Contemporary wildfires in Sonoran desert scrub are mostly ignited by humans [7,49,112], tend to be larger and more widespread following periods of above-average precipitation (e.g., [7,44,112]), and are often and increasingly fueled by nonnative invasive plants [26,28,44,48,54,68,104,109,114]. Buffelgrass and red brome are the most common nonnative plant fuels in Sonoran desert scrub communities, although other nonnative plants may be locally abundant and/or widespread in adjacent communities, and thus also responsible for increasing the likelihood of fire ignition and spread on sites where dense and continuous populations of these plants occur.

Although the details of presettlement fire regimes in North American deserts remain uncertain, available evidence suggests that the size and frequency of contemporary wildfires are outside the historical range of variability for Sonoran desert scrub communities. As fires become more common in desert scrub communities, fuel characteristics may be further altered by postfire changes in plant community composition and structure, with a greater dominance of nonnative, fire-adapted grasses, and thus increasing the probability of additional fires (see Fire Management Considerations: Invasive Plant/Fire Cycle).

Role of Nonnative Invasive Plants

Disturbances from human settlements and land use activities coupled with an increase in cool-season precipitation variability over the past half century have facilitated invasion by nonnative plants on many sites throughout the Sonoran and other North American deserts [1,58]. Nonnative plant invasions can have a variety of impacts. Of most concern to fire managers are changes in fuel characteristics and fire regimes. Other impacts might include decreases in total native herb canopy, herb species richness, shrub density and canopy, and insect and bird diversity, as was found on sites invaded by lovegrass in southeastern Arizona (Bock et al. 1986, cited in [54]).

Nonnative invasive plants can add abundant and novel fuels to desert scrub ecosystems, increasing fuel biomass and continuity, and thus increasing the potential for uncharacteristically large and frequent wildfires. By altering fuel characteristics, nonnative plant invasions—especially grasses—can increase the range of conditions favorable for fire ignition and increase the potential for fire spread and high fire intensity [28,55,71,83,104,114,118]. High-intensity fires may be more likely than low-intensity fires to severely damage or kill cacti [83,118] and other dominant native desert scrub perennials and may also limit resprouting (see Contemporary Fire Severity). Nonnative invasive plants may also increase the likelihood of repeated fire by recovering quickly and providing sufficient fuels to carry another fire soon after burning [28,55,71]. The long-term effect of nonnative plant invasions on the seasonality, frequency, and severity of desert wildfires threatens the sustainability of desert ecosystems on many sites [28].

Species of Concern

Photo Credit

A) Photo © Eric Hough, some rights reserved (CC BY-NC). B) U.S. Geological Survey photo by Molly McCormick. C and D) Photo © Steve Jones, some rights reserved (CC BY-NC).

Figure 6—Invasive annual grasses cover large areas in the Sonoran Desert. A) A cluster of red brome plants. B) Dried red brome plants in desert scrub interspaces. C) A cluster of common Mediterranean grass plants. D) A large, dense patch of common Mediterranean grass plants. 

Nonnative species of greatest concern with regard to altered fuel characteristics and increases in fire occurrence in Sonoran desert scrub communities include the annual grasses red brome and Mediterranean grass (fig. 6), the perennial grasses buffelgrass, fountaingrass (especially, crimson fountaingrass) (fig. 7), and lovegrass (Lehmann lovegrass and weeping lovegrass), and the annual forbs Sahara mustard and arugula (fig. 8) [28,54,71,103,104,114], although several other nonnative species, including soft feather pappusgrass, are spreading and of increasing concern (e.g., [16,54,77,130,134]). Presence of red brome, Mediterranean grass, buffelgrass, fountaingrass, and lovegrass has been linked, at least anecdotally, to fire occurrence in Sonoran desert scrub communities [55,103,104] and other Sonoran and Mojave desert ecosystems [44,48,55,71,103,114]. Of these, red brome and buffelgrass are most widespread and problematic in Sonoran desert scrub [28,44,48,49,71,121]. Other species may be abundant in localized areas of Sonoran desert scrub (e.g., arugula [28,44], lovegrass [103], fountaingrass [54,134], and soft feather pappusgrass [16,77,130]) or in adjacent creosotebush communities (e.g., Mediterranean grass [28,48,88] and Sahara mustard [28,63]) and could become more widespread [28,44].

The invasive annual grasses red brome and Mediterranean grass are the most abundant plants below 1,250 m elevation over large areas in the northern Sonoran Desert. Red brome is especially invasive in the Arizona Upland subdivision where Sonoran desert scrub communities are common, while Mediterranean grass is more invasive in creosotebush communities in the Lower Colorado River Valley subdivision. Both can dominate herbaceous layers and have the potential to establish more horizontally continuous, flammable, and persistent fuel beds than native annual plants [28,48,104].

Photo Credit

National Park Service photos.

Figure 7—Nonnative invasive perennial grasses form dense patches of fine fuels. A) Buffelgrass, and B) fountaingrass under Sonoran paloverde–mixed cacti desert scrub in Saguaro National Park.

Nonnative invasive perennial grasses are increasingly prevalent in the monsoonal regions of the Sonoran Desert, and land managers and scientists share a growing concern over their spread and effects on fuel characteristics and fire regimes in invaded communities [47,134]. Buffelgrass, for example, is especially invasive in the Arizona Upland Subdivision, often in Sonoran desert scrub communities [8,44,77,128]. Buffelgrass populations in these communities have rapidly increased in extent since its introduction [49], filling in bare areas between native plants and spreading rapidly to displace native vegetation [83,95,96]. Rapid spread of buffelgrass has been observed in Saguaro National Park [83,95,96], in the Santa Catalina Mountains [95], at the Desert Laboratory in Tucson [19], and at Organ Pipe Cactus National Monument [111]. See the FEIS Species Review on buffelgrass for more details [63]. Crimson fountaingrass, lovegrass, and soft feather pappusgrass also have the potential to invade Sonoran desert scrub communities and alter fuel characteristics [16,130]. Crimson fountaingrass is an ornamental perennial and a close relative of buffelgrass. It has escaped cultivation and is abundant and widespread at Organ Pipe Cactus National Monument, and it has spread into both districts of Saguaro National Park [54,134]. Lovegrass was widely used for postfire reclamation on the Tonto National Forest in the 1960s and 1970s, and subsequently established in saguaro communities on the Tonto National Monument, where they likely helped carry repeated fires [103]. Soft feather pappusgrass was first observed in the Finger Rock Canyon Watershed of the Santa Catalina Mountains, Arizona, in 1990, and has since been increasing rapidly in distribution and abundance [16].

Photo Credit

A) Photo © Kevin W Smith, some rights reserved (CC BY-NC). B) Photo © Eric Hough, some rights reserved (CC BY-NC). C) Photo © kdouglas8787, some rights reserved (CC BY-NC). D) Photo © Mike Plagens, some rights reserved (CC BY-NC).

Figure 8—Nonnative invasive annual forbs can form dense patches of fine fuels in the Sonoran Desert. A) A Sahara mustard plant. B) Dense Sahara mustard plants under a large shrub. C) Scattered, small arugula plants. D) A tall, dense patch of arugula in desert scrub. 

Relationships with Fire Regimes

Sonoran desert scrub communities with an invasive grass component may have greater fine-fuel density and continuity and a higher likelihood of frequent, high-intensity, and/or large fires than uninvaded communities [71]; however, data supporting this relationship are limited to a few species and locations [10] and are lacking for Sonoran desert scrub communities in particular. A 2020 meta-analysis of published literature on the effects of native versus nonnative species on fire frequency, fire intensity, flammability, fuels quantity, and fire spatial extent showed “a strongly significant difference between fire metrics associated with nonnative and native species, with nonnative species linked to enhanced fire effects and risk.” However, this analysis also highlighted the limited scope of our understanding—only a small number of species (22) and ecosystems had been quantitatively examined. Only 30 studies met criteria for inclusion in the analysis, and only two included information relevant to Sonoran desert scrub [10].

While anecdotal evidence suggests a strong potential for negative impacts that may not yet be detectable, an analysis of the effects of 12 invasive grass species on fire regime characteristics in U.S. ecoregions showed significant impacts for some but not all invasive species in the Sonoran Desert. In the Sonoran Basin and Range and Sonoran Desert ecoregions, fire was significantly more likely to occur in areas with invasive populations of buffelgrass than in uninvaded areas, and fire frequency in buffelgrass-invaded areas was more than twice that in uninvaded areas. Also in those ecoregions, fire was significantly more likely to occur in areas invaded by common Mediterranean grass than in uninvaded areas [55]; however, that most likely reflects fires in creosotebush-white bursage desert scrub, where Mediterranean grass is most invasive [28,48,88]. Across five ecoregions, including the Sonoran Basin and Range (but excluding the Sonoran Desert ecoregion, where red brome abundance was limited) no significant differences in fire occurrence or fire size was detected between areas invaded by red brome and uninvaded areas. Insufficient spatial data were available to examine effects of Lehmann lovegrass, weeping lovegrass, and crimson fountaingrass on fire regime characteristics [55].

Although data confirming effects on fire regime characteristics are lacking for red brome, lovegrass, and crimson fountaingrass [55], observations suggest that these grasses have helped fuel wildfires that burned areas of Sonoran desert scrub. Anecdotally, reported impacts of red brome—primarily from the Mojave Desert—include increased fuel load and persistent flammability; and increased frequency, occurrence, and spread of low-intensity fire (plants are 0.1–0.7 m tall) (e.g., [24,25]). Observations also indicate that several wildfires in Sonoran desert scrub since 1979 were fueled by red brome, including the 1994 Mother’s Day Fire in Saguaro National Park, which burned 138 ha of Arizona Upland desert scrub [47,49]; and a 6,500-ha fire in the Harcuvar Mountains in 1999 [48]. Lehmann lovegrass, weeping lovegrass, and crimson fountaingrass are reported to be fire promoters in some ecosystems. For example, repeated wildfires in upper Cave Creek on the Tonto National Monument were likely fueled by lovegrass that spread onto the monument from the adjacent Tonto National Forest, where it had been planted to stabilize soils after the 1964 Schultz Fire [103].

Buffelgrass-fueled fires can be exceptionally detrimental to Sonoran desert scrub communities because they burn at high temperatures and intensities (see Contemporary Fire Severity), spread quickly, and can occur any time of year [49,83]. In addition, buffelgrass recovers quickly after fire, so it has the potential to carry very frequent fires (see Fire Management Considerations: Invasive Plant/Fire Cycle).

Contemporary Fuel Characteristics

Fine surface fuel layers in contemporary desert scrub communities are increasingly dominated by nonnative invasive grasses and forbs, which provide novel fuels and contribute to fine fuel load, height, continuity, and persistence on many sites in the Sonoran and Mojave deserts [2,28,47,49,83,88,112,114]. Across 2,530 plots in the Sonoran Desert in southwestern Arizona, plots with nonnative plants had higher total fine fuel biomass estimates (about 33 kg/ha) than plots with only native plants (about 9 kg/ha). Biomass estimates were extremely low across the study area for both native and nonnative plants because field work took place during years with below-average precipitation (2011 and 2012) [114]. Sonoran desert scrub sites invaded by perennial buffelgrass are likely to have considerably higher fine fuel biomass than uninvaded sites or sites dominated by native and nonnative annuals (table 2), although data are limited. Nonnative herbaceous fuels may burn during different seasons and at greater intensity and spread fires faster and farther than native fuels, although fuel characteristics differ among nonnative species, and information on this topic is lacking.

Surface fine fuel beds in uninvaded Sonoran desert scrub communities are typically dominated by annual plants, therefore fine fuel loads and continuity can change dramatically from year to year, depending on precipitation patterns [114]. Nonnative annual plant invasions may amplify this fuel–climate relationship in the Sonoran Desert [44,48,58]. Like native annual plants in desert scrub communities, nonnative annual plants can establish dense populations in understories and interspaces following periods of above-average precipitation [7,48,104,112,114,121] and form nearly continuous beds of fine surface fuels that dry out in late spring and persist during the summer lightning season or longer [48,104,133]. For example, annuals were so dense in spring 2005 in some areas that large amounts of fine fuel persisted into the middle of 2007 [50,58,114,127]. These fine fuels can be sufficient to sustain a wildfire and thereby increase the likelihood of fire ignition and spread [48,104,133].

Unlike native annual plants, which are mostly forbs, nonnative annual plants are mostly grasses. Grasses have fine leaves with large surface area to volume ratios that can dry out quickly and provide fuels that carry fire under a broader range of conditions than do woody fuels and surface leaf litter [7,25,43]. Nonnative annual grass litter and standing dead material can accumulate rapidly and decompose very slowly in desert climates [25,28,48,104], persisting for 2 years or longer [28,48].

Nonnative annual forbs (e.g., Sahara mustard and arugula) are often large and can also produce abundant, persistent litter on sites where they invade [28,29]. In comparison, native annual fuels tend to be small and decompose rapidly after they die [28,48,62,104].

Nonnative perennial grass invasions can form large, continuous fine fuel loads [48,49,114] that persist with less interannual variation than sites dominated by annual plants [7,29,83]. For example, fine-fuel biomass in buffelgrass-dominated desert scrub sites can be orders of magnitude greater than similar sites dominated by red brome [47,49], and it may approach or exceed that reported for nearby native and nonnative desert grassland sites (table 2). In the absence of intense grazing, buffelgrass can fill in bare interspaces between native perennial plants [22,51,83] and provide continuous fine fuel—particularly, but not exclusively, after high rainfall years [34,42,79,104]. Fine fuel loads on buffelgrass-invaded sites fluctuate much less than fine fuel loads on sites dominated by native and/or nonnative annuals in the Sonoran Desert [47,77,83,94,119], which can increase the probability of fire ignition and spread in any season of any given year [7,29,47,83]. Ungrazed buffelgrass accumulates flammable material every growing season, and after several years of growth it has been described as an “almost-woody subshrub” [46]. See table S1 in Wilder (2021) for information on fuel models used in desert scrub ecosystems, including adjustments made for buffelgrass fuels [133].

Although contemporary fires are often fueled by nonnative plant invasions, native plants also helped fuel large wildfires in some years (e.g., [50,58,84,114]). The King Valley Fire, for example, was carried largely by desert Indianwheat and other annual fine fuels and by traveling along xeroriparian desert wash ecosystems with "substantial tree cover" composed of species such as desert ironwood, blue paloverde, yellow paloverde, desert willow, and/or smoketree [127,130]. However, it is unlikely that fuels were exclusively native plants in most contemporary fires, given that nonnative plants were already present in Arizona and often locally abundant.

Contemporary Fire History

Fires were rarely if ever documented in Sonoran desert scrub communities until several large wildfires occurred near Tucson and Phoenix during the late 1970s and early 1980s that included areas of Sonoran desert scrub (e.g., [44,80,89]). These fires prompted several studies of contemporary fire regimes and their effects on native plants and animals (table 3). Seemingly unprecedented "intense and dramatic" fires continued to occur throughout the northern Sonoran Desert in the late 20th and early 21st centuries [48,114]. From the early 1970s through 2021, years with high fire activity in the Sonoran Desert typically followed periods of above-average precipitation that resulted in periodic abundance of both native and nonnative annual plants [32,120]. The number of fires and total area burned in and around the Sonoran Desert in Arizona were especially high in 2004, 2005, 2019, and 2020 (based on data from Short (2022) [115]). Descriptions of a few of these fires are available in the literature (see Contemporary Fire Frequency), although data and analyses are lacking, and it is unclear how much of the total area burned in these fires was in Sonoran desert scrub communities.

Several sources indicate large areas of Arizona Upland vegetation or saguaro-dominated communities burned in contemporary fires. On and near the Tonto National Forest, for example, "a considerable portion" of mature saguaro-shrub habitat burned in the late 20th century [89], including an estimated 30% of saguaro habitat on the Mesa Ranger District between about 1970 and 1995 [7,89,107,135]. Esque (2004) describes several "major" fires in the Arizona Uplands including fires in the Santa Catalina Mountains in 1986 and the Mother's Day Fire in Saguaro National Park in 1994 [49]. One of the largest fires in 1979, the Granite Fire, south of Florence, Arizona, was said to have burned >10,000 ha of Arizona Upland, including paloverde-mixed cacti communities [49,84,127]. The 2021 Telegraph Fire started in desert scrub, burned 3,326 ha of desert scrub on the Tonto National Forest, spread into other plant communities, and grew to 73,150 ha [72].

Frequent, large, and severe fires in contemporary Sonoran desert scrub are attributed mostly to increased human access (i.e., increased road and trail density in and adjacent to wildlands) and subsequent human-caused ignitions (see Contemporary Fire Ignition and Season, below), in addition to increases in receptive fine fuel loads and continuity brought about by nonnative species invasions. These fuels can increase the likelihood of fire ignition and spread (see Contemporary Fuel Characteristics) that may result in larger patches of high-severity effects (see Contemporary Fire Size and Pattern). Rapid recovery of herbaceous fuels—especially nonnative grasses—may lead to repeated fires on a site (see Fire Management Considerations: Invasive Plant/Fire Cycle). Repeated and/or large patches of high-severity fire can lead to vegetation type conversions that may persist indefinitely (see Fire Management Considerations: Plant Community Conversion).

Contemporary Fire Ignition and Season

Contemporary fires in the Sonoran Desert are mostly ignited by humans and can occur in any month of the year. Fires are most frequent from late spring to mid-summer, when both human- and lightning-caused ignitions are most common. From 1955 to 2000, nearly 70% of fires on the Tonto National Forest were human-caused. Most fires occurred in May (26.9%) and June (25.2%), largely before the summer monsoon and lightning season, followed by July (21.6%), August (8.8%), and September (5.3%). About 6% occurred in fall and winter (October–March), when lightning is uncommon [6]. From 1955 to 1983, all fires on the Tonto National Forest during the cooler months of November through March were human-caused, and only during July and August were lightning-caused fires more frequent than human-caused fires. Fires were numerous from May through August, but the area burned was greater during June than all other months combined [112]. The number of fires on the Tonto National Forest from 1955 to 2000 increased with the population of Maricopa County and with traffic along major highways in the forest. However, there was no relationship between these factors and area burned, presumably because human-caused fires were mostly smaller than lightning-caused fires [6].

In the Sonoran and Mojave deserts, lightning-ignited fires are most common when lightning is most frequent—during the summer monsoon. From 1985 to 2001, 78% of lightning strikes in the desert bioregion of California occurred from July to September [28]. Lightning-ignited fires generally occur from mid-May to September in the Sonoran Desert and are most frequent and burn the greatest area from June through August [103]. From 1942 to 1991 all fires in Sonoran desert scrub communities on the Tonto National Monument were started by lightning between mid-May and mid-August—most frequently in July. Most area was burned by fires that started in late June or July, including four large fires (1947, 1964, 1970, and 1980) that came onto the monument from the national forest to the south during conditions of extreme heat and, in some cases, high winds [103]. The 2019 Mercer Fire was ignited in August when lightning struck a dense saguaro-paloverde stand on a south-facing slope heavily infested with buffelgrass. It burned 10 ha before running out of fuel (buffelgrass) and dying out [133].

Contemporary fires in Sonoran desert scrub communities are often ignited in adjacent communities (e.g., grassland, interior chaparral, or pine-oak woodland) and spread into desert scrub through ecotone areas with high densities of fine fuels. For example, the 2020 Bighorn Fire was ignited in June by lightning in semidesert grassland, on a steep and rugged, north-facing slope at about 1,200 m in the Santa Catalina Mountains. Fueled primarily by dead winter annuals (about 50% cover on unburned plots), as well as native bunchgrasses and nonnative lovegrass, the fire backed downslope from its ignition point into transitional, Sonoran mid-elevation desert scrub communities dominated by sotol, ocotillo, velvet mesquite, pricklypear, and hopbush. The fire spread into areas of paloverde-saguaro desert scrub on northern aspects, where fuel loads from dried winter annuals were higher than on other aspects, and in areas where fuels consisted of patches of nonnative perennial grasses. Other paloverde-saguaro communities were mostly unburned in the Bighorn Fire [133]. In some areas and years, dense patches of fine fuels along roadsides can extend into native plant communities and potentially carry fire into wildlands [7].

Nonnative invasive plants may increase the potential for fire in any season by providing receptive fuels nearly year-round [49,78,83]. For example, although nonnative perennial buffelgrass may be particularly dry and flammable during the monsoon season when lightning is most frequent [47], it can burn in any season [49], even when green [46]. See the buffelgrass Species Review for more information.

Contemporary Fire Frequency

Fires have become much more common across large parts of North American warm deserts, and they can pose significant land management challenges in localized areas [28]. Limited analyses from a small number of fire history studies (table 3) suggest that fire frequency in Sonoran desert scrub communities generally increased during the late 20th century. Most fires were small, human-ignited fires along roads (e.g., [6,49,112,114]). Arizona Bureau of Land Management records show that 210 fires burned 36,621 ha in the Arizona Upland and Lower Colorado River Valley subdivisions between 1973 and 1979, suggesting fire cycles shorter than the life span of long-lived dominant species [109]. In contrast, fire frequency was still sufficiently low (fire cycle of about 274–294 years) in the approximately 391,000-ha Sonoran Desert portion of the Tonto National Forest (dominated by Arizona Upland desert scrub) from 1955 to 1983, to support the idea that fires in the Sonoran Desert had been historically rare or infrequent, despite a trend of increasing fire frequency and annual area burned during that time [112]. Increased fire frequency was attributed to increased human-caused ignitions, changes in fine fuels—which would have been more abundant during the wetter than normal winters in the late 1970s—and/or improvements in fire detection and reporting [112]. Annual number of fires in that area remained high through 2000, especially during relatively wet periods [6,7]. These early studies (e.g., [112]) are often cited to support the contention that fire frequency and extent are “increasing” in the Sonoran Desert. An updated analysis including fires from 2000 to present and identifying area burned in specific plant communities within the Sonoran Desert is needed to validate and clarify trends [24].

Because fine fuel loads and continuity vary dramatically from year to year (e.g., [44,58,114]), contemporary wildfires in the Sonoran Desert are most common (and larger, see below) after abundant precipitation produces dense patches of native and/or nonnative herbaceous vegetation that cures during dry, hot conditions in late spring and forms relatively continuous fine fuel beds [7,28,48,58]. For example, on the Tonto National Forest, both number of fires and area burned from 1955 to 2000 were greater after two and three consecutive winters with above-average precipitation [6,7]. In dry shrub communities in southern Arizona, area burned by wildfires from 1980 to 2000 was greatest 1 to 2 years after periods of above-average precipitation [132]. While lacking analyses, additional observations from the early 1970s through 2021 support this pattern. For example:

• During late spring and summer of 1974, "numerous" fires occurred in the Arizona Upland after above-average precipitation in the winter of 1973–1974 resulted it a high density of annual plants—mostly the nonnative annuals redstem stork’s bill and red brome—on many sites [130].
• During spring and summer of 1979, several fires occurred in the desert regions of Arizona [49,84,127] and in creosotebush desert scrub along the western margins of the Sonoran and Mojave deserts of southern California [32], following two consecutive winters of above-average precipitation.
• In 2005, at least 25 fires >400 ha occurred in the Sonoran Desert [44,88,115,130] after 3 years of above-average precipitation, including exceptionally high rainfall in the winter of 2004–2005 (nearly 4 times the mean in some areas) [127]. In spring 2005, large areas of the Sonoran Desert were dominated by dense stands of the native forb, desert Indianwheat, and lesser amounts of nonnative Mediterranean grass [44,50,88,130].
• In 2019, 2020, and 2021, large fires in Arizona and California that included areas of desert scrub (fig. 9) were preceded by periods of above-average precipitation. Heavy rains in October 2018 and 2 to 3 inches of snow in January 2019 led to dense patches of red brome, which fueled the 50,100-ha Woodbury Fire in June 2019 [72,115]. The following wet winter and early spring (2019–2020) resulted in abundant winter annuals in the Sonoran Desert—about half of which were native species and half red brome. This was followed by a delayed and dry monsoon season, periodic strong winds, and the hottest summer in Tucson’s 125-year weather record [133], as well as the 48,600-ha Bighorn Fire [133] and the 78,300-ha Bush Fire [72,115] in Arizona, and the 17,500-ha Dome Fire in southeastern California [133]. In June 2021, the Telegraph Fire was ignited in desert scrub and eventually burned a total of 73,150 ha, including >3,200 ha of desert scrub on the Tonto National Forest [72].

Photo Credit

U.S. Geological Survey photos by Molly McCormick.

Figure 9—In 2019 and 2020, fires carried by the invasive red brome burned hundreds of thousands of acres of Sonoran desert scrub and resulted in some of the biggest fires in Arizona history. Burned areas are highly vulnerable to vegetation type conversion and further degradation. The Restoration Assessment and Monitoring Program for the Southwest (RAMPS) mapped the spread of red brome to help land managers make decisions on how to prevent large fires in the Sonoran Desert. 

Contemporary Fire Severity

Clear accounts of fire severity in Sonoran desert scrub communities are not common in the available literature. While some accounts generally indicate high levels of mortality and high-severity effects (see Fire Adaptations and Postfire Recovery), others suggest more of a mosaic of unburned patches and burned patches with varied fire severity within fire perimeters (i.e., mixed-severity fires) (e.g., the 2021 Telegraph Fire [72] and 2020 Bighorn Fire [133]). Nonetheless, high-severity fire effects are very likely more widespread in contemporary Sonoran desert scrub communities than they were historically, in part because nonnative invasive grass fuels are generally taller and/or more continuous than native herbaceous fuels. Therefore, fires in nonnative fuels are likely to burn at higher intensity and spread farther than those in native fuels, resulting in larger burned patches with more severe effects [83,133]. Greater frequency of extreme fire weather can also contribute to greater frequency and patch size of high-severity fire in desert scrub.

Few postfire data are available, and those may be biased toward areas with high-severity effects or repeated fires because these are the areas that fire scientists have focused on. Fires on and around the Tonto National Forest in the late 20th century caused considerable mortality to saguaro and other dominant plants (e.g., [7,89,107]) and were likely considered high-severity fires, although severity data were not reported or were not comparable. The 1979 Granite Fire was considered a high-severity fire because nearly all trees and shrubs were top-killed in burned areas, and overall mortality was high (50%–92%) for five dominant trees and shrubs 19 months after the fire. Mortality was especially high among cacti, and very few individuals sprouted; 91% of hedgehog cactus, 59% of barrel cactus, 98% of globe cactus, and 98% of cholla were dead after the fire. Although many plants survived in small, unburned patches within burned areas, and some individuals of each species sprouted after fire (25% of yellow paloverde, 32% of velvet mesquite, 1% of triangle bur ragweed, 9% of Nevada jointfir, and 37% of creosotebush), no seedlings of woody perennials were observed on burned transects either 7 or 19 months after fire [84], and mortality was delayed for some species. While initial mortality of saguaro was estimated at 11%, remeasurements [107] and repeat photographs [130] indicate delayed mortality for saguaro and codominant trees, eventually reaching 100% in some areas. Within 6 years after the fire, "an open plant community of widely scattered low shrubs is all that remains of the former arborescent community". Native dominants were still mostly absent 20 years later [130]. Observations of delayed mortality in paloverdes were similar after the 2005 King Valley Fire, and resprouting of woody dominants seemed less common than after the Granite Fire. Rainfall was well below average during the 2 years after the King Valley Fire, which may have inhibited resprouting [130].

While historical fires in Sonoran desert scrub communities are classified by LANDFIRE as stand-replacing (i.e., high-severity), observations of many contemporary fires suggest that mixed-severity fires may have been more likely. For example, data from the 2021 Telegraph Fire suggest a mixed-severity fire in the 3,326 ha of desert scrub within the fire perimeter on the Tonto National Forest, with most area (65%) burned at high and moderate severity. Fire severity maps based on postfire changes in vegetation show 1,140 ha burned at high severity, 1,016 ha at moderate severity, 839 ha at low severity, and 331 ha remained unburned within the fire perimeter [72]. The 1994 Mother’s Day Fire in Saguaro National Park may also have been classified as mixed severity, based on 24% mortality of saguaro and 73% of paloverde in 2000 [47,49]. Small, unburned patches of vegetation were found throughout severely burned areas after the 1979 Granite Fire [84]. The King Valley Fire was carried, in part, by woody plants in desert wash ecosystems, which were more severely burned than surrounding creosotebush desert scrub [130].

Fires fueled by nonnative annual grasses may burn at greater intensity and spread faster and farther than those fueled by native plants—and therefore have different effects—although fuel characteristics differ among nonnative species, and information on this topic is lacking for most species. Fires in Mediterranean grass tend to burn with low intensity and negligible soil heating, and they leave most woody shrubs unburned. In contrast, nonnative annual brome grasses produce greater fuel loads and depths than do Mediterranean grasses, and accordingly produce longer flame lengths and more intense fires [28] and are likely to have higher-severity effects. Laboratory experiments suggest that Mediterranean grass is a “spreader” of fire, carrying fire quickly (1.32 cm/second) and burning for short duration (0.5 minute) at low temperatures. In contrast, native plants such as creosotebush, white bursage, and Menzies' fiddleneck spread fire slowly (0.12 cm/second) and burn up to eight times longer (4 minutes) and at higher temperatures [53], suggesting higher severity effects from these native fuels.

Fires fueled by nonnative perennial grasses like buffelgrass may burn at higher intensity and have higher severity effects (i.e., greater postfire mortality of dominant plants and less postfire resprouting) than those fueled by native or nonnative annual plants. For example, in paloverde-mixed cacti communities burned in the 2020 Bighorn Fire, paloverde, brittlebush, and saguaro were "burned more severely" (i.e., scorched and/or killed) in areas where the fire was fueled by nonnative perennial grasses (mostly buffelgrass, although fountaingrass was also common) than in areas where the fire was fueled by winter annuals [133]. Based on fuel characteristics, buffelgrass-fueled fires likely have longer flame lengths, more rapid rates of spread, higher temperatures, and higher-severity effects (i.e., cause greater mortality of native flora and fauna) than fires fueled by native and nonnative annuals [47,83]. Mean maximum temperatures generated by buffelgrass-fueled fires in the Avra Valley, Arizona—where buffelgrass was planted in the late 1970s, and its biomass ranged from 6,710 to 12,100 kg/ha—ranged from 568 °C at the soil surface to 799 °C at 30 cm above ground, with a maximum recorded temperature of 871 °C (height not given). The authors noted that even thin but continuous patches of buffelgrass could carry fire and produce temperatures lethal to native plants [83]. In comparison, using similar but not identical methods, mean maximum temperatures during a June prescribed fire in an uninvaded paloverde-saguaro community on the Tonto National Forest—where fuel loads ranged from 700 kg/ha to 3,200 kg/ha—were 76 °C in interspaces, 167 °C under yellow paloverdes, and 210 °C within triangle bur ragweeds at 30 cm above ground. Maximum temperatures were 88 °C in interspaces, 299 °C under yellow paloverdes, and 405 °C within triangle bur ragweeds—all recorded at 1 cm above ground [99].

Fires that occur during periods with higher fuel moisture may have lower severity effects than those occurring during more extreme fire weather (i.e., hotter, drier, and/or windier), regardless of fuel type. For example, the 2019 Mercer Fire burned 10 ha of a dense saguaro-paloverde community and was fueled by buffelgrass; however, it occurred during a period of high humidity, which "raised dead fuel moisture and dampened fire behavior". The authors suggest that this may explain the strong spatial variation in saguaro mortality within the fire perimeter, and why most other perennial species "exhibited at least some regrowth four months after fire". Although many species (e.g., yellow paloverde, hopbush, saguaro, and most other cacti) had “high rates” of fire injury and postfire mortality during the first postfire year, and the authors anticipate additional, delayed mortality [133], data for saguaro alone suggest a mixed-severity rather than a stand-replacement fire: Of 95 saguaros sampled in the burned area, 89 (94%) had some fire injury; 39 (41%) had >50% scorching; and 28 (29%) had >80% scorching; 17 individuals (18%) died in the first year, most of which were young plants, and the authors estimated that 28 (36%) of the 78 living saguaros had a 50% or greater chance of dying in the future [133].

Contemporary Fire Size and Pattern

Most contemporary fires in Sonoran desert scrub communities are small (<20 ha), and fire size is generally smaller in the Sonoran and Mojave deserts—where 97% of total area burned from 1992¬ to 2016 was by fires <50 ha—compared to other North American desert regions—where 97% of total area burned from 1992 to 2016 was by fires <300 ha (Short 2017, cited in [59]). Nonetheless, several large fires occurred in and around the Sonoran Desert during the late 20th and early 21st centuries, and many exceeded 400 ha in total size (based on data from Short (2022) [115]). Large fires are more likely following periods of above-average precipitation that result in continuous patches of fine fuels. Many contemporary fires in the Sonoran Desert burned areas of desert scrub, including Sonoran desert scrub; however, data on area burned by plant community or BpS were not reported in the literature reviewed. Information on fire pattern is similarly lacking.

Fine surface fuel density and connectivity can limit or enhance fire spread and thus, fire size and patchiness in desert scrub communities. Fires in general, and especially large fires and large area burned in the Sonoran Desert, typically coincide with periods of above-average precipitation that result in increased continuity of fine surface fuels. In and around the Sonoran Desert in Arizona, maximum fire size, total area burned, and number of large fires (>400 ha) were high in 2004, 2005, 2019 and 2020 (based on data from Short (2022) [115]), coinciding with periods of above-average precipitation (see Contemporary Fire Frequency). From 1989 to 2010, 316 small (<20 ha) and 79 large (≥20 ha) fires occurred in the Sonoran Desert in Arizona [114]. Most large fires occurred (n = 36) and most area burned (51,700 ha) in 2005 [44]. These fires were fueled by dense patches of native and nonnative annual plants that had established in typically bare interspaces and connected patches of native perennial plants [44,88,114]. Median size of large fires in the Sonoran Desert in 2005 was 95 ha, compared to a median size of 60 ha over the 22-year study period [44,114]. Across a 45,100-km2 study area in the Sonoran Desert of Arizona (mostly within the Lower Colorado River subdivision), the probability of large fires (≥20 ha) from 1989 to 2010 was greatest in years with high maximum annual normalized difference vegetation index (NDVI, which represents patterns of vegetation greenness), and in areas at low elevations and with low road density [44,58]. NDVI is used as a proxy for annual fine fuel loads to estimate fire risk over large, contiguous extents (see Fire Management Considerations: Fire Prediction and Fuels Management). Twenty hectares was used as a cut off for large fires because it “represents a low-end estimate for large fire size in desert fuels" (W. Reaves, BLM, personal communication cited in [58]).

Published observations of contemporary fires suggest that because fire spread (and by extension, fire size) is limited by patchy fuels, the largest burned areas in Sonoran desert scrub communities tend to occur on relatively mesic sites and/or sites with dense populations of invasive grasses. For example, burned areas within the 2020 Bighorn Fire had dense patches of winter annuals and/or perennial grasses, and these patches varied with elevation and slope characteristics. Patches of buffelgrass, fountaingrass, and Lehmann’s lovegrass were most dense and common at the semidesert grassland/desert scrub ecotone, and patches were generally small and discontinuous along rocky hillsides. Fire spread was limited along those hillsides, and by areas of bare ground, hiking or game trails, and rock outcrops, which helped prevent fire spread into areas dominated by yellow paloverde, brittlebush, and saguaro that occurred downslope. Remnant winter annuals were much less dense (<5% total cover) [133], and nonnative perennial grasses were presumably absent from these unburned Sonoran desert scrub sites. Continued spread and coalescence of buffelgrass patches would increase fuel continuity and fire spread on these sites [133].

Large fires are most often ignited by lightning and driven by extreme fire weather. Lightning-caused fires tended to be larger on average, and they burned approximately twice the area burned by human-caused fires on the Tonto National Forest from 1955 to 1983 [112]. The larger size of lightning-caused fires might be attributed to the timing of dry thunderstorms at the end of the spring drought when vegetation is most flammable, and to the strong, gusty winds that accompany them. Lightning-caused fires are also more likely to occur in remote, inaccessible regions where they are not discovered or suppressed before growing large [112].

In contrast, human-caused fires typically occur in accessible places such as along roads and near campsites, where there is greater access for both human-caused ignitions and faster detection and suppression [112]. As of this writing (2023), these analyses are over 30 years old, and none has been published since, so it is unclear if this relationship has persisted. However, of 78 fires larger than 400 ha reported on the Tonto National Forest (in all vegetation types) from 1993 through 2020, 52 were lightning-caused, 25 were human-caused, and 1 had an undetermined cause (data from Short (2022) [115]), suggesting that this relationship persists.

While observations are limited, patchy fuels suggest that fires in desert scrub burn in a mosaic of varied severities, including unburned patches within the fire perimeter. One of the largest fires in 1979, the Granite Fire, south of Florence, Arizona, was said to have burned >10,000 ha of Arizona Upland, including paloverde-mixed cacti communities [49,84,127]. One observation suggests this was a patchy fire, with small, unburned patches of vegetation in the fire perimeter [84]. The 2021 Telegraph Fire started in desert scrub and grew to 73,150 ha. The fire burned through 3,326 ha of desert scrub (including but not limited to Sonoran desert scrub) on the Tonto National Forest, which resulted in a mosaic of high-severity (1,140 ha), moderate-severity (1,016 ha), low-severity (839 ha), and unburned patches (331 ha) [72].