\(~\)

Introduction

In this exercise, you will:

  • Familiarize yourself with a variety of spatial environmental data resources
  • Download and manage spatial environmental data

Note that this exercise gives only a very quick overview. For more detailed background and practice, I highly recommend these lessons from Software Carpentry:

\(~\)

Getting Started

As with species occurrence data, it is critical to begin by thinking carefully about the variables you will use to build your models. Key questions include:

  • What is the biological process I am trying to understand?
  • What are the environmental conditions that might affect this process?
  • What is the spatial / temporal scale that is relevant for my question?

Clear answers to these questions will help you immensely in the modeling process and beyond.

\(~\)

Data availability

There is a wide range of environmental variables readily available for download. The list below mentions some of the most commonly used sources (that cover the global scale) but remember that there may be higher resolution or higher quality data sets available for a particular study area. Take a little time to explore the links that seem most relevant to your interests and then proceed to the next section, Working with raster data in R.

  • Climate
    • Chelsa: interpolated climate variables at ~1km resolution, includes climate layers for various time periods and variables, ranging from the Last glacial maximum, to current times, to several future scenarios.
    • WorldClim: interpolated climate variables at variable resolutions, including historical climate and weather as well as climate under several future scenarios.
    • ENVIREM: numerous derived climatic and topographic variables (e.g., annual potential evapotranspiration, growing degree days).
  • Elevation / topography
    • STRM elevation: Data from the Shuttle Radar Topography Mission used to generate elevation data with a global resolution of 3 arc-seconds (~90 m).
  • Land-use / land-cover
  • Soils
    • The Harmonized World Soil Database is a 30 arc-second raster database with over 15000 different soil mapping units that combines existing regional and national updates of soil information worldwide.
    • SoilGrids is a system for global digital soil mapping that uses machine learning methods to map the global distribution of soil properties.

*Note: working with large and high resolution raster data in R can lead to computational challenges, which are outside the scope of this exercise. Some of the datasets listed above are very large and so you need to select a certain region before downloading.

**Note: The resources above are heavily biased towards terrestrial ecosystems. For marine variables, see this video lecture by Hannah Owens (data sources are covered in the 2nd half).

In general, the resources above can be downloaded from the web, saved to your computer, and later read into R for visualization and analysis. Alternatively, there are several R packages that allow you to download some of these data sources directly in R. For example:

\(~\)

Working with raster data in R

For the rest of this exercise, you should work through this tutorial from NEON. To start, you will need to:

  • Make sure you have the raster and rgdal packages installed
  • Download the data for the exercise here

\(~\)

Other resources

The following series of video lectures (from the ENM2020 course organized by T. Peterson) provide much more background on environmental data for SDMs/ENMs: