Increasingly, land managers have attempted to use extreme prescribed fire as a method to address woody plant encroachment in savanna ecosystems. The effect that these fires have on herbaceous vegetation is poorly understood. We experimentally examined immediate (<24hr) bud response of two dominant graminoids, a C₃ caespitose grass, Nassella leucotricha, and a C₄ stoloniferous grass, Hilaria belangeri, following fires of varying energy (J/m²) in a semi-arid savanna in the Edwards Plateau ecoregion of Texas. Treatments included high- and low-energy fires determined by contrasting fuel loading and a no burn (control) treatment. Belowground axillary buds were counted and their activities classified to determine immediate effects of fire energy on bud activity, dormancy, and mortality. High-energy burns resulted in immediate mortality of N. leucotricha and H. belangeri buds (P < 0.05). Active buds decreased following high-energy and low-energy burns for both species (P < 0.05). In contrast, bud activity, dormancy, and mortality remained constant in the control. In the high-energy treatment, 100% (n=24) of N. leucotricha individuals resprouted while only 25% (n=24) of H. belangeri individuals resprouted (P < 0.0001) three weeks following treatment application. Bud depths differed between species and may account for this divergence, with average bud depths for N. leucotricha 1.3 cm deeper than H. belangeri (P < 0.0001). Synthesis and applications: Our results suggest that fire energy directly affects bud activity and mortality through soil heating for these two species. It is imperative to understand how fire energy impacts the bud banks of grasses to better predict grass response to increased use of extreme prescribed fire in land management.

Location:
Sonora Texas A&M Agrilife Research Station (SARS)
Western edge of the Edwards Plateau ecoregion in Texas (-100.574°, 30.251°)
Semiarid savanna ecosystem encroached by Prosopis glandulosa and Juniperus spp.

Experimental Design:
Fire treatments arranged in a randomized design with three treatments (no burn, low fire energy, and high fire energy) replicated 12 times for a total of 36 experimental plots. Each plot was 10 x 10 m and centered on a mature mesquite shrub (Prosopis glandulosa).

Two grass species were selected for this study: Nassella leucotricha and Hilaria belangeri.
    Nassella leucotricha: C3 bunchgrass
    Hilaria belangeri: C4 stoloniferous grass

Within each 100 m² plot, two 1 m² subplots were demarcated with steel posts. One of these subplots was created around a patch of N. leucotricha, and the other around a patch of H. belangeri, and both served as a reference group for tiller collections described in the next section. 

Reemergence:
Within each of these subplots, two individuals of the focal species were marked and monitored for regrowth of tillers three weeks following fire application.

Bud Activity:
Tillers were harvested 24hrs before and after fires from three randomly determined individuals of each grass species inside each plot. These tillers were collected from individuals in similar phenological stages as the permanently marked individuals using the classification system of Moore et al. (1991). All collected tillers were from current year growth and all vegetative. Plants visibly damaged by herbivores, insects, or pathogens were excluded. The buds associated with these tillers were counted and their activity classified as either active, dormant, or dead using the Tetrazolium and Evans Blue staining procedures established by Busso (1989).

Bud Depth:
The day before the fire treatments were applied, bud depth was measured. Two random individuals of each species from each large plot were selected based on similar size and phenological stage as our permanently marked individuals. A hole dug at the base of each individual grass exposed the deepest buds. Bud depth was recorded as the distance between the mineral soil surface and the base of each tiller (approximately at the beginning of the root system).