Fig 1.
Biophysical and anthropogenic determinants of wildfire.
Interactions of wildfire requirements as regulated by biophysical and anthropogenic influences. Anthropogenic determinants are shaded in yellow and biophysical influences in orange.
Table 1.
Description of the 10 variables considered to estimate fire occurrence and burned area.
Fig 2.
Mean fire return interval (mFRI) estimate 1976–2000.
Mean fire return interval (mFRI) estimate 1976–2000 (A), and its change from the historical period (1951–1975) (B). In (a) low mFRIs (red) designate short intervals between fire events, and high mFRI designate long intervals (green). In (B) positive values (green) indicate an increase in mFRI (i.e., decrease in the number of fires) between the estimation and previous period (1976–2000 less 1951–1975), whereas negative values (red) indicate a decreasing mFRI or an increase in the number of fires. (C) Varying constraint map of fire resource and conditions 1976–2000: Four categories of fire resources as proxied by AET, and fire conditions proxied by CWD. Color is determined by the unique combinations of resources and conditions. Where increasingly favorable fire conditions are indicated with deepening red hues, and increasingly favorable resource classes are indicated with deeper hues of blue.
Table 2.
Results from Zero Inflated Negative Binomial Model (Y = Wildfire Count).
Table 3.
P-values from chi-squared test on the difference of log likelihoods.
Fig 3.
Contribution to 25 year expected fire count by natural and human factors.
Contribution of natural (A) and human (B) variables, ceteris paribus, to the expected fire count predicted during the 1976–2000 period. Natural variables are defined as all variants of AET, CWD, lighting, elevation and slope. Human variables include distance to population centers, neighboring maximum housing density, distance to populated places, and distance to campsites.
Fig 4.
Area burned (Km2) by modeling period and climate model.
Grey bars depict the area burned over the historical (1951–1975) and estimation (1976–2001) periods, whereas red and green bars depict the expected area burned into the future under the Geophysical Fluid Laboratory (GFDL) and Parallel Climate Model (PCM), respectively.
Fig 5.
Change in mFRI by 2001–2025 and 2026–2050 under the GFDL climate scenario.
Panel (A)depicts change in mFRI from the estimation period 1976–2000 to the 2001–2025 forecast period, whereas (B) show the expected change in mFRI by 2026–2050. Note that positive values indicate increasing mFRI or fewer fires, and the inverse for negative values.
Fig 6.
Change in MFRI 2001–2025, 2026–2050 PCM compared to 1976–2000.
Change in MFRI for the PCM climate from the base period 1976–2000 to the first forecast period 2001–2025 (A), the change in MFRI from the base period 1976–2000 to the second forecast period 2026–2050 (B). Positive values indicate increasing MFRI or fewer fires, and the inverse for negative values.
Fig 7.
Model agreement for change in MFRI 2001–2025 and 2026–2050 compared to 1976–2000.
Model agreement between the GFDL and PCM A2 scenarios for the (A) 2001–2025 and (B) 2026–2050 periods. Green indicates areas where model estimations from both GFDL and PCM models indicate increasing MFRI, or fewer fires. Red indicates areas where model estimations from both indicate a decreasing MFRI, or more fires. Grey indicates areas of disagreement.