The unusual U.S. climatic conditions of the historic year called “The Big Burn” were not matched until the devastating fire year of 2012.
The Great Fire of 1910 (also commonly referred to as the Big Blowup or the Big Burn) was a wildfire that burned about three million acres (12,000 km2, approximately the size of Connecticut) in northeast Washington, northern Idaho (the panhandle), and western Montana. The area burned included parts of the Bitterroot, Cabinet, Clearwater, Coeur d'Alene, Flathead, Kaniksu, Kootenai, Lewis and Clark, Lolo, and St. Joe National Forests. The firestorm burned over two days (20 and 21 August 1910) and killed 87 people, including 78 firefighters. It is believed to be the largest, although not the deadliest, fire in recorded U.S. history. The “Big Blowup” in the summer of 1910 was a singular event in the history of the U.S. Forest Service, shaping fire management strategies and policies from that time to today (Pyne 2008; Egan 2009).
Relatively dry conditions prevailed across the western United States during the spring and summer months of that year, but the northern Rocky Mountains were especially dry. Prior to the establishment of the U.S. Forest Service (in 1905) land use practices on public and private forestlands were commonly laissez-faire, resulting in over-harvesting of timber and accumulated slash fuels in some areas. During and following timbering activity, people set fire to the slash to dispose of it. Accidental fires were also common, especially from sparks along railways from wood-burning locomotives. The fledgling Forest Service had a tiny force of rangers with the responsibility for detecting and suppressing wildfires over enormous and remote areas. The national forests increased by 16 million acres in 1907 by executive order of Theodore Roosevelt in the last days of his presidency (Egan 2009). Given the regional dryness, abundant slash fuels near frontier logging settlements, ubiquitous ignitions from human sources, and the lack of fire detection and firefighting capacity by the Forest Service, the conditions were ripe for the Big Blowup.
Although there have been various descriptions of the human and natural history of this episode (Pyne 2008; Egan 2009), and some climatological analyses encompassing the 1910 year and this region (e.g., Morgan et al. 2008), we show here that warm weather conditions in 1910 were highly anomalous, and more so than previously reported or evaluated. Further, we identify and evaluate analogous and different spatial climate conditions in other large regional fire years in the northern Rockies, including spring 2012 (as we write this paper), and we discuss the implications of these observations and patterns for upcoming fire seasons. Last, we also illustrate an example of extensive wildfire and drought synchrony across western North America during the eighteenth century using a recently compiled network of tree-ring-based reconstructions.
DATA SOURCES.
Climate data were accessed from the National Oceanic and Atmospheric Administration (NOAA)'s National Climatic Data Center in Asheville, North Carolina (www.ncdc.noaa.gov/climate-monitoring/index.php) and the Earth System Research Laboratory of NOAA in Boulder, Colorado (www.esrl.noaa.gov/psd/psd1/). Data accessed included U.S. surface temperature and precipitation, Palmer drought severity index (PDSI) data, and upper-level data from the National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) reanalysis (NRA) (Kistler et al. 2001) and the historical reanalysis (HRA) dataset (Compo et al. 2006Compo et al. 2011). Tree-ring width and fire scar data used to illustrate drought and fire synchrony come from the North American Drought Atlas dataset (Cook et al. 2004), and the International Multiproxy Paleofire Database (www.ncdc.noaa.gov/paleo/impd/paleofire.html) and T. W. Swetnam et al. (2011, unpublished manuscript).
ANALYSIS RESULTS.
Plummer's (1912) map of the 1910 fires across the western United States (Fig. 1) illustrates the widespread occurrence of wildfires in that year over the whole region. The largest burned areas were in the northern Rockies and particularly in Idaho, where most fatalities occurred. March 1910 was an exceptionally warm month, as illustrated by the spatial pattern and magnitude of the temperature departure from the long-term average and the time series of area-weighted mean temperature over the contiguous United States (Fig. 2). The warmth during March 1910 was not exceeded in the climate record until 2012, and it was particularly unusual in the early part of the twentieth century, when generally cooler temperatures prevailed in the United States. These extreme warm conditions likely contributed to the extent and magnitude of the wildfires during that year.
Map of the occurrence and extent of the 1910 wildfires across the western United States (after Plummer 1912).
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
(a) Map of surface temperature anomaly (°F, for Mar 1910 for the contiguous United States) and (b) time series of Mar mean temperature anomalies (°F) for the period of record. Red line highlights the value for 1910. Note that the monthly record stood for 102 years until Mar 2012. (Source: National Climatic Data Center, NOAA.)
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
The extreme warm conditions during March 1910 set the stage for the great fires later that summer. (We will discuss this interpretation and likely mechanisms in more detail later in this paper, along with findings from other studies showing seasonal temperature and drought associations with wildfire activity in the western United States.) The anomaly field of the 500-mb geopotential height surface based on the HRA dataset is consistent with the record warmth recorded for the Lower 48 during that month (Fig. 3). The anomalous warmth persisted throughout the nominal spring season of March–May (Fig. 4), which was then followed by a rather dry summer, particularly in the state of Idaho, with departures exceeding minus two standard deviations (Fig. 5).
Map of the anomalous 500-mb geopotential height field (m) for Mar 1910 (Compo et al. 2006). (Data source: HRA.)
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
(a) As in Fig. 2a, but for the 3-month average of Mar–May 1910. (b) As in Fig. 2b, but for Mar–May averages. Spring temperature record of 1910 was broken by a substantial margin in 2012.
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
(left) Standardized precipitation anomalies for the summer season (Jun–Aug) of 1910. (right) PDSI for Aug 1910. Note drought index values of −3 and below in Idaho.
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
We examined several sequences of daily weather maps available from the NOAA Central Library site (http://docs.lib.noaa.gov/rescue/dwm/data_rescue_daily_weather_maps.html). Daily warm anomalies exceeding 10°C were common during the 1910 spring months. The generally warm and dry 6-month period preceding the wildfire is consistent with drought conditions present in the Northwest (Idaho, Oregon, and Washington) at the end of the summer of 1910 (Fig. 5). Interestingly, a sharp moisture gradient across the northern Rockies is also indicated, as moist conditions were prevalent in Montana at that time. However, humidity was generally below normal during the summer months in the region of the Northwest affected by the fires. This is evident in the July–August 1910 surface relative humidity anomaly field in the HRA dataset (Fig. 6), indicating drier-than-normal conditions in the West in general. A similar map computed for the months of July and August of 1910 at the 850-mb surface (close to the average ground elevation for Idaho, map not shown) gives essentially the same picture. However, during the days when the wildfires were at their peak, around the third week of August, relative humidity was extremely low, with values in the areas most affected around 20% or lower (top two panels in Fig. 7).
Map illustrating the anomalous surface relative humidity field (%) for Jul–Aug 1910 (Compo et al. 2006). (Data source: HRA.)
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
(top left) Mean relative humidity; (top right) corresponding anomaly field (%) at 800 mb; (bottom left) mean vector wind; and (bottom right) anomaly field (m s–1) at 750 mb for the period 20–22 Aug 1910 (Compo et al. 2006). (Data source: HRA.)
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
One other factor that appeared to be important in the extent and rate of increase in the size of the wildfires during the Big Burn was the occurrence of anomalously high wind speeds. Figure 7 (bottom two panels) shows the pattern of anomalous vector mean winds at 750 mb (~2,300 m—7,500 ft more typical of the mountainous terrain that experienced the most widespread burning) during the same time interval in August when the wildfires flared up (Pyne 2008). The daily weather maps from this period in August 1910 clearly show the passage of a disturbance and frontal system through the region. Figure 8 illustrates the strong sea level pressure gradient present over western North America for 21 August 1910. It is inferred that preexisting dry soils and woody fuels from the long string of exceptionally warm months, with relatively low humidity and stronger-than-normal near-surface winds during the month of the fires were critical factors leading to the exceptional nature of the Big Burn forest fires.
Sea level pressure map for 21 Aug 1910. A strong pressure gradient of ~20 hPa across the Northwest is evident (Compo et al. 2006). (Data source: HRA.)
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
Longer-term perspectives of climate and fire conditions across the western United States are provided by tree-ring reconstructions. Summer and other seasonal drought reconstructions based on extensive networks of tree-ring width chronologies have been very useful in recent years in identifying and mapping westwide drought years and decades (Cook et al. 2004). Fire history reconstructions from networks of thousands of tree-ring dated, fire-scarred trees in this region also show a regional synchrony of extensive fire years and westwide drought years (e.g., Kitzberger et al. 2007; Williams et al. 2012). Recent compilations of the largest fire scar data network yet assembled (T. W. Swetnam et al. 2011, unpublished manuscript) from 1,248 sites in western North America confirm this pattern. The most extensive fire year in the past 250 years in this network was 1748, which corresponds with westwide drought, as indicated in the North American Drought Atlas (http://iridl.ldeo.columbia.edu/SOURCES/.LDEO/.TRL/.NADA2004/.pdsi-atlas.html) reconstructions (Fig. 9). Further study of these spatial and temporal paleofire and climate reconstructions (including new, gridded temperature reconstructions; e.g., Wahl and Smerdon 2012) may help identify interannual to decadal patterns useful for improving our understanding of the fire climatology of western North America.
Western North American fire scar chronology network includes 1,248 sites, where typically 10 or more trees were sampled and dated at each site, providing exact fire dates and percentages of trees scarred each year over the past 400 years (T. W. Swetnam et al. 2011, unpublished manuscript). Year 1748 is the most active fire year, with >200 sites recording fire in that year. 1910 was a relatively light fire year in this reconstruction because fire suppression and livestock grazing were already reducing surface fire extent.
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
DISCUSSION AND CURRENT CONTEXT.
There are two reasonably close analogs to the anomalous spring and summer of 1910 that are notable. One occurred in 1988 and was associated with extreme summer drought in the Midwest United States and extensive fires in the Yellowstone National Park (YNP) region in the northern Rockies (Balling et al. 1992). The 1988 fire was one of the largest wildfire events in the recorded history of YNP. As in most major wildfire episodes, smaller individual fires quickly grew out of control with increasing winds aided by severe drought and combined into one large conflagration, which burned for several months. A total of 793,880 acres (3,213 km2), or 36% of the park, were affected by the wildfires (Schullery 1989). Figure 10 illustrates the anomalous anticyclonic conditions prevailing from late spring through summer of 1988. The anomalous ridging centered in midcontinent led to extreme drought from the upper Midwest to the northern Rockies (Fig. 11) and to extreme forest fire conditions in the northern Rockies.
As in Fig. 3, but for the period of May–Aug 1988.
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
PDSI values for Jul–Aug 1988.
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
As noted above, by an interesting coincidence, during the writing of this article an extreme monthly anomaly for the month of March, strikingly similar in magnitude to that of March 1910, occurred in 2012 (Fig. 12, top panel; cf. Fig. 3). Substantially above-normal temperatures continued through the rest of the spring months of 2012 and major forest fires broke out with high loss of property and some lives in New Mexico, Colorado, and Utah. For comparison, the 500-mb anomaly for March–May 2012 (Fig. 12, bottom panel) and the associated PDSI values for May 2012 (Fig. 13, top panel) are presented.
(top) As in Fig. 3, but for Mar 2012. (bottom) As in (top), but for the nominal spring season (Mar–May) of 2012.
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
PDSI for (top) May and (bottom) Jul 2012.
Citation: Bulletin of the American Meteorological Society 94, 9; 10.1175/BAMS-D-12-00150.1
The anomalous warm March 2012 set numerous records across the United States, and edged out 1910 as the warmest year since the start of state records in 1895 by about 0.5°F. Considering the warming of about 1°F over the last 100 years for the contiguous United States, it is likely that March 1910 represents a larger deviation from the prevailing mean than March 2012. March 2012 was followed by much above-average temperatures in April and May, with spring of 2012 again beating the previous record set in 1910 (Fig. 4). As documented here, there is considerable resemblance in the anomalous upper-level circulation and surface patterns between the spring and summer of 1910 and 2012; also the fast start and large extent of the wildfire season in 2012 appear to be more than the equal of the 1910 Big Blowup, though thankfully without the large loss of life. The summer months of 2012 (as we write this in late August), however, are playing out differently from 1910, with continued and even increasing extreme drought conditions in the central Rockies region (and extending through the Midwest; see Fig. 13, bottom panel). Meanwhile, the Pacific Northwest is one of the few areas in the country to have reached midsummer with near-normal moisture.
Wildfire activity in montane forests of the western United States has greatly increased in recent years (since circa 1986), as measured by the number of large events (more than a fourfold increase in the number of fires >400 ha in size) and increasing length of the fire season (more than 75 days' increase; Westerling et al. 2006; Swetnam and Betancourt 1998). Fire severity (i.e., proportion of burned area causing tree mortality) has also increased in some subregions and vegetation types (Holden et al. 2007; Williams et al. 2010; Dillon et al. 2011). Moreover, the largest wildfires in state histories (>100 years) have occurred in Arizona, New Mexico, and Colorado in just the past two years (2011 and 2012). In some cases, these record-breaking wildfires have exceeded the previous largest documented wildfires (before 1980) by an order of magnitude. It is likely that increasing areas burned and higher severity fires in some low- and midelevation drier forest types (such as ponderosa pine forests, dry mixed conifer forests, and pine savannas) are associated with a combination of factors. These may include disruption of frequent surface fire regimes that previously existed in some low- and midelevation forest types due to firefighting by government agencies and extensive livestock grazing around the turn of nineteenth to twentieth centuries (Arno 1980; Steele et al. 1986; Keane et al. 1990; Allen et al. 2002; Heyerdahl et al. 2008). Higher-elevation forests (such as mesic lodgepole pine and spruce–fir stands) have probably not been altered by twentieth-century fire suppression effects because these forests generally did not sustain surface fires, and burned only at 100–150-plus-year intervals in the past (Schoennagle et al. 2004).
A comparison of the 1910 “blow up” event to modern extreme wildfire events is useful because it also was associated with preceding seasonal climate conditions (i.e., warm springs extending into warm summers), resulting in wildfires exceeding in size all previous known events (and still the largest fires on record in Idaho forest areas). The fact that antecedent conditions are important suggests there is potential for developing predictive models (e.g., Westerling et al. 2003) at least a season ahead. Interestingly, 1910 was also a westwide high fire occurrence year (see Fig. 1; Plummer 1912), suggesting that changes in fuels and forest structure may have been less important than during recent events, which are taking place following a century of fire suppression efforts, other land use practices, and the extensive spread of invasive grass species in some areas.
Our findings generally concur with other studies of seasonal climate–wildfire patterns in the western United States, wherein warmer springs, reduced snowpacks, and consequent longer drying periods leading into the peak wildfire season (from mid- to late summer) are especially important factors in extensive wildfire outbreaks in montane forests of the central and northern Rockies (Balling et al. 1992; Westerling et al. 2006; Morgan et al. 2008; Littell et al. 2009). The drying (and subsequent combustion) occurs across a broad range of scales, from tree needles and grasses to small branches, whole tree stems (logs and snags), and entire forest canopies and watersheds. Moisture content of dead fuels is particularly important in fire ignition and initial spread rates; however, live fuel moistures (e.g., tree needle moisture content) may be a more important factor in some “crown fire” type conflagrations, where very high-intensity burning occurs as a consequence of increased volatility of these fuels.
Reduced live fuel moistures can be important in promoting crown fire behavior and are an optional variable in some models (e.g., Scott and Reinhardt 2001), but the precise (and changing) relationships between live fuel moisture and crown fire behavior are mostly theory based (rather than observed and calibrated). Presumably this is because of the difficulties in measuring live fuel moisture content and related fire intensities (and spread rates, etc.) at the requisite scales and in different forest types. Spectacular burning of combustible gases emitted from burning forest canopies are commonly visible in intense conflagrations as briefly burning vertical or tilted shafts of flame, or longer-lasting “fire whirls,” reaching heights of 100 m or more above canopies (Forthofer and Goodrick 2011). These observations indicate that combustible gases can be emitted in large quantities and ignited as living tree canopies are heated and burned, thereby extending flaming fronts (especially when wind driven) and contributing to very rapid spread rates. We hypothesize that the extreme warming and drying conditions over periods of months and seasons (as in 1910 and 2012), which cause lower live fuel moistures at leaf to landscape scales, in turn lead to greater volatility (and hence combustibility) of fuels at all scales. This is not a particularly new idea (e.g., Simard and Donoghue 1987), but we emphasize it here as potentially a factor of greater importance than previously recognized (or modeled) in triggering extraordinary wildfire extent.
Another factor of common importance in driving large wildfire events, both in 1910 and during many recent very large (>50,000 ha) wildfires, are surface and near-surface winds. The first-person accounts of the 1910 fires include many lurid descriptions of high winds blowing fire and large burning embers far in advance of the burning front. Likewise, winds were a key factor during the largest southwestern wildfires in recent years, with maximum burn rates of more than 20,000 ha in less than 24 hours (e.g., Rodeo– Chediski Fire 2002, Wallow Fire 2011, Las Conchas Fire 2011). These very rapid fire runs resulted in total or near-total tree canopy mortality in long, linear strips, aligned with prevailing winds and extending for 15 km or more (T. W. Swetnam et al. 2011, unpublished manuscript). An important distinction here is that “wind driven” fire runs may be associated with synoptic weather patterns (e.g., passage of frontal systems, or “jet” wind currents at the surface or near surface; e.g., Crimmins 2006; Wirth 2011) versus “plume dominated” fires with runs caused by local, down-drafting (and horizontal) winds from collapsing pyro-convection columns (Potter 2011). Both types of wind-related fire behaviors typically occurred during recent large wildfires because they lasted days or weeks, encompassing both kinds of conditions. There is a need for much more study of the role of synoptic climate patterns in generating wind profiles that are characteristic of extreme crown fire behaviors.
Very large wildfire events, like those of 1910 and recent years, are a consequence of many contributing factors operating across a broad range of spatial and temporal scales. The extended multivariate ENSO index (MEI) series available from NOAA's Earth System Research Laboratory (ESRL; Wolter and Timlin 2011) indicates the occurrence of a strong La Niña during 1910, and it may have contributed to the strong and persistent anomalous anticyclonic pattern over North America in the spring and summer of that year. Weather events (such as regional surface winds) are superimposed upon seasonal and longer climate patterns (such as drought and secular warming), just as forest fuel drying and changing volatility/combustibility conditions are superimposed upon long-term trends of forest fuel growth and accumulation (e.g., the effects of a century of surface fire suppression, or recently induced tree mortality from bark beetle and drought). Improved understanding of the slower, long-term changes and rapid, extreme events we are witnessing today, and potentially predicting them in the future, will depend in large measure upon how thoroughly we can identify, measure, and model these interacting factors. It is increasingly apparent that spring temperature conditions are likely to be a key variable in such models of wildfire hazard and behavior in montane forest areas of western North America.
ACKNOWLEDGMENTS
The authors thank three anonymous reviewers and the editor for their helpful comments. Partial support for TWS's time was provided by the U.S. interagency Joint Fire Sciences Program. We thank E. Bigio, M. Hall, E. Vasquez, and D. Falk at The University of Arizona for help in compiling the North American fire scar chronology network shown in Fig. 9. We also thank the dozens of data contributors to this North American fire scar network, which will be described and analyzed in detail in forthcoming multiauthored papers.
REFERENCES
Allen, C. D., and Coauthors, 2002: Ecological restoration of southwestern ponderosa pine ecosystems: A broad perspective. Ecol. Appl., 12, 1418–1433.
Arno, S. F., 1980: Forest fire history in the northern Rockies. J. For., 78, 460–465.
Balling, R. C., Jr., G. A. Meyer, and S. G. Wells, 1992: Relation of surface climate and burned area in Yellowstone National Park. Agric. For. Meteor., 60, 285–293.
Compo, G. P., J. S. Whitaker, and P. D. Sardeshmukh, 2006: Feasibility of a 100-year reanalysis using only surface pressure data. Bull. Amer. Meteor. Soc., 87, 175–190.
Compo, G. P., and Coauthors, 2011: The Twentieth Century Reanalysis Project. Quart. J. Roy. Meteor. Soc., 137, 1–28.
Cook, E. R., C. A. Woodhouse, C. M. Eakin, D. M. Meko, and D. W. Stahle, 2004: Long-term aridity changes in the western United States. Science, 306, 1015–1018.
Crimmins, M. A., 2006: Synoptic climatology of extreme fire-weather conditions across the Southwest United States. Int. J. Climatol., 26, 1001–1016.
Dillon, G. K., Z. A. Holden, P. Morgan, M. A. Crimmins, E. K. Heyerdahl, and C. H. Luce, 2011: Both topography and climate affected forest and woodland burn severity in two regions of the western US, 1984 to 2006. Ecosphere, 2, Article 130, http://dx.doi.org/10.1890/ES11-00271.1.
Egan, T., 2009: The Big Burn: Teddy Roosevelt and the Fire That Saved America. Houghton Mifflin Harcourt, 336 pp.
Forthofer, J. A., and S. L. Goodrick, 2011: Vortices and wildland fire. Synthesis of knowledge of extreme fire behavior: Volume I for fire managers, P. A. Werth et al., Eds., Pacific Northwest Research Station General Tech. Rep. PNW-GTR-854, U.S. Department of Agriculture, Forest Service, 89–105.
Heyerdahl, E. K., P. Morgan, and J. P. Riser II, 2008: Crossdated fire histories (1650 to 1900) from ponderosa pine–dominated forests of Idaho and western Montana. Rocky Mountain Research Station General Tech. Rep. RMRS-GTR-214WWW, U.S. Department of Agriculture, Forest Service, 83 pp.
Holden, Z. A., P. Morgan, M. A. Crimmins, R. K. Steinhorst, and A. M. S. Smith., 2007: Fire season precipitation variability influences fire extent and severity in a large southwestern wilderness area, United States. Geophys. Res. Lett., 34, L16708, doi:10.1029/2007GL030804.
Keane, R. E., S. F. Arno, and J. K. Brown, 1990: Simulating cumulative fire effects in ponderosa pine/Douglas-fir forests. Ecology, 71, 189–203.
Kistler, R., and Coauthors, 2001: The NCEP–NCAR 50- Year Reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82, 247–267.
Kitzberger, T., P. M. Brown, E. K. Heyerdahl, T. W. Swetnam, and T. T. Veblen, 2007: Contingent Pacific-Atlantic Ocean influence on multi-century wildfire synchrony over western North America. Proc. Natl. Acad. Sci. USA, 104, 543–548.
Littell, J. S., D. L. McKenzie, D. L. Peterson, and A. L. Westerling, 2009: Climate and wildfire area burned in the western U.S. ecoprovinces, 1916–2003. Ecol. Appl., 19, 1003–1021.
Morgan, P., E. K. Heyerdahl, and C. E. Gibson, 2008: Multi-season climate synchronized forest fires throughout the 20th century, northern Rockies, U.S.A. Ecology, 89, 717–728.
Plummer, F. G., 1912: Forest fires: Their causes, extent and effects, with a summary of recorded destruction and loss. USDA Forest Service Bull. 117, U.S. Government Printing Office, 39 pp.
Potter, B. E., 2011: Column/plume dynamics. Synthesis of knowledge of extreme fire behavior: Volume I for fire managers, P. A. Werth et al., Eds., Pacific Northwest Research Station General Tech. Rep. PNW-GTR-854, U.S. Department of Agriculture, Forest Service, 73–80.
Pyne, S. J., 2008: Year of the Fires: The Story of the Great Fires of 1910. Mountain Press Publishing Company, 352 pp.
Schoennagle, T., T. T. Veblen, and W. H. Romme, 2004: The interaction of fire, fuels, and climate across Rocky Mountain forests. Bioscience, 54, 661–676.
Schullery, P., 1989: Yellowstone fires: A preliminary report. Northwest Sci., 63, 44–54.
Scott, J. H., and E. D. Reinhardt, 2001: Assessing crown fire potential by linking models of surface and crown fire behavior. Rocky Mountain Research Station Research Paper RMRS-RP-29, U.S. Department of Agriculture, Forest Service, 59 pp.
Simard, A. J., and L. R. Donoghue, 1987: Wildland Fire Prevention: Today, Intuition--Tomorrow, Management. Proc. Symp. on Wildland Fire 2000, South Lake Tahoe, California, U.S. Department of Agriculture, 187–198.
Steele, R., S. F. Arno, and K. Geier-Hayes, 1986: Wildfire patterns change in central Idaho's Ponderosa Pine-Douglas-fir forest. Western J. Appl. Forestry, 1, 16–18.
Swetnam, T. W., and J. L. Betancourt, 1998: Mesoscale disturbance and ecological response to decadal climatic variability in the American Southwest. J. Climate, 11, 3128–3147.
Wahl, E. R., and J. E. Smerdon, 2012: Comparative performance of paleoclimatic field and index reconstructions derived from climate proxies and noise-only predictors. Geophys. Res. Lett., 39, L06703, doi:10.1029/2012GL051086.
Westerling, A. L., A. Gershunov, T. J. Brown, D. R. Cayan, and M. D. Dettinger, 2003: Climate and wildfire in the western United States. Bull. Amer. Meteor. Soc., 84, 595–604.
Westerling, A. L., H. G. Hidalgo, D. R. Cayan, and T. W. Swetnam, 2006: Warming and earlier spring increase western U.S. wildfire activity. Science, 313, 940–943.
Williams, A. P., C. D. Allen, C. I. Millar, T. W. Swetnam, J. Michaelsen, C. J. Still, and S. W. Leavitt, 2010: Forest responses to increasing aridity and warmth in the southwestern United States. Proc. Natl. Acad. Sci. USA, 107, 21 289–21 294.
Williams, A. P., and Coauthors, 2012: Temperature as a potent driver of regional forest drought stress and tree mortality. Nat. Climate Change, 3, 292–297, doi:10.1038/nclimate1693.
Wirth, P. A., 2011: Critical fire weather patterns. Synthesis of knowledge of extreme fire behavior: Volume I for fire managers, P. A. Werth et al., Eds., Pacific Northwest Research Station General Tech. Rep. PNW-GTR-854, U.S. Department of Agriculture, Forest Service, 25–48.
Wolter, K., and M. S. Timlin, 2011: El Niño/Southern Oscillation behaviour since 1871 as diagnosed in an extended multivariate ENSO index (MEI.ext). Int. J. Climatol., 31, 1074–1087.
