Researchers estimated the fate of 26,403 giant sequoias across 19 groves struck by California's 2020 Castle Fire and 2021 KNP Complex Fire. Trees in areas burned by managers during the preceding ten years had 77% lower odds of mortality than untreated trees, making them nearly four times as likely to survive. [1] [2]
The result is large. It is also observational and modeled. Nature Communications first published the paper on July 11; a University of California, Davis release distributed through ScienceDaily made it newly prominent on July 26. [1] [2] Neither date marks a randomized intervention, and the later publicity does not turn an association into a universal rule.
That discipline is central to the paper's lost-science position: preserve the population, method, date and evidence stage. "More fire saves forests" is memorable. It also discards which fire, where, when and under what conditions.
A natural comparison, not a trial
Giant sequoias evolved with periodic, relatively low-intensity fire. Smaller fires clear vegetation, recycle nutrients and help create conditions for new trees. A century of fire exclusion allowed combustible material to accumulate in parts of the Sierra Nevada. Sequoia and Kings Canyon National Parks, meanwhile, have used prescribed fire since the 1960s, leaving a patchwork of different treatment histories. [1]
That patchwork gave researchers a natural comparison when extreme wildfire arrived. The team combined airborne lidar, three-meter PlanetScope satellite imagery and field observations in a deep-learning classifier. It estimated 7,974 deaths, with a 95% Bayesian credible interval from 7,555 to 8,430. Estimated mortality was 30.2%, with an interval from 28.6% to 31.9%. [2]
The 77% figure describes lower mortality odds associated with prescribed burns completed no more than ten years before the wildfires. Its credible interval ran from 69% to 83%. [2] Odds are not the same as a randomized reduction in observed deaths. Treatment was not assigned by lottery; managers chose where and when to burn, and treated places may differ from untreated places in ways a model cannot fully remove.
Researchers then simulated counterfactuals. They estimated that prior prescribed burns prevented at least 1,888 deaths, with an interval from 1,487 to 2,302. Universal treatment in the affected groves might have prevented another 3,888, with an interval from 3,236 to 4,580. [2] Those trees were not observed surviving. They are model estimates of what could have happened under different treatment.
Strong evidence has boundaries
The analysis improves on anecdote. Twenty-six thousand trees, remote sensing and a stated statistical framework provide more than photographs of one successful grove. The comparison also concerns extraordinary fires that killed ancient trees at rates the authors describe as unseen for millennia. [2]
But implementation demands records beyond efficacy. A prescribed burn has a season, intensity, perimeter, crew, weather window and smoke burden. It competes for staff and money. It may draw on Indigenous fire practice and local ecological knowledge. The right treatment for giant sequoias on the Sierra Nevada's western slopes is not automatically the right treatment for another species, forest or community.
Timing matters particularly. The study's reported association concerns burns within ten years. It does not establish one permanent treatment, the best interval inside that window, or what happens when drought, fuels and extreme weather change faster than management cycles.
The evidence also needs its full methods tested. Field-confirmation rates, classifier error, topography, fire severity and variation among groves could affect the estimates. Replication after other fires would show whether the measured advantage persists across different weather and fuel conditions.
ScienceDaily's headline offers an attractive reversal: the best way to save trees may be more fire. [1] The more useful conclusion is narrower. Carefully managed fire was associated with substantially better survival for giant sequoias facing two extreme Sierra Nevada wildfires. That finding supports action, but action must retain the conditions that made the finding intelligible.
No authorized X post was recovered. Online certainty for or against prescribed burning therefore remains unmeasured. The evidence does not need a synthetic platform fight. Its real divergence is between a clean slogan and a difficult management program.
The study gives forest managers a strong receipt, not a blank prescription pad. The next records should show feasible acreage, treatment timing, smoke and safety costs, staffing, Indigenous participation, field validation and survival after later fires. Fire may protect these trees. The method decides when that sentence is true.
-- DARA OSEI, London