Growing human and ecological costs due to increasing wildfire are an urgent concern in policy and management, particularly given projections of worsening fire conditions under climate change. Thus, understanding the relationship between climatic variation and fire activity is a critically important scientific question. Different factors limit fire behavior in different places and times, but most fire-climate analyses are conducted across broad spatial extents that mask geographical variation. This could result in overly broad or inappropriate management and policy decisions that neglect to account for regionally specific or other important factors driving fire activity. We developed statistical models relating seasonal temperature and precipitation variables to historical annual fire activity for 37 different regions across the continental United States and asked whether and how fire-climate relationships vary geographically, and why climate is more important in some regions than in others. Climatic variation played a significant role in explaining annual fire activity in some regions, but the relative importance of seasonal temperature or precipitation, in addition to the overall importance of climate, varied substantially depending on geographical context. Human presence was the primary reason that climate explained less fire activity in some regions than in others. That is, where human presence was more prominent, climate was less important. This means that humans may not only influence fire regimes but their presence can actually override, or swamp out, the effect of climate. Thus, geographical context as well as human influence should be considered alongside climate in national wildfire policy and management.

Chaparral shrublands burn in large high-intensity crown fires. Managers interested in how these wildfires affect ecosystem processes generally rely on surrogate measures of fire intensity known as fire severity metrics. In shrublands burned in the autumn of 2003, a study of 250 sites investigated factors determining fire severity and ecosystem responses. Using structural equation modeling we show that stand age, prefire shrub density, and the shortest interval of the prior fire history had significant direct effects on fire severity, explaining > 50% of the variation in severity. Fire severity per se is of interest to resource managers primarily because it is presumed to be an indicator of important ecosystem processes such as vegetative regeneration, community recovery, and erosion. Fire severity contributed relatively little to explaining patterns of regeneration after fire. Two generalizations can be drawn: fire severity effects are mostly shortlived, i.e., by the second year they are greatly diminished, and fire severity may have opposite effects on different functional types. Species richness exhibited a negative relationship to fire severity in the first year, but fire severity impacts were substantially less in the second postfire year and varied by functional type. Much of this relationship was due to alien plants that are sensitive to high fire severity; at all scales from 1 to 1000 m2, the percentage of alien species in the postfire flora declined with increased fire severity. Other aspects of disturbance history are also important determinants of alien cover and richness as both increased with the number of times the site had burned and decreased with time since last fire. A substantial number of studies have shown that remote-sensing indices are correlated with field measurements of fire severity. Across our sites, absolute differenced normalized burn ratio (dNBR) was strongly correlated with field measures of fire severity and with fire history at a site but relative dNBR was not. Despite being correlated with fire severity, absolute dNBR showed little or no relationship with important ecosystem responses to wildfire such as shrub resprouting or total vegetative regeneration. These findings point to a critical need for further research on interpreting remote sensing indices as applied to postfire management of these shrublands.

A substantial portion of chaparral shrublands in the southern part of California’s Sierra Nevada Mountain Range has never had a recorded fire since record keeping began in 1910. We hypothesised that such long periods without fire are outside the historical range of variability and that when such areas burn, postfire recovery is weaker than in younger stands. We predicted that long fire-free periods will result in loss of shrub species and deterioration of soil seed banks, which, coupled with higher fire intensities from the greater accumulation of dead biomass, will lead to poorer postfire regeneration. The 2002 McNally Fire burned ancient stands that were as much as 150 years old, as well as much younger (mature) stands. Based on shrub skeletons in the burned area as a surrogate for prefire density, we found that ancient stands change in structure, owing primarily to the loss of obligate seeding Ceanothus cuneatus; other species appear to have great longevity. Despite the reduction in C. cuneatus, postfire shrub–seedling recruitment remained strong in these ancient stands, although some seed bank deterioration is suggested by the three-quarters lower seedling recruitment than recorded from mature stands. Total diversity and the abundance of postfire endemic annuals are two other response variables that suggest that these ancient stands are recovering as well as mature stands. The one area of some concern is that non-native species richness and abundance increased in the ancient stands, suggesting that these are more open to alien colonisers. It is concluded that chaparral more than a century old is resilient to such long fire-free periods and fire severity impacts are indistinguishable from those in younger chaparral stands.

Blue oak woodlands in California have been a focus of conservation concern for many years. Numerous studies have found that existing seedling and sapling numbers are inadequate to sustain current populations, and recent work has suggested that blue oak woodlands might be particularly vulnerable to a warming climate. California has recently experienced a drought of historically unprecedented severity, resulting in the mortality of tens of millions of trees, including an apparent spike in mortality in oak communities. Here we present the results of a survey of tree mortality and composition in blue oak woodlands in Sequoia National Park. We found that 18% (95% CI = 14-24,) of all standing trees and 23% (95% CI = 17-30) of standing Quercus douglasii Hook. & Arn. (blue oak) were dead, substantially higher than proportions of dead trees recorded in pre-drought datasets, which showed 4% (95% CI = 2-9) standing dead for all trees and 5% (95% CI = 4-7) dead or 8% (95% CI = 4-16) standing dead for blue oak. Furthermore, much of this mortality appeared to be recent. Based on foliage or fine twig retention, 19% (95% CI = 14-26) of blue oak and 23% (95% CI = 16-31) of Quercus wislizeni A. DC. (interior live oak) appear to have died recently. In contrast, only 5% (95% CI = 3-8) of Aesculus californica (Spach) Nutt. (California buckeye) and 5% (95% CI = 2-11) of Fraxinus dipetala Hook. & Arn. (California ash) appear to have died recently. Even after such high mortality, with blue oak basal area dropping by 26% (from 9.5 m*/ha [95% CI = 7.4-11.6] to 7.0 m?/ha [95% CI = 5.3-8.7]), blue oak remains the dominant species in these ecosystems. However, given the lack of recruitment and the apparent vulnerability to extreme drought, blue oak populations may be at risk for severe decline if such mortality events become more frequent.

No abstract