CHRONOSPHERE: EARTH SYSTEM HISTORY VARIABLES

Researchers focusing on global-scale ecological and evolutionary patterns are faced with the overwhelmingly increasing volume and diversity of data compilations. The complexity of the relevant research questions necessitates the interconnected use of multiple datasets (the source variables) that are often regularly updated. Finding these data and incorporating them in the analytical workflow can be time consuming and often lead to difficulties and conflicts in version control.

The goal of the ‘chronosphere’ project is to offer a portal to these data and to allow their immediate use by compiling frequently accessed representations on a server that keeps track of different versions. These data are available through an R package of the same name (available on the CRAN servers), that downloads a specific version of a dataset and attaches it to the R environment. The project focuses on data used in biogeography, macroecology and macroevolution, from large datasets to small tables published as supplementary information to research articles. Currently available data include major sources such as the Paleobiology Database, the PaleoReefs Database, and the paleogeographic reconstructions of the PaleoMAP project. ‘chronosphere’ will also be used to distribute the output of BRIDGE HadCM3L family deep-time climate models.

To facilitate the use of deep-time spatial data, the R package includes functions that utilize either the GPlates Web Service or the GPlates desktop application to reconstruct the positions of present-day coordinates and spatial features using a user-specified reconstruction model. An efficient container class is also implemented to facilitate the work with raster and vector data layers that are organized in multi-dimensional arrays.

These features were proven especially useful in the classroom environment, where the fast download and transformation of data are important. Providing data in a consistent form within the R programming framework, ‘chronosphere’ allows the fast reproducibility, and ultimately the tracing of scientific results with the same codebase.