The field naturalist vision is being increasingly abandoned in palaeontology, favouring desktop study ahead of fieldwork and leading to virtual palaeoworlds. Even though the numerical dimension may alienate scientists and disconnect them from the actual parameters of their problems, the use of numerical simulations can allow the testing of hypotheses that depend on multiple variables that putatively control the dynamic evolution of a system (Cleland 2001). Moreover, quantitative visualization offers illustrations or animations that may convey ideas more effectively. This paper investigates how powerful Interactive Data Language® (IDL) tools can be used to interactively visualize, analyse, and discuss results from large palaeontological data sets.

This work presents two examples of palaeobiogeographic reconstructions, one Triassic and one Cenozoic, using the IDL programming environment. We begin by examining the Triassic example, a general 2D palaeobiogeographic model based on the constrained spreading of random phylogenies (Brayard 2002). Simulations are carried out in the biogeographic, thermal, and ocean current setting of the Early Triassic. Modelled and present-day Pacific Ocean Sea Surface Temperatures (SST) and Sea Surface Currents (SSC) are applied to discuss the recovery and distribution of planktonic species following the Permo-Triassic crisis. In the second part, we focus on IDL's Geographic Information Systems (GIS) facilities that can be used to plot mammal diversity and ecomorphologic data on Neogene palaeogeographic maps.

Both models are written with IDL 5.2 and run on Windows 98 and XP, Unix/Linux, and MacOS systems. IDL represents a complete computing and programming system for interactive analysis and visualization of data sets (Marschallinger 2001). IDL is a programming meta-language implemented in ENVI (Environment for Visualizing Images®) by Research Systems Inc. This meta-language is especially dedicated to numerical analysis and 2D/3D image processing. Several pre-packaged graphic modules are able to interact with each other, giving IDL a major advantage in time required for application development compared with other, more popular languages like FORTRAN or C. Yet IDL keeps its compatibility with standard programming languages through specialized functions.

IDL allows the user to create custom procedures, functions, or applications using simple matrix representation without using explicit loop structures to process matrix data, further reducing development time. IDL also allows the creation of graphic user interfaces. Results can be viewed immediately after processing using many pre-packaged graphic tools. All IDL functions, additional functions, and examples can be downloaded on the Research Systems Inc. homepage. An open-source clone (PyDL) of IDL is also developed in Python for Linux.

In geosciences, this language has been used primarily for satellite image analysis and 2D/3D object modelling in applications such as the spatiotemporal analysis of basin history. The use of IDL in palaeontology is still rare and usually limited to the reconstruction and analysis of fossil morphology (Marschallinger 2001). In this work, IDL is used as a tool for calculating and visualizing palaeobiogeographic maps. The palaeobiological implications of this study will not be discussed in great detail; rather we concentrate on the construction of the two models and the benefits of IDL.