# Classes¶

The package is built around a number of “classes” of which the RasterLayer, RasterBrick, and RasterStack are the most important. When discussing methods that can operate on objects of all three of these classes, they are referred to as Raster* objects.

## RasterLayer¶

A RasterLayer represents single-layer (variable) raster data. A RasterLayer object always stores a number of fundamental parameters that describe it. These include the number of columns and rows, the coordinates of its spatial extent (‘bounding box’), and the coordinate reference system (the ‘map projection’). In addition, a RasterLayer can store information about the file in which the raster cell values are stored (if there is such a file). A RasterLayer can also hold the raster cell values in memory.

## RasterStack and RasterBrick¶

It is quite common to analyze raster data using single-layer objects. However, in many cases multi-variable raster data sets are used. The raster package has two classes for multi-layer data the RasterStack and the RasterBrick. The principal difference between these two classes is that a RasterBrick can only be linked to a single (multi-layer) file. In contrast, a RasterStack can be formed from separate files and/or from a few layers (‘bands’) from a single file.

In fact, a RasterStack is a collection of RasterLayer objects with the same spatial extent and resolution. In essence it is a list of RasterLayer objects. A RasterStack can easily be formed form a collection of files in different locations and these can be mixed with RasterLayer objects that only exist in memory.

A RasterBrick is truly a multilayered object, and processing a RasterBrick can be more efficient than processing a RasterStack representing the same data. However, it can only refer to a single file. A typical example of such a file would be a multi-band satellite image or the output of a global climate model (with e.g., a time series of temperature values for each day of the year for each raster cell). Methods that operate on RasterStack and RasterBrick objects typically return a RasterBrick.

## Other Classes¶

Below is some more detail, you do not need to read or understand this section to use the raster package.

The three classes described above inherit from the raster class (that means they are derived from this more basic ‘parent’ class by adding something to that class) which itself inherits from the BasicRaster class. The BasicRaster only has a few properties (referred to as ‘slots’ in S4 speak): the number of columns and rows, the coordinate reference system (which itself is an object of class CRS, which is defined in package sp) and the spatial extent, which is an object of class Extent.

An object of class Extent has four slots: xmin, xmax, ymin, and ymax. These represent the minimum and maximum x and y coordinates of the of the Raster object. These would be, for example, -180, 180, -90, and 90, for a global raster with longitude/latitude coordinates. Note that raster uses the coordinates of the extremes (corners) of the entire raster (unlike some files/programs that use the coordinates of the center of extreme cells).

raster is a virtual class. This means that it cannot be instantiated (you cannot create objects from this class). It was created to allow the definition of methods for that class. These methods will be dispatched when called with a descendent of the class (i.e. when the method is called with a RasterLayer, RasterBrick or RasterStack object as argument). This allows for efficient code writing because many methods are the same for any of these three classes, and hence a single method for raster suffices.

RasterStackBrick is a class union of the RasterStack and RasterBrick class. This is a also a virtual class. It allows defining methods (functions) that apply to both RasterStack and RasterBrick objects.