Building Palaeogeographic Maps in R

6 minute read

At the end of last year, I was looking at data from the Palaeobiology Database (PBDB) and wanted to draw a map of where fossils were found. I made such maps for Jurassic ichthyosaurs before, which appeared in Moon & Kirton (2018, Fig. 46). These I made by hand, tracing maps from Blakey (2008, 2014) then overlaying palaeo-positions of various ichthyosaur finds. As I remember, the tracing itself was not too time consuming, but wasn’t the most interesting thing I’ve ever done.

Now I wanted to create maps with many more points and show a series of time slices as the position of the continents change. This would definitely benefit from a more automated method rather than me getting annoyed doing it all manually.

Fortunately there are ways to do this in R with code and packages that are readily available. Here’s a little intro to how I’ve made a map with data from the PBDB, including:

  • plotting continent and coastal outlines
  • adding points to show occurrence locations
  • separating different groups into subplots.

In this case I’ll plot data of Toarcian (182.7–174.1 million years ago, Ma) ichthyosaurs. Let’s dive in.

Data gathering

The few things that we need are:

  • Occurrence data from the Palaeobiology Database
  • Palaeogeographic outlines of the continents at a desired time.
Fossil occurrences
Finds of species at a single time and place. In the Palaeobiology Database these are linked to collections, which are groups of fossil finds from a single effort, like a single fossil dig.
Palaeogeography
The ancient configuration of continents, coastlines, mountains and seas changing with continental drift and sea level change.
Palaeocoordinate
The ancient position of modern locations. Modern locations have a latitude and longitude value, while palaeocoordinates have palaeolatitude and palaeolongitude. These can be reconstructed using the relative movement of continental plates from the present back through the past.

The PBDB has a Web API, which I find the easiest way to download data. There is also the package ropensci/paleobiodb, but I haven’t used it myself. Or you can download a CSV file direct from the portal.

To download occurrences of ichthyosaurs from the Toarcian I used the following code in R.

library(tidyverse)

pbdb_url <-
  "https://paleobiodb.org/data1.2/occs/list.csv?base_name=Ichthyosauromorpha&interval=Toarcian&show=paleoloc"

occ_toarcian_ichthyosaurs <-
  readr::read_csv(pbdb_url)

In looking for outlines of ancient continents, I found a few sources. First, I came across the R package NonaR/paleoMap, which includes map data and functions for organising them within the package. But this was last updated in 2015, so may be a little out of date now or may use functions not in the more recent versions of R. (I admittedly did not take too long to check.)

After, I found LunaSare/gplatesr. gplatesr downloads data from the GPlates Web Service, and can reconstruct palaeocoordinates – something I look forward to reading up on and using more.

Getting this data into R is straightforward: using code from the package gplatesr as a cue, these following lines download outline data for the continents at 182 Ma, in the early Toarcian. In this case, I wanted both the positions of the continental plates (static polygons) and the coastlines. The package rgdal reads in the data and organises it into a structure that R can use and understand.

library(rgdal)

coastline_gws_url <-
  "http://gws.gplates.org/reconstruct/coastlines/?time=182&model=GOLONKA"
polygons_gws_url <-
  "http://gws.gplates.org/reconstruct/static_polygons/?time=182&model=GOLONKA"

toarcian_coastlines <-
  rgdal::readOGR(coastline_gws_url)
toarcian_polygons <-
  rgdal::readOGR(polygons_gws_url)

GPlates is software built to reconstruct palaeogeography using models of continent trajectories through time. By default GPlates uses reconstructions from Matthews et al. (2016), but it can interpolate between known continental configurations. This means that you can get the positions of the continents at any time in the past and export this into a format for R. This makes the GPlates software useful for producing reconstructions of several times or producing animations. The web service may be more time-consuming, having to download lots of data, or restrict you if you try to grab too much too quickly.

Incidentally, ‘GOLONKA’ in the URLs above refers to the Golonka (2007) model of continental movement (Vérard 2019).

Plotting the map

To plot the map data, I found that using ggplot is the most convenient as this provides the geom_map function exactly for this purpose. I modified the following code from the gplatesr vignette to plot the base map, ready to add points later.

Here, I use geom_map twice: first to add the position of the continental regions in grey, then to overlay outlines of the modern coastlines. The extra geom_rect, coord_map and theme_map lines change the look to make the map a little more appealing.

library(broom)
library(ggplot2)
library(ggthemes)

toarcian_coastlines <-
  broom::tidy(toarcian_coastlines)
toarcian_polygons <-
  broom::tidy(toarcian_polygons)

toarcian_map <-
  ggplot() +
    geom_map(
      data = toarcian_polygons, map = toarcian_polygons,
      aes(x = long, y = lat, map_id = id),
      size = 0.15, fill = "#d8d8d8"
    ) +
    geom_map(
      data = toarcian_coastlines, map = toarcian_coastlines,
      aes(x = long, y = lat, map_id = id),
      size = 0.15, fill = NA, colour = "grey30"
    ) +
    geom_rect(
      data = data.frame(xmin = -180, xmax = 180, ymin = -90, ymax = 90),
      aes(xmin = xmin, xmax = xmax, ymin = ymin, ymax = ymax),
      color = 1, fill = NA, size = 0.3
    ) +
    coord_map("mollweide") +
    ggthemes::theme_map()

toarcian_map +
  labs(
    title = "Palaeogeographical map of continental plate (grey) arrangement\nin the Toarcian (182 Ma) with modern coastlines outlined above."
  )
Palaeogeographical map of the Toarcian
Palaeogeographical map of continental plate (grey) arrangement in the Toarcian (182 Ma) with modern coastlines outlined above.

It’s a decent map. Aside from a few colour changes, from personal preference, this should cover most of what we want from a base layer. The early Toarcian is at the earlier end of the separation of Pangaea – the Atlantic has only just started to open.

Plotting occurrence locations

The data downloaded from the PBDB include pre-calculated palaeocoordinates when using show=paleoloc in the URL. Conveniently, this is often using GPlates, and the palaeomodel column shows the origin of this. We can just stick the PBDB data straight onto our palaeogeographical map. Neat.

toarcian_map +
  geom_point(
    data = occ_toarcian_ichthyosaurs,
    aes(x = paleolng, y = paleolat)
  ) +
  labs(
    title = "Palaeogeographical map of ichthyosaur occurrences\nin the Toarcian (182 Ma)."
  )
Ichthyosaur occurrences in the Toarcian
Palaeogeographical map of ichthyosaur occurrences in the Toarcian (182 Ma).

Ah … well I was hoping for a little more spread! But it seems that just about all Toarcian ichthyosaurs come from western Europe. This is not surprising – most specimens are either from the Posidonia Shale Formation of Germany or the Yorkshire Lias. And this region was a shallow sea at the time – well suited to many different marine reptile groups. It’s good to know there are more out there though, covering France up to Norway apparently.

Separating out different groups is also easy to do using facets. In this case I separated out the different identification ranks for the occurrences.

toarcian_map +
  geom_point(
    data = occ_toarcian_ichthyosaurs,
    aes(x = paleolng, y = paleolat)
  ) +
  facet_wrap(vars(identified_rank)) +
  theme(legend.position = "none") +
  labs(
    title = "Palaeogeographic map of ichthyosaur occurrences\nin the Toarcian (182 Ma), separated by identified rank."
  )
Toarcian ichthyosaur occurrences separated by identified rank
Palaeogeographical map of ichthyosaur occurrences in the Toarcian (182 Ma) separated by identified rank.

Conclusion

Hopefully you’ll find these code snippets useful to make a map and add occurrence palaeolocations to this. I was pleasantly surprised by how readily available the data are and how easy it is to plot in R. Next I’m looking to add in maps for multiple times and occurrences of different taxa for a project that, perhaps surprisingly, doesn’t include ichthyosaurs.

References

Blakey, R. 2008. Gondwana paleogeography from assembly to breakup—a 500 m.y. odyssey. Special Paper 441: Resolving the Late Paleozoic Ice Age in Time and Space 441: 1–28. doi:10.1130/2008.2441(01)

Blakey, R. 2014. Library of paleogeography. https://jan.ucc.nau.edu/~rcb7/

Golonka, J. 2007. Phanerozoic paleoenvironment and paleolithofacies maps: Mesozoic. Geologia 33: 211–264.

Matthews, K.J., Maloney, K.T., Zahirovic, S., Williams, S.E., Seton, M. and Müller, R.D. 2016. Global plate boundary evolution and kinematics since the late Paleozoic. Global and Planetary Change 146: 226–250. doi:10.1016/j.gloplacha.2016.10.002

Moon, B.C. and Kirton, A.M. 2018. Ichthyosaurs of the British Middle and Upper Jurassic. Part 2, Brachypterygius, Nannopterygius, Macropterygius, and Taxa Invalida. Monograph of the Palaeontographical Society 172 (650): 85–176. doi:10.1080/02693445.2018.1468139

Vérard, C. 2019. Plate tectonic modelling: review and perspectives. Geological Magazine 156: 208–241. doi:10.1017/S0016756817001030

Leave a comment