Warm Fire

Background

 Pretty ForestOn June 8th, 2006, a lightning strike ignited the Warm Fire on the northeastern edge of the Kaibab Plateau. The Kaibab National Forest had recognized the need to manage fire and fuels at a landscape-scale and developed a Wildland Fire Use (WFU) program with the Warm Fire being the first large implementation of that plan. The fire was initially managed as a WFU fire where it burned with a mix of low, moderate and high severity fire across approximately 7900 ha (19,500 acres) of predominately ponderosa pine forests (USDA 2007). The initial plan was altered on June 25th, when weather conditions changed considerably, and the Warm Fire exceeded its maximum manageable area and was declared a wildland fire. The wildland fire portion of the fire was predominately high severity, but did burn in low and moderate severity and burned in all three vegetative types: mixed conifer forests, ponderosa pine forests and pinyon juniper woodlands. In total, the fire burned 24,000 ha (59,000 acres) with several very large high severity patches. The Warm Fire served as the first large application of the WFU program and had unintended consequences which resulted in controversial second-guessing of fire policies within northern Arizona and especially across the North Kaibab ranger district. Fire suppression and post-fire rehabilitation was done in the moderate and high severity burn areas of the wildfire section. Post-fire mitigation consisted of seeding with ryegrass for the purpose of reducing flood risk, soil erosion, and invasion of non-native plant species.

This portion of the study falls into two sections of the study and addressed two of the main objectives:

A. Fire severity effects on the overstory composition and structure, understory vegetation and fuels

1. Characterize understory vegetation response to fire and determine correlation between fire severity, physiographic variables, and changes in understory vegetation characteristics.

B. Post-fire rehabilitation effects on understory vegetation

2. Characterize understory vegetation response to post-fire seeding.

Our study was located in the Warm Fire on the Kaibab Plateau in the Kaibab National Forest in northern Arizona, USA (Fig. 1). Plot locations range in elevation from 2300 to 2590 m and at the landscape-scale have a similar disturbance history in terms of grazing and logging. The fire burned across three vegetation communities: higher elevation mixed conifer; mid-elevation ponderosa pine dominated interspersed with quaking aspen and lower elevation pinyon-juniper woodlands. This study was conducted in ponderosa pine dominated overstory with an understory composed of common grasses, such as muttongrass, squirreltail, and Junegrass and common forbs including small leaf pussytoes, Fendler's sandwort, and woolly cinquefoil.


Map JFSP All Plots  

Figure 1. Location of study site. Perimeter of the Warm Fire; the northern portion of the fire (no crosshatch) was managed as a WFU while the southern portion (crosshatched) was managed as a wildfire and was subjected to post-fire mitigation. The fire encompassed 24,000 ha across three vegetation types. Burn severity is indicated by varying shades of grey. Unburned controls are indicated by green circles, low severity plots by blue triangles, high severity non-seeded by yellow squares, and high severity seeded plots by red circles.


We sampled a total of 102 plots in 2008 and 2009 with 25 unburned controls; 27 low severity; 25 high severity non-seeded; and 25 high severity seeded. We sampled understory plant cover, biomass, species richness, and species composition during August and early September. We measured forest floor cover by estimating cover of bare soil, rock, wood, litter, duff, lichen, moss, and scat. Overstory canopy cover was measured with a densitometer in 2009 by counting the number of overstory hits at 33 points along the transect lines. Topographic variables recorded at each site were slope (degrees), aspect, and elevation (m).

Key findings

A. Fire severity effects on the overstory composition and structure, understory vegetation and fuels

1. Characterize understory vegetation response to fire and determine correlation between fire severity, physiographic variables, and changes in understory vegetation characteristics.

Fire in ponderosa pine forests stimulates the growth and diversity of the understory plant community. We observed this trend in our study as vegetation cover, biomass and species richness increased with increasing fire severity. High severity plots averaged the highest vegetation cover, biomass and greatest species richness with low severity plots a close second in species richness. There was however, a lack of response in vegetative cover and biomass in the low severity plots. We attribute this to the unnatural accumulation of litter and duff from lack of fire in the last century. We saw that as litter cover increased, total vegetative cover and biomass decreased and conversely, when soil cover increased, so did vegetative cover and biomass. On low severity sites litter cover was still 53% two years post-fire while high severity sites had only 32% cover. Decreasing unnaturally high levels of litter through fire creates an opportunity for understory growth. Comparisons of pre-fire data collected on 16 of the plots indicate similar plant composition and structure to our unburned controls. When the community is separated into functional groups, we observed that the pre-fire cover of graminoids, forbs and shrubs is most similar to the unburned controls and low severity plots in 2009. Exotic species also appeared to represent about 10% of the total species in the pre-fire community.

RainyDayWorkThe plant community composition was significantly different between all three treatments; unburned, low severity and high severity sites. The differences in the community were primarily driven by factors of fire severity: litter cover, canopy cover and exposed soil/rock. Indicator species analysis showed significantly more annual and biennial forbs strongly associated with high severity and perennial forbs and trees as indicators of unburned controls and low severity. There were 23 species associated with high severity and of those 18 are annuals/biennials, 5 exotics, and 9 ruderals. Reference sites in ponderosa pine forests on the North Rim of Grand Canyon National Park that have not been subjected to fire suppression indicate that annual and biennial forbs are of great importance to the post-fire plant community for species richness and vegetative cover. Overall, we observed greater frequency and abundance of exotics as fire severity increased (unburned, low severity and high severity) and the total cover of exotics was less than 2% for all three treatments. Prickly lettuce, yellow salsify, and dandelion were consistent in high severity plots but had low average cover (<1%). All three species are listed as noxious weeds in at least one state and are well adapted to disturbed areas as they are prolific seed producers with wind-dispersed seeds that can colonize from off-site. Cheatgrass and mullein were also more prevalent in high severity sites and cheatgrass in particular should be monitored to prevent spread across the landscape. Overall, we observed greater changes, both positive and negative within the high severity burn areas.

B. Post-fire rehabilitation effects on understory vegetation

2. Characterize understory vegetation response to post-fire seeding.

Total vegetation cover is used to determine the success of seeding and a recent review of published seeding studies concludes that while seeding is a common choice, it is not effective at producing the required 60% ground cover to reduce the amount of bare soil and prevent erosion and exotic invasions. The results of this study support this review in that the seeded species averaged only 6% cover one-year post-fire and seed was most prevalent on gradual slopes. Seeded sites actually had less total cover and biomass than non-seeded sites in the Warm Fire although it was not a significant difference in either case. Bare soil was also significantly higher in seeded sites hence seeding was not effective at decreasing the exposed bare soil. Seeded areas also did not differ significantly from non-seeded sites in the cover of exotics species and the impacts of seeding appear to be relatively marginal on that front as well. The presence of cheatgrass in both seeded and non-seeded plots increased from 43% of all plots to 50% in 2009 with no significant difference in cover between treatments. Cheatgrass cover is low (< 2%) but the observed increase warrants concern and continued monitoring is strongly recommended. Ryegrass was still present across the landscape 3 years post-seeding and while the percent cover is decreasing, the presence of an exotic species occupies species space and usurps resources that would otherwise be available for native plant species.

The plant community composition was significantly different between seeded and non-seeded sites as confirmed by both statistical analysis and ordinations and the differences associated with seeding are disconcerting. Although ryegrass cover was relatively low, there is evidence that it competes for space and may displace natives thus altering the initial community composition and possibly the community trajectory. The presence of ryegrass was associated with significantly less cover of three dominant native bunchgrasses (squirreltail, mountain muhly and muttongrass) and less cover annual and biennial forbs cover in seeded plots. Annual and biennial forbs play an important role in post-fire restoration as they add significantly to the species richness and vegetative cover in burned areas in ponderosa pine forests on the Kaibab Plateau. We also observed significantly fewer ponderosa pine seedlings in seeded sites and acknowledge that this may be due to environmental variation as well as propagule pressure. We analyzed of the rate of plant community change in a treatment between years and results indicated that these communities are changing at similar rates, but potentially in different directions as per our indicator species analysis. Analysis of biomass data reflected similar trends with less biomass of native bunchgrasses and fewer annual and biennial forbs as indicator species in the seeded plots. The long-term persistence of this intentionally introduced ryegrass is currently unknown and managers and stakeholders will need to continue monitoring.


Last Updated ( Tuesday, 15 March 2011 )