Climate models: warming and greater variability threaten some Mount Tamalpais forests and raise wildfire risk

Pepperwood Preserve’s climate vulnerability review, presented by Kai Hennepin, framed future conditions for Mount Tamalpais using a set of climate-model scenarios and a probabilistic vegetation model. Hennepin told the webinar that "all models agree that the climate is getting warmer," but he stressed that models diverge on precipitation and that timing and intensity of rainfall — critical for aquatic species — remain difficult to predict.

The presentation used a four-square framing (warm/wet, hot/wet, warm/dry, hot/dry) and summarized mid- and late-century projections. Hennepin reported a broad range in mid-century annual precipitation change by model (from roughly a 1% decrease up to about a 37% increase), and said late-century projections could show up to a 50% change depending on the model. He warned that greater variability means both more intense flood-producing atmospheric-river events and longer, more severe drought intervals.

On wildfire, Hennepin cited regional assessments indicating a projected rise in wildfire probability and area burned: "The fourth climate assessment projects an increase of about 10% in the probability of wildfire by the end of [the] century with an increase in area burned of about 42% mid-century and then 97%" in later projections, he said, underscoring that models indicate rising risk though they cannot predict precise timing or ignition sources.

Using the probabilistic vegetation model (PBM), the team assessed likely shifts in suitability for 22 vegetation types. Some chaparral and drought-tolerant species (for example, chamise, coyote brush, California bay) are likely to remain stable or expand under many scenarios, while sensitive species (coast redwood and Douglas fir in particular settings) are predicted to decline in suitability and may require targeted protection in moist refugia.

Hennepin emphasized that models generate hypotheses for where to prioritize monitoring and management, and that local, repeated observations are critical to evaluate model projections. He noted specific limitations: aquatic species’ responses are highly sensitive to rainfall timing and duration — a factor models do not reliably capture — and fog data at the coast are a known gap in the regional modeling.

The climate results were presented as inputs to stewardship planning; speakers repeated that monitoring and adaptive management will be essential to test model-driven hypotheses and guide interventions such as fuel reduction, refugia protection and strategic use of prescribed fire.