Floodplain Forest Seedling Survival through Dormant Season Flooding: A Controlled Experiment

These data are associated with a controlled experiment that subjected approximately seven-month-old seedlings of <i>Nyssa aquatica</i>, <i>Nyssa ogeche</i>, <i>Fraxinus caroliniana</i>, <i>Carya aquatica</i>, and <i>Quercus lyrata </i>to treatments ranging between one and four months of complete submergence in water. The treatments and timing of the experiment simulate winter flooding commonly experienced in the Apalachicola River floodplain of Florida, USA where these species dominate neighboring forest types (swamp and low bottomland) whose distributions compete due to climatic and anthropogenic influences on hydrologic regimes. Following prior experimentation in 2022 that aimed to investigate differences in first-year seedling performance of swamp and low bottomland hardwood species from the Apalachicola floodplain under varying moisture regimes, we utilized the remaining approximately seven-month-old seedlings from a low-stress treatment. All seedlings were grown from seeds collected directly from the floodplain except for <i>Nyssa aquatica</i>, whose seeds were purchased from the Louisiana Forest Seed Company. One-gallon GroPro® black plastic grow bags were used to contain the seedlings. The soil column for each bag consisted of 7.5 cm of commercially available peaty topsoil in the top half and another 7.5 cm of commercially available loamy mix topsoil in the bottom half of the column. See the "Prior Seedling Experiment Data" file to view seedling heights above the root collar before experimentation. Treatments occurred in a single circular Intex 28131EH plastic/PVC above-ground pool measuring 3.65 m in diameter and 76.2 cm in height with a capacity of 5,621 liters. The pool was filled with local untreated well water and monitored four times periodically for pH, dissolved oxygen concentration (mg/L), and turbidity (NTU) using an Aqua TROLL 500. See the "Water Quality Monitoring" file for observations over the experiment period. The experiment occurred in an open environment where ambient temperatures and precipitation could influence the treatments. See the "PNS_airport_weather" file for local weather data. The experiment involved three replicates of four test treatments and two replicates of a control treatment. The control treatments were only used to check for unexpected mortality during the experiment, and analysis was only performed on the four test treatments resulting in a randomized complete block design. Each species had varying sample sizes (n) in each treatment and block due to a lack of seedling availability. Only general observations were made for <i>Nyssa ogeche</i> in all treatments and <i>Nyssa aquatica</i> in treatments two and three. Any seedlings whose terminal buds could not be covered during test treatments were also omitted from the analysis. See the "Dormant Season Flooding_Seedling Data" file for all seedling observations throughout the experiment. The total survival of each species in each treatment and block was calculated at the end of the experiment. A pairwise Fisher exact test was then performed to identify significant differences among the average survival of each species and each treatment at the end of the experiment. The cumulative number of individuals flushing leaves was tallied each week after treatments and until the end of the experiment to identify potential signs of stress or delayed flushing due to the treatments. To compare the effect of treatments on the timing and rate of leaf-flushing individuals among species, multiple dose-response curve models were constructed to find the best fit to the data. Models were ranked using Akaike's information criterion and once the best-fit model was selected, a likelihood ratio test was used to assess the effect of treatment on cumulative flushing individuals. If a significant p-value was returned, then pairwise comparisons of species using the Tukey adjustment for multiple comparisons were constructed. The hill slope parameter was used for comparisons of leaf flushing rates and the half-maximal effective concentration, or EC50 parameter, was used for comparisons of timing to reach 50% of total flushing individuals. All analyses were performed in R. See the "Dormant Season Flooding_Seedling Experiment" R file for analyses using the data files.These files are associated with a manuscript prepared for submission to a peer-reviewed journal.