Climate change and wildfire Research at the PNW Station: past, present,

future

Don McKenzie (TCM/FERA)

with contributions from

PNW Science Day March 12, 2014

• Paul Hessburg• Becky Flitcroft

• Sim Larkin• John Kim

Rationale✤ It’s getting warm down here.

✤ No “hiatus” in hot extremes over land.

✤ More area is expected to burn.

✤ What we care about is how fire climatology translates to the issues and scales relevant to land management.

• Fire effects: tree mortality, smoke and air quality, habitat structure and pattern, regeneration and forest succession.

• Time domains: immediate (to 2020s), next generation (2040s), long-term (2060s and beyond). Uncertainties grow non-linearly over time.

• Space domains: cross-scale, from watersheds (“landscapes”) to the region.

• Specificity: fire regimes. It’s not about individual fires or “my favorite pixel”.

Seneviratne et al. (2014)

(though not so much as on this map)

Past and present research (1)

✤ Drivers of area burned

✤ Fire-climate models at the scales of ecosections project that the West will burn up.

✤ Expectation breaks down because there are limitations.

• Fire area can’t keep increasing because fires will run out of real estate.

• “Hotter and drier = more fire” doesn’t work everywhere. Best in the dark green ecosections.

Expectation: Hotter and drier = more fire!

Correlations between annual area burned and water-balance deficit

Temperate rain forests: extreme weather causes rare wildfires.

Transitional forests: drought stress will increase fire extent and severity.

Arid forests: fire extent and severity may actually decrease.

Past and present research (2)

✤ Smoke and air quality (AirFire/FERA — Larkin/McKenzie)

✤ Smoke modeling framework (BlueSky) that accepts either observed or simulated (i.e., future) fires.

✤ Stochastic fire simulator tuned to the spatio-temporal domains of air-quality models.

McKenzie et al. (2014)

Past and present research (3)

✤ Future megafires (AirFire/FERA — Larkin/McKenzie)

✤ Expectation of more extreme events based on projections of future fire weather.

✤ Representing all the factors that combine to produce a megafire.

• Weather pre-ignition conditions fuels.• Weather on the day or hour of the fire.• Escapes initial attack? (hard to model but a big

source of uncertainty)• Weather in days or weeks following fire.

✤ How will this change in a warming climate?• Downscaling climate models.• Different regions will see different fire weather

(not always hotter and drier).

Stavros et al. (2014)

Past and present research (4)

✤ Fire and landscape dynamics (EPF/CLI — Hessburg)

✤ Patch-size distributions associated with future climate.• Topographic controls based on terrain patch structure.• Endogenous vs. exogenous controls on fire & other disturbance.

✤ Restore and maintain ecosystem function in future climate.• Use topography as a template.• Patch structure and tree density tuned to “climate analog” reference conditions

rather than HRV.• Anticipate patterns of fire severity and seral stages.

Past and present research (5)

✤ Fire, climate change, and bull trout vulnerability (LWM/AEM — Flitcroft)

✤ Patch-size distributions associated with future climate.

✤ Habitat extent of cold water aquatic species is vulnerable to climate change.

✤ Climate change may isolate small patches of habitat, often in the headwaters of a watershed.

✤ Wildfire may compound the negative effects of climate change for cold water species.

✤ Some management action to reduce wildfire effects may serve to protect some cold water aquatic refugia.

Past and present research (6)

✤ Process-based modeling of climate, vegetation, fire (EPF/CLI — Kim)

✤ MC2 DGVM simulates vegetation-fire interactions at multiple scales.• Global, CONUS, regional.

• Currently studying R6, R5, R4, R1, and Blue Mountains Ecoregion.

✤ MC1-based Seasonal Drought and Fire Forecasting System creates 7-month fire and drought forecasts, updated monthly.

✤ Downscaled output from CMIP5 GCM projections used to drive DGVM and predict changes in fire.

Projections of biomass consumed by wildfire:1951-2000. vs. 2050-2099

Future research (1): categories

• Field & remote-sensing studies‣ Fire and succession

‣ Fire and other disturbances

‣ Fire and carbon

• Theory‣ Conceptual models

‣ Scaling

‣ Extreme events and thresholds

• Models‣ Landscape projections

‣ Process AND empirical

‣ “As simple as possible, but no simpler”

Fire

Subalpine fire and succession (Cansler 2014)

Kellogg et al. (2008)

Future research (2): questions

• How much, how quickly?‣ High-severity patches

‣ Carbon source

‣ Air quality

• Where?‣ Vulnerable landscapes

‣ Thresholds for species and life forms (e.g., forest shrubland)➛

‣ Thresholds for processes (e.g., habitat connectivity)

• What can we do?‣ Resistance (short-term)

‣ Resilience (mid-term)

‣ Adaptation (start now)

?

The end