---
title: "BIEN tutorial"
author: "Brian Maitner"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
%\VignetteIndexEntry{BIEN tutorial}
%\VignetteEngine{knitr::rmarkdown}
%\VignetteEncoding{UTF-8}
---
```{r setup, include = FALSE}
knitr::opts_chunk$set(echo = TRUE)
```
# Setup
```{r load-packages, message = F,warning = FALSE, results = 'hide'}
library(BIEN)
library(ape) #Package for working with phylogenies in R
library(maps) #Useful for making quick maps of occurrences
library(sf) # A package for spatial data
```
# An overview of the package
We try to make this package as easy and intuitive to use as possible, but it is still often easiest to start with our vignette. Particularly useful are the "Function Names" and "Function Directory" sections.
```{r load-vignette, eval = FALSE}
vignette("BIEN")
```
The function names follow a consistent naming strategy, and mostly consist of 3 parts:
1. The prefix "BIEN_"
2. The type of data being accessed, e.g. "occurrence_"
3. How you'll be querying the data. For example, the suffix "state" refers to functions that return data for a specified state.
As a complete example, the function `BIEN_occurrence_species` returns occurrence records for a given species (or set of species).
# Function Families
Currently we have 9 function families in BIEN. These are sets of functions that access a given type of data.
1. occurrence records (`BIEN_occurrence_...`)
2. range maps (`BIEN_ranges_...`)
3. plot data (`BIEN_plot_...`)
4. trait data (`BIEN_trait_...`)
5. taxonomic information (`BIEN_taxonomy_...`)
6. phylogenies (`BIEN_phylogeny_...`)
7. stem data (`BIEN_stem_...`)
8. species lists (`BIEN_list_...`)
9. metadata (`BIEN_metadata_...`)
We'll walk through each of the function families and take a look at some the options available within each.
# Occurrence records
These functions begin with the prefix `BIEN_occurrence_...` and allow you to query occurrences by either taxonomy or geography. Functions include:
1. `BIEN_occurrence_country` Returns all occurrence records within a given country
2. `BIEN_occurrence_state` Returns all occurrences records within a given state/province
3. `BIEN_occurrence_county` Returns all occurrences records within a given state/province
4. `BIEN_occurrence_family` Returns all occurrence records for a specified family
5. `BIEN_occurrence_genus` Returns all occurrence records for a specified genus
6. `BIEN_occurrence_species` Returns all occurrence records for a specified species
Each of these functions has a number of different arguments that modify your query, either refining your search criteria or returning more data for each record. These arguments include:
1. `cultivated` If `TRUE`, records known to be cultivated will be returned.
2. `new.world` If `TRUE`, records returned are limited to those in North and South America, where greater data cleaning and validation has been done. IF `FALSE`, records will be limited to the Old World. If `NULL` (the default), global records will be returned.
* Note that the arguments cultivated and new.world may change the number of records returned.
3. `all.taxonomy` If `TRUE`, the query will return additional taxonomic data, including the uncorrected taxonomic information for those records.
4. `native.status` If `TRUE`, additional information will be returned regarding whether a species is native in a given region.
5. `natives.only` If `TRUE`, the default, information for occurrences flagged as introduced will not be returned.
6. `observation.type` If `TRUE`, the query will return whether each record is from either a plot or a specimen. This may be useful if a user believes one type of information may be more accurate.
7. `political.boundaries` If `TRUE`, the query will return information on which country, state, etc. that an occurrence is found within.
8. `collection.info` If `TRUE`, the quest will return additional information about the collection and identification of that specimen.
**Example 1: Occurrence records for a species**
Okay, enough reading. Let's get some data.
Let's say we're interested in the species *Xanthium strumarium* and we'd like some occurrence data. We'll use the function `BIEN_occurrence_species` to grab the occurrence data.
```{r xs-occurrences}
Xanthium_strumarium <- BIEN_occurrence_species(species = "Xanthium strumarium")
```
Take a moment and view the dataframe and take a look at the structure
```{r view-xs-occurrences}
str(Xanthium_strumarium)
head(Xanthium_strumarium)
```
The default data that is returned consists of the latitude, longitude and date collected, along with a set of attribution data. The meaning of some of these columns is obvious (e.g. latitude, longitude), however others may be less clear. The meanings of these columns and the information within is explained in more detail in our data dictionary, which can be accessed using the function `BIEN_metadata_data_dictionaries`.
If we want more information on these occurrences, we just need to change the arguments:
```{r xs-occurrences-pt-2, eval = FALSE}
Xanthium_strumarium_full <- BIEN_occurrence_species(species = "Xanthium strumarium",
cultivated = TRUE,
all.taxonomy = TRUE,
native.status = TRUE,
observation.type = TRUE,
political.boundaries = TRUE)
```
We now have considerably more information.
If we want to take a quick look at where those occurrences are we could use:
```{r map-xs, eval = FALSE}
# Make a quick map to plot our points on
map('world', fill = TRUE, col= "grey", bg = "light blue")
# Plot the points from the full query in red
points(cbind(Xanthium_strumarium_full$longitude,
Xanthium_strumarium_full$latitude),
col = "red",
pch = 20,
cex = 1)
# Plot the points from the default query in blue
points(cbind(Xanthium_strumarium$longitude,
Xanthium_strumarium$latitude),
col = "blue",
pch = 20,
cex = 1)
```
**Example 2: Occurrence records for a country**
Since we may be interested in a particular geographic area, rather than a particular set of species, there are also options to easily extract data by political region as well.
We'll choose a relatively small region, the Bahamas, for our demonstration.
```{r occs-bahamas, eval = FALSE}
Bahamas <- BIEN_occurrence_country(country = "Bahamas")
#Let's see how many species we have
length(unique(Bahamas$scrubbed_species_binomial))
#About 400 species with valid occurrence records.
#Now, let's take a look at where those occurrences are:
map(regions = "Bahamas" ,
fill = TRUE ,
col= "grey",
bg = "light blue")
points(cbind(Bahamas$longitude,Bahamas$latitude),
col = "blue",
pch = 20,
cex = 1)
#Looks like some islands are considerably better sampled than others.
```
# Range maps
These functions begin with the prefix `BIEN_ranges_...` and return (unsurprisingly) species ranges. Most of these functions work by saving the downloaded ranges to a specified directory in shapefile format, rather than by loading them into the R environment.
Functions include:
1. `BIEN_ranges_species` Downloads range maps for given species and save them to a specified directory.
2. `BIEN_ranges_genus` Saves range maps for all species within a genus to a specified directory.
3. `BIEN_ranges_load_species` This function returns the ranges for a set of species as a sf object.
4. `BIEN_range_sf` This function returns all ranges that intersect a user-specified sf object.
The range functions have different arguments than we have seen so far, including:
1. `directory` This is where the function will be saving the shapefiles you download
2. `matched` If `TRUE`, the function will return a dataframe listing which species ranges were downloaded and which weren't.
3. `match_names_only` If `TRUE`, the function will check whether a map is available for each species without actually downloading it
4. `include.gid` If `TRUE`, the function will append a unique gid number to each range map's filename. This argument is designed to allow forward compatibility when BIEN contains multiple range maps for each species.
**Example 3: Range maps and occurrence points**
If we have a species we're interested in, and would like to load the range map into the environment, we can use the function `BIEN_ranges_load_species`. Let's try this for *Xanthium strumarium*.
```{r xs-range}
Xanthium_strumarium_range <- BIEN_ranges_load_species(species = "Xanthium strumarium")
```
The range map is now in our global environment as an sf object. Let's plot the map and see what it looks like.
```{r xs-range-map}
#First, let's add a base map so that our range has some context:
map('world', fill = TRUE ,
col= "grey",
bg = "light blue",
xlim = c(-180, -20),
ylim = c(-60, 80))
#Now, we can add the range map:
plot(Xanthium_strumarium_range[1],
col = "green",
add = TRUE)
```
Now, let's add those occurrence points from earlier to this map:
```{r xs-range-and-points}
map('world',
fill = TRUE ,
col = "grey",
bg = "light blue",
xlim = c(-180, -20),
ylim = c(-60, 80))
plot(Xanthium_strumarium_range[1],
col = "green",
add = TRUE)
points(cbind(Xanthium_strumarium$longitude,Xanthium_strumarium$latitude),
col = "blue",
pch = 20,
cex = 1)
```
# Plot data
These functions begin with the prefix "BIEN_plot_" and return ecological plot data. Functions include:
1. `BIEN_plot_list_sampling_protocols` Returns the different plot sampling protocols found in the BIEN database.
2. `BIEN_plot_list_datasource` Returns the different datasources that are available in the BIEN database.
* These first two functions are useful for identifying plots with comparable sampling methods.
3. `BIEN_plot_sampling_protocol` Downloads data for a specified sampling protocol
4. `BIEN_plot_datasource` Downloads data for a specific datasource
* These next two function are then useful for downloading datasets with consistent methodology.
5. `BIEN_plot_country`
6. `BIEN_plot_state`
7. `BIEN_plot_dataset` Downloads data for a given dataset (which is nested within a datasource)
8. `BIEN_plot_name` Downloads data for a specific plot name (these are nested within a given dataset)
Again we have some of the same arguments available for these queries that we saw for the occurrence functions. We also have the new argument *`all.metadata`*, which causes the functions to return more metadata for each plot.
**Example 4: Plot data by plot name**
Let's take a look at the data for an individual plot.
```{r luq-occs, eval = FALSE}
LUQUILLO <- BIEN_plot_name(plot.name = "LUQUILLO")
head(LUQUILLO)
```
We can see that this is a 0.1 hectare transect where stems >= 2.5 cm diameter at breast height were included. If we'd like more detail, we can use additional arguments:
```{r, eval = FALSE}
LUQUILLO_full <- BIEN_plot_name(plot.name = "LUQUILLO",
cultivated = TRUE,
all.taxonomy = TRUE,
native.status = TRUE,
political.boundaries = TRUE,
all.metadata = TRUE)
```
The dataframe `LUQUILLO_full` contains more useful information, including metadata on which taxa were included, which growth forms were included and information on whether species are known to be native or introduced.
# Trait data
These functions begin with the prefix `BIEN_trait_...` and access the BIEN trait database. Note that the spelling of the trait names must be precise, so we recommend using the function `BIEN_trait_list` first. Traits names are standardized to follow https://www.top-thesaurus.org/ where available. Trait units have been standardized for each trait.
Functions include:
1. `BIEN_trait_list` Start with this. It returns a dataframe of the traits available.
2. `BIEN_trait_family` Returns a dataframe of all trait data for a given family (or families).
3. `BIEN_trait_genus`
4. `BIEN_trait_species`
5. `BIEN_trait_trait` Downloads all records of a specified trait (or traits).
6. `BIEN_trait_mean` Estimates species mean trait values using genus or family level means where species-level data is absent.
7. `BIEN_trait_traitbyfamily` Downloads data for a given family (or families) and trait(s).
8. `BIEN_trait_traitbygenus`
9. `BIEN_trait_traitbyspecies`
**Example 5: Accessing trait data**
If you're interested in accessing all traits for a taxon, say the genus *Salix*, just go ahead and use the corresponding function:
```{r salix-traits, eval = FALSE}
Salix_traits <- BIEN_trait_genus(genus = "Salix")
```
If instead we're interested in a particular trait, the first step is to check if that trait is present and verify the spelling using the function `BIEN_trait_list`.
```{r trait-list, eval = FALSE}
BIEN_trait_list()
```
If we're interested in leaf area, we see that this is indeed called "leaf area" in the database. Now that we know the proper spelling, we can use the function `BIEN_trait_trait` to download all observations of that trait.
```{r leaf-area, eval = FALSE}
leaf_area <- BIEN_trait_trait(trait = "leaf area")
```
Note that the units have been standardized and that there is a full set of attribution data for each trait.
# Taxonomy Data
While there are existing packages that query taxonomic data (e.g. those included in the excellent taxize package), the BIEN taxonomy functions access the taxonomic information that underlies the BIEN database, ensuring consistency.
1. `BIEN_taxonomy_family` Downloads all taxonomic information for a given family.
2. `BIEN_taxonomy_genus`
3. `BIEN_taxonomy_species`
**Example 6: Taxonomic data**
Let's say we're interested in the genus *Asclepias*, and we'd like to get an idea of how many species there are in this genus and what higher taxa it falls within.
```{r taxonomy}
Asclepias_taxonomy <- BIEN_taxonomy_genus(genus = "Asclepias")
#We see that the genus Asclepias falls within the family Apocynaceae and the order Gentianales.
#You'll also notice that a given species may appear more than once (due to multiple circumscriptions, some of which may be illegitimate).
#If we'd just like to know all the speciess that aren't illegitimate:
Asclepias_species <- unique(Asclepias_taxonomy$scrubbed_species_binomial[Asclepias_taxonomy$scrubbed_taxonomic_status %in% c("accepted", "no opinion")])
```
# Phylogenies
The BIEN database currently contains 101 phylogenies for new world plants. This includes 100 replicated phylogenies that include a large fraction of New World plant species ("complete phylogenies") and 1 phylogeny containing only those New World plant species for which molecular data were available ("conservative phylogeny"). Currently, there are only 2 functions available:
1. `BIEN_phylogeny_complete` This function will return a specified number of the replicated "complete" phylogenies. Note that each phylogeny is several Mb in size, so downloading many may take a while on slow connections.
2. `BIEN_phylogeny_conservative` This function returns the conservative phylogeny.
Arguments: The function `BIEN_phylogeny_complete` has a few arguments that are worth explaining:
1. `n_phylogenies` This is the number of replicated phylogenies that you want to download (between 1 and 100)
2. `seed` This function sets the seed for the random number generator before randomly drawing the phylogenies to be downloaded. This is useful for replicating analyses.
3. `replicates` This function allows you to specify WHICH of the 100 phylogenies to download, rather than having them selected randomly.
**Example 7: Phylogenies**
Let's say we want to download the conservative phylogeny.
```{r phylogeny, eval = FALSE}
phylo <- BIEN_phylogeny_conservative()
#Let's make sure it looks alright
plot.phylo(x = phylo, show.tip.label = FALSE)
#If we just want to see which species are included
phylo_species <- phylo$tip.label
```
# Stem Data
The BIEN database contains stem data associated with many of the plots. This is typically either diameter at breast height or diameter at ground height. At present, there is only one stem function (although expect more in the future):
1. `BIEN_stem_species` This function downloads all of the stem data for a given species (or set of species)
2. `BIEN_stem_genus`
3. `BIEN_stem_family`
4. `BIEN_stem_datasource` This function downloads all of the stem data for a given datasource.
Arguments:
The arguments for this function are the same that we have seen in the occurrence and plot functions.
**Example 8: Stem data**
If we'd like stem data for the species *Cupressus arizonica*
```{r stems, eval = FALSE}
Cupressus_arizonica_stems <- BIEN_stem_species("Cupressus arizonica")
```
# Species lists
These functions begin with the prefix `BIEN_list_` and allow you to quickly get a list of all the species in a geographic unit. Functions include:
1. `BIEN_list_country` Returns all species found within a country.
2. `BIEN_list_state` Returns all species found within a given state/province or other 2nd level political division.
3. `BIEN_list_county` Returns all species found within a given county/parish/or other 3rd level political division.
Some of the same arguments we saw in the occurrence functions appear here as well, including `cultivated` and `new.world`.
**Example 9: Species list for a country**
Let's return to our previous example. What if we just need a list of the species in the Bahamas, rather than the specific details of each occurrence record? We can instead use the function `BIEN_list_country` to download a list of species, which should be much faster than using `BIEN_occurrence_country` to get a species list.
```{r lists, eval = FALSE}
Bahamas_species_list <- BIEN_list_country(country = "Bahamas")
#Notice that we find many more species listed than we found occurrence records for. What happened? There are many records coming from the Bahamas that lack coordinates. These records are used used in the "_list_" functions, but not the occurrence functions.
```
If we wanted to retrieve the results for multiple countries at once, that is simple as well. We just need to supply a vector of countries.
```{r lists-multi-country, eval = FALSE}
country_vector <- c("Haiti","Dominican Republic")
Haiti_DR <- BIEN_list_country(country = country_vector)
```
We can also use political division codes (from geonames.org) instead of writing out the full country names.
```{r lists-pol-divs}
#To see all of the political division names, and associated codes, we can use this function:
political_names <- BIEN_metadata_list_political_names()
#Let's take a look at what the dataframe contains:
head(political_names)
#In addition to the standardized country, state (state_province_ascii) and county (county_parish_ascii) names, we have the associated codes that can be used in BIEN functions.
#Note that 'state' refers to any primary political division (e.g. province), and 'county' refers to any secondary political division (e.g. parish).
#Looking at the political_names dataframe, we see that the Dominican Republic has country code "DO", and Haiti has country code "HT"
Haiti_DR_from_codes <- BIEN_list_country(country.code = c("HT","DO"))
```
# Metadata
The BIEN metadata functions start with the prefix `BIEN_metadata_...` and provide useful metadata for the BIEN database.
1. `BIEN_metadata_database_version` Returns the current version number of the BIEN database and the release date.
2. `BIEN_metadata_match_data` Rudimentary function to check for changed records between old and current queries.
3. `BIEN_metadata_citation` Function to generate bibtex citations for use in reference managers.
4. `BIEN_metadata_list_political_names` Returns a dataframe containing political division names and associate codes.
**Example 10: Metadata**
To check what the current version of the BIEN database is (which we recommend reporting when using BIEN data):
```{r md-1}
BIEN_metadata_database_version()
```
**Example 11: Citations**
One of the more innovative features of the BIEN package is that it will generate custom attribution data for you based on what data you downloaded through the package.
Let's say we're interested in *Selaginella selaginoides*, and we'd like to download some occurrence data:
```{r selaginella-occs, eval = FALSE}
Selaginella_selaginoides_occurrences <- BIEN_occurrence_species("Selaginella selaginoides", new.world = NULL)
```
If we plan on using those data in a publication, we'll need proper attribution. We can use `BIEN_metadata_citation` to do this for us:
```{r selaginella-md, eval = FALSE}
citation_info <- BIEN_metadata_citation(dataframe = Selaginella_selaginoides_occurrences)
```
citation_info is a list that contains 3 elements:
1. A bit of general information on how to use the list.
2. A set of bibtex formatted references.
3. Acknowledgement text.
To make things even easier on ourselves, we can use some of the additional functionality of the `BIEN_metadata_citation` function:
```{r md-2, eval = FALSE}
temp_dir <- file.path(tempdir(), "BIEN_temp") #Set a temporary working directory
citation_info <- BIEN_metadata_citation(dataframe = Selaginella_selaginoides_occurrences,
bibtex_file = file.path(temp_dir,"selaginella_selaginoides.bib"),
acknowledgement_file = file.path(temp_dir,"selaginella_selaginoides.txt"))
```
Now, we have a bibtex file, `selaginella_selaginoides.bib`, that can be loaded into a reference manager (e.g. Endnote, Paperpile, etc.), and a text file, `selaginella_selaginoides.txt`, containing text that can be pasted into the acknowledgements section of a publication.
What if we also have some trait data? No problem there, the code handles that as well:
```{r md-3, eval = FALSE}
#First, let's get some trait data:
selaginella_selaginoides_traits <- BIEN_trait_species(species = "Selaginella selaginoides")
#Now, we just need to modify our previous bit of code to include the trait data as well:
temp_dir <- file.path(tempdir(), "BIEN_temp")
citation_info <- BIEN_metadata_citation(dataframe = Selaginella_selaginoides_occurrences,
trait.dataframe = selaginella_selaginoides_traits,
bibtex_file = file.path(temp_dir,"selaginella_selaginoides.bib"),
acknowledgement_file = file.path(temp_dir,"selaginella_selaginoides.txt"))
```
The updated citation information will now contain references for both trait and occurrence records.
# Combining Queries
**Example 11: Putting it all together **
Coming soon!