Construct a Species-Area Relationship with ssarp

Introduction

A species-area relationship (SAR) visualizes the relationship between species richness (the number of species) and the area of the land mass on which the species live. The observation that species richness increases with increasing area is a fundamental law of ecology, and a disruption in this relationship may be associated with habitat loss, habitat fragmentation, and increasing numbers of non-native species. Creating SARs for island-dwelling species helps researchers understand how trends in biodiversity across archipelagos are changing due to these effects.

The goal of this vignette is to use the ssarp R package to create a SAR for Anolis, a well-studied genus of lizards. We will focus on Anolis occurrence records from the Caribbean Islands. More information about the ssarp package and a comparison to a previously published SAR for Anolis can be found in the manuscript associated with the package.

In order to construct a species-area relationship with ssarp, we will:

• Gather occurrence data from GBIF
• Filter out invalid occurrence records
• Find areas of pertinent land masses
• Create a species-area relationship

At the end of this vignette, there are additional sections covering the use of presence-absence matrices (PAMs) to estimate a species-area relationship and detailing the ways the user can customize inputs to ssarp::find_areas().

Gathering Occurrence Data

GBIF (Global Biodiversity Information Facility) provides an easy method for gathering occurrence data for taxa of interest. ssarp uses functions from the rgbif package to gather occurrence records associated with a given taxon. The user may also provide their own data for use in creating a SAR, but we will use GBIF in this example.

A tutorial for gathering occurrence records from GBIF can be found in the “Get Occurrence Records from GBIF” vignette here. This example will use rgbif::occ_search() for simplicity, but please note that rgbif::occ_download() is more appropriate for gathering data used in research. Here, we will gather the first 10000 occurrence records for island-dwelling Anolis lizards in the Caribbean restricted by a WKT polygon (see the vignette linked above for more information).

# Load packages
library(rgbif)
library(ssarp)

query <- "Anolis"
rank <- "Genus"

suggestions <- name_suggest(q = query, rank = rank)

key <- as.numeric(suggestions$data[1,1])

limit <- 10000

occurrences <- occ_search(taxonKey = key, 
                          hasCoordinate = TRUE, 
                          limit = limit,
                          geometry = 'POLYGON((-84.8 23.9, -84.7 16.4, -65.2 13.9, -63.1 11.0, -56.9 15.5, -60.5 21.9, -79.3 27.8, -79.8 24.8, -84.8 23.9))')

dat <- occurrences$data

Finding Land Mass Names and Areas

Once the occurrence data is returned, we will use each occurrence record’s GPS point to determine the land mass on which the species was found and find the area associated with that land mass using a database of island areas and names from ssarp.

# Find land mass names
land_dat <- find_land(occurrences = dat)

# Print first 5 lines of land_dat
head(land_dat, n = 5)

##                                                   SpeciesName  Genus      Species  Longitude  Latitude
## 1 Anolis hispaniolae (Köhler, Zimmer, Mcgrath & Hedges, 2019) Anolis  hispaniolae -70.597156 19.098515
## 2                                 Anolis distichus Cope, 1861 Anolis    distichus  -68.40635  18.67363
## 3                            Anolis roquet (Bonnaterre, 1789) Anolis       roquet -60.893013  14.77053
## 4                            Anolis roquet (Bonnaterre, 1789) Anolis       roquet -60.893013  14.77053
## 5                  Anolis cristatellus Duméril & Bibron, 1837 Anolis cristatellus -66.123847 18.471217
##                first second third                           datasetKey
## 1 Dominican Republic   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7
## 2 Dominican Republic   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7
## 3         Martinique   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7
## 4         Martinique   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7
## 5               <NA>   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7

The locality information is split across three columns: “first,” “second,” and “third.” The mapping utilities that ssarp uses sometimes output different levels of specificity for locality information (up to three different levels), so these columns provide space for these different levels. The island name that we are interested in will be in the last filled-in column of the three. For example, if there are two columns of locality information for a given occurrence record, the island name will be in the second. If there is only one column of locality information, it will contain the island name (as with Puerto Rico and Antigua above). If all columns have NA, the occurrence record is invalid and will be filtered out in the next step.

Now that we have determined the names of the land masses associated with each occurrence record, we will find the area associated with each land mass.

# Use the land mass names to get their areas
area_dat <- find_areas(occs = land_dat)

# Print first 5 lines of area_dat
head(area_dat, n = 5)

##                                                   SpeciesName  Genus     Species  Longitude  Latitude
## 1 Anolis hispaniolae (Köhler, Zimmer, Mcgrath & Hedges, 2019) Anolis hispaniolae -70.597156 19.098515
## 2                                 Anolis distichus Cope, 1861 Anolis   distichus  -68.40635  18.67363
## 3                            Anolis roquet (Bonnaterre, 1789) Anolis      roquet -60.893013  14.77053
## 4                            Anolis roquet (Bonnaterre, 1789) Anolis      roquet -60.893013  14.77053
## 8                            Anolis evermanni Stejneger, 1904 Anolis   evermanni -66.314592  18.29657
##                first second third                           datasetKey       areas
## 1 Dominican Republic   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7 83104562500
## 2 Dominican Republic   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7 83104562500
## 3         Martinique   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7  1190000000
## 4         Martinique   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7  1190000000
## 8        Puerto Rico   <NA>  <NA> 50c9509d-22c7-4a22-a47d-8c48425ef4a7  9710687500

Now, our occurrence record dataframe includes records with GPS points that are associated with a land mass, along with the areas of those land masses (in m^2).

The ssarp::remove_continents() function removes any continental occurrence records, which is useful when the user is only interested in island-dwelling species (as we are in this example). While the data obtained by using the rgbif::occ_search() function was geographically restricted, potential user error in specifying the polygon in WKT format often leads to accidental continental records that will be removed by using this function.

nocont_dat <- remove_continents(occs = area_dat)

Create Species-Area Relationship

Finally, we will generate the SAR using the ssarp::create_sar() function (can also be run with create_SAR()). The ssarp::create_sar() function creates multiple regression objects with breakpoints up to the user-specified “npsi” parameter. For example, if “npsi” is two, ssarp::create_sar() will generate regression objects with zero (linear regression), one, and two breakpoints. The function will then return the regression object with the lowest AIC score. The “npsi” parameter will be set to one in this example. Note that if linear regression (zero breakpoints) is better-supported than segmented regression with one breakpoint, the linear regression will be returned instead. We set the “visualize” parameter to TRUE so that the function outputs the plot of the SAR.

ssarp::create_sar(occurrences = nocont_dat, npsi = 1, visualize = TRUE)

## 
##  ***Regression Model with Segmented Relationship(s)***
## 
## Call: 
## segmented.lm(obj = linear, seg.Z = ~x, npsi = 1, control = segmented::seg.control(display = FALSE))
## 
## Estimated Break-Point(s):
##           Est. St.Err
## psi1.x 22.429  0.482
## 
## Coefficients of the linear terms:
##             Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.08563    1.00429  -0.085    0.933
## x            0.04440    0.05412   0.820    0.419
## U1.x         0.92566    0.20603   4.493       NA
## 
## Residual standard error: 0.5381 on 28 degrees of freedom
## Multiple R-Squared: 0.6864,  Adjusted R-squared: 0.6528 
## 
## Boot restarting based on 9 samples. Last fit:
## Convergence attained in 2 iterations (rel. change 2.8018e-09)

Figure 2. This is the species-area relationship (SAR) for Anolis including island-based occurrences within a polygon around Caribbean islands from the first 10000 records for the genus in GBIF! The best-fit model was a segmented regression with one breakpoint.

The ssarp::create_sar() function will also output the summary for the best-fit model for the data (displayed above). This summary includes a few important pieces of information to highlight. First, the Est column of psi1.x displays the location of the breakpoint on the x-axis. Next, the table of Coefficients of the linear terms includes estimates (located in the Estimate column) for the following terms: the y-intercept ((Intercept)), the slope before the breakpoint (x), and the slope after the breakpoint (U1.x). In this example, to three signifcant figures, we would report that the breakpoint is at 22.4, the slope before the breakpoint is 0.0444, and the slope after the breakpoint is 0.926.

Using a Presence-Absence Matrix to Infer a SAR

In a presence-absence matrix (PAM), the column names are species names and the row names are island names. Within each species column, a 1 represents the presence of that species on the island corresponding to the given row, and a 0 represents the absence of that species on the island corresponding to the given row.

Using a PAM for species richness data in the ssarp workflow allows the user to skip the data curation steps discussed in the “Gathering Occurrence Data” section here. The following code reads an example PAM that is built into the ssarp package. Then, it uses that PAM as an input to the ssarp::find_pam_areas() function to generate the same style of dataframe that the ssarp::find_areas() function created above.

# Load ssarp
library(ssarp)

# Read example PAM file
pam <- read.csv(system.file("extdata",
                            "example_pam.csv",
                            package = "ssarp"))

# Create dataframe that includes areas
area_dat <- ssarp::find_pam_areas(pam = pam)

# Print first 5 lines of area_dat
head(area_dat, n = 5)

##   genericName specificEpithet    third first second     areas
## 1      Anolis       distichus  Hog Cay  <NA>   <NA> 3.125e+06
## 2      Anolis          sagrei  Hog Cay  <NA>   <NA> 3.125e+06
## 3      Anolis          sagrei Anguilla  <NA>   <NA> 2.200e+07
## 4      Anolis        allisoni     Cuba  <NA>   <NA> 1.220e+11
## 5      Anolis          sagrei     Cuba  <NA>   <NA> 1.220e+11

In the code above, the ssarp::find_pam_areas() function was used with the example PAM to create a dataframe (area_dat) that can be used directly with ssarp::create_sar() in the same way as in the “Finding Land Mass Names and Areas” section.

create_sar(occurrences = area_dat, npsi = 1, visualize = FALSE)

## 
##  ***Regression Model with Segmented Relationship(s)***
## 
## Call: 
## segmented.lm(obj = linear, seg.Z = ~x, npsi = 1, control = segmented::seg.control(display = FALSE))
## 
## Estimated Break-Point(s):
##           Est. St.Err
## psi1.x 21.463   0.66
## 
## Coefficients of the linear terms:
##             Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.17060    1.25280  -0.136    0.893
## x            0.02970    0.06606   0.450    0.658
## U1.x         0.69862    0.16744   4.172       NA
## 
## Residual standard error: 0.4165 on 18 degrees of freedom
## Multiple R-Squared: 0.8449,  Adjusted R-squared: 0.8191 
## 
## Boot restarting based on 6 samples. Last fit:
## Convergence attained in 2 iterations (rel. change 2.7106e-09)

Please note that while the example PAM represents Anolis occurrence data, it does not include exactly the same species, so the breakpoint and slopes here are different compared to the first SAR inferred in this vignette.

Customizing Inputs to find_areas()

The ssarp::find_areas() function provides multiple options for users to customize how ssarp finds land mass areas. By default, the ssarp::find_areas() function uses a built-in database of island names and areas that is queried by island names present in the user’s occurrence records. Users may access this database directly by running the ssarp::get_island_areas() function as written below:

# Access the island area database (returns a dataframe)
island_areas <- get_island_areas()

# The first 5 rows of island_areas
head(island_areas, n = 5)

##   OBJECTID_1 COUNT    AREA MAX           Name
## 1       6614   123 7687500  22         ‘Uqbān
## 2      80233    10  625000   4         ‘Ushsh
## 3      87448     2  125000   0         A Chau
## 4       5277    33 2062500  30 A Illa da Toxa
## 5      12313    27 1687500 109       A Quatre

The island_areas object includes 5 columns:

• OBJECTID_1: The object ID for the island within the global island shapefiles used to compile the area database (see Martinet et al. 2025 for more details).
• COUNT: The number of polygons covered by the island as calculated by ArcGIS (see Martinet et al. 2025 for more details)
• AREA: The area of the island
• MAX: The maximum elevation of the island (in meters)
• Name: The name of the island

If you notice inaccuracies in this built-in island dataset, please follow the steps outlined in ssarp’s contributing guidelines to fix those inaccuracies. Thank you for helping make our package more accurate for everybody!

# A vector of names
names <- c("Kanto", "Kansai", "Kyushu")

# A vector of matching areas
areas <- c(32424, 33124, 36782)

# Create a dataframe of names and areas
area_custom <- as.data.frame(cbind(names, areas))

# Add required column names
colnames(area_custom) <- c("Name", "AREA")

# Ensure that the AREA column is numeric
area_custom$AREA <- as.numeric(area_custom$AREA)