Cusco/Wayqecha, Peru - March 2020
Group 3: Trait & taxonomic community response to fire and elevation
Contact: dagmar.egelkraut@uib.no
For repo access contact Tanya either on Slack or shoot her an email: tanya.strydom@icloud.com
Repository for a proposed manuscript focusing on the effects of fire and elevation on species composition and functional trait in the high Andean moist Puna grasslands. This repository is used to store code associated with the data analysis portion of the manuscript. The proposal can be found here
This can be done by running the scripts/0_data_import.R file. This will then
download the data as well as filter/clean the data so that we only have the data
that we will be using for our analyses.
The raw datasets can be retrieved from osf using {dataDownloader} -
you will need to install this using devtools::install_github("Between-the-Fjords/dataDownloader")
if you haven't done so previously. These data will be stored in the
newly created data/raw folder.
You can 'run'/call this script by calling
source(here::here(path = "scripts/0_data_import.R")) at the
beginning of your script. which will also add the two cleaned/filtered
datasets to your environment as traits (for functional traits) and
species (for community composition) as well as ensuring that you have
the most up to date dataset stored on you machine.
🚀 RELEASE NOTE calling
source(here::here(path = "scripts/0_data_import.R")) now also incorporates
the CSR traits as well as temporary chemical traits. This will be updated in
such a way that when we get the new chemical traits data it should all
seamlessly integrate and none of the scripts should break - she says
hopefully...
Note if you do call
source(here::here(path = "scripts/0_data_import.R"))at the beginning of your script it does mean you will always be using the most complete/updated dataset if we do end up changing which data we want to keep or exclude from analyses.
The scripts/0_data_import.R code will also filter out the the sites
relevant for our analyses. As we are downloading the entire Puna
dataset form osf we need to filter the data so that we have the following
sites and treatments from the following years:
| Site | Treatment | Dry Season | Wet Season |
|---|---|---|---|
| QUE | C | 2019 | 2019 |
| QUE | NB | 2020 | |
| TRE | C | 2019 | 2018, 2019, 2020 |
| TRE | NB | 2019 | 2019, 2020 |
| ACJ | C | 2019 | 2018, 2019, 2020 |
| ACJ | NB | 2019 | 2020 |
⚠️ Note thatdata/is in the.gitignorethis means you have to run the data import script at least once regardless and that anything you export to the data folder will not be pushed to the repo and others will not be able to use that output unless you tell them which script to run to source that output.
Each subtask
related to/needing a
coding workflow should be contained within its own script and should be
named starting with the subtask number and a brief descriptive name.
Scripts should be placed in the scripts/ folder. This means we can
keep track of each task separately.
Ideally each subtask should be on a new branch. This means that each subtask
can be turned into a pull request (PR) allowing us to easily see the full
commit history for that subtask and also allows subgroup members to request
reviews/feedback from each other as well as have conversation threads. PRs can
initially be marked as drafts and once ready (i.e. completed) it can be
marked as ready for review and then merged into the master branch.
branches should be named after the subtask code - same for the PR (although this can be a bit more comprehensive/descriptive).
Code for scraping and appending chemical traits data to the traits dataset can
be found in scripts/DC2_chem_trait_database.R. running this script will result
in having the traits df in your R environment also containing said chemical
traits.
We have scraped traits data from the BIEN database and TRY. Not all traits data are available in BIEN and/or TRY so we do not have any data for N15 isotopes as well as N:P. Of the other chemical traits data we do not have full coverage (even at the family level) for all of the traits. In addition we've used the same trait mean/value across the dataset i.e. there is no variation between species/genera/families - mostly because this is placeholder data it seems wasteful to expend energy on trying to generate variation across sites
All data accessed from TRY are in data/TRY/ if anyone is interested in looking
at them and the dataframe with the placeholder chemical data is in data/processed/
and is currently called into your working environment when calling scripts/0_data_import.R as traits. This will be changed once we have the actual
chemical traits (and as has been said many times should happen seamlessly and
not break anything that has already been scaffolded).
A quick overview of trait distribution featuring scraped chem traits:
CSR scores were calculated for each individual based on trait values
using StrateFY. This process
is not automated so the output is saved in data/processed/ and has been appended
to the original leaf traits dataset. it has also been integrated into
scripts/0_data_import.R so calling that script will automatically 'add' the CSR
traits to the traits df to your environment
🚧 still under construction 🚧
The idea is to centralise and standardise plotting themes and colour
palettes using scripts/plotting_aesthetics.R. This script will call a
custom {ggplot2} theme that you can call at the end of your plotting
code chunk that will standardise aesthetics across the board. You will
still be able to include your own theme tweaks after this by just
calling theme() again after. There will also be a colour theme which
will standadise colours used for different treatments and sites which
you can then call when you specify scale_*_manual - example pending...
Calling this script using source(here::here(path = "scripts/plotting_aesthetics.R"))
will import the needed themes etc.
A preliminary figure of trait distributions because why not! 😉
