aestears/PlantTracker
Extract Demographic and Competition Data from Fine-Scale Maps

Package index
Search the aestears/PlantTracker package
Vignettes
Functions
21
Source code
26
Man pages
13
Home
/
GitHub
/
aestears/PlantTracker
/
README.md

README.md
In aestears/PlantTracker: Extract Demographic and Competition Data from Fine-Scale Maps

plantTracker

CRAN status check-standard test-coverage

License: MIT DOI

Welcome to plantTracker! This package was designed to transform long-term quadrat maps that show plant occurrence and size into demographic data that can be used to answer questions about population and community ecology.

Table of Contents

Installation

Install plantTracker from CRAN:

install.packages("plantTracker")

Alternatively, you can install the current version of plantTracker from GitHub:

install.packages("devtools")
devtools::install_github("aestears/plantTracker")

Contributing

Please report any problems that you encounter while using plantTracker as “issues” on (our GitHub repository)[https://github.com/aestears/plantTracker/issues/]. Help us make this package better!

License

This package is licensed under MIT License Copyright (c) 2022 Alice Stears

Contact

Questions about plantTracker can be forwarded to Alice Stears, the package maintainer, at alice.e.stears@gmail.com.

Interactive tools to explore plantTracker functions

Follow this link to a shiny app that lets you change arguments in the trackSpp() function, and see how that impacts individual ID assignment in an example dataset: https://astears.shinyapps.io/shinyapp/

How to use the plantTracker R package

The material below explains how to use plantTracker, starting with formatting your data correctly. This information is also available in the ‘Suggested plantTracker Workflow’ vignette, which is included in the package.

1. Prepare data

The functions in plantTracker require data in a specific format. plantTracker includes an example dataset that consists of two pieces: grasslandData and grasslandInventory. You can load these example datasets into your global environment by calling data(grasslandData) and data(grasslandInventory). You can view the documentation for these datasets by calling?grasslandData and ?grasslandInventory.

Most plantTracker functions require two data objects. The first is a data frame that contains the location and metadata for each mapped individual, which we from now on will call dat. The second is a list that contains a vector of years in which each quadrat was sampled, which we from now on will cal inv.

Below are the basic requirements for these data objects.

1.1 The dat data frame must . . .

Here are the first few rows of a possible dat input data.frame:

#> Simple feature collection with 6 features and 6 fields
#> Geometry type: POLYGON
#> Dimension:     XY
#> Bounding box:  xmin: -0.000160084 ymin: 0.4334812 xmax: 0.286985 ymax: 0.9419673
#> CRS:           NA
#>                 Species Type Site Quad Year sp_code_6
#> 1 Heteropogon contortus poly   AZ  SG2 1922    HETCON
#> 2 Heteropogon contortus poly   AZ  SG2 1922    HETCON
#> 3 Heteropogon contortus poly   AZ  SG2 1922    HETCON
#> 4 Heteropogon contortus poly   AZ  SG2 1922    HETCON
#> 5 Heteropogon contortus poly   AZ  SG2 1922    HETCON
#> 6 Heteropogon contortus poly   AZ  SG2 1922    HETCON
#>                         geometry
#> 1 POLYGON ((0.237747 0.908835...
#> 2 POLYGON ((0.2833037 0.85959...
#> 3 POLYGON ((0.008583123 0.449...
#> 4 POLYGON ((0.1480142 0.46983...
#> 5 POLYGON ((0.03573306 0.5259...
#> 6 POLYGON ((0.2441894 0.52689...

Here’s what some of the example dat data (from the “SG2” quadrat at the “AZ” site in 1922) look like when plotted spatially:

#> Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
#> ℹ Please use `linewidth` instead.
#> This warning is displayed once every 8 hours.
#> Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
#> generated.
Figure 1.1 : Spatial map of a subset of example 'dat' dataset

Figure 1.1 : Spatial map of a subset of example ‘dat’ dataset

It’s important to note that, while plantTracker was designed to be used with small-scale maps of plant occurrence in quadrats, it is conceivably possible to use other styles of map data in plantTracker functions. All that is required is a single mapped basal area (or point location converted to a small polygon) at each time point for each organism (or ramet), and is accompanied by the required metadata detailed above. For example, plantTracker functions could be used to estimate tree demographic rates at the scale of 100 m x 50 m plots.

1.1.1 Get your data into the sf data format

As mentioned above, plantTracker uses the sf R package to deal with spatial data. The map data that plantTracker was built to analyze is inherently spatial, so you need to know how to the basics of dealing with spatial data in R if you want to use plantTracker! There are many good resources to help you orient yourself to working with spatial data in R generally:

And the sf package more specifically:

These resources provide a great orientation, and while I recommend looking over them if you’re new to working with spatial data in R, I’ve included a brief tutorial for uploading shapefiles into R as sf data frames.

Most of the published chart-quadrat datasets have the map data stored as shapefiles in complex file structures, which can be a bit confusing to navigate. plantTracker requires all of your data (for all species, plots and years) to be in one single data frame. This example shows how you might navigate through a complex file structure to to pull out shapefiles and put them into one single sf data frame. for further analysis with plantTracker. For this example, I’ll use a subset of the data from the Santa Rita Experimental Range in Arizona, which has been published in this data paper. In this dataset, shapefiles for each quadrat are stored in their own folder. Within that folder there are two shapefiles for each year: one that contains map data for polygons, and one that contains data for points. The following code reads in those shapefiles, transforms the points to polygons of a fixed radius, and puts all the data into one sf data frame. If you want to follow along, download the “shapefiles.zip” file from the data paper, un-zip it, and name it “AZ_shapefiles”. The dataset that is the result of this example is the same as part of the “grasslandData” dataset included in plantTracker.

#  save a character vector of the file names in the file that contains the 
# shapefiles (in this case, called "CO_shapefiles"), each of which is a quadrat
# note: 'wdName' is a character string indicating the path of the directory 
# containing the 'AZ_shapefiles' folder
quadNames <- list.files(paste0(wdName,"AZ_shapefiles/"))
# trim the quadrats down to 2, for the sake of runtime in this example 
quadNames <- quadNames[quadNames %in% c("SG2", "SG4")]

# now we'll loop through the quadrat folders to download the data
for (i in 1:2){#length(quadNames)) {
  # get the names of the quadrat for this iteration of the loop
  quadNow <- quadNames[i]
  #  get a character vector of the unique quad/Year combinations of data in 
  # this folder that contain polygon data
  quadYears <- quadYears <-  unlist(strsplit(list.files(
    paste0(wdName, "AZ_shapefiles/",quadNow,"/"),
    pattern = ".shp$"), split = ".shp"))
  # loop through each of the years in this quadrat
  for (j in 1:length(quadYears)) {
    # save the name of this quadYear combo
    quadYearNow <- quadYears[j]
    # read in the shapefile for this quad/year combo as an sf data frame 
    # using the 'st_read()' function from the sf package 
    shapeNow <- sf::st_read(dsn = paste0(wdName,"AZ_shapefiles/",quadNow), 
                            #  the 'dsn' argument is the folder that 
                            # contains the shapefile files--in this case, 
                            # the folder for this quadrat
                            layer = quadYearNow) # the 'layer' argument has the 
    # name of the shapefile, without the filetype extension! This is because each 
    # shapefile consists of at least three separate files, each of which has a 
    # unique filetype extension. 
    # the shapefiles in this dataset do not have all of the metadata we 
    # need, and have some we don't need, so we'll remove what we don't need and 
    # add columns for 'site', 'quad', and 'year'
    shapeNow$Site <- "AZs"
    shapeNow$Quad <- quadNow
    # get the Year for the shapefile name--in this case it is the last for 
    # numbers of the name
    shapeNow$Year <- as.numeric(strsplit(quadYearNow, split = "_")[[1]][2]) + 1900
    # determine if the 'current' quad/year contains point data or polygon data
    if (grepl(quadYearNow, pattern = "C")) { # if quadYearNow has point data
      # remove the columns we don't need
      shapeNow <- shapeNow[,!(names(shapeNow)
                              %in% c("Clone", "Seedling", "Area", "Length", "X", "Y"))]
      # reformat the point into a a very small polygon 
      # (a circle w/ a radius of .003 m)
      shapeNow <- sf::st_buffer(x = shapeNow, dist = .003)
      # add a column indicating that this observation was originally 
      # mapped as a point
      shapeNow$type <- "point"
    } else { # if quadYearNow has polygon data
      # remove the columns we don't need
      shapeNow <- shapeNow[,!(names(shapeNow) %in% c("Seedling", "Canopy_cov", "X", "Y", "area"))]
      # add a column indicating that this observation was originally 
      # mapped as a polygon
      shapeNow$type <- "polygon"
    }
    # now we'll save this sf data frame 
    if (i == 1 & j == 1) { # if this is the first year in the first quadrat
      dat <- shapeNow
    } else { # if this isn't the first year in the first quadrat, simply rbind 
      # the shapeNow sf data frame onto the previous data 
      dat <- rbind(dat, shapeNow)
    }
  }
}

# Now, all of the spatial data are in one sf data frame!
# for the sake of this example, we'll remove data for some species and years in order to make the example run faster (and to make this 'dat' data.frame identical to the "grasslandData" dataset included in this R pakcage).
dat <- dat[dat$Species %in% c("Heteropogon contortus", "Bouteloua rothrockii", "Ambrosia artemisiifolia", "Calliandra eriophylla", "Bouteloua gracilis", "Hesperostipa comata", "Sphaeralcea coccinea", "Allium textile"),]
dat <- dat[  (dat$Quad %in% c("SG2", "SG4") & 
             dat$Year %in% c(1922:1927)),]

In some spatial datasets, observations that were measured as “points” in the field are still stored as “points” in the shapefiles. plantTracker requires all observations to be stored as “polygon” geometry in order to streamline functions, so we need to translate “points” into small polygons of a fixed area. In this case, we’ll transform them into circles with a radius of 1 cm (.01, since this dataset measures area in meters). ‘dat’ has a column called “type.” A value of “point” in this column will tell us that, even though the geometry of the “point” data is now in “polygon” format, the values for basal area and growth are not indicative of the true size of the plant.

# We use the function "st_buffer()" to add a buffer of our chosen radius (.01) around each point observation, which will transform each observation into a circle of the "polygon" format with a radius of .01. 
dat_1 <- st_buffer(x = dat[st_is(x = dat, type = "POINT"),], dist = .01)
dat_2 <- dat[!st_is(x = dat, type = "POINT"),]
dat <- rbind(dat_1, dat_2)

If you don’t want to download the data and format it into an sf data.frame, you can also use a subset of the “grasslandData” data object stored in this R package. You will just need to subset it to include only the data from the “AZ” site. Code to do this is below:

dat <- grasslandData[grasslandData$Site == "AZ",]

1.2 The inv list must . . .

Here is an example of an inv argument that corresponds to the example dat argument above. The quadrats that have data in dat are “SG2” and “SG4”, so there are elements in inv that correspond to each of these quadrats.

#> $SG2
#> [1] 1922 1923 1924 1925 1926 1927
#> 
#> $SG4
#> [1] 1922 1923 1924 1925 1926 1927

If you already have a quadrat inventory as a data frame, it isn’t complicated to reformat it to work with plantTracker functions. For example, if your quadrat inventory data frame looks like this… :

#>   quad1 quad2 quad3
#> 1  2000  2000  2000
#> 2  2001  2001    NA
#> 3    NA  2002  2002
#> 4  2003  2003  2003
#> 5  2004  2004  2004
#> 6  2005  2005  2005
#> 7  2006  2006  2006
#> 8  2007  2007  2007

… then do the following to get it into a format ready for plantTracker:

quadInv_DF <- data.frame("quad1" = c(2000, 2001, NA, 2003, 2004, 2005, 2006, 2007),
           "quad2" = c(2000:2007), 
           "quad3" = c(2000, NA, 2002, 2003, 2004, 2005, 2006, 2007))

# use the 'as.list()' function to transform your data frame into a named list
quadInv_list <- as.list(quadInv_DF)
# we still need to remove the 'NA' values, which we can do using the 
# 'lapply()' function
(quadInv_list <- lapply(X = quadInv_list, FUN = function(x) x[is.na(x) == FALSE]))
#> $quad1
#> [1] 2000 2001 2003 2004 2005 2006 2007
#> 
#> $quad2
#> [1] 2000 2001 2002 2003 2004 2005 2006 2007
#> 
#> $quad3
#> [1] 2000 2002 2003 2004 2005 2006 2007

1.3 Check the inv and dat arguments using checkDat()

The generic checkDat() function:

checkDat(dat, inv = NULL, species = "Species", site = "Site", quad = "Quad",
  year = "Year", geometry = "geometry", reformatDat = FALSE, ...)

This step is optional, but can be useful if you’re unsure whether your dat and inv arguments are in the correct format. The plantTracker function checkDat() takes dat and inv as arguments for the arguments dat andinv, and will return informative error messages if either argument is not in the correct format.

Additional optional arguments to checkDat() are species, site, quad, year, geometry, and reformatDat.

2 Track individuals through time using trackSpp()

Now it’s time to transform your raw dataset into demographic data! This is accomplished using the trackSpp() function. This function follows individual plants from year to year in the same quadrat to determine survival, size in the next year, age, and some additional potentially-useful demographic data. It does this by comparing quadrat maps from sequential years. If there is overlap of individuals of the same species in consecutive years, then the rows in dat that contain data for those overlapping individuals are given the same “trackID”, or unique identifier.

Here is the generic trackSpp() function:

trackSpp(dat, inv, dorm, buff, buffGenet, clonal, species = "Species",
  site = "Site", quad = "Quad", year = "Year", geometry = "geometry",
  aggByGenet = TRUE, printMessages = TRUE, flagSuspects = FALSE,
  shrink = 0.1, dormSize = 0.05, ...)

2.1 Function arguments

trackSpp() takes the following arguments:

Consider the following example: There is a polygon of species “A” in year 1, which is our “focal individual”. In year 2, there is not a polygon of species “A” that overlaps with our focal individual. In year 3, there is a polygon of species “A” that is in the same location as our focal individual. If dorm = 0, then our focal individual would get a 0 in the survival column, and the polygon of species “A” in year 3 would be considered a new recruit and get a new trackID. If dorm = 1, because there is overlap between two polygons of the same species with only a 1-year gap between when they occur, these two polygons will be considered the same genetic individual, will have the same trackID, and our focal individual will have a “1” in the survival column. In an alternative scenario, in years 3 and 4 there are not polygons of species “A” that are in the same location as our focal individual, but there is a polygon in year 4 that overlaps our focal individual. If dorm = 1, then our focal individual would get a “0” for survival, but if dorm = 2, then it would get a 1 for survival.

Figure 2.1: A visualization of the 'dormancy' scenario described above.

Figure 2.1: A visualization of the ‘dormancy’ scenario described above.

If you’d like to be more specific and perhaps biologically accurate, you can also specify the dorm argument uniquely for each species. For example, it might be that you are confident that your data collectors did not accidentally “miss” any individuals, and your dat data frame contains observations for shrubs or trees, which are very unlikely to go dormant, and small forbs, which are much more likely to go dormant for one or two years. In order to disallow dormancy for trees and shrubs, but to allow dormancy for forbs, you will provide a data frame to the dorm argument instead of a single positive integer value. There will be two columns: 1) a “Species” column that has the species name for each species present in dat, and 2) a column called “dorm” that has positive integer values indicating the dormancy you’d like to allow for each species. Make sure that if you are following the data frame approach, you must provide a dormancy argument for every species that has data in dat. Make sure that the species names in the dorm data frame are spelled exactly the same as they are in dat. The data frame should look something like this:

#>   Species dorm
#> 1  tree A    0
#> 2 shrub B    0
#> 3  tree C    0
#> 4  forb D    1
#> 5  forb E    2
#> 6  forb F    1
Figure 2.2: With a 10 cm buffer, these polygons in 1922 and 1923 overlap and will be identified by trackSpp() as the **same** individual and receive the same trackID.

Figure 2.2: With a 10 cm buffer, these polygons in 1922 and 1923 overlap and will be identified by trackSpp() as the **same** individual and receive the same trackID.

Figure 2.3: With a 3 cm buffer, these polygons in 1922 and 1923 don't quite overlap, so will be identified by trackSpp() as **different** individuals and receive different trackIDs.

Figure 2.3: With a 3 cm buffer, these polygons in 1922 and 1923 don’t quite overlap, so will be identified by trackSpp() as **different** individuals and receive different trackIDs.

The following arguments to trackSpp() are only required in certain contexts.

These are all of the possible arguments to trackSpp()!

2.2 Function output

Below is an example of a potential function call to trackSpp(), using the example dat and inv data we’ve used so far. :

datTrackSpp <- plantTracker::trackSpp(dat = dat, inv = inv,
         dorm = 1,
         buff = .05,
         buffGenet = .005,
         clonal = data.frame("Species" = c("Heteropogon contortus",
                                           "Bouteloua rothrockii",
                                           "Ambrosia artemisiifolia",
                                           "Calliandra eriophylla"),
                             "clonal" = c(TRUE,TRUE,FALSE,FALSE)),
         aggByGenet = TRUE,
         printMessages = FALSE
         )

And here’s what the output of this call to trackSpp() looks like:

#> Simple feature collection with 477 features and 12 fields
#> Geometry type: GEOMETRY
#> Dimension:     XY
#> Bounding box:  xmin: -0.001386579 ymin: -0.001017592 xmax: 1.000536 ymax: 1.001267
#> CRS:           NA
#> First 10 features:
#>    Site Quad                 Species        trackID Year  type    basalArea
#> 1    AZ  SG2 Ambrosia artemisiifolia  AMBART_1922_1 1922 point 2.461883e-05
#> 2    AZ  SG2 Ambrosia artemisiifolia AMBART_1922_10 1922 point 2.461883e-05
#> 3    AZ  SG2 Ambrosia artemisiifolia AMBART_1922_11 1922 point 2.461883e-05
#> 4    AZ  SG2 Ambrosia artemisiifolia AMBART_1922_12 1922 point 2.461883e-05
#> 5    AZ  SG2 Ambrosia artemisiifolia AMBART_1922_13 1922 point 2.461883e-05
#> 6    AZ  SG2 Ambrosia artemisiifolia AMBART_1922_14 1922 point 2.461883e-05
#> 7    AZ  SG2 Ambrosia artemisiifolia AMBART_1922_15 1922 point 2.461883e-05
#> 8    AZ  SG2 Ambrosia artemisiifolia AMBART_1922_16 1922 point 2.461883e-05
#> 9    AZ  SG2 Ambrosia artemisiifolia AMBART_1922_17 1922 point 2.461883e-05
#> 10   AZ  SG2 Ambrosia artemisiifolia AMBART_1922_18 1922 point 2.461883e-05
#>    recruit survives_t+1 age size_t+1 nearEdge                       geometry
#> 1       NA            0  NA       NA     TRUE POLYGON ((0.350604 0.021361...
#> 2       NA            0  NA       NA    FALSE POLYGON ((0.7172048 0.25834...
#> 3       NA            0  NA       NA    FALSE POLYGON ((0.1845598 0.38566...
#> 4       NA            0  NA       NA    FALSE POLYGON ((0.759387 0.399850...
#> 5       NA            0  NA       NA    FALSE POLYGON ((0.1696044 0.40291...
#> 6       NA            0  NA       NA    FALSE POLYGON ((0.7290925 0.41097...
#> 7       NA            0  NA       NA    FALSE POLYGON ((0.6255546 0.45737...
#> 8       NA            0  NA       NA    FALSE POLYGON ((0.5872073 0.46925...
#> 9       NA            0  NA       NA    FALSE POLYGON ((0.8863168 0.52256...
#> 10      NA            0  NA       NA    FALSE POLYGON ((0.5212498 0.52831...

If you did not allow any species to be clonal (clonal = 0) or if aggByGenet = TRUE in your call to trackSpp(), then your output data frame will have one row for each genet, and is ready for demographic analysis! If your output data frame is not yet aggregated by genet (i.e. you use aggByGenet = FALSE), then you need to transform your data frame so that each genet is represented by only one row of data. You can use the aggregateByGenet() function from plantTracker (see this function’s documentation for guidance), or your own method of choice.

You can stop here and proceed to your own analyses using the demographic data you generated, or you can proceed with other plantTracker functions outlined below for some additional useful data.

3 Calculate local neighborhood density using getNeighbors()

It is often useful in demographic analyses to have some idea of the competition (or facilitation) that an individual organism is dealing with. Interactions between individuals can have a profound impact on whether an organism survives and grows. Spatial datasets of plant occurrence allow us to generate an estimate of the interactions an individual plant has with other plants by determining how many other individuals occupy the “local neighborhood” of each focal plant. While this isn’t a direct measure of competition or facilitation, it gives us an estimate that we can include in demographic models.

Here is the generic getNeighbors() function:

getNeighbors(dat, buff, method, compType = "allSpp", output = "summed",
  trackID = "trackID", species = "Species", quad = "Quad", year = "Year",
  site = "Site", geometry = "geometry", ...)

3.1 Function options and arguments

The getNeighbors() function in plantTracker calculates local neighborhood density for each unique individual in your dataset. A user-specified buffer is drawn around each individual, and then the function counts the number of other plants within this buffer.This function can only be run on a dataset where each unique individual (genet) is represented by only one row of data. If the genet consists of multiple polygons, then they must be aggregated into one sf MULTIPOLYGON object. If your dataset has multiple rows for each genet, then you can use the aggregateByGenet() function to get it ready to use in getNeighbors(). Additionally, getNeighbors() requires your dataset to have a column containing a unique identifier for each genet. Across multiple years, that genet must have the same unique identifier. If you are using this function right after trackSpp(), your dataset will already have this unique identifier in a column called “trackID”.

getNeighbors() has several options that allow you to customize how local neighborhood density is calculated.

Figure 3.1: This individual outlined in pink is a focal individual, and the pale pink shows a 10 cm buffer around it.

Figure 3.1: This individual outlined in pink is a focal individual, and the pale pink shows a 10 cm buffer around it.

Figure 3.2: The 10cm buffer around the focal individual overlaps with 5 other unique individuals of two species. These overlapping individuals are outlined in dark grey. Using the 'count' method in `getNeighbors()`, we would get an intraspecific competition value of 3, and an interspecific competition value of 5.

Figure 3.2: The 10cm buffer around the focal individual overlaps with 5 other unique individuals of two species. These overlapping individuals are outlined in dark grey. Using the ‘count’ method in `getNeighbors()`, we would get an intraspecific competition value of 3, and an interspecific competition value of 5.

Figure 3.3: The 10cm buffer around the focal individual overlaps with 5 other unique individuals of two species. The overlapping area is shaded in grey. Using the 'area' method in `getNeighbors()`, we would get an intraspecific competition metric of 0.0454, and an interspecific competition metric of 0.0462.

Figure 3.3: The 10cm buffer around the focal individual overlaps with 5 other unique individuals of two species. The overlapping area is shaded in grey. Using the ‘area’ method in `getNeighbors()`, we would get an intraspecific competition metric of 0.0454, and an interspecific competition metric of 0.0462.

Below are the arguments in the getNeighbors() function.

3.2 Function outputs

The output of getNeighbors() is an sf data frame that is identical to the input dat, but with either one or two additional columns. If method = "area", there are two columns added called “nBuff_area” and “neighbors_area”. The first contains the area of the buffer zone around each focal individual. The second contains the basal area of neighbors that overlap with a focal individual’s buffer zone. If method = "count", there is only one additional column added to the output, called “neighbors_count.” This column contains a count of the neighbors that occur within a focal individual’s buffer zone.

Here’s an example of a getNeighbors() function call using the resulting data from the example in section 2.2, as well as the resulting data frame. Note that method = "area", so two columns are added to the returned data frame:

datNeighbors <- plantTracker::getNeighbors(dat = datTrackSpp,
             buff = .15,
             method = "area",
             compType = "allSpp")

#> Simple feature collection with 477 features and 14 fields
#> Geometry type: GEOMETRY
#> Dimension:     XY
#> Bounding box:  xmin: -0.001386579 ymin: -0.001017592 xmax: 1.000536 ymax: 1.001267
#> CRS:           NA
#> First 10 features:
#>                    Species Site Quad        trackID Year neighbors_area
#> 1  Ambrosia artemisiifolia   AZ  SG2  AMBART_1922_1 1922   0.0011328591
#> 2  Ambrosia artemisiifolia   AZ  SG2 AMBART_1922_10 1922   0.0040518140
#> 3  Ambrosia artemisiifolia   AZ  SG2 AMBART_1922_11 1922   0.0055774468
#> 4  Ambrosia artemisiifolia   AZ  SG2 AMBART_1922_12 1922   0.0029129080
#> 5  Ambrosia artemisiifolia   AZ  SG2 AMBART_1922_13 1922   0.0050297901
#> 6  Ambrosia artemisiifolia   AZ  SG2 AMBART_1922_14 1922   0.0037646754
#> 7  Ambrosia artemisiifolia   AZ  SG2 AMBART_1922_15 1922   0.0026270846
#> 8  Ambrosia artemisiifolia   AZ  SG2 AMBART_1922_16 1922   0.0013297048
#> 9  Ambrosia artemisiifolia   AZ  SG2 AMBART_1922_17 1922   0.0022690175
#> 10 Ambrosia artemisiifolia   AZ  SG2 AMBART_1922_18 1922   0.0006338116
#>    nBuff_area basalArea basalArea_genet survives_t+1 survives_tplus1 size_t+1
#> 1  0.04344697     point    2.461883e-05           NA               0       NA
#> 2  0.07329218     point    2.461883e-05           NA               0       NA
#> 3  0.07329218     point    2.461883e-05           NA               0       NA
#> 4  0.07329218     point    2.461883e-05           NA               0       NA
#> 5  0.07329218     point    2.461883e-05           NA               0       NA
#> 6  0.07329218     point    2.461883e-05           NA               0       NA
#> 7  0.07329218     point    2.461883e-05           NA               0       NA
#> 8  0.07329218     point    2.461883e-05           NA               0       NA
#> 9  0.06848976     point    2.461883e-05           NA               0       NA
#> 10 0.07329218     point    2.461883e-05           NA               0       NA
#>    size_tplus1 nearEdge                       geometry
#> 1           NA     TRUE POLYGON ((0.350604 0.021361...
#> 2           NA    FALSE POLYGON ((0.7172048 0.25834...
#> 3           NA    FALSE POLYGON ((0.1845598 0.38566...
#> 4           NA    FALSE POLYGON ((0.759387 0.399850...
#> 5           NA    FALSE POLYGON ((0.1696044 0.40291...
#> 6           NA    FALSE POLYGON ((0.7290925 0.41097...
#> 7           NA    FALSE POLYGON ((0.6255546 0.45737...
#> 8           NA    FALSE POLYGON ((0.5872073 0.46925...
#> 9           NA    FALSE POLYGON ((0.8863168 0.52256...
#> 10          NA    FALSE POLYGON ((0.5212498 0.52831...

The example above uses the option output = "summed", which is the default for the getNeighbors() function. With this option, the “neighbors_area” or “neighbors_count” column (depending on the method argument) contains just one value that sums the neighbor count or area across all neighbor species. However, if output = "bySpecies", the “neighbors_count” or “neighbors_area” column contains a list with the counts or areas broken down by species. The output argument is described in more detail in section 3.1. If you want to use the getNeighbors() function to determine how the effect of neighbors differs according to the species identity of those neighbors, setting output = "bySpecies" allows you to do this. However, it is likely that your subsequent analyses will need the by-species neighbor data in a matrix or data frame format, rather than in a list nested inside a data frame, which is how getNeighbors() returns the data.

Here is some code that turns the data returned by getNeighbors(output = "bySpecies") into a matrix where each row has data for one focal individual, and each column has data for one species. This format should be easier to work with for further analysis.

# save the output of the getNeighbors() function
datNeighbors_bySpp <- plantTracker::getNeighbors(dat = datTrackSpp,
             buff = .15, method = "area", compType = "allSpp", output = "bySpecies")

# determine all of the possible species that can occupy the buffer zone
compSpp <- unique(datTrackSpp$Species)

temp <- lapply(X = datNeighbors_bySpp$neighbors_area,  FUN = function(x) {
  tmp <- unlist(x)
  tmp2 <- tmp[compSpp]
  }
)

for (i in 1:length(temp)) {
  # fix the column names
  names(temp[[i]]) <- compSpp
  # save the data in a matrix
  if (i == 1) {
    datOut <- temp[[i]]
  } else {
    datOut <- rbind(datOut, temp[[i]])
  }
}
# make the rownames of the matrix correspond to the trackID of the focal individual
rownames(datOut) <- datNeighbors_bySpp$trackID

# show the first few rows of the datOut data frame: 
datOut[1:5,]
#>                Ambrosia artemisiifolia Bouteloua rothrockii
#> AMBART_1922_1             4.923767e-05         5.216672e-04
#> AMBART_1922_10            8.437672e-05         1.250163e-03
#> AMBART_1922_11            2.461883e-05         2.310445e-04
#> AMBART_1922_12            9.353468e-05         7.408342e-04
#> AMBART_1922_13            2.461883e-05         1.906272e-08
#>                Calliandra eriophylla Heteropogon contortus
#> AMBART_1922_1                     NA          0.0005619542
#> AMBART_1922_10                    NA          0.0027172740
#> AMBART_1922_11                    NA          0.0053217835
#> AMBART_1922_12                    NA          0.0020785392
#> AMBART_1922_13                    NA          0.0050051522

4 Next steps

At this point, this dataset should be ready for you to use in any applications wish! There are a few additional functions that may help you in your analyses, and these are outlined in this section.

Specific instructions for how to use each of these functions can be found in their documentation!



aestears/PlantTracker documentation built on July 20, 2023, 1:52 p.m.

R Package Documentation

rdrr.io home R language documentation Run R code online

Browse R Packages

CRAN packages Bioconductor packages R-Forge packages GitHub packages

We want your feedback!

Note that we can't provide technical support on individual packages. You should contact the package authors for that.
GitHub issue tracker
ian@mutexlabs.com
Personal blog

 
Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.

Close