Palmetto data
Study sites
Flatwood and scrubby flatwood habitats were selected in Archbold's West Section in 1981. Flatwood sites included Ridge Road, WSP1 and WS30 and were compared to scrubby flatwoods sites Ridge Road, WSP2 and Copse Road. The West Section of Archbold had not burned within the last decade, but historically had experienced a more natural fire return interval (5 - 20 years). All other sites are on Archbold's Red Hill in either sand pine scrub or sandhill, and had been fire suppressed, with the last known fire occurring in 1927. These latter two habitats fall in either Control management units or Experimental Burn management units. At the time this study was initiated, the fire management plan for Red Hill included restoring fire in some management units (Experimental Burn) while leaving others long unburned as reference sites (Controls). Experimental manipulations of clipping and fertilizing took place in sand pine scrub control management units.
The sandhill experimental site was roller chopped in 2013 as part of a sandhill restoration effort, dislodging aluminum stakes so they could not be matched up with their respective palmetto. Therefore, data were not collected in 2017.
A wildfire in May 2013 burned several of our study plants in both sandhill and sand pine scrub. Notes of whether or not the plant burned were recorded as comments in 2017. Because this was a wildfire, a plow was used to contain the escape and subsequently plowed through and killed 53 of our plants (31 S. repens and 22 S. etonia). These notes were also recorded in 2017. By 2017, out of 940 palmettos, 168 (17.9%) had died or were lost from the study but only 15 (1.6%) were not human induced. All others were mowed or plowed or died from some other human-induced management activity.
Plant selection
Plants of S. repens and S. etonia >30 cm in height (S. repens <30 cm in height were excluded as these do not flower) were selected in pairs to minimize microsite differences. We selected individuals of the least dominant species first and then found the closest pair of the more common species. Dominance varied by habitat and management unit. Each palmetto was marked with an aluminum angle stake with an attached aluminum number either pop-riveted or wired to the aluminum stake. Stakes were driven into the ground adjacent to the palmetto. For S. repens, the stake was driven next to the rhizome because rhizomes grow horizontally and thus move spatially. To avoid confusion of which palmetto the stake was marking, notes were taken on the S. repens' direction of rhizome growth. Stakes were always placed alongside the rhizome and never "in front" of the rhizome's growth path. During the 2017 census, we moved a number of S. repens stakes closer to their growing apex to avoid future confusion. Since S. etonia doesn't move spatially, stake position relative to apex location was not an issue.
Clipping and Fertilizing Experiment
This experiment was designed to examine what triggers post-fire flowering responses in palmettos. Treatments included: (1) Clipping to removing all leaves and their petioles to simulate leaf loss by fire (we took care to not physically damage the unfolded leaf), (2) Fertilizing to simulate enhanced nutrient availability following fire, and (3) applying both the clipping and fertilizing treatments to individuals.
Control plants for this study were the nearby sand pine scrub control plants that were established in 1985. We first numbered 300 aluminum stakes using 150 S. repens and S. etonia each (stakes 501-650 were used for S. repens and stakes 651-800 were used for S. etonia). Next, we randomly assigned each numbered stake within a species to the clipping, fertilizing, or both treatment (N=50 for each treatment). Palmettos were selected regardless of size so both small and large are represented.
We mixed 80 lbs. of commercial Palm Fertilizer (10-5-5) and 40 lbs. of Dolomite lime to make 120 lbs. of fertilizer/lime mix. This was added at a rate of 1 cup/fertilizer treatment palmetto or approximately 1/4 lb. (actual 156.6 g) of mix per treated palmetto. This mix was scattered evenly under the crown of the palmetto. Fertilized plants were well separated spatially from unfertilized plants.
Several stakes and/or number tags were replaced during the 2017 census because the tags and/or stakes had melted during a May 2013 escaped sand pine scrub fire. Damage to stakes and mortality to palmettos occurred primarily where downed sand pines burned creating higher temperatures for longer periods.
Seedling plant selection
For seedlings, we used a 2-m wide belt transect running through the WSP1 flatwoods marked palmetto area and a second 2-m wide belt transect near the WSP2 scrubby flatwoods. We marked every putative palmetto seedling we could locate using a wire flag inserted through the hole of a metal number tag that laid on the surface of the ground. At WSP1, we marked 100 putative seedlings using numbers 1-100 and at WSP2 we marked putative seedlings with numbers 101-188. At both sites, we recorded which quadrats on the belt transect each putative seedling occurred in.
Harvested palmetto plant selection
We destructively harvested 33 palmettos (17 S. repens and 16 S. etonia) from three sites in 1985. Plants (5 of each species) were harvested on 18 January 1985 on the southwest corner at the intersection of Culvert and Ridge Road. A second harvest was made (5 S. repens and 6 S. etonia) on 21 January 1985 on the southwest corner at the intersection of Ridge Road and North Seven Road. A final harvest was made on 22 January 1985 (7 S. repens and 5 S. etonia) on Red Hill in the sand pine scrub control area. Thirty-two palmettos were used to develop biomass regression models (17 S. repens and 15 S. etonia).
Data collection
Data were collected annually, starting in 1981 (or when plants were initially marked) in January/February for all plants and sites. Data collection included maximum standing height (cm), maximum canopy length (cm), maximum canopy width (cm, perpendicular to the length), number of green leaves and number of inflorescence scapes. In 1982, the number of new leaves produced between annual censuses were counted (the newest leaf of a given year was banded with a plastic-coated wire so new leaves could be determined). For green leaves and new leaves, a value of 0.5 was given if a new leaf had emerged and was free from the apex but the leaf blade had not yet expanded. In 1990, we also began to calculate leaf longevity, which is the working leaf life span, as described by Corner (1966), Chazdon (1986), and Cunningham (1997). This estimate is calculated by dividing the number of living, fully expanded leaves in the crown by the number of new leaves produced during that year. This leaf longevity measure does not include the "nonworking" leaf development time before leaf expansion, which according to Corner (1966) would roughly double the life spans estimated.
Canopy cover for each palmetto was measured in January 1993 with a densitometer pointed to the four cardinal directions, then averaged. Biomass for each palmetto species was calculated for each year of measurements using regression models developed from 32 destructively harvested palmettos. Leaf longevity was also calculated by dividing the number of living, fully expanded leaves in the crown by the number of new leaves produced during that year.
For seedlings, starting in 1989, only height and maximum crown diameter (cm) were recorded. In 1992, we started measuring the base of the plants using calipers (cm). Data were collected continuously from 1989 through 1997, then again in 2001 and for a final census in 2008.
For harvested palmetto biomass, we took measures (in cm) at three levels: whole plant, largest leaf and petiole of largest leaf. Height, length and width were taken across the entire canopy and on the single, largest leaf. Length and width of the petiole for the largest leaf were also recorded. We also counted the number of green leaves per plant. Data were collected on 33 individuals (17 S. repens and 16 S. etonia), but only 32 were used in developing biomass regressions (17 S. repens and 15 S. etonia).
Biomass from the harvested palmettos is the estimated dry mass (g) of each harvested palmetto. The regression models used for S. repens and S. etonia were generated from 32 destructively harvested palmettos of a wide range of sizes excavated from two scrubby flatwoods stands and one sand pine scrub stand. Plants were measured for basal diameter, crown height, crown maximum width, crown maximum width perpendicular to the maximum width, number of living leaves, leaf-blade width and length, and petiole length and width of the largest leaf, and number of inflorescences. Palmettos were divided into leaf and stem (both above- and below-ground), but roots were not harvested since they grow to depths of several meters, making recovery of all root tissues virtually impossible for fresh-mass determination. Subsamples of fresh mass were oven dried at 80C to constant mass for estimation of dry mass. The best regression models to estimate dry mass (g) are:
Serenoa repens dry mass = Exp (0.637*LN (MINXX) + 2.3*LN (GLVSXX) + 0.254
Sabal etonia dry mass = (10.71*(MINXX)) + (332.5*(GLVSXX)) - 826.3
Where: MINXX = crown "minimum" in year XX and GLVSXX = number of green living leaves in year XX.
The S. repens model for N = 17 had a r^2 = 0.84, F =36.7, df = 2,14, P < 0.001
The S. etonia model for N = 15 had a r^2 =0.86, F = 38.2, df = 2,12, P <0.001
Some biomass estimates are negative due to error in the regression models (Abrahamson 1995). This mostly occurs for smaller plants. These values have been retained and the data user may choose to include the values, remove the values or modify the values with additional data and a better fitting regression model.
