Point Fire 1995: Thunderstorm Winds, Radio Communication and Entrapment

Historical wildfire analysis · Idaho, United States · 28 July 1995
Sources reviewed: 25 October 2026.

The Point Fire killed Kuna Rural Fire District firefighters Bill Buttram and Josh Oliver when their disabled engine was overrun. The incident is valuable for understanding how a change in fire behavior, vehicle mobility and communications can interact during mop-up. A warning issued somewhere in the response system is not evidence that every crew received it. [1]

From initial attack to renewed spread

The investigation overview places the initial report at 18:29 MDT, northeast of Initial Point, about 16 miles southwest of Boise. Arriving resources encountered sagebrush and dense cheatgrass, an estimated 60–65 acres of fire, and moderate spread. It records west winds of 4–6 mph and flank flame lengths of 3–5 feet at that stage. Those observations describe an early phase; they should not be carried forward unchanged into the later entrapment. [2]

NWCG’s chronology reports that spread had stopped at approximately 120 acres by 20:10. At 20:22, a Red Flag warning forecast thunderstorm gusts up to 50 mph. Around 20:46, southerly thunderstorm winds drove renewed spread across the northern perimeter. [3]

The critical sequence

Time, 28 July	Event reported by NWCG	What the timestamp establishes
20:10	Spread reported stopped at 120 acres.	A condition at one stage of the response.
20:22	Weather warning issued.	Issuance, rather than confirmed receipt by every resource.
About 20:46	Fire escapes the northern perimeter.	An approximate start of renewed spread.
20:49	Engine 620 reports approaching fire and loss of mobility.	A recorded distress communication.
About 20:50	The engine reports being overtaken.	The final reported transmission in this sequence.

Timeline source: NWCG’s Day in History account. [3]

Four minutes is an interval, not an escape allowance

Subtracting 20:46 from 20:50 gives four minutes. That reproduces the approximate interval in the published chronology; it does not establish when the occupants first recognized the threat, which movements remained possible or how much usable response time they had.

Likewise, the 24 minutes between warning issuance and renewed spread cannot be labelled as warning time received by this crew. A defensible reconstruction needs the message, channel, recipient, acknowledgment and any subsequent change in assignment. An incident-level timeline and an individual crew’s information timeline are different records.

Communications: channel access and message receipt

NWCG describes attempts to contact the stationary engine on the BLM tactical channel without a response, followed by Engine 620’s distress call on a local non-federal frequency. [3] This supports examining how information moved between organizations. It does not establish that silence meant a message had been heard and ignored.

For an evidence review, a communications diagram should therefore distinguish technical ability to transmit, the channel actually monitored and confirmed receipt. Drawing a line between two agencies simply because both used radios would conceal the question that matters.

Vehicle movement and the limits of inferred intent

The Lessons Learned Center describes Engine 620 turning north onto a two-track road and then travelling cross-country through heavy sagebrush before becoming disabled. It explicitly leaves the reason for the turn unknown. [1]

That uncertainty should remain visible. A reconstructed route can establish movement without establishing motive. Explaining the route as a deliberate choice to reach a particular refuge would require evidence beyond the summary. Similarly, a report of overheating is not by itself a complete mechanical diagnosis.

A weather forecast is not a local wind measurement

The forecast maximum of 50 mph is an upper gust prediction, not a demonstrated observation at Engine 620. Comparing it directly with the earlier 4–6 mph wind description would mix a forecast with an observation from a different stage. [2] [3]

Before calculating a wind-speed increase, an analyst needs observations with compatible locations, heights, averaging periods and times. The published sequence supports a major change in conditions; it does not justify inventing a precise multiplication factor at the fatality site.

What this case contributes

NWCG’s RT-130 resource places Point Fire within communication, decision-making and cooperation training. It notes that BLM and Kuna personnel had worked together earlier that season despite the absence of a mutual-aid agreement at the time. [4] Prior cooperation is relevant context, but it cannot substitute for evidence of how information and assignments were coordinated during this incident.

This retrospective reviews agency summaries and the investigation overview. It does not claim a new mechanical investigation, a complete reconstruction of radio traffic or a determination of individual responsibility.