Frequent burning is a common land practice in many grasslands worldwide, and this land use strategy has large impacts on a wide variety of ecosystem functions and services. Fire in tallgrass prairie, in the absence of grazing, alters plant community composition, decreases richness, and increases plant production. Proposed mechanisms for the changes in community composition and function are that fire decreases N availability (through volatilization) and removes litter (thereby increasing light availability and decreasing soil moisture). However, few experiments explicitly test these mechanisms, and those that do monitor short-term effects.Yet, the strength of these mechanisms likely differ over longer time scales, as other ecosystem attributes (e.g., plant community composition) change through time. Ghost Fire aims to determine the mechanisms behind community and ecosystem differences between annually burned grassland and 20-year burned grassland (hereafter called unburned) by experimentally manipulating N availability and litter. We impose litter and N conditions found in unburned grassland onto annually burned grassland, and litter and N conditions typically found in annually burned grassland onto unburned grassland. Importantly, Ghost Fire monitors both below-and above-ground plant community and ecosystem dynamics as well other dimensions of the ecosystem including microbial and mycorrhizal communities and insect community composition and biomass.
Experimental Methods: Ghost Fire plots are located in the uplands of 2 annually burned watersheds (1D and SpB) and 2 unburned watersheds (20B and 20C). On each watershed there are 3 blocks (4x12m) of 6 plots (2x4) each. The corners of each block are marked with a tall (2m conduit) marker. Each block has a split plot design (Figure 1) with the litter present (L+) treatment covering half the block and the litter absent (L-) treatment covering the other half.
In the unburned watershed, L+ is the control, and L- is achieved removing all the litter from the plot in spring (just before average burning time) via a combination of weed whacking and raking. In the annual burn watershed, L+ is achieved by adding approximately 400 g/m2 of hay from Konza which has been solarized to remove seeds immediately following burning. L- will be the control. Then nested within each litter treatment, each of the three plots is assigned one of three N treatments - Added Carbon, Control, Added Nitrogen. The added C treatment decreases N availabity. The Added Carbon treatment is achieved by adding 250 g/m2 of sugar to the plots on the first of the month for four months (May, June, July, August). The Added Nitrogen treatment is achieved by adding 10 g/m2 of nitrogen to the plots on May 1st. Theoretically due to differences in starting soil N levels due to the long-term burning history, we add enough nitrogen to the annual burned Nitrogen plots to cause N levels to be equivalent with those found in the unburned Control plots, and nitrogen content in the Carbon treatment in the unburned will be similar to the Control in the annual burn (Figure 2). These treatments began in the spring of 2015. The corners of each plot are marked with a mid-height (1m conduit) marker.
Each plot consists of three 1x1m subplots. Subplot A is the permanent species composition plot. No destructive sampling will be allowed in this subplot. Subplot A is surveyed for plant species composition twice yearly, once in the spring and once in the fall. Subplot A is marked each spring with flags. Subplot B is used for yearly ANPP sampling, as well as any other aboveground tissue sampling (e.g., for C:N leaf tissue). Subplot C is used for destructive belowground sampling of plant roots, microbial community, and soil nutrient levels. Other researchers are allowed to use subplots B and C as long as the level of disturbance caused by the sampling is deemed appropriate.
Data Collected: A range of community and ecosystem measurements are taken in Ghost Fire. Table 1 shows the data available from the first three years of sampling. Note that 2014 is pretreatment data. Not all of the samples collected have been processed yet. For example, the mycorrhizal recolonization from 2014 and the soil organic matter samples from all years have not been processed yet. Everything else from 2014 and 2015 is complete, and much of the 2016 data is still be processed and entered at this time (December 2016).
Ideally in each year we will sample soil enzyme activity, plant species composition, plant stem density, ANPP, aboveground biomass (disc pasture meter), belowground standing crop (soil cores), light, and C:N of soil. Every 5 years we plan to conduct more intensive sampling including belowground productivity, mycorrhizal biomass, microbial biomass and community composition, insect biomass and community composition, and resin bags. Treatments and sampling will be maintained until significant plant community shifts have been observed.
Data Collection Methods: Soil (Resins): Soil available N was measured using two resin bags per plot installed early May-September. Bags were made of fine (less tan 1 mm), undyed nylon mesh and contained 5 g each of anion exchange resin (Cl- form; Dowex 1X8-100, 50-100 mesh) and cation exchange resin (H+ form; 50Wx8-40, Dowex HCR-W2, 8% cross linking, 16-40 mesh). The day prior to installation, bags were soaked for one hour in 0.6 N HCl then rinsed with de-ionized water three times and stored at 4oC. Bags were buried 10 cm deep at the edge of Subplot C. After extraction of bags in September, measurement of available N (nitrate and ammonium) bound to resins was accomplished by first extracting N by shaking resin bags in 100 ml of 2 M KCl at 200 rpm for 2 hours and processing the solution through polycarbonate filters. Concentration of N in extracts was then measured using an Alpkem Flow Solution 4 Automated Wet Chemistry System (O.I. Analytical, College Station, TX, USA). Values from each resin bag were averaged per plot.
Soil Organic Matter (SOM): Soil organic matter is measured using the loss on ignition (LOI) method. A known weight of field-moist soil is dried overnight at 105oC, then weighed again to measure the dry weight of soil. Then the soil subsample is place in a muffle furnace at 400oC for 2h to combust all organic material and weighed again. The soil organic matter lost on ignition is reported as a proportion (w/w) of the total dry soil mass.
Soil Enzyme Activity: Hydrolytic enzyme potential activities were measured using fluorometric substrates (MUB and methylcoumarinmethylumbelliferone / MUB and methylcoumarin) and oxidative enzyme potential activities were measured using a colorimetric substrate (L-3,4-dihydroxyphenylalanine L-DOPA) in 96-well (6-8 technical replicates) plate assays. Hydrolytic enzyme assays included phosphatase (Phos; EC 184.108.40.206, 4-MUB-phosphate), leucyl aminopeptidase (LAP; EC 220.127.116.11, L-leucine-7-amido-4-methylcoumarin), cellobiohydrolase (CBH; EC 18.104.22.168, 4-MUB-B-D-cellobioside), B-glucosidase (bG; EC 22.214.171.124, 4-MUB-B-D-glucoside), and β-N-acetylglucosaminidase (NAG; EC 126.96.36.199, 4-MUB-N-acetyl-β-D-glucosaminide), and were run at a final substrate concentration of 40 µM. Oxidative enzyme assays included peroxidase (Perox; EC 188.8.131.52, L-3,4-dihydroxyphenylalanine and H2O2) and phenol oxidase (Phenox; EC 184.108.40.206, L-3,4-dihydroxyphenylalanine) and were run at a final substrate concentration of 5 mM. -glucosidase 'G':EC220.127.116.11,4-MUB - Dglucoside,leucylaminopeptidase 'LAP':EC18.104.22.168,L-leucine-7-amido-4-methylcoumarin,phosphatase'Phos':EC22.214.171.124,4-MUB-phosphate,cellobiohydrolase 'CBH':EC126.96.36.199,4-MUB--D-cellobioside,-N-acetylglucosaminidase 'NAG':EC188.8.131.52,4-MUB-N-acetyl--D-glucosaminideandoxidativeenzymeassayswereperoxidase 'Perox':EC184.108.40.206,L-3,4-DOPAandH2O2andphenoloxidase 'Phenox':EC220.127.116.11,L-3,4-DOPA. All assays were run at 24oC in 50 mM sodium acetate buffer (pH 5) for 2 (bG and Phos), 4 (NAG and CBH), or 18h (LAP, Perox and Phenox), with appropriate blank and quench controls, and final activities were standardized to nmol substrate degraded (hydrolyzed or oxidized) g-1 dry soil h-1.
Mycorrhizal Root Colonization: Measures of root colonization by mycorrhizal fungi are obtained from 3 pooled soil cores taken in each plot in August. The soil cores are pooled, roots cleaned and stained with trypan blue, and the percentage mycorrhizal colonized roots are determined microscopically using the gridline intersect method.
Belowground Standing Crop: Three 2 cm diameter x 15 cm depth soil cores were taken in each plot in August or September to estimate belowground root standing crop. Samples were taken randomly within the designated soil sampling subplots. Samples were kept at 4 C until processed, elutriated to remove roots from soil, and dried at 60C for 48 hours. SOM particles and dead roots (identified by color/texture) were sorted out of the sample, then the remaining live root portion was weighed. Finally, samples were combusted at 450 C for 4 hours to obtain ash masses, which were subtracted from live root mass to obtain ash free dry mass (AFDM).
BNPP: In May, two root ingrowth cores were installed in each plot. Cores were 5 cm diameter x 15cm deep, and constructed of 2 mm mesh screening. Holes were drilled into the soil using a 7" diameter bucket auger, cores were installed and both the cores and surrounding hole was filled with sieved soil from the site. Soil in and around cores was packed to approximate field conditions. Ingrowth cores were removed in September and kept at 4 C until processing. Samples were elutriated to remove roots from soil, and root contents dried at 60C for 48 hours. SOM particles were sorted out of the sample, then the remaining root samples were weighed. Finally, samples were combusted at 450 C for 4 hours to obtain ash masses, which were subtracted from root mass to obtain ash free dry mass (AFDM).
CO2 Flux: Soil CO2 efflux measurements were taken weekly from early June to mid-August in 2016. Measurements were taken from 2 PVC collars (10 cm diameter x 8 cm deep, buried 6 cm in the ground) per plot using a Li-Cor 6400 portable gas exchange system with a soil CO2 chamber attachment. Collars were placed at a standardized location in all plots except when soil depth was less than 6 cm. Measurements were taken between 11 am and 2 pm and all aboveground living biomass was clipped prior to taking measurements. We also measured the headspace within the collars monthly and adjusted measurements to reflect this volume. CO2 efflux measurements were recorded once values had stabilized (usually about 1 minute after attaching the instrument to the collar). Additionally, soil moisture and temperature were taken at each time of soil CO2 sampling.
Plant Species Composition: Within each plot, permanent 1x1m species composition plots are designated as Subplot A. Percent aerial cover is estimated to the nearest 1% for each species that was rooted in the plot in early June and late August.
Plant Stem Density: Within each species composition plot (Subplot A), a permanent 0.1m2 (20x50cm) stem density quadrat was established. Each year in early June, number of stems of each species are counted.
ANPP (Above-ground Net Primary Production): Each year between late August and early September, above-ground biomass is clipped to ground level within 2 20x50cm quadrats located in each plot's designated destructive sampling subplot (Subplot B) and sorted into graminoids (grasses and sedges), non-graminoid forbs, woody plants, and previous year's dead. Detailed clipping locations are recorded with Subplot B to make sure that quadrat locations are clipped as infrequently as possible. After clipping, biomass was dried at 60 oC for ca. 48 hours and then weighed. Previous year's dead biomass (unburned years) was not included in estimates of annual productivity.
Disc Pasture Meter: Disc Pasture Meters allow for coarse estimates of aboveground biomass in a non-destructive way. Four measurements are taken in Subplot B (before ANPP is clipped), with each measurement occurring in a different corner of the plot. These are then averaged across the four subplot and using an established allometric equation, converted to g/m2 of aboveground biomass.
Light: Light, sampled using a ceptometer, provides a percentage of photosynthetically active radiation that reaches the soil surface. Light is measured each year in the Spring (early June) in Subplot A. The June measure helps to determine the efficacy of the litter treatment. Additionally, this time period shows the biggest discrepancies between treatments in PAR hitting the soil surface (late in the growing season plant growth in all treatments drastically limits PAR at the soil surface). As personnel are available and time permits, other time points throughout the growing season are measured. However, these midseason and late season measurements are sporadic. Light is measured by taking one reading above the canopy and 4 readings at the soil surface. The four measurements are then averaged and the percentage of above canopy light hitting the surface calculated.
Invertebrate Collection (Biomass and Species Composition): Invertebrates were vacuum sampled from the permanently marked productivity subplots (Subplot B) at peak invertebrate densities (August) using a modified leaf blower set. Each plot was sampled for 60 seconds. Visual counts of grasshoppers and katydids that hopped out of the plot during vacuum sampling were recorded. Sampled invertebrate communities were frozen at -20C until processed. Invertebrates from each sampled were identified to family and counted. Once identified, each invertebrate sample was dried at 60C for at least 48 hours and weighed to the nearest 0.001g. Grasshoppers and katydids were dried and weighed separately from the rest of the sample to estimate biomass of Orthoptera that hopped out of the plots, which were counted but not able to be weighed.