Raster data model

The raster data model is best for modeling properties which vary continuously across the landscape, and can be measured at any location

Examples of raster features from field surveys or remote sensing:

• Soil type or vegetation class
• Elevation
• Plant density
• Ice depth
• Vegetation productivity
• Suitability or risk

A raster dataset is composed of a uniform grid of identically-sized cells (usually square), with each cell (or pixel) holding a single numeric value

Raster attributes

Raster attributes are about the variable at that location, stored as a cell value. Each grid cell can only store a single value, so if multiple attributes need to be stored (e.g. elevation and slope), multiple rasters are needed, which are sometimes bound together in a raster stack

Rasters values can be:

• Quantities - continuous variables such as rainfall or density. The DEM we’ll use below is an example of a continuous raster
• Thematic - categories such as vegetation class, presence/absence of a species, represented by a numerical code

Our LandCover dataset is an example of a thematic raster - each pixel contains a code indicating the landcover class at that location. These are the classes that occur in and around Che Tao Nature Reserve:¹

Code	Landcover class
20	Shrubs
30	Herbaceous vegetation
40	Cropland
50	Urban
80	Permanent water bodies
90	Herbaceous wetland
111	Closed forest, evergreen needle leaf
112	Closed forest, evergreen broad leaf
114	Closed forest, deciduous broad leaf
116	Closed forest, unknown
122	Open forest, evergreen broad leaf
126	Open forest, unknown

Quiz: Raster data model

Here’s another quick quiz!

---