This is the final group project repository for EDA, Spring '22.
In this project we explore the effects of defaunation, depletion of animals due to hunting, on tropical forest structure and composition. To answer these questions we use a dataset provided by the Poulsen Tropical Ecology lab. This dataset contains tree measurements within forest plots along a defaunation gradient in Ivindo National Park, Gabon. The major takeaways from this data are that there were no apparent relationships between basal area and species richness and the defaunation gradient. These results were unexpected and are most likely a result of faulty data or unit conversion errors.
Aubrey Knier^1^, Israel Golden^1^, Mishka Malinowski^2^, and Tasha Griffiths^1^ 1 - Duke Nicholas School Masters of Environmental Management Candidates 2 - Duke Nicholas School of the Environment - University Program in Ecology PhD Student
Gabon, defaunation, tropical ecology, species composition, gradient, forest structure, basal area, DBH
Data in the repository was collected from the Poulsen Ecology Lab. The data represent forest measurement data from twenty, 50 x 50m forest plots along a defaunation gradient in Ivindo National Park, Gabon. Data collection began in late 2020 and is still ongoing.
Folder structure -> Data (Processed, Raw, Spatial) -> R_Markdown_Scripts -> GoldenGriffithsKnierMalinowski_ENV872_EDA_FinalProject -> Scratch (misc. things)
File formats
Data folder -> Processed (csv) -> Raw (csv, xlsx) -> Spatial (two folders Africa and Gabon) -> Africa (shapefile and additional necessary files) -> Gabon (shapefiles and additional necessary files and csv files with coordinates)
R_Markdown_Scripts -> files in .Rmd format
Naming Conventions For R markdown scripts -> Last name_analysis type
Final Project -> Last names of all members_class_FinalProject
All data can be found in the Data folder. Raw data is found in the Raw folder within the Data folder and contains a dataset with 21 columns and 45,681 rows. Each row represents a sample or a tree that was measured within the forest plots. Below is a table that described the columns within the dataset:
| Column Name | Description | Unit | Class |
|---|---|---|---|
| E | Field expedition season | Season-Year | Nominal |
| Data_entry | Name of person inputting data to Excel | Name | Nominal |
| Date..dd.mm.yyyy. | Date of Excel data entry | Date/Month/Year | Date |
| File_name | Photo file name of field data sheet | .JPG | Nominal |
| Date..dd.mm.yyyy..1 | Date of field data collection | Date/Month/Year | Date |
| Note_taker | Name of individual recording field data | Name | Nominal |
| Project | Forest type (defaunated/intact forest) | Category | Nominal |
| Plot | Unique plot identification | Category | Nominal |
| Grid | Within-plot grid | Category | Nominal |
| TAG_SUM | The most unique sample identifier | Plot-Grid combo | Nominal |
| Plant_tag | Identifer assigned to each sample | Letter-Number combo | Nominal |
| X_coord | X coordinate of sample location | Meters | Continuous |
| Y_coord | Y coordinate of sample location | Meters | Continuous |
| Tool | Tool used to measure diameter | Category | Nominal |
| POM | Point of measurement for diameter | Meters | Continuous |
| DBH.mm | Diameter at breast height (DBH) | Millimeters | Continuous |
| Height..meters. | Height of plant | Meters | Continuous |
| Type_Field | Vegetation type or size class of plant | Category | Nominal |
| Note_Field | Miscellaneous field notes | Phrase | Nominal |
| ID | Latin species identification | Name | Nominal |
| Treatment | Plot treatments (fungicide/insecticide) | Category | Nominal |
Processed data is stored in the processed folder within the data folder and was created from wrangling the raw dataset and creating variables. It contains 6,279 rows and 8 columns. A description of the columns is below:
| Column Name | Description | Unit | Class |
|---|---|---|---|
| Project_Plot | Combo of project and plot ID | Category | Nominal |
| DBH.mm | Diameter at breast height (DBH) | Millimeters | Continuous |
| Height..meters. | Height of plant | Meters | Continuous |
| Plant_tag | Identifer assigned to each sample | Letter-Number combo | Nominal |
| Veg_Field | Vegetation type or size class of plant | Category | Nominal |
| ID | Latin species identification | Name | Nominal |
| Status | Indicates defaunated/intact | Name | Nominal |
| Distance_km | Calculated distance for Makokou | Kilometers | Continuous |
Spatial data used to create maps of the study site are found in the spatial folder within the data folder. Spatial files are then broken up into more folders - Gabon Country Mask (shapefile of Gabon), National Parks (shapefile of Gabon National Parks), and Plots (shapefile of the forest plots). Additionally there is a csv file labeled Makokou_Coordinates.csv which gives the latitude and longitude of Makokou the closest large town to Ivindo National Park.
| Column Name | Description | Class | Units |
|---|---|---|---|
| City | Name of city | nominal | N/A |
| Longitude | coordinates | continuous | degrees |
| Latitude | coordinates | continuous | degrees |
Final project script and code -> GoldenGriffithsKnierMalinowski_ENV872_EDA_FinalProject.Rmd
All other scripts are in R_Markdown_Scripts folder and are as follows:
Golden_LM.Rmd -> Exploratory analysis for research questions 2 & 3 and linear model synthesis Griffiths_Wrangling_&Visualization_DBH.Rmd -> Exploratory analysis for research question 1 Knier&Malinowski_Data_Cleaning&_Exploration.Rmd -> Raw data wrangling and cleaning Malinowski_Spatial.Rmd -> Creation of maps for the study site
Within raw data wrangling and cleaning removed unreasonable samples and those with NAs. Extensive data exploration and application of ecological knowledge of the area. Also met with group members and reviewed each others code.