in command at the time
of the fatal accident (Hackett, pp. 4-5, Hafenfeld, p.
- The decentralized structure of fire management in the forest placed
responsibility on the District to manage fires up to a Type 2 incident
(see first heading, "Decentralized" under paragraph
- No FMO was assigned to the North Fork District at the time of the
Cramer Fire (Hackett, pp. 4-5);
- Most FMO duties were then to fall to the Assistant FMO, Alan Hackett,
who was also the Type 3 Incident Commander for the Cramer Fire (Dudley,
p. 18-19;, Hackett, pp. 4-5);
- The District Ranger for the North Fork was also the Middle Fork District
Ranger, which were logistically separated by a considerable geographic
distance (Goheen, p. 21). This Ranger was also
preoccupied with "higher priority" fires elsewhere on her
districts and was unable to provide the oversight as designed in the
decentralized fire structure. (NOTE: 85% of fire activity normally occurs
in her North Fork and Middle Fork Districts; Bates, p.17;
- By the beginning of August 2003, 28 fires had occurred in the Salmon-
Challis National Forest, 14 of which occurred in the North Fork Ranger
District with no Fire. Management Officer, and with the District Ranger
also overseeing the Middle Fork Ranger District (Bates
OSHA recorded interview, 32m15s). On July 21 and 22, the District
Ranger was occupied with the Crystal and Dutch Lake fires in the Middle
Fork District (Bates statements, p. 6-7 plus);
- The North Zone Duty Officer for the day (also the Salmon-Cobalt DFMO)
was not providing oversight or keeping in regular communication with
the Cramer IC ([————], pp. 2-8,
- The Supervisors Office and Central Idaho Dispatch did not provide
oversight to the fire, even though they were aware of the extreme conditions,
fire growth, and concerns about competency had been brought to their
attention (Hackett statement, pp. 18-l9, 88-89; Hafenfe1d,
pp. 9, 34-35, 65-68; [————], pp. 7-10, 13-14,
25-26, 29-37; Mills, pp. 3, 5, 8; Sever, pp. 9-11, 29-32).
Management failed to ensure that the evolution of the Incident Command
System was commensurate with the fire threat. The decentralized fire management
structure in the Forest placed most of the oversight responsibilities
on the District. In short, the incident commander for the fire was essentially
left to his own devices without any form of formal oversight.
District FMO positions had also not been filled on the Middle Fork and
Yankee Fork Ranger Districts (Matejko, p. 18; Forest Org
Charts). The prolonged vacancies of key fire management positions
in the districts appear to conflict with the Forest's goal of decentralized
fire management to the district level.
Almost as soon as ground forces arrived at the fire, the potential for
spread was noted and a Type 3 Incident Commander was ordered (Dispatch
Log, McCall Jumper Ship J41, 07/20/03, 1708, and Moyer Helitack, 1938;
[————] p. 5; Sever, pp. 6-7). Each day
the fire grew significantly and the potential for significant spread was
observed. The potential for blow-up and significant growth should have
been extremely evident to those working on the fire as well as to the
zone duty officer (not involved), dispatch center, District Ranger (did
not attempt to be involved), the Forest FMO and Forest Operations Staff
The Salmon-Challis National Forest Fire Management Plan (FMP) requires
"If initial action is unsuccessful, a WFSA (Wildland Fire Situational
Analysis) will be prepared to determine the next set of management responses"
(Sect III, p 10).
Under these requirements, a WFSA would have been necessary no later than
the afternoon of Monday, July 21. A WFSA was not created for the Cramer
Fire until Wednesday, July 23, 2003 at 0800. At that time the fire size
was approximately 5,400 acres after more than three days of suppression
activities, two to three operational periods after the fire had gone from
initial attack to extended attack.
D. Firefighting Culture
The wildland firefighting community still has a long way to go before
they truly have a zero tolerance for infractions of firefighting safety
standards and procedures. Except after a tragic event, it appears upper
level management has rarely been held accountable for safety on the fireline.
The most basic safety rules for federal wildland firefighters, the Ten
Standard Fire Orders, were written in the 1957 and have been in place
since then with only minor adjustments (http://www.fs.fed.us/fire/safety/10_18/10_18.html).
Yet men and women are killed due to violations of these basic rules.
Employees and supervisors are rewarded for aggressively fighting fires
and taking serious risks. Firefighters are often provided hero status
for their exposure to hazardous working conditions. The can-do
culture throughout the fire fighting ranks is likely to be a very significant
contributor to these unfortunate incidents. Personnel are often promoted
to higher prestige positions that infer risk in their position titles,
such as "Hotshots." In contrast, there still appears to be no
effective incentives for safe behavior.
Another example of the risk taking culture is the current tendency among
Forest Service "regulars" (federal Forest Service fire fighting
employees) to criticize private contract firefighting crews for being
too willing to withdraw from a perceived dangerous situation (Hackett,
p 10, 51-52: Sever, p. 9, 18-20). This begs the question, Are contract
crews overly-cautious, or are 'regular' crews too willing to routinely
accept serious risk? Most fireline crews strongly prefer working long
hours on a fireline versus idling waiting. The typical fireline crew member
has significant financial incentives when actively working a fireline,
such as overtime, per diem, and hazard pay.
One of the measures used to evaluate the effectiveness of wildland firefighting
activities is the percentage of fires suppressed during initial attack
and prevented from escaping and becoming large. Another measure has to
do with the percentage of threatened structures in the wildland urban
interface that were protected or destroyed (http://www.nifc.gov/fireinfo/2002/summary.html).
It is understandable that other pressures to suppress fires are very real
and valid. Uncontrolled wildfire can be very costly and damaging for a
wide variety of interests. However, this merely is another example of
the motivations of fire fighters, supervisors, and management.
VII. Other Investigators
A. Interagency Accident Investigation Team (AIT)
Position on Team
Team Technical Support
B. USDA Office of Inspector General