Tornado Effects on Temperate Deciduous Forest Mortality:
A Macro Report of the Temple Forest Observatory

Faisal Shaikh*, Mariana Bonfim, Mary Cortese, Katherine Stevenson, Amy L. Freestone

Abstract

After the Temple Ambler Field Station was hit by an EF-2 tornado, opportunities to study forest resilience opened. As the Temple Forest Observatory (TFO) was 63% censused before the storm hit in September 2021, we shifted our focus to finding what remained in the forest. As the field station interns were able to fully census TFO as of July 2022, we then compared forest census data pre- and post- tornado. Having data from before and after the storm allows us to start examining rates of loss and predict potential forest regeneration trajectories. With the help of exploratory analysis in R, we were able to look at numerous patterns of the deceased species and learn why the forest looks the way it looks today.

Introduction

“Resilience is a measure of the forest’s adaptability to a range of stresses and reflects the functional integrity of the ecosystem.” (1)

Forest ‘resilience is a useful concept to understand ecosystem change.’ (2)

Objective: Give a macro report of what is still in the forest and why they belong where they belong. Specifically, this report provides maps of all woody stems before and after the storm, discusses the most common deceased species, and the densities of each of the deceased growth habits. These patterns can help us better understand the forest resilience at TFO.

Methods

• Forest census was collected by Temple Ambler Field Station interns and staff during Summer 2021 (63% of total area) and 2022 (100%, 4 ha).
  All living woody stems were tagged, measured and identified.

• Deceased stems were obtained by concluding that whichever stem tags were present in the pre-storm census and absent in the post-storm census were likely lost due to storm damage.

• Dead stems in the post-storm census (2022) that were not missing and marked as dead were added to the dead stems dataset.

#Storm census with removed dead tree. Now they are "missing" so now you can anti-join with prestorm cenaua to find ALL missing trees
Only_Alive_TFO_2022_updated <- subset(TFO_2022census_cleaninginprogress_excel_modified, TFO_2022census_cleaninginprogress_excel_modified$Status.x != 'DS')

#Drop all tags in pre-storm census that match tags in post storm census. This shows all tags aka all woody stems that all are missing, therefore, dead
#dead_trees is a dataframe of ALL trees that died
dead_trees_updated <- anti_join(TFO_prestormcensus,Only_Alive_TFO_2022_updated, by = "Stem_tag")

code creating dead stems dataset in R

• Exploratory data analyses were performed in R v 4.2.1. Focal variables included:
  • Number of each species
  • Growth Habits
  • Relative X and Y coordinates for all tags

Results

Figure 1 Forest stand in TFO post storm for 100% of the forest plot

Figure 2 Forest stand in TFO pre- and post- storm for 63% of the forest plot

——————————————–Prestorm————————————————————————–Poststorm————————————

.

TFO forest Pre-storm (left) and Post-storm (right)

Figure 3 Number the 10 Most Common Species at TFO Post-storm.

Lindera benzoin is the most common species

Figure 4 Number of the 5 Most Common Deceased Species at TFO Post-storm.

Lindera benzoin is the most common deceased species

Figure 5 Density of Each Growth Habit of Deceased Plants at TFO Post-storm.

Shrubs are the highest deceased plants

#create a dataset 'num_Trees_updated' that have total counts of all three growth habits
num_Trees_updated <- allDS_with_trees_gh3 %>%
  group_by("growth_habit") %>% 
  count("growth_habit")

#Calculates density of each growth habit 
num_Trees_updated$dead_trees_species_updated <- as.numeric(num_Trees_updated$freq) / nrow(allDS_with_trees_gh3)

creating densities of each growth habit in R

Discussion

• There is a spatial zonation with lower number of stems in certain areas of the forest. Spatial distribution of loss, however, was substantial throughout. (Figure 1, Figure 2)

• Lindera benzoin (spicebush) stems are more than 4x more common than the 2nd most common species, Ligustrum obstusifolium (border privet) (Figure 3)

• The deaths of individuals of Lindera benzoin (spicebush) was more than 6x higher than the 2nd highest death species, Fagus grandifolia (american beech) (Figure 4)

• Shrubs are responsible for over 70% of the deceased plants (Figure 5)

Future Direction

• As this macro report informs us of the forest resilience at TFO, there is still much more exploratory analysis needed to fully understand the forest resilience. Looking at scatterplots we can hypothesize the correlations of trees height/canopy density of specific species vs the growth rate of shrubs or study the indirect effects of the disturbance on scrub mortality

• The goal is to upload our census to a national forest database so that we can share TFO’s story nationally and internationally

• As Temple Ambler Field Station interns recently completed censusing Robin’s Park, a nearby older growth forest patch that remained relatively unharmed after the tornado, performing data analysis on this forest as a reference site can help us better understand recovery trajectories and resilience at TFO

Citation

1. Kerlin, Katherine E. “Just What Is a ‘Resilient’ Forest, Anyway?” UC Davis, 19 Jan. 2022, https://www.ucdavis.edu/climate/news/just-what-resilient- forest-anyway.
2. Reyer, Christopher. “Forest Resilience, Tipping Points and Global Change Processes”, 7 Jan. 2015, https://besjournals.onlinelibrary.wiley.com/doi/10.1111/136 5-2745.13165.

Acknowledgements

I would like to thank Temple Ambler Field Station and the interns I worked with for the past six months. I would also like to thank Amy L. Freestone for introducing me to TFO, Mary Cortese for helping me in the field, and Dr. Bonfim for working with me individually and guiding me in the data science field. Thanks to the CST Undergraduate Research Program for supporting this research experience

This webpage was entirely created in RStudios with R markdown