Experiment 012 was established by Johannes Knops in 1983 to study the impact of four different fire frequencies on vegetation structure infertile grassland. Aboveground biomass (sorted by species or litter) was sampled four times over the course of the ongoing experiment, with the most recent data dating from 2000. In 2000, percent cover was also measured to estimate non-vascular plant change (lichens, mosses, bare soil, etc.) Light penetration was sampled periodically.
The 24 experimental burn plots are located in an abandoned agricultural field (Field B.) Starting in 1984, four treatments were implemented in replicate: annual burning, burning every other year, burning every four years or un-burned control. Burns take place in the early spring, March or April, depending on the snow melt.
Experiment 012 adds an important dimension to the study of fire's role in ecology by studying its effect on a secondary prairie with nutrient poor, sandy soils.
Fire's impact on prairie productivity and species composition strongly depends on the ecosystem's productivity. Long-term fire studies conducted in more productive mesic prairies and grasslands such as Konza prairie, have shown that increased fire frequency can increase productivity in frequently burned watersheds (Briggs and Knapp 1995; Towne and Knapp 1996; Knapp et al. 1998) through the removal of accumulated litter, which strongly decreases light availability at the soil level (Knapp and Seastedt 1986; Collins et al. 1998). Experiment 133, which was carried out in a more productive forest-grassland continuum at Cedar Creek, found that 32 years of different fire frequencies showed in a shift from 90% dominance by trees to 80% dominance by grasses with increasing fire frequency, accompanied by large changes in nitrogen cycling and vegetation composition (Reich et al 2001).
Experiment 012 illustrates that without litter accumulation (which could lead to light limitation) and tree invasion, the impacts of 17 years of different fire treatments on vegetation composition are minor (Fig 1; Knops 2006). The most consistent change was a decrease of C3 grass abundance with increasing fire frequency (Fig 2; Knops 2006), which was driven by the loss of Poa pratensis (Fig 1). Poa, an introduced species, was the most abundant C3 grass, contributing 80% of the biomass in the control burned plots. Other C3 grasses showed no significant effect.